Author response:

eLife Assessment

This manuscript introduces a useful protein-stability-based fitness model for simulating protein evolution and unifying non-neutral models of molecular evolution with phylogenetic models. The model is applied to four viral proteins that are of structural and functional importance. The justification of some hypotheses regarding fitness is incomplete, as well as the evidence for the model's predictive power, since it shows little improvement over neutral models in predicting protein evolution.

We thank for the constructive comments that helped improve our study. Regarding the comment about justification of fitness, we will include in the revised manuscript additional information to support the relevance of modeling protein evolution accounting for protein folding stability. We agree that increasing the parameterization of the developed birth-death model is interesting, if it does not lead to overfitting. The model presented considers the fitness of protein variants to determine their reproductive success through the corresponding birth and death rates, varying among lineages, and it is biologically meaningful and technically correct (Harmon 2019). Following a suggestion of the first reviewer to allow variation of the global birth-death rate among lineages, we will additionally incorporate this aspect into the model and evaluate its performance with the data for the evaluation of the models. The integration of structurally constrained substitution models of protein evolution, as Markov models, into the birth-death process was made following standards approaches of molecular evolution in population genetics (Yang 2006; Carvajal-Rodriguez 2010; Arenas 2012; Hoban, et al. 2012) and we will provide more information about it in the revised manuscript. Regarding the predictive power, our study showed good accuracy in predicting the real folding stability of forecasted protein variants. On the other hand, predicting the exact sequences proved to be more challenging, indicating needs in the field of substitution models of molecular evolution. Altogether, we believe our findings provide a significant contribution to the field, as accurately forecasting the folding stability of future real proteins is fundamental for predicting their protein function and enabling a variety of applications. Additionally, we implemented the models into a freely available computer framework, with detailed documentation and diverse practical examples.

Public Reviews:

Reviewer #1 (Public review):

Summary:

Ferreiro et al. present a method to simulate protein sequence evolution under a birth-death model where sequence evolution is constrained by structural constraints on protein stability. The authors then use this model to explore the predictability of sequence evolution in several viral structural proteins. In principle, this work is of great interest to molecular evolution and phylodynamics, which have struggled to couple non-neutral models of sequence evolution to phylodynamic models like birth-death. Unfortunately, though, the model shows little improvement over neutral models in predicting protein evolution, and this ultimately appears to be due to fundamental conceptual problems with how fitness is modeled and linked to the phylodynamic birth-death model.

We thank the reviewer for the positive comments about our work.

Regarding predictive power, the study showed a good accuracy in predicting the real folding stability of forecasted protein variants under a selection model, but not under a neutral model. However, predicting the exact sequences was more challenging. For example, amino acids with similar physicochemical properties can result in similar folding stability while differ in the specific sequence, more accurate substitution models of molecular evolution are required in the field. We consider that forecasting the folding stability of future real proteins is an important advancement in forecasting protein evolution, given the essential role of folding stability in protein function and its variety of applications. Regarding the conceptual concerns related to fitness modeling, we clarify this issue in detail in our responses to the specific comments below.

Major concerns:

(1) Fitness model: All lineages have the same growth rate r = b-d because the authors assume b+d=1. But under a birth-death model, the growth r is equivalent to fitness, so this is essentially assuming all lineages have the same absolute fitness since increases in reproductive fitness (b) will simply trade off with decreases in survival (d). Thus, even if the SCS model constrains sequence evolution, the birth-death model does not really allow for non-neutral evolution such that mutations can feed back and alter the structure of the phylogeny.

We thank the reviewer for this comment that aims to improve the realism of our model. In the model presented (but see later for another model derived from the proposal of the reviewer and that we are now implementing into the framework and applying to the data used for the evaluation of the models), the fitness predicted from a protein variant is used to obtain the corresponding birth rate of that variant. In this way, protein variants with high fitness have high birth rates leading to overall more birth events, while protein variants with low fitness have low birth rates resulting in overall more extinction events, which has biological meaning for the study system. The statement “All lineages have the same growth rate r = b-d” in our model is incorrect because, in our model, b and d can vary among lineages according to the fitness. For example, a lineage might have b=0.9, d=0.1, r=0.8, while another lineage could have b=0.6, d=0.4, r=0.2. Indeed, the statement “this is essentially assuming all lineages have the same absolute fitness” is incorrect. Clearly, assuming that all lineages have the same fitness would not make sense, in that situation the folding stability of the forecasted protein variants would be similar under any model, which is not the case as shown in the results. In our model, the fitness affects the reproductive success, where protein variants with a high fitness have higher birth rates leading to more birth events, while those with lower fitness have higher death rates leading to more extinction events. This parameterization is meaningful for protein evolution because the fitness of a protein variant can affect its survival (birth or extinction) without necessarily affecting its rate of evolution. While faster growth rate can sometimes be associated with higher fitness, a variant with high fitness does not necessarily accumulate substitutions at a faster rate. Regarding the phylogenetic structure, the model presented considers variable birth and death events across different lineages according to the fitness of the corresponding protein variants, and this alters the derived phylogeny (i.e., protein variants selected against can go extinct while others with high fitness can produce descendants). We are not sure about the meaning of the term “mutations can feed back” in the context of our system. Note that we use Markov models of evolution, which are well-stablished in the field (despite their limitations), and substitutions are fixed mutations, which still could be reverted later if selected by the substitution model (Yang 2006). Altogether, we find that the presented birth-death model is technically correct and appropriate for modeling our biological system. Its integration with structurally constrained substitution (SCS) models of protein evolution, as Markov models, is correct following general approaches of molecular evolution in population genetics (Yang 2006; Carvajal-Rodriguez 2010; Arenas 2012; Hoban, et al. 2012). We will provide a more detailed description of the model in the revised manuscript.

Apart from these clarifications about the birth-death model used, we understand the point of the reviewer and following the suggestion we are now incorporating an additional birth-death model that accounts for variable global birth-death rate among lineages. Specifically, we are following the model proposed by Neher et al (2014), where the death rate is considered as 1 and the birth rate is modeled as 1 + fitness. In this model, the global birth-death rate varies among lineages. We are now implementing this model into the computer framework and applying it to the data used for the evaluation of the models. Preliminary results, which will be finally presented in the revised manuscript, indicate that this model yields similar predictive accuracy compared to the previous birth-death model. If this is confirmed, accounting for variability in the global birth-death rate does not appear to play a major role in the studied systems of protein evolution. We will present this additional birth-death model and its results in the revised manuscript.

(2) Predictive performance: Similar performance in predicting amino acid frequencies is observed under both the SCS model and the neutral model. I suspect that this rather disappointing result owes to the fact that the absolute fitness of different viral variants could not actually change during the simulations (see comment #1).

The study shows similar performance in predicting the sequences of the forecasted proteins under both the SCS model and the neutral model, but shows differences in predicting the folding stability of the forecasted proteins between these models. Indeed, as explained in the previous answer, the birth-death model accounts for variation in fitness among lineages, leading to differences among lineages in reproductive success. The new birth-death model that we are now implementing, which incorporates variation of the global birth-death rate among lineages, is producing similar preliminary results. In addition to these considerations, it is known that SCS models applied to phylogenetics (such as ancestral molecular reconstruction) can model protein evolution with high accuracy in terms of folding stability. However, inferring sequences (i.e., ancestral sequences) is considerably more challenging even for ancestral molecular reconstruction (Arenas, et al. 2017; Arenas and Bastolla 2020). The observed sequence diversity is much greater than the observed structural diversity (Illergard, et al. 2009; Pascual-Garcia, et al. 2010), and substitutions among amino acids with similar physicochemical properties can result in protein variants with similar folding stability but different specific amino acid sequences; further work is demanded in the field of substitution models of molecular evolution. We will expand the discussion of this aspect in the revised manuscript.

(3) Model assessment: It would be interesting to know how much the predictions were informed by the structurally constrained sequence evolution model versus the birth-death model. To explore this, the authors could consider three different models: 1) neutral, 2) SCS, and 3) SCS + BD. Simulations under the SCS model could be performed by simulating molecular evolution along just one hypothetical lineage. Seeing if the SCS + BD model improves over the SCS model alone would be another way of testing whether mutations could actually impact the evolutionary dynamics of lineages in the phylogeny.

In the present study, we compare the neutral model + birth-death (BD) with the SCS model + BD. Markov substitution models Q are applied upon an evolutionary time (i.e., branch length, t) and this allows to determine the probability of substitution events during that time period [P(t) = exp (Qt)]. This approach is traditionally used in phylogenetics to model the incorporation of substitutions over time. Therefore, to compare the neutral and SCS models, an evolutionary time is required, in this case it is provided by the birth-death process. The suggestions 1) and 2) cannot be compared without an underlined evolutionary history. However, comparisons in terms of likelihood, and other aspects, between models that ignore the protein structure and the implemented SCS models are already available in our previous studies based on coalescent simulations or given phylogenetic trees (Arenas, et al. 2013; Arenas, et al. 2015). There, SCS models produced proteins with more realistic folding stability than models that ignore evolutionary constraints from the protein structure, and those findings are consistent with the results from the present study where we explore the application of these models to forecasting protein evolution. We would like to emphasize that forecasting the folding stability of future real proteins is a significant and novel finding, folding stability is fundamental to protein function and has diverse implications. While accurately forecasting the exact sequences would indeed be ideal, this remains a challenging task with current substitution models. In this regard, we will discuss in the revised manuscript the need of developing more accurate substitution models.

(4) Background fitness effects: The model ignores background genetic variation in fitness. I think this is particularly important as the fitness effects of mutations in any one protein may be overshadowed by the fitness effects of mutations elsewhere in the genome. The model also ignores background changes in fitness due to the environment, but I acknowledge that might be beyond the scope of the current work.

This comment made us realize that more information about the features of the implemented SCS models should be included in the manuscript. In particular, the implemented SCS models consider a negative design based on the observed residue contacts in nearly all proteins available in the Protein Data Bank (Arenas, et al. 2013; Arenas, et al. 2015). This data is provided as an input file and it can be updated to incorporate new structures (see the framework documentation and the practical examples). Therefore, the prediction of folding stability is a combination of positive design (direct analysis of the target protein) and negative design (consideration of background proteins to reduce biases), thus incorporating background molecular diversity. This important feature was not sufficiently described in the manuscript, and we will add more details in the revised version. Regarding the fitness caused by the environment, we agree with the reviewer. This is a challenge for any method aiming to forecast evolution, as future environmental shifts are inherently unpredictable and may impact the accuracy of the predictions. Although one might attempt to incorporate such effects into the model, doing so risks overparameterization, especially when the additional factors are uncertain or speculative. We will include a discussion in the revised manuscript about our perspective on the potential effects of environmental changes on forecasting evolution.

(5) In contrast to the model explored here, recent work on multi-type birth-death processes has considered models where lineages have type-specific birth and/or death rates and therefore also type-specific growth rates and fitness (Stadler and Bonhoeffer, 2013; Kunhert et al., 2017; Barido-Sottani, 2023). Rasmussen & Stadler (eLife, 2019) even consider a multi-type birth-death model where the fitness effects of multiple mutations in a protein or viral genome collectively determine the overall fitness of a lineage. The key difference with this work presented here is that these models allow lineages to have different growth rates and fitness, so these models truly allow for non-neutral evolutionary dynamics. It would appear the authors might need to adopt a similar approach to successfully predict protein evolution.

We agree with the reviewer that robust birth-death models have been developed applying statistics and, in many cases, the primary aim of those studies is the development and refinement of the model itself. Regarding the study by Rasmussen and Stadler 2019, it incorporates an external evaluation of mutation events where the used fitness is specific for the proteins investigated in that study, which may pose challenges for users interested in analyzing other proteins. In contrast, our study takes a different approach. We implement a fitness function that can be predicted and evaluated for any type of protein (Goldstein 2013), making it broadly applicable. In addition, we provide a freely available and well-documented computational framework to facilitate its use. The primary aim of our study is not the development of novel or complex birth-death models. Rather, we aim to explore the integration of a standard birth-death model with structurally constrained substitution models for the purpose of predicting protein evolution. In the context of protein evolution, substitution models are a critical factor (Liberles, et al. 2012; Wilke 2012; Bordner and Mittelmann 2013; Echave, et al. 2016; Arenas, et al. 2017; Echave and Wilke 2017), and their combination with a birth-death model constitutes a first approximation upon which next studies can build to better understand this biological system. We will include these considerations in the revised manuscript.

Reviewer #2 (Public review):

Summary:

In this study, "Forecasting protein evolution by integrating birth-death population models with structurally constrained substitution models", David Ferreiro and co-authors present a forward-in-time evolutionary simulation framework that integrates a birth-death population model with a fitness function based on protein folding stability. By incorporating structurally constrained substitution models and estimating fitness from ΔG values using homology-modeled structures, the authors aim to capture biophysically realistic evolutionary dynamics. The approach is implemented in a new version of their open-source software, ProteinEvolver2, and is applied to four viral proteins from HIV-1 and SARS-CoV-2.

Overall, the study presents a compelling rationale for using folding stability as a constraint in evolutionary simulations and offers a novel framework and software to explore such dynamics. While the results are promising, particularly for predicting biophysical properties, the current analysis provides only partial evidence for true evolutionary forecasting, especially at the sequence level. The work offers a meaningful conceptual advance and a useful simulation tool, and sets the stage for more extensive validation in future studies.

We also thank this reviewer for the positive comments on our study. Regarding the predictive power, our results showed good accuracy in predicting the folding stability of the forecasted protein variants. However, predicting the specific sequences of these variants is more challenging. For example, forecasting in amino acids with similar physicochemical properties can result in different sequences but in similar folding stability. We believe that these findings are realistic and interesting as they indicate that while forecasting folding stability is feasible, forecasting the specific sequence evolution is more complex that one could anticipate.

Strengths:

The results demonstrate that fitness constraints based on protein stability can prevent the emergence of unrealistic, destabilized variants - a limitation of traditional, neutral substitution models. In particular, the predicted folding stabilities of simulated protein variants closely match those observed in real variants, suggesting that the model captures relevant biophysical constraints.

We agree with the reviewer and appreciate the consideration that forecasting the folding stability of future real proteins is a relevant finding. For instance, folding stability is fundamental for protein function and affects several other molecular properties.

Weaknesses:

The predictive scope of the method remains limited. While the model effectively preserves folding stability, its ability to forecast specific sequence content is not well supported.

It is known that structurally constrained substitution (SCS) models applied to phylogenetics (such as ancestral molecular reconstruction) can model protein evolution with high accuracy in terms of folding stability, while inferring sequences (i.e., ancestral sequences) remains considerably more challenging (Arenas, et al. 2017; Arenas and Bastolla 2020). The observed sequence diversity is much higher than the observed structural diversity (Illergard, et al. 2009; Pascual-Garcia, et al. 2010), and substitutions between amino acids with similar physicochemical properties can result in protein variants with similar folding stability but with different specific amino acid composition. We will expand the discussion of this aspect in the manuscript.

Only one dataset (HIV-1 MA) is evaluated for sequence-level divergence using KL divergence; this analysis is absent for the other proteins. The authors use a consensus Omicron sequence as a representative endpoint for SARS-CoV-2, which overlooks the rich longitudinal sequence data available from GISAID. The use of just one consensus from a single time point is not fully justified, given the extensive temporal and geographical sampling available. Extending the analysis to include multiple timepoints, particularly for SARS-CoV-2, would strengthen the predictive claims. Similarly, applying the model to other well-sampled viral proteins, such as those from influenza or RSV, would broaden its relevance and test its generalizability.

The evaluation of forecasting evolution using real datasets is complex due to several conceptual and practical aspects. In contrast to traditional phylogenetic reconstruction of past evolutionary events and ancestral sequences, forecasting evolution often begins with a variant that is evolved forward in time and requires a rough fitness landscape to select among possible future variants (Lässig, et al. 2017). Another concern for validating the method is the need to know the initial variant that gives rise to the corresponding forecasted variants, and it is not always known. Thus, we investigated systems where the initial variant, or a close approximation, is known, such as scenarios of in vitro monitored evolution. In the case of SARS-CoV-2, the Wuhan variant is commonly used as the starting variant of the pandemic. Next, since forecasting evolution is highly dependent on the used model of evolution, unexpected external factors can be dramatic for the predictions. For this reason, systems with minimal external influences provide a more controlled context for evaluating forecasting evolution. For instance, scenarios of in vitro monitored virus evolution avoid some external factors such as host immune response. Another important aspect is the availability of data at two (i.e., present and future) or more time points along the evolutionary trajectory, with sufficient genetic divergence between them to identify clear evolutionary signatures. Additionally, using consensus sequences can help mitigate effects from unfixed mutations, which should not be modeled by a substitution model of evolution. Altogether, not all datasets are appropriate to properly evaluate forecasting evolution. We will include these considerations in the revised manuscript.

Sequence comparisons based on the KL divergence require, at the studied time point, an observed distribution of amino acid frequencies among sites and an estimated distribution of amino acid frequencies among sites. In the study datasets, this is only the case for the HIV-1 MA dataset, which belongs to a previous study from one of us and collaborators where we obtained at least 20 independent sequences at each sampling point (Arenas, et al. 2016). We will provide additional information on this aspect in the manuscript.

Regarding the Omicron dataset, we used 384 curated sequences of the Omicron variant of concern to construct the study dataset and we believe that it is a representative sample. The sequence used for the initial time point was the Wuhan variant (Wu, et al. 2020), which is commonly assumed to be the origin of the pandemic in SARS-CoV-2 studies. As previously indicated, the use of consensus sequences is convenient to avoid variants with unfixed mutations. Regarding extending the analysis to other timepoints (other variants of concern), we kindly disagree because Omicron is the variant of concern with the highest genetic distance to the Wuhan variant, and a high genetic distance is required to properly evaluate the prediction method. We noted that earlier variants of concern show a small number of fixed mutations in the study proteins, despite the availability of large numbers of sequences in databases such as GISAID.

Additionally, we investigated the evolutionary trajectories of HIV-1 protease (PR) in 12 intra-host viral populations.

Next, following the proposal of the reviewer, we will incorporate the analysis of an additional viral dataset (probably influenza following the suggestion of the reviewer) to further assess the generalizability of the method. Still, as previously indicated, not all datasets are suitable for a proper evaluation of forecasting evolution. Factors such as the shape of the fitness landscape and the amount of genetic variation over time can influence the accuracy of predictions. We will present the results of the analysis of the new data in the revised manuscript.

It would also be informative to include a retrospective analysis of the evolution of protein stability along known historical trajectories. This would allow the authors to assess whether folding stability is indeed preserved in real-world evolution, as assumed in their model.

Our present study is not focused on investigating the evolution of the folding stability over time, although it provides this information indirectly at the studied time points. Instead, the present study shows that the folding stability of the forecasted protein variants is similar to the folding stability of the corresponding real protein variants for diverse viral proteins, which is an important evaluation of the method. Next, the folding stability can indeed vary over time in both real and modeled evolutionary scenarios, and our present study is not in conflict with this. In that regard, which is not the aim of our present study, some previous phylogenetic-based studies have reported temporal fluctuations in folding stability for diverse data (Arenas, et al. 2017; Olabode, et al. 2017; Arenas and Bastolla 2020; Ferreiro, et al. 2022).

Finally, a discussion on the impact of structural templates - and whether the fixed template remains valid across divergent sequences - would be valuable. Addressing the possibility of structural remodeling or template switching during evolution would improve confidence in the model's applicability to more divergent evolutionary scenarios.

This is an important point. For the datasets that required homology modeling (in several cases it was not necessary because the sequence was present in a protein structure of the PDB), the structural templates were selected using SWISS-MODEL, and we applied the best-fitting template. We will include additional details about the parameters of the homology modeling in the revised version. Indeed, our method assumes that the protein structure is maintained over the studied evolutionary time, which can be generally reasonable for short timescales where the structure is conserved (Illergard, et al. 2009; Pascual-Garcia, et al. 2010). Over longer evolutionary timescales, structural changes may occur, and in such cases, modeling the evolution of the protein structure would be necessary. To our knowledge, modeling the evolution of the protein structure remains a challenging task that requires substantial methodological developments. Recent advances in artificial intelligence, particularly in protein structure prediction from sequence, may offer promising tools for addressing this challenge. However, we believe that evaluating such approaches in the context of structural evolution would be difficult, especially given the limited availability of real data with known evolutionary trajectories involving structural change. In any case, this is probably an important direction for future research. We will include this discussion in the revised manuscript.

Cited references

Arenas M. 2012. Simulation of Molecular Data under Diverse Evolutionary Scenarios. PLoS Comput Biol 8:e1002495.

Arenas M, Bastolla U. 2020. ProtASR2: Ancestral reconstruction of protein sequences accounting for folding stability. Methods Ecol Evol 11:248-257.

Arenas M, Dos Santos HG, Posada D, Bastolla U. 2013. Protein evolution along phylogenetic histories under structurally constrained substitution models. Bioinformatics 29:3020-3028.

Arenas M, Lorenzo-Redondo R, Lopez-Galindez C. 2016. Influence of mutation and recombination on HIV-1 in vitro fitness recovery. Molecular Phylogenetics and Evolution 94:264-270.

Arenas M, Sanchez-Cobos A, Bastolla U. 2015. Maximum likelihood phylogenetic inference with selection on protein folding stability. Molecular Biology and Evolution 32:2195-2207.

Arenas M, Weber CC, Liberles DA, Bastolla U. 2017. ProtASR: An Evolutionary Framework for Ancestral Protein Reconstruction with Selection on Folding Stability. Systematic Biology 66:1054-1064.

Bordner AJ, Mittelmann HD. 2013. A new formulation of protein evolutionary models that account for structural constraints. Molecular Biology and Evolution 31:736-749.

Carvajal-Rodriguez A. 2010. Simulation of genes and genomes forward in time. Current Genomics 11:58-61.

Echave J, Spielman SJ, Wilke CO. 2016. Causes of evolutionary rate variation among protein sites. Nature Reviews Genetics 17:109-121.

Echave J, Wilke CO. 2017. Biophysical Models of Protein Evolution: Understanding the Patterns of Evolutionary Sequence Divergence. Annu Rev Biophys 46:85-103.

Ferreiro D, Khalil R, Gallego MJ, Osorio NS, Arenas M. 2022. The evolution of the HIV-1 protease folding stability. Virus Evol 8:veac115.

Goldstein RA. 2013. Population Size Dependence of Fitness Effect Distribution and Substitution Rate Probed by Biophysical Model of Protein Thermostability. Genome Biol Evol 5:1584-1593.

Harmon LJ. 2019. Introduction to birth-death models. In. Phylogenetic Comparative Methods. p. https://lukejharmon.github.io/pcm/chapter10_birthdeath/.

Hoban S, Bertorelle G, Gaggiotti OE. 2012. Computer simulations: tools for population and evolutionary genetics. Nature Reviews Genetics 13:110-122.

Illergard K, Ardell DH, Elofsson A. 2009. Structure is three to ten times more conserved than sequence--a study of structural response in protein cores. Proteins 77:499-508.

Lässig M, Mustonen V, Walczak AM. 2017. Predicting evolution. Nature Ecology & Evolution 1:0077.

Liberles DA, Teichmann SA, Bahar I, Bastolla U, Bloom J, Bornberg-Bauer E, Colwell LJ, de Koning AP, Dokholyan NV, Echave J, et al. 2012. The interface of protein structure, protein biophysics, and molecular evolution. Protein Science 21:769-785.

Neher RA, Russell CA, Shraiman BI. 2014. Predicting evolution from the shape of genealogical trees. Elife 3.

Olabode AS, Kandathil SM, Lovell SC, Robertson DL. 2017. Adaptive HIV-1 evolutionary trajectories are constrained by protein stability. Virus Evol 3:vex019.

Pascual-Garcia A, Abia D, Mendez R, Nido GS, Bastolla U. 2010. Quantifying the evolutionary divergence of protein structures: the role of function change and function conservation. Proteins 78:181-196.

Wilke CO. 2012. Bringing molecules back into molecular evolution. PLoS Comput Biol 8:e1002572.

Wu F, Zhao S, Yu B, Chen YM, Wang W, Song ZG, Hu Y, Tao ZW, Tian JH, Pei YY, et al. 2020. A new coronavirus associated with human respiratory disease in China. Nature 579:265-269.

Yang Z. 2006. Computational Molecular Evolution. Oxford, England.: Oxford University Press.

Author response:

Reviewer #1 (Public review):

Summary:

In this work, the authors have developed SPLASH+, a micro-assembly and biological interpretation framework that expands on their previously published reference-free statistical approach (SPLASH) for sequencing data analysis.

Thank you for this thorough overview of our work.

Strengths:

(1) The methodology developed by the authors seems like a promising approach to overcome many of the challenges posed by reference-based single-cell RNA-seq analysis methods.

Thank you for your positive comment on the potential of our approach to address the limitations of reference-based methods for scRNA-Seq analysis.

(2) The analysis of the RNU6 repetitive small nuclear RNA provides a very compelling example of a type of transcript that is very challenging to analyze with standard reference-based methods (e.g., most reads from this gene fail to align with STAR, if I understood the result correctly).

We thank the reviewer for their positive comment. We agree that the variation in RNU6 detected by SPLASH+ underscores the potential of our reference-free method to make discoveries in cases where reference-based approaches fall short.

Weaknesses:

(1) The manuscript presents a number of case studies from very diverse domains of single-cell RNA-seq analysis. As a result, the manuscript has been challenging to review, because it requires domain expertise in centromere biology, RNA splicing, RNA editing, V(D)J transcript diversity, and repeat polymorphisms.

We appreciate the reviewer’s effort in thoroughly evaluating this manuscript, especially given the broad range of biological domains discussed. Our main goal in presenting a wide range of applications was to highlight the key strength of the SPLASH+ framework: its ability to unify diverse biological discoveries within a single method that operates directly on sequencing reads.

(2) Although the paper focuses on SmartSeq2 full-length single-cell RNA-seq data analysis, the vast majority of single-cell RNA-seq data that is currently being generated comes from droplet-based methods (e.g., 10x Genomics) that sequence only the 3' or 5' ends of transcripts. As a result, it is unclear if SPLASH+ is also applicable to these types of data.

We thank the reviewer for this comment. Due to the specific data format of barcoded single-cell sequencing platforms such as 10x Genomics, extending the SPLASH framework to support 10x analysis required engineering a specialized preprocessing tool. We have addressed this in a recent work, which is now available as a preprint (https://doi.org/10.1101/2024.12.24.630263).

(3) The criteria used for the selection of the 10 'core genes' have not been sufficiently justified.

We chose these genes as SPLASH+ detected regulated splicing for them in nearly all tissues (18 out of 19)  analyzed in our study (i.e., identifying anchors classified as splicing anchors in those tissues). Our subsequent analysis showed that all these genes are involved in either splicing regulation or histone modification. We will further clarify this selection criterion in the revision. 

(4) It is currently unclear how the splicing diversity discovered in this paper relates to the concept of noisy splicing (i.e., there are likely many low-frequency transcripts and splice junctions that are unlikely to have a significant functional impact beyond triggering nonsense-mediated decay).

In our analysis, to ensure sufficient read coverage, we considered significant anchors supported by more than 50 reads and detected in over 10 cells. Additionally, our downstream analyses (including splicing analysis) are based on assembled sequences (compactors) generated through our micro-assembly step. This process effectively acts as a denoising step by filtering out sequences likely caused by sequencing errors or with very low read support. However, we agree that the detected splice variants have not been fully functionally characterized, and further functional experiments may be needed.

(5) The paper presents only a very superficial discussion of the potential weaknesses of the SPLASH+ method.

We discussed two potential limitations of SPLASH+ in the Conclusions section: (1) it is not suitable for differential gene expression analysis, and (2) although we provide a framework for interpreting and analyzing SPLASH results, further work is still needed to improve the annotation of calls lacking BLAST matches. We will add more discussion for these in the revision. 

(6) The cursory mention of metatranscriptome in the conclusion of the paper is confusing, as it might suggest the presence of microbial cells in sterile human tissues (which has recently been discredited in cancer, see e.g. https://www.science.org/content/article/journal-retracts-influential-cancer-microbiome-paper).

We will remove the mention of metatranscriptome in the revised manuscript.

Reviewer #2 (Public review):

The authors extend their SPLASH framework with single-cell RNA-seq in mind, in two ways. First, they introduce "compactors", which are possible paths branching out from an anchor. Second, they introduce a workflow to classify compactors according to the type of biological sequence variation represented (splicing, SNV, etc). They focus on simulated data for fusion detection, and then focus on analyzing the Tabula sapiens Smart-seq2 data, showing extensive results on alternative splicing analysis, VDJ, and repeat elements.

This is strong work with an impressive array of biological investigations and results for a methods paper. I have various concerns about terminology and comparisons, as follows (in a somewhat arbitrary order, apologies).

Thank you for this thorough overview of our work and your positive comment on the strength of our work.

(1) The discussion of the weaknesses of the consensus sequence approach of SPLASH is an odd way to motivate SPLASH+ in my opinion, in that SPLASH is not yet so widely used, so the baseline for SPLASH+ is really standard alignment-based approaches. It is fine to mention consensus sequence issues briefly, but it felt belabored.

We thank the reviewer and agree that the primary comparison for SPLASH+ is with reference-based methods. However, since SPLASH+ builds upon SPLASH, we also aimed to highlight the limitations of the consensus step in original SPLASH and how SPLASH+ addresses them. To maintain the main focus of the paper on comparison with reference-based methods and biological investigations, this discussion with consensus was provided in a Supplementary Figure. We will shorten this discussion in the revision.

(2) Regarding compactors reducing alignment cost: the comparison should really be between compactor construction and alignment vs read alignment (and maybe vs modern contig construction algorithms and alignment).

Since the SPLASH framework is fundamentally reference-free and does not require read alignment, we compared the number of sequence alignments for compactors to the total read alignments required by a reference-based method to show that while compactors are aligned to the reference, the number of alignments needed is still orders of magnitude less than a reference-based approach requiring alignment of all the reads.

(3) The language around "compactors" is a bit confusing, where the authors sometimes refer to the tree of possibilities from an anchor as a "compactor", and sometimes a compactor is a single branch. Presumably, ideally, compactors should be DAGs, not trees, i.e., they can connect back together. Perhaps the authors could comment on whether this matters/would be a valuable extension.

We thank the reviewer for their comment. We refer to each generated assembled sequence as “a compactor”, and we attempted to make this clear in the paper. We will review the text further to ensure this definition is clear in the revised version.

(4) The main oddness of the splicing analysis to me is not using cell-type/state in any way in the statistical testing. This need not be discrete cell types: psiX, for example, tested whether exonic PSI was variable with reference to a continuous gene expression embedding. Intuitively, such transcriptome-wide signal should be valuable for a) improving power and b) distinguishing cell-type intrinsic/"noisy" from cell-type specific splicing variation. A straightforward way of doing this would be pseudobulking cell types. Possibly a more sophisticated hierarchical model could be constructed also.

We appreciate the reviewer’s concern regarding SPLASH+ not using cell type metadata. SPLASH, which performs the core statistical inference in SPLASH+, is an unsupervised tool specifically designed to make biological discoveries without relying on metadata (such as cell type annotations in scRNA-Seq). This is particularly useful in scRNA-seq, where cell type labels could be missing, imprecise, or may miss important within-cell-type variation. As shown in the paper, even without using metadata, SPLASH+ demonstrated improved performance than both SpliZ and Leafcutter (two metadata-dependent tools) in terms of achieving higher concordance and identifying more differentially spliced genes. Regarding pseudobulking, as has been shown in the SpliZ paper (https://doi.org/10.1038/s41592-022-01400-x), pseudobulking requires multiple pseudobulked replicates per cell type for reliable inference, which is often not feasible in scRNA-seq settings, making such methods statistically suboptimal for single-cell studies. We will add a discussion on pseudobulking in the revision. 

(5) A secondary weakness is that some informative reads will not be used, for example, unspliced reads aligning to an alterantive exons. This relates to the broader weakness of SPLASH that it is blind to changes in coverage that are not linked to a specific anchor (which should be acknowledged somewhere, maybe in the Discussion). In the deeply sequenced SS2 data, this is likely not an issue, but might be more limiting in sparser data. A related issue is that coverage change indicative of, e.g., alternative TSS or TES (that do not also include a change in splice junction use) will not be detected. In fairness, all these weaknesses are shared by LeafCutter. It would be valuable to have a comparison to a more "traditional" splicing analysis approach (pick your favorite of rMATS, MISO, SUPPA).

We thank the reviewer for their comment. As noted in the Conclusion, the SPLASH framework is not designed for differential gene expression analysis, which relies on quantifying read coverage. Rather, it focuses on detecting differential sequence diversity arising from mechanisms like alternative splicing or RNA editing. We will clarify this limitation further in the revised Conclusion. 

Regarding splicing evaluation, we have performed extensive comparisons with two widely used and recent methods—SpliZ and Leafcutter—for both bulk and single-cell splicing analysis. While we appreciate the reviewer’s suggestion to include an additional method, given the current length of the paper and the fact that leafcutter has previously been shown to outperform rMATS, MAJIQ, and Cufflinks2

(https://www.nature.com/articles/s41588-017-0004-9), we believe the current comparisons provide sufficient support for the evaluation of the splicing detection by SPLASH+.

(6) "We should note that there is no difference between gene fusions and other RNA variants (e.g., RNA splicing) from a sequence assembly viewpoint". Maybe this is true in an abstract sense, but I don't think it is in reality. AS can produce hundreds of isoforms from the same gene, and be variable across individual cells. Gene fusions are generally less numerous/varied and will be shared across clonal populations, so the complexity is lower. That simplicity is balanced against the challenge that any genes could, in principle, fuse.

We selected the fusion benchmarking dataset solely to evaluate how well compactors reconstruct sequences. Since our goal was to assess the accuracy of reconstructed compactor sequences, we needed a benchmarking dataset with ground truth sequences, which this dataset provides. We had explained our main reason and purpose for selecting fusion dataset in the text, but we will clarify it further in the revision.

(7) For the fusion detection assessment, SPLASH+ is given the correct anchor for detection. This feels like cheating since this information wouldn't usually be available. Can the authors motivate this? Are the other methods given comparable information? Also, TPM>100 seems like a very high expression threshold for the assessment.

We agree with the reviewer that the fusion benchmarking dataset should not be used to assess the entire SPLASH+ framework. In fact, we did not use this dataset to evaluate SPLASH+; it was used exclusively to evaluate the performance of compactors as a standalone module. Specifically, we tested how well compactors can reconstruct fusion sequences when provided with seed sequences corresponding to fusion junctions. This aligns with our expectation from compactors in SPLASH+, that they should correctly reconstruct the sequence context for the detected anchors. As noted in our previous response, since our goal was to assess the accuracy of reconstructed compactor sequences, we required a benchmarking dataset with ground truth sequences, which this dataset provides. We will clarify this further in the revision.

We appreciate the reviewer’s concern that a TPM of 100 is high. In Figure 1C, we presented the full TPM distribution for fusions missed or detected by compactors. The 100 threshold was an arbitrary benchmark to illustrate the clear difference in TPM profiles between these two sets of fusions. We will clarify this point in the revised manuscript.

(8) Why are only 3'UTRs considered and not 5'? Is this because the analysis is asymmetric, i.e., only considering upstream anchors and downstream variation? If so, that seems like a limitation: how much additional variation would you find if including the other direction?

We thank the reviewer for their comment. SPLASH+ can, in principle, detect variation in 5’ UTR regions, as demonstrated by the variations observed in the 5’ UTRs of the genes ANPC16 and ARPC2. If sequence variation exists in the 5′ UTR, SPLASH+ can still detect it by identifying an anchor upstream of the variable region, as it directly parses sequencing reads to find anchors with downstream sequence diversity. Even when the variation occurs near the 5′ end of the 5′ UTR, SPLASH+ can still capture this diversity if the user selects a shorter anchor length.

(9) I don't find the theoretical results very meaningful. Assuming independent reads (equivalently binomial counts) has been repeatedly shown to be a poor assumption in sequencing data, likely due to various biases, including PCR. This has motivated the use of overdispersed distributions such as the negative Binomial and beta binomial. The theory would be valuable if it could say something at a specified level of overdispersion. If not, the caveat of assuming no overdispersion should be clearly stated.

We appreciate the reviewer’s comment. We will clarify this in the revised paper.

Author response:

The following is the authors’ response to the original reviews

We thank the reviewers for the careful review of our manuscript. Overall, they were positive about our use of cutting-edge methods to identify six inversions segregating in Lake Malawi. Their distribution in ~100 species of Lake Malawi species demonstrated that they were differentially segregating in different ecogroups/habitats and could potentially play a role in local adaptation, speciation, and sex determination. Reviewers were positive about our finding that the chromosome 10 inversion was associated with sex-determination in a deep benthic species and its potential role in regulating traits under sexual selection. They agree that this work is an important starting point in understanding the role of these inversions in the amazing phenotypic diversity found in the Lake Malawi cichlid flock.

There were two main criticisms that were made which we summarize:

(1) Lack of clarity. It was noted that the writing could be improved to make many technical points clearer. Additionally, certain discussion topics were not included that should be.

We will rewrite the text and add additional figures and tables to address the issues that were brought up in a point-by-point response. We will improve/include (1) the nomenclature to understand the inversions in different lineages, (2) improved descriptions for various genomic approaches, (3) a figure to document the samples and technologies used for each ecogroup, and 4) integration of LR sequences to identify inversion breakpoints to the finest resolution possible.

(2) We overstate the role that selection plays in the spread of these inversions and neglect other evolutionary processes that could be responsible for their spread.

We agree with the overarching point. We did not show that selection is involved in the spread of these inversions and other forces can be at play. Additionally, there were concerns with our model that the inversions introgressed from a Diplotaxodon ancestor into benthic ancestors and incomplete lineage sorting or balancing selection (via sex determination) could be at play. Overall, we agree with the reviewers with the following caveats. 1. Our analysis of the genetic distance between Diplotaxodons and benthic species in the inverted regions is more consistent with their spread through introgression versus incomplete lineage sorting or balancing selection. 2. Further the role of these inversions is likely different in different species. For example, the inversion of 10 and 11 play a role in sex determination in some species but not others and the potential pressures acting on the inverted and non-inverted haplotypes will be very different. These are very interesting and important questions booth for understanding the adaptive radiations in Lake Malawi and in general, and we are actively studying crosses to understand the role of these inversions in phenotypic variation between two species. We will modify the text to make all of these points clearer.

Public Reviews:

Reviewer #1 (Public review):

Summary:

Using high-quality genomic data (long-reads, optical maps, short-reads) and advanced bioinformatic analysis, the authors aimed to document chromosomal rearrangements across a recent radiation (Lake Malawi Cichlids). Working on 11 species, they achieved a high-resolution inversion detection and then investigated how inversions are distributed within populations (using a complementary dataset of short-reads), associated with sex, and shared or fixed among lineages. The history and ancestry of the inversions is also explored.

On one hand, I am very enthusiastic about the global finding (many inversions well-characterized in a highly diverse group!) and impressed by the amount of work put into this study. On the other hand, I have struggled so much to read the manuscript that I am unsure about how much the data supports some claims. I'm afraid most readers may feel the same and really need a deep reorganisation of the text, figures, and tables. I reckon this is difficult given the complexity brought by different inversions/different species/different datasets but it is highly needed to make this study accessible.

The methods of comparing optical maps, and looking at inversions at macro-evolutionary scales can be useful for the community. For cichlids, it is a first assessment that will allow further tests about the role of inversions in speciation and ecological specialisation. However, the current version of the manuscript is hardly accessible to non-specialists and the methods are not fully reproducible.

Strengths:

(1) Evidence for the presence of inversion is well-supported by optical mapping (very nice analysis and figure!).

(2) The link between sex determination and inversion in chr 10 in one species is very clearly demonstrated by the proportion in each sex and additional crosses. This section is also the easiest to read in the manuscript and I recommend trying to rewrite other result sections in the same way.

(3) A new high-quality reference genome is provided for Metriaclima zebra (and possibly other assemblies? - unclear).

(4) The sample size is great (31 individuals with optical maps if I understand well?).

(5) Ancestry at those inversions is explored with outgroups.

(6) Polymorphism for all inversions is quantified using a complementary dataset.

Weaknesses:

(1) Lack of clarity in the paper: As it currently reads, it is very hard to follow the different species, ecotypes, samples, inversions, etc. It would be useful to provide a phylogeny explicitly positioning the samples used for assembly and the habitat preference. Then the text would benefit from being organised either by variant or by subgroups rather than by successive steps of analysis.

We have extensively rewritten the paper to improve the clarity. With respect to this point, we moved Figure 6 to Figure 1, which places the phylogeny of Lake Malawi cichlids at the beginning of the paper. We incorporated information about samples/technologies by ecogroup into this figure to help the reader gain an overview of the technologies involved. We added information about habitat for each ecogroup as well. While we considered a change to the text organization suggested here, we thought it was clearer to keep the original headings.

(2) Lack of information for reproducibility: I couldn't find clearly the filters and parameters used for the different genomic analyses for example. This is just one example and I think the methods need to be re-worked to be reproducible. Including the codes inside the methods makes it hard to follow, so why not put the scripts in an indexed repository?

We now provide a link to a github repository (https://github.com/ptmcgrat/CichlidSRSequencing/tree/Kumar_eLife) containing the scripts used for the major analysis in the paper. Because our data is behind a secure Dropbox account, readers will not be able to run the analysis, however, they can see the exact programs, filters, and parameters used for manuscript embedded within each script.

(3) Further confirmation of inversions and their breakpoints would be valuable. I don't understand why the long-reads (that were available and used for genome assembly) were not also used for SV detection and breakpoint refinement.

We did use long reads to confirm the presence of the inversions by creating five new genome assemblies from the PacBio HiFi reads: two additional Metriaclima zebra samples and three Aulonocara samples. Alignment of these five genomes to the MZ_GT3 reference is shown in Figures S2 – S7. These genome assemblies were also used to identify the breakpoints of the inversions. However, because of the extensive amount of repetitive DNA at the breakpoints (which is known to be important for the formation of large inversions), our ability to resolve the breakpoints was limited.

(4) Lack of statistical testing for the hypothesis of introgression: Although cichlids are known for high levels of hybridization, inversions can also remain balanced for a long time. what could allow us to differentiate introgression from incomplete lineage sorting?

The coalescent time between the inversions between Diplotaxodons and benthics should allow us to distinguish these two mechanisms. Our finding that the genetic distance, which is related to coalescent time, is closer within the inversions than the whole genome is supportive of introgression. However, we did not perform any simulations or statistical tests. We make it clearer in the text that incomplete lineage sorting remains a possible mechanism for the distribution of inversions within these ecogroups.

(5) The sample size is unclear: possibly 31 for Bionano, 297 for short-reads, how many for long-reads or assemblies? How is this sample size split across species? This would deserve a table.

We have included this information in the new Figure 1.

(6) Short read combines several datasets but batch effect is not tested.

We do not test for batch effect. However, we do note that all of the datasets were analyzed by the same pipeline starting from alignment so batch effects would be restricted to aspects of the reads themselves. Additionally, samples from the different data sets clustered as expected by lineage and inferred inversion, so for these purposes unlikely to have affected analysis.

(7) It is unclear how ancestry is determined because the synteny with outgroups is not shown.

Ancestry analysis was determined using the genome alignments of two outgroups from outside of Lake Malawi. This is shown in Figure S8.

(8) The level of polymorphism for the different inversions is difficult to interpret because it is unclear whether replicated are different species within an eco-group or different individuals from the same species. How could it be that homozygous references are so spread across the PCA? I guess the species-specific polymorphism is stronger than the ancestral order but in such a case, wouldn't it be worth re-doing the PCa on a subset?

The genomic PCA plots reflect the evolutionary histories that are observed in the whole genome phylogenies. Because the distribution of the inverted alleles violate the species tree, they form separate clusters on the PCA plots that can be used to genotype specific species. We have also performed this analysis on benthics (utaka/shallow benthics/deep benthics) and the distribution matches the expectation.

Reviewer #2 (Public review):

Summary:

Chromosomal inversions have been predicted to play a role in adaptive evolution and speciation because of their ability to "lock" together adaptive alleles in genomic regions of low recombination. In this study, the authors use a combination of cutting-edge genomic methods, including BioNano and PacBio HiFi sequencing, to identify six large chromosomal inversions segregating in over 100 species of Lake Malawi cichlids, a classic example of adaptive radiation and rapid speciation. By examining the frequencies of these inversions present in species from six different linages, the authors show that there is an association between the presence of specific inversions with specific lineages/habitats. Using a combination of phylogenetic analyses and sequencing data, they demonstrate that three of the inversions have been introduced to one lineage via hybridization. Finally, genotyping of wild individuals as well as laboratory crosses suggests that three inversions are associated with XY sex determination systems in a subset of species. The data add to a growing number of systems in which inversions have been associated with adaptation to divergent environments. However, like most of the other recent studies in the field, this study does not go beyond describing the presence of the inversions to demonstrate that the inversions are under sexual or natural selection or that they contribute to adaptation or speciation in this system.

Strengths:

All analyses are very well done, and the conclusions about the presence of the six inversions in Lake Malawi cichlids, the frequencies of the inversions in different species, and the presence of three inversions in the benthic lineages due to hybridization are well-supported. Genotyping of 48 individuals resulting from laboratory crosses provides strong support that the chromosome 10 inversion is associated with a sex-determination locus.

Weaknesses:

The evidence supporting a role for the chromosome 11 inversion and the chromosome 9 inversion in sex determination is based on relatively few individuals and therefore remains suggestive. The authors are mostly cautious in their interpretations of the data. However, there are a few places where they state that the inversions are favored by selection, but they provide no evidence that this is the case and there is no consideration of alternative hypotheses (i.e. that the inversions might have been fixed via drift).

We have removed mention of chromosome 9’s potential role in sex determination from the paper. While our analysis of sex association with chromosome 11 was limited compared to our analysis of chromosome 10, it was still statistically significant, and we believe it should be left in the paper. The role of 11 (and 9 and 10) in sex determination was also demonstrated using an independent dataset by Blumer et al (https://doi.org/10.1101/2024.07.28.605452)

We agree that we did not properly consider alternative hypothesis in the original submission and have rewritten the Discussion substantially to consider various alternative hypothesis.

Reviewer #3 (Public review):

This is a very interesting paper bringing truly fascinating insight into the genomic processes underlying the famous adaptive radiation seen in cichlid fishes from Lake Malawi. The authors use structural and sequence information from species belonging to distinct ecotypic categories, representing subclades of the radiation, to document structural variation across the evolutionary tree, infer introgression of inversions among branches of the clade, and even suggest that certain rearrangements constitute new sex-determining loci. The insight is intriguing and is likely to make a substantial contribution to the field and to seed new hypotheses about the ecological processes and adaptive traits involved in this radiation.

I think the paper could be clarified in its prose, and that the discussion could be more informative regarding the putative roles of the inversions in adaptation to each ecotypic niche. Identifying key, large inversions shared in various ways across the different taxa is really a great step forward. However, the population genomics analysis requires further work to describe and decipher in a more systematic way the evolutionary forces at play and their consequences on the various inversions identified.

The model of evolution involving multiple inversions putatively linking together co-adapted "cassettes" could be better spelled out since it is not entirely clear how the existing theory on the recruitment of inversions in local adaptation (e.g. Kirkpatrick and Barton) operates on multiple unlinked inversions. How such loci correspond to distinct suites of integrated traits, or not, is not very easy to envision in the current state of the manuscript.

This is a very interesting point, and we agree creates complications for a simple model of local adaptation. We imagine though that the actual evolutionary history was much more complicated than a single Rhamphochromis-type species separating from a single Diplotaxodon-type species and could have occurred sequentially involving multiple species that are now extinct. A better understanding of the role each of these inversions play in phenotypic diversity could potentially help us determine if different inversions carry variation that could be linked to distinct habit differences. We have added a line to the discussion.

The role of one inversion in sex determination is apparent and truly intriguing. However, the implication of such locus on ecological adaptation is somewhat puzzling. Also, whether sex determination loci can flow across species via introgression seems quite important as a route to chromosomal sex determination, so this could be discussed further.

Another very interesting point. If the inversions are involved in ecological adaptation (an important caveat), then potentially the inverted and non-inverted haplotypes play dual roles in the Aulonocara animals with the inverted haplotype carrying adaptive alleles to deep water and the non-inverted haplotype carrying alleles resolving sexual conflict. We have broadened our discussion about their function at the origin including non-adaptive roles.

Recommendations for the authors:

Reviewer #2 (Recommendations for the authors):

Overall, the paper is well-written and clear. I do have a few suggestions for changes that would help the reader:

(1) Figure 1: the figure legend could be expanded here to help the reader; what are the blue and yellow lines? Why are there two lines for the GT3a assembly? And, I had to somehow read the legend a few times to understand that the top line is the UMD2a reference assembly, and the next line is the new Bionano map.

Fixed in what is now Figure 2

(2) Paragraph starting on line 133: you use the word "test" to refer to the Bionano analyses; it is not clear whether anything is being tested. Perhaps "analyse the maps" or just "map" would be more clear? Or more explanation?

The text has been modified to address this point

(3) L145-146: perhaps change "a single inversion" and "a double inversion" to "single inversions" and "double inversions".

The text has been modified to address this point

(4) L157: suppression of recombination in inversion heterozygotes is "textbook" material and perhaps does not need a reference. Or, you could reference an empirical paper that demonstrates this point. Though I love the Kirkpatrick and Barton paper, it certainly is not the correct reference for this point.

The Kirkpatrick reference was incorrectly included here. The correct reference was an empirical demonstration (Conte) that there were regions of suppressed recombination that have been observed in the location of the inversions. We have also moved this reference further up in the sentence to a more appropriate position

(5) L173: how do you know this is an assembly error and not polymorphism?

The text has been modified to address this point

(6) L277(?): "currently growing in the lab" is probably unnecessary.

The text has been modified to address this point

(7) L298: "the inversion on 10 acts as an XY sex determiner": the inversion itself is not the sex determination gene; rather, it is linked. I think it would be more precise, here and throughout the paper, to say that these inversions likely harbor the sex determination locus (for example, the wording on lines 369-370 is misleading).

We agree with the larger point that the inversion might not be causal for sex determination, however, it could still be causal through positional effects. We have modified the text to make it clear that it could also carry the causal locus (or loci).

(8) Figure 6: overall, this figure is very helpful! However, it contains several problematic statements. In no case do you have evidence that these inversions are "favored by selection"; such statements should be deleted. Also, in point 3, you state that inversions 9, 11, and 20 are transferred to benthic lineages, and then that these inversions are involved in sex determination. But, your data suggests that it is chromosomes 9, 10, and 11 that are linked to sex determination.

This figure is now Figure 1. We have remove these problematic statements.

(9) L356-360: I would move the references that are currently at the end of the sentence to line 357 after the statement about the previous work on hybridization. Otherwise, it reads as if these previous papers demonstrated what you have demonstrated in your work.

The text has been modified to address this point

(10) Overall, the discussion focuses completely on adaptive explanations for your results, and I would like to see at least an acknowledgement that drift could also be involved unless you have additional data to support adaptive explanations.

We have rewritten the text to account for the possibility of drift (line 404 and 405).

Reviewer #3 (Recommendations for the authors):

The paper utilizes heterogeneous datasets coming from different sources, and it is not always clear which specimens were used to generate structural information (bionano) or sequence information. A diagram summarizing the sequence data, methodologies, and research questions would be beneficial for the reader to navigate in this paper.

Much of this information has been added to what is now Figure 1. All of this data is also found in Table S2.

The authors performed genome alignments to analyze and homologize inversion, but this process is not clearly described. For the PCA, SNP information likely involves mapping onto a common reference genome. However, it is not clear how this was achieved given the different species and varying divergence times involved.

We now include a link to the github that contains the commands that were run. Because the overall level of sequence divergence between cichlid species is quite low (2*10^-3 – Milansky et al), mapping different species onto a common reference is commonly performed in Lake Malawi cichlids.

The introgression scenario is very intriguing but its role in local adaptation of the ecogroup types is not easy to understand. I understand this is still an outstanding question, but it is unclear how the directionality of introgressions was estimated. This can be substantiated using tree topology analysis, comparative estimates of sequence divergence, and accumulation of DNA insertions. The diagram does not clearly indicate which ones are polymorphic. In some cases, polymorphic inversions could result from the coexistence of native and introgressed haplotypes.

We agree that this analysis would be interesting but is beyond the scope of this paper.

The alternative model of introgression proposed in the cited preprint is interesting and should deserve a formal analysis here. The authors consider unclear what would drive "back" introgressions of non-inverted haplotypes, but this would depend on the selection regimes acting on the inversions themselves, which can include forms of balancing selection and a role for recessive lethals (heterozygote advantage). For instance, a standard haplotype could be favored if it shelters deleterious mutations carried by an inversion. Testing the introgression history over a wider range of branches and directions would provide further insights.

We agree that this analysis would be interesting but is beyond the scope of this paper.

The prose in the paper is occasionally muddled and somewhat unclear. Referring to chromosomes solely by their numbers (e.g.. "inversion on 11") complicates readability.

This is the standard way to refer to chromosomes in cichlids and we believe while it complicates readability, any other method would be inconsistent with other papers. Changes to nomenclature might improve the readability of this paper, but would make it more difficult to compare results for these chromosomes from other papers with what we have found.

I believe that harassment is never justified. Harassment involves actions like online insults, cyberstalking, and invasion of personal information to harm a user. While some people may think that harassment is acceptable when directed at extremists such as racists, white supremacists, or sexists. While I strongly disagree, there are clearly better ways to address such issues than resorting to harassment. For example, we can use facts and logic to refute their views instead of launching personal attacks, or report their behavior through legal and official channels.

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

The Authors investigated the anatomical features of the excitatory synaptic boutons in layer 1 of the human temporal neocortex. They examined the size of the synapse, the macular or the perforated appearance and the size of the synaptic active zone, the number and volume of the mitochondria, the number of the synaptic and the dense core vesicles, also differentiating between the readily releasable, the recycling and the resting pool of synaptic vesicles. The coverage of the synapse by astrocytic processes was also assessed, and all the above parameters were compared to other layers of the human temporal neocortex. The Authors conclude that the subcellular morphology of the layer 1 synapses is suitable for the functions of the neocortical layer, i.e. the synaptic integration within the cortical column. The low glial coverage of the synapses might allow the glutamate spillover from the synapses enhancing synaptic crosstalk within this cortical layer.

Strengths:

The strengths of this paper are the abundant and very precious data about the fine structure of the human neocortical layer 1. Quantitative electron microscopy data (especially that derived from the human brain) are very valuable, since this is a highly time- and energy consuming work. The techniques used to obtain the data, as well as the analyses and the statistics performed by the Authors are all solid, strengthen this manuscript, and mainly support the conclusions drawn in the discussion.

Comments on latest version:

The corrected version of the article titled “Ultrastructural sublaminar specific diversity of excitatory synaptic boutons in layer 1 of the adult human temporal lobe neocortex" has been improved thanks to the comments and suggestions of the reviewers. The Authors implemented several of my comments and suggestions. However, many of them were not completed. It is understandable that the Authors did not start a whole new series of experiments investigating inhibitory synapses (as it was a misunderstanding affecting 2 reviewers from the three). But the English text is still very hard to understand and has many mistakes, although I suggested to extensively review the use of English. Furthermore, my suggestion about avoiding many abbreviations in the abstract, analyse and discuss more the perforated synapses, the figure presentation (Figure 3) and including data about the astrocytic coverage in the Results section were not implemented. My questions about the number of docked vesicles and p10 vesicles, as well as about the different categories of the vesicle pools have not been answered neither. Many other minor comments and suggestions were answered, corrected and implemented, but I think it could have been improved more if the Authors take into account all of the reviewers' suggestions, not only some of them. I still have several main and minor concerns, with a few new ones as well I did not realize earlier, but still think it is important.

We would like to thank the reviewer for the comments.

- We worked on the English again and tried to improve the language.

- We avoided to use too many abbreviations in the Abstract and reduced them to a minimum.

- We included a small paragraph about non-perforated vs. perforated active zones in both the Results and Discussion sections. However, since the majority of active zones in all cortical layers of the human TLN were of the macular type, we concluded that it is not relevant to describe their function in more detail.

- In Figure 3 A-C we added contour lines to the boutons to make their outlines more visible.

- We completed the data about the astrocytic coverage in the Results section (see also below).

- Concerning the vesicle pools please see below.

Main concerns:

(1) Epileptic patients:

As all patients were epileptic, it is not correct to state in the abstract that non-epileptic tissue was investigated. Even if the seizure onset zone was not in the region investigated, seizures usually invade the temporal lobe in TLE. If you can prove that no spiking activity occurred in the sample you investigated and the seizures did not invade that region, then you can write that it is presumably non-epileptic. I would suggest to write “L1 of the human temporal lobe neocortical biopsy tissue". See also Methods lines 608-612. Write only “non-epileptic" or “non-affected" if you verified it with EcoG. If this was the case, please write a few sentences about it in the Methods.

We rephrased Material and Methods concerning this point and added that patients were monitored with EEG, MRI and multielectrode recordings. In addition, we stated that the epileptic focus was always far away from the neocortical tissue samples. Furthermore, we added a small paragraph that functional studies using the same methodology have shown that neocortical access tissue samples taken from epilepsy surgery do not differ in electrophysiological properties and synaptic physiology when compared with acute slice preparations in experimental animals and we quoted the relevant papers.

We hope that the reviewer is now convinced that our tissue samples can be regarded as non-affected.

(2) About the inhibitory/excitatory synapses.

Since our focus was on excitatory synaptic boutons as already stated in the title we have not analyzed inhibitory SBs. Now, I do understand that only excitatory synapses were investigated. Although it was written in the title, I did not realized, since all over the manuscript the Authors were writing synapses, and were distinguishing between inhibitory and excitatory synapses in the text and showing numerous excitatory and inhibitory synapses on Figure 2 and discussing inhibitory interneurons in the Discussion as well. Maybe this was the reason why two reviewers out of the three (including myself) thought you investigated both types of synapses but did not differentiated between them. So, please, emphasize in the Abstract (line 40), Introduction (for ex. line 92-97) and the Discussion (line 369) that only excitatory synaptic boutons were investigated.

As this paper investigated only excitatory synaptic boutons, I think it is irrelevant to write such a long section in the Discussion about inhibitory interneurons and their functions in the L1 of the human temporal lobe neocortex. Same applies to the schematic drawing of the possible wiring of L1 (Figure 7). As no inhibitory interneurons were examined, neither the connection of the different excitatory cells, only the morphology of single synaptic boutons without any reference on their origin, I think this figure does not illustrate the work done in this paper. This could be a figure of a review paper about the human L1, but is inappropriate in this study.

We followed the reviewer’s suggestion and pointed out explicitly that we only investigated excitatory synaptic boutons. We also changed the Discussion and focused more on circuitry in L1 and the role of CR-cells.

(3) Perforated synapses

The findings of the Geinismann group suggesting that perforated synapses are more efficient than non-perforated ones is nowadays very controversially discussed” I did not ask the Authors to say that perforated synapses are more efficient. However, based on the literature (for ex. Harris et al, 1992; Carlin and Siekievitz, 1982; Nieto-Sampedro et al., 1982) the presence of perforated synapses is indeed a good sign of synapse division/formation - which in turn might be coupled to synaptic plasticity (Geinisman et al, 1993), increased synaptic activity (Vrensen and Cardozo, 1981), LTP (Geinisman et al, 1991, Harris et al, 2003), pathological axonal sprouting (Frotscher et al, 2006), etc. I think it is worth mentioning this at least in the Discussion.

We agree with the reviewer and added a small paragraph in the Results section about the two types of AZs in L1 of the human TLN. We pointed out that there are both types, macular non-perforated and perforated AZs, but the majority in all layers were of the non-perforated type. In the Discussion we added some paper pointing out the role of perforated synapses.

(4) Question about the vesicle pools

Results, Line 271: Still not understandable, why the RRP was defined as {less than or equal to}10 nm and {less than or equal to}20nm. Why did you use two categories? One would be sufficient (for example {less than or equal to}20nm). Or the vesicles between 10 and 20nm were considered to be part of RRP? In this case there is a typo, it should be {greater than or equal to}10 nm and {less than or equal to}20nm.

The answer of the Authors was to my question raised: We decided that also those very close within 10 and 20 nm away from the PreAZ, which is less than a SV diameter may also contribute to the RRP since it was shown that SVs are quite mobile.

This does not clarify why did you use two categories. Furthermore, I did not receive answer (such as Referee #2) for my question on how could you have 3x as many docked vesicles than vesicles {less than or equal to}10nm. The category {less than or equal to}10nm should also contain the docked vesicles. Or if this is not the case, please, clarify better what were your categories.

We thank the reviewer for pointing out that mentioning two distance criteria (p10 and p20) to define one physiological entity (RRP) is somewhat confusing and we acknowledge that the initial response to the reviewers falls short of explaining this choice. This is indeed only understandable in the context of the original paper by Sätzler et al. 2002, where these criteria were first introduced. We therefore referenced this publication more prominently in the paragraph in question.

So to explain this, we first would like to clarify the definition of the two RRP classification criteria used (p10 and p20), which has caused some confusion amongst the reviewers as to which vesicles where included or not:

- p10 criterion: p£10 nm (SVs have a minimum distance less than or equal to 10 nm from the PreAZ), including ‘docked’ vesicles which have a distance of zero or less (p0)

- p20 criterion: p£20 nm (SVs have a minimum distance less than or equal to 20 nm from the PreAZ), including vesicles of the p10 criterion.

As mentioned, these criteria were introduced first in Sätzler et al. 2002 looking at the Calyx of Held synapse. In that paper, we tried to establish a morphological correlate to existing physiological measurements, which included the RRP. As there is no known marker that would allow to discriminate between vesicles that contribute to the RRP anatomically, we looked at existing physiological experiments such as Schneggenburger et al. 1999; Wu and Borst 1999; Sun and Wu 2001 and compared their total numbers to our measurements. As the number of docked vesicles (p0, see above) was on the lower side of these physiological estimates, we also looked at vesicles close to the AZ, which we think could be recruited within a short time (£ 10 msec). Comparing with existing literature, we found that at p20 we get pool sizes comparable to midrange estimates of reported RRP sizes. In order to account for the variability of the observed physiological pool sizes, we reported all three measurements (p0, p10, p20) not only in the original Calyx of Held, but in all subsequent studies of different CNS synapses of our group since then.

As it remains uncertain if such correlate indeed exists, we therefore followed the suggestion to rephrase RRP and RP to putative RRP and putative RP (see also Rollenhagen et al. 2007). We thank both reviewers for pointing out this omission.

Concerning the difference between ‘docked’ vesicles and vesicles within the p10 perimeter criterion. First of all, the reviewer is right in saying that the category p10 ({less than or equal to}10nm) should also contain the docked vesicles (see above). The fact to have 3x as many ‘docked’ vesicles in our TEM tomography than in the p10 distance analysis could be partly explained, on the one hand, by a very high variability between patients (as expressed by the high SD, table 1) and, on the other hand, by a high intraindividual synaptic bouton variability. In both sublayers, there is a huge difference in the number of vesicles within the p10 criterion of individual synaptic boutons ranging from 0 to ~40 with a mean value of ~1 to ~4 (calculated per patient), the upper level being close to the values calculated with TEM tomography for the ‘docked’ vesicles.

(5) Astrocytic coverage

On Fig. 6 data are presented on the astrocytic coverage derived from L1 and L4. In my previous review I asked to include this in the text of the Results as well, but I still do not see it. It is also lacking from the Results how many samples from which layer were investigated in this analysis. Only percentages are given, and only for L1 (but how many patients, L1a and/or L1b and/or L4 is not provided). In contrast, Figure 6 and Supplementary Table 2 (patient table) contains the information that this analysis has been made in L4 as well. Please, include this information in the text as well (around lines 348-360).

In our previous revised version, we had included the values shown in Fig. 6 for both L1 and L4 in the Results section (L4: lines 352 – 355: ‘The findings in L1…’). However, we agree with the reviewer and have now also added the number of patients and synapses investigated (now lines 359 – 365).

About how to determine glial elements. I cannot agree with the Authors that glial elements can be determined with high certainty based only on the anatomical features of the profiles seen in the EM. “With 25 years of experience in (serial) EM work" I would say, that glial elements can be very similar to spine necks and axonal profiles.

All in all, if similar methods were used to determine the glial coverage in the different layers of the human neocortex, than it can be compared (I guess this is the case). However, I would say in the text that proper determination would need immunostaining and a new analysis. This only gives an estimation with the possibility of a certain degree of error.

We do not entirely agree with the reviewer on this point. As stated in the text, there are structural criteria to identify astrocytic elements (see citations quoted). These golden standard criteria are commonly used also by other well-known groups (DeFelipe and co-workers, Francisco Clasca and co-workers; Michael Frotscher the late and co-workers etc.). However, in a past paper about astrocytic coverage of synaptic complexes in L5 of the human TLN, immunohistochemistry against glutamine synthetase, a key enzyme in astrocytes, was carried out to describe the coverage. This experiment supports our findings in the other cortical layers of the human TLN. As the reviewer might know, immunohistochemistry always led to a reduction in ultrastructural preservation, so we decided not to use immunohistochemistry for the further publications of the other cortical layers. We added a short notice on this in the Material and Methods section.

(6) Large interindividual differences in the synapse density should be discussed in the Discussion.

As suggested by the reviewer we have included a sentence in the Discussion that interindividual differences can be either related to differences in age, gender and the use of different methodology as suggested by DeFelipe and co-workers (1999)

Reviewer #2 (Public review):

Summary:

The study of Rollenhagen et al examines the ultrastructural features of Layer 1 of human temporal cortex. The tissue was derived from drug-resistant epileptic patients undergoing surgery, and was selected as further from the epilepsy focus, and as such considered to be non-epileptic. The analyses has included 4 patients with different age, sex, medication and onset of epilepsy. The MS is a follow-on study with 3 previous publications from the same authors on different layers of the temporal cortex:

Layer 4 - Yakoubi et al 2019 eLife

Layer 5 - Yakoubi et al 2019 Cerebral Cortex,

Layer 6 - Schmuhl-Giesen et al 2022 Cerebral Cortex

They find, the L1 synaptic boutons mainly have single active zone a very large pool of synaptic vesicles and are mostly devoid of astrocytic coverage.

Strengths:

The MS is well written easy to read. Result section gives a detailed set of figures showing many morphological parameters of synaptic boutons and surrounding glial elements. The authors provide comparative data of all the layers examined by them so far in the Discussion. Given that anatomical data in human brain are still very limited, the current MS has substantial relevance.

The work appears to be generally well done, the EM and EM tomography images are of very good quality. The analyses is clear and precise.

Weaknesses:

The authors made all the corrections required, answered most of my concerns, included additional data sets, and clarified statements where needed.

My remaining points are:

Synaptic vesicle diameter (that has been established to be ~40nm independent of species) can properly be measured with EM tomography only, as it provides the possibility to find the largest diameter of every given vesicle. Measuring it in 50 nm thick sections result in underestimation (just like here the values are ~25 nm) as the measured diameter will be smaller than the true diameter if the vesicle is not cut in the middle, (which is the least probable scenario). The authors have the EM tomography data set for measuring the vesicle diameter properly.

We thank the reviewer for the helpful comments. We followed the recommendation to measure the vesicle diameter using our TEM tomography tilt series, but came to similar results concerning this synaptic parameter. As stated in our Material and Methods section, we only counted (measured) clear ring-link structures according to a paper by Abercrombie (1963). Since our results are similar for both methods, we do believe that our measurements are correct. Even random single measurements on the original 3D tilt-series yielded comparable results (Lübke and co-workers, personal observation). Furthermore, our results are within ranges, although with high variability, also described by other groups (see discussion lines 436 - 449). We therefore hope that the reviewer will now accept our measurements.

It is a bit misleading to call vesicle populations at certain arbitrary distances from the presynaptic active zone as readily releasable pool, recycling pool and resting pool, as these are functional categories, and cannot directly be translated to vesicles at certain distances. Even it is debated whether the morphologically docked vesicles are the ones, that are readily releasable, as further molecular steps, such as proper priming is also a prerequisite for release.

It would help to call these pools as "putative" correlates of the morphological categories.

We followed the suggestion by the reviewer and renamed our vesicle pools as putative RRP, putative RP and putative resting pools.

Reviewer #3 (Public review):

Summary:

Rollenhagen at al. offer a detailed description of layer 1 of the human neocortex. They use electron microscopy to assess the morphological parameters of presynaptic terminals, active zones, vesicle density/distribution, mitochondrial morphology and astrocytic coverage. The data is collected from tissue from four patients undergoing epilepsy surgery. As the epileptic focus was localized in all patients to the hippocampus, the tissue examined in this manuscript is considered non-epileptic (access) tissue.

Strengths:

The quality of the electron microscopic images is very high, and the data is analyzed carefully. Data from human tissue is always precious and the authors here provide a detailed analysis using adequate approaches, and the data is clearly presented.

Weaknesses:

The text connects functional and morphological characteristics in a very direct way. For example, connecting plasticity to any measurement the authors present would be rather difficult without any additional functional experiments. References to various vesicle pools based on the location of the vesicles is also more complex than it is suggested in the manuscript. The text should better reflect the limitations of the conclusions that can be drawn from the authors' data.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Astrocytic coverage

On Fig. 6 data are presented on the astrocytic coverage derived from L1 and L4. In my previous review I asked to include this in the text of the Results as well, but I still do not see it. It is also lacking from the Results how many samples from which layer were investigated in this analysis. Only percentages are given, and only for L1 (but how many patients, L1a and/or L1b and/or L4 is not provided). In contrast, Figure 6 and Supplementary Table 2 (patient table) contains the information that this analysis has been made in L4 as well. Please, include this information in the text as well (around lines 348-360).

See above.

About how to determine glial elements. I cannot agree with the Authors that glial elements can be determined with high certainty based only on the anatomical features of the profiles seen in the EM. “With 25 years of experience in (serial) EM work" I would say, that glial elements can be very similar to spine necks and axonal profiles. Please, see the photos below, out of the 16 circled profiles (2nd picture, very similar to each other) only 3 belong to an astroglial cell (last picture, purple profiles-purple cell), 10 are spines/spine necks/small caliber dendrites of pyramidal cells, 3 are axonal profiles (last but one picture, blue profiles, marked with arrows on the right side). If you follow in your serial sections those elements which you think are glial processes and indeed they are attached to a confidently identifiable glial cell, I agree, it is a glial process. But identifying small, almost empty profiles without any specific staining, from one single EM section, as glial process is very uncertain. Please, check the database of the Allen Institute made from the V1 visual cortex of a mouse. It is a large series of EM sections where they reconstructed thousands of neurons, astroglial and microglial cells. It is possible to double click on the EM picture on a profile and it will show the cell to which that profile belongs. https://portal.brain-map.org/connectivity/ultrastructural-connectomics Pictures included here: https://elife-rp.msubmit.net/eliferp_files/2024/11/25/00132644/02/132644_2_attach_21_29456_convrt.pdf

All in all, if similar methods were used to determine the glial coverage in the different layers of the human neocortex, than it can be compared (I guess this is the case). However, I would say in the text that proper determination would need immunostaining and a new analysis. This only gives an estimation with the possibility of a certain degree of error.

As stated above, we carried out glutamine synthetase immunohistochemistry in L5 of the human TLN and came to the same results. However, we added a sentence on this in the chapter on astrocytic coverage in the Material and Methods section. Additionally, we modified this chapter according to the reviewer’s suggestion.

Minor comments

Introduction: Last sentence is not understandable (lines 101-103), please rephrase. (contribute to understand or contribute in understanding or contribute to the understanding of..., but definitely not contribute to understanding). The authors should check and review extensively for improvements to the use of English, or use a program such as Grammarly.

Results: Grammar (line 107): L1 in the adult mammalian neocortex represents a relatively...

Line 173: “Some SBs in both sublaminae were seen to establish either two or three SBs on the same spine, spines 173 of other origin or dendritic shafts." - Some SBs established two or three SBs? I would write Some SBs established two or three synapses on...

Line 243: “The synaptic cleft size were slightly, but non-significantly different"

Line 260: “DCVs play an important role in endo- and exocytosis, the build-up of PreAZs by releasing Piccolo and Bassoon (Schoch and Gundelfinger 2006; Murkherjee et al. 2010)," - please, correct this.

We have done corrections as suggested by the reviewer.

Line 374: No point at the end of the last phrase.

Discussion:

Lines 400-404: “The majority of SBs in L1 of the human TLN had a single at most three AZs that could be of the non perforated macular or perforated type comparable with results for other layers in the human TLN but by ~1.5-fold larger than in rodent and non-human primates." - What is comparable with the other layers, but different from animals? Please rephrase this sentence, it is not understandable. I already mentioned this sentence in my previous review, but nothing happened.

Lines 435-437: “Remarkably, the total pool sizes in the human TLN were significantly larger by more than 6-fold (~550 SVs/AZ), and ~4.7-fold (~750 SVs/AZ;) than those in L4 and L5 (Yakoubi et al. 2019a, b; see also Rollenhagen et al. 2018) in rats." Please rethink what you wished to say and compare to the sentence meaning. I think you wanted to compare human TLN L1 pool size to L4 and L5 in the human TLN (Yakoubi 2019a and b) and to rat (Rollenhagen 2018). Instead, you compared all layers of the human TLN to L4 and L5 in rats (with partly wrong references). Please rephrase this. Lines 483-484: “Astrocytes serve as both a physical barrier to glutamate diffusion and as mediate neurotransmitter uptake via transporters".

This sentence is grammatically incorrect, please rephrase.

We corrected the sentences as suggested by the reviewer.

Methods:

In the text, there are only 4 patients (lines 603-604), but in the supplementary table there are 9 patients (5 new included for L4 astrocytic coverage). Please, correct it in the text.

Lines 608-609: “neocortical access tissue samples were resected to control the seizures for histological inspection by neuropathologists." - What is the meaning of this? Please, rephrase.

We thank the reviewer for the comment and included the 5 patients used for L4 to the Material and Methods section, as well as in the Results section.

The reviewer is right, and we rephrased and corrected the sentence concerning the inspection by neuropathologists.

Figures

Figures 5B: The legend says “SB (sb) synapsing on a stubby spine (sp) with a prominent spine apparatus (framed area) and a thick dendritic segment (de) in L1b" - In my opinion this is not one synaptic bouton, but two. Clearly visible membranes separate them, close to the spine.

Supplemental Table 2 (patient table). If there is no information about Hu_04 patient's epilepsy, please write N/A (=non available) instead of - (which means it does not exist).

The reviewer is right, and we corrected the figure and the legend, as well as the table accordingly.

Reviewer #2 (Recommendations for the authors):

The authors addressed almost all of my concern, only this one remained:

If there is, however, relevant literature on "methods based on EM tomography" and "stereological methods to estimate both types of error" (over- and underestimates) that we are missing out on, we would appreciate the reviewer providing us with the corresponding references so that we can include such calculations in our paper.

There is a very detailed new study on calculating correction for TEM 2D 3D, Rothman et al 2023 PLOS One. That addresses most of these issues.

We thank the reviewer for drawing our attention to the publication by Rothman et al. 2023, which is a very detailed and comprehensive study looking at accurately estimating distributions of 3D size and densities of particles from 2D measurements using – amongst others – ET and TEM images as well as synaptic vesicles for validating their method. However, we do not see how this would be relevant to the reported mean diameters and their corresponding variances. And even if we would have reported on vesicle size/diameter distributions (referred to as G(d) in Rothmann et al. 2023), the authors themselves state that “… the results from our ET and TEM image analysis highlight the difficulty in computing a complete G(d) of MFT vesicles due to their small size…

This section really made me think about how LGBTQ topics are often framed around danger, risk, or trauma. While those realities are important, it's limiting if that’s all we focus on. I like how the reading reminds us that LGBTQ communities also have resilience, joy, and rich cultural histories. Including those aspects in education helps move the conversation from tolerance to genuine respect.

This reaffirms that sexuality and gender are far more slippery and complex than categories can imply. It reinforces that even when schools try to place "normal" expectations upon them, people's experiences of identity cannot be constrained within firm boxes. By inquiring how sexuality intersects with race, class, and gender, the book highlights that safe schools for LGBTQ students require responding to broader systems of oppression rather than discrete cases of bullying and discrimination. It challenges us to examine more thoroughly how all students, regardless of identity, do well when schools push back on narrow definitions of what is "normal."

Author response:

The following is the authors’ response to the original reviews.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Using a cross-modal sensory selection task in head-fixed mice, the authors attempted to characterize how different rules reconfigured representations of sensory stimuli and behavioral reports in sensory (S1, S2) and premotor cortical areas (medial motor cortex or MM, and ALM). They used silicon probe recordings during behavior, a combination of single-cell and population-level analyses of neural data, and optogenetic inhibition during the task.

Strengths:

A major strength of the manuscript was the clarity of the writing and motivation for experiments and analyses. The behavioral paradigm is somewhat simple but well-designed and wellcontrolled. The neural analyses were sophisticated, clearly presented, and generally supported the authors' interpretations. The statistics are clearly reported and easy to interpret. In general, my view is that the authors achieved their aims. They found that different rules affected preparatory activity in premotor areas, but not sensory areas, consistent with dynamical systems perspectives in the field that hold that initial conditions are important for determining trial-based dynamics.

Weaknesses:

The manuscript was generally strong. The main weakness in my view was in interpreting the optogenetic results. While the simplicity of the task was helpful for analyzing the neural data, I think it limited the informativeness of the perturbation experiments. The behavioral read-out was low dimensional -a change in hit rate or false alarm rate- but it was unclear what perceptual or cognitive process was disrupted that led to changes in these read-outs. This is a challenge for the field, and not just this paper, but was the main weakness in my view. I have some minor technical comments in the recommendations for authors that might address other minor weaknesses.

I think this is a well-performed, well-written, and interesting study that shows differences in rule representations in sensory and premotor areas and finds that rules reconfigure preparatory activity in the motor cortex to support flexible behavior.

Reviewer #2 (Public Review):

Summary:

Chang et al. investigate neuronal activity firing patterns across various cortical regions in an interesting context-dependent tactile vs visual detection task, developed previously by the authors (Chevee et al., 2021; doi: 10.1016/j.neuron.2021.11.013). The authors report the important involvement of a medial frontal cortical region (MM, probably a similar location to wM2 as described in Esmaeili et al., 2021 & 2022; doi: 10.1016/j.neuron.2021.05.005; doi: 10.1371/journal.pbio.3001667) in mice for determining task rules.

Strengths:

The experiments appear to have been well carried out and the data well analysed. The manuscript clearly describes the motivation for the analyses and reaches clear and well-justified conclusions. I find the manuscript interesting and exciting!

Weaknesses:

I did not find any major weaknesses.

Reviewer #3 (Public Review):

This study examines context-dependent stimulus selection by recording neural activity from several sensory and motor cortical areas along a sensorimotor pathway, including S1, S2, MM, and ALM. Mice are trained to either withhold licking or perform directional licking in response to visual or tactile stimulus. Depending on the task rule, the mice have to respond to one stimulus modality while ignoring the other. Neural activity to the same tactile stimulus is modulated by task in all the areas recorded, with significant activity changes in a subset of neurons and population activity occupying distinct activity subspaces. Recordings further reveal a contextual signal in the pre-stimulus baseline activity that differentiates task context. This signal is correlated with subsequent task modulation of stimulus activity. Comparison across brain areas shows that this contextual signal is stronger in frontal cortical regions than in sensory regions. Analyses link this signal to behavior by showing that it tracks the behavioral performance switch during task rule transitions. Silencing activity in frontal cortical regions during the baseline period impairs behavioral performance.

Overall, this is a superb study with solid results and thorough controls. The results are relevant for context-specific neural computation and provide a neural substrate that will surely inspire follow-up mechanistic investigations. We only have a couple of suggestions to help the authors further improve the paper.

(1) We have a comment regarding the calculation of the choice CD in Fig S3. The text on page 7 concludes that "Choice coding dimensions change with task rule". However, the motor choice response is different across blocks, i.e. lick right vs. no lick for one task and lick left vs. no lick for the other task. Therefore, the differences in the choice CD may be simply due to the motor response being different across the tasks and not due to the task rule per se. The authors may consider adding this caveat in their interpretation. This should not affect their main conclusion.

We thank the Reviewer for the suggestion. We have discussed this caveat and performed a new analysis to calculate the choice coding dimensions using right-lick and left-lick trials (Fig. S3h) on page 8. 

“Choice coding dimensions were obtained from left-lick and no-lick trials in respond-to-touch blocks and right-lick and no-lick trials in respond-to-light blocks. Because the required lick directions differed between the block types, the difference in choice CDs across task rules (Fig. S4f) could have been affected by the different motor responses. To rule out this possibility, we did a new version of this analysis using right-lick and left-lick trials to calculate the choice coding dimensions for both task rules. We found that the orientation of the choice coding dimension in a respond-to-touch block was still not aligned well with that in a respond-to-light block (Fig. S4h;  magnitude of dot product between the respond-to-touch choice CD and the respond-to-light choice CD, mean ± 95% CI for true vs shuffled data: S1: 0.39 ± [0.23, 0.55] vs 0.2 ± [0.1, 0.31], 10 sessions; S2: 0.32 ± [0.18, 0.46] vs 0.2 ± [0.11, 0.3], 8 sessions; MM: 0.35 ± [0.21, 0.48] vs 0.18 ± [0.11, 0.26], 9 sessions; ALM: 0.28 ± [0.17, 0.39] vs 0.21 ± [0.12, 0.31], 13 sessions).”

We also have included the caveats for using right-lick and left-lick trials to calculate choice coding dimensions on page 13.

“However, we also calculated choice coding dimensions using only right- and left-lick trials. In S1, S2, MM and ALM, the choice CDs calculated this way were also not aligned well across task rules (Fig. S4h), consistent with the results calculated from lick and no-lick trials (Fig. S4f). Data were limited for this analysis, however, because mice rarely licked to the unrewarded water port (# of licksunrewarded port  / # of lickstotal , respond-to-touch: 0.13, respond-to-light: 0.11). These trials usually came from rule transitions (Fig. 5a) and, in some cases, were potentially caused by exploratory behaviors. These factors could affect choice CDs.”

(2) We have a couple of questions about the effect size on single neurons vs. population dynamics. From Fig 1, about 20% of neurons in frontal cortical regions show task rule modulation in their stimulus activity. This seems like a small effect in terms of population dynamics. There is somewhat of a disconnect from Figs 4 and S3 (for stimulus CD), which show remarkably low subspace overlap in population activity across tasks. Can the authors help bridge this disconnect? Is this because the neurons showing a difference in Fig 1 are disproportionally stimulus selective neurons?

We thank the Reviewer for the insightful comment and agree that it is important to link the single-unit and population results. We have addressed these questions by (1) improving our analysis of task modulation of single neurons  (tHit-tCR selectivity) and (2) examining the relationship between tHit-tCR selective neurons and tHit-tCR subspace overlaps.  

Previously, we averaged the AUC values of time bins within the stimulus window (0-150 ms, 10 ms bins). If the 95% CI on this averaged AUC value did not include 0.5, this unit was considered to show significant selectivity. This approach was highly conservative and may underestimate the percentage of units showing significant selectivity, particularly any units showing transient selectivity. In the revised manuscript, we now define a unit as showing significant tHit-tCR selectivity when three consecutive time bins (>30 ms, 10ms bins) of AUC values were significant. Using this new criterion, the percentage of tHittCR selective neurons increased compared with the previous analysis. We have updated Figure 1h and the results on page 4:

“We found that 18-33% of neurons in these cortical areas had area under the receiver-operating curve (AUC) values significantly different from 0.5, and therefore discriminated between tHit and tCR trials (Fig. 1h; S1: 28.8%, 177 neurons; S2: 17.9%, 162 neurons; MM: 32.9%, 140 neurons; ALM: 23.4%, 256 neurons; criterion to be considered significant: Bonferroni corrected 95% CI on AUC did not include 0.5 for at least 3 consecutive 10-ms time bins).”

Next, we have checked how tHit-tCR selective neurons were distributed across sessions. We found that the percentage of tHit-tCR selective neurons in each session varied (S1: 9-46%, S2: 0-36%, MM:25-55%, ALM:0-50%). We examined the relationship between the numbers of tHit-tCR selective neurons and tHit-tCR subspace overlaps. Sessions with more neurons showing task rule modulation tended to show lower subspace overlap, but this correlation was modest and only marginally significant (r= -0.32, p= 0.08, Pearson correlation, n= 31 sessions). While we report the percentage of neurons showing significant selectivity as a simple way to summarize single-neuron effects, this does neglect the magnitude of task rule modulation of individual neurons, which may also be relevant. 

In summary, the apparent disconnect between the effect sizes of task modulation of single neurons and of population dynamics could be explained by (1) the percentages of tHit-tCR selective neurons were underestimated in our old analysis, (2) tHit-tCR selective neurons were not uniformly distributed among sessions, and (3) the percentages of tHit-tCR selective neurons were weakly correlated with tHit-tCR subspace overlaps. 

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

For the analysis of choice coding dimensions, it seems that the authors are somewhat data limited in that they cannot compare lick-right/lick-left within a block. So instead, they compare lick/no lick trials. But given that the mice are unable to initiate trials, the interpretation of the no lick trials is a bit complicated. It is not clear that the no lick trials reflect a perceptual judgment about the stimulus (i.e., a choice), or that the mice are just zoning out and not paying attention. If it's the latter case, what the authors are calling choice coding is more of an attentional or task engagement signal, which may still be interesting, but has a somewhat different interpretation than a choice coding dimension. It might be worth clarifying this point somewhere, or if I'm totally off-base, then being more clear about why lick/no lick is more consistent with choice than task engagement.

We thank the Reviewer for raising this point. We have added a new paragraph on page 13 to clarify why we used lick/no-lick trials to calculate choice coding dimensions, and we now discuss the caveat regarding task engagement.  

“No-lick trials included misses, which could be caused by mice not being engaged in the task. While the majority of no-lick trials were correct rejections (respond-to-touch: 75%; respond-to-light: 76%), we treated no-licks as one of the available choices in our task and included them to calculate choice coding dimensions (Fig. S4c,d,f). To ensure stable and balanced task engagement across task rules, we removed the last 20 trials of each session and used stimulus parameters that achieved similar behavioral performance for both task rules (Fig. 1d; ~75% correct for both rules).”

In addition, to address a point made by Reviewer 3 as well as this point, we performed a new analysis to calculate choice coding dimensions using right-lick vs left-lick trials. We report this new analysis on page 8:

“Choice coding dimensions were obtained from left-lick and no-lick trials in respond-to-touch blocks and right-lick and no-lick trials in respond-to-light blocks. Because the required lick directions differed between the block types, the difference in choice CDs across task rules (Fig. S4f) could have been affected by the different motor responses. To rule out this possibility, we did a new version of this analysis using right-lick and left-lick trials to calculate the choice coding dimensions for both task rules. We found that the orientation of the choice coding dimension in a respond-to-touch block was still not aligned well with that in a respond-to-light block (Fig. S4h;  magnitude of dot product between the respond-to-touch choice CD and the respond-to-light choice CD, mean ± 95% CI for true vs shuffled data: S1: 0.39 ± [0.23, 0.55] vs 0.2 ± [0.1, 0.31], 10 sessions; S2: 0.32 ± [0.18, 0.46] vs 0.2 ± [0.11, 0.3], 8 sessions; MM: 0.35 ± [0.21, 0.48] vs 0.18 ± [0.11, 0.26], 9 sessions; ALM: 0.28 ± [0.17, 0.39] vs 0.21 ± [0.12, 0.31], 13 sessions).” 

We added discussion of the limitations of this new analysis on page 13:

“However, we also calculated choice coding dimensions using only right- and left-lick trials. In S1, S2, MM and ALM, the choice CDs calculated this way were also not aligned well across task rules (Fig. S4h), consistent with the results calculated from lick and no-lick trials (Fig. S4f). Data were limited for this analysis, however, because mice rarely licked to the unrewarded water port (# of licksunrewarded port  / # of lickstotal , respond-to-touch: 0.13, respond-to-light: 0.11). These trials usually came from rule transitions (Fig. 5a) and, in some cases, were potentially caused by exploratory behaviors. These factors could affect choice CDs.”

The authors find that the stimulus coding direction in most areas (S1, S2, and MM) was significantly aligned between the block types. How do the authors interpret that finding? That there is no major change in stimulus coding dimension, despite the change in subspace? I think I'm missing the big picture interpretation of this result.

That there is no significant change in stimulus coding dimensions but a change in subspace suggests that the subspace change largely reflects a change in the choice coding dimensions.

As I mentioned in the public review, I thought there was a weakness with interpretation of the optogenetic experiments, which the authors generally interpret as reflecting rule sensitivity. However, given that they are inhibiting premotor areas including ALM, one might imagine that there might also be an effect on lick production or kinematics. To rule this out, the authors compare the change in lick rate relative to licks during the ITI. What is the ITI lick rate? I assume pretty low, once the animal is welltrained, in which case there may be a floor effect that could obscure meaningful effects on lick production. In addition, based on the reported CI on delta p(lick), it looks like MM and AM did suppress lick rate. I think in the future, a task with richer behavioral read-outs (or including other measurements of behavior like video), or perhaps something like a psychological process model with parameters that reflect different perceptual or cognitive processes could help resolve the effects of perturbations more precisely.

Eighteen and ten percent of trials had at least one lick in the ITI in respond-to-touch and  respond-tolight blocks, respectively. These relatively low rates of ITI licking could indeed make an effect of optogenetics on lick production harder to observe. We agree that future work would benefit from more complex tasks and measurements, and have added the following to make this point (page 14):

“To more precisely dissect the effects of perturbations on different cognitive processes in rule-dependent sensory detection, more complex behavioral tasks and richer behavioral measurements are needed in the future.”

Reviewer #2 (Recommendations For The Authors):

I have the following minor suggestions that the authors might consider in revising this already excellent manuscript :

(1) In addition to showing normalised z-score firing rates (e.g. Fig 1g), I think it is important to show the grand-average mean firing rates in Hz.

We thank the Reviewer for the suggestion and have added the grand-average mean firing rates as a new supplementary figure (Fig. S2a). To provide more details about the firing rates of individual neurons, we have also added to this new figure the distribution of peak responses during the tactile stimulus period (Fig. S2b).

(2) I think the authors could report more quantitative data in the main text. As a very basic example, I could not easily find how many neurons, sessions, and mice were used in various analyses.

We have added relevant numbers at various points throughout the Results, including within the following examples:

Page 3: “To examine how the task rules influenced the sensorimotor transformation occurring in the tactile processing stream, we performed single-unit recordings from sensory and motor cortical areas including S1, S2, MM and ALM (Fig. 1e-g, Fig. S1a-h, and Fig. S2a; S1: 6 mice, 10 sessions, 177 neurons, S2: 5 mice, 8 sessions, 162 neurons, MM: 7 mice, 9 sessions, 140 neurons, ALM: 8 mice, 13 sessions, 256 neurons).”

Page 5: “As expected, single-unit activity before stimulus onset did not discriminate between tactile and visual trials (Fig. 2d; S1: 0%, 177 neurons; S2: 0%, 162 neurons; MM: 0%, 140 neurons; ALM: 0.8%, 256 neurons). After stimulus onset, more than 35% of neurons in the sensory cortical areas and approximately 15% of neurons in the motor cortical areas showed significant stimulus discriminability (Fig. 2e; S1: 37.3%, 177 neurons; S2: 35.2%, 162 neurons; MM: 15%, 140 neurons; ALM: 14.1%, 256 neurons).”

Page 6: “Support vector machine (SVM) and Random Forest classifiers showed similar decoding abilities

(Fig. S3a,b; medians of classification accuracy [true vs shuffled]; SVM: S1 [0.6 vs 0.53], 10 sessions, S2

[0.61 vs 0.51], 8 sessions, MM [0.71 vs 0.51], 9 sessions, ALM [0.65 vs 0.52], 13 sessions; Random

Forests: S1 [0.59 vs 0.52], 10 sessions, S2 [0.6 vs 0.52], 8 sessions, MM [0.65 vs 0.49], 9 sessions, ALM [0.7 vs 0.5], 13 sessions).”

Page 6: “To assess this for the four cortical areas, we quantified how the tHit and tCR trajectories diverged from each other by calculating the Euclidean distance between matching time points for all possible pairs of tHit and tCR trajectories for a given session and then averaging these for the session (Fig. 4a,b; S1: 10 sessions, S2: 8 sessions, MM: 9 sessions, ALM: 13 sessions, individual sessions in gray and averages across sessions in black; window of analysis: -100 to 150 ms relative to stimulus onset; 10 ms bins; using the top 3 PCs; Methods).” 

Page 8: “In contrast, we found that S1, S2 and MM had stimulus CDs that were significantly aligned between the two block types (Fig. S4e; magnitude of dot product between the respond-to-touch stimulus CDs and the respond-to-light stimulus CDs, mean ± 95% CI for true vs shuffled data: S1: 0.5 ± [0.34, 0.66] vs 0.21 ± [0.12, 0.34], 10 sessions; S2: 0.62 ± [0.43, 0.78] vs 0.22 ± [0.13, 0.31], 8 sessions; MM: 0.48 ± [0.38, 0.59] vs 0.24 ± [0.16, 0.33], 9 sessions; ALM: 0.33 ± [0.2, 0.47] vs 0.21 ± [0.13, 0.31], 13 sessions).”  Page 9: “For respond-to-touch to respond-to-light block transitions, the fractions of trials classified as respond-to-touch for MM and ALM decreased progressively over the course of the transition (Fig. 5d; rank correlation of the fractions calculated for each of the separate periods spanning the transition, Kendall’s tau, mean ± 95% CI: MM: -0.39 ± [-0.67, -0.11], 9 sessions, ALM: -0.29 ± [-0.54, -0.04], 13 sessions; criterion to be considered significant: 95% CI on Kendall’s tau did not include 0).

Page 11: “Lick probability was unaffected during S1, S2, MM and ALM experiments for both tasks, indicating that the behavioral effects were not due to an inability to lick (Fig. 6i, j; 95% CI on Δ lick probability for cross-modal selection task: S1/S2 [-0.18, 0.24], 4 mice, 10 sessions; MM [-0.31, 0.03], 4 mice, 11 sessions; ALM [-0.24, 0.16], 4 mice, 10 sessions; Δ lick probability for simple tactile detection task: S1/S2 [-0.13, 0.31], 3 mice, 3 sessions; MM [-0.06, 0.45], 3 mice, 5 sessions; ALM [-0.18, 0.34], 3 mice, 4 sessions).”

(3) Please include a clearer description of trial timing. Perhaps a schematic timeline of when stimuli are delivered and when licking would be rewarded. I may have missed it, but I did not find explicit mention of the timing of the reward window or if there was any delay period.

We have added the following (page 3): 

“For each trial, the stimulus duration was 0.15 s and an answer period extended from 0.1 to 2 s from stimulus onset.”

(4) Please include a clear description of statistical tests in each figure legend as needed (for example please check Fig 4e legend).

We have added details about statistical tests in the figure legends:

Fig. 2f: “Relationship between block-type discriminability before stimulus onset and tHit-tCR discriminability after stimulus onset for units showing significant block-type discriminability prior to the stimulus. Pearson correlation: S1: r = 0.69, p = 0.056, 8 neurons; S2: r = 0.91, p = 0.093, 4 neurons; MM: r = 0.93, p < 0.001, 30 neurons; ALM: r = 0.83, p < 0.001, 26 neurons.” 

Fig. 4e: “Subspace overlap for control tHit (gray) and tCR (purple) trials in the somatosensory and motor cortical areas. Each circle is a subspace overlap of a session. Paired t-test, tCR – control tHit: S1: -0.23, 8 sessions, p = 0.0016; S2: -0.23, 7 sessions, p = 0.0086; MM: -0.36, 5 sessions, p = <0.001; ALM: -0.35, 11 sessions, p < 0.001; significance: ** for p<0.01, *** for p<0.001.”  

Fig. 5d,e: “Fraction of trials classified as coming from a respond-to-touch block based on the pre-stimulus population state, for trials occurring in different periods (see c) relative to respond-to-touch → respondto-light transitions. For MM (top row) and ALM (bottom row), progressively fewer trials were classified as coming from the respond-to-touch block as analysis windows shifted later relative to the rule transition. Kendall’s tau (rank correlation): MM: -0.39, 9 sessions; ALM: -0.29, 13 sessions. Left panels: individual sessions, right panels: mean ± 95% CI. Dash lines are chance levels (0.5). e, Same as d but for respond-to-light → respond-to-touch transitions. Kendall’s tau: MM: 0.37, 9 sessions; ALM: 0.27, 13 sessions.”

Fig. 6: “Error bars show bootstrap 95% CI. Criterion to be considered significant: 95% CI did not include 0.”

(5) P. 3 - "To examine how the task rules influenced the sensorimotor transformation occurring in the tactile processing stream, we performed single-unit recordings from sensory and motor cortical areas including S1, S2, MM, and ALM using 64-channel silicon probes (Fig. 1e-g and Fig. S1a-h)." Please specify if these areas were recorded simultaneously or not.

We have added “We recorded from one of these cortical areas per session, using 64-channel silicon probes.”  on page 3.  

(6) Figure 4b - Please describe what gray and black lines show.

The gray traces are the distance between tHit and tCR trajectories in individual sessions and the black traces are the averages across sessions in different cortical areas. We have added this information on page 6 and in the Figure 4b legend. 

Page 6: “To assess this for the four cortical areas, we quantified how the tHit and tCR trajectories diverged from each other by calculating the Euclidean distance between matching time points for all possible pairs of tHit and tCR trajectories for a given session and then averaging these for the session (Fig. 4a,b; S1: 10 sessions, S2: 8 sessions, MM: 9 sessions, ALM: 13 sessions, individual sessions in gray and averages across sessions in black; window of analysis: -100 to 150 ms relative to stimulus onset; 10 ms bins; using the top 3 PCs; Methods).

Fig. 4b: “Distance between tHit and tCR trajectories in S1, S2, MM and ALM. Gray traces show the time varying tHit-tCR distance in individual sessions and black traces are session-averaged tHit-tCR distance (S1:10 sessions; S2: 8 sessions; MM: 9 sessions; ALM: 13 sessions).”

(7) In addition to the analyses shown in Figure 5a, when investigating the timing of the rule switch, I think the authors should plot the left and right lick probabilities aligned to the timing of the rule switch time on a trial-by-trial basis averaged across mice.

We thank the Reviewer for suggesting this addition. We have added a new figure panel to show the probabilities of right- and left-licks during rule transitions (Fig. 5a).

Page 8: “The probabilities of right-licks and left-licks showed that the mice switched their motor responses during block transitions depending on task rules (Fig. 5a, mean ± 95% CI across 12 mice).” 

(8) P. 12 - "Moreover, in a separate study using the same task (Finkel et al., unpublished), high-speed video analysis demonstrated no significant differences in whisker motion between respond-to-touch and respond-to-light blocks in most (12 of 14) behavioral sessions.". Such behavioral data is important and ideally would be included in the current analysis. Was high-speed videography carried out during electrophysiology in the current study?

Finkel et al. has been accepted in principle for publication and will be available online shortly. Unfortunately we have not yet carried out simultaneous high-speed whisker video and electrophysiology in our cross-modal sensory selection task.

Reviewer #3 (Recommendations For The Authors):

(1) Minor point. For subspace overlap calculation of pre-stimulus activity in Fig 4e (light purple datapoints), please clarify whether the PCs for that condition were constructed in matched time windows. If the PCs are calculated from the stimulus period 0-150ms, the poor alignment could be due to mismatched time windows.

We thank the Reviewer for the comment and clarify our analysis here. We previously used timematched windows to calculate subspace overlaps. However, the pre-stimulus activity was much weaker than the activity during the stimulus period, so the subspaces of reference tHit were subject to noise and we were not able to obtain reliable PCs. This caused the subspace overlap values between the reference tHit and control tHit to be low and variable (mean ± SD, S1:  0.46± 0.26, n = 8 sessions, S2: 0.46± 0.18, n = 7 sessions, MM: 0.44± 0.16, n = 5 sessions, ALM: 0.38± 0.22, n = 11 sessions).  Therefore, we used the tHit activity during the stimulus window to obtain PCs and projected pre-stimulus and stimulus activity in tCR trials onto these PCs. We have now added a more detailed description of this analysis in the Methods (page 32). 

“To calculate the separation of subspaces prior to stimulus delivery, pre-stimulus activity in tCR trials (100 to 0 ms from stimulus onset) was projected to the PC space of the tHit reference group and the subspace overlap was calculated. In this analysis, we used tHit activity during stimulus delivery (0 to 150 ms from stimulus onset) to obtain reliable PCs.”   

We acknowledge this time alignment issue and have now removed the reported subspace overlap between tHit and tCR during the pre-stimulus period from Figure 4e (light purple). However, we think the correlation between pre- and post- stimulus-onset subspace overlaps should remain similar regardless of the time windows that we used for calculating the PCs. For the PCs calculated from the pre-stimulus period (-100 to 0 ms), the correlation coefficient was 0.55 (Pearson correlation, p <0.01, n = 31 sessions). For the PCs calculated from the stimulus period (0-150 ms), the correlation coefficient was 0.68 (Figure 4f, Pearson correlation, p <0.001, n = 31 sessions). Therefore, we keep Figure 4f.  

(2) Minor point. To help the readers follow the logic of the experiments, please explain why PPC and AMM were added in the later optogenetic experiment since these are not part of the electrophysiology experiment.

We have added the following rationale on page 9.

“We recorded from AMM in our cross-modal sensory selection task and observed visually-evoked activity (Fig. S1i-k), suggesting that AMM may play an important role in rule-dependent visual processing. PPC contributes to multisensory processing51–53 and sensory-motor integration50,54–58.  Therefore, we wanted to test the roles of these areas in our cross-modal sensory selection task.”

(3) Minor point. We are somewhat confused about the timing of some of the example neurons shown in figure S1. For example, many neurons show visually evoked signals only after stimulus offset, unlike tactile evoked signals (e.g. Fig S1b and f). In addition, the reaction time for visual stimulus is systematically slower than tactile stimuli for many example neurons (e.g. Fig S1b) but somehow not other neurons (e.g. Fig S1g). Are these observations correct?

These observations are all correct. We have a manuscript from a separate study using this same behavioral task (Finkel et al., accepted in principle) that examines and compares (1) the onsets of tactile- and visually-evoked activity and (2) the reaction times to tactile and visual stimuli. The reaction times to tactile stimuli were slightly but significantly shorter than the reaction times to visual stimuli (tactile vs visual, 397 ± 145 vs 521 ± 163 ms, median ± interquartile range [IQR], Tukey HSD test, p = 0.001, n =155 sessions). We examined how well activity of individual neurons in S1 could be used to discriminate the presence of the stimulus or the response of the mouse. For discriminability for the presence of the stimulus, S1 neurons could signal the presence of the tactile stimulus but not the visual stimulus. For discriminability for the response of the mouse, the onsets for significant discriminability occurred earlier for tactile compared with visual trials (two-sided Kolmogorov-Smirnov test, p = 1x10-16, n = 865 neurons with DP onset in tactile trials, n = 719 neurons with DP onset in visual trials).

Author Response

The following is the authors’ response to the original reviews.

eLife assessment

This important study combines a range of advanced ultrastructural imaging approaches to define the unusual endosomal system of African trypanosomes. Compelling images show that instead of a distinct set of compartments, the endosome of these protists comprises a continuous system of membranes with functionally distinct subdomains as defined by canonical markers of early, late and recycling endosomes. The findings suggest that the endocytic system of bloodstream stages has evolved to facilitate the extraordinarily high rates of membrane turnover needed to remove immune complexes and survive in the blood, which is of interest to anyone studying infectious diseases.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Bloodstream stages of the parasitic protist, Trypanosoma brucei, exhibit very high rates of constitutive endocytosis, which is needed to recycle the surface coat of Variant Surface Glycoproteins (VSGs) and remove surface immune complexes. While many studies have shown that the endo-lysosomal systems of T. brucei BF stages contain canonical domains, as defined by classical Rab markers, it has remained unclear whether these protists have evolved additional adaptations/mechanisms for sustaining these very high rates of membrane transport and protein sorting. The authors have addressed this question by reconstructing the 3D ultrastructure and functional domains of the T. brucei BF endosome membrane system using advanced electron tomography and super-resolution microscopy approaches. Their studies reveal that, unusually, the BF endosome network comprises a continuous system of cisternae and tubules that contain overlapping functional subdomains. It is proposed that a continuous membrane system allows higher rates of protein cargo segregation, sorting and recycling than can otherwise occur when transport between compartments is mediated by membrane vesicles or other fusion events.

Strengths:

The study is a technical tour-de-force using a combination of electron tomography, super-resolution/expansion microscopy, immune-EM of cryo-sections to define the 3D structures and connectivity of different endocytic compartments. The images are very clear and generally support the central conclusion that functionally distinct endocytic domains occur within a dynamic and continuous endosome network in BF stages.

Weaknesses:

The authors suggest that this dynamic endocytic network may also fulfil many of the functions of the Golgi TGN and that the latter may be absent in these stages. Although plausible, this comment needs further experimental support. For example, have the authors attempted to localize canonical makers of the TGN (e.g. GRIP proteins) in T. brucei BF and/or shown that exocytic carriers bud directly from the endosomes?

We agree with the criticism and have shortened the discussion accordingly and clearly marked it as speculation. However, we do not want to completely abandon our hypothesis.

The paragraph now reads:

Lines 740 – 751:

“Interestingly, we did not find any structural evidence of vesicular retrograde transport to the Golgi. Instead, the endosomal ‘highways’ extended throughout the posterior volume of the trypanosomes approaching the trans-Golgi interface. It is highly plausible that this region represents the convergence point where endocytic and biosynthetic membrane trafficking pathways merge. A comparable merging of endocytic and biosynthetic functions has been described for the TGN in plants. Different marker proteins for early and recycling endosomes were shown to be associated and/ or partially colocalized with the TGN suggesting its function in both secretory and endocytic pathways (reviewed in Minamino and Ueda, 2019). As we could not find structural evidence for the existence of a TGN we tentatively propose that trypanosomes may have shifted the central orchestrating function of the TGN as a sorting hub at the crossroads of biosynthetic and recycling pathways to the endosome. Although this is a speculative scenario, it is experimentally testable.”

Furthermore, we removed the lines 51 - 52, which included the suggestion of the TGN as a master regulator, from the abstract.

Reviewer #2 (Public Review):

The authors suggest that the African trypanosome endomembrane system has unusual organisation, in that the entire system is a single reticulated structure. It is not clear if this is thought to extend to the lysosome or MVB. There is also a suggestion that this unusual morphology serves as a trans-(post)Golgi network rather than the more canonical arrangement.

The work is based around very high-quality light and electron microscopy, as well as utilising several marker proteins, Rab5A, 11 and 7. These are deemed as markers for early endosomes, recycling endosomes and late or pre-lysosomes. The images are mostly of high quality but some inconsistencies in the interpretation, appearance of structures and some rather sweeping assumptions make this less easy to accept. Two perhaps major issues are claims to label the entire endosomal apparatus with a single marker protein, which is hard to accept as certainly this reviewer does not really even know where the limits to the endosomal network reside and where these interface with other structures. There are several additional compartments that have been defined by Rob proteins as well, and which are not even mentioned. Overall I am unconvinced that the authors have demonstrated the main things they claim.<br /> The endomembrane system in bloodstream form T. brucei is clearly delimited. Compared to mammalian cells it is tidy and confined to the posterior part of the spindleshaped cell. The endoplasmic reticulum is linked to one side of the longitudinal cell axis, marked by the attached flagellum, while the mitochondrion locates to the opposite side. Glycosomes are easily identifiable as spheres, as are acidocalcisomes, which are smaller than glycosomes and – in electron micrographs – are characterized by high electron density. All these organelles extend beyond the nucleus, which is not the case for the endosomal compartment, the lysosome and the Golgi. The vesicles found in the posterior half of the trypanosome cell are quantitatively identifiable as COP1, CCVI or CCVII vesicles, or exocytic carriers. The lysosome has a higher degree of morphological plasticity, but this is not topic of the present work. Thus, the endomembrane system in T. brucei is comparatively well structured and delimited, which is why we have chosen trypanosomes as cell biological model.

We have published EP1::GFP as marker for the endosome system and flagellar pocket back in 2004. We have defined the fluid phase volume of the trypanosome endosome in papers published between 2002 and 2007. This work was not intended to represent the entirety of RAB proteins. We were only interested in 3 canonical markers for endosome subtypes. We do not claim anything that is not experimentally tested, we have clearly labelled our hypotheses as such, and we do not make sweeping assumptions.

The approaches taken are state-of-the-art but not novel, and because of the difficulty in fully addressing the central tenet, I am not sure how much of an impact this will have beyond the trypanosome field. For certain this is limited to workers in the direct area and is not a generalisable finding.

To the best of our knowledge, there is no published research that has employed 3D Tokuyasu or expansion microscopy (ExM) to label endosomes. The key takeaway from our study, which is the concept that "endosomes are continuous in trypanosomes" certainly is novel. We are not aware of any other report that has demonstrated this aspect.

The doubts formulated by the reviewer regarding the impact of our work beyond the field of trypanosomes are not timely. Indeed, our results, and those of others, show that the conclusions drawn from work with just a few model organisms is not generalisable. We are finally on the verge of a new cell biology that considers the plethora of evolutionary solutions beyond ophistokonts. We believe that this message should be widely acknowledged and considered. And we are certainly not the only ones who are convinced that the term "general relevance" is unscientific and should no longer be used in biology.

Reviewer #3 (Public Review):

Summary:

As clearly highlighted by the authors, a key plank in the ability of trypanosomes to evade the mammalian host’s immune system is its high rate of endocytosis. This rapid turnover of its surface enables the trypanosome to ‘clean’ its surface removing antibodies and other immune effectors that are subsequently degraded. The high rate of endocytosis is likely reflected in the organisati’n and layout of the endosomal system in these parasites. Here, Link et al., sought to address this question using a range of light and three-dimensional electron microscopy approaches to define the endosomal organisation in this parasite.

Before this study, the vast majority of our information about the make-up of the trypanosome endosomal system was from thin-section electron microscopy and immunofluorescence studies, which did not provide the necessary resolution and 3D information to address this issue. Therefore, it was not known how the different structures observed by EM were related. Link et al., have taken advantage of the advances in technology and used an impressive combination of approaches at the LM and EM level to study the endosomal system in these parasites. This innovative combination has now shown the interconnected-ness of this network and demonstrated that there are no ‘classical’ compartments within the endosomal system, with instead different regions of the network enriched in different protein markers (Rab5a, Rab7, Rab11).

Strengths:

This is a generally well-written and clear manuscript, with the data well-presented supporting the majority of the conclusions of the authors. The authors use an impressive range of approaches to address the organisation of the endosomal system and the development of these methods for use in trypanosomes will be of use to the wider parasitology community.

I appreciate their inclusion of how they used a range of different light microscopy approaches even though for instance the dSTORM approach did not turn out to be as effective as hoped. The authors have clearly demonstrated that trypanosomes have a large interconnected endosomal network, without defined compartments and instead show enrichment for specific Rabs within this network.

Weaknesses:

My concerns are:

i) There is no evidence for functional compartmentalisation. The classical markers of different endosomal compartments do not fully overlap but there is no evidence to show a region enriched in one or other of these proteins has that specific function. The authors should temper their conclusions about this point.

The reviewer is right in stating that Rab-presence does not necessarily mean Rabfunction. However, this assumption is as old as the Rab literature. That is why we have focused on the 3 most prominent endosomal marker proteins. We report that for endosome function you do not necessarily need separate membrane compartments. This is backed by our experiments.

ii) The quality of the electron microscopy work is very high but there is a general lack of numbers. For example, how many tomograms were examined? How often were fenestrated sheets seen? Can the authors provide more information about how frequent these observations were?

The fenestrated sheets can be seen in the majority of the 37 tomograms recorded of the posterior volume of the parasites. Furthermore, we have randomly generated several hundred tiled (= very large) electron micrographs of bloodstream form trypanosomes for unbiased analyses of endomembranes. In these 2D-datasets the “footprint” of the fenestrated flat and circular cisternae is frequently detectable in the posterior cell area.

We now have included the corresponding numbers in all EM figure legends.

iii) The EM work always focussed on cells which had been processed before fixing. Now, I understand this was important to enable tracers to be used. However, given the dynamic nature of the system these processing steps and feeding experiments may have affected the endosomal organisation. Given their knowledge of the system now, the authors should fix some cells directly in culture to observe whether the organisation of the endosome aligns with their conclusions here.

This is a valid criticism; however, it is the cell culture that provides an artificial environment. As for a possible effect of cell harvesting by centrifugation on the integrity and functionality of the endosome system, we consider this very unlikely for one simple reason. The mechanical forces acting in and on the parasites as they circulate in the extremely crowded and confined environment of the mammalian bloodstream are obviously much higher than the centrifugal forces involved in cell preparation. This becomes particularly clear when one considers that the mass of the particle to be centrifuged determines the actual force exerted by the g-forces. Nevertheless, the proposed experiment is a good control, although much more complex than proposed, since tomography is a challenging technique. We have performed the suggested experiment and acquired tomograms of unprocessed cells. The corresponding data is now included as supplementary movie 2, 3 and 4. We refer to it in lines 202 – 206: To investigate potential impacts of processing steps (cargo uptake, centrifugation, washing) on endosomal organization, we directly fixed cells in the cell culture flask, embedded them in Epon, and conducted tomography. The resulting tomograms revealed endosomal organization consistent with that observed in cells fixed after processing (see Supplementary movie 2, 3, and 4).

We furthermore thank the reviewer for the experiment suggestion in the acknowledgments.

iv) The discussion needs to be revamped. At the moment it is just another run through of the results and does not take an overview of the results presenting an integrated view. Moreover, it contains reference to data that was not presented in the results.

We have improved the discussion accordingly.

Recommendations for the authors:

The reviewers concurred about the high calibre of the work and the importance of the findings.

They raised some issues and made some suggestions to improve the paper without additional experiments - key issues include

(1) Better referencing of the trypanosome endocytosis/ lysosomal trafficking literature.

The literature, especially the experimental and quantitative work, is very limited. We now provide a more complete set of references. However, we would like to mention that we had cited a recent review that critically references the trypanosome literature with emphasis on the extensive work done with mammalian cells and yeast.

(2) Moving the dSTORM data that detracts from otherwise strong data in a supplementary figure.

We have done this.

(3) Removal of the conclusion that the continuous endosome fulfils the functions of TGN, without further evidence.

As stated above, this was not a conclusion in our paper, but rather a speculation, which we have now more clearly marked as such. Lines 740 to 751 now read:

“Interestingly, we did not find any structural evidence of vesicular retrograde transport to the Golgi. Instead, the endosomal ‘highways’ extended throughout the posterior volume of the trypanosomes approaching the trans-Golgi interface. It is highly plausible that this region represents the convergence point where endocytic and biosynthetic membrane trafficking pathways merge. A comparable merging of endocytic and biosynthetic functions was already described for the TGN in plants. Different marker proteins for early and recycling endosomes were shown to be associated and/ or partially colocalized with the TGN suggesting its function in both secretory and endocytic pathways (reviewed in Minamino and Ueda, 2019). As we could not find structural evidence for the existence of a TGN we tentatively propose that trypanosomes may have shifted the central orchestrating function of the TGN as a sorting hub at the crossroads of biosynthetic and recycling pathways to the endosome. Although this is a speculative scenario, it is experimentally testable.”

(4) Broader discussion linking their findings to other examples of organelle maturation in eukaryotes (e.g cisternal maturation of the Golgi)

We have improved the discussion accordingly.

Reviewer #1 (Recommendations For The Authors):

What are the multi-vesicular vesicles that surround the marked endosomal compartments in Fig 1. Do they become labelled with fluid phase markers with longer incubations (e.g late endosome/ lysosomal)?

The function of MVBs in trypanosomes is still far from being clear. They are filled with fluid phase cargo, especially ferritin, but are devoid of VSG. Hence it is likely that MVBs are part of the lysosomal compartment. In fact, this part of the endomembrane system is highly dynamic. MVBs can be physically connected to the lysosome or can form elongated structures. The surprising dynamics of the trypanosome lysosome will be published elsewhere.

Figure 2. The compartments labelled with EP1::Halo are very poorly defined due to the low levels of expression of the reporter protein and/or sensitivity of detection of the Halo tag. Based on these images, it would be hard to conclude whether the endosome network is continuous or not. In this respect, it is unclear why the authors didn't use EP1-GFP for these analyses? Given the other data that provides more compelling evidence for a single continuous compartment, I would suggest removing Fig 2A.

We have used EP1::GFP to label the entire endosome system (Engstler and Boshart, 2004). Unfortunately, GFP is not suited for dSTORM imaging. By creating the EP1::Halo cell line, we were able to utilize the most prominent dSTORM fluorescent dye, Alexa 647. This was not primarily done to generate super resolution images, but rather to measure the dynamics of the GPI-anchored, luminal protein EP with single molecule precision. The results from this study will be published separately. But we agree with the reviewer and have relocated the dSTORM data to the supplementary material.

The observation that Rab5a/7 can be detected in the lumen of lysosome is interesting. Mechanistically, this presumably occurs by invagination of the limiting membrane of the lysosome. Is there any evidence that similar invagination of cytoplasmic markers occurs throughout or in subdomains of the endocytic network (possibly indicative of a 'late endosome' domain)?

So far, we have not observed this. The structure of the lysosome and the membrane influx from the endosome are currently being investigated.

The authors note that continuity of functionally distinct membrane compartments in the secretory/endocytic pathways has been reported in other protists (e.g T. cruzi). A particular example that could be noted is the endo-lysosomal system of Dictyostelium discoideum which mediates the continuous degradation and eventual expulsion of undigested material.

We tried to include this in the discussion but ultimately decided against it because the Dictyostelium system cannot be easily compared to the trypanosome endosome.

Reviewer #2 (Recommendations For The Authors):

Abstract

Not sure that 'common' is the correct term here. Frequent, near-universal..... it would be true that endocytosis is common across most eukaryotes.

We have changed the sentence to “common process observed in most eukaryotes” (line 33).

Immune evasion - the parasite does not escape the immune system, but does successfully avoid its impact, at least at the population level.

We have replaced the word “escape” with “evasion” (line 35).

The third sentence needs to follow on correctly from the second. Also, more than Igs are internalised and potentially part of immune evasion, such as C3, Factor H, ApoL1 etcetera.

We believe that there may be a misunderstanding here. The process of endocytic uptake and lysosomal degradation has so far only been demonstrated in the context of VSGbound antibodies, which is why we only refer to this. Of course, the immune system comprises a wide range of proteins and effector molecules, all of which could be involved in immune evasion.

I do not follow the logic that the high flux through the endocytic system in trypanosomes precludes distinct compartmentalisation - one could imagine a system where a lot of steps become optimised for example. This idea needs expanding on if it is correct.

Membrane transport by vesicle transfer between several separate membrane compartments would be slower than the measured rate of membrane flux.

Again I am not sure 'efficient' on line 40. It is fast, but how do you measure efficiency? Speed and efficiency are not the same thing.

We have replaced the word “efficient” with “fast” (line 42).

The basis for suggesting endosomes as a TGN is unclear. Given that there are AP complexes, retromer, exocyst and other factors that are part of the TGN or at least post-G differentiation of pathways in canonical systems, this seems a step too far. There really is no evidence in the rest of the MS that seems to support this.

Yes, we agree and have clarified the discussion accordingly. We have not completely removed the discussion on the TGN but have labelled it more clearly as speculation.

I am aware I am being pedantic here, but overall the abstract seems to provide an impression of greater novelty than may be the case and makes several very bold claims that I cannot see as fully valid.

We are not aware of any claim in the summary that we have not substantiated with experiments, or any hypothesis that we have not explained.

Moreover, the concept of fused or multifunctional endosomes (or even other endomembrane compartments) is old, and has been demonstrated in metazoan cells and yeast. The concept of rigid (in terms of composition) compartments really has been rejected by most folks with maturation, recycling and domain structures already well-established models and concepts.

We agree that the (transient) presence of multiple Rab proteins decorating endosomes has been demonstrated in various cell types. This finding formed the basis for the endosomal maturation model in mammals and yeast, which has replaced the previous rigid compartment model.

However, we do not appreciate attempts to question the originality of our study by claiming that similar observations have been made in metazoans or yeast. This is simply wrong. There are no reports of a functionally structured, continuous, single and large endosome in any other system. The only membrane system that might be similar was described in the American parasite Trypanosoma cruzi, however, without the use of endosome markers or any functional analysis. We refer to this study in the discussion.

In summary, the maturation model falls short in explaining the intricacies of the membrane system we have uncovered in trypanosomes. Therefore, one plausible interpretation of our data is that the overall architecture of the trypanosome endosomes represents an adaptation that enables the remarkable speed of plasma membrane recycling observed in these parasites. In our view, both our findings and their interpretation are novel and worth reporting. Again, modern cell biology should recognize that evolution has developed many solutions for similar processes in cells, about whose diversity we have learned almost nothing because of our reductionist view. A remarkable example of this are the Picozoa, tiny bipartite eukaryotes that pack the entire nutritional apparatus into one pouch and the main organelles with the locomotor system into the other. Another one is the “extreme” cell biology of many protozoan parasites such as Giardia, Toxpoplasma or Trypanosoma.

Higher plants have been well characterised, especially at the level of Rab/Arf proteins and adaptins.

We now mention plant endosomes in our brief discussion of the trypanosome TGN. Lines 744 – 747:

“A comparable merging of endocytic and biosynthetic functions was already described for the TGN in plants. Different marker proteins for early and recycling endosomes were shown to be associated and/ or partially colocalized with the TGN suggesting its function in both secretory and endocytic pathways (reviewed in Minamino and Ueda, 2019).”

The level of self-citing in the introduction is irritating and unscholarly. I have no qualms with crediting the authors with their own excellent contributions, but work from Dacks, Bangs, Field and others seems to be selectively ignored, with an awkward use of the authors' own publications. Diversity between organisms for example has been a mainstay of the Dacks lab output, Rab proteins and others from Field and work on exocytosis and late endosomal systems from Bangs. These efforts and contributions surely deserve some recognition?

This is an original article and not a review. For a comprehensive overview the reviewer might read our recent overview article on exo- and endocytic pathways in trypanosomes, in which we have extensively cited the work of Mark Field, Jay Bangs and Joel Dacks. In the present manuscript, we have cited all papers that touch on our results or are otherwise important for a thorough understanding of our hypotheses. We do not believe that this approach is unscientific, but rather improves the readability of the manuscript. Nevertheless, we have now cited additional work.

For the uninitiated, the posterior/anterior axis of the trypanosome cell as well as any other specific features should be defined.

In lines 102 - 110 we wrote:

“This process of antibody clearance is driven by hydrodynamic drag forces resulting from the continuous directional movement of trypanosomes (Engstler et al., 2007). The VSG-antibody complexes on the cell surface are dragged against the swimming direction of the parasite and accumulate at the posterior pole of the cell. This region harbours an invagination in the plasma membrane known as the flagellar pocket (FP) (Gull, 2003; Overath et al., 1997). The FP, which marks the origin of the single attached flagellum, is the exclusive site for endo- and exocytosis in trypanosomes (Gull, 2003; Overath et al., 1997). Consequently, the accumulation of VSG-antibody complexes occurs precisely in the area of bulk membrane uptake.”

We think this sufficiently introduces the cell body axes.

I don't understand the comment concerning microtubule association. In mammalian cells, such association is well established, but compartments still do not display precise positioning. This likely then has nothing to do with the microtubule association differences.

We have clarified this in the text (lines 192 – 199). There is no report of cytoplasmic microtubules in trypanosomes. All microtubules appear to be either subpellicular or within the flagellum. To maintain the structure and position of the endosomal apparatus, they should be associated either with subpellicular microtubules, as is the case with the endoplasmic reticulum, or with the more enigmatic actomyosin system of the parasites. We have been working on the latter possibility and intend to publish a follow-up paper to the present manuscript.

The inability to move past the nucleus is a poor explanation. These compartments are dynamic. Even the nucleus does interesting things in trypanosomes and squeezes past structures during development in the tsetse fly.

The distance between the nucleus and the microtubule cytoskeleton remains relatively constant even in parasites that squeeze through microfluidic channels. This is not unexpected as the nucleus can be highly deformed. A structure the size of the endosome will not be able to physically pass behind the nucleus without losing its integrity. In fact, the recycling apparatus is never found in the anterior part of the trypanosome, most probably because the flagellar pocket is located at the posterior cell pole.

L253 What is the evidence that EP1 labels the entire FP and endosomes? This may be extensive, but this claim requires rather more evidence. This is again suggested at l263. Again, please forgive me for being pedantic, but this is an overstatement unless supported by evidence that would be incredibly difficult to obtain. This is even sort of acknowledged on l271 in the context of non-uniform labelling. This comes again in l336.

The evidence that EP1 labels the entire FP and endosomes is presented here: Engstler and Boshart, 2004; 10.1101/gad.323404).

Perhaps I should refrain from comments on the dangers of expansion microscopy, or asking what has actually been gained here. Oddly, the conclusion on l290 is a fair statement that I am happy with.

An in-depth discussion regarding the advantages and disadvantages of expansion microscopy is beyond the manuscript's intended scope. Our approach involved utilizing various imaging techniques to confirm the validity of our findings. We appreciate that our concluding sentence is pleasing.

F2 - The data in panel A seem quite poor to me. I also do not really understand why the DAPI stain in the first and second columns fails to coincide or why the kinetoplast is so diffuse in the second row. The labelling for EP1 presents as very small puncta, and hence is not evidence for a continuum. What is the arrow in A IV top? The data in panel B are certainly more in line with prior art, albeit that there is considerable heterogeneity in the labelling and of the FP for example. Again, I cannot really see this as evidence for continuity. There are gaps.... Albeit I accept that labelling of such structures is unlikely to ever be homogenous.

We agree that the dSTORM data represents the least robust aspect of the findings we have presented, and we concur with relocating it to the supplementary material.

F3 - Rather apparent, and specifically for Rab7, that there is differential representation - for example, Cell 4 presents a single Rab7 structure while the remaining examples demonstrate more extensive labelling. Again, I am content that these are highly dynamic strictures but this needs to be addressed at some level and commented upon. If the claim is for continuity, the dynamics observed here suggest the usual; some level of obvious overlap of organellar markers, but the representation in F3 is clever but not sure what I am looking at. Moreover, the title of the figure is nothing new. What is also a bit odd is that the extent of the Rab7 signal, and to some extent the other two Rabs used, is rather variable, which makes this unclear to me as to what is being detected. Given that the Rab proteins may be defining microdomains or regions, I would also expect a region of unique straining as well as the common areas. This needs to at least be discussed.

The differences in the representation result from the dynamics of the labelled structures. Therefore, we have selected different cells to provide examples of what the labelling can look like. We now mention this in the results section.

The overlap of the different Rab signals was perhaps to be expected, but we now have demonstrated it experimentally. Importantly, we performed a rigorous quantification by calculating the volume overlaps and the Pearson correlation coefficients.

In previous studies the data were presented as maximal intensity projections, which inherently lack the complete 3D information.

We found that Rab proteins define microdomains and that there are regions of unique staining as well as common areas, as shown in Figure 3. The volumes do not completely overlap. This is now more clearly stated in lines 315 – 319:

“These objects showed areas of unique staining as well as partially overlapping regions. The pairwise colocalization of different endosomal markers is shown in Figure 3 A, XI - XIII and 3 B. The different cells in Figure 3 B were selected to represent the dynamic nature of the labelled structures. Consequently, the selected cells provide a variety of examples of how the labelling can appear.”

This had already been stated in lines 331 – 336:

“In summary, the quantitative colocalization analyses revealed that on the one hand, the endosomal system features a high degree of connectivity, with considerable overlap of endosomal marker regions, and on the other hand, TbRab5A, TbRab7, and TbRab11 also demarcate separated regions in that system. These results can be interpreted as evidence of a continuous endosomal membrane system harbouring functional subdomains, with a limited amount of potentially separated early, late or recycling endosomes.”

F4-6 - Fabulous images. But a couple of issues here; first, as the authors point out, there is distance between the gold and the antigen. So, this of course also works in the z-plane as well as the x/y-planes and some of the gold may well be associated with membraneous figures that are out of the plane, which would indicate an absence of colinearity on one specific membrane. Secondly, in several instances, we have Rab7 essentially mixed with Rab11 or Rab5 positive membrane. While data are data and should be accepted, this is difficult to reconcile when, at least to some level, Rab7 is a marker for a late-endosomal structure and where the presence of degradative activity could reside. As division of function is, I assume, the major reason for intracellular compartmentalisation, such a level of admixture is hard to rationalise. A continuum is one thing but the data here seem to be suggesting something else, i.e. almost complete admixture.

We are grateful for the positive feedback regarding the image quality. It is true that the "linkage error," representing the distance between the gold and the antigen, also functions to some extent in the z-axis. However, it's important to note that the zdimension of the section in these Figures is 55 nm. Nevertheless, it's interesting to observe that membranes, which may not be visible within the section itself but likely the corresponding Rab antigen, is discernible in Figure 4C (indicated by arrows).

We have clarified this in lines 397 – 400:

“Consequently, gold particles located further away may represent cytoplasmic TbRab proteins or, as the “linkage error” can also occur in the z-plane, correspond to membranes that are not visible within the 55 nm thickness of the cryosection (Figure 4, panel C, arrows). “

The coexistence of different Rabs is most likely concentrated in regions where transitions between different functions are likely. Our focus was primarily on imaging membranes labelled with two markers. We wanted to show that the prevailing model of separate compartments in the trypanosome literature is not correct.

F7 - Not sure what this adds beyond what was published by Grunfelder.

First, this figure is an important control that links our results to published work (Grünfelder et al. (2003)). Second, we include double staining of cargo with Rab5, Rab7, and Rab11, whereas Grünfelder focused only on Rab11. Therefore, our data is original and of such high quality that it warrants a main figure.

F8 - and l583. This is odd as the claim is 'proof' which in science is a hard thing to claim (and this is definitely not at a six sigma level of certainty, as used by the physics community). However, I am seeing structures in the tomograms which are not contiguous - there are gaps here between the individual features (Green in the figure).

We have replaced the term "proof". It is important to note that the structures in individual tomograms cannot all be completely continuous because the sections are limited to a thickness of 250 nm. Therefore, it is likely that they have more connectivity above and below the imaged section. Nevertheless, we believe that the quality of the tomograms is satisfactory, considering that 3D Tokuyasu is a very demanding technique and the production of serial Tokuyasu tomograms is not feasible in practice.

Discussion - Too long and the self-citing of four papers from the corresponding author to the exclusion of much prior work is again noted, with concerns about this as described above. Moreover, at least four additional Rab proteins are known associated with the trypanosome endosomal system, 4, 5B, 21 and 28. These have been completely ignored.

We have outlined our position on referencing in original articles above. We also explained why we focused on the key marker proteins associated with early (Rab5), late (Rab7) and recycling endosomes (Rab11). We did not ignore the other Rabs, we just did not include them in the present study.

Overall this is disappointing. I had expected a more robust analysis, with a clearer discussion and placement in context. I am not fully convinced that what we have here is as extreme as claimed, or that we have a substantial advance. There is nothing here that is mechanistic or the identification of a new set of gene products, process or function.

We do not think that this is constructive feedback.

This MS suggests that the endosomal system of African trypanosomes is a continuum of membrane structures rather than representing a set of distinct compartments. A combination of light and electron microscopy methods are used in support. The basic contention is very challenging to prove, and I'm not convinced that this has been. Furthermore, I am also unclear as to the significance of such an organisation; this seems not really addressed.

We acknowledge and respect varying viewpoints, but we hold a differing perspective in this matter. We are convinced that the data decisively supports our interpretation. May future work support or refute our hypothesis.

Reviewer #3 (Recommendations For The Authors):

Line 81 - delete 's

Done.

Generally, the introduction was very well written and clearly summarised our current understanding but the paragraph beginning line 134 felt out of place and repeated some of the work mentioned earlier.

We have removed this paragraph.

For the EM analysis throughout quantification would be useful as highlighted in the public review. How many tomograms were examined, and how often were types of structures seen? I understand the sample size is often small but this would help the reader appreciate the diversity of structures seen.

We have included the numbers.

Following on from this how were the cells chosen for tomogram analysis? For example, the dividing cell in 1D has palisades associating with the new pocket - is this commonly seen? Does this reflect something happening in dividing cells. This point about endosomal division was picked up in the discussion but there was little about in the main results.

This issue is undoubtedly inherent to the method itself, and we have made efforts to mitigate it by generating a series of tomograms recorded randomly. We have refrained from delving deeper into the intricacies of the cell cycle in this manuscript, as we believe that it warrants a separate paper.

As the authors prosecute, the co-localisation analysis highlights the variable nature of the endosome and the overlap of different markers. When looking at the LM analysis, I was struck by the variability in the size and number of labelled structures in the different cells. For example, in 3A Rab7 is 2 blobs but in 3B Cell 1 it is 4/5 blobs. Is this just a reflection of the increase in the endosome during the cell cycle?

The variability in representation is a direct consequence of the dynamic nature of the labelled structures. For this reason, we deliberately selected different cells to represent examples of how the labelling can look like. We have decided not to mention the dynamics of the endosome during the cell cycle. This will be the subject of a further report.

Moreover, Rab 11 looks to be the marker covering the greatest volume of the endosomal system - is this true? I think there's more analysis of this data that could be done to try and get more information about the relative volumes etc of the different markers that haven't been drawn out. The focus here is on the co-localisation.

Precisely because we recognize the importance of this point, we intend to turn our attention to the cell cycle in a separate publication.

I appreciate that it is an awful lot of work to perform the immuno-EM and the data is of good quality but in the text, there could be a greater effort to tie this to the LM data. For example, from the Rab11 staining in LM you would expect this marker to be the most extensive across the networks - is this reflected in the EM?

For the immuno-EM there were no numbers, the authors had measured the position of the gold but what was the proportion of gold that was in/near membranes for each marker? This would help the reader understand both the number of particles seen and the enrichment of the different regions.

Our original intent was to perform a thorough quantification (using stereology) of the immuno-EM data. However, we later realized that the necessary random imaging approach is not suitable for Tokuyasu sections of trypanosomes. In short, the cells are too far apart, and the cell sections are only occasionally cut so that the endosomal membranes are sufficiently visible. Nevertheless, we continue to strive to generate more quantitative data using conventional immuno-EM.

The innovative combination of Tokuyasu tomograms with immuno-EM was great. I noted though that there was a lack of fenestration in these models. Does this reflect the angle of the model or the processing of these samples?

We are grateful to the referee, as we have asked ourselves the same question. However, we do not attribute the apparent lack of fenestration to the viewing angle, since we did not find fenestration in any of the Tokuyasu tomograms. Our suspicion is more directed towards a methodological problem. In the Tokuyasu workflow, all structures are mainly fixed with aldehydes. As a result, lipids are only effectively fixed through their association with membrane proteins. We suggest that the fenestration may not be visible because the corresponding lipids may have been lost due to incomplete fixation.

We now clearly state this in the lines 563 – 568.

“Interestingly, these tomograms did not exhibit the fenestration pattern identified in conventional electron tomography. We suspect that this is due to methodological reasons. The Tokuyasu procedure uses only aldehydes to fix all structures. Consequently, effective fixation of lipids occurs only through their association with membrane proteins. Thus, the lack of visible fenestration is likely due to possible loss of lipids during incomplete fixation.”

The discussion needs to be reworked. Throughout it contains references to results not in the main results section such as supplementary movie 2 (line 735). The explicit references to the data and figures felt odd and more suited to the results rather than the discussion. Currently, each result is discussed individually in turn and more effort needs to be made to integrate the results from this analysis here but also with previous work and the data from other organisms, which at the moment sits in a standalone section at the end of the discussion.

We have improved the discussion and removed the previous supplementary movies 2 and 3. Supplementary movie 1 is now mentioned in the results section.

Line 693 - There was an interesting point about dividing cells describing the maintenance of endosomes next to the old pocket. Does that mean there was no endosome by the new pocket and if so where is this data in the manuscript? This point relates back to my question about how cells were chosen for analysis - how many dividing cells were examined by tomography?

The fate of endosomes during the cell cycle is not the subject of this paper. In this manuscript we only show only one dividing cell using tomography. An in-depth analysis focusing on what happens during the cell cycle will be published separately.

Line 729 - I'm unclear how this represents a polarization of function in the flagellar pocket. The pocket I presume is included within the endosomal system for this analysis but there was no specific mention of it in the results and no marker of each position to help define any specialisation. From the results, I thought the focus was on endosomal co-localisation of the different markers. If the authors are thinking about specialisation of the pocket this paper from Mark Field shows there is evidence for the exocyst to be distributed over the entire surface of the pocket, which is relevant to the discussion here. Boehm, C.M. et al. (2017) The trypanosome exocyst: a conserved structure revealing a new role in endocytosis. PLoS Pathog. 13, e1006063

We have formulated our statement more cautiously. However, we are convinced that membrane exchange cannot physically work without functional polarization of the pocket. We know that Rab11, for example, is not evenly distributed on the pocket. By the way, in Boehm et al. (2017) the exocyst is not shown to cover the entire pocket (as shown in Supplementary Video 1).

We now refer to Boehm et al. (Lines 700 – 703):

“Boehm et al (2017) report that in the flagellar pocket endocytic and exocytic sites are in close proximity but do not overlap. We further suggest that the fusion of EXCs with the flagellar pocket membrane and clathrin-mediated endocytosis take place on different sites of the pocket. This disparity explains the lower colocalization between TbRab11 and TbRab5A.”

Line 735 - link to data not previously mentioned I think. When I looked at this data I couldn't find a key to explain what all the different colours related to.

We have removed the previous supplementary movies 2 and 3. We now reference supplementary movie 1 in the results section.

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

Du et al. address the cell cycle-dependent clearance of misfolded protein aggregates mediated by the endoplasmic reticulum (ER) associated Hsp70 chaperone family and ER reorganisation. The observations are interesting and impactful to the field.

Strength:

The manuscript addresses the connection between the clearance of misfolded protein aggregates and the cell cycle using a proteostasis reporter targeted to ER in multiple cell lines. Through imaging and some biochemical assays, they establish the role of BiP, an

Hsp70 family chaperone, and Cdk1 inactivation in aggregate clearance upon mitotic exit.

Furthermore, the authors present an initial analysis of the role of ER reorganisation in this clearance. These are important correlations and could have implications for ageingassociated pathologies. Overall, the results are convincing and impactful to the field.

Weakness:

The manuscript still lacks a mechanistic understanding of aggregate clearance. Even though the authors have provided the role of different cellular components, such as BiP, Cdk1 and ATL2/3 through specific inhibitors, at least an outline establishing the sequence of events leading to clearance is missing. Moreover, the authors show that the levels of ERFlucDM-eGFP do not change significantly throughout the cell cycle, indicating that protein degradation is not in play. Therefore, addressing/elaborating on the mechanism of disassembly can add value to the work. Also, the physiological relevance of aggregate clearance upon mitotic exit has not been tested, nor have the cellular targets of this mode of clearance been identified or discussed.

Thank you for your suggestions. 

We have added descriptions about the sequence of events leading to clearance in the abstract (line 33) and discussion (line 316). 

We have commented on the future work that could address the molecular mechanisms behind the aggregate clearance in the discussion (line 388). 

It has been difficult to address the physiological relevance of aggregate clearance during cell division, as the inhibition of BiP or depletion of ATL2/3 that prevent aggregate clearance cause cellular consequences not specific to aggregate clearance. Future work that lead to understanding of aggregate clearance at the molecular level may allow us to address this more specifically. Furthermore, we have commented about the potential defects that could arise in cells expressing ER-FlucDM-eGFP that have a perturbed cellular health based on the proteomic analysis (line 359). 

To identify pathological targets that undergo clearance as the ER-FlucDM-eGFP, we tested three pathological mutants (CFTR-∆F508, AAT S and Z variants) that are known to mis-fold and accumulate in the ER. Unfortunately, expression of these mutants did not result in the confinement of aggregates in the nucleus. The data related to this have been added as Figure S1E and S1F (line 102) in this revised manuscript. We have also commented in the discussion that pathological targets are yet to be identified and could be a part of future work (line 392).

Reviewer #2 (Public review):

This paper describes an interesting observation that ER-targeted misfolded proteins are trapped within vesicles inside nucleus to facilitate quality control during cell division. This work supports the concept that transient sequestration of misfolded proteins is a fundamental mechanism of protein quality control. The authors satisfactorily addressed several points asked in the review of first submission. The manuscript is improved but still unable to fully address the mechanisms.

Strengths:

The observations in this manuscript are very interesting and open up many questions on proteostasis biology.

Weaknesses:

Despite inclusions of several protein-level experiments, the manuscript remained a microscopy-driven work and missed the opportunity to work out the mechanisms behind the observations.

Thank you for your suggestions. We believe that our study has provided a genetic basis for the involvement of ER reorganization and BiP during cell division in aggregate clearance, which is a new observation. We have also commented in this revised manuscript about the future work that could address the molecular mechanisms behind the aggregate clearance in the discussion (line 388).  

Reviewer #3 (Public review):

This paper describes a new mechanism for the clearance of protein aggregates associated to endoplasmic reticulum re-organization that occurs during mitosis.

Experimental data showing clearance of protein aggregates during mitosis is solid, statistically significant, and very interesting. The authors made several new experiments included in the revised version to address the concerns raised by reviewers. A new proteomic analysis, co-localization of the aggregates with the ER membrane Sec61beta protein, expression of the aggregate-prone protein in the nucleus does not result in accumulation of aggregates, detection of protein aggregates in the insoluble faction after cell disruption and mostly importantly knockdown of ATL proteins involved in the organization of ER shape and structure impaired the clearance mechanism. This last observation addresses one of the weakest points of the original version which was the lack of experimental correlation between ER structure capability to re-shape and the clearance mechanism.

In conclusion, this new mechanism of protein aggregate clearance from the ER was not completely understood in this work but the manuscript presented, particularly in the revised version, an ensemble of solid observations and mechanistic information to scaffold future studies that clarify more details of this mechanism. As stated by the authors: "How protein aggregates are targeted and assembled into the intranuclear membranous structure waits for future investigation". This new mechanism of aggregate clearance from the ER is not expected to be fully understood in a single work but this paper may constitute one step to better comprehend the cell capability to resolve protein aggregates in different cell compartments.

We thank the reviewer for the comments.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

The manuscript presents a very interesting set of observations that could have significant implications on age-related protein misfolding and aggregate clearance. There are a few places in the manuscript that still need more clarity. Some are listed below, which I think can improve the manuscript.

- The new data associated with proteomic analysis is appreciated, but the information gained has not been explored or elaborated sufficiently in the manuscript. Based on the differential expression of cell cycle proteins, how the authors interpret cellular health is unclear. Also, the physiological role of this mode of aggregate clearance remains unclear.

We have added our interpretation of perturbed cellular health in cells expressing ERFlucDM-eGFP in the discussion (line 359). 

It has been difficult to address the physiological relevance of aggregate clearance during cell division, as the inhibition of BiP or depletion of ATL2/3 that prevent aggregate clearance cause cellular consequences not specific to aggregate clearance. Future work that lead to understanding of aggregate clearance at the molecular level may allow us to address this more specifically.

- In Figure 3A, have the authors measured the total GFP intensity from interphase through early G1? Even though the number and area of the aggregates decrease significantly, the cytoplasmic GFP signal does not seem to increase. Considering new CHX chase experiments and total Fluorescence intensity calculations (Figure S7D), which indicate no difference, one would expect an increase in cytoplasmic signal upon the disassembly of aggregates. Therefore, the data from Figures 3A and 7D seem contradictory. Can the authors please explain?

We apologized for the confusion. The images in Figure 3A were derived from fixed cells. So, different cells were shown in every cell cycle phases and were not suitable for quantification. Fluorescence intensity changes could be better appreciated in Figure 3C or 4D as these were time-lapse microscopy images of live cells progressing through mitosis and cytokinesis. Data used in the quantification of fluorescence intensity in Figure S7D were derived from live cells taken from specific time points to avoid unwanted fluorescence bleaching during time-lapse microscopy. 

- Do the authors expect a similar clearance of pathological aggregates such as mutant FUS or TDP43 condensates? Showing aggregate disassembly of disease-relevant aggregates would be an excellent addition to the manuscript, but it might be beyond the scope of the current version. However, the authors can comment/speculate how their study might extend to pathological condensates.

We tested three pathological mutants (CFTR-∆F508, AAT S and Z variants) that are known to mis-fold and accumulate in the ER. Unfortunately, expression of these mutants did not result in the confinement of aggregates in the nucleus. The data related to this have been added as Figure S1E and S1F (line 102) in this revised manuscript. We have commented that pathological targets are yet to be identified and could be a part of future work (line 392).

- The presence of ER membrane around these aggregates is an interesting observation. This membrane is retained even after nuclear membrane breakdown. What could be the relevance of membrane-bound aggregates, especially since the membrane can limit the access of chaperones involved in disassembly? This observation becomes more important since the depletion of ER membrane fusion proteins also leads to the accumulation of aggregates. Are the membranes a beacon for disassembly? The authors may comment/ speculate. This could also be an important aspect of the mechanism of clearance.

We think that the ER membranes around the aggregates are disassembled when the ER networks reorganize during mitotic exit and this may allow accessibility of BiP to disaggregate the aggregates. We have added this in the discussion (line 316).

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

__Reviewer #1 (Evidence, reproducibility and clarity (Required)): __ Summary In this work, the authors present a careful study of the lattice of the indirect flight muscle (IFM) in Drosophila using data from a morphometric analysis. To this end, an automated tool is developed for precise, high-throughput measurements of sarcomere length and myofibril width, and various microscopy techniques are used to assess sub-sarcomeric structures. These methods are applied to analyze sarcomere structure at multiple stages in the process of myofibrillogenesis. In addition, the authors present various factors and experimental methods that may affect the accurate measurement of IFM structures. Although the comprehensive structural study is appreciated, there are major issues with the presentation/scope of the work that need to be addressed: Major Comments 1. The main weakness of the paper is in its claim of presenting a model of the sarcomere. Indeed, the paper reports a structural study that is drawn onto a 3D schematic. There is no myofibrillogenesis model that would provide insights into mechanisms. Therefore, the use of the word model is grossly overstated.

In biology, the term “model” is used in various contexts, but it generally refers to a simplified representation of a biological system, a structure or a process. Accordingly, we consider “model” the most fitting phrase for what we present in Figure 4 (Figure 7 in the revised manuscript). These are not arbitrary 3D schematics; they are scaled representations in which the length, the number and the relative three-dimensional arrangement of thin and thick filaments are based on measurements. These measurements are primarily based on our own data (presented in the main text and provided in the supplementary materials), as published data were either lacking or inconsistent. Moreover, we would like to highlight that we do not claim to present a conceptual or mechanistic model of myofibrillogenesis, but we do present structural reconstructions or models for four developmental time points. Therefore, we disagree with the remark that “the use of the word model is grossly overstated”, as our wording fully corresponds to the common sense.

In general, the major focus and contribution of the work is unclear. How does the comprehensive nature of the measurements contribute to existing literature?

We significantly revised the text to highlight the main points more firmly, and added an additional section to help non-specialist readers to better understand our aims and findings.

Figure labels are often rather confusing - for example it is unclear why there is a B, B', B' etc instead of B,C,D, etc.

The figure labels have been revised in accordance with the reviewer’s recommendation.

Some comments in the text are not clearly tied to the figures. For example, in lines 108-109, are the authors referring to the shadow along the edges of the myofibril when saying they are not clearly defined (Figure 1C)?

The lines refer to the fact that identifying the boundary of an “object” in a fluorescence microscopy image is inherently challenging - even under ideal conditions where the object’s image is not affected by nearby signals or background noise. To improve clarity, we revised this section and now it reads: The other key parameter - myofibril diameter - is typically measured using phalloidin staining. However, accurately delineating their boundaries in micrographs is difficult - even under optimal conditions (high signal‑to‑noise ratio, no overlapping fibers, etc.; Fig. 1C). This limitation arises from the fundamental nature of light microscopy as the image produced is a blurred version of the actual structure, due to convolution with the microscope’s point spread function.

In line 116, it is unclear what "surrounding structures" the authors are referring to if the myofibrils are isolated.

We revised the text for clarity. It now states: Once isolated, myofibrils lie flat on the coverslip, aligning with the focal plane of the objective lens. This orientation allows for high-resolution, undistorted imaging and accurate two-dimensional measurements, free from interference by neighboring biological structures (e.g.: other myofibrils).

In lines 141-142, there is no reference of data to back up the claim of validation.

We addressed this mistake by including a reference to Fig. S1E (Fig. S1D in the revised manuscript).

In line 170, the authors mention the mef2-Gal4/+ strain as a Gal4 driver line but do not clearly state how this strain is different from the wildtypes or how this impacts their results.

Mef2-Gal4 is a muscle-specific Gal4 driver, often used in Drosophila muscle studies. It is a convention between Drosophila geneticists that presence of a transgene (i.e. Mef2-Gal4) changes the genetic background, and although it does not necessariliy cause any phenotypic effect, it is clearly distinguished from the wild type situation, and whenever relevant, Mef2-Gal4/+ is the preferred choice (if not the correct choice) as a control instead of wild type. As clear from our data, presence of the Mef2-Gal4 driver line does not affect the length or width of IFM sarcomeres as compared to wild type.

In lines 182-185, the authors discuss the effects of tissue embedding on morphometrics. Were factors such as animal sex, age, fiber type, etc. conserved in these experiments? If not, any differences in results may be confounding.

We fully agree with the reviewer that when testing the effect of a single variable, all other variables should remain constant. This is actually one of the main points emphasized in the results section. Additionally, this information is already provided in the Source Data files for each panel.

In lines 199-201, the authors discuss results of myofibril diameter using different preparation methods, yet no data is cited to support the claims. In line 220, the phrase "6 independent experiments" is unclear. Is each independent experiment performed using a different animal? Furthermore, are 6 experiments performed for each time point?

We substantially revised the relevant paragraphs and ensured that the corresponding data (Figure 2A in the revised manuscript) is cited each time when it is discussed. We conducted six independent experiments at each time point. This is consistently indicated in the figures and can be verified in the SourceData files (specifically, Fig3SourceData in this case). To clarify what we mean by "independent experiments," we added the following sentence to the Methods section: Experiments were considered independent when specimens came from different parental crosses, and each experiment included approximately six animals to capture individual variability.

In line 254, the authors refer to "number of sarcomeres". It must be clearly stated if this refers to sarcomeres per myofibril, image area, etc.

It is now clearly stated as: "number of sarcomeres per myofibril".

In line 274, the authors refer to "myofilament number". It must be clearly stated if this refers to myofilaments per myofibril, image area, etc.

We counted the number of myofilaments in developing myofibrils, and this is now clearly stated in the text and in the legend of Figure 3 (Figure 4 in the revised manuscript).

In line 299, the authors mention that thin filaments measured less than 560 nm in length, yet no data is cited to support this.

The previously missing reference to Figure 4 (Figure 7 in the revised manuscript) has now been added in addition to the revised Supplementary Figure 5.

In the "Quantifying sarcomere growth dynamics" section of the summary (starting from line 402) the authors introduce data that would be more naturally placed in the results and discussion section.

As suggested by the reviewer, we incorporated the key aspects of sarcomere growth dynamics into the Results and Discussion section.

In lines 422-423, it is not mentioned what the controls are for.

This was already explained in the main text between lines 167 and 173.

In the caption of Figure 1C, it is not mentioned what the red dashed lines in the microscope images represent.

The caption has been updated to include the following clarification: The red dashed lines border the ROI used for generating the intensity profiles.

In the caption of Figure 1D, the difference between the lighter and darker grey points is not mentioned.

This was already explained in each relevant figure legend. In this specific case, it is stated between lines 850 and 852: “Light gray dots represent individual measurements of sarcomere length and myofibril diameter, while the larger dots indicate the mean values from independent experiments.”

In line 849, the stated p-value (0.003) does not match that mentioned in the figure (0.0003).

We thank the reviewer for noticing this small mistake; correction was made to display the accurate p-value of 0.0003 at both places.

In line 874, it is not clear what an "independent experiment" refers to (different animal, etc.?).

We refer the reviewer to point 9, where this question has already been addressed.

Figure 2A is hard to read. Using different colored dots for different time points might help.

As suggested by the reviewer, we generated a plot with the individual points color-coded by time.

The significant figures presented in Figure 4 give a completely inaccurate representation of the variability of the measurements achieved with these techniques.

Certainly, each measured parameter exhibits inherent biological and technical variability. We have made all the raw data available to the reader through the SourceData files, and this variability is also evident in Figures 1, 2, 3, Supplementary Figure 1, 3, and 5 (Figure 1, 2, 3, 4, 6, and Supplementary Figure 1 in the revised manuscript). Also we have included an additional plot (Supplementary Figure 5 in the revised manuscript) that presents the calculated thin and thick filament lengths and their uncertainty. However, in Figure 4 (Figure 7 in the revised manuscript), our goal was to present an easily understandable visual representation of the sarcomeric structures for each time point, based on the averages of the relevant measurements.

In line 877, it should be mentioned that the number of filaments is counted per myofibril. The y-axes in the figure should also be adjusted to clarify this.

As suggested by the reviewer, both the figure legend and the plot have been updated to clearly indicate that the filament count refers to the number per myofibril.

In line 883, it is not clear what an "independent experiment" refers to (different animal, etc.?).

We refer the reviewer to point 9, where this question has already been addressed.

The statement of sample sizes in all figures is a little confusing.

Following general guidelines, we used SuperPlots to effectively present the data, as nicely demonstrated in the JCB viewpoint article by Lord et al., 2020 (PMID: 32346721). Individual measurements are shown as pooled data points, allowing readers to appreciate the spread, distribution and number of measurements. Overlaid on these pooled dot plots are the mean values from each independent experiment, with error bars representing variability between independent experiments. Sample sizes are provided for both individual measurements and independent experiments. This is now clearly explained in the Materials and Methods section, and we corrected the legends to improve clarity (“n” indicates the number of independent experiments/individual measurements).

In lines 1007-1008, the authors imply that the lattice model is needed for calculation of myofilament length. However, from the equations and previous data, it seems that this can be estimated using the confocal and dSTORM images.

As the reviewer correctly noted, myofilament length can be estimated using measurements from confocal and dSTORM images, following the equations provided. However, constructing even a simplified model requires multiple constraints to be defined and applied in a specific order. In practice, one must first determine the number and arrangement of myofilaments in a cross-sectional view of an “average sarcomere” before attempting to build a longitudinal model, where length calculations become relevant. This is now clarified in the text.

A more specific discussion of future directions is needed to put this paper in context. For example: Can anything from the overall process be used to better understand sarcomere dynamics in larger animals/humans? Can this be applied to disease modelling?

To address these questions, we have added a section titled STUDY LIMITATIONS, which states: “Our study is focused on describing the growth of IFM sarcomeres during myofibrillogenesis at the level of individual myofilaments. Additionally, we developed a user-friendly software tool for precise sarcomere size measurements and demonstrate that these measurements are sensitive to varying conditions. Whereas, this tool can be used successfully on whole muscle fiber preparations as well, our pipeline was intentionally optimized for individual IFM myofibrils ensuring higher measurement precision in our hands than other type of preparations. Thus, we predict that future work will be required to extend it to sarcomeres from other muscle tissues or species. Nevertheless, our study exemplifies a workflow how to measure sarcomere dimensions precisely. With some variations, it should be possible to adopt it for other muscles, including vertebrate and human striated muscles. To facilitate this and to enhance the accessibility and usability of this dataset, we welcome any feedback and suggestions from researchers in the field.”

One of the major claims of the paper is that there is a measurable variability with sex and other parameters. However, this data is never clearly summarized, presented (except for supplement), or discussed for its implications.

We followed the suggestion of the reviewer, and we moved this supplementary data into a main figure, and thoroughly revised the corresponding paragraphs to present and discuss the findings more clearly.

Minor Comments: 1. Lines 60-65 seem to break the flow of the introduction. As the authors discuss existing methods in literature for IFM analysis in the previous couple sentences, the following sentences should clearly state the limitations of existing methods/current gap in literature and a general idea of what the current work is contributing.

We agree with this remark, and we substantially revised the Introduction to clearly define the existing gap in the literature and to articulate how our work addresses this gap.

In line 104, the acronym for ZASPs is not spelled out.

The acronym has now been spelled out for clarity.

**Referee Cross-commenting**

I agree as well.

Reviewer #1 (Significance (Required)):

In summary, this paper provides a multi-scale characterization of Drosophila flight muscle sarcomere structure under a variety of conditions, which is potentially a significant contribution for the field. However, the paper scope is overstated in that it does not provide an actual sarcomere model. Further, there are multiple issues with data presentation that impact the readability of the manuscript.

Although it is somewhat unclear what would be “an actual sarcomere model” for the reviewer, but we cannot accept that we made on overstatement by using the word “model”, because one of the main outcomes of our work are indeed the myofilament level sarcomere models depicted in Figure 4 (Figure 7 in the revised manuscript). As said above, we do not claim that these would be molecular models, or mechanistic models or developmental models, but it makes absolutely nonsense (even in common terms!) that our scaled graphical representations (based on a wealth of measurements) should not be or cannot be called models.

As to the comment with data presentation, we thank the reviewer for the numerous suggestions, and we substantially revised the manuscript to increase clarity and overall readability.

__Reviewer #2 (Evidence, reproducibility and clarity (Required)): __ Summary: In this manuscript titled "A myofilament lattice model of Drosophila flight muscle sarcomeres based on multiscale morphometric analysis during development," Görög et al. perform a detailed analysis of morphological parameters of the indirect flight muscle (IFM) of D. melanogaster. The authors start by illustrating the range of measurements reported in the literature for mature IFM sarcomere length and width, showing a need to revisit and determine a standardized measurement. They develop a new Python-based tool, IMA, to analyze sarcomere lengths from confocal micrographs of isolated myofibrils stained with phalloidin and a z-disc marker. Using this tool, they demonstrate that sample preparation (especially mounting medium), as well as fiber type, sex, and age influence sarcomere measurements. Combining IMA, TEM, and STORM data, they measure sarcomere parameters across development, providing a comprehensive and up-to-date set of "standardized" sarcomere measurements. Using these data, they generate a model integrating all of the parameters to model sarcomeres at four discrete timepoints of development, recapitulating key phases of sarcomere formation and growth.

Major comments: Line 200 & 901 - Figure S1B - The authors make a strong statement about the use of liquid versus hardening media, and it is clear from the image provided in Figure S1 that there is a difference in the apparent sarcomere width. The identity of the "liquid media" versus the "hardening media" should be clearly identified in the Results, in addition to the legend for Figure S1. The authors show that "glycerol-based solutions" increase sarcomere width, but the Materials only list 90% glycerol and PBS. However, a frequently used liquid mounting media is Vectashield. Based on the literature, measurements in liquid Vectashield show diameters significantly less than 2.2 microns observed here with presumably 90% glycerol or PBS. Can the authors qualify this statement, or provide data that all forms of liquid mounting media cause this effect? Does this also apply to hemi-thorax and sectioned preparations, or just isolated myofibrils?

We used a PBS-based solution containing 90% glycerol as our liquid medium, as now stated in the main text. In response to the reviewer’s suggestion, we also tested a non-hardening version of Vectashield (H-1000). Myofibrils in Vectashield were significantly thicker than those in ProLong Gold but still thinner than those in the 90% glycerol–PBS solution, shown in Figure 2B. The mechanisms that could potentially explain these observations have been described in several studies (Miller et al., 2008; Tanner et al., 2011, 2012). Briefly, IFM is a densely packed macromolecular assembly. Upon removal of the cell membrane, myofibrillar proteins attract water, leading to overhydration of the myofilament lattice. This increases the spacing between filaments, resulting in an expansion of overall myofibril diameter. The extent of hydration depends on the osmolarity of the surrounding medium, as the system eventually reaches osmotic equilibrium. While both liquid media induced significant swelling, the observed differences likely reflect variations in their osmotic properties. In contrast, dehydration - an essential step in electron microscopy sample preparation - reduces the spacing between filaments, making myofibrils appear thinner. This explains why EM micrographs consistently show significantly smaller myofibril diameters (Chakravorty et al., 2017).

        Hardening media such as ProLong Gold introduce additional artifacts: during polymerization, these media shrink, exerting compressive forces on the tissue (Jonkman et al., 2020). We therefore propose that isolated myofibrils first expand due to overhydration in the dissection solution, and are then compressed back toward their *in vivo* dimensions during incubation in ProLong Gold. The average *in vivo* diameter of IFM myofibrils can be estimated without direct measurements, as it is determined by two key factors: (i) the number of myofilaments, which has been quantified in EM cross-sections in several studies (Fernandes & Schöck, 2014; Shwartz et al., 2016; Chakravorty et al., 2017) including our own, and (ii) the spacing between filaments, which can be measured by X-ray diffraction even in live *Drosophila* or under various experimental conditions (Irving & Maughan, 2000; Miller et al., 2008; Tanner et al., 2011, 2012). Our findings suggest that the effects of lattice overhydration and media-induced shrinkage are most pronounced in isolated myofibrils. In larger tissue preparations, the inter-myofibrillar space likely acts as a mechanical and osmotic buffer, reducing the extent of such distortions

Can the authors comment on whether the length of fixation or fixation buffer solution, in addition to the mounting medium, make a difference on sarcomere length and diameter measurements? This is another source of variation in published protocols.

The effect of fixation time on sarcomere morphometrics in whole-mount IFM preparations has been previously demonstrated by DeAguero et al. (2019), as briefly noted in our manuscript. To extend these findings, we performed a comparison using isolated myofibrils, assessing morphometric parameters after fixation for 10, 20 (standard) and 60 minutes. We found no difference between the 10- and 20-minute fixation conditions; however, fixation for 60 minutes resulted in significantly increased myofibril diameter (and these data are now shown in Supplementary Figure 1C). A comparable increase in thickness was also observed when using a glutaraldehyde-based fixative. These results suggest that more extensively fixed myofibrils may better resist the compressive forces exerted by hardening media.

Line 237-238. The authors conclude that premyofibrils are much thinner than previously measured. The use of Airyscan to more accurately measure myofibril width at this timepoint is a good contribution, as indeed diffraction and light scatter likely contribute to increased width measured in light microscopy images. I also wonder, though, how well the IMP software performs in measuring width at 36h APF, given how irregular the isolated myofibrils at this stage look (wide z-lines but thinner and weaker H and I bands as shown in Fig. 2B)?

The reviewer is correct that measurements during the early stages of myofibrillogenesis require additional effort. However, in addition to its automatic mode, IMA can also operate in semi-automatic or manual modes, ensuring complete control over the measurements. Myofibril width is determined from the phalloidin channel at the Z-line (as described in the software’s User Guide and Supplementary Figure 2), where it is at its thickest.

Also, how much of the difference in sarcomere width arises due to effects of "stripping" components off of the sarcomere at the earliest timepoint (for example alpha-actinin or Zasp proteins)?

A comparison between isolated myofibrils and those from microdissected muscles (Supplementary Figure 3B, Figure 3C in the revised manuscript) shows that the isolation process does not alter the morphometric measurements of sarcomeres. Moreover, the measured myofibril width aligns well with what we expect based on the number of myofilaments observed in TEM cross-sections of myofibrils at 36 hours APF (Figure 3A, now Figure 4A in the revised manuscript), supporting the consistency of our model.

Myofibrils at early timepoints do contain more than 4-12 sarcomeres in a line (they extend the full length of the myofiber), so it is possible they are breaking due to the detergent and mechanical disruption induced by the isolation method.

The reviewer is correct - myofibrils likely span the full length of the myofiber from the onset of myofibrillogenesis. However, during the isolation of individual myofibrils, they often break, and even mature myofibrils typically fragment into pieces of about 300 µm in length (illustrated in Figure 1E, now Figure 2A in the revised manuscript). Importantly, our measurements show that this fragmentation does not affect the assessed sarcomere length or width (as shown in Supplementary Figure 3B, now Figure 3C in the revised manuscript).

Line 312 - What does "stable association" mean in this context? The authors mention early timepoints lack stable association of alpha-Actinin or Zasp52, and they reference Fig. S4C, but this figure only shows 72h and 24 AE, not 36h and 48 h APF. Previous reports have seen localization of both alpha-Actinin and Zasp52, so presumably the detergent or mechanical isolation is stripping these components off of the isolated myofibrils up until 72h.

In agreement with previous reports, we also detected both α-Actinin (as shown in former Supplementary Figure 3B, now Figure 3C) and Zasp52 in microdissected IFM starting from 36 hours APF. However, these markers were largely absent from the isolated myofibrils of young pupae (36 to 60 hours APF). By 60 hours APF, strong α-Actinin and Zasp52 staining became evident in isolated myofibrils, whereas dTitin epitopes were clearly detectable from the earliest time point examined. This indicates that some proteins, such as α-Actinin and Zasp52, can be lost during the isolation process, whereas others like dTitin are retained and this differential sensitivity appears to depend on developmental stage. A likely explanation is that α-Actinin and Zasp52 are recruited early to Z-bodies but are only fully incorporated as more mature Z-disks form between 48 and 60 hours APF. This incomplete incorporation at the earlier stages could account for their loss during the isolation process. This interpretation is supported by our morphological analysis of the Z-discs, as shown in the dSTORM dataset (former Figure 3B, B’’, now Figure 4C, E) and in longitudinal TEM sections (former Supplementary Figure 5B, now in Figure 6B). Because α-Actinin and Zasp52 are not detected in isolated myofibrils at 36 and 48 hours APF, they are not included in Figure S4C (Figure 5C in the revised manuscript). This is explained in the updated figure legend.

This same type of issue comes up again in Lines 325-334, where the authors talk about 3E8 and MAC147. They state that 3E8 signal significantly declines in later stages and that MAC147 is not suitable to label myofibrils in young pupae, but they only show data from 72 APF and 24 AE (which looks to have decent staining for both 3E8 and MAC147). A clearer explanation here would be helpful.

To put it simply: we used one myosin antibody to label the A-band in the IFM of 36h APF and 48h APF animals, and a different antibody for the 72h APF and 24h AE stages. In more detail: Myosin 3E8 is a monoclonal antibody targeting the myosin heavy chain and labels the entire length of mature thick filaments except for the bare zone (former Supplementary Figure 4D, now in Figure 5D), suggesting its epitope is near the head domain. As a result, we expect a uniform A-band staining - excluding the bare zone - which is exactly what we observe in the IFM of young pupae (36h APF and 48h APF; formerly Figure 3B, now Figure 4C in the revised manuscript). However, at 72h APF and 24h AE, Myosin 3E8 produces a different staining pattern: two narrow stripes flanking the bare zone and two broader, more diffuse stripes near the A/I band junction (former Supplementary Figure 4D, now Figure 5D). This change is likely due to restricted antigen accessibility at these later developmental stages - a common issue in the densely packed IFM - making this antibody unsuitable for reliably measuring thick filament length in these stages.

MAC147 is another monoclonal antibody against Mhc that recognizes an epitope near the head domain. However, it only works reliably in more mature myofibrils (72h APF and 24h AE; formerly Figure 3B, now Figure 4C in the revised manuscript), likely due to its specificity for a particular Mhc isoform. This is why we do not include images from earlier developmental stages using this antibody. We added a revised, concise explanation in the main text for general readers, and provided a more detailed description for specialist readers in the legend of Supplementary Figure 4D (updated as Figure 5D in the revised manuscript).

Figure 3B. The authors show the H, Z, and I lengths in B', B', and B' and discuss these lengths in the text (lines 305-320). It would also be nice to actually have the plots showing the measured/calculated lengths for thin and thick filaments. These are mentioned in the results, but I cannot find the plots in the figures and there is no panel reference.

A summary table of the measured and calculated parameters is provided in Fig4SourceData (Fig7Source Data in the revised manuscript). However, following the reviewer’s suggestion, we also generated an additional plot (Supplementary Figure 5 in the revised manuscript) that displays the calculated thin and thick filament lengths.

Line 400. Does the model in Figure 4 actually have molecular resolution as the authors claim? From these views, thick and thin filaments appear to be represented by cylindrical objects. Localization of specific molecules would require further modeling with individual proteins. Or do the authors mean localization from STORM imaging relative to the ends of the thick and/or thin filaments? The model itself is a useful contribution, but based on Figure 4, resolution of individual molecules is not evident.

The reviewer is correct; and we fully agree that we do not present a molecular model of sarcomeres in this study - nor do we claim to. Instead we present a myofilament level model. Nevertheless, the scaled myofilament lattice model we introduce could serve as a geometric constraint when constructing supramolecular models of sarcomeres. As the reviewer rightly notes, implementing such an approach would require additional effort.

The main Results section of the text is condensed into 4 figures. However, I found myself flipping back and forth between the main figures and the supplement continuously, especially parts of Supplemental Figures 1, 3, 4, and 5. With such large amounts of detail in the Results relying on the supplement, it may be worth considering reorganizing the main and supplemental figures, and having 7 main figures, to include important panels that are currently in the supplement (esp. Fig S1B, S1C, S1D, S3B, S4, S5).

We found it a very useful suggestion, and we substantially reorganized the figures in the revised manuscript according to the recommendations of the reviewer.

Minor comments: On the plots in Fig. S1B, D, and F, it is hard to see the color of the dots because the red error bars are on top of them. Can the other distribution dots be tinted the correct color or the x-axis labels be added, so it is clear which dataset is which?

We significantly enlarged the dots to enhance visual clarity.

Line 142 needs a reference to Figure S1, Panel E, which shows the accuracy and precision measurements.

The requested panel reference has now been included in the revised manuscript.

Lines 198 - is this range from the above publications? Needs to be clearly cited.

The range has indeed been estimated using measurements from the aforementioned publications, and this point is now further clarified in the revised text.

Figure S3B is confusing - why do the blow-ups overlap both the top (presumably microdissected) and the bottom (presumably isolated) images? The identity of microdissected images should be labeled, as they are hard to see underneath of the blown-up images and the identity of individual image planes wasn't immediately obvious.

We refined the panel structure of Figure S3B (Figure 3C in the revised manuscript) to enhance clarity as the reviewer suggested.

Line 298. By "misaligned," do the authors mean the pointed ends are not uniformly anchored in the z-disc, leading to the wide z-disc measurements? At this early stage, I'm not sure "misaligned" is the right word - perhaps "were not yet aligned in register at the z-disc" or something similar.

We revised the text for clarity. It now reads: At 36 hours APF, thin filaments had not yet aligned in perfect register at the Z-disc, with most measuring less than 560 nm in length - and exhibiting considerable variability.

Figure S6 - spelling mistake in label of panel A, "sarcomer" should be "sarcomere"

The typo is corrected.

Line 487. Spelling "Zaps52" should be "Zasp52"

The typo is corrected.

Line 887. Spelling "Myofilement" should be "Myofilament"

The typo is corrected.

Line 946-947. In the legend for Supp. Fig. 3., the authors should specify which published datasets on sarcomere length are shown in the figure by including the references in the legend. Presumably the "isolated individual myofibrils" are the blue "this study" lines, leaving the "microdissected muscles" as the magenta "previous reports" on the figure. Without the reference, it is not clear if these are microdissected, isolated myofibrils, hemi-thorax sections, cryosections, or another preparation method for the "previous reports" data.

The references have now been added to both the figure and its legend.

**Referee Cross-commenting**

I agree with the comments from the other reviewers. Many of the major themes are consistent across the reviews, including regarding the model, preparation methods, and the software tool.

Reviewer #2 (Significance (Required)):

Strengths: This manuscript is an important contribution to the field of sarcomere development. The authors use modern technologies to revisit variation in morphometric measurements in the literature, and they identify parameters that influence this variation. Notably, sex-specific differences, DLM versus DVM measurements, and mouting media are potential contributors to the variability. Combining TEM and STORM with a confocal timecourse of isolated myofibrils, they refine previously published values of sarcomere length and width, and add more comprehensive data for filament length, number and spacing. This highly accurate timecourse demonstrates continual growth of sarcomeres after 48 h APF, and correct some inconsistencies from previous large-scale timecourse datasets. These data are very valuable to the field, especially Drosophila muscle biologists, and will serve as a comparative resource for future studies. Weaknesses: At early timepoints, loss of sarcomere components through mechanical or detergent-mediated artifacts may influence the authors' measurements. In addition, isolating myofibrils is not always the most ideal approach, as it loses information on myofiber structure as well as organization and structure of the myofibrils in vivo.

We believe that the control experiments we presented here adequately demonstrate that sarcomere measurements are not affected by the myofibril isolation process at early timepoints (Figure 3C). Nevertheless, we certainly agree with the reviewer that isolated myofibrils alone cannot capture the entire complexity of muscle tissues, and additional approaches should also be applied in complex projects. Yet, we are confident that our approach offers the most reliable and efficient method for precise morphometric analysis of the sarcomeres, and although alone it is very unlikely to be sufficient to address all questions of a muscle development project, it can still be applied as a very useful and robust tool.

The point regarding liquid versus hardening mounting media is valuable, but remains to be tested and validated with the diverse liquid and hardening media used by other labs.

Whereas it would not be feasible for us to test all possible liquid and hardening media used by others in all possible conditions, we tested the effect of Vectashield (the most commonly used liquid media) according to the suggestion of the reviewer, and the results are now included in the manuscript. We think that this is a valuable extension of the list of the materials and conditions we tested, although we need to point out that our primary goal was not necessarily to test as many conditions as possible (because the number of those conditions is virtually endless), rather to raise awareness among colleagues that these variables can significantly impact the data obtained and affect their comparability.

The IMA software seems to be designed specifically for analysis of isolated myofibrils, and it is unclear if it would work for other types of IFM preparations.

As stated in the manuscript, IMA is a specialized tool designed for the analysis of individual myofibrils. While it can also process other types of IFM preparations in semi-automatic or manual modes, we believe these approaches compromise both efficiency and accuracy. This is further clarified in the revised manuscript.

A last point is that TEM and STORM may not be available on a regular basis to many labs, hindering wide implementation of the approach used in this manuscript to generate very accurate and detailed measurements of sarcomere morphometrics.

Regarding the availability of TEM and STORM, we acknowledge that these techniques are not universally accessible. However, that is exactly one major value of our work that our open-source software tool now allows researchers to generate valuable data using only a confocal microscope in combination with our published datasets.

Audience: Scientists who study sarcomerogenesis or Drosophila muscle biology.

My expertise: I study muscle development in the Drosophila model.

__Reviewer #3 (Evidence, reproducibility and clarity (Required)): __ Summary: This manuscripts presents a computational tool to quantify sarcomere length and myofibril width of the Drosophila indirect flight muscles, including developmental samples. This tool was applied to confocal and STORM super-resolution images of isolated myofibrils from adult and developing flight muscles. Thick filament numbers per myofibril were counted during development of flight muscles. A myofilament model of developing flight muscle myofibrils is presented that remains speculative for the early developmental stages.

Major comments: 1. The title of the manuscript appears unclear. What is a lattice model? Lattice is an ordered array. The filament array parameters for mature flight muscles was aready measured. It appears that the authors speculate how this order might be generated during sarcomere assembly, which is not studied in this manuscript as it is limited to periodic arrays after 36h APF.

As the reviewer correctly points out, a lattice refers to an ordered array - in the case of IFM sarcomeres, this includes both thin and thick filaments. Therefore, the phrase "myofilament lattice model of Drosophila flight muscle sarcomeres" specifically describes a model representing the spatial organization of these filament arrays within the sarcomere. To provide additional clarity for readers, we have revised the title to include more context. It now reads: Developmental Remodeling of Drosophila Flight Muscle Sarcomeres: A Scaled Myofilament Lattice Model Based on Multiscale Morphometrics

To create a model of these arrays, three essential pieces of information are required:

1) The length of the filaments,

2) The number of filaments, and

3) The relative position of the filaments.

While some direct measurements are available in the literature, and others can be used to calculate the necessary values, available data is often contradictory or simply different from each other (as described in our ms) making them unsuitable for constructing scaled models of the myofilament arrays. In contrast to that, here we present a comprehensive and consistent set of measurements that enabled us to build models not only of mature sarcomeres but also of sarcomeres at three other significant developmental time points.

Regarding the mention of "sarcomere assembly" in line 37, we intended it to refer to the growth of the sarcomeres, not their initial formation. We do not speculate about sarcomere assembly anywhere in the text. In fact, we have clearly stated multiple times that our focus is on the growth of the IFM myofilament array during myofibrillogenesis. Nevertheless, to avoid confusion, we revised the phrase in line 37 to "sarcomere growth".

The authors review the flight muscle sarcomere length literature and conclude it is variable because of imprecise measurements. Likely this is partially true, however, more importantly is that the sarcomere length and width changes during isolation methods of the myofibrils, as well as by various embedding methods, as the authors show here as well in Figure 1B-E.

We dedicated two sections of the Results - “An automated method to accurately measure sarcomeric parameters” and “IFM sarcomere morphometrics are affected by sex, age, fiber type, and sample preparation” - to exploring potential sources of variability in published IFM sarcomere measurements. Based on these analyses, we conclude that such variability stems from both measurement imprecision and biological or technical factors, including sex, age, fiber type and, of foremost, sample preparation. Because it is difficult to quantify the relative impact of each variable across published studies, we have refrained from speculations about the relative contribution of the different factors in the revised manuscript.

Hence, I find the strongly claims the authors make here surprising, while they are isolating the myofibrils. Hence, these myofibrils are ruptured at the ends, relaxed or contracted, depending on buffer choice and passive tension is released. On page 8, the authors correctly state that the embedding medium causes shrinkage of the myofibrils. While isolation is state of the art for electron microscopy techniques, other methods including sectioning or even whole mount preparation have been developed for high resolution microscopy of IFMs that avoid these artifacts. Unfortunately, this manuscript only uses isolated myofibrils that were fixed and then mechanically dissociated by pipetting. This method likely induces variations as seen by the large spread of sarcomere length reported in Figure 1C (2.8-3.9µm?) and even bigger spreads for myofibril widths. Are these also seen in tissue without dissections? Unfortunately, no comparision to intact flight muscles are reported with the here presented quantification tool. The sarcomere length spread in the developmental samples is even larger.

The major issue raised in this paragraph is the use of isolated myofibril versus intact flight muscle preparations. The reviewer claims that the latter might be superior because the isolated myofibrils are ruptured at their ends. Clearly, the intact IFMs cannot be imaged in vivo by light microscopy because the adult fly cuticle is opaque. To visualize these muscles, one must open the thorax, but neither microdissection nor sectioning preserves them perfectly, even the cleanest longitudinal cuts sever some myofibrils, and dissection itself can damage the tissue. Although published images often show only the most pristine regions, the practice of selective cropping cannot be taken as a scientific argument. Here, by comparing sarcomere lengths measured in isolated myofibrils with those from whole-mount longitudinal DLM sections and microdissected IFM myofibers, we demonstrate that isolation does not alter sarcomere length (Figure 1E, now Figure 2A in the revised manuscript). As to myofibril width, it is determined by two parameters: the number of myofilaments and the spacing between them. In vivo filament spacing has been measured directly, and filament counts can be obtained from EM cross-sections of DLM fibers. Combining these values gives an expected in vivo myofibril diameter. While isolated myofibrils measure thinner than those in whole-mount or microdissected samples (Figure 1E, now Figure 2A in the revised manuscript), their diameter closely matches this in vivo estimate (see manuscript, lines 187–198). Therefore, we conclude that isolated myofibrils (even if it seems counterintuitive for this reviewer) are superior for sarcomere measurements than whole-mount preparations - and that is why we primarily rely on them here.

Despite that, we certainly recognize that isolated myofibrils cannot recapitulate every aspect of an IFM fiber, and the need for whole-mount preparations during our IFM studies is not questioned by us.

        In addition to this general answer to the issues raised in the above paragraph of the reviewer, we would like to specifically reflect for some of the remarks:

„Unfortunately, this manuscript only uses isolated myofibrils that were fixed and then mechanically dissociated by pipetting.”

This is a false statement that “this manuscript only uses isolated myofibrils” as we used different preparation methods for initial comparisons (see Figure 1E, now Figure 2A in the revised manuscript). Additionally, unlike the reviewer assumed, the myofibrils were first dissociated and then fixed, and not vice versa (as described in the Materials and Methods section).

„This method likely induces variations as seen by the large spread of sarcomere length reported in Figure 1C (2.8-3.9µm?) and even bigger spreads for myofibril widths. Are these also seen in tissue without dissections?”

This remark makes absolutely no sense, as we do not report sarcomere length values in Figure 1C at all. By assuming that the reviewer meant to refer to Figure 1B, it still remains a misunderstanding or a false statement, because that panel refers to the variations found in published data (not in our current data), and this is clearly explained both in the figure legend and the main text. Regardless of that, the stated spread does not appear unusual. In the article by Spletter et al. (2018), the authors report a similar spread (2.576–3.542 µm) for sarcomere length in mature IFM using whole-mount DLM cross-sections. As to the second question here, we do observe a comparable spread in other preparations as well (see Figure 1E, now Figure 2A in the revised manuscript), which is again the opposite conclusion as compared to the (clearly false) assumption of the reviewer.

„Unfortunately, no comparision to intact flight muscles are reported with the here presented quantification tool. „

This is also a false statement; as we do report comparison to whole mount cross sections which we belive the reviewer considers „intact” in Figure 1E (Figure 2A in the revised manuscript).

„The sarcomere length spread in the developmental samples is even larger.”

The spread is not larger at all than in previous reports, as clearly shown in Supplementary Figure 3A.

The authors suggest that there are sex differences in sarcomere length and pupal development duration. This is potentially interesting, unfortunately they then use mixed sex samples to analyse sarcomeres during flight muscle development.

In the revised manuscript, we now provide a more detailed description of a subtle post-eclosion difference in IFM sarcomere metrics between male and female Drosophila. We attribute this variation to the well-established observation that female pupae develop slightly faster than males, a property that may last till shortly after eclosion. Confirming this experimentally would require considerable effort with limited scientific benefit. Nonetheless, the subtle nature of this sex-linked variation reinforced our decision to include IFM sarcomeres from both male and female flies in our comprehensive developmental analysis.

The IMA software tool lacks critical assessment of its performance compared to other tools and the validation presented is too limited. IMA seems to generate systematic errors, based on Fig S1E, as it does not report the ground truth. These have to be discussed and compared to available tools. The principles of fitting used in IMA seem well adapted to IFM myofibrils in low noise conditions, but may not be usable in other situations. This should be assessed and discussed.

IMA is a specialized software tool developed to address a specific need, notably, to accurately and efficiently measure sarcomere length and myofibril diameter in individual IFM myofibril images labeled with both phalloidin and Z-disc markers. For our purposes, it remains the most suitable and reliable option, and we are confident that IMA outperforms all other available tools. To demonstrate this, we have included a table comparing the few alternatives (MyofibrilJ, SarcGraph, and sarcApp) capable of both measurements, which further supports our conclusion. Given IMA's focused application, extensive validation under artificially low signal-to-noise conditions is unnecessary. While IMA may introduce minor systematic errors (~0.01 µm for sarcomere length and ~0.03 µm for myofibril diameter), these are negligible errors relative to the limitations of the simulated ground truth data used for benchmarking. This point is now addressed in the manuscript.

It is claimed that validation was achieved on simulated IFM images: do the authors rather mean simulated isolated IFM myofibril images? This is not quite the same in terms of algorithm complexity and this should be corrected if this is the case.

Indeed, we used simulated individual IFM myofibril images, where both phalloidin labeling and Z-disc labeling are present. This is clearly shown in Supplementary Figure 1A, and stated in the text when first introduced: „we generated artificial images of IFM myofibrils with known dimensions, simulating the image formation process”

The authors need to revise their comparison to other tools. It is incomplete and seemingly incorrect. It should be clearly stated that IMA is limited to isolated myofibrils, which is a far easier segmentation task than what other tools can do, such as sarcApp (Neininger-Castro et al. 2023, PMID: 37921850). Defining the acronym would be valuable in that sense. The claim line 129-130 "none can adequately measure myofibril diameter from regular side view images" is unclear. What do the authors refer to as "side view images"? Sarc-Graph from Zhao et al 2021, PMID: 34613960, and sarcApp from Neininger-Castro et al. 2023 provide sarcomere width, in conditions that are very similar to what IMA does, e.g. on xy images based on the documentation provided on github. A performance comparison with these tools would be valuable. Does installation and use of IMA require computational skills?

Motivated by the reviewer’s comments, we revised the section introducing IMA. However, we chose not to include an extensive comparison with other software tools, as this would divert the manuscript’s focus without impacting the main conclusions. Instead, we added a summary table highlighting the key requirements for analyzing IFM sarcomere morphometrics from Z-stacks of phalloidin- and Z-line-labeled individual myofibrils and compared the available tools accordingly. In our experience, most software tools are developed to address very specific problems, even those marketed as general-purpose solutions. Consequently, applying them beyond their intended scope often results in reduced efficiency and suboptimal performance. Although sarcApp was initially available as a free tool, one of its dependencies (PySimpleGUI 5) has since adopted a commercial license model. Using a trial version of PySimpleGUI 5, we evaluated sarcApp on our dataset. The software is limited to single-plane image input, hence raw image stacks must be preprocessed into a suitable format, which is a time consuming step. Furthermore, implementation requires basic programming proficiency, as parameter adjustments must be performed directly within the source code to accommodate dataset-specific configurations. Once appropriately configured, sarcApp reliably quantifies both sarcomere length and myofibril width with accuracy comparable to that of IMA. However, it lacks built-in diagnostic feedback or visualization tools to facilitate measurement verification or troubleshooting during batch processing. SarcGraph also supports only single-plane image inputs and requires prior image preprocessing. Additionally, images must be loaded manually one by one, which further reduces processing efficiency. Parameter optimization relies on direct code modification through a trial-and-error process, demanding a certain level of programming proficiency. Even with these adjustments, the software frequently introduces artifacts - such as Z-line splitting - when applied to our dataset. Even when segmentation is successful, sarcomere length is often overestimated, whereas myofibril diameter is consistently underestimated. As compared to these issues, IMA was designed for ease of use and does not require any programming experience to install or operate. It can automatically handle raw microscopic image formats without the need for preprocessing. Segmentation is fully automated, with no requirement for parameter tuning. The tool provides visual feedback during both the segmentation and fitting steps, allowing users to confidently assess and validate the results. IMA produces accurate and precise measurements of sarcomere length and diameter. Batch processing is enabled by default, significantly improving efficiency when analyzing multiple images. Finally, unlike the reviewer stated, IMA is not limited to isolated myofibrils. It is optimized for isolated myofibrils (i.e. full performance is achieved on these samples), but it can also work on whole-mount preparations in semi-automatic and manual mode, which still allow precise measurements (with some reduction in processing efficiency).

As to the minor comments, the acronym IMA was already defined in lines 541 and 917–918 of the original submission, as well as on the software’s GitHub page. Additionally, we replaced the phrase "side view images" with "longitudinal myofibril projections" to improve clarity.

How do the authors know that the bright phallodin signal visible that the Z-disc at 36h and 48h APF is due to actin filament overlap, as suggested? An alternative solution are more short actin filaments at the early Z-discs.

It is widely accepted that the bright phalloidin signal at the Z-line in mature sarcomeres reflects actin filament overlap (e.g., Littlefield and Fowler, 2002; PMID: 11964243). Accordingly, in slightly stretched myofibrils, this bright signal diminishes, and in more significantly stretched myofibrils, a small gap appears (e.g., Kulke et al., 2001; PMID: 11535621). The width of this bright phalloidin signal corresponds to the electron-dense band seen in longitudinal EM sections (Figure 3B and Supplementary Figure 5B, now Figure 4B and Figure 6B in the revised manuscript) and matches the actin filament overlap observed in Z-disc cryo-EM reconstructions from other species (Yeganeh et al., 2023; Rusu et al., 2017), where individual thin filaments can be resolved. By extension, we interpret the bright phalloidin signals at the Z-discs observed at 36 h and 48 h APF as arising from similar actin filament overlaps, given their comparable width to the electron-dense Z-bodies described both in our study (Supplemantary Figure 5B, now Figure 6B in the revised manuscript) and by Reedy and Beall (1993). While we cannot fully rule out the reviewer’s alternative interpretation, for the time being it remains a bold speculation without supporting evidence, and therefore we prefer to stay with the conventional view.

The authors seem to doubt their own interpretation that actin filaments shrink when reading line 304 and following. This is obviously critical for the "model" presented.

Unlike the reviewer implies, we certainly do not doubt our own interpretation, but to avoid confusion we revised the corresponding paragraph in the manuscript and provided more details on our explanation, and we also provide a brief overview of it here. Between 36 h and 48 h APF we observe a pronounced structural transition in the IFM sarcomeres. In EM cross-sections, the previously irregular myofilament lattice becomes organized into a regular hexagonal pattern (Figure 3A, now Figure 4A in the revised manuscript) with filament spacing typical of mature myofibrils (Supplementary Figure 5A, now Figure 6A in the revised manuscript). In longitudinal EM sections, the elongated, amorphous Z-bodies condense along the myofibril axis to form well-defined, adult-like Z-discs (Supplementary Figure 5B, now Figure 6B in the revised manuscript). Similarly, dSTORM imaging shows that the Z-disc associated D-Titin epitopes become more compact and organized during this period (Supplementary Figure 4E, now Figure 5E in the revised manuscript). The edges of the thick filament arrays also become more sharply defined, and the appearance of a distinct bare zone indicates the establishment of a regular register (Figure 3B, now Figure 4B in the revised manuscript). By assuming that a similar reorganization occurs within the thin filament array, the apparent length of the thin filament array would decrease—not due to shortening of individual filaments, rather due to improved alignment. Although we cannot directly resolve single thin filaments, this reorganization offers the most plausible explanation for the observed change.

Minor comments: 1. Figure S1B is not called out in the text.

The reviewer might have missed this, but in fact, it is explicitly called out in line 181.

Fig. 1: Please state whenever images are simulations?

We appreciate the reviewer’s observation that the simulated IFM myofibril images are indistinguishable from the real ones, as this confirms the adequacy of these images for testing our software tool. However, this is already clearly indicated: Figure 1B features simulated images, as noted in the figure legend (line 824), and Supplementary Figure 1A similarly shows simulated images, as stated both in the legend (line 886) and in the figure.

Fig. 2: Length-width correlation - please provide individual points color-coded by time point?

As suggested by the reviewer, we generated a plot with the individual points color-coded by time.

"newly eclosed males and females, we observed that males have slightly shorter sarcomeres and narrower myofibrils". Please provide a statistical test supporting the difference.

In the revised manuscript, we compared sarcomere length and myofibril width between males and females from 0 to 96 hours AE using a two-way ANOVA with Sidak’s multiple comparisons test. We expanded our description of these observations in the main text, and details of the statistical analysis are now included in the revised figure legend (Figure 1E). Briefly, newly eclosed males showed slightly shorter sarcomeres than females - a consistent but non-significant trend (p = 0.9846) - which resolved by 12 h AE, with sarcomere lengths remaining similar thereafter (p = 0.1533; Figure 1E). In contrast, myofibril width was significantly narrower in the newly eclosed males (p = 0.0374), but this difference disappeared between 24 and 48 h AE as myofibrils expanded in diameter during post-eclosion development (p

Were statistical tests performed using animals as sample numbers? Please clarify in the images what are animal and what are sarcomere numbers.

Following standard guidelines, statistical tests were performed using the means of independent experiments, as noted in the figure legends. For each experiment, we used approximately 6 animals, and this information is now included in the Materials and Methods section.

mef2-Gal4 should be spelled Mef2-GAL4 according to Flybase.

This has been corrected in the revised text and figures.

Are the images shown in Figure 2B representative? 96h AE appears thicker than 24h AE but the graph reports no difference.

We aimed to show representative images, however, in the case of 96h APF we may have selected a wrong example. We now changed the image for a more appropriate one.

The authors only found Zasp52 and alpha-Actinin at the Z-discs from 72h APF onwards, which is different to what others have reported.

Similarly to former reports, we detected both α-Actinin (see Supplementary Figure 3B, now Figure 3C in the revised manuscript) and Zasp52 in microdissected IFMs as early as 36 hours APF. However, these markers were largely absent in isolated myofibrils from the early pupal stages (36–60 hours APF). By 60 hours APF, strong α-Actinin and Zasp52 signals were clearly visible in isolated myofibrils (the closest timepoint captured by dSTORM is 72h APF). As discussed in the manuscript, a likely explanation is that α-Actinin and Zasp52 are recruited to developing Z-bodies early on but are only fully incorporated into mature Z-discs between 48 and 60 hours APF. Their incomplete integration at earlier stages may lead to their loss during the isolation procedure.

Thick filament length during development has also been estimated by Orfanos and Sparrow, which should be cited (PMID: 23178940)

Contrary to the reviewer’s claim, the article 'Myosin isoform switching during assembly of the Drosophila flight muscle thick filament lattice' does not provide any measurements or estimates of thick filament length; it only includes a schematic illustration where the length of the thick filaments is not based on empirical data.

**Referee Cross-commenting**

I also agree with my colleagues comments, which are largely consistent.

Reviewer #3 (Significance (Required)):

This paper introduces a tool to measure sarcomere length. Easy to use tools that do this as well already exist. The tool can also measure sarcomere width, which it claims as unique point, which is not the case, see above comment.

We are aware that other tools exist to measure sarcomere parameters (and we did not claim the opposite in our ms), nevertheless, we need to emphasize that based on our comparisons, IMA is superior to all three alternatives. Three software tools could, in principle, be used to measure both sarcomere length and myofibril diameter: MyofibrilJ, SarcGraph, and sarcApp. However, two of them - MyofibrilJ and SarcGraph - consistently under- or overestimate these values. The only tool capable of performing these measurements reliably, sarcApp, is no longer freely available, it requires programming expertise, and it does not support raw image file formats, making it difficult to use in practice (see above comments for more details). In contrast, IMA is user-friendly and does not require any programming expertise to install or operate. It can automatically process raw microscopic image formats without the need for preprocessing. Segmentation is fully automated, and no parameter tuning is necessary. The tool offers visual feedback on both the segmentation and fitting processes, enabling users to validate results with confidence. IMA delivers accurate and precise measurements of sarcomere length and diameter. Additionally, batch processing is enabled by default, significantly enhancing workflow efficiency.

This manuscript shows that depending on the isolation and embedding media sarcomere and myofibrils width changes and hence artifacts can be introduced. While this is not suprising, it has not been well controlled in a number of previous publications.

Furthermore, this paper measures sarcomere length and width during flight muscle development and consolidates what was already known from previous publications. Sarcomeres are added until 48 h APF, then they grow in diameter. Despite strong claims in the text, I do not see any significant novel findings how sarcomeres grow in length or width or any significant deviations from what has been published before. This is even documented in the supplementary graphs by comparing to published data. It is close to identical.

The overall process has been quantitatively described in four previous studies (Reedy and Beall, 1993, Orfanos et al., 2015, Spletter et al., 2018, Nikonova et al., 2024). While there is general agreement on the pattern of sarcomere development, significant discrepancies exist among these datasets; differences that become particularly problematic when attempting to build structural models. More specifically: Reedy and Beall (1993) report substantially shorter sarcomeres compared to all other datasets, including ours. This discrepancy likely stems from two factors: (i) their use of longitudinal EM sections, where sample preparation is known to cause considerable tissue shrinkage; and (ii) the maintenance of their flies at 23 °C, a temperature that clearly delays development relative to the more commonly used 25 °C. Interestingly, Spletter et al. (2018) and Nikonova et al. (2024) conducted their experiments at 27 °C, which also deviates from standard conditions and may complicate comparisons. Orfanos et al. (2015) suggested that mature sarcomere length is reached by approximately 88 hours after puparium formation (APF). In contrast, our measurements show that sarcomeres continue to elongate beyond this point, reaching mature length between 12 and 24 hours post-eclosion. All four earlier studies report a mature sarcomere length around 3.2-3.3 µm, only slightly longer than the ~3.2 µm length of thick filaments (Katzemich et al., 2012; Gasek et al., 2016). This would imply an I-band length below ~100 nm, which is an implausibly short distance. In contrast, our data, along with several recent studies (González-Morales et al., 2019; Deng et al., 2021; Dhanyasi et al., 2020; DeAguero et al., 2019), support a mature sarcomere length of approximately 3.45 µm, placing the length of the I-band at around 250 nm. This estimate is more consistent with high-resolution structural observations from longitudinal EM sections and fluorescent nanoscopy (Szikora et al., 2020; Schueder et al., 2023). Although Reedy and Beall (1993) provide limited data on myofibril diameter during myofibrillogenesis, a more detailed quantitative analysis is presented by Spletter et al. (2018) and by Nikonova et al. (2024). Interestingly, Spletter et al. report two separate datasets - one based on longitudinal sections and another on cross-sections of DLM fibers. While the measurements are consistent during early pupal stages, they diverge significantly in mature IFMs (1.116 ± 0.1025 µm vs. 1.428 ± 0.0995 µm), a discrepancy that is not addressed in their publication. Nikonova et al. (2024) report even narrower myofibril widths (0.9887 ± 0.1273 µm). Moreover, the reported diameters of early myofibrils in all three datasets are nearly twice as large as those reported by Reedy and Beall (1993) and in our own measurements, directly contradicting the reviewer's claim that the values are “close to identical.” Finally, our data clearly demonstrate that both the length and diameter of IFM sarcomeres reach a plateau in young adults, which is a key developmental feature not examined in previous studies.

In summary, we did not and we do not intend to claim that our conclusions are novel as to the general mechanisms of myofibril and sarcomere growth. Rather, our contribution lies in providing a high-precision, robust analysis of the growth process using a state-of-the-art toolkit, resulting in a comprehensive description that aligns with structural data obtained from TEM and dSTORM. We therefore believe that expert readers will recognize numerous valuable aspects of our approaches that will advance research in the field.

Counting the total number of thick filaments during myofibril development is nice, however, this also has been done (REEDY, M. C. & BEALL, C. 1993, PMID: 8253277). In this old study, the authors reported the amount of filament across one myofibril. How does this compare to the new data here counting all filaments? Unfortunatley, this is not discussed.

Indeed, the study by Reedy and Beall (1993) was primarily based on longitudinal DLM sections, which were used to estimate myofibril width and count the number of thick filaments on this lateral view images (e.g., ~15 thick filaments wide at 75 hours APF), but total thick filament numbers were not provided. While such data could theoretically be used to estimate the number of myofilaments per myofibril, these estimations would depend on the unverified assumption that the section includes the full width of the myofibril. Additionally, the study did not provide standard deviations or the number of measurements, limiting the interpretability and reproducibility of their findings. These points highlight the need for a more rigorous and quantitative approach. For these reasons, we chose to quantify myofilament number using cross-sections, providing more accurate and reliable assessments.

Besides the difference between the lateral versus cross sections, a direct comparison of our studies is further complicated by differences in the developmental time points and experimental conditions used. Reedy and Beall (1993) reports data from pupae aged 42, 60, 75 and 100 hours, as well as from adults, whereas we present data from 36, 48, and 72 hours APF, and from 24 hours after eclosion, which corresponds to approximately 124 hours APF. Moreover, their experiments were carried out at 23 °C, a temperature that somewhat slows down pupal development and results in adult eclosion at around 112 hours APF, as stated in their study. In contrast, our experiments were carried out at the more commonly used 25 °C, where adults typically emerge around 100 hours APF.

Collectively, these differences prevented meaningful comparisons between the two datasets, and therefore we preferred to avoid lengthy discussions on this issue.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2025-02953

Corresponding author(s): Andreas, Villunger

1. General Statements [optional]

*We would like to thank the reviewers for their constructive input and overall support. We appreciate to provide a provisional revision plan, as outlined here, and are happy to engage in additional communication with journal editors via video call, in case further clarifications are needed. *

2. Description of the planned revisions

Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.

Reviewer #1

__Evidence, reproducibility and clarity __

Summary: This manuscript by Leone et al describes the role of the PIDDosome in cardiomyocytes. Using a series of whole body and cardiomyocyte specific knockouts, the authors show that the PIDDosome maintains correct ploidy in these cells. It achieves this through inducing cell cycle arrest in cardiomyocytes in a p53 dependent manner. Despite this effect on ploidy, PIDDosome-deficient hearts show no structural or functional defects. Statistics and rigor appear to be adequate.

We thank this referee for taking the time to evaluate our work and their valuable comments. We assume that this reviewer by mistake indicates that the phenomenon we describe, depends on p53. As outlined in the abstract and throughout the manuscript, the effect is independent of p53, but may additionally still involve p21, acting along or parallel to the PIDDosome.

Major comments: 1. Figure 1 uses fluorescent intensity of a nuclear stain to determine ploidy per nucleus and they further separate the results into mononucleated, binucleated or multinucleated cells. It is hard to know how to interpret these results without further information or controls. Is there a good positive control that can be used to help to show whether this assay is quantitative? The differences are larger with the Raidd and caspase-2 knockouts than with the Pidd knockouts but this is not addressed.

*We appreciate this concern. Regarding a "good positive control" we can say that we follow state-of the art in the cardiomyocyte field and studies by the Evans (PMID: 36622904), Kuhn (PMID: 32109383), Bergmann (PMID: 26544945) and Patterson labs (PMID: 28783163, 36912240) all use the identical approach to discriminate 2n from 4n nuclei in microscopy images at the cellular level. The fact that the majority of rodent CM nuclei is indeed diploid (PMID: 31175264, 31585517 and 32078450) and a large number of nuclei has been evaluated to assess their mean fluorescence intensity (MFI) reduces the risk of a systematic bias in our analysis. Moreover, we have used an orthogonal approach that is indeed quantitative to define DNA content, i.e,. flow-cytometry based evaluation of DNA content in isolated CM nuclei (Fig. 1C). We hence are confident our assays are quantitative. *

Regarding the fact that loss of Pidd1 causes a more saddle phenotype, we can offer to discuss this in light of the fact that Pidd1 has additional functions, outside the PIDDosome (PMID: 35343572), and that we made similar observations when analyzing ploidy in hepatocytes (PMID: *31983631). Given the fact that all components of the PIDDosome show a similar phenotype, and that this phenotype is mimicked by loss of the protein that connects PIDD1 and centrosomes, ANKRD26 (Fig. 4a), we are confident that this biological variation in our analysis is not affecting our conclusions. *

On line 459 the authors state that the increase in polyploidy in PIDDosome knockouts occurs in adult hood but this is not directly tested. In fact, in the next section the polyploidy is assessed in early postnatal development. This statement should be explained or removed.

We see that we have made an unclear statement here. In fact, we first noted increases in ploidy in adult heart and then define the time window in development when this happens. This sentence will be rephrased.

In Figure 4. The authors obtained RNAseq data for P1, P7 and P14 but only show the differences with and without caspase-2 at P7. Given that the differences in ploidy are more significant at P14 (Fig 3D), all the comparisons should be shown along with analysis of whether the same genes/gene families are altered in the absence of caspase-2.

The reason why we focus on postnatal day 7 (P7) is that data from Alkass et al (PMID: 26544945) and other labs (PMID: 31175264 ) document that on this day the initial wave of binucleation peaks. Hence, we hypothesized that the PIDDosome must be active in most CM, which aligns well with the increased mRNA levels of all of its components (Figure 3). Interestingly, it seems that its action is tightly regulated, as mRNA of PIDDosome components drop on P10, suggesting PIDDosome shut-down or downregulation. Similar findings have been noted in the liver (PMID: *31983631). Alkass and colleagues also show that very few CMs enter another round of DNA synthesis between P7 and P14, and hence possible transcriptome changes in the absence of the PIDDosome will be strongly diluted. *

Please note that on P1, there is no difference between genotypes to be expected as all CM are mononucleated diploids and cytokinesis competent, as previously demonstrated (PMID: *26544945). Moreover, PIDDosome expression levels are extremely low (Fig. 3A). As such, no difference between genotypes are expected on P1. In addition, on P14 the ploidy phenotype observed in PIDDosome knockout mice reaches the maximum and ploidy increases are comparable to adult tissue. Thus, at this time the trigger for PIDDosome activation (cytokinesis failure) is no longer observed as the majority of CMs are post-mitotic, (PMID: 26247711). As such the impact of PIDDosome activation on the P14 transcriptome is most likely negligible. However, if desired, we can expand our bioinformatics analysis summarizing findings made related to DEGs over time in wt animals by comparing genotypes also on day 1 and day 14. In light of the above, analysis between genotypes on P7 holds still appears as the one most meaningful. *

Some validation of the RNAseq and/or proteomics results would be an important addition to this study

We agree with this notion and propose to validate key candidates related to cardiomyocyte proliferation and polyploidization, some of which we found to be differentially expressed at the mRNA level on day 7in the RNAseq data (e.g., p21, Foxm1, Kif18a, Lin37 and others)

Regarding the proteomics results, we face the challenge that we can only try to confirm if candidate proteins are likely caspase substrates in silico using DeepCleave*, and potentially pick one or two candidates linked to CM differentiation for further analysis in vitro and in heterologous cell based assays (e.g. 293T cells), as no bona-fide ventricular cardiomyocyte cell lines exist. Primary postnatal CMs are extremely difficult to transfect, nor they proliferate without drug-treatment, or fail cytokinesis ex vivo. *

Figure 4D: the authors make the conclusion that p21 is downstream of PIDD (et p53 independent). However, this is not supported by the data because the increase in 4N cells/decrease in 2N cells, although statistically significant, is nowhere near that of caspase-2 KO and caspase-2/p21 KO. Statistics should also compare p32KO with c2KO. In the absence of any other data, the more likely conclusion is that p21 is not involved.

*We agree that the findings related to the impact seen upon loss of p21 suggest that it is not the only effector involved in ploidy control and it may not even be an effector engaged by caspase-2, as C2/p21 DKO mice have an even higher ploidy increase, albeit not statistically significant. However, it is important to highlight that p21 (Cdkn1a) was found to be downregulated in our transcriptomic analysis suggesting an involvement in the caspace2-cascade. We are happy to highlight this when presenting the results and in the discussion. *

*We assume that this referee refers to p73 KO data that should be compared to Casp2 KO data (could be read as p73 or p53, but the latter we compare side by side with Casp2 in Fig. 4 already). As p73 KO mice were not found to be viable beyond day 7 (our attempt to find animals on day 10 failed, in line with published literature (PMID: 24500610, 10716451)), we can only offer to compare this data set to the data presented in Figure 3C, where we have analyzed ploidy increases on day 7 from wt and PIDDosome mutant mice. This re-analysis will show that only Caspase-2 mutant mice display a significant ploidy increase on P7, when compared to wt or p73 mutant animals, while no difference are noted between wt and p73 mutant mice (to be included in new Suppl. Fig. 3C) *

Minor comments: Suggest moving Figure 4A to Figure 3 as it seems to fit better there based on the citation of this figure in the text

*We can see some benefit in this recommendation and included panel 4A now in an updated version of Figure 3. *

Recommend enhancing the brightness of microscopy images in Figure 1E and 2D

We will try to improve image quality, may have been due to PDF conversion

Significance

This study provides interesting information for the role of the PIDDosome in protecting from polyploidy and adds to the body of work by this same group studying this pathway in the liver.

The main weakness in terms of significance is the lack of a phenotype in the hearts of these animals. Therefore, it is clear that ploidy (or at least PIDDosome dependent ploidy) has minimal impact on cardiac development.

We respectfully disagree with the comment that the lack of impact on cardiac function constitutes a weakness of our findings. Several studies on ploidy control in the liver (PMID 34228992) but importantly also heart (PMID: 36622904) have failed to document a clear impact of increased ploidy on organ function. This does not infer insignificance, but maybe rather that the context where this becomes relevant has not been identified. We are happy expand on this in our discussion

The authors mention that they have not tried giving these mice an myocardial infarct (MI) or inducing any other type of cardiac damage. Although it is understood that these experiments are likely outside of the scope of the present study, without this information the impact of this study is moderate. I recommend expanding the discussion to provide a more in-depth possible rationale as to why ploidy perturbations do not lead to structural changes like in the liver.

Despite this, the insights to the pathway itself are interesting to investigators in the caspase-2 field if a little underdeveloped, especially concerning the role of p21.

My expertise is in cell death and caspase biology (especially caspase-2). I have sufficient expertise to evaluate all parts of this paper.

*As mentioned above, we will amend our conclusions on p21, in light of potential findings made when validating DEG candidates, as stated above. *

*We hope that the changes and amendments proposed here will be satisfactory to this referee to recommend publication of a revised manuscript. *

Reviewer #2

__Evidence, reproducibility and clarity: __

__Summary: __

In this study, the authors investigated the role of the PIDDosome during cardiomyocyte polyploidization. PIDDosome is a multi-protein complex activating the endopeptidase Caspase-2, and shown to be involved in eliminating cells with extra centrosomes or in response to genotoxic stress (Burigotto & Fava, 2021, Sladky and Villunger, 2020). In both cases, the PIDDosome is recruited in a ANKRD26-dependent manner at the centrosomes leading to p53 stabilization and cell death (Burigotto & Fava, 2021; Evans et al., 2020; Burigotto et al., 2021).

Here, by studying mouse cardiomyocyte differentiation, the authors showed that PIDDosome is imposing ploidy restriction during cardiomyocyte differentiation. Importantly, in contrast to a previous report in the liver (Sladky et al., 2020), they showed that PIDDosome acts in a p53-independent manner in cardiomyocytes. Indeed, they suggested that PIDDosome controls ploidy in cardiomyocytes through p21 activation.

We want to thank this reviewer for the time taken to evaluate our work and provide critical feedback that will help to improve our revised manuscript.

__Major comments: __

In general the conclusions of the authors are well supported by the experiments. However, I would suggest the following experiments/analysis to strengthen the paper:

The authors should improve the Figure 1 to help the readers who are not familiar with cardiomyocyte polyploidization. For instance, I would suggest to add a scheme to summarize cardiomyocyte polyploidization (in terms of nuclear size, mono vs multi and so on).

We agree that a visual summary of the postnatal timing of CM polyploidization will be helpful for the generalist not familiar with the topic and have added a scheme, adapted from a study by Alkass et al. (PMID: *26544945), who elegantly defined the timing of this process during postnatal mice life (now Fig. 1A). *

Based on the images they presented in 1B, the authors should also measure the nuclear area or volume in the different conditions in which components of the PIDDosome were depleted. Indeed, these two parameters should be easier to conceptualize for the readers (instead of the fluorescence nuclear intensity). This could help to understand if the nuclear size is maintained between the different conditions and if this is comparable between mono, bi or multinucleated cardiomyocytes.

We have acquired this data and it can be used to provide additional information on nuclear area and/or volume. We propose to focus on re-analyzing data from wt, Casp2 and XMLC2CRE/Casp2f/f mice. The additional information can be included in Figures 1 & 2, respectively.

• In Figure 2A, the authors presented cross section of heart from animals showing that PIDDosome depletion has no effect on heart size. This is a surprising result since cardiomyocytes have higher ploidy levels and this could have an effect on their function. Since the importance of this observation, the authors should present a quantification of the heart size in the different conditions shown in Figure 2A.

We agree with this comment. We can measure the heart vs. body weight ratio or tibia length in adult Casp2-/- vs. WT (3 month old) in order to indirectly evaluate possible increases in CM size linked to increased ploidy.

Also, the percentage of cardiomyocytes presenting higher levels of ploidy seems quite low. The authors should discuss this point. In particular because this could explain the absence of consequences on heart size and function at steady state.

We agree with this conclusion and will expand on this in our discussion. It is important to note that as opposed to findings made in liver (PMID: *31983631), genetic manipulation of ploidy regulators such as E2f7/8 (PMID: 36622904), only led to modest changes in CM ploidy, suggesting that either a small band-width compatible with normal heart function exists, or that additional mechanisms exist that take control when these thresholds set by the PIDDosome or E2f7/8 are exceeded. These mechanisms could involve Cyclin G (PMID: 20360255), or TNNI3K (PMID: 31589606). Importantly, a recent publication has shown that overexpression of Plk1(T210D) and Ect2 from birth causes increased heart weight coupled with a minor decrease in CM size. These mice undergo to premature death (PMID: 39912233) suggesting that CM polyploidization is a tight regulated process regulated by several independent mechanisms during heart development. *

In Figure 2D, the authors measured the cardiomyocyte cross-sectional area and concluded that removing PIDDosome components have no effect on cardiomyocyte cell size. Since it has been shown that ploidy increase is normally associated with an increase in cell area, the authors should measure cell area of cardiomyocytes analyzed in Figure 1B. It could be then interesting to establish a correlation with nuclear area and the mono, bi or multinucleated status. This will strengthen the results showing that ploidy increases without affecting cell area.

Indeed, studies in PIDDosome deficient livers suggest that tissue is containing fewer but bigger cells (PMID: *31983631). As opposed to the liver the percentage of cardiomyocytes presenting higher levels of ploidy is relatively low. Thus, a possible increase in CM size in PIDDosome deficient mice may be masked in heart cross-sections. In order to better correlate the ploidy with cell size, we propose to reanalyze our microscopy images used to extract the data displayed in Fig. 1D. We may run into the problem though that the number of cells acquired may become limiting to achieve sufficient statistical power. In this case we could pool data from different PIDDosome mutant CM to increase statistical power. Again, we propose to initially prioritize wt vs. Casp2 vs. XMLC2/Casp2f/f mice. In addition, we can offer to quantify heart to body weight ratio or tibia length as an additional read-out (see answer to a previous reviewer comment). *

The authors should discuss the fact that PIDDosome depletion lead only to a mild increase in ploidy levels (4N) in a small percentage of cardiomyocyte. If the PIDDosome is controlling ploidy, one could expect that removing it should lead to a drastic increase in the ploidy levels. Is PIDDosome depletion leading to cell death in some cardiomyocyte? The authors should discuss this point in the discussion or if relevant show a staining with an apoptosis marker. Is another mechanism compensating to prevent higher ploidy levels in cardiomyocytes?

These are valid thoughts, some of which we contemplated before. In part, we have addressed them in our response to Reviewer#1, above, discussing similar findings made in E2f7/8 deficient hearts (PMID: 36622904), or Cyclin G overexpressing hearts (PMID: 20360255), where also only modest changes in ploidy were achieved. Together these observations are suggesting alternative control mechanism able to act, or limited tolerance towards larger shifts in ploidy, incompatible with proper cell function and survival. Towards this end, we can offer to test if we find increased signs of cell death in PIDDosome mutant hearts by TUNEL staining of histological sections. Of note, we did not find evidence for such a phenomenon in the liver (PMID: 31983631).

Even if the authors presented RNAseq data suggesting that the PIDDosome is activated during cardiomyocyte differentiation, they should clearly demonstrate this point to strengthen the message of the paper. Indeed, the conclusions are based on the absence of PIDDosome components triggering higher ploidy in cardiomyocytes. However, we don't know whether (and when) the PIDDosome is activated during cardiomyocyte differentiation to control their ploidy levels. I would suggest to analyze PIDDosome activation markers by immunofluorescence in *cardiomyocytes at different developmental stages. *

*We agree with this referee that direct proof of PIDDosome activation would be helpful and that we only infer back from loss of function phenotypes when and where the PIDDosome becomes activated. However, several technical issues prevent us from collecting more direct evidence of PIDDosome activation in the developing heart. 1) Polyploidization in heart CM appears to happen gradually in CM from day 3 on with a peak at day 7 (PMID: 26544945). Hence, this is not a synchronous process, where we could pinpoint simultaneous activation of the PIDDosome in all cells at the same time, which would facilitate biochemical analysis, e.g., by western blotting for signs of Caspase-2 activation (i.e. the loss of its pro-form, PMID: 28130345). 2) Our most reliable readout, MDM2 cleavage by caspase-2 giving rise to specific fragments detectable in western, is not applicable to mouse tissue, as the antibody we use only detects human MDM2 (PMID: 28130345) and no other MDM2 Ab we tested gave satisfactory results. Independent of that, 3) we do not see involvement of p53 in CM ploidy control (arguing against a role of MDM2). *

*As such, we can only offer to look at extra centrosome clustering in postnatal binucleated CM (as also suggested further below), as a putative trigger for PIDDosome activation. However, this has been published by the first author of this study before (PMID 31301302). Given that we have made the significant effort to time resolve the increase in ploidy in postnatal mice (please note that several hearts needed to be pooled for each time point, analyzed in multiple biological replicates), we think that our conclusions are well-justified based on the genetic data provided. *

Concerning the methods, the authors must add the references for each product they used and not only the origin. When relevant, the RRID should be indicated. Without this information the method and the data cannot be reproduced.

We will update this information where relevant to reproduce our results

Minor comments:

In general, the text and the figures are clear. Nevertheless, I would suggest the following changes:

• Figures 1B, 2B and 2C: the y-axis must start at 0.

We will adopt axes accordingly

Figure 4A: The authors should stain centrosomes in cardiomyocytes. This should strengthen the conclusion taken by the authors based on the results obtained in mice depleted for ANKRD26. Indeed, for the moment they are insufficient to conclude about the role of the centrosomes. The authors should show that centrosomes cluster in cardiomyocytes (a condition necessary for PIDDosome activation in polyploid cells) and if possible that component of the PIDDosome are recruited here.

*This point is well taken and addressed in part above. Clustering of extra centrosomes has been documented and published by the first author of this study in rat polyploid cardiomyocytes (PIMID; cited). We can offer to show clustering of centrosomes in mouse CM isolated from day 7 hearts, but while PIDD1 can be detected well in MEF, we repeatedly failed to stain fro PIDD1 in primary CMs. *

Figure 4F: I would suggest to modify the working model to emphasize more the differences between WT and PIDDosome KO.

We will aim to improve this cartoon/graphical abstract

The prior studies are referenced appropriately.

Reviewer #2 (Significance (Required)):

How polyploid cells control their ploidy levels during differentiation remains poorly understood. The data presented here represent thus an advance concerning this question. The actual model concerning PIDDosome activation relies on the presence of extra centrosomes that drives the ANKDR26-dependent recruitment of the PIDDosome. Then, Caspase 2 is activated leading to a p53-p21 dependent cell cycle arrest (Burigotto & Fava, 2021, Sladky and Villunger, 2020; Janssens & Tinel, 2012; Evans et al., 2020; Burigotto et al., 2021). In this study, the authors showed that similar pathway takes place during cardiomyocyte differentiation to control ploidy levels. These data are reminiscent of previous work showing PIDDosome involvement during hepatocyte polyploidization (Sladky et al. 2020). Together, these data highlight the prominent role of the PIDDosome complex in controlling ploidy levels in physiological context. Importantly, this study identified that the classical p53-dependent cell cycle arrest described after PIDDosome activation is not involved here. Instead, the data established that independently of p53, p21 contribute to control cardiomyocyte ploidy. In consequence, this study extends the initial pathway associated with PIDDosome activation and suggest that other mechanisms could take place to restrain cell proliferation upon PIDDosome activation. Overall, this makes this paper significant and of interest for the following fields: polyploidy, heart/cardiomyocyte development and PIDDosome.

My field of expertise includes polyploidy, cell cycle and genetic instability.

We thank this reviewer for the time taken and the positive feedback provided.

3. Description of the revisions that have already been incorporated in the transferred manuscript

Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. If no revisions have been carried out yet, please leave this section empty.

N/A

4. Description of analyses that authors prefer not to carry out

Please include a point-by-point response explaining why some of the requested data or additional analyses might not be necessary or cannot be provided within the scope of a revision. This can be due to time or resource limitations or in case of disagreement about the necessity of such additional data given the scope of the study. Please leave empty if not applicable.

• *As outlined above, limited tools are available to validate putative caspase-2 substrates, identified in proteomics analysis, in an impactful manner. *
• *Also, as discussed above, we deem myocardial infarction experiments in mice as unsuitable to improve our work, as with all likely-hood, they will yield negative results. *

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2025-02953

Corresponding author(s): Andreas, Villunger

[The “revision plan” should delineate the revisions that authors intend to carry out in response to the points raised by the referees. It also provides the authors with the opportunity to explain their view of the paper and of the referee reports.

• *

The document is important for the editors of affiliate journals when they make a first decision on the transferred manuscript. It will also be useful to readers of the reprint and help them to obtain a balanced view of the paper.

• *

If you wish to submit a full revision, please use our "Full Revision" template. It is important to use the appropriate template to clearly inform the editors of your intentions.]

1. General Statements [optional]

*We would like to thank the reviewers for their constructive input and overall support. We appreciate to provide a provisional revision plan, as outlined here, and are happy to engage in additional communication with journal editors via video call, in case further clarifications are needed. *

2. Description of the planned revisions

Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.

• *

Reviewer #1

__Evidence, reproducibility and clarity __

Summary: This manuscript by Leone et al describes the role of the PIDDosome in cardiomyocytes. Using a series of whole body and cardiomyocyte specific knockouts, the authors show that the PIDDosome maintains correct ploidy in these cells. It achieves this through inducing cell cycle arrest in cardiomyocytes in a p53 dependent manner. Despite this effect on ploidy, PIDDosome-deficient hearts show no structural or functional defects. Statistics and rigor appear to be adequate.

We thank this referee for taking the time to evaluate our work and their valuable comments. We assume that this reviewer by mistake indicates that the phenomenon we describe, depends on p53. As outlined in the abstract and throughout the manuscript, the effect is independent of p53, but may additionally still involve p21, acting along or parallel to the PIDDosome.

Major comments: 1. Figure 1 uses fluorescent intensity of a nuclear stain to determine ploidy per nucleus and they further separate the results into mononucleated, binucleated or multinucleated cells. It is hard to know how to interpret these results without further information or controls. Is there a good positive control that can be used to help to show whether this assay is quantitative? The differences are larger with the Raidd and caspase-2 knockouts than with the Pidd knockouts but this is not addressed.

*We appreciate this concern. Regarding a “good positive control” we can say that we follow state-of the art in the cardiomyocyte field and studies by the Evans (PMID: 36622904), Kuhn (PMID: 32109383), Bergmann (PMID: 26544945) and Patterson labs (PMID: 28783163, 36912240) all use the identical approach to discriminate 2n from 4n nuclei in microscopy images at the cellular level. The fact that the majority of rodent CM nuclei is indeed diploid (PMID: 31175264, 31585517 and 32078450) and a large number of nuclei has been evaluated to assess their mean fluorescence intensity (MFI) reduces the risk of a systematic bias in our analysis. Moreover, we have used an orthogonal approach that is indeed quantitative to define DNA content, i.e,. flow-cytometry based evaluation of DNA content in isolated CM nuclei (Fig. 1C). We hence are confident our assays are quantitative. *

Regarding the fact that loss of Pidd1 causes a more saddle phenotype, we can offer to discuss this in light of the fact that Pidd1 has additional functions, outside the PIDDosome (PMID: 35343572), and that we made similar observations when analyzing ploidy in hepatocytes (PMID: *31983631). Given the fact that all components of the PIDDosome show a similar phenotype, and that this phenotype is mimicked by loss of the protein that connects PIDD1 and centrosomes, ANKRD26 (Fig. 4a), we are confident that this biological variation in our analysis is not affecting our conclusions. *

On line 459 the authors state that the increase in polyploidy in PIDDosome knockouts occurs in adult hood but this is not directly tested. In fact, in the next section the polyploidy is assessed in early postnatal development. This statement should be explained or removed.

We see that we have made an unclear statement here. In fact, we first noted increases in ploidy in adult heart and then define the time window in development when this happens. This sentence will be rephrased.

In Figure 4. The authors obtained RNAseq data for P1, P7 and P14 but only show the differences with and without caspase-2 at P7. Given that the differences in ploidy are more significant at P14 (Fig 3D), all the comparisons should be shown along with analysis of whether the same genes/gene families are altered in the absence of caspase-2.

The reason why we focus on postnatal day 7 (P7) is that data from Alkass et al (PMID: 26544945) and other labs (PMID: 31175264 ) document that on this day the initial wave of binucleation peaks. Hence, we hypothesized that the PIDDosome must be active in most CM, which aligns well with the increased mRNA levels of all of its components (Figure 3). Interestingly, it seems that its action is tightly regulated, as mRNA of PIDDosome components drop on P10, suggesting PIDDosome shut-down or downregulation. Similar findings have been noted in the liver (PMID: *31983631). Alkass and colleagues also show that very few CMs enter another round of DNA synthesis between P7 and P14, and hence possible transcriptome changes in the absence of the PIDDosome will be strongly diluted. *

Please note that on P1, there is no difference between genotypes to be expected as all CM are mononucleated diploids and cytokinesis competent, as previously demonstrated (PMID: *26544945). Moreover, PIDDosome expression levels are extremely low (Fig. 3A). As such, no difference between genotypes are expected on P1. In addition, on P14 the ploidy phenotype observed in PIDDosome knockout mice reaches the maximum and ploidy increases are comparable to adult tissue. Thus, at this time the trigger for PIDDosome activation (cytokinesis failure) is no longer observed as the majority of CMs are post-mitotic, (PMID: 26247711). As such the impact of PIDDosome activation on the P14 transcriptome is most likely negligible. However, if desired, we can expand our bioinformatics analysis summarizing findings made related to DEGs over time in wt animals by comparing genotypes also on day 1 and day 14. In light of the above, analysis between genotypes on P7 holds still appears as the one most meaningful. *

Some validation of the RNAseq and/or proteomics results would be an important addition to this study

We agree with this notion and propose to validate key candidates related to cardiomyocyte proliferation and polyploidization, some of which we found to be differentially expressed at the mRNA level on day 7in the RNAseq data (e.g., p21, Foxm1, Kif18a, Lin37 and others)

Regarding the proteomics results, we face the challenge that we can only try to confirm if candidate proteins are likely caspase substrates in silico using DeepCleave*, and potentially pick one or two candidates linked to CM differentiation for further analysis in vitro and in heterologous cell based assays (e.g. 293T cells), as no bona-fide ventricular cardiomyocyte cell lines exist. Primary postnatal CMs are extremely difficult to transfect, nor they proliferate without drug-treatment, or fail cytokinesis ex vivo. *

Figure 4D: the authors make the conclusion that p21 is downstream of PIDD (et p53 independent). However, this is not supported by the data because the increase in 4N cells/decrease in 2N cells, although statistically significant, is nowhere near that of caspase-2 KO and caspase-2/p21 KO. Statistics should also compare p32KO with c2KO. In the absence of any other data, the more likely conclusion is that p21 is not involved.

*We agree that the findings related to the impact seen upon loss of p21 suggest that it is not the only effector involved in ploidy control and it may not even be an effector engaged by caspase-2, as C2/p21 DKO mice have an even higher ploidy increase, albeit not statistically significant. However, it is important to highlight that p21 (Cdkn1a) was found to be downregulated in our transcriptomic analysis suggesting an involvement in the caspace2-cascade. We are happy to highlight this when presenting the results and in the discussion. *

*We assume that this referee refers to p73 KO data that should be compared to Casp2 KO data (could be read as p73 or p53, but the latter we compare side by side with Casp2 in Fig. 4 already). As p73 KO mice were not found to be viable beyond day 7 (our attempt to find animals on day 10 failed, in line with published literature (PMID: 24500610, 10716451)), we can only offer to compare this data set to the data presented in Figure 3C, where we have analyzed ploidy increases on day 7 from wt and PIDDosome mutant mice. This re-analysis will show that only Caspase-2 mutant mice display a significant ploidy increase on P7, when compared to wt or p73 mutant animals, while no difference are noted between wt and p73 mutant mice (to be included in new Suppl. Fig. 3C) *

Minor comments: Suggest moving Figure 4A to Figure 3 as it seems to fit better there based on the citation of this figure in the text

*We can see some benefit in this recommendation and included panel 4A now in an updated version of Figure 3. *

Recommend enhancing the brightness of microscopy images in Figure 1E and 2D

We will try to improve image quality, may have been due to PDF conversion

Significance

This study provides interesting information for the role of the PIDDosome in protecting from polyploidy and adds to the body of work by this same group studying this pathway in the liver.

The main weakness in terms of significance is the lack of a phenotype in the hearts of these animals. Therefore, it is clear that ploidy (or at least PIDDosome dependent ploidy) has minimal impact on cardiac development.

We respectfully disagree with the comment that the lack of impact on cardiac function constitutes a weakness of our findings. Several studies on ploidy control in the liver (PMID 34228992) but importantly also heart (PMID: 36622904) have failed to document a clear impact of increased ploidy on organ function. This does not infer insignificance, but maybe rather that the context where this becomes relevant has not been identified. We are happy expand on this in our discussion

• *

The authors mention that they have not tried giving these mice an myocardial infarct (MI) or inducing any other type of cardiac damage. Although it is understood that these experiments are likely outside of the scope of the present study, without this information the impact of this study is moderate. I recommend expanding the discussion to provide a more in-depth possible rationale as to why ploidy perturbations do not lead to structural changes like in the liver.

Despite this, the insights to the pathway itself are interesting to investigators in the caspase-2 field if a little underdeveloped, especially concerning the role of p21.

My expertise is in cell death and caspase biology (especially caspase-2). I have sufficient expertise to evaluate all parts of this paper.

*As mentioned above, we will amend our conclusions on p21, in light of potential findings made when validating DEG candidates, as stated above. *

*We hope that the changes and amendments proposed here will be satisfactory to this referee to recommend publication of a revised manuscript. *

• *

Reviewer #2

__Evidence, reproducibility and clarity: __

__Summary: __

In this study, the authors investigated the role of the PIDDosome during cardiomyocyte polyploidization. PIDDosome is a multi-protein complex activating the endopeptidase Caspase-2, and shown to be involved in eliminating cells with extra centrosomes or in response to genotoxic stress (Burigotto & Fava, 2021, Sladky and Villunger, 2020). In both cases, the PIDDosome is recruited in a ANKRD26-dependent manner at the centrosomes leading to p53 stabilization and cell death (Burigotto & Fava, 2021; Evans et al., 2020; Burigotto et al., 2021).

Here, by studying mouse cardiomyocyte differentiation, the authors showed that PIDDosome is imposing ploidy restriction during cardiomyocyte differentiation. Importantly, in contrast to a previous report in the liver (Sladky et al., 2020), they showed that PIDDosome acts in a p53-independent manner in cardiomyocytes. Indeed, they suggested that PIDDosome controls ploidy in cardiomyocytes through p21 activation.

We want to thank this reviewer for the time taken to evaluate our work and provide critical feedback that will help to improve our revised manuscript.

__Major comments: __

In general the conclusions of the authors are well supported by the experiments. However, I would suggest the following experiments/analysis to strengthen the paper:

The authors should improve the Figure 1 to help the readers who are not familiar with cardiomyocyte polyploidization. For instance, I would suggest to add a scheme to summarize cardiomyocyte polyploidization (in terms of nuclear size, mono vs multi and so on).

We agree that a visual summary of the postnatal timing of CM polyploidization will be helpful for the generalist not familiar with the topic and have added a scheme, adapted from a study by Alkass et al. (PMID: *26544945), who elegantly defined the timing of this process during postnatal mice life (now Fig. 1A). *

Based on the images they presented in 1B, the authors should also measure the nuclear area or volume in the different conditions in which components of the PIDDosome were depleted. Indeed, these two parameters should be easier to conceptualize for the readers (instead of the fluorescence nuclear intensity). This could help to understand if the nuclear size is maintained between the different conditions and if this is comparable between mono, bi or multinucleated cardiomyocytes.

We have acquired this data and it can be used to provide additional information on nuclear area and/or volume. We propose to focus on re-analyzing data from wt, Casp2 and XMLC2CRE/Casp2f/f mice. The additional information can be included in Figures 1 & 2, respectively.

• In Figure 2A, the authors presented cross section of heart from animals showing that PIDDosome depletion has no effect on heart size. This is a surprising result since cardiomyocytes have higher ploidy levels and this could have an effect on their function. Since the importance of this observation, the authors should present a quantification of the heart size in the different conditions shown in Figure 2A.

We agree with this comment. We can measure the heart vs. body weight ratio or tibia length in adult Casp2-/- vs. WT (3 month old) in order to indirectly evaluate possible increases in CM size linked to increased ploidy.

Also, the percentage of cardiomyocytes presenting higher levels of ploidy seems quite low. The authors should discuss this point. In particular because this could explain the absence of consequences on heart size and function at steady state.

We agree with this conclusion and will expand on this in our discussion. It is important to note that as opposed to findings made in liver (PMID: *31983631), genetic manipulation of ploidy regulators such as E2f7/8 (PMID: 36622904), only led to modest changes in CM ploidy, suggesting that either a small band-width compatible with normal heart function exists, or that additional mechanisms exist that take control when these thresholds set by the PIDDosome or E2f7/8 are exceeded. These mechanisms could involve Cyclin G (PMID: 20360255), or TNNI3K (PMID: 31589606). Importantly, a recent publication has shown that overexpression of Plk1(T210D) and Ect2 from birth causes increased heart weight coupled with a minor decrease in CM size. These mice undergo to premature death (PMID: 39912233) suggesting that CM polyploidization is a tight regulated process regulated by several independent mechanisms during heart development. *

• *

In Figure 2D, the authors measured the cardiomyocyte cross-sectional area and concluded that removing PIDDosome components have no effect on cardiomyocyte cell size. Since it has been shown that ploidy increase is normally associated with an increase in cell area, the authors should measure cell area of cardiomyocytes analyzed in Figure 1B. It could be then interesting to establish a correlation with nuclear area and the mono, bi or multinucleated status. This will strengthen the results showing that ploidy increases without affecting cell area.

Indeed, studies in PIDDosome deficient livers suggest that tissue is containing fewer but bigger cells (PMID: *31983631). As opposed to the liver the percentage of cardiomyocytes presenting higher levels of ploidy is relatively low. Thus, a possible increase in CM size in PIDDosome deficient mice may be masked in heart cross-sections. In order to better correlate the ploidy with cell size, we propose to reanalyze our microscopy images used to extract the data displayed in Fig. 1D. We may run into the problem though that the number of cells acquired may become limiting to achieve sufficient statistical power. In this case we could pool data from different PIDDosome mutant CM to increase statistical power. Again, we propose to initially prioritize wt vs. Casp2 vs. XMLC2/Casp2f/f mice. In addition, we can offer to quantify heart to body weight ratio or tibia length as an additional read-out (see answer to a previous reviewer comment). *

The authors should discuss the fact that PIDDosome depletion lead only to a mild increase in ploidy levels (4N) in a small percentage of cardiomyocyte. If the PIDDosome is controlling ploidy, one could expect that removing it should lead to a drastic increase in the ploidy levels. Is PIDDosome depletion leading to cell death in some cardiomyocyte? The authors should discuss this point in the discussion or if relevant show a staining with an apoptosis marker. Is another mechanism compensating to prevent higher ploidy levels in cardiomyocytes?

These are valid thoughts, some of which we contemplated before. In part, we have addressed them in our response to Reviewer#1, above, discussing similar findings made in E2f7/8 deficient hearts (PMID: 36622904), or Cyclin G overexpressing hearts (PMID: 20360255), where also only modest changes in ploidy were achieved. Together these observations are suggesting alternative control mechanism able to act, or limited tolerance towards larger shifts in ploidy, incompatible with proper cell function and survival. Towards this end, we can offer to test if we find increased signs of cell death in PIDDosome mutant hearts by TUNEL staining of histological sections. Of note, we did not find evidence for such a phenomenon in the liver (PMID: 31983631).

Even if the authors presented RNAseq data suggesting that the PIDDosome is activated during cardiomyocyte differentiation, they should clearly demonstrate this point to strengthen the message of the paper. Indeed, the conclusions are based on the absence of PIDDosome components triggering higher ploidy in cardiomyocytes. However, we don't know whether (and when) the PIDDosome is activated during cardiomyocyte differentiation to control their ploidy levels. I would suggest to analyze PIDDosome activation markers by immunofluorescence in *cardiomyocytes at different developmental stages. *

*We agree with this referee that direct proof of PIDDosome activation would be helpful and that we only infer back from loss of function phenotypes when and where the PIDDosome becomes activated. However, several technical issues prevent us from collecting more direct evidence of PIDDosome activation in the developing heart. 1) Polyploidization in heart CM appears to happen gradually in CM from day 3 on with a peak at day 7 (PMID: 26544945). Hence, this is not a synchronous process, where we could pinpoint simultaneous activation of the PIDDosome in all cells at the same time, which would facilitate biochemical analysis, e.g., by western blotting for signs of Caspase-2 activation (i.e. the loss of its pro-form, PMID: 28130345). 2) Our most reliable readout, MDM2 cleavage by caspase-2 giving rise to specific fragments detectable in western, is not applicable to mouse tissue, as the antibody we use only detects human MDM2 (PMID: 28130345) and no other MDM2 Ab we tested gave satisfactory results. Independent of that, 3) we do not see involvement of p53 in CM ploidy control (arguing against a role of MDM2). *

*As such, we can only offer to look at extra centrosome clustering in postnatal binucleated CM (as also suggested further below), as a putative trigger for PIDDosome activation. However, this has been published by the first author of this study before (PMID 31301302). Given that we have made the significant effort to time resolve the increase in ploidy in postnatal mice (please note that several hearts needed to be pooled for each time point, analyzed in multiple biological replicates), we think that our conclusions are well-justified based on the genetic data provided. *

Concerning the methods, the authors must add the references for each product they used and not only the origin. When relevant, the RRID should be indicated. Without this information the method and the data cannot be reproduced.

We will update this information where relevant to reproduce our results

Minor comments:

In general, the text and the figures are clear. Nevertheless, I would suggest the following changes:

• Figures 1B, 2B and 2C: the y-axis must start at 0.

We will adopt axes accordingly

Figure 4A: The authors should stain centrosomes in cardiomyocytes. This should strengthen the conclusion taken by the authors based on the results obtained in mice depleted for ANKRD26. Indeed, for the moment they are insufficient to conclude about the role of the centrosomes. The authors should show that centrosomes cluster in cardiomyocytes (a condition necessary for PIDDosome activation in polyploid cells) and if possible that component of the PIDDosome are recruited here.

*This point is well taken and addressed in part above. Clustering of extra centrosomes has been documented and published by the first author of this study in rat polyploid cardiomyocytes (PIMID; cited). We can offer to show clustering of centrosomes in mouse CM isolated from day 7 hearts, but while PIDD1 can be detected well in MEF, we repeatedly failed to stain fro PIDD1 in primary CMs. *

Figure 4F: I would suggest to modify the working model to emphasize more the differences between WT and PIDDosome KO.

We will aim to improve this cartoon/graphical abstract

The prior studies are referenced appropriately.

Reviewer #2 (Significance (Required)):

How polyploid cells control their ploidy levels during differentiation remains poorly understood. The data presented here represent thus an advance concerning this question. The actual model concerning PIDDosome activation relies on the presence of extra centrosomes that drives the ANKDR26-dependent recruitment of the PIDDosome. Then, Caspase 2 is activated leading to a p53-p21 dependent cell cycle arrest (Burigotto & Fava, 2021, Sladky and Villunger, 2020; Janssens & Tinel, 2012; Evans et al., 2020; Burigotto et al., 2021). In this study, the authors showed that similar pathway takes place during cardiomyocyte differentiation to control ploidy levels. These data are reminiscent of previous work showing PIDDosome involvement during hepatocyte polyploidization (Sladky et al. 2020). Together, these data highlight the prominent role of the PIDDosome complex in controlling ploidy levels in physiological context. Importantly, this study identified that the classical p53-dependent cell cycle arrest described after PIDDosome activation is not involved here. Instead, the data established that independently of p53, p21 contribute to control cardiomyocyte ploidy. In consequence, this study extends the initial pathway associated with PIDDosome activation and suggest that other mechanisms could take place to restrain cell proliferation upon PIDDosome activation. Overall, this makes this paper significant and of interest for the following fields: polyploidy, heart/cardiomyocyte development and PIDDosome.

My field of expertise includes polyploidy, cell cycle and genetic instability.

We thank this reviewer for the time taken and the positive feedback provided.

• *

• *

3. Description of the revisions that have already been incorporated in the transferred manuscript

Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. If no revisions have been carried out yet, please leave this section empty.

• *

N/A

4. Description of analyses that authors prefer not to carry out

Please include a point-by-point response explaining why some of the requested data or additional analyses might not be necessary or cannot be provided within the scope of a revision. This can be due to time or resource limitations or in case of disagreement about the necessity of such additional data given the scope of the study. Please leave empty if not applicable.

• *

• *As outlined above, limited tools are available to validate putative caspase-2 substrates, identified in proteomics analysis, in an impactful manner. *

• *Also, as discussed above, we deem myocardial infarction experiments in mice as unsuitable to improve our work, as with all likely-hood, they will yield negative results. *

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This study investigates the potential of targeting specific regions within the RNA genome of the Porcine Epidemic Diarrhea Virus (PEDV) for antiviral drug development. The authors used SHAPE-MaP to analyze the structure of the PEDV RNA genome in infected cells. They categorized different regions of the genome based on their structural characteristics, focusing on those that might be good targets for drugs or small interfering RNAs (siRNAs).

They found that dynamic single-stranded regions can be stabilized by compounds (e.g., to form G-quadruplexes), which inhibit viral proliferation. They demonstrated this by targeting a specific G4-forming sequence with a compound called Braco-19. The authors also describe stable (structured) single-stranded regions that they used to design siRNAs showing that they effectively inhibited viral replication.

Strengths:

There are a number of strengths to highlight in this manuscript.

(1) The study uses a sophisticated technique (SHAPE-MaP) to analyze the PEDV RNA genome in situ, providing valuable insights into its structural features.

(2) The authors provide a strong rationale for targeting specific RNA structures for antiviral development.

(3) The study includes a range of experiments, including structural analysis, compound screening, siRNA design, and viral proliferation assays, to support their conclusions.

(4) Finally, the findings have potential implications for the development of new antiviral therapies against PEDV and other RNA viruses.

Overall, this interesting study highlights the importance of considering RNA structure when designing antiviral therapies and provides a compelling strategy for identifying promising RNA targets in viral genomes.

Weaknesses:

I have some concerns about the utility of the 3D analyses, the effects of their synonymous mutants on expression/proliferation, a potentially missed control for studies of mutants, and the therapeutic utility of the compound they tested vs. Gquadruplexes.

We thank the reviewer for their positive assessment and insightful comments. Below, we address each point of concern:

(1) The utility of the 3D analyses:

In the revised manuscript, we have toned down this discussion and moved Figure 3A to the supplementary materials to reduce any sense of fragmentation in the overall story. While SHAPE-MaP technology is mature and convenient to use and can indeed capture some RNA structural elements with special functions in certain case; we acknowledge that its application for 3D analyses requires further validation. We believe this approach will become more prevalent in future research.

(2) The effects of synonymous mutants on expression/proliferation:

In the PEDV genome, the PQS1 mutation site encodes lysine (AAG). Given that lysine has only two codons (AAG and AAA), the G3109A synonymous mutation represented our sole viable option. Published studies (Ding et al., 2024) confirm that neither AAG nor AAA are classified as rare or dominant codons in mammalian cells. Therefore, the observed changes in viral proliferation levels are likely to stem from alterations in RNA secondary structure rather than codon usage effects.

REFERENCES:

Ding W, Yu W, Chen Y, et al. Rare codon recoding for efficient noncanonical amino acid incorporation in mammalian cells. Science. 2024;384(6700):1134-1142. 

(3) Potentially missed control for studies of mutants:

In the revised manuscript, we have incorporated additional control experiments evaluating Braco-19's therapeutic effects on the PQS3 mutant strain (Figure 4 – figure supplement 3)：

(4) The therapeutic utility of Braco-19 vs. G-quadruplexes:

While Braco-19 is indeed a broad-spectrum G4 ligand, our data clearly show that not all PQSs in the viral genome can form G4 structures. Our findings primarily provide proof-of-concept that sequences with high G4-forming potential in viral genomes represent viable targets for antiviral therapy. Future studies could leverage SHAPEguided structural insights to design ligands with enhanced specificity for viral G4s, potentially improving therapeutic utility while minimizing off-target effects.

Reviewer #2 (Public review):

Summary:

Luo et. al. use SHAPE-MaP to find suitable RNA targets in Porcine Epidemic Diarrhoea Virus. Results show that dynamic and transient structures are good targets for small molecules, and that exposed strand regions are adequate targets for siRNA. This work is important to segment the RNA targeting.

Strengths:

This work is well done and the data supports its findings and conclusions. When possible, more than one technique was used to confirm some of the findings.

Weaknesses:

The study uses a cell line that is not porcine (not the natural target of the virus).

We thank the reviewer for their insightful comments and recognition of our study's value. The most commonly employed cell models for in vitro PEDV studies are monkey-derived Vero E6 cells and porcine PK1 cells. However, PEDV (particularly our strain) exhibits significantly lower replication efficiency in PK1 cells compared to Vero cells, and no cytopathic effects were observed in PK1 cells. In our preliminary attempts to perform SHAPE-MaP experiments using infected PK1 cells, the sequencing data showed less than 0.03% alignment to the PEDV genome, rendering subsequent analysis and downstream experiments unfeasible.

Reviewer #3 (Public review):

Summary:

This manuscript by Luo et al. applied SHAPE-Map to analyze the secondary structure of the Porcine Epidemic Diarrhoea Virus (PEDV) RNA genome in infected cells. By combining SHAPE reactivity and Shannon entropy, the study indicated that the folding of the PEDV genomic RNA was nonuniform, with the 5' and 3' untranslated regions being more compactly structured, which revealed potentially antiviral targetable RNA regions. Interestingly, the study also suggested that compounds bound to well-folded RNA structures in vitro did not necessarily exhibit antiviral activity in cells, because the binding of these compounds did not necessarily alter the functions of the well-folded RNA regions. Later in the manuscript, the authors focus on guanine-rich regions, which may form G-quadruplexes and be potential targets for small interfering RNA (siRNA). The manuscript shows the binding effect of Braco-19 (a G-quadruplex-binding ligand) to a predicted G4 region in vitro, along with the inhibition of PEDV proliferation in cells. This suggests that targeting high SHAPE-high Shannon G4 regions could be a promising approach against RNA viruses. Lastly, the manuscript identifies 73 singlestranded regions with high SHAPE and low Shannon entropy, which demonstrated high success in antiviral siRNA targeting.

Strengths:

The paper presents valuable data for the community. Additionally, the experimental design and data analysis are well documented.

Weakness:

The manuscript presents the effect of Braco-19 on PQS1, a single G4 region with high SHAPE and high Shannon entropy, to suggest that "the compound can selectively target the PQS1 of the high SHAPE-high Shannon region in cells" (lines 625-626). While the effect of Braco-19 on PQS1 is supported by strong evidence in the manuscript, the conclusion regarding the G4 region with high SHAPE and high Shannon entropy is based on a single target, PQS1.

We thank the reviewer for their positive assessment of our methodology and dataset. We propose that dynamic RNA structures in high SHAPE-high Shannon regions, when stabilized by small molecules, can serve as viable targets for antiviral therapy. Gquadruplexes represent a characteristic type of such dynamic structures that compete with local stem-loop formations in the genome. While we identified seven highly conserved PQSs in the PEDV genome, only PQS1 was located within a high SHAPEhigh Shannon region. To further validate this concept, we have supplemented the revised manuscript with Thioflavin T (ThT) fluorescence turn-on assays (Figures 3D, 3E, and Figure 3 – figure supplement 6), which provide additional evidence for the differential G4-forming capabilities of PQSs across regions with distinct structural features.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Major Comments:

(1) It could be valuable for the authors to spend some more effort comparing their approach to siRNA target discovery and design to current methods for siRNA design. It would be good to highlight which components are novel, and which might offer superior performance with respect to other existing methods.

We thank the reviewer for highlighting this important point. In response, we have rewritten the relevant section in the discussion:

“Our approach uniquely integrates in situ RNA structural data (SHAPE reactivity and Shannon entropy) to prioritize siRNA targets within stable single-stranded regions (high SHAPE reactivity, low Shannon entropy), which are experimentally validated as accessible in infected cells. This represents a significant departure from traditional siRNA design methods that rely primarily on sequence conservation, thermodynamic rules (e.g., Tuschl rules), or in vitro structural predictions (Ali Zaidi et al., 2023; Qureshi et al., 2018; Tang and Khvorova, 2024),which may not accurately reflect intracellular RNA accessibility. Bowden-Reid et al. designed 39 antiviral siRNAs against various SARS-CoV-2 variants based on sequence conservation, ultimately identifying 8 highly effective sequences (Bowden-Reid et al., 2023). Notably, five of these effective sequences targeted regions that were located in high SHAPE-high Shannon regions according to SARS-CoV-2 SHAPE datasets (Supplementary Table 8) (Manfredonia et al., 2020). This independent finding aligns perfectly with our conclusions and demonstrates that SHAPE-based siRNA design outperforms sequence/structureagnostic approaches, at least in terms of significantly improving antiviral siRNA screening efficiency. Given the growing availability of SHAPE datasets for numerous viruses, we are confident that our methodology will facilitate more precise design of antiviral siRNAs.”

(2) The section targeting their discovered G4 structure with Braco-19 is interesting, particularly showing effects on viral proliferation; however, it's not clear to me how this compound could be used therapeutically against PEDV, as it is a non-selective binder of G4 structures. Their results are good support for the presence and functionality of a G4 structure in PEDV, but I don't see any strategy outlined in the manuscript on how this could be specifically targeted with Braco-19.

While Braco-19 is indeed a broad-spectrum G4 ligand, our data demonstrate that not all PQSs in the viral genome can form G4 structures under physiological conditions. Our results specifically show that Braco-19 exerts its anti-PEDV activity by targeting PQS1, which is located in a high SHAPE-high Shannon entropy region. This target specificity was further confirmed by the complete resistance of the PQS1mut strain (lacking G4-forming ability) to Braco-19 treatment in our in vitro assays. 

Additionally, previous studies have reported that during rapid viral replication, viral RNA accumulates to levels that significantly exceed host RNA concentrations. This "concentration advantage" suggests that G4 ligands like Braco-19 would preferentially bind viral G4 structures over host targets, thereby enhancing their antiviral specificity in vivo. In summary, our data provide proof-of-concept that viral genomic regions with high G4-forming potential - particularly those in high SHAPE-high Shannon entropy regions - represent promising targets for antiviral therapy.

(3) The section where they proposed 3D RNA structures based on sequence similarity feels "tacked on" and I don't see how it adds to the overall story. The authors identify a short RNA hairpin in the PEDV genome with some sequence similarity to the CPEB3 nuclease P4 hairpin. However, they don't provide any evidence that this motif functions in a similar way or that it's important for the virus's life cycle. They also don't explain how this similarity could be exploited for antiviral drug development. It's not clear whether targeting this motif would have any effect on the virus. It's interesting that these two sequences share nucleotides, but it's unlikely that they share any homology...perhaps they convergently evolved (or were captured), but the similarity could also be coincidental.

We appreciate the reviewer's insightful observation regarding this section. While our intention was to demonstrate that flexible conformations in high SHAPE-high Shannon regions could potentially be targeted, we acknowledge that extensive discussion of these motifs' functions would exceed the scope of this study, resulting in some disconnection from the main narrative. In response to this valuable feedback, we have consequentially removed it from the manuscript.

(4) The authors should consider the optimality of the synonymous mutation (G3109A) that they introduced, as G3109A could swap a rare codon for a more optimal one. Even though the protein sequence is unaffected, the translation rate (and ability to proliferate) could be very different due to altered codon optimality. Additionally, to show the inactivity of the PQS3 mutant, the Braco-19 treatment studies performed on the PQS1 mutants could be repeated with PQS3 - using this as a control for these experiments.

We appreciate the reviewer's insightful comment regarding codon optimization. In the PEDV genome, the PQS1 mutation site encodes lysine (AAG). Since lysine has only two codons (AAG and AAA), the G3109A synonymous mutation was our only viable option. Published literature (Ding et al. 2024) confirms that neither AAG nor AAA are classified as either preferred or rare codons in mammalian cells. Therefore, this substitution should have minimal direct impact on translation efficiency. Compared to nonsynonymous mutations that would alter amino acid sequences, we believe this synonymous mutation represents the optimal approach for maintaining native protein function while introducing the desired structural modification.

REFERENCES:

Ding W, Yu W, Chen Y, et al. Rare codon recoding for efficient noncanonical amino acid incorporation in mammalian cells. Science. 2024;384(6700):1134-1142.

In the revised version, we have added control experiments showing the inhibitory activity of Braco-19 against the PQS3 mutant strain (Figure 4—figure supplement 3C) and discussed it in the results section.

“Furthermore, as a control, we observed nearly identical inhibitory activity of Braco19 against both the PQS3 mutant strain (AJ1102-PQS3mut) and wild-type virus (Figure 4—figure supplement 3C), demonstrating the specificity of Braco-19's action on PQS1.”

Minor Comments:

(5) The authors' description of the Shannon Entropy could be improved. The current description makes it seem like the Shannon Entropy only provides information on base pairing, however, the Shannon entropy quantifies the uncertainty of structural states at each position and is calculated based on the probabilities of the different states (paired or unpaired) that a nucleotide can adopt.

We have revised the description of Shannon entropy in the manuscript：

"The pairing probability of each nucleotide derived from SHAPE reactivities was subsequently used to calculate Shannon entropy. Regions with high Shannon entropy may adopt alternative conformations, while those with low Shannon entropy correspond to either well-defined RNA structures or persistently single-stranded regions (MATHEWS, 2004; Siegfried et al., 2014)."

(6) The overall writing of the manuscript is very good, but there are some minor grammatical issues throughout, e.g., here are some of the ones that I caught:

a) Lines 71-3: "various types of RNA structures such as hairpin structure, RNA singlestrand, RNA pseudoknot and RNA G-quadruplex (G4)" - the examples should be plural and, rather than "hairpins" (or in addition), perhaps add "helixes" to be more generically correct(?).

We have revised the relevant description: 

"various types of RNA structures such as stem-loop structures (with double-helical stems), RNA single-strand, RNA pseudoknot and RNA G-quadruplex (G4)"

b) Lines 74-5: "Of these, RNA G4 has shown considerable promise because of the high stability and modulation by small molecules" should be "Of these, RNA G4 has shown considerable promise because of its high stability and ability for modulation by small molecules."

We have revised the sentence:

“Of these, RNA G4 has shown considerable promise because of its high stability and ability for modulation by small molecules.”

c) Line 76: "have" should be "has".

We have revised the sentence.

d) Lines 104-5 (and elsewhere): "frameshift stimulation element (FSE)" should be "frameshift stimulatory element (FSE)".

We have revised the sentence.

e) Lines 428-9: following the Manfredonia's methods" should be "following Manfredonia's method" or "following the Manfredonia method".

We have made the appropriate edit.

These edits ensure grammatical accuracy and consistency with standard scientific terminology. We appreciate the reviewer's attention to detail, which has significantly improved the clarity of our manuscript.

Reviewer #2 (Recommendations for the authors):

(1) There are some important references missing, on shape-seq from Julius Lucks.

We have added citations to the foundational work by Lucks et al. (2011, PNAS) that pioneered in vitro RNA structure probing using SHAPE-seq.

(2) Describe the acronym "SHAPE",

We have now included the full name of SHAPE:“Selective 2’-Hydroxyl Acylation and Primer Extension”.

(3) Line 81: 2"-hydroxyl-selective - the prime is incorrect.

We thank the reviewer for catching this technical error. We have corrected "2"hydroxyl" to "2'-hydroxyl".

(4) Explaining a bit better how shape reagent works would be beneficial (one sentence should suffice).

We have revised the Introduction section:

“SHAPE reagents like NAI selectively modify flexible, unpaired 2′-OH groups in RNA, and these modifications are detected as mutations during reverse transcription, enabling precise mapping of RNA secondary structures through sequencing.”

(5) Line 128: cite the paper that introduced NAI.

We have now properly cited the original publication introducing NAI（Spitale et al., 2012）.

(6) Line 243: Can you describe what the compound is?

The compound is Braco-19. This has now been included in the methods section. 

(7) Line 272: describe what 3Dpol is and the source of it.

We have supplemented the relevant information as follows:

"3Dpol (recombinant RNA-dependent RNA polymerase; Abcam, ab277617, 0.02 mg/reaction)"

(8) Figure 1 legend: For both C and D, the explanation of the G4 structure and the RISC complex should be added, otherwise, it becomes unclear why they are there.

We have revised the captions for Figure 1 as follows:

"(A) Well-folded regions (low SHAPE reactivity and low Shannon entropy; 26.40% of genome). These regions represent stably folded RNA structures with minimal conformational flexibility, likely serving as structural scaffolds or functional elements in viral replication. (B) Dynamic structured regions (low SHAPE reactivity and high Shannon entropy; 11.70% of genome). These conformationally plastic domains likely mediate regulatory switches between alternative secondary structures during infection. (C) Dynamic unpaired regions (high SHAPE reactivity and high Shannon entropy; 26.90% of genome). These regions are prone to form non-canonical nucleic acid structures (e.g., G-quadruplexes), which can be stabilized by small-molecule ligands to inhibit viral replication. (D) Persistent unpaired regions (high SHAPE reactivity and low Shannon entropy; 9.67% of genome). These regions are more accessible for siRNA binding, facilitating recruitment of Argonaute proteins and Dicer to form the RNAinduced silencing complex (RISC) for targeted cleavage."

(9) Figure S2 panel A should be in Figure 1. This is a nice picture showing the backbone of the research.

In the revised manuscript, we have reorganized Figure 1 and Figure S2 by incorporating the SHAPE-MaP workflow diagram (previously Figure S2A) into Figure 1 as panel (A): 

(10) Please add the citation to Braco-19.

We have now added the appropriate citation for Braco-19 (Gowan et al., 2002) in the revised manuscript.

(11) Figure 5 legend: could you add in parenthesis the what ds means (and call Figure S28).

We appreciate the reviewer's attention to detail. In the revised manuscript, we have clarified the abbreviations in the Figure 5 legend: ss (single-stranded targeting siRNAs); ds (dual-stranded targeting siRNAs). 

(12) Line 107: I would argue that the "stabilization of a G4" inhibited viral proliferation. And that supports the point of the paper, that a small molecule that stabilizes the G4 can be used to reduce viral replication. I suggest emphasizing this thorough the paper.

We fully concur with the reviewer's insightful perspective. In the revised manuscript, we have comprehensively strengthened the point of 'G4 stabilization' as an antiviral mechanism through the following enhancements:

(1) In the Results section: We present Thioflavin T (ThT) fluorescence assays demonstrating the G4-forming capability of PQSs in the full-length PEDV genomic RNA context:

“These findings indicate that although most PQSs can form G4 structures in vitro, PQS1—located in the high SHAPE-high Shannon entropy region—demonstrates the most robust G4-forming capability when competing with local secondary structures in the genomic context.”

(2) In the Results section: The inclusion of Braco-19 inhibition assays using PQS3 mutant virus as control provides robust evidence that Braco-19 exerts its antiviral effects specifically through PQS1 stabilization:

“Furthermore, as a control, we observed nearly identical inhibitory activity of Braco-19 against both the PQS3 mutant strain (AJ1102-PQS3mut) and wild-type virus, demonstrating the specificity of Braco-19's action on PQS1.”

(3) In the Discussion section: We have rewritten the mechanistic interpretation to emphasize: 

"Crucially, Braco-19 showed no inhibitory activity against the PQS1-mutant strain while maintaining potent activity against the PQS3-mutant strain (Figure 4E, Figure 4—figure supplement 3C). This suggests that the compound can selectively target the PQS1 of the high SHAPE-high Shannon region in cells." 

(13) For PQS1, it's suggested that it is indeed a competing and transient conformation that forms the G4. I wonder if using an extended PQS1 (perhaps what is shown in Figure 3E) and using fluorescence, and/or K+ vs Li+, and/or in-vitro SHAPE could tell us more about this dynamic structure. Thioflavin T or any other fluorescent molecule that binds to G4s could be easily used to show how the formation of G4 may happen or not. In addition, how Braco-19 could really lock the dynamic structure in-vitro as well. I think the field would benefit from a deeper investigation of it.

To address the dynamic competition between G4 and alternative RNA conformations, we performed Thioflavin T (ThT) fluorescence turn-on assay (now in Figure 3D-E and Figure 3—figure supplement 6) under physiological K⁺ conditions (100 mM), with PRRSV-G4 RNA as a positive control. This reads as:

“To validate whether SHAPE analysis could reflect the competitive conformational folding of PQSs in the PEDV genome, we performed in vitro transcription to obtain local intact structures containing PQSs within dynamic single-stranded regions and stable double-stranded regions (Table S6). Thioflavin T (ThT) fluorescence turn-on assays were conducted under physiological K⁺ conditions (100 mM), with the G4 sequence of porcine reproductive and respiratory syndrome virus (PRRSV) serving as a positive control (Control-G4)(Fang et al., 2023). The results demonstrated that for short PQSs sequences containing only G4-forming motifs (Table S7), PQS1, PQS3, PQS4, and PQS6 all induced significant ThT fluorescence enhancement (Figure 3D-E, Figure 3—figure supplement 6), confirming their ability to form G4 structures. However, in long RNA fragments encompassing PQSs and their flanking sequences, only PQS1 and PQS4 exhibited pronounced ThT fluorescence responses (Figure 3DE), whereas PQS2, PQS3, and PQS6 showed negligible signals (Figure 3E, Figure 3— figure supplement 6). Notably, the PQS1-long chain displayed the strongest fluorescence signal, while its mutant counterpart (PQS1mut-long chain) exhibited the lowest background fluorescence (Figure 3D). These findings indicate that although most PQSs can form G4 structures in vitro, PQS1—located in the high SHAPE-high Shannon entropy region—demonstrates the most robust G4-forming capability when competing with local secondary structures in the genomic context. Therefore, PQS1 was selected for further structural and functional validation.”

(14) Figure S29 is nice and informative. Consider moving it to the main text.

We appreciate the reviewer's positive assessment of Figure S29. Now we have renamed this figure as "Figure 5—Supplement 2".

Reviewer #1 (Public review):

This is a very interesting paper addressing the hierarchical nature of the mammalian auditory system. The authors use an unconventional technique to assess brain responses -- functional ultrasound imaging (fUSI). This measures blood volume in the cortex at a relatively high spatial resolution. They present dynamic and stationary sounds in isolation and together, and show that the effect of the stationary sounds (relative to the dynamic sounds) on blood volume measurements decreases as one ascends the auditory hierarchy. Since the dynamic/stationary nature of sounds is related to their perception as foreground/background sounds (see below for more details), this suggests that neurons in higher levels of the cortex may be increasingly invariant to background sounds.

The study is interesting, well conducted, and well written. I am broadly convinced by the results. However, I do have some concerns about the validity of the results, given the unconventional technique. fUSI is convenient because it is much less invasive than electrophysiology, and can image a large region of the cortex in one go. However, the relationship between blood volume and neuronal activity is unclear, and blood volume measurements are heavily temporally averaged relative to the underlying neuronal responses. I am particularly concerned about the implications of this for a study on dynamic/stationary stimuli in auditory cortical hierarchy, because the time scale of the dynamic sounds is such that much of the dynamic structure may be affected by this temporal averaging. Also, there is a well-known decrease in temporal following rate that is exhibited by neurons at higher levels of the auditory system. This means that results in different areas will be differently affected by the temporal averaging. I would like to see additional control models to investigate the impact of this.

I also think that the authors should address several caveats: the fact that their measurements heavily spatially average neuronal responses, and therefore may not accurately reflect the underlying neuronal coding; that the perceptual background/foreground distinction is not identical to the dynamic/stationary distinction used here; and that ferret background/foreground perception may be very different from that in humans.

Major points

(1) Changes in blood volume due to brain activity are indirectly related to neuronal responses. The exact relationship is not clear, however, we do know two things for certain: (a) each measurable unit of blood volume change depends on the response of hundreds or thousands of neurons, and (b) the time course of the volume changes are are slow compared to the potential time course of the underlying neuronal responses. Both of these mean that important variability in neuronal responses will be averaged out when measuring blood changes. For example, if two neighbouring neurons have opposite responses to a given stimulus, this will produce opposite changes in blood volume, which will cancel each other out in the blood volume measurement due to (a). This is important in the present study because blood volume changes are implicitly being used as a measure of coding in the underlying neuronal population. The authors need to acknowledge that this is a coarse measure of neuronal responses and that important aspects of neuronal responses may be missing from the blood volume measure.

(2) More importantly for the present study, however, the effect of (b) is that any rapid changes in the response of a single neuron will be cancelled out by temporal averaging. Imagine a neuron whose response is transient, consisting of rapid excitation followed by rapid inhibition. Temporal averaging of these two responses will tend to cancel out both of them. As a result, blood volume measurements will tend to smooth out any fast, dynamic responses in the underlying neuronal population. In the present study, this temporal averaging is likely to be particularly important because the authors are comparing responses to dynamic (nonstationary) stimuli with responses to more constant stimuli. To a first approximation, neuronal responses to dynamic stimuli are themselves dynamic, and responses to constant stimuli are themselves constant. Therefore, the averaging will mean that the responses to dynamic stimuli are suppressed relative to the real responses in the underlying neurons, whereas the responses to constant stimuli are more veridical. On top of this, temporal following rates tend to decrease as one ascends the auditory hierarchy, meaning that the comparison between dynamic and stationary responses will be differently affected in different brain areas. As a result, the dynamic/stationary balance is expected to change as you ascend the hierarchy, and I would expect this to directly affect the results observed in this study.

It is not trivial to extrapolate from what we know about temporal following in the cortex to know exactly what the expected effect would be on the authors' results. As a first-pass control, I would strongly suggest incorporating into the authors' filterbank model a range of realistic temporal following rates (decreasing at higher levels), and spatially and temporally average these responses to get modelled cerebral blood flow measurements. I would want to know whether this model showed similar effects as in Figure 2. From my guess about what this model would show, I think it would not predict the effects shown by the authors in Figure 2. Nevertheless, this is an important issue to address and to provide control for.

(3) I do not agree with the equivalence that the authors draw between the statistical stationarity of sounds and their classification as foreground or background sounds. It is true that, in a common foreground/background situation - speech against a background of white noise - the foreground is non-stationary and the background is stationary. However, it is easy to come up with examples where this relationship is reversed. For example, a continuous pure tone is perfectly stationary, but will be perceived as a foreground sound if played loudly. Background music may be very non-stationary but still easily ignored as a background sound when listening to overlaid speech. Ultimately, the foreground/background distinction is a perceptual one that is not exclusively determined by physical characteristics of the sounds, and certainly not by a simple measure of stationarity. I understand that the use of foreground/background in the present study increases the likely reach of the paper, but I don't think it is appropriate to use this subjective/imprecise terminology in the results section of the paper.

(4) Related to the above, I think further caveats need to be acknowledged in the study. We do not know what sounds are perceived as foreground or background sounds by ferrets, or indeed whether they make this distinction reliably to the degree that humans do. Furthermore, the individual sounds used here have not been tested for their foreground/background-ness. Thus, the analysis relies on two logical jumps - first, that the stationarity of these sounds predicts their foreground/background perception in humans, and second, that this perceptual distinction is similar in ferrets and humans. I don't think it is known to what degree these jumps are justified. These issues do not directly affect the results, but I think it is essential to address these issues in the Discussion, because they are potentially major caveats to our understanding of the work.

Author response:

Reviewer #1:

(1) Changes in blood volume due to brain activity are indirectly related to neuronal responses. The exact relationship is not clear, however, we do know two things for certain: (a) each measurable unit of blood volume change depends on the response of hundreds or thousands of neurons, and (b) the time course of the volume changes are slow compared to the potential time course of the underlying neuronal responses. Both of these mean that important variability in neuronal responses will be averaged out when measuring blood changes. For example, if two neighbouring neurons have opposite responses to a given stimulus, this will produce opposite changes in blood volume, which will cancel each other out in the blood volume measurement due to (a). This is important in the present study because blood volume changes are implicitly being used as a measure of coding in the underlying neuronal population. The authors need to acknowledge that this is a coarse measure of neuronal responses and that important aspects of neuronal responses may be missing from the blood volume measure.

The reviewer is correct: we do not measure neuronal firing, but use blood volume as a proxy for bulk local neuronal activity, which does not capture the richness of single neuron responses. We will highlight this point in the manuscript. This is why the paper focuses on large-scale spatial representations as well as cross-species comparison. For this latter purpose, fMRI responses are on par with our fUSI data, with both neuroimaging techniques showing the same weakness.

(2) More importantly for the present study, however, the effect of (b) is that any rapid changes in the response of a single neuron will be cancelled out by temporal averaging. Imagine a neuron whose response is transient, consisting of rapid excitation followed by rapid inhibition. Temporal averaging of these two responses will tend to cancel out both of them. As a result, blood volume measurements will tend to smooth out any fast, dynamic responses in the underlying neuronal population. In the present study, this temporal averaging is likely to be particularly important because the authors are comparing responses to dynamic (nonstationary) stimuli with responses to more constant stimuli. To a first approximation, neuronal responses to dynamic stimuli are themselves dynamic, and responses to constant stimuli are themselves constant. Therefore, the averaging will mean that the responses to dynamic stimuli are suppressed relative to the real responses in the underlying neurons, whereas the responses to constant stimuli are more veridical. On top of this, temporal following rates tend to decrease as one ascends the auditory hierarchy, meaning that the comparison between dynamic and stationary responses will be differently affected in different brain areas. As a result, the dynamic/stationary balance is expected to change as you ascend the hierarchy, and I would expect this to directly affect the results observed in this study.

It is not trivial to extrapolate from what we know about temporal following in the cortex to know exactly what the expected effect would be on the authors' results. As a first-pass control, I would strongly suggest incorporating into the authors' filterbank model a range of realistic temporal following rates (decreasing at higher levels), and spatially and temporally average these responses to get modelled cerebral blood flow measurements. I would want to know whether this model showed similar effects as in Figure 2. From my guess about what this model would show, I think it would not predict the effects shown by the authors in Figure 2. Nevertheless, this is an important issue to address and to provide control for.

We understand the reviewer’s concern about potential differences in response dynamics in stationary vs non-stationary sounds. In particular, it seems that the reviewer is concerned that responses to foregrounds may be suppressed in non-primary fields because foregrounds are not stationary, and non-primary regions could struggle to track and respond to these sounds. Nevertheless, we  observed the contrary, with non-primary regions over-representing non-stationary (dynamic) sounds, over stationary ones. For this reason, we are inclined to think that this explanation cannot falsify our findings.

Furthermore, background sounds are not completely constant: they are still dynamic sounds, but their temporal modulation rates are usually faster (see Figure 3B). Similarly, neural responses to these two types of sounds are dynamic (see for example Hamersky et al., 2025, Figure 1).  Thus, we are not sure that blood volume would transform the responses to these types of sounds non-linearly.

We understand the comment that temporal following rates might differ across regions in the auditory hierarchy and agree. In fact, we show that tuning to temporal rates differ across regions and partly explains the differences in background invariance we observe. We think the reviewer’s suggestion is already implemented by our spectrotemporal model, which incorporates the full range of realistic temporal following rates (up to 128 Hz). The temporal averaging is done as we take the output of the model (which varies continuously through time) and average it in the same window as we used for our fUSI data. When we fit this model to the ferret data, we find that voxels in non-primary regions, especially VP (tertiary auditory cortex), tend to be more tuned to low temporal rates (Figure 2F, G), and that background invariance is stronger in voxels tuned to low rates. This is, however, not true in humans, suggesting that background invariance in humans rely on different computational mechanisms.

(3) I do not agree with the equivalence that the authors draw between the statistical stationarity of sounds and their classification as foreground or background sounds. It is true that, in a common foreground/background situation - speech against a background of white noise - the foreground is non-stationary and the background is stationary. However, it is easy to come up with examples where this relationship is reversed. For example, a continuous pure tone is perfectly stationary, but will be perceived as a foreground sound if played loudly. Background music may be very non-stationary but still easily ignored as a background sound when listening to overlaid speech. Ultimately, the foreground/background distinction is a perceptual one that is not exclusively determined by physical characteristics of the sounds, and certainly not by a simple measure of stationarity. I understand that the use of foreground/background in the present study increases the likely reach of the paper, but I don't think it is appropriate to use this subjective/imprecise terminology in the results section of the paper.

We appreciate the reviewer’s comment that the classification of our sounds into foregrounds and backgrounds is not verified by any perceptual experiments. We use those terms to be consistent with the literature, including the paper we derived this definition from (Kell et al., 2019). These terms are widely used in studies where no perceptual or behavioral experiments are included, and even when animals are anesthetized. However, we will emphasize the limits of this definition when introducing it, as well as in the discussion.

(4) Related to the above, I think further caveats need to be acknowledged in the study. We do not know what sounds are perceived as foreground or background sounds by ferrets, or indeed whether they make this distinction reliably to the degree that humans do. Furthermore, the individual sounds used here have not been tested for their foreground/background-ness. Thus, the analysis relies on two logical jumps - first, that the stationarity of these sounds predicts their foreground/background perception in humans, and second, that this perceptual distinction is similar in ferrets and humans. I don't think it is known to what degree these jumps are justified. These issues do not directly affect the results, but I think it is essential to address these issues in the Discussion, because they are potentially major caveats to our understanding of the work.

We agree with the reviewer that the foreground-background distinction might be different in ferrets. In anticipation of that issue, we had enriched the sound set with more ecologically relevant sounds, such as ferret and other animal vocalizations. Nevertheless, the point remains valid and is already raised in the discussion. We will emphasize this limitation in addition to the limitation of our definition of foregrounds and backgrounds.

Reviewer #2:

(1) Interpretation of the cerebral blood volume signal: While the results are compelling, more caution should be exercised by the authors in framing their results, given that they are measuring an indirect measure of neural activity, this is the difference between stating "CBV in area MEG was less background invariant than in higher areas" vs. saying "MEG was less background invariant than other areas". Beyond framing, the basic properties of the CBV signal should be better explored:

a) Cortical vasculature is highly structured (e.g. Kirst et al.( 2020) Cell). One potential explanation for the results is simply differences in vasculature and blood flow between primary and secondary areas of auditory cortex, even if fUS is sensitive to changes in blood flow, changes in capillary beds, etc (Mace et al., 2011) Nat. Methods.. This concern could be addressed by either analyzing spontaneous fluctuations in the CBV signal during silent periods or computing a signal-to-noise ratio of voxels across areas across all sound types. This is especially important given the complex 3D geometry of gyri and sulci in the ferret brain.

We agree with the reviewers that there could be differences in vasculature across subregions of the auditory cortex. We will run analyses providing comparisons of basic signal properties across our different regions of interest. We note that this point would also be valid for the human fMRI data, for which we cannot run these controls. Nevertheless, this should not affect our analyses and results, which should be independent of local vascular density. First, we normalize the signal in each voxel before any analysis, so that the absolute strength of the signal, or blood volume in a given voxel, does not matter. Second, we do see sound-evoked responses in all regions (Figure S2) and only focus on reliable voxels in each region. Third, our analysis mostly relies on voxel-based correlation across sounds, which is independent of the mean and variance of the voxel responses. Thus, we believe that differences in vascular architecture across regions are unlikely to affect our results.

b) Figure 1 leaves the reader uncertain what exactly is being encoded by the CBV signal, as temporal responses to different stimuli look very similar in the examples shown. One possibility is that the CBV is an acoustic change signal. In that case, sounds that are farther apart in acoustic space from previous sounds would elicit larger responses, which is straightforward to test. Another possibility is that the fUS signal reflects time-varying features in the acoustic signal (e.g. the low-frequency envelope). This could be addressed by cross-correlating the stimulus envelope with fUS waveform. The third possibility, which the authors argue, is that the magnitude of the fUS signal encodes the stimulus ID. A better understanding of the justification for only looking at the fUS magnitude in a short time window (2-4.8 s re: stimulus onset) would increase my confidence in the results.

We thank the reviewer for raising that point as it highlights that the layout of Figure 1 is misleading. While Figure 1B shows an example snippet of our sound streams, Figure 1D shows the average timecourse of CBV time-locked to a change in sound (foreground or background, isolated or in a mixture). This is the average across all voxels and sounds, and the point is just to illustrate the dynamics for the three broad categories. In Figure 1E however, we show the cross-validated cross-correlation of CBV  across sounds (and different time lags). To obtain this, we compute for each voxel the response to each sound at each time lag, thus obtaining two vector of size number of sounds per lag, one per repeat. Then, we correlate all these vectors across the two repeats, obtaining one cross-correlation matrix per neuron. We finally average these matrices across all neurons. The fact that you see red squares demonstrates that the signal encodes sound identity, since CBV is more similar across two repeats of the same sound (for e.g., in the foreground only matrix, 0-5 s vs 0-5 s), than two different sounds (0-5 s vs. 7-12 s). We will modify the figure layout as well as the legend to improve clarity.

(2) Interpretation of the human data: The authors acknowledge in the discussion that there are several differences between fMRI and fUS. The results would be more compelling if they performed a control analysis where they downsampled the Ferret fUS data spatially and temporally to match the resolution of fMRI and demonstrated that their ferret results hold with lower spatiotemporal resolution.

We agree with the reviewer that the use of different techniques might come in the way of cross-species comparison. We will add additional discussion on this point. We already control for the temporal aspect by using the average of stimulus-evoked activity across time (note that due to scanner noise, sounds are presented cut into small pieces in the fMRI experiments). Regarding the spatial aspect, there are several things to consider. First, both species have brains of very different sizes, a factor that is conveniently compensated for by the higher spatial resolution of fUSI compared to fMRI (0.1 vs 2 mm). Downsampling to fMRI resolution would lead to having one voxel per region per slice, which is not feasible. We also summarize results with one value per region, which is a form of downsampling that is fairer across species. Furthermore, we believe that we already established in a previous study (Landemard et al, 2021 eLife) that fUSI and fMRI data are comparable signals. We indeed could predict human fMRI responses to most sounds from ferret fUSI responses to the same identical sounds.

Reviewer #3:

As mentioned above, interpretation of the invariance analyses using predictions from the spectrotemporal modulation encoding model hinges on the model's ability to accurately predict neural responses. Although Figure S5 suggests the encoding model was generally able to predict voxel responses accurately, the authors note in the introduction that, in human auditory cortex, this kind of tuning can explain responses in primary areas but not in non-primary areas (Norman-Haignere & McDermott, PLOS Biol. 2018). Indeed, the prediction accuracy histograms in Figure S5C suggest a slight difference in the model's ability to predict responses in primary versus non-primary voxels. Additional analyses should be done to a) determine whether the prediction accuracies are meaningfully different across regions and b) examine whether controlling for prediction accuracy across regions (i.e., sub-selecting voxels across regions with matched prediction accuracy) affects the outcomes of the invariance analyses.

The reviewer is correct: the spectrotemporal model tends to perform less well in human non-primary cortex. We believe this does not contradict our results but goes in the same direction: while there is a gradient in invariance in both ferrets and humans, this gradient is predicted by the spectrotemporal model in ferrets, but not in humans (possibly indeed because predictions are less good in human non-primary auditory cortex). Regardless of the mechanism, this result points to a difference across species. We will clarify these points by quantifying potential differences in prediction accuracy in both species and comment on those in the manuscript.

A related concern is the procedure used to train the encoding model. From the methods, it appears that the model may have been fit using responses to both isolated and mixture sounds. If so, this raises questions about the interpretability of the invariance analyses. In particular, fitting the model to all stimuli, including mixtures, may inflate the apparent ability of the model to "explain" invariance, since it is effectively trained on the phenomenon it is later evaluated on. Put another way, if a voxel exhibits invariance, and the model is trained to predict the voxel's responses to all types of stimuli (both isolated sounds and mixtures), then the model must also show invariance to the extent it can accurately predict voxel responses, making the result somewhat circular. A more informative approach would be to train the encoding model only on responses to isolated sounds (or even better, a completely independent set of sounds), as this would help clarify whether any observed invariance is emergent from the model (i.e., truly a result of low-level tuning to spectrotemporal features) or simply reflects what it was trained to reproduce.

We thank the reviewer for this suggestion and will run an additional prediction using only the sounds presented in isolation. This will be included in the next version of the manuscript.

Finally, the interpretation of the foreground invariance results remains somewhat unclear. In ferrets (Figure 2I), the authors report relatively little foreground invariance, whereas in humans (Figure 5G), most participants appear to show relatively high levels of foreground invariance in primary auditory cortex (around 0.6 or greater). However, the paper does not explicitly address these apparent cross-species differences. Moreover, the findings in ferrets seem at odds with other recent work in ferrets (Hamersky et al. 2025 J. Neurosci.), which shows that background sounds tend to dominate responses to mixtures, suggesting a prevalence of foreground invariance at the neuronal level. Although this comparison comes with the caveat that the methods differ substantially from those used in the current study, given the contrast with the findings of this paper, further discussion would nonetheless be valuable to help contextualize the current findings and clarify how they relate to prior work.

We thank the reviewer for this point. We will indeed add further discussion of the  difference between ferrets and humans in foreground invariance in primary auditory cortex. In addition, while we found a trend for higher background invariance than foreground invariance in ferret primary auditory cortex, this difference was not significant and many voxels exhibit similar levels of background and foreground invariance (for example in Figure 2D, G). Thus, we do not think our results are inconsistent with Hamersky et al., 2025, though we agree the bias towards background sounds is not as strong in our data. This might indeed reflect differences in methodology, both in the signal that is measured (blood volume vs spikes), and the sound presentation paradigm. We will add this point to our discussion.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

This is a very interesting paper investigating the fitness and cellular effects of mutations that drive dihedral protein complex into forming filaments. The Levy group have previously shown that this can happen relatively easily in such complexes and this paper now investigates the cellular consequences of this phenomenon. The study is very rigorous biophysically and very surprisingly comes up empty in terms of an effect: apparently this kind of self-assembly can easily be tolerated in yeast, which was certainly not my expectation. This is a very interesting result, because it implies that such assemblies may evolve neutrally because they fulfill the two key requirements for such a trajectory: They are genetically easily accessible (in as little as a single mutation), and they have perhaps no detrimental effect on fitness. This immediately poses two very interesting questions: Are some natural proteins that are known to form filaments in the cell perhaps examples of such neutral trajectories? And if this trait is truly neutral (as long as it doesn't affect the base biochemical function of the protein in question), why don't we observe more proteins form these kinds of ordered assemblies.

I have no major comments about the experiments as I find that in general very carefully carried out. I have two more general comments:

1. The fitness effect of these assemblies, if one exists, seems very small. I think it's worth remembering that even very small fitness effects beyond even what competition experiments can reveal could in principle be enough to keep assembly-inducing alleles at very low frequencies in natural populations. Perhaps this could be acknowledged in the paper somewhere.
2. The proteins used in this study I think were chosen such that they do not have an important function in yeast that could be disrupted by assembly This allows the effect of the large scale assemblies to be measured in isolation. If I deduced this correctly, this should probably be pointed out agin in this paper (I apologise if I missed this).
3. The model system in which these effects were tested for is yeast. This organism has a rigid cell wall and I was wondering if this makes it more tolerant to large scale assemblages than wall-less eukaryotes. Could the authors comment on this?

Minor points:

In Figure 2D, what are the fits? And is there any analysis that rules out expression effects on the mutant caused by higher levels of the wild-type? The error bars in Figure 2E are not defined.

Significance

This is a remarkably rigours paper that investigates whether self-assembly into large structures has any fitness effect on a single celled organism. This is very relevant, because a landmark paper from the Levy group showed that many proteins are very close in genetic terms to forming such assemblies. The general expectation I think would have been that this phenomenon is pretty harmful. This would have explained why such filaments are relatively rare as far as we know. This paper now does a large number of highly rigours experiments to first prove beyond doubt that a range of model proteins really can be coaxed into forming such filaments in yeast cells through a very small number of mutations. Its perhaps most surprising result is that this does not negatively affect yeast cells.

From an evolutionary perspective, this is a very interesting and highly surprising result. It forces us to rethink why such filaments are not more common in Nature. Two possible answers come to mind: First, it's possible that filamentation is not directly harmful to the cell, but that assembling proteins into filaments can interfere with their basic biochemical function (which was not tested for here).

Second, perhaps assembly does cause a fitness defect, but one so small that it is hard to measure experimentally. Natural selection is very powerful, and even fitness coefficients we struggle to measure in the laboratory can have significant effects in the wild. If this is true, we might expect such filaments to be more common in organisms with small effective population sizes, in which selection is less effective.

A third possibility is of course that the prevalence of such self-assembly is under-appreciated. Perhaps more proteins than we currently know assemble into these structures under some conditions without any benefit or detriment to the organism.

These are all fascinating implications of this work that straddle the fields of evolutionary genetics and biochemistry and are therefore relevant to a very wide audience. My own expertise is in these two fields. I also think that this work will be exciting for synthetic biologists, because it proves that these kinds of assemblies are well tolerated inside cells.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This paper is an elegant, mostly observational work, detailing observations that polysome accumulation appears to drive nucleoid splitting and segregation. Overall I think this is an insightful work with solid observations.

Thank you for your appreciation and positive comments. In our view, an appealing aspect of this proposed biophysical mechanism for nucleoid segregation is its self-organizing nature and its ability to intrinsically couple nucleoid segregation to biomass growth, regardless of nutrient conditions.

Strengths:

The strengths of this paper are the careful and rigorous observational work that leads to their hypothesis. They find the accumulation of polysomes correlates with nucleoid splitting, and that the nucleoid segregation occurring right after splitting correlates with polysome segregation. These correlations are also backed up by other observations:

(1) Faster polysome accumulation and DNA segregation at faster growth rates.

(2) Polysome distribution negatively correlating with DNA positioning near asymmetric nucleoids.

(3) Polysomes form in regions inaccessible to similarly sized particles.

These above points are observational, I have no comments on these observations leading to their hypothesis.

Thank you!

Weaknesses:

It is hard to state weaknesses in any of the observational findings, and furthermore, their two tests of causality, while not being completely definitive, are likely the best one could do to examine this interesting phenomenon.

It is indeed difficult to prove causality in a definitive manner when the proposed coupling mechanism between nucleoid segregation and gene expression is self-organizing, i.e., does not involve a dedicated regulatory molecule (e.g., a protein, RNA, metabolite) that we could have eliminated through genetic engineering to establish causality. We are grateful to the reviewer for recognizing that our two causality tests are the best that can be done in this context.

Points to consider / address:

Notably, demonstrating causality here is very difficult (given the coupling between transcription, growth, and many other processes) but an important part of the paper. They do two experiments toward demonstrating causality that help bolster - but not prove - their hypothesis. These experiments have minor caveats, my first two points.

(1) First, "Blocking transcription (with rifampicin) should instantly reduce the rate of polysome production to zero, causing an immediate arrest of nucleoid segregation". Here they show that adding rifampicin does indeed lead to polysome loss and an immediate halting of segregation - data that does fit their model. This is not definitive proof of causation, as rifampicin also (a) stops cell growth, and (b) stops the translation of secreted proteins. Neither of these two possibilities is ruled out fully.

That’s correct; cell growth also stops when gene expression is inhibited, which is consistent with our model in which gene expression within the nucleoid promotes nucleoid segregation and biomass growth (i.e., cell growth), inherently coupling these two processes. This said, we understand the reviewer’s point: the rifampicin experiment doesn’t exclude the possibility that protein secretion and cell growth drive nucleoid segregation. We are assuming that the reviewer is envisioning an alternative model in which sister nucleoids would move apart because they would be attached to the membrane through coupled transcription-translation-protein secretion (transertion) and the membrane would expand between the separating nucleoids, similar to the model proposed by Jacob et al in 1963 (doi:10.1101/SQB.1963.028.01.048). There are several observations arguing against cell elongation/transertion acting a predominant mechanism of nucleoid segregation.

(1) For this alternative mechanism to work, membrane growth must be localized at the middle of the splitting nucleoids (i.e., midcell position for slow growth and ¼ and ¾ cell positions for fast growth) to create a directional motion. To our knowledge, there is no evidence of such localized membrane incorporation. Furthermore, even if membrane growth was localized at the right places, the fluidity of the cytoplasmic membrane (PMID: 6996724, 20159151, 24735432, 27705775) would be problematic. To circumvent the membrane fluidity issue, one could potentially evoke an additional connection to the rigid peptidoglycan, but then again, peptidoglycan growth would have to be localized at the middle of the splitting nucleoid. However, peptidoglycan growth is dispersed early in the cell division cycle when the nucleoid splitting happens in fast growing cells and only appears to be zonal after the onset of cell constriction (PMID: 35705811, 36097171, 2656655).

(2) Even if we ignore the aforementioned caveats, Paul Wiggins’s group ruled out the cell elongation/transertion model by showing that the rate of cell elongation is slower than the rate of chromosome segregation (PMID: 23775792). In our revised manuscript, we clarify this point and provide confirmatory data showing that the cell elongation rate is indeed slower than the nucleoid segregation rate (Figure 1H and Figure 1 - figure supplement 5A), indicating that it cannot be the main driver.

(3) The asymmetries in nucleoid compaction that we described in our paper are predicted by our model. We do not see how they could be explained by cell growth or protein secretion.

(4) We also show that polysome accumulation at ectopic sites (outside the nucleoid) results in correlated nucleoid dynamics, consistent with our proposed mechanism. It is not clear to us how such nucleoid dynamics could be explained by cell growth or protein secretion (transertion).

(1a) As rifampicin also stops all translation, it also stops translational insertion of membrane proteins, which in many old models has been put forward as a possible driver of nucleoid segregation, and perhaps independent of growth. This should at last be mentioned in the discussion, or if there are past experiments that rule this out it would be great to note them.

It is not clear to us how the attachment of the DNA to the cytoplasmic membrane could alone create a directional force to move the sister nucleoids. We agree that old models have proposed a role for cell elongation (providing the force) and transertion (providing the membrane tether). Please see our response above for the evidence (from the literature and our work) against it. This was mentioned in the Introduction and Results section, but we agree that this was not well explained. We have now put emphasis on the related experimental data (Figure 1H, Figure 1 – figure supplement 5A, ) and revised the text (lines 199 - 210) to clarify these points.

(1b) They address at great length in the discussion the possibility that growth may play a role in nucleoid segregation. However, this is testable - by stopping surface growth with antibiotics. Cells should still accumulate polysomes for some time, it would be easy to see if nucleoids are still segregated, and to what extent, thereby possibly decoupling growth and polysome production. If successful, this or similar experiments would further validate their model.

We reviewed the literature and could not find a drug that stops cell growth without stopping gene expression. Any drug that affects the integrity or potential of the membrane depletes cells of ATP; without ATP, gene expression is inhibited. However, our experiment in which we drive polysome accumulation at ectopic sites decouples polysome accumulation from cell growth. In this experiment, by redirecting most of chromosome gene expression to a single plasmid-encoded gene, we reduce the rate of cell growth but still create a large accumulation of polysomes at an ectopic location. This ectopic polysome accumulation is sufficient to affect nucleoid dynamics in a correlated fashion. In the revised manuscript, we have clarified this point and added model simulations (Figure 7 – figure supplement 2) to show that our experimental observations are predicted by our model.

(2) In the second experiment, they express excess TagBFP2 to delocalize polysomes from midcell. Here they again see the anticorrelation of the nucleoid and the polysomes, and in some cells, it appears similar to normal (polysomes separating the nucleoid) whereas in others the nucleoid has not separated. The one concern about this data - and the differences between the "separated" and "non-separated" nuclei - is that the over-expression of TagBFP2 has a huge impact on growth, which may also have an indirect effect on DNA replication and termination in some of these cells. Could the authors demonstrate these cells contain 2 fully replicated DNA molecules that are able to segregate?

We have included new flow cytometry data of fluorescently labeled DNA to show that DNA replication is not impacted.

(3) What is not clearly stated and is needed in this paper is to explain how polysomes do (or could) "exert force" in this system to segregate the nucleoid: what a "compaction force" is by definition, and what mechanisms causes this to arise (what causes the "force") as the "compaction force" arises from new polysomes being added into the gaps between them caused by thermal motions.

They state, "polysomes exert an effective force", and they note their model requires "steric effects (repulsion) between DNA and polysomes" for the polysomes to segregate, which makes sense. But this makes it unclear to the reader what is giving the force. As written, it is unclear if (a) these repulsions alone are making the force, or (b) is it the accumulation of new polysomes in the center by adding more "repulsive" material, the force causes the nucleoids to move. If polysomes are concentrated more between nucleoids, and the polysome concentration does not increase, the DNA will not be driven apart (as in the first case) However, in the second case (which seems to be their model), the addition of new material (new polysomes) into a sterically crowded space is not exerting force - it is filling in the gaps between the molecules in that region, space that needs to arise somehow (like via Brownian motion). In other words, if the polysome region is crowded with polysomes, space must be made between these polysomes for new polysomes to be inserted, and this space must be made by thermal (or ATP-driven) fluctuations of the molecules. Thus, if polysome accumulation drives the DNA segregation, it is not "exerting force", but rather the addition of new polysomes is iteratively rectifying gaps being made by Brownian motion.

We apologize for the understandable confusion. In our picture, the polysomes and DNA (conceptually considered as small plectonemic segments) basically behave as dissolved particles. If these particles were noninteracting, they would simply mix. However, both polysomes and DNA segments are large enough to interact sterically. So as density increases, steric avoidance implies a reduced conformational entropy and thus a higher free energy per particle. We argue (based on Miangolarra et al. 2021 PMID: 34675077 and Xiang et al. 2021 PMID: 34186018) that the demixing of polysomes and DNA segments occurs because DNA segments pack better with each other than they do with polysomes. This raises the free energy cost associated with DNA-polysome interactions compared to DNA-DNA interactions. We model this effect by introducing a term in the free energy χ_np, which refers to as a repulsion between DNA and polysomes, though as explained above it arises from entropic effects. At realistic cellular densities of DNA and polysomes, this repulsive interaction is strong enough to cause the DNA and polysomes to phase separate.

This same density-dependent free energy that causes phase separation can also give rise to forces, just in the way that a higher pressure on one side of a wall can give rise to a net force on the wall. Indeed, the “compaction force” we refer to is fundamentally an osmotic pressure difference. At some stages during nucleoid segregation, the region of the cell between nucleoids has a higher polysome concentration, and therefore a higher osmotic pressure, than the regions near the poles. This results in a net poleward force on the sister nucleoids that drives their migration toward the poles. This migration continues until the osmotic pressure equilibrates. Therefore, both phase separation (due to the steric repulsion described above) and nonequilibrium polysome production and degradation (which creates the initial accumulation of polysomes around midcell) are essential ingredients for nucleoid segregation.

This has been clarified in the revised text, with the support of additional simulation results showing how the asymmetry in polysome distribution causes a compaction force (Figure 4A).

The authors use polysome accumulation and phase separation to describe what is driving nucleoid segregation. Both terms are accurate, but it might help the less physically inclined reader to have one term, or have what each of these means explicitly defined at the start. I say this most especially in terms of "phase separation", as the currently huge momentum toward liquid-liquid interactions in biology causes the phrase "phase separation" to often evoke a number of wider (and less defined) phenomena and ideas that may not apply here. Thus, a simple clear definition at the start might help some readers.

In our case, phase separation means that the DNA-polysome steric repulsion is strong enough to drive their demixing, which creates a compact nucleoid. As mentioned in a previous point, this effect is captured in the free energy by the χ_np term, which is an effective repulsion between DNA and polysomes, though it arises from entropic effects.

In the revised manuscript, we now illustrate this with our theoretical model by initializing a cell with a diffuse nucleoid and low polysome concentration. For the sake of simplicity, we assume that the cell does not elongate. We observe that the DNA-polysome steric repulsion is sufficient to compact the nucleoid and place it at mid-cell (new Figure 4A).

(4) Line 478. "Altogether, these results support the notion that ectopic polysome accumulation drives nucleoid dynamics". Is this right? Should it not read "results support the notion that ectopic polysome accumulation inhibits/redirects nucleoid dynamics"?

We think that the ectopic polysome accumulation drives nucleoid dynamics. In our theoretical model, we can introduce polysome production at fixed sources to mimic the experiments where ectopic polysome production is achieved by high plasmid expression. The model is able to recapitulate the two main phenotypes observed in experiments (Figure 7). These new simulation results have been added to the revised manuscript (Figure 7 – figure supplement 2).

(5) It would be helpful to clarify what happens as the RplA-GFP signal decreases at midcell in Figure 1- is the signal then increasing in the less "dense" parts of the cell? That is, (a) are the polysomes at midcell redistributing throughout the cell? (b) is the total concentration of polysomes in the entire cell increasing over time?

It is a redistribution—the RplA-GFP signal remains constant in concentration from cell birth to division (Figure 1 – Figure Supplement 1E). This is now clarified in the revised text.

(6) Line 154. "Cell constriction contributed to the apparent depletion of ribosomal signal from the mid-cell region at the end of the cell division cycle (Figure 1B-C and Movie S1)" - It would be helpful if when cell constriction began and ended was indicated in Figures 1B and C.

Good idea. We have added markers in Figure 1C to indicate the average start of cell constriction. This relative time from birth to division was estimated as described in the new Figure 1 – figure supplement 2. We have also indicated that cell birth and division correspond to the first and last images/timepoint in Figure 1B and C, respectively. The two-imensional average cell projections presented in Figure 3D also indicate the average timing of cell constriction, consistent with our analysis in Figure 1 – figure supplement 2.

(7) In Figure 7 they demonstrate that radial confinement is needed for longitudinal nucleoid segregation. It should be noted (and cited) that past experiments of Bacillus l-forms in microfluidic channels showed a clear requirement role for rod shape (and a given width) in the positing and the spacing of the nucleoids.

Wu et al, Nature Communications, 2020. "Geometric principles underlying the proliferation of a model cell system" https://dx.doi.org/10.1038/s41467-020-17988-7

Good point! We have revised the text to mention this work. Thank you.

(8) "The correlated variability in polysome and nucleoid patterning across cells suggests that the size of the polysome-depleted spaces helps determine where the chromosomal DNA is most concentrated along the cell length. This patterning is likely reinforced through the displacement of the polysomes away from the DNA dense region"

It should be noted this likely functions not just in one direction (polysomes dictating DNA location), but also in the reverse - as the footprint of compacted DNA should also exclude (and thus affect) the location of polysomes

We agree that the effects could go both ways at this early stage of the story. We have revised the text accordingly.

(9) Line 159. Rifampicin is a transcription inhibitor that causes polysome depletion over time. This indicates that all ribosomal enrichments consist of polysomes and therefore will be referred to as polysome accumulations hereafter". Here and throughout this paper they use the term polysome, but cells also have monosomes (and 2 somes, etc). Rifampicin stops the assembly of all of these, and thus the loss of localization could occur from both. Thus, is it accurate to state that all transcription events occur in polysomes? Or are they grouping all of the n-somes into one group?

In the original discussion, we noted that our term “polysomes” also includes monosomes for simplicity, but we agree that the term should have been defined much earlier. The manuscript has been revised accordingly. Furthermore, in the revised manuscript, we have included additional simulation results with three different diffusion coefficients that reflect different polysome sizes to show that different polysome species with less or more ribosomes give similar results (Figure 4 – figure supplement 4). This shows that the average polysome description in our model is sufficient.

Thank you for the valuable comments and suggestions!

Reviewer #2 (Public review):

Summary:

The authors perform a remarkably comprehensive, rigorous, and extensive investigation into the spatiotemporal dynamics between ribosomal accumulation, nucleoid segregation, and cell division. Using detailed experimental characterization and rigorous physical models, they offer a compelling argument that nucleoid segregation rates are determined at least in part by the accumulation of ribosomes in the center of the cell, exerting a steric force to drive nucleoid segregation prior to cell division. This evolutionarily ingenious mechanism means cells can rely on ribosomal biogenesis as the sole determinant for the growth rate and cell division rate, avoiding the need for two separate 'sensors,' which would require careful coupling.

Terrific summary! Thank you for your positive assessment.

Strengths:

In terms of strengths; the paper is very well written, the data are of extremely high quality, and the work is of fundamental importance to the field of cell growth and division. This is an important and innovative discovery enabled through a combination of rigorous experimental work and innovative conceptual, statistical, and physical modeling.

Thank you!

Weaknesses:

In terms of weaknesses, I have three specific thoughts.

Firstly, my biggest question (and this may or may not be a bona fide weakness) is how unambiguously the authors can be sure their ribosomal labeling is reporting on polysomes, specifically. My reading of the work is that the loss of spatial density upon rifampicin treatment is used to infer that spatial density corresponds to polysomes, yet this feels like a relatively indirect way to get at this question, given rifampicin targets RNA polymerase and not translation. It would be good if a more direct way to confirm polysome dependence were possible.

The heterogeneity of ribosome distribution inside E. coli cells has been attributed to polysomes by many labs (PMID: 25056965, 38678067, 22624875, 31150626, 34186018, 10675340). The attribution is also consistent with single-molecule tracking experiments showing that slow-moving ribosomes (polysomes) are excluded by the nucleoid whereas fast-diffusing ribosomes (free ribosomal subunits) are distributed throughout the cytoplasm (PMID: 25056965, 22624875). These points are now mentioned in the revised manuscript.

Second, the authors invoke a phase separation model to explain the data, yet it is unclear whether there is any particular evidence supporting such a model, whether they can exclude simpler models of entanglement/local diffusion (and/or perhaps this is what is meant by phase separation?) and it's not clear if claiming phase separation offers any additional insight/predictive power/utility. I am OK with this being proposed as a hypothesis/idea/working model, and I agree the model is consistent with the data, BUT I also feel other models are consistent with the data. I also very much do not think that this specific aspect of the paper has any bearing on the paper's impact and importance.

We appreciate the reviewer’s comment, but the output of our reaction-diffusion model is a bona fide phase separation (spinodal decomposition). So, we feel that we need to use the term when reporting the modeling results. Inside the cell, the situation is more complicated. As the reviewer points out, there are likely entanglements (not considered in our model) and other important factors (please see our discussion on the model limitations). This said, we have revised our text to clarify our terms and proposed mechanism.

Finally, the writing and the figures are of extremely high quality, but the sheer volume of data here is potentially overwhelming. I wonder if there is any way for the authors to consider stripping down the text/figures to streamline things a bit? I also think it would be useful to include visually consistent schematics of the question/hypothesis/idea each of the figures is addressing to help keep readers on the same page as to what is going on in each figure. Again, there was no figure or section I felt was particularly unclear, but the sheer volume of text/data made reading this quite the mental endurance sport! I am completely guilty of this myself, so I don't think I have any super strong suggestions for how to fix this, but just something to consider.

We agree that there is a lot to digest. We could not come up with great ideas for visuals others than the schematics we already provide. However, we have revised the text to clarify our points and added a simulation result (Figure 4A) to help explain biophysical concepts.

Reviewer #3 (Public review):

Summary:

Papagiannakis et al. present a detailed study exploring the relationship between DNA/polysome phase separation and nucleoid segregation in Escherichia coli. Using a combination of experiments and modelling, the authors aim to link physical principles with biological processes to better understand nucleoid organisation and segregation during cell growth.

Strengths:

The authors have conducted a large number of experiments under different growth conditions and physiological perturbations (using antibiotics) to analyse the biophysical factors underlying the spatial organisation of nucleoids within growing E. coli cells. A simple model of ribosome-nucleoid segregation has been developed to explain the observations.

Weaknesses:

While the study addresses an important topic, several aspects of the modelling, assumptions, and claims warrant further consideration.

Thank you for your feedback. Please see below for a response to each concern.

Major Concerns:

Oversimplification of Modelling Assumptions:

The model simplifies nucleoid organisation by focusing on the axial (long-axis) dimension of the cell while neglecting the radial dimension (cell width). While this approach simplifies the model, it fails to explain key experimental observations, such as:

(1) Inconsistencies with Experimental Evidence:

The simplified model presented in this study predicts that translation-inhibiting drugs like chloramphenicol would maintain separated nucleoids due to increased polysome fractions. However, experimental evidence shows the opposite-separated nucleoids condense into a single lobe post-treatment (Bakshi et al 2014), indicating limitations in the model's assumptions/predictions. For the nucleoids to coalesce into a single lobe, polysomes must cross the nucleoid zones via the radial shells around the nucleoid lobes.

We do not think that the results from chloramphenicol-treated cells are inconsistent with our model. Our proposed mechanism predicts that nucleoids will condense in the presence of chloramphenicol, consistent with experiments. It also predicts that nucleoids that were still relatively close at the time of chloramphenicol treatment could fuse if they eventually touched through diffusion (thermal fluctuation) to reduce their interaction with the polysomes and minimize their conformational energy. Fusion is, however, not expected for well-separated nucleoids since their diffusion is slow in the crowded cytoplasm. This is consistent with our experimental observations: In the presence of a growth-inhibitory concentration of chloramphenicol (70 μg/mL), nucleoids in relatively close proximity can fuse, but well-separated nucleoids condense and do not fuse. Since the growth rate inhibition is not immediate upon chloramphenicol treatment, many cells with well-separated condensed nucleoids divide during the first hour. As a result, the non-fusion phenotype is more obvious in non-dividing cells, achieved by pre-treating cells with the cell division inhibitor cephalexin (50μg/mL). In these polyploid elongated cells, well-separated nucleoids condensed but did not fuse, not even after an hour in the presence of chloramphenicol. We have revised the manuscript to add these data (illustrative images + a quantitative analysis) in Figure 4 – figure supplement 1.

(2) The peripheral localisation of nucleoids observed after A22 treatment in this study and others (e.g., Japaridze et al., 2020; Wu et al., 2019), which conflicts with the model's assumptions and predictions. The assumption of radial confinement would predict nucleoids to fill up the volume or ribosomes to go near the cell wall, not the nucleoid, as seen in the data.

The reviewer makes a good point that DNA attachment to the membrane through transertion could contribute to the nucleoid being peripherally localized in A22 cells. We have revised the text to add this point. However, we do not think that this contradicts the proposed nucleoid segregation mechanism described in our model. On the contrary, by attaching the nucleoid to the cytoplasmic membrane along the cell width, transertion might help reduce the diffusion and thus exchange of polysomes across nucleoids. We have revised the text to discuss transertion over radial confinement.

(3) The radial compaction of the nucleoid upon rifampicin or chloramphenicol treatment, as reported by Bakshi et al. (2014) and Spahn et al. (2023), also contradicts the model's predictions. This is not expected if the nucleoid is already radially confined.

We originally evoked radial confinement to explain the observation that polysome accumulations do not equilibrate between DNA-free regions. We agree that transertion is an alternative explanation. Thank you for bringing it to our attention. However, please note that this does not contradict the model. In our view, it actually supports the 1D model by providing a reasonable explanation for the slow exchange of polysomes across DNA-free regions. The attachment of the nucleoid to the membrane along the cell width may act as diffusion barrier. We have revised the text and the title of the manuscript accordingly.

(4) Radial Distribution of Nucleoid and Ribosomal Shell:

The study does not account for well-documented features such as the membrane attachment of chromosomes and the ribosomal shell surrounding the nucleoid, observed in super-resolution studies (Bakshi et al., 2012; Sanamrad et al., 2014). These features are critical for understanding nucleoid dynamics, particularly under conditions of transcription-translation coupling or drug-induced detachment. Work by Yongren et al. (2014) has also shown that the radial organisation of the nucleoid is highly sensitive to growth and the multifork nature of DNA replication in bacteria.

We have revised the manuscript to discuss the membrane attachment. Please see the previous response.

The omission of organisation in the radial dimension and the entropic effects it entails, such as ribosome localisation near the membrane and nucleoid centralisation in expanded cells, undermines the model's explanatory power and predictive ability. Some observations have been previously explained by the membrane attachment of nucleoids (a hypothesis proposed by Rabinovitch et al., 2003, and supported by experiments from Bakshi et al., 2014, and recent super-resolution measurements by Spahn et al.).

We agree—we have revised the text to discuss membrane attachment in the radial dimension. See previous responses.

Ignoring the radial dimension and membrane attachment of nucleoid (which might coordinate cell growth with nucleoid expansion and segregation) presents a simplistic but potentially misleading picture of the underlying factors.

Please see above.

This reviewer suggests that the authors consider an alternative mechanism, supported by strong experimental evidence, as a potential explanation for the observed phenomena:

Nucleoids may transiently attach to the cell membrane, possibly through transertion, allowing for coordinated increases in nucleoid volume and length alongside cell growth and DNA replication. Polysomes likely occupy cellular spaces devoid of the nucleoid, contributing to nucleoid compaction due to mutual exclusion effects. After the nucleoids separate following ter separation, axial expansion of the cell membrane could lead to their spatial separation.

This “membrane attachment/cell elongation” model is reminiscent to the hypothesis proposed by Jacob et al in 1963 (doi:10.1101/SQB.1963.028.01.048). There are several lines of evidence arguing against it as the major driver of nucleoid segregation:

(Below is a slightly modified version of our response to a comment from Reviewer 1—see page 3)

(1) For this alternative model to work, axial membrane expansion (i.e., cell elongation) would have to be localized at the middle of the splitting nucleoids (i.e., midcell position for slow growth and ¼ and ¾ cell positions for fast growth) to create a directional motion. To our knowledge, there is no evidence of such localized membrane incorporation. Furthermore, even if membrane growth was localized at the right places, the fluidity of the cytoplasmic membrane (PMID: 6996724, 20159151, 24735432, 27705775) would be problematic. To go around this fluidity issue, one could potentially evoke a potential connection to the rigid peptidoglycan, but then again, peptidoglycan growth would have to be localized at the middle of the splitting nucleoid to “push” the sister nucleoid apart from each other. However, peptidoglycan growth is dispersed prior to cell constriction (PMID: 35705811, 36097171, 2656655).

(2) Even if we ignore the aforementioned caveats, Paul Wiggins’s group ruled out the cell elongation/transertion model by showing that the rate of cell elongation is slower than the rate of chromosome segregation (PMID: 23775792). In the revised manuscript, we confirm that the cell elongation rate is indeed overall slower than the nucleoid segregation rate (see Figure 1 - figure supplement 5A where the subtraction of the cell elongation rate to the nucleoid segregation rate at the single-cell level leads to positive values).

(3) Furthermore, our correlation analysis comparing the rate of nucleoid segregation to the rate of either cell elongation or polysome accumulation argues that polysome accumulation plays a larger role than cell elongation in nucleoid segregation. These data were already shown in the original manuscript (Figure 1I and Figure 1 – figure supplement 5B) but were not highlighted in this context. We have revised the text to clarify this point.

(4) The membrane attachment/cell elongation model does not explain the nucleoid asymmetries described in our paper (Figure 3), whereas they can be recapitulated by our model.

(5) The cell elongation/transertion model cannot predict the aberrant nucleoid dynamics observed when chromosomal expression is largely redirected to plasmid expression (Figure 7). In the revised manuscript, we have added simulation results showing that these nucleoid dynamics are predicted by our model (Figure 7 – figure supplement 2).

Based on these arguments, we do not believe that a mechanism based on membrane attachment and cell elongation is the major driver of nucleoid segregations. However, we do believe that it may play a complementary role (see “Nucleoid segregation likely involves multiple factors” in the Discussion). We have revised the text to clarify our thoughts and mention the potential role of transertion.

Incorporating this perspective into the discussion or future iterations of the model may provide a more comprehensive framework that aligns with the experimental observations in this study and previous work.

As noted above, we have revised the text to mention transertion.

Simplification of Ribosome States:

Combining monomeric and translating ribosomes into a single 'polysome' category may overlook spatial variations in these states, particularly during ribosome accumulation at the mid-cell. Without validating uniform mRNA distribution or conducting experimental controls such as FRAP or single-molecule measurements to estimate the proportions of ribosome states based on diffusion, this assumption remains speculative.

Indeed, for simplicity, we adopt an average description of all polysomes with an average diffusion coefficient and interaction parameters, which is sufficient for capturing the fundamental mechanism underlying nucleoid segregation. To illustrate that considering multiple polysome species does not change the physical picture, we have considered an extension of our model, which contains three polysome species, each with a different diffusion coefficient (D_P = 0.018, 0.023, or 0.028 μm²/s), reflecting that polysomes with more ribosomes will have a lower diffusion coefficient. Simulation of this model reveals that the different polysome species have essentially the same concentration distribution, suggesting that the average description in our minimal model is sufficient for our purposes. We present these new simulation results in Figure 4 – figure supplement 4 of the revised manuscript.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) Does the polysome density correlate with the origins? If the majority of ribosomal genes are expressed near the origins,

This is indeed an interesting point that we mention in the discussion. The fact that the chromosomal origin is surrounded by highly expressed genes (PMID: 30904377) and is located near the middle of the nucleoid prior to DNA replication (PMID: 15960977, 27332118, 34385314, 37980336) can only help the model that we propose by increasing the polysome density at the mid-nucleoid position.

(2) Red lines in 3C are hard to resolve - can the authors make them darker?

Absolutely. Sorry about that.

Reviewer #2 (Recommendations for the authors):

The authors use rifampicin treatment as a mechanism to trigger polysome disassembly and show this leads to homogenous RplA distribution. This is a really important experiment as it is used to link RplA localization to polysomes, and tp argue that RplA density is reporting on polysomes. Given rifampicin inhibits RNA polymerase, and given the only reference of the three linking rifampicin to polysome disassembly is the 1971 Blundell and Wild ref), it would perhaps be useful to more conclusively show that polysome depletion (as opposed to inhibition of mRNA synthesis, which is upstream of polysome assembly) by using an alternative compound more commonly linked to polysome disassembly (e.g., puromycin) and show timelapse loss of density as a function of treatment time. This is not a required experiment, but given the idea that RplA density reports on polysomes is central to the authors' interpretation, it feels like this would be a thing worth being certain of. An alternative model is that ribosomes undergo self-assembly into local storage depots when not being used, but those depots are not translationally active/lack polysomes. I don't know if I think this is likely, but I'm not convinced the rifampicin treatment + waiting for a relatively long period of time unambiguously excludes other possible mechanisms given the large scale remodeling of the intracellular environment upon mRNA inhibition. I 100% buy the relationship between ribosomal distribution and nucleoid segregation (and the ectopic expression experiments are amazing in this regard), so my own pause for thought here is "do we know those ribosomes are in polysomes in the ribosome-dense regions". I'm not sure the answer to this question has any bearing on the impact and importance of this work (in my mind, it doesn't, but perhaps there's a reason it does?). The way to unambiguously show this would really be to do CryoET and show polysomes in the dense ribosomal regions, but I would never suggest the authors do that here (that's an entire other paper!).

We agree that mRNAs play a role, as mRNAs are major components of polysomes and most mRNAs are expected to be in the form of polysomes (i.e., in complex with ribosomes). In addition, as mentioned above, the enrichments of ribosome distribution are known to be associated with polysomes (PMID: 25056965, 38678067, 22624875, 31150626, 34186018, 10675340). The attribution is consistent with single-molecule tracking experiments showing that slow-moving ribosomes (polysomes) are excluded by the nucleoid whereas fast-diffusing ribosomes (free ribosomal subunits) are distributed throughout the cytoplasm (PMID: 25056965, 22624875). This is also consistent with cryo-ET results that we actually published (see Figure S5, PMID: 34186018). We have added this information to the revised manuscript. Thank you for alerting us of this oversight.

On line 320 the authors state "Our single-cell studies provided experimental support that phase separation between polysomes and DNA contributes to nucleoid segregation." - this comes pretty out of left field? I didn't see any discussion of this hypothesis leading up to this sentence, nor is there evidence I can see that necessitates phase separation as a mechanistic explanation unless we are simply using phase separation to mean cellular regions with distinct cellular properties (which I would advise against). If the authors really want to pursue this model I think much more support needs to be provided here, including (1) defining what the different phases are, (2) providing explicit description of what the attractive/repulsive determinants of these different phases could be/are, and (3) ruling out a model where the behavior observed is driven by a combination of DNA / polysome entanglement + steric exclusion; if this is actually the model, then being much more explicit about this being a locally arrested percolation phenomenon would be essential. Overall, however, I would probably dissuade the authors from pursuing the specific underlying physics of what drives the effects they're seeing in a Results section, solely because I think ruling in/out a model unambiguously is very difficult. Instead, this would be a useful topic for a Discussion, especially couched under a "our data are consistent with..." if they cannot exclude other models (which I think is unreasonably difficult to do).

Thank you for your advice. We have revised the text to more carefully choose our words and define our terms.

Minor comments:

The results in "Cell elongation may also contribute to sister nucleoid migration near the end of the division cycle" are really interesting, but this section is one big paragraph, and I might encourage the authors to divide this paragraph up to help the reader parse this complex (and fascinating) set of results!

We have revised this section to hopefully make it more accessible.

Reviewer #3 (Recommendations for the authors):

Technical Controls:

The authors should conduct a photobleaching control to confirm that the perceived 'higher' brightness of new ribosomes at the mid-cell position is not an artefact caused by older ribosomes being photobleached during the imaging process. Comparing results at various imaging frequencies and intensities is necessary to address this issue.

The ribosome localization data across 30 nutrient conditions (Figure 2, Figure 1 – figure supplement 6, Figure 2 – Figure supplement 1, Figure 2 – Figure supplement 3 and Figure 5) are from snapshot images, which do not have any photobleaching issue. They confirm the mid-cell accumulation seen by time-lapse microscopy. We have revised the text to clarify this point.

Novelty of Experimental Measurements:

While the scale of the study is unprecedented, claims of novelty (e.g., line 142) regarding ribosome-nucleoid segregation tracking are overstated. Similar observations have been made previously (e.g., Bakshi et al., 2012; Bakshi et al., 2014; Chai et al., 2014).

Our apologies. The text in line 142 oversimplified our rationale. This has been corrected in the revised manuscript.

I think this paragraph is really true and powerful. We often think that "identity" is a label that others put on us, but in fact, we ourselves are constantly participating in, responding to, and even internalizing these labels to some extent. Lydia's sentence "I hate it!" really made me feel the conflict - she was given a label that "helped" her, but her feelings were not really understood. It is too easy for us to use words like "special needs" as neutral words, but ignore the oppression that may be brought to the person involved. I like the author's reminder that identity is complex and dynamic, not a fixed definition.

Sometimes the advantages we have are "invisible". Things like skin color, social class and language, which we may not pay much attention to in our daily life, do quietly influence our experiences at school, such as whether we are misunderstood or easily understood and supported by teachers. The author's admission of her privilege is not to deny the difficulties she is facing, but to present a more comprehensive and honest educational perspective. I think this kind of self-awareness is also very important in the school environment, especially for us students. Only by learning to recognize our own position can we better understand the situation of others.

Author Response

The following is the authors’ response to the original reviews.

Reviewer 1:

Public review:

In this study, Porter et al report on outcomes from a small, open-label, pilot randomized clinical trial comparing dornase-alfa to the best available care in patients hospitalized with COVID-19 pneumonia. As the number of randomized participants is small, investigators describe also a contemporary cohort of controls and the study concludes about a decrease of inflammation (reflected by CRP levels) aJer 7 days of treatment but no other statistically significant clinical benefit.

Suggestions to the authors:

• The RCT does not follow CONSORT statement and reporting guidelines

We thank you for this suggestion and have now amended the order and content of the manuscript to follow the CONSORT statement as closely as possible.

• The authors have chosen a primary outcome that cannot be at least considered as clinically relevant or interesting. AJer 3 years of the pandemic with so much research, why investigate if a drug reduces CRP levels as we already have marketed drugs that provide beneficial clinical outcomes such as dexamethasone, anakinra, tocilizumab and baricitinib.

We thank the reviewer for bringing up this central topic. The answer to this question has both a historical and practical component. This trial was initiated in June of 2020 and was completed in June of 2021. At that time there were no known treatments for the severe immune pathology of COVID19 pneumonia. In June 2020, dexamethasone data came out and we incorporated dexamethasone into the study design. It took much longer for all other anti-inflammatories to be tested. Hence, our decision to trial an approved endonuclease was based purely on basic science work on the pathogenic role of cell-free chromatin and NETs in murine sepsis and flu models and the ability of DNase I to clear them and reduce pathology in these animal models. In addition, evidence for the presence of cell-free chromatin components in COVID-19 patient plasma had already been communicated in a pre-print. Finally, several studies had reported the anti-inflammatory effects of dornase treatment in CF patients. Hence there was a strong case for a cheap, safe, pulmonary noninvasive treatment that could be self-administered outside the clinical se]ng.

The Identification of novel/repurposed treatments effective for COVID-19 were hampered by patient recruitment to competing studies during a pandemic. This resulted in small studies with inconclusive or contrary findings. In general, effective treatments were only picked up in very large RCTs. For example, demonstrating dexamethasone as effective in COVID-19 required recruitment of 6,425 patients into the RECOVERY study. Multiple trials with anti-IL-6 gave conflicting evidence until RECOVERY recruited 4116 adults with COVID-19 (n=2022, tocilizumab and 2094, control) similar for Baracitinib (4,148 randomised to treatment and 4,008 to standard care). Anakinra is approved for patients with elevated suPAR, based on data from one randomized clinical trial of 594 patients, of whom 405 had active treatment (PMID: 34625750). However, a systematic review analysing over 1,627 patients (anakinra 888, control 739) with COVID-19 showed no benefit (PMID: 36841793). Regarding the choice of the primary endpoint, there is a wealth of clinical evidence to support the relevance of CRP as a prognostic marker for COVID-19 pneumonia patients and it is a standard diagnostic and prognostic clinical parameter in infectious disease wards. This choice in March 2020 was supported by evidence of the prognostic value of IL-6; CRP is a surrogate of IL-6. We also provide our own data from a large study of severe COVID-19 pneumonia in figure 1, showing how well CRP correlates with survival.

In summary, our data suggest that Dornase yields an anti-inflammatory effect that is comparable or potentially superior to cytokine-blocking monotherapies at a fraction of the cost and potentially without the additional adverse effects such as the increase for co-infections.

We now provide additional justification on these points in the introduction on pg.4 as follows:

“The trial was ini.ated in June 2020 and was completed in September of 2021. At the start of the trial only dexamethasone had been proven to benefit hospitalized COVID-19 pneumonia pa.ents and was thus included in both arms of the trial. To increase the chance of reaching significance under challenging constraints in pa.ent access, we opted to increase our sample size by using a combina.on of randomized individuals and available CRP data from matched contemporary controls (CC) hospitalized at UCL but not recruited to a trial. These approaches demonstrated that when combined with dexamethasone, nebulized DNase treatment was an effec.ve an.-inflammatory treatment in randomized individuals with or without the implementa.on of CC data.”

We also added the following explanation in the discussion on pg. 16:

“Our study design offered a solution to the early screening of compounds for inclusion in larger platform trials. The study took advantage of frequent repeated measures of quantifiable CRP in each patient, to allow a smaller sample size to determine efficacy/futility than if powered on clinical outcomes. We applied a CRP-based approach that was similar to the CATALYST and ATTRACT studies. CATALYST showed in much smaller groups (usual care, 54, namilumab, 57 and infliximab, 35) that namilumab that is an antibody that blocks the cytokine GM-CSF reduced CRP even in participants treated with dexamethasone whereas infliximab that targets TNF-α had no significant effect on CRP. This led to a suggestion that namilumab should be considered as an agent to be prioritised for further investigation in the RECOVERY trial. A direct comparison of our results with CATALYST is difficult due to the different nature of the modelling employed in the two studies. However, in general Dornase alfa exhibited comparable significance in the reduction in CRP compared to standard of care as described for namilumab at a fraction of the cost. Furthermore, endonuclease therapies may prove superior to cytokine blocking monotherapies, as they are unlikely to increase the risk for microbial co-infections that have been reported for antibody therapies that neutralize cytokines that are critical for immune defence such as IL-1β, IL-6 or GM-CSF. “

• Please provide in Methods the timeframe for the investigation of the primary endpoint

This information is provided in the analysis on pg. 8:

“The primary outcome was the least square (LS) mean CRP up to 7 days or at hospital discharge whichever was sooner.”

• Why day 35 was chosen for the read-out of the endpointt?

We now state on pg. 8 that “Day 35 was chosen as being likely to include most early mortality due to COVID-19 being 4 weeks after completion of a week of treatment. ( i.e. d7 of treatment +28 (4 x 7 days))”

• The authors performed an RCT but in parallel chose to compare also controls. They should explain their rationale as this is not usual. I am not very enthusiastic to see mixed results like Figures 2c and 2d.

We initially aimed at a fully randomized trial. However, the swiJ implementation of trial prioritization strategies towards large and pre-established trial plamorms in the UK made the recruitment COVID19 patients to small studies extremely challenging. Thus, we struggled to gain access to patients. Our power calculations suggested that a mixed trial with randomized and contemporary controls was the best way forward under these restrictions in patient access that could provide sufficient power.

That being said, we also provide the primary endpoint (CRP) results in Fig. 3B as well as the results for the length of hospitalization (Fig. S3D) for the randomized subjects only.

• Analysis is performed in mITT; this is a major limitation. The authors should provide at least ITT results. And they should describe in the main manuscript why they chose mITT analysis.

We apologize if this point was confusing. We performed the analysis on the ITT as defined in our SAP: “The primary analysis population will be all evaluable patients randomised to BAC + dornase alfa or BAC only who have at least one post-baseline CRP measurement, as well as matched historical comparators.”

We understand that the reason this might be mistaken as an mITT is because the N in the ITT (39) doesn’t match the number randomised and because we had stated on pg. 8 that “ Efficacy assessments of primary and secondary outcomes in the modified inten.on-to-treat popula.on were performed.”

However, we did randomise 41 participants, but:

One participant in the DA arm never received treatment. The individual withdrew consent and was replaced. We also have no CRP data for this participant in the database, so they were unevaluable, and we couldn’t include them in the baseline table even if we wanted to. In addition, 1 participant in BAC only had a baseline CRP measurement available. Hence not evaluable as we only have a baseline CRP measurement for this participant.

We have corrected the confusing statement on pg. 8 and added an additional explanation.

“Efficacy assessments of primary and secondary outcomes in the inten.on-to-treat (ITT) popula.on were performed on all randomised par.cipants who had received at least one dose of dornase alfa if randomized to treatment. For full details see Sta.s.cal Analysis Plan. The ITT was adjusted to mi.gate the following protocol viola.ons where one par.cipant in the BAC arm and one in the DA arm withdrew before they received treatment and provided only a baseline CRP measurement available. The par.cipant in the DA arm was replaced with an addi.onal recruited pa.ent. Exploratory endpoints were only available in randomised par.cipants and not in the CC. In this case, a post hoc within group analysis was conducted to compare baseline and post-baseline measurements.”

• It is also not usual to exclude patients from analysis because investigators just do not have serial measurements. This is lost to follow up and investigators should have pre-decided what to do with lost-to-follow-up.

Our protocol pre-specified that the primary analysis population should have at least one postbaseline CRP measurement (pg. 13 of protocol). The patient that was excluded was one that initially joined the trial but withdrew consent after the first treatment but before the first post-treatment blood sample could be drawn. Hence, the pre-treatment CRP of this patient alone provided no useful information.

• In Table 1 I would like to see all randomized patients (n=39), which is missing. There are also baseline characteristics that are missing, like which other treatments as BAT received by those patients except for dexamethasone.

Table 1 includes all 39 patients plus 60 CCs.<br /> Table 2 shows additional treatments given for COVID-19 as part of BAC.

• In the first paragraph of clinical outcomes, the authors refer to a cohort that is not previously introduced in the manuscript. This is confusing. And I do not understand why this analysis is performed in the context of this RCT although I understand its pilot nature.

One of the main criticisms we have encountered in this study has been the choice of the primary endpoint. The best way respond to these questions was to provide data to support the prognostic relevance of CRP in COVID-19 pneumonia from a separate independent study where no other treatments such as dexamethasone, anakinra or anti-IL6 therapies were administered. We think this is very useful analysis and provides essential context for the trial and the choice of the primary endpoint, indicating that CRP has good enough resolution to predict clinical outcomes.

• Propensity-score selected contemporary controls may introduce bias in favor of the primary study analysis, since controls are already adjusted for age, sex and comorbidities.

The contemporary controls were selected to best match the characteristics of the randomized patients including that the first CRP measurement upon admission surpassed the trial threshold, so we do not see how this selection process introduces biases, as it was blinded with regards to the course of the CRP measurements. Given that this was a small trial, matching for baseline characteristics is necessary to minimize confounding effects.

• The authors do not clearly present numerically survivors and non-survivors at day 34, even though this is one of the main secondary outcomes.

We now provide the mortality numbers in the following paragraph on pg. 13.

“Over 35 days follow up, 1 person in the BAC + dornase-alfa group died, compared to 8 in the BAC group. The hazard ra.o observed in the Cox propor.onal hazards model (95% CI) was 0.47 (0.06, 3.86), which es.mates that throughout 35 days follow-up, there was a 53% reduced chance of death at any given .mepoint in the BAC + dornase-alfa group compared to the BAC group, though the confidence intervals are wide due to a small number of events. The p-value from a log-rank test was 0.460, which does not reach sta.s.cal significance at an alpha of 0.05.”

• It is unclear why another cohort (Berlin) was used to associate CRP with mortality. CRP association with mortality should (also) be performed within the current study.

As we explained above, the Berlin cohort CRP data serve to substantiate the relevance of CRP as a primary endpoint in a cohort that experienced sufficient mortality as this cohort did not receive any approved anti-inflammatory therapy. Mortality in our COVASE trial was minimal, since all patients were on dexamethasone and did not reach the highest severity grade, since we opted to treat patients before they deteriorated further. The overall mortality was 8% across all arms of our study, which does not provide enough events for mortality measurements. In contrast the Berlin cohort did not receive dexamethasone and all patients had reached a WHO severity grade 7 category with mortality at 30%.

My other concerns are:

• This report is about an RCT and the authors should follow the CONSORT reporting guidelines. Please amend the manuscript and Figure 1b accordingly and provide a CONSORT checklist.

We now provide a CONSORT checklist and have amended the CONSORT diagram accordingly.

• Please provide in brief the exclusion criteria in the main manuscript

We have now included the exclusion criteria in the manuscript on pg. 6.

“1.1.1 Exclusion criteria

1. Females who are pregnant, planning pregnancy or breasmeeding

2. Concurrent and/or recent involvement in other research or use of another experimental inves.ga.onal medicinal product that is likely to interfere with the study medica.on within (specify .me period e.g. last 3 months) of study enrolment 3. Serious condi.on mee.ng one of the following:

a. Respiratory distress with respiratory rate >=40 breaths/min

b. oxygen satura.on<=93% on high-flow oxygen

1. Require mechanical invasive or non-invasive ven.la.on at screening

2. Concurrent severe respiratory disease such as asthma, COPD and/or ILD

3. Any major disorder that in the opinion of the Inves.gator would interfere with the evalua.on of the results or cons.tute a health risk for the trial par.cipant

4. Terminal disease and life expectancy <12 months without COVID-19

5. Known allergies to dornase alfa and excipients

6. Par.cipants who are unable to inhale or exhale orally throughout the en.re nebulisa.on period So briefly Patients were excluded if they were:

7. pregnant, planning pregnancy or breasmeeding

8. Serious condition meeting one of the following:

a. Respiratory distress with respiratory rate >=40 breaths/min

b. oxygen satura.on<=93% on high-flow oxygen

1. Require ven.la.on at screening

2. Concurrent severe respiratory disease such as asthma, COPD and/or ILD

3. Terminal disease and life expectancy <12 months without COVID-19

4. Known allergies to dornase alfa and excipients

5. Participants who are unable to inhale or exhale orally throughout the en.re nebulisa.on period”

• "The final trial visit occurred at day 35." "Analysis included mortality at day 35". I am not sure I understand why. In clinicaltrials.gov all endpoints are meant to be studies at day 7 except for mortality rate day 28. Why day 35 was chosen? Please be consistent.

Thank you for identifying this inconsistency. We have amended the record on clinicaltrials.gov to read ‘’the time to event data was censored at 28 days post last dose (up to d35) for the randomised participants and at the date of the last electronic record for the CC.”

• Please provide in Methods the timeframe for the investigation of the primary endpoint

• The authors performed an RCT but in parallel chose to compare also controls. They should explain their rationale as this is not usual. I am not very enthusiastic to see mixed results like Figures 2c and 2d.

• Analysis is performed in mITT; this is a major limitation. The authors should provide at least ITT results. And they should describe in the main manuscript why they chose mITT analysis.

• It is also not usual to exclude patients from analysis because investigators just do not have serial measurements. This is lost to follow up and investigators should have pre-decided what to do with lost-to-follow-up.

• Figure 1b as in CONSORT statement, please provide reasons why screened patients were not enrolled.

• In Table 1 I would like to see all randomized patients (n=39), which is missing. There are also baseline characteristics that are missing, like which other treatment as BAT received those patients except for dexamethasone.

• In the first paragraph of clinical outcomes, the authors refer to a cohort that is not previously introduced in the manuscript. This is confusing. And I do not understand why this analysis is performed in the context of this RCT although I understand its pilot nature.

• In Figure 2 the authors draw results about ITT although in methods describe that they performed an mITT analysis. Please be consistent.

Please see answers provided to these queries above.

Reviewer #2 (Recommendations For The Authors):

1) Suppl Figure 2B would be more informative if presented as a Table with N of patients with per day sampling

We now provide the primary end point daily sampling table in Table 3.

2) The numbers at risk should figure under the KM curves

The numbers at risk for figures 1E, 2C, 2D have been added as graphs either in the main figures or in the supplement.

3) HD in Supplementary figure 3 should be explained

We apologize for this omission. We now provide a description for the healthy donor samples that we used in the cell-free DNA measurements in figure S3B on pg. 14:

“Compared to the plasma of anonymized healthy donors volunteers at the Francis Crick ins.tute (HD), plasma cf-DNA levels were elevated in both BAC and DA-treated COVASE par.cipants.

4) Presentation is inappropriate for Table S4

We thank the reviewer for pointing this issue. We have now formaxed Table S4 to be consistent with all other tables.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript is a focused investigation of the phosphor-regulation of a C. elegans kinesin-2 motor protein, OSM-3. In C-elegans sensory ciliary, kinesin-2 motor proteins Kinesin-II complex and OSM-3 homodimer transport IFT trains anterogradely to the ciliary tip. Kinesin-II carries OSM-3 as an inactive passenger from the ciliary base to the middle segment, where kinesin-II dissociates from IFT trains and OSM-3 gets activated and transports IFT trains to the distal segment. Therefore, activation/inactivation of OSM-3 plays an essential role in its ciliary function.

Strengths:

In this study, using mass spectrometry, the authors have shown that the NEKL-3 kinase phosphorylates a serine/threonine patch at the hinge region between coiled coils 1 and 2 of an OSM-3 dimer, referred to as the elbow region in ubiquitous kinesin-1. Phosphomimic mutants of these sites inhibit OSM-3 motility both in vitro and in vivo, suggesting that this phosphorylation is critical for the autoinhibition of the motor. Conversely, phospho-dead mutants of these sites hyperactivate OSM-3 motility in vitro and affect the localization of OSM3 in C. elegans. The authors also showed that Alanine to Tyrosine mutation of one of the phosphorylation rescues OS-3 function in live worms.

Weaknesses:

Collectively, this study presents evidence for the physiological role of OSM-3 elbow phosphorylation in its autoregulation, which affects ciliary localization and function of this motor. Overall, the work is well performed, and the results mostly support the conclusions of this manuscript. However, the work will benefit from additional experiments to further support conclusions and rule out alternative explanations, filling some logical gaps with new experimental evidence and in-text clarifications, and improving writing before I can recommend publication.

We appreciate Reviewer #1’s comments and suggestions. We have now provided additional evidences and discussions to further support our conclusions and fill the logical gaps. We have also provided alternative explanations to our data and improved writing.

Reviewer #2 (Public review):

Summary:

The regulation of kinesin is fundamental to cellular morphogenesis. Previously, it has been shown that OSM-3, a kinesin required for intraflagellar transport (IFT), is regulated by autoinhibition. However, it remains totally elusive how the autoinhibition of OSM-3 is released. In this study, the authors have shown that NEKL-3 phosphorylates OSM-3 and releases its autoinhibition.

The authors found NEKL-3 directly phosphorylates OSM-3 (although the method is not described clearly) (Figure 1). The phophorylated residue is the "elbow" of OSM-3. The authors introduced phospho-dead (PD) and phospho-mimic (PM) mutations by genome editing and found that the OSM-3(PD) protein does not form cilia, and instead, accumulates to the axonal tips. The phenotype is similar to another constitutive active mutant of OSM-3, OSM-3(G444A) (Imanishi et al., 2006; Xie et al., 2024). osm-3(PM) has shorter cilia, which resembles with loss of function mutants of osm-3 (Figure 3). The authors did structural prediction and showed that G444E and PD mutations change the conformation of OSM-3 protein (Figure 3). In the single-molecule assays G444E and PD mutations exhibited increased landing rate (Figure 4). By unbiased genetic screening, the authors identified a suppressor mutant of osm-3(PD), in which A489T occurs. The result confirms the importance of this residue. Based on these results, the authors suggest that NEKL-3 induces phosphorylation of the elbow domain and inactivates OSM-3 motor when the motor is synthesized in the cell body. This regulation is essential for proper cilia formation.

Strengths:

The finding is interesting and gives new insight into how the IFT motor is regulated.

Weaknesses:

The methods section has not presented sufficient information to reproduce this study.

We appreciate that Reviewer #2 is also positive to our study. We have now provided sufficient information in the revised Methods section.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Major Concerns

(1) Why do the authors think that NEKL-3 phosphorylates OSM-3 in the first place? This seems to come out of nowhere and prior evidence indicating that NEKL-3 may be phosphorylating OSM-3 is not even mentioned in the Introduction.

We thank the Reviewer for raising this important point. Our hypothesis that NEKL-3 phosphorylates OSM-3 stems from prior findings in our lab. In a previous study (Yi et al., Traffic, 2018, PMID: 29655266), we identified NEKL-4, a member of the NIMA kinase family, as a suppressor of the OSM-3(G444E) hyperactive mutation. This discovery prompted us to explore the broader role of NIMA kinases in regulating OSM3. Subsequent genetic screens (Xie et al., EMBO J, 2024, PMID: 38806659) revealed that both NEKL-3 and NEKL-4 suppress multiple OSM-3 mutations, further supporting their functional interaction. Given the established role of NIMA kinases in phosphorylation-dependent processes (Fry et al., JCS, 2012, PMID: 23132929; Chivukula et al., Nat. Med., 2020, PMID: 31959991; Thiel, C. et al. Am. J. Hum. Genet. 2011, PMID: 21211617; Smith, L. A. et al., J. Am. Soc. Nephrol., 2006, PMID: 16928806), we hypothesized that NEKL-3/4 may directly phosphorylate OSM-3 to modulate its activity.

To test this hypothesis, we expressed recombinant C. elegans NEKL-3 and OSM-3 proteins and conducted in vitro phosphorylation assays. While we were unable to obtain active recombinant NEKL-4 (limitations noted in the revised text), our experiments with NEKL-3 revealed phosphorylation at residues 487-490 (YSTT motif) in OSM-3’s tail region, as confirmed by mass spectrometry. These findings are now explicitly contextualized in the Introduction and Results sections of the revised manuscript.

Page #4, Line #11:

“...In our previous study (Yi et al., Traffic, 2018, PMID: 29655266), a genetic screen targeting the OSM-3(G444E) hyperactive mutation identified NEKL-4, a member of the NIMA kinase family, as a suppressor of this phenotype. This finding, combined with reports that NIMA kinases regulate ciliary processes independently of their canonical mitotic roles (Fry et al., JCS, 2012, PMID: 23132929; Chivukula et al., Nat. Med., 2020, PMID: 31959991; Thiel, C. et al. Am. J. Hum. Genet. 2011, PMID: 21211617; Smith, L. A. et al., J. Am. Soc. Nephrol., 2006, PMID: 16928806), prompted us to investigate whether NIMA kinases modulate OSM-3-driven intraflagellar transport. We hypothesized that NEKL-3/4, as paralogs within this family, might directly phosphorylate OSM-3 to regulate its motility...”

Page #4, line #26:  

“... To determine whether NIMA kinase family members could directly phosphorylate

OSM-3, we purified prokaryotic recombinant C. elegans NEKL-3/NEKL-4 and OSM3 protein in order to perform in vitro phosphorylation assays. We were able to obtain active recombinant NEKL-3 but not NEKL-4. The in vitro phosphorylation assays showed that NEKL-3, directly phosphorylates OSM-3 (Fig. 1A-B, Appendix Table S1). Subsequent mass spectrometric analysis revealed phosphorylation at residues 487-490, which localize to the conserved "YSTT" motif within OSM-3’s C-terminal tail region ...”

(2) The authors need to characterize the proteins they expressed and purified for in vitro ATPase and motility assays. Are these proteins monomers or dimers?

For our in vitro ATPase and motility assays, OSM-3 was expressed in E. coli BL21(DE3) and purified using established protocols (Xie et al., EMBO J, 2024, PMID: 38806659; Imanishi et al., JCB, 2006, PMID: 17000874). To confirm its oligomeric state, we analyzed recombinant OSM-3 by size-exclusion chromatography coupled with multiangle light scattering (SEC-MALS). As reported in Xie et al. (2024), OSM-3 (~80 kDa monomer) elutes with a molecular weight of 173–193 kDa under physiological buffer conditions, consistent with a homodimeric assembly. These findings confirm that the functional unit used in our assays is the biologically relevant dimer. This characterization has been added to the revised manuscript on Page #35, Line #7.

“…OSM-3 was expressed in E. coli BL21(DE3) and purified for in vitro assays using established protocols (REFs). Size-exclusion chromatography coupled with multiangle light scattering (SEC-MALS) (Xie et al., EMBO J., 2024) confirmed that recombinant OSM-3 forms a homodimer (173–193 kDa) under physiological conditions, ensuring its dimeric state remained intact....” 

(3) The authors primarily used PD and PM mutations, which affect all four amino acids in the region. This may or may not be physiologically relevant. Figure 5 indicates that T489 is a critical regulatory site. However, this conclusion is undermined by reliance on PD mutations, which affect all four amino acids. Creating PM (T489E) and PD (T489A) mutations based on WT OSM-3 would better reflect physiological relevance. In vitro assays with a single phosphomimic or phosphor-dead mutation at residue 489 are missing at the end of this story. This would better link Figure 5 with the rest of the manuscript.

We thank the reviewer for this constructive critique. Below, we address the concerns and integrate new data to strengthen the link between T489 and autoinhibition:

To probe the regulatory role of T489 phosphorylation, we generated osm-3(T489E) (phosphomimetic, PM) and osm-3(T489A) (phospho-dead, PD) mutant animals. Strikingly, both mutants formed axonal puncta (Figure S7), recapitulating the hyperactive phenotype of the OSM-3G444E mutant. While the similar puncta formation in PM and PD mutants initially appeared paradoxical, this observation underscores the necessity of dynamic phosphorylation cycling at T489 for proper autoinhibition. Specifically, the PD mutant (T489A) likely disrupts phosphorylationdependent autoinhibition stabilization, leading to constitutive activation, where as the PM mutant (T489E) may mimic a "locked" phosphorylated state, preventing dephosphorylation-dependent release of autoinhibition in cilia and trapping OSM-3 in an aggregation-prone conformation. These results highlight T489 as a structural linchpin whose post-translational modification dynamically regulates motor activity. While the precise molecular mechanism—such as how phosphorylation modulates tailmotor domain interactions—remains to be elucidated, our data conclusively demonstrate that perturbing T489 (even in isolation) destabilizes autoinhibition, driving puncta formation and the constitutive activity.

We have integrated the above paragraph in the revised manuscript on page #8, line #27.

(4) There seems to be a disconnect between the MT gliding assays in Figure 4C and single molecule motility assays in Figure 4E. The gliding assays show that all constructs can glide microtubules at near WT speeds. Yet, the motility assays show that WT and PM cannot land or walk on MTs. The authors need to explain why this is the case. Is this because surface immobilization of kinesin from its tail disrupts autoinhibition? Alternatively, the protein preparation may include monomers that cannot be autoinhibited and cannot land and processively walk on surface-immobilized microtubules (because they only have one motor domain) but can glide microtubules when immobilized on the surface from their tail.

The surface immobilization of OSM-3 via its tail domain disrupts autoinhibition, a phenomenon previously observed in other kinesins such as kinesin-1 (Nitzsche et al, Methods Cell Biol., 2010, PMID: 20466139). In our assays, OSM-3 was nonspecifically immobilized on glass surfaces, enabling microtubule gliding by motors whose autoinhibition was relieved through tail anchoring. Critically, the PD and PM mutations reside in the tail region and do not alter the intrinsic properties of the motor head domain. Consequently, once autoinhibition is released via immobilization, the gliding velocities reflect the conserved motor head activity, which is expected to remain comparable across all constructs. While we cannot entirely rule out the presence of monomeric OSM-3 in solution, several lines of evidence argue against this possibility. First, the mutations are located in the elbow region, which is dispensable for motor dimerization. Second, SEC-MALS analysis from prior studies confirms that purified OSM-3 exists predominantly as dimers in solution. 

We have discussed these issues in the revised text on page #10, line #18: 

“…In our gliding assays, OSM-3PM has an increased gliding speed of 0.69 ± 0.07 μm/s (Fig. 4 C-D), similar to PD mutant. PD and PM mutations are confined to the elbow region, leaving the motor head’s mechanochemical properties intact. Upon tail immobilization—which releases autoinhibition—the gliding speeds reflect motor head activity. Single-molecule assays, however, directly resolve their native regulatory states: PD mutants are constitutively active, whereas PM mutants persist in an autoinhibited state (Fig. 4E-G). Although monomeric OSM-3 could theoretically mediate singlemotor gliding, the previous SEC-MALS data demonstrate that OSM-3 purifies as stable dimers (Xie et al., EMBO J, 2024, PMID: 38806659). Thus, dimeric OSM-3 is perhaps the predominant functional species in our assays…”

(5) An alternative explanation for the data is that both PD and PM mutations result in loss-of-function effects, disrupting OSM-3 activity. For instance:

a) In Figure 2C, both mutations cause shorter cilia than the wild type (WT).

b) In Figure 4A, both mutations result in higher ATPase activity than WT.

c) In Figure 4D, both mutations show increased gliding velocity compared to WT. These results suggest the observed effects could stem from loss of function rather than phosphorylation-specific regulation.

Although PD and PM mutations exhibit superficially similar "loss-of-function" phenotypes in certain assays, they mechanistically disrupt motor regulation in distinct ways:

a) Ciliary Length (Figure 2C) PD Mutants: Hyperactivation causes OSM-3-PD to prematurely aggregate into axonal puncta, preventing ciliary entry. Consequently, cilia are built solely by the weaker Kinesin-II motor, which only constructs shorter middle segments.

PM Mutants: OSM-3-PM retains autoinhibition during transport (enabling ciliary entry) but cannot be dephosphorylated in cilia. This blocks activation, leaving OSM-3-PM partially functional and resulting in cilia intermediate in length between WT and PD.

We have discussed this issue in the revised text on page #5, line #30:

“…These findings indicate that OSM-3-PM is in an autoinhibited state capable of ciliary delivery, yet fails to achieve full activation due to defective dephosphorylation. This incomplete activation results in suboptimal motor function and intermediate ciliary length phenotypes (Fig.2 B-C). In contrast, OSM-3-PD exhibits constitutive activation leading to aggregation into axonal puncta, which completely abolishes its ciliary entry capacity (Fig.2 A-B)...”

b) ATPase Activity (Figure 4A)

PD Mutants: Fully autoinhibition-released (98.15% of KHC ATPase activity), consistent with constitutive activation.

PM Mutants: Show partial ATPase activity (34.28% of KHC), reflecting imperfect phosphomimicry. While the DDEE substitution introduces negative charges, it fails to fully replicate the steric/kinetic effects of phosphorylated tyrosine (Y486; phenyl ring absent), resulting in incomplete autoinhibition stabilization. Despite this, the residual inhibition is sufficient to phenocopy shorter cilia in vivo.

We have discussed this issue in the revised text on page #7, line#19:

“…The PM mutant’s partial ATPase activity (34.28% of KHC) might arise from imperfect phosphomimicry—while the DDEE substitution introduces negative charges, it lacks the steric bulk of phosphorylated tyrosine (pY487). And this incomplete mimicry allows residual autoinhibition, sufficient to limit ciliary construction in vivo...”

c) Microtubule Gliding Velocity (Figure 4D)

Gliding Assay Limitation: Tail immobilization artificially releases autoinhibition, masking regulatory differences. Thus, all constructs (PD, PM) exhibit similar velocities (~0.7 µm/s), reflecting conserved motor head activity.

Single-Molecule Assay (Figure 4E): Directly resolves native autoinhibition states:

PD mutants show robust motility (autoinhibition released).

PM mutants remain largely inactive (autoinhibition retained).

We have discussed this issue in the revised text on page #10, line#18:

“…In our gliding assays, OSM-3PM has an increased gliding speed of 0.69 ± 0.07 μm/s (Fig. 4 C-D), similar to PD mutant. PD and PM mutations are confined to the elbow region, leaving the motor head’s mechanochemical properties intact. Upon tail immobilization—which releases autoinhibition—the gliding speeds reflect motor head activity. Single-molecule assays, however, directly resolve their native regulatory states: PD mutants are constitutively active, whereas PM mutants persist in an autoinhibited state (Fig. 4E-G)...”

Minor Suggestions and Concerns

(1) Lines 60-66: References that support these observations are missing from this section.

We have added the relevant references.

(2) Lines 66-67: I would revise this sentence as "It remains unclear how OSM-3 becomes enriched...".

We have made the changes.

(3) Line 85: The authors should describe how they perform these assays (i.e. recombinantly expressed NEKL-3 and OSM-3, are these C. elegans proteins, and which expression system was used...).

We have described them in the main text and methods

Page #4 line #26

“...To determine whether NIMA kinase family members could directly phosphorylate OSM-3, we purified prokaryotic recombinant C. elegans NEKL-3/NEKL-4 and OSM-3 protein in order to perform in vitro phosphorylation assays...”

Page #35 line#12

“...Basically, point mutations was introduced in to pET.M.3C OSM-3-eGFP-His6 plasmid for prokaryotic expression. Plasmid transformed E. coli (BL21) was cultured at 37°C and induced overnight at 23°C with 0.2 mM IPTG. Cells were lysed in lysis buffer (50 mM NaPO4 pH8.0, 250 mM NaCl, 20 mM imidazole, 10 mM bME, 0.5 mM ATP, 1 mM MgCl¬2, Complete Protease Inhibitor Cocktail (Roche)) and Ni-NTA beads were applied for affinity purification. After incubation, beads were washed with wash buffer (50 mM NaPO4 pH6.0, 250 mM NaCl, 10 mM bME, 0.1 mM ATP, 1 mM MgCl¬2) and eluted with elute buffer (50 mM NaPO4 pH7.2, 250 mM NaCl, 500 mM imidazole, 10 mM bME, 0.1 mM ATP, 1 mM MgCl¬2). Protein concentration was determined by standard Bradford assay. C elegans nekl-3 cDNA was cloned in to pGEX-6P GST vector and expressed in E. coli BL21 (DE3) and purified for in vitro phosphorylation assays. Plasmid transformed E. coli (BL21) was cultured at 37°C and induced overnight at 18°C with 0.5 mM IPTG. Cells were lysed in lysis buffer (50 mM NaPO4 pH8.0, 250 mM NaCl, 1 mM DTT, Complete Protease Inhibitor Cocktail (Roche)) and GST beads were applied for affinity purification. After incubation, beads were washed with wash buffer (50 mM NaPO4 pH6.0, 250 mM NaCl, 1 mM DTT) and eluted with elute buffer (50 mM NaPO4 pH7.2, 150 mM NaCl, 10 mM GSH, 1 mM DTT). Purified proteins were dialyzed against storge buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl). Protein concentration was determined by standard Bradford assay...”

(4) Line 141: The first sentence of this paragraph lacks motivation. I would start this sentence with "To directly observe the effects of phosphor mutants in the elbow region in microtubule binding and motility of OSM-3, we...".

We have made the change.

(5) Figure 1B: The mass spectrometry data in Figure 1B lacks adequate explanation. The Methods section should detail the experimental protocol, data interpretation, and any databases used. Additionally, the manuscript should list all identified phosphorylation sites on OSM-3 to provide context, including whether Y487_T490 is the major site.

We have provided the detailed experimental protocol, data interpretation, and databases used in methods. We have provided all identified sites as Appendix table S1.

(6) Figure 1C: Is it possible to model the effect of PM and PD mutations using AlphaFold? The authors should also show PAE or pLDDT scores of their model.

AlphaFold cannot well model the effect of mutants, but we conducted the Rosetta relax to capture their possible conformational changes, as shown in the revised Figure 3. We have provided PAE and pLDDT as a new figure, Figure S2.

(7) Figure 2D: The unit for speed should use a lowercase "s" for seconds.

We have fixed it.

(8) Figure 3: I am not sure whether this figure stands for a main text figure on its own, as it is only a Rosetta prediction and is not supported by any experimental data. In addition, it remains unclear what the labels on the x-axis mean.

We have updated the figure and explain the labels on the x-axis in Figure S4 to make it more reader-friendly.

(9) Figure 4: NEKL-3-treated OSM-1 should be included as a positive control in the in vitro experiments.

We suspect that the Reviewer asked for NEKL-3-treated OSM-3. 

In our other study which has just been accepted by the Journal of Cell Biology, NEKL3-treated OSM-3 significantly reduced the affinity between OSM-3 motor and microtubules and showed very low ATPase activity. We have cited and discussed this in the revised text on page #10, line #28: 

“…As demonstrated in our recent study (Huang et al., JCB, 2025, In press, attached), phosphorylation of OSM-3 by NEKL-3 at two distinct regions—Ser96 and the conserved "elbow" motif—differentially regulates its activity and localization. Phosphorylation at Ser96 reduces OSM-3’s ATPase activity and alters its ciliary distribution from the distal segment to a uniform localization, while elbow phosphorylation induces autoinhibition, retaining OSM-3 in the cell body. Strikingly, in vitro phosphorylation of OSM-3 by NEKL-3 significantly reduces its microtubulebinding affinity, likely arising from combined modifications at both sites. We propose a model wherein elbow phosphorylation ensures anterograde ciliary transport, while Ser96 phosphorylation fine-tunes distal segment targeting. This multistep regulation may involve distinct phosphatases to reverse phosphorylation at specific sites, a hypothesis warranting further investigation….”

(10) Figure 4C, D, and F: The unit of velocity is wrong. The authors should use the same units they used in the table shown in Figure 4B.

We have fixed these errors

(11) Figure 4F: The velocity of PD is a lot lower than G444E. Therefore, it would be more appropriate to refer to PD as partially active, rather than hyperactive.

We have made the change. 

(12) Figure 5: There is too much genetics jargon on this figure (EMF, F2, 100%Dyf,...). How are the alleles numbered? Is it OK to refer to them as Alleles 1 and 2 for simplicity?

According to the established C. elegans allele nomenclature, each worm allele has a unique number named after the lab code for identification. We have simplified the labels and updated the figure to make it more reader-friendly.

(13) Figure 5E: A plot would be more reader-friendly than a table. Additionally, the legend for Fig. 5E mistakenly refers to it as "D."

We have changed the table to a plot and fixed the mistakes. We thank the Reviewer for pointing them out.

Reviewer #2 (Recommendations for the authors):

(1) The model appears as if NEKL-3 induces dephosphorylation of OSM-3 (Figure 6). This is not consistent with the conclusions described in the Discussion and is confusing.

We have updated the model figure and fixed the error.

(2) It should be described why the authors hypothesized NEKL-3 phosphorylates OSM3. Was there genetic evidence? Did the authors screened cilia-related kinases? or Did the authors identify it incidentally? Providing this information would help readers to understand the context of the research.

We appreciate both Reviewers for pointing out this issue. 

Our hypothesis that NEKL-3 phosphorylates OSM-3 stems from prior findings in our lab. In a previous study (Yi et al., Traffic, 2018, PMID: 29655266), we identified NEKL-4, a member of the NIMA kinase family, as a suppressor of the OSM-3(G444E) hyperactive mutation. This discovery prompted us to explore the broader role of NIMA kinases in regulating OSM-3. Subsequent genetic screens (Xie et al., EMBO J, 2024, PMID: 38806659) revealed that both NEKL-3 and NEKL-4 suppress multiple OSM-3 mutations, further supporting their functional interaction. Given the established role of NIMA kinases in phosphorylation-dependent processes (Fry et al., JCS, 2012, PMID: 23132929; Chivukula et al., Nat. Med., 2020, PMID: 31959991; Thiel, C. et al. Am. J. Hum. Genet. 2011, PMID: 21211617; Smith, L. A. et al., J. Am. Soc. Nephrol., 2006, PMID: 16928806), we hypothesized that NEKL-3/4 may directly phosphorylate OSM3 to modulate its activity.

To test this hypothesis, we expressed recombinant C. elegans NEKL-3 and OSM-3 proteins and conducted in vitro phosphorylation assays. While we were unable to obtain active recombinant NEKL-4 (limitations noted in the revised text), our experiments with NEKL-3 revealed phosphorylation at residues 487-490 (YSTT motif) in OSM-3’s tail region, as confirmed by mass spectrometry. These findings are now explicitly contextualized in the Introduction and Results sections of the revised manuscript.

Page #4, Line #11:

“... In our previous study (Yi et al., Traffic, 2018, PMID: 29655266), a genetic screen targeting the OSM-3(G444E) hyperactive mutation identified NEKL-4, a member of the NIMA kinase family, as a suppressor of this phenotype. This finding, combined with reports that NIMA kinases regulate ciliary processes independently of their canonical mitotic roles (Fry et al., JCS, 2012, PMID: 23132929; Chivukula et al., Nat. Med., 2020, PMID: 31959991; Thiel, C. et al. Am. J. Hum. Genet. 2011, PMID: 21211617; Smith, L. A. et al., J. Am. Soc. Nephrol., 2006, PMID: 16928806), prompted us to investigate whether NIMA kinases modulate OSM-3-driven intraflagellar transport. We hypothesized that NEKL-3/4, as paralogs within this family, might directly phosphorylate OSM-3 to regulate its motility...”

Page #4, line #26: 

“... To determine whether NIMA kinase family members could directly phosphorylate OSM-3, we purified prokaryotic recombinant C. elegans NEKL-3/NEKL-4 and OSM3 protein in order to perform in vitro phosphorylation assays. We were able to obtain active recombinant NEKL-3 but not NEKL-4. The in vitro phosphorylation assays showed that NEKL-3, directly phosphorylates OSM-3 (Fig. 1A-B, Appendix Table S1). Subsequent mass spectrometric analysis revealed phosphorylation at residues 487-490, which localize to the conserved "YSTT" motif within OSM-3’s C-terminal tail region...”

(3) It is curious the authors have not addressed the cilia phenotype and the localization of OSM-3 in nekl-3 mutant. Regardless of whether these observations agrees with the proposed mechanisms, it is essential for the authors to show and discuss the cilia phenotype and OSM-3 localization in nekl-3 mutants.

We thank the Reviewer for highlighting this critical point. Indeed, nekl-3 null mutants are inviable due to essential mitotic roles (Barstead et al., 2012, PMID: 23173093), precluding direct analysis of ciliary phenotypes. To bypass this limitation, we recently generated nekl-3 conditional knockouts (cKOs) in ciliated neurons (Huang et al., JCB, 2025 in press, attached). In these mutants, OSM-3—which is normally enriched in the ciliary distal segment—becomes uniformly distributed along the cilium. This redistribution correlates with premature activation of OSM-3-driven anterograde motility in the ciliary middle region, consistent with our proposed model where NEKL3 phosphorylation suppresses OSM-3 activity. We have now integrated this result and discussion into the revised manuscript, reinforcing the physiological relevance of NEKL-3-mediated regulation in ciliary transport. 

Page #6 line #10

“… While nekl-3 null mutants are inviable due to essential mitotic roles (Barstead et al., 2012, PMID: 23173093), conditional knockout (cKO) of nekl-3 in ciliated neurons (Huang et al., JCB, 2025 in press, attached) revealed its critical role in regulating OSM3 dynamics. In nekl-3 cKO animals, OSM-3—normally enriched in the ciliary distal segment—redistributed uniformly along the cilium, concomitant with premature activation of anterograde motility in the middle ciliary region. This phenotype aligns with our model wherein NEKL-3 phosphorylation suppresses OSM-3 activity, ensuring spatiotemporal regulation of IFT.…”

(4) The methods section lacks some information, which is critical to reproducing this study.

We have now provided detailed information in the methods section in the revised manuscript.

(a) It is not described how the authors determined phosphorylation of OSM-3 by NEKL-3. In methods, nothing is described about the assay.

We performed in vitro phosphorylation assays using recombinant OSM-3 and NEKL3 purified from bacteria. We then used LC-MS/MS for identification of phosphorylation sites. We have now updated the methods section to include all the information.

Page #4 line #26

“... To determine whether NIMA kinase family members could directly phosphorylate OSM-3, we purified prokaryotic recombinant C. elegans NEKL-3/NEKL-4 and OSM3 protein in order to perform in vitro phosphorylation assays. We were able to obtain active recombinant NEKL-3 but not NEKL-4. The in vitro phosphorylation assays showed that NEKL-3, directly phosphorylates OSM-3 (Fig. 1A-B, Appendix Table S1). Subsequent mass spectrometric analysis revealed phosphorylation at residues 487-490, which localize to the conserved "YSTT" motif within OSM-3’s C-terminal tail region...”

Page #36, line #19

“In vitro phosphorylation assay 20 μM purified OSM-3 was incubated with 1 μM GST-NEKL-3 at 30 °C in 100 μL reaction buffer (50 mM Tris-HCl pH 8.0, 10 mM MgCl2, 150 mM NaCl, and 2 mM ATP) for 30 min. The reaction was terminated by boiling for 5 min with an SDS-sample buffer.

Mass spectrometry

Following NEKL-3 treatment, OSM-3 proteins were resolved by SDS-PAGE and visualized with Coomassie Brilliant Blue staining. Protein bands corresponding to OSM-3 were excised and subjected to digestion using the following protocol: reduction with 5 mM TCEP at 56°C for 30 min; alkylation with 10 mM iodoacetamide in darkness for 45 min at room temperature, and tryptic digestion at 37°C overnight with a 1:20 enzyme-to-protein ratio. The resulting peptides were subjected to mass spectrometry analysis. Briefly, the peptides were analyzed using an UltiMate 3000 RSLCnano system coupled to an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific). We applied an in-house proteome discovery searching algorithm to search the MS/MS data against the C. elegans database. Phosphorylation sites were determined using PhosphoRS algorithm with manual validation of MS/MS spectra.”

(b) The method of structural prediction by Alfafold2 and LocalColabFold needs clarification. In general, the prediction gives several candidates. How did the authors choose one of these candidates?

We generated five candidate models and all of them showed similar conformation. We thus chose the model with the highest confidence. We have provided PAE and pLDDT as additional data in Figure S2 and discussed them in the revised text on, Page #4, line #32: 

“...To gain structural insights from this motif, we employed LocalColabFold based on AlphaFold2 to predict the dimeric structure of OSM-3 (Evans et al., 2022; Jumper et al., 2021; Mirdita et al., 2022). The highest-confidence model was selected for further analysis (Fig. 1C, Fig. S2)...”

(c) The methods to predict conformational changes by introducing various point mutations are interesting (Figure 3). However, the methods require more detailed descriptions. In the current form, the manuscript only lists the tools used. The pipelines and parameters need to be described. This information is important because AlphaFoldbased predictions often give folded conformations because the training data are mainly composed of folded proteins. It is surprising that the methods applied here give open conformations induced by point mutations.

We have described the pipelines in the revised Methods section on page#34, line#25: 

“…OSM-3 model was predicted using LocalColabFold (Evans et al., 2022; Jumper et al., 2021; Mirdita et al., 2022). Mutated proteins were designed by Pymol 2.6, choosing the rotamer of the mutated residues in G444E, PM and PD models with the least clash as the initial conformation. To predict mutation-induced conformational changes, the initial models were subjected to Pyrosetta (Chaudhury et al., 2010). The energies of pre-relaxed models were evaluated with Rosetta Energy Function 2015 (Alford et al., 2017), and then the relax procedure were applied to the models with default parameters to obtain the relaxed models visualized by Pymol to minimize the energy of these models. In detail, to obtain the relaxed models visualized by Pymol and minimize the energy of these models, the classic relax mover was used in the procedure mentioned above with default settings. The relax script has been uploaded to Github: https://github.com/young55775/RosettaRelax_for_OSM3...”

(5) The authors have purified proteins. Do they show different properties in gel filtration that are consistent with the structural prediction? It is anticipated that open-form mutants are eluted from earlier than closed forms.

We thank the reviewer for this insightful suggestion. Indeed, our recent study supported that the open-from of the active OSM-3 G444E mutation were eluted earlier than the wild-type closed form (Xie et al., EMBO J., 2024). While the current study did not perform gel filtration chromatography (SEC) to directly compare the hydrodynamic properties of the OSM-3 mutants, our functional assays provide robust evidence for conformational changes predicted by structural modeling. For example: ATPase activity assays revealed that the open-state mutants (e.g., G444E and PD muatnts) exhibited significantly enhanced enzymatic activity (Figure 4A), consistent with structural predictions of an active, destabilized autoinhibitory interface (Figure 3A). These functional readouts collectively validate the predicted structural states. While SEC could further corroborate these findings by distinguishing compact (closed) versus extended (open) conformations, we prioritized assays that directly link structural predictions to in vitro enzymatic activity and in vivo ciliary transport dynamics. Future studies incorporating SEC or cryo-EM will provide additional biophysical validation of these states.

We have revised the text in the manuscript (Page #7, Lines #22): 

“…Notably, the open-state OSM-3 mutants (e.g., G444E) displayed elevated ATPase activity, consistent with structural predictions of autoinhibition release (Fig. 3A, Fig. 4A) (Xie et al., 2024). While hydrodynamic profiling (e.g., SEC) could further resolve conformational states, our functional assays directly connect predicted structural changes to altered biochemical and cellular activity...”

Minor point

(1) Line 85 "MIMA kinase family" should be "NIMA kinase family".

We have corrected the typo and appreciate that the Reviewer for pointing it out. 

(2) M.S. and D.S. need to be defined in Figure 2D.

We have updated the figures.

Author Response:

Reviewer #1 (Public Review):

Summary:<br /> The global decline of amphibians is primarily attributed to deadly disease outbreaks caused by the chytrid fungus, Batrachochytrium dendrobatidis (Bd). It is unclear whether and how skin-resident immune cells defend against Bd. Although it is well known that mammalian mast cells are crucial immune sentinels in the skin and play a pivotal role in the immune recognition of pathogens and orchestrating subsequent immune responses, the roles of amphibian mast cells during Bd infections are largely unknown. The current study developed a novel way to enrich X. laevis skin mast cells by injecting the skin with recombinant stem cell factor (SCF), a KIT ligand required for mast cell differentiation and survival. The investigators found an enrichment of skin mast cells provides X. laevis substantial protection against Bd and mitigates the inflammation-related skin damage resulting from Bd infection. Additionally, the augmentation of mast cells leads to increased mucin content within cutaneous mucus glands and shields frogs from the alterations to their skin microbiomes caused by Bd.

Strengths:<br /> This study underscores the significance of amphibian skin-resident immune cells in defenses against Bd and introduces a novel approach to examining interactions between amphibian hosts and fungal pathogens.

Weaknesses:<br /> The main weakness of the study is the lack of functional analysis of X. laevis mast cells. Upon activation, mast cells have the characteristic feature of degranulation to release histamine, serotonin, proteases, cytokines, and chemokines, etc. The study should determine whether X. laevis mast cells can be degranulated by two commonly used mast cell activators IgE and compound 48/80 for IgE-dependent and independent pathways. This can be easily done in vitro. It is also important to assess whether in vivo these mast cells are degranulated upon Bd infection using avidin staining to visualize vesicle releases from mast cells. Figure 3 only showed rSCF injection caused an increase in mast cells in naïve skin. They need to present whether Bd infection can induce mast cell increase and rSCF injection under Bd infection causes a mast cell increase in the skin. In addition, it is unclear how the enrichment of mast cells provides protection against Bd infection and alternations to skin microbiomes after infection. It is important to determine whether skin mast cells release any contents mentioned above.

We would like to thank the reviewer for taking the time to review our work and for providing us with valuable feedback.

Please note that amphibians do not possess the IgE antibody isotype1.

To our knowledge there have been no published studies using approaches for studying mammalian mast cell degranulation to examine amphibian mast cells. Notably, several studies suggest that amphibian mast cells lack histamine2, 3, 4, 5 and serotonin2, 6. While there are commercially available kits and reagents for examining mammalian mast cell granule content, most of these reagents may not cross-react with their amphibian counterparts. This is especially true of cytokines and chemokines, which diverged quickly with evolution and thus do not share substantial protein sequence identity across species as divergent as frogs and mammals. Respectfully, while following up on these findings is possible, it would involve considerable additional work to find reagents that would detect amphibian mast cell contents.

We would also like to respectfully point out that while mast cell degranulation is a feature most associated with mammalian mast cells, this is not the only means by which mammalian mast cells confer their immunological effects. While we agree that defining the biology of amphibian mast cell degranulation is important, we anticipate that since the anti-Bd protection conferred by enriching frog mast cells is seen after 21 days of enrichment, it is quite possible that degranulation may not be the central mechanism by which the mast cells are mediating this protection.

As noted in our manuscript, frog mast cells upregulate their expression of interleukin-4 (IL4), which is a hallmark cytokine associated with mammalian mast cells7. We are presently exploring the role of the frog IL4 in the observed mast cell anti-Bd protection. Should we generate meaningful findings in this regard, we will add them to the revised version of this manuscript.

We are also exploring the heparin content of frog mast cells and capacities of these cells to degranulate in vitro in response to compound 48/80. In addition, we are exploring in vivo mast cell degranulation via histology and avidin-staining. Should these studies generate significant findings, we will include them in the revised version of this manuscript.

Per the reviewer’s suggestion, in our revised manuscript we also plan to include data showing whether Bd infections affect skin mast cell numbers and how rSCF injection impacts skin mast cell numbers in the context of Bd infections.

In regard to how mast cells impact Bd infections and skin microbiomes, our data indicate that mast cells are augmenting skin integrity during Bd infections and promoting mucus production, as indicated by the findings presented in Figure 4A-C and Figure 5A-C, respectively. There are several mammalian mast cell products that elicit mucus production. In mammals, this mucus production is mediated by goblet cells while the molecular control of amphibian skin mucus gland content remains incompletely understood. Interleukin-13 (IL13) is the major cytokine associated with mammalian mucus production8, while to our knowledge this cytokine is either not encoded by amphibians or else has yet to be identified and annotated in these animals’ genomes. IL4 signaling also results in mucus production9 and we are presently exploring the possible contribution of the X. laevis IL4 to skin mucus gland filling. Any significant findings on this front will be included in the revised manuscript. Histamine release contributes to mast cell-mediated mucus production10, but as we outline above, several studies indicate that amphibian mast cells may lack histamine2, 3, 4, 5. Mammalian mast cell-produced lipid mediators also play a critical role in eliciting mucus secretion11 and our transcriptomic analysis indicates that frog mast cells express several enzymes associated with production of such mediators. We will highlight this observation in our revised manuscript.

We anticipate that X. laevis mast cells influence skin integrity, microbial composition and Bd susceptibility in a myriad of ways. Considering the substantial differences between amphibian and mammalian evolutionary histories and physiologies, we anticipate that many of the mechanisms by which X. laevis mast cells confer anti-Bd protection will prove to be specific to amphibians and some even unique to X. laevis. We are most interested in deciphering what these mechanisms are but foresee that they will not necessarily reflect what one would expect based on what we know about mammalian mast cells in the context of mammalian physiologies.

Reviewer #2 (Public Review):

Summary:<br /> In this study, Hauser et al investigate the role of amphibian (Xenopus laevis) mast cells in cutaneous immune responses to the ecologically important pathogen Batrachochytrium dendrobatidis (Bd) using novel methods of in vitro differentiation of bone marrow-derived mast cells and in vivo expansion of skin mast cell populations. They find that bone marrow-derived myeloid precursors cultured in the presence of recombinant X. laevis Stem Cell Factor (rSCF) differentiate into cells that display hallmark characteristics of mast cells. They inject their novel (r)SCF reagent into the skin of X. laevis and find that this stimulates the expansion of cutaneous mast cell populations in vivo. They then apply this model of cutaneous mast cell expansion in the setting of Bd infection and find that mast cell expansion attenuates the skin burden of Bd zoospores and pathologic features including epithelial thickness and improves protective mucus production and transcriptional markers of barrier function. Utilizing their prior expertise with expanding neutrophil populations in X. laevis, the authors compare mast cell expansion using (r)SCF to neutrophil expansion using recombinant colony-stimulating factor 3 (rCSF3) and find that neutrophil expansion in Bd infection leads to greater burden of zoospores and worse skin pathology.

Strengths: <br /> The authors report a novel method of expanding amphibian mast cells utilizing their custom-made rSCF reagent. They rigorously characterize expanded mast cells in vitro and in vivo using histologic, morphologic, transcriptional, and functional assays. This establishes solid footing with which to then study the role of rSCF-stimulated mast cell expansion in the Bd infection model. This appears to be the first demonstration of the exogenous use of rSCF in amphibians to expand mast cell populations and may set a foundation for future mechanistic studies of mast cells in the X. laevis model organism. 

We thank the reviewer for recognizing the breadth and extent of the undertaking that culminated in this manuscript. Indeed, this manuscript would not have been possible without considerable reagent development and adaptation of techniques that had previously not been used for amphibian immunity research. In line with the reviewer’s sentiment, to our knowledge this is the first report of using molecular approaches to augment amphibian mast cells, which we hope will pave the way for new areas of research within the fields of comparative immunology and amphibian disease biology.

Weaknesses:<br /> The conclusions regarding the role of mast cell expansion in controlling Bd infection would be stronger with a more rigorous evaluation of the model, as there are some key gaps and remaining questions regarding the data. For example:

1. Granulocyte expansion is carefully quantified in the initial time courses of rSCF and rCSF3 injections, but similar quantification is not provided in the disease models (Figures 3E, 4G, 5D-G). A key implication of the opposing effects of mast cell vs neutrophil expansion is that mast cells may suppress neutrophil recruitment or function. Alternatively, mast cells also express notable levels of csfr3 (Figure 2) and previous work from this group (Hauser et al, Facets 2020) showed rG-CSF-stimulated peritoneal granulocytes express mast cell markers including kit and tpsab1, raising the question of what effect rCSF3 might have on mast cell populations in the skin. Considering these points, it would be helpful if both mast cells and neutrophils were quantified histologically (based on Figure 1, they can be readily distinguished by SE or Giemsa stain) in the Bd infection models.

We thank the reviewer for this insightful suggestion. We are performing a further examination of skin granulocyte content during Bd infections and plan on including any significant findings in our revised manuscript.

We predict that rSCF administration results in the accumulation of mast cells that are polarized such that they ablate the inflammatory response elicited by Bd infection. Mammalian mast cells, including peritonea-resident mast cells, express csf3r12, 13. Although the X. laevis animal model does not permit nearly the degree of immune cell resolution afforded by mammalian animal models, we do know that the adult X. laevis peritonea contain heterogenous leukocyte populations. We anticipate that the high kit expression reported by Hauser et al., 2020 in the rCSF3-recruited peritoneal leukocytes reflects the presence of mast cells therein. As such and in acknowledgement of the reviewer’s suggestion, we also think that the cells recruited by rCSF3 into the skin may include not only neutrophils but also mast cells. Possibly, these mast cells have distinct polarization states from those enriched by rSCF. While the lack of antibodies against frog neutrophils or mast cells has limited our capacity to address this question, we will attempt to reexamine by histology the proportions of skin neutrophils and mast cells in the skins of frogs under the conditions described in our manuscript. Any new findings in this regard will be included in the revised version of this work.

2. Epithelial thickness and inflammation in Bd infection are reported to be reduced by rSCF treatment (Figure 3E, 5A-B) or increased by rCSF3 treatment (Figure 4G) but quantification of these critical readouts is not shown.

We thank the reviewer for this suggestion. We will score epithelial thickness under the distinct conditions described in our manuscript and present the quantified data in the revised paper.

3. Critical time points in the Bd model are incompletely characterized. Mast cell expansion decreases zoospore burden at 21 dpi, while there is no difference at 7 dpi (Figure 3E). Conversely, neutrophil expansion increases zoospore burden at 7 dpi, but no corresponding 21 dpi data is shown for comparison (Figure 4G). Microbiota analysis is performed at a third time point,10 dpi (Figure 5D-G), making it difficult to compare with the data from the 7 dpi and 21 dpi time points. Reporting consistent readouts at these three time points is important to draw solid conclusions about the relationship of mast cell expansion to Bd infection and shifts in microbiota.

Because there were no significant effects of mast cell enrichment at 7 days post Bd infection, we chose to look at the microbiome composition in a subsequent experiment at 10 days and 21 days post Bd infection, with 10 days being a bit more of a midway point between the initial exposure and day 21, when we see the effect on Bd loads. We will clarify this rationale in the revised manuscript.

The enrichment of neutrophils in frog skins resulted in prompt (12 hours post enrichment) skin thickening (in absence of Bd infection) and increased frog Bd susceptibility by 7 days of infection. Conversely, mast cell enrichment stabilized skin mucosal and symbiotic microbial environment, presumably accounting at least in part for the lack of further Bd growth on mast cell-enriched animals by 21 days of infection. Our question regarding the roles of inflammatory granulocytes/neutrophils during Bd infections was that of ‘how’ rather ‘when’ these cells affect Bd infections. Because the central focus of this work was mast cells and not other granulocyte subsets, when we saw that rCSF3-recruited granulocytes adversely affected Bd infections at 7 days post infection, we did not pursue the kinetics of these responses further. We plan to explore the roles of inflammatory mediators and disparate frog immune cell subsets during the course of Bd infections, but we feel that these future studies are more peripheral to the central thesis of the present manuscript regarding the roles of frog mast cells during Bd infections.

4. Although the effect of rSCF treatment on Bd zoospores is significant at 21 dpi (Figure 3E), bacterial microbiota changes at 21 dpi are not (Figure S3B-C). This discrepancy, how it relates to the bacterial microbiota changes at 10 dpi, and why 7, 10, and 21 dpi time points were chosen for these different readouts (Figure 5F-G), is not discussed.

Our results indicate that after 10 days of Bd infection, control Bd-challenged animals exhibited reduced microbial richness, while skin mast cell-enriched Bd-infected frogs were protected from this disruption of their microbiome. The amphibian microbiome serves as a major barrier to these fungal infections14, and we anticipate that Bd-mediated disruption of microbial richness and composition facilitates host skin colonization by this pathogen. Control and mast cell-enriched animals had similar skin Bd loads at 10 days post infection. However, by 21 days of Bd infection the mast cells-enriched animals maintained their Bd loads to levels observed at 10 days post infection, whereas the control animals had significantly greater Bd loads. Thus, we anticipate that frog mast cells are conferring the observed anti-Bd protection in part by preventing microbial disassembly and thus interfering with optimal Bd colonization and growth on frog skins. In other words, maintained microbial composition at 10 days of infection may be preventing additional Bd colonization/growth, as seen when comparing skins of control and mast cell-enriched frogs at 21 days post infection. By 21 days of infection, control animals rebounded from the Bd-mediated reduction in bacterial richness seen at 10 days. Considering that after 21 days of infection control animals also had significantly greater Bd loads than mast-cell enriched animals suggests that there may be a critical earlier window during which microbial composition is able to counteract _Bd_growth. 

While the current draft of our manuscript has a paragraph to this effect (see below), we appreciate the reviewer conveying to us that our perspective on the relationship between skin mast cells and the kinetics of microbial composition and _Bd_loads could be better emphasized. We plan to revise our manuscript to include the above discussion points. 

Bd infections caused major reductions in bacterial taxa richness, changes in composition and substantial increases in the relative abundance of Bd-inhibitory bacteria early in the infection. Similar changes to microbiome structure occur during experimental Bd infections of red-backed salamanders and mountain yellow-legged frogs15, 16. In turn, progressing Bd_infections corresponded with a return to baseline levels of _Bd-inhibitory bacteria abundance and rebounding microbial richness, albeit with dissimilar communities to those seen in control animals. These temporal changes indicate that amphibian microbiomes are dynamic, as are the effects of Bd infections on them. Indeed, Bd infections may have long-lasting impacts on amphibian microbiomes15. While Bd infections manifested in these considerable changes to frog skin microbiome structure, mast cell enrichment appeared to counteract these deleterious effects to their microbial composition. Presumably, the greater skin mucosal integrity and mucus production observed after mast cell enrichment served to stabilize the cutaneous environment during Bd infections, thereby ameliorating the Bd-mediated microbiome changes. While this work explored the changes in established antifungal flora, we anticipate the mast cell-mediated inhibition of Bd may be due to additional, yet unidentified bacterial or fungal taxa. Intriguingly, while mammalian skin mast cell functionality depends on microbiome elicited SCF production by keratinocytes17, our results indicate that frog skin mast cells in turn impact skin microbiome structure and likely their function. It will be interesting to further explore the interdependent nature of amphibian skin microbiomes and resident mast cells.

5. The time course of rSCF or rCSF3 treatments relative to Bd infection in the experiments is not clear. Were the treatments given 12 hours prior to the final analysis point to maximize the effect? For example, in Figure 3E, were rSCF injections given at 6.5 dpi and 20.5 dpi? Or were treatments administered on day 0 of the infection model? If the latter, how do the authors explain the effects at 7 dpi or 21 dpi given mast cell and neutrophil numbers return to baseline within 24 hours after rSCF or rCSF3 treatment, respectively?

Please find the schematic of the immune manipulation, Bd infection, and sample collection times below. We will include a figure like this in our revised manuscript.

The title of the manuscript may be mildly overstated. Although Bd infection can indeed be deadly, mortality was not a readout in this study, and it is not clear from the data reported that expanding skin mast cells would ultimately prevent progression to death in Bd infections.

We acknowledge this point. The revised manuscript will be titled: “Amphibian mast cells: barriers to chytrid fungus infections”.

Reviewer #3 (Public Review):

Summary:<br /> Hauser et al. provide an exceptional study describing the role of resident mast cells in amphibian epidermis that produce anti-inflammatory cytokines that prevent Batrachochytrium dendrobatidis (Bd) infection from causing harmful inflammation, and also protect frogs from changes in skin microbiomes and loss of mucin in glands and loss of mucus integrity that otherwise cause changes to their skin microbiomes. Neutrophils, in contrast, were not protective against Bd infection. Beyond the beautiful cytology and transcriptional profiling, the authors utilized elegant cell enrichment experiments to enrich mast cells by recombinant stem cell factor, or to enrich neutrophils by recombinant colony-stimulating factor-3, and examined respective infection outcomes in Xenopus.

Strengths:<br /> Through the use of recombinant IL4, the authors were able to test and eliminate the hypothesis that mast cell production of IL4 was the mechanism of host protection from Bd infection. Instead, impacts on the mucus glands and interaction with the skin microbiome are implicated as the protective mechanism. These results will press disease ecologists to examine the relative importance of this immune defense among species, the influence of mast cells on the skin microbiome and mucosal function, and open the potential for modulating mucosal defense.

We thank the reviewer for recognizing the significance and utility of the findings presented in our manuscript.

Weaknesses:<br /> A reduction of bacterial diversity upon infection, as described at the end of the results section, may not always be an "adverse effect," particularly given that anti-Bd function of the microbiome increased. Some authors (see Letourneau et al. 2022 ISME, or Woodhams et al. 2023 DCI) consider these short-term alterations as encoding ecological memory, such that continued exposure to a pathogen would encounter an enriched microbial defense. Regardless, mast cell-initiated protection of the mucus layer may negate the need for this microbial memory defense.

We thank the reviewer their insightful comment. We will revise our discussion to include this possible interpretation.

While the description of the mast cell location in the epidermal skin layer in amphibians is novel, it is not known how representative these results are across species ranging in chytridiomycosis susceptibility. No management applications are provided such as methods to increase this defense without the use of recombinant stem cell factor, and more discussion is needed on how the mast cell component (abundance, distribution in the skin) of the epidermis develops or is regulated.

We appreciate the reviewer’s comment and would like to point out that the work presented in our manuscript was driven by comparative immunology questions more than by conservation biology.

We thank the reviewer for suggesting expanding our discussion to include potential management applications and potential mechanisms for regulating frog skin mast cells. While any content to these effects would be highly speculative, we agree that it may spark new interest and pave new avenues for research. To this end, our revised manuscript will include a paragraph to this effect.

References:

1. Flajnik, M.F. A cold-blooded view of adaptive immunity. Nat Rev Immunol 18, 438-453 (2018).

2. Mulero, I., Sepulcre, M.P., Meseguer, J., Garcia-Ayala, A. & Mulero, V. Histamine is stored in mast cells of most evolutionarily advanced fish and regulates the fish inflammatory response. Proc Natl Acad Sci U S A 104, 19434-19439 (2007).

3. Reite, O.B. A phylogenetical approach to the functional significance of tissue mast cell histamine. Nature 206, 1334-1336 (1965).

4. Reite, O.B. Comparative physiology of histamine. Physiol Rev 52, 778-819 (1972).

5. Takaya, K., Fujita, T. & Endo, K. Mast cells free of histamine in Rana catasbiana. Nature 215, 776-777 (1967).

6. Galli, S.J. New insights into "the riddle of the mast cells": microenvironmental regulation of mast cell development and phenotypic heterogeneity. Lab Invest 62, 5-33 (1990).

7. Babina, M., Guhl, S., Artuc, M. & Zuberbier, T. IL-4 and human skin mast cells revisited: reinforcement of a pro-allergic phenotype upon prolonged exposure. Archives of dermatological research 308, 665-670 (2016).

8. Lai, H. & Rogers, D.F. New pharmacotherapy for airway mucus hypersecretion in asthma and COPD: targeting intracellular signaling pathways. J Aerosol Med Pulm Drug Deliv 23, 219-231 (2010).

9. Rankin, J.A. et al. Phenotypic and physiologic characterization of transgenic mice expressing interleukin 4 in the lung: lymphocytic and eosinophilic inflammation without airway hyperreactivity. Proc Natl Acad Sci U S A 93, 7821-7825 (1996).

10. Church, M.K. Allergy, Histamine and Antihistamines. Handb Exp Pharmacol 241, 321-331 (2017).

11. Nakamura, T. The roles of lipid mediators in type I hypersensitivity. J Pharmacol Sci 147, 126-131 (2021).

12. Aponte-Lopez, A., Enciso, J., Munoz-Cruz, S. & Fuentes-Panana, E.M. An In Vitro Model of Mast Cell Recruitment and Activation by Breast Cancer Cells Supports Anti-Tumoral Responses. Int J Mol Sci 21 (2020).

13. Jamur, M.C. et al. Mast cell repopulation of the peritoneal cavity: contribution of mast cell progenitors versus bone marrow derived committed mast cell precursors. BMC Immunol 11, 32 (2010).

14. Walke, J.B. & Belden, L.K. Harnessing the Microbiome to Prevent Fungal Infections: Lessons from Amphibians. PLoS Pathog 12, e1005796 (2016).

15. Jani, A.J. et al. The amphibian microbiome exhibits poor resilience following pathogen-induced disturbance. ISME J 15, 1628-1640 (2021).

16. Muletz-Wolz, C.R., Fleischer, R.C. & Lips, K.R. Fungal disease and temperature alter skin microbiome structure in an experimental salamander system. Mol Ecol 28, 2917-2931 (2019).

17. Wang, Z. et al. Skin microbiome promotes mast cell maturation by triggering stem cell factor production in keratinocytes. J Allergy Clin Immunol 139, 1205-1216 e1206 (2017).

This is interesting to see that a majority of people are lurkers because as someone you uses social media it seems like there are many people who contribute but this actually shocks me because I guess you have to look at it in the scale of numbers if there are 100 million users of an app even 10% of that is 10 million which is many users and thats why people may think that there are a lot of engagers when in fact a good chunk of us are just lurkers.

Author response:

The following is the authors’ response to the original reviews

Thank you for your valuable comments, which helped us improve our manuscript. We will make the following modifications in the revised manuscript:

(1) In the first paragraph of the Result section, we will provide a summary of trimeric G proteins in Ciona and explain how we focused on Gαs and Gαq in the initial phase of this study.

We added a summary of trimeric G proteins in Ciona in the initial part of the Results section (page 6, line 23 to page 8, line 5). In this summary, we added the following sentence explaining the reason we focused on Gas and Gaq in the initial phase of this study: "Among them, we prioritized examining the Gα proteins having an excitatory function (Gαq and Gαs) rather than inhibitory roles since previous studies suggested that excitatory events like Ca²⁺ transient and neuropeptide secretion occur when Ciona metamorphose."

(2) As the reviewer 1 suggests, the polymodal roles of papilla neurons are interesting. Although we could not address this through functional analyses in this study, we will add a discussion regarding this aspect. The sentences will be something like the following:

“The recent study (Hoyer et al., 2024) provided several lines of evidence suggesting that PSNs can serve as the sensors of several chemicals in addition to the mechanical stimuli. This finding and our model could be mutually related because these chemicals could modify Ca²⁺ and cAMP production. The use of G protein signaling allows Ciona to reflect various environmental stimuli to initiate metamorphosis in the appropriate situation, both mechanically and chemically.”

We added a discussion related to the recent publication by Hoyer and colleagues on page 23, lines 13-18: " A recent study[19] provided several lines of evidence suggesting that PNs can serve as the sensors of several chemicals in addition to mechanical stimuli. This finding and our model could be mutually related because these chemicals could modify Ca²⁺ and cAMP production. G protein signaling allows Ciona to reflect various environmental stimuli to initiate metamorphosis either mechanically or chemically according to the situation."

(3) As both reviewers suggested, imaging cAMP on the backgrounds of some G protein knockdowns is essential, and we will conduct the experiments.

We added the data on cAMP imaging in Gas, Gaq, and dvGai_Chr2 knockdown larvae in Supplementary Figure S4C-D and Figure 6E.

(4) We carefully modify the text throughout the manuscript so that the descriptions suitably reflect the results.

We modified the descriptions of experimental results so that the text reflects the results more precisely.

Reviewer #1:

Pg1 - need to add an additional '6' to the author list to clarify which two or more authors contributed equally.

We added a 6 as suggested. Thank you for pointing this out.

Pg3 - note that larval adhesive organ applies to not all benthic adults, but to benthic sessile adults this makes it sound like the adhesive organ can trigger metamorphosis but has that been shown? In Ciona or others? Need to specify the role of cells secreting adhesive, vs sensory cells that trigger metamorphosis?

We divided the corresponding sentence into two to clearly state that adhesion and triggering metamorphosis are related but could be different events. Moreover, we modified the sentence to state that physical contact is one example of a cue triggering metamorphosis. We then added another example of a factor triggering metamorphosis—i.e., chemicals from the organisms surrounding the adherence site (page 3, lines 16-20 of the revised version):

"Many marine invertebrates exhibit a benthic lifestyle at the adult stage[4]. Their planktonic larvae have an adhesive organ that secretes adhesives and adheres to a substratum. The cues associated with the adhesion, such as the physical contact with the substratum and a chemical from organisms surrounding the adherence site, can trigger their metamorphosis."

Pg 4 - although mechanosensation is the focus here, could there also be chemoreception/chemoreceptors involved in Ciona metamorphosis? For example, Hoyer et al. 2024 (Current Biology 34(6):1168-1182) concluded that some palp sensory neurons were multimodal and could be both chemo- and mechano-sensory.

We added statements about this recent finding in the Introduction and Discussion sections. In the Introduction (page 4, lines 16-18), however, we also stated that a mechanical stimulus can trigger metamorphosis in the lab without the need to supply these chemicals. This is to emphasize that the mechanical stimulus is the focus of this study. In the Discussion, we added a statement that G-protein signaling could also be used to receive the chemical stimuli (page 23, lines 13-18).

Pg 6 - Before starting functional characterizations, it would be useful to give an overview (table?) of the G proteins found in papillae, and what receptor they are suspected of binding to, or if this is completely unknown, and which downstream pathways they likely activate. That is, to show some results about which G proteins are found in Ciona, and which are found in papillae. In this way, it will make more sense for readers when the Gai is suddenly introduced later, following the sections of Gaq and Gas.

Thank you for your idea to improve the readability of this manuscript. In the initial part of the Results section (page 6, line 22 to page 8, line 5), we added descriptions of the repertoire of trimeric G-proteins in Ciona, including phylogenetic analyses, and expression in the papillae based on RNA-seq data, followed by the reason why we initially focused on Gaq and Gas. The data are displayed in Supplementary Figure S1. The phylogenetic analyses were modified from those shown in Supplementary Figure S5 of the previous version. We also added the general downstream activities of Gas, Gai and Gaq in the Introduction section (page 6, lines 10-12). Considering the contents, the general function of Ga12/13 was stated in the Results section (page 8, lines 2-3).

We did not add the information about their partner receptors in this early section. This is because there are many candidates, and we could not pick some of them. Instead, we described our current suppositions about their possible partners in the Discussion (page 23, line 22 to page 24, line 19). However, we suspect that there are more candidates, and we wish to promote unbiased research in the future.

Pg 9 - would be good to know the timing of this PF fluorescence increase and the timing of stimulation in the text here, relevant to the 30-min gap before metamorphosis initiation

We added the start times for the cAMP reduction and re-upregulation in the following sentence (page 11, lines 17-18): "The cAMP reduction and increase respectively started at 35 seconds and 4 min 40 seconds after stimulation on average."

Pg 28 - Phylogenetic analysis: Given that the results may be of interest to metamorphosis in other marine invertebrates as discussed in the last paragraph of the paper, it would be useful to include G proteins from these other animal phyla where available in the phylogenetic tree. Similarly, in Figure S5A it would be useful to highlight further all the different Ciona G proteins, and the different protein families, through the use of additional colour/labelling (regardless of whether this remains Fig S5A, or becomes part of the main figures)

We drew a phylogenetic tree of G-proteins including those in some sessile and benthic animals (barnacle, sea anemone, hydra, sponge, sea urchin and shell). However, we decided not to add the tree in the revised version because, unfortunately, the bootstrap values of many branches were not high enough to have confidence in the results. We hope you understand our decision. Ciona divergent G-proteins are likely to be specific to Ciona.

According to your comment, we highlighted all Ciona G alpha proteins in red in Figure S5A, which is now Figure S1A in the revised version.

Figure 3E and Figure S3 - is the data shown as an average of all larvae measured (n=5 and n=4) or is it data from one representative larva out of the 4-5 measured? This needs clarification.

The original graphs in Figure 3E and Figure S3 are typical examples. We added the graphs summarizing data of all larvae in each experimental condition in Supplementary Figure S4 (corresponding to Supplementary Figure S3 of the original version). Figure 3E remains as a typical example of the result of a single larva to explain our data analysis in detail.

Experimental suggestion - As mentioned above, one missing detail seems to be the need for evidence that cAMP is elevated in the papillae directly as a result of Gs activation- this could be shown with measurement of cAMP via PF in Gs knockdown larvae that are mechanically stimulated compared to wildtype stimulated and non-stimulated?

Thank you for your suggestion. The experiments are indeed important. We added the data of Pink Flamindo imaging in the Gas, Gaq and dvGai_Chr2 knockdown conditions. The results of Gas and Gaq knockdowns are described in page 11, line 24 to page 12, line 5, and are displayed in Supplementary Figure S4C-D. The result of dvGai_Chr2 knockdown is given on page 16, lines 20-22 and shown in Figure 6E.

In order to insert the data of cAMP imaging of dvGai_Chr2 knockdown larvae, we transferred some panels of Figure 6 to Supplementary Figure S6. In addition, the knockdown data of dvGαi_Chr4 and double knockdowns of Gai genes are also included in Supplementary Figure S6.

Reviewer #2:

Page 6, line 3-4 in the first paragraph of the "Results"; the authors state "Neither morphant showed any signature of metamorphosis even though both were allowed to adhere to the base of culture dishes...". However, judging from Fig. 1E, "the percentage of metamorphosis initiation" (indicated by the initiation of tail regression) in Gαq morphans is not close to 0 (average about 40%), thus I am not convinced this observation can be described as "Neither morphant showed any signature of metamorphosis..." in this sentence.

Thank you for your suggestion. In writing the original text, we oversimplified some of the descriptions when trying to improve the readability. We agree this resulted in imprecision in places. We have revised all these passages in our revision. In this particular case, we softened the overly emphatic statement to better reflect the results, changing “... any signature of metamorphosis...” to “... reduced rate of metamorphosis initiation...” In addition, we stated that the effect of G_α_q MO was weaker than that of G_α_s MO on page 8, lines 10-12. The weaker effect of Gaq MO was due to the redundant role of the Gi pathway, which is shown on page 17, lines 10-17, and in Figure 6G-H.

Similarly, in the next paragraph describing the knockdown of PLCβ1/2/3, PLCβ4, and IP3R genes, the authors appear to neglect there is a weaker effect of the PLCβ4 MO, and simply described the results as "The knockdown larvae of these three genes failed to start metamorphosis". Based on Fig. 1H, about 30% of the PLCβ4 MO-injected animals still initiated tail regeneration. This difference may have some biological meanings and thus should be described more precisely.

We added the following sentence on page 8, lines 18-19 of the revised version: “The effect of PLCβ4 MO was weaker than those of the other MOs, suggesting that this PLC plays an auxiliary role.”

Page 7, second paragraph, on the description of GCaMP8 fluorescence and also at the end of Fig. 1O legend, the citation to "Figure S1" is confusing; Fig. S1 is the phylogenetic tree of PLCβ proteins. Is there additional data regarding this Gαq MO plus GCaMP8 mRNA injection experiment?

Figure S1 of the original version corresponds to Figure S2 of the revised version. To avoid confusion, we deleted this citation from the legend of Figure 1O. By this modification, the sentence stating the repertoire of PLCb and IP3R in Ciona (page 8, lines 15-16) is the only sentence citing Figure S2 in the revised version.

Page 8, first sentence; The purpose of theophylline treatment is not to prevent larvae from adhesion, thus I would suggest modifying this sentence to: "We treated wild-type larvae with theophylline after tail amputation, and we observed that most theophylline-treated larvae completed tail regression without adhesion (Figure 2D-F)".

We modified the sentence according to your comment. Thank you for your suggestion.

Page 9, second paragraph; judging from the data presented in Fig. 3C, I think this description: "when papillae were removed from larvae, theophylline failed to induce metamorphosis" is not accurate, because about ~30% of the Papilla cut +Theophylline-treated larvae still initiated their tail regression. This needs to be explained clearly.

We modified the sentence (page 11, lines 2-3) as follows: “...the average rate of metamorphosis induction by theophylline was reduced from 100% to 30%...”

Similarly in the next few sentences regarding the results presented in Fig, 3D, the effects of overexpressing those genes are not uniform. While amputation of papillae in larvae overexpressing caPLCβ1/2/3 could inhibit metamorphosis almost completely, papilla cut seems to have a weaker effect on caGαq, caGαs, and bPAC-overexpressing larvae.

We added a description explaining that caPLCβ1/2/3 was the most sensitive to papilla amputation, and the possibility that PLCβ1/2/3 works specifically in the papillae (page 11, lines 9-11): “Among these experiments, caPLCβ1/2/3 overexpression was the most sensitive to papilla amputation, suggesting that PLCβ1/2/3 acts specifically in the papillae during metamorphosis.”

Page 9, the paragraph on using the fluorescent cAMP indicator; there is a discrepancy between the described developmental time when the authors conducted this experiment and the metamorphosis competent timing (after 24hpf) described on page 7. On page 26, the authors describe "The Pink Flamindo mRNA-injected larvae were immobilized on Poly L lysine-coated glass bottom dishes at 20-21 hpf...". Did the authors start stimulating the larvae to observe the fluorescent signal soon after immobilization, or wait several hours until the larvae passed 24hpf and then conduct the experiment?

The latter is the case. The immobilized larvae were kept until they acquired the competence for metamorphosis and then stimulation/recording was carried out. This point is described in the Materials and Methods section of the revised version (page 29, lines 16-18):

"The Pink Flamindo mRNA-injected larvae were immobilized on Poly L lysine-coated glass-bottom dishes at 20-21 hpf, and stimulated their adhesive papillae around 25 hpf."

Page 10, the description "...Gαq morphants initiated metamorphosis when caGαs was overexpressed in the nervous system (Figure 4F)". It should be noted that the result is only a partial rescue. To be precise, this description needs to be modified.

We changed the sentence to reflect the results more precisely (page 14, lines 2-3): “Moreover, caGαs overexpression in the nervous system significantly, although not perfectly, ameliorated the effect of Gαq MO (Figure 4F).”

Page 12-13, This description and the figure 5E presented is a bit confusing to me. The figure legend for 5E: "GABA is necessary for Ca2+ transient in the adhesive papillae (arrow)" But the arrow in this image points to a place with no fluorescent signal, and on the upper corner it labeled as "29% (n=17)". Does that mean the proportion of "no Ca2+ increase after stimulation" was 29% among the 17 samples examined? Or actually, is the other way around that 81% of the examined larvae did not show Ca2+ signal increase after stimulation?

The latter is the case. We added a caption explaining this clearly in the Figure legend: “The percentage and number exhibit the rate of animals showing Ca²⁺ transient in the papillae.”

Page 13, second paragraph; I do not agree with the overly simplified description that "GABA significantly ameliorated the metamorphosis-failed phenocopies of Gαq, PLCβ, and Gαs morphants". As shown in Fig. 5F-H, adding GABA exerts different levels of partial rescue effect on each morphant, and thus should be described clearly.

When the outliers are neglected, the effect of GABA is most evident in Gαs knockdowns. This suggests that the target(s) of GABA signaling is more likely to be Gq pathway components. We added the following sentence to the revised version (page 15, lines 14-16):

“Among the three morphants, GABA exhibited the most effective rescues in Gαs knockdowns than Gαq and PLCβ.”

In addition, we think this sentence establishes a more logical connection with the sentence that follows it: “These results could be explained by assuming enhancement of the Gq pathway by GABA through PLCβ and another GABA-mediated metamorphic pathway bypassing Gq components.” Thank you for your suggestion.

The section "Contribution of Gi to metamorphosis" confirmed the possibility that GABA signaling targets Gq pathway components.

Page 13, the first paragraph on "Contribution of Gi to metamorphosis"; the description that "The knockdown of this gene (Gαi) exhibited a significantly reduced rate of metamorphosis;..." is misleading. I would suggest modifying the entire sentence as "The knockdown of this gene (Gαi) exhibited a moderate (although statistically significant) reduction of metamorphosis rate, suggesting the presence of another Gαi regulating metamorphosis".

Thank you for your suggestion. We modified the sentence (page 16, lines 2-4 in the revised version) as recommended. We believe the description is much improved.

Page 20, the last sentence about Ciona papilla neurons expressing transcription factor Islet; the authors seem to attempt to make some comparison with the vertebrate pancreatic beta cells in this paragraph, but the comparison and the argument are not fully developed in this current format.

To deepen this discussion, we added the following sentence (page 23, lines 10-12): “The atypical secretion of GABA might depend on the transcription factor like Islet shared between Ciona papilla neurons and vertebrate beta cells.”

However, we would like to limit the depth of our discussion on this point, as we hope to expand on it further in future studies.

Other suggestions:

Page 3, second paragraph: as they become unable to "move" after metamorphosis -> "relocate"

We corrected the word as suggested.

Page 4, second paragraph: In the first sentence, the author states the current understanding of chordate phylogeny and cites Delsuc et al. 2006 Nature paper at the end of this sentence. However, in this paper cephalochordates were erroneously grouped with echinoderms, and thus chordates did not form a monophyletic clade. A later paper by Bourlat et al, (Nature 444:85-88, 2006) corrected this problem, and subsequently Dulsuc et al. also published another paper (genesis, 46:592-604, 2008) with broader sampling to overcome this problem. These later publications need to be included for the sake of correctness.

We added this reference.

Page 14, regarding the redundant function of the typical Gαi protein in the papillae; the authors may try double KD of Gαi and dvGαi_Chr2 in their experimental system to test this idea.

We carried out double knockdown of typical Gai and dvGαi_Chr2. However, we could not address their redundant role sufficiently because most of the double knockdown larvae exhibited severe shape malformation.

dvGαi_Chr4 is also expressed in the papillae. We carried out knockdown of this gene, to find that the knockdown resulted in very minor but statistically significant reduction of the metamorphosis rate, suggesting that this Gai also plays a supportive role in metamorphosis. We also carried out double knockdown of dvGαi_Chr2 and dvGαi_Chr4. The double KD larvae exhibited responsiveness to GABA, probably because of the presence of typical Gai.

These results are described on page 16, lines 2-18, and the data are shown in Supplementary Figure S6A-D of the revised version.

Responses to the Reviewing editor's comments:

"Larvae of the ascidian Ciona initiate metamorphosis tens of minutes after adhesion to a substratum via its adhesive organ." - Larvae is plural so change to 'via their adhesive organ'

The sentence was corrected as suggested.

"Metamorphosis is a widespread feature of animal development that allows them" - revise the sentence, e.g. "Metamorphosis is a widespread feature of development that allows animals"

The sentence was corrected as suggested.

"GABA synthase (GAD)" GAD is not called GABA synthase but glutamate decarboxylase - clarify, e.g. encoding the enzyme synthesizing GABA called glutamate decarboxylase (GAD)

This part was corrected exactly as suggested. Thank you.

"IP3 is received by its receptor on the endoplasmic reticulum (ER) and releases calcium ion (Ca2+ )" revise to "IP3 is received by its receptor on the endoplasmic reticulum (ER) that releases calcium ion (Ca2+ )"

The sentence was corrected as suggested.

"Moreover, GPCR is implicated as the mediator of settlement" - GPCRs are implicated

This sentence was modified as suggested.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

We thank the reviewers for providing valuable comments and suggestions for improving the manuscript.

Response to reviewer comments:

Reviewer-1

Comment 1: Major concern is the study lacks rigor in several areas where n=2, results are not quantified with statistics. They need to run power analysis and increase their samples sizes. Please include statistics on all measurements. Filamentous actin staining and alpha-sma is used to visualize mechanosensing but also in other cell activities such as cell contractility for movement, cell to substrate adhesion, cell division, etc. They need to query more mechanosensing related pathways (Piezo1/2, Yap/taz-Hippo, integrin-Focal Adhesion Kinase, etc) to show that mechanosensing changed.

Response: We have increased the sample size to a minimum of n=3 in most cases. However, a few experiments will require more time to increase sample size, as mentioned below.

Our data emphasized the role of Rac1 and SRF. We understand that other molecular players may also be involved in sensing or responding to mechanical forces, but surveying multiple families of candidates without a specific hypothesis or functional experiment is beyond the scope of this study.

__Comment 2: __Fig. 1: In panel E, the cranial bone area measurement is not normalized to mitigate the possible effect of individual differences.

Response: We have re-quantified the data with normalization to the length of the skull.

__Comment 3: __In Fig. 2 the authors mentioned many phenotypical changes (bone length changes, gap thickness change, apex thickness change, etc.) based on histology stain, none of them are quantified to show a significant difference between Rac1-WT and Rac1-KO.

Response: In Fig. 2A, we present the gross morphology of the Rac1-KO embryos and only discuss the tissue defects like edema, hematoma, and hypoplasia, confirmed through H&E as shown in Fig. 2C. We also show the apical limits of the intact calvaria in Fig. 2D, consistent with the calvaria defects observed at birth. In fact, we do not discuss any “bone length changes, gap thickness, or apex thickness change” in this section as suggested by the reviewer. To address the request for more quantification we have added measurement of the edematous area of the apical mesenchyme at E14.5 (Fig. 2C), and this is now shown in Suppl. Fig. 1E. We also added quantification of embryo genotypes and Chi-square tests, now shown in Suppl. Fig. 1D.

Comment 4: Fig. 2 In panel D, with only 2 embryos per group is not enough for quantitation

Response: We plan to increase the number of embryos during the revision period.

Comment 5: Fig. 2 In panel D, the two arrows in the Rac1-KO mutants are not easy to catch.

Response: We made the arrows bigger and bolder.

Comment 6: Fig. 3 The thickness quantification is not performed.

Response: We added quantification in Fig. 3D.

Comment 7: Fig. 3 The images show an obvious curve change of the apex between the control and mutant. Such change is not discussed in the results. Is it due to histology issue?

Response: We do not think it is due to technical issues but reflects a real change in the shape of the apex of the head. We modified the graphical representation in Figure 3E to reflect this change in curvature. We also added the following sentence to the results on page 7: “We also noted a loss of curvature in the apex of the Rac1-KO head at E13.5, which correlated with loss of aSMA+ mesenchymal cells and thinning of the EMM (Fig. 3E).”

__Comment 8: __The merged layer did not show S100a6. While the authors are showing apical expansion of the mesenchyme toward the dermis and meninges, it is hard to track where they are without a merged image.

Response: We added merged images.

Comment 9: Fig. 4 In panel B, 2 biological replicates per genotype are very low.

__Response: __The effect of Rac1-KO on cell cycle is already known (Moore et al. 1997; Nikolova et al. 2007; Gahankari et al. 2021), and our result is supported by in vivo quantification of Tom+Edu+ cells in different regions of the embryonic head shown in Fig. 4A. We prefer not to repeat this assay.

Comment 10: Fig. 4 There is no cell death data.

Response: We will generate data on cell death during the revision period.

__Comment 11: __Fig. 5 In panel B, the GAPDH western plot bands in the mutants seem to be thinner than those of controls.

Response: We verified equal loading with a Ponceau stain, so this minor change in the GAPDH level could be due to biological differences in the protein level. Nevertheless, by our estimation this minor difference does not explain away the major difference in Rac1 and Srf levels.

__Comment 12: __Though the immunostain showed a decrease in signal intensity, it is hard to know whether the decrease is significant enough across all Rac1-KO mutants. They need to measure the fluorescence intensity and perform statistics.

Response: We will generate better images of SRF staining and quantify the difference between Rac1-WT and Rac1-KO during the revision period.

Comment 13: Fig. 6: Similar as Fig. 2, there is no quantification and n=1 per genotype is not enough

Response: During the revision period we will increase the number of E12.5 Srf-KO and Srf-WT embryos to n=3 for Figure 6G. All other panels currently have n=7 or greater.

Comment 14: Fig. 7: Need quantification between Srf-KO and Rac1-KO with statistics to show they are not different, but both significantly different from WTs

Response: In Figure 7D we have added quantification of aSMA area in Srf-KO and Rac1-KO. These results show that both mutants have a similar phenotype with reduced aSMA expression compared to their respective WT littermates, which supports the conclusion that they work in the same pathway. We do not agree with the reviewer that the two mutants should show no statistical difference, because Rac1 and Srf are different genes with overlapping but also non-overlapping functions. During the revision period we will add more Srf-KO embryos and repeat the statistical analysis.

Comment 15: Supplement Fig.2: No image showing the time point before E11.5.

Response: We will add an E10.5 time point during the revision period.

Comment 16: Supplement Fig.3: The ventral view of Rac1-WT does not have the same angle as it shows in Rac1-KO. Makes harder to see the difference between control and mutant.

Response: We adjusted the brightness/contrast to make the difference clearer.

Comment 17: Supplement Fig.4 &7: The alkaline phosphatase stained area needs to be normalized to some other metric because the embryos could be different size.

Response: We normalized to the width of the eye and is now represented in Suppl. Fig. 4 and 7.

Comment 18: Supplement Fig 6 A: The legend and figure don't match. Is it E13.5 or 14.5. Panel 6B needs better images without curling of the tissue.

Response: This has been fixed. The immunostaining images in Suppl. Fig. 6A is E14.5. Panel B is now replaced with better images in the revised manuscript.

Reviewer-2

__Comment 1.1: __In Fig. 5, links between Rac1, SRF, αSMA, and contractility in mesenchymal cells are shown. Molecular analyses (Western blot and qPCR) were performed using primary cultured mesenchymal cells (prepared after freed from the epidermal population). Although use of cells prepared from E18.5 embryos may have been chosen by the authors for the safe isolation of the mesenchymal population without contamination of epidermal cells, this reviewer finds that anti-SRF immunoreactivity is weaker at E13.5 than at E12.5 (throughout the section including the mesencephalic wall) and therefore wonder whether SRF expression changes in a stage-dependent manner. So, simply borrowing results obtained from E18.5-derived cells for describing the scenario around E12.5 and E13.5 is a little disappointing point found only here in this study.

Response: In fact, the reason we chose E18.5 was to get enough cells to do the experiments in Figure 5A-D without extensive passaging and/or immortalization, which would undoubtedly cause the cells to deviate from their in vivo character as they become adapted to growing on plastic with 10% serum. Therefore, we prefer not to change the cells as suggested by the reviewer.

__Comment 1.2: __In Fig. 5F, it is difficult to clearly see "reduction" of SRF immunoreactivity in Rac1-KO. Therefore, quantification of %SRF+/totalTomato+ would be desired.

Response: __We will generate better images of SRF staining and quantify the difference between Rac1-WT and Rac1-KO during the __revision period.

__Comment 1.3: __Separately, direct comparison of spontaneous centripetal shrinkage of the apical/dorsal scalp tissues, which will occur in 30 min when prepared at E12.5 or E13.5 (Tsujikawa et al., 2022), between WT and Rac1-KO would strengthen the results in Fig. 5D. As KO is specific to the mesenchyme, the authors do not have to worry about removal of the epidermal layer (which would be much more difficult at E12.5-13.5 than E18.5). If the degree of centripetal shrinkage of the "epidermis plus mesenchyme" layers were smaller in Rac1-KO, it would be interpreted to be mainly due to poorer recoiling activity and contractility of the Rac1-KO mesenchymal tissue.

Response: __We will try to perform the centripetal shrinkage assays as shown by Tsujikawa et al., during the __revision period.

Comment 2: The authors favor "apical" vs. "basolateral" to tell the relative positions in the embryonic head, not only in the adult head. But "apical" vs. "basolateral" should be accompanied with dorsal vs. ventral at least at the first appearance. Apical-to-basal axis or apex vs. basolateral by itself can provide, in many contexts, impressions that epithelial layers/cells are being discussed. Please note that the authors also use "caudal" (in the embryonic head). Usually, a universally defined anatomical axis perpendicular to the rostral-to-caudal axis is the dorsal-to-ventral axis.

Response: Apologies for confusing terminology. The terminology is now defined uniformly according to the anatomical axis.

Comment 3: One of the authors' statements in ABSTRACT "In control embryos, α-smooth muscle actin (αSMA) expression was spatially restricted to the apical mesenchyme, suggesting a mechanical interaction between the growing brain and the overlying mesenchyme" and a similar one in RESULTS "αSMA was not detected in the basolateral mesenchyme of either genotype from E12.5-E14.5 (Suppl. Fig. 4A), suggesting restriction of the mechanosensitive cell state to the apical mesenchyme" need to be at least partly revised, taking previous publication about the normal αSMA pattern in the embryonic head into account more carefully. Tsujikawa et al. (2022) described "Low-magnification observations showed superficial immunoreactivity for alpha smooth muscle actin (αSMA), which has been suggested to function in cells playing force-generating and/or constricting roles; this immunoreactivity was continuously strong throughout the dorsal (calvarial) side of the head but not ventrally toward the face, producing a staining pattern similar to a cap (Figure 2A)" . Therefore, in this new paper, descriptions like "we observed ...., consistent with ....(2022)" or "we confirmed .... (2022)" would be more accurate and appropriate regarding this specific point. Such a minor change does not reduce this study's overall novelty at all.

Response: Thank you for the correction. We have replaced the terminology and cited the article (Tsujikawa et al., 2022) appropriately, crediting their finding.

Comment 4: It would be very helpful if the authors provide a schematic illustration in which physiological and pathological scenarios (at the molecular, cellular, and tissue levels found or suggested by this study) are shown.

Response: We have added a schematic representation of the molecular changes happening in the apical head development because of Rac1- and Srf-KO, and it is represented in Suppl. Fig. 7C.

---

Comment 5: Despite being put in the title, "mechanosensing" by mesenchymal cells is not directly assessed in this study. If appropriate, something like "mechano-functioning" would be closer to what the authors demonstrated.

__Response: __We changed the title to refer to “mechano-responsive mesenchyme”. We think this is appropriate because the cells of interest have reduced aSMA and reduced proliferation, both of which are known to occur, at least in part, as responses to mechanical inputs.

Reviewer-3

Comment 1: Prrx1-Cre targets calvarial mesenchyme and Suzuki et al., 2009 showed that Prrx1-Cre mediated loss of Rac1 lead to calvarial bone phenotype due to incomplete fusion of the skull. While this phenotype was not studied in detail, the statement in the intro and discussion that the calvarial phenotype has not been recapitulated in mice is incorrect.

Response: Suzuki et al showed incomplete fusion of the skull. Although the skull is a tissue that is affected in AOS, it is not akin to the scalp and calvaria aplasia that typifies AOS. Our result stands apart from this. We clarified our position as such:

Introduction (page 4): “Nevertheless, the calvaria phenotype seen in AOS individuals has not been explored in detail or fully recapitulated in mice.”

Discussion (page 11): “Previous studies have demonstrated the role of Rac1 in mesenchyme-derived tissues, but they did not recapitulate AOS phenotypes.”

Comment 2: The authors show that Pdgfra-Cre induced knockout of Rac1 leads to lower-than-expected numbers of Rac1-cKO embryos at E18.5 and P1. Phenotypic analysis shows that the earliest phenotype is blebbing and hematoma in the nasal region at E11.5/12.5. It is stated that this was resolved at E18.5. It is unclear if this is truly a resolution of the phenotype or that these embryos fail to survive until E18.5. Do 100% of the Rac1-cKO embryos exhibit the blebbing/hematoma at E11.5/12.5? What is the observed number/percentage of Rac1-cKO embryos at E11.5/12.5? If the observed percentage of Rac1-cKO is similar to that at E18.5 (lower than the expected 25%), this would support resolution. If the observed ratio is as expected at E11.5/12.5, then this would support embryonic loss before E18.5 rather than phenotypic resolution.

Response: Please note that 100% (n=12) of E12.5 Rac1-KO embryos displayed nasal and mild caudal edema as exhibited in Fig. 2A, but none (n=16) had blebbing/hematoma by E18.5. We added tables for the number of embryos recovered at E12.5 and E18.5 to Supplemental Figure 1. These results show that the percentage of mutants at E12.5 was 21.42%, not significantly different from the expected frequency (p = 0.5371). At E18.5, the percentage dropped slightly to 18.3%, but still not significantly different from expected (p = 0.1545). The significant change in frequency of blebbing/hematoma from E12.5 to E18.5, without any significant change in the frequency of mutants, supports phenotypic resolution of the early blebbing/hematoma.

Comment 3: It is stated that brain shape is altered in Rac1-cKO embryos at E14.5 and E18.5 and concluded that these shape differences are secondary to the cranial defects. Pdgfra+ cells gives rise to the meninges and if the Pdgfra-Cre line recapitulates this expression, then loss of the ubiquitously expressed Rac1 in the meninges could lead to a primary defect in the brain, which may lead to secondary defects in the calvarium and scalp. Their conclusion should recognize other possibilities.

Response: We agree it is possible that there are meninges defects that secondarily change the shape of the brain, and we added a mention of this possibility. It is highly unlikely that scalp defects are only secondary to brain changes because the first observable phenotypes are in the EMM that gives rise to the scalp.

Comment 4: The TdTom staining in wholemount at E13.5 (Supplemental Figure 2B) is difficult to appreciate in the image shown.

Response: At E11.5 there is good contrast between labeled cranial structures and non-labeled body. At E13.5, Tomato appears in most of the mesenchymal cells in the embryo, so there is not as much contrast. The lack of contrast at E13.5 may cause the reviewer think there is something wrong with the image.

Comment 5: The idea that the EMM laminates into the meninges and scalp layers is not new and should be properly cited (Vu et al., 2021, Scientific Reports). The following paper should also be cited on the use of alpha-SMA (Acta2) as a marker of the anterior calvaria mesenchyme: Holms et al., 2020 Cell Reports.

Response: Thank you. We are happy to add those citations.

Comment 6: It is concluded that meningeal development is maintained in the cKO; however, this conclusion was based on a single marker (S100a6) that is both expressed in the presumptive meninges and dermis and greatly reduced overall in the cKO. This conclusion should be softened or other markers used to show that the meninges is indeed normal.

Response: We softened the conclusion on the meninges in the revised manuscript, as this part of the phenotype is was not our focus but it would be a good thing to look at in the future.

Comment 7: The overlap of S100a6 and alpha-SMA is difficult to appreciate in the images shown in Figure 3. Since this is important to the conclusion, co-staining should be done. If co-staining cannot be done due to the primary antibodies' origins, then ISH should be done.

Response: We added merged images.

Comment 8: It is concluded that reduced alpha-SMA suggests an early failure of Rac-cKO cells to respond to the mechanical environment. While this is one possibility, the reduction of alpha-SMA may simply be due to a reduction of these cells resulting from failed differentiation, decreased proliferation, or increased apoptosis.

Response: We think the fact that aSMA is downregulated in cultured cells strongly argues against it being a trivial consequence of reduce proliferation etc. Nevertheless, we softened our conclusion to allow for some of these things to also contribute to the reduced aSMA expression. We will check apoptosis during the revision period.

Comment 9: The conclusion that alpha-SMA is a transient population only present in apical cranial mesenchyme between E12.5-14.5 is not consistent with prior studies: Holms et al., 2020 Cell Reports; Holms et al., 2021 Nature Communications; Farmer et al., 2021 Nature Communications; Takeshita et al., 2016 JBMR.

Response: There is no contradiction. Our statements are based on antibody staining where it is very evident that a-SMA-expressing cells are detectable throughout the apical mesenchyme between E12.5 and E14.5. But at E18.5 we do not see this kind of broad aSMA expression the apical head, suggesting a transient and spatially restricted population of cells in the apical mesenchyme. This is consistent with the studies from Tsujikawa et al., 2022 and Angelozzi et al., 2022. The papers mentioned by the reviewer are only focused on the suture mesenchyme. They do not claim there is broad aSMA/Acta2 expression in the apical head, but only in a spatially restricted subpopulation of suture mesenchymal cells.

Comment 10: In the SRF immunostaining results in control and Rac1-cKO embryos, it is difficult to appreciate the nuclear localization at E12.5 in Figure 5E, as the DAPI is over saturated, and the image quality is poor. The image quality is also poor in Figure 5F.

Response: We will generate better images of SRF staining and quantify the difference between Rac1-WT and Rac1-KO during the revision period.

Comment 11: To what extent is the expression/localization of MRTF, the transcriptional co-activator of SRF, altered in the calvarial mesenchyme of Rac1-cKO embryos? Changes in MRTF would strengthen the link between Rac1 and SRF.

Response: We do not know how MRTF expression/localization changes in the embryo tissue, but western blot data on Rac1-KO fibroblasts revealed a reduction in expression/nuclear localization of MRTF-A/B that mirrored the changes in SRF. We added these blots to Figure 5A. However, as noted at the end of the discussion, MRTF is not always required for SRF function in vivo ( Dinsmore, Elife 2022). The MRTFA/B-KO is a possibility for future work.

Comment 12: Hypoplasia of the apical mesenchyme (Figure 6G, inset 1) in Srf-cKO is difficult to see.

Response: During the revision period we will increase the number of E12.5 Srf-KO and Srf-WT embryos to n=3 for Figure 6G and replace the picture with a better one.

Comment 13: Generally, the organization of the data into many main and supplemental Figures makes the flow difficult to follow.

__Response____: __We understand the concern, but we have tried our best to organize the most important data into main figures and the relevant but less essential data into supplemental figures.

Comment 14: SFR interacts with Pdgfra interacts genetically with Srf in neural crest cells in craniofacial development, with Srf being a target of PDGFRa signaling (Vasudevan and Soriano, 2015, Dev Cell). Since the Pdgfra-Cre line used here is hemizygous, is important that the control used to look at SRF expression in the Rac1-cKO is Pdgfra-Cre+.

Response: It is standard practice to include some Cre+ mice in the control set to reveal whether Cre has toxic effects in the cells of interest. To the reviewer’s concern about genetic interactions between the Pdgfra gene and Srf, this should not be relevant here because the Pdgfra-Cre used in our study is a transgene and does not affect the endogenous Pdgfra gene.

Comment 15: The text size in all figures is too small and varies throughout, making it difficult to read.

Response: To fit the panel in the Word document, the figure is resized. This should not be an issue in the final manuscript.

Comment 16: Details about the pulse-chase timing of the EdU experiments should be included in the results. Also, does n = 3 for each stage and each genotype? I would be helpful to include a representative section for a control and cKO littermate pair.

Response: The details are now included in the methods section. Yes, n=3 in each stage and genotype (Fig. 4A). The representative images are also included.

Comment 17: The relative sizing of the panels within and between figures is haphazard. Some are very large and others very small (Figure 2, 6, Supplemental Figure 1, 2, 6, 7).

Response: The image panels are fixed in the revised manuscript.

Comment 18: In Figure 5A and F, the titles "E12.5" and "E13.5" are in italics.

Response: The fonts for the figures are fixed in the revised manuscript.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Recommendations for the authors):

We appreciate the reviewers' thoughtful comments and suggestions. Below, we provide point-by-point responses to the recommendations and outline the updates made to the manuscript.

(1) Discussion, "the obvious experiment is to manipulate a neuron's anatomical embedding while leaving stimulus information intact."] The epiphenomenon can arise from the placement and types of a neuron's neurotransmitters and neuromodulators, too.

The content of vesicles released by a neuron is obviously of great importance in determining postsynaptic impact. However, we’re suggesting that (assuming vesicular content is held constant) the anatomically-relevant patterning of spiking might additionally affect the postsynaptic neuron’s integration of the presynaptic input. To avoid confusion, we updated the text accordingly: “the obvious experiment is to manipulate a neuron's anatomical embedding while minimally impacting external and internal variables, such as stimulus information and levels of neurotransmitters or neuromodulators” (Line 594 - 596).

(2) “In all conditions, the slope of the input duration versus sensitivity line was still positive at 1,800 seconds (Fig. 3B)". This may suggest that the estimate of the calculated statistics (ISI, PSTH) is more reliable with more data, rather than (or in addition to) specific information being extracted from faraway time points. Another potential confound is the training statistics were calculated from all training data, so the test data is a better match to training data when test statistics are calculated from more data. Overall, the validity of the conclusions following this observation is not clear to me.

This is a great point. Accordingly, we revised the text to include this possibility: “Because the training data were of similar duration, this could be explained by either of two possibilities. First, the signal is relatively short, but noisy—in this case, extended sampling will increase reliability. Second, the anatomical signal is, itself, distributed over time scales of tens to hundreds of seconds.” (Line 252 - 255).

(3) "This further suggests that there is a latent neural code for anatomical location embedded within the spike train, a feature that could be practically applied to determining the brain region of a recording electrode without the need for post-hoc histology". The performance of the model at the subregion level, which is a typical level of desired precision in locating cells, does not seem to support such a practical application. Please clarify to avoid confusion.

The current model should not be considered a replacement for traditional methods, such as histology. Our intention is to convey that, with the inclusion of multimodal data and additional samples, a computational approach to anatomical localization has great promise. We updated the manuscript to clarify this point: “While significantly above chance, the structure-level model still lacks the accuracy for immediate practical application. However, it is highly likely that the incorporation of datasets with diverse multi-modal features and alternative regions from other research groups will increase the accuracy of such a model. In addition, a computational approach can be combined with other methods of anatomical reconstruction.” (Line 355 - 359).

Additionally, we directly addressed this point in our original manuscript (Discussion section: Line 498 - 505 in the current version). Furthermore, following the release of our preprint, independent efforts have adopted a multimodal strategy with qualitatively similar results (Yu et al., 2024). Other recent work expands on the idea of utilizing single-neuron features for brain region/structure characterization (La Merre et al., 2024).

Yu, H., Lyu, H., Xu, E. Y., Windolf, C., Lee, E. K., Yang, F., ... & Hurwitz, C. (2024). In vivo cell-type and brain region classification via multimodal contrastive learning. bioRxiv, 2024-11.

Le Merre, P., Heining, K., Slashcheva, M., Jung, F., Moysiadou, E., Guyon, N., ... & Carlén, M. (2024). A Prefrontal Cortex Map based on Single Neuron Activity. bioRxiv, 2024-11.

(4) "These results support the notion the meaningful computational division in murine visuocortical regions is at the level of VISp versus secondary areas.". The use of the word "meaningful" is vague and this conclusion is not well justified because it is possible that subregions serve different functional roles without having different spiking statistics.

Precisely! It is well established that different subregions serve different functional purposes - but they do not necessitate different regional embeddings. It is important to note the difference between stimulus encoding and the embedding that we are describing. As a rough analogy, the regional embedding might be considered a language, while the stimulus is the content of the spoken words. However, to avoid vague words, we revised the sentence to “These results suggest that the computational differentiability of murine visuocortical regions is at the level of VISp versus secondary areas.” (Line 380 - 381)

(5) Figure 3D left/right halves look similar. A measure of the effect size needs to accompany these p-values.

We assume the reviewer is referring to Figure 3E. Although some of the violin plots in Figure 3E look similar, they are not identical. In the revision, we include effect sizes in the caption.

(6) Figure 3A, 3F: Could uncertainty estimates be provided?

Yes. We added uncertainty estimates to the text (Line 272 - 294) and to the caption of Figure S2, which displays confusion matrices corresponding to Figure 3A. The inclusion of similar estimates for 3F would be so unwieldy as to be a disservice to the reader—there are 240 unique combinations of stimulus parameters and structures. In the context of the larger figure, 3F serves to illustrate a relationship between stimulus, region, and the anatomical embedding.

(7) Page 21. "semi-orthogonal". Please reword or explain if this usage is technical.

We replaced “semi-orthogonal” with “dissociable” (Line 549).

(8) Page 11, "This approach tested whether..."] Unclear sentence. Please reword.

We changed “This approach tested whether the MLP’s performance depended on viewing the entire ISI distribution or was enriched in a subset of patterns” to “This approach identified regions of the ISI distribution informative for classification” (Line 261).

Reviewer #2 (Recommendations for the authors):

We appreciate the reviewer’s comments and summary of the results. We agree that the introductory results (Figs. 1-3) are not particularly compelling when considered in isolation. They provide a baseline of comparison for the subsequent results. Our intention was to approach the problem systematically, progressing from well-established, basic methods to more advanced approaches. This allows us to clearly test a baseline and avoid analytical leaps or untested assumptions. Specifically:

● Figure 1 provides an evaluation of the standard dimensionality reduction methods. As expected, these methods yield minimal results, serving as a clear baseline. This is consistent, for example, with an understanding of single units as rate-varying Poisson processes.

● Figures 2 and 3 then build upon these results with spiking features frequent in neuroscience literature such as firing rate, coefficient of variation, etc using linear supervised and more detailed spiking features such as ISI distribution using nonlinear supervised machine learning methods.

By starting from the standpoint of the status quo, we are better able to contextualize the significance of our later findings in Figures 4–6.

Response to Specific Points in the Summary

(6) Separability of VISp vs. Secondary Visual Areas

I found the entire argument about visual areas somewhat messy and unclear. The stimuli used might not drive the secondary visual areas particularly well and might necessitate task engagement.

We appreciate your feedback that the dissection of visual cortical structures is unclear. To summarize, as shown in the bottom three rows of Figure 6, there is a notable lack of diagonality in visuocortical structures. This means that our model was unable to learn signatures to reliably predict these classes. In contrast, visuocortical layer is returned well above chance, and superstructures (primary and secondary areas) are moderately well identified, albeit still well above chance.

Consider a thought experiment, if Charlie Gross had not shown faces to monkeys to find IT, or Newsome and others shown motion to find MT and Zeki and others color stimuli to find V4, we would conclude that there are no differences.

The thought experiment is misleading. The results specifically do not arise from stimulus selectivity—much of Newsome’s own work suggests that the selectivity of neurons in IT etc. is explained by little more than rate varying Poisson processes. In this case, there should be no fundamental anatomical difference in the “language” of the neurons in V4 and IT, only a difference in the inputs driving those neurons. In contrast, our work suggests that the “language” of neurons varies as a function of some anatomical divisions. In other words, in contrast to a Poisson rate code, our results predict that single neuron spike patterns might be remarkably different in MT and IT— and that this is not a function of stimulus selectivity. Notably, the anatomical (and functional) division between V1 and secondary visual areas does not appear to manifest in a different “language”, thus constituting an interesting result in and of itself.

We regret a failure to communicate this in a tight and compelling fashion on the first submission, but hope that the revision is limpid and accessible.

Barberini, C. L., Horwitz, G. D., & Newsome, W. T. (2001). A comparison of spiking statistics in motion sensing neurones of flies and monkeys. Motion Vision: Computational, Neural, and Ecological Constraints, 307-320.

Bair, W., Zohary, E., & Newsome, W. T. (2001). Correlated firing in macaque visual area MT: time scales and relationship to behavior. Journal of Neuroscience, 21(5), 1676-1697.

Similarly, why would drifting gratings be a good example of a stimulus for the hippocampus, an area thought to be involved in memory/place fields?

The results suggest that anatomical “language” is not tied to stimuli. It is imperative to recall that neurons are highly active absent experimentally imposed stimuli, such as when an animal is at rest, when an animal is asleep, and when an animal is in the dark (relevant to visual cortices). With this in mind, also recall that, despite the lack of stimuli tailored to the hippocampus, neurons therein were still reliably separable from neurons in seven nuclei in the thalamus, 6 of which are not classically considered visual regions. Should these regions (including hippocampus) have been inert during the presentation of visual stimuli, there would have been very little separability.

(7) Generalization across laboratories

“[C]omparison across laboratories was somewhat underwhelming. It does okay but none of the results are particularly compelling in terms of performance.

Any result above chance is a rejection of the null hypothesis: that a model trained on a set of animals in Laboratory A will be ineffective in identifying brain regions when tested on recordings collected in Laboratory B (in different animals and under different experimental conditions). As an existence proof, the results suggest conserved principles (however modest) that constrain neuronal activity as a function of anatomy. That models fail to achieve high accuracy (in this context) is not surprising (given the limitations of available recordings)---that models achieve anything above chance, however, is.

Thus, after reading the paper many times, I think part of the problem is that the study is not cohesive, and the authors need to either come up with a tool or demonstrate a scientific finding.

We demonstrate that neuronal spike trains carry robust anatomical information. We developed an ML architecture for this and that architecture is publicly available.

They try to split the middle and I am left somewhat perplexed about what exact scientific problem they or other researchers are solving.

We humbly suggest that the question of a neurons “language” is highly important and central to an understanding of how brains work. From a computational perspective, there is no reason for a vast diversity of cell types, nor a differentiation of the rules that dictate neuronal activity in one region versus another. A Turing Complete system can be trivially constructed from a small number of simple components, such as an excitatory and inhibitory cell type. This is the basis of many machine learning tools.

Please do not confuse stimulus specificity with the concept of a neuron’s language. Neurons in VISp might fire more in response to light, while those in auditory cortex respond to sound. This does not mean that these neurons are different - only that their inputs are. Given the lack of a literature describing our main effect—that single neuron spiking carries information about anatomical location—it is difficult to conclude that our results are either commonplace or to be expected.

I am also unsure why the authors think some of these results are particularly important.

See above.

For instance, has anyone ever argued that brain areas do not have different spike patterns?

Yes. In effect, by two avenues. The first is a lack of any argument otherwise (please do not conflate spike patterns with stimulus tuning), and the second is the preponderance of, e.g., rate codes across many functionally distinct regions and circuits.

Is that not the premise for all systems neuroscience?

No. The premise for all systems neuroscience (from our perspective) is that the brain is a) a collection of interacting neurons and b) the collective system of neurons gives rise to behavior, cognition, sensation, and perception. As stated above, these axiomatic first principles fundamentally do not require that neurons, as individual entities, obey different rules in different parts of the brain.

I could see how one could argue no one has said ISIs matter but the premise that the areas are different is a fundamental part of neuroscience.

Based on logic and the literature, we fundamentally disagree. Consider: while systems neuroscience operates on the principle that brain regions have specialized functions, there is no a priori reason to assume that these functions must be reflected in different underlying computational rules. The simplest explanation is that a single language of spiking exists across regions, with functional differences arising from processing distinct inputs rather than fundamentally different spiking rules. For example, an identical spike train in the amygdala and Layer 5 of M1 would have profoundly different functional impacts, yet the spike timing itself could be identical (even as stimulus response). Until now, evidence for region-specific spiking patterns has been lacking, and our work attempts to begin addressing this gap. There is extensive further work to be conducted in this space, and it is certain that models will improve, rules will be clarified, and mechanisms will be identified.

Detailed major comments

(1) Exploratory trends in spiking by region and structure across the population:

The argument in this section is that unsupervised analyses might reveal subtle trends in the organization of spiking patterns by area. The authors show 4 plots from t-SNE and claim to see subtle organization. I have concerns. For Figure 1C, it is nearly impossible to see if a significant structure exists that differentiates regions and structures. So this leads certain readers to conclude that the authors are looking at the artifactual structure (see Chari et al. 2024) - likely to contribute to large Twitter battles. Contributing to this issue is that the hyperparameter for tSNE was incorrectly chosen. I do think that a different perplexity should be used for the visualization in order to better show the underlying structure; the current visualization just looks like a single "blob". The UMAP visualizations in the supplement make this point more clearly. I also think the authors should include a better plot with appropriate perplexity or not include this at all. The color map of subtle shades of green and yellow is hard to see as well in both Figure S1 and Figure 1.

In response to the feedback, we replaced t-SNE/UMAP with LDA, while keeping PCA for dimensionality reduction.

As stated in the original methods, t-SNE/UMAP hyperparameters were chosen based on the combination that led to the greatest classifiable separability of the regions/structures in the space (across a broad range of possible combinations). It just so happens that the maximally separable structure from a regions/structures perspective is the “blob”. This suggests that perhaps the predominant structure the t-SNE finds in the data is not driven by anatomy. If we selected hyperparameters in some other way that was not based specifically on regions/structures (e.g. simple visual inspection of the plots) the conformation would of course be different and not blob-like. However, we removed the t-SNE and UMAP to avoid further confusion.

The “muddy appearance” is not an issue with the color map. As seen in Figure 1B, the chosen colors are visibly distinct. Figure 1C (previous version) appeared muddy yellow/green because of points that overlap with transparency, resulting in a mix of clearly defined classes (e.g., a yellow point on top of a blue point creating green). This overlap is a meaningful representation of the separability observed in this analysis. We also tried using 2D KDE for visualization, but it did not improve the impression of visual separability.

We are removing p-values from the figures because they lead to the impression that we over-interpret these results quantitatively. However, we calculated p-values based on label permutation similar to the way R2 suggests (see previous methods). The conflation with the Wasserstein distances is an understandable misunderstanding. These are unrelated to p-values and used for the heatmaps in S1 only (see previous methods).

Instead of p-values, we now use the adjusted rand index, which measures how accurately neurons within the same region are clustered together (see Line 670 - 671, Figure 1C, and Figure S1) (Hubert & Arabie 1985). This quantifies the extent to which the distribution of points in dimensionally-reduced space is shaped by region/structure.

Hubert, L., & Arabie, P. (1985). Comparing partitions. Journal of Classification, 2(1), 193–218. https://doi.org/10.1007/BF01908075

(2) Logistic classifiers:

The results in this section are somewhat underwhelming. Accuracy is around 40% and yes above chance but I would be very surprised if someone is worried about separating visual structures from the thalamus. Such coarse brain targeting is not difficult. If the authors want to include this data, I recommend they show it as a control in the ISI distribution section. The entire argument here is that perhaps one should not use derived metrics and a nonlinear classifier on more data is better, which is essentially the thrust of the next section.

As outlined above, our work systematically increases in model complexity. The logistic result is an intermediate model, and it returns intermediate results. This is an important stepping stone between the lack of a result based on unsupervised linear dimensionality reduction and the performance of supervised nonlinear models.

From a purely utilitarian perspective, the argument could be framed as “one should not use derived metrics, and a nonlinear classifier on more data is better.” However, please see all of our notes above.

(3) MLP classifiers:

Even in this section, I was left somewhat underwhelmed that a nonlinear classifier with large amounts of data outperforms a linear classifier with small amounts of data. I found the analysis of the ISIs and which timescales are driving the classifier interesting but I think the classifier with smoothing is more interesting. So with a modest chance level decodability of different brain areas in the visual system, I found it somewhat grandiose to claim a "conserved" code for anatomy in the brain. If there is conservation, it seems to be at the level of the coarse brain organization, which in my opinion is not particularly compelling.

The sample size used for both the linear and nonlinear classifiers is the same; however, the nonlinear classifier leverages the detailed spiking time information from ISIs. Our goal here was to systematically evaluate how classical spike metrics compare to more detailed temporal features in their ability to decode brain areas. We chose a linear classifier for spike metrics because, with fewer features, nonlinear methods like neural networks often offer very modest advantages over linear methods, less interpretability, and are prone to overfitting.

Respectfully, we stand by our word choice. The term “conserved” is appropriate given that our results hold appreciably, i.e., statistically above chance, across animals.

(4) Generalization section:

The authors suggest that a classifier learned from one set of data could be used for new data. I was unsure if this was a scientific point or the fact that they could use it as a tool.

It can be both. We are more driven by the scientific implications of a rejection of the null.

Is the scientific argument that ISIs are similar across areas even in different tasks?

It appears so - despite heterogeneity in the tuning of single neurons, their presynaptic inputs, and stimuli, there is identifiable information about anatomical location in the spike train.

Why would one not learn a classifier from every piece of available data: like LFP bands, ISI distributions, and average firing rates, and use that to predict the brain area as a comparison?

Because this would obfuscate the ability to conclude that spike trains embed information about anatomy.

Considering all features simultaneously and adding additional data modalities—such as LFP bands and spike waveforms—has potential to improve classification accuracy at the cost of understanding the contribution of each feature. The spike train as a time series is the most fundamental component of neuronal communication. As a result, this is the only feature of neuronal activity of concern for the present investigation.

Or is the argument that the ISIs are a conserved code for anatomy? Unfortunately, even in this section, the data are underwhelming.

We appreciate the reviewer’s comments, but arrive at a very different conclusion. We were quite surprised to find any generalizability whatsoever.

Moreover, for use as a tool, I think the authors need to seriously consider a control that is either waveforms from different brain areas or the local field potentials. Without that, I am struggling to understand how good this tool is. The authors said "because information transmission in the brain arises primarily from the timing of spiking and not waveforms (etc)., our studies involve only the timestamps of individual spikes from well-isolated units ". However, we are not talking about information transmission and actually trying to identify and assess brain areas from electrophysiological data.

While we are not blind to the “tool” potential that is suggested by our work, this is not the primary motivation or content in any section of the paper. As stated clearly in the abstract, our motivation is to ask “whether individual neurons [...] embed information about their own anatomical location within their spike patterns”. We go on to say “This discovery provides new insights into the relationship between brain structure and function, with broad implications for neurodevelopment, multimodal integration, and the interpretation of large-scale neuronal recordings. Immediately, it has potential as a strategy for in-vivo electrode localization.” Crucially, the last point we make is a nod to application. Indeed, our results suggest that in-vivo electrode localization protocols may benefit from the incorporation of such a model.

In light of the reviewer’s concerns, we have further dampened the weight of statements about our model as a consumer-ready tool.

Example 1: The final sentence of the abstract now reads: “Computational approximations of anatomy have potential to support in-vivo electrode localization.”

Example 2: The results sections now contains the following text: “While significantly above chance, the structure-level model still lacks the accuracy for immediate practical application. However, it is highly likely that the incorporation of datasets with diverse multi-modal features and alternative regions from other research groups will increase the accuracy of such a model. In addition, a computational approach can be combined with other methods of anatomical reconstruction.” (Line 355 - 359).

Example 3: We replaced the phrase "because information transmission in the brain arises primarily from the timing of spiking and not waveforms (etc) " with the phrase “because information is primarily encoded by the firing rate or the timing of spiking and not waveforms (etc)” (Line 116 - 118).

(5) Discussion section:

In the discussion, beginning with "It is reasonable to consider . . ." all the way to the penultimate paragraph, I found the argumentation here extremely hard to follow. Furthermore, the parts of the discussion here I did feel I understood, I heavily disagreed with. They state that "recordings are random in their local sampling" which is almost certainly untrue when it comes to electrophysiology which tends to oversample task-modulated excitatory neurons (https://elifesciences.org/articles/69068). I also disagree that "each neuron's connectivity is unique, and vertebrate brains lack 'identified neurons' characteristic of simple organisms. While brains are only eutelic and "nameable" in only the simplest organisms (C. elegans), cell types are exceedingly stereotyped in their connectivity even in mammals and such connectivity defines their computational properties. Thus I don't find the premise the authors state in the next sentence to be undermined ("it seems unlikely that a single neuron's happenstance imprinting of its unique connectivity should generalize across stimuli and animals"). Overall, I found this subsection to rely on false premises and in my opinion it should be removed.

At the suggestion of R2, we removed the paragraph in question. However, we would like to address some points of disagreement:

We agree that electrophysiology, along with spike-sorting, quality metrics, and filtering of low-firing neurons, leads to oversampling of task-modulated neurons. However, when we stated that recordings are random in their local sampling, we were referring to structural (anatomical) randomness, not functional randomness. In other words, the recorded neurons were not specifically targeted (see below).

Electrode arrays, such as Neuropixels, record from hundreds of neurons within a small volume relative to the total number of neurons and the volume of a given brain region. For instance, the paper R2 referenced includes a statement supporting this: “... assuming a 50-μm ‘listening radius’ for the probes (radius of half-cylinder around the probe where the neurons’ spike amplitude is sufficiently above noise to trigger detection) …, the average yield of 116 regular-spiking units/probe (prior to QC filtering) would imply a density of 42,000 neurons/mm³, much lower than the known density of ~90,000 neurons/mm³ for excitatory cells in mouse visual cortex….”

If we take the estimated volume of V1 to be approximately 3 mm³, this region could theoretically be subdivided into multiple cylinders with a 100-μm diameter. While stereotaxic implantation of the probe mitigates some variability, the natural anatomical variability across individual animals introduces spatially random sampling. This was the randomness we were referring to, and thus, we disagree with the assertion that our claim is “almost certainly untrue.”

Additionally, each cortical pyramidal neuron is understood to have ~ 10,000 presynaptic partners. It is highly unlikely that these connections are entirely pre-specified, perfectly replicated within the same animal, and identical across all members of species. Further, there is enormous diversity in the activity properties of even neighboring cells of the same type. Consider pyramidal neurons in V1. Single neuron firing rates are log normally distributed, there are many of combinations of tuning properties (i.e., direction, orientation) that must occupy each point in retinotopic space, and there is powerful experience dependent change in the connectivity of these cells. We suggest that it is inconceivable that any two neurons, even within a small region of V1, have identical connectivity.

Minor Comments:

(1) Although the description of confusion matrices is good from a didactic perspective, some of this could be moved to methods to simplify the paper.

We thank the reviewer for the suggestion. However, given the broad readership of eLife, we gently suggest that confusion matrices are not a trivial and universally appreciated plotting format. For the purpose of accessibility, a brief and didactic 2-sentence description will make the paper far more comprehensible to many readers at little cost to experts.

(2) Figure 3A: It is concluded in their subsequent figure that the longer the measured amount of time, the better the decoding performance. Thus it makes sense why the average PSTHs do not show significant decoding of areas or structures

That is a good observation. However, all features were calculated from the same duration of data, except in Figure 3B, where we tested the effect of duration. The averaged PSTH was calculated from the same length of data as the ISI distribution and binned to have the same number of feature lengths as the ISI distribution (refer to Methods section). Therefore, we interpreted this as an indication of information degradation through averaging, rather than an effect of data length (Line 234 - 237).

(3) Figure 3D: A Gaussian is used to fit the ISI distributions here but ISI distributions do not follow a normal distribution, they follow an inverse gamma distribution.

We agree with the reviewer and we are familiar with the literature that the ISI distribution is best fitted by a gamma family distribution (as a recent, but not earliest example: Li et al. 2018). However, we did not fit a gaussian (or any distribution) to the data, we just calculated the sample mean and variance. Reporting sample mean and variance (or standard deviation) is not something that is only done for Gaussian distributions. They are broadly used metrics that simply have additional intrinsic meaning for Gaussian distributions. We used the schematic illustration in Fig 3D because mean and variance are much more familiar in Gaussian distribution context, but ultimately that does not affect our analyses in Fig 3 E-F. Alternatively, the alpha and beta intrinsic parameters of a gamma distribution could have been used, but they are known by a much smaller portion of neuroscientists.

Li, M., Xie, K., Kuang, H., Liu, J., Wang, D., Fox, G. E., ... & Tsien, J. Z. (2018). Spike-timing pattern operates as gamma-distribution across cell types, regions and animal species and is essential for naturally-occurring cognitive states. Biorxiv, 145813(10.1101), 145813.

(4) Figure 3G: Something is wrong with this figure as each vertical bar is supposed to represent a drifting grating onset but yet, they are all at 5 hz despite the PSTH being purportedly shown at many different frequencies from 1 to 15 hz.

We appreciate your attention to detail, but we are not representing the onset of individual drifting gratings in this. We just meant to represent the overall start\end of the drifting grating session. We did not intend to signal the temporal frequency of the drifting gratings (or the spatial frequency, orientation, or contrast).

Author response:

The following is the authors’ response to the previous reviews

Reviewer #1 (Public review):

The mechanism, as I understand it, is different from what the authors described before in the RNN with tonic gain changes. As uncertainty increases, the network enters a regime in which the two excitatory populations start to oscillate. My intuition is that this oscillation arises from the feedback loop created by the new gain control mechanism. If my intuition is correct, I think it would be worth to explain this mechanism in the paper more explicitly.

While interesting, this intuition is not correct. The oscillations are generated by the interaction between excitatory and inhibitory nodes in the network and occur in the model even with stationary gain. All of the plots in figure 3 exploring the dynamical regime of the network at different input x gain combinations (i.e., where the oscillatory regime is characterised) are simulations run with stationary gain.

To ensure that this intuition is more clearly presented in the manuscript, we have edited the description in the text.

P. 12: “Because of the large size of the network, we could not solve for the fixed points or study their stability analytically. Instead, we opted for a numerical approach and characterised the dynamical regime (i.e. the location and existence of approximate fixed-point attractors) across all combinations of (static) gain and  visited by the network.”

Reviewer #2 (Public review):

- The demonstration of the causal role of gain modulation in perceptual switches is partial. This causality is clearly demonstrated in the simulation work with the RNN. However, it is not fully demonstrated in the pupil analysis and the fMRI analysis. One reason is that this work is correlative (which is already very informative). An analysis of the timing of the effect might have overcome this limitation. For example, in a previous study, the same group showed that fMRI activity in the LC region precedes changes in the energy landscape of fMRI dynamics, which is a step towards investigating causal links between gain modulation, changes in the energy landscape and perceptual switches.

Thank you for the suggestion, which we considered in detail. Unfortunately, the  temporal and spatial resolution of the fMRI data collected for this study precluded the same analyses we’ve run in previous work, however this is an important question for future work.

- Some effects may reflect the expectation of a perceptual switch rather than the perceptual switch itself. To mitigate this risk, the design of the fMRI task included catch trials, in which no switch occurs, to reduce the expectation of a switch. The pupil study, however, did not include such catch trials.

We agree that this is a limitation of the current study, which we previously highlighted in the methods section.

- The paper uses RNN-based modelling to provide mechanistic insight into the role of gain modulation in perceptual switches. However, the RNN solves a task that differs markedly from that performed by human participants, which may limit the explanatory value of the model. The RNN is provided with two inputs characterising the sensory evidence supporting the first and last image category in the sequence (e.g. plane and shark). In contrast, observers in the task were naïve as to the identity of the last image at the beginning of the sequence. The brain first receives sensory evidence about the image category (e.g. plane) with which the sequence begins, which is very easy to recognise, then it sees a sequence of morphed images and has to discover what the final image category will be. To discover the final image category, the brain has to search a vast space of possible second images (it is a shark?, a frog?, a bird?, etc.), rather than comparing the likelihood of just two categories. This search process and the perceptual switch in the task appear to be mechanistically different from the competition between two inputs in the RNN.

We appreciate the critical analysis of the experimental paradigm but disagree with the reviewers conclusions for two keys reasons: 1) Participants prior exposure to the images, such that they could create an expectation about what stimulus category a particular image would transition into (i.e., the image could not switch into any possible category); and 2) even if the reviewers’ concern was founded, models of K winner-take-all decision making are structured identically irrespective of whether the options are 2 or K options all that changes is the simulated reaction times which depend linearly on the K (for an example model see Hugh Wilson’s textbook Spikes, Decisions, and Actions, 1999, p.89-91). For these reasons, we maintain that the RNN is a sensible representation of the behavioural task.

- Another aspect of the motivation for the RNN model remains unclear. The authors introduce dynamic gain modulation in the RNN, but it is not clear what the added value of dynamic gain modulation is. Both static (Fig. S1) and dynamic (Fig. 2F) gain modulation lead to the predicted effect: faster switching when the gain is larger.

While we agree that the effect is observable with both static and dynamic gain, the stronger construct validity associated with the dynamic approach, including a stronger link with the observed pupil dynamics and a rich literature associated with modelling the behavioural consequences of surprise/uncertainty led us to the conclusion that the dynamical approach was a better representation of our hypothesis.

- Fig 1C: I don't see a "top grey bar" indicating significance.

Thank you for catching this, the caption has been amended. The text was from an older version of the manuscript.

- p. 10, reference to fig 3F seems incorrect: there is Fig 3F upper and Fig 3F lower, and nothing on Fig 3 and its legend mention the lesion of units

This has been amended. We meant to refer to 2F.

- In the response letter you mention a MATLAB tutorial, but I could not find it.

This has been amended. Github repository can be found at https://github.com/ShineLabUSYD/AmbiguousFigures

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

The manuscript reports that expression of the E. coli operon topAI/yjhQ/yjhP is controlled by the translation status of a small open reading frame, that authors have discovered and named toiL, located in the leader region upstream of the operon. Authors propose the following model for topAI activation: Under normal conditions, toiL is translated but topAI is not expressed because of Rho-dependent transcription termination within the topAI ORF and because its ribosome binding site and start codon are trapped in an mRNA hairpin. Ribosome stalling at various codons of the toiL ORF, prompted in this work by some ribosome-targeting antibiotics, triggers an mRNA conformational switch which allows translation of topAI and, in addition, activation of the operon's transcription because presence of translating ribosomes at the topAI ORF blocks Rho from terminating transcription. The model is appealing and several of the experimental data mainly support it. However, it remains unanswered what is the true trigger of the translation arrest at toiL and what is the physiological role of the induced expression of the topAI/yjhQ/yjhP operon.

Reviewer #2 (Public review):

Summary:

Baniulyte and Wade describe how translation of an 8-codon uORF denoted toiL upstream of the topAI-yjhQP operon is responsive to different ribosome-targeting antibiotics, consequently controlling translation of the TopAI toxin as well as Rho-dependent termination with the gene.

Strengths:

The authors used multiple different approaches such as a genetic screen to identify factors such as 23S rRNA mutations that affect topA1 expression and ribosome profiling to examine the consequences of various antibiotics on toiL-mediated regulation.

Weaknesses:

Future experiments will be needed to better understand the physiological role of the toiL-mediated regulation and elucidate the mechanism of specific antibiotic sensing.

The results are clearly described, and the revisions have helped to improve the presentation of the data.

Reviewer #3 (Public review):

In this revised manuscript, the authors provide convincing data to support an elegant model in which ribosome stalling by ToiL promotes downstream topAI translation and prevents premature Rho-dependent transcription termination. However, the physiological consequences of activating topAI-yjhQP expression upon exposure to various ribosome-targeting antibiotics remain unresolved. The authors have satisfactorily addressed all major concerns raised by the reviewers, particularly regarding the SHAPE-seq data. Overall, this study underscores the diversity of regulatory ribosome-stalling peptides in nature, highlighting ToiL's uniqueness in sensing multiple antibiotics and offering significant insights into bacterial gene regulation coordinated by transcription and translation.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

- Showing the ribosome density profiles of topAI/yjhQP and toiL in control and tetracycline treated cells is necessary to support that ribosome arrest at toiL increases translation of topAI/yjhQP.

Figure 7B shows ribosome density around the start of toiL. Ribosome density increases across topAI in the presence of tetracycline, but we have opted not to show this region because we cannot say whether the increase in ribosome occupancy (represented in Figure 7A) is due to an increase in translation efficiency, RNA level, or both.

- The subinhibitory antibiotic concentrations used in the reporter assays were based on MICs reported in the literature. This is not appropriate since MICs can greatly vary between strains, antibiotic solution stocks, and experimental conditions.

Reported MICs were used as an initial guide for selecting antibiotic concentrations to test in our reporter assays. We have added text to indicate this, and to highlight that MICs vary considerably between strains.

- toiL sequence may have evolved to maintain base-pairing with the topAI upstream region rather than, as authors suggest in Discussion, to respond to antibiotic-mediated arrest in an amino acid sequence specific manner.

We have chosen to frame this as speculation.

- Authors may consider commenting on the possibility that chloramphenicol does not induce because ToiL lacks alanine residues, whose presence at specific places of a nascent protein have been shown to promote chloramphenicol action (2016 PNAS 113:12150; 2022 NSMB 29:152).

This is a great point as none of our stalling reporters included an ORF with alanine. We now include a short paragraph in the Discussion section to raise this possibility.

- Tetracycline was added at the "subinhibitory concentration" of 8 ug/mL for the reporter assays but at 1 ug/mL for the ribosome profiling experiments. Authors should explain what was the rational for this.

We think the reviewer is mixing up the epidemiological cut-off value of 8 ug/mL with the concentration used in experiments (0.5-1 ug/mL for reporter assays and ribosome profiling). The text was confusing, so we have added a sentence to the Methods section to indicate that epidemiological cut-off values and MICs were only a guide for selecting antibiotic concentrations to test.

Reviewer #2 (Recommendations for the authors):

I wish the authors had been slightly less dismissive of the reviewers' comments. At a minimum, it would be nice if the authors could be consistent about the ribosome representation throughout the manuscript;

We apologize if our previous responses gave the impression of being dismissive. That was certainly not our intention. We greatly value the reviewers' feedback, and we appreciate the opportunity to clarify any misunderstandings. We believe the reviewer is referring to the different shape and color of the ribosome in Figures 8 and 9, and Figure 8 figure supplement 2, which we have now corrected.

Author response:

The following is the authors’ response to the current reviews.

eLife Assessment

This important and creative study finds that the uplift of the Qinghai-Tibet Plateau-via its resultant monsoon system rather than solely its high elevation-has shifted avian migratory directions from a latitudinal to a longitudinal orientation. However, the main claims are incomplete and only partially supported, as the reliance on eBird data-which lacks the resolution to capture population-specific teleconnections-combined with a limited tracking dataset covering only seven species leaves key aspects of the argument underdetermined, and the critical assumption of niche conservatism is not sufficiently foregrounded in the manuscript. More clearly communicating these limitations would significantly enhance the interpretability of the results, ensuring that the major conclusions are presented in the context of these essential caveats.

We appreciate your positive comments and constructive suggestions. We fully acknowledge your concerns about clearly communicating the limitations associated with the data used and analytical assumptions. We will try to get more satellite tracking data of birds migrating across the plateau. We will carefully consider the insights that our paper can deliver and make sure the limitations of our datasets and the critical assumption of niche conservatism are clearly presented. By explicitly clarifying these caveats, we believe the transparency and interpretability of the findings will be much improved.

Public Reviews:

Reviewer #1 (Public review):

The authors have done a good job of responding to the reviewer's comments, and the paper is now much improved.

Again, we thank the reviewer for constructive comments during review.

Reviewer #2 (Public review):

I would like to thank the authors for the revision and the input they invested in this study.

We are grateful for your thoughtful feedback and enthusiasms, which will help us improve our manuscript.

With the revised text of the study, my earlier criticism holds, and your arguments about the counterfactual approach are irrelevant to that. The recent rise of the counterfactual approach might likely mirror the fact that there are too many scientists behind their computers, and few go into the field to collect in situ data. Studies like the one presented here are a good intellectual exercise but the real impact is questionable.

We understand your question about the relevance of the counterfactual approach used in our study. Our intent in using a counterfactual scenario (reconstructing migration patterns assuming pre-uplift conditions on the QTP) was to isolate the potential influence of the plateau’s geological history on current migration routes. We agree that such an approach must be used properly. In the revision, we will explicitly clarify why this counterfactual comparison is useful – namely, it provides a theoretical baseline to test how much the QTP’s uplift (and the associated monsoon system) might have redirected migration paths. We acknowledge that the counterfactual results are theoretical and will explicitly emphasise the assumptions involved (e.g. species–environment relationships hold between pre- and post- lift environments) in the main text. Nonetheless, we defend the approach as a valuable study design: it helps generate testable hypotheses about migration (for instance, that the plateau’s monsoon-driven climate, rather than just its elevation, introduces an east–west shift en route). We will also tone down the language around this analysis to avoid overstating its real-world relevance. In summary, we will clarify that the counterfactual analysis is meant to complement, not replace, empirical observations, and we will discuss its limitations so that its role is appropriately bounded in the paper.

All your main conclusions are inferred from published studies on 7! bird species. In addition, spatial sampling in those seven species was not ideal in relation to your target questions. Thus, no matter how fancy your findings look, the basic fact remains that your input data were for 7 bird species only! Your conclusion, “our study provides a novel understanding of how QTP shapes migration patterns of birds” is simply overstretching.

Thank you for your comments. We apologise for any confusion regarding the scope of our dataset. Our main conclusions are not solely derived from seven bird species. Rather, we integrated a full list of 50 bird species that migrate across the QTP and analysed their migratory patterns with eBird data. We studied the factors influencing their choices of migratory routes with seven species that were among the few with available tracking data across the QTP. In this revision, we will clarify the role of these seven species and the rationale for their selection. Additionally, we attempt to include more satellite tracking data to improve spatial coverage, as recommended by the reviewer and editor. Based on discussions with potential collaborators, we will hopefully include a number of at least 10 more species with available tracking data.

The way you respond to my criticism on L 81-93 is something different than what you admit in the rebuttal letter. The text of the ms is silent about the drawbacks and instead highlights your perspective. I understand you; you are trying to sell the story in a nice wrapper. In the rebuttal you state: “we assume species' responses to environments are conservative and their evolution should not discount our findings.” But I do not see that clearly stated in the main text.

Thanks, as suggested we will clearly state the assumptions of niche conservatism in the Introduction.

In your rebuttal, you respond to my criticism of "No matter how good the data eBird provides is, you do not know population-specific connections between wintering and breeding sites" when you responded: ... "we can track the movement of species every week, and capture the breeding and wintering areas for specific populations" I am having a feeling that you either play with words with me or do not understand that from eBird data nobody will be ever able to estimate population-specific teleconnections between breeding and wintering areas. It is simply impossible as you do not track individuals. eBird gives you a global picture per species but not for particular populations. You cannot resolve this critical drawback of your study.

We agree that inferring population-specific migratory connections (teleconnections) from eBird data is challenging and inherently limited. eBird provides occurrence records for species, but it generally cannot distinguish which breeding population an individual bird came from or exactly where it goes for winter. However, in this study we intend to infer broad-scale movement patterns (e.g. general directions and stopover regions) rather than precise one-to-one population linkages. In the revision, we will carefully rephrase those sections to make clear that our inferences are at the species level and at large spatial scales. We will also explicitly state in the Discussion that confirming population connectivity would require targeted tracking or genetic studies, and that our eBird-based analysis can only suggest plausible routes and region-to-region linkages. We will contrast migratory routes identified by using eBird data and satellite tracking for the same species to check their similarity. We argue that, even with its limits, the eBird dataset can still yield useful insights (such as identifying major flyway corridors over the QTP).

I am sorry that you invested so much energy into this study, but I see it as a very limited contribution to understanding the role of a major barrier in shaping migration.

Thank you for recognising our efforts in the study. By integrating both satellite tracking and community-contributed data, we explored how the uplift of the QTP could shape avian migration across the area. We believe our findings provide important insights of how birds balance their responses to large-scale climate change and geological barrier, which yields the most comprehensive picture to date of how the QTP uplift shapes migratory patterns of birds. We will also acknowledge the study’s limitations to ensure that readers understand the context and constraints of our findings.

My modest suggestion for you is: go into the field. Ideally use bird radars along the plateau to document whether the birds shift the directions when facing the barrier.

We appreciate your suggestions to incorporate field tracking or radar studies to strengthen our results. All coauthors have years of field experiences, even on the QTP and Arctic. For example, the tracking data of peregrine falcons (Falco peregrinus) that we will incorporate in the revision are collected with during our own fieldwork in the Arctic for more than six years. We agree that more direct tracking (through GPS tagging or radar) would be an ideal way to validate migration pathways and population connectivity. In this revision, as stated above we will try to more species with satellite tracking data. We will also note that future studies should build on our findings by using dedicated tracking of more individual birds and radar monitoring of migration over the QTP. We will cite recent advances in these techniques and suggest that incorporating more tracking data could further test the hypotheses generated by our analyses.

Recommendations for the authors:

Reviewer #2 (Recommendations for the authors):

L55 "an important animal movement behaviour is.." Is there any unimportant animal movement? I mean this sentence is floppy, empty.

We will rewrite this sentence to remove any ambiguous phrasing.

L 152-154 This sentence is full of nonsense or you misinterpretation. First of all, the issue of inflexible initiation of migration was related to long-distance migrants only! The way you present it mixes apples and oranges (long- and short-distance migrants). It is not "owing to insufficient responses" but due to inherited patterns of when to take off, photoperiod and local conditions.

We will remove the sentence to avoid misinterpretation.

L 158 what is a migration circle? I do not know such a term.

We will amend it as “annual migration cycle”, which is a more common way to describe the yearly round-trip journey between breeding and wintering grounds of birds.

L 193 The way you present and mix capital and income breeding theory with your simulation study is quite tricky and super speculative.

We will present this idea as an inference rather than a conclusion: “This pattern could be consistent with a ‘capital breeding’ strategy — where birds rely on energy reserves acquired before breeding — rather than an ‘income’ strategy that depends on food acquired during breeding. However, we note that this interpretation would require further study.” By adding this caution, we will make it clear that we are not asserting this link as proven fact, only suggesting it as one possible explanation. We will also double-check that the rest of the discussion around this point is framed appropriately.

The following is the authors’ response to the previous reviews

eLife Assessment

This study addresses a novel and interesting question about how the rise of the Qinghai-Tibet Plateau influenced patterns of bird migration, employing a multi-faceted approach that combines species distribution data with environmental modeling. The findings are valuable for understanding avian migration within a subfield, but the strength of evidence is incomplete due to critical methodological assumptions about historical species-environment correlations, limited tracking data, and insufficient clarity in species selection criteria. Addressing these weaknesses would significantly enhance the reliability and interpretability of the results.

We would like to thank you and two anonymous reviewers for your careful, thoughtful, and constructive feedback on our manuscript. These reviews made us revisit a lot of our assumptions and we believe the paper is much improved as a result. In addition to minor points, we have made three main changes to our manuscript in response to the reviews. First, we addressed the concerns on the assumptions of historical species-environment correlations from perspectives of both theoretical and empirical evidence. Second, we discussed the benefits and limitations of using tracking data in our study and demonstrate how the findings of our study are consolidated with results of previous studies. Third, we clarified our criteria for selecting species in terms of both eBird and tracking data.

Below, we respond to each comment in turn. Once again, we thank you all for your feedback.

Public Reviews:

Reviewer #1 (Public review):

Strengths:

This is an interesting topic and a novel theme. The visualisations and presentation are to a very high standard. The Introduction is very well-written and introduces the main concepts well, with a clear logical structure and good use of the literature. The methods are detailed and well described and written in such a fashion that they are transparent and repeatable.

We are appreciative of the reviewer’s careful reading of our manuscript, encouraging comments and constructive suggestions.

Weaknesses:

I only have one major issue, which is possibly a product of the structure requirements of the paper/journal. This relates to the Results and Discussion, line 91 onwards. I understand the structure of the paper necessitates delving immediately into the results, but it is quite hard to follow due to a lack of background information. In comparison to the Methods, which are incredibly detailed, the Results in the main section reads as quite superficial. They provide broad overviews of broad findings but I found it very hard to actually get a picture of the main results in its current form. For example, how the different species factor in, etc.

Yes, it is the journal request to format in this way (Methods follows the Results and Discussion) for the article type of short reports. As suggested, in the revision we have elaborated on details of our findings, in terms of (i) shifts of distribution of avian breeding and wintering areas under the influence of the uplift of the Qinghai-Tibet Plateau (Lines 102-116), and (ii) major factors that shape current migration patterns of birds in the plateau (Lines 118-138). We have also better referenced the approaches we used in the study.

Reviewer #2 (Public review):

Summary:

The study tries to assess how the rise of the Qinghai-Tibet Plateau affected patterns of bird migration between their breeding and wintering sites. They do so by correlating the present distribution of the species with a set of environmental variables. The data on species distributions come from eBird. The main issue lies in the problematic assumption that species correlations between their current distribution and environment were about the same before the rise of the Plateau. There is no ground truthing and the study relies on Movebank data of only 7 species which are not even listed in the study. Similarly, the study does not outline the boundaries of breeding sites NE of the Plateau. Thus it is absolutely unclear potentially which breeding populations it covers.

We are very grateful for the careful review and helpful suggestions. We have revised the manuscript carefully in response to the reviewer’s comments and believe that it is much improved as a result. Below are our point-by-point replies to the comments.

Strengths:

I like the approach for how you combined various environmental datasets for the modelling part.

We appreciate the reviewer’s encouragement.

Weaknesses:

The major weakness of the study lies in the assumption that species correlations between their current distribution and environments found today are back-projected to the far past before the rise of the Q-T Plateau. This would mean that species responses to the environmental cues do not evolve which is clearly not true. Thus, your study is a very nice intellectual exercise of too many ifs.

This is a valid concern. We have addressed this from both the perspectives of the theoretical design of our study and empirical evidence.

First, we agree with the reviewer that species responses to environmental cues might vary over time. Nonetheless, the simulated environments before the uplift of the plateau serve as a counterfactual state in our study. Counterfactual is an important concept to support causation claims by comparing what happened to what would have happened in a hypothetical situation: “If event X had not occurred, event Y would not have occurred” (Lewis 1973). Recent years have seen an increasing application of the counterfactual approach to detect biodiversity change, i.e., comparing diversity between the counterfactual state and real estimates to attribute the factors causing such changes (e.g., Gonzalez et al. 2023). Whilst we do not aim to provide causal inferences for avian distributional change, using the counterfactual approach, we are able to estimate the influence of the plateau uplift by detecting the changes of avian distributions, i.e., by comparing where the birds would have distributed without the plateau to where they currently distributed. We regard the counterfactual environments as a powerful tool for eliminating, to the extent possible, vagueness, as opposed to simply description of current distributions of birds. Therefore, we assume species’ responses to environments are conservative and their evolution should not discount our findings. We have clarified this in the Introduction (Lines 81-93).

Second, we used species distribution modelling to contrast the distributions of birds before and after the uplift of the plateau under the assumption that species tend to keep their ancestral ecological traits over time (i.e., niche conservatism). This indicates a high probability for species to distribute in similar environments wherever suitable. Particularly, considering bird distributions are more likely to be influenced by food resources and vegetation distributions (Qu et al. 2010, Li et al. 2021, Martins et al. 2024), and the available food and vegetation before the uplift can provide suitable habitats for birds (Jia et al. 2020), we believe the findings can provide valuable insights into the influence of the plateau rise on avian migratory patterns. Having said that, we acknowledge other factors, e.g., carbon dioxide concentrations (Zhang et al. 2022), can influence the simulations of environments and our prediction of avian distribution. We have clarified the assumptions and evidence we have for the modelling in Methods (Lines 362-370).

The second major drawback lies in the way you estimate the migratory routes of particular birds. No matter how good the data eBird provides is, you do not know population-specific connections between wintering and breeding sites. Some might overwinter in India, some populations in Africa and you will never know the teleconnections between breeding and wintering sites of particular species. The few available tracking studies (seven!) are too coarse and with limited aspects of migratory connectivity to give answer on the target questions of your study.

We agree with the reviewer that establishing interconnections for birds is important for estimating the migration patterns of birds. We employed a dynamic model to assess their weekly distributions. Thus, we can track the movement of species every week, and capture the breeding and wintering areas for specific populations. That being said, we acknowledge that our approach can be subjected to the patchy sampling of eBird data. In contrast, tracking data can provide detailed information of the movement patterns of species but are limited to small numbers of species due to the considerable costs and time needed. We aimed to adopt the tracking data to examine the influence of focal factors on avian migration patterns, but only seven species, to the best of our ability, were acquired. Moreover, similar results were found in studies that used tracking data to estimate the distribution of breeding and wintering areas of birds in the plateau (e.g., Prosser et al. 2011, Zhang et al. 2011, Zhang et al. 2014, Liu et al. 2018, Kumar et al. 2020, Wang et al. 2020, Pu and Guo 2023, Yu et al. 2024, Zhao et al. 2024). We believe the conclusions based on seven species are rigour, but their implications could be restricted by the number of tracking species we obtained. We have better demonstrated how our findings on breeding and wintering areas of birds are reinforced by other studies reporting the locations of those areas. We have also added a separate caveat section to discuss the limitations stated above (Lines 202-215).

Your set of species is unclear, selection criteria for the 50 species are unknown and variability in their migratory strategies is likely to affect the direction of the effects.

In this revision, we have clarified the selection criteria for the 50 species and outlined the boundaries of the breeding areas of all birds (Lines 243-249). Briefly, we first obtained a full list of birds in the plateau from Prins and Namgail (2017). We then extracted species identified as full migrants in Birdlife International (https://datazone.birdlife.org/species/spcdistPOS) from the full list. Migratory birds may follow a capital or income migratory strategy depending on how much birds ingest endogenous reserved energy gained prior to reproduction. We have added discussions on how these migratory strategies might influence the effects of environment on migratory direction (Lines 183-200).

In addition, the position of the breeding sites relative to the Q-T plate will affect the azimuths and resulting migratory flyways. So in fact, we have no idea what your estimates mean in Figure 2.

We calculated the azimuths not only by the angles between breeding sites and wintering sites but also based on the angles between the stopovers of birds. Therefore, the azimuths are influenced by the relative positions of breeding, wintering and stopover sites. This would minimize the possible errors by just using breeding areas such as the biases caused by relative locations of breeding areas to the QTP as the reviewer pointed. We have better explained this both in the Introduction, Methods and legend of Figure 2.

There is no way one can assess the performance of your statistical exercises, e.g. performances of the models.

As suggested, we have reported Area Under the Curve (AUC) of the Receiver Operator Characteristic (ROC)assess the performances of the models (Table S1). AUC is a threshold-independent measurement for discrimination ability between presence and random points (Phillips et al. 2006). When the AUC value is higher than 0.75, the model was considered to be good (Elith et al. 2006). (Lines 379-383).

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

This is an interesting topic and a novel theme. The visualisations and presentation are to a very high standard. The Introduction is very well-written and introduces the main concepts well, with a clear logical structure and good use of the literature. The Methods are detailed and well described and written in such a fashion that they are transparent and repeatable.

I only have one major issue, which is possibly a product of the structure requirements of the paper/journal. With the Results and Discussion, line 91 onwards. I understand the structure of the paper necessitates delving immediately into the results, but it is quite hard to follow due to a lack of background information. In comparison to the Methods, which are incredibly detailed, the Results in the main section read quite superficial. They provide broad overviews of broad findings but I found it very hard to actually get a picture of the main results in its current form. For example, how the different species factor in, etc.

Please see our responses above.

Reviewer #2 (Recommendations for the authors):

Methodological issues:

Line 219 Why have you selected only 64 species and what were the selection criteria?

We have clarified the selection criteria (Lines 243-248). Briefly, we first obtained a full list of birds in the plateau from Prins and Namgail (2017). We then extracted species identified as full migrants in Birdlife International (https://datazone.birdlife.org/species/spcdistPOS) from the full list.

Minor:

Line 219 eBird has very uneven distribution, especially in vast areas of Russia. How can your exercise on Lines 232-238 overcome this issue?

Yes, eBird data can be biased due to patchy sampling and variation of observers’ skills in identifying species. To address this issue, we have developed an adaptive spatial-temporal modelling (stemflow; Chen et al. 2024) to correct the imbalance distribution of data and modelled the observer experience to address the bias in recognising species. The stemflow was developed based on a machine learning modelling framework (AdaSTEM) which leverages the spatio-temporal adjacency information of sample points to model occurrence or abundance of species at different scales. It has been frequently used in modelling eBird data (Fink et al. 2013, Johnston et al. 2015, Fink et al. 2020) and has been proven to be efficient and advanced in multi-scale spatiotemporal data modelling. We have better explained this (Lines 251-270; Lines 307-321).

Line 54 This sentence sounds very empty and in fact does not tell us much.

We have adjusted this sentenced to “Animal movement underpins species’ spatial distributions and ecosystem processes”.

Line 55 Again a sentence that implies a causality of the annual cycle to make the species migrate. It does not make sense.

We have revised this sentence as “An important animal movement behaviour is migrating between breeding and wintering grounds”.

Line 58 How is our fascination with migratory journeys related to the present article? I think this line is empty.

We have changed this sentence to “Those migratory journeys have intrigued a body of different approaches and indicators to describe and model migration, including migratory direction, speed, timing, distance, and staging periods”.

Figure 1 - ABC insets are OK, but a combination of lati- and longitudinal patterns is possible, e.g. in species with conservative strategies or for whatever other reason.

Thank you for the suggestion. We kept the ABC insets rather than combining them together as we believe this can deliver a clear structure of influence of QTP uplift under different scenarios.

The legend to Figure 2 is not self-explanatory. Please make it clear what the response variable is and its units. The first line of the legend should read something like The influence of environmental factors on the direction of avian migration.

Thank you. We have amended the legends of Figure 2 as suggested:

“Figure 2. The influence of environmental factors on the direction of avian migration.  Migratory directions are calculated based on the azimuths between each adjacent stopover, breeding and wintering areas for each species. We employ multivariate linear regression models under the Bayesian framework to measure the correlation between environmental factors and avian migratory directions. Wind represents the wind cost calculated by wind connectivity. Vegetation is measured by the proportion of average vegetation cover in each pixel (~1.9° in latitude by 2.5° in longitude). Temperature is the average annual temperature. Precipitation is the average yearly precipitation. All environmental layers are obtained using the Community Earth System Model. West QTP, central QTP, and East QTP denote areas in the areas west (longitude < 73°E), central (73°E ≤ longitude < 105°E), and east of (longitude ≥ 105°E) the Qinghai-Tibet Plateau, respectively.”

References

Chen, Y., Z. Gu, and X. Zhan. 2024. stemflow: A Python Package for Adaptive Spatio-Temporal Exploratory Model. Journal of Open Source Software 9:6158.

Elith, J., C. H. Graham, R. P. Anderson, M. Dudík, S. Ferrier, A. Guisan, R. J. Hijmans, F. Huettmann, J. R. Leathwick, A. Lehmann, J. Li, L. G. Lohmann, B. A. Loiselle, G. Manion, C. Moritz, M. Nakamura, Y. Nakazawa, J. McC. M. Overton, A. Townsend Peterson, S. J. Phillips, K. Richardson, R. Scachetti-Pereira, R. E. Schapire, J. Soberón, S. Williams, M. S. Wisz, and N. E. Zimmermann. 2006. Novel methods improve prediction of species' distributions from occurrence data. Ecography 29:129-151.

Fink, D., T. Auer, A. Johnston, V. Ruiz-Gutierrez, W. M. Hochachka, and S. Kelling. 2020. Modeling avian full annual cycle distribution and population trends with citizen science data. Ecological Applications 30:e02056.

Fink, D., T. Damoulas, and J. Dave. 2013. Adaptive Spatio-Temporal Exploratory Models: Hemisphere-wide species distributions from massively crowdsourced eBird data. Pages 1284-1290 in Proceedings of the AAAI Conference on Artificial Intelligence.

Gonzalez, A., J. M. Chase, and M. I. O'Connor. 2023. A framework for the detection and attribution of biodiversity change. Philosophical Transactions of the Royal Society B: Biological Sciences 378.

Jia, Y., H. Wu, S. Zhu, Q. Li, C. Zhang, Y. Yu, and A. Sun. 2020. Cenozoic aridification in Northwest China evidenced by paleovegetation evolution. Palaeogeography, Palaeoclimatology, Palaeoecology 557:109907.

Johnston, A., D. Fink, M. D. Reynolds, W. M. Hochachka, B. L. Sullivan, N. E. Bruns, E. Hallstein, M. S. Merrifield, S. Matsumoto, and S. Kelling. 2015. Abundance models improve spatial and temporal prioritization of conservation resources. Ecological Applications 25:1749-1756.

Kumar, N., U. Gupta, Y. V. Jhala, Q. Qureshi, A. G. Gosler, and F. Sergio. 2020. GPS-telemetry unveils the regular high-elevation crossing of the Himalayas by a migratory raptor: implications for definition of a “Central Asian Flyway”. Scientific Reports 10:15988.

Lewis, D. 1973. Counterfactuals. Oxford: Blackwell.

Li, S.-F., P. J. Valdes, A. Farnsworth, T. Davies-Barnard, T. Su, D. J. Lunt, R. A. Spicer, J. Liu, W.-Y.-D. Deng, J. Huang, H. Tang, A. Ridgwell, L.-L. Chen, and Z.-K. Zhou. 2021. Orographic evolution of northern Tibet shaped vegetation and plant diversity in eastern Asia. Science Advances 7:eabc7741.

Liu, D., G. Zhang, H. Jiang, and J. Lu. 2018. Detours in long-distance migration across the Qinghai-Tibetan Plateau: individual consistency and habitat associations. PeerJ 6:e4304.

Martins, L. P., D. B. Stouffer, P. G. Blendinger, K. Böhning-Gaese, J. M. Costa, D. M. Dehling, C. I. Donatti, C. Emer, M. Galetti, R. Heleno, Í. Menezes, J. C. Morante-Filho, M. C. Muñoz, E. L. Neuschulz, M. A. Pizo, M. Quitián, R. A. Ruggera, F. Saavedra, V. Santillán, M. Schleuning, L. P. da Silva, F. Ribeiro da Silva, J. A. Tobias, A. Traveset, M. G. R. Vollstädt, and J. M. Tylianakis. 2024. Birds optimize fruit size consumed near their geographic range limits. Science 385:331-336.

Phillips, S. J., R. P. Anderson, and R. E. Schapire. 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling 190:231-259.

Prins, H. H. T., and T. Namgail. 2017. Bird migration across the Himalayas : wetland functioning amidst mountains and glaciers. Cambridge University Press, Cambridge.

Prosser, D. J., P. Cui, J. Y. Takekawa, M. Tang, Y. Hou, B. M. Collins, B. Yan, N. J. Hill, T. Li, Y. Li, F. Lei, S. Guo, Z. Xing, Y. He, Y. Zhou, D. C. Douglas, W. M. Perry, and S. H. Newman. 2011. Wild Bird Migration across the Qinghai-Tibetan Plateau: A Transmission Route for Highly Pathogenic H5N1. Plos One 6:e17622.

Pu, Z., and Y. Guo. 2023. Autumn migration of black-necked crane (Grus nigricollis) on the Qinghai-Tibetan and Yunnan-Guizhou plateaus. Ecology and Evolution 13:e10492.

Qu, Y., F. Lei, R. Zhang, and X. Lu. 2010. Comparative phylogeography of five avian species: implications for Pleistocene evolutionary history in the Qinghai-Tibetan plateau. Molecular Ecology 19:338-351.

Wang, Y., C. Mi, and Y. Guo. 2020. Satellite tracking reveals a new migration route of black-necked cranes (Grus nigricollis) in Qinghai-Tibet Plateau. PeerJ 8:e9715.

Yu, X., G. Song, H. Wang, Q. Wei, C. Jia, and F. Lei. 2024. Migratory flyways and connectivity of Brown Headed Gulls (Chroicocephalus brunnicephalus) revealed by GPS tracking. Global Ecology and Conservation 56:e03340.

Zhang, G.-G., D.-P. Liu, Y.-Q. Hou, H.-X. Jiang, M. Dai, F.-W. Qian, J. Lu, T. Ma, L.-X. Chen, and Z. Xing. 2014. Migration routes and stopover sites of Pallas’s Gulls Larus ichthyaetus breeding at Qinghai Lake, China, determined by satellite tracking. Forktail 30:104-108.

Zhang, G.-G., D.-P. Liu, Y.-Q. Hou, H.-X. Jiang, M. Dai, F.-W. Qian, J. Lu, Z. Xing, and F.-S. Li. 2011. Migration Routes and Stop-Over Sites Determined with Satellite Tracking of Bar-Headed Geese (Anser indicus) Breeding at Qinghai Lake, China. Waterbirds 34:112-116, 115.

Zhang, R., D. Jiang, C. Zhang, and Z. Zhang. 2022. Distinct effects of Tibetan Plateau growth and global cooling on the eastern and central Asian climates during the Cenozoic. Global and Planetary Change 218:103969.

Zhao, T., W. Heim, R. Nussbaumer, M. van Toor, G. Zhang, A. Andersson, J. Bäckman, Z. Liu, G. Song, M. Hellström, J. Roved, Y. Liu, S. Bensch, B. Wertheim, F. Lei, and B. Helm. 2024. Seasonal migration patterns of Siberian Rubythroat (Calliope calliope) facing the Qinghai–Tibet Plateau. Movement Ecology 12:54.

Author response:

We thank the reviewers for their thoughtful and generous assessment of our work. Overall, the reviewers found our work to be novel and relevant. In particular: reviewer #1 found that our manuscript “It is timely and highly valuable for the telomere field” reviewer #2 stated, “Overall, I find this manuscript worthy of publication, as the optimized END-seq methods described here will likely be widely utilized in the telomere field.” Reviewer #3 stated that “The study is original, the experiments were well-controlled and excellently executed.”

We are extremely grateful for these comments and want to thank all the reviewers and the editors for their time and effort in reviewing our work.

The reviewers had a number of suggestions to improve our work. We have addressed all the points as highlighted in the point-by-point responses below.

Reviewer 1:

One minor question would be whether the authors could expand more on the application of END-Seq to examine the processive steps of the ALT mechanism? Can they speculate if the ssDNA detected in ALT cells might be an intermediate generated during BIR (i.e., is the ssDNA displaced strand during BIR) or a lesion? Furthermore, have the authors assessed whether ssDNA lesions are due to the loss of ATRX or DAXX, either of which can be mutated in the ALT setting?

We appreciate the reviewer’s insightful questions regarding the application of our assays to investigate the nature of the ssDNA detected in ALT telomeres. Our primary aim in this study was to establish the utility of END-seq and S1-END-seq in telomere biology and to demonstrate their applicability across both ALT-positive and -negative contexts. We agree that exploring the mechanistic origins of ssDNA would be highly informative, and we anticipate that END-seq–based approaches will be well suited for such future studies. However, it remains unclear whether the resolution of S1-END-seq is sufficient to capture transient intermediates such as those generated during BIR. We have now included a brief speculative statement in the revised discussion addressing the potential nature of ssDNA at telomeres in ALT cells.

Reviewer #2:

How can we be sure that all telomeres are equally represented? The authors seem to assume that END-seq captures all chromosome ends equally, but can we be certain of this? While I do not see an obvious way to resolve this experimentally, I recommend discussing this potential bias more extensively in the manuscript.

We thank the reviewer for raising this important point. END-seq and S1-END-seq are unbiased methods designed to capture either double-stranded or single-stranded DNA that can be converted into blunt-ended double-stranded DNA and ligated to a capture oligo. As such, if a subset of telomeres cannot be processed using this approach, it is possible that these telomeres may be underrepresented or lost. However, to our knowledge, there are no proposed telomeric structures that would prevent capture using this method. For example, even if a subset of telomeres possesses a 5′ overhang, it would still be captured by END-seq. Indeed, we observed the consistent presence of the 5′-ATC motif across multiple cell lines and species (human, mouse, and dog). More importantly, we detected predictable and significant changes in sequence composition when telomere ends were experimentally altered, either in vivo (via POT1 depletion) or in vitro (via T7 exonuclease treatment). Together, these findings support the robustness of the method in capturing a representative and dynamic view of telomeres across different systems.

That said, we have now included a brief statement in the revised discussion acknowledging that we cannot fully exclude the possibility that a subset of telomeres may be missed due to unusual or uncharacterized structures

I believe Figures 1 and 2 should be merged.

We appreciate the reviewer’s suggestion to merge Figures 1 and 2. However, we feel that keeping them as separate figures better preserves the logical flow of the manuscript and allows the validation of END-seq and its application to be presented with appropriate clarity and focus. We hope the reviewer agrees that this layout enhances the clarity and interpretability of the data.

Scale bars should be added to all microscopy figures.

We thank the reviewer for pointing this out. We have now added scale bars to all the microscopy panels in the figures and included the scale details in the figure legends.

Reviewer #3:

Overall, the discussion section is lacking depth and should be expanded and a few additional experiments should be performed to clarify the results.

We thank the reviewer for the suggestions. Based on this reviewer’s comments and comments for the other reviewers, we incorporated several points into the discussion. As a result, we hope that we provide additional depth to our conclusions.

(1) The finding that the abundance of variant telomeric repeats (VTRs) within the final 30 nucleotides of the telomeric 5' ends is similar in both telomerase-expressing and ALT cells is intriguing, but the authors do not address this result. Could the authors provide more insight into this observation and suggest potential explanations? As the frequency of VTRs does not seem to be upregulated in POT1-depleted cells, what then drives the appearance of VTRs on the C-strand at the very end of telomeres? Is CST-Pola complex responsible?

The reviewer raises a very interesting and relevant point. We are hesitant at this point to speculate on why we do not see a difference in variant repeats in ALT versus non-ALT cells, since additional data would be needed. One possibility is that variant repeats in ALT cells accumulate stochastically within telomeres but are selected against when they are present at the terminal portion of chromosome ends. However, to prove this hypothesis, we would need error-free long-read technology combined with END-seq. We feel that developing this approach would be beyond the scope of this manuscript.

(2) The authors also note that, in ALT cells, the frequency of VTRs in the first 30 nucleotides of the S1-END-SEQ reads is higher compared to END-SEQ, but this finding is not discussed either. Do the authors think that the presence of ssDNA regions is associated with the VTRs? Along this line, what is the frequency of VTRs in the END-SEQ analysis of TRF1-FokI-expressing ALT cells? Is it also increased? Has TRF1-FokI been applied to telomerase-expressing cells to compare VTR frequencies at internal sites between ALT and telomerase-expressing cells?

Similarly to what is discussed above, short reads have the advantage of being very accurate but do not provide sufficient length to establish the relative frequency of VTRs across the whole telomere sequence. The TRF1-FokI experiment is a good suggestion, but it would still be biased toward non-variant repeats due to the TRF1-binding properties. We plan to address these questions in a future study involving long-read sequencing and END-seq capture of telomeres.

Finally, in these experiments (S1-END-SEQ or END-SEQ in TRF1-Fok1), is the frequency of VTRs the same on both the C- and the G-rich strands? It is possible that the sequences are not fully complementary in regions where G4 structures form.

We thank the reviewer for this observation. While we do observe a higher frequency of variant telomeric repeats (VTRs) in the first 30 nucleotides of S1-END-seq reads compared to END-seq in ALT cells, we are currently unable to determine whether this difference is significant, as an appropriate control or matched normalization strategy for this comparison is lacking. Therefore, we refrain from overinterpreting the biological relevance of this observation.

The reviewer is absolutely correct. Our calculation did not exclude the possibility of extrachromosomal DNA as a source of telomeric ssDNA. We have now addressed this point in our discussion.

The reviewer is correct in pointing out that we still do not know what causes ssDNA at telomeres in ALT cells. Replication stress seems the most logical explanation based on the work of many labs in the field. However, our data did not reveal any significant difference in the levels of ssDNA at telomeres in non-ALT cells based on telomere length. We used the HeLa1.2.11 cell line (now clarified in the Materials section), which is the parental line of HeLa1.3 and has similarly long telomeres (~20 kb vs. ~23 kb). Despite their long telomeres and potential for replication-associated challenges such as G-quadruplex formation, HeLa1.2.11 cells did not exhibit the elevated levels of telomeric ssDNA that we observed in ALT cells (Figure 4B). Additional experiments are needed to map the occurrence of ssDNA at telomeres in relation to progression toward ALT.

(3) Based on the ratio of C-rich to G-rich reads in the S1-END-SEQ experiment, the authors estimate that ALT cells contain at least 3-5 ssDNA regions per chromosome end. While the calculation is understandable, this number could be discussed further to consider the possibility that the observed ratios (of roughly 0.5) might result from the presence of extrachromosomal DNA species, such as C-circles. The observed increase in the ratio of C-rich to G-rich reads in BLM-depleted cells supports this hypothesis, as BLM depletion suppresses C-circle formation in U2OS cells. To test this, the authors should examine the impact of POLD3 depletion on the C-rich/G-rich read ratio. Alternatively, they could separate high-molecular-weight (HMW) DNA from low-molecular-weight DNA in ALT cells and repeat the S1-END-SEQ in the HMW fraction.

The reviewer is absolutely correct. Our calculation did not exclude the possibility of extrachromosomal DNA as a source of telomeric ssDNA. We have now addressed this point in our discussion.

(4) What is the authors' perspective on the presence of ssDNA at ALT telomeres? Do they attribute this to replication stress? It would be helpful for the authors to repeat the S1-END-SEQ in telomerase-expressing cells with very long telomeres, such as HeLa1.3 cells, to determine if ssDNA is a specific feature of ALT cells or a result of replication stress. The increased abundance of G4 structures at telomeres in HeLa1.3 cells (as shown in J. Wong's lab) may indicate that replication stress is a factor. Similar to Wong's work, it would be valuable to compare the C-rich/G-rich read ratios in HeLa1.3 cells to those in ALT cells with similar telomeric DNA content.

The reviewer is correct in pointing out that we still do not know what causes ssDNA at telomeres in ALT cells. Replication stress seems the most logical explanation based on the work of many labs in the field. However, our data did not reveal any significant difference in the levels of ssDNA at telomeres in non-ALT cells based on telomere length. We used the HeLa1.2.11 cell line (now clarified in the Materials section), which is the parental line of HeLa1.3 and has similarly long telomeres (~20 kb vs. ~23 kb). Despite their long telomeres and potential for replication-associated challenges such as G-quadruplex formation, HeLa1.2.11 cells did not exhibit the elevated levels of telomeric ssDNA that we observed in ALT cells (Figure 4B). Additional experiments are needed to map the occurrence of ssDNA at telomeres in relation to progression toward ALT.

Finally, Reviewer #3 raises a list of minor points:

(1) The Y-axes of Figure 4 have been relabeled to account for the G-strand reads.

(2) Statistical analyses have been added to the figures where applicable.

(3) The manuscript has been carefully proofread to improve clarity and consistency throughout the text and figure legends.

(4) We have revised the text to address issues related to the lack of cross-referencing between the supplementary figures and their corresponding legends.

This rubric is supposed to highlight the basic needs of all students and how well a certain e-learning tool fits into these needs. I think this rubric is a good start for analyzing digital tools that you may bring into the classroom, but the true test is seeing how much your students have learned from these tools after they are used. Just because a tool may fit perfectly in this rubric does not mean it will educate students perfectly in the classroom.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

This study examined the interaction between two key cortical regions in the mouse brain involved in goal-directed movements, the rostral forelimb area (RFA) - considered a premotor region involved in movement planning, and the caudal forelimb area (CFA) - considered a primary motor region that more directly influences movement execution. The authors ask whether there exists a hierarchical interaction between these regions, as previously hypothesized, and focus on a specific definition of hierarchy - examining whether the neural activity in the premotor region exerts a larger functional influence on the activity in the primary motor area than vice versa. They examine this question using advanced experimental and analytical methods, including localized optogenetic manipulation of neural activity in either region while measuring both the neural activity in the other region and EMG signals from several muscles involved in the reaching movement, as well as simultaneous electrophysiology recordings from both regions in a separate cohort of animals.

The findings presented show that localized optogenetic manipulation of neural activity in either RFA or CFA resulted in similarly short-latency changes in the muscle output and in firing rate changes in the other region. However, perturbation of RFA led to a larger absolute change in the neural activity of CFA neurons. The authors interpret these findings as evidence for reciprocal, but asymmetrical, influence between the regions, suggesting some degree of hierarchy in which RFA has a greater effect on the neural activity in CFA. They go on to examine whether this asymmetry can also be observed in simultaneously recorded neural activity patterns from both regions. They use multiple advanced analysis methods that either identify latent components at the population level or measure the predictability of firing rates of single neurons in one region using firing rates of single neurons in the other region. Interestingly, the main finding across these analyses seems to be that both regions share highly similar components that capture a high degree of variability of the neural activity patterns in each region. Single units' activity from either region could be predicted to a similar degree from the activity of single units in the other region, without a clear division into a leading area and a lagging area, as one might expect to find in a simple hierarchical interaction. However, the authors find some evidence showing a slight bias towards leading activity in RFA. Using a two-region neural network model that is fit to the summed neural activity recorded in the different experiments and to the summed muscle output, the authors show that a network with constrained (balanced) weights between the regions can still output the observed measured activities and the observed asymmetrical effects of the optogenetic manipulations, by having different within-region local weights. These results put into question whether previous and current findings that demonstrate asymmetry in the output of regions can be interpreted as evidence for asymmetrical (and thus hierarchical) inputs between regions, emphasizing the challenges in studying interactions between any brain regions.

Strengths:

The experiments and analyses performed in this study are comprehensive and provide a detailed examination and comparison of neural activity recorded simultaneously using dense electrophysiology probes from two main motor regions that have been the focus of studies examining goal-directed movements. The findings showing reciprocal effects from each region to the other, similar short-latency modulation of muscle output by both regions, and similarity of neural activity patterns without a clear lead/lag interaction, are convincing and add to the growing body of evidence that highlight the complexity of the interactions between multiple regions in the motor system and go against a simple feedforward-like network and dynamics. The neural network model complements these findings and adds an important demonstration that the observed asymmetry can, in theory, also arise from differences in local recurrent connections and not necessarily from different input projections from one region to the other. This sheds an important light on the multiple factors that should be considered when studying the interaction between any two brain regions, with a specific emphasis on the role of local recurrent connections, that should be of interest to the general neuroscience community.

Weaknesses:

While the similarity of the activity patterns across regions and lack of a clear leading/lagging interaction are interesting observations that are mostly supported by the findings presented (however, see comment below for lack of clarity in CCA/PLS analyses), the main question posed by the authors - whether there exists an endogenous hierarchical interaction between RFA and CFA - seems to be left largely open. 

The authors note that there is currently no clear evidence of asymmetrical reciprocal influence between naturally occurring neural activity patterns of the two regions, as previous attempts have used non-natural electrical stimulation, lesions, or pharmacological inactivation. The use of acute optogenetic perturbations does not seem to be vastly different in that aspect, as it is a non-natural stimulation of inhibitory interneurons that abruptly perturbs the ongoing dynamics.

We do believe that our optogenetic inactivation identifies a causal interaction between the endogenous activity patterns in the excitatory projection neurons, which we have largely silenced, and the downstream endogenous activity that is perturbed. The effect in the downstream region results directly from the silencing of activity in the excitatory projection neurons that mediate each region’s interaction with other regions. Here we have performed a causal intervention common in biology: a loss-of-function experiment. Such experiments generally reveal that a causal interaction of some sort is present, but often do not clarify much about the nature of the interaction, as is true in our case. By showing that a silencing of endogenous activity in one motor cortical region causes a significant change to the endogenous activity in another, we establish a causal relationship between these activity patterns. This is analogous to knocking out the gene for a transcription factor and observing causal effects on the expression of other genes that depend on it. 

Moreover, our experiments are, to our knowledge, the first that localize a causal relationship to endogenous activity in motor cortex at a particular point during a motor behavior. Lesion and pharmacological or chemogenetic inactivation have long-lasting effects, and so their consequences on firing in other regions cannot be attributed to a short-latency influence of activity at a particular point during movement. Moreover, the involvement of motor cortex in motor learning and movement preparation/initiation complicates the interpretation of these consequences in relation to movement execution, as disturbance to processes on which execution depends can impede execution itself. Stimulation experiments generate spiking in excitatory projection neurons that is not endogenous.

That said, we would agree that the form of the causal interaction between RFA and CFA remains unaddressed by our results. These results do not expose how the silenced activity patterns affect activity in the downstream region, just as knocking out a transcription factor gene does not expose how the transcription factor influences the expression of other genes. To show evidence for a specific type of interaction dynamics between RFA and CFA, a different sort of experiment would be necessary. See Jazayeri and Afraz, Neuron, 2017 for more on this issue.

Furthermore, the main finding that supports a hierarchical interaction is a difference in the absolute change of firing rates as a result of the optogenetic perturbation, a finding that is based on a small number of animals (N = 3 in each experimental group), and one which may be difficult to interpret. 

Though N = 3, we do show statistical significance. Moreover, using three replicates is not uncommon in biological experiments that require a large technical investment.

As the authors nicely demonstrate in their neural network model, the two regions may differ in the strength of local within-region inhibitory connections. Could this theoretically also lead to a difference in the effect of the artificial light stimulation of the inhibitory interneurons on the local population of excitatory projection neurons, driving an asymmetrical effect on the downstream region? 

We (Miri et al., Neuron, 2017) and others (Guo et al., Neuron, 2014) have shown that the effect of this inactivation on excitatory neurons in CFA is a near-complete silencing (90-95% within 20 ms). There thus is not much room for the effects on projection neurons in RFA to be much larger. We have measured these local effects in RFA as part of other work (Kristl et al., biorxiv, 2025), verifying that the effects on RFA projection neuron firing are not larger.

Moreover, the manipulation was performed upon the beginning of the reaching movement, while the premotor region is often hypothesized to exert its main control during movement preparation, and thus possibly show greater modulation during that movement epoch. It is not clear if the observed difference in absolute change is dependent on the chosen time of optogenetic stimulation and if this effect is a general effect that will hold if the stimulation is delivered during different movement epochs, such as during movement preparation.

We agree that the dependence of RFA-CFA interactions on movement phase would be interesting to address in subsequent experiments. While a strong interpretation of lesion results might lead to a hypothesis that premotor influence on primary motor cortex is local to, or stronger during, movement preparation as opposed to execution, at present there is to our knowledge no empirical support from interventional experiments for this hypothesis. Moreover, existing results from analysis of activity in these two regions have produced conflicting results on the strength of interaction between these regions during preparation. Compare for example BachschmidRomano et al., eLife, 2023 to Kaufman et al., Nature Neuroscience, 2014.

That said, this lesion interpretation would predict the same asymmetry we have observed from perturbations at the beginning of a reach - a larger effect of RFA on CFA than vice versa.

Another finding that is not clearly interpretable is in the analysis of the population activity using CCA and PLS. The authors show that shifting the activity of one region compared to the other, in an attempt to find the optimal leading/lagging interaction, does not affect the results of these analyses. Assuming the activities of both regions are better aligned at some unknown groundtruth lead/lag time, I would expect to see a peak somewhere in the range examined, as is nicely shown when running the same analyses on a single region's activity. If the activities are indeed aligned at zero, without a clear leading/lagging interaction, but the results remain similar when shifting the activities of one region compared to the other, the interpretation of these analyses is not clear.

Our results in this case were definitely surprising. Many share the intuition that there should be a lag at which the correlations in activity between regions may be strongest. The similarity in alignment across lags we observed might be expected if communication between regions occurs over a range of latencies as a result of dependence on a broad diversity of synaptic paths that connect neurons. In the Discussion, we offer an explanation of how to reconcile these findings with the seemingly different picture presented by DLAG.

Reviewer #2 (Public review):

Summary:

While technical advances have enabled large-scale, multi-site neural recordings, characterizing inter-regional communication and its behavioral relevance remains challenging due to intrinsic properties of the brain such as shared inputs, network complexity, and external noise. This work by Saiki-Ishkawa et al. examines the functional hierarchy between premotor (PM) and primary motor (M1) cortices in mice during a directional reaching task. The authors find some evidence consistent with an asymmetric reciprocal influence between the regions, but overall, activity patterns were highly similar and equally predictive of one another. These results suggest that motor cortical hierarchy, though present, is not fully reflected in firing patterns alone.

Strengths:

Inferring functional hierarchies between brain regions, given the complexity of reciprocal and local connectivity, dynamic interactions, and the influence of both shared and independent external inputs, is a challenging task. It requires careful analysis of simultaneous recording data, combined with cross-validation across multiple metrics, to accurately assess the functional relationships between regions. The authors have generated a valuable dataset simultaneously recording from both regions at scale from mice performing a cortex-dependent directional reaching task.

Using electrophysiological and silencing data, the authors found evidence supporting the traditionally assumed asymmetric influence from PM to M1. While earlier studies inferred a functional hierarchy based on partial temporal relationships in firing patterns, the authors applied a series of complementary analyses to rigorously test this hierarchy at both individual neuron and population levels, with robust statistical validation of significance.

In addition, recording combined with brief optogenetic silencing of the other region allowed authors to infer the asymmetric functional influence in a more causal manner. This experiment is well designed to focus on the effect of inactivation manifesting through oligosynaptic connections to support the existence of a premotor to primary motor functional hierarchy.

Subsequent analyses revealed a more complex picture. CCA, PLS, and three measures of predictivity (Granger causality, transfer entropy, and convergent cross-mapping) emphasized similarities in firing patterns and cross-region predictability. However, DLAG suggested an imbalance, with RFA capturing CFA variance at a negative time lag, indicating that RFA 'leads' CFA. Taken together these results provide useful insights for current studies of functional hierarchy about potential limitations in inferring hierarchy solely based on firing rates.

While I would detail some questions and issues on specifics of data analyses and modeling below, I appreciate the authors' effort in training RNNs that match some behavioral and recorded neural activity patterns including the inactivation result. The authors point out two components that can determine the across-region influence - 1) the amount of inputs received and 2) the dependence on across-region input, i.e., the relative importance of local dynamics, providing useful insights in inferring functional relationships across regions.

Weaknesses:

(1) Trial-averaging was applied in CCA and PLS analyses. While trial-averaging can be appropriate in certain cases, it leads to the loss of trial-to-trial variance, potentially inflating the perceived similarities between the activity in the two regions (Figure 4). Do authors observe comparable degrees of similarity, e.g., variance explained by canonical variables? Also, the authors report conflicting findings regarding the temporal relationship between RFA and CFA when using CCA/PLS versus DLAG. Could this discrepancy be due to the use of trial-averaging in former analyses but not in the latter?

We certainly agree that the similarity in firing patterns is higher in trial averages than on single trials, given the variation in single-neuron firing patterns across trials. Here, we were trying to examine the similarity of activity variance that is clearly movement dependent, as trial averages are, and to use an approach aligned with those applied in the existing literature. We would also agree that there is more that can be learned about interactions from trial-by-trial analysis. It is possible that the activity components identified by DLAG as being asymmetric somehow are not reflected strongly in trial averages. In our Discussion we offer another potential explanation that is based on other differences in what is calculated by DLAG and CCA/PLS.

We also note here that all of the firing pattern predictivity analysis we report (Figure 6) was done on single trial data, and in all cases the predictivity was symmetric. Thus, our results in aggregate are not consistent with symmetry purely being an artifact of trial averaging.

(2) A key strength of the current study is the precise tracking of forelimb muscle activity during a complex motor task involving reaching for four different targets. This rich behavioral data is rarely collected in mice and offers a valuable opportunity to investigate the behavioral relevance of the PM-M1 functional interaction, yet little has been done to explore this aspect in depth. For example, single-trial time courses of inter-regional latent variables acquired from DLAG analysis can be correlated with single-trial muscle activity and/or reach trajectories to examine the behavioral relevance of inter-regional dynamics. Namely, can trial-by-trial change in inter-regional dynamics explain behavioral variability across trials and/or targets? Does the inter-areal interaction change in error trials? Furthermore, the authors could quantify the relative contribution of across-area versus within-area dynamics to behavioral variability. It would also be interesting to assess the degree to which across-area and within-area dynamics are correlated. Specifically, can acrossarea dynamics vary independently from within-area dynamics across trials, potentially operating through a distinct communication subspace?

These are all very interesting questions. Our study does not attempt to parse activity into components predictive of muscle activity and others that may reflect other functions. Distinct components of RFA and CFA activity may very well rely on distinct interactions between them.

(3) While network modeling of RFA and CFA activity captured some aspects of behavioral and neural data, I wonder if certain findings such as the connection weight distribution (Figure 7C), across-region input (Figure 7F), and the within-region weights (Figure 7G), primarily resulted from fitting the different overall firing rates between the two regions with CFA exhibiting higher average firing rates. Did the authors account for this firing rate disparity when training the RNNs?

The key comparison in Figure 7 is shown in 7F, where the firing rates are accounted for in calculating the across-region input strength. Equalizing the firing rates in RFA and CFA would effectively increase RFA rates. If the mean firing rates in each region were appreciably dependent on across-region inputs, we would then expect an off-setting change in the RFA→CFA weights, such that the RFA→CFA distributions in 7F would stay the same. We would also expect the CFA→RFA weights would increase, since RFA neurons would need more input. This would shift the CFA→RFA (blue) distributions up. Thus, if anything, the key difference in this panel would only get larger. 

We also generally feel that it is a better approach to fit the actual firing rates, rather than normalizing, since normalizing the firing rates would take us further from the actual biology, not closer.

(4) Another way to assess the functional hierarchy is by comparing the time courses of movement representation between the two regions. For example, a linear decoder could be used to compare the amount of information about muscle activity and/or target location as well as time courses thereof between the two regions. This approach is advantageous because it incorporates behavior rather than focusing solely on neural activity. Since one of the main claims of this study is the limitation of inferring functional hierarchy from firing rate data alone, the authors should use the behavior as a lens for examining inter-areal interactions.

As we state above, we agree that examining interactions specific to movement-related activity components could reveal interesting structure in interregional interactions. Since it remains a challenge to rigorously identify a subset of neural activity patterns specifically related to driving muscle activity, any such analysis would involve an additional assumption. It remains unclear how well the activity that decoders use for predicting muscle activity matches the activity that actually drives muscle activity in situ.

To address this issue, which related to one raised by Reviewer #3 below, we have added an additional paragraph to the Discussion (see “Manifestations of hierarchy in firing patterns”).

Reviewer #3 (Public review):

This study investigates how two cortical regions that are central to the study of rodent motor control (rostral forelimb area, RFA, and caudal forelimb area, CFA) interact during directional forelimb reaching in mice. The authors investigate this interaction using

(1) optogenetic manipulations in one area while recording extracellularly from the other, (2) statistical analyses of simultaneous CFA/RFA extracellular recordings, and (3) network modeling.

The authors provide solid evidence that asymmetry between RFA and CFA can be observed, although such asymmetry is only observed in certain experimental and analytical contexts.

The authors find asymmetry when applying optogenetic perturbations, reporting a greater impact of RFA inactivation on CFA activity than vice-versa. The authors then investigate asymmetry in endogenous activity during forelimb movements and find asymmetry with some analytical methods but not others. Asymmetry was observed in the onset timing of movement-related deviations of local latent components with RFA leading CFA (computed with PCA) and in a relatively higher proportion and importance of cross-area latent components with RFA leading than CFA leading (computed with DLAG). However, no asymmetry was observed using several other methods that compute cross-area latent dynamics, nor with methods computed on individual neuron pairs across regions. The authors follow up this experimental work by developing a twoarea model with asymmetric dependence on cross-area input. This model is used to show that differences in local connectivity can drive asymmetry between two areas with equal amounts of across-region input.

Overall, this work provides a useful demonstration that different cross-area analysis methods result in different conclusions regarding asymmetric interactions between brain areas and suggests careful consideration of methods when analyzing such networks is critical. A deeper examination of why different analytical methods result in observed asymmetry or no asymmetry, analyses that specifically examine neural dynamics informative about details of the movement, or a biological investigation of the hypothesis provided by the model would provide greater clarity regarding the interaction between RFA and CFA.

Strengths:

The authors are rigorous in their experimental and analytical methods, carefully monitoring the impact of their perturbations with simultaneous recordings, and providing valid controls for their analytical methods. They cite relevant previous literature that largely agrees with the current work, highlighting the continued ambiguity regarding the extent to which there exists an asymmetry in endogenous activity between RFA and CFA.

A strength of the paper is the evidence for asymmetry provided by optogenetic manipulation. They show that RFA inactivation causes a greater absolute difference in muscle activity than CFA interaction (deviations begin 25-50 ms after laser onset, Figure 1) and that RFA inactivation causes a relatively larger decrease in CFA firing rate than CFA inactivation causes in RFA (deviations begin <25ms after laser onset, Figure 3). The timescales of these changes provide solid evidence for an asymmetry in the impact of inactivating RFA/CFA on the other region that could not be driven by differences in feedback from disrupted movement (which would appear with a ~50ms delay).

The authors also utilize a range of different analytical methods, showing an interesting difference between some population-based methods (PCA, DLAG) that observe asymmetry, and single neuron pair methods (granger causality, transfer entropy, and convergent cross mapping) that do not. Moreover, the modeling work presents an interesting potential cause of "hierarchy" or "asymmetry" between brain areas: local connectivity that impacts dependence on across-region input, rather than the amount of across-region input actually present.

Weaknesses:

There is no attempt to examine neural dynamics that are specifically relevant/informative about the details of the ongoing forelimb movement (e.g., kinematics, reach direction). Thus, it may be preemptive to claim that firing patterns alone do not reflect functional influence between RFA/CFA. For example, given evidence that the largest component of motor cortical activity doesn't reflect details of ongoing movement (reach direction or path; Kaufman, et al. PMID: 27761519) and that the analytical tools the authors use likely isolate this component (PCA, CCA), it may not be surprising that CFA and RFA do not show asymmetry if such asymmetry is related to the control of movement details. 

An asymmetry may still exist in the components of neural activity that encode information about movement details, and thus it may be necessary to isolate and examine the interaction of behaviorally-relevant dynamics (e.g., Sani, et al. PMID: 33169030).

To clarify, we are not claiming that firing patterns in no way reflect the asymmetric functional influence that we demonstrate with optogenetic inactivation. Instead, we show that certain types of analysis that we might expect to reflect such influence, in fact, do not. Indeed, DLAG did exhibit asymmetries that matched those seen in functional influence (at least qualitatively), though other methods we applied did not.

As we state above, we do think that there is more that can be gleaned by looking at influence specifically in terms of activity related to movement. However, if we did find that movement-related activity exhibited an asymmetry following functional influence, our results imply that the remaining activity components would exhibit an opposite asymmetry, such that the overall balance is symmetric. This would itself be surprising. We also note that the components identified by CCA and PLS do show substantial variation across reach targets, indicating that they are not only reflecting condition-invariant components. These analyses were performed on components accounting for well over 90% of the total activity variance, suggesting that both conditiondependent and condition-invariant components should be included.

To address the concern about condition-dependent and condition-invariant components, we have added a sentence to the Results section reporting our CCA and PLS results: “Because our results here involve the vast majority of trial-averaged activity variance, we expect that they encompass both components of activity that vary for different movement conditions (condition-dependent), and those that do not (condition-invariant).” To address the general concerns about potential differences in activity components specifically related to muscle activity, we have also added an additional paragraph to the Discussion (see “Manifestations of hierarchy in firing patterns”).

The idea that local circuit dynamics play a central role in determining the asymmetry between RFA and CFA is not supported by experimental data in this paper. The plausibility of this hypothesis is supported by the model but is not explored in any analyses of the experimental data collected. Given the focus on this idea in the discussion, further experimental investigation is warranted.

While we do not provide experimental support for this hypothesis, the data we present also do not contradict this hypothesis. Here we used modeling as it is often used - to capture experimental results and generate hypotheses about potential explanation. We do feel that our Discussion makes clear where the hypothesis derives from and does not misrepresent the lack of experimental support. We expect readers will take our engagement with this hypothesis with the appropriate grain of salt. The imaginable experiments to support such a hypothesis would constitute another substantial study, requiring numerous controls - a whole other paper in itself.

Recommendations for the authors:  

Reviewer #1 (Recommendations for the authors):

(1) There are a few small text/figure caption modifications that can be made for clarity of reading:

(2) Unclear sentence in the second paragraph of the introduction: "For example, stimulation applied in PM has been shown to alter the effects on muscles of stimulation in M1 under anesthesia, both in monkeys and rodents."

This sentence has been rephrased for clarity: “For example, in anesthetized monkeys34 and rodents35, stimulation in PM alters the effects of stimulation in M1 on muscles.”

(3) The first section of the results presents the optogenetic manipulation. However, the critical control that tests whether this was strictly a local manipulation that did not affect cells in the other region is introduced only much later. It may be helpful to add a comment in this section noting that such a control was performed, even if it is explained in detail later when introducing the recordings.

We have added the following to the first Results section: “we show below that direct optogenetic effects were only seen in the targeted forelimb area and not the other.”

(4) Figure 1D - I imagine these averages are from a single animal, but this is not stated in the figure caption.

“For one example mouse,” has been added to the beginning of the Figure 1D legend.

(5) Figure 2F - N=6 is not stated in the panel's caption (though it can make it clearer), while it is stated in the caption of 2H.

“n = 6 mice” has been added to the Figure 2F legend.

(6) There's some inconsistency with the order of RFA/CFA in the figures, sometimes RFA is presented first (e.g., Figure 1D and 1F), and sometimes CFA is presented first (e.g., panels of Figure 2).

We do not foresee this leading to confusion.

(7) "As expected, the majority of recorded neurons in each region exhibited an elevated average firing rate during movement as compared to periods when forelimb muscles were quiescent (Figure 2D,E; Figure S1A,B)" - Figure S1A,B show histograms of narrow vs. wide waveforms, is this the relevant figure here?

We apologize for the cryptic reference. The waveform width histograms were referred to here because they enabled the separation of narrow- and wide-waveform cells shown in Figure 2D,E. We have added the following clause to the referenced sentence to make this explicit:  “, both for narrow-waveform, putative interneurons and wide-waveform putative pyramidal neurons.”

(8) Figure 2I caption - "The fraction of activity variance from 150 ms before reach onset to 150 ms after it that occurs before reach onset" - this sentence is not clear.

The Figure 2I legend has been updated to “The activity variance in the 150 ms before muscle activity onset, defined as a fraction of the total activity variance from 150 ms before to 150 ms after muscle activity onset, for each animal (circles) and the mean across animals (black bars, n = 6 mice).”

(9) Figure 4B-G - is this showing results across the 6 animals? Not stated clearly.

Yes - the 21 sessions we had referred to are drawn from all six mice. We have updated the legend here to make this explicit.

(10) DLAG analysis - is there any particular reasoning behind choosing four across-region and four within-region components?

In actuality, we completed this analysis for a broad range of component numbers and obtained similar results in all cases. Four fell in the center of our range, and so we focused the illustrations shown in the figure on this value. In general, the number of components is arbitrary. The original paper from Gokcen et al. describes a method for identifying a lower bound on the number of distinct components the method can identify. However, this method yields different results for each individual recording session. For the comparisons we performed, we needed to use the same range of values for each session.

(11) Figure 5A seems to show 11 across-session components, it's unclear from the caption but I imagine this should show 12 (4 components times 3 sessions?)

As we state in the Methods, any across-region latent variable with a lag that failed to converge between the boundary values of ±200 ms was removed from the analysis. In the case illustrated in this panel, the lag for one of the components failed to converge and is not shown. We have now clarified this both in the relevant Results paragraph and in the figure legend.

(12) Figure 5B - is each marker here the average variance explained by all across/within components that were within the specified lag criteria across sessions per mouse? In other words, what does a single marker here stand for?

We apologize for the lack of clarity here. These values reflect the average across sessions for each mouse. We have updated the legend to make this explicit.

Reviewer #2 (Recommendations for the authors):

As I have addressed most of my major recommendations in the public review, I will use this section to include relatively minor points for the authors to consider.

(1) The EMG data in Figure 1C shows distinct patterns across spouts, both in the magnitude and complexity of muscle activations. It would be interesting to investigate whether these differences in muscle activity lead to behavioral variations (e.g., reaction time, reach duration) and how they relate to the relative involvement of the two areas.

We agree that it would be interesting to examine how the interactions between areas vary as behavior varies. While the differences between reaches here are limited, we have addressed this question for two substantially different motor behaviors (reaching and climbing) in a follow-up study that was recently preprinted (Kristl et al., biorxiv, 2025).

(2) How do the authors account for the lingering impact of RFA inactivation on muscle activity, which persists for tens of milliseconds after laser offset? Could this effect be due to compensatory motor activity following the perturbation? A further illustration of how the raw limb trajectories and/or muscle activity are perturbed and recovered would help readers better understand the impact of motor cortical inactivation.

To clarify the effects of inactivation on a longer timescale, we have added a new supplemental figure showing the plots from Figure 1D over a longer time window extending to 500 ms after trial onset (new Figure S1). Lingering effects do persist, at least in certain cases. In general, we find it hard to ascertain the source of optogenetic effects on longer timescales like this. On the shortest timescales, effects will be mediated by relatively direct connections between regions. However, on these longer timescales, effects could be due to broader changes in brain and behavioral state that can influence muscle activity. For example, attempts to compensate for the initial disturbance to muscle activity could cause divergence from controls on these longer timescales. Muscle tissue itself is also known to have long timescale relaxation dynamics, and it would not be surprising if the relevant control circuits here also had long timescales dynamics, such that we would not expect an immediate return to control when the light pulse ends. Because of this ambiguity, we generally avoid interpretation of optogenetic effects on these longer timescales.

Reviewer #3 (Recommendations for the authors):

(1) Page 9: ". We measured the time at which the activity state deviated from baseline preceding reach onset," - I cannot find how this deviation was defined (neither the baseline nor the threshold).

We have added text to the Figure 2G legend that explicitly states how the baseline and activity onset time were defined.

(2) Given the shape of the curves in Figure 2G, the significance of this result seems susceptible to slight modifications of what defines a baseline or a deviation threshold. For example, it looks like the circle for CFA has a higher y-axis value, suggesting the baseline deviance is higher, but it is unclear why that would be from the plot. If the threshold for deviation in neural activity state were held uniform between CFA and RFA is the difference still significant across animals?

We have repeated the analysis using the same absolute threshold for each region. We used the higher of the two thresholds from each region. The difference remains significant. This is now described in the last paragraph of the Results section for Figure 2.

(3) Since summed deviation of the top 3 PCs is used to show a difference in activity onset between CFA/RFA, but only a small proportion of variance is explained pre-movement (<2% in most animals), it seems relevant to understand what percentage of CFA/RFA neuron activity actually is modulated and deviates from baseline prior to movement and to show the distribution of activity onsets at the single neuron level in CFA/RFA. Can an onset difference only be observed using PCA? 

Because many neurons have low firing rates, estimating the time at which their firing rate begins to rise near reach onset is difficult to do reliably. It is also true that not all neurons show an increase around onset - some show a decrease and others show no discernible change. Using PCs to measure onset avoids both of these problems, since they capture both increases and decreases in individual neuron firing rates and are much less noisy than individual neuron firing rates. 

However, based on this comment, we have repeated this analysis on a single-neuron level using only neurons with relatively high average firing rates. Specifically, we analyzed neurons with mean firing rates above the 90th percentile across all sessions within an animal. Neurons whose activity never crossed threshold were excluded. Results matched those using PCs, with RFA neurons showing an earlier average activity onset time. This is now described in the last paragraph of the Results section for Figure 2.

(4) It is stated that to study the impact of inactivation on CFA/RFA activity, only the 50 highest average firing rate neurons were used (and maybe elsewhere too, e.g., convergent cross mapping). It is unclear why this subselection is necessary. It is justified by stating that higher firing rate neurons have better firing rate estimates. This may be supportable for very low firing rate units that spike sorting tools have a hard time tracking, but I don't think this is supported by data for most of the distribution of firing rates. It therefore seems like the results might be biased by a subselection of certain high firing rate neuron populations. It would be useful to also compute and mention if the results for all neurons/neuron pairs are the same. If there is worry about low-quality units being those with low firing rates, a threshold for firing rate as used elsewhere in the paper (at least 1 spike / 2 trials) seems justified.

The issue here is that as firing rates decrease and firing rate estimates get noisier, estimates of the change in firing rate get more variable. Here we are trying to estimate the fraction of neurons for which firing rates decreased upon inactivation of the other region. Variability in estimates of the firing rate change will bias this estimate toward 50%, since in the limit when the change estimates are entirely based on noise, we expect 50% to be decreases. As expected, when we use increasingly liberal thresholds for this analysis, the fraction of decreases trends closer to 50%. 

As a consequence of this, we cannot easily distinguish whether higher firing rate neurons might for some reason have a greater tendency to exhibit decreases in firing compared to lower firing rate neurons. However, we see no positive reason to expect such a difference. We have added a sentence noting this caveat in interpreting our findings to the relevant paragraph of the Results.

The lack of min/max axis values in Figure 3B-F makes it hard to interpret - are these neurons almost silent when near the bottom of the plot or are they still firing a substantial # of spikes?

To aid interpretation of the relative magnitude of firing rate changes, we have added minimum firing rates for the averages depicted in Figure 3B,C,E and F to the legend. Our original thinking was that the plots in Figure 3G and H would provide an indication of the relative changes in firing.

It would be interesting to know if the impact of optogenetic stimulation changed with exposure to the manipulation. Are all results presented only from the first X number of sessions in each animal? Or is the effect robust over time and (within the same animal) you can get the same results of optogenetic inactivation over time? This information seems critical for reproducibility.

We have now performed brief optogenetic inactivations in several brain areas in several different behavioral paradigms, and have found that inactivation effects are stable both within and across sessions, almost surprisingly so. This includes cases where the inactivations were more frequent (every ~1.25 s on average) and more numerous (>15,000 trials per animal) than in the present manuscript. Thus we did not restrict our analysis here to the first X sessions or trials within a session. We have added additional plots as Figure S3T-AA showing the stability of optogenetic effects both within and across sessions.

Given that it can be difficult to record from interneurons (as the proportion of putative interneurons in Figure S1 attests), the SALT analyses would be more convincing if a few recordings had been performed in the same region as optogenetic stimulation to show a "positive control" of what direct interneuron stimulation looks like. Could also use this to validate the narrow/wide waveform classification.

We have verified that using SALT as we have in the present manuscript does detect vGAT+ interneurons directly responding to light. This is included in a recent preprint from the lab (Kristl et al., biorxiv, 2025). We (Warriner et al., Cell Reports, 2022) and others (Guo et al., Neuron, 2014) have previously used direct ChR2 activation to validate waveform-based classification.

Simultaneous CFA/RFA recordings during optogenetic perturbation would also allow for time courses of inhibition to be compared in RFA/CFA. Does it take 25ms to inhibit locally, and the cross-area impact is fast, or does it inactivate very fast locally and takes ~25ms to impact the other region?

Latencies of this sort are difficult to precisely measure given the statistical limits of this sort of data, but there does appear to be some degree of delay between local and downstream effects. We do not have a statistical foundation as of yet for concluding that this is the case. It will be interesting to examine this issue more rigorously in the future.

Given the difference in the analytical methods, the authors should share data in a relatively unprocessed format (e.g., spike times from sorted units relative to video tracking + behavioral data), along with analysis code, to allow others to investigate these differences.

We plan to post the data and code to our lab’s Github site once the Version of Record is online.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:  

Reviewer #1 (Public Review):

Summary:

This paper reports an intracranial SEEG study of speech coordination, where participants synchronize their speech output with a virtual partner that is designed to vary its synchronization behavior. This allows the authors to identify electrodes throughout the left hemisphere of the brain that have activity (both power and phase) that correlates with the degree of synchronization behavior. They find that high-frequency activity in the secondary auditory cortex (superior temporal gyrus) is correlated to synchronization, in contrast to primary auditory regions. Furthermore, activity in the inferior frontal gyrus shows a significant phase-amplitude coupling relationship that is interpreted as compensation for deviation from synchronized behavior with the virtual partner.

Strengths:

(1) The development of a virtual partner model trained for each individual participant, which can dynamically vary its synchronization to the participant's behavior in real-time, is novel and exciting.

(2) Understanding real-time temporal coordination for behaviors like speech is a critical and understudied area.

(3) The use of SEEG provides the spatial and temporal resolution necessary to address the complex dynamics associated with the behavior.

(4) The paper provides some results that suggest a role for regions like IFG and STG in the dynamic temporal coordination of behavior both within an individual speaker and across speakers performing a coordination task.

We thank the Reviewer for their positive comments on our manuscript.

Weaknesses:

(1) The main weakness of the paper is that the results are presented in a largely descriptive and vague manner. For instance, while the interpretation of predictive coding and error correction is interesting, it is not clear how the experimental design or analyses specifically support such a model, or how they differentiate that model from the alternatives. It's possible that some greater specificity could be achieved by a more detailed examination of this rich dataset, for example by characterizing the specific phase relationships (e.g., positive vs negative lags) in areas that show correlations with synchronization behavior. However, as written, it is difficult to understand what these results tell us about how coordination behavior arises.

We understand the reviewer’s comment. It is true that this work, being the first in the field using real-time adapting synchronous speech and intracerebral neural data, is a descriptive work, that hopefully will pave the way for further studies. We have now added more statistical analyses (see point 2) to go beyond a descriptive approach and we have also rewritten the discussion to clarify how this work can possibly contribute to disentangle different models of language interaction. Most importantly we have also run new analyses taking into account the specific phase relationship, as suggested.

We already had an analysis using instantaneous phase difference in the phase-amplitude coupling approach, that bridges phase of behaviour to neural responses (amplitude in the high-frequency range). However, this analysis, as the reviewer noted, does not distinguish between positive and negative lags, but rather uses the continuous fluctuations of coordinative behaviour. Following the reviewer’s suggestion, we have now run a new analysis estimating the average delay (between virtual partner speech and patient speech) in each trial, using a cross-correlation approach. This gives a distribution of delays across trials that can then be “binned” as positive or negative. We have thus rerun the phase-amplitude coupling analyses on positive and negative trials separately, to assess whether the phase amplitude relationship depends upon the anticipatory (negative lags) or compensatory (positive lags) behaviour. Our new analysis (now in the supplementary, see figure below) does not reveal significant differences between positive and negative lags. This lack of difference, although not easy to interpret, is nonetheless interesting because it seems to show that the IFG does not have a stronger coupling for anticipatory trials. Rather the IFG seems to be strongly involved in adjusting behaviour, minimizing the error, independently of whether this is early or late.

We have updated the “Coupling behavioural and neurophysiological data” section in Materials and methods as follows:  

“In the third approach, we assessed whether the phase-amplitude relationship (or coupling) depends upon the anticipatory (negative delays) or compensatory (positive delays) behaviour between the VO and the patients’ speech. We computed the average delay in each trial using a cross-correlation approach on speech signals (between patient and VP) with the MATLAB function xcorr. A median split (patient-specific ; average median split = 0ms, average sd = 24ms) was applied to conserve a sufficient amount of data, classifying trials below the median as “anticipatory behaviour” and trials above the median as “compensatory behaviour”. Then we conducted the phase-amplitude coupling analyses on positive and negative trials separately.”

We also added a paragraph on this finding in the Discussion:

“Our results highlight the involvement of the inferior frontal gyrus (IFG) bilaterally, in particular the BA44 region, in speech coordination. First, trials with a weak verbal coordination (VCI) are accompanied by more prominent high frequency activity (HFa, Fig.4; Fig.S4). Second, when considering the within-trial time-resolved dynamics, the phase-amplitude coupling (PAC) reveals a tight relation between the low frequency behavioural dynamics (phase) and the modulation of high-frequency neural activity (amplitude, Fig.5B ; Fig.S5). This relation is strongest when considering the phase adjustments rather than the phase of speech of the VP per se : larger deviations in verbal coordination are accompanied by increase in HFa. Additionally, we also tested for potential effects of different asynchronies (i.e., temporal delay) between the participant's speech and that of the virtual partner but found no significant differences (Fig.S6). While lack of delay-effect does not permit to conclude about the sensitivity of BA44 to absolute timing of the partner’s speech, its neural dynamics are linked to the ongoing process of resolving phase deviations and maintaining synchrony.”

(2) In the results section, there's a general lack of quantification. While some of the statistics reported in the figures are helpful, there are also claims that are stated without any statistical test. For example, in the paragraph starting on line 342, it is claimed that there is an inverse relationship between rho-value and frequency band, "possibly due to the reversed desynchronization/synchronization process in low and high frequency bands". Based on Figure 3, the first part of this statement appears to be true qualitatively, but is not quantified, and is therefore impossible to assess in relation to the second part of the claim. Similarly, the next paragraph on line 348 describes optimal clustering, but statistics of the clustering algorithm and silhouette metric are not provided. More importantly, it's not entirely clear what is being clustered - is the point to identify activity patterns that are similar within/across brain regions? Or to interpret the meaning of the specific patterns? If the latter, this is not explained or explored in the paper.

The reviewer is right. We have now added statistical analyses showing that:

(1) the ratio between synchronization and desynchronization evolves across frequencies (as often reported in the literature).

(2) the sign of rho values also evolves across frequencies.

(3) the clustering does indeed differ when taking into account behaviour. We have also clarified the use of clustering and the reasoning behind it.

We have updated the Materials and methods section as follows:

“The statistical difference between spatial clustering in global effect and brain-behaviour correlation was estimated with linear model using the R function lm (stat package), post-hoc comparisons were corrected for multiple comparisons using the Tukey test (lsmeans R package ; Lenth, 2016). The statistical difference between clustering in global effect and behaviour correlation across the number of clusters was estimated using permutation tests (N=1000) by computing the silhouette score difference between the two conditions.” We have updated the Results section as follows:

(1) “This modulation between synchronization and desynchronization across frequencies was significant (F(5) = 6.42, p < .001 ; estimated with linear model using the R function lm).”

(2) “The first observation is a gradual transition in the direction of correlations as we move up frequency bands, from positive correlations at low frequencies to negative ones at high frequencies (F(5) = 2.68, p = .02). This effect, present in both hemispheres, mimics the reversed desynchronization/synchronization process in low and high frequency bands reported above.”

(3) “Importantly, compared to the global activity (task vs rest, Fig 3A), the neural spatial profile of the behaviour-related activity (Fig 3B) is more clustered, in the left hemisphere. Indeed, silhouette scores are systematically higher for behaviour-related activity compared to global activity, indicating greater clustering consistency across frequency bands (t(106) = 7.79, p < .001, see Figure S3). Moreover, silhouette scores are maximal, in particular for HFa, for five clusters (p < .001), located in the IFG BA44, the IPL BA 40 and the STG BA 41/42 and BA22 (see Figure S3).”

(3) Given the design of the stimuli, it would be useful to know more about how coordination relates to specific speech units. The authors focus on the syllabic level, which is understandable. But as far as the results relate to speech planning (an explicit point in the paper), the claims could be strengthened by determining whether the coordination signal (whether error correction or otherwise) is specifically timed to e.g., the consonant vs the vowel. If the mechanism is a phase reset, does it tend to occur on one part of the syllable?

Thank you for this thoughtful feedback. We agree that the relationship between speech coordination and specific speech units, such as consonants versus vowels, is an intriguing question. However, in our study, both interlocutors (the participant and the virtual partner) are adapting their speech production in real-time. This interactive coordination makes it difficult to isolate neural signatures corresponding to precise segments like consonants or vowels, as the adjustments occur in a continuous and dynamic context.

The VP's ability to adapt depends on its sensitivity to spectral cues, such as the transition from one phonetic element to another. This is likely influenced by the type of articulation, with certain transitions being more salient (e.g., between a stop consonant like "p" and a vowel like "a") and others being less distinct (e.g., between nasal consonants like "m" and a vowel). Thus, the VP’s spectral adaptation tends to occur at these transitions, which are more prominent in some cases than in others.

For the participants, previous studies have shown a greater sensitivity during the production of stressed vowels (Oschkinat & Hoole, 2022; Li & Lancia, 2024), which may reflect a heightened attentional or motor adjustment to stressed syllables.

Here, we did not specifically address the question of coordination at the level of individual linguistic units. Moreover, even if we attempted to focus on this level, it would be challenging to relate neural dynamics directly to specific speech segments. The question of how synchronization at the level of individual linguistic units might relate to neural data is complex. The lack of clear, unit-specific predictions makes it difficult to parse out distinct neural signatures tied to individual segments, particularly when both interlocutors are continuously adjusting their speech in relation to one another.

Therefore, while we recognize the potential importance of examining synchronization at the level of individual phonetic elements, the design of our task and the nature of the coordination in this interactive context (realtime bidirection adaptation) led us to focus more broadly on the overall dynamics of speech synchronization at the syllabic level, rather than on specific linguistic units.

We now state at the end of the Discussion section:

“It is worth noting that the influence of specific speech units, such as consonants versus vowels, on speech coordination remains to be explored. In non-interactive contexts, participants show greater sensitivity during the production of stressed vowels, possibly reflecting heightened attentional or motor adjustments (Oschkinat & Hoole, 2022; Li & Lancia, 2024). In this study, the VP’s adaptation relies on sensitivity to spectral cues, particularly phonetic transitions, with some (e.g., formant transitions) being more salient than others. However, how these effects manifest in an interactive setting remains an open question, as both interlocutors continuously adjust their speech in real time. Future studies could investigate whether coordination signals, such as phase resets, preferentially align with specific parts of the syllable.” References cited:

– Oschkinat, M., & Hoole, P. (2022). Reactive feedback control and adaptation to perturbed speech timing in stressed and unstressed syllables. Journal of Phonetics, 91, 101133.

– Li, J., & Lancia, L. (2024). A multimodal approach to study the nature of coordinative patterns underlying speech rhythm. In Proc. Interspeech, 397-401.

(4) In the discussion the results are related to a previously-described speech-induced suppression effect. However, it's not clear what the current results have to do with SIS, since the speaker's own voice is present and predictable from the forward model on every trial. Statements such as "Moreover, when the two speech signals come close enough in time, the patient possibly perceives them as its own voice" are highly speculative and apparently not supported by the data.

We thank the reviewer for raising thoughtful concerns about our interpretation of the observed neural suppression as related to speaker-induced suppression (SIS). We agree that our study lacks a passive listening condition, which limits direct comparisons to the original SIS effect, traditionally defined as the suppression of neural responses to self-produced speech compared to externally-generated speech (Meekings & Scott, 2021).

In response, we have reconsidered our terminology and interpretation. In the revised Discussion section, we refer to our findings as a "SIS-related phenomenon specific to the synchronous speech context". Unlike classic SIS paradigms, our interactive task involves simultaneous monitoring of self- and externally-generated speech, introducing additional attentional and coordinative demands.

The revised Discussion also incorporates findings by Ozker et al. (2022, 2024), which link SIS and speech monitoring, suggesting that suppressing responses to self-generated speech facilitates error detection. We propose that the decrease in high-frequency activity (HFa) as verbal coordination increases reflects reduced error signals due to closer alignment between perceived and produced speech. Conversely, HFa increases with reduced coordination may signify greater prediction error.

Additionally, we relate our findings to the "rubber voice" effect (Zheng et al., 2011; Lind et al., 2014; Franken et al., 2021), where temporally and phonetically congruent external speech can be perceived as self-generated. We speculate that this may occur in synchronous speech tasks when the participant's and VP's speech signals closely align. However, this interpretation remains speculative, as no subjective reports were collected to confirm this perception. Future studies could include participant questionnaires to validate this effect and relate subjective experience to neural measures of synchronization.

Overall, our findings extend the study of SIS to dynamic, interactive contexts and contribute to understanding internal forward models of speech production in more naturalistic scenarios.

We have now added these points to the discussion as follows:

“The observed negative correlation between verbal coordination and high-frequency activity (HFa) in STG BA22 suggests a suppression of neural responses as the degree of behavioural synchrony increases. This result is reminiscent of findings on speaker-induced suppression (SIS), where neural activity in auditory cortex decreases during self-generated speech compared to externally-generated speech (Meekings & Scott, 2021; Niziolek et al., 2013). However, our paradigm differs from traditional SIS studies in two critical ways: (1) the speaker's own voice is always present and predictable from the forward model, and (2) no passive listening condition was included. Therefore, our findings cannot be directly equated with the original SIS effect.

Instead, we propose that the suppression observed here reflects a SIS-related phenomenon specific to the synchronous speech context. Synchronous speech requires simultaneous monitoring of self- and externallygenerated speech, a task that is both attentionally demanding and coordinative. This aligns with evidence from Ozker et al. (2024, 2022), showing that the same neural populations in STG exhibit SIS and heightened responses to feedback perturbations. These findings suggest that SIS and speech monitoring are related processes, where suppressing responses to self-generated speech facilitates error detection. In our study, suppression of HFa as coordination increases may reflect reduced prediction errors due to closer alignment between perceived and produced speech signals. Conversely, increased HFa during poor coordination may signify greater mismatch, consistent with prediction error theories (Houde & Nagarajan, 2011; Friston et al., 2020). Furthermore, when self- and externally-generated speech signals are temporally and phonetically congruent, participants may perceive external speech as their own. This echoes the "rubber voice" effect, where external speech resembling self-produced feedback is perceived as self-generated (Zheng et al., 2011; Lind et al., 2014; Franken et al., 2021). While this interpretation remains speculative, future studies could incorporate subjective reports to investigate this phenomenon in more detail.” References cited:

– Franken, M. K., Hartsuiker, R. J., Johansson, P., Hall, L., & Lind, A. (2021). Speaking With an Alien Voice: Flexible Sense of Agency During Vocal Production. Journal of Experimental Psychology-Human perception and performance, 47(4), 479-494. https://doi.org/10.1037/xhp0000799

– Houde, J. F., & Nagarajan, S. S. (2011). Speech production as state feedback control. Frontiers in human neuroscience, 5, 82.

– Lind, A., Hall, L., Breidegard, B., Balkenius, C., & Johansson, P. (2014). Speakers' acceptance of real-time speech exchange indicates that we use auditory feedback to specify the meaning of what we say. Psychological Science, 25(6), 1198-1205. https://doi.org/10.1177/0956797614529797

– Meekings, S., & Scott, S. K. (2021). Error in the Superior Temporal Gyrus? A Systematic Review and Activation Likelihood Estimation Meta-Analysis of Speech Production Studies. Journal of Cognitive Neuroscience, 33(3), 422-444. https://doi.org/10.1162/jocn_a_01661

– Niziolek C. A., Nagarajan S. S., Houde J. F (2013) What does motor efference copy represent? Evidence from speech production Journal of Neuroscience 33:16110–16116Ozker M., Doyle W., Devinsky O., Flinker A (2022) A cortical network processes auditory error signals during human speech production to maintain fluency PLoS Biology 20.

– Ozker, M., Yu, L., Dugan, P., Doyle, W., Friedman, D., Devinsky, O., & Flinker, A. (2024). Speech-induced suppression and vocal feedback sensitivity in human cortex. eLife, 13, RP94198. https://doi.org/10.7554/eLife.94198

– Zheng, Z. Z., MacDonald, E. N., Munhall, K. G., & Johnsrude, I. S. (2011). Perceiving a Stranger's Voice as Being One's Own: A 'Rubber Voice' Illusion? PLOS ONE, 6(4), e18655.

(5) There are some seemingly arbitrary decisions made in the design and analysis that, while likely justified, need to be explained. For example, how were the cutoffs for moderate coupling vs phase-shifted coupling (k ~0.09) determined? This is noted as "rather weak" (line 212), but it's not clear where this comes from. Similarly, the ROI-based analyses are only done on regions "recorded in at least 7 patients" - how was this number chosen? How many electrodes total does this correspond to? Is there heterogeneity within each ROI?

The reviewer is correct, we apologize for this missing information. We now specify that the coupling values were empirically determined on the basis of a pilot experiment in order to induce more or less synchronization, but keeping the phase-shifted coupling at a rather implicit level.  

Concerning the definition of coupling as weak, one should consider that, in the Kuramoto model, the strength of coupling (k) is relative to the spread of the natural frequencies (Δω) in the system. In our study, the natural frequencies of syllables range approximately from 2 Hz to 10Hz, resulting in a frequency spread of Δω = 8 Hz. For coupling to strongly synchronize oscillators across such a wide range, k must be comparable to or exceed Δω. Thus, since k = 0.1 is far much smaller than Δω, it is therefore classified as weak coupling.

We have now modified the Materials and methods section as follows:

“More precisely, for a third of the trials the VP had a neutral behaviour (close to zero coupling: k = +/- 0.01). For a third it had a moderate coupling, meaning that the VP synchronised more to the participant speech (k = -0.09). And for the last third of the trials the VP had a moderate coupling but with a phase shift of pi/2, meaning that it moderately aimed to speak in between the participant syllables (k = + 0.09). The coupling values were empirically determined on the basis of a pilot experiment in order to induce more or less synchronization but keeping the phase-shifted coupling at a rather implicit level. In other terms, while participants knew that the VP would adapt, they did not necessarily know in which direction the coupling went.”

Regarding the criterion of including regions recorded in at least 7 patients, our goal was to balance data completeness with statistical power. Given our total sample of 16 patients, this threshold ensures that each included region is represented in at least ~44% of the cohort, reducing the likelihood of spurious findings due to extremely small sample sizes. This choice also aligns with common neurophysiological analysis practices, where a minimum number of subjects (at least 2 in extreme cases) is required to achieve meaningful interindividual comparisons while avoiding excessive data exclusion. Additionally, this threshold maintains a reasonable tradeoff between maximizing patient inclusion and ensuring that statistical tests remain robust.

We have now added more information in the Results section “Spectral profiles in the language network are nuanced by behaviour” on this point as follows:

“To balance data completeness and statistical power, we included only brain regions recorded in at least 7 patients (~44% of the cohort) for the left hemisphere and at least 5 patients for the right hemisphere (~31% of the cohort), ensuring sufficient representation while minimizing biases due to sparse data.”

Reviewer #2 (Public Review):

Summary:

This paper investigates the neural underpinnings of an interactive speech task requiring verbal coordination with another speaker. To achieve this, the authors recorded intracranial brain activity from the left hemisphere in a group of drug-resistant epilepsy patients while they synchronised their speech with a 'virtual partner'. Crucially, the authors were able to manipulate the degree of success of this synchronisation by programming the virtual partner to either actively synchronise or desynchronise their speech with the participant, or else to not vary its speech in response to the participant (making the synchronisation task purely one-way). Using such a paradigm, the authors identified different brain regions that were either more sensitive to the speech of the virtual partner (primary auditory cortex), or more sensitive to the degree of verbal coordination (i.e. synchronisation success) with the virtual partner (secondary auditory cortex and IFG). Such sensitivity was measured by (1) calculating the correlation between the index of verbal coordination and mean power within a range of frequency bands across trials, and (2) calculating the phase-amplitude coupling between the behavioural and brain signals within single trials (using the power of high-frequency neural activity only). Overall, the findings help to elucidate some of the left hemisphere brain areas involved in interactive speaking behaviours, particularly highlighting the highfrequency activity of the IFG as a potential candidate supporting verbal coordination.

Strengths:

This study provides the field with a convincing demonstration of how to investigate speaking behaviours in more complex situations that share many features with real-world speaking contexts e.g. simultaneous engagement of speech perception and production processes, the presence of an interlocutor, and the need for inter-speaker coordination. The findings thus go beyond previous work that has typically studied solo speech production in isolation, and represent a significant advance in our understanding of speech as a social and communicative behaviour. It is further an impressive feat to develop a paradigm in which the degree of cooperativity of the synchronisation partner can be so tightly controlled; in this way, this study combines the benefits of using prerecorded stimuli (namely, the high degree of experimental control) with the benefits of using a live synchronisation partner (allowing the task to be truly two-way interactive, an important criticism of other work using pre-recorded stimuli). A further key strength of the study lies in its employment of stereotactic EEG to measure brain responses with both high temporal and spatial resolution, an ideal method for studying the unfolding relationship between neural processing and this dynamic coordination behaviour.

We sincerely appreciate the Reviewer's thoughtful and positive feedback on our manuscript.

Weaknesses:

One major limitation of the current study is the lack of coverage of the right hemisphere by the implanted electrodes. Of course, electrode location is solely clinically motivated, and so the authors did not have control over this. However, this means that the current study neglects the potentially important role of the right hemisphere in this task. The right hemisphere has previously been proposed to support feedback control for speech (likely a core process engaged by synchronous speech), as opposed to the left hemisphere which has been argued to underlie feedforward control (Tourville & Guenther, 2011). Indeed, a previous fMRI study of synchronous speech reported the engagement of a network of right hemisphere regions, including STG, IPL, IFG, and the temporal pole (Jasmin et al., 2016). Further, the release from speech-induced suppression during a synchronous speech reported by Jasmin et al. was found in the right temporal pole, which may explain the discrepancy with the current finding of reduced leftward high-frequency activity with increasing verbal coordination (suggesting instead increased speech-induced suppression for successful synchronisation). The findings should therefore be interpreted with the caveat that they are limited to the left hemisphere, and are thus likely missing an important aspect of the neural processing underpinning verbal coordination behaviour.

We have now included, in the supplementary materials, data from the right hemisphere, although the coverage is a bit sparse (Figures S2, S4, S5, see our responses in the ‘Recommendation for the authors’ section, below). We have also revised the Discussion section to add the putative role of right temporal regions (see below as well).

A further limitation of this study is that its findings are purely correlational in nature; that is, the results tell us how neural activity correlates with behaviour, but not whether it is instrumental in that behaviour. Elucidating the latter would require some form of intervention such as electrode stimulation, to disrupt activity in a brain area and measure the resulting effect on behaviour. Any claims therefore as to the specific role of brain areas in verbal coordination (e.g. the role of the IFG in supporting online coordinative adjustments to achieve synchronisation) are therefore speculative.

We appreciate the reviewer’s observation regarding the correlational nature of our findings and agree that this is a common limitation of neuroimaging studies. While elucidating causal relationships would indeed require intervention techniques such as electrical stimulation, our study leverages the unique advantages of intracerebral recordings, offering the best available spatial and temporal resolution alongside a high signal-tonoise ratio. These attributes ensure that our data accurately reflect neural activity and its temporal dynamics, providing a robust foundation for understanding the relationship between neural processes and behaviour. Therefore, while causal claims are beyond the scope of this study, the precision of our methodology allows us to make well-supported observations about the neural correlates of synchronous speech tasks.

Recommendations for the authors:

Reviewing Editor Comment:

After joint consultation, we are seeing the potential for the report to be strengthened and the evidence here to be deemed ultimately at least 'solid': to us (editors and reviewers) it seems that this would require both (1) clarifying/acknowledging the limitations of not having right hemisphere data, and (2) running some of the additional analyses the reviewers suggest, which should allow for richer examination of the data e.g. phase relationships in areas that correlate with synchronisation.

We have now added data on the right hemisphere (RH) that we did not previously report due to a rather sparse sampling of the RH. These results are now reported in the Results section as well as in the Supplementary section, where we put all right hemisphere figures for all analyses (Figure S2, S4, S5). We have also run additional analyses digging into the phase relationship in areas that correlate with synchronisation (Figure S6). These additional analyses allowed us to improve the Discussion section as well.

Reviewer #1 (Recommendations For The Authors):

In some sections, the writing is a bit unclear, with both typos and vague statements that could be fixed with careful proofreading.

We thank the reviewer for pointing out areas where the writing could be improved. We carefully proofread the manuscript to address typos and clarify any vague statements. Specific sections identified as unclear have been rephrased for better precision and readability.

In Figure 1, the colors repeat, making it impossible to tell patients apart.

We have now updated Figure 1 colormap to avoid redundancy and added the right hemisphere.

Line 132: "16 unilateral implantations (9 left, 7 bilateral implantations)". Should this say 7 right hemisphere? If so, the following sentence stating that there was "insufficient cover [sic] of the right hemisphere" is unclear, since the number of patients between LH and RH is similar.

The confusion was due to the fact that the lateralization refers to the presence/absence of electrodes in the Heschl’s gyrus (left : H’ ; right : H) exclusively.

We have thus changed this section as follows:

“16 patients (7 women, mean age 29.8 y, range 17 - 50 y) with pharmacoresistant epilepsy took part in the study. They were included if their implantation map covered at least partially the Heschl's gyrus and had sufficiently intact diction to support relatively sustained language production.” The relevant part (previously line 132) now states:

“Sixteen patients with a total of 236 electrodes (145 in the left hemisphere) and 2395 contacts (1459 in the left hemisphere, see Figure 1). While this gives a rather sparse coverage of the right hemisphere, we decided, due to the rarity of this type of data, to report results for both hemispheres, with figures for the left hemisphere in the main text and figures for the right hemisphere in the supplementary section.”

Reviewer #2 (Recommendations For The Authors):

(1) To address the concern regarding the absence of data from the right hemisphere, I would advise the authors to directly acknowledge this limitation in their Discussion section, citing relevant work suggesting that the right hemisphere has an important role to play in this task (e.g. Jasmin et al., 2016). You should also make this clear in your abstract e.g. you could rewrite the sentence in line 40 to be: "Then, we recorded the intracranial brain activity of the left hemisphere in 16 patients with drug-resistant epilepsy...".

We are grateful to the reviewer for this comment that incited us to look into the right hemisphere data. We have now included results in the right hemisphere, although the coverage is a bit sparse. We have also revised the Discussion section to add the putative role of right temporal regions. Interestingly, our results show, as suggested by the reviewer, a clear involvement of the RH in this task.

First, the full brain analyses show a very similar implication of the RH as compared to the LH (see Figure below). We have now added in the Results section:

“As expected, the whole language network is strongly involved, including both dorsal and ventral pathways (Fig 3A). More precisely, in the left temporal lobe the superior, middle and inferior temporal gyri, in the left parietal lobe the inferior parietal lobule (IPL) and in the left frontal lobe the inferior frontal gyrus (IFG) and the middle frontal gyrus (MFG). Similar results are observed in the right hemisphere, neural responses being present across all six frequency bands with medium to large modulation in activity compared to baseline (Figure S2A) in the same regions. Desynchronizations are present in the theta, alpha and beta bands while the low gamma and HFa bands show power increases.”

As to compared to the left hemisphere, assessing brain-behaviour correlations in the right hemisphere does not provide the same statistical power, because some anatomical regions have very few electrodes. Nonetheless, we observe a strong correlation in the right IFG, similar to the one we previously reported in the left hemisphere, and we now report in the Results section:

“The decrease in HFa along the dorsal pathway is replicated in the right hemisphere (Figure S4). However, while both the right STG BA41/42 and STG BA22 present a power increase (compared to baseline) — with a stronger increase for the STG BA41/42 — neither shows a significant correlation with verbal coordination (t(45)=-1.65, p=.1 ; t(8)=-0.67, p=.5 ; Student’s T test, FDR correction). By contrast, results in the right IFG BA44 are similar to the one observed in the left hemisphere with a significant power increase associated with a negative brainbehaviour correlation (t(17) = -3.11, p = .01 ; Student’s T test, FDR correction).”

Interestingly, the phase-amplitude coupling analysis yields very similar results in both hemispheres (exception made for BA22). We have thus updated the Results section as follows:

“Notably, when comparing – within the regions of interest previously described – the PAC with the virtual partner speech and the PAC with the phase difference, the coupling relationship changes when moving along the dorsal pathway: a stronger coupling in the auditory regions with the speech input, no difference between speech and coordination dynamics in the IPL and a stronger coupling for the coordinative dynamics compared to speech signal in the IFG (Figure 5B ). When looking at the right hemisphere, we observe the same changes in the coupling relationship when moving along the dorsal pathway, except that no difference between speech and coordination dynamics is present in the right secondary auditory regions (STG BA22; Figure S5).”

We also included in the Discussion section the right hemisphere results also mentioning previous work of Guenther and the one of Jasmin. On the section “Left secondary auditory regions are more sensitive to coordinative behaviour” one can read:

“Furthermore, the absence of correlation in the right STG BA22 (Figure S4) seems in first stance to challenge influential speech production models (e.g. Guenther & Hickok, 2016) that propose that the right hemisphere is involved in feedback control. However, one needs to consider the the task at stake heavily relied upon temporal mismatches and adjustments. In this context, the left-lateralized sensitivity to verbal coordination reminds of the works of Floegel and colleagues (2020, 2023) suggesting that both hemispheres are involved depending on the type of error: the right auditory association cortex monitoring preferentially spectral speech features and the left auditory association cortex monitoring preferentially temporal speech features. Nonetheless, the right temporal pole seems to be sensitive to speech coordinative behaviour, confirming previous findings using fMRI (Jasmin et al., 2016) and thus showing that the right hemisphere has an important role to play in this type of tasks (e.g. Jasmin et al., 2016).”

References cited:

– Floegel, M., Fuchs, S., & Kell, C. A. (2020). Differential contributions of the two cerebral hemispheres to temporal and spectral speech feedback control. Nature Communications, 11(1), 2839.

– Floegel, M., Kasper, J., Perrier, P., & Kell, C. A. (2023). How the conception of control influences our understanding of actions. Nature Reviews Neuroscience, 24(5), 313-329.

– Guenther, F. H., & Hickok, G. (2016). Neural models of motor speech control. In Neurobiology of language (pp. 725-740). Academic Press.

(2) When discussing previous work on alignment during synchronous speech, you may wish to include a recently published paper by Bradshaw et al (2024); this manipulated the acoustics of the accompanist's voice during a synchronous speech task to show interactions between speech motor adaptation and phonetic convergence/alignment.

We thank the reviewer for pointing to this recent and interesting paper. We added the article as reference as follows

“Furthermore, synchronous speech favors the emergence of alignment phenomena, for instance of the fundamental frequency or the syllable onset (Assaneo et al., 2019 ; Bradshaw & McGettigan, 2021 ; Bradshaw et al., 2023; Bradshaw et al., 2024).”

(3) Line 80: "Synchronous speech resembles to a certain extent to delayed auditory feedback tasks"- I think you mean "altered auditory feedback tasks" here.

In the case of synchronous speech it is more about timing than altered speech signals, that is why the comparison is done with delayed and not altered auditory feedback. Nonetheless, we understand the Reviewer’s point and we have now changed the sentence as follows:

“Synchronous speech resembles to a certain extent to delayed/altered auditory feedback tasks”

(4) When discussing superior temporal responses during such altered feedback tasks, you may also want to cite a review paper by Meekings and Scott (2021).

We thank the reviewer for this suggestion, indeed this was a big oversight!

The paper is now quoted in the introduction as follows:

“Previous studies have revealed increased responses in the superior temporal regions compared to normal feedback conditions (Hirano et al., 1997 ; Hashimoto & Sakai, 2003 ; Takaso et al., 2010 ; Ozerk et al., 2022 ; Floegel et al., 2020 ; see Meekings & Scott, 2021 for a review of error-monitoring and feedback control in the STG during speech production).”

Furthermore, we updated the discussion part concerning the speaker-induced suppression phenomenon (see below our response to the point 10).

(5) Line 125: "The parameters and sound adjustment were set using an external low-latency sound card (RME Babyface Pro Fs)". Can you please report the total feedback loop latency in your set-up? Or at the least cite the following paper which reports low latencies with this audio device.

Kim, K. S., Wang, H., & Max, L. (2020). It's About Time: Minimizing Hardware and Software Latencies in Speech Research With Real-Time Auditory Feedback. Journal of Speech, Language, and Hearing Research, 63(8), 25222534. https://doi.org/10.1044/2020_JSLHR-19-00419

We now report the total feedback loop latency (~5ms) and also cite the relevant paper (Kim et al., 2020).

(6) Line 127 "A calibration was made to find a comfortable volume and an optimal balance for both the sound of the participant's own voice, which was fed back through the headphones, and the sound of the stimuli." What do you mean here by an 'optimal balance'? Was the participant's own voice always louder than the VP stimuli? Can you report roughly what you consider to be a comfortable volume in dB?

This point was indeed unlcear. We have now changed as follows:

“A calibration was made to find a comfortable volume and an optimal balance for both the sound of the participant's own voice, which was fed back through the headphones, and the sound of the stimuli. The aim of this procedure was that the patient would subjectively perceive their voice and the VP-voice in equal measure. VP voice was delivered at approximately 70dB.”

(7) Relatedly, did you use any noise masking to mask the air-conducted feedback from their own voice (which would have been slightly out of phase with the feedback through the headphones, depending on your latency)?

Considering the low-latency condition allowed with the sound card (RME Babyface Pro Fs), we did not use noise masking to mask the air-conducted feedback from the self-voice of the patients.

(8) Line 141: "four short sentences were pre-recorded by a woman and a man." Did all participants synchronise with both the man and woman or was the VP gender matched to that of the participant/patient?

We thank the reviewer for this important missing detail. We know changed the text as follows:

“Four stimuli corresponding to four short sentences were pre-recorded by both a female and a male speaker. This allowed to adapt to the natural gender differences in fundamental frequency (i.e. so that the VP gender matched that of the patients). All stimuli were normalised in amplitude.”

(9) Can you clarify what instructions participants were given regarding the VP? That is, were they told that this was a recording or a real live speaker? Were they naïve to the manipulation of the VP's coupling to the participant?

We have now added this information to the task description as follows:

“Participants, comfortably seated in a medical chair, were instructed that they would perform a real-time interactive synchronous speech task with an artificial agent (Virtual Partner, henceforth VP, see next section) that can modulate and adapt to the participant’s speech in real time.”

“The third step was the actual experiment. This was identical to the training but consisted of 24 trials (14s long, speech rate ~3Hz, yielding ~1000 syllables). Importantly, the VP varied its coupling behaviour to the participant. More precisely, for a third of the sequences the VP had a neutral behaviour (close to zero coupling : k = +/- 0.01). For a third it had a moderate coupling, meaning that the VP synchronised more to the participant speech (k = - 0.09). And for the last third of the sequences the VP had a moderate coupling but with a phase shift of pi/2, meaning that it moderately aimed to speak in between the participant syllables (k = + 0.09). The coupling values were empirically determined on the basis of a pilot experiment in order to induce more or less synchronization, but keeping the phase-shifted coupling at a rather implicit level. In other terms, while participants knew that the VP would adapt, they did not necessarily know in which direction the coupling went.”  

(10) The paragraph from line 438 entitled "Secondary auditory regions are more sensitive to coordinative behaviour" includes an interesting discussion of the relation of the current findings to the phenomenon of speech-induced suppression (SIS). However, the authors appear to equate the observed decrease in highfrequency activity as speech coordination increases with the phenomenon of SIS (in lines 456-457), which is quite a speculative leap. I would encourage the authors to temper this discussion by referring to SIS as a potentially related phenomenon, with a need for more experimental work to determine if this is indeed the same phenomenon as the decreases in high-frequency power observed here. I believe that the authors are arguing here for an interpretation of SIS as reflecting internal modelling of sensory input regardless of whether this is self-generated or other-generated; if this is indeed the case, I would ask the authors to be more explicit here that these ideas are not a standard part of the traditional account of SIS, which only includes internal modelling of self-produced sensory feedback.

As stated in the public review, we thank both reviewers for raising thoughtful concerns about our interpretation of the observed neural suppression as related to speaker-induced suppression (SIS). We agree that our study lacks a passive listening condition, which limits direct comparisons to the original SIS effect, traditionally defined as the suppression of neural responses to self-produced speech compared to externally-generated speech (Meekings & Scott, 2021).

In response, we have reconsidered our terminology and interpretation. In the revised discussion, we refer to our findings as a "SIS-related phenomenon specific to the synchronous speech context." Unlike classic SIS paradigms, our interactive task involves simultaneous monitoring of self- and externally-generated speech, introducing additional attentional and coordinative demands.

The revised discussion also incorporates findings by Ozker et al. (2024, 2022), which link SIS and speech monitoring, suggesting that suppressing responses to self-generated speech facilitates error detection. We propose that the decrease in high-frequency activity (HFa) as verbal coordination increases reflects reduced error signals due to closer alignment between perceived and produced speech. Conversely, HFa increases with reduced coordination may signify greater prediction error.

Additionally, we relate our findings to the "rubber voice" effect (Zheng et al., 2011; Lind et al., 2014; Franken et al., 2021), where temporally and phonetically congruent external speech can be perceived as self-generated. We speculate that this may occur in synchronous speech tasks when the participant's and VP's speech signals closely align. However, this interpretation remains speculative, as no subjective reports were collected to confirm this perception. Future studies could include participant questionnaires to validate this effect and relate subjective experience to neural measures of synchronization.

Overall, our findings extend the study of SIS to dynamic, interactive contexts and contribute to understanding internal forward models of speech production in more naturalistic scenarios.

We have now added these points to the discussion as follows:

“The observed negative correlation between verbal coordination and high-frequency activity (HFa) in STG BA22 suggests a suppression of neural responses as the degree of synchrony increases. This result aligns with findings on speaker-induced suppression (SIS), where neural activity in auditory cortex decreases during self-generated speech compared to externally-generated speech (Meekings & Scott, 2021; Niziolek et al., 2013). However, our paradigm differs from traditional SIS studies in two critical ways: (1) the speaker's own voice is always present and predictable from the forward model, and (2) no passive listening condition was included. Therefore, our findings cannot be directly equated with the original SIS effect.

Instead, we propose that the suppression observed here reflects a SIS-related phenomenon specific to the synchronous speech context. Synchronous speech requires simultaneous monitoring of self- and externally generated speech, a task that is both attentionally demanding and coordinative. This aligns with evidence from Ozker et al. (2024, 2022), showing that the same neural populations in STG exhibit SIS and heightened responses to feedback perturbations. These findings suggest that SIS and speech monitoring are related processes, where suppressing responses to self-generated speech facilitates error detection.

In our study, suppression of HFa as coordination increases may reflect reduced prediction errors due to closer alignment between perceived and produced speech signals. Conversely, increased HFa during poor coordination may signify greater mismatch, consistent with prediction error theories (Houde & Nagarajan, 2011; Friston et al., 2020).”

(11) Within this section, you also speculate in line 460 that "Moreover, when the two speech signals come close enough in time, the patient possibly perceives them as its own voice." I would recommend citing studies on the 'rubber voice' effect to back up this claim (e.g. Franken et al., 2021; Lind et al., 2014; Zheng et al., 2011).

We are grateful to the Reviewer for this interesting suggestion. Directly following the previous comment, the section now states:

“Furthermore, when self- and externally-generated speech signals are temporally and phonetically congruent, participants may perceive external speech as their own. This echoes the "rubber voice" effect, where external speech resembling self-produced feedback is perceived as self-generated (Zheng et al., 2011; Lind et al., 2014; Franken et al., 2021). While this interpretation remains speculative, future studies could incorporate subjective reports to investigate this phenomenon in more detail.”

(12) As noted in my public review, since your methods are correlational, you need to be careful about inferring the causal role of any brain areas in supporting a specific aspect of functioning e.g. line 501-504: "By contrast, in the inferior frontal gyrus, the coupling in the high-frequency activity is strongest with the input-output phase difference (input of the VP - output of the speaker), a metric that reflects the amount of error in the internal computation to reach optimal coordination, which indicates that this region optimises the predictive and coordinative behaviour required by the task." I would argue that the latter part of this sentence is a conclusion that, although consistent with, goes beyond the current data in this study, and thus needs tempering.

We agree with the Reviewer and changed the sentence as follows:

“By contrast, in the inferior frontal gyrus, the coupling in the high-frequency activity is strongest with the inputoutput phase difference (input of the VP - output of the speaker), a metric that could possibly reflect the amount of error in the internal computation to reach optimal coordination. This indicates that this region could have an implication in the optimisation of the predictive and coordinative behaviour required by the task.”

I think it’s a great question of should we still allow people to use instagram if it’s bad for us. If Meta’s proves that instagram can be bad to teenage girls. Shouldn’t we find ways to let it be beneficial instead of just ban it. I would considered that as a method to stop instagram from taking away customers from Facebook. Also the study indicated that in general this could be harmful. But there’s people benefits from this platform. If unfair to close it just because this may be bad for the tonnage girls mental health. Finding ways to make it beneficial to mental health would be the right solution.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2025-02888

Corresponding author(s): Christian, Fankhauser

General Statements

We were pleased to see that the three reviewers found our work interesting and provided supportive and constructive comments.

Our answers to their comments and/or how we propose to address them in a revised manuscript are included in bold.

1. Description of the planned revisions

Reviewer #1 (Evidence, reproducibility and clarity (Required)):

Summary: Plant systems sense shading by neighbors via the phytochrome signaling system. In the shade, PHYTOCHROME-INTERACTING FACTORS (PIFs) accumulate and are responsible for transcriptional reprogramming that enable plants to mobilize the "shade-avoidance response". Here, the authors have sought to examine the role of chromatin in modulating this response, specifically by examining whether "open" or "closed" chromatin regions spanning PIF target genes might explain the transcriptional output of these genes. They used a combination of ATAC-seq/CoP-qPCR (to detect open regions of chromatin), ChIP (to assay PIF binding) and RNA-seq (to measure transcript abundance) to understand how these processes may be mechanistically linked in Arabidopsis wild-type and pif mutant lines. They found that some chromatin accessibility changes do occur after LRFR (shade) treatment (32 regions after 1h and 61 after 25 h). While some of these overlap with PIF-binding sites, the authors found no correlation between open chromatin states and high levels of transcription. Because auxin is an important component of the shade-avoidance response and has been shown to control chromatin accessibility in other contexts, they examined whether auxin might be required for opening these regions of chromatin. They find that in an auxin biosynthesis mutant, there is a small subset of PIF target genes whose chromatin accessibility seems altered relative to the wild-type. Likewise, they found that chromatin accessibility for certain PIF targets is altered in phyB and pif mutant and propose that PIFs are necessary for changing the accessibility of chromatin in these genes. The authors thus propose that PIF occupancy of already open regions, rather than increased accessibility, underly the increase in transcript of abundance of these target genes in response to shade.

Major comments: *• I find that the data generally support the hypothesis presented in the manuscript that chromatin accessibility alone does not predict transcription of PIF target genes in the shade. That said, I think that a paragraph from the discussion (lines 321-332) would benefit from some careful rephrasing. I think it is perfectly reasonable to propose that PIF occupancy is more predictive of shade-induced transcriptional output than chromatin accessibility, but I think that calling PIF occupancy "the key drivers" (line 323) or "the main driving force" (line 76) risks ignoring the observation that levels of PIF occupancy specifically do not predict expression levels of PIF target genes (Pfeiffer et al., 2014, Mol Plant). For PIL1 and HFR1, the authors have shown that PIF promoter occupancy and transcript levels are correlated, but the central finding of Pfeiffer et al. was that this pattern does not apply to the majority of PIF direct target genes. Finding factors (i.e. histone marks) that convert PIF-binding information into transcriptional output appears to have been the impetus for the experiments devised in Willige et al., 2021 and Calderon et al., 2022. It is great that the authors have outlined in the discussion that there are a number of factors that modulate PIF transcriptional activating activity but I think that the emphasis on PIF-binding explaining transcript abundance should be moderated in the text. *

We appreciate the reviewers’ comments and will address it by introducing appropriate changes to the discussion. One element that should be pointed out is that the study of Willige et al., 2021 allows us to look at sites where PIF7 is recruited in response to the shade stimulus (a low R/FR treatment) and relate this to higher transcript abundance of the nearby genes. The study of Pfeiffer et al., 2014 which analyses PIF ChIP studies from several labs does not include this dynamic view of PIF recruitment in response to a stimulus. For example, this study re-analyses data from our lab, Hornitschek et al., 2012, in which we did PIF5 ChIP in low R/FR, but we did not compare that to high R/FR to enable an analysis of sites where we see recruitment of PIF5 in response to a shade cue. In the revised manuscript we will also include a new figure comparing PIF7 recruitment and changes in gene expression at direct PIF target genes.

• I think that the hypothesis could be further supported by incorporating the previously published ChIP-seq data on PIF1, PIF3 and PIF5 binding. Given these data are published/publicly available, I think it would be helpful to note which of the 72 DARs are bound by PIF1, PIF3 and/or PIF5. Especially so given that PIF5 (Lorrain et al., 2008, Plant J) and PIF1/PIF3 (Leivar et al., 2012, Plant Cell) contribute at least in some capacity to transcriptional regulation in response to shade. At the very least, it might help explain some of the observed increases in nucleosome accessibility observed for genes that don't have PIF4 or PIF7-binding.* This is a thoughtful suggestion. Our choice to focus on PIF7 target genes is dictated by two reasons. First, the finding that amongst all tested PIFs, PIF7 is the major contributor to the control of low R/FR (neighbor proximity) induced responses in seedlings (e.g. Li et al., 2012; de Wit et al., 2016; Willige et al., 2021). In addition, the PIF7 ChIP-seq and gene expression data from the Willige et al., 2021 paper was obtained using growth conditions very similar to the ones we used, hence allowing us to compare it to our data. As the reviewer suggests, other PIFs also contribute to the low R/FR response and hence looking at ChIP-seq for those PIFs in publicly available data is also informative. One limitation of this data is that ChIP-seq was not always done in seedlings grown in conditions directly comparable to the conditions we used (except for PIF5, see above). Nevertheless, we have performed this analysis with the available data suggested by the reviewer and intend to include the results in the revised version of the manuscript, presumably updated Figure 4B.

• In the manuscript, there are several instances where separate col-0 (wild type) controls have been used for identical experiments. Specifically, qPCR (Fig 3C, Fig S7C/D and Fig S8C/D), CoP-qPCR (Fig 5B/5C and Fig S8E/F) and hypocotyl measurements (Fig S7A/B and Fig S8A/B). In the cases of the hypocotyl measurements, there appear to be hardly any differences between col-0 controls indicating the measurements can be confidently compared between genotypes.

We appreciate this comment but to be comprehensive, we like to include a Col-0 control for each experiment (whenever possible) and hence also include the data when available.

• In some cases of qPCR and CoP-qPCR experiments however, the differences in values obtained from col-0 samples that underwent identical experimental treatments appear to vary significantly. In Figure 3C for example, the overall trend for PIL1 expression in col-0 is the same (e.g. HRFR levels are low, LRFR1 levels are much higher and LRFR25 levels drop down to some intermediate level) but the expression levels themselves appear to differ almost two-fold for the LRFR 1h timepoint (~110 on the left panel vs ~60 for the right panel). Given the size of the error bars, it appears that these data represent the mean from only one biological replicate. PIL1 expression levels at LRFR 1h as reported in Fig S7C and D also show similar ~2-fold differences. __This is a good comment. Having looked at PIL1 gene induction by low R/FR in dozens of similar experiments made us realize that indeed while the PIL1 induction is always massive, the extent is somewhat variable. Based on the data that we have (including from RNA-seq) we are convinced that this is due to the very low level of expression of PIL1 in high R/FR conditions. Given that induction by low R/FR is expressed as fold increase relative to baseline high R/FR expression, small changes in the lowly expressed PIL1* in high R/FR leads to seemingly significant differences in its induction by low R/FR across experiments.__

All qPCR data is represented by three biological replicates, and the variation between them per experiment is low, which is reflected in the size of the SD error bars. Data on technical and biological replicates in each panel will be clearly indicated in the revised figure legends.

• I would recommend that the authors explicitly describe the number of biological replicates used for each experiment in the methods section. If indeed these experiments were only performed once, I think the authors should be very careful in the language used in describing their conclusions and in assigning statistical significance. One possibility that could also be helpful would be normalizing LRFR 1h and LRFR 25h values to HRFR values and plotting these data somewhere in the supplemental data. If, for example, the relative levels of PIL1 are different between replicates but the fold-induction between HRFR and LRFR 1h are the same, this would at least allay any concerns that the experimental treatments were not the same. I understand that doing so precludes comparison between genotypes, but I do think it's important to show that at least the control data are comparable between experiments.

* All qPCR and CoP-qPCR experiments have been performed with three 3 biological replicates as described in Materials and Methods section, and these are represented in the Figures. Relative gene expression in the qPCR experiments was normalized to two housekeeping genes YLS8 and UBC21 and afterwards to one biological replicate of Col-0 control in HRFR. As indicated for the previous comment information about replicates will be included in the updated figure legends.

• Similarly, for the CoP-qPCR experiments presented in Fig 5B and 5C, the col-0 values for region P2 between Fig 5B and 5C shows that while HRFR and LRFR 1h look comparable, the values presented for LRFR 25h are quite different.

* This comment of the reviewer prompted us to propose a different way of representing the data that is clearer (new Figure 5B and 5C). We believe that this facilitates the comparison between the genotypes. Enrichment over the input was calculated for the chromatin accessibility of each region. Chromatin accessibility was further normalized against two open control regions on the promoters of ACT2 (AT3G18780, region chr3:6474579: 6474676) and RNA polymerase II transcription elongation factor (AT1G71080 region chr1:26811833:26811945). The difference between previous representation is that the regions are not additionally subtracted to Col-0 in HRFR. We will update the Materials and Methods and figure legend sections with this information.

Minor comments: • Presentation of Supplemental Figure 7A/7B and Supplemental Figure 8A/8B could be changed to make the data more clear (i.e. side-by-side rather than superimposed).

We propose changing the presentation of the hypocotyl length data to show the values for days side-by-side as the Reviewer suggests.

• I think that the paragraph introducing auxin (lines 25-37) could be reduced to 1-2 sentences and merged into a separate introductory paragraph given that the SAV3 work makes up a relatively minor component of the manuscript.

* We agree with the reviewer and will reduce the paragraph about auxin and merge it with the previous paragraph about transcription.

• • For Figure 3A, I would strongly encourage the authors to show some of the raw western blot data for PIF4, PIF5 and PIF7 protein abundance (and loading control), not just the normalized values. This could be put into supplemental data, but I think it should accompany the manuscript.

* We agree that presenting the raw data that was used for quantification is important. We will include the western blots used for quantifying PIF4, PIF5 and PIF7 protein abundance (and loading control DET3). This information will presumably be included to the Supplementary Figure 3C (figure number to be confirmed once we decide on all new data to be presented).

• Lines 145-147 "we observed a strong correlation between PIF4 protein levels (Figure 3A) and PIL1 promoter occupancy (Figure 3B), and a similar behavior was seen with PIF7 (Figure 3B)." According to Fig 3A, there is no statistically significant increase in PIF7 abundance after 1h shade. There is an apparent increase in PIF7 promoter occupancy, but the variation appears too large for it to be statistically significant. I agree that qualitatively there is a correlation for PIF4 but I think the description of the behavior of PIF7 should be rephrased.

* __As suggested by the reviewer, we will rephrase this paragraph to more accurately account for our data and also what was reported by others (e.g. Willige et al, 2021, in Li et al, 2012) regarding the regulation PIF7 levels and phosphorylation in response to a low R/FR treatment. __

• There appear to be issues in the coloring of the labels (light blue dots vs dark blue dots) for the PIF7 panels of Fig 3B and Supplemental Fig 3B.*

We thank the reviewer for pointing this out. This will be clarified by appropriate changes in the figure to avoid confusion in the revised version of Figure 3B.

Reviewer #1 (Significance (Required)):

This authors here have sought to examine the possibility that the transcriptional responses to shade mediated by the phy-PIF system might involve large-scale opening or closing of chromatin regions. This is an important and unanswered question in the field despite several studies that have looked at the role of histone variants (H2A.Z) and modifications (H3K4me3 and H3K9ac) in modulating PIF transcriptional activating activity. The authors have shown that, at least in the case of the transcriptional response to shade mediated by PIF7 (and to an extent PIF4), large-scale changes in chromatin accessibility are not occurring in response to shade treatment.

The results presented in this study support the hypothesis that large-scale changes in chromatin accessibility may have already occurred before plants see shade. This opens the possibility that perhaps the initial perception of light by etiolated (dark-grown seedlings) might trigger changes in chromatin accessibility, opening up chromatin in regions encoding "shade-specific" genes and/or closing chromatin in regions encoding "dark-specific" genes.

The audience for this particular manuscript encompasses a fairly broad group of biologists interested in understanding how environmental stimuli can trigger changes in chromatin reorganization and transcription. The results here are important in that they rule out chromatin accessibility changes as underlying the changes in transcription between the short-term and long-term shade responses. They also reveal that there are a few cases in which chromatin accessibility does change in a statistically-significant manner in response to shade. These regions, and the molecular players which regulate their accessibility, merit further exploration.

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

The study by Paulisic et al. explores the variations in chromatin accessibility landscape induced by plant exposure to light with low red/far-red ratios (LRFR), which mimicks neighbor shade perception. The authors further compare these changes with the genome association of PIF4 and PIF7 transcription factors - two major actors of gene expression regulation in response to LRFR. While this is not highlighted in the main text, the analyses of chromatin accessibility are performed on INTACT-mediated nucleus sorting, presumably to ensure proper and clean isolation of nuclei.

Major comments

• Why is the experimental setup exposing plants to darkness overnight? Does this affect the response to LRFR, by a kind of reset of phytochrome signaling? I guess this choice was made to maintain a strong circadian rhythm. Yet, given that PIF genes are clock-regulated, I am afraid that this choice complicates data interpretation concerning the specific effects of LRFR exposure.

There appears to be some confusion which prompts us to better explain our protocol both by changing Figure 1A (that outlines the experimental conditions) and in the text.

Seedlings are grown in long day conditions because this is more physiologically relevant than growing them in constant light, which is a rather unnatural condition.

The reviewer is correct that PIF transcription is under circadian control and the shade avoidance response is gated by the circadian clock (e.g. Salter et al., 2003). To prevent conflating circadian and light quality effects, all samples that are compared are harvested at the same ZT (circadian time – hours after dawn). This allows us to focus our analysis on light quality effects specifically. We are therefore convinced that our protocol does not complicate the interpretation of the LRFR effects reported here.

• As a result of this setup, the 1h exposure to LRFR immediately follows HRFR while the 3h final LRFR exposure of the « 25h LRFR » samples immediately follows a long period of darkness. Can this explain why in several instances (e.g., at the ATHB2 gene) 1h LRFR seems to have stronger effects than 25h LRFR on chromatin accessibility?* Please check the explanation above. Both samples are harvested at the same ZT (ZT3, meaning 3 hours after dawn). The 1h LRFR seedlings went through the night, had 2 hours of HRFR then 1h of LRFR. The 25h are harvested at the very same ZT, meaning 3h after dawn. Importantly, the HRFR control was also harvested at ZT3, meaning 3h after dawn. As indicated above this protocol allows us to focus on the light quality effects by comparing samples that are all harvested at the same ZT.

We expect that the changes in Fig. 1A and associated text changes will clarify this issue.

• Lane 42 cites the work by Calderon et al 2022 as « Transcript levels of these genes increase before the H3K4me3 levels, implying that H3K4me3 increases as a consequence of active transcription ». Despite this previous study being reviewed and published, such a strong conclusion should be taken cautiously, and I disagree with it. The study by Calderon et al compares RNA-seq with ChIP-seq data, two methodologies with very different sensitivity, especially when employing bulk cells/whole seedlings as starting materials. For example, a gene strongly induced in a few cells will give a good Log2FC in RNA-seq data analysis (as new transcripts are produced after a low level of transcripts before shade) but, even though its chromatin variations would follow the same temporality or would even precede gene induction, this would be invisible in bulk ChIP-seq data analysis (which averages the signal of all cells together). I understand the rationale for relying on the conclusions made in an excellent lab with strong expertise in light signaling, but I recommend being cautious when relying on these conclusions to interpret new data.* We agree with this comment, and we will change the text to reflect this.

• The problem is that the same issue holds true when comparing ATAC-seq and RNA-seq data. ATAC signals reflect average levels over all cells while RNA-seq data can be influenced by a few cell highly expressing a given gene. Even though authors carefully sorted nuclei using an INTACT approach, this should be discussed, in particular when gene clusters (such as cluster C-D) show no match between chromatin accessibility and transcript level variations. In this regard, is PIF7 expressed in many cells or a small niche of cells upon LRFR exposure? The conclusions on its role in chromatin accessibility, analyzed here as mean levels of many different seedling cells, could be affected by PIF7 activity pattern (e.g., at lane 293). __This is a good comment. PIF7 is expressed in the cotyledons and leaves in LD conditions (Kidokoro et al, 2009, Galvao et al, 2019), and few available scRNA-seq datasets indicate an enrichment of PIF7 in the epidermis (Kim et al, 2021, Lopez-Anido et al, 2021). LRFR exposure only mildly represses PIF7* expression as seen in Figure 3A and also in our bulk RNA-seq study (Table S4). We will discuss this potential limitation to our study in a revised version of the manuscript.__

• Lane 89, the conclusion linking DNA methylation and DNA accessibility is unclear to me, this may be rephrased. Also, it should be noted that in gene-rich regions, most DNA methylation is located along the body of moderately to highly transcribing genes (gene-body methylation) while promoters of active and inactive genes are most frequently un-methylated.* We will rephrase to better reflect the presence or absence of DNA methylation on promoter regions of shade regulated genes that contain accessible sites.

• Figure 3B shows a few ChIP-qPCR results with important conclusions. Why not sequencing the ChIPped DNA to obtain a genome-wide view of the PIF4-PIF7 relationships at chromatin, and also consequently a more robust genome-wide normalization?

* Several studies have shown that in the conditions that we studied here: transfer of seedlings from high R/FR (simulated sun) to low R/FR (neighbor proximity), amongst all PIFs, PIF7 is the one that plays the most dominant function (e.g. Li et al., 2012; de Wit et al., 2016; Willige et al., 2021). PIF4 and PIF5 also contribute but to a lesser extent. Given that Willige et al., 2021 did extensive ChIP-seq studies for PIF7 using similar conditions to the ones we used, we decided to rely on their data (that we re-analyzed), rather than performing our own PIF7 ChIP-seq analysis. While also performing a ChIP-seq analysis for PIF4 in similar conditions might be useful (this data is not available as far as we know), we are not convinced that doing that experiment would substantially modify the message. In the revised version we will also include analysis of the data from Pfeiffer et al., 2014, which comprises a ChIP-seq. dataset for PIF5 (the closest paralog of PIF4) initially performed by Hornitschek et al., in 2012 in low R/FR conditions (see comment to reviewer 1 above). For new ChIP-seq, we would have to make this experiment from scratch with substantially more material than what we used for the targeted ChIP-qPCR analyses. We thus do not feel that such an investment (time and money) is warranted.

• • Given the known functional interaction between PIF7 and INO80, it would be relevant to test whether changes in chromatin accessibility at ATHB2 and other genes are affected in ino80 mutant seedlings. __We agree with the reviewer that this is potentially an interesting experiment. This will allow us to determine whether the nucleosome histone composition has an influence on nucleosome positioning at selected shade-regulated genes (e.g. ATHB2). We note that according to available data, the effect of INO80 would be expected once PIF7 started transcribing shade-induced genes. We therefore propose comparing the WT with an ino80 mutant for their seedling growth phenotype, expression of selected shade marker gene (e.g. ATHB2*) and chromatin accessibility before (high R/FR) and after low R/FR treatment at selected shade marker genes. This will allow us to determine whether INO80 influences chromatin accessibility prior to a low R/FR treatment and/or once the treatment started. Our plan is to include this data in a revised version of the manuscript. __
• On the same line, it would be interesting to test whether PIF7 target regions with pre-existing accessible chromatin would exist in ino80 mutant plants. In other words, testing a model in which chromatin remodeling by INO80 defines accessibility under HRFR to enable rapid PIF recruitment and DNA binding upon LRFR exposure.*

See our answer just above.

Minor comments

*• In Figure 1C, it seems that PIF7 target genes do not match the set of LRFR-downregulated genes (even less than at random). Why not exclude these 4 genes from the analyses? *

This is correct. There are indeed only 4 downregulated PIF7 target genes as we define them. Removing these genes from the analyses does not change our interpretation of the data and hence for completeness we propose keeping them in a revised version of the manuscript

• Figure 3A shows the quantification of protein blots, but I did not find the corresponding images. These should be shown in the figure or as a supplementary figure with proper controls.

* We will include the raw Westen blots used for quantification of PIF4, PIF5 and PIF7 in the revised version of the manuscript

• • Lane 102, it is unclear why PIF7 target genes were defined as the -3kb/TSS domains while Arabidopsis intergenic regions are on average much shorter. Gene regulatory regions, or promoters, are typically called within -1kb/TSS regions to avoid annotating a ChIP peak to the upstream gene or TE. A better proxy of PIF7 typical binding sites in gene regulatory regions could be determined by analysing the mean distance between PIF7 peak coordinates and the closest TSS. Typically, a gene meta-plot would give this information. __We agree that the majority of PIF7 binding peaks are close to the 5’ of the TSS based on the PIF7 binding distribution meta-plot. But several known PIF binding sites are actually further upstream than 1kb 5’ of the TSS (e.g. ATHB2 and HFR1). However, we re-analyzed the data using your suggestion with -2kb/TSS and -1kb/TSS and while the number of target genes is reduced, it does not change our conclusions about PIF7 binding sites being located on accessible chromatin regions. Importantly, some well characterized LRFR induced genes such as HFR1* would not be annotated correctly if only peaks closest to the gene TSS were taken into account, without flanking genes. In this case only the neighboring AT1G02350 would be annotated, hence missing some important PIF7 target genes. Taking this into consideration we will not modify this part of the analysis in a revised manuscript.__
• Figure 4B, what's represented in the ATAC-seq heatmap: does a positive z-score represent high accessibility?*

On the ATAC-seq heatmap we have represented z-scores of the average CPM (counts per million) for accessible chromatin regions. Z-scores are calculated by subtracting the average CPM from the median of averaged CPMs for each accessible chromatin region and then divided by the standard deviation (SD) of those averaged CPMs across all groups per accessible region (in our case a group is an average of three biological replicates for either HRFR, 1h or 25h of LRFR). In that sense, z-score indicates a change in accessibility, where higher z-score indicates opening of the region and lower z-score indicates a region becoming more closed when compared among the three light treatments (HRFR, 1h or 25h of LRFR). We will make sure that this is clear in the revised manuscript. Reviewer #2 (Significance (Required)):

Contradicting the naive hypothesis that PIFs may target shade-inducible genes to « open » chromatin of shade-inducible genes with the help of chromatin remodelers, such as INO80, the study highlights that PIF7 typically associates with pre-existing accessible chromatin states. Thus, even though this is not stated, results from this study indicate that PIF7 is not a pioneer transcription factor. The data seem very robust, and while some conclusions need clarification, it should be of great interest to the community of scientists studying plant light signaling and shade responses.

Reviewer #3 (Evidence, reproducibility and clarity (Required)):

In their manuscript, Paulisic et al. investigate whether the transcriptional response of Arabidopsis seedlings to shade depends on chromatin accessibility, with a specific focus on PIF7-regulated genes. To this end, they perform ATAC-seq and RNA-seq, along with other experiments, on seedlings exposed to short and long shade and correlate the results with previously reported PIF7 and PIF4 ChIP-seq data. Based on their findings, they propose that shade-mediated transcriptional regulation may not require extensive remodeling of DNA accessibility. Specifically, they suggest that the open chromatin conformation allows PIFs to easily access and recognize their binding motifs, rapidly initiating gene expression in response to shade. This transcriptional response primarily depends on a transient increase in PIF stability and gene occupancy, with changes in chromatin accessibility occurring in only a small number of genes.

Major comments: * • I have one issue that, in my opinion, requires more attention. To define the PIF7 target genes, which were later used to estimate whether PIF7 binds to open or closed chromatin and affects DNA accessibility after its binding, the authors compared the 4h LRFR data point from Willige et al. (2021) ChIP-seq with their 1h RNA-seq data point. This comparison might have missed early genes where PIF7 binds before the 1h time point but is no longer present on DNA at 4h. I understand the decision to choose the 4h Willige et al. ChIP-seq data point, performed under LD conditions, as it matches the data in this study, rather than the 5min-30min data points, which were conducted in constant light. However, if possible, it would be interesting to also compare the RNA-seq data with the early PIF7 binding genes to assess how many additional PIF7 target genes could be identified based on that comparison and whether this might alter the conclusions. If the authors do not agree with this point, it should at least be emphasized that the ChIP-seq data and the RNA-seq/ATAC-seq data were performed under different LRFR conditions (R/FR 0.6 vs. 0.1), which may lead to the misidentification of PIF7 target genes in the manuscript.*

1) This is an interesting suggestion, we therefore reanalyzed 5, 10 and 30 min ChIP-seq timepoints from Willige et al, 2021 and compared them to 4h of LRFR (ZT4). We have crossed these lists of potential PIF7 targets with our 1h LRFR PIF457 dependent genes based on our RNA-seq. While some PIF7 targets appear only in early time points 5-10 min of LRFR exposure, overall, the number and composition of PIF7 target genes is rather constant across these timepoints. We propose to include these additional analyses in a revised version of the manuscript as a supplemental figure. However, these additional analyses do not influence our general conclusions.

2) The comment regarding the R/FR ratio is important. We will point this out although the conditions used by Willige et al., 2021 and the ones we used are similar, they are not exactly the same in terms of R/FR ratio. Importantly, in both studies the early transcriptional response largely depends on the same PIFs, many of the same response genes are induced (e.g. PIL1, AtHB2, HFR1, YUC8, YUC9 and many others) and the physiological response (hypocotyl elongation) is similar. This shows that this low R/FR response yields robust responses.

Minor comments: • In Fig. 1D, please describe the meaning of the blue shaded areas and the blue lines under the ATAC-seq peaks, as they do not always correlate.

The shaded areas and the bars define the extension of the ATAC-seq accessible chromatin peaks. We will add the meaning of the shaded areas and the blue bars in the Figure legend and correct the colors in a revised manuscript

• In Fig. 1E, it could be helpful to note that the 257 peaks in the right bar correspond to the peaks associated with the 177 genes in the left bar.* We will update Figure 1E and Figure legends for better understanding as the Reviewer suggested.

• In lines 116, 119, and 122, I believe it should read "Fig. 2" instead of "Fig. 2A."* We thank the Reviewer for noticing the error that we will correct.

• Lines 138-139: "PIF7 total protein levels were overall more stable, and only a mild and non-significant increase of PIF7 levels was seen at 1 h of LRFR." Since PIF7 usually appears as two bands in HRFR and only one band in LRFR, how was the protein level of PIF7 quantified in Fig. 3A? Additionally, I was wondering about the authors' thoughts on the discrepancy with Willige et al. (2021, Extended Data Fig. 1d), where PIF7 abundance seems to increased after 30 min and 2 h of LRFR.* PIF7 protein levels were quantified by considering both the upper and the lower band in HRFR (total PIF7) and normalizing its levels to DET3 loading control. We still observe an increase in the total PIF7 protein levels at 1h of LRFR, however this change was not statistically significant in these experiments. In our conditions as in Willige et al, 2021, the increase in PIF7 protein levels to short term shade seems consistent as is the pronounced shift or disappearance of the upper band (phosphorylated form) on the Western blots (raw data will be available in the revised manuscript). We will introduce text changes referring to the phosphorylation status of PIF7 in our conditions.

• Line 150: "... many early PIF target genes (Figure 3C)." Since only PIL1 is shown in Fig. 3C, I would recommend revising this sentence. Alternatively, the data could be presented, as in Fig. 2, for all the PIF7 target genes with transient expression patterns.

* We will introduce changes in the text to reflect that we only show PIL1 in the main Figure 3C.

• Line 204: I'm not sure if Supplementary Fig. 7C-D is correct here. If it is, could the order of the figures be changed so that Supplementary Fig. 7C-D becomes Supplementary Fig. 7A-B?*

The order of the panels A-B in the Supplementary Figure 7 follows the order of the text in the manuscript and is mentioned before panels C-D. It refers to the sentence “Overexpression of phyB resulted in a strong repression of hypocotyl elongation in both HRFR and LRFR, while the absence of phyB promoted hypocotyl elongation (Supplementary Figure 7A-B).”

• • Line 208: "In all three cases...". Please clarify what the three cases refer to. __We will change the text to more explicitly refer to the differentially accessible regions (DARs) of the genes ATHB2 and HFR1* shown in Figure 5A.__
• Line 231: Should Fig. 5C also be cited here in addition to Supplementary Fig. 7?* We will add the reference to Figure 5C that was missing.

*• In Supplementary Table 3, more information is needed. For example, it could mention: "This data is presented in Fig. 3 and is based on datasets from ChIP-seq, RNA-seq, etc."

*

The table will be updated with more information as suggested by the Reviewer.

• In the figure legend of Fig. 4B, please check the use of "( )".*

We will correct the error and include the references inside the parenthesis.

Reviewer #3 (Significance (Required)):

Paulisic et al. present novel discoveries in the field of light signaling and shade avoidance. Their findings extend our understanding of how DNA organization, prior to shade, affects PIF binding and how PIF binding remodels DNA accessibility. The data presented support the conclusions well and are backed by sufficient experimental evidence.

2. Description of the revisions that have already been incorporated in the transferred manuscript

The manuscript has not been modified yet.

3. Description of analyses that authors prefer not to carry out

• *

Reviewer 2 asked for new ChIP-seq analyses for PIF7 and PIF4. For reasons that we outlined above, we believe that such analyses are not required, and we currently do not intend performing these experiments.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary:

Plant systems sense shading by neighbors via the phytochrome signaling system. In the shade, PHYTOCHROME-INTERACTING FACTORS (PIFs) accumulate and are responsible for transcriptional reprogramming that enable plants to mobilize the "shade-avoidance response". Here, the authors have sought to examine the role of chromatin in modulating this response, specifically by examining whether "open" or "closed" chromatin regions spanning PIF target genes might explain the transcriptional output of these genes. They used a combination of ATAC-seq/CoP-qPCR (to detect open regions of chromatin), ChIP (to assay PIF binding) and RNA-seq (to measure transcript abundance) to understand how these processes may be mechanistically linked in Arabidopsis wild-type and pif mutant lines. They found that some chromatin accessibility changes do occur after LRFR (shade) treatment (32 regions after 1h and 61 after 25 h). While some of these overlap with PIF-binding sites, the authors found no correlation between open chromatin states and high levels of transcription. Because auxin is an important component of the shade-avoidance response and has been shown to control chromatin accessibility in other contexts, they examined whether auxin might be required for opening these regions of chromatin. They find that in an auxin biosynthesis mutant, there is a small subset of PIF target genes whose chromatin accessibility seems altered relative to the wild-type. Likewise, they found that chromatin accessibility for certain PIF targets is altered in phyB and pif mutant and propose that PIFs are necessary for changing the accessibility of chromatin in these genes. The authors thus propose that PIF occupancy of already open regions, rather than increased accessibility, underly the increase in transcript of abundance of these target genes in response to shade.

Major comments:

I find that the data generally support the hypothesis presented in the manuscript that chromatin accessibility alone does not predict transcription of PIF target genes in the shade. That said, I think that a paragraph from the discussion (lines 321-332) would benefit from some careful rephrasing. I think it is perfectly reasonable to propose that PIF occupancy is more predictive of shade-induced transcriptional output than chromatin accessibility, but I think that calling PIF occupancy "the key drivers" (line 323) or "the main driving force" (line 76) risks ignoring the observation that levels of PIF occupancy specifically do not predict expression levels of PIF target genes (Pfeiffer et al., 2014, Mol Plant). For PIL1 and HFR1, the authors have shown that PIF promoter occupancy and transcript levels are correlated, but the central finding of Pfeiffer et al. was that this pattern does not apply to the majority of PIF direct target genes. Finding factors (i.e. histone marks) that convert PIF-binding information into transcriptional output appears to have been the impetus for the experiments devised in Willige et al., 2021 and Calderon et al., 2022. It is great that the authors have outlined in the discussion that there are a number of factors that modulate PIF transcriptional activating activity but I think that the emphasis on PIF-binding explaining transcript abundance should be moderated in the text.

I think that the hypothesis could be further supported by incorporating the previously published ChIP-seq data on PIF1, PIF3 and PIF5 binding. Given these data are published/publicly available, I think it would be helpful to note which of the 72 DARs are bound by PIF1, PIF3 and/or PIF5. Especially so given that PIF5 (Lorrain et al., 2008, Plant J) and PIF1/PIF3 (Leivar et al., 2012, Plant Cell) contribute at least in some capacity to transcriptional regulation in response to shade. At the very least, it might help explain some of the observed increases in nucleosome accessibility observed for genes that don't have PIF4 or PIF7-binding.

In the manuscript, there are several instances where separate col-0 (wild type) controls have been used for identical experiments. Specifically, qPCR (Fig 3C, Fig S7C/D and Fig S8C/D), CoP-qPCR (Fig 5B/5C and Fig S8E/F) and hypocotyl measurements (Fig S7A/B and Fig S8A/B). In the cases of the hypocotyl measurements, there appear to be hardly any differences between col-0 controls indicating the measurements can be confidently compared between genotypes.

In some cases of qPCR and CoP-qPCR experiments however, the differences in values obtained from col-0 samples that underwent identical experimental treatments appear to vary significantly. In Figure 3C for example, the overall trend for PIL1 expression in col-0 is the same (e.g. HRFR levels are low, LRFR1 levels are much higher and LRFR25 levels drop down to some intermediate level) but the expression levels themselves appear to differ almost two-fold for the LRFR 1h timepoint (~110 on the left panel vs ~60 for the right panel). Given the size of the error bars, it appears that these data represent the mean from only one biological replicate. PIL1 expression levels at LRFR 1h as reported in Fig S7C and D also show similar ~2-fold differences.

I would recommend that the authors explicitly describe the number of biological replicates used for each experiment in the methods section. If indeed these experiments were only performed once, I think the authors should be very careful in the language used in describing their conclusions and in assigning statistical significance. One possibility that could also be helpful would be normalizing LRFR 1h and LRFR 25h values to HRFR values and plotting these data somewhere in the supplemental data. If, for example, the relative levels of PIL1 are different between replicates but the fold-induction between HRFR and LRFR 1h are the same, this would at least allay any concerns that the experimental treatments were not the same. I understand that doing so precludes comparison between genotypes, but I do think it's important to show that at least the control data are comparable between experiments.

Similarly, for the CoP-qPCR experiments presented in Fig 5B and 5C, the col-0 values for region P2 between Fig 5B and 5C shows that while HRFR and LRFR 1h look comparable, the values presented for LRFR 25h are quite different.

Minor comments:

Presentation of Supplemental Figure 7A/7B and Supplemental Figure 8A/8B could be changed to make the data more clear (i.e. side-by-side rather than superimposed).

I think that the paragraph introducing auxin (lines 25-37) could be reduced to 1-2 sentences and merged into a separate introductory paragraph given that the SAV3 work makes up a relatively minor component of the manuscript.

For Figure 3A, I would strongly encourage the authors to show some of the raw western blot data for PIF4, PIF5 and PIF7 protein abundance (and loading control), not just the normalized values. This could be put into supplemental data, but I think it should accompany the manuscript.

Lines 145-147 "we observed a strong correlation between PIF4 protein levels (Figure 3A) and PIL1 promoter occupancy (Figure 3B), and a similar behavior was seen with PIF7 (Figure 3B)." According to Fig 3A, there is no statistically significant increase in PIF7 abundance after 1h shade. There is an apparent increase in PIF7 promoter occupancy, but the variation appears too large for it to be statistically significant. I agree that qualitatively there is a correlation for PIF4 but I think the description of the behavior of PIF7 should be rephrased.

There appear to be issues in the coloring of the labels (light blue dots vs dark blue dots) for the PIF7 panels of Fig 3B and Supplemental Fig 3B.

Significance

This authors here have sought to examine the possibility that the transcriptional responses to shade mediated by the phy-PIF system might involve large-scale opening or closing of chromatin regions. This is an important and unanswered question in the field despite several studies that have looked at the role of histone variants (H2A.Z) and modifications (H3K4me3 and H3K9ac) in modulating PIF transcriptional activating activity. The authors have shown that, at least in the case of the transcriptional response to shade mediated by PIF7 (and to an extent PIF4), large-scale changes in chromatin accessibility are not occurring in response to shade treatment.

The results presented in this study support the hypothesis that large-scale changes in chromatin accessibility may have already occurred before plants see shade. This opens the possibility that perhaps the initial perception of light by etiolated (dark-grown seedlings) might trigger changes in chromatin accessibility, opening up chromatin in regions encoding "shade-specific" genes and/or closing chromatin in regions encoding "dark-specific" genes.

The audience for this particular manuscript encompasses a fairly broad group of biologists interested in understanding how environmental stimuli can trigger changes in chromatin reorganization and transcription. The results here are important in that they rule out chromatin accessibility changes as underlying the changes in transcription between the short-term and long-term shade responses. They also reveal that there are a few cases in which chromatin accessibility does change in a statistically-significant manner in response to shade. These regions, and the molecular players which regulate their accessibility, merit further exploration.

My fields of expertise are photobiology, photosynthesis and early seedling development.

Reviewer #1 (Public review):

This is a well-designed and very interesting study examining the impact of imprecise feedback on outcomes in decision-making. I think this is an important addition to the literature, and the results here, which provide a computational account of several decision-making biases, are insightful and interesting.

I do not believe I have substantive concerns related to the actual results presented; my concerns are more related to the framing of some of the work. My main concern is regarding the assertion that the results prove that non-normative and non-Bayesian learning is taking place. I agree with the authors that their results demonstrate that people will make decisions in ways that demonstrate deviations from what would be optimal for maximizing reward in their task under a strict application of Bayes' rule. I also agree that they have built reinforcement learning models that do a good job of accounting for the observed behavior. However, the Bayesian models included are rather simple, per the author's descriptions, applications of Bayes' rule with either fixed or learned credibility for the feedback agents. In contrast, several versions of the RL models are used, each modified to account for different possible biases. However, more complex Bayes-based models exist, notably active inference, but even the hierarchical Gaussian filter. These formalisms are able to accommodate more complex behavior, such as affect and habits, which might make them more competitive with RL models. I think it is entirely fair to say that these results demonstrate deviations from an idealized and strict Bayesian context; however, the equivalence here of Bayesian and normative is, I think, misleading or at least requires better justification/explanation. This is because a great deal of work has been done to show that Bayes optimal models can generate behavior or other outcomes that are clearly not optimal to an observer within a given context (consider hallucinations for example) but which make sense in the context of how the model is constructed as well as the priors and desired states the model is given.

As such, I would recommend that the language be adjusted to carefully define what is meant by normative and Bayesian and to recognize that work that is clearly Bayesian could potentially still be competitive with RL models if implemented to model this task. An even better approach would be to directly use one of these more complex modelling approaches, such as active inference, as the comparator to the RL models, though I would understand if the authors would want this to be a subject for future work.

Abstract:

The abstract is lacking in some detail about the experiments done, but this may be a limitation of the required word count. If word count is not an issue, I would recommend adding details of the experiments done and the results.<br /> One comment is that there is an appeal to normative learning patterns, but this suggests that learning patterns have a fixed optimal nature, which may not be true in cases where the purpose of the learning (e.g. to confirm the feeling of safety of being in an in-group) may not be about learning accurately to maximize reward. This can be accommodated in a Bayesian framework by modelling priors and desired outcomes. As such, the central premise that biased learning is inherently non-normative or non-Bayesian, I think, would require more justification. This is true in the introduction as well.

Introduction:

As noted above, the conceptualization of Bayesian learning being equivalent to normative learning, I think requires further justification. Bayesian belief updating can be biased and non-optimal from an observer perspective, while being optimal within the agent doing the updating if the priors/desired outcomes are set up to advantage these "non-optimal" modes of decision making.

Results:

I wonder why the agent was presented before the choice, since the agent is only relevant to the feedback after the choice is made. I wonder if that might have induced any false association between the agent identity and the choice itself. This is by no means a critical point, but it would be interesting to get the authors' thoughts.

The finding that positive feedback increases learning is one that has been shown before and depends on valence, as the authors note. They expanded their reinforcement learning model to include valence, but they did not modify the Bayesian model in a similar manner. This lack of a valence or recency effect might also explain the failure of the Bayesian models in the preceding section, where the contrast effect is discussed. It is not unreasonable to imagine that if humans do employ Bayesian reasoning that this reasoning system has had parameters tuned based on the real world, where recency of information does matter; affect has also been shown to be incorporable into Bayesian information processing (see the work by Hesp on affective charge and the large body of work by Ryan Smith). It may be that the Bayesian models chosen here require further complexity to capture the situation, just like some of the biases required updates to the RL models. This complexity, rather than being arbitrary, may be well justified by decision-making in the real world.

The methods mention several symptom scales- it would be interesting to have the results of these and any interesting correlations noted. It is possible that some of the individual variability here could be related to these symptoms, which could introduce precision parameter changes in a Bayesian context and things like reward sensitivity changes in an RL context.

Discussion:

(For discussion, not a specific comment on this paper): One wonders also about participants' beliefs about the experiment or the intent of the experimenters. I have often had participants tell me they were trying to "figure out" a task or find patterns even when this was not part of the experiment. This is not specific to this paper, but it may be relevant in the future to try and model participant beliefs about the experiment especially in the context of disinformation, when they might be primed to try and "figure things out".

As a general comment, in the active inference literature, there has been discussion of state-dependent actions, or "habits", which are learned in order to help agents more rapidly make decisions, based on previous learning. It is also possible that what is being observed is that these habits are at play, and that they represent the cognitive biases. This is likely especially true given, as the authors note, the high cognitive load of the task. It is true that this would mean that full-force Bayesian inference is not being used in each trial, or in each experience an agent might have in the world, but this is likely adaptive on the longer timescale of things, considering resource requirements. I think in this case you could argue that we have a departure from "normative" learning, but that is not necessarily a departure from any possible Bayesian framework, since these biases could potentially be modified by the agent or eschewed in favor of more expensive full-on Bayesian learning when warranted.

Indeed, in their discussion on the strategy of amplifying credible news sources to drown out low-credibility sources, the authors hint at the possibility of longer-term strategies that may produce optimal outcomes in some contexts, but which were not necessarily appropriate to this task. As such, the performance on this task- and the consideration of true departure from Bayesian processing- should be considered in this wider context.

Another thing to consider is that Bayesian inference is occurring, but that priors present going in produce the biases, or these biases arise from another source, for example, factoring in epistemic value over rewards when the actual reward is not large. This again would be covered under an active inference approach, depending on how the priors are tuned. Indeed, given the benefit of social cohesion in an evolutionary perspective, some of these "biases" may be the result of adaptation. For example, it might be better to amplify people's good qualities and minimize their bad qualities in order to make it easier to interact with them; this entails a cost (in this case, not adequately learning from feedback and potentially losing out sometimes), but may fulfill a greater imperative (improved cooperation on things that matter). Given the right priors/desired states, this could still be a Bayes-optimal inference at a social level and, as such, may be ingrained as a habit that requires effort to break at the individual level during a task such as this.

The authors note that this task does not relate to "emotional engagement" or "deep, identity-related issues". While I agree that this is likely mostly true, it is also possible that just being told one is being lied to might elicit an emotional response that could bias responses, even if this is a weak response.

Reviewer #3 (Public review):

Summary

This paper investigates how disinformation affects reward learning processes in the context of a two-armed bandit task, where feedback is provided by agents with varying reliability (with lying probability explicitly instructed). They find that people learn more from credible sources, but also deviate systematically from optimal Bayesian learning: They learned from uninformative random feedback, learned more from positive feedback, and updated too quickly from fully credible feedback (especially following low-credibility feedback). Overall, this study highlights how misinformation could distort basic reward learning processes, without appeal to higher-order social constructs like identity.

Strengths

(1) The experimental design is simple and well-controlled; in particular, it isolates basic learning processes by abstracting away from social context.

(2) Modeling and statistics meet or exceed the standards of rigor.

(3) Limitations are acknowledged where appropriate, especially those regarding external validity.

(4) The comparison model, Bayes with biased credibility estimates, is strong; deviations are much more compelling than e.g., a purely optimal model.

(5) The conclusions are interesting, in particular the finding that positivity bias is stronger when learning from less reliable feedback (although I am somewhat uncertain about the validity of this conclusion)

Weaknesses

(1) Absolute or relative positivity bias?

In my view, the biggest weakness in the paper is that the conclusion of greater positivity bias for lower credible feedback (Figure 5) hinges on the specific way in which positivity bias is defined. Specifically, we only see the effect when normalizing the difference in sensitivity to positive vs. negative feedback by the sum. I appreciate that the authors present both and add the caveat whenever they mention the conclusion (with the crucial exception of the abstract). However, what we really need here is an argument that the relative definition is the *right* way to define asymmetry....

Unfortunately, my intuition is that the absolute difference is a better measure. I understand that the relative version is common in the RL literature; however previous studies have used standard TD models, whereas the current model updates based on the raw reward. The role of the CA parameter is thus importantly different from a traditional learning rate - in particular, it's more like a logistic regression coefficient (as described below) because it scales the feedback but *not* the decay. Under this interpretation, a difference in positivity bias across credibility conditions corresponds to a three-way interaction between the exponentially weighted sum of previous feedback of a given type (e.g., positive from the 75% credible agent), feedback positivity, and condition (dummy coded). This interaction corresponds to the non-normalized, absolute difference.

Importantly, I'm not terribly confident in this argument, but it does suggest that we need a compelling argument for the relative definition.

(2) Positivity bias or perseveration?

A key challenge in interpreting many of the results is dissociating perseveration from other learning biases. In particular, a positivity bias (Figure 5) and perseveration will both predict a stronger correlation between positive feedback and future choice. Crucially, the authors do include a perseveration term, so one would hope that perseveration effects have been controlled for and that the CA parameters reflect true positivity biases. However, with finite data, we cannot be sure that the variance will be correctly allocated to each parameter (c.f. collinearity in regressions). The fact that CA- is fit to be negative for many participants (a pattern shown more strongly in the discovery study) is suggestive that this might be happening. A priori, the idea that you would ever increase your value estimate after negative feedback is highly implausible, which suggests that the parameter might be capturing variance besides that it is intended to capture.

The best way to resolve this uncertainty would involve running a new study in which feedback was sometimes provided in the absence of a choice - this would isolate positivity bias. Short of that, perhaps one could fit a version of the Bayesian model that also includes perseveration. If the authors can show that this model cannot capture the pattern in Figure 5, that would be fairly convincing.

(3) Veracity detection or positivity bias?

The "True feedback elicits greater learning" effect (Figure 6) may be simply a re-description of the positivity bias shown in Figure 5. This figure shows that people have higher CA for trials where the feedback was in fact accurate. But, assuming that people tend to choose more rewarding options, true-feedback cases will tend to also be positive-feedback cases. Accordingly, a positivity bias would yield this effect, even if people are not at all sensitive to trial-level feedback veracity. Of course, the reverse logic also applies, such that the "positivity bias" could actually reflect discounting of feedback that is less likely to be true. This idea has been proposed before as an explanation for confirmation bias (see Pilgrim et al, 2024 https://doi.org/10.1016/j.cognition.2023.105693 and much previous work cited therein). The authors should discuss the ambiguity between the "positivity bias" and "true feedback" effects within the context of this literature....

The authors get close to this in the discussion, but they characterize their results as differing from the predictions of rational models, the opposite of my intuition. They write:

Alternative "informational" (motivation-independent) accounts of positivity and confirmation bias predict a contrasting trend (i.e., reduced bias in low- and medium credibility conditions) because in these contexts it is more ambiguous whether feedback confirms one's choice or outcome expectations, as compared to a full-credibility condition.

I don't follow the reasoning here at all. It seems to me that the possibility for bias will increase with ambiguity (or perhaps will be maximal at intermediate levels). In the extreme case, when feedback is fully reliable, it is impossible to rationally discount it (illustrated in Figure 6A). The authors should clarify their argument or revise their conclusion here.

(4) Disinformation or less information?

Zooming out, from a computational/functional perspective, the reliability of feedback is very similar to reward stochasticity (the difference is that reward stochasticity decreases the importance/value of learning in addition to its difficulty). I imagine that many of the effects reported here would be reproduced in that setting. To my surprise, I couldn't quickly find a study asking that precise question, but if the authors know of such work, it would be very useful to draw comparisons. To put a finer point on it, this study does not isolate which (if any) of these effects are specific to *disinformation*, rather than simply _less information._ I don't think the authors need to rigorously address this in the current study, but it would be a helpful discussion point.

(5) Over-reliance on analyzing model parameters

Most of the results rely on interpreting model parameters, specifically, the "credit assignment" (CA) parameter. Exacerbating this, many key conclusions rest on a comparison of the CA parameters fit to human data vs. those fit to simulations from a Bayesian model. I've never seen anything like this, and the authors don't justify or even motivate this analysis choice. As a general rule, analyses of model parameters are less convincing than behavioral results because they inevitably depend on arbitrary modeling assumptions that cannot be fully supported. I imagine that most or even all of the results presented here would have behavioral analogues. The paper would benefit greatly from the inclusion of such results. It would also be helpful to provide a description of the model in the main text that makes it very clear what exactly the CA parameter is capturing (see next point).

(6) RL or regression?

I was initially very confused by the "RL" model because it doesn't update based on the TD error. Consequently, the "Q values" can go beyond the range of possible reward (SI Figure 5). These values are therefore *not* Q values, which are defined as expectations of future reward ("action values"). Instead, they reflect choice propensities, which are sometimes notated $h$ in the RL literature. This misuse of notation is unfortunately quite common in psychology, so I won't ask the authors to change the variable. However, they should clarify when introducing the model that the Q values are not action values in the technical sense. If there is precedent for this update rule, it should be cited.

Although the change is subtle, it suggests a very different interpretation of the model.

Specifically, I think the "RL model" is better understood as a sophisticated logistic regression, rather than a model of value learning. Ignoring the decay term, the CA term is simply the change in log odds of repeating the just-taken action in future trials (the change is negated for negative feedback). The PERS term is the same, but ignoring feedback. The decay captures that the effect of each trial on future choices diminishes with time. Importantly, however, we can re-parameterize the model such that the choice at each trial is a logistic regression where the independent variables are an exponentially decaying sum of feedback of each type (e.g., positive-cred50, positive-cred75, ... negative-cred100). The CA parameters are simply coefficients in this logistic regression.

Critically, this is not meant to "deflate" the model. Instead, it clarifies that the CA parameter is actually not such an assumption-laden model estimate. It is really quite similar to a regression coefficient, something that is usually considered "model agnostic". It also recasts the non-standard "cross-fitting" approach as a very standard comparison of regression coefficients for model simulations vs. human data. Finally, using different CA parameters for true vs false feedback is no longer a strange and implausible model assumption; it's just another (perfectly valid) regression. This may be a personal thing, but after adopting this view, I found all the results much easier to understand.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public Review):

Summary: 

The manuscript by Nicoletti et al. presents a minimal model of habituation, a basic form of non-associative learning, addressing both from dynamical and information theory aspects of how habituation can be realized. The authors identify that negative feedback provided with a slow storage mechanism is sufficient to explain habituation.

Strengths: 

The authors combine the identification of the dynamical mechanism with information-theoretic measures to determine the onset of habituation and provide a description of how the system can gain maximum information about the environment.

We thank the reviewer for highlighting the strength of our work and for their comments, which we believe have been instrumental in significantly improving our work and its scope. Below, we address all their concerns.

Weaknesses: 

I have several main concerns/questions about the proposed model for habituation and its plausibility. In general, habituation does not only refer to a decrease in the responsiveness upon repeated stimulation but as Thompson and Spencer discussed in Psych. Rev. 73, 16-43 (1966), there are 10 main characteristics of habituation, including (i) spontaneous recovery when the stimulus is withheld after response decrement; dependence on the frequency of stimulation such that (ii) more frequent stimulation results in more rapid and/or more pronounced response decrement and more rapid spontaneous recovery; (iii) within a stimulus modality, the less intense the stimulus, the more rapid and/or more pronounced the behavioral response decrement; (iv) the effects of repeated stimulation may continue to accumulate even after the response has reached an asymptotic level (which may or may not be zero, or no response). This effect of stimulation beyond asymptotic levels can alter subsequent behavior, for example, by delaying the onset of spontaneous recovery. 

These are only a subset of the conditions that have been experimentally observed and therefore a mechanistic model of habituation, in my understanding, should capture the majority of these features and/or discuss the absence of such features from the proposed model. 

We are really grateful to the reviewer for pointing out these aspects of habituation that we overlooked in the previous version of our manuscript. Indeed, our model is able to capture most of these 10 observed behaviors, specifically: 1) habituation; 2) spontaneous recovery; 3) potentiation of habituation; 4) frequency sensitivity; 5) intensity sensitivity; 6) subliminal accumulation. Here, we are following the same terminology employed in Eckert et al., Current Biology 34, 5646–5658 (2024), the paper highlighted by the reviewer. We have dedicated a section of the revised version of the manuscript to these hallmarks, substantiating the validity of our framework as a minimal model to have habituation. We remark that these are the sole hallmarks that can be discussed by considering one single external stimulus and that can be identified without ambiguity in a biochemical context. This observation is again in line with Eckert et al., Current Biology 34, 5646–5658 (2024).

In the revised version, we employ the same strategy of the aforementioned work to determine when the system can be considered “habituated”. Indeed, we introduce a response threshold that is now discussed in the manuscript. We also included a note in the discussions stating that, since any biochemical model will eventually reach a steady state, subliminal accumulation, for example, can only be seen with the use of a threshold. The introduction of different storage mechanisms, ideally more detailed at a molecular level, can shed light on this conceptual gap. This is an interesting direction of research.

Furthermore, the habituated response in steady-state is approximately 20% less than the initial response, which seems to be achieved already after 3-4 pulses, the subsequent change in response amplitude seems to be negligible, although the authors however state "after a large number of inputs, the system reaches a time-periodic steady-state". How do the authors justify these minimal decreases in the response amplitude? Does this come from the model parametrization and is there a parameter range where more pronounced habituation responses can be observed? 

The reviewer is correct, but this is solely a consequence of the specific set of parameters we selected. We made this choice solely for visualization purposes in the previous version. In the revised version, in the section discussing the hallmarks of habituation, we also show other parameter choices when the response decrement is more pronounced. Moreover, we remark that the contour plot of \Delta⟨U> clearly shows that the decrement can largely exceed the 20% threshold presented in the previous version.

In the revised version, also in light of the works highlighted by the reviewer, we decided to move the focus of the manuscript to the information-theoretic advantage of habituation. As such, we modified several parts of the main text. Also, in the region of optimal information gain, habituation is at an intermediate level. For this reason, we decided to keep the same parameter choice as the previous version in Figure 2.

We stated that the time-periodic steady-state is reached “after a large number of stimuli” from a mathematical perspective. However, by using a habituation threshold, as done in Eckert et al., Current Biology 34, 5646–5658 (2024), we can state that the system is habituated after a few stimuli for each set of parameters. This aspect is highlighted in the revised version of the manuscript (see also the point above).

The same is true for the information content (Figure 2f) - already at the first pulse, IU, H ~ 0.7 and only negligibly increases afterwards. In my understanding, during learning, the mutual information between the input and the internal state increases over time and the system extracts from these predictions about its responses. In the model presented by the authors, it seems the system already carries information about the environment which hardly changes with repeated stimulus presentation. The complexity of the signal is also limited, and it is very hard to clarify from the presented results, whether the proposed model can actually explain basic features of habituation, as mentioned above. 

As for the response decrement of the readout, we can certainly choose a set of parameters for which the information gain is higher. In the revised version, we also report the information at the first stimulation and when the system is habituated to give a better idea of the range of these quantities. At any rate, as the referee correctly points out, it is difficult to give an intuitive interpretation of the information in our minimal model.

It is also important to remark that, since the readout population and the receptor both undergo fast dynamics (with appropriate timescales as discussed in the text), we are not observing the transient gain of information associated with the first stimulus. As such, the mutual information presents a discontinuous behavior that resembles the dynamics of the readout, thereby starting at a non-zero value already at the first stimulus.

Additionally, there have been two recent models on habituation and I strongly suggest that the authors discuss their work in relation to recent works (bioRxiv 2024.08.04.606534; arXiv:2407.18204).

We thank the reviewer for pointing out these relevant references. In the revised version, we highlighted that we discuss the information-theoretic aspects of habituation, while the aforementioned references focus on the dynamics of this phenomenon.

Reviewer #1 (Recommendations for the authors):

I would also like to note here the simplification of the proposed biological model - in particular, that the receptor can be in an active/passive state, as well as proposing the Nf-kB signaling module as a possible molecular realization. Generally, a large number of cell surface receptors including RTKs of GPCRs have much more complex dynamics including autocatalytic activation that generally leads to bistability, and the Nf-kB has been demonstrated to have oscillatory even chaotic dynamics (works of Savas Tsay, Mogens Jensen and others). Considering this, the authors should at least discuss under which conditions these TNF-Alpha signaling could potentially serve as a molecular realisation for habituation. 

We thank the reviewer for bringing this to our attention. In the previous version, we reported the TNF signaling network only to show a similar coarse-grained modular structure. However, following a suggestion of reviewer #2, we decided to change Figure 1 to include a simplified molecular scheme of chemotaxis rather than TNF signaling, to avoid any source of confusion about this issue.

Also, a minor point: Figures 2d-e are cited before 2a-c. 

We apologize for the oversight. The structure of the Figures and their order is now significantly different, and they are now cited in the correct order. 

Reviewer #2 (Public review):

In this study, the authors aim to investigate habituation, the phenomenon of increasing reduction in activity following repeated stimuli, in the context of its information-theoretic advantage. To this end, they consider a highly simplified three-species reaction network where habituation is encoded by a slow memory variable that suppresses the receptor and therefore the readout activity. Using analytical and numerical methods, they show that in their model the information gain, the difference between the mutual information between the signal and readout after and before habituation, is maximal for intermediate habituation strength. Furthermore, they demonstrate that the Pareto front corresponds to an optimization strategy that maximizes the mutual information between signal and readout in the steady state, minimizes some form of dissipation, and also exhibits similar intermediate habituation strength. Finally, they briefly compare predictions of their model to whole-brain recordings of zebrafish larvae under visual stimulation. 

The author's simplified model might serve as a solid starting point for understanding habituation in different biological contexts as the model is simple enough to allow for some analytic understanding but at the same time exhibits all basic properties of habituation in sensory systems. Furthermore, the author's finding of maximal information gain for intermediate habituation strength via an optimization principle is, in general, interesting. However, the following points remain unclear or are weakly explained: 

We thank the reviewer for deeming our work interesting and for considering it a solid starting point for understanding habituation in biological systems.

(1) Is it unclear what the meaning of the finding of maximal information gain for intermediate habituation strength is for biological systems? Why is information gain as defined in the paper a relevant quantity for an organism/cell? For instance, why is a system with low mutual information after the first stimulus and intermediate mutual information after habituation better than one with consistently intermediate mutual information? Or, in other words, couldn't the system try to maximize the mutual information acquired over the whole time series, e.g., the time series mutual information between the stimulus and readout?

This is a delicate aspect to discuss and we thank the referee for the comment. In the revised version, we report information gain, initial and final information, highlighting that both gain and final information are higher in regions where habituation is present. They have qualitatively similar behavior and highlight a clear information-theoretic advantage of this dynamical phenomenon. An important point is that, to determine the optimal Pareto front, we consider a prolonged stimulus and its associated steady-state information. Therefore, from the optimization point of view, there is no notion of “information gain” or “final information”, which are intrinsically dynamical quantities. As a result, the fact that optimal curve lies in the region of optimal information gain is a-priori not expected and hints at the potential crucial role of this feature. In the revised version, we elucidate this aspect with several additional analyses.

We would like to add that, from a naive perspective, while the first stimulation will necessarily trigger a certain (non-zero) mutual information, multiple observations of the same stimulus have to reflect into accumulated information that consequently drives the onset of observed dynamical behaviors, such as habituation.

(2) The model is very similar to (or a simplification of previous models) for adaptation in living systems, e.g., for adaptation in chemotaxis via activity-dependent methylation and demethylation. This should be made clearer.

We apologize for having missed this point. Our choice has been motivated by the fact that we wanted to avoid confusion between the usual definition of (perfect) adaptation and habituation. However, we now believe that this is not the case for the revised manuscript, and we now include chemotaxis as an example in Figure 1.

(3) It remains unclear why this optimization principle is the most relevant one. While it makes sense to maximize the mutual information between stimulus and readout, there are various choices for what kind of dissipation is minimized. Why was \delta Q_R chosen and not, for instance, \dot{\Sigma}_int or the sum of both? How would the results change in that case? And how different are the results if the mutual information is not calculated for the strong stimulation input statistics but for the background one?

We thank the reviewer for the suggestion. We agree that a priori, there is no reason to choose \delta Q_R or a function of the internal energy flux J_int (that, in the revised version, we are using in place of \dot\Sigma_int following the suggestion of reviewer #3). The rationale was to minimize \delta Q_R since this dissipation is unavoidable and stems from the presence of the storage inhibiting the receptor through the internal pathway. Indeed, considering the existence of two different pathways implementing sensing and feedback, the presence of any input will result in a dissipation produced by the receptor. This energy consumption is reflected in \delta Q_R.

In the revised version, we now include in the optimization principle two energy contributions (see Eq. (14) of the revised manuscript): \delta Q_R and E_int, which is the energy consumption associated with the driven storage production per unit energy. All Figures have been updated accordingly. The results remain similar, as \delta Q_R still represents the main contribution, especially at high \beta.

Furthermore, in the revised version, we include examples of the Pareto optimization for different values of input strength. As detailed both in the main text and the Supplementary Information, changing the value of ⟨H⟩ moves the Pareto frontier in the (\beta, \sigma) space, since the signal needs to be strong enough for the system to distinguish it from the intrinsic thermal noise (controlled by beta). We also show that if the system is able to tune the inhibition strength \kappa, the Pareto frontiers at different ⟨H⟩ collapse into a single curve. This shows that, although the values of, e.g., the mutual information, depend on ⟨H⟩, the qualitative behavior of the system in this regime is effectively independent of it. We also added more details about this in the Supplementary Information.

(4) The comparison to the experimental data is not too strong of an argument in favor of the model. Is the agreement between the model and the experimental data surprising? What other behavior in the PCA space could one have expected in the data? Shouldn't the 1st PC mostly reflect the "features", by construction, and other variability should be due to progressively reduced activity levels? 

The agreement between data and model is not surprising - we agree on this - since the data exhibit habituation. However, we believe that the fact that our minimal model is able to capture the features of a complex neural system just by looking at the PCs, without any explicit biological details, is non-trivial. We also stress that the 1st PC only reflects the feature that captures most of the variance of the data and, as such, it is difficult to have a-priori expectations on what it should represent. In the case of the data generated from the model, most of the variance of the activity comes from the switching signal, and similar considerations can be made for the looming stimulations in the data. We updated the manuscript to clarify this point.

Reviewer #2 (Recommendations for the authors):

(1) The abstract makes it sound like a new finding is that habituation is due to a slow, negative feedback mechanism. But, as mentioned in the introduction, this is a well-known fact. 

We agree with the reviewer. We have revised the abstract.

(2) Figure 2c Why does the range of Delta Delta I_f include negative values if the corresponding region is shaded (right-tilted stripes)? 

The negative values in the range are those attained in the shaded region with right-tilted stripes. We decided to include them in the colorbar for clarity, since Delta Delta I_f is also plotted in the region where it attains negative values.

(3) What does the Pareto front look like if the optimization is done for input statistics given by ⟨H⟩_min? 

In the revised version, we include examples of the Pareto optimization for different values of input strength. As detailed both in the main text and the Supplementary Information, changing the value of ⟨H⟩ moves the Pareto frontier in the (\beta, \sigma) space, since the strength of the signal is crucial for the system to discriminate input and thermal noise (see also the answers above).

In particular, in Figure 4 we explicitly compare the results of the Pareto optimization (which is done with a static input of a given statistics) with the dynamics of the model for different values of ⟨H⟩ in two scenarios, i.e., adaptive and non-adaptive inhibition strength (see answers above for details).

We also remark that ⟨H⟩_min represents the background signal that the system is not trying to capture, which is why we never used it for optimization.

(4) From the main text, it is rather difficult to understand how the comparison to the experimental data was performed. How was the PCA done exactly? What are the "features" of the evoked neural response? 

The PCA on data is performed starting from the single-neuron calcium dynamics. To perform a far comparison, we reconstruct a similar but extremely simplified dynamics using our model as explained in Methods to perform the PCA on analogous simulated data. We added a comment on this in the revised version. While these components capture most of the variance in the data, their specific interpretation is usually out of reach and we believe that it lies beyond the scope of this theoretical work. We also remark that the model does not contain all these biological details - a strong aspect in our opinion - and, as such, it cannot capture specific biological features.

Reviewer #3 (Public review):

The authors use a generic model framework to study the emergence of habituation and its functional role from information-theoretic and energetic perspectives. Their model features a receptor, readout molecules, and a storage unit, and as such, can be applied to a wide range of biological systems. Through theoretical studies, the authors find that habituation (reduction in average activity) upon exposure to repeated stimuli should occur at intermediate degrees to achieve maximal information gain. Parameter regimes that enable these properties also result in low dissipation, suggesting that intermediate habituation is advantageous both energetically and for the purpose of retaining information about the environment. 

A major strength of the work is the generality of the studied model. The presence of three units (receptor, readout, storage) operating at different time scales and executing negative feedback can be found in many domains of biology, with representative examples well discussed by the authors (e.g. Figure 1b). A key takeaway demonstrated by the authors that has wide relevance is that large information gain and large habituation cannot be attained simultaneously. When energetic considerations are accounted for, large information gain and intermediate habituation appear to be a favorable combination. 

We thank the reviewer for this positive assessment of our work and its generality.

While the generic approach of coarse-graining most biological detail is appealing and the results are of broad relevance, some aspects of the conducted studies, the problem setup, and the writing lack clarity and should be addressed: 

(1) The abstract can be further sharpened. Specifically, the "functional role" mentioned at the end can be made more explicit, as it was done in the second-to-last paragraph of the Introduction section ("its functional advantages in terms of information gain and energy dissipation"). In addition, the abstract mentions the testing against experimental measurements of neural responses but does not specify the main takeaways. I suggest the authors briefly describe the main conclusions of their experimental study in the abstract.

We thank the reviewer for raising this point. In the revised version, we have changed the abstract to reflect the reviewer’s points and the new structure and results of the manuscript.

(2) Several clarifications are needed on the treatment of energy dissipation. 

-   When substituting the rates in Eq. (1) into the definition of δQ_R above Eq. (10), "σ" does not appear on the right-hand side. Does this mean that one of the rates in the lower pathway must include σ in its definition? Please clarify.

We apologize to the reviewer for this typo. Indeed, \sigma sets the energy scale of feedback and, as such, it appears in the energetic driving given by the feedback on the receptor, i.e., in Eq. (1) together with \kappa. This typo has been corrected in the revised manuscript, and all subsequent equations are consistent.

-   I understand that the production of storage molecules has an associated cost σ and hence contributes to dissipation. The dependence of receptor dissipation on ⟨H⟩, however, is not fully clear. If the environment were static and the memory block was absent, the term with ⟨H⟩ would still contribute to dissipation. What would be the nature of this dissipation?

In the spirit of building a paradigmatic minimal model with a thermodynamic meaning, we considered H to act as an external thermodynamic driving. Since this driving acts on a different pathway with respect to the one affected by the storage, the receptor is driven out of equilibrium by its presence.

By eliminating the memory block, we would also be necessarily eliminating the presence of the pathway associated with the storage effect (“internal pathway” in the manuscript), since its presence is solely due to the existence of a storage population. Therefore, in this case, the receptor would be a 2-state, 1-pathway system and, as such, it would always satisfy an effective detailed balance. As a consequence, the definition of \delta Q_R reported in the manuscript would not hold anymore and the receptor would not exhibit any dissipation. Thus, in a static environment and without a memory block, no receptor dissipation would be present. We would also like to stress that our choice to model two different pathways has been motivated by the observation that the negative feedback acts along a different pathway in several biochemical and biological examples. We made some changes to the model description in the revised version and we hope that this aspect has been clarified.

-   Similarly, in Eq. (9) the authors use the ratio of the rates Γ_{s → s+1} and Γ_{s+1 → s} in their expression for internal dissipation. The first-rate corresponds to the synthesis reaction of memory molecules, while the second corresponds to a degradation reaction. Since the second reaction is not the microscopic reverse of the first, what would be the physical interpretation of the log of their ratio? Since the authors already use σ as the energy cost per storage unit, why not use σ times the rate of producing S as a metric for the dissipation rate? 

We agree with the referee that the reverse reaction we considered is not the microscopic reverse of the storage production. In the case of a fast readout population, we employed a coarse-grained view to compute this entropy production. To be more precise, we gladly welcomed the referee’s suggestion in the revised version and modified the manuscript accordingly. As suggested, we now employ the energy flux associated with the storage production to estimate the internal dissipation (see new Fig. 3). 

In the revised version, we also use this quantity in the optimization procedure in combination with \deltaQ_R (see new Fig. 4) to have a complete characterization of the system’s energy consumption. The conclusions are qualitatively identical to before, but we believe that now they are more solid from a theoretical perspective. For this important advance in the robustness and quality of our work, we are profoundly grateful to the referee.

(3) Impact of the pre-stimulus state. The plots in Figure 2 suggest that the environment was static before the application of repeated stimuli. Can the authors comment on the impact of the pre-stimulus state on the degree of habituation and its optimality properties? Specifically, would the conclusions stay the same if the prior environment had stochastic but aperiodic dynamics? 

The initial stimulus is indeed stochastic with an average constant in time and mimics the background (small) signal. We apply the (strong) stimulation when the system already reached a stationary state with respect to the background. As it can be appreciated in Fig. 2 of the revised version, the model response depends on the pre-stimulus level, since it sets the storage concentration before the stimulation arrives and, as such, the subsequent habituation dynamics. This dependence is important from a dynamical perspective. The information-theoretic picture has been developed, as said above, by letting the system relax before the first stimulus. This eliminates this arbitrary dependence and provides a clearer idea of the functional advantages of habituation. Moreover, the optimization procedure is performed in a completely different setting, with no pre-stimulus at all, since we only have one prolonged stimulation. We hope that the revised version is clearer on all these points.

(4) Clarification about the memory requirement for habituation. Figure 4 and the associated section argue for the essential role that the storage mechanism plays in habituation. Indeed, Figure 4a shows that the degree of habituation decreases with decreasing memory. The graph also shows that in the limit of vanishingly small Δ⟨S⟩, the system can still exhibit a finite degree of habituation. Can the authors explain this limiting behavior; specifically, why does habituation not vanish in the limit Δ⟨S⟩ -> 0?

We apologize for the lack of clarity and we thank the reviewer for spotting this issue. In Figure 4 (now Figure 5 in the revised manuscript) Δ⟨S⟩ is not exactly zero, but equal to 0.15% at the final point. It appeared as 0% in the plot due to an unwanted rounding in the plotting function that we missed. This has been fixed in the revised version, thank you.

Reviewer #3 (Recommendations for the authors):

(1) Page 2 | "Figure 1b-e" should be "Figure 1b-d" since there is no panel (e) in Figure 1. 

(2) Figure 1a | In the top schematic, the symbol "k" is used, while in the rest of the text, the proportionality constant is denoted by κ. 

We thank the reviewer for pointing this out. Figure 1 has been revised and the panels are now consistent. The proportionality constant (the inhibition strength) has also been fixed.

(3) Figure 1a | I find the upper part of the schematic for Storage hard to perceive. I understand the lower part stands for the degradation reaction for storage molecules. The upper part stands for the synthesis reaction catalyzed by the readout population. I think the bolded upper arrow would explain it sufficiently well; the left/right arrows, together with the crossed green circle make that part of the figure confusing. Consider simplifying. 

We decided to remove the left/right arrows, as suggested by the reviewer, as we agree that they were unnecessarily complicating the schematic. We hope that the revised version will be easier to understand.

(4)Page 3 | It would be helpful to tell what the temporal statistics of the input signal $p_H(h,t)$ is, i.e. <h(t) h(t')>. Looking at the example trajectory in Figure 1a, consecutive signal values do not seem correlated. 

We agree with the reviewer that this is an important detail and worth mentioning. We now explicitly state that consecutive values are not correlated, for simplicity. 

(5)Figure 2 | I believe the label "EXTERNAL INPUT" refers to the *average* external input, not one specific realization (similar to panels (d) and (e) that report on average metrics). I suggest you indicate this in the label, or, what may be even better, add one particular realization of the stochastic input to the same graph.

We thank the reviewer for spotting this. We now write that what we show is the average external signal. We prefer this solution rather than showing a realization of the stochastic input, since it is more consistent with the rest of the plots, where we always show average quantities. We also note that Figure 2 is now Figure 3 in the revised manuscript.

(6)Figure 2d | The expression of Δ⟨U⟩ is the negative of the definition in Eq. (5). It should be corrected. 

In the revised version, both the definitions in Figure 2 (now Figure 3) and in the text (now Eq. (11)) are consistent.

(7) Figure 3(d-e) caption | "where ⟨U⟩ starts to be significantly smaller than zero." There, it should be Δ⟨U⟩ instead of ⟨U⟩. 

Thanks again, we corrected this typo.

The starting point can no longer be a draft, must be a conversation?

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public review): 

Summary: 

Wang et al. identify Hamlet, a PR-containing transcription factor, as a master regulator of reproductive development in Drosophila. Specifically, the fusion between the gonad and genital disc is necessary for the development of continuous testes and seminal vesicle tissue essential for fertility. To do this, the authors generate novel Hamlet null mutants by CRISPR/Cas9 gene editing and characterize the morphological, physiological, and gene expression changes of the mutants using immunofluorescence, RNA-seq, cut-tag, and in-situ analysis. Thus, Hamlet is discovered to regulate a unique expression program, which includes Wnt2 and Tl, that is necessary for testis development and fertility. 

Strengths: 

This is a rigorous and comprehensive study that identifies the Hamlet-dependent gene expression program mediating reproductive development in Drosophila. The Hamlet transcription targets are further characterized by Gal4/UAS-RNAi confirming their role in reproductive development. Finally, the study points to a role for Wnt2 and Tl as well as other Hamlet transcriptionally regulated genes in epithelial tissue fusion. 

We appreciate that the reviewer thinks our study is rigorous.

Weaknesses: 

The image resolution and presentation of figures is a major issue in this study. As a nonexpert, it is nearly impossible to see the morphological changes as described in the results. Quantification of all cell biological phenotypes is also lacking therefore reducing the impact of this study to those familiar with tissue fusion events in Drosophila development. 

In the revised version, we have improved the image presentation and resolution. For all the images with more than two channels, we included single-channel images, changed the green color to lime and the red to magenta, highlighted the testis (TE) and seminal vescicles to make morphological changes more visible.  

We had quantification for marker gene expression in the original version, and now also included quantification for cell biological phenotypes which are generally with 100% penetrance.  

Reviewer #2 (Public review): 

Strengths: 

Wang and colleagues successfully uncovered an important function of the Drosophila PRDM16/PRDM3 homolog Hamlet (Ham) - a PR domain-containing transcription factor with known roles in the nervous system in Drosophila. To do so, they generated and analyzed new mutants lacking the PR domain, and also employed diverse preexisting tools. In doing so, they made a fascinating discovery: They found that PR-domain containing isoforms of ham are crucial in the intriguing development of the fly genital tract. Wang and colleagues found three distinct roles of Ham: (1) specifying the position of the testis terminal epithelium within the testis, (2) allowing normal shaping and growth of the anlagen of the seminal vesicles and paragonia and (3) enabling the crucial epithelial fusion between the seminal vesicle and the testis terminal epithelium. The mutant blocks fusion even if the parts are positioned correctly. The last finding is especially important, as there are few models allowing one to dissect the molecular underpinnings of heterotypic epithelial fusion in development. Their data suggest that they found a master regulator of this collective cell behavior. Further, they identified some of the cell biological players downstream of Ham, like for example E-Cadherin and Crumbs. In a holistic approach, they performed RNAseq and intersected them with the CUT&TAG-method, to find a comprehensive list of downstream factors directly regulated by Ham. Their function in the fusion process was validated by a tissue-specific RNAi screen. Meticulously, Wang and colleagues performed multiplexed in situ hybridization and analyzed different mutants, to gain a first understanding of the most important downstream pathways they characterized, which are Wnt2 and Toll. 

This study pioneers a completely new system. It is a model for exploring a process crucial in morphogenesis across animal species, yet not well understood. Wang and colleagues not only identified a crucial regulator of heterotypic epithelial fusion but took on the considerable effort of meticulously pinning down functionally important downstream effectors by using many state-of-the-art methods. This is especially impressive, as the dissection of pupal genital discs before epithelial fusion is a time-consuming and difficult task. This promising work will be the foundation future studies build on, to further elucidate how this epithelial fusion works, for example on a cell biological and biomechanical level. 

We appreciate that the reviewer thinks our study is orginal and important.

Weaknesses: 

The developing testis-genital disc system has many moving parts. Myotube migration was previously shown to be crucial for testis shape. This means, that there is the potential of non-tissue autonomous defects upon knockdown of genes in the genital disc or the terminal epithelium, affecting myotube behavior which in turn affects fusion, as myotubes might create the first "bridge" bringing the epithelia together. The authors clearly showed that their driver tools do not cause expression in myoblasts/myotubes, but this does not exclude non-tissue autonomous defects in their RNAi screen. Nevertheless, this is outside the scope of this work. 

We thank the reviewer’s consideration of non-tissue autonomous defects upon gene knockdown. The driver, hamRSGal4, drives reporter gene expression mainly in the RS epithelia, but we did observe weak expression of the reporter in the myoblasts before they differentiate into myotubes. Thus, we could not rule out a non-tissue autonomou effect in the RNAi screen. So we now included a statement in the result, “Given that the hamRSGal4 driver is highly expressed in the TE and SV epithelia, we expect highly effective knockdown occurs only in these epithelial cells. However, hamRSGal4 also drives weak expression in the myoblasts before they differentiated into myotubes (Supplementary Fig. 5B), which may result in a non-tissue autonomous effect when knocking down the candidate genes expressed in myoblasts.”

However, one point that could be addressed in this study: the RNAseq and CUT&TAG experiments would profit from adding principal component analyses, elucidating similarities and differences of the diverse biological and technical replicates. 

Thanks for the suggestion. We now have included the PCA analyses in supplementary figure 6A-B and the corresponding description in the text. The PCA graphs validated the consistency between biological replicates of the RNA-seq samples. The Cut&Tag graphs confirm the consistency between the two biological replicates from the GFP samples, but show a higher variability between the w1118 replicates. Importantly, we only considered the overlapped peaks pulled by the GFP antibody from the ham_GFP genotype and the Ham antibody from the wildtype (w1118) sample as true Ham binding sites. 

Recommendations for the authors:  

Reviewer #1 (Recommendations for the authors): 

Major Concern: 

(1) The image resolution and presentation of figures (Figures 2, 5, 6, and 7) is a major issue in this study. As a non-expert, it is nearly impossible to see the morphological changes as described in the results. Images need to be captured at higher resolution and zoomed in with arrows denoting changes as described. Individual channels, particularly for intensity measurement need to be shown in black and white in addition to merged images. Images also need pseudo-colored for color-blind individuals (i.e. no red-green staining). 

The images were captured at a high resolution, but somehow the resolution was drammaticlly reduced in the BioRxiv PDF. We try to overcome this by directly submitting the PDF in the Elife submission system. In the revised version, we have included single-channel images, changed the green and red colors to lime and magenta for color blindness. We also highlighted the testis (TE) and seminal vescicle structures in the images to make morphological changes more visible.  

(2) The penetrance of morphological changes observed in RT development is also unclear and needs to be rigorously quantified for data in Figures 2, 5, and 7. 

We now included quantification for cell biological phenotypes which are generally with 100% penetrance. The percentage of the penetrance and the number of animals used are indicated in each corresponding image.  

Reviewer #2 (Recommendations for the authors): 

Major Points 

(1) Lines 193- 220 I would strongly suggest pointing out the obvious shape defects of the testes visible in Figure 2A ("Spheres" instead of "Spirals"). These are probably a direct consequence of a lack in the epithelial connection that myotubes require to migrate onto the testis (in a normal way) as depicted in the cartoons, allowing the testis to adopt a spiral shape through myotube-sculpting (Bischoff et al., 2021), further confirming the authors' findings! 

Good point. In the revised text, we have added more description of the testis shape defects and pointed out a potential contribution from compromised myotube migration.   

(2) Line 216: "Often separated from each other". Here it would be important to mention how often. If the authors cannot quantify that from existing data, I suggest carrying it out in adult/pharate adult genital tracts (if there is no strong survivor bias due to the lethality of stronger affected animals), as this is much easier than timing prepupae. This should be a quick and easy experiment. 

Because it is hard to tell whether the separation of the SV and TE was caused by developmental defects or sometimes could be due to technical issues (bad dissection), we now change the description to, “control animals always showed connected TE and SV, whereas ham mutant TE and SV tissues were either separated from each other, or appeared contacted but with the epithelial tubes being discontinuous (Fig. 2B).” Additionally, we quantified the disconnection phenotype, which is 100% penetrance in 18 mutant animals. This quantification is now included in the figure. 

(3) Lines 289-305, Figure 3. I could only find how many replicates were analyzed in the RNAseq/CUT&Tag experiments in the Material & Methods section. I would add that at least in the figure legends, and perhaps even in the main text. Most importantly, I would add a Principal Component Analysis (one for RNAseq and one for the CUT&TAG experiment), to demonstrate the similarity of biological replicates (3x RNaseq, 4x Cut&Tag) but also of the technical replicates (RNAseq: wt & wt/dg, ham/ham & ham/df, GD & TE; CUT&TAG: Antibody & GFP-Antibody, TG&TE...). This should be very easy with the existing data, and clearly demonstrate similarities & differences in the different types of replicates and conditions. 

Principle component analysis and its description are now added to Supplementary Fig 6 and the main text respectively. 

(4) Line 321; Supplementary Table 1: In the table, I cannot find which genes are down- or upregulated - something that I think is very important. I would add that, and remove the "color" column, which does not add any useful information. 

In Supplementary table 1, the first sheet includes upregulated genes while the second sheet includes downregulated genes. We removed the column “color” as suggested.  

(5) Line 409: SCRINSHOT was carried out with candidate genes from the screen. One gene I could not find in that list was the potential microtubule-actin crosslinker shot. If shot knockdown caused a phenotype, then I would clearly mention and show it. If not, I would mention why a shot is important, nonetheless. 

shot is one of the candidate target genes selected from our RNA-seq and Cut&Tag data. However, in the RNAi screen, knocking down shot with the available RNAi lines did not cause any obvious phenotype. These could be due to inefficient RNAi knockdown or redundancy with other factors. We anyway wanted to examine shot expression pattern in the developing RS, give the important role of shot in epithelial fusion (Lee S., 2002). Using SCRINSHOT, we could detect epithelial-specific expression of shot, implying its potential function in this context. We now revised the text to clarify this point. 

Minor points 

(1) Cartoons in Figure 1: The cartoons look like they were inspired by the cartoon from Kozopas et al., 1998 Fig. 10 or Rothenbusch-Fender et al., 2016 Fig 1. I think the manuscript would greatly profit from better cartoons, that are closer to what the tissue really looks like (see Figure 1H, 2G), to allow people to understand the somewhat complicated architecture. The anlagen of the seminal vesicles/paragonia looks like a butterfly with a high columnar epithelium with a visible separation between paragonia/seminal vesicles (upper/lower "wing" of the "butterfly"). Descriptions like "unseparated" paragonia/seminal vesicle anlagen, would be much easier to understand if the cartoons would for example reflect this separation. It would even be better to add cartoons of the phenotypic classes too, and to put them right next to the micrographs. (Another nitpick with the cartoons: pigment cells are drastically larger and fewer in number (See: Bischoff et al., 2021 Figure 1E & MovieM1).) 

Thanks for the suggestion. We have updated Figure 1 by adding additional illustrations showing the accessory gland and seminal vesicle structures in the pupal stage and changing the size of pigment cells.

(2) Line 95-121 I would also briefly introduce PR domains, here. 

We have added a brief descripition of the PR domains.

(3) Line 152, 158, 160, 162. When first reading it, I was a bit confused by the usage of the word sensory organ. I would at least introduce that bristles are also known as external mechanosensory organs. 

We have now revised the description to “mechano-sensory organ”.

eg. Line 184, 194, and many more. Most times, the authors call testis muscle precursors "myoblasts". This is correct sometimes, but only when referring to the stage before myoblast-fusion, which takes place directly before epithelial fusion (28 h APF). Postmyoblast-fusion (eg. during migration onto the testis), these cells should be called myotubes or nascent myotubes, as the fly muscle community defined the term myoblast as the singlenuclei precursors to myotubes. 

We have now revised the description accordingly.  

(4) Line 217/Figure 2B. It looks like there is a myotube bridge between the testis and the genital disc. I would point that out if it's true. If the authors have a larger z-stack of this connection, I suggest creating an MIP, and checking if there are little clusters of two/three/four nuclei packed together. This would clearly show that the cells in between are indeed myotubes (granted that loss of ham does not introduce myoblast-fusion-defects). 

We do not have a Z-stack of this connection, and thus can not confirm whether the cells in this image are myotubes. However, we found that mytubes can migrate onto the testis and form the muscular sheet in the ham mutant despite reduced myotube density. At the junction there are myotubes, suggesting that loss of ham does not introduce myoblast-fusion defects. These results are now included in the revised manuscript, supplementary Fig. 5 C-D.

(5) Line 231/Supplementary Fig. 3C-G: I would add to the cartoons, where the different markers are expressed. 

We have added marker gene expression in the cartoons.

(6) Line 239. I don't see what Figure 1A/1H refers to, here. I would perhaps just remove it. 

Yes, we have removed it.

(7) Line 232. I would rephrase the beginning of the sentence to: Our data suggest Ham to be... 

Yes, we have revised it.

(8) Line 248-250/Figure 2F. Clonal analyses are great, but I think single channels should be shown in black and white. Also, a version without the white dashed line should be shown, to clearly see the differences between wt and ham-mutant cells. 

Now single channel images from the green and red images are presented in Supplementary Figures. This particular one is in Supplementary Figure 3B. 

(9) Line 490. The Toll-9 phenotype was identified on the sterility effect/lack-of-spermphenotype alone, and it was deduced, that this suggests connection defects. By showing the right focus plane in Fig S8B (lower right), it should be easy to directly show whether there is a connection defect or not. Also, one would expect clearer testis-shaping defects, like in ham-mutants, as a loss of connection should also affect myotube migration to shape the testis. This is just a minor point, as it only affects supplementary data with no larger impact on the overall findings, even if Toll-9 is shown not to have a defect, after all. 

We find that scoring defects at the junction site at the adult stage is difficult and may not be always accurate. Instead, we score the presence of sperms in the SV, which indirectly but firmly suggests successful connection between the TE and SV. We have now included a quantification graph, showing the penetrance of the phentoype in the new Supplementary Fig.14C. There were indeed morphological defects of TE in Toll-9 RNAi animals. We now included the image and quantification in the new Supplementary Fig.14B.

Author response:

The following is the authors’ response to the original reviews

Response to the public reviews:

We are very pleased to see these positive reviews of our preprint.

Reviewers 1 and 3 raise issues around PIP-PP1 interactions.

(1) Role of the “RVxF-ΦΦ-R-W string”

Most PIPs interact with the globular PP1 catalytic core through short linear interaction motifs (SLiMs) and Choy et al (PNAS 2014) previously showed that many PIPs interact with PP1 through conserved trio of SLiMs, RVxF-ΦΦ-R, which is also present in the Phactrs.

Previous structural analysis showed the trajectory of the PPP1R15A/B, Neurabin/Spinphilin (PPP1R9A/B), and PNUTS (PPP1R10) PIPs across the PP1 surface encompasses not only the RVxF-ΦΦ-R trio, but also additional sequences C-terminal to it (Chen et al, eLife, 2015). This extended trajectory is maintained in the Phactr1-PP1 complex (Fedoryshchak et al, eLife (2020). Based on structural alignment we proposed the existence of an additional hydrophobic “W” SLiM that interacts with the PP1 residues I133 and Y134.

The extended “RVxF-ΦΦ-R-W” interaction brings sequences C-terminal to the “W” SLiM into the vicinity of the hydrophobic groove that adjoins the PP1 catalytic centre. In the Phactr1/PP1 complex, these sequences remodel the groove, generating a novel pocket that facilitates sequence-specific substrate recognition.

This raises the possibility that sequences C-terminal to the extended “RVxF-ΦΦ-R-W string” in the other complexes also confer sequence-specific substrate recognition, and our study aims to test this hypothesis. Indeed, the hydrophobic groove structures of the Neurabin/Spinophilin/PP1 and Phactr1/PP1 complexes differ significantly (Ragusa et al, 2010; see Fedoryshchak et al 2020, Fig2 FigSupp1).

(2) Orientation of the W side chain

Reviewer 1 points out that in the substrate-bound PP1/PPP1R15A/Actin/eIF2 pre-dephosphorylation complex the W sidechain is inverted with respect to its orientation in  PP1-PPP1R15B complex (Yan et al, NSMB 2021). The authors proposed that this may reflect the role of actin in assembly of the quaternary complex. This does not necessarily invalidate the notion that sequences C-terminal to the “W” motif might play a role in actin-independent substrate recognition, and we therefore consider our inclusion of the R15A/B fusions in our analysis to be reasonable.

(3) Conservation of W

The motif ‘W’ does not mandate tryptophan - Phactrs and PPP1R15A/B indeed have W at this position but Neurabin/spinophilin contain VDP, which makes similar interactions. Similarly the “RVxF” motifs in Phactr1, Neurabin/Spinophilin, PPP1R15A/B and PNUTS are LIRF, KIKF, KV(R/T)F and TVTW respectively.

In our revision, we will present comparisons of the differentially remodelled/modified PP1 hydrophobic groove in the various complexes, discuss the different orientations of the tryptophan in the previously published PPP1R15A/PP1 and PPP1R15B/PP1 structures. We will also address the other issues raised by the referees.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Comments and suggestions for revisions

(1) The authors do not provide strong evidence that the interactions of the 'W' of the RVxF- øø -R-W string with the hydrophobic groove of PP1 is conserved in PIPs. Whereas the RVxF motif is well conserved and validated since its discovery in 1997, as are the øø - (an extension of the RVxF motif), and the 'R', the conservation of the Trp residue in the RVxF-øø-R-W string is not conserved.

We did not mean to imply that the W motif is conserved amongst all PIPs.

Most PIPs interact with the globular PP1 catalytic core through short linear interaction motifs (SLiMs). Choy et al (PNAS 2014) previously showed that many PIPs interact with PP1 through a conserved trio of SLiMs, RVxF-ΦΦ-R, which is also present in the Phactrs.

Previous structural analysis showed that the PPP1R15A/B, Neurabin/Spinophilin (PPP1R9A/B), and PNUTS (PPP1R10) PIPs share a trajectory across the PP1 surface that encompasses not only the RVxF-ΦΦ-R SLIMs, but also additional sequences C-terminal to the R SLIM (Chen et al, eLife, 2015). This trajectory is also shared by the Phactr1-PP1 complex (Fedoryshchak et al, eLife, 2020). Based on this structural alignment we proposed the existence of an additional hydrophobic “W” SLiM that interacts with the PP1 residues I133 and Y134 (See Fedoryshchak et al, 2020, Figure 1 figure supplement 2).

Introduction, paragraph 2 is rewritten to make this clearer.

The sequence and positions of W differ in amino acid type and position relative to the RVxF-øø-R string.

The motif ‘W’ does not mandate tryptophan, it is our name for a common structurally aligned motif: although the Phactrs and PPP1R15A/B indeed have W at this position, Neurabin and spinophilin contain VDP, which nevertheless makes similar interactions. Similarly the _“_RVxF” motifs in Phactr1, Neurabin/Spinophilin, PPP1R15A/B and PNUTS are LIRF, KIKF, KV(R/T)F and TVTW respectively.

In the Discussion the authors state that the hydrophobic groove of PP1 is remodelled by Neurabin. However, details of this are not described or shown in the manuscript.

The shared trajectory determined by the RVxF-øø-R-W string brings the sequences C-terminal to the W SLIM into the vicinity of the PP1 hydrophobic groove. In the Phactr1/PP1 holoenzyme this generates a novel pocket required for substrate recognition (Fedoryshchak et al, 2020). These observations raised the possibility that sequences C-terminal to the “W” motif in the other RVxF-øø-R-W PIPs also play a role in substrate recognition.

Introduction paragraph 3 now cites a new Figure 1-S2, which shows how the hydrophobic groove is remodelled in the various different PIP/PP1 complexes. A revised Figure 1A now indicates the hydrophobic residues defining the hydrophobic groove by grey shading.

(2) To add to the confidence of the structure, the authors should include a 2Fo-Fc simulated annealing omit map, perhaps showing the R and W interactions of the RVxF-øø-R-W string.

This is now included as new Figure 6 Figure supplement 1. Note that in Neurabin, the W motif is VDP, where the valine and proline sidechains interact similarly to the tryptophan (see also new Figure 1-S2G,H).

We also add a new supplementary Figure 6-S1 comparing our PBM-liganded Neurabin PDZ domain with the previously published unliganded structure (Ragusa et al 2010).

(3) Page 16. The authors state that spinophilin remodels the PP1 hydrophobic groove differently from Phactrs. Arguably spinophilin does not remodel the PP1 hydrophobic groove at all. There are no contacts between spinophilin and the PP1 hydrophobic groove in the spinophilin-PP1 structure, correlating with the absence of 'W" in the RVxF-øø-R-W string in spinophilin.

The VDP sequence corresponding to the W motif in spinophilin and neurabin makes analogous contacts to those made by the W in Phactr1 (see Fedoryshchak et al 2020).

Remodelling is meant in the sense of altering the structure of the major groove by bringing new sequences into its vicinity rather than necessarily directly interacting with it. The spinophilin/PP1 and Phactr/PP1 hydrophobic grooves are compared in new Figure 1-S2 (see also Fedoryshchak et al 2020, Figure 2 figure supplement 1)

(4) Page 8. For the cell-based/proteomics-dephosphorylation assay in Figure 2, it isn't clear why there were no dephosphorylation sites detected for the PPP1R15A/B-PP1 fusion (except PPP6R1 S531 for PPP1R15B). One might have expected a correlation with PP1 alone. Does this imply that PPP1R15A/B are inhibiting PP1 catalytic activity? Was the activity tested in vitro?

The R15A/B data are compared to average abundance of all the phosphosites in the dataset, including those of PP1.

We have not tested for a general inhibitory effect of R15A/B on PP1 activity. Many PIPs including R15A/B do occlude one or more of the PP1 substrate groove and therefore generally act as inhibitors of PP1 activity against some potential substrates, while enhancing activities against others.

Other points 

(4) Figure S1: Colour sequence similarities/identities.

Done

(6) Figures: Structure figures lacked labels:

Figure 1A, label PP1, Phactrs etc.

Done

Figure 6, label PP1, Neurabin, previous Neurabin structure (Fig. 6C), hydrophobic groove, PDZ domain, etc.

Done

(7) Statistical analysis. p values should be shown for data in:

Figure 5.

To avoid cluttering the Figure, a new sheet, “statistical significance” has been added to Supplementary Table 3, summarizing the analysis.

Figure 1.

Figure amended (now figure 1-S1).

(8) Some inconsistency with labels, eg '34-WT' used in Fig. 5C, whereas '34A-WT' (better) in Methods.

Now changed to 34A etc where used.

(9) Page 6. PPP1R9A/B is not shown in Figure 1A and Figure S1A.

PPP1R9A/B are Neurabin and spinophilin - now clarified in Introduction paragraph 2, Results paragraph 1, Discussion paragraph 1.

(10) Page 7: lines 4, 'site' not 'side'.

Done

(11) Page 9: DTL and CAMSAP3 were found to be dephosphorylated in the PP1-Neurabin/spinophilin screen. Are these PDZ-binding proteins?

Neither DTL nor CAMSAP3 contain C-terminal hydrophobic residues characteristic of classical PBMs. Sentence added in Discussion, paragraph 5

(12) Page 12 and Figure 5 and S5: The synthetic p4E-BP1 and IRSp53WT peptides with PBM should be given more specific names to indicate the presence of the PBM.

We have renamed 4E-BP1^WT and IRSp53^WT to 4E-BP1^PBM and  IRSp53^PBM respectively, emphasising the inclusion of the wildtype or mutated PBM from 4E-BP1 on these peptides.

Text, Figure 5, and Figure S5 all revised accordingly.

(13) Give PDB code for spinophilin-PP1 complex coordinates shown in Figure 6C.

PDB codes for the various PIP/PP1 complexes now given in new Figure 1-S2 and revised Figure 6C.

Reviewer #2 (Recommendations for the authors):

The work undertaken by the authors is extensive and robust, however, I believe that some improvement in the writing and some detailed explanation of certain results sections would help with the presentation of the work and clarity for the readers.

(1) The introduction should contain more information about the interaction between PP1 and Neurabin, given that this is the focus of the paper. This would give the reader the necessary background required to follow the paper.

Introduction paragraph 2 revised to describe the different SLIMs in more detail. New Figure 1-S2 shows detail of the different remodelled hydrophobic grooves in the various PIP/PP1 complexes.

(2) More information on PP1-IRSp53L460A has to be added before discussing results in S1B.

Sentence explaining that IRSp53 L460 docks with the remodelled PP1 hydrophobic groove in the Phactr1/PP1 holoenzyme added in Results paragraph 2.

(3) Page 6: "as expected, the +5 residue L460A mutation, which impairs dephosphorylation by the intact Phactr1/PP1 holoenzyme, impaired sensitivity to all the fusions, indicating that they recognise phosphorylated IRSp53 in a similar way (Figure S1B)". Statistics between IRSp53 and IRSp53L460A across PP1-PIPs need to be conducted before concluding the above. From the graph and the images, the impairment to dephosphorylation is not convincing.

For each of the four PP1-Phactr fusions, the IRSp53 L460A peptide shows significantly less reactivity than the IRSp53WT peptide (p<0.05 for each fusion).

Since the proteomics studes in Figure 2 show that the substrate specificity of the four PP1-Phactr1 fusions is virtually identical, we combined the data for the four different fusions. The IRSp53 L460A peptide shows significantly less reactivity than the IRSp53WT peptide in this analysis (p< 0.0001). This result shown in revised Figure S1B and legend.

(4) mCherry-4E-BP1(118+A), in which an additional C-terminal alanine should still allow TOSmediated phosphorylation, but prevent PDZ interaction. Does 4EBP1 (118+A) actually prevent interaction between PP1-Neurabin? This interaction needs to be validated, especially since spinophilin was shown to bind to multiple regions of PP1.

It is not clear what the referee is asking for here. The biochemical analysis in Figure 4C shows that the C-terminus of 4E-BP1 constitutes a classical PBM. The X-ray crystallography in Figure 6 confirms this, demonstrating H-bond interactions between the 4E-BP1 C-terminal carboxylate and main chain amides of L514, G515 and I516.

We consider the possibility that the 4E-BP1(118+A) mutant inhibits the activity of PP1-neurabin via a mechanism other than direct blocking 4E-BP1 / PDZ interaction to be unlikely for the following reasons:

(1) Addition of a C-terminal alanine will disrupt the PBM interaction because the extra residue sterically blocks access to the PBM-binding groove. This is the most parsimonious explanation, and is based on our solid structural and biochemical evidence that the 4E-BP1 C-terminus is a classical PBM.

(2) Alphafold3 modelling predicts Neurabin PDZ / 4E-BP1 PBM interaction with high confidence (shown in Figure 6-S2E), but it does not predict any PDZ interaction with 4E-BP1(118+A). Note added in Figure 6-S2 legend.

(3) Recognition of the 4E-BP1(118+A) mutation without loss of binding affinity would require that the mutant becapable of binding formally equivalent to recognition of an “internal” PDZ-binding peptide. Recognition of such “internal peptides” is dependent on their adopting a specifically constrained conformation, which typically requires reorganisation of the PDZ carboxylate-binding GLGF loop. Such “internal site” recognition typically involves more than one residue C-terminal to the conventional PDZ “0” position (see Penkert et al NSMB 2004, doi:10.1038/nsmb839; Gee et al JBC 1998, DOI: 10.1074/jbc.273.34.21980; Hillier et al 1999, Science PMID: 10221915).

(5) It is nice to see that the various PP1-Phactr fusions have around 60% substrate overlap between them. Would it be possible to compare these results with previously published mass spec data of Phactr1XXX from the group? There is mention of some substrates being picked up, but a comparison much like in Figure 2E would be more informative about the extent to which the described method captures relevant information.

This is difficult to do directly as the PP1-Phactr fusion data are from human cells while that in Fedoryshchak et al 2020 is from mouse.

However, manual curation shows that of the 28 top hits seen in our previous analysis of Phactr1XXX in NIH3T3 cells, 18 were also detectable in the HEK293 system; of these, 13 were also detected as as PP1-Phactr fusion hits. Data summarised in new Figure 2-S1C. Text amended in Results, “Proteomic analysis...”, paragraph 2.

(6) Figure 3D Why are the levels of pT70, pT37/46 and total protein in vector controls much lower as compared to 0nM Tet in PP1-Neurabin conditions? It is also weird that given total protein is so low, why are the pS65/101 levels high compared to the rest?

We think it likely these phenomena reflect a low level expression of PP1-Neurabin expression in uninduced cells. Now noted in Figure 3D legend, basal PP1-Neurabin expression shown in new Figure 3-S1C. This alters the relative levels of the different species detected by the total 4E-BP1 antibody in favour of the faster migrating forms, which are less phosphorylated than the slower ones, and the total amount increases about 2-fold (Figure 3D, compare 0nM Tet lanes).

The altered p65/101-pT70 ratio is also likely to reflect the leaky PP1-Neurabin expression, since the relative intensities of the various phosphorylated species are dependent on both the relative rates of phosphorylation and dephosphorylation. Expression of a phosphatase would therefore be expected to differentially affect the phosphorlyation levels of different sites according to their reactivity.

(7) Figure 3E: Does inhibiting mTORC further reduce translation when PP1-Neurabin is expressed? If this is the case, this might suggest that they might not necessarily be mTORC inhibitors?

We have not done this experiment. Since Rapamycin cannot be guaranteed to completely block 4E-BP1 phosphorylation, and PP1-Neurabin cannot be guaranteed to completely dephosphorylate 4E-BP1, any further reduction upon their combination would be hard to interpret.

(8) Substrate interactions with the remodelled PP1 hydrophobic groove do not affect PP1-Neurabin specificity. Is there evidence that PP1-Neurabin remodels the hydrophobic groove? Is it not possible that Neurabin does not remodel the PP1 groove to begin with and hence there is no effect observed with the various mutants? If this is not the case, it should be explained in a bit more detail.

Comparison of the Neurabin/PP1 and Phactr1/PP1 structures shows that the hydrophobic groove is remodelled differently in the two complexes. Now shown in new Figure 1-S2B,C,G.

(9) Figure 5B has a lot of interesting information, which I believe has not been discussed at all in the results section.

To help interpretation of the enzymology in Figure 5 we have renamed 4E-BP1WT and IRSp53WT to 4E-BP1PBM and IRSp53PBM respectively, emphasising the inclusion of the wildtype or mutated PBM from 4E-BP1 on these peptides. Text in Results, “PDZ domain interaction…”, paragraph 1, and Figures 5 and S5 revised accordingly.

Why does the 4E-BP1Mut affect catalytic efficiency of PP1 alone when compared with WT, while no difference is observed with IRSp53WT and mutant?

We do not understand the basis for the differential reactivity of 4E-BP1PBM and 4E-BP1MUT with PP1 alone; we suspect that it reflects the hydrophobicity change resulting from the MDI -> SGS substitution. However this is unlikely to be biologically significant as PP1 is sequestered in PIP-PP1 complexes.

Importantly, the two PP1 fusion proteins behave consistently in this assay – the presence of the intact PBM increases reactivity with PP1-Neurabin, but has no effect on dephosphorylation by PP1-Phactr1.

Why does PP1 alone not have a difference between IRSp53WT and mutant, while PP1-Neurabin does have a difference?

This is due to the presence of the PBM in IRSp53WT (now renamed IRSp53PBM), which affects increases affinity for PP1 Neurabin, but not PP1 alone. Likewise, PP1-Phactr1, which does not possess a PDZ domain, is also unaffected by the integrity of the PBM.

(7) “Strikingly, alanine substitutions at +1 and +2 in 4E-BP1WT increased catalytic efficiency by both fusions, perhaps reflecting changes at the catalytic site itself (Figure 5E, Figure S5E)”. This could be expanded upon, because this suggests a mechanism that makes the substrate refractory to PDZ/hydrophobic groove remodelling?

We favour the idea that this reflects a requirement to balance dephosphorylation rates between the multiple 4E-BP1 phosphorylation sites, especially if multiple rounds of dephosphorylation occur for each PBM—PDZ interaction. Additional sentences added in Discussion paragraph 7.

(8) Typographical errors and minor comments:

a) PIPs can target PP1 to specific subcellular locations, and control substrate specificity through autonomous substrate-binding domains, occupation or extension of the substrate grooves, or modification of PP1 surface electrostatics.

b) Phosphophorylation side site abundances within triplicate samples from the same cell line were comparable between replicates (Figure 2B).

c) While the alanine substitutions had little effect, conversion of +4 to +6 to the IRSp534E-BP1 sequence LLD increased catalytic efficiency some 20-fold (Figure 5C, Figure S5C). 

d) Figure 3E labels are not clear. The graph can be widened to make the labels of the conditions clearer.

All corrected

Reviewer #3 (Recommendations for the authors):

This was a very well-written manuscript.

However, I was looking for a summary mechanistic figure or cartoon to help me navigate the results.

I noted a few typos in the text.

New summary Figure 5-S2 added, cited in results, and discussed in Discussion paragraph 6,7.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This article presents a meta-analysis that challenges established abundance-occupancy relationships (AORs) by utilizing the largest known bird observation database. The analysis yields contentious outcomes, raising the question of whether these findings could potentially refute AORs.

We thank the Reviewer for their positive comments.

Strengths:

The study employed an extensive aggregation of datasets to date to scrutinize the abundance-occupancy relationships (AORs).

We thank the Reviewer for their positive comments.

Weaknesses:

While the dataset employed in this research holds promise, a rigorous justification of the core assumptions underpinning the analytical framework is inadequate. The authors should thoroughly address the correlation between checklist data and global range data, ensuring that the foundational assumptions and potential confounding factors are explicitly examined and articulated within the study's context.

We thank the Reviewer for these comments. We agree that more justification and transparency is needed of the core assumptions that form the foundation of our methods. In our revised version, we have taken the following steps to achieve this:

- Altered the title to be more explicit about the core assumptions, which now reads: “Local-scale relative abundance is decoupled from global range size”

- We have added more details on why and how we treat global range size as a measure of ‘occupancy.’

- We have added a section that discusses the limitations of using eBird relative abundance

Reviewer #2 (Public Review):

Summary:

The goal is to ask if common species when studied across their range tend to have larger ranges in total. To do this the authors examined a very large citizen science database which gives estimates of numbers, and correlated that with the total range size, available from Birdlife. The average correlation is positive but close to zero, and the distribution around zero is also narrow, leading to the conclusion that, even if applicable in some cases, there is no evidence for consistent trends in one or other direction.

We thank the Reviewer for these comments.

Strengths:

The study raises a dormant question, with a large dataset.

We thank the Reviewer for these comments. We intended to take a longstanding question and attempt to apply novel datasets that were not available mere decades ago. While we do not imply that we have ‘solved’ the question, we hope this work highlights the potential for further interrogation using these large datasets.

Weaknesses:

This study combines information from across the whole world, with many different habitats, taxa, and observations, which surely leads to a quite heterogeneous collection.

We agree that there is a heterogeneous collection of data across many habitats, taxa, and observations. However, rather than as a weakness, we see this as a significant strength. Our work assumes we are averaging over this variability to assess for a large-scale pattern in the relationship - something that was potentially a limitation of previous work, as these large datasets were often focused on particular contexts (e.g., much work focused solely on the UK), which we believe could limit some of the generalizability of the previous work. However, the reviewer makes a fair point in regard to the heterogeneity of data collection. We have now added some text in the discussion which is explicit about this - see the new section named “Potential limitations of current work and future work –-although our findings challenge some long-held assumptions about the consistency of the abundance-occupancy relationship, our work only deals with interspecific AORs among birds, synthesizing observations of potentially heterogeneous locations, context and quality”.

First, scale. Many of the earlier analyses were within smaller areas, and for example, ranges are not obviously bounded by a physical barrier. I assume this study is only looking at breeding ranges; that should be stated, as 40% of all bird species migrate, and winter limitation of populations is important. Also are abundances only breeding abundances or are they measured through the year? Are alien distributions removed?

Second, consider various reasons why abundance and range size may be correlated (sometimes positively and sometimes negatively) at large scales. Combining studies across such a large diversity of ecological situations seems to create many possibilities to miss interesting patterns. For example:

(1) Islands are small and often show density release.

See comment below.

(2) North temperate regions have large ranges (Rapoport's rule) and higher population sizes than the tropics.

See comment below.

(3) Body size correlates with global range size (I am unsure if this has recently been tested but is present in older papers) and with density. For example, cosmopolitan species (barn owl, osprey, peregrine) are relatively large and relatively rare.

See comment below.

(4) In the consideration of alien species, it certainly looks to me as if the law is followed, with pigeon, starling, and sparrow both common and widely distributed. I guess one needs to make some sort of statement about anthropogenic influences, given the dramatic changes in both populations and environments over the past 50 years.

See comment below. We also added a sentence in the methods that highlighted we did not remove alien ranges and provided reasons why. Still, we do acknowledge the dramatic changes in populations and environments over the past 50 years (see the new section  “Potential limitations of current work and futur work”)

(5) Wing shape correlates with ecological niche and range size (e.g. White, American Naturalist). Aerial foraging species with pointed wings are likely to be easily detected, and several have large ranges reflecting dispersal (e.g. barn swallow).

We agree that all of the points above are interesting data explorations. As said above, our main purpose was to highlight the potential for further interrogation using these large datasets. However, we have added some additional text in the discussion that explicitly mentions/encourages these additional data explorations. We hope people will pick up on the potential for these data and explore them further.

Third, biases. I am not conversant with ebird methodology, but the number appearing on checklists seems a very poor estimate of local abundance. As noted in the paper, common species may be underestimated in their abundance. Flocking species must generate large numbers, skulking species few. The survey is often likely to be in areas favorable to some species and not others. The alternative approach in the paper comes from an earlier study, based on ebird but then creating densities within grids and surely comes with similar issues.

We agree that if we were interested in the absolute abundance of a given species, the local number on an eBird checklist would be a poor representation. However, our study aims not to estimate absolute abundance but to examine relative abundance among species on each checklist. By focusing on relative abundance, we leverage eBird data's strengths in detecting the presence and frequency of species across diverse locations and times, thereby capturing community composition trends that can provide meaningful insights despite individual checklist biases. This approach allows us to assess the comparative prominence of species in the community as reported by the observer, providing a consistent metric of relative abundance. Despite detectability biases, the structure of eBird checklists reflects the observer’s encounter rates with each species under similar conditions, offering a valuable snapshot of relative species composition across sites and times. The key to our assumption is that these biases discussed are not directional and, therefore, random throughout the sampling process, which would translate to no ‘real’ bias in our effect size of interest.

Range biases are also present. Notably, tropical mountain-occupying species have range sizes overestimated because holes in the range are not generally accounted for (Ocampo-Peñuela et al., Nature Communications). These species are often quite rare, too.

We thanks the reviewer for pointing to this issue and reference. We included a discussion on these biases in our limitations section and reference Ocampo-Peñuela et al. to emphasize the need for improved spatial resolution in range data for more accurate AOR assessments.”More precise range-size estimates would also improve the accuracy of AOR assessments, since species range data are often overestimated due to the failure to capture gaps in actual distributions ”

Fourth, random error. Random error in ebird assessments is likely to be large, with differences among observers, seasons, days, and weather (e.g. Callaghan et al. 2021, PNAS). Range sizes also come with many errors, which is why occupancy is usually seen as the more appropriate measure.

If we consider both range and abundance measurements to be subject to random error in any one species list, then the removal of all these errors will surely increase the correlation for that list (the covariance shouldn't change but the variances will decrease). I think (but am not sure) that this will affect the mean correlation because more of the positive correlations appear 'real' given the overall mean is positive. It will definitely affect the variance of the correlations; the low variance is one of the main points in the paper. A high variance would point to the operation of multiple mechanisms, some perhaps producing negative correlations (Blackburn et al. 2006).

We agree random errors can affect estimates, but as we wrote above, random errors, regardless of magnitudes, would not bias estimates. After accounting for sampling error (a part of random errors), little variance is left to be explained as we have shown in the MS. This suggests that many of the random errors were part of the sampling errors. And this is where meta-analysis really shines.

On P.80 it is stated: "Specifically, we can quantify how AOR will change in relation to increases in species richness and sampling duration, both of which are predicted to reduce the magnitude of AORs" I haven't checked the references that make this statement, but intuitively the opposite is expected? More species and longer durations should both increase the accuracy of the estimate, so removing them introduces more error? Perhaps dividing by an uncertain estimate introduces more error anyway. At any rate, the authors should explain the quoted statement in this paper.

It would be of considerable interest to look at the extreme negative and extreme positive correlations: do they make any biological sense?

Extremely high correlations would not make any biological sense if these observations were based on large sample sizes. However, as shown in Figure 2, all extreme correlations come from small sample sizes (i.e., low precision), as sampling theory expects (actually our Fig 2 a text-book example of the funnel shape). Therefore, we do not need to invoke any biological explanations here.

Discussion:

I can see how publication bias can affect meta-analyses (addressed in the Gaston et al. 2006 paper) but less easily see how confirmation bias can. It seems to me that some of the points made above must explain the difference between this study and Blackburn et al. 2006's strong result.

We agree. Now, we extended an explanation of why confirmation bias could result in positive AOR. Yet, we point out confirmation bias is a very common phenomena which we cite relevant citations in the original MS. The only way to avoid confirmation bias is to conduct a study blind but this is not often possible in ecological work.

“Meta-research on behavioural ecology identified 79 studies on nestmate recognition, 23 of which were conducted blind. Non-blind studies confirmed a hypothesis of no aggression towards nestmates nearly three times more often. It is possible that confirmation bias was at play in earlier AOR studies.”

Certainly, AOR really does seem to be present in at least some cases (e.g. British breeding birds) and a discussion of individual cases would be valuable. Previous studies have also noted that there are at least some negative and some non-significant associations, and understanding the underlying causes is of great interest (e.g. Kotiaho et al. Biology Letters).

We agree. And yes, we pointed out these in our introduction.

Reviewer #3 (Public Review):

Summary:

This paper claims to overturn the longstanding abundance occupancy relationship.

Strengths:

(1) The above would be important if true.

(2) The dataset is large.

We have clarified this point by changing the title to emphasize that we do not suggest overturning AORs entirely but instead provide a refined view of the relationship at a global scale. Our results suggest a weaker and more context-dependent AOR than previously documented. We hope our revised title and additional clarifications in the text convey our intent to contribute to a more nuanced understanding rather than a whole overturning of the AOR framework.

Weaknesses:

(1) The authors are not really measuring the abundance-occupancy relationship (AOR). They are measuring abundance-range size. The AOR typically measures patches in a metapopulation, i.e. at a local scale. Range size is not an interchangeable notion with local occupancy.

We have refined this in our revision to be more explicitly focused on global range size. However, we note that the classic paper by Bock and Richlefs (1983, Am Nat) also refers to global (species entire) range size in the context of the AOR. Importantly, Bock and Richlefs pointed out the importance of using species’ entire ranges; without such uses, there will be sampling artifacts creating positive AORs when using arbitrary geographical ranges, which were used in some studies of AORs. So we highlight that our work is well in line with the previous work, allowing us to question the longstanding macroecological work. One of the issues of AOR has been how to define occupancy and global range size, which provides a relatively ambiguous measure, which is why we used this measure.

(2) Ebird is a poor dataset for this. The sampling unit is non-standard. So abundance can at best be estimated by controlling for sampling effort. Comparisons across space are also likely to be highly heterogenous. They also threw out checklists in which abundances were too high to be estimated (reported as "X"). As evidence of the biases in using eBird for this pattern, the North American Breeding Bird Survey, a very similar taxonomic and geographic scope but with a consistent sampling protocol across space does show clear support for the AOR.

Yes, we agree the sampling unit is non-standard. However, this is a significant strength in that it samples across much heterogeneity (as discussed in response to Reviewer 2, above). We were interested in relative abundance and not direct absolute abundance per se, which is accurate, especially since we did control for sampling effort.

We appreciate the reviewer’s attention to our data selection criteria. We excluded checklists containing ‘X’ entries to minimize biases in our abundance estimates. The 'X' notation is often used for the most common species, reflecting the observer's identification of presence without specifying a count. This approach was chosen to avoid disproportionately inflating presence data for these abundant species, which could distort the relative abundance calculations in our analysis. By excluding such checklists, we aimed to retain consistency and ensure that local abundance estimates were representative across all species on each checklist. We have revised our manuscript to clarify this methodological choice and hope this explanation addresses the reviewer’s concern. We modified our text in the methods to make the entries ‘X’ clearer (see the Method section).

(3) In general, I wonder if a pattern demonstrated in thousands of data sets can be overturned by findings in one data set. It may be a big dataset but any biases in the dataset are repeated across all of those observations.

Overturning a major conclusion requires careful work. This paper did not rise to this level.

We appreciate the reviewer’s caution regarding broad conclusions based on a single dataset, even one as large as eBird. Our intention was not to definitively overturn the abundance-occupancy relationship (AOR) but to re-evaluate it with the most extensive and globally representative dataset currently available. We recognise that potential biases in citizen science data, such as observer variation, may influence our findings, and we have taken steps to address these in our methodology and limitations sections. We see this work as a contribution to an ongoing discourse, suggesting that AOR may be less universally consistent than previously believed, mainly when tested with large-scale citizen science data. We hope this study will encourage additional research that tests AORs using other expansive datasets and approaches, further refining our understanding of this classic macroecological relationship. However, we have left our broad message about instigating credible revolution and also re-examining ecological laws.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

(1) The investigation focuses solely on interspecific relationships among birds; thus, the extrapolation of these conclusions to broader ecological contexts requires further validation.

We have now added this point to our new section: “Although our findings challenge some long-held assumptions about the consistency of the abundance-occupancy relationship, our work only deals with interspecific AORs among birds, so we hope this work serves as a foundation for further investigations that utilize such comprehensive datasets.”

(2) The rationale for combining data from eBird - a platform predominantly representing individual observations from urban North America - with the more globally comprehensive BirdLife International database needs to be substantiated. The potential underrepresentation of global abundance in the eBird checklist data could introduce a sampling bias, undermining the foundational premises of AORs.

We agree with the limitation of ebird sampling coverage, but it should not bias our results. In statistical definitions, bias is directional, and if not directional, it will become statistical noise, making it difficult to detect the signal. In fact, our meta-analyses adjust what statisticians call sampling bias and it is the strength of meta-analysis.

(3) In the full mixed-effect model, checklist duration and sampling variance (inversely proportional to sample size N) are treated as fixed effects. However, these variables are likely to be negatively correlated, which could introduce multicollinearity, inflating standard errors and diminishing the statistical significance of other factors, such as the intercept. This calls into question the interpretation of insignificance in the results.

Multicollinearity is an issue with sample sizes. For example, with small datasets, correlations of 0.5 could be an issue, and such an issue would usually show up as a large SE. We do not have such an issue with ~ 17 million data points. Please refer to this paper.

Freckleton, Robert P. "Dealing with collinearity in behavioural and ecological data: model averaging and the problems of measurement error." Behavioral Ecology and Sociobiology 65 (2011): 91-101.

(4) The observed low heterogeneity may stem from discrepancies in sampling for abundance versus occupancy, compounded by uncertainties in reporting behavior.

If we assume everybody underreports common species or overreports rare species, this could happen. However, such an assumption is unlikely. If some people report accurately (but not others), we should see high heterogeneity, which we do not observe).  We have touched upon this point in our original MS.

(5) The contribution and implementation of phylogenetic comparative analysis remain ambiguous and were not sufficiently clarified within the study.

We need to add more explanation for the global abundance analysis

“To statistically test whether there was an effect of abundance and occupancy at the macro-scale, we used phylogenetic comparative analysis.  This analysis also addresses the issue of positive interspecific AORs potentially arising from not accounting for phylogenetic relatedness among species examined ”

(6) The use of large N checklists could skew the perceived rarity or commonality of species, potentially diminishing the positive correlation observed in AORs. A consistent observer effect could lead to a near-zero effect with high precision.

Regardless of the number of N species in checklists (seen in Fig 2), correlations are distributed around zero. This means there is nothing special about large N checklists. 

(7) The study should acknowledge and discuss any discrepancies or deviations from previous literature or expected outcomes.

We felt we had already done this as we discussed the previous meta-analysis and what we expected from this meta-analysis.  Nevertheless, we have added some relevant sentences in the new version of MS.

In addition to these major points, there are several minor concerns:

(1) Figure 2B lacks discussion, and the metric for the number of observations is not clarified. Furthermore, the labeling of the y-axis appears to be incorrect.

Thank you very much for pointing out this shortcoming. Now, the y-axis label has been fixed and we mention 2B in the main text.

(2) The study should provide a clear, mathematical expression of the multilevel random effect models for greater transparency.

Many thanks for this point, and now we have added relevant mathematical expressions in Table S6.

(3) On Line 260, the term "number of species" should be refined to "number of species in a checklist," ideally represented by a formula for precision.

This ambiguity has been mended as suggested.

Please provide the data and R code linked to the outputs.

The referee must have missed the link (https://github.com/itchyshin/AORs) in our original MS. In addition to our GitHub repository link, we now have added a link to our Zenodo repository (https://doi.org/10.5281/zenodo.14019900).

Reviewer #3 (Recommendations For The Authors):

The authors cite Rabinowitz's 7 forms of rarity paper as a suggestion that previous findings also break the AOR. In fact empirical studies of the 7 forms of rarity typically find that all three forms of rareness vs commonness are heavily correlated (e.g. Yu & Dobson 2000).

We thank the reviewer for drawing attention to Yu & Dobson (2000) and similar studies that find positive correlations among the axes of rarity. Ref 3 is correct in that Rabinowitz’s (1981) framework does not require that local abundance and geographic range size be uncorrelated for every species; instead, it highlights conceptual scenarios where a species may be common locally yet have a restricted distribution (or vice versa).

Empirical analyses such as Yu & Dobson (2000) show that, on average, these axes can be correlated, which may align with conventional AOR findings in some taxonomic groups. However, Rabinowitz’s key insight was that exceptions do occur, so these exceptions demonstrate that strong positive AORs may not be universally applicable. Our results do not claim that Rabinowitz’s framework “breaks” the AOR outright; instead, we use it to underscore that local abundance can, in principle, be “decoupled” from global occupancy.  Whether the correlation found by Yu & Dobson (2000) implies a positive AOR, requires a detailed simulation study, which is an interesting avenue for future research. 

Thus, citing Rabinowitz serves to highlight the potential heterogeneity and complexity of abundance–occupancy relationships rather than to refute every positive correlation reported in the literature. Our findings suggest that when examined at large spatiotemporal scales (with unbiased sampling), the overall AOR signal may be less robust than traditionally believed. This is consistent with Rabinowitz’s view that local abundance and global range can vary along independent axes. Now we added

“Although studies using her framework found positive correlations between species range and local abundance.”

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript assesses the utility of spatial image correlation spectroscopy (ICS) for measuring physiological responses to DNA damage. ICS is a long-established (~1993) method similar to fluorescence correlation spectroscopy, for deriving information about the fluorophore density that underlies the intensity distributions of images. The authors first provide a technical but fairly accessible background to the theory of ICS, then compare it with traditional spot-counting methods for its ability to analyze the characteristics of γH2AX staining. Based on the degree of aggregation (DA) value, the authors then survey other markers of DNA damage and uncover some novel findings, such as that RPA aggregation inversely tracks the sensitivity to PARP inhibitors of different cell lines.

The need for a more objective and standardized tool for analyzing DNA damage has long been felt in the field and the authors argue convincingly for this. The data in the manuscript are in general well-supported and of high quality, and show promise of being a robust alternative to traditional focus counting. However, there are a number of areas where I would suggest further controls and explanations to strengthen the authors' case for the robustness of their ICS method.

Strengths:

The spatial ICS method the authors describe and demonstrate is easy to perform and applicable to a wide variety of images. The DDR was well-chosen as an arena to showcase its utility due to its well-characterized dose-responsiveness and known variability between cell types. Their method should be readily useable by any cell biologist wanting to assess the degree of aggregation of fluorescent tags of interest.

Weaknesses:

The spatial ICS method, though of longstanding history, is not as intuitive or well-known as spot-based quantitation. While the Theory section gives a standard mathematical introduction, it is not as accessible as it could be. Additionally, the values of TNoP and DA shown in the Results are not discussed sufficiently with regard to their physical and physiological interpretation.

We agree that a major limitation in adaption of this approach is a deeper understanding of the theory and results. We have updated the theory section to include further discussion (Page 4 line 132)

The correlation of TNoP with γH2AX foci is high (Figure 2) and suggestive that the ICS method is suitable for measuring the strength of the DDR. The authors correctly mention that the number of spots found using traditional means can vary based on the parameters used for spot detection. They contrast this with their ICS detection method; however, the actual robustness of spatial ICS is not given equal consideration.

We found it difficult to give equal consideration of robustness to ICS. The major limitation of traditional approaches is proper selection of an intensity threshold that is necessary to define and separate foci from background intensity. However, ICS does not employ a threshold, therefore we could not test different thresholding applications in ICS as we did with traditional methods. In our view the absence of the need for a threshold is profoundly advantageous. The only inputs we employ in the ICS analysis are used to segment cell nuclei, yet these have no impact on the ICS calculation and are necessary for any analysis of the DDR.

Reviewer #2 (Public review):

Summary:

Immunostaining of chromatin-associated proteins and visualization of these factors through fluorescence microscopy is a powerful technique to study molecular processes such as DNA damage and repair, their timing, and their genetic dependencies. Nonetheless, it is well-established that this methodology (sometimes called "foci-ology") is subject to biases introduced during sample preparation, immunostaining, foci visualization, and scoring. This manuscript addresses several of the shortcomings associated with immunostaining by using image correlation spectroscopy (ICS) to quantify the recruitment of several DNA damage response-associated proteins following various types of DNA damage.

The study compares automated foci counting and fluorescence intensity to image correlation spectroscopy degree of aggregation study the recruitment of DNA repair proteins to chromatin following DNA damage. After validating image correlation spectroscopy as a reliable method to visualize the recruitment of γH2AX to chromatin following DNA damage in two separate cell lines, the study demonstrates that this new method can also be used to quantify RPA1 and Rad51 recruitment to chromatin following DNA damage. The study further shows that RPA1 signal as measured by this method correlates with cell sensitivity to Olaparib, a widely-used PARP inhibitor.

Strengths:

Multiple proof-of-concept experiments demonstrate that using image correlation spectroscopy degree of aggregation is typically more sensitive than foci counting or foci intensity as a measure of recruitment of a protein of interest to a site of DNA damage. The sensitivity of the SKOV3 and OVCA429 cell lines to MMS and the PARP inhibitors Olaparib and Veliparib as measured by cell viability in response to increasing amounts of each compound is a valuable correlate to the image correlation spectroscopy degree of aggregation measurements.

Weaknesses:

The subjectivity of foci counting has been well-recognized in the DNA repair field, and thus foci counts are usually interpreted relative to a set of technical and biological controls and across a meaningful time period. As such:

(1) A more detailed description of the numerous prior studies examining the immunostaining of proteins such as γH2AX, RAD51, and RPA is needed to give context to the findings presented herein.

We apologize for not providing enough detail. We have added further references and discussion. γH2AX foci counting, in particular, has been used in thousands of previous studies. (Pages 18 line 513 and 517)

(2) The benefits of adopting image correlation spectroscopy should be discussed in comparison to other methods, such as super-resolution microscopy, which may also offer enhanced sensitivity over traditional microscopy.

Thank you for raising this point. We have added this discussion (page 19 line 553). The limiting factor that ICS addresses is the partition coefficient of signal in a foci or cluster versus outside the cluster. Super-resolution will not necessarily improve this unless it is resolved down to single molecule counting. However, one would still need to evaluate how to define a cluster or foci in the background of non-cluster distribution.

(3) Additional controls demonstrating the specificity of their antibodies to detection of the proteins of interest should be added, or the appropriate citations validating these antibodies included.

We have added text stating that we only use validated antibodies (page 6 line 193). One thing to note is that we are measuring differences between treatment conditions, thus, if an antibody has non-specific labeling of proteins of cellular structures that do not change upon treatment, our approach would overcome this limitation.

Reviewer #3 (Public review):

Summary:

This paper described a new tool called "Image Correlation Spectroscopy; ICS) to detect clustering fluorescence signals such as foci in the nucleus (or any other cellular structures). The authors compared ICS DA (degree of aggregation) data with Imaris Spots data (and ImageJ Find Maxima data) and found a comparable result between the two analyses and that the ICS sometimes produced a better quantification than the Imaris. Moreover, the authors extended the application of ICS to detect cell-cycle stages by analyzing the DAPI image of cells. This is a useful tool without the subjective bias of researchers and provides novel quantitative values in cell biology.

Strengths:

The authors developed a new tool to detect and quantify the aggregates of immunofluorescent signals, which is a center of modern cell biology, such as the fields of DNA damage responses (DDR), including DNA repair. This new method could detect the "invisible" signal in cells without pre-extraction, which could prevent the effect of extracted materials on the pre-assembled ensembles, a target for the detection. This would be an alternative method for the quantification of fluorescent signals relative to conventional methods.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Major comments:

(1) The ICS theory section is essential and based on an excellent review from one of the authors. It would benefit greatly from a diagram showing where the quantities 𝒈(𝟎, 𝟎), 𝝎𝟎, and 𝒈inf come from in the 2D Gaussian fit, ideally for two cases where these quantities differ (i.e., how they correspond to different DA or TNoP values). In my opinion, this addition would greatly increase the manuscript's accessibility for DDR researchers. The citation of the review at the beginning would also be a plus.

We have added the review citation at the front of the theory section (page 3 line 87).We have highlighted where g(0,0), the most critical measurement for determination of TNoP and DA, derives from in Figure 2D. However, it is difficult to describe all the curve fit parameters in an image as they have some interdependency on each other and thus labeling one in a single image would not independently capture how they might be observed in a different curve fit.

(2) The TNoP measured in Figure 2 is a quantity about 2000-3000 times greater than the number of "traditionally detected" foci by both methods and the linear relations have very low Y intercepts. Can the authors comment explicitly on the physical interpretation of this number - are 2 to 3 thousand independent particles present within each "focus" detected by traditional means? If so, then what might one "particle" correspond to? (a single secondary antibody or fluorophore? a nucleosome?). In a similar vein, the X intercepts lie at around 25 foci, meaning that in images with fewer than that number of foci detected by ImageJ or Imaris, the ICS method should detect zero TNoP - is this in line with the authors' predictions? Is it possible that a first-order line fit is not the most appropriate relation between the two methods?

We apologize for our brevity here. Since DA proved to be a more useful metric we did not spend much effort discussing TNoP. TNoP correlates to the number of clustered particles, or non-diffuse fluorophores. TNoP is the inverse of the number of individual particles per nucleus, but the value is not a direct measure of foci. If a sample had no clustering at all, the number of individual particles would be at a maximum and the TNoP would be at a minimum. However, as fluorophores cluster, the number of individual particles (i.e. non-clustered fluorophores) decreases, which increases the TNoP value. Therefore, TNoP has a correlation to the number of foci detected through traditional measurements, as we found here. Yet, TNoP is a relative measurement and cannot be compared across different conditions. Similar to foci counting, TNoP is unable to factor the size or intensity of each cluster, thus DA is a more appropriate quantification of the DNA damage response.

The value of TNoP is dependent on the fitted point spread function and the area of the nucleus. The y=0 intercept of TNoP is defined by the optical setup and is not expected to necessarily go through x=0. Intriguingly, other groups have found that some foci identified through traditional measurements are actually clusters of multiple smaller foci, thus the concept of what a foci represents is difficult to interpret. Thus, here we aimed to show a general correlation of TNoP with foci count through traditional methods to reflect how ICS is similar to foci counting, then employed DA to overcome the limitations of defining a foci.

We have tried to clarify this in the text (page 8, line 266)

(3) Some suggestions to address the robustness of ICS:

For a given sample (i.e. one segmented nucleus), the calculation of DA and TNoP should be similar between different images of that same nucleus taken at different times, similar to how the number of traditionally detected foci would be fairly invariant. In particular, it should be shown that these values are not just scaling with the higher normalized intensity seen in stronger DDR responses. In the same vein, the linear relationship between TNoP and "foci" should not change even if the confocal settings are slightly different (i.e., higher/lower illumination intensity) as long as the condition stipulated by the authors in the Discussion holds ("ICS can be implemented on any fluorescence image as long as the square relative fluorescence intensity fluctuations are detectable above noise fluctuations."). To show, as the title states, that spatial ICS is a robust tool, it would be desirable to demonstrate this with a series of images of the same cell at the same or varying excitation intensities.

Thank you for your suggestions. Indeed, the calculation will be the same over sequential images of the same cell. Observations of dose dependent DA that does not correlate with intensity for RPA1 and RAD51 results (Fig. S5) directly demonstrates that DA does not just scale with intensity.

We would not expect the TNoP to change with confocal setting, however we show in Figure 1 that the number of foci does indeed change with intensity settings as captured by thresholds. Therefore, any interpretation of TNoP vs. foci count would be very difficult to make at different microscope settings. To ensure we are fairly comparing ICS to existing analysis we keep the settings the same and measure changes between conditions.

(4) More information is needed on how intensity normalization was performed. The Methods states "Measurements across experiments were normalized by the control in each dataset." The DMSO (0mM drug) plots all appear to have a mean of 1.0, so it appears the values for each set of control nuclei were divided by their own mean, and then the values for each set of experimental nuclei were divided by the mean value of all 3 controls as an aggregate; is this correct?

We apologize for not being more clear. Thank you for raising this point. We normalized data to a control from each experimental group. Thus, in figures 3,4 and 5 data were collected over multiple experiments with one control per experiment and each treatment condition included in each experiment. Therefore, we normalized each result to the corresponding control from that imaging session. However, in Figure 8 we ran experiments at much higher throughput with multiple controls per experiment, thus the data were normalized to the overall average of the controls, which is why the control averages are not all at a value of 1. We have clarified this in the text. (Page 7 line 218).

(5) Some more information about the ICS analysis should be given if the full code is not provided - in particular, how the nucleus mask was implemented on the "signal" channel (were the edges abruptly set to zero or was a window function introduced to avoid edge effects in the discrete FFT?

Thank you for raising this point. We have added the code to GitHub - github.com/ dubachLab/ics. The signal region was established by simply applying the nuclear mask from the DAPI channel to the IF channel. Each region is padded with average intensity value at the edges for 2x the dimensions of the ROI to remove edge effects in the FFT.

Minor comments:

(1) Figure 3, 4, 5: I think it would aid figure readability if channels were labeled in the images themselves, not just in the legend.

Thank you for the suggestion, we tried doing this and struggle to fit a label with the layout of the images. We were also concerned about interpretation of data in each column and the potential to assign data to each figure if they were so prominently labeled.

(2) Supplemental Figures are mislabeled; the order given in the legends is S1, S2, S3, S2, S3. S4 is called out in the main text where it should be S5.

Thank you for catching this error. We have made the necessary corrections. S4 contains data on cellular response to the drugs, while S5 contains intensity data in response to MMS.

(3) It should be stated for each Figure what kind of microscopy was performed - I assume that it is confocal for everything except when widefield is explicitly stated, but for clarity please add this information.

Indeed, this is correct, we have indicated which microscopy was used for each figure.

(4) The MATLAB code and full (uncropped) Western blots should be provided as supplemental data if possible.

We have included a GitHub link for the code and un-cropped western blots.

(5) The p values from significance tests should indicate whether multiple comparisons correction was necessary (if suggested by Prism) and performed.

Apologies for a lack of clarity but this was not necessary, significance was calculated vs. the next lower dose (e.g. 10 micromolar vs. 1 micromolar). We have clarified this in the methods (page 7 line 221).

Reviewer #2 (Recommendations for the authors):

Major points:

In addition to the weaknesses noted above, to encourage widespread adoption of this method, the authors should make the tools that they used for their analysis publicly available. In a few instances (e.g., compare Figures 3J and 3L), other methods outperform DA. It would be meaningful to discuss when especially DA may be a better measure than others (such as intensity or number of foci).

We have made code available on Github. We expect results, such as those in Figures 3J and 3L where intensity is significantly higher at the highest concentration but DA is not are reflective of the underlying biology and this may be interpreted differently under different experimental conditions. Imaris spots (Fig. 3K) also does not capture a significant increase at the highest dose of olaparib, suggesting that intensity may raise but it doesn’t not generate more foci. These results are likely highly dependent on the mechanism of olaparib at such a high concentration and the DDR response. We are hesitant to draw biological conclusions from these results and instead would like to highlight the capacity of ICS to evaluate the DDR, therefore we don’t want to make any broad comments about different applications.

Minor points:

(1) Pg. 12: "We used MMS to induce DNA damage in SKOV3 and OVCA429 cells. As expected, normalized intensity for RPA1 and RAD51 values (Figure S5) did not display a dose dependence on MMS concentration."

Please provide a citation for the claim that RPA1 and RAD51 normalized intensities do not display a dose dependence on MMS concentration.

These were data that we generated. We were not expecting an intensity change as that would presumably require increased protein generation in response to MMS, compared to gH2AX where the phospho-specific H2AX is generated in the DDR.

(2) Pg. 12: "Similar to RPA1, RAD51 does not form distinguishable foci in the nuclei in cells without preextraction (Fig. 5)." Please provide a citation for this claim.

We did not do pre-extraction and our results don’t produce changes in distinguishable foci. We provided citations discussing how, without pre extraction, foci formation for these proteins is not obvious (REF 38 and 39).

(3) I noted that the authors cite one paper [38] apparently showing that RPA and Rad51 do not always form foci, however, this is in the C. elegans germline in response to micro irradiation, therefore I am not sure that it is applicable to human cells.

We apologize for referencing a paper on C elegans. Most papers looking at RPA and RAD51 in the DDR use pre-extraction as it seems necessary to observe foci. Therefore, there are not as many papers, that we could find, that do not use pre-extraction. Reference 39 is in Hela cells.

Reviewer #3 (Recommendations for the authors):

Major points:

(1) Page 8, the second paragraph: In the Result section, it is better to describe how the authors carried out immuno-staining (without pre-extract subtraction) and ICS briefly, although the method is described in detail in the Method section.

Thank you for the suggestion, we have added this description (page 8, line 259)

(2) In Figure 5K-P: The authors analyzed "invisible" RAD51 foci on the image (Fig. 5L, M, O, and P) without pre-extraction. As a control experiment, it is useful to check whether pre-extraction would provide "visible" RAD51 foci and to examine the similar MMS concentration dependency shown in Figure 5R (or 5T). This would strengthen the power of the ICS analysis.

Thank you for the suggestion. In our hands, pre-extraction is extremely subjective. We have tried performing pre-extraction but find highly variable results depending on conditions. Therefore, we did not include any pre-extraction here. We expect that performing these experiments may or may not agree with results in Figure 5 largely because we are unable to achieve repeatable pre-extraction foci counting.

(3) Figure 6D (and 6C) looks very interesting. It would be important to show the interpretation of this correlation shown in the graph. Although the authors argued that ICS analysis results shown in the graph could provide new insight into the DDR (page 14, last line 5), as shown in another part, it is important to carry out the same analysis by using Imaris Spots. Moreover, it is interesting to apply the analysis to RAD51 foci (shown in Figure 5), given that the PARPi effect is enhanced in the absence of RAD51mediated recombination.

We completely agree that this analysis may generate interesting results to help interpret the DDR response to PARP inhibition. These experiments are part of an ongoing follow up study where we extend the use of ICS to other parts of the DDR and investigate protein clustering across several proteins with impact on PARPi response. Therefore, since the focus of this manuscript is introducing ICS as a tool to study the DDR, we believe that omitting those data here does not deter from the central points of the manuscript. We including results in Figure 6 because we wanted to show how ICS could impact DDR research. Furthermore, combined with our advances shown in Figures 7 and 8, we are currently working on adapting ICS to be high-throughput and much simpler than Imaris spots for handling large datasets needed to generate results like those in Figure 6.

Minor points:

(1) Figure 1I, blue arrows: These showed an area with a higher background. Because of a low magnification, it is very hard to see the difference from the other areas of the background. It is better to show a magnified image of the representative region with a higher background.

We hope that readers can see the higher intensity in the diffuse area. We attempted to construct a zoomed in area, but that either blocked a significant portion of the nonzoomed image or added complexity to the figure. We have noted that images in Figure S1 are larger and more obviously capture an increase in background intensity.

(2) Figure 2 legend, line 5, the same as "A)": This should be "B".

Here, the number of independent particle clusters is intended to be the same as A, the difference is that the independent particles are clusters in C and individual fluorophores in A.

(3) Page 9, the first paragraph, last line, foci formation, and foci composition: These should be "focus formation and focus composition".

We have changed this.

(4) Page 15, the first paragraph, line 5, palbociclib, camptothecin, or etoposide: please explain what kinds of the drugs are.

We have added that these drugs cause cells to stall at different cell cycle stages. Explaining the drugs would take considerable room in the text.

(5) Page 16, the first paragraph, line 1, bleomycin: Please explain what this drug is.

Similar to above, we have stated that this drug causes DNA damage, going into detail would take several sentences.

Author response:

The following is the authors’ response to the original reviews

Public reviews:

Reviewer #1:

(1) This manuscript introduces a useful curation pipeline of antibody-antigen structures downloaded from the PDB database. The antibody-antigen structures are presented in a new database called AACDB, alongside annotations that were either corrected from those present in the PDB database or added de-novo with a solid methodology. Sequences, structures, and annotations can be very easily downloaded from the AACDB website, speeding up the development of structure-based algorithms and analysis pipelines to characterize antibody-antigen interactions. However, AACDB is missing some key annotations that would greatly enhance its usefulness.

Here are detailed comments regarding the three strengths above:

I think potentially the most significant contribution of this database is the manual data curation to fix errors present in the PDB entries, by cross-referencing with the literature. However, as a reviewer, validating the extent and the impact of these corrections is hard, since the authors only provided a few anecdotal examples in their manuscript.

I have personally verified some of the examples presented by the authors and found that SAbDab appears to fix the mistakes related to the misidentification of antibody chains, but not other annotations.

(a) "the species of the antibody in 7WRL was incorrectly labeled as "SARS coronavirus B012" in both PDB and SabDab" → I have verified the mistake and fix, and that SAbDab does not fix is, just uses the pdb annotation.

(b) "1NSN, the resolution should be 2.9 , but it was incorrectly labeled as 2.8" → I have verified the mistake and fix, and that sabdab does not fix it, just uses the PDB annotation.

(c) "mislabeling of antibody chains as other proteins (e.g. in 3KS0, the light chain of B2B4 antibody was misnamed as heme domain of flavocytochrome b2)" → SAbDab fixes this as well in this case.

(d) "misidentification of heavy chains as light chains (e.g. both two chains of antibody were labeled as light chain in 5EBW)" → SAbDab fixes this as well in this case.

I personally believe the authors should make public the corrections made, and describe the procedures - if systematic - to identify and correct the mistakes. For example, what was the exact procedure (e.g. where were sequences found, how were the sequences aligned, etc.) to find mutations? Was the procedure run on every entry?

We appreciate the reviewer’s valuable feedback. Our correction procedures combined manual curation with systematic sequence analysis. While most metadata discrepancies were resolved through cross-referencing original literature, we implemented a structured approach for identifying mutations in specific cases. For PDB entries labeled as variants (e.g., "Bevacizumab mutant" or "Ipilimumab variant Ipi.106") where the "Mutation(s)" field was annotated as "NO," we retrieved the canonical therapeutic antibody sequence from Thera-SAbDab, then performed pairwise sequence alignment against the PDB entry using BLAST program to identified mutated residues.

This procedure was not applied to all entries, as mutations are context-dependent. Therapeutic antibodies have well-defined reference sequences, enabling systematic alignment. For antibodies lacking unambiguous wild-type references (e.g., research-grade or non-therapeutic antibodies), mutation annotations were directly inherited from the PDB or literature.

All corrections have been publicly archived in AACDB. We have added a detailed discussion of this issue in the section “2.3 Metadata” of revised manuscript.

(2) I believe the splitting of the pdb files is a valuable contribution as it standardizes the distribution of antibody-antigen complexes. Indeed, there is great heterogeneity in how many copies of the same structure are present in the structure uploaded to the PDB, generating potential artifacts for machine learning applications to pick up on. That being said, I have two thoughts both for the authors and the broader community. First, in the case of multiple antibodies binding to different epitopes on the same antigen, one should not ignore the potentially stabilizing effect that the binding of one antibody has on the complex, thereby enabling the binding of the second antibody. In general, I urge the community to think about what is the most appropriate spatial context to consider when modeling the stability of interactions from crystal structure data. Second, and in a similar vein, some antigens occur naturally as homomultimers - e.g. influenza hemagglutinin is a homotrimer. Therefore, to analyze the stability of a full-antigen-antibody structure, I believe it would be necessary to consider the full homo-trimer, whereas, in the current curation of AACDB with the proposed data splitting, only the monomers are present.

We sincerely appreciate the reviewer’s insightful comments regarding the splitting of PDB files and we appreciate the opportunity to address the reviewer’s thoughtful concerns.

Firstly, when two antibodies bind to distinct epitopes on the same antigen, we would like to clarify that this scenario can be divided into two cases based on the experimental context: Case1: When two antibodies bind to distinct epitopes on the same antigen, and their complexes are determined in separate structures. For example, SAR650984 (PDB: 4CMH) and daratumumab (PDB: 7DHA) target CD38 at non-overlapping epitopes. These two antibody-antigen complexes were determined independently, and their structures do not influence each other. Case 2 : When the crystal structure contains a ternary complex with two antibodies and an antigen, as in the example of 6OGE discussed in Section 2.2 of our manuscript. After reviewing the original literature, the experiment confirmed that the order of Fab binding does not affect the formation of the ternary complex, and the binding of one antibody does not enhance the binding of the other. This supports the rationale for splitting 6OGE into two separate structures. However, we acknowledge that not all ternary complexes in the PDB provide such detailed experimental descriptions in their original literature. We agree with the reviewer that in some cases, one antibody may stabilize the structure to facilitate the binding of a second antibody. For instance, in 3QUM, the 5D5A5 antibody stabilizes the structure, enabling the binding of the 5D3D11 antibody to human prostate-specific antigen. Such sandwich complexes are indeed valuable for identifying true epitopes and paratopes. Importantly, splitting the structure does not alter the interaction sites.

Secondly, we fully agree with the reviewer that for antigens that naturally exist as homomultimers (e.g., influenza hemagglutinin as a homotrimer), the full multimeric structure should be considered when analyzing stability. In such cases, users can directly utilize the original PDB structures provided in their multimeric form. Our splitting approach is intended to provide an additional option for cases where monomeric analysis is sufficient or preferred, but it does not preclude the use of the original multimeric structures when necessary.

(3) I think the manuscript is lacking in justification about the numbers used as cutoffs (1A^2 for change in SASA and 5A for maximum distance for contact) The authors just cite other papers applying these two types of cutoffs, but the underlying physico-chemical reasons are not explicit even in these papers. I think that, if the authors want AACDB to be used globally for benchmarks, they should provide direct sources of explanations of the cutoffs used, or provide multiple cutoffs. Indeed, different cutoffs are often used (e.g. ATOM3D uses 6A instead of 5A to determine contact between a protein and a small molecule https://datasets-benchmarks-proceedings.neurips.cc/paper/2021/hash/c45147dee729311ef5b5c3003946c48f-Abstract-round1.html). I think the authors should provide a figure with statistics pertaining to the interface atoms. I think showing any distribution differences between interface atoms determined according to either strategy (number of atoms, correlation between change in SASA and distance...) would be fundamental to understanding the two strategies. I think other statistics would constitute an enhancement as well (e.g. proportion of heavy vs. light chain residues).

Some obvious limitations of AACDB in its current form include:

AACDB only contains entries with protein-based antigens of at most 50 amino acids in length. This excludes non-protein-based antigens, such as carbohydrate- and nucleotide-based, as well as short peptide antigens.

AACDB does not include annotations of binding affinity, which are present in SAbDab and have been proven useful both for characterizing drivers of antibody-antigen interactions (cite https://www.sciencedirect.com/science/article/pii/S0969212624004362?via%3Dihub) and for benchmarking antigen-specific antibody-design algorithms (cite https://www.biorxiv.org/content/10.1101/2023.12.10.570461v1)).

We thank the reviewer for raising this critical point about the cutoff values used in AACDB. In the current study, the selection of the threshold value is very objective; the threshold chosen in the manuscript is summarized based on existing literature, and we have provided more literature support in the manuscript. The criteria for defining interacting amino acids in established tools, typically do not set the ΔSASA exceed 1 Å2 and the distance exceed 6 Å. While our manuscript emphasizes widely accepted thresholds for consistency with prior benchmarks, AACDB explicitly provides raw ΔSASA and distance values for all annotated residues. Users can dynamically filter the data from downloaded files by excluding entries exceeding their preferred thresholds (e.g., selecting 5Å instead of 6Å). This ensures adaptability to diverse research needs. In the revised version, we reset the distance threshold to 6 Å and calculated the interacting amino acids in order to give the user a wider range of choices. In the section “3.2 Database browse and search” of revised manuscript, we provide a description of the flexible choice of thresholds for practical use.

Furthermore, distance and ΔSASA are two distinct metrics for evaluating interactions. Distance directly quantifies spatial proximity between atoms, reflecting physical contacts such as van der Waals interactions or hydrogen bonds, and is ideal for identifying direct spatial adjacency. ΔSASA, on the other hand, measures changes in solvent accessibility of residues during binding, capturing the contribution of buried surfaces to binding free energy. Even for residues not in direct contact, reduced SASA due to conformational changes may indicate indirect functional roles.

As demonstrated through comparisons on the detailed information pages, the sets of interacting amino acids defined by these two methods differ by only a few residues, with no significant variation in their overall distributions. However, since interaction patterns vary significantly across different complexes, analyzing residue distributions across all structures using both criteria is not feasible.

We thank the reviewer for highlighting these limitations. AACDB currently focuses on protein-based antigens ≤50 amino acids to prioritize structural consistency, which excludes non-protein antigens and shorter peptides. While affinity annotations are critical for benchmarking antibody design tools, these data were not integrated in this release due to insufficient data verification caused by internal team constraints. We acknowledge these gaps and plan to expand antigen diversity and incorporate affinity metrics in future updates.

Reviewer #2:

Summary:

Antibodies, thanks to their high binding affinity and specificity to cognate protein targets, are increasingly used as research and therapeutic tools. In this work, Zhou et al. have created, curated, and made publicly available a new database of antibody-antigen complexes to support research in the field of antibody modelling, development, and engineering.

Strengths:

The authors have performed a manual curation of antibody-antigen complexes from the Protein Data Bank, rectifying annotation errors; they have added two methods to estimate paratope-epitope interfaces; they have produced a web interface that is capable of both effective visualisation and of summarising the key useful information in one page. The database is also cross-linked to other databases that contain information relevant to antibody developability and therapeutic applications.

Weaknesses:

The database does not import all the experimental information from PDB and contains only complexes with large protein targets.

Thank you for the valuable feedback. As previously responded to Reviewer 1, due to limitations within our team, comprehensive data integration from PDB has not been achieved in the current version. We acknowledge the significance of expanding the database to encompass a broader range of experimental information and complexes with diverse target sizes. Regrettably, immediate updates to address these limitations are not feasible at this time. Nevertheless, we are committed to enhancing the database in upcoming upgrades to provide users with a more comprehensive and inclusive resource

Recommendations for the authors:

Reviewer #1:

(1) Line 194: "produce" → "produced"

We thank the reviewer for the feedback. We have checked the grammar and spelling carefully in the revised manuscript.

(2) As mentioned in the public review, I think adding binding affinity annotations would greatly enhance the use cases for the database.

We thank the reviewer for the suggestion. As the response in “Public review”. Due to team constraints, these data are not integrated into this release but are being collated. We recognize these gaps and plan to expand antigenic diversity and incorporate affinity metrics in future updates.

(3) I think adding a visualization of interface atoms and contacts on an entry's webpage would be useful for someone exploring specific entries. It also would be useful if the authors provided a pymol command to select interface residues since that's a procedure any structural biologist is likely to do.

We sincerely appreciate the reviewer’s constructive suggestions. In response to the request for enhanced visualization and accessibility of interface residue information, we have implemented the following improvements: (1) Web Interface Visualization. On the entry-specific webpage, we have added an interactive visualization window that highlights the antigen-antibody interaction interface using distinct colors. The interaction interface visualization has been incorporated into Figure 5 of the revised manuscript, with a detailed description. (2) PyMOL Command Accessibility. The “Help” page now provides step-by-step PyMOL commands to select and visualize interface residues.

(4) I think the authors should provide headers to the files containing interface residues according to the change-in-SASA criterion, as they do for those computed according to contact. This would avoid unnecessary confusion - however slight - and make parsing easier. I was initially confused by the meaning of the last column, though after a minute I understood it to be the change in SASA.

We thank the reviewer for providing such detailed feedback. We thank the reviewer for the comment and the suggestion. We have provided headers for the files of the interacting residues defined by ΔSASA.

(5) Line 233: "AACDB's data processing pipeline supports mmCIF files" → The meaning and implications of this statement are not obvious to me, and are mentioned nowhere else in the paper. Do you mean that in AACDB there are structure entries that the RCSB PDB database only has in mmCIF file format, and not .pdb format? So, effectively, there are some entries in AACDB that are not in any other antibody-specific database?I checked and, as of Dec 3rd, 2024, there are 41 structures in AACDB that are NOT in SAbDab. Manually checking 5 of those 41 structures, none are mmCIF-only structures.

We thank the reviewer for the valuable comment. Because of the size of the structures within certain entries, representing them in a single PDB format data file is not feasible due to the excessive number of atoms and polymer chains they contain. As a result, PDB stores these structures in “mmcif” format files. In AACDB, 47 entries, such as 7SOF, 7NKT, 7B27, and 6T9D, are only available in the “mmCIF” format from the PDB. The “.pdb” and “.cif” files contain atomic coordinates in distinct text formats, and the segmentation of these structure files is automatically conducted based on manually annotated antibody-antigen chains. To accommodate this, we have incorporated these considerations into our file processing pipeline, thereby enabling a fully automated file segmentation process. Additionally, we employed Naccess to calculate interatomic distances. However, since this software only accepts .pdb format files as input, we also converted all split .cif files into .pdb format within our fully automated pipeline. We apologize for the lack of clarity in the original manuscript and have included a more detailed explanation in the "2.2 PDB Splitting" section of the revised manuscript.

Reviewer #2:

(1) In SabDab and PDB, experimental binding affinities are also reported: could the authors comment on whether they also imported this information and double-checked it against the original paper? If it wasn't imported, that might discourage some users and should be considered as an extension for the future.

We thank the reviewer for the comment and the suggestion. As the response in “Public review”. Due to current resource constraints, quantitative affinity data has not been incorporated into this release but is undergoing systematic curation. We explicitly recognize these limitations and propose a two-pronged strategy for future iterations: (1) broadening antigen diversity coverage through expanded structural sampling, and (2) integrating quantitative binding affinity measurements. In the Discussion section, we have included description outlining the planned enhancements.

(2) Line 49-50: the references mentioned in connection to deep learning methods for antibody-antigen predictions seem a bit limited given the amount of articles in this field, with 3 of 4 references on one method only (SEPPA), could the authors expand this list to reflect a bit more the state of the art?

We thank the reviewer for the suggestion. We agree that more relevant studies should be listed and therefore more references are provided in the revised manuscript.

When mentioning the limitations of the existing databases, it feels a bit that the criticism is not fully justified. For instance:

Line 52-53: could the authors elaborate on the reasons why such an identification is challenging? (Isn't it possible to make an efficient database-filtered search? Or rather, should one highlight that a more focussed resource is convenient and why?)

Thank you for feedback. In this study, the keywords "antibody complex," "antigen complex," and "immunoglobulin complex," were employed during data collection. PDB returned over 30,000 results, of which only one-tenth met our criteria after rigorous filtering. This demonstrates that keyword searches, while useful, inherently limit result precision and introduce substantial redundancy, likely due to the PDB's search mechanism. That’s why we illustrated the significant challenges in identifying antibody-antigen complexes from general protein structures in the PDB.

Line 55: reading the website http://www.abybank.org/abdb/, it would be fairer to say that the web interface lacks updates, as the database and the code have gone through some updates. Could the authors provide a concrete example of the reason why: 'The AbDb database currently lacks proper organization and management of this valuable data.'?

We thank the reviewer for highlighting this issue. In our original manuscript, the statement that the AbDb database "lacks proper organization and management" was based on the absence of explicit statement regarding data updates on its official website at the time of submission, even though internal updates to its content may have occurred. We fully respect the long-standing contributions of AbDb to antibody structural research, and our comments were solely directed at the specific state of the database at that time. As the reviewer noted, following the release of our preprint, we have also taken note of AbDb's recent updates. To reflect the latest developments and avoid potential misinterpretation, we have revised the original statement in revised manuscript.

Also 'this rapid updating process may inadvertently overlook a significant amount of information that requires thorough verification,': it's difficult for me to understand what this means in practice. Could the authors clarify if they simply mean that SabDab collects information from PDB and therefore tends to propagate annotation errors from there? If yes, I think it's enough to state it in these terms, and for sure I agree that the reason is that correcting these annotation errors requires a substantial amount of work.

We thank the reviewer for providing such detailed feedback on the manuscript. We acknowledge that SabDab represents a highly valuable contribution to the field, and its rapid update mechanism has significantly advanced related research areas. However, as stated by the reviewer, we aim to clarify that SabDab primarily relies on automated metadata extraction from the PDB for annotation, and its rapid update process inherently inherits raw data from upstream sources. According to their paper, manual curation is only applied when the automated pipeline fails to resolve structural ambiguities. This workflow—dependent on PDB annotations with limited manual verification—may propagate errors provided by PDB. Examples include species misannotation and mutation status misinterpretation. We fully agree with the reviewer's observation that correcting errors in such cases necessitates labor-intensive manual curation, which is a core motivation for our study.

Line 86: why 'Structures that consisted solely of one type of antibody were excluded'? Why exclude complexes with antigens shorter than 50 amino acids? These complexes are genuine antibody-antigen complexes.

We thank the reviewer for the valuable question. The AACBD database is dedicated to curating structural data of antigen-antibody complexes. Structures featuring only a single antibody type are classified as free antibodies and systematically excluded from the database due to the absence of protein-bound partners. During data screening , we retained sequences shorter than 50 amino acids by categorizing them as peptides rather than eliminating them outright. The current release exclusively encompasses complexes with protein-based antigens. Meanwhile, complexes involving peptide, haptens, and nucleic acid antigens are undergoing systematic curation, with planned inclusion in future updates to broaden antigen category representation.

Line 96 needs a capital letter at the beginning.

Line 107: 'this would generate' → 'this generates' (given it is something that has been implemented, correct?).

Line 124: missing an 'of'.

Line 163: inspiring by -> inspired by.

Thank you for feedback. All of the above grammatical or spelling errors have been revised in the manuscript.

Line 109-111: apart from the example, it would be good to spell out the general rule applied to anti-idiotypic antibodies.

We thank the reviewer for the valuable feedback. For anti-idiotypic antibodies complex. the partner antibody is treated as a dual-chain antigen, , necessitating individual evaluation of heavy chain and light chain interactions with the anti-idiotypic component. We have given a general rule for anti-idiotypic antibodies in section “2.2 PDB splitting” of revised manuscript.

Line 155-159: could the authors provide references for the two choices (based on sasa and any-atom distance) that they adopted to define interacting residues?

We thank the reviewer for the comment and the suggestion. As the same as the response to reviewer #1 in Public review. The interacting residues definition and the threshold chosen in the manuscript is summarized based on existing literature. We have added additional references for support in section “1.Introduction”. Our resource does not provide a fixed amino acid list. Instead, all interacting residues are explicitly documented alongside their corresponding ΔSASA (solvent-accessible surface area changes) and intermolecular distances, allowing researchers to flexibly select residue pairs based on customized thresholds from downloadable datasets. Furthermore, aligning with widely adopted criteria in current literature—where interactions are defined by ΔSASA >1 Ų and atomic distances <6 Å, we have recalibrated our analysis in the revised version. Specifically, we replaced the previous 5 Å distance threshold with a 6 Å cutoff to recalculate interacting residues.

Line 176-178: could the authors re-phrase this sentence to clarify what they mean by 'change in the distribution'?

We thank the reviewer for the suggestion. Our search was conducted with an end date of November 2023. However, Figure 3B includes an entry dated 2024. Upon reviewing this record, we identified that the discrepancy arises from the supersession of the 7SIX database entry (originally released in December 2022) by the 8TM1 version in January 2024. This version update explains the apparent chronological inconsistency. We regret any lack of clarity in our original description and have revised the corresponding section in the manuscript to explicitly clarify this change of database.

Caption Figure 3: please spell out all the acronyms in the figure. Provide the date when the last search was performed (i.e., the date of the last update of these statistics).

We thank the reviewer for the comment. We have systematically expanded all acronyms and included update dates for statistics in the legend of Figure 3. Corresponding changes have also been made to the statistical pages on the website.

Finally, it would be advisable to do a general check on the use of the English language (e.g. I noted a few missing articles). In Figure 5 DrugBank contains typos.

We sincerely appreciate the reviewer's meticulous attention to linguistic precision. We have corrected the typographical error in Figure 5 and conducted a comprehensive review of the entire manuscript to ensure accuracy and clarity.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

The authors investigate how the viscoelasticity of the fingertip skin can affect the firing of mechanoreceptive afferents and they find a clear effect of recent physical skin state (memory), which is different between afferents. The manuscript is extremely well-written and well-presented. It uses a large dataset of low threshold mechanoreceptive afferents in the fingertip, where it is particularly noteworthy that the SA-2s have been thoroughly analyzed and play an important role here. They point out in the introduction the importance of the non-linear dynamics of the event when an external stimulus contacts the skin, to the point at which this information is picked up by receptors. Although clearly correlated, these are different processes, and it has been very well-explained throughout. I have some comments and ideas that the authors could think about that could further improve their already very interesting paper. Overall, the authors have more than achieved their aims, where their results very much support the conclusions and provoke many further questions. This impact of the previous dynamics of the skin affecting the current state can be explored further in so many ways and may help us to better understand skin aging and the effects of anatomical changes of the skin.

At the beginning of the Results, it states that FA-2s were not considered as stimuli did not contain mechanical events with frequency components high enough to reliably excite them. Was this really the case, did the authors test any of the FA-2s from the larger dataset? If FA-2s were not at all activated, this is also relevant information for the brain to signal that it is not a relevant Pacinian stimulus (as they respond to everything). Further, afferent receptive fields that were more distant to the stimulus were included, which likely fired very little, like the FA-2s, so why not consider them even if their contribution was low?

Thank you for bringing this up, we have now clarified in the text that while FA-2s did respond at a low rate during the experiment, their responses were not reliably driven by the force stimuli. In the Methods section we have included the following text:

“Initially, 10 FA-2 neurons were also included in the analysis. But their responsiveness during the experiment was remarkably low, and unlike the other neuron types, their responses were rarely affected by force stimuli. Specifically, only one of the observed FA-2 neurons responded during the force protraction phases. Due to the lack of clear stimulus-driven responses, FA-2 neurons were subsequently excluded from further analysis.”

One question that I wondered throughout was whether you have looked at further past history in stimulation, i.e. not just the preceding stimulus, but 2 or 3 stimuli back? It would be interesting to know if there is any ongoing change that can be related back further. I do not think you would see anything as such here, but it would be interesting to test and/or explore in future work (e.g. especially with sticky, forceful, or sharp indentation touch). However, even here, it could be that certain directions gave more effects.

This is a very interesting question! A discernible effect from the previous stimulus could persist at the end of the current stimulation (see Figure 4C), potentially influencing the next one—a 2-stimuli-back effect. Unfortunately, our experimental design did not allow for rigorous testing of this effect. While all possible pairs of stimulus directions were included in immediately consecutive trials, this was not the case for pairs separated by additional trials. Hence, the combination of a likely weak effect and limited variation in history precluded a thorough analysis of a 2-stimuli-back effect. Future work should delve into the time course of the viscoelastic effect in greater detail.

Did the authors analyze or take into account the difference between receptive field locations? For example, did afferents more on the sides have lower responses and a lesser effect of history?

An investigation into the potential impact of the relationship between the receptive field location on the fingertip skin and the primary contact site of the stimulus surface revealed no discernible influence for SA-1 and SA-2 neurons. In contrast, FA-1 neurons, particularly those predominantly sensitive to the previous stimulation or displaying mixed sensitivity, exhibited a tendency to terminate near the primary stimulation site. We have added these observations to the text:

“We found no straightforward relationship between a neuron's sensitivity to current and previous stimulation and its termination site in fingertip skin. Specifically, there was no statistically significant effect of the distance between a neuron's receptive field center and the primary contact site of the stimulus surface on whether neurons signaled current, prior, or mixed information for SA-1 (Kruskal-Wallis test H(2)=3.86, p= 0.15) or SA-2 neurons (H(2)=0.75, p=0.69). However, a significant difference emerged for FA-1 neurons (H(2)=8.66, p=0.01), indicating that neurons terminating closer to the stimulation site on the flat part of the fingertip were more likely to signal past or mixed information.”

Was there anything different in the firing patterns between the spontaneous and non-spontaneously active SA-2s? For example, did the non-spontaneous show more dynamic responses?

The firing patterns of both spontaneously and non-spontaneously active SA-2 neurons shared similarities in terms of adaptation and range of firing rate modulation in response to force stimuli, i.e., ‘dynamic response’. The distinction lay in the pattern of modulation of the firing rate associated with stimulus presentations. For spontaneously active SA-2 neurons, this modulation occurred around a significant background discharge, implying that a force stimulus could either decrease or increase the firing rate, depending on how it deformed the fingertip. This characteristic is well illustrated by the firing pattern of the neuron depicted in the lower panels of Figure 3D. Conversely, in non-spontaneously active SA-2 neurons, a force stimulus could only induce an increase in the firing rate or no change. Although the neuron depicted in the upper panels of Figure 3D exhibited some background activity, it serves to exemplify this characteristic. In the text, we have elucidated the dynamics of the SA-2 neuron response by highlighting that force stimulation can either decrease or increase the firing rate in neurons with spontaneous activity through the following addition/change:

“This increased variability was most evident during the force protraction phase where most neurons exhibited the most intense responses. Increased variability was also observed in instances where the dynamic response to force stimulation involved a decrease in the firing rate (lower panels of Figure 3D). This phenomenon was observed in SA-2 neurons that maintained an ongoing discharge during intertrial periods (cf. Fig. 2A). In these cases, the response to a force stimulus constituted a modulation of the firing rate around the background discharge, signifying that a force stimulus could either decrease or increase the firing rate depending on the prevailing stimulus direction.”

Were the spontaneously active SA-2 afferents firing all the time or did they have periods of rest - and did this relate to recent stimulation? Were the spontaneously active SA-2s located in a certain part of the finger (e.g. nail) or were they randomly spread throughout the fingertip? Any distribution differences could indicate a more complicated role in skin sensing.

SA-2 neurons, in general, are well-known for undergoing significant post-stimulation depression (e.g., Knibestöl and Vallbo, 1970; Chambers et al., 1972; Burgess and Perl, 1973). In our force stimulations, this post-excitatory depression manifested as a reduced or absent response during the latter part of the stimulus retraction period for stimuli in directions that markedly excited the neuron. The excitability recovered when the fingertip relaxed during the subsequent intertrial period, and for "spontaneously active" neurons, the firing resumed (see examples in Figure 7A). Furthermore, some “spontaneously active” neurons could be silenced or exhibit a near-silent period during force stimulation for certain force directions, while the spontaneous firing returned during the upcoming intertrial period when the fingertip shape recovered (for example, see responses to stimulation in the proximal and especially ulnar directions in the top panel in Figure 7A).

Regarding the location of the receptive field centres of spontaneously active and non-spontaneously active SA-2 neurons on the fingertip we did not observe any obvious spatial segregation. To illustrate this, we have revised Figure 1A by color-marking SA-2 neurons that exhibited ongoing activity in intertrial periods, and the figure caption has been modified accordingly:

“Figure 1. Experimental setup. A. Receptive field center locations shown on a standardized fingertip for all first-order tactile neurons included in the study, categorized by neuron type. Purple symbols denote spontaneously active SA-2 neurons exhibiting ongoing activity without external stimulation.”

Did the authors look to see if the spontaneous firing in SA-2s between trials could predict the extent to which the type 1 afferents encode the proceeding stimulus? Basically, does the SA-2 state relate to how the type 1 units fire?

We found no clear indications that the responses of FA-1 and SA-1 could be readily anticipated based on the firing patterns of SA-2 neurons.

In the discussion, it is stated that "the viscoelastic memory of the preceding loading would have modulated the pattern of strain changes in the fingertip differently depending on where their receptor organs are situated in the fingertip". Can the authors expand on this or make any predictions about the size of the memory effect and the distance from the point of stimulation?

We have explored this topic further in the text, referring to recent studies modeling essential aspects of fingertip mechanics. However, in our view, current models lack the capability to predict the specific nature sought by the reviewer. These models should include a detailed understanding of the intricate networks of collagen fibers anchoring the pulp tissue at the distal phalangeal bone and the nail. They should also consider potential inherent directional preferences of the receptor organs, attributed to their microanatomy. The text modifications are as follows:

“In addition to the receptor organ locations, the variation in sensitivity among neurons to fingertip deformations in response to both previous and current loadings would stem from the fingertip’s geometry and its complex composite material properties. Possible inherent directional preferences of the receptor organs, attributed to their microanatomy, could also be significant. However, mechanical anisotropy, particularly within the viscoelastic subcutaneous tissue of the fingertip induced by intricately oriented collagen fiber strands forming fat columns in the pulp (Hauck et al., 2004), are likely to play a crucial role. This anisotropy would shape the dynamic pattern of strain changes at neurons' receptor sites, intricately influencing a neuron's sensitivity not only to current but also to preceding loadings. Indeed, recent modeling efforts suggest that such mechanical anisotropy strongly influences the spatiotemporal distribution of stresses and strains across the fingertip (Duprez et al., 2024).”

Relatedly, we have included additional text to provide a more comprehensive explanation of the “bulk deformation” of the fingertip that occurs during the loadings:

“As pressure increases in the pulp, the pulp tissue bulges at the end and sides of the fingertip. Simultaneously, the tangential force component amplifies the bulging in the direction of the force while stretching the skin on the opposite side.”

In the discussion, it would be good if the authors could briefly comment more on the diversity of the mechanoreceptive afferent firing and why this may be useful to the system.

The diversity in responses among neurons is instrumental in enhancing the information transmitted to the brain by averting redundancy in information acquisition. This diversity thereby contributes to an overall increase in information. We've included a brief statement, along with several references, underscoring this concept:

"The resulting diversity in the sensitivities of neurons might enhance the overall information collected and relayed to the brain by the neuronal population, facilitating the discrimination between tactile stimuli or mechanical states of the fingertip (see Rongala et al., 2024; Corniani et al., 2022; Tummala et al., 2023, for more extensive explorations of this idea)."

Also, the authors could briefly discuss why this memory (or recency) effect occurs - is it useful, does it serve a purpose, or it is just a by-product of our skin structure? There are examples of memory in the other senses where comparisons could be drawn. Is it like stimulus adaptation effects in the other senses (e.g. aftereffects of visual motion)?

We have expanded the concluding paragraph of the discussion, specifically delving into the question of whether the mechanical memory effect serves a deliberate purpose or is simply an incidental byproduct of our skin structure:

“In any case, the viscoelastic deformability of the fingertips plays a pivotal role in supporting the diverse functions of the fingers. For example, it allows for cushioned contact with objects featuring hard surfaces and allows the skin to conform to object shapes, enabling the extraction of tactile information about objects' 3D shapes and fine surface properties. Moreover, deformability is essential for the effective grasping and manipulation of objects. This is achieved, among other benefits, by expanding the contact surface, thereby reducing local pressure on the skin under stronger forces and enabling tactile signaling of friction conditions within the contact surface for control of grasp stability. Throughout, continuous acquisition of information about various aspects of the current state of the fingertip and its skin by tactile neurons is essential for the functional interaction between the brain and the fingers. In light of this, the viscoelastic memory effect on tactile signaling of fingertip forces can be perceived as a by-product of an overall optimization process within prevailing biological constraints.”

One point that would be nice to add to the discussion is the implications of the work for skin sensing. What would you predict for the time constant of relaxation of fingertip skin, how long could these skin memory effects last? Two main points to address here may be how the hydration of the skin and anatomical skin changes related to aging affect the results. If the skin is less viscoelastic, what would be the implications for the firing of mechanoreceptors?

It is likely that the time constant depends to some extent on mechanical factors of the skin, which will likely change due to age or environmental factors. However, while these questions are intriguing, they fall outside the scope of the current study and we are not aware of studies that have addressed these issues directly in experiments either.

How long does it take for the effect to end? Again, this will likely depend on the skin's viscoelasticity. However, could the authors use it in a psychophysical paradigm to predict whether participants would be more or less sensitive to future stimuli? In this way, it would be possible to test whether the direction modifies touch perception.

Time constants for tissue viscoelasticity have been estimated to extend up to several seconds (see citations in the introduction). While direct perceptual effects could indeed be explored through psychophysical experimental paradigms, we are currently unaware of any studies specifically addressing the type of effect described in this study. In addition to the statement that, concerning manipulation and haptic tasks, "to our knowledge, a possible influence of fingertip viscoelasticity on task performance has not been systematically investigated," we have now also addressed tactile psychophysical tasks conducted during passive touch with the following sentence in the text:

“Similarly, there is a lack of systematic investigation of potential effects of fingertip viscoelasticity on performance in tactile psychophysical tasks conducted during passive touch.”

Reviewer #2 (Public Review):

Summary:

The authors sought to identify the impact skin viscoelasticity has on neural signalling of contact forces that are representative of those experienced during normal tactile behaviour. The evidence presented in the analyses indicates there is a clear effect of viscoelasticity on the imposed skin movements from a force-controlled stimulus. Both skin mechanics and evoked afferent firing were affected based on prior stimulation, which has not previously been thoroughly explored. This study outlines that viscoelastic effects have an important impact on encoding in the tactile system, which should be considered in the design and interpretation of future studies. Viscoelasticity was shown to affect the mechanical skin deflections and stresses/strains imposed by previous and current interaction force, and also the resultant neuronal signalling. The result of this was an impaired coding of contact forces based on previous stimulation. The authors may be able to strengthen their findings, by using the existing data to further explore the link between skin mechanics and neural signalling, giving a clearer picture than demonstrating shared variability. This is not a critical addition, but I believe would strengthen the work and make it more generally applicable.

Strengths:

- Elegant design of the study. Direct measurements have been made from the tactile sensory neurons to give detailed information on touch encoding. Experiments have been well designed and the forces/displacements have been thoroughly controlled and measured to give accurate measurements of global skin mechanics during a set of controlled mechanical stimuli.

- Analytical techniques used. Analysis of fundamental information coding and information representation in the sensory afferents reveals dynamic coding properties to develop putative models of the neural representation of force. This advanced analysis method has been applied to a large dataset to study neural encoding of force, the temporal dynamics of this, and the variability in this.

Weaknesses:

- Lack of exploration of the variation in neural responses. Although there is a viscoelastic effect that produces variability in the stimulus effects based on prior stimulation, it is a shame that the variability in neural firing and force-induced skin displacements have been presented, and are similarly variable, but there has been no investigation of a link between the two. I believe with these data the authors can go beyond demonstrating shared variability. The force per se is clearly not faithfully represented in the neural signal, being masked by stimulation history, and it is of interest if the underlying resultant contact mechanics are.

Thank you for this suggestion. We have added a new section investigating the link between skin deformation and neural firing in more depth via a simple neural model. Please see our answer below in the ‘Recommendations’ section for further details.

Validity of conclusions:

The authors have succeeded in demonstrating skin viscoelasticity has an impact on skin contact mechanics with a given force and that this impacts the resultant neural coding of force. Their study has been well-designed and the results support their conclusions. The importance and scope of the work is adequately outlined for readers to interpret the results and significance.

Impact:

This study will have important implications for future studies performing tactile stimulation and evaluating tactile feedback during motor control tasks. In detailed studies of tactile function, it illustrates the necessity to measure skin contact dynamics to properly understand the effects of a force stimulus on the skin and mechanoreceptors.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

(Very) minor comments

- The authors say at the beginning of the Results that, "The fourth type of tactile neurons in the human glabrous skin, fast adapting type II neurons...". Although generally written that there are four types of afferent in the glabrous skin, it would be better to state that these are low-threshold A-beta myelinated mechanoreceptive afferents, at least one time, as there are other types of afferent in the glabrous skin that respond to mechanical stimulation (e.g. low and high threshold C-fibers).

This is now clarified at the start of the Results section:

“We recorded action potentials in the median nerve of individual low-threshold A-beta myelinated first-order human tactile neurons innervating the glabrous skin of the fingertip…”

- Fig. 3: Could you add '(N)' as the measurement of force for Fig. 3A for Fz, Fy, and Fz? Also, please change 'Data was recorded' to 'Data were recorded' in the legend.

Fixed.

- At the beginning of the Methods, you say that your study conforms to the Declaration of Helsinki, which actually requires pre-registration in a database. If you did not pre-register your study, please can you add '... in accordance with the Declaration of Helsinki, apart from pre-registration in a database'.

Thanks for making us aware of this. We have added the suggested qualifier to the ethics statement.

Reviewer #2 (Recommendations For The Authors):

The neural representation/encoding of the actual displacement vectors would be a useful addition to the analyses. These vectors have been demonstrated to systematically change with the condition in the irregular series (Figure 2E) and will thus significantly act on the dynamics of induced mechanical changes in the skin with a given interaction force. Thus, it could be examined how the neurons code the magnitude of displacements as well as their direction. An evaluation of the extent to which the imposed displacement magnitudes are encoded in the neural responses would be a useful addition in explaining the signalling of the force events and how the central nervous system decodes these. Evaluating an alternative displacement encoding for comparison to pure force encoding may reveal more about how contact events are represented in the tactile system, which must decode these variable afferent signals to reconstruct a percept of the interaction. It could then be explored how the central nervous system may then scale the dynamic afferent responses based on the background viscoelastic state likely to be present in the SA-II afferent signals (Figure 7) for a context in which to evaluate the dynamic contact forces. This may of course be a complex relationship for the type-I afferents, where the underlying mechanical events evoking the firing (microslips not represented in global forces) have not been measured here. Such a model could be more widely applicable, as the skin viscoelasticity and displacement magnitudes are a straightforward measurement metric and could perhaps be used as a better proxy for neural signalling. This would allow the investigation of a wider variety of forces, and the study of the timing of the viscoelastic effect, both of which have been fixed here. This would give the work a broader impact, rather than just highlighting that this effect produces variability, it could reveal if this mechanical feature is structured in the neural representation. The categorical encoding/decoding tested here is specific to the stimuli used (magnitudes, intervals), but there is the possibility that this may be more generally applicable (within the bounds of forces/speeds) if the underlying basis of the variability in the signalling produced by the viscoelasticity is identified. Since the time course of the viscoelasticity has not been measured here (fixed forces and intervals), further study is required to fully understand the implications this has for a wider variety of situations.

We agree that a better understanding of how the mechanical deformations are reflected in the resulting spike trains would be valuable. While ultimately a full understanding will need precise measurements of skin deformation across the whole fingertip to account for mechanical propagation to mechanoreceptor locations, relating the deformations at the contact location with neural firing patterns directly can provide useful hints into which aspects of deformation are encoded and how. To this end, we ran a new analysis that aimed to predict the time-varying neural responses directly from the recorded mechanical movements of the contactor.

Below we have reproduced the new results and methods text along with the additional figures for this analysis. Note that we have also added text in the Discussion to interpret these findings in the context of our other results.

New section in Results titled Predicting neural responses from contactor movements: “The similarity in the history-dependent variation in neural firing and fingertip deformation at a given force stimulus suggests that neuronal firing is determined by how the fingertip deforms rather than the applied force itself. However, this similarity does not clarify the relationship between fingertip deformation dynamics and neural signaling. To investigate further, we fit cross-validated multiple linear regression models to evaluate how well distinct aspects of contactor movement could predict the time-varying firing rates of individual neurons during the protraction phases of the irregular sequence. The models used predictors based on (1) the three-dimensional position of the contactor, (2) its three-dimensional velocity, (3) a combination of position and velocity signals, and, finally, (4) position and velocity signals along with all possible two-way interactions between them, capturing potentially complex relationship between fingertip deformations and neural signaling.

Comparing the variance explained (R²) by each regression model for each neuron type revealed clear differences between the models (Figure 5A). A two-way mixed design ANOVA, with regression model as within-group effects and neuron type as a between-group effect revealed a main effect of model on variance explained (F(3,462) = 815.5, p < 0.001, η_p² = 0.84). Model prediction accuracy overall increased with the number of predictors, with the two-way interaction model outperforming all others (p < 0.001 for all comparisons, Tukey’s HSD). Additionally, a significant main effect of neuron type (F(2,154) = 29.8, p < 0.001, η_p² = 0.28) and a significant interaction between regression model and neuron type were observed (F(6,462) = 50.8, p < 0.001, η_p² = 0.40).

For neuron type, model predictions were most accurate for SA-2 neurons, followed by SA-1 neurons, with FA-1 neurons showing the lowest accuracy (p < 0.003 for all comparisons, Tukey’s HSD). The interaction between model and neuron type revealed distinct patterns. For SA-1 and SA-2 neurons, position-only and velocity-only models had similar prediction accuracy (p ≥ 0.996, Tukey’s HSD) with no significant differences between these neuron types (p ≥ 0.552, Tukey’s HSD). FA-1 neurons performed poorly with the position-only model but showed higher accuracy with the velocity-only model (p < 0.001, Tukey’s HSD) and better than SA-1 neurons (p = 0.006, Tukey’s HSD). Models combining position and velocity predictors (without interactions) surpassed both position-only and velocity-only models for SA-1 and SA-2 neurons (p < 0.001, Tukey’s HSD). Overall, the differences between neuron types broadly match their tuning to static and dynamic stimulus properties.

The two-way interaction model, accounting for most variance in neural responses, produced mean R² values of 0.75 for FA-1, 0.88 for SA-1, and 0.91 for SA-2 neurons (Figure 5A). To evaluate the contribution of the different predictors, we ranked them using the permutation feature importance method, focusing on the six most important ones. Regression analyses using only these variables explained almost all of the variance explained by the full model, with a median R² reduction of just 0.055 across all neurons. Across all neuron types, at least half included all three velocity components (dPx, dPy, dPz) among the top six, with FA-1 neurons showing the highest prevalence (Figure 5B). Interactions between normal position (Pz) and each velocity component were also frequently observed, while interactions involving tangential position and velocity components were less common. Interactions among velocity components were relatively well represented, followed by interactions limited to position components. Position signals were generally less represented, except for normal position (Pz) in slowly adapting neurons, where it appeared in 50% of SA-1 and 68% of SA-2 neurons. Despite these broad trends, important predictors varied widely across ranks even within a given neuron class (see Figure 5-figure supplement 1), and even the most frequent variables appeared in only a subset of cases, suggesting broad variability in sensitivity across neurons.”

New methods paragraph titled Predicting time-varying firing rates from skin deformations:

“This analysis was conducted in Python (v3.13) with pandas for data handling, numpy for numerical operations, and scikit-learn for model fitting and evaluation.

To assess how well individual neurons' time-varying firing rates could be predicted from simultaneous contactor movements, we fitted multiple linear regression models (see Khamis et al., 2015, for a similar approach}. This analysis focused on the force protraction phase of the irregular sequence, where neurons were most responsive and sensitive to stimulation history. Data from 100 ms before to 100 ms after the protraction phase (between -0.100 s and 0.225 s relative to protraction onset) were included for each trial. Neurons were included if they fired at least two action potentials during the force protraction phase and the following 100 ms in at least five of the 25 trials. This ensured sufficient variability in firing rates for meaningful regression analysis, resulting in 68 SA-1, 38 SA-2, and 51 FA-1 neurons being included.

Contractor position signals digitized at 400 Hz were linearly interpolated to 1000 Hz. Instantaneous firing rates, derived from action potentials sampled at 12.8 kHz, were resampled at 1000 Hz to align with position signals. A Gaussian filter (σ = 10 ms, cutoff ~16 Hz) was applied to the firing rate as well as to the position signals before differentiation. To account for axonal conduction (8–15 ms) and sensory transduction delays (1–5 ms), firing rates were advanced by 15 ms to align approximately with independent variables.

Regressions were performed using scikit-learn's Ridge and RidgeCV regressors, which apply L2 regularization to mitigate overfitting. Hyperparameter tuning for the regularization parameter (alpha) was performed using GridSearchCV with a predefined range (0.001–1000.0), incorporating five-fold cross-validation to select the best value. To minimize overfitting risks, model performance was further validated with independent five-fold cross-validation (KFold), and R² scores were computed using cross_val_score.

We constructed four linear regression models with increasing complexity: (1) Position-only, using three-dimensional contactor positions (Px, Py, Pz); (2) Velocity-only, using three-dimensional velocities (dPx, dPy, dPz); (3) Combined, including all position and velocity signals (6 predictors); and (4) Interaction, including all signals and their two-way interactions (21 predictors). All features were standardized using StandardScaler to improve regularization and model convergence. PolynomialFeatures generated second-order interaction terms for the interaction model. Feature importance was evaluated with permutation_importance, and simpler models were built using the most important features. These models were validated through cross-validation to assess retained explanatory power.”

Minor:

- It would be useful to add a brief description of the material aspects of the contactor tip to the methods (as per Birznieks 2001).

We have added the following statement:

“To ensure that friction between the contactor and the skin was sufficiently high to prevent slips, the surface was coated with silicon carbide grains (50–100 μm), approximating the finish of smooth sandpaper.”

- The axes labelling on Figure 3A and legend description is ambiguous, probably placing the Px, Py, and Pz labels on the far left axes and the Fx, Fy, and Fz on the right side of the far right axes would make this clearer.

Label placement has been improved along with some other minor fixes.

- For the quasi-static phase analysis, the phrase "absence of loading" used in reference to the interstimulus period and SA-II afferents does not seem to be a correct description. The finger is still loaded (at least in the normal direction), with a magnitude of imposed displacement that counteracts the viscoelastic force exerted by the skin mechanics of the fingertip. Although there is a zero net-force load, a mechanical stimulus is still being actively applied to the skin.

We have changed the wording throughout the text and now consistently refer either to the “interstimulus period” directly or to an “absence of externally applied stimulation” to avoid confusion.

This passage really reflects the "tacit understanding" in many classrooms - if students don't cause trouble, the teacher can easily finish the class, and everyone doesn't make things difficult for each other. I used to feel this atmosphere in high school. Some students in the class didn't study much, but as long as they didn't disturb others, the teacher would let them "slack off" by default. It's like an unspoken rule of "we don't undermine each other." The author quoted Sizer's "Let's Make a Deal" to satirize this seemingly calm but lacking in-depth communication in education. I think this "transactional" teaching atmosphere may seem to be easy in the short term, but in the long run, it will make teaching lose its real challenge and meaning.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This short report shows that the transcription factor gene mirror is specifically expressed in the posterior region of the butterfly wing imaginal disk, and uses CRISPR mosaic knock-outs to show it is necessary to specify the morphological features (scales, veins, and surface) of this area.

Strengths:

The data and figures support the conclusions. The article is swiftly written and makes an interesting evolutionary comparison to the function of this gene in Drosophila. Based on the data presented, it can now be established that mirror likely has a similar selector function for posterior-wing identity in a plethora of insects.

We thank the reviewer for their feedback.

Weaknesses:

This first version has minor terminological issues regarding the use of the terms "domains" and "compartment".

We acknowledge that the terminologies “domains” and “compartments” might lead to confusion. To avoid confusion we have removed the term “compartment” from the manuscript.

Reviewer #2 (Public Review):

This is a short and unpretentious paper. It is an interesting area and therefore, although much of this area of research was pioneered in flies, extending basic findings to butterflies would be worthwhile. Indeed, there is an intriguing observation but it is technically flawed and these flaws are serious.

The authors show that mirror is expressed at the back of the wing in butterflies (as in flies). They present some evidence that is required for the proper development of the back of the wing in butterflies (a region dubbed the vannus by the ancient guru Snodgrass). But there are problems with that evidence. First, concerning the method, using CRISP they treat embryos and the expectation is that the mirror gene will be damaged in groups of cell lineages, giving a mosaic animal in which some lines of cells are normal for mirror and others are not. We do not know where the clones or patches of cells that are defective for mirror are because they are not marked. Also, we do not know what part of the wing is wild type and what part is mutant for mirror. When the mirror mutant cells colonise the back of the wing and that butterfly survives (many butterflies fail to develop), the back of the wing is altered in some selected butterflies. This raises a second problem: we do not know whether the rear of the wing is missing or transformed. From the images, the appearance of the back of the wing is clearly different from the wild type, but is that due to transformation or not? And then I believe we need to know specifically what the difference is between the rear of the wing and the main part. What we see is a silvery look at the back that is not present in the main part, is it the structure of the scales? We are not told.

Thank you for this feedback. We appreciate that many readers may not accustomed to looking at mosaic knockouts. As discussed in a previous review article (Zhang & Reed 2017), we rely on a combination of contralateral asymmetry and replicates to infer mutant phenotypes. For many genes (e.g. pigmentation enzymes) mutant clones are obvious, but for other types of genes (e.g. ligands) clone boundaries are sometimes not directly diagnosable. It is simply a limitation of our study system. Nonetheless, you see for yourself that “the back of the wing is altered in some butterflies” – the effects of deleting mirror are clear and repeatable.

In terms of interpreting mutant phenotypes, we agree that that paper would benefit from a better description of the specific effects. Therefore, we have included an improved, more systematic description of phenotypes, along with better-annotated figures showing changes in wing shape and venation, scale coloration, and color pattern transformation (e.g. posterior elongation of the orange marginal stripes).

There are other problems. Mirror is only part of a group of genes in flies and in flies both iroquois and mirror are needed to make the back of the wing, the alula (Kehl et al). What is known about iro expression in butterflies?

In Drosophila mirror, araucan, and caupolican comprise the so-called Iroqouis Complex of genes. As denoted in Figure S4 and in Kerner et al (doi: https://doi.org/10.1186/1471-2148-9-74) the divergence of araucan and caupolican into two separate paralogs is restricted to Drosophila. As in most insects, butterflies have only two Iroquois Complex genes: araucan and mirror. We tested the role of araucan in Junonia coenia as shown in our pre-print: https://doi.org/10.1101/2023.11.21.568172. Its expression appears to be restricted to early pupal wings where it is transcribed in all scale-forming cells. Mosaic araucan KOs resulted in a change in scale iridescent coloration associated with changes in the laminar thickness of scale cells.  

In flies, mirror regulates a late and local expression of dpp that seems to be responsible for making the alula. What happens in butterflies? Would a study of the expression of Dpp in wildtype and mirror compromised wings be useful?

We thank the reviewer for the proposal and agree that a future study comparing Dpp in wild-type versus mirror KO butterflies would be useful to clarify the mechanism of Dpp signalling in wing development. It is not clear, however, that the results of a Dpp experiment would change the conclusions of our current study therefore we decided not to undertake these additional experiments for our revision.

Thus, I find the paper to be disappointing for a general journal as it does little more than claim what was discovered in Drosophila is at least partly true in butterflies. 

We respect that the reviewer does not have a strong interest in the comparative aspects of this study. Fair enough. This report is primarily aimed at biologists interested in the evolutionary history of insect wings.

Also, it fails to explain what the authors mean by "wing domains" and "domain specification". They are not alone, butterfly workers, in general, appear vague about these concepts, their vagueness allowing too much loose thinking.

A domain is “a region distinctively marked by some physical feature”. This term is used extensively in the developmental biology literature (e.g. “expression domain”, “embryonic domain”, “tissue domain”, “domain specification”) and is found throughout popular textbooks (e.g. Alberts et al. “The Cell”, Gilbert “Developmental Biology”). We prefer the term “domain” because of its association in the Drosophila literature with transcription factors that define fields of cells. We specifically avoided using the term “compartment” because of its association with cell lineage, which we have not tested. 

Since these matters are at the heart of the purpose and meaning of the work reported here, we readers need a paper containing more critical thought and information. I would like to have a better and more logical introduction and discussion.

We would like the very same thing, of course, and we hope the reviewer finds our revised manuscript to be more satisfying to read.

The authors do define what they mean by the vannus of the wing. In flies the definition of compartments is clear and abundantly demonstrated, with gene expression and requirement being limited precisely to sets of cells that display lineage boundaries. It is true that domains of gene expression in flies, for example of the iroquois complex, which includes mirror, can only be related to patterns with difficulty. Some recap of what is known plus the opinion of the authors on how they interpret papers on possible lineage domains in butterflies might also be useful as the reader, is no wiser about what the authors might mean at the end of it!

We thank the reviewer for this suggestion. However, our experiments have little to contribute to the topic of cell lineage compartmentalization. We have therefore opted to avoid speculating on this topic to prevent confusion and to keep the manuscript focused on our experimental results.

The references are sometimes inappropriate. The discovery of the AP compartments should not be referred to Guillen et al 1995, but to Morata and Lawrence 1975. Proofreading is required.

We thank the reviewer for suggesting this important reference. We have included it in our revision.

Reviewer #3 (Public Review):

Summary:

The manuscript by Chatterjee et al. examines the role of the mirror locus in patterning butterfly wings. The authors examine the pattern of mirror expression in the common buckeye butterfly, Junonia coenia, and then employ CRISPR mutagenesis to generate mosaic butterflies carrying clones of mirror mutant cells. They find that mirror is expressed in a well-defined posterior sector of final-instar wing discs from both hindwings and forewings and that CRISPR-injected larvae display a loss of adult wing structures presumably derived from the mirror expressing region of hindwing primordium (the case for forewings is a bit less clear since the mirror domain is narrower than in the hindwing, but there also do seem to be some anomalies in posterior regions of forewings in adults derived from CRISPR injected larvae). The authors conclude that the wings of these butterflies have at least three different fundamental wing compartments, the mirror domain, a posterior domain defined by engrailed expression, and an anterior domain expressing neither mirror nor engrailed. They speculate that this most posterior compartment has been reduced to a rudiment in Drosophila and thus has not been adequately recognized as such a primary regional specialization.

Critique:

This is a very straightforward study and the experimental results presented support the key claims that mirror is expressed in a restricted posterior section of the wing primordium and that mosaic wings from CRISPR-injected larvae display loss of adult wing structures presumably derived from cells expressing mirror (or at least nearby). The major issue I have with this paper is the strong interpretation of these findings that lead the authors to conclude that mirror is acting as a high-level gene akin to engrailed in defining a separate extreme posterior wing compartment. To place this claim in context, it is important in my view to consider what is known about engrailed, for which there is ample evidence to support the claim that this gene does play a very ancestral and conserved function in defining posterior compartments of all body segments (including the wing) across arthropods.

(1) Engrailed is expressed in a broad posterior domain with a sharp anterior border in all segments of virtually all arthropods examined (broad use of a very good panspecies anti-En antibody makes this case very strong).

(2) In Drosophila, marked clones of wing cells (generated during larval stages) strictly obey a straight anterior-posterior border indicating that cells in these two domains do not normally intermix, thus, supporting the claim that a clear A/P lineage compartment exists.

In my opinion, mirror does not seem to be in the same category of regulator as engrailed for the following reasons:

(1) There is no evidence that I am aware of, either from the current experiments, or others that the mirror expression domain corresponds to a clonal lineage compartment. It is also unclear from the data shown in this study whether engrailed is co-expressed with mirror in the posterior-most cells of J. coenia wing discs. If so, it does not seem justified to infer that mirror acts as an independent determinant of the region of the wing where it is expressed.

(2) Mirror is not only expressed in a posterior region of the wing in flies but also in the ventral region of the eye. In Drosophila, mirror mutants not only lack the alula (derived approximately from cells where mirror is expressed), but also lack tissue derived from the ventral region of the eye disc (although this ventral tissue loss phenotype may extend beyond the cells expressing mirror).

In summary, it seems most reasonable to me to think of mirror as a transcription factor that provides important development information for a diverse set of cells in which it can be expressed (posterior wing cells and ventral eye cells) but not that it acts as a high-level regulator as engrailed.

Recommendation:

While the data provided in this succinct study are solid and interesting, it is not clear to me that these findings support the major claim that mirror defines an extreme posterior compartment akin to that specified by engrailed. Minimally, the authors should address the points outlined above in their discussion section and greatly tone down their conclusion regarding mirror being a conserved selector-like gene dedicated to establishing posterior-most fates of the wing. They also should cite and discuss the original study in Drosophila describing the mirror expression pattern in the embryo and eye and the corresponding eye phenotype of mirror mutants: McNeill et al., Genes & Dev. 1997. 11: 1073-1082; doi:10.1101/gad.11.8.1073.

We thank the reviewer for their summary, critique, and recommendations. We agree with everything the reviewer says. Honestly, however, we were surprised by these comments because we took great care in the paper to never refer to mirror as a compartmentalization gene or claim it has a function in cell lineage compartmentalization like engrailed. As pointed out, we lack clonal analyses to test for compartmentalization. This is why we used the term “domain” instead of “compartment” in the title and throughout the manuscript. Nevertheless, we have recrafted the discussion in the manuscript, including completely removing the term “compartment”, to better avoid implications that mirror plays a role in cell lineage compartmentalization. 

We also thank the reviewer for recommending the paper about the role of mirror in eye development. For the sake of keeping the paper focused, however, we decided not to broach the topic of mirror functions outside the context of wing development.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

I have minor comments for improvement.

The abstract and introductions are terminologically problematic when they refer to the concept of compartment and compartment boundaries. Allegedly this confusion has previously propagated in several articles related to butterfly wing development, which keeps alienating this literature from being taken seriously by fly specialists, for example. So it is important to use the right terms. I will try to explain point by point here, but I would appreciate it if the authors could undertake a significant rewrite taking these comments into account. The authors use the terms compartment and compartment boundary. This has a very specific use in developmental genetics: mitotic clones never cross a boundary (or compartment). I think the authors can keep referring to the equivalent of the A-P boundary, which is situated somewhere between M1-M2 based on unpublished data from the Patel Lab, and is not very well defined (Engrailed expression moves a little bit during development in this area). Domain is a looser term and can be used more liberally to describe genetically defined regions.

- "Classical morphological work subdivides insect wings into several distinct domains along the antero-posterior (AP) axis, each of which can evolve relatively independently." Yes. This concept of domain and individuation seems important. You could make a proposed link to selector genes here.

- "There has been little molecular evidence, however, for AP subdivision beyond a single compartment boundary described from Drosophila melanogaster." Incorrect, and this conflates "domain" and "compartment".

Flies have wing AP domains too, that pattern their veins (see the cited Banerjee et al). 

- "Our results confirm that insect wings can have more than one posterior developmental domain, and support models of how selector genes may facilitate evolutionarily individuation of distinct AP domains in insect wings". Yes, and I like the second part of the sentence. Still, I would recommend simply deleting "confirm that insect wings can have more than one posterior developmental domain, and" because this is neglecting previous work on AP genetic regionalization in both flies (vein literature) and butterflies (e.g. McKenna and Nijhout, Banerjee et al).

- "Analyses of wing pattern diversity across butterflies, considering both natural variation and genetic mutants, suggest that wings can be subdivided into at least five AP domains, bounded by the M1, M3, Cu2, and 2A veins respectively, within each of which there are strong correlations in color pattern variation and wing morphology (Figure 1A)". Yes, and I would recommend emphasizing they correspond to welldefined gene expression domains as mentioned in Banerjee et al, or McKenna and Nijhout.

- "The anterior-most of these domains, bordered by the M1 vein, appears to correspond to an AP compartment boundary originally described by cell lineage tracing in Drosophila melanogaster, and later supported in butterfly wings by expression of the Engrailed transcription factor. Interestingly, however, D. melanogaster work has yet to reveal clear evidence for additional AP domain boundaries in the wing." Confusingly, because the first sentence is about compartments while the second is about AP domains. I also think the claim that Dmel has no other known AP domains is dubious because Spalt is highly regionalized in flies.

- "Previous authors have proposed the existence of such individuated domains, and speculated that they may be specified by selector genes.5,10 Our data provide experimental support for this model, and now motivate us to identify factors that specify other domain boundaries between the M1 and A2 veins." Yes, I completely agree with this way to emphasize the selector effect, and to link it to the concept of "individuated domain"

We cannot thank the reviewer enough for the time and thought they devoted to giving helpful suggestions to improve our manuscript. We have applied all of the above recommendations to the revision.

Fig. S1: the field needs to move away from Red/Green microscopy images, for accessibility reasons.

The easiest fix here would be to change the red channels to magenta.

Green/Magenta provides excellent contrast and accessibility in general in 2-channel images.

We thank the reviewer for this suggestion. We have improved the color accessibility of Fig. S1.

definition memex

One cannot hope thus to equal the speed and flexibility with which the mind follows an associative trail, but it should be possible to beat the mind decisively in regard to the permanence and clarity of the items resurrected from storage.

selection by association may be mechanized

atlantic

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Kv2 subfamily potassium channels contribute to delayed rectifier currents in virtually all mammalian neurons and are encoded by two distinct types of subunits: Kv2 alpha subunits that have the capacity to form homomeric channels (Kv2.1 and Kv2.2), and KvS or silent subunits (Kv5,6,8.9) that can assemble with Kv2.1 or Kv2.2 to form heteromeric channels with novel biophysical properties. Many neurons express both types of subunits and therefore have the capacity to make both homomeric Kv2 channels and heteromeric Kv2/KvS channels. Determining the contributions of each of these channel types to native potassium currents has been very difficult because the differences in biophysical properties are modest and there are no Kv2/KvS-specific pharmacological tools. The authors set out to design a strategy to separate Kv2 and Kv2/KvS currents in native neurons based on their observation that Kv2/KvS channels have little sensitivity to the Kv2 pore blocker RY785 but are blocked by the Kv2 VSD blocker GxTx. They clearly demonstrate that Kv2/KvS currents can be differentiated from Kv2 currents in native neurons using a two-step strategy to first selectively block Kv2 with RY785, and then block both with GxTx. The manuscript is beautifully written; takes a very complex problem and strategy and breaks it down so both channel experts and the broad neuroscience community can understand it.

Strengths:

The compounds the authors use are highly selective and unlikely to have significant confounding cross-reactivity to other channel types. The authors provide strong evidence that all Kv2/KvS channels are resistant to RY785. This is a strength of the strategy - it can likely identify Kv2/KvS channels containing any of the 10 mammalian KvS subunits and thus be used as a general reagent on all types of neurons. The limitation then of course is that it can't differentiate the subtypes, but at this stage, the field really just needs to know how much Kv2/KvS channels contribute to native currents and this strategy provides a sound way to do so.

Weaknesses:

The authors are very clear about the limitations of their strategy, the most important of which is that they can't differentiate different subunit combinations of Kv2/KvS heteromers. This study is meant to be a start to understanding the roles of Kv2/KvS channels in vivo. As such, this is a minor weakness, far outweighed by the potential of the strategy to move the field through a roadblock that has existed since its inception.

The study accomplishes exactly what it set out to do: provide a means to determine the relative contributions of homomeric Kv2 and heteromeric Kv2/KvS channels to native delayed rectifier K+ currents in neurons. It also does a fabulous job laying out the case for why this is important to do.

Reviewer #2 (Public Review):

Summary:

Silent Kv subunits and the channels containing these Kv subunits (Kv2/KvS heteromers) are in the process of discovery. It is believed that these channels fine-tune the voltage-activated K+ currents that repolarize the membrane potential during action potentials, with a direct effect on cell excitability, mostly by determining action potentials firing frequency.

Strengths:

What makes silent Kv subunits even more important is that, by being expressed in specific tissues and cell types, different silent Kv subunits may have the ability to fine-tune the delayed rectifying voltage-activated K+ currents that are one of the currents that crucially determine cell excitability in these cells. The present manuscript introduces a pharmacological method to dissect the voltage-activated K+ currents mediated by Kv2/KvS heteromers as a means of starting to unveil their importance, together with Kv2-only channels, to the cells where they are expressed.

Weaknesses:

While the method is effective in quantifying these currents in any isolated cell under an electric voltage clamp, it is ineffective as a modulating maneuver to perhaps address these currents in an in vivo experimental setting. This is an important point but is not a claim made by the authors.

We agree. We have now stated in the introduction that this study does not address the roles of Kv2/KvS currents in an in vivo setting.

Manuscript revisions:

While this study does not address the impact of GxTX or RY785 on action potentials or in vivo, the distinct pharmacology of Kv2/KvS heteromers presented here suggests that KvS conductances could be targeted to selectively modulate discrete subsets of cell types.  

There are other caveats with the methods and data:

(i) The need for a 'cocktail' of blockers to supposedly isolate Kv2 homomers and Kv2/KvS heteromers' currents from others may introduce errors in the quantification Kv2/KvS heteromers-mediated K+ currents and that is due to possible blockers off targets.

We now point out that is possible that off target effects of blockers may introduce errors, include references that identify the selectivity of the blockers used in the cocktail, and specifically note that 4-aminopyridine in the cocktail is expected to block 2% of Kv2 homomers yet have a lesser impact Kv2/KvS heteromers. Additionally, to test whether the KvS isolation strategy requires the cocktail in neurons, we performed new experiments on a different subclass of nociceptors without the blocker cocktail and identified a substantial KvS-like component (new Fig 7 Supplement 3).

Manuscript revisions:

“After whole-cell voltage clamp was established, non-Kv2/KvS conductances were suppressed by changing to an external solution containing a cocktail of inhibitors: 100 nM alpha-dendrotoxin (Alomone) to block Kv1 (Harvey and Robertson, 2004), 3 μM AmmTX3 (Alomone) to block Kv4 (Maffie et al., 2013; Pathak et al., 2016), 100 μM 4-aminopyridine to block Kv3 (Coetzee et al., 1999; Gutman et al., 2005), 1 μM TTX to block TTX sensitive Nav channels, and 10 μM A803467 (Tocris) to block Nav1.8 (Jarvis et al., 2007). It is possible that off target effects of blockers may introduce errors in the quantification Kv2/KvS heteromer-mediated K⁺ currents. For example, 4-aminopyridine is expected to block a small fraction, 2%, of Kv2 homomers and have a lesser impact on Kv2/KvS heteromers (Post et al., 1996; Thorneloe and Nelson, 2003; Stas et al., 2015) which could result in a slight overestimation of the ratio of Kv2/KvS heteromers to Kv2 homomers.”

“We also tested the other major mouse C-fiber nociceptor population, peptidergic nociceptors, to determine if this subpopulation also has conductances resistant to RY785 yet sensitive to GxTX. We voltage clamped DRG neurons from a CGRP^GFP mouse line that expresses GFP in peptidergic nociceptors (Gong et al., 2003). Deep sequencing has identified mRNA transcripts for Kv6.2, Kv6.3, Kv8.1 and Kv9.3 present in GFP+ neurons, an overlapping but distinct set of KvS subunits from the Mrgprd^GFP non-peptidergic population (Zheng et al., 2019). In GFP+ neurons from CGRP^GFP mice, we found that a fraction of outward current was inhibited by 1 µM RY785 and additional current inhibited by 100 nM GxTX (Fig 7 Supplement 3 A-C). In these experiments, 58 ± 2% (mean ± SEM) was KvS-like (Fig 7 Supplement 3 D) identifying that KvSlike conductances are present in these peptidergic nociceptors. For CGRP^GFP neurons we did not include the Kv1, Kv3, Kv4, Nav and Cav channel inhibitor cocktail used for other neuron experiments, indicating that the cocktail of inhibitors is not required to identify KvS-like conductances.”

(ii) During the electrophysiology experiments, the authors use a holding potential that is not as negative as it is needed for the recording of the full population of the Kv2/KvS channels. Depolarized holding potentials lead to a certain level of inactivation of the channels, that vary according to the KvS involved/present in that specific population of channels. As a reminder, some KvS promote inactivation and others prevent inactivation. Therefore, the data must be interpreted as such.

We agree. We now point out that the physiological holding potentials used are insufficiently negative to relieve inactivation from all Kv2/KvS heteromeric channels. We also note that the ratio of Kv2-like to KvS-like conductance is expected to vary with voltage protocols.

Manuscript revisions:

“Neurons were held at a membrane potential of –74 mV to mimic a physiological resting potential. KvS subunits can profoundly shift the voltage-inactivation relation (Salinas et al., 1997a; Kramer et al., 1998; Kerschensteiner and Stocker, 1999) and this potential is likely insufficiently negative to relieve inactivation from all Kv2/KvS heteromeric channels. Also, the activation membrane potential is close to the half-maximal point of Kv2/KvS conductances. Thus the ratio of Kv2-like to KvS-like conductance is expected to vary with voltage protocols.”

(iii) The analysis of conductance activation by using tail currents is only accurate when dealing with non-inactivating conductances. Also, in dealing with a heterogenous population of Kv2/KvS heteromers, heterogenous K+ conductance deactivation kinetics is a must. Indeed, different KvS may significantly relate to different deactivation kinetics as well.

We now discuss that the bi-exponential fit of tail currents is likely inadequate to capture the deactivation kinetics of all underlying components of a heterogenous population of Kv2/KvS heteromers.

Manuscript revisions:

“We note that the analysis of conductance activation by using tail currents is only accurate when dealing with non-inactivating conductances. We expect that inactivation of Kv2/KvS conductances during the 200 ms pre-pulse is minimal (Salinas et al., 1997a; Kramer et al., 1998; Kerschensteiner and Stocker, 1999) and did not notice inactivation during the activation pulse. Also, deactivation kinetics can vary in a heterogenous population of Kv2/KvS heteromers. While analysis of tail currents could skew the quantification of total Kv2 like and KvS-like conductances, our data supports that mouse nociceptors and human neurons have tail currents that are resistant to RY785 and sensitive to GxTX consistent with the presence of Kv2/KvS heteromers.”

(iv) Silent Kv subunits may be retained in the ER, in heterologous systems like CHO cells. This aspect may subestimate their expression in these systems. Nevertheless, the authors show similar data in CHO cells and in primary neurons.

We agree. We now note that in heterologous systems, including CHO cells, transfection of KvS subunits can result in KvS subunits that are retained intracellularly.

Manuscript revisions:

“While a fraction of KvS subunits appear to be retained intracellularly, immunofluorescence for Kv5.1, Kv9.3 and Kv2.1 also appeared localized to the perimeter of transfected Kv2.1-CHO cells (Figure 1 Supplement).”

(v) The hallmark of silent Kv subunits is their effect on the time inactivation of K+ currents. As such, data should be shown throughout, preferably, from this perspective, but it was only done so in Figure 4G.

Indeed, effects on inactivation are a hallmark of KvS subunits. However, quantifying inactivation of Kv2/KvS channels requires steps to positive voltages for approximately 10 seconds. In neurons steps this long usually resulted in irreversible changes in leak currents/input resistance that degraded the accuracy of RY785/GxTX subtraction currents. Consequently, we did not acquire inactivation data in neurons, and we now explain in the manuscript why such data was not obtained.

Manuscript revisions:

“While changes in inactivation are prominent with KvS subunits, we did not investigate inactivation in neurons because the lengthy depolarizations required often resulted in irreversible leak current increases that degraded the accuracy of RY785/GxTX subtraction current quantification.”

(vi) Functional characterization of currents only, as suggested by the authors as a bona fide of Kv2 and Kv2/KvS currents, should not be solely trusted to classify the currents and their channel mediators.

We agree, and now state explicitly that functional characterization cannot be trusted to classify their channel mediators of conductances, and we try to be clear about this throughout the manuscript by using soft terms such as "KvS-like" when identity is uncertain.

Manuscript revisions:

“As functional characterization alone cannot be trusted to classify their channel mediators of conductances, we define conductances consistent with Kv2/KvS heteromers as 'KvS-like' and conductances consistent with Kv2 homomers as 'Kv2-like'.”

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

There is not a lot to do here - this was a real pleasure to read and very easy to understand, as written. Here are a few minor things to consider:

(1) The naming of the KvS subunits has always been confusing - it is not clear that Kv5,6,8,9 are members of the Kv2 subfamily from the names. KvS does a good job of differentiating them by assembly phenotype and has been used a lot in the literature, but it doesn't solve the misconception of what subfamily they belong to. This might not matter so much for mammals, where all KvS channels are in the Kv2 subfamily, but it makes it impossible to extend the naming system to other animals where subunits requiring heteromeric assembly are common in most subfamilies. How about trying the name Kv2S? It would have continuity with KvS in the reader's mind, make it clear that they are Kv2 subfamily, and make a naming system that could be extended beyond vertebrates. This is not a problem the authors created - just a completely optional suggestion on how to solve it if so inclined.

We agree that naming conventions for these subunits are problematic, and agonized quite a bit about nomenclature. In the end we chose to stick with the precedent of KvS.

(2) Another naming issue they should definitely change is the use of "subfamily" for the different KvS subtypes (Kv5, Kv6, Kv8, and Kv9). This really creates confusion with the higher-order subfamilies that have a very clear functional definition: a subfamily of Kv genes is a group of related genes that have assembly compatibility. Those are Kv1, Kv2, Kv3 and Kv4. KvS genes are assembly compatible with Kv2, evolutionarily derived from the Kv2 lineage, and thus clearly a part of the Kv2 subfamily. Using a subfamily for the next lower level of the naming hierarchy confuses this. The authors should use different terms like sub-type or class or subgroups for the divisions within KvS.

Thank you. We have standardized to Kv2/KvS as a subfamily; Kv5, Kv6, Kv8, and Kv9 as subtypes; and individual proteins, e.g. Kv8.1, as subunits.

(3) When you discuss whether the KvS subunit directly disrupts Ry785 binding in the pore or works allosterically and you said you know which KvS residues point into the pore from models, I thought that maybe you could tell from a sequence alignment whether the KvS channels you didn't test look the same in the conduction pathway as the ones you did test. If so, you could mention that if the binding site is the pore, they should all be resistant. Alternatively, if one you didn't test looks fundamentally more similar to the Kv2s in this region, then maybe it could be fingered as a possible exception that needs to be tested later.

Great ideas. We now assess sequence KvS variability near the proposed RY785 binding site in all KvS subunits. We generated structural models of RY785 docking to Kv2.1 and Kv2.1/Kv8.1 and found that residues near RY785 are different in all KvS subunits.

Manuscript revisions:

“We analyzed computational structural models of RY785 docked to a Kv2.1 homomer and a 3:1 Kv2.1:Kv8.1 heteromer (Fig 9) to gain structural insight into how KvS subunits might interfere with RY785 binding. We used Rosetta to dock RY785 to a cryo-EM structure of a Kv2.1 homomer in an apparently open state (Fernández-Mariño et al., 2023). The top-scoring docking pose has RY785 positioned below the selectivity filter and off-axis of the pore (Fig 9 A), similar to a stable pose observed in molecular dynamic simulations (Zhang et al., 2024). In this pose, RY785 contacts a collection of Kv2.1 residues that vary in every KvS subtype (Fig 9 B,D,E). Notably, RY785 bound similarly to a 3:1 model of Kv2.1/Kv8.1, in contact with the three Kv2.1 subunits, yet avoided the Kv8.1 subunit (Fig 9C). This is consistent with RY785 binding less well to Kv2.1/Kv8.1 heteromers, and also suggests that a 3:1 Kv2:KvS channel could retain a RY785 binding site when open.”

(4) Future suggestion or tip - not for this paper. Your data shows your isolation strategy works really well on Kv6 channels, and these are also the Kv2/KvS channels that have the most pronounced biophysical changes. Working on neurons that have a prominent Kv2/Kv6 component would really show how well the strategy outlined here works to describe the physiology of native neurons. The highest KvS expression I have seen in public data in a wellstudied cell type is Kv6.4 in spinal motor neurons.

Wonderful tip, thank you. We are indeed very interested in Kv6.4 in spinal motor neurons.

Reviewer #2 (Recommendations For The Authors):

The manuscript makes a good contribution to the identification of Kv2/KvS channels in primary cells. The pharmacological method proposed by the authors to dissect the currents in an experimental setting seems proper. Although meritorious in themselves, the findings are heavily phenomenological in the opinion of this reviewer. The manuscript should be improved with some level of mechanistic data and/or the demonstration of different levels of expression in different cell types.

Thank you for the suggestions. This manuscript now demonstrates strikingly higher levels of the KvS-like component of Kv2 currents in somatosensory (DRG nonpeptidergic and peptidergic nociceptor) versus autonomic (SCG) neuron types. The mechanistic question of what electrophysiological properties the KvS subunits are providing to the neuronal circuit is an exciting one that we are pursuing separately.

Manuscript revisions:

“While we found only RY785-sensitive Kv2-like conductances in SCG neurons, Kv2/KvS heteromer-like conductances were dominant in DRG neurons.”

At present, the manuscript says that the combination of RY785 and guangxitoxin-1E can be used to define Kv2/KvS-mediated K+ currents. Importantly, this method cannot be used in a way that one can functionally determine the function of Kv2/KvS channels, since it depends on the pre-blocking of Kv2-mediated K+ currents prior. In the opinion of this reviewer, this fact decreases the attention of a potential reader.

Indeed, our study is focused on revealing KvS heteromers by voltage clamp, and we now clarify in the introduction that we do not determine the function of Kv2/KvS channels in this study, so as not to lead the reader to expect studies of neuronal signaling.

However, the selective pharmacology we identify suggests RY785 application could reveal the function of Kv2 homomers, and for RY785-insensitive signaling, GxTX application of could reveal the function of Kv2/KvS heteromers. We now mention these possible applications in the Discussion.

Manuscript revisions:

“While this study does not address the impact of GxTX or RY785 on action potentials or in vivo, the distinct pharmacology of Kv2/KvS heteromers presented here suggests that KvS conductances could be targeted to selectively modulate discrete subsets of cell types.”

Please find below suggestions for improving the manuscript:

(1) The term "Kv2/KvS heteromers" should be used throughout instead of variations such as "Kv2/KvS channels", "Kv2/KvS" and others. Standardization of the term to refer to heteromers would make the manuscript easier to read.

Thank you. We have standardized terms to consistently refer to Kv2/KvS heteromers.

(2) Confusing terms like KvS conductances, KvS-like conductances, KvS-like (RY785-resistant, GxTX-sensitive) currents, and KvS channels should be avoided because they disregard the current belief that KvS cannot form functional homomeric channels. The term KvS-containing channels, and Kv2/KvS channels, seem more accurate. Uniformization in this regard will also make the manuscript more easily readable.

Thank you. We have standardized terms to Kv2/KvS heteromers and KvS-containing channels when channel subunits are known and the use terms KvS-like and Kv2-like for functionally identified endogenous conductances with unknown channel subunits.

(3) Referring to KvS as a regulatory subunit is inaccurate. It is clear that KvS is part of, and it makes up the alpha pore. KvS therefore is a part of the conductive pathway and not a regulatory (suggesting accessory) subunit. KvS take part in selectivity filter (fully conserved), but they also make up an important part of the conducting pathway with non-conserved amino acid residues.

We felt it important to include the descriptor “regulatory” to connect our nomenclature with prior use of the descriptor in the literature, and now only use the term at the start of the introduction.

Manuscript revisions:

“A potential source of molecular diversity for Kv2 channels are a group of Kv2-related proteins which have been referred to as regulatory, silent, or KvS subunits.”

(4) The use of a cocktail of channel inhibitors may affect the quantification of Kv2/KvS heteromers-mediated K+ currents because they may interact with RY785 and/or GxTx or they may even interact with the sites for these two drugs on Kv2-containing channels.

This is an interesting point worth considering, thank you. We now alert readers to this possibility in the discussion when considering the limitations of our approach.

Manuscript revisions:

“Also, the cocktail of inhibitors used in most neuron experiments here could potentially alter RY785 or GxTX action against KvS/Kv2 channels.”

(5) The graphical representation of fractional blocking and other parameters (e.g., Fig 1D), is hard to read in these slim plots. In my opinion, tall bars would be more meaningfully visualized.

Thank you for pointing out that the graphs were hard to read, we have made the graph easier to read and added tall bars.

(6) Vehicle control for IHC and electrophysiology. Please state what is the vehicle used in the electrophysiology experiments.

Thank you. The composition of vehicle has now been stated in the methods.

Manuscript revisions:

“All RY785 solutions contained 0.1% DMSO. Vehicle control solutions also contained 0.1% DMSO but lacked RY785.”

“Sections were incubated in vehicle solution (4% milk, 0.2% triton diluted in PB) for 1 hr at RT.”

(7) The reference Trapani & Korn, 2003 (?) is not included in the list. This reference is important since it sets what are the Kv2.1-CHO cells. In this regard it is also important to mention, even better to address, the expressing qualities of this system in the face of a co-expression with a plasmid-based expression of silent Kv subunits. Are these two ways of expressing Kv subunits, meant to come together (or not) in heteromers, balanced? This question is critical here. Still, in regard to Kv2.1-CHO cells, it was not clear in the manuscript if the term "transfection" refers only to the plasmids used to temporarily induce the expression of silent Kv subunits and potentially Kv channels accessory subunits.

We now include the Trapani & Korn, 2003 reference (thank you for pointing out this accidental omission), and better explain expression methods. The benefit of the inducible Kv2.1 expression is control of Kv conductance densities which can otherwise become so large as to be refractory to voltage clamp. The beauty of the expression system is that cells recently transfected with KvS subunits can be induced to express just enough Kv2.1 to get a substantial but not clampoverwhelming RY785-resistant Kv2/KvS conductance. We also discuss that our expression methods are distinct from past studies. We stop short of comparing the expression systems, as this is beyond the scope of what we set out to study.

Manuscript revisions: See next response

(8) Kv2.1-CHO cells transfection procedures, induction, and validation are unclear. This validation is important here.

We have clarified transfection procedures, induction, and validation in the methods section.

Manuscript revisions:

“The CHO-K1 cell line transfected with a tetracycline-inducible rat Kv2.1 construct (Kv2.1-CHO) (Trapani and Korn, 2003) was cultured as described previously (Tilley et al., 2014).”

Transfections were achieved with Lipofectamine 3000 (Life Technologies, L3000001). 1 μl Lipofectamine was diluted, mixed, and incubated in 25 μl of Opti-MEM (Gibco, 31985062).”

“Concurrently, 0.5 μg of KvS or AMIGO1 or Navβ2, 0.5 μg of pEGFP, 2 μl of P3000 reagent and 25 μl of Opti-MEM were mixed. DNA and Lipofectamine 3000 mixtures were mixed and incubated at room temperature for 15 min. This transfection cocktail was added to 1 ml of culture media in a 24 well cell culture dish containing Kv2.1-CHO cells and incubated at 37 °C in 5% CO2 for 6 h before the media was replaced. Immediately after media was replaced, Kv2.1 expression was induced in Kv2.1-CHO cells with 1 μg/ml minocycline (Enzo Life Sciences, ALX380-109-M050), prepared in 70% ethanol at 2 mg/ml. Voltage clamp recordings were performed 12-24 hours later. We note that the expression method of Kv2/KvS heteromers used here is distinct from previous studies which show that the KvS:Kv2 mRNA ratio can affect the expression of functional Kv2/KvS heteromers (Salinas et al., 1997b; Pisupati et al., 2018). We validated the functional Kv2/KvS heteromer expression using voltage clamp to establish distinct channel kinetics and the presence of RY785-resistant conductance in KvS-transfected cells and using immunohistochemistry to label apparent surface localization of KvS subunits (Figure 4, Figure 1 Supplement, Figure 1 and Figure 5).”

(9) It is important for readers to add some context to Kv2.1/Kv8.1 channels (and other Kv2/KvS heteromers) used to test the combination of RY785 and GxTx. In my opinion, this enriches the discussion.

Good idea. We have added context about each of the KvS subunits we test.

Manuscript revisions:

“To test the pharmacological response of KvS we began with Kv8.1, a subunit that creates heteromers with biophysical properties distinct from Kv2 homomers (Salinas et al., 1997a), and modulates motor neuron vulnerability to cell death (Huang et al., 2024).

Each of these KvS subunits create Kv2/KvS heteromers that have distinct biophysical properties (Kramer et al., 1998; Kerschensteiner and Stocker, 1999; Bocksteins et al., 2012). Kv5.1/Kv2.1 heteromers play an important role in controlling the excitability of mouse urinary bladder smooth muscle (Malysz and Petkov, 2020), mutations in Kv6.4 have been shown to influence human labor pain (Lee et al., 2020b), and deficiency of Kv9.3 disrupts parvalbumin interneuron physiology in mouse prefrontal cortex (Miyamae et al., 2021).”

(10) In general, the membrane potential used to activate Kv2 only channels and Kv2/KvS channels is too close to the activation V1/2. In case the comparing curves are displaced in their relative voltage dependence and voltage sensitivity, using that range of membrane potential may introduce a crucial error in the estimation of the conductance's relative amplitudes.

We now note that the relative conductances of Kv2-only vs Kv2/KvS channels are expected to vary with voltage protocol, as KvS inclusion results in channels with altered voltage responses.

Manuscript revisions:

“…the activation membrane potential is close to the half-maximal point of Kv2/KvS conductances. Thus the ratio of Kv2-like to KvS-like conductance is expected to vary with voltage protocols.”

(11) The use of tail currents to estimate conductance is problematic if i) lack of current inactivation is not assured, and ii) if the different currents, with possible different deactivation kinetics at the used membrane potential (e.g., mV), are not assured. Why was the activation peak used at times, and at different elapsed times the tail currents were used instead? These aspects of conductance's amplitude estimation methods should be well defined.

In CHO cells peak currents were analyzed because outward currents seem to offer the best signal/noise. In neurons, we restricted analysis to tail currents at elapsed times to minimize complications from non-Kv2 endogenous voltage-gated channels which deactivate more quickly. We have clarified this analysis in the methods section.

Manuscript revisions:

“In CHO cells peak currents were analyzed because outward currents seem to offer the best signal/noise. In neurons, we restricted analysis to tail currents at elapsed times to minimize complications from non-Kv2 endogenous voltage-gated channels which deactivate more quickly. In neurons, voltage gated currents remained in the toxin cocktail + RY785 and GxTX, that were sometimes unstable. To minimize complications from these currents, we restricted analysis of RY785 and GxTX subtraction experiments to tail currents at elapsed times to minimize complications from non-Kv2 endogenous voltage-gated channels which deactivate more quickly. We note that the analysis of conductance activation by using tail currents is only accurate when dealing with non-inactivating conductances. We expect that inactivation of Kv2/KvS conductances during the 200 ms pre-pulse is minimal (Salinas et al., 1997a; Kramer et al., 1998; Kerschensteiner and Stocker, 1999) and did not notice inactivation during the activation pulse. Also, deactivation kinetics can vary in a heterogenous population of Kv2/KvS heteromers. While analysis of tail currents could skew the quantification of total Kv2 like and KvS-like conductances, our data supports that mouse nociceptors and human neurons have tail currents that are resistant to RY785 and sensitive to GxTX consistent with the presence of Kv2/KvS heteromers.”

(12) Were the experiments including different conditions such as control, RY, and RY+GxTx done pair-wised? This could potentially better the statistics and strengthen the data and the conclusions drawn from them.

The control, RY, and RY+GxTX in neurons were done pairwise and the statistical tests performed for these experiments were pairwise tests. We have clarified this in the figure legends.

Manuscript revisions:

“Wilcoxon rank tests were paired, except the comparison of RY785 to vehicle which was unpaired.”

(13) The holding potential of the experiments, mostly -89 mV, may be biasing the estimation of Kv2 only channels vs. Kv2/KvS channels conductances. Figure 4I exemplifies this concern.

We agree. Figure 4I reveals that a holding potential of -89 mV vs -129 mV reduces conductance of Kv2.1/Kv8.1 heteromers vs Kv2.1 homomers in CHO cells by ~20%. We have now alerted readers that the ratio of Kv2 only channels vs. Kv2/KvS conductances can vary with holding voltage.

Manuscript revisions:

“Under these conditions, 58 ± 3 % (mean ± SEM) of the delayed rectifier conductance was resistant to RY785 yet sensitive to GxTX (KvS-like) (Fig 7 F). We note that the ratio of KvS- to Kv2-like conductances is expected to vary with holding potential, as KvS subunits can change the degree and voltage-dependence of steady state inactivation (e.g. Fig 4I).”

(14) It is possible that Figure 6A (control trace) and Figure 6C ("Kv2-like" trace) are the same, by mistake, since their noise pattern looks too similar.