Reviewer #5 (Public Review):

Summary:

The manuscript by Walker et al., nicely demonstrated a role of TMEM127 in regulating membrane dynamics of RET, a receptor tyrosine kinase and oncogene for multiple cancers, particularly in pheochromocytoma (PCC). They provided compelling cellular and biochemical evidence on how TMEM127 deficiency reduces RET internalization and degradation in specific cancer cell lines, thus stabilizing cell surface RET and promoting its signaling related to cell proliferation. Moreover, TMEM127 may have a broad function beyond RET, and may affect other surface protein activity such as EGFR etc. These findings provided novel mechanisms and key insights to the field of cancer biology.

Strengths:

Very interesting findings that nicely explained the mechanistic link between TMEM127 and tumorigenesis by RET regulation...the biochemical analysis was quite thorough and impressive.... the general messages delivered by this study are considered as important to the field of TMEM127 biology and tumorigenesis.

Weaknesses:

As acknowledged by the authors, the role of TMEM127 can be conditional and beyond RET modulation, the authors may need to include more discussion that why the loss of TMEM127 is more linked to the development of PCC compared to other cancer types, and why TMEM127 can have such a broad effects in those membrane molecules...in addition, TMEM127 deficiency has been recently linked to enhanced MHC-I-mediated tumor immunity and tumor eradication, in some cancer types...it is then worthwhile to discuss the dual effects of TMEM127 in tumor control in the context of immunity...<br /> Moreover, the authors may need to tune down their "ligand independent" claim on the loss of TMEM127 in driving RET signaling, which should be more related to the level of RET expression/strength of clustering...

eLife assessment

This fundamental study advances our understanding of how Notch signaling activates transcription by analyzing dynamics of the Mastermind transcriptional co-activator and its role in the activation complex. The evidence is compelling and based on state of the art methods with precise quantitative measurements.

Reviewer #2 (Public Review):

The manuscript from deHaro-Arbona et al, entitled "Dynamic modes of Notch transcription hubs conferring memory and stochastic activation revealed by live imaging the co-activator Mastermind", uses single molecule microscopy imaging in live tissues to understand the dynamics and molecular determinants of transcription factor recruitment to the E(spl)-C locus in Drosophila salivary gland cells under Notch-ON and -OFF conditions. Previous studies have identified the major players that are involved in transcription regulation in the Notch pathway, as well as the importance of general transcriptional coregulators, such as CBP/P300 and the Mediator CDK module, but the detailed steps and dynamics involved in these processes are poorly defined. The authors present a wealth of single molecule data that provides significant insights into Notch pathway activation, including:

(1) Activation complexes, containing CSL and Mam, have slower dynamics than the repressor complexes, containing CSL and Hairless.<br /> (2) Contribution of CSL, NICD, and Mam IDRs to recruitment.<br /> (3) CSL-Mam slow-diffusing complexes are recruited and form a hub of high protein concentrations around the target locus in Notch-ON conditions.<br /> (4) Mam recruitment is not dependent on transcription initiation or RNA production.<br /> (5) CBP/P300 or its associated HAT activity is not required for Mam recruitment<br /> (6) Mediator CDK module and CDK8 activity is required for Mam recruitment, and vice-versa, but not CSL recruitment.<br /> (7) Mam is not required for chromatin accessibility but is dependent on CSL and NICD.<br /> (8) CSL recruitment and increased chromatin accessibility persist after NICD removal and loss of Mam, which confers a memory state that enables rapid re-activation in response to subsequent Notch activation<br /> (9) Differences in the proportions of nuclei with both Pol II and with Mam enrichment, which results in transcription being probabilistic/stochastic. These data demonstrate that presence of Mam-complexes is not sufficient to drive all the steps required for transcription in every Notch-ON nucleus.<br /> (10) The switch from more stochastic to robust transcription initiation was elicited when ecdysone was added.

Overall, the manuscript is well written, concise, and clear, and makes significant contributions to the Notch field, which are also important for a general understanding of transcription factor regulation and behavior in the nucleus. The authors have satisfactorily addressed all my criticisms of their initial submission and therefore congratulate the authors on an excellent paper.

Reviewer #3 (Public Review):

Summary:

DeHaro-Arbona and colleagues investigate the in vivo dynamics of Notch-dependent transcriptional activation with a focus on the role of the Mastermind (MAM) transcriptional co-activator. They use GFP and HALO-tagged versions of the CSL DNA-binding protein and MAM to visualize the complex, and Int/ParB to visualize the site of Notch-dependent E(Spl)-C transcription. They make several conclusions. First, MAM accumulates at E(Spl)-C when Notch signaling is active, just like CSL. Second, MAM recruits the CDK module of Mediator but does not initiate chromatin accessibility. Third, after signaling is turned off, MAM leaves the site quickly but CSL and chromatin accessibility are retained. Fourth, RNA pol II recruitment, Mediator recruitment and active transcription were similar and stochastic. Fifth, ecdysone enhance the probability of transcriptional initiation.

Strengths:

The conclusions are well supported by multiple lines of extensive data that is carefully executed and controlled. A major strength is the strategic combination of Drosophila genetics, imaging and quantitative analyses to conduct compelling and easily interpretable experiments. A second major strength is the focus on MAM to gain insights into dynamics of transcriptional activation specifically.

Weaknesses:

Weaknesses were minor. and have been addressed in the revised manuscript.

eLife assessment

This useful study describes expression profiling by scRNA-seq of thousands of cells of recombinant yeast genotypes from a system that models natural genetic variation. The rigorous new method presented here holds promise for improving the efficiency of genotype-to-phenotype mapping in yeast, providing convincing evidence for its efficacy. This manuscript focuses on overcoming technical challenges with this approach. It currently offers somewhat limited new biological insights to place the work within the broader context of the field. Doing so would broaden the interest of the work for all geneticists and evolutionary biologists.

Reviewer #1 (Public Review):

In this paper, N'Guessan et al report a study of expression QTL (eQTL) mapping in yeast using single cells. The authors make use of advances in single-cell RNAseq (scRNAseq) in yeast to increase the efficiency with which this type of analysis can be undertaken. Building on prior research led by the senior author that entailed genotyping and fitness profiling of almost 100,000 cells derived from a cross between two yeast strains (BY and RM) they performed scRNAseq on a subset of 4,489 individual cells. To address the sparsity of genotype data in the expression profiling they used a Hidden Markov Model (HMM) to infer genotypes and then identify the most likely known lineage genotype from the original dataset. To address the relationship between variance in fitness and gene expression the authors partition the variance to investigate the sources of variation. They then perform eQTL mapping and study the relationship between eQTL and fitness QTL identified in the earlier study.

This paper seeks to address the challenging question of how quantitative trait variation and expression variation are related. scRNAseq represents an appealing approach to eQTL mapping as it is possible to simultaneously genotype individual cells and measure expression in the same cell. As eQTL mapping requires large sample sizes to identify statistical relationships, this approach is likely to dramatically increase the statistical power of such studies. However, there are several technical challenges associated with scRNAseq and the authors' study is focused on addressing those challenges. Although the authors present results suggesting the feasibility of the approach there are limitations in the conclusions that can be drawn in the current study owing to the lack of clarity in the presentation of the results. Ultimately, this study presents a proof of concept with limited novel biological insights that would nonetheless make a useful contribution to the literature if the following major points were addressed:

(1) There is insufficient information provided about the nature of data. At a minimum, the following information should be provided to enable assessment of the study: What is the total library size, how many genes are identified per cell, how many UMIs are found per cell, what is the doublet rate, and how are doublets identified (e.g. on the basis of heterozygous calls at polymorphic loci?), how many times is each genotype observed, and how many polymorphic sites are identified per cell that are the basis of genotype inferences?

(2) The prior study analyzed 18 different conditions, whereas this study only assays expression in a single condition. However, the power of the authors' approach is that its efficiency enables testing eQTLs in multiple conditions. The study would be greatly strengthened through analysis of at least one more condition, and ideally several more conditions. The previous fitness study would be a useful guide for choosing additional conditions as identifying those conditions that result in the greatest contrasts in fitness QTL would be best suited to testing the generalizations that can be drawn from the study.

(3) Alternatively, the authors could demonstrate the power of their approach by applying it to a cross between two other yeast strains. As the cross between BY and RM has been exhaustively studied, applying this approach to a different cross would increase the likelihood of making novel biological discoveries.

(4) Figure 1 is misleading as A presents the original study from 2022 without important details such as how genotypes were identified. It is unclear what the barcode is in this study and how it is used in the analysis. Is the barcode for each lineage transcribed so that it is identified in the scRNAseq data? Or, does the barcode in B refer to the cell index barcode? A clearer presentation and explanation of terms are needed to understand the method.

(5) The rationale for the analysis reported in Figure 2B is unclear. The fitness data are from the previous study and the goal is to estimate the heritability using the genotyping data from the scRNAseq data. What is the explanation for why the data don't agree for only one condition, i.e. 37C? And, what are we to understand from the overall result?

(6) Figure 3 presents an analysis of variance partitioning as a Venn diagram. This summarized result is very hard to understand in the absence of any examples of what the underlying raw data look like. For example, what does trait variation look like if only genotype explains the variance or if only gene expression explains the variance? The presented highly summarized data is not intuitive and its presentation is poor - the result that is currently provided would be easier to read in a table format, but the reader needs more information to be able to interpret and understand the result.

(7) I am concerned about the conclusions that can be drawn about expression heritability. The authors claim that expression heritability is correlated with expression levels. It seems likely that this reflects differing statistical power. How can this possibility be excluded?

(8) Conversely, the authors claim that the genes with the lowest heritability are genes involved in the cell cycle. However, uniquely in scRNAseq, cell cycle regulated genes appear to have the highest variance in the data as they are only expressed in a subset of cells. Without incorporating this fact one would erroneously conclude that the variation is not heritable. To test the heritability of cell cycle regulation genes the authors should partition the cells into each cell cycle stage based on expression.

(9) I do not understand Figure S5 and how eQTL sites are assigned to these specific classes given that the authors say that causative variation cannot be resolved because of linkage disequilibrium.

(10) The paragraph starting at line 305 is very confusing. In particular, the authors state that they identify a hotspot of regulation at the mating type locus. It is not obvious why this would be the case. Moreover, they claim that they find evidence for both MATa and MATalpha gene expression. Information is not provided about how segregants were isolated, but assuming that the authors did not dissect 25,000 tetrads to obtain 100,000 segregants I would infer that random spore using SGA was used. In that case, all cells should be MATa. The authors should clarify and explain this observation.

(11) Ultimately, it is not clear what new biological findings the authors have made. There are no novel findings with respect to causative variation underlying eQTLs and I would encourage the authors to make clearer statements in their abstract, introduction, and conclusion about the key discoveries. E.g. What are the "new associations between phenotypic and transcriptomic variations" mentioned in the abstract?

The following minor points should be addressed:

(1) The segregants should be referred to as F2 segregants as they are derived from an F1 cross.

(2) The connections to eQTLs in other organisms should be addressed in the introduction and conclusion. For example, in humans, there has been little evidence for trans eQTLs in contrast to what has been found in yeast.

(3) The authors state that an advantage of scRNAseq over bulk is that it captures rare cell populations (line 79), but this advantage is not exploited in this study.

(4) The authors use ~5% of the lineages from the original study. There is no rationale for why this is an appropriate sample size. Is there an argument for using more cells in eQTL mapping or conversely could the authors ask if fewer cells would provide similar conclusions by downsampling?

(5) I do not agree that the use of UMIs overcomes the challenges of low sequencing depth. UMIs mitigate the possible technical artifacts due to massive PCR amplification.

(6) There is an inadequate reference to prior work on scRNAseq in yeast that established the methods used by the authors and eQTL mapping in human cells using scRNAseq.

(7) The use of empty quotes in Figure 4A is confusing and an alternative presentation method should be used.

(8) The authors speculate about the use of predicted fitness instead of observed fitness, but this is something they could explicitly address in their current study.

Reviewer #2 (Public Review):

Summary:

The experiments and analysis appear to be carefully done. My concerns center on the impact of the work in its current form on the research community.

The focal yeast cross here has been the subject of many previous publications (for smaller sets of recombinant progeny), by the last author and others, including phenotyping, genotyping, transcriptomics, and proteomics. This mini-literature has proven relevant to the community because it has empirically pinpointed exactly how many variants underlie a given trait, both molecular and cellular. That is, whereas in more complex organisms we try our best to estimate/infer the full genetic architecture of varying traits from the results of mapping of necessarily weaker power, the highly-powered yeast system can access a more comprehensive mapping of the dozens of loci impinging on a given trait and learn from it. The question is what exactly we learn from the current study?

Strengths and weaknesses:

Most of the figures center on methods development and validation for the authors' single-cell RNA-seq in the yeast cross, including generating the large raw data set; analysis pipelines for mapping and genotyping (Figure 1); and higher-level analyses that recapitulate previously reported trends in heritability (Figure 2) and eQTL mapping (Figure 3 and Figure 4B-C). One potential novelty of the study is the methods per se: that is, showing that scRNA-seq works for concomitant genotyping and gene expression profiling in the natural variation context. The authors' rigor and effort notwithstanding: in my view, this can be described as modest in terms of principles. That is, the authors did a good job putting the scRNA-seq idea into practice, but their success is perhaps not surprising or highly relevant for work outside of yeast (as the discussion says). The more substantive claim by the authors for the impact of the study is that they make new observations about the role of expression in phenotype (lines 333-335). The major display item of the manuscript on this theme is Figure 4A, reporting which loci that control growth phenotype (from an earlier paper) also control expression. This is solid but I regret to say that the results strike me as modest. The discussion makes some perhaps fairly big claims that the work has helped "bridge understanding of how genetic variation influences transcriptomic variation" and ultimately cellular phenotype. But with the data as they stand, the authors have missed an opportunity to crystallize exactly how a given variant affects expression (perhaps in waves of regulators affecting targets that affect more regulators) and then phenotype, except for the speculations in the text on lines 305-319. The field started down this road years ago with Bayesian causality inference methods applied to eQTL and phenotype mapping (via e.g. the work of Eric Schadt). The authors could now try Mendelian randomization-type fine-grained detailed models for more firepower toward the same end, and/or experimental tests of the genotype-to-expression-to-phenotype relationship. I would see these directions, motivated by fundamental questions that are relevant to the field at large, as leading to a major advance for this very crowded field. As it stands, I felt their absence in this manuscript especially if the authors are selling principles about linking expression and phenotype as their take-home. I also wonder whether the co-mapping of expression and growth traits in Figure 4A would have been possible with e.g. the bulk RNA-seq from Albert et al., 2018, and I recommend that the authors repeat the Figure 4A-type analyses with the latter to justify their statement that their massive scRNA data set would actually be necessary for them to bear fruit (lines 386-388).

I also read the discussion of the manuscript as bringing to the fore some of the challenges a reader has in judging the current state of the results to be of actionable impact. The discussion, and the manuscript, will be improved if the authors can put the work in context, posing concrete questions from the field and stating how they are addressed here and what's left to do.

eLife assessment

This important study presents a new quantitative imaging pipeline that describes with high temporal precision and throughput the movements of late-stage Drosophila embryos, a critical moment when motion first appears. A new approach is used to explore the role of miRNAs in motion onset and presents solid evidence that shows a role for miR-2b-1 and its target Janus in embryonic motion. The data are well supported but do not provide mechanistic insight into the emergence of movement while the writing inflates the importance of the conclusions. The authors must change the name of Janus which is already used in Drosophila genetics.

Reviewer #1 (Public Review):

Summary:

This is an experimentally soundly designed work and a very well-written manuscript. There is a very clear logic that drives the reader from one experiment to the next, the experimental design is clearly explained throughout and the relevance of the acquired data is well analyzed and supports the claims made by the authors. The authors made an evident effort to combine imaging, genetic, and molecular data to describe previously unknown early embryonic movement patterns and to identify regulatory mechanisms that control several aspects of it.

Strengths:

The authors develop a new method to analyze, quantitatively, the onset of movement during the latter embryonic stages of Drosophila development. This setup allows for a high throughput analysis of general movement dynamics based on the capture of variations of light intensity reflected by the embryo. This setup is capable of imaging several embryos simultaneously and provides a detailed measure of movement over time, which proves to be very useful for further discoveries in the manuscript. This setup already provides a thorough and quantifiable description of a process that is little known and identifies two different phases during late embryonic movements: a myogenic phase and a neurogenic phase, which they elegantly prove is dependent on neuronal activity by knocking down action potentials across the nervous system.

However, in this system, movement is detected as a whole, and no further description of the type of movement is provided beyond frequency and amplitude; it would be interesting to know from the authors if a more precise description of the movements that take place at this stage can be achieved with this method (e.g. motion patterns across the A-P body axis).

Importantly, this highly quantitative experimental setup is an excellent system for performing screenings of motion regulators during late embryonic development, and its use could be extended to search for different modulators of the process, beyond miRNAs (genetic mutants, drugs, etc.).

Using their newly established motion detection pipeline, the authors identify miR-2b-1 as required for proper larval and embryonic motion, and identify an overall reduction in the quantity of both myogenic and neurogenic movements, as well as an increased frequency in neurogenic movement "pulses".

Focusing on the neurogenic movement phenotype the authors use in situ probes and perform RT-PCR on FACS-sorted CNS cells to unambiguously detect miR-2b-1 expression in the embryonic nervous system. The neurogenic motion defects observed in miR-2b-1 mutant embryos and early larvae can be completely rescued by the expression of ectopic miR-2b-1 specifically in the nervous system, providing solid evidence of the requirement and sufficiency of miR-2b-1 expressed in the nervous system to regulate these phases of movement.

To explore the mechanism through which miR-2b-1 impacts embryonic movement, the authors use a state-of-the-art bioinformatic approach to identify potential targets of miR-2b-1, and find that the expression levels of an uncharacterized gene, CG3638, are indeed regulated by miR-2b-1. Furthermore, they prove that by knocking down the expression of CG3638 in a miR-2b-1 mutant background, the neurogenic embryonic movement defects are rescued, pointing that the repression of CG3638 by miR-2b-1 is necessary for correct motion patterns in wild-type embryos. Therefore, this paper provides the first functional characterization of CG3638, and names this gene Janus.

Finally, the authors aim to discriminate which elements of the embryonic motor system miR-2b-1/Janus are required. Using directed overexpression of miR-2b-1 and Janus knockdown in the motor neurons and the chordotonal (sensory) organs, they prove that the miR-2b-1/Janus regulatory axis is specifically required in the sensory organs to promote normal embryonic and larval movement.

Weaknesses:

The authors do not describe properly how the miRNA screening was performed and just claim that only miR-2b-1 mutants presented a defective motion phenotype in early L1. How many miRNAs were tested, and how candidates were selected is never explicitly mentioned in the text or the Methods section.

The initial screening to identify miRNAs involved in motion behaviors is performed in early larval movement. The logic presented by the authors is clear - it is assumed that early larval movement cannot proceed normally in the absence of previous embryonic motion - and ultimately helped them identify a miRNA required for modulation of embryonic movement. However, it is possible that certain miRNAs play a role in the modulation of embryonic movement while being dispensable for early L1 behaviors. Such regulators might have been missed with the current screening setup.

Although similar changes to those described for the neurogenic phase of embryonic movement are described for the myogenic phase in miR-2b-1 mutants (reduction in motion amplitude), this phenotype goes unexplored. This is not a big issue, as the authors convincingly demonstrate later that miR-2b-1 is specifically required in the nervous system for proper embryonic and larval movement, and the effects of miR-2b-1 on myogenic movement might as well be the focus of future work. However, it will be interesting to discuss here the implications of a reduced myogenic movement phase, especially as miR-2b-1 is specifically involved in regulating the activity of the chordotonal system - which precisely detects early myogenic movements.

FACS-sorting of neuronal cells followed by RT-PCR convincingly detects the presence of miR-2b-1 in the embryonic CNS. However, control of non-neuronal cells would be required to explore whether miR-2b-1 is not only present but enriched in the nervous system compared to other tissues. This is also the case in the miR-2b-1 and Janus expression analysis in the chordotonal organs: a control sample from the motor neurons would help discriminate whether miR-2b-1/Janus regulatory axis is specifically enriched in chordotonal organs or whether both genes are expressed throughout the CNS but operate under a different regulation or requirements for the movement phenotypes.

Reviewer #2 (Public Review):

Summary:<br /> The manuscript, "A microRNA that controls the emergence of embryonic movement" by Menzies, Chagas, and Alonso provides evidence that Drosophila miR-2b-1 is expressed in neurons and controls the expression of the predicted chloride channel CG3638, here named "Janus". Loss of the miRNA leads to movement phenotypes that can be rescued by downregulation of Janus; using specific drivers, the authors show that a larval movement phenotype (slower movement) can be rescued by knockdown of Janus in the chordotonal organs, suggesting that the increase in Janus found in the chordotonal organs is likely the root of the movement defects. Overall, I found the data presented in the manuscript of reasonable quality and are well enough supported by the presented data. That being said, I do have a few problems with the manuscript, mostly stemming from what I feel is an inflated presentation of the importance of the findings.

Strengths:<br /> The genetic and phenotypic analysis seems to be correct. The nicest part of the manuscript is the connection between the loss of a miRNA and finding its likely target in generating a phenotype. The authors also develop some protocols for the analysis of the movement phenotypes which may be useful for others.

Weaknesses:<br /> As I mentioned above, I felt the presentation was a bit overstated. The authors present their data in a way that focuses on movement, the emergence of movement, and how their miRNA of interest is at the center of this topic. I only point to the title and name that they wish to give the target of their miRNA to emphasize this point. "Janus" the god of movement and change. The results and discussion section starts with a paragraph saying, "Movement is the main output of the nervous system... how developing embryos manage to organise the necessary molecular, cellular, and physiological processes to initiate patterned movement is still unknown. Although it is clear that the genetic system plays a role, how genes control the formation, maturation and function of the cellular networks underlying the emergence of motor control remains poorly understood." While there is nothing inherently untrue about these statements, it is a question of levels of understanding. One can always argue that something in biology is still unknown at a certain level. However, one could also argue that much is known about the molecular nature of movement. Next, I am not sure how much this work impacts the area of study regarding the emergence of movement. The authors show that a reduction of a miRNA can affect something about certain neurons, that affects movement. The early movements, although slightly diminished, still emerge. Thus, their work only suggests that the function of some neurons, or perhaps the development of these neurons may impact the early movements. This is not new as it was known already from early work from the Bate lab.

Later larval movements were also shown to be modified in the miRNA mutants and were traced to "janus" overexpression in the chordotonal organs. As neurons are quite sensitive to the levels of Cl- and Janus is thought to be a Cl- channel, this could lead to a slight dysfunction of the chordotonal neurons. So, based on this, the work suggests that dysfunction of the chordotonal organs could impact larval movement. This was, of course, already known. The novelty of this work is in the genes being studied (important or not). We now know that miR 2b-1 and Janus are expressed in the early neurons and larval chordotonal neurons and their removal is consistent with a role for these genes in the functioning of these neurons. This is not to trivialize these findings, simply to state that these results are not significantly changing our overall understanding of movement and the emergence of movement. I would call it a stretch to say that this miRNA 'controls' the emergence of movement, as in the title.

Finally, the name Janus should be changed as it is already being used. A quick scan of flybase shows that there is a Janus A and B in flies (phosphatases) and I am surprised the authors did not check this. I was initially worried about the Janus kinase (JAK) when I performed the search. While I understand that none are only called Janus, studies of the jan A and B genes refer to the locus as the janus region, which could lead to confusion. The completely different molecular functions of the genes relative to CG3638 add to the confusion. Thus, I ask that the authors change the name of CG3638 to something else.

eLife assessment

This study presents valuable findings on two distinct modes of endosomal fusion and the roles of actin dynamics in this process. The evidence supporting the authors' claims is solid, although the underlying molecular mechanisms and whether the proposed fusion modes are applicable in other cell types remain unclear. The work will be of interest to cell biologists and biophysicists working on the cytoskeleton and organelles.

Reviewer #1 (Public Review):

Summary:

This manuscript employs yolk sac visceral endoderm cells as a novel model for studying endosomal fusion, observing two distinct fusion behaviors: quick homotypic fusion between late endosomes, and slower heterotypic fusion between late endosomes and lysosomes. The mathematical modeling suggests that vesicle size critically influences the mode of fusion. Further investigations reveal that actin filaments are dynamically associated with late endosomal membranes, and are oriented in the x-y plane and along the apical-basal axis. Actin and Arf2/3 were shown to appear at the rear end of the endosomes along the moving direction suggesting polymerization of actin may provide force for the movement of endosomes. Additionally, the authors found that actin dynamics regulate homotypic and heterotypic fusion events in a different manner. The authors also provide evidence to suggest that Cofilin-dependent actin dynamics are involved in late endosome fusion.

Strengths:

The unique feature of this study is that the authors use yolk sac visceral endoderm cells to study endosomal fusion. Yolk sac visceral endoderm cells have huge endocytic vesicles, endosomes, and lysosomes, offering an excellent system to explore endosomal fusion dynamics and the assembly of cellular factors on membranes. The manuscript provides a valuable and convincing observation of the modes of endosomal fusion and the roles of actin dynamics in this process, and the conclusions of the study are justified by the data.

Weaknesses:

While the study offers compelling observations, it falls short of delivering clear mechanistic insights. Key questions remain unaddressed, such as the functional significance of actin filaments that extend apically in positioning late endosomes, the ways in which actin dynamics influence fusion events, and the functional implications of the slower bridge fusion process.

Reviewer #2 (Public Review):

Summary:

Seiichi Koike et al. studied two fusion models, explosive fusion, and bridge fusion, utilizing yolk sac visceral endoderm cells. They elucidated these two fusion models in vivo by employing mathematical modeling and incorporating fluctuations derived from actin dynamics as a key regulator for rapid homotypic fusion between late endosomes.

Strengths:

This study uncovered the role of actin dynamics in regulating the transition of fusion models in homotypic fusion between late endosomes and introduced a method for observing the fusion of single vesicles with two different targets. The role of actin dynamics in vesicle fusion in other systems has been extensively studied. This study could offer useful insights for research on vesicle fusion.

Weaknesses:<br /> The physiological significance of different fusion models is lacking.

eLife assessment

This important study addresses the mechanisms by which mutations in the PURA protein, a regulator of gene transcription and mRNA transport and translation, cause the neurodevelopmental PURA syndrome. Based on convincing evidence from structural biology, molecular dynamics simulation, biochemical, and cell biological analyses, the authors show that the PURA structure is very dynamic, rendering it generally sensitive to structure-altering mutations that affect its folding, DNA-unwinding activity, RNA binding, dimerization, and partitioning into processing bodies. These findings are of substantial importance to cell biology, neurogenetics, and neurology alike, because they provide first insights into how very diverse PURA mutations can cause similar and penetrant molecular, cellular, and clinical defects.

eLife assessment

This study by Sheng and colleagues provides valuable insights into the mechanism of competitive inhibitors of P2X receptors. The structural and functional evidence supporting the subtype specificity of pyridoxal-5'-phosphate derivatives is compelling and provides information for designing drugs that selectively target different subtypes of P2X receptor channels. The work will be of interest to biochemists, structural biologists, and pharmacologists.

Reviewer #1 (Public Review):

This work provides new mechanistic insights into the competitive inhibition in the mammalian P2X7 receptors using structural and functional approaches. The authors solved the structure of panda (pd) P2X7 in the presence of the classical competitive antagonists PPNDS and PPADS. They find that both the drugs bind to the orthosteric site employed by the physiological agonist ATP. However, owing to the presence of a single phosphate group, they prevent movements in the flipper domain required for channel opening. The authors performed structure based mutational analysis together with electrophysiological characterization to understand the subtype specific binding of these drugs. It is known from previous studies that P2X1 and P2X3 are more sensitive to these drugs as compared to P2X7, hence, the residues adjacent to the ATP binding site in pdP2X7 were mutated to those present in P2X1. They observed that mutations of Q143, I214 and Q248 into lysine (hP2X1) increased the P2X7 sensitivity to PPNDS, whereas in P2X1, mutations of these lysines to alanine reduced sensitivity to PPNDS, suggesting that these key residues contribute to the subunit specific sensitivity to these drugs. Similar experiments were done in hP2X3 to demonstrate its higher sensitivity to PPNDS. This preprint provides a useful framework for developing subtype specific drugs for the family of P2X receptor channels, an area that is currently relatively unexplored.<br /> The conclusions of the paper are well supported.

The revised manuscript is well written and presents its data with more clarity.

Reviewer #2 (Public Review):

Summary:

P2X receptors play pivotal roles in physiological processes such as neurotransmission and inflammation, making them promising drug targets. This study, through cryo-EM and functional experiments, reveals the structural basis of the competitive inhibition of the PPNDS and PPADS on mammalian P2X7 receptors. Key findings include the identification of the orthosteric site for these antagonists, the revelation of how PPADS/PPNDS binding impedes channel-activating conformational changes, and the pinpointing of specific residues in P2X1 and P2X3 subtypes that determine their heightened sensitivity to these antagonists. These insights present a comprehensive understanding that could guide the development of improved drugs targeting P2X receptors. This work will be a valuable addition to the field.

Strengths:

The combination of structural experiments and mutagenesis analyses offers a deeper understanding of the mechanism. While the inclusion of MD simulation is appreciated, providing more insights from the simulation might further strengthen this already compelling story.

eLife assessment

This important study used Voltage Sensitive Dye Imaging (VSDI) to measure neural activity in the primary visual cortex of monkeys trained to detect an oriented grating target that was presented either alone or against an oriented mask. The authors show convincingly that the initial effect of the mask ran counter to the behavioral effects of the mask, a pattern that reversed in the latter phase of the response. They interpret these results in terms of influences from the receptive field center, and although an alternative view that emphasizes the role of the receptive field surround also seems reasonable, this study stands as an interesting contribution to our understanding of mechanisms of visual perception.

Reviewer #1 (Public Review):

Summary:

The manuscript describes the crystal structures of Streptococcus pneumoniae NOXs. Crystals were obtained for the wild-type and mutant dehydrogenase domain, as well as for the full-length protein comprising the membrane domain. The manuscript further carefully studies the enzyme's kinetics and substrate-specificity properties. Streptococcus pneumoniae NOX is a non-regulated enzyme, and therefore, its structure should provide a view of the NOX active conformation. The structural and biochemical data are discussed on this ground.

Strengths:

This is very solid work. The protein chemistry and biochemical analysis are well executed and carefully described. Similarly, the crystallography must be appreciated given the difficulty of obtaining good enzyme preparations and the flexibility of the protein. Even if solved at medium resolution, the crystal structure of the full-length protein conveys relevant information. The manuscript nicely shows that the domain rotations are unlikely to be the main mechanistic element of NOX regulation. It rather appears that the NADPH-binding conformation is pivotal to enzyme activation. The paper extensively refers to the previous literature and analyses the structures comprehensively with a comparison to previously reported structures of eukaryotic and prokaryotic NOXs.

Reviewer #2 (Public Review):

The authors describe the structure of the S. pneumoniae Nox protein (SpNOX). This is a first. The relevance of it to the structure and function of eukaryotic Noxes is discussed in depth.

One of the strengths of this work is the effort put into preparing a pure and functionally active SpNOX preparation. The protein was expressed in E. coli and the purification and optimization of its thermostability and activity are described in detail, involving salt concentration, glycerol concentration, and pH.

Comments on revised version:

This reviewer would like to compliment the authors for the conscientious revision of the manuscript. Their response to the comments and the detailed explanations of the issues that did not appear clear enough to the reviewer are much appreciated. Their reaction to the review was not only superbly competent but also prominently good natured.

The revised version is perfect and represents a major contribution to our understanding of the molecular details of Nox function. As for the questions not yet answered, I shall quote the authors: "Time will tell".

eLife assessment

In this manuscript, the authors investigate the properties of prokaryotic NADPH oxidases (NOX) and discuss the implications for NOX regulation and function. The structure of the S. pneumoniae Nox protein is an important step forward in our understanding of procaryotic NOX enzymes and the characterization and interpretation are convincing. The results will be of interest to structural biologists as well as biochemists focusing on enzymatic functions.

eLife assessment

This important study identifies the gene mamo as a new regulator of pigmentation in the silkworm Bombyx mori, a function that was previously unsuspected based on extensive work on Drosophila where the mamo gene is involved in gamete production. The evidence supporting the role of Bm-nano in pigmentation is convincing, including high-resolution linkage mapping of two mutant strains, expression profiling, and reproduction of the mutant phenotypes with state-of-the-art RNAi and CRISPR knock-out assays. The work will be of interest to evolutionary biologists and geneticists studying color patterns and evolution of gene networks.

Author Response

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

eLife assessment

This important study identifies the gene mamo as a new regulator of pigmentation in the silkworm Bombyx mori, a function that was previously unsuspected based on extensive work on Drosophila where the mamo gene is involved in gamete production. The evidence supporting the role of Bm-nano in pigmentation is convincing, including high-resolution linkage mapping of two mutant strains, expression profiling, and reproduction of the mutant phenotypes with state-of-the-art RNAi and CRISPR knock-out assays. While the discussion about genetic changes being guided or accelerated by the environment is extremely speculative and has little relevance for the findings presented, the work will be of interest to evolutionary biologists and geneticists studying color patterns and evolution of gene networks.

Response: Thank you very much for your careful work. In the revised version, we conducted a comparative genomic analysis of the upstream regions of the Bm-mamo gene in 51 wild silkworms and 171 domesticated local silkworms. The analysis of nucleotide diversity (pi) and the fixation index (FSTs) of the Bm-mamo genome sequences in the wild and domesticated silkworm populations were also performed. The results showed that the Bm-mamo genome sequence of local silkworms was relatively conserved, while the upstream sequence of wild silkworms exhibited high nucleotide diversity. This finding suggested a high degree of variability in the regulatory region of the Bm-mamo gene, in wild strains. Additionally, the sequence in this region may have been fixed by domestication selection. We have optimized the description in the discussion section.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This papers performs fine-mapping of the silkworm mutants bd and its fertile allelic version, bdf, narrowing down the causal intervals to a small interval of a handful of genes. In this region, the gene orthologous to mamo is impaired by a large indel, and its function is later confirmed using expression profiling, RNAi, and CRISPR KO. All these experiments are convincingly showing that mamo is necessary for the suppression of melanic pigmentation in the silkworm larval integument.

The authors also use in silico and in vitro assays to probe the potential effector genes that mamo may regulate.

Strengths:

The genotype-to-phenotype workflow, combining forward (mapping) and reverse genetics (RNAi and CRISPR loss-of-function assays) linking mamo to pigmentation are extremely convincing.

This revision is a much improved manuscript and I command the authors for many of their edits.

Response: Thank you very much for your careful work. With the help of reviewers and editors, we have revised the manuscript to improve its readability.

I find the last part of the discussion, starting at "It is generally believed that changes in gene expression patterns are the result of the evolution of CREs", to be confusing.

In this section, I believe the authors sequentially:

• emphasize the role of CRE in morphological evolution (I agree)

• emphasize that TF, and in particular their own CRE, are themselves important mutational targets of evolution (I agree, but the phrasing need to insist the authors are here talking about the CRE found at the TF locus, not the CRE bound by the TF).

• use the stickleback Pel enhancer as an example, which I think is a good case study, but the authors also then make an argument about DNA fragility sites, which is hard to connect with the present study.

• then continue on "DNA fragility" using the peppered moth and butterfly cortex locus. There is no evidence of DNA fragility at these loci, so the connection does not work. "The cortex gene locus is frequently mutated in Lepidoptera", the authors say. But a more accurate picture would be that the cortex locus is repeatedly involved in the generation of color pattern variants. Unlike for Pel fragile enhancer, we don't know if the causal mutations at this locus are repeatedly the same, and the haplotypes that have been described could be collateral rather than causal. Overall, it is important to clarify the idea that mutation bias is a possible factor explaining "genetic hotspots of evolution" (or genetic parallelism sensu 10.1038/nrg3483), but it is also possible that many genetic hotspots are repeated mutational targets because of their "optimal pleiotropy" (e.g. hub position in GRNs, such as mamo might be), or because of particularly modular CRE region that allow fine-tuning. Thus, I find the "fragility" argument misleading here. In fact the finding that "bd" and "bdf" alleles are different in nature is against the idea of a fragility bias (unless the authors can show increased mutation rates at this locus in a wild silkmoth species?). These alleles are also artificially-selected ie. they increased in frequency by breeding rather than natural selection in the wild, so while interesting for our understand of the genotype-phenotype map, they are not necessarily representative of the mutations that may underlie evolution in the wild.

Response: Thank you very much for your careful work. DNA fragility is an interesting topic, but some explanations for DNA fragility are confusing. One study measured the rate of DNA double-strand breaks (DSBs) in yeast artificial chromosomes (YACs), which are chromosomes containing marine Pel that broke ~25 to 50 times more frequently than did the control. These authors believe that the increase in the mutation rate is caused by DNA sequence characteristics, particularly TG-dinucleotide repeats. Moreover, they found that adding a replication origin on the opposite side of Pel did not cause the fungus to switch fragile, making the forward sequence stable and the reverse complement fragile. Thus, Pel fragility is also dependent on the direction of DNA replication. In summary, they suggested that the special DNA sequence is the cause of DNA fragility. In addition, the sequence features associated with DNA fragility in the Pel region are also found in thousands of other positions in the stickleback and human genomes (Xie KT et al, 2019, science).

In yeast artificial chromosomes (YACs), the characteristics of DNA sequences, such as TG-dinucleotide repeat sequences, may be important reasons for DNA fragility, and these breaks occur during DNA replication. However, the inserted sequence of YAC often undergoes deletion or recombination during cultivation and passage. In addition, yeast is a single-celled organism. Therefore, the results in yeast cannot represent the situation in multicellular organisms. If multicellular organisms are like this, there are several issues as follows:

(1) The DNA replication process occurs separately in different multicellular organisms. Because DNA breakage and repair are independent, they can lead to the presence of different alleles in different cells. This can potentially lead to the occurrence of extensive chimeric organisms. However, we have not found such a situation in the genome sequencing of many multicellular organisms.

(2) If the DNA sequence, TG-dinucleotide repeats, is the determining factor, the mutations near the sequence lose their strong correlation with environmental changes. The researchers conducted yeast artificial chromosome experiments in the same environment and found that the frequency of DNA breaks containing TG dinucleotide repeat sequences was 25 to 50 times greater than that of the control group. This means that, whether in the marine population or the lake population, this part of the sticklebacks’ genome has undergone frequent mutations. However, according to related research, populations of lake sticklebacks, rather than marine populations, often exhibit a decrease in the pelvic phenotype.

(3) Researchers have found thousands of loci in the genome of sticklebacks and humans that contain such sequences (TG-dinucleotide repeats). This means that thousands of sites undergo frequent mutations during DNA replication. Unless these sites do not possess functionality, they will have some impact on the organism, even causing damage. Even if they are not functional sequences, these sequences will gradually be discarded or replaced during frequent mutations rather than being present in large quantities in the genome.

Therefore, the study of DNA fragility in yeast cannot explain the situation in multicellular organisms.

As you noted, we want to express that the frequent variation in the cortex gene should be regulated by targeted regulation involving the GRN in Lepidoptera. In addition, studies on specific epigenetic modifications discovered through the referenced fragile DNA sites suggest that DNA fragility is not determined by the DNA sequence (Ji F, 2020, Cell Res) but rather by other factors, such as epigenetic factors. The sequence features discovered at fragile DNA sites are traces of frequent mutations, not causes.

In this revision, we analyzed the nucleotide diversity of the mamo genome in 51 wild and 171 domestic silkworms. We found high nucleic acid diversity from the third exon to the upstream region of this gene in wild silkworms. We randomly selected 12 wild silkworms and 12 domestic silkworms and compared their upstream sequences to approximately 1 kb. In wild silkworms, there is significant diversity in their upstream sequences. In domestic silkworms, the sequences are highly conserved, but in some silkworms, a long interspersed nuclear element (LINE) is inserted. This finding suggested that there is frequent variation in the sequence of this region in wild silkworms, while fixation occurs in domesticated silkworms. These genomic data are sourced from the pangenome of silkworms (Tong X, 2022, Nat Commun.). In the pangenomic research, 1078 strains (205 local strains, 194 improved strains, 632 mutant strains, and 47 wild silkworms), which included 545 third-generation sequencing genomes, were obtained. An online website was built to utilize these data (http://silkmeta.org.cn/). We warmly welcome you to use these data.

In summary, for clearer expression, we have rewritten this section.

Xie KT, Wang G, Thompson AC, Wucherpfennig JI, Reimchen TE, MacColl ADC, Schluter D, Bell MA, Vasquez KM, Kingsley DM. DNA fragility in the parallel evolution of pelvic reduction in stickleback fish. Science. 2019 Jan 4;363(6422):81-84. doi: 10.1126/science.aan1425.

Ji F, Liao H, Pan S, Ouyang L, Jia F, Fu Z, Zhang F, Geng X, Wang X, Li T, Liu S, Syeda MZ, Chen H, Li W, Chen Z, Shen H, Ying S. Genome-wide high-resolution mapping of mitotic DNA synthesis sites and common fragile sites by direct sequencing. Cell Res. 2020 Nov;30(11):1009-1023. doi: 10.1038/s41422-020-0357-y.

Tong X, Han MJ, Lu K, Tai S, Liang S, Liu Y, Hu H, Shen J, Long A, Zhan C, Ding X, Liu S, Gao Q, Zhang B, Zhou L, Tan D, Yuan Y, Guo N, Li YH, Wu Z, Liu L, Li C, Lu Y, Gai T, Zhang Y, Yang R, Qian H, Liu Y, Luo J, Zheng L, Lou J, Peng Y, Zuo W, Song J, He S, Wu S, Zou Y, Zhou L, Cheng L, Tang Y, Cheng G, Yuan L, He W, Xu J, Fu T, Xiao Y, Lei T, Xu A, Yin Y, Wang J, Monteiro A, Westhof E, Lu C, Tian Z, Wang W, Xiang Z, Dai F. High-resolution silkworm pan-genome provides genetic insights into artificial selection and ecological adaptation. Nat Commun. 2022 Sep 24;13(1):5619. doi: 10.1038/s41467-022-33366-x.

Lu K, Pan Y, Shen J, Yang L, Zhan C, Liang S, Tai S, Wan L, Li T, Cheng T, Ma B, Pan G, He N, Lu C, Westhof E, Xiang Z, Han MJ, Tong X, Dai F. SilkMeta: a comprehensive platform for sharing and exploiting pan-genomic and multi-omic silkworm data. Nucleic Acids Res. 2024 Jan 5;52(D1):D1024-D1032. doi: 10.1093/nar/gkad956.

Curiously, the last paragraph ("Some research suggests that common fragile sites...") elaborate on the idea that some sites of the genome are prone to mutation. The connection with mamo and the current article are extremely thin. There is here an attempt to connect meiotic and mitotic breaks to Bm-mamo, but this is confusing: it seems to propose Bm-mamo as a recruiter of epigenetic modulators that may drive higher mutation rates elsewhere. Not only I am not convinced by this argument without actual data, but this would not explain how the mutations at the Bm-mamo itself evolved.

Response: Thank you very much for your careful work. This section mainly illustrates that DNA fragility is not determined by sequence but is regulated by other factors in animals. In fruit flies, they found that mamo is an important candidate gene for recombination hotspot setting in meiosis. First, we evaluated PRDM9, which plays an important role in setting recombination hotspots during meiosis. Our purpose in mentioning this information is to illustrate that chromosome recombination is a process of programmed double strand breaks and to answer another reviewer's question about programmed events in the genome. In summary, we suggest that some variations in DNA sequences are procedural results. We have optimized the description of this section in this version.

On a more positive note, I find it fascinating that the authors identified a TF that clearly articulates or orchestrate larval pattern development, and that when it is deleted, can generate healthy individuals. In other words, while it is a TF with many targets, it is not too pleiotropic. This idea, that the genetically causal modulators of developmental evolution are regulatory genes, has been described elsewhere (e.g. Fig 4c in 10.1038/s41576-020-0234-z, and associated refs). To me, the beautiful findings about Bm-mamo make sense in the general, existing framework that developmental processes and regulatory networks "shape" the evolutionary potential and trajectories of organisms. There is a degree of "programmability" in the genomes, because some loci are particularly prone to modulate a given type of trait. Here, Bm-mamo, as a potentially regulator of both CPs and melanin pathway genes, appear to be a potent modulator of epithelial traits. Claiming that there are inherent mutational biases behind this is unwarranted.

Response: Thank you very much for your careful work. I completely agree with your statement that the genome exhibits a certain degree of programmability. On the one hand, some transcription factors can precisely control the spatiotemporal expression levels of some structural genes (such as pigment synthesis genes). On the other hand, these transcription factors are also subject to strict expression regulation. Because the color pattern is complex, changes in single or minority structural genes result in incomplete or imprecise changes in coloring patterns. Nevertheless, several regulatory factors can regulate multiple downstream target genes. Changes in their expression patterns can lead to holistic and significant changes in color patterns. There are long intergenic regions upstream of many important transcription factors, dozens of kilobase pairs (Kb) to hundreds of Kb, which may contain many different regulatory elements for better control of their expression patterns. Therefore, gene regulatory networks can directly regulate transcription factors to modulate a given type of trait. Transcription factors and their downstream target genes can form a functional module, which is similar to a functional module in software or operating systems. This regulation of transcription factors is simpler in terms of steps, which are similar to a single click switch button. The gene regulatory network regulates these modules in response to environmental changes and is widely recognized.

Some people do not agree that genetic variations can also be regulated. They claim that this is completely random. The infinite monkey theorem (Félix-Édouard-Justin-Émile Borel, 1909) states that if an infinite number of monkeys were given typewriters and an infinite amount of time, they would eventually produce the complete works of Shakespeare. Although this theory advocates randomness on the surface, its conclusions are full of inevitability (tail event). In nature, some things we observe do not have obvious regularity because they involve relatively complex factors, and the underlying logic is obscure and difficult to understand. We often name them random. However, as we gradually understand the logic behind this complex event, we can also recognize the procedural nature of this randomness.

Previously, chromosomal recombination during meiosis was believed to be a random event. However, currently, it is believed that the process is procedural. The occurrence of meiotic recombination mentioned earlier indicates that the genome has the ability to self-set the position of double-strand breaks to form new allelic forms. Because meiotic recombination is programmed, transcription factors that recognize DNA sites, enzymes that cleave double strands, and DNA repair systems exist, programming can also introduce genetic variation. A study in plants has provided insights into this programmed mutation (Monroe JG, 2023, nature). Frequent changes in the expression patterns of some transcription factors occur between and/or within species. In this article, we only discuss the possible reasons for variations in the expression patterns of some transcription factors in a general manner and simple reasoning. We have added an analysis of the response of wild silkworms and improved the relevance of the discussion.

Monroe JG, Srikant T, Carbonell-Bejerano P, Becker C, Lensink M, Exposito-Alonso M, Klein M, Hildebrandt J, Neumann M, Kliebenstein D, Weng ML, Imbert E, Ågren J, Rutter MT, Fenster CB, Weigel D. Mutation bias reflects natural selection in Arabidopsis thaliana. Nature. 2022 Feb;602(7895):101-105. doi: 10.1038/s41586-021-04269-6. Epub 2022 Jan 12. Erratum in: Nature. 2023 Aug;620(7973):

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

• Please structure your Discussion with section headers.

Response: Thank you very much for your careful work. We have added relevant section headers.

• As explained in my public review, I found the two last sections of the Discussion to be dispersed and confusing. I also must say that I carefully read the Response to Reviewers on this, which helped me to better understand the authors' intentions here. Please consider the revision of this Discussion as this feels extremely speculative difficult to connect with Bm-mamo.

Response: Thank you very much for your careful work. We have rewritten this part of the content.

• typo: were found near the TTS of yellow --> TSS

Response: Thank you very much for your careful work. We have made these modifications.

• l. 234 :"expression level of the 18 CP genes in the integument". Consider adding a mention of Figure 7 here, as only Fig. S10 is cited here.

Response: Thank you very much for your careful work. We have made these modifications.

• Editorial comment on the second half of the Abstract:

Wu et al : "We found that Bm-mamo can comprehensively regulate the expression of related pigment synthesis and cuticular protein genes to form color patterns. This indicates that insects have a genetic basis for coordinate regulation of the structure and shape of the cuticle, as well as color patterns. This genetic basis provides the possibility for constructing the complex appearances of some insects. This study provides new insight into the regulation of color patterns."

I respectfully suggest a more accurate rephrasing, where the methods are mentioned, and where the logical argument is more straightforward. For example

"Using RNAi and CRISPR we show that Bm-mamo is a repressor or dark melanin patterns in the larval epithelium. Using in-vitro binding assays and gene expression profiling in wild-type and mutant larvae, we also show that Bm-mamo likely regulate the expression of related pigment synthesis and cuticular protein genes in a coordinated manner to mediate its role in color pattern formation. This mechanism is consistent with a dual role of this transcription factor in regulating both the structure and shape of the cuticle and pigments that are embedded within it. This study provides new insight into the regulation of color patterns as well as in the construction more complex epithelial features in some insects."

I hope this let the ideas of the original version transpire as the authors intended.

Response: Thank you very much for your careful work. We have made these modifications.

Joint Public Review:

This papers performs fine-mapping of the silkworm mutants bd and its fertile allelic version, bdf, narrowing down the causal intervals to a small interval of a handful of genes. In this region, the gene orthologous to mamo is impaired by a large indel, and its function is later confirmed using expression profiling, RNAi, and CRISPR KO. All these experiments are convincingly showing that mamo is necessary for the suppression of melanic pigmentation in the silkworm larval integument.

The authors also use in silico and in vitro assays to probe the potential effector genes that mamo may regulate.

The genotype-to-phenotype workflow, combining forward (mapping) and reverse genetics (RNAi and CRISPR loss-of-function assays) linking mamo to pigmentation are extremely convincing.

Comments on latest version:

This second revision took into account all the reviewers' comments. The authors added an interesting analysis of nucleotide diversity at the Bm-mamo locus, using available sequence data from 51 wild silkworms and 171 domesticated silkworms.<br /> The last paragraph added to the discussion, starting with "It has often been believed that changes in CREs are caused by random mutations", is speculative. There is currently no evidence that the mutation rate is biased at the Bm-mamo locus.

Author Response

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

Public Reviews:

Reviewer #2 (Public Review):

I would like to express my appreciation for the authors' dedication to revising the manuscript. It is evident that they have thoughtfully addressed numerous concerns I previously raised, significantly contributing to the overall improvement of the manuscript.

Response: We appreciate the reviewers’ recognition of our efforts in revising the manuscript.

My primary concern regarding the authors' framing of their findings within the realm of habitual and goal-directed action control persists. I will try explain my point of view and perhaps clarify my concerns. While acknowledging the historical tendency to equate procedural learning with habits, I believe a consensus has gradually emerged among scientists, recognizing a meaningful distinction between habits and skills or procedural learning. I think this distinction is crucial for a comprehensive understanding of human action control. While these constructs share similarities, they should not be used interchangeably. Procedural learning and motor skills can manifest either through intentional and planned actions (i.e., goal-directed) or autonomously and involuntarily (habitual responses).

Response: We would like to clarify that, contrary to the reviewer’s assertion of a scientific consensus on this matter, the discussion surrounding the similarities and differences between habits and skills remains an ongoing and unresolved topic of interest among scientists (Balleine and Dezfouli, 2019; Du and Haith, 2023; Graybiel and Grafton, 2015; Haith and Krakauer, 2018; Hardwick et al., 2019; Kruglanski and Szumowska, 2020; Robbins and Costa, 2017). We absolutely agree with the reviewer that “Procedural learning and motor skills can manifest either through intentional and planned actions (i.e., goal-directed) or autonomously and involuntarily (habitual responses)”. But so do habits. Some researchers also highlight the intentional/goal-directed nature of habits (e.g., Du and Haith, 2023, “Habits are not automatic” (preprint) or Kruglanski and Szumowska, 2020, “Habitual behavior is goal-driven”: “definitions of habits that include goal independence as a foundational attribute of habits are begging the question; they effectively define away, and hence dispose of, the issue of whether habits are goal-driven (p 1258).” Therefore, there is no clear consensus concerning the concept of habit.

While we acknowledge the meaningful distinctions between habits and skills, we also recognize a substantial body of literature supporting the overlap between these concepts (cited in our manuscript), particularly at the neural level. The literature clearly indicates that both habits and skills are mediated by subcortical circuits, with a progressive disengagement of cognitive control hubs in frontal and cingulate cortices as repetition evolves. We do not use these concepts interchangeably. Instead, we simply present evidence supporting the assertion that our trained app sequences meet several criteria for their habitual nature.

Our choice of Balleine and Dezfouli (2018)'s criteria stemmed from the comprehensive nature of their definitions, which effectively synthesized insights from various researchers (Mazar and Wood, 2018; Verplanken et al., 1998; Wood, 2017, etc). Importantly, their list highlights the positive features of habits that were previously overlooked. However, these authors still included a controversial criterion ("habits as insensitive to changes in their relationship to their individual consequences and the value of those consequences"), even though they acknowledged the problems of using outcome devaluation methods and of relying on a null-effect. According to Kruglanski and Szumowska (2020), this criterion is highly problematic as “If, by definition, habits are goalindependent, then any behavior found to be goal-dependent could not be a habit on sheer logical grounds” (p. 1257). In their definition, “habitual behavior is sensitive to the value of the reward (i.e., the goal) it is expected to mediate and is sensitive to the expectancy of goal attainment (i.e., obtainment of the reward via the behavior, p.1265). In fact, some recent analyses of habitual behavior are not using devaluation or revaluation as a criterion (Du and Haith, 2023). This article, for example, ascertains habits using different criteria and provides supporting evidence for trained action sequences being understood as skills, with both goal-directed and habitual components.

In the discussion of our manuscript, we explicitly acknowledge that the app sequences can be considered habitual or goal-directed in nature and that this terminology does not alter the fact that our overtrained sequences exhibit clear habitual features.

Watson et al. (2022) aptly detailed my concerns in the following statements: "Defining habits as fluid and quickly deployed movement sequences overlaps with definitions of skills and procedural learning, which are seen by associative learning theorists as different behaviors and fields of research, distinct from habits."

"...the risk of calling any fluid behavioral repertoire 'habit' is that clarity on what exactly is under investigation and what associative structure underpins the behavior may be lost." I strongly encourage the authors, at the very least, to consider Watson et al.'s (2022) suggestion: "Clearer terminology as to the type of habit under investigation may be required by researchers to ensure that others can assess at a glance what exactly is under investigation (e.g., devaluationinsensitive habits vs. procedural habits)", and to refine their terminology accordingly (to make this distinction clear). I believe adopting clearer terminology in these respects would enhance the positioning of this work within the relevant knowledge landscape and facilitate future investigations in the field.

Response: We would like to highlight that we have indeed followed Watson et al (2022)’s recommendations on focusing on other features/criteria of habits at the expense of the outcome devaluation/contingency degradation paradigm, which has been more controversial in the human literature. Our manuscript clearly aligns with Watson et al. (2022) ‘s recommendations: “there are many other features of habits that are not captured by the key metrics from outcome devaluation/contingency degradation paradigms such as the speed at which actions are performed and the refined and invariant characteristics of movement sequences (Balleine and Dezfouli, 2019). Attempts are being made to develop novel behavioral tasks that tap into these positive features of habits, and this should be encouraged as should be tasks that are not designed to assess whether that behavior is sensitive to outcome devaluation, but capture the definition of habits through other measures”.

Regarding the authors' use of Balleine and Dezfouli's (2018) criteria to frame recorded behavior as habitual, as well as to acknowledgment the study's limitations, it's important to highlight that while the authors labelled the fourth criterion (which they were not fulfilling) as "resistance to devaluation," Balleine and Dezfouli (2018) define it as "insensitive to changes in their relationship to their individual consequences and the value of those consequences." In my understanding, this definition is potentially aligned with the authors' re-evaluation test, namely, it is conceptually adequate for evaluating the fourth criterion (which is the most accepted in the field and probably the one that differentiate habits from skills). Notably, during this test, participants exhibited goaldirected behavior.

The authors characterized this test as possibly assessing arbitration between goal-directed and habitual behavior, stating that participants in both groups "demonstrated the ability to arbitrate between prior automatic actions and new goal-directed ones." In my perspective, there is no justification for calling it a test of arbitration. Notably, the authors inferred that participants were habitual before the test based on some criteria, but then transitioned to goal-directed behavior based on a different criterion. While I agree with the authors' comment that: "Whether the initiation of the trained motor sequences in experiment 3 (arbitration) is underpinned by an action-outcome association (or not) has no bearing on whether those sequences were under stimulus-response control after training (experiment 1)." they implicitly assert a shift from habit to goal-directed behavior without providing evidence that relies on the same probed mechanism. Therefore, I think it would be more cautious to refer to this test as solely an outcome revaluation test. Again, the results of this test, if anything, provide evidence that the fourth criterion was tested but not met, suggesting participants have not become habitual (or at least undermines this option).

Response: In our previously revised manuscript, we duly acknowledged that the conventional (perhaps nowadays considered outdated) goal devaluation criterion was not met, primarily due to constraints in designing the second part of the study. We did cite evidence from another similar study that had used devaluation app-trained action sequences to demonstrate habitual qualities (but the reviewer ignored this).

The reviewer points out that we did use a manipulation of goal revaluation in one of the follow-up tests conducted (although this was not a conventional goal revaluation test inasmuch that it was conducted in a novel context). In this test, please note that we used 2 manipulations: monetary and physical effort. Although we did show that subjects, including OCD patients, were apparently goaldirected in the monetary reward manipulation, this was not so clear when goal re-evaluation involved the physical effort expended. In this effort manipulation, participants were less goaloriented and OCD patients preferred to perform the longer, familiar, to the shorter, novel sequence, thus exhibiting significantly greater habitual tendencies, as compared to controls. Hence, we cannot decisively conclude that the action sequence is goal-directed as the reviewer is arguing. In fact, the evidence is equivocal and may reflect both habitual and goal-directed qualities in the performance of this sequence, consistent with recent interpretations of skilled/habitual sequences (Du and Haith, 2023). Relying solely on this partially met criterion to conclude that the app-trained sequences are goal-directed, and therefore not habitual, would be an inaccurate assessment for several reasons: 1) the action sequences did satisfy all other criteria for being habitual; 2) this approach would rest on a problematic foundation for defining habits, as emphasized by Kruglanski & Szumowska (2020); and 3) it would succumb to the pitfall of subscribing to a zero-sum game perspective, as cautioned by various researchers, including the review by Watson et al. (2022) cited by the referee, thus oversimplifying the nuanced nature of human behavior.

While we have previously complied with the reviewer’s suggestion on relabelling our follow-up test as a “revaluation test” instead of an “arbitration test”, we have now explicitly removed all mentions of the term “arbitration” (which seems to raise concerns) throughout the manuscript. As the reviewer has suggested, we now use a more refined terminology by explicitly referring to the measured behavior as "procedural habits", as he/she suggested. We have also extensively revised the discussion section of our manuscript to incorporate the reviewer’s viewpoint. We hope that these adjustments enhance the clarity and accuracy of our manuscript, addressing the concerns raised during this review process.

In essence, this is an ontological and semantic matter, that does not alter our findings in any way. Whether the sequences are consider habitual or goal directed, does not change our findings that 1) Both groups displayed equivalent procedural learning and automaticity attainment; 2) OCD patients exhibit greater subjective habitual tendencies via self-reported questionnaires; 3) Patients who had elevated compulsivity and habitual self-reported tendencies engaged significantly more with the motor habit-training app, practiced more and reported symptom relief at the end of the study; 4) these particular patients also show an augmented inclination to attribute higher intrinsic value to familiar actions, a possible mechanism underlying compulsions.

Reviewer #2 (Recommendations For The Authors):

A few more small comments (with reference to the point numbers indicated in the rebuttal):

(14) I am not entirely sure why the suggested analysis is deemed impractical (i.e., why it cannot be performed by "pretending" participants received the points they should have received according to their performance). This can further support (or undermine) the idea of effect of reward on performance rather than just performance on performance.

Response: We have now conducted this analysis, generating scores for each trial of practices after day 20, when participants no longer gained points for their performance. This analysis assesses whether participants trial-wise behavioral changes exhibit a similar pattern following simulated relative increases or decrease in scores, as if they had been receiving points at this stage. Note that this analysis has fewer trials available, around 50% less on average.

Before presenting our results, we wish to emphasize the importance of distinguishing between the effects of performance on performance and the effects of reward on performance. In response to a reviewer's suggestion, we assessed the former in the first revision of our manuscript. We normalized the movement time variable and evaluated how normalized behavioral changes responded to score increments and decrements. The results from the original analyses were consistent with those from the normalized data.

Regarding the phase where participants no longer received scores, we believe this phase primarily helps us understand the impact of 'predicted' or 'learned' rewards on performance. Once participants have learned the simple association between faster performance and larger scores, they can be expected to continue exhibiting the reward sensitivity effects described in our main analysis. We consider it is not feasible to assess the effects of performance on performance during the reward removal phase, which occurs after 20 days. Therefore, the following results pertain to how the learned associations between faster movement times and scores persist in influencing behavior, even when explicit scores are no longer displayed on the screen.

Results: The main results of the effect of reward on behavioral changes persist, supporting that relative increases or decreases in scores (real or imagined/inferred) modulate behavioral adaptations trial-by-trial in a consistent manner across both cohorts. The direction of the effects of reward is the same as in the main analyses presented in the manuscript: larger mean behavioral changes (smaller std) following ∆R- . First, concerning changes in “normalized” movement time (MT) trial-by-trial, we conducted a 2 x 2 factorial analysis of the centroid of the Gaussian distributions with the same factors Reward, Group and Bin. This analysis demonstrated a significant main effect of Reward (P = 2e-16), but not of Group (P = 0.974) or Bin (P = 0.281). There were no significant interactions between factors. The main Reward effect can be observed in the top panel of the figure below. The same analysis applied to the spread (std) of the Gaussian distributions revealed a significant main effect of Reward (P = 0.000213), with no additional main effects or interactions.

Author response image 1.

Next, conducting the same 2 x 2 factorial analyses on the centroid and spread of the Gaussian distributions fitted to the Consistency data, we also obtained a robust significant main effect of Reward. For the centroid variable, we obtained a significant main effect of Reward (P = 0.0109) and Group (P = 0.0294), while Bin and the factor interactions were non-significant. See the top panel of the figure below.

On the other hand, Reward also modulated significantly the spread of the Gaussian distributions fitted to the Consistency data, P = 0.00498. There were no additional significant main effects or interactions. See the bottom panel in the figure below.

Note that here the factorial analysis was performed on the logarithmic transformation of the std.

Author response image 2.

(16) I find this result interesting and I think it might be worthwhile to include it in the paper.

Response: We have now included this result in our revised manuscript (page 28)

(18) I referred to this sentence: "The app preferred sequence was their preferred putative habitual sequence while the 'any 6' or 'any 3'-move sequences were the goal-seeking sequences." In my understanding, this implies one choice is habitual and another indicates goal-directedness.

One last small comment:
In the Discussion it is stated: "Moreover, when faced with a choice between the familiar and a new, less effort-demanding sequence, the OCD group leaned toward the former, likely due to its inherent value. These insights align with the theory of goal-direction/habit imbalance in OCD (Gillan et al., 2016), underscoring the dominance of habits in particular settings where they might hold intrinsic value."

This could equally be interpreted as goal-directed behavior, so I do not think there is conclusive support for this claim.

Response: The choice of the familiar/trained sequence, as opposed to the 'any 6' or 'any 3'-move sequences cannot be explicitly considered goal-directed: firstly, because the app familiar sequences were associated with less monetary reward (in the any-6 condition), and secondly, because participants would clearly need more effort and time to perform them. Even though these were automatic, it would still be much easier and faster to simply tap one finger sequentially 6 times (any6) or 3 times (any-3). Therefore, the choice for the app-sequence would not be optimal/goaldirected. In this sense, that choice aligns with the current theory of goal-direction/habit imbalance of OCD. We found that OCD patients prefer to perform the trained app sequences in the physical effort manipulation (any-3 condition). While this, on one hand cannot be explicitly considered a goal-directed choice, we agree that there is another possible goal involved here, which links to the intrinsic value associated to the familiar sequence. In this sense the action could potentially be considered goal-directed. This highlights the difficulty of this concept of value and agrees with: 1) Hommel and Wiers (2017): “Human behavior is commonly not driven by one but by many overlapping motives . . . and actions are commonly embedded into larger-scale activities with multiple goals defined at different levels. As a consequence, even successful satiation of one goal or motive is unlikely to also eliminate all the others(p. 942) and 2) Kruglanski & Szumowska (2020)’s account that “habits that may be unwanted from the perspective of an outsider and hence “irrational” or purposeless, may be highly wanted from the perspective of the individual for whom a habit is functional in achieving some goal” (p. 1262) and therefore habits are goal-driven.

References:

Balleine BW, Dezfouli A. 2019. Hierarchical Action Control: Adaptive Collaboration Between Actions and Habits. Front Psychol 10:2735. doi:10.3389/fpsyg.2019.02735

Du Y, Haith A. 2023. Habits are not automatic. doi:10.31234/osf.io/gncsf Graybiel AM, Grafton ST. 2015. The Striatum: Where Skills and Habits Meet. Cold Spring Harb Perspect Biol 7:a021691. doi:10.1101/cshperspect.a021691

Haith AM, Krakauer JW. 2018. The multiple effects of practice: skill, habit and reduced cognitive load. Current Opinion in Behavioral Sciences 20:196–201. doi:10.1016/j.cobeha.2018.01.015

Hardwick RM, Forrence AD, Krakauer JW, Haith AM. 2019. Time-dependent competition between goal-directed and habitual response preparation. Nat Hum Behav 1–11. doi:10.1038/s41562019-0725-0

Hommel B, Wiers RW. 2017. Towards a Unitary Approach to Human Action Control. Trends Cogn Sci 21:940–949. doi:10.1016/j.tics.2017.09.009

Kruglanski AW, Szumowska E. 2020. Habitual Behavior Is Goal-Driven. Perspect Psychol Sci 15:1256– 1271. doi:10.1177/1745691620917676

Mazar A, Wood W. 2018. Defining Habit in Psychology In: Verplanken B, editor. The Psychology of Habit: Theory, Mechanisms, Change, and Contexts. Cham: Springer International Publishing. pp. 13–29. doi:10.1007/978-3-319-97529-0_2

Robbins TW, Costa RM. 2017. Habits. Current Biology 27:R1200–R1206. doi:10.1016/j.cub.2017.09.060

Verplanken B, Aarts H, van Knippenberg A, Moonen A. 1998. Habit versus planned behaviour: a field experiment. Br J Soc Psychol 37 ( Pt 1):111–128. doi:10.1111/j.2044-8309.1998.tb01160.x

Watson P, O’Callaghan C, Perkes I, Bradfield L, Turner K. 2022. Making habits measurable beyond what they are not: A focus on associative dual-process models. Neurosci Biobehav Rev 142:104869. doi:10.1016/j.neubiorev.2022.104869

Wood W. 2017. Habit in Personality and Social Psychology. Pers Soc Psychol Rev 21:389–403. doi:10.1177/1088868317720362

eLife assessment

This study provides solid evidence for differences in habit-learning in obsessive-compulsive disorder versus controls. Contrary to previous studies that employed a single laboratory session to study habit-learning, here a smartphone app delivered motor-sequence tasks daily for a month. These results have important implications for our understanding of goal-directed versus habit learning in obsessive-compulsive disorder.

Reviewer #1 (Public Review):

It is known that aberrant habit formation is a characteristic of obsessive-compulsive disorder (OCD). Habits can be defined according to the following features (Balleine and Dezfouli, 2019): rapid execution, invariant response topography, action 'chunking' and resistance to devaluation.

The extent to which OCD behavior is derived from enhanced habit formation relative to deficits in goal-directed behavior is a topic of debate in the current literature. This study examined habit-learning specifically (cf. deficits in goal-directed behavior) by regularly presenting, via smartphone, sequential learning tasks to patients with OCD and healthy controls. Participants engaged in the tasks every day over the course of a month. Automaticity, including the extent to which individual actions in the sequence become part of a unified 'chunk', was an important outcome variable. Following the 30 days of training, in-laboratory tasks were then administered to examine 1) if performing the learned sequences themselves had become rewarding 2) differences in goal-directed vs. habitual behavior.

Several hypotheses were tested, including:<br /> Patients would have impaired procedural learning vs. healthy volunteers (this was not supported, possibly because there were fewer demands on memory in the task used here)<br /> Once the task had been learned, patients would display automaticity faster (unexpectedly, patients were slower to display automaticity)<br /> Habits would form faster under a continuous (vs. variable) reinforcement schedule

Exploratory analyses were also conducted: an interesting finding was that OCD patients with higher self-reported symptoms voluntarily completed more sessions with the habit-training app and reported a reduction in symptoms.

Strengths

This paper is well situated theoretically within the habit learning/OCD literature.<br /> Daily training in a motor-learning task, delivered via smartphone, was innovative, ecologically valid and more likely to assay habitual behaviors specifically. Daily training is also more similar to studies with non-humans, making a better link with that literature. The use of a sequential-learning task (cf. tasks that require a single response) is also more ecologically valid.<br /> The in-laboratory tests (after the 1 month of training) allowed the researchers to test if the OCD group preferred familiar, but more difficult, sequences over newer, simpler sequences.

Weaknesses

The authors were not able to test one criterion of habits, namely resistance to devaluation, due to the nature of the task.<br /> The sample size was relatively small. Some potentially interesting individual differences within the OCD group could have been examined more thoroughly with a bigger sample (e.g., preference for familiar sequences). A larger sample may have allowed the statistical testing of any effects due to medication status.

The authors achieved their aims in that two groups of participants (patients with OCD and controls) engaged with the task over the course of 30 days. The repeated nature of the task meant that 'overtraining' was almost certainly established, and automaticity was demonstrated. This allowed the authors to test their hypotheses about habit learning. The results are supportive of the author's conclusions.

This article is likely to be impactful -- the delivery of a task across 30 days to a patient group is innovative and represents a new approach for the study of habit learning that is superior to an in-laboratory approach.

An interesting aspect of this manuscript is that it prompts a comparison with previous studies of goal-directed/habitual responding in OCD that used devaluation protocols, and which may have had their effects due to deficits in goal-directed behavior and not enhanced habit learning per se.

Author Response

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

eLife assessment

This manuscript represents a cleanly designed experiment for assessing biological motion processing in children (mean age = 9) with and without ADHD. The group differences concerning accuracy in global and local motion processing abilities are solid, but the analyses suggesting dissociable relationships between global and local processing and social skills, age, and IQ need further interrogation. The results are useful in terms of understanding ADHD and the ontogenesis of different components of the processing of biological motion.

We thank the editors for the positive assessment of our manuscript. We have carefully considered the reviewers’ constructive and helpful comments and revised our manuscript accordingly. To address the question about the dissociable relationships between global and local BM processing, we have provided more evidence and additional analyses in this revised version.

Reviewer #1 (Public Review):

Summary:

The paper presents a nice study investigating differences in biological motion perception in participants with ADHD in comparison with controls. Motivated by the idea that there is a relationship between biological motion perception and social capabilities, the authors investigated local and global (holistic) biological motion perception, the group, and several additional behavioral variables that are affected in ADHS (IQ, social responsiveness, and attention/impulsivity). As well as local global biological motion perception is reduced in ADHD participants. In addition, the study demonstrates a significant correlation between local biological motion perception skills and the social responsiveness score in the ADHD group, but not the controls. A path analysis in the ADHD data suggests that general performance in biological motion perception is influenced mainly by global biological motion perception performance and attentional and perceptual reasoning skills.

Strengths:

It is true that there exists not much work on biological motion perception and ADHD. Therefore, the presented study contributes an interesting new result to the biological motion literature and adds potentially also new behavioral markers for this clinical condition. The design of the study is straightforward and technically sound, and the drawn conclusions are supported by the presented results.

Thank you for your positive assessment of our work.

Weaknesses:

Some of the claims about the relationship between genetic factors and ADHD and the components of biological motion processing have to remain speculative at this point because genetic influences were not explicitly tested in this paper.

We agree that the relationship between genetic factors and BM processing in ADHD needs more investigation, We have modified our statement in Discussion section as following:

“Using the classical twin method, Wang et al. found that the distinction between local and global BM processing may stem from the dissociated genetic bases. The former, to a great degree, seems to be acquired phylogenetically20,21,59,60, while the latter is primarily obtained through individual development19.” (lines 421 - 425),

Reviewer #2 (Public Review):

Summary:

Tian et al. aimed to assess differences in biological motion (BM) perception between children with and without ADHD, as well as relationships to indices of social functioning and possible predictors of BM perception (including demographics, reasoning ability and inattention). In their study, children with ADHD showed poorer performance relative to typically developing children in three tasks measuring local, global, and general BM perception. The authors further observed that across the whole sample, performance in all three BM tasks was negatively correlated with scores on the social responsiveness scale (SRS), whereas within groups a significant relationship to SRS scores was only observed in the ADHD group and for the local BM task. Local and global BM perception showed a dissociation in that global BM processing was predicted by age, while local BM perception was not. Finally, general (local & global combined) BM processing was predicted by age and global BM processing, while reasoning ability mediated the effect of inattention on BM processing.

Strengths:

Overall, the manuscript is presented in a relatively clear fashion and methods and materials are presented with sufficient detail so the study could be reproduced by independent researchers. The study uses an innovative, albeit not novel, paradigm to investigate two independent processes underlying BM perception. The results are novel and have the potential to have wide-reaching impact on multiple fields.

We appreciate your positive assessment of our work.

Weaknesses:

Except for the main analysis, it is unclear what the authors' specific predictions are regarding the three different tasks they employ. The three BM tasks are used to probe different processes underlying BM perception, but it is difficult to gather from the introduction why these three specific tasks were chosen and what predictions the authors have about the performance of the ADHD group in these tasks. Relatedly, the authors do not report whether (and if so, how) they corrected for multiple comparisons in their analyses. As the number of tests one should control for depends on the theoretical predictions (http://daniellakens.blogspot.com/2016/02/why-you-dont-need-to-adjust-you-alpha.html), both are necessary for the reader to assess the statistical validity of the results and any inferences drawn from them. The same is the case for the secondary analyses exploring relationships between the 3 individual BM tasks and social function measured by the social responsivity scale (SRS).

We appreciate these constructive suggestions. In response, we have included a detailed description in the Introduction section explaining why we employed three different tasks and our predictions about the performance in ADHD:

“Despite initial indications, a comprehensive investigation into BM perception in ADHD is warranted. We proposed that it is essential to deconstruct BM processing into its multiple components and motion features, since treating them as a single entity may lead to misleading or inconsistent findings31. To address this issue, we employed a carefully designed behavioral paradigm used in our previous study19, making slight adjustments to adapt for children. This paradigm comprises three tasks. Task 1 (BM-local) aimed to assess the ability to process local BM cues. Scrambled BM sequences were displayed and participants could use local BM cues to judge the facing direction of the scrambled walker. Task 2 (BM-global) tested the ability to process the global configuration cues of the BM walker. Local cues were uninformative, and participants used global BM cues to determine the presence of an intact walker. Task 3 (BM-general) tested the ability to process general BM cues (local + global cues). The stimulus sequences consisted of an intact walker and a mask containing similar target local cues, so participants could use general BM cues (local + global cues) to judge the facing direction of the walker.” (lines 116 - 130)

“In Experiment 1, we examined three specific BM perception abilities in children with ADHD. As mentioned earlier, children with ADHD also show impaired social interaction, which implies atypical social cognition. Therefore, we speculated that children with ADHD performed worse in the three tasks compared to TD children.” (lines 131 - 134)

Additionally, we have reported the p values corrected for multiple comparisons (false discovery rate, FDR) in the revised manuscript wherever it was necessary to adjust the alpha (lines 310 - 316; Table 2). The pattern of the results remained unchanged.

In relation to my prior point, the authors could provide more clarity on how the conclusions drawn from the results relate to their predictions. For example, it is unclear what specific conclusions the authors draw based on their findings that ADHD show performance differences in all three BM perception tasks, but only local BM is related to social function within this group. Here, the claim is made that their results support a specific hypothesis, but it is unclear to me what hypothesis they are actually referring to (see line 343 & following). This lack of clarity is aggravated by the fact that throughout the rest of the discussion, in particular when discussing other findings to support their own conclusions, the authors often make no distinction between the two processes of interest. Lastly, some of the authors' conclusions related to their findings on local vs global BM processing are not logically following from the evidence: For instance, the authors conclude that their data supports the idea that social atypicalities are likely to reduce with age in ADHD individuals. However, according to their own account, local BM perception - the only measure that was related to social function in their study - is understood to be age invariant (and was indeed not predicted by age in the present study).

Thank you for pointing out this issue. We have carefully revised the Discussion section about our findings to clarify these points:

“Our study contributes several promising findings concerning atypical biological motion perception in ADHD. Specifically, we observe the atypical local and global BM perception in children with ADHD. Notably, a potential dissociation between the processing of local and global BM information is identified. The ability to process local BM cues appears to be linked to the traits of social interaction among children with ADHD. In contrast, global BM processing exhibits an age-related development. Additionally, general BM perception may be affected by factors including attention.” (lines 387 - 393)

We have provided a detailed discussion on the two processes of interest to clarify their potential differences and the possible reasons behind the difference of the divergent developmental trajectories between local and global BM processing:

“BM perception is considered a multi-level phenomenon56-58. At least in part, processing information of local BM and global BM appears to involve different genetic and neural mechanisms16,19. Using the classical twin method, Wang et al. found that the distinction between local and global BM processing may stem from the dissociated genetic bases. The former, to a great degree, seems to be acquired phylogenetically20,21,59,60, while the latter is primarily obtained through individual development19. The sensitivity to local rather than global BM cues seems to emerge early in life. Visually inexperienced chicks exhibit a spontaneous preference for the BM stimuli of hen, even when the configuration was scrambled20. The same finding was reported in newborns. On the contrary, the ability to process global BM cues rather than local BM cues may be influenced by attention28,29 and shaped by experience24,56.” (lines 419 - 430)

“We found that the ability to process global and general BM cues improved significantly with age in both TD and ADHD groups, which imply the processing module for global BM cues tends to be mature with development. In the ADHD group, the improvement in processing general and global BM cues is greater than that in processing local BM cues, while no difference was found in TD group. This may be due to the relatively higher baseline abilities of BM perception in TD children, resulting in a relatively milder improvement. These findings also suggest a dissociation between the development of local and global BM processing. There seems to be an acquisition of ability to process global BM cues, akin to the potential age-related improvements observed in certain aspects of social cognition deficits among individuals with ADHD5, whereas local BM may be considered an intrinsic trait19.” (lines 438 -449)

In addition, we have rephased some inaccurate statements in revised manuscript. Another part of social dysfunction might be stable and due to the atypical local BM perception in ADHD individuals, although some studies found a part of social dysfunction would reduce with age in ADHD individuals. One reason is that some factors related to social dysfunction would improve with age, like the symptom of hyperactivity.

Results reported are incomplete, making it hard for the reader to comprehensively interpret the findings and assess whether the conclusions drawn are valid. Whenever the authors report negative results (p-values > 0.05), the relevant statistics are not reported, and the data not plotted. In addition, summary statistics (group means) are missing for the main analysis.

Thanks for your comments. We have provided the complete statistical results in the revised manuscript (lines 309 - 316) and supplementary material, which encompass relevant statistics and plots of negative results (Figure 4, Figure S2 and S3), in accordance with our research questions. And we have also included summary statistics in the Results section (lines 287 - 293).

Some of the conclusions/statements in the article are too strong and should be rephrased to indicate hypotheses and speculations rather than facts. For example, in lines 97-99 the authors state that the finding of poor BM performance in TD children in a prior study 'indicated inferior applicability' or 'inapplicable experimental design'. While this is one possibility, a perhaps more plausible interpretation could be that TD children show 'poor' performance due to outstanding maturation of the underlying (global) BM processes (as the authors suggest themselves that BM perception can improve with age). There are several other examples where statements are too strong or misleading, which need attention.

We thank you for pointing out the issue. We have toned down and rephrased the strong statements and made the necessary revisions.

“Another study found that children with ADHD performed worse in BM detection with moderate ratios of noise34. This may be due to the fact that BM stimuli with noise dots will increase the difficulty of identification, which highlights the difference in processing BM between the two groups33,35.” (lines 111 - 115)

Reviewer #3 (Public Review):

Summary:

The authors presented point light displays of human walkers to children (mean = 9 years) with and without ADHD to compare their biological motion perception abilities and relate them to IQ, social responsiveness scale (SRS) scores and age. They report that children with ADHD were worse at all three biological motion tasks, but that those loading more heavily on local processing related to social interaction skills and global processing to age. The important and solid findings are informative for understanding this complex condition, as well as biological motion processing mechanisms in general. However, I am unsure that these differences between local and global skills are truly supported by the data and suggest some further analyses.

Strengths:

The authors present clear differences between the ADHD and TD children in biological motion processing, and this question has not received as much attention as equivalent processing capabilities in autism. They use a task that appears well controlled. They raise some interesting mechanistic possibilities for differences in local and global motion processing, which are distinctions worth exploring. The group differences will therefore be of interest to those studying ADHD, as well as other developmental conditions, and those examining biological motion processing mechanisms in general.

We appreciate your positive feedback. In revised manuscript, we have added more analyses to support the differences between local and global motion processing. Please refer to our response to the point #3 you mentioned below.

Weaknesses:

I am unsure that the data are strong enough to support claims about differences between global and local processing wrt social communication skills and age. The mechanistic possibilities for why these abilities may dissociate in such a way are interesting, but do not seem so plausible to me. I am also concerned about gender, and possible autism, confounds when examining the effect of ADHD. Specifics:

Gender confound. There are proportionally more boys in the ADHD than TD group. The authors appear to attempt to overcome this issue by including gender as a covariate. I am unsure if this addresses the problem. The vast majority of participants in the ADHD group are male, and gender is categorically, not continuously, defined. I'm pretty sure this violates the assumptions of ANCOVA.

We appreciate your comments. We concur with you that although we observed a clear difference between local and global BM processing in ADHD, the evidence is to some extent preliminary. The mechanistic possibilities for why these abilities may dissociate have been discussed in revised manuscript. Please refer to the response to reviewer 2’s point #2. To further examine if gender played a role in the observed results, we used a statistical matching technique to obtain a sub-dataset. The pattern of results remained with the more balanced dataset (see Supplementary Information part 1). According to your suggestion, we have also presented the results without using gender as a covariate in main text and also separated the data of boys and girls on the plots (see Figure 1 and Figure S1). There were indeed no signs of a gender effect.

Autism. Autism and ADHD are highly comorbid. The authors state that the TD children did not have an autism or ADHD diagnosis, but they do not state that the ADHD children did not have an autism diagnosis. Given the nature of the claims, this seems crucial information for the reader.

Thanks for your suggestion. We have confirmed that all children with ADHD in our study were not diagnosed with autism. We used a semi-structured interview instrument (K-SADSPL-C) to confirm every recruited child with ADHD but not with ASD. The exclusion criteria for both groups were mentioned in the Materials and methods section:

“Exclusion criteria for both groups were: (a) neurological diseases; (b) other neurodevelopmental disorders (e.g., ASD, Mental retardation, and tic disorders), affective disorders and schizophrenia…” (lines 158 - 162)

Conclusions. The authors state frequently that it was the local BM task that related to social communication skills (SRS) and not the global tasks. However, the results section shows a correlation between SRS and all three tasks. The only difference is that when looking specifically within the ADHD group, the correlation is only significant for the local task. I think that if the authors wish to make strong claims here they must show inferential stats supporting (1) a difference between ADHD and TD SRS-Task 1 correlations, and (2) a difference in those differences for Task 2 and 3 relative to Task 1. I think they should also show a scatterplot of this correlation, with separate lines of best fit for the two groups, for Tasks 2 and 3 as well. I.e. Figure 4 should have 3 panels. I would recommend the same type of approach for age. Currently, they have small samples for correlations, and are reading much of theoretical significance between some correlations passing significance threshold and others not. It would be incredibly interesting if the social skills (as measured by SRS) only relate to local BM abilities, and age only to global, but I think the data are not so clear with the current information. I would be surprised if all BM abilities did not improve with age. Even if there is some genetic starter kit (and that this differs according to particular BM component), most abilities improve with learning/experience/age.

Thank you for this recommendation. We have added more statistics to test differences between the correlations (a difference between ADHD and TD in SRS-Task 1 correlations (see the first paragraph of Supplementary Information part 2), a difference in SRS-response accuracy correlations for Task 2 and 3 relative to Task 1(see the second paragraph of Supplementary Information part 2), and a difference in age-response accuracy correlations for Task 2 and 3 relative to Task 1 in ADHD group (see Supplementary Information part 3)). Additionally, we have included scatterplots for SRS-Task1, SRS-Task2, SRS-Task3 (with separate lines of best fit for the two groups in each, see Figure 4), SRS-ADHD, SRS-TD, age-ADHD and age-TD (with separate lines of best fit for the three tasks in each, see Figure S2 and S3) to make a clear demonstration. Detailed results have been presented in the revised manuscript and Supplementary Information. We expect these further analyses would strengthen our conclusions.

Theoretical assumptions. The authors make some sweeping statements about local vs global biological motion processing that need to be toned down. They assume that local processing is specifically genetically whereas global processing is a product of experience. The fact their global, but not local, task performance improves with age would tend to suggest there could be some difference here, but the existing literature does not allow for this certainty. The chick studies showing a neonatal preference are controversial and confounded - I cannot remember the specifics but I think there an upper vs lower visual field complexity difference here.

Thank you for pointing out this issue. We have toned down rephrased our claims that the difference between local and global BM processing according to your suggestion:

“These findings suggest that local and global mechanisms might play different roles in BM perception, though the exact mechanisms underlying the distinction remain unclear. Exploring the two components of BM perception will enhance our understanding of the difference between local and global BM processing, shedding light on the psychological processes involved in atypical BM perception.” (lines 87 - 92)

Reviewer #1 (Recommendations For The Authors):

I have only a number of minor points that should be addressed prior to publication:

L. 95ff: What is meant by 'inapplicability of experimental designs' ? This paragraph is somewhat unclear.

In revised manuscript, we have clarified this point (lines 111 - 115).

L. 146: The groups were not perfectly balanced for sex. Would results change fundamentally in a more balanced design, or can arguments be given that gender does not play a role, like it seems to be the case for some functions in biological motion perception (e.g. Pavlova et al. 2015; Tsang et al 2018). One could provide a justification that this disbalance does not matter or test for subsampled balanced data sets maybe.

This point is similar to the point #1 from reviewer 3, and we have addressed this issue in our response above.

L. 216 f.: In this paragraph it does not become very clear that the mask for the global task consisted of scrambles generated from walkers walking in the same direction. The mask for the local task then should consist of a balanced mask that contains the same amount of local motion cues indicating right and leftwards motion. Was this the case? (Not so clear from this paragraph.)

Regarding the local task, the introduction of mask would make the task too difficult for children. Therefore, in the local task, we only displayed a scrambled walker without a mask, which was more suitable for children to complete the task. We have made clear this point in the corresponding paragraph (lines 232 - 241).

L. 224 ff.: Here it would be helpful to see the 5 different 'facing' directions of the walkers. What does this exactly mean? Do they move on oblique paths that are not exactly orthogonal to the viewing directions, and how much did these facing directions differ?

Out of the five walkers we used, two faced straight left or right, orthogonal to the viewing directions. Two walked with their bodies oriented 45 degrees from the observer, to the left or right. The last one walked towards the observer. We have included a video (Video 4) to demonstrate the 5 facing directions.

L. 232: How was the number of 5 practicing trials determined/justified?

As mentioned in main text, global BM processing is susceptible to learning. Therefore, too many practicing trials would increase BM visual experience and influence the results. We determined the number of training trials to be 5 based on the results of the pilot experiment. During this phase, we observed that nearly all children were able to understand the task requirements well after completing 5 practicing trials.

L 239: Apparently no non-parametric statistics was applied. Maybe it would be good to mention in the Statistics section briefly why this was justified.

We appreciate your suggestion and have cited two references in the Statistics section (Fagerland et al. 2012, Rochon et al. 2012). Fagerland et al., mentioned that when the sample size increases, the t-test is more robust. According to the central limit theorem, when the sample size is greater than 30, the sampling distribution of the mean can be safely assumed to be normal.

(http://www2.psychology.uiowa.edu/faculty/mordkoff/GradStats/part%201/I.07%20normal.p df). In fact, we also ran non-parametric statistics for our data and found the results to be robust.

L 290: 'FIQ' this abbreviation should be defined.

Regarding the abbreviation ’FIQ’, it stands for the abbreviation of the full-scale intellectual quotient, which was mentioned in Materials and methods section:

“Scores of the four broad areas constitute the full-scale intellectual quotient (FIQ).”

L. 290 ff.: These model 'BM-local = age + gender etc ' is a pretty sloppy notation. I think what is meant that a GLM was used that uses the predictors gender etc. time appropriate beta_i values. This formula should be corrected or one just says that a GLM was run with the predictors gender ....

The same criticism applies to these other models that follow.

We thank you for pointing this out. We have modified all formulas accordingly in the revised manuscript (see part3 of the Results section).

All these models assume linearity of the combination of the predictors.was this assumption verified?

We referred to the previous study of BM perception in children. They found main predictor variables, including IQ (Rutherford et al., 2012; Jones et al., 2011) and age (Annaz et al., 2010; van et al., 2016), have a linear relation with the ability of BM processing.

L. 296ff.: For model (b) it looks like general BM performance is strongly driven by the predictor global BM performance in the group of patients. Does the same observation also apply to the normals?

The same phenomenon was not observed in TD children. We have briefly discussed this point in the Discussion section of the revised manuscript (lines 449 - 459).

Reviewer #2 (Recommendations For The Authors):

(1) Please add public access to the data repository so data availability can be assessed.

The data of the study will be available at https://osf.io/37p5s/.

(2) Although overall, the language was clear and understandable, there are a few parts where language might confuse a reader and lead to misconceptions. For instance, line 52: Did the authors mean to refer to 'emotions and intentions' instead of 'emotions and purposes'? See also examples where rephrasing may help to reflect a statement is speculation rather than fact.

Thanks for the comments. We have carefully checked the full text and rephrased the confused statements.

(3) Line 83/84: Autism is not a 'mental disorder' - please change to something like 'developmental disability'. Authors are encouraged to adapt their language according to terms preferred by the community (e.g., see Fig. 5 in this article:

https://onlinelibrary.wiley.com/doi/10.1002/aur.2864)

Suggestion well taken. We have changed the wording accordingly:

“In recent years, BM perception has received significant attention in studies of mental disorders (e.g., schizophrenia30) and developmental disabilities, particularly in ASD, characterized by deficits in social communication and social interaction31,32.” (lines 93 - 95)

(4) Please report how the sample size for the study was determined.

In the Materials and methods section (lines 168 - 173), we explained how the sample size was determined.

Line 94: It would be helpful to have a brief description of what neurophysiological differences have been observed upon BM perception in children with ADHD.

Thanks for the comment. We have added a brief description of neurophysiological findings in children with ADHD (lines 108 - 111).

(6) Line 106/107 and 108/109: please add references.

We have revised this part, and the relevant findings and references are in line with the revised manuscript (lines 77, 132 - 133).

(7) Line 292: Please add what order the factors were entered into each regression model.

Regarding this issue, we used SPSS 26 for the main analysis. SPSS utilizes the Type III sum of squares (default) to evaluate models. Regardless of the order in the GLM, we will obtain the same result. For more information, please refer to the documentation of SPSS 26 (https://www.ibm.com/docs/en/spss-statistics/26.0.0?topic=features-glm-univariate-analysis).

Reviewer #3 (Recommendations For The Authors)

(1) Task specifics. It is key to understanding the findings, as well as the dissociation between tasks, that the precise nature of the stimuli is clear. I think there is room for improvement in description here. Task 1 is described as involving relocating dots within the range of the intact walker. Of course, PLWs are created by presenting dots at the joints, so relocation can involve either moving to another place on the body, or random movement within the 2D spatial array (which likely involves moving it off the body). Which was done? It is said that Ps must indicate the motion direction, but what was the display of the walker? Sagittal? Task 2 requires detecting whether there is an intact walker amongst scrambled walkers. Were all walkers completely overlaid? Task 3 requires detecting the left v right facing of an intact walker at different orientations, presented amongst noise. So Task 3 requires determining facing direction and Task 1 walking direction. Are these tasks the same but described differently? Or can walkers ever walk backwards? Wrt this point, I also think it would help the reader if example videos were uploaded.

We appreciate you for bringing this to our attention. With regards to Task 1, it appears that your second speculation is correct. We scrambled the original dots and randomly presented them within the 2D spatial array (which likely involved moving them off the body). As a result, the global configuration of the 13 dots was completed disrupted while preserving the motion trajectory of each individual dot. This led to the display of scrambled dots on the monitor (which does not resemble a human). In practice, these local BM cues contain information about motion direction. In Task 2, the target walkers completely overlaid by a mask that is approximately 1.44 times the size of the intact walker. The task requirements of Task 1 and Task3 are same, which is judging the motion (walking) direction. The difference is that Task 1 displayed a scrambled walker while Task 3 displayed an intact walker within a mask. We have clarified these points and improved our descriptions in Procedure section and created example videos for each task, which we believe will be helpful for the readers to understand each task.

(2) Gender confound (see above). I think that the authors should present the results without gender as a covariate. Can they separate boys and girls on the plots with different coloured individual datapoints, such that readers can see whether it's actually a gender effect driving the supposed ADHD effect? And show that there are no signs of a gender effect in their TD group?

This point is similar to the point #1 you mentioned. Please refer to our response to that point above.

(3) Autism possible confound (see above). I think the authors must report whether any of the ADHD group had an autism diagnosis.

Please refer to the response for the point #2 your mentioned.

(4) Conclusions concerning differences between the local and global tasks wrt SRS and age (see above). I believe the authors should add stats demonstrating differences between the correlations to support such claims, as well as demonstrating appropriate scatterplots for SRS-Task 1, SRS-Task 2, SRS-Task 3 and age-Task 1, age-Task2 and age-Task 3 (with separate lines of best fit for the two groups in each).

Please refer to the response for the point #3 your mentioned.

(5) Theoretical assumptions (see above). I would suggest rephrasing all claims here to outline that these discussed mechanistic differences between local and global BM processing are only possibilities and not known on the basis of existing data.

Please refer to the response for the point #4 your mentioned.

eLife assessment

This manuscript represents a cleanly designed experiment for assessing biological motion processing in children (mean age = 9) with and without ADHD. The group differences concerning accuracy in global and local motion processing abilities are solid, but the analyses suggesting dissociable relationships between global and local processing and social skills, age, and IQ are inconclusive. The results are useful in terms of understanding ADHD and the ontogenesis of different components of the processing of biological motion.

Reviewer #1 (Public Review):

Summary:

The paper presents a nice study investigating the impairments of biological motion perception in individuals with ADHD in comparison with neurotypical controls. Motivated by the idea that there is a relationship between biological motion perception and social capabilities, the authors investigated the impairments of local and global (holistic) biological motion perception, the diagnosis status, and several additional behavioral variables that are affected in ADHS (IQ, social responsiveness, and attention / impulsivity). As well local as global biological motion perception is impaired in ADHD individuals. In addition, the study demonstrates a significant correlation between local biological motion perception skills and the social responsiveness score in the ADHD group, but not in controls. A path analysis in the ADHD group suggests that general performance in<br /> biological motion perception is influenced mainly by global biological motion perception performance and attentional and perceptual reasoning skills.

Strengths:

It is true that there exists not much work on biological motion perception and ADHD. Therefore, the presented study contributes an interesting new result to the biological motion literature, and adds potentially also new behavioral markers for this clinical group. The design of the study is straightforward and technically sound, and the drawn conclusions are supported by the presented results.

Weaknesses:

Some of the claims about the relationship between genetic factors and ADHD and the components of biological motion processing have to remain speculative at this point because genetic influences were not explicitly tested in this paper. Specifically, the hypothesis that the perception of human social interaction is critically based on a local mechanism for the detection of asymmetry in foot trajectories of walkers (this is what 'BL-local' really measures), or on the detection of live agents in cluttered scenes seems not very plausible.

Based on my last comments, now the discussion has been changed in a way that tries to justify the speculative claims by citing a lot of other speculative papers, which does not really address the problem. For example, the fact that chicks walk towards biological motion stimuli is interesting. To derive that this verifies a fundamental mechanism in human biological motion processing is extremely questionable, given that birds do not even have a cortex. Taking the argumentation of the authors serious, one would have to assume that the 'Local BM' mechanism is probably located in the mesencephalon in humans, and then would have to interact in some way with social perception differences of ADHD children. To me all this seems to make very strong (over-)claims. I suggest providing a much more modest interpretation of the interesting experimental result, based on what has been really experimentally shown by the authors and closely related other data, rather than providing lots of far-reaching speculations.

In the same direction, in my view, go claims like 'local BM is an intrinsic trait' (L. 448) , which is not only imprecise (maybe better 'mechanisms of processing of local BM cues') but also rather questionable. Likely, this' local processing of BM' is a lower level mechanisms, located probably in early and mid-levels of the visual cortex, with a possible influence of lower structures. It seems not really plausible that this is related to a classical trait variables in the sense of psychology, like personality, as seems to be suggested here. Also here I suggest a much more moderate and less speculative interpretation of the results.

Reviewer #2 (Public Review):

Summary:

Tian et al. aimed to assess differences in biological motion (BM) perception between children with and without ADHD, as well as relationships to indices of social functioning and possible predictors of BM perception (including demographics, reasoning ability and inattention). In their study, children with ADHD showed poorer performance relative to typically developing children in three tasks measuring local, global, and general BM perception. The authors further observed that across the whole sample, performance in all three BM tasks was negatively correlated with scores on the social responsiveness scale (SRS), whereas within groups a significant relationship to SRS scores was only observed in the ADHD group and for the local BM task. Local and global BM perception showed a dissociation in that global BM processing was predicted by age, while local BM perception was not. Finally, general (local & global combined) BM processing was predicted by age and global BM processing, while reasoning ability mediated the effect of inattention on BM processing.

Strengths:

Overall, the manuscript is presented in a clear fashion and methods and materials are presented with sufficient detail so the study could be reproduced by independent researchers. The study uses an innovative, albeit not novel, paradigm to investigate two independent processes underlying BM perception. The results are novel and have the potential to have wide-reaching impact on multiple fields.

Weaknesses:

The manuscript has greatly improved in clarity and methodological considerations in response to the review. There are only a few minor points which deserve the authors' attention:

When outlining the moviation for the current study, results from studies in ADHD and ASD are used too interchangeably. The authors use a lack of evidence for contributing (psychological/developmental) factors on BM processing in ASD to motivate the present study and refer to evidence for differences between typical and non-typical BM processing using studies in both ASD and ADHD. While there are certainly overlapping features between the two conditions/neurotypes, they are not to be considered identical and may have distinct etiologies, therefore the distinction between the two should be made clearer.

In the first/main analysis, is unclear to me why in the revised manuscript the authors changed the statistical method from ANOVA/ANCOVA to independent samples t-tests (unless the latter were only used for post-hoc comparisons, then this needs to be stated). Furthermore, although p-values look robust, for this analysis too it should be indicated whether and how multiple comparison problems were accounted for.

Reviewer #3 (Public Review):

Strengths:

The authors present differences between ADHD and TD children in biological motion processing, and this question has not received as much attention as equivalent processing capabilities in autism. They use a task that appears well controlled. They raise some interesting mechanistic possibilities for differences in local and global motion processing, which are distinctions worth exploring. The group differences will therefore be of interest to those studying ADHD, as well as other developmental conditions, and those examining biological motion processing mechanisms in general.

Weaknesses:

The data are not strong enough to support claims about differences between global and lobal processing wrt social communication skills and age. The mechanistic possibilities for why these abilities may dissociate in such a way are interesting, but the crucial tests of differences between correlations do not present a clear picture. Further empirical work would be needed to test the authors' claims. Specifics:

The authors state frequently that it was the local BM task that related to social communication skills (SRS) and not the global tasks. However, the results section shows a correlation between SRS and all three tasks. The only difference is that when looking specifically within the ADHD group, the correlation is only significant for the local task. The supplementary materials demonstrate that tests of differences between correlations present an incomplete picture. Currently they have small samples for correlations, so this is unsurprising.

Theoretical assumptions. The authors make some statements about local vs global biological motion processing that should still be made more tentatively. They assume that local processing is specifically genetically whereas global processing is a product of experience. These data in newborn chicks are controversial and confounded - I cannot remember the specifics but I think there an upper vs lower visual field complexity difference here.

Readability. The manuscript needs very careful proofreading and correction for grammar. There are grammatical errors throughout.

Author Response

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

General response:

We thank the reviewers for their thorough evaluation of our manuscript. Working on the raised concerns has improved the manuscript greatly. Specifically, the recommendations to clarify the adopted assumptions in the study strengthened the motivation for the study; further, following up some of the reviewers’ concerns with additional analyses validated our chosen measures and strengthened the compatibility of the findings with the predictions of the dynamic attending framework. Below, you will find our detailed point-by-point responses, along with information on specific revisions.

The reviewers pointed out that study assumptions were unclear, some of the measures we chose were not well motivated, and the findings were not well enough explained considering possible alternatives. As suggested, we reformulated the introduction, explained the common assumptions of entrainment models that we adopted in the study, and further clarified how our chosen measures for the properties of the internal oscillators relate to these assumptions.

We realized that the initial emphasis on the compatibility of the current findings with predictions of entrainment models might have led to the wrong impression that the current study aimed to test whether auditory rhythmic processing is governed by timekeeper or oscillatory mechanisms. However, testing these theoretical models to explain human behavior necessitates specific paradigms designed to compare the contrasting predictions of the models. A number of studies do so by manipulating regularity in a stimulus sequence or expectancy of stimulus onsets, or assessing the perceived timing of targets that follow a stimulus rhythm. Such paradigms allow testing the prediction that an oscillator, underlying perceptual timing, would entrain to a regular but not an irregular sequence. This would further lead to stronger expectancies at the peak of the oscillation, where 'attentional energy' is the highest. These studies report 'rhythmic facilitation', where targets that align with the peaks of the oscillation are better detected than those that do not (see Henry and Herrmann (2014) and Haegens and Zion Golumbic (2018) for reviews). Additionally, unexpected endings of standard intervals, preceded by a regular entraining sequence, lead to a biased estimation of subsequent comparison intervals, due to the contrast between the attentional oscillator's phase and a deviating stimulus onset (Barnes & Jones, 2000; Large & Jones, 1999; McAuley & Jones, 2003). Even a sequence rate that is the multiple of the to-be-judged standard and comparison intervals give rise to rhythmic facilitation (McAuley & Jones, 2003), and the expectancy of a stimulus onset modulates duration judgments. These findings are not compatible with predictions of timekeeper models as time intervals in these models are represented arbitrarily and are not affected by expectancy violations.

In the current study, we adopted an entrainment approach to timing, rather than testing predictions of competing models. This choice was motivated by several aspects of entrainment models that align better with the aims of the current study. First, our focus was on understanding perception and production of rhythms, for which perception is better explained by entrainment models than by timekeeper models, which excel at explaining perception of isolated time intervals (McAuley, 2010). Moreover, we wanted to leverage the fact that entrainment models elegantly include parameters that can explain different aspects of timing abilities, and these parameters can be estimated in an individualized manner. For instance, the flexibility property of oscillators can be linked to the ability to adapt to changes in external context, while timekeeper or Bayesian timing approaches lack a specific mechanism to quantify temporal adaptation across perceptual and motor domains. Finally, that entrainment is observed across theoretical, behavioral, and neural levels renders entrainment models useful in explaining and generalizing behavior across different domains. Nevertheless, some results showed partial compatibility with predictions of the timekeeper models, such as the modulation of 'bestperformance rates' by the temporal context, observed in Experiment 1’ random-order sessions, where stimulus rates maximally differed across consecutive trials. However, given that the mean, standard deviation, and range of stimulus rates were identical across sessions, and timekeeper models assume no temporal adaptation in duration perception, we should have observed similar results across these sessions. Conversely, we found significant accuracy differences, biased duration judgments, and harmonic relationships between the best-performance rates. We elaborate more on these results with respect to their compatibility with the contrasting models of human temporal perception in the revised discussion.

Responses to specific comments:

(1.1) At times, I found it challenging to evaluate the scientific merit of this study from what was provided in the introduction and methods. It is not clear what the experiment assumes, what it evaluates, and which competing accounts or predictions are at play. While some of these questions are answered, clear ordering and argumentative flow is lacking. With that said, I found the Abstract and General Discussion much clearer, and I would recommend reformulating the early part of the manuscript based on the structure of those segments.

Second, in my reading, it is not clear to what extent the study assumes versus demonstrates the entrainment of internal oscillators. I find the writing somewhat ambiguous on this count: on the one hand, an entrainment approach is assumed a priori to design the experiment ("an entrainment approach is adopted") yet a primary result of the study is that entrainment is how we perceive and produce rhythms ("Overall, the findings support the hypothesis that an oscillatory system with a stable preferred rate underlies perception and production of rhythm..."). While one could design an experiment assuming X and find evidence for X, this requires testing competing accounts with competing hypotheses -- and this was not done.

We appreciate the reviewer’s concerns and suggestion to clarify the assumptions of the study and how the current findings relate to the predictions of competing accounts. To address these concerns:

• We added the assumptions of the entrainment models that we adopted in the Introduction section and reformulated the motivation to choose them accordingly.

• We clarified in the Introduction that the study’s aim was not to test the entrainment models against alternative theories of rhythm perception.

• We added a paragraph in the General Discussion to further distinguish predictions from the competing accounts. Here we discussed the compatibility of the findings with predictions of both entrainment and timekeeper models.

• We rephrased reasoning in the Abstract, Introduction, and General Discussion to further clarify the aims of the study, and how the findings support the hypotheses of the current study versus those of the dynamic attending theory.

(1.2) In my view, more evidence is required to bolster the findings as entrainment-based regardless of whether that is an assumption or a result. Indeed, while the effect of previous trials into the behaviour of the current trial is compatible with entrainment hypotheses, it may well be compatible with competing accounts as well. And that would call into question the interpretation of results as uncovering the properties of oscillating systems and age-related differences in such systems. Thus, I believe more evidence is needed to bolster the entrainment hypothesis.

For example, a key prediction of the entrainment model -- which assumes internal oscillators as the mechanism of action -- is that behaviour in the SMT and PTT tasks follows the principles of Arnold's Tongue. Specifically, tapping and listening performance should worsen systematically as a function of the distance between the presented and preferred rate. On a participant-by-participant, does performance scale monotonically with the distance between the presented and preferred rate? Some of the analyses hint at this question, such as the effect of 𝚫IOI on accuracy, but a recontextualization, further analyses, or additional visualizations would be helpful to demonstrate evidence of a tongue-like pattern in the behavioural data. Presumably, non-oscillating models do not follow a tongue-like pattern, but again, it would be very instructive to explicitly discuss that.

We thank the reviewer for the excellent suggestion of assessing 'Arnold's tongue' principles in timing performance. We agree that testing whether timing performance forms a pattern compatible with an Arnold tongue would further support our assumption that the findings related to preferred rate stem from an entrainment-based mechanism. We rather refer to the ‘entrainment region’, (McAuley et al., 2006) that corresponds to a slice in the Arnold tongue at a fixed stimulus intensity that entrains the internal oscillator. In both representations of oscillator behavior across a range of stimulus rates, performance should systematically increase as the difference between the stimulus rate and the oscillator's preferred rate, namely, 'detuning' decreases. In response to the reviewer’s comment, we ran further analyses to test this key prediction of entrainment models. We assessed performance at stimulus rates that were faster and slower than an individual's preferred rate estimates from in Experiment 1. To do so, we ran logistic regression models on aggregated datasets from all participants and sessions, where normalized IOI, in trials where the stimulus rate was faster than the preferred rate estimate, and in those where it was slower, predicted accuracy. Stimulus IOIs were normalized within each direction (faster- versus slower-than-preferred rate) using z-score transformation, and the direction was coded as categorical in the model. We reasoned that a positive slope for conditions with stimulus rates faster than IOI, and a negative slope from conditions with slower rates, should indicate a systematic accuracy increase toward the preferred rate estimate. This is exactly what we found. These results revealed significant main effect for the IOI and a significant interaction between IOI and direction, indicating that accuracy increased towards the preferred rate at fast rates and decreased as the stimulus rate diverged from the preferred rate at slow rates. We added these results to the respective subsections of Experiment 1 Methods and Results, added a plot showing the slices of the regression surfaces to Figure 2B and elaborated on the results in Experiment 1 Discussion. As the number of trials in Experiment 2 was much lower (N = 81), we only ran these additional analyses in Experiment 1.

(1.3) Fourth, harmonic structure in behaviour across tasks is a creative and useful metric for bolstering the entrainment hypothesis specifically because internal oscillators should display a preference across their own harmonics. However, I have some doubts that the analyses as currently implemented indicate such a relationship. Specifically, the main analysis to this end involves summing the residuals of the data closest to y=x, y=2*x and y=x/2 lines and evaluating whether this sum is significantly lower than for shuffled data. Out of these three dimensions, y=x does not comprise a harmonic, and this is an issue because it could by itself drive the difference of summed residuals with the shuffled data. I am uncertain whether rerunning the same analysis with the x=y dimension excluded constitutes a simple resolution because presumably there are baseline differences in the empirical and shuffled data that do not have to do with harmonics that would leak into the analysis. To address this, a simulation with ground truths could be helpful to justify analyses, or a different analysis that evaluates harmonic structure could be thought of.

We thank the reviewer for pointing out the weakness of the permutation test we developed to assess the harmonic relationship between Experiment 1’s preferred rate estimates. Datapoints that fall on the y=x line indeed do not represent harmonic relationships. They rather indicate one-to-one correspondence between the axes, which is a stronger indicator of compatibility between the estimates. Maybe speaking to the reviewer’s point, standard correlation analyses were not significant, which would have been expected if the permutation results were being driven by the y=x relationship. This was the reason we developed the permutation test to include integer-ratio datapoints could also contribute.

Based on reviewer’s comment, we ran additional analyses to assess the harmonic relationships between the estimates. The first analysis involved a circular approach. We first normalized each participant’s estimates by rescaling the slower estimate with respect to the faster one by division; and converted the values to radians, since a pair of values with an integer-ratio relationship should correspond to the same phase on a unit circle. Then, we assessed whether the resulting distribution of normalized values differed from a uniform distribution, using Rayleigh’s test, which was significant (p = .004). The circular mean of the distribution was 44 (SD = 53) degrees (M = 0.764, SD = 0.932 radians), indicating that the slower estimates were slightly slower than the fast estimate or its duplicates. As this distribution was skewed toward positive values due to the normalization procedure, we did not compare it against zero angle. Instead, we ran a second test, which was a modular approach. We first calculated how much the slower estimate deviated proportionally from the faster estimate or its multiples (i.e., subharmonics) by normalizing the estimates from both sessions by the faster estimate. The outcome measure was the modulus of the slower, relative to the faster estimate, divided by the faster estimate. Then, we ran a permutation test, shuffling the linear-order session estimates over 1000 iterations and taking the median percent deviation values for each iteration. The test statistic was significant (p = .004), indicating that the harmonic relationships we observed in the estimates were not due to chance or dependent on the assessment method. We added these details of additional analyses to assess harmonic relationships between the Experiment 1 preferred rate estimates in the Supplementary Information.

(2.1) The current study is presented in the framework of the ongoing debate of oscillator vs. timekeeper mechanisms underlying perceptual and motor timing, and authors claim that the observed results support the former mechanism. In this line, every obtained result is related by the authors to a specific ambiguous (i.e., not clearly related to a biophysical parameter) feature of an internal oscillator. As pointed out by an essay on the topic (Doelling & Assaneo, 2021), claiming that a pattern of results is compatible with an "oscillator" could be misleading, since some features typically used to validate or refute such mechanisms are not well grounded on real biophysical models. Relatedly, a recent study (Doelling et al., 2022) shows that two quantitatively different computational algorithms (i.e., absolute vs relative timing) can be explained by the same biophysical model. This demonstrates that what could be interpreted as a timekeeper, or an oscillator can represent the same biophysical model working under different conditions. For this reason, if authors would like to argue for a given mechanism underlying their observations, they should include a specific biophysical model, and test its predictions against the observed behavior. For example, it's not clear why authors interpret the observation of the trial's response being modulated by the rate of the previous one, as an oscillator-like mechanism underlying behavior. As shown in (Doelling & Assaneo, 2021) a simple oscillator returns to its natural frequency as soon as the stimulus disappears, which will not predict the long-lasting effect of the previous trial. Furthermore, a timekeeper-like mechanism with a long enough integration window is compatible with this observation.

Still, authors can choose to disregard this suggestion, and not testing a specific model, but if so, they should restrict this paper to a descriptive study of the timing phenomena.

We thank the reviewer for their valuable suggestion of to include a biophysical model to further demonstrate the compatibility of the current findings with certain predictions of the model. While we acknowledge the potential benefits of implementing a biophysical model to understand the relationships between model parameters and observed behavior, this goes beyond the scope of the current study.

We note that we have employed a modeling approach in a subsequent study to further explore how the properties and the resulting behavior of an oscillator map onto the patterns of human behavior we observed in the current study (Kaya & Henry, 2024, February 5). In that study, we fitted a canonical oscillator model, and several variants thereof, separately to datasets obtained from random-order and linear-order sessions of Experiment 1 of the current submission. The base model, adapted from McAuley and Jones (2003), assumed sustained oscillations within the trials of the experiment, and complete decay towards the preferred rate between the trials. We introduced a gradual decay parameter (Author response image 1A) that weighted between the oscillator's concurrent period value at the time of decay and its initial period (i.e., preferred rate). This parameter was implemented only within trials, between the standard stimulus sequence and comparison interval in Variant 1, between consecutive trials in Variant 2, and at both temporal locations in Variant 3. Model comparisons (Author response image 1B) showed that Variant 3 was the best-fitting model for both random- and linear-order datasets. Crucially, estimates for within- and between-trial decay parameters, obtained from Variant 3, were positively correlated, suggesting that oscillators gradually decayed towards their preferred rate at similar timescales after cessation of a stimulus.

Author response image 1.

(A) Illustration of the model fitted to Experiment 1 datasets and (B) model comparison results. In each trial, the model is initialized with a phase (ɸ) and period (P) value. A At the offset of each stimulus interval i, the model updates its phase (pink arrows) and period (blue arrows) depending on the temporal contrast (C) between the model state and stimulus onset and phase and period correction weights, Wɸ and Wp. Wdecaywithin updates the model period as a weighted average between the period calculated for the 5th interval, P5, and model’s preferred rate, P0. C, calculated at the offset of the comparison interval. Wdecaybetween parameter initializes the model period at the beginning of a new trial as a weighted average between the last period from the previous trial and P0. The base model’s assumptions are marked by asterisks, namely sustained oscillation during the silence (i=5), and complete decay between trials. B Left: The normalized probability of each model having the minimum BIC value across all models and across participants. Right: AICc, calculated from each model’s fit to participants’ single-session datasets. In both panels, random-order and linear-order sessions were marked in green and blue, respectively. B denotes the base model, and V1, V2 and V3 denote variants 1, 2 and 3, respectively.

Although our behavioral results and modeling thereof must necessarily be interpreted as reflecting the mechanics of an attentional, but not a neural oscillator, these findings might shed light on the controversy in neuroscience research regarding the timeline of entrainment decay. While multiple studies show that neural oscillations can continue at the entrained rate for a number of cycles following entrainment (Bouwer et al., 2023; Helfrich et al., 2017; Lakatos et al., 2013; van Bree et al., 2021), different modeling approaches reveal mixed results on this phenomenon. Whereas Doelling and Assaneo (2021) show that a Stuart-Landau oscillator returns immediately back to its preferred rate after synchronizing to an external stimulus, simulations of other oscillator types suggest gradual decay toward the preferred rate (Large, 1994; McAuley, 1995; Obleser et al., 2017) or self-sustained oscillation at the external stimulus rate (Nachstedt et al., 2017).

While the Doelling & Assaneo study (2021) provides insights on entrainment and behavior of the Stuart-Landau oscillator under certain conditions, the internal oscillators hypothesized by the dynamic attending theory might have different forms, therefore may not adhere to the behavior of a specific implementation of an oscillator model. Moreover, that a phase-coupled oscillator does not show gradual decay does not preclude that models with period tracking behave similarly. Adaptive frequency oscillators, for instance, are able to sustain the oscillation after the stimulus ceases (Nachstedt et al., 2017). Alongside with models that use Hebbian learning (Roman et al., 2023), the main implementations of the dynamic attending theory have parameters for period tracking and decay towards the preferred rate (Large, 1994; McAuley, 1995). In fact, the u-shaped pattern of duration discrimination sensitivity across a range of stimulus rates (Drake & Botte, 1993) is better explained by a decaying than a non-decaying oscillator (McAuley, 1995). To conclude, the literature suggests that the emergence of decay versus sustain behavior of the oscillators and the timeline of decay depend on the particular model used as well as its parameters and does therefore not offer a one-for-all solution.

Reviewer #2 (Recommendations For The Authors):

• Are the range, SD and mean of the random-order and linear-order sessions different? If so, why?

Information regarding the SD and mean of the random-order and linear-order sessions was added to Experiment 1 Methods section.

“While the mean (M = 599 ms), standard deviation (SD = 231 ms) and range (200, 998 ms) of the presented stimulus IOIs were identical between the sessions, the way IOI changed from trial to trial was different.“ (p. 5)

• Perhaps the title could mention the age-related flexibility effect you demonstrate, which is an important contribution that without inclusion in the title could be missed in literature searches.

We have changed the title to include age-related changes in oscillator flexibility. Thanks for the great suggestion.

• Is the statistical analysis in Figure 4A between subjects? Shouldn't the analyses be within subjects?

We have now better specified that the statistical analyses of Experiment 2’s preferred rate estimates were across the tasks, in Figure 4 caption.

"Vertical lines above the box plots represent within-participants pairwise comparisons." (p. 17)

• It says participants' hearing thresholds were measured using standard puretone audiometry. What threshold warranted participant exclusion and how many participants were excluded on the basis of hearing skills?

We have now clarified that hearing threshold was not an exclusion criterion.

"Participants were not excluded based on hearing threshold." (p. 11)

• "Tapping rates from 'fastest' and 'slowest' FMT trials showed no difference between pre- and postsession measurements, and were additionally correlated across repeated measurements" - could you point to the statistics for this comparison?

Table 2 includes the results from both experiments’ analyses on unpaced tapping. (p. 10)

“The results of the pairwise comparisons between tapping rates from all unpaced tapping tasks across measurements are provided in Table 2.” (p. 15)

• How was the loudness (dB) of the woodblock stimuli determined on a participant-by-participant basis? Please ignore if I missed this.

Participants were allowed to set the volume to a comfortable level.

"Participants then set the sound volume to a level that they found comfortable for completing the task." (p. 4)

• Please spell out IOI, DEV, and other terms in full the first time they are mentioned in the manuscript.

We added the descriptions of abbreviations before their initial mention.

"In each experimental session, 400 unique trials of this task were presented, each consisting of a combination of the three main independent variables: the inter-onset interval, IOI; amount of deviation of the comparison interval from the standard, DEV, and the amount of change in stimulus IOI between consecutive trials, 𝚫IOI. We explain each of these variables in detail in the next paragraphs." (p. 4)

• Small point: In Fig 1 sub-text, random order and linear order are explained in reverse order from how they are presented in the figure.

We fixed the incompatibility between of Figure 1 content and caption.

• Small point: I found the elaborate technical explanation of windowing methods, including alternatives that were not used, unnecessary.

We moved the details of the smoothing analysis to the Supplementary Information.

• With regard to the smoothing explanation, what is an "element"? Is this a sample? If so, what was the sampling rate?

We reworded ‘element’ as ‘sample’. In the smoothing analyses, the sampling rate was the size of the convolution window, which was set to 26 for random-order, 48 for linear-order sessions.

• Spelling/language error: "The pared-down", "close each other", "always small (+4 ms), than".

We fixed the spelling errors.

Reviewer #3 (Recommendations For The Authors):

• My main concern is the one detailed as a weakness in the public review. In that direction, if authors decide to keep the mechanistic interpretation of the outcomes (which I believe is a valuable one) here I suggest a couple of models that they can try to adapt to explain the pattern of results:

a. Roman, Iran R., et al. "Hebbian learning with elasticity explains how the spontaneous motor tempo affects music performance synchronization." PLOS Computational Biology 19.6 (2023): e1011154.

b. Bose, Amitabha, Áine Byrne, and John Rinzel. "A neuromechanistic model for rhythmic beat generation." PLoS Computational Biology 15.5 (2019): e1006450.

c. Egger, Seth W., Nhat M. Le, and Mehrdad Jazayeri. "A neural circuit model for human sensorimotor timing." Nature Communications 11.1 (2020): 3933.

d. Doelling, K. B., Arnal, L. H., & Assaneo, M. F. (2022). Adaptive oscillators provide a hard-coded Bayesian mechanism for rhythmic inference. bioRxiv, 2022-06

Thanks for the suggestion! Please refer to our response (2.1.) above. To summarize, although we considered a full, well-fleshed-out modeling approach to be beyond the scope of the current work, we are excited about and actively working on exactly this. Our modeling take is available as a preprint (Kaya & Henry, 2024, February 5).

• Since the authors were concerned with the preferred rate they circumscribed the analysis to extract the IOI with better performance. Would it be plausible to explore how is the functional form between accuracy and IOI? This could shed some light on the underlying mechanism.

Unfortunately, we were unsure about what the reviewer meant by the functional form between accuracy and IOI. We interpret it to mean a function that takes IOI as input and outputs an accuracy value. In that case, while we agree that estimating this function might indeed shed light on the underlying mechanisms, this type of analysis is beyond the scope of the current study. Instead, we refer the reviewer and reader to our modeling study (please see our response (2.1.) above) that includes a model which takes the stimulus conditions, including IOI, and model parameters for preferred rate, phase and period correction and within- and between-trial decay and outputs predicted accuracy for each trial. We believe that such modeling approach, as compared to a simple function, gives more insights regarding the relationship between oscillator properties and duration perception.

• Is the effect caused by the dIOI modulated by the distance to the preferred frequency?

We thank the reviewer for the recommendation. We measured flexibility by the oscillator's ability to adapt to on-line changes in the temporal context (i.e., effect of 𝚫IOI on accuracy), rather than by quantifying the range of rates with improved accuracy. Nevertheless, we acknowledge that distance to the preferred rate should decrease accuracy, as this is a key prediction of entrainment models. In fact, testing this prediction was recommended also by the other reviewer, in response to which we ran additional analyses. These analyses involved assessment of the relationship between accuracy and detuning. Specifically, we assessed accuracy at stimulus rates that were faster and slower than an individual's preferred rate estimates from in Experiment 1. We ran logistic regression models on aggregated datasets from all participants and sessions, where accuracy was predicted by z-scored IOI, from trials where the stimulus rate was faster than the preferred rate estimate, and in those where it was slower. The model had a significant main effect of IOI and an interaction between IOI and direction (i.e., whether stimulus rate was faster or slower than the preferred rate estimate), indicating that accuracy increased towards the preferred rate at fast rates and decreased as the stimulus rate diverged from the preferred rate at slow rates. We added information regarding this analysis to the respective subsections of Experiment 1 Methods and Results, added a plot showing the slices of the regression surfaces to Figure 2B and elaborated on the results in Experiment 1 Discussion. As the number of trials in Experiment 2 was insufficient, we only ran these additional analyses in Experiment 1. We agree that a range-based measure of oscillator flexibility would also index the oscillators’ adaptive abilities. However, the current paradigms were designed for assessment of temporal adaptation. Thus, comparison of the two approaches to measuring oscillator flexibility, which can be addressed in future studies, is beyond the scope of the current study.

• Did the authors explore if the "motor component" (the difference between the motor and perceptual rates) is modulated by the participants age?

In response to the reviewer’s comment, we correlated the difference between the motor and perceptual rates with age, which was nonsignificant.

• Please describe better the slider and the keypress tasks. For example, what are the instructions given to the participant on each task, and how they differ from each other?

We added the Experiment 2 instructions in Appendix A.

• Typos: The caption in figure one reads 2 ms, while I believe it should say 200. Page 4 mentions that there are 400 trials and page 5 says 407.

We fixed the typos.

References

Barnes, R., & Jones, M. R. (2000). Expectancy, attention, and time. Cogn Psychol, 41(3), 254-311. https://doi.org/10.1006/cogp.2000.0738

Bouwer, F. L., Fahrenfort, J. J., Millard, S. K., Kloosterman, N. A., & Slagter, H. A. (2023). A Silent Disco: Differential Effects of Beat-based and Pattern-based Temporal Expectations on Persistent Entrainment of Low-frequency Neural Oscillations. J Cogn Neurosci, 35(6), 9901020. https://doi.org/10.1162/jocn_a_01985

Doelling, K. B., Arnal, L. H., & Assaneo, M. F. (2022). Adaptive oscillators provide a hard-coded Bayesian mechanism for rhythmic inference. bioRxiv. https://doi.org/10.1101/2022.06.18.496664

Doelling, K. B., & Assaneo, M. F. (2021). Neural oscillations are a start toward understanding brain activity rather than the end. PLoS Biol, 19(5), e3001234. https://doi.org/10.1371/journal.pbio.3001234

Drake, C., & Botte, M. C. (1993). Tempo sensitivity in auditory sequences: evidence for a multiplelook model. Percept Psychophys, 54(3), 277-286. https://doi.org/10.3758/bf03205262

Haegens, S., & Zion Golumbic, E. (2018). Rhythmic facilitation of sensory processing: A critical review. Neurosci Biobehav Rev, 86, 150-165. https://doi.org/10.1016/j.neubiorev.2017.12.002

Helfrich, R. F., Huang, M., Wilson, G., & Knight, R. T. (2017). Prefrontal cortex modulates posterior alpha oscillations during top-down guided visual perception. Proc Natl Acad Sci U S A, 114(35), 9457-9462. https://doi.org/10.1073/pnas.1705965114

Henry, M. J., & Herrmann, B. (2014). Low-Frequency Neural Oscillations Support Dynamic Attending in Temporal Context. Timing & Time Perception, 2(1), 62-86. https://doi.org/10.1163/22134468-00002011

Kaya, E., & Henry, M. J. (2024, February 5). Modeling rhythm perception and temporal adaptation: top-down influences on a gradually decaying oscillator. https://doi.org/10.31234/osf.io/q9uvr

Lakatos, P., Musacchia, G., O'Connel, M. N., Falchier, A. Y., Javitt, D. C., & Schroeder, C. E. (2013). The spectrotemporal filter mechanism of auditory selective attention. Neuron, 77(4), 750-761. https://doi.org/10.1016/j.neuron.2012.11.034

Large, E. W. (1994). Dynamic representation of musical structure. The Ohio State University.

Large, E. W., & Jones, M. R. (1999). The dynamics of attending: How people track time-varying events. Psychological Review, 106(1), 119-159. https://doi.org/Doi 10.1037/0033295x.106.1.119

McAuley, J. D. (1995). Perception of time as phase: Toward an adaptive-oscillator model of rhythmic pattern processing Indiana University Bloomington].

McAuley, J. D. (2010). Tempo and Rhythm. In Music Perception (pp. 165-199). https://doi.org/10.1007/978-1-4419-6114-3_6

McAuley, J. D., & Jones, M. R. (2003). Modeling effects of rhythmic context on perceived duration: a comparison of interval and entrainment approaches to short-interval timing. J Exp Psychol Hum Percept Perform, 29(6), 1102-1125. https://doi.org/10.1037/0096-1523.29.6.1102

McAuley, J. D., Jones, M. R., Holub, S., Johnston, H. M., & Miller, N. S. (2006). The time of our lives: life span development of timing and event tracking. J Exp Psychol Gen, 135(3), 348-367. https://doi.org/10.1037/0096-3445.135.3.348

Nachstedt, T., Tetzlaff, C., & Manoonpong, P. (2017). Fast Dynamical Coupling Enhances Frequency Adaptation of Oscillators for Robotic Locomotion Control. Front Neurorobot, 11, 14. https://doi.org/10.3389/fnbot.2017.00014

Obleser, J., Henry, M. J., & Lakatos, P. (2017). What do we talk about when we talk about rhythm? PLoS Biol, 15(9), e2002794. https://doi.org/10.1371/journal.pbio.2002794

Roman, I. R., Roman, A. S., Kim, J. C., & Large, E. W. (2023). Hebbian learning with elasticity explains how the spontaneous motor tempo affects music performance synchronization. PLoS Comput Biol, 19(6), e1011154. https://doi.org/10.1371/journal.pcbi.1011154<br /> van Bree, S., Sohoglu, E., Davis, M. H., & Zoefel, B. (2021). Sustained neural rhythms reveal endogenous oscillations supporting speech perception. PLoS Biol, 19(2), e3001142. https://doi.org/10.1371/journal.pbio.3001142

eLife assessment

This valuable study has practical implications for understanding rhythm perception and production in human cognition. The evidence for individual frequency preferences and a deterioration in frequency adaptation with age is solid. These findings may inform existing models of rhythm perception and production, and the reported effects of age may have clinical implications.

Reviewer #1 (Public Review):

Summary:

This study assumes but also demonstrates that auditory rhythm processing is produced by internal oscillating systems and evaluates the properties of internal oscillators across individuals. The authors designed an experiment and performed analyses that address individuals' preferred rate and flexibility, with a special focus on how much past rhythms influence subsequent trials. They find evidence for such historical dependence and show that we adapt less well to new rhythms as we age. While I have some doubts about the entrainment-based interpretation of the results, this work offers a useful contribution to our understanding of individual differences in rhythm processing regardless.

Strengths:

The inclusion of two tasks -- a tapping and a listening task -- complement each other methodologically. By analysing both the production and tracking of rhythms, the authors emphasize the importance of the characteristics of the receiver, the external world, and their interplay. The relationship between the two tasks and components within tasks are explored using a range of analyses. The visual presentation of the results is very clear. The age-related changes in flexibility are useful and compelling.

The paper includes a discussion of the study assumptions, and it contextualizes itself more explicitly as taking entrainment frameworks as a starting point. As such, even if the entrainment of oscillators cannot be decisively shown, it is now clear that this is nevertheless adopted as a useful theoretical lens.

Weaknesses:

The newly included analyses that justify an entrainment or oscillator-based interpretation of the result could be presented in a clearer manner so that readers can parse their validity better. For example, in line with an entrainment interpretation, the regression lines in Figure 2B show accuracy increases as the IOI moves towards the preferred rate -- but then beyond the preferred rate, accuracy appears to increase further still. Furthermore, the additional analyses on harmonic relationships could be enriched with justification and explanation of each of its steps.

Reviewer #2 (Public Review):

Summary:

The current work describes a set of behavioral tasks to explore individual differences in the preferred perceptual and motor rhythms. Results show a consistent individual preference for a given perceptual and motor frequency across tasks and, while these were correlated, the latter is slower than the former one. Additionally, the adaptation accuracy to rate changes is proportional to the amount of rate variation and, crucially, the amount of adaptation decreases with age.

Strengths:

Experiments are carefully designed to measure individual preferred motor and perceptual tempo. Furthermore, the experimental design is validated by testing the consistency across tasks and test-retest, what makes the introduced paradigm a useful tool for future research.<br /> The obtained data is rigorously analyzed using a diverse set of tools, each adapted to the specificities across the different research questions and tasks.<br /> This study identifies several relevant behavioral features: (i) each individual shows a preferred and reliable motor and perceptual tempo and, while both are related, the motor is consistently slower than the pure perceptual one; (ii) the presence of hysteresis in the adaptation to rate variations; and (iii) the decrement of this adaptation with age. All these observations are valuable for the auditory-motor integration field of research, and they could potentially inform existing biophysical models to increase their descriptive power.

Weaknesses:

To get a better understanding of the mechanisms underlying the behavioral observations, it would have been useful to compare the observed pattern of results with simulations done with existing biophysical models. However, this point is addressed if the current study is read along with this other publication of the same research group: Kaya, E., & Henry, M. J. (2024, February 5). Modeling rhythm perception and temporal adaptation: top-down influences on a gradually decaying oscillator. https://doi.org/10.31234/osf.io/q9uvr

Author Response

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

Reviewer #1 (Recommendations For The Authors):

I only have a few minor suggestions:

Abstract: I really liked the conclusion (that IM and VWM are two temporal extremes of the same process) as articulated in lines 557--563. (It is always satisfying when the distinction between two things that seem fundamentally different vanishes). If something like this but shorter could be included in the Abstract, it would highlight the novel aspects of the results a little more, I think.

Thank you for this comment. We have added the following to the abstract:

“A key conclusion is that differences in capacity classically thought to distinguish IM and VWM are in fact contingent upon a single resource-limited WM store.”

L 216: There's an orphan parenthesis in "(justifying the use".

Fixed.

L 273: "One surprising result was the observed set size effect in the 0 ms delay condition". In this paragraph, it might be a good idea to remind the reader of the difference between the simultaneous and zero-delay conditions. If I got it right, the results differ between these conditions because it takes some amount of processing time to interpret the cue and free the resources associated with the irrelevant stimuli. Recalling that fact would make this paragraph easier to digest.

That is correct. However, at this point in the text, we have not yet fitted the DyNR model to the data. Therefore, we believe that introducing cue processing and resource reallocation as concepts that differentiate between those two conditions would disrupt the flow of this paragraph. We address these points soon after, in a paragraph starting on line 341.

Figures 3, 5: The labels at the bottom of each column in A would be more clear if placed at the top of each column instead. That way, the x-axis for the plots in A could be labeled appropriately, as "Error in orientation estimate" or something to that effect.

We edited both figures, now Figure 4 and Figure 6, as suggested.

L 379: It should be "(see Eq 6)", I believe.

That is correct, line 379 (currently line 391) should read ‘Eq 6’. Fixed.

L 379--385: I was a bit mystified as to why the scaled diffusion rate produced a worse fit than a constant rate. I imagine the scaled version was set to something like

sigma^2_diff_scaled = sigma^2_base + K*(N-1)

where N is the set size and sigma^2_base and K are parameters. If this model produced a similar fit as with a constant diffusion rate, the AIC would penalize it because of the extra parameter. But why would the fit be worse (i.e., not match the pattern of variability)? Shouldn't the fitter just find that the K=0 solution is the best? Not a big deal; the Nelder-Mead solutions can wobble when that many parameters are involved, but if there's a simple explanation it might be worth commenting on.

The scaled diffusion was implemented by extending Eq 6 in the following way:

σ(t)2 = (t-toffset) * σ̇ 2diff * N

where N is set size. Therefore, the scaling was not associated with a free parameter that could become 0 if set size did not affect diffusion rate, but variability rather mandatory increased with set size. We now clarify this in the text:

“The second variant was identical to the proposed model, except that we replaced the constant diffusion rate with a set size scaled diffusion rate by multiplying the right side of Eq 6 by N.“

Figure 4 is not mentioned in the main text. Maybe the end of L 398 would be a good place to point to it. The paragraph at L 443-455 would also benefit from a couple of references to it.

Thank you for this suggestion. Figure 4 (now Figure 5) was previously mentioned on line 449 (previously line 437), but now we have included it on line 410 (previously line 398), within the paragraph spanning lines 455-467 (previously 443-455), and also on line 136 where we first discuss masking effects.

L 500: Figure S7 is mentioned before Figures S5 and S6. Quite trivial, I know....

Thank you for this comment. There was no specific reason for Figure S7 to appear after S5 & S6, so we simply swapped their order to be consistent with how they are referred to in the manuscript (i.e., S7 became S5, S5 became S6, and S6 became S7).

Reviewer #2 (Recommendations For The Authors):

(1) One potential weakness is that the model assumes sensory information is veridical. However, this isn't likely the case. Acknowledging noise in sensory representations could affect the model interpretation in a couple of different ways. First, neurophysiological recordings have shown normalization affects sensory representations, even when a stimulus is still present on the screen. The DyNR model partially addresses this concern because reports are drawn from working memory, which is normalized. However, if sensory representations were also normalized, then it may improve the model variant where subjects draw directly from sensory representations (an alternative model that is currently described but discarded).

Thank you for this suggestion. We can consider two potential mechanisms through which divisive normalization might be incorporated into sensory processing within the DyNR model.

The first possibility involves assuming that normalization is pre-attentive. In this scenario, the sensory activity of each object would be rescaled at the lowest level of sensory processing, occurring before the allocation of attentional or VWM resources. One strong prediction of such an implementation is that recall error in the simultaneous cue condition (Experiment 1) should vary with set size. However, this prediction is inconsistent with the observed data, which failed to show a significant difference between set sizes, and is more closely aligned with the hypothesis of no-difference (F(2,18) = 1.26, p = .3, η2 = .04, BF10 = 0.47). On that basis, we anticipate that introducing normalization as a pre-attentive mechanism would impair the model fit.

An alternative scenario is to consider normalization as post-attentive. In the simultaneous cueing condition, only one item is attended (i.e., the cued one), regardless of the displayed set size. Here, we would expect normalized activity for a single item, regardless of the number of presented objects, which would then be integrated into VWM. This expanded DyNR model with post-attentive normalization would make exactly the same predictions as the proposed DyNR for recall fidelity, so distinguishing between these models would not be possible based on working memory experiments.

To acknowledge the possibility that sensory signals could undergo divisive normalization and to motivate future research, we have added the following to our manuscript:

“As well as being implicated in higher cognitive processes including VWM (Buschman et al, 2011; Sprague et al., 2014), divisive normalization has been shown to be widespread in basic sensory processing (Bonin et al., 2005; Busse et al., 2009; Ni et al., 2017). The DyNR model presently incorporates the former but not the latter type of normalization. While the data observed in our experiments do not provide evidence for normalization of sensory signals (note comparable recall errors across set size in the simultaneous cue condition of Experiment 1), this may be because sensory suppressive effects are localized and our stimuli were relatively widely separated in the visual field: future research could explore the consequences of sensory normalization for recall from VWM using, e.g., centre-surround stimuli (Bloem et al., 2018).”

Bloem, I. M., Watanabe, Y. L., Kibbe, M. M., & Ling, S. (2018). Visual Memories Bypass Normalization. Psychological Science, 29(5), 845–856. https://doi.org/10.1177/0956797617747091

Bonin, V., Mante, V., & Carandini, M. (2005). The Suppressive Field of Neurons in Lateral Geniculate Nucleus. The Journal of Neuroscience, 25(47), 10844–10856. https://doi.org/10.1523/JNEUROSCI.3562-05.2005

Buschman, T. J., Siegel, M., Roy, J. E., & Miller, E. K. (2011). Neural substrates of cognitive capacity limitations. Proceedings of the National Academy of Sciences, 108(27), 11252–11255. https://doi.org/10.1073/pnas.1104666108

Busse, L., Wade, A. R., & Carandini, M. (2009). Representation of Concurrent Stimuli by Population Activity in Visual Cortex. Neuron, 64(6), 931–942. https://doi.org/10.1016/j.neuron.2009.11.004

Ni, A. M., & Maunsell, J. H. R. (2017). Spatially tuned normalization explains attention modulation variance within neurons. Journal of Neurophysiology, 118(3), 1903–1913. https://doi.org/10.1152/jn.00218.2017

Sprague, T. C., Ester, E. F., & Serences, J. T. (2014). Reconstructions of Information in Visual Spatial Working Memory Degrade with Memory Load. Current Biology, 24(18), 2174–2180. https://doi.org/10.1016/j.cub.2014.07.066

Second, visual adaptation predicts sensory information should decrease over time. This would predict that for long stimulus presentation times, the error would increase. Indeed, this seems to be reflected in Figure 5B. This effect is not captured by the DyNR model.

Indeed, neural responses in the visual cortex have been observed to quickly adapt during stimulus presentation, showing reduced responses to prolonged stimuli after an initial transient (Groen et al., 2022; Sawamura et al., 2006; Zhou et al., 2019). This adaptation typically manifests as 1) reduced activity towards the end of stimulus presentation and 2) a faster decay towards baseline activity after stimulus offset.

In the DyNR model, we use an idealized solution in which we convolve the presented visual signal with a response function (i.e., temporal filter). At the longest presentation durations, in DyNR, the sensory signal plateaus and remains stable until stimulus offset. Because our psychophysical data does not allow us to identify the exact neural coding scheme that underlies the sensory signal, we tend to favour this simple implementation, which is broadly consistent with some previous attempts to model temporal dynamics in sensory responses (e.g., Carandini and Heeger, 1994). However, we agree with the reviewer that some adaptation of the sensory signal with prolonged presentation would also be consistent with our data.

We have added the following to the manuscript:

“In Experiment 2, the longest presentation duration shows an upward trend in error at set sizes 4 and 10. While this falls within the range of measurement error, it is also possible that this is a meaningful pattern arising from visual adaptation of the sensory signal, whereby neural populations reduce their activity after prolonged stimulation. This would mean less residual sensory signal would be available after the cue to supplement VWM activity, predicting a decline in fidelity at higher set sizes. Visual adaptation has previously been successfully accounted for by a type of delayed normalization model in which the sensory signal undergoes a series of linear and nonlinear transformations (Zhou et al., 2019). Such a model could in future be incorporated into DyNR and validated against psychophysical and neural data.”

Carandini, M., & Heeger, D. J. (1994). Summation and division by neurons in primate visual cortex. Science, 264(5163), 1333–1336. https://doi.org/10.1126/science.8191289

Groen, I. I. A., Piantoni, G., Montenegro, S., Flinker, A., Devore, S., Devinsky, O., Doyle, W., Dugan, P., Friedman, D., Ramsey, N. F., Petridou, N., & Winawer, J. (2022). Temporal Dynamics of Neural Responses in Human Visual Cortex. The Journal of Neuroscience, 42(40), 7562–7580. https://doi.org/10.1523/JNEUROSCI.1812-21.2022

Sawamura, H., Orban, G. A., & Vogels, R. (2006). Selectivity of Neuronal Adaptation Does Not Match Response Selectivity: A Single-Cell Study of the fMRI Adaptation Paradigm. Neuron, 49(2), 307–318. https://doi.org/10.1016/j.neuron.2005.11.028

Zhou, J., Benson, N. C., Kay, K., & Winawer, J. (2019). Predicting neuronal dynamics with a delayed gain control model. PLOS Computational Biology, 15(11), e1007484. https://doi.org/10.1371/journal.pcbi.1007484

(2) A second potential weakness is that, in Experiment 1, the authors briefly change the sensory stimulus at the end of the delay (a 'phase shift', Fig. 6A). I believe this is intended to act as a mask. However, I would expect that, in the DyNR model, this should be modeled as a new sensory input (in Experiment 2, 50 ms is plenty of time for the subjects to process the stimuli). One might expect this change to disrupt sensory and memory representations in a very characteristic manner. This seems to make a strong testable hypothesis. Did the authors find evidence for interference from the phase shift?

The phase shift was implemented with the intention of reducing retinal after-effects, essentially acting as a mask for retinal information only; crucially the orientation of the stimulus is unchanged by the phase shift, so from the perspective of the DyNR model, it transmits the same orientation information to working memory as the original stimulus.

If our objective were to model sensory input at the level of individual neurons and their receptive fields, we would indeed need to treat this phase shift as a novel input. Nevertheless, for DyNR, conceived as an idealization of a biological system for encoding orientation information, we can safely assume that visual areas in biological organisms have a sufficient number of phase-sensitive simple cells and phase-indifferent complex cells to maintain the continuity of input to VWM.

When comparing conditions with and without the phase shift of stimuli (Fig S1B), we found performance to be comparable in the perceptual condition (simultaneous presentation) and with the longest delay (1 second), suggesting that the phase shift did not change the visibility or encoding of information into VWM. In contrast, we found strong evidence that observers had access to an additional source of information over intermediate delays when the phase shift was not used. This was evident through enhanced recall performance from 0 ms to 400 ms delay. Based on this, we concluded that the additional source of information available in the absence of a phase shift was accessible immediately following stimulus offset and had a brief duration, aligning with the theoretical concept of retinal afterimages.

(3) It seems odd that the mask does not interrupt sensory processing in Experiment 2. Isn't this the intended purpose of the mask? Should readers interpret this as all masks not being effective in disrupting sensory processing/iconic memory? Or is this specific to the mask used in the experiment?

Visual masks are often described as instantly and completely halting the visual processing of information that preceded the mask. We also anticipated the mask would entirely terminate sensory processing, but our data indicate the effect was not complete (as indicated by model variants in Experiment 2). Nevertheless, we believe we achieved our intended goal with this experiment – we observed a clear modulation of response errors with changing stimulus duration, indicating that the post-stimulus information that survived masking did not compromise the manipulation of stimulus duration. Moreover, the DyNR model successfully accounted for the portion of signal that survived the mask.

We can identify two possible reasons why masking was incomplete. First, it is possible that the continuous report measure used in our experiments is more sensitive than the discrete measures (e.g., forced-choice methods) commonly employed in experiments that found masks to be 100% effective. Second, despite using a flickering white noise mask at full contrast, it is possible that it may not have been the most effective mask; for instance, a mask consisting of many randomly oriented Gabor patches matched in spatial frequency to the stimuli could prove more effective. We decided against such a mask because we were concerned that it could potentially act as a new input to orientation-sensitive neurons, rather than just wiping out any residual sensory activity.

(4) I apologize if I missed it, but the authors did not compare the DyNR model to a model without decaying sensory information for Experiment 1.

We tested two DyNR variants in which the diffusion process was solely responsible for memory fidelity dynamics. These models assumed that the sensory signal terminates abruptly with stimuli offset, and the VWM signal encoding the stimuli was equal to the limit imposed by normalization, independent of the delay duration.

As variants of this model failed to account for the observed response errors both quantitatively (see 'Fixed neural signal' under Model variants) and qualitatively (Figure S3), we decided not to test any more restrictive variants, such as the one without sensory decay and diffusion.

(5) In the current model, selection is considered to be absolute (all or none). However, this need not be the case (previous work argues for graded selection). Could a model where memories are only partially selected, in a manner that is mediated by load, explain the load effects seen in behavior?

Thank you for this point. If attentional selection was partial, it would affect the observers’ efficiency in discarding uncued objects to release allocated resources and encode additional information about the cued item. We and others have previously examined whether humans can efficiently update their VWM when previous items become obsolete. For example, Taylor et al. (2023) showed that observers could efficiently remove uncued items from VWM and reallocate the released resources to new visual information. These findings align with results from other studies (e.g., Ecker, Oberauer, & Lewandowsky, 2014; Kessler & Meiran, 2006; Williams et al., 2013).

Based on these findings, we feel justified in assuming that observers in our current task were capable of fully removing all uncued objects, allowing them to continue the encoding process for the cued orientation that was already partially stored in VWM, such that the attainable limit on representational precision for the cued item equals the maximum precision of VWM.

Partial removal could in principle be modelled in the DyNR model by introducing an additional plateau parameter specifying a maximum attainable precision after the cue. Our concern would be that such a plateau parameter would trade off with the parameter associated with Hick’s law (i.e., cue interpretation time). The former would control the amount of information that can be encoded into VWM, while the latter regulates the amount of sensory information available for encoding. We are wary of adding additional parameters, and hence flexibility, to the model where we do not have the data to sufficiently constrain them.

Ecker, U. K. H., Oberauer, K., & Lewandowsky, S. (2014b). Working memory updating involves item-specific removal. Journal of Memory and Language, 74, 1–15. https://doi.org/10.1016/j.jml. 2014.03.006

Kessler, Y., & Meiran, N. (2006). All updateable objects in working memory are updated whenever any of them are modified: Evidence from the memory updating paradigm. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32, 570–585. https://doi.org/10.1037/0278-7393.32.3.570

Taylor, R., Tomić, I., Aagten-Murphy, D., & Bays, P. M. (2023). Working memory is updated by reallocation of resources from obsolete to new items. Attention, Perception, & Psychophysics, 85(5), 1437–1451. https://doi.org/10.3758/s13414-022-02584-2

Williams, M., & Woodman, G. F. (2012). Directed forgetting and directed remembering in visual working memory. Journal of Experimental Psychology. Learning, Memory, and Cognition, 38(5), 1206–1220. https://doi.org/10.1037/a0027389

(6) Previous work, both from the authors and others, has shown that memories are biased as if they are acted on by attractive/repulsive forces. For example, the memory of an oriented bar is biased away from horizontal and vertical and biased towards diagonals. This is not accounted for in the current model. In particular, this could be one mechanism to generate a non-uniform drift rate over time. As noted in the paper, a non-uniform drift rate could capture many of the behavioral effects reported.

The reviewer is correct that the model does not currently include stimulus-specific effects, although our work on that topic provides a clear template for incorporating them in future (e.g. Taylor & Bays, 2018). Specifically on the question of generating a non-uniform drift, we have another project that currently looks at this exact question (cited in our manuscript as Tomic, Girones, Lengyel, and Bays; in prep.). By examining various datasets with varying memory delays, including the Additional Dataset 1 reported in the Supplementary Information, we found that stimulus-specific effects on orientation recall remain constant with retention time. Specifically, although there is a clear increase in overall error over time, estimation biases remain constant in direction and amplitude, indicating that the bias does not manifest in drift rates (see also Rademaker et al., 2018; Figure S1).

Taylor, R., & Bays, P. M. (2018). Efficient coding in visual working memory accounts for stimulus-specific variations in recall. The Journal of Neuroscience, 1018–18. https://doi.org/10.1523/JNEUROSCI.1018-18.2018

Rademaker, R. L., Park, Y. E., Sack, A. T., & Tong, F. (2018). Evidence of gradual loss of precision for simple features and complex objects in visual working memory. Journal of Experimental Psychology: Human Perception and Performance. https://doi.org/10.1037/xhp0000491

(7) Finally, the authors use AIC to compare many different model variants to the DyNR model. The delta-AICs are high (>10), indicating a strong preference for the DyNR model over the variants. However, the overall quality of fit to the data is not clear. What proportion of the variance in data was the model able to explain? In particular, I think it would be helpful for the reader if the authors reported the variance explained on withheld data (trials, conditions, or subjects).

Thank you for this comment.

Below we report the estimates of r2, representing the goodness of fit between observed data (i.e., RMSE) and the DyNR model predictions.

In Experiment 1, the r2 values between observations and predictions were computed across delays for each set size, yielding the following estimates: r2ss1 = 0.60; r2ss4 = 0.87; r2ss10 = 0.95. Note that lower explained variance for set size 1 arises from both data and model predictions having near-constant precision.

In Experiment 2, we calculated r2 between observations and predictions across presentation durations, separately for each set size, resulting in the following estimates: r2ss1 = 0.88; r2ss4 = 0.71; r2ss10 = 0.70. Note that in this case the decreasing percentage of explained variance with set size is a consequence of having less variability in both data and model predictions with larger set sizes.

While these estimates suggest that the DyNR model effectively fits the psychophysical data, a more rigorous validation approach would involve cross-validation checks across all conditions with a withheld portion of trials. Regrettably, due to the large number of conditions in each experiment, we could only collect 50 trials per condition. We are sceptical that fitting the model to even fewer trials, as necessary for cross-validation, would provide a reliable assessment of model performance.

Minor: It isn't clear to me why the behavioral tasks are shown in Figure 6. They are important for understanding the results and are discussed earlier in the manuscript (before Figure 3). This just required flipping back and forth to understand the task before I could interpret the results.

Thank you for this comment. We have now moved the behavioural task figure to appear early in the manuscript (as Figure 3).

Reviewer #3 (Recommendations For The Authors):

(1) Dynamics of sensory signals during perception

I believe that the modeled sensory signal is a reasonable simplification and different ways to model the decay function are discussed. I would like to ask the authors to discuss the implications of slightly more complex initial sensory transients such as the ones shown in Teeuwen (2021). Specifically for short exposure times, this might be particularly relevant for the model fits as some of the alternative models diverge from the data for short exposures. In addition, the role of feedforward (initial transient?) and feedback signaling (subsequent "plateau" activity) could be discussed. The first one might relate more strongly to sensory signals whereas the latter relates more to top-down attention/recurrent processing/VWM.

Particularly, this latter response might also be sensitive to the number of items present on the screen which leads to a related question pertaining to the limitations of attention during perception. Some work suggests that perception is similarly limited in the amount of information that can be represented concurrently (Tsubomi, 2013). Could the authors discuss the implications of this hypothesis? What happens if maximum sensory amplitude is set as a free parameter in the model?

Tsubomi, H., Fukuda, K., Watanabe, K., & Vogel, E. K. (2013). Neural limits to representing objects still within view. Journal of Neuroscience, 33(19), 8257-8263.

Thank you for this question. Below, we unpack it and answer it point by point.

While we agree our model of the sensory response is justified as an idealization of the biological reality, we also recognise that recent electrophysiological recordings have illuminated intricacies of neuronal responses within the striate cortex, a critical neural region associated with sensory memory (Teeuwen et al, 2021). Notably, these recordings reveal a more nuanced pattern where neurons exhibit an initial burst of activity succeeded by a lower plateau in firing rate, and stimulus offset elicits a second small burst in the response of some neurons, followed by a gradual decrease in activity after the stimulus disappears (Teeuwen et al, 2021).

In general, asynchronous bursts of activity in individual neurons will tend to average out in the population making little difference to predictions of the DyNR model. Synchronized bursts at stimulus onset could affect predictions for the shortest presentations in Exp 2, however the model appears to capture the data very well without including them. We would be wary of incorporating these phenomena into the model without more clarity on their universality (e.g., how stimulus-dependent they are), their significance at the population level (as opposed to individual neurons), and most importantly, their prominence in visual areas outside striate cortex. Specifically, while Teeuwen et al. (2021) described activity in V1, our model does not make strong assumptions about which visual areas are the source of the sensory input to WM. Based on these uncertainties we believe the idealized sensory response is justified for use in our model.

Next, thank you for the comment on feedforward and feedback signals. We have added the following to our manuscript:

“Following onset of a stimulus, the visual signal ascends through visual areas via a cascade of feedforward connections. This feedforward sweep conveys sensory information that persists during stimulus presentation and briefly after it disappears (Lamme et al., 1998). Simultaneously, reciprocal feedback connections carry higher-order information back towards antecedent cortical areas (Lamme and Roelfsema, 2000). In our psychophysical task, feedback connections likely play a critical role in orienting attention towards the cued item, facilitating the extraction of persisting sensory signals, and potentially signalling continuous information on the available resources for VWM encoding. While our computational study does not address the nature of these feedforward and feedback signals, a challenge for future research is to describe the relative contributions of these signals in mediating transmission of information between sensory and working memory (Semedo et al., 2022).”

Lamme, V. A., Supèr, H., & Spekreijse, H. (1998). Feedforward, horizontal, and feedback processing in the visual cortex. Current Opinion in Neurobiology, 8(4), 529–535. https://doi.org/10.1016/S0959-4388(98)80042-1

Lamme, V. A. F., & Roelfsema, P. R. (2000). The distinct modes of vision offered by feedforward and recurrent processing. Trends in Neurosciences, 23(11), 571–579. https://doi.org/10.1016/S0166-2236(00)01657-X

Semedo, J. D., Jasper, A. I., Zandvakili, A., Krishna, A., Aschner, A., Machens, C. K., Kohn, A., & Yu, B. M. (2022). Feedforward and feedback interactions between visual cortical areas use different population activity patterns. Nature Communications, 13(1), 1099. https://doi.org/10.1038/s41467-022-28552-w

Finally, both you and Reviewer 2 raised a similar interesting question regarding capacity limitations of attention during perception Such a limitation could be modelled by freely estimating sensory amplitude and implementing divisive normalization to that signal, similar to how VWM is constrained. We can consider two potential mechanisms through which divisive normalization might be incorporated into sensory processing within the DyNR model.

The first possibility involves assuming that normalization is pre-attentive. In this scenario, the sensory activity of each object would be rescaled at the lowest level of sensory processing, occurring before the allocation of attentional or VWM resources. One strong prediction of such an implementation is that recall error in the simultaneous cue condition (Experiment 1) should vary with set size. However, this prediction is inconsistent with the observed data, which failed to show a significant difference between set sizes, and is more closely aligned with the hypothesis of no-difference (F(2,18) = 1.26, p = .3, η2 = .04, BF10 = 0.47). On that basis, we anticipate that introducing normalization as a pre-attentive mechanism would impair the model fit.

An alternative scenario is to consider normalization as post-attentive. In the simultaneous cueing condition, only one item is attended (i.e., the cued one), regardless of the displayed set size. Here, we would expect normalized activity for a single item, regardless of the number of presented objects, which would then be integrated into VWM. This expanded DyNR model with post-attentive normalization would make exactly the same predictions as the proposed DyNR for recall fidelity, so distinguishing between these models would not be possible based on working memory experiments.

To acknowledge the possibility that sensory signals could undergo divisive normalization and to motivate future research, we have added the following to our manuscript:

“As well as being implicated in higher cognitive processes including VWM (Buschman et al, 2011; Sprague et al., 2014), divisive normalization has been shown to be widespread in basic sensory processing (Bonin et al., 2005; Busse et al., 2009; Ni et al., 2017). The DyNR model presently incorporates the former but not the latter type of normalization. While the data observed in our experiments do not provide evidence for normalization of sensory signals (note comparable recall errors across set size in the simultaneous cue condition of Experiment 1), this may be because sensory suppressive effects are localized and our stimuli were relatively widely separated in the visual field: future research could explore the consequences of sensory normalization for recall from VWM using, e.g., centre-surround stimuli (Bloem et al., 2018).”

Bloem, I. M., Watanabe, Y. L., Kibbe, M. M., & Ling, S. (2018). Visual Memories Bypass Normalization. Psychological Science, 29(5), 845–856. https://doi.org/10.1177/0956797617747091

Bonin, V., Mante, V., & Carandini, M. (2005). The Suppressive Field of Neurons in Lateral Geniculate Nucleus. The Journal of Neuroscience, 25(47), 10844–10856. https://doi.org/10.1523/JNEUROSCI.3562-05.2005

Buschman, T. J., Siegel, M., Roy, J. E., & Miller, E. K. (2011). Neural substrates of cognitive capacity limitations. Proceedings of the National Academy of Sciences, 108(27), 11252–11255. https://doi.org/10.1073/pnas.1104666108

Busse, L., Wade, A. R., & Carandini, M. (2009). Representation of Concurrent Stimuli by Population Activity in Visual Cortex. Neuron, 64(6), 931–942. https://doi.org/10.1016/j.neuron.2009.11.004

Ni, A. M., & Maunsell, J. H. R. (2017). Spatially tuned normalization explains attention modulation variance within neurons. Journal of Neurophysiology, 118(3), 1903–1913. https://doi.org/10.1152/jn.00218.2017

Sprague, T. C., Ester, E. F., & Serences, J. T. (2014). Reconstructions of Information in Visual Spatial Working Memory Degrade with Memory Load. Current Biology, 24(18), 2174–2180. https://doi.org/10.1016/j.cub.2014.07.066

(2) Effectivity of retro-cues at long delays

Can the authors discuss how cues presented at long delays (>1000 ms) can still lead to increased memory fidelity when sensory signals are likely to have decayed? A list of experimental work demonstrating this can be found in Souza & Oberauer (2016).

Souza, A. S., & Oberauer, K. (2016). In search of the focus of attention in working memory: 13 years of the retro-cue effect. Attention, Perception, & Psychophysics, 78, 1839-1860.

The increased memory fidelity observed with longer delays between memory array offset and cue does not result from integrating available sensory signals into VWM because the sensory signal would have completely decayed by that time. Instead, research so far has indicated several alternative mechanisms that could lead to higher recall precision for cued items, and we can briefly summarize some of them, which are also reviewed in more detail in Souza and Oberauer (2016).

One possibility is that, after a highly predictive retro-cue indicates the to-be-tested item, uncued items can simply be removed from VWM. This could result in decreased interference for the cued item, and consequently higher recall precision. Secondly, the retro-cue could also indicate which item can be selectively attended to, and thereby differentially strengthening it in memory. Furthermore, the retro-cue could allow evidence to accumulate for the target item ahead of decision-making, and this could increase the probability that the correct information will be selected for response. Finally, the retro-cued stimulus could be insulated from interference by subsequent visual input, while the uncued stimuli may remain prone to such interference.

A neural account of this retro-cue effect based on the original neural resource model has been proposed in Bays & Taylor, Cog Psych, 2018. However, as we did not use a retro-cue design in the present experiments, we have decided not to elaborate on this in the manuscript.

(3) Swap errors

I am somewhat surprised by the empirically observed and predicted pattern of swap errors displayed in Figure S2. For set size 10, swap probability does not consistently increase with the duration of the retention interval, although this was predicted by the author's model. At long intervals, swap probability is significantly higher for large compared to small set sizes, which also seems to contrast with the idea of shared, limited VWM resources. Can the authors provide some insight into why the model fails to reproduce part of the behavioral pattern for swap errors? The sentence in line 602 might also need some reconsideration in this regard.

Determining the ground truth for swap errors poses a challenge. The prevailing approach has been to employ a simpler model that estimates swap errors, such as a three-component mixture model, and use those estimates as a proxy for ground truth. However, this method is not without its shortcomings. For example, the variability of swap frequency estimates tends to increase with variability in the report feature dimension (here, orientation). This is due to the increasing overlap of response probability distributions for swap and non-swap responses. Consequently, the discrepancy between any two methods of swap estimation is most noticeable when there is substantial variability in orientation reports (e.g., 10 items and long delay or short exposure).

When modelling swap frequency in the DyNR model, our aim was to provide a parsimonious account of swap errors while implementing similar dynamics in the spatial (cue) feature as in the orientation (report) feature. This parametric description captured the overall pattern of swap frequency with set size and retention and encoding time, but is still only an approximation of the predictions if we fully modelled memory for the conjunction of cue and report features (as in e.g. Schneegans & Bays, 2017; McMaster et al, 2020).

We expanded the existing text in the section ‘Representational dynamics of cue-dimension features’ of our manuscript:

“… Although we did not explicitly model the neural signals representing location, the modelled dynamics in the probability of swap errors were consistent with those of the primary memory feature. We provided a more detailed neural account of swap errors in our earlier works that is theoretically compatible with the DyNR model (McMaster et al., 2020; Schneegans & Bays, 2017).

The DyNR model successfully captured the observed pattern of swap frequencies (intrusion errors). The only notable discrepancy between DyNR and the three-component mixture model (Fig. S2) arises with the largest set size and longest delay, although with considerable interindividual variability. As the variability in report-dimension increases, the estimates of swap frequency become more variable due to the growing overlap between the probability distributions of swap and non-swap responses. This may explain apparent deviations from the modelled swap frequencies with the highest set size and longest delay where orientation response variability was greatest. “

McMaster, J. M. V., Tomić, I., Schneegans, S., & Bays, P. M. (2022). Swap errors in visual working memory are fully explained by cue-feature variability. Cognitive Psychology, 137, 101493. https://doi.org/10.1016/j.cogpsych.2022.101493

Schneegans, S., & Bays, P. M. (2017). Neural Architecture for Feature Binding in Visual Working Memory. The Journal of Neuroscience, 37(14), 3913–3925. https://doi.org/10.1523/JNEUROSCI.3493-16.2017

(4) Direct sensory readout

The model assumes that readout from sensory memory and from VWM happens with identical efficiency. Currently, we don't know if these two systems are highly overlapping or are fundamentally different in terms of architecture and computation. In the case of the latter, it might be less reasonable to assume that information readout would happen at similar efficiencies, as it is currently assumed in the manuscript. Perhaps the authors could briefly discuss this possibility.

In the direct sensory read-out model, we did not explicitly model the efficiency of readout from either sensory or VWM store. However, the distinctive prediction of this model is that the precision of recall changes exponentially with delay at every set size, including one item. This prediction does not depend on the relative efficiency of readout from sensory and working memory, but only on the principle that direct readout from sensory memory bypasses the capacity limit on working memory. This prediction is inconsistent with the pattern of results observed in Experiment 1, where early cues did not show a beneficial effect on recall error for set size 1. While the proposal raised by the reviewer is intriguing, even if we were to model the process of readout from both the sensory and VWM stores with different efficiencies, the direct read-out model could not account for the near-constant recall error with delay for set size one.

(5) Encoding of distractors

One of the model assumptions is that, for simultaneous presentations of memory array and cue only the cued feature will be encoded. Previous work has suggested that participants often accidentally encode distractors even when they are cued before memory array onset (Vogel 2005). Given these findings, how reasonable is this assumption in the authors' model?

Vogel, E. K., McCollough, A. W., & Machizawa, M. G. (2005). Neural measures reveal individual differences in controlling access to working memory. Nature, 438(7067), 500-503.

Although previous research suggested that observers can misinterpret the pre-cue and encode one of the uncued items, our results argue against this being the case in the current experiment. Such encoding failures would manifest in overall recall error, resulting in a gradient of error with set size, owing to the presence of more adjacent distractors in larger set sizes. However, when we compared recall errors between set sizes in the simultaneous cue condition, we did not find a significant difference between set sizes, and moreover, our results were more likely under the hypothesis of no-difference (F(2,18) = 1.26, p = .3, η2 = .04, BF10 = 0.47). If observers occasionally encoded and reported one of the uncued items in the simultaneous cue condition, those errors were extremely infrequent and did not affect the overall error distributions.

eLife assessment

This study presents important insights into the dynamical process whereby sensory information is converted from stimulus-driven activity to a working memory representation from which the information can be recalled later. The evidence supporting the claims is convincing, using detailed fits and model-comparison techniques applied to new and existing psychophysical data sets to evaluate a wide variety of potential mechanisms. The overall conclusion, that iconic memory and working memory are not distinct mechanisms but rather two slightly different regimes of the same circuitry, will be of interest to neuroscientists and psychologists studying sensory systems and/or working memory.

Reviewer #2 (Public Review):

Summary:

Previous work has shown subjects can use a form of short-term sensory memory, known as 'iconic memory', to accurately remember stimuli over short periods of time (several hundred milliseconds). Working memory maintains representations for longer periods of time but is strictly limited in its capacity. While it has long been assumed that sensory information acts as the input to working memory, a process model of this transfer has been missing. To address this, Tomic and Bays present the Dynamic Neural Resource (DyNR) model. The DyNR model captures the dynamics of processing sensory stimuli, transferring that representation into working memory, the diffusion of representations within working memory, and the selection of a memory for report.

The DyNR model captures many of the effects observed in behavior. Most importantly, psychophysical experiments found the greater the delay between stimulus presentation and the selection of an item from working memory, the greater the error. This effect also depended on working memory load. In the model, these effects are captured by the exponential decay of sensory representations (i.e., iconic memory) following the offset of the stimulus. Once the selection cue is presented, residual information in iconic memory can be integrated into working memory, improving the strength of the representation and reducing error. This selection process takes time, and is slower for larger memory loads, explaining the observation that memory seems to decay instantly. The authors compare the DyNR model to several variants, demonstrating the importance of persistence of sensory information in iconic memory, normalization of representations with increasing memory load, and diffusion of memories over time.

Strengths:

The manuscript provides a convincing argument for the interaction of iconic memory and working memory, as captured by the DyNR model. The analyses are cutting-edge and the results are well captured by the DyNR model. In particular, a strength of the manuscript is the comparison of the DyNR model to several alternative variants.

The results provide a process model for interactions between iconic memory and working memory. This will be of interest to neuroscientists and psychologists studying working memory. I see this work as providing a foundation for understanding behavior in continuous working memory tasks, from either a mechanistic, neuroscience, perspective or as a high-water mark for comparison to other psychological process models.

Finally, the manuscript is well written. The DyNR model is nicely described and an intuition for the dynamics of the model are clearly shown in Figures 2 and 4.

Weaknesses:

The manuscript appropriately acknowledges and addresses several minor weaknesses that are due to the limited ability of the approach to disambiguate underlying neural mechanisms. Nevertheless, the manuscript adds significant value to the literature on working memory.

Reviewer #3 (Public Review):

Summary

The authors set out to formally contrast several theoretical models of working memory, being particularly interested in comparing the models regarding their ability to explain cueing effects at short cue durations. These benefits are traditionally attributed to the existence of a high capacity, rapidly decaying sensory storage which can be directly read out following short latency retro-cues. Based on the model fits, the authors alternatively suggest that cue-benefits arise from a freeing of working memory resources, which at short cue latencies can be utilized to encode additional sensory information into VWM.

A dynamic neural population model consisting of separate sensory and VWM populations was used to explain temporal VWM fidelity of human behavioral data collected during several working memory tasks. VWM fidelity was probed at several timepoints during encoding, while sensory information was available and maintenance, when sensory information was no longer available. Furthermore, set size and exposure durations were manipulated to disentangle contributions of sensory and visual working memory.

Overall, the model explained human memory fidelity well, accounting for set size, exposure time, retention time, error distributions and swap errors. Crucially the model suggests that recall at short delays is due to post-cue integration of sensory information into VWM as opposed to direct readout from sensory memory. The authors formally address several alternative theories, demonstrating that models with reduced sensory persistence, direct readout from sensory memory, no set-size dependent delays in cue processing and constant accumulation rate provide significantly worse fits to the data.

I congratulate the authors for this rigorous scientific work. All my remarks were thoroughly addressed.

Author Response

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Zhang et al., investigated the relationship between monocular and binocular responses of V1 superficial-layer neurons using two-photon calcium imaging. They found a strong relationship in their data: neurons that exhibited a greater preference for one eye or the other (high ocular dominance) were more likely to be suppressed under binocular stimulation, whereas neurons that are more equivalently driven by each other (low ocular dominance) were more likely to be enhanced by binocular stimulation. This result chiefly demonstrates the relationship between ocular dominance and binocular responses in V1, corroborating what has been shown previously using electrophysiological techniques but now with greater spatial resolution (albeit less temporal resolution). The binocular responses were well-fitted by a model that institutes divisive normalization between the eyes that accounts for both the suppression and enhancement phenomena observed in the subpopulation of binocular neurons. In so doing, the authors reify the importance of incorporating ocular dominance in computational models of binocular combination.

The conclusions of this paper are mostly well supported by the data, but there are some limitations of the methodology that need to be clarified, and an expansion of how the results relate to previous work would better contextualize these important findings in the literature.

Strengths:

The two-photon imaging technique used to resolve the activity of individual neurons within intact brain tissue grants a host of advantages. Foremost, two-photon imaging confers considerably high spatial resolution. As a result, the authors were able to sample and analyze the activity from thousands of verified superficial-layer V1 neurons. The animal model used, awake macaques, is also highly relevant for the study of binocular combination. Macaques, like humans, are binocular animals, meaning they have forward-facing eyes that confer overlapping visual fields. Importantly, macaque V1 is organized into cortical columns that process specific visual features from the separate eyes just like in humans. In combination with a powerful imaging technique, this allowed the authors to evaluate the monocular and binocular response profiles of V1 neurons that are situated within neighboring ocular dominance columns, a novel feat. To this aim, the approach was well-executed and should instill further confidence in the notion that V1 neurons combine monocular information in a manner that is dependent on the strength of their ocular dominance.

Weaknesses:

While two-photon imaging provides excellent spatial resolution, its temporal resolution is often lower compared to some other techniques, such as electrophysiology. This limits the ability to study the fast dynamics of neuronal activity, a well-understood trade-off of the method. The issue is more so that the authors draw comparisons to electrophysiological studies without explicit appreciation of the temporal difference between these techniques. In a similar vein, two-photon imaging is limited spatially in terms of cortical depth, preferentially sampling from neurons in layers 2/3. This limitation does not invalidate any of the interpretations but should be considered by readers, especially when making comparisons to previous electrophysiological reports using microelectrode linear arrays that sample from all cortical layers. Indeed, it is likely that a complete picture of early cortical binocular processing will require high spatial resolution (i.e., sampling from neurons in neighboring ocular dominance columns, from pia mater to white matter) at the biophysically relevant timescales (1ms resolution, capturing response dynamics over the full duration of the stimulus presentation, including the transient onset and steady-state periods).

To address the same concern from all three reviewers, we discussed the technical limitations of two photon calcium imaging at the end of Discussion, including limited imaging depth, low temporal resolution, and nonlinearity. The relevant texts are copied here:

(Ln 304) “Limitations of the current study

Although capable of sampling a large number of neurons at cellular resolution and with low sampling bias, two-photon calcium imaging has its known limitations that may better make it a complementary research tool to electrophysiological recordings.

For example, two-photon imaging can only sample neurons from superficial-layers, while binocular neurons also exist in deeper layers, and even neurons in the input layer are affected by feedback from downstream binocular neurons to exhibit binocular response properties (Dougherty, Cox, Westerberg, & Maier, 2019). Furthermore, calcium signals are relatively slow and cannot reveal the fast dynamics of neuronal responses. Due to these spatial and temporal limitations, a more complete picture of the neuronal mechanisms underlying binocular combination of monocular responses may come from studies using both technologies.

In addition, calcium signals may exaggerate the nonlinear properties of neurons. Although calcium signals indicated by GCaMP5, our favored choice of calcium indicator, displays a linear relationship to neuronal spike rates within a range of 10-150 Hz (Li, Liu, Jiang, Lee, & Tang, 2017), weak and strong signals out of this range are more nonlinear, and may appear poorer and stronger, respectively, than electrode-recorded effects. Consequently, the differences in population responses between monocular and binocular stimulations revealed by this study might be less pronounced.”

(Recommendations For The Authors):

Overall, my main suggestion for the authors to improve the paper is to revise some of the interpretations of their results in relation to previous research. The purpose of the present study was to illustrate a more complete picture of the binocular combination of monocular responses by taking into consideration the ocular dominance of V1 cells (lines 34-36). A study published earlier this year had an identical purpose (Mitchell et al., Current Biology, 2023) and arrived at a highly similar conclusion (and also applied divisive normalization to fit their data). I would ask that this paper be mentioned in the introduction and discussed.

The Mitchell et al 2023 paper is added to the Introduction and Discussion:

(Ln 50) “In addition (to the Dougherty et al 2019 paper from the same group), Mitchell, Carlson, Westerberg, Cox, and Maier (2023) reported that binocular combination of monocular stimuli with different contrasts is also affected by neurons’ eye preference.”

(Ln 286) “The critical roles of ocular dominance have been largely overlooked by extant binocular vision models to our knowledge, except that Anderson and Movshon (1989) demonstrated that a model consisting of multiple ocular dominance channels can better explain their psychophysical adaptation data, and that Mitchell et al. (2023) revealed that binocular combination of different contrasts presented to different eyes are affected by neurons’ ocularity preference.”

Nevertheless, the results of the present study are very valuable. They add substantial spatial resolution and sophisticated relational analysis of monocular and binocular responses that Mitchell et al., 2023 did not include. Therefore, my suggestion is to emphasize the advantages of two-photon imaging in the introduction, focusing on the ability to image neurons in neighboring ocular dominance columns. The rigorous modeling of the relationship between nearby neurons with a range of eye preferences, in tandem with the incredible yield of two-photon imaging, is what sets this paper apart from previous electrophysiological work.

The finding that binocular responses were dependent on ocular dominance is largely consistent with previous electrophysiological results. However, there should be a paragraph in the discussion section that speaks to the limitations of comparing two-photon imaging data to electrophysiological data. Namely, there are two limitations:

(1) These two techniques confer different temporal resolutions. It is conceivable that some of the electrophysiology relationships (for example, described by Dougherty et al., 2019) may be dependent on the temporal window over which the data was averaged, typically over 50-100ms around stimulus onset, or 100-250ms comprising the neurons' sustained response to the stimulus. This possible explanation of the difference in obtained results would be especially useful for the discussion paragraph starting at line 232. It would also be helpful to readers for there to be some mention of the advantage of having high temporal resolution (i.e., the benefits of electrophysiology) since (a) recent work has distinguished between sequential stages of binocular combination (Cox et al., 2019) and (b) modern models of V1 neurons emphasize recurrent feedback to explain V1 temporal dynamics (see Heeger et al., 2019; Rubin et al., 2015), which could prove to be relevant for combination of stimuli in the two eyes (Fleet et al., 1997).

Our discussion regarding the technical limitations of 2-p calcium imaging has been listed earlier. Specific to the Dougherty et 2019 paper, we added the following discussion to address the issue of temporal resolution difference between two technologies.

(Ln 266) “In addition, it is unclear whether the discrepancies are caused by different temporal resolutions of electrode recording and calcium imaging. The results of Dougherty et al. (2019) represent changes of neuronal spike activities over a period of approximately 50-200 ms after the stimulus onset, which may reflect the sustained neuronal responses to the stimulus and possible feedback signals. Calcium signals are much slower and indicative of the aggregated neuronal responses over a longer period (up to 1000 ms in the current study). They should have smeared, rather than exaggerated, the differences between monocular and binocular responses, although we cannot exclude the possibility that some neuronal response changes beyond 200 ms are responsible for the discrepancies.”

(2) The sample of V1 neurons in this study is limited to cells in the most superficial layers of the cortex (layers 2/3). This limitation is, of course, well understood, but it should be mentioned at least in the context of studying the formative mechanisms of binocular combination in V1 (since we know that binocular neurons also exist in layers 5/6, and there is now substantial evidence that even layer 4 neurons are not as "monocular" as we previously thought (Dougherty et al., 2019)).

See our discussion regarding the technical limitations of 2-p calcium imaging listed earlier.

In short, I believe the paper would be improved by (1) adding the above citations in the appropriate places, (2) acknowledging in the introduction that this question has been investigated electrophysiologically but emphasizing the advantages of two-photon imaging, and (3) adding a paragraph to the discussion section that discusses the temporal and spatial limitations when using two-photon imaging to study binocular combination, particularly when comparing the results to electrophysiology.

Reviewer #2 (Public Review):

Summary:

This study examines the pattern of responses produced by the combination of left-eye and right-eye signals in V1. For this, they used calcium imaging of neurons in V1 of awake, fixating monkeys. They take advantage of calcium imaging, which yields large populations of neurons in each field of view. With their data set, they observe how response magnitude relates to ocular dominance across the entire population. They analyze carefully how the relationship changed as the visual stimulus switched from contra-eye only, ipsi-eye only, and binocular. As expected, the contra-eye-dominated neurons responded strongly with a contra-eye-only stimulus. The ipsi-eye-dominated neurons responded strongly with an ipsi-eye-only stimulus. The surprise was responses to a binocular stimulus. The responses were similarly weak across the entire population, regardless of each neuron's ocular dominance. They conclude that this pattern of responses could be explained by interocular divisive normalization, followed by binocular summation.

Strengths:

A major strength of this work is that the model-fitting was done on a large population of simultaneously recorded neurons. This approach is an advancement over previous work, which did model-fitting on individual neurons. The fitted model in the manuscript represents the pattern observed across the large population in V1, and washes out any particular property of individual neurons. Given the large neuronal population from which the conclusion was drawn, the authors provide solid evidence supporting their conclusion. They also observed consistency across 5 fields of view.

The experiments were designed and executed appropriately to test their hypothesis. Their data support their conclusion.

Weaknesses:

One weakness of their study is that calcium signals can exaggerate the nonlinear properties of neurons. Calcium imaging renders poor responses poorer and strong responses stronger, compared to single-unit recording. In particular, the dramatic change in the population response between monocular stimulation and binocular stimulation could actually be less pronounced when measured with single-unit recording methods. This means their choice of recording method could have accidentally exaggerated the evidence of their finding.

We discussed the nonlinearity of calcium signals as part of the technical limitations of 2-p imaging calcium. The calcium indicator we use, GCaMP5, has a reasonable range of linear relationship with spike rates. But out of this range, the nonlinearity is indeed a concern.

(Ln 314) “In addition, calcium signals may exaggerate the nonlinear properties of neurons. Although signals indicated by GCaMP5, our favored choice of calcium indicator, displays a linear relationship to neuronal spike rate within a range of 10-150 Hz (Li et al., 2017), weak and strong signals out of this range are more nonlinear, and may appear poorer and stronger, respectively, than electrode-recorded effects. Consequently, the changes in population responses between monocular and binocular stimulations revealed by this study might be less pronounced.”

The implication of their finding is that strong ocular dominance is the result of release from interocular suppression by a monocular stimulus, rather than the lack of binocular combination as many traditional studies have assumed. This could significantly advance our understanding of the binocular combination circuitry of V1. The entire population of neurons could be part of a binocular combination circuitry present in V1.

This is a very good insight. We added the following sentences to the end of the first paragraph of Discussion:

(Ln 242) “These findings implicate that at least for neurons in superficial layers of V1, significant ocular dominance may result from a release of interocular suppression during monocular stimulation, an unusual viewing condition as our vision is typically binocular, rather than a lack of binocular combination of inputs from upstream monocular neurons.”

(Recommendations For The Authors):

Line 150: "To model interocular response suppression, responses from each eye in Eq. 2 were further normalized by an interocular suppression factor wib or wcb," I recommend the authors improve their explanation of how they arrived at Eq. 3 from Eq. 2. As it stands, my impression is that they have one model for the responses to monocular stimulation, and another model for the responses to binocular stimulation. What I think is missing is that both equations are derived from the same model. Monocular stimulation is a situation in which the stimulus in one eye's contrast is zero. Could the authors clarify whether this situation produces an interocular suppression of zero, and how that leads to Eq. 2?

We rewrote the modeling part to show that Equations 1-3 are sequential steps of development for the same model. We also added a brief paragraph to discuss how Eq. 3 could lead to Eq. 2 under monocular viewing:

(Ln 166) “Although not shown in Eq. 3, we also assumed that the nonlinear exponent b also depends on the contrast of the stimulus presented to the other eye (i.e., Sc or Si). Consequently, when Sc or Si = 0 under monocular stimulation, Rc or Ri = 0 (Eq. 1), and interocular suppression wib or wcb = 1, so Eq. 3 changes back to Eq. 2. It is only when Sc and Si are equal and close to 1, as in the current study, that interocular suppression and binocular combination would be in the current Eq. 3 format.”

Line 225: "However, individually, compared to monocular responses, responses of monocular neurons more preferring the stimulated eye are actually suppressed, and only responses of binocular neurons are increased by binocular stimulation." This sentence is difficult to follow. I recommend the authors improve clarity by breaking up the sentence into several sentences. If I understand correctly, they summarize the pattern in the data that is indicative of interocular divisive normalization, i.e., their final conclusion.

This sentence no longer exists in the Discussion.

Line 426: "Third, for those showing significant orientation difference, the trial-based orientation responses of each neuron were fitted with a Gaussian model with a MATLAB nonlinear least squares function:" The choice of using a Gaussian function to fit orientation tuning was probably suboptimal. A Gaussian function provides an adequate fit only for neurons whose tuning is very sharp. The responses outside of the peak fall down to the baseline and the two ends meet. Otherwise, the two ends do not meet. An adequate fit would be achieved with a function of a circular variable, which wraps around 180 deg. I recommend using a Von Mises function for fitting orientation tuning.

We agree with the reviewer that the Von Mises function is more accurate than Gaussian for fitting orientation tuning functions. Indeed we are using it to fit orientation tuning of V4 neurons, many of which have two peaks. For the current V1 data, the differences between Von Mises and Gaussian fittings are very small, as shown in the orientation functional maps from three macaques below. Because we also use the same Gaussian fitting of orientation tuning in several published and current under-review papers, we prefer to keep the Gaussian fitting results in the manuscript.

Author response image 1.

Reviewer #3 (Public Review):

The authors have made simultaneous recordings of the responses of large numbers of neurons from the primary visual cortex using optical two-photon imaging of calcium signals from the superficial layers of the cortex. Recordings were made to compare the responses of the cortical neurons under normal binocular viewing of a flat screen with both eyes open and monocular viewing of the same screen with one eye's view blocked by a translucent filter. The screen displayed visual stimuli comprising small contrast patches of Gabor function distributions of luminance, a stimulus that is known to excite cortical neurons.

This is an important data set, given the large numbers of neurons recorded. The authors present a simple model to explain the binocular combination of neuronal signals from the right and left eyes.

The limitations of the paper as written are as follows. These points can be addressed with some additional analysis and rewriting of sections of the paper. No new experimental data need to be collected.

(1) The authors should acknowledge the fact that these recordings arise from neurons in the superficial layers of the cortex. This limitation arises from the usual constraints on optical imaging in the macaque cortex. This means that the sample of neurons forming this data set is not fully representative of the population of binocular neurons within the visual cortex. This limitation is important in comparing the outcome of these experiments with the results from other studies of binocular combination, which have used single-electrode recording. Electrode recording will result in a sample of neurons that is drawn from many layers of the cortex, rather than just the superficial layers.

See our discussion regarding the technical limitations of 2-p calcium imaging listed earlier.

(2) Single-neuron recording of binocular neurons in the primary visual cortex has shown that these neurons often have some spontaneous activity. Assessment of this spontaneous level of firing is important for accurate model fitting [1]. The paper here should discuss the level of spontaneous neuronal firing and its potential significance.

We have noticed previously that at non-optimal spatial frequencies, calcium responses to a moving Gabor grating are close to zero (Guan et al., Prog Neurobiology, 2021, Fig. 1B), but we cannot tell whether this is due to calcium response nonlinearity, or a close-to-zero level of spontaneous neuronal activity. Prince et al (2002) reported low spontaneous responses of V1 neurons with moving grating stimuli (e.g., about 3 spikes/sec in one exemplar neuron, their Fig. 1B), so this appears not a big effect. In our data fitting, we do have an orientation-unspecific component in the Gaussian model, which represents the neuronal response at a non-preferred orientation, but not necessarily the spontaneous activity.

(3) The arrangements for visual stimulation and comparison of binocular and monocular responses mean that the stereoscopic disparity of the binocular stimuli is always at zero or close to zero. The animal's fixation point is in the centre of a single display that is viewed binocularly. The fixation point is, by definition, at zero disparity. The other points on the flat display are also at zero disparity or very close to zero because they lie in the same depth plane. There will be some small deviations from exactly zero because the geometry of the viewing arrangements results in the extremities of the display being at a slightly different distance than the centre. Therefore, the visual stimulation used to test the binocular condition is always at zero disparity, with a slight deviation from zero at the edges of the display, and never changes. [There is a detail that can be ignored. The experimenters tested neurons with visual stimulation at different real distances from the eyes, but this is not relevant here. Provided the animals accurately converged their eyes on the provided binocular fixation point, then the disparity of the visual stimuli will always be at or close to zero, regardless of viewing distance in these circumstances.] However, we already know from earlier work that neurons in the visual cortex exhibit a range of selectivity for binocular disparity. Some neurons have their peak response at non-zero disparities, representing binocular depths nearer than the fixation depth or beyond it. The response of other neurons is maximally suppressed by disparities at the depth of the fixation point (so-called Tuned Inhibitory [TI] neurons). The simple model and analysis presented in the paper for the summation of monocular responses to predict binocular responses will perform adequately for neurons that are tuned to zero disparity, so-called tuned excitatory neurons [TE], but is necessarily compromised when applied to neurons that have other, different tuning profiles. Specifically, when neurons are stimulated binocularly with a non-preferred disparity, the binocular response may be lower than the monocular response[2, 3]. This more realistic view of binocular responses needs to be considered by the authors and integrated into their modelling.

We agree and include the following texts when discussing the future work:

(Ln 298) “In addition, in our experiments, binocular stimuli were presented with zero disparity, which best triggered the responses of neurons with zero-disparity tuning. A more realistic model of binocular combination also requires the consideration of neurons with other disparity-tuning profiles.”

(4) The data in the paper show some features that have been reported before but are not captured by the model. Notably for neurons with extreme values of ocular dominance, the binocular response is typically less than the larger of the two monocular responses. This is apparent in the row of plots in Figure 2D from individual animals and in the pooled data in Figure 2E. Responses of this type are characteristic of tuned inhibitory [TI] neurons[2]. It is not immediately clear why this feature of the data does not appear in the summary and analysis in Figure 3.

This difference is indeed captured by the model, which can be more easily appreciated in Fig. 4A where monocular and binocular model simulations are plotted in the same panel. In the text, we also wrote: (Ln 195) “It is apparent that binocular responses cannot be explained by the sum of monocular responses, as binocular responses are substantially lower than the summed monocular responses for both monocular and binocular neurons. Nor can binocular responses be explained by the responses to the preferred eye, as binocular responses are also lower than those to the preferred eye (the larger of the two monocular responses) for monocular neurons.”

The paper text states that the responses were "first normalized by the median of the binocular responses". This will certainly get rid of this characteristic of the data, but this step needs better justification, or an amendment to the main analysis is needed.

The relevant sentence has been rewritten as “Monocular and binocular data of each FOV/depth, as well as the pooled data, were first normalized by the respective median of the binocular responses of all neurons in the same FOV/depth.” This normalization would render the overall binocular responses to be around unity, for the purpose of facilitating comparisons among all FOV/depth, but it would not affect the overall characteristic of the data.

In the present form, the model and analysis do not appear to fit the data in Figure 2 as accurately as needed.

Thanks for pointing out the problem, as data fitting for FOV C_270 and the pooled data were especially inaccurate. The issue has been mostly fixed when each datum was weighted by its standard deviation (please see the updated Fig. 3).

eLife assessment

Overall, the reviewers found the significance of the work valuable to the field of visual neuroscience, particularly given the large data set and strength of the method used that allowed for spatial analysis of neuronal responses in macaque V1. The evidence was deemed compelling, owing in part to the consistency of responses across animals and the fitness of modeling. The authors have addressed the major comments from reviewers and improved the manuscript through relation to prior literature and addressing specific limitations of the method used.

Reviewer #1 (Public Review):

Summary:

Zhang et al., investigated the relationship between monocular and binocular responses of V1 superficial-layer neurons using two-photon calcium imaging. They found a strong relationship in their data: neurons that exhibited a greater preference for one eye or the other (high ocular dominance) were more likely to be suppressed under binocular stimulation, whereas neurons that are more equivalently driven by each other (low ocular dominance) were more likely to be enhanced by binocular stimulation. This result chiefly demonstrates the relationship between ocular dominance and binocular responses in V1, corroborating what has been shown previously using electrophysiological techniques with now much finer spatial resolution. The binocular responses were well-fitted by a model that institutes divisive normalization between the eyes that accounts for both the suppression and enhancement phenomena observed in the subpopulation of binocular neurons. In so doing, the authors reify the importance of incorporating ocular dominance in computational models of binocular combination.

The conclusions of this paper are well supported by the data. The authors deftly contextualize these important findings in the literature while also acknowledging the limitations of the methodology employed. Future work would do well to combine the spatial power of 2P imaging with the temporal power of electrophysiology to assess ocular dominance-dependent binocular combination across the V1 laminar microcircuit.

Strengths:

The two-photon imaging technique used to resolve the activity of individual neurons within intact brain tissue grants a host of advantages. Foremost, two-photon imaging confers considerably high spatial resolution. As a result, the authors were able to sample and analyze the activity from thousands of verified superficial-layer V1 neurons. The animal model used, awake macaques, is also highly relevant for the study of binocular combination. Macaques, like humans, are binocular animals, meaning they have forward-facing eyes that confer overlapping visual fields. Importantly, macaque V1 is organized into cortical columns that process specific visual features from the separate eyes just like in humans. In combination with a powerful imaging technique, this allowed the authors to evaluate the monocular and binocular response profiles of V1 neurons that are situated within neighboring ocular dominance columns, a novel feat. To this aim, the approach was well-executed and should instill confidence in the notion that V1 neurons combine monocular information in a manner that is dependent on the strength of their ocular dominance.

Weaknesses:

This study suffers no major weaknesses. The authors address the limitations of the methodology and have calibrated the interpretations accordingly.

Reviewer #2 (Public Review):

Summary:

This study examines the pattern of responses produced by the combination of left-eye and right-eye signals in V1. For this, they used calcium imaging of neurons in V1 of awake, fixating monkeys. They take advantage of calcium imaging, which yields large populations of neurons in each field of view. With their data set, they observe how response magnitude relates to ocular dominance across the entire population. They analyze carefully how the relationship changed as the visual stimulus switched from contra-eye only, ipsi-eye only, and binocular. As expected, the contra-eye dominated neurons responded strongly with a contra-eye only stimulus. The ipsi-eye dominated neurons responded strongly with an ipsi-eye only stimulus. The surprise was responses to a binocular stimulus. The responses were similarly weak across the entire population, regardless of each neuron's ocular dominance. They conclude that this pattern of responses could be explained by interocular divisive normalization, followed by binocular summation.

Strengths:

A major strength of this work is that the model-fitting was done on a large population of simultaneously recorded neurons. This approach is an advancement over previous work, which did model-fitting on individual neurons. The fitted model in the manuscript represents the pattern observed across the large population in V1, and washes out any particular property of individual neurons. Given the large neuronal population from which the conclusion was drawn, the authors provide solid evidence supporting their conclusion. They also observed consistency across 5 field of views.

The experiments were designed and executed appropriately to test their hypothesis. Their data support their conclusion.

Weaknesses:

The nonlinear interocular suppression found in this study, could potentially be partially exaggerated by the nonlinear properties of calcium signals. One of the authors of this study has previously reported that the particular GCaMP used in this study has a nice proportional relationship with firing rate of a neuron. So the concern of exaggeration probably does not apply to this particular study. The concern would apply to others who try similar measurements with other versions of GCaMP.

The implication of their finding is that strong ocular dominance is the result of release from interocular suppression by a monocular stimulus, rather than the lack of binocular combination as many traditional studies have assumed. This could significantly advance our understanding of the binocular combination circuitry of V1. The entire population of neurons could be part of a binocular combination circuitry present in V1.

Reviewer #3 (Public Review):

Summary

The authors have made simultaneous recordings of the responses of large numbers of neurons from the primary visual cortex of macaque monkeys using optical two-photon imaging of calcium signals from the superficial layers of the cortex. Recordings were made to compare the responses of the cortical neurons under normal binocular viewing of a flat screen with both eyes open and monocular viewing of the same screen with one eye's view blocked by a translucent filter. The screen displayed visual stimuli comprising small contrast patches of Gabor function distributions of luminance, a stimulus that is known to excite cortical neurons.

Strengths

This is an important data set, given the large number of neurons recorded. The authors present a simple model to explain binocular combination of neuronal signals from the right and left eyes. The work advances the use of two-photon imaging in the cerebral neocortex. The research design adds valuable information to our understanding of the organization of binocular vision in macaque monkeys, which are the only realistic animal model of human vision for the study of binocular interactions.

Limitations and Weaknesses

(1) Given that these recordings are made optically, these results reflect primarily activations of neurons in the superficial layers of the cortex. This limitation arises from the usual constraints (depth of cortex, degree of myelination) on optical imaging in the macaque cortex. This means that the sample of neurons forming this data set is not fully representative of the population of binocular neurons within the visual cortex. This limitation is important in comparing the outcome of these experiments with the results from other studies of binocular combination, which have used single-electrode recording. Electrode recording will result in a sample of neurons that is drawn from many layers of the cortex, rather than just the superficial layers, noting that electrode recordings also carry different risks of sampling bias.

(2) Single neuron recording of binocular neurons in the primary visual cortex has shown that these neurons often have some spontaneous activity. Assessment of this spontaneous level of firing is important for accurate model fitting [1]. The present imaging approach works exclusively with differential measurements of neuronal signals, so assessment of the level of spontaneous activity is not feasible.

(3) The arrangements for visual stimulation and comparison of binocular and monocular responses mean that the stereoscopic disparity of the binocular stimuli is always at zero or close to zero. The consequence is that the experimental design does not test the cortical response over a range of different binocular depths.

The animal's fixation point is in the centre of a single display that is viewed binocularly. The fixation point is, by definition, at zero disparity.. Provided that the animals accurately converged their eyes on the binocular fixation point, then the disparity of the visual stimuli across the whole display will always be at or close to zero. However, we already know from earlier work that neurons in the visual cortex exhibit a range of selectivity for binocular disparity. Some neurons have their peak response at non-zero disparities, representing binocular depths nearer than the fixation depth or beyond it.

There are also other neurons whose response is maximally suppressed by disparities at the depth of the fixation point (so-called Tuned Inhibitory [TI] neurons). The simple model and analysis presented in the paper for the summation of monocular responses to predict binocular responses will perform adequately for neurons that are tuned to zero disparity, so-called tuned excitatory neurons [TE], but is necessarily compromised when applied to neurons that have other, different tuning profiles for binocular disparity. Specifically, when neurons are stimulated binocularly with a non-preferred disparity, the binocular response may be lower than the monocular response [2, 3]. The same limitation applies to another recent paper [4].

This more realistic view of binocular responses needs to be considered further to gain a full picture of the operation of the visual cortex in responding to binocular depth

Citations

1. Prince, S.J.D., Pointon, A.D., Cumming, B.G., and Parker, A.J., (2002). Quantitative analysis of the responses of V1 neurons to horizontal disparity in dynamic random-dot stereograms. Journal of Neurophysiology, 87: 191-208.

2. Prince, S.J.D., Cumming, B.G., and Parker, A.J., (2002). Range and mechanism of encoding of horizontal disparity in macaque V1. Journal of Neurophysiology, 87: 209-221.

3. Poggio, G.F. and Fischer, B., (1977). Binocular interaction and depth sensitivity in striate and prestriate cortex of behaving rhesus monkey. Journal of Neurophysiology, 40: 1392-1405 doi 10.1152/jn.1977.40.6.1392.

4. B. A. Mitchell, K. Dougherty, J. A. Westerberg, B. M. Carlson, L. Daumail, A. Maier, et al. (2022) Stimulating both eyes with matching stimuli enhances V1 responses.<br /> iScience 2022 Vol. 25 Issue 5 DOI: 10.1016/j.isci.2022.104182

Author Response

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In this manuscript, Zeng and Staley provide a valuable analysis of the molecular requirements for the export of a reporter mRNA that contains a lariat structure at its 5' end in the budding yeast S. cerevisiae. The authors provide evidence that this is regulated by the main mRNA export machinery (Yra1, Mex67, Nab2, Npl3, Tom1, and Mlp1). Of note, Mlp1 has been mainly implicated in the nuclear retention of unspliced pre-mRNA (i.e. quality control), and relatively little has been done to investigate its role in mRNA export in budding yeast.

Strengths:

There is relatively little information in the current literature about the nuclear export of splicing intermediates. This paper provides one of the first analyses of this process and dissects the molecular components that promote this form of RNA export. Overall, the strength of the data presented in the manuscript is solid. The paper is well written and the message is clear and of general interest to the mRNA community.

We thank the reviewer for highlighting these strengths.

Weaknesses:

There are three problems with the paper, although these are not major and likely would not affect the final model as most aspects of the molecular details are confirmed by multiple complementary assays.

(1) The brG reporter produces both unspliced pre-mRNA and a lariat-containing intermediate RNA. Based on the primer extension assay the authors claim that only 33% of the final product is in pre-mRNA form and that this "is insufficient to account for the magnitude of the cytoplasmic signal from the brG reporter (83%)". Nevertheless, it is possible that primer extension is incomplete or that the lariat-containing RNA is inaccessible for smFISH. The authors could easily perform a dual smFISH experiment (similar to Adivarahan et l., Molecular Cell 2018) where exon 1 is labelled with probes of one color, and the region that overlaps the lariat-containing intermediate is labelled with probes of a second color. If the authors are correct, then one-third of the smFISH foci should have both labels and the rest would have only the second label. This would also confirm that the latter (i.e. the lariat-containing RNAs) are exported to the cytoplasm. Using this approach, the authors could then show that MLP1-depletion (or depletion of any of the other factors) affect(s) one pool of RNAs (i.e. those that are lariat-containing) but not the other (i.e. pre-mRNA). Including these experiments would make the evidence for their model more convincing.

We appreciate the reviewer’s comments and suggestions. Concerning the primer extension analysis, we are considering alternative assays to quantitate the pre-mRNA and lariat intermediate levels. Concerning the accessibility of the lariat intermediate in smRNA-FISH, in a dbr1∆ strain the only major species from the UAc reporter that is detected by primer extension is the lariat intermediate (Fig. S3), and this reporter is readily detected by smRNA-FISH, indicate that the lariat intermediate is accessible to smRNA-FISH. Concerning discriminating between pre-mRNA and lariat intermediate by smRNA-FISH, we agree with the reviewer that a dual smFISH experiment would directly distinguish between the signals of these species. The brG reporter we used in most smRNA-FISH experiments has a 5’ exon that is too short for smRNA-FISH probes, as is typical of most budding yeast 5’ exons. We have tried to replace the 5’ exon with a longer sequence (GFP) to allow for smRNA-FISH; however, this substitution inhibited splicing. Therefore, to distinguish signals from pre-mRNA versus lariat intermediate, we used additional reporters: G1c and brC reporters, which accumulate pre-mRNA essentially exclusively (Fig. S2A-C), and the UAc reporter, which accumulates lariat intermediate exclusively, in a dbr1∆ strain (Fig. S3). Whereas the mlp1 deletion did not change beta-galactosidase activities of the G1c and brC pre-mRNA-accumulating reporters (Fig. S2E), the mlp1 deletion in a dbr1∆ background did reduce the beta-galactosidase activities of the UAc lariat intermediate-accumulating reporter (Fig. 3D) and did increase smRNA-FISH signal of this reporter in the nucleus (Fig. 3E). These observations corroborate our interpretation based on the brG reporter that Mlp1p is required for efficient export of lariat intermediates but not pre-mRNAs.

(2) In some cases, the number of smFISH foci appears to change drastically depending on the genetic background. This could either be due to the stochastic nature of mRNA expression between cells or reflect real differences between the genetic backgrounds that could alter the interpretation of the other observations.

We thank the reviewer for raising this point. We will review our data to distinguish between these possibilities.

(3) The authors state in the discussion that "the general mRNA export pathway transports discarded lariat intermediates into the cytoplasm". Although this appears to be the case for the reporters that are investigated in this paper, I don't think that the authors should make such a broad sweeping claim. It may be that some discarded lariat intermediates are exported to the cytoplasm while others are targeted for nuclear retention and/or decay.

The reviewer’s point is well-taken. We will revise the wording accordingly.

Reviewer #2 (Public Review):

In this report, Zeng and Staley have used an elegant combination of RNA imaging approaches (single molecule FISH), RNA co-immunoprecipitations, and translation reporters to characterize the factors and pathways involved in the nuclear export of splicing intermediates in budding yeast. Their study notably involves the use of specific reporter genes, which lead to the accumulation of pre-mRNA and lariat species, in a battery of mutants impacting mRNA export and quality control.

The authors convincingly demonstrate that mRNA species expressed from such reporters are exported to the cytoplasm in a manner depending on the canonical mRNA export machinery (Mex67 and its adaptors) and the nuclear pore complex (NPC) basket (Mlp1). Interestingly, they provide evidence that the export of splicing intermediates requires docking and subsequent undocking at the nuclear basket, a step possibly more critical than for regular mRNAs.

We thank the reviewer for this overall positive assessment.

However, their assays do not always allow us to define whether the impacted mRNA species correspond to lariats and/or pre-mRNAs. This is all the more critical since their findings apparently contradict previous reports that supported a role for the nuclear basket in pre-mRNA quality control. These earlier studies, which were similarly based on the use of dedicated yet distinct reporters, had found that the nuclear basket subunit Mlp1, together with different cofactors, prevents the export of unspliced mRNA species. It would be important to clarify experimentally and discuss the possible reasons for these discrepancies.

It is true that we did not assess export of all reporters in all mutant strains by smFISH; however, we did validate the key conclusion that the export of lariat intermediates requires the nuclear basket gene MLP1: the export of both the brG reporter (mostly lariat intermediate) and the UAc reporter (exclusively lariat intermediate) showed a dependence on MLP1 (Fig. 3). Further, by beta-galactosidase activity, we tested in total five separate reporters – three that accumulated lariat intermediate and two that accumulated exclusively pre-mRNA; only the three reporters accumulating lariat intermediate showed a dependence of export on MLP1 (Fig. 4B,D; Fig S2D); the reporters accumulating pre-mRNA did not show a dependence on MLP1 (Fig. S2E), further validating our main conclusion. We are considering additional experiments to validate this key conclusion even further. Also, see response to comment 1 from reviewer 1.

We agree that the main conclusion from this manuscript differs from earlier studies. A key difference is that prior studies monitored exclusively pre-mRNA. In our study, we monitored pre-mRNA and lariat intermediate species and in doing so revealed a role for MLP1 in the export of lariat intermediates. This study, our previous study, as well as the previous studies of others have all provided evidence for efficient export of pre-mRNA; all of these studies are in conflict with the studies purporting a general role for the nuclear basked in retaining immature mRNA. Still, these past apparently conflicting studies can be re-interpreted in the context of our model that the export of such species requires docking at the nuclear basket, followed by undocking. In a revised manuscript, we will discuss the possibility that pre-mRNA apparently “retained” by the nuclear basket are stalled in export at the undocking stage.

Reviewer #3 (Public Review):

Summary:

Zeng and Stanley show that in yeast, intron-lariat intermediates that accumulated due to defects in pre-mRNA splicing, are transported to the cytoplasm using the canonical mRNA export pathway. Moreover, they demonstrate that export requires the nuclear basket, a sub-structure of the nuclear pore complex previously implicated with the retention of immature mRNAs. These observations are important as they put into question a longstanding model that the main role of the nuclear basket is to ensure nuclear retention of immature or faulty mRNAs.

Strengths:

The authors elegantly combine genetic, biochemical, and single-molecule resolution microscopy approaches to identify the cellular pathway that mediates the cytoplasmic accumulation of lariat intermediates. Cytoplasmic accumulation of such splicing intermediates had been observed in various previous studies but how these RNAs reach the cytoplasm had not yet been investigated. By using smFISH, the authors present compelling, and, for the first time, direct evidence that these intermediates accumulate in the cytoplasm and that this requires the canonical mRNA export pathway, including the RNA export receptor Mex67 as well as various RNA-binding proteins including Yra1, Npl3 and Nab2. Moreover, they show that the export of lariat intermediates, but not mRNAs, requires the nuclear basket (Mlp1) and basket-associated proteins previously linked to the mRNP rearrangements at the nuclear pore. This is a surprising and important observation with respect to a possible function of the nuclear basket in mRNA export and quality control, as it challenges a longstanding model that the role of the basket in mRNA export is primarily to act as a gatekeeper to ensure that immature mRNAs are not exported. As discussed by the authors, their finding suggests a role for the basket in promoting the export of certain types of RNAs rather than retention, a model also supported by more recent studies in mammalian cells. Moreover, their findings also collaborate with a recent paper showing that in yeast, not all nuclear pores contain a basket (PMID: 36220102), an observation that also questioned the gatekeeper model of the basket, as it is difficult to imagine how the basket can serve as a gatekeeper if not all nuclear pore contain such a structure.

We thank the reviewer for highlighting the importance and surprising nature of our findings.

Weaknesses:

One weakness of this study is that all their experiments rely on using synthetic splicing reporter containing a lacZ gene that produces a relatively long transcript compared to the average yeast mRNA.

We are considering repeating some of our experiments to monitor export of RNAs with more average lengths.

The rationale for using a reporter containing the brG (G branch point) resulting in more stable lariat intermediates due to them being inefficient substrates for the debranching enzyme Dbr1 could be described earlier in the manuscript, as this otherwise only becomes clear towards the end, what is confusing.

We thank the reviewer for this comment. We will revise the text to explain sooner the rationale for using the brG reporter to assess the export of lariat intermediates.

Discussion of their observation in the context that, in yeast, not all pores contain a basket would be useful.

Thanks for this suggestion. We will raise this point that a nuclear basket is not present on all nuclear pores and discuss the implications.

eLife assessment

This is an important study that demonstrates that RNA intermediates arising from improper splicing are exported out of the nucleus via the canonical mRNA export machinery and the nuclear pore basket. The authors provide convincing evidence that the role of the nuclear basket rather than retaining the transcripts is stimulating their export, putting into question the current model of the role of the basket. The conclusions are in line with recent studies in mammalian cells that suggest that the basket's role in mRNA export and quality control has to be revised.

Reviewer #1 (Public Review):

Summary:

In this manuscript, Zeng and Staley provide a valuable analysis of the molecular requirements for the export of a reporter mRNA that contains a lariat structure at its 5' end in the budding yeast S. cerevisiae. The authors provide evidence that this is regulated by the main mRNA export machinery (Yra1, Mex67, Nab2, Npl3, Tom1, and Mlp1). Of note, Mlp1 has been mainly implicated in the nuclear retention of unspliced pre-mRNA (i.e. quality control), and relatively little has been done to investigate its role in mRNA export in budding yeast.

Strengths:

There is relatively little information in the current literature about the nuclear export of splicing intermediates. This paper provides one of the first analyses of this process and dissects the molecular components that promote this form of RNA export. Overall, the strength of the data presented in the manuscript is solid. The paper is well written and the message is clear and of general interest to the mRNA community.

Weaknesses:

There are three problems with the paper, although these are not major and likely would not affect the final model as most aspects of the molecular details are confirmed by multiple complementary assays.

(1) The brG reporter produces both unspliced pre-mRNA and a lariat-containing intermediate RNA. Based on the primer extension assay the authors claim that only 33% of the final product is in pre-mRNA form and that this "is insufficient to account for the magnitude of the cytoplasmic signal from the brG reporter (83%)". Nevertheless, it is possible that primer extension is incomplete or that the lariat-containing RNA is inaccessible for smFISH. The authors could easily perform a dual smFISH experiment (similar to Adivarahan et l., Molecular Cell 2018) where exon 1 is labelled with probes of one color, and the region that overlaps the lariat-containing intermediate is labelled with probes of a second color. If the authors are correct, then one-third of the smFISH foci should have both labels and the rest would have only the second label. This would also confirm that the latter (i.e. the lariat-containing RNAs) are exported to the cytoplasm. Using this approach, the authors could then show that MLP1-depletion (or depletion of any of the other factors) affect(s) one pool of RNAs (i.e. those that are lariat-containing) but not the other (i.e. pre-mRNA). Including these experiments would make the evidence for their model more convincing.

(2) In some cases, the number of smFISH foci appears to change drastically depending on the genetic background. This could either be due to the stochastic nature of mRNA expression between cells or reflect real differences between the genetic backgrounds that could alter the interpretation of the other observations.

(3) The authors state in the discussion that "the general mRNA export pathway transports discarded lariat intermediates into the cytoplasm". Although this appears to be the case for the reporters that are investigated in this paper, I don't think that the authors should make such a broad sweeping claim. It may be that some discarded lariat intermediates are exported to the cytoplasm while others are targeted for nuclear retention and/or decay.

Reviewer #2 (Public Review):

In this report, Zeng and Staley have used an elegant combination of RNA imaging approaches (single molecule FISH), RNA co-immunoprecipitations, and translation reporters to characterize the factors and pathways involved in the nuclear export of splicing intermediates in budding yeast. Their study notably involves the use of specific reporter genes, which lead to the accumulation of pre-mRNA and lariat species, in a battery of mutants impacting mRNA export and quality control.

The authors convincingly demonstrate that mRNA species expressed from such reporters are exported to the cytoplasm in a manner depending on the canonical mRNA export machinery (Mex67 and its adaptors) and the nuclear pore complex (NPC) basket (Mlp1). Interestingly, they provide evidence that the export of splicing intermediates requires docking and subsequent undocking at the nuclear basket, a step possibly more critical than for regular mRNAs.

However, their assays do not always allow us to define whether the impacted mRNA species correspond to lariats and/or pre-mRNAs. This is all the more critical since their findings apparently contradict previous reports that supported a role for the nuclear basket in pre-mRNA quality control. These earlier studies, which were similarly based on the use of dedicated yet distinct reporters, had found that the nuclear basket subunit Mlp1, together with different cofactors, prevents the export of unspliced mRNA species. It would be important to clarify experimentally and discuss the possible reasons for these discrepancies.

Reviewer #3 (Public Review):

Summary:

Zeng and Stanley show that in yeast, intron-lariat intermediates that accumulated due to defects in pre-mRNA splicing, are transported to the cytoplasm using the canonical mRNA export pathway. Moreover, they demonstrate that export requires the nuclear basket, a sub-structure of the nuclear pore complex previously implicated with the retention of immature mRNAs. These observations are important as they put into question a longstanding model that the main role of the nuclear basket is to ensure nuclear retention of immature or faulty mRNAs.

Strengths:

The authors elegantly combine genetic, biochemical, and single-molecule resolution microscopy approaches to identify the cellular pathway that mediates the cytoplasmic accumulation of lariat intermediates. Cytoplasmic accumulation of such splicing intermediates had been observed in various previous studies but how these RNAs reach the cytoplasm had not yet been investigated. By using smFISH, the authors present compelling, and, for the first time, direct evidence that these intermediates accumulate in the cytoplasm and that this requires the canonical mRNA export pathway, including the RNA export receptor Mex67 as well as various RNA-binding proteins including Yra1, Npl3 and Nab2. Moreover, they show that the export of lariat intermediates, but not mRNAs, requires the nuclear basket (Mlp1) and basket-associated proteins previously linked to the mRNP rearrangements at the nuclear pore. This is a surprising and important observation with respect to a possible function of the nuclear basket in mRNA export and quality control, as it challenges a longstanding model that the role of the basket in mRNA export is primarily to act as a gatekeeper to ensure that immature mRNAs are not exported. As discussed by the authors, their finding suggests a role for the basket in promoting the export of certain types of RNAs rather than retention, a model also supported by more recent studies in mammalian cells. Moreover, their findings also collaborate with a recent paper showing that in yeast, not all nuclear pores contain a basket (PMID: 36220102), an observation that also questioned the gatekeeper model of the basket, as it is difficult to imagine how the basket can serve as a gatekeeper if not all nuclear pore contain such a structure.

Weaknesses:

One weakness of this study is that all their experiments rely on using synthetic splicing reporter containing a lacZ gene that produces a relatively long transcript compared to the average yeast mRNA.

The rationale for using a reporter containing the brG (G branch point) resulting in more stable lariat intermediates due to them being inefficient substrates for the debranching enzyme Dbr1 could be described earlier in the manuscript, as this otherwise only becomes clear towards the end, what is confusing.

Discussion of their observation in the context that, in yeast, not all pores contain a basket would be useful.

Author Response

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This work describes the mechanism of protein disaggregation by the ClpL AAA+ protein of Listeria monocytogenes. Using several model subtrate proteins the authors first show that ClpL possesses a robust disaggregase activity that does not further require the endogenous DnaK chaperone in vitro. In addition, they found that ClpL is more thermostable than the endogenous L. monocytogenes DnaK and has the capacity to unfold tightly folded protein domains. The mechanistic basis for the robust disaggregase activity of ClpL was also dissected in vitro and in some cases, supported by in vivo data performed in chaperonedeficient E. coli strains. The data presented show that the two AAA domains, the pore-2 site and the N-terminal domain (NTD) of ClpL are critical for its disaggregase activity. Remarkably, grafting the NTD of ClpL to ClpB converted ClpB into an autonomous disaggregase, highlighting the importance of such a domain in the DnaK-independent disaggregation of proteins. The role of the ClpL NTD domain was further dissected, identifying key residues and positions necessary for aggregate recognition and disaggregation. Finally, using sets of SEC and negative staining EM experiments combined with conditional covalent linkages and disaggregation assays the authors found that ClpL shows significant structural plasticity, forming dynamic hexameric and heptameric active single rings that can further form higher assembly states via their middle domains.

Strengths:

The manuscript is well-written and the experimental work is well executed. It contains a robust and complete set of in vitro data that push further our knowledge of such important disaggregases. It shows the importance of the atypical ClpL N-terminal domain in the disaggregation process as well as the structural malleability of such AAA+ proteins. More generally, this work expands our knowledge of heat resistance in bacterial pathogens.

Weaknesses:

There is no specific weakness in this work, although it would have helped to have a drawing model showing how ClpL performs protein disaggregation based on their new findings. The function of the higher assembly states of ClpL remains unresolved and will need further extensive research. Similarly, it will be interesting in the future to see whether the sole function of the plasmid-encoded ClpL is to cope with general protein aggregates under heat stress.

We thank the reviewer for the positive evaluation. We agree with the reviewer that it will be important to test whether ClpL can bind to and process non-aggregated protein substrates. Our preliminary analysis suggests that the disaggregation activity of ClpL is most relevant in vivo, pointing to protein aggregates as main target.

We also agree that the role of dimers or tetramers of ClpL rings needs to be further explored. Our initial analysis suggests a function of ring dimers as a resting state. It will now be important to study the dynamics of ClpL assembly formation and test whether substrate presence shifts ClpL assemblies towards an active, single ring state.

Reviewer #2 (Public Review):

The manuscript by Bohl et al. is an interesting and carefully done study on the biochemical properties and mode of action of potent autonomous AAA+ disaggregase ClpL from Listeria monocytogenes. ClpL is encoded on plasmids. It shows high thermal stability and provides Listeria monocytogenes food-pathogen substantial increase in resistance to heat. The authors show that ClpL interacts with aggregated proteins through the aromatic residues present in its N-terminal domain and subsequently unfolds proteins from aggregates translocating polypeptide chains through the central pore in its oligomeric ring structure. The structure of ClpL oligomers was also investigated in the manuscript. The results suggest that mono-ring structure and not dimer or trimer of rings, observed in addition to mono-ring structures under EM, is an active species of disaggregase.

Presented experiments are conclusive and well-controlled. Several mutants were created to analyze the importance of a particular ClpL domain.

The study's strength lies in the direct comparison of ClpL biochemical properties with autonomous ClpG disaggregase present in selected Gram-negative bacteria and well-studied E. coli system consisting of ClpB disaggregase and DnaK and its cochaperones. This puts the obtained results in a broader context.

We thank the reviewer for the detailed comments. There are no specific weaknesses indicated in the public review.

Reviewer #3 (Public Review):

Summary:

This manuscript details the characterization of ClpL from L. monocytogenes as a potent and autonomous AAA+ disaggregase. The authors demonstrate that ClpL has potent and DnaKindependent disaggregase activity towards a variety of aggregated model substrates and that this disaggregase activity appears to be greater than that observed with the canonical DnaK/ClpB co-chaperone. Furthermore, Lm ClpL appears to have greater thermostability as compared to Lm DnaK, suggesting that ClpL-expressing cells may be able to withstand more severe heat stress conditions. Interestingly, Lm ClpP can provide thermotolerance to E. coli that have been genetically depleted of either ClpB or in cells expressing a mutant DnaK103. The authors further characterized the mechanisms by which ClpL interacts with protein aggregates, identifying that the N-terminal domain of ClpL is essential for disaggregase function. Lastly, by EM and mutagenesis analysis, the authors report that ClpL can exist in a variety of larger macromolecular complexes, including dimer or trimers of hexamers/heptamers, and they provide evidence that the N-terminal domains of ClpL prevent dimer ring formation, thus promoting an active and substrate-binding ClpL complex. Throughout this manuscript the authors compare Lm ClpL to ClpG, another potent and autonomous disaggregase found in gram-negative bacteria that have been reported on previously, demonstrating that these two enzymes share homologous activity and qualities. Taken together this report clearly establishes ClpL as a novel and autonomous disaggregase.

Strengths:

The work presented in this report amounts to a significant body of novel and significant work that will be of interest to the protein chaperone community. Furthermore, by providing examples of how ClpL can provide in vivo thermotolerance to both E. coli and L. gasseri the authors have expanded the significance of this work and provided novel insight into potential mechanisms responsible for thermotolerance in food-borne pathogens.

Weaknesses:

The figures are clearly depicted and easy to understand, though some of the axis labeling is a bit misleading or confusing and may warrant revision. While I do feel that the results and discussion as presented support the authors' hypothesis and overall goal of demonstrating ClpL as a novel disaggregase, interpretation of the data is hindered as no statistical tests are provided throughout the manuscript. Because of this only qualitative analysis can be made, and as such many of the concluding statements involving pairwise comparisons need to be revisited or quantitative data with stats needs to be provided. The addition of statistical analysis is critical and should not be difficult, nor do I anticipate that it will change the conclusions of this report.

We thank the reviewer for the valid criticism. We addressed the major concern of the reviewer and added the requested statistical analysis to all relevant figures. The analysis confirms our conclusions. We also followed the advice of the reviewer and revised axis labeling to increase clarity.

Reviewer #1 (Recommendations For The Authors):

• It would really help to have a model showing how ClpL performs protein disaggregation based on their findings.

We show that ClpL exerts a threading activity that is fueled by ATP hydrolysis in both AAA domains and executed by pore-located aromatic residues. The basic disaggregation mechanism of ClpL therefore does not differ from ClpB and ClpG disaggregases. Similarly, the specificity of ClpL towards protein aggregates is based on simultaneous interactions of multiple N-terminal domains with the aggregate surface. We could recently describe a similar mode of aggregate recognition for ClpG [1]. We therefore prefer not to add a model to the manuscript. We are currently in preparation of a review that includes the characterization of the novel bacterial disaggregases and will present models there as we consider a review article as more appropriate for such illustrations.

• AAA2 domain of ClpL in Fig 3E should be the same color as in Fig 1A.

We used light grey instead of dark grey for the ClpL AAA2 domain in Fig 3E, to distinguish between ClpL and ClpB AAA domains. This kind of illustration allows for clearer separation of both AAA+ proteins and the fusion construct LN-ClpB*. We therefore prefer keeping the color code.

• Partial suppression of the dnaK mutant could be added in the main manuscript Figure.

The main figure 3 is already very dense and we therefore prefer showing respective data as part of a supplementary figure.

• It would have been interesting to know if the robust autonomous disaggregation activity of ClpL would be sufficient to rescue the growth of more severe E. coli chaperone mutants, like dnaK tig for example. Did the authors test this?

We tested whether expression of clpL can rescue growth of E. coli dnaK103 mutant cells at 40°C on LB plates. This experiment is different from the restoration of heat resistance in dnaK103 cells (Figure 3, figure supplement 2A), as continuous growth at elevated temperatures (40°C) is monitored instead of cell survival upon abrupt severe heat shock (49°C). We did not observe rescue of the temperature-sensitive growth phenotype (40°C) of dnaK103 cells upon clpL expression, though expression of clpG complemented the temperature-sensitive growth phenotype (see Author response image 1 below). This finding points to differences in chaperone activities of ClpL and ClpG. It also suggests that ClpL activity is largely restricted to heat-shock generated protein aggregates, enabling ClpL to complement the missing disaggregation function of DnaK but not other Hsp70 activities including folding and targeting of newly synthesized proteins. We believe that dissecting the molecular reasons for differences in ClpG and ClpL complementation activities should be part of an independent study and prefer showing the growth-complementation data only in the response letter.

Author response image 1.

Serial dilutions (10-1 – 10-6) of E. coli dnaK103 mutant cells expressing E. coli dnaK, L. monocytogenes clpL or P. aeruginosa clpG were spotted on LB plates including the indicated IPTG concentrations. Plates were incubated at 30°C or 40°C for 24 h. p: empty vector control.

Reviewer #2 (Recommendations For The Authors):

Based on results presented in Fig. 2B the authors conclude "that stand-alone disaggregases ClpL and ClpG but not the canonical KJE/ClpB disaggregase exhibit robust threading activities that allow for unfolding of tightly folded domains" (page 5 line 209). In this experiment, the threading power of disaggregases was assessed by monitoring YFP fluorescence during the disaggregation of aggregates formed by fusion luciferase-YFP protein. In my opinion, the results of the experiment depend not only on the threading power of disaggregases but also on the substrate recognition by analyzed disaggregating systems and/or processivity of disaggregases. N-terminal domain in the case of ClpL and KJE chaperones in the case of the KJE/ClpB system are involved in recognition. This is not discussed in the manuscript and the obtained result might be misinterpreted. The authors have created the LN-ClpB* construct (N-terminal domain of ClpL fused to derepressed ClpB) (Fig. 3 E and F). In my opinion, this construct should be used as an additional control in the experiment in Fig. 2 B. It possesses the same substrate recognition domain and therefore the direct comparison of disaggregases threading power might be possible.

We performed the requested experiment (new Figure 3 - figure supplement 2D). We did not observe unfolding of YFP by LN-ClpB. Sínce ClpL and LN-ClpB do not differ in their aggregate targeting mechanisms, this finding underlines the differences in threading power between ClpL and activated (derepressed) ClpB. It also suggests that the AAA threading motors and the aggregate-targeting NTD largely function independently.

Presented results suggest that tetramer and dimer of rings might be a "storage form" of disaggregase. It would be interesting to analyze the thermotolerance and/or phenotype of ClpL mutants that do not form tetramer and dimer (E352A). This variant possesses similar to WT disaggregation activity but does not form dimers and tetramers. If in vivo the differences are observed (for example toxicity of the mutant), the "storage form" hypothesis will be probable.

When testing expression of clpL-MD mutants (E352A, F354A), which cannot form dimers and tetramers of ClpL rings, in E. coli ∆clpB cells, we observed reduced production levels as compared to ClpL wildtype and speculated that reduced expression might be linked to cellular toxicity. We therefore compared spotting efficiencies of E. coli ∆clpB cells expression clpL, ∆NclpL or the clpL-MD mutants at different temperatures. Expression of clpL at high levels abrogated colony formation at 42°C (new Figure 6 - figure supplement 3). ClpL toxicity was dependent on its NTD as no effect was observed upon expression of ∆N-clpL. ClpL-MD mutants (E352A, F354A) were expressed at much lower levels and exhibited strongly increased toxicity as compared to ClpL-WT when produced at comparable levels (new Figure 6 – figure supplement 3). This implies a protective role of ClpL ring dimers and tetramers in the cellular environment by downregulating ClpL activity. We envision that the formation of ClpL assemblies restricts accessibility of the ClpL NTDs and reduces substrate interaction. Increased toxicity of ClpL-E352A and ClpL-F354A points to a physiological relevance of the dimers and tetramers of ClpL rings and is in agreement with the proposed function as storage forms. We added this potential role of ClpL ring assemblies to the discussion section. Due to the strongly reduced production levels of ClpL MD mutants and their enhanced toxicity at elevated temperatures we did not test for their ability to restore thermotolerance in E. coli ∆clpB cells.

Figure 6G and Figure 6 -figure supplement 2 - it is not clear what is the difference in the preparation of WT and WTox forms of ClpL.

ClpL WT was purified under reduced conditions (+ 2 mM DTT), whereas WTox was purified in absence of DTT, thus serving as control for ClpL-T355C, which forms disulfide bonds upon purification without DTT. We have added respective information to the figure legend and the materials and methods section.

Page 5 line 250 - wrong figure citation. Instead of Figure 1 - Figure Supplement 2A should be Figure 3 - Figure Supplement 2A.

Page 5 line 251 - wrong figure citation. Instead of Figure 1 - Figure Supplement 2B/C should be Figure 3 - Figure Supplement 2B/C.

Page 7 line 315 - wrong figure citation. Instead of Figure 4F, it should be Figure 4G Figure 1 - Figure Supplement 2E - At first glance, this Figure does not correspond to the text and is confusing. It would be nice to have bars for Lm ClpL activity in the figure. Alternatively, the description of the y-axis might be changed to "relative to Lm ClpL disaggregation activity" instead of "relative disaggregation activity". One has to carefully read the figure legend to find out that 1 corresponds to Lm ClpL activity.

We have corrected all mistakes and changed the description of y-axis (Figure 1 - figure Supplement 2E) as suggested.

Reviewer #3 (Recommendations For The Authors):

(1) While the authors make many experimental comparisons throughout their study, no statistical tests are described or presented with their results or figures, nor are these statistical tests described in the methods. While the data as presented does appear to support the author's conclusions, without these statistical tests no meaningful conclusions from paired analysis can be drawn. Critically, please report these statistical tests. As a general suggestion please include the statistics (p-values) in the results section when presenting this data, as well as in the figure legends, as this will allow the reader to better understand the authors' presentation and interpretation of the data.

We have added statistical tests to all relevant figures. The analysis is confirming our former statements. We have further clarified our approach for the statistical analysis in the methods section. We report p-values in the results section, however, due to the volume of comparisons we did not add individual p-values to the figure legends but used standard labeling with stars.

(2) Some of the axis labels for the presented graphs are a bit misleading or confusing. Many describe a relative (%) disaggregation rate, but it is not clear from the methods or figure legends what this rate is relative to. Is it relative to non-denatured substrates, to no chaperone conditions, etc.? Is it possible to present the figures with the raw data rates/activity (ex. luciferase activity / time) vs. relative rates? I think that labeling these figure axes with "disaggregation rate" is a bit misleading as none of these experiments measure the actual rate of disaggregation of these model substrates per se (say by SEC-MALS or other biophysical measurements), but instead infer the extent of disaggregation by measuring a property of these substrates, i.e. luciferase activity or fluorescence intensity over time. Thus, labeling these figures with the appropriate axis for what is being measured, and then clarifying in the methods and results what is being inferred by these measurements, will help solidify the author's conclusions.

Relative (%) disaggregation rate usually refers to the disaggregation activity of ClpL wildtype serving as reference. We clarified this point in the revised text and respective figure legends. We now also refer to the process measured (e.g. relative refolding activity of aggregated Luciferase instead of relative disaggregation activity) as suggested by the reviewer and added clarifications to text and materials and methods.

Since we have many measurements for our most frequently used assays and have a reasonable estimate for the general variance within these assays, we found it reasonable to show activity data in relation to fixed controls. This reduces the impact of unspecific variance and thereby makes more accurate comparisons between different repetitions. The reference is now indicated in the axis title.

(3) The figures are well presented, clutter-free, and graphically easy to understand. Figure legends have sufficient information aside from the aforementioned statistical information and should include the exact number of independent replicates for each panel/experiment (ex. n=4), not just a greater than 3. While the figures do show each data point along with the mean and error, in some figures it is difficult to determine the number of replicate data points. Example figures 2c, 2d, and 3a. Also, please state whether the error is std. error or SEM.

While we agree, that this is valuable information, we fear that overloading the figure legends with information may take a toll on the readability. We therefore decided to append the number of replicates for each experiment in a separate supplementary table (Table S2). The depicted error is showing the SD and not the SEM, which we also specified in the figure legends.

(4) There are various examples throughout the results where qualitative descriptors are used to describe comparisons. Examples of this are "hardly enhanced" (Figure 1) and "partially reduced" (Figure 6). While this is not necessarily wrong, qualitative descriptions of comparisons in this manner would require further explanation. What is the definition of "hardly" or "partially"? My recommendation is to just state the data quantitatively, such as "% enhanced" or "reduced by x", this way there is no misinterpretation. Examples of this can be found in Figures 6C-G. This would require a full statistical overview and presentation of these stats in the results.

We followed the reviewer`s advice and no longer use the terms criticized (e.g. “hardly enhanced”). We instead provide the requested quantifications in the text.

Questions for Figures:

Figures 1B and 1C:

(1) Is the disaggregase activity of ClpL towards heat-denatured luciferase and GFP ATPdependent? While the authors later in the manuscript show that mutations within the Walker B domains dramatically impair reactivation (disaggregation) of denatured luciferase, this does not rule out an ATP-independent effect of these mutations. Thus, the authors should test whether disaggregase activity is observed when wild-type ClpL is incubated with denatured substrates without ATP present or in the presence of ADP only.

We tested for ClpL disaggregation activity in absence of nucleotide and presence of ADP only (new Figure 1 – figure supplement 2A). We did not observe any activity, demonstrating that ClpL activity depends on ATP binding and hydrolysis (see also Figure 3 – figure supplement 1D: ATPase-deficient ClpL-E197A/E530A is lacking disaggregation activity).

(2) The authors suggest that a reduction in disaggregase activity observed in samples combining Lm ClpL and KJE (Figure 1C, supp. 1C-E) could be due to competition for protein aggregate binding as observed previously with ClpG. Did the authors test this directly by pulldown assay or another interaction-based assay? While ClpL and ClpG appear to work in a similar manner, it would be good to confirm this. Also, clarification on how this competition operates would be useful. Is it that ClpL prevents aggregates from interacting with KJE, or vice versa?

We probed for binding of ClpL to aggregated Malate Dehydrogenase in the presence of L. monocytogenes or E. coli Hsp70 (DnaK + respective J-domain protein DnaJ) by a centrifugation-based assay. Here, we used the ATPase-deficient ClpL-E197A/E530A (ClpLDWB) mutant, ensuring stable substrate interaction in presence of ATP. We observe reduced binding of ClpL-DWB to protein aggregates in presence of DnaK/DnaJ (new Figure 1 – figure supplement 2G). This finding indicates that both chaperones compete for binding to aggregated proteins and explains inhibition of ClpL disaggregation activity in presence of Hsp70.

(3) Related to the above, while incubation of aggregated substrates with ClpL and KJE does appear to reduce aggregase activity towards GFP (Figure 1c), α-glucosidase (Supp. 1C), and MDH (Supp. 1D), this doesn't appear to be the case towards luciferase (Figure 1b, Supp. 1b). Furthermore, ClpL aggregase activity is reduced towards luciferase when combined with E. coli KJE (Supp. 1e) but not with Lm KJE (Figure 1b). The authors provide no commentary or explanation for these observations. Furthermore, these results complicate the concluding statement that "combining ClpL with Lm KJE always led to a strong reduction in disaggregation activity ... ".

We suggest that the differing inhibitory degrees of the KJE system on ClpL disaggregation activities reflect diverse binding affinities of KJE and ClpL to the respective aggregates. While we usually observe strong inhibition of ClpL activity in presence of KJE, this is different for aggregated Luciferase. This points to specific structural features of Luciferase aggregates or the presence of distinct binding sites on the aggregate surface that favour ClpL binding. We have added a respective comment to the revised manuscript.

The former statement that “combining ClpL with Lm KJE always led to a strong reduction in disaggregation activity” referred to aggregated GFP, MDH and α-Glucosidase for which a strong inhibition of ClpL activity was observed. We have specified this point.

Figures 1D and 1E:

(1) The authors conclude that the heat sensitivity of ΔClpL L. gasseri cells is because they do not express the canonical ClpB disaggregase. A good test to validate this would be to express KJE/ClpB in these Lg ΔClpL cells to see if heat-sensitivity could be fully or partially rescued.

We agree that such experiment would further strengthen the in vivo function of ClpL as alternative disaggregase. However, such approach would demand for co-expression of E. coli ClpB with the authentic E. coli DnaK chaperone system (KJE), as ClpB and DnaK cooperate in a species-specific manner [2-4]. This makes the experiment challenging, also because the individual components need to be expressed at a correct stochiometry. Furthermore, the presence of the authentic L. gasseri KJE system, which is likely competing with the E. coli KJE system for aggregate binding, will hamper E. coli KJE/ClpB disaggregation activity in L. gasseri. In view of these limitations, we would like to refrain from conducting such an experiment.

(2) The rationale for investigating Lg ClpL, and the aggregase activity assays are compelling and support the hypothesis that ClpL contributes to thermotolerance in multiple grampositive species. Though, from Figure 1d, why was only Lg ClpL investigated? It appears that S. thermophilus also lacks the canonical ClpB disaggregase and demonstrates ΔClpL heat sensitivity. There is also other Lactobacillus sp. presented that lack ClpB but were not tested for heat sensitivity. Why only test and move forward with L. gasseri? Lastly, L. mesenteroides is ClpB-negative but doesn't demonstrate ΔClpL heat sensitivity. Why?

We wanted to document high, partner-independent disaggregation activity for another ClpL homolog. We chose L. gasseri, as (i) this bacterial species lacks a ClpB homolog and (ii) a ∆clpL mutant exhibit reduced survival upon severe heat shock (thermotolerance phenotype), which is associated with defects in cellular protein disaggregation. The characterization of L. gasseri ClpL as potent disaggregase in vitro represents a proof-of-concept and allows to generalize our conclusion. We therefore did not further test S. thermophilus ClpL. L. mesenteroides encodes for ClpL but not ClpB, yet, a ∆clpL mutant has not yet been characterized in this species to the best of our knowledge. As we wanted to link ClpL in vitro activity with an in vivo phenotype, we did not characterize L. mesenteroides ClpL.

We agree with the reviewer that the characterization of additional ClpL homologs is meaningful and interesting, however, we strongly believe that such analysis should be part of an exhaustive and independent study.

Figures 2A and 2B:

(1) Figure 2B demonstrates that both ClpL and ClpG, but not the canonical KJE/ClpB, are able to unfold YFP during the luciferase disaggregation process, suggesting that ClpL and ClpG exhibit stronger threading activity. A technical question, can luciferase activity be measured alongside in the same assay sample? If so, would you expect to observe a concomitant increase in luciferase activity as YFP fluorescence decreases?

KJE/ClpB can partially disaggregate and refold aggregated Luciferase-YFP without unfolding YFP during the disaggregation reaction [5]. YFP unfolding is therefore not linked to refolding of aggregated Luciferase-YFP. On the other hand, unfolding of YFP during disaggregation can hamper the refolding of the fused Luciferase moiety as observed for the AAA+ protein ClpC in presence of its partner MecA [5]. These diverse effects make the interpretation of LuciferaseYFP refolding experiments difficult as the degree of YFP unfolding activity does not necessarily correlate with the extend of Luciferase refolding. We therefore avoided to perform the suggested experiment.

Figure 2C and 2D:

(1) Thermal shift assays for ClpL, ClpG, and DnaK were completed with various nucleotides. Were these experiments also completed with samples in their nucleotide-free apo state? Also, while all these chaperones are ATPases, the nucleotides used differ, but no explanation is provided. Comparison should be made of these ATPases bound to the same molecules.

We did not monitor thermal stabilities of chaperones without nucleotide as such state is likely not relevant in vivo. We used ATPγS in case of ClpL to keep the AAA+ protein in the ATPconformation. ATP would be rapidly converted to ADP due to the high intrinsic ATPase activity of ClpL. In case of DnaK ATPγS cannot be used as it does not induce the ATP conformation [6]. The low intrinsic ATPase activity of DnaK allows determining the thermal stability of its ATP conformation in presence of ATP. This is confirmed by calculating a reduced thermal stability of ADP-bound DnaK.

(2) The authors suggest that incubation at 55⁰C will cause unfolding of Lm DnaK, but not ClpL, providing ClpL-positive Lm cells disaggregase activity at 55⁰C. While the thermal shift assays in Figures 2C and 2D support this, an experiment to test this would be to heat-treat Lm DnaK and ClpL at 55⁰C then test for disaggregase activity using either aggregated luciferase or GFP as in Figure 1.

We followed the suggestion of the reviewer and incubated Lm ClpL and DnaK at 55-58°C in presence of ATP for 15 min prior to their use in disaggregation assays. We compared the activities of pre-heated chaperones with controls that were incubated at 30°C for 15 min. Notably, we did not observe a loss of DnaK disaggregation activity, suggesting that thermal unfolding of DnaK at this temperature is reversible. We provide these data as Figure 2 -figure supplement 1 and added a respective statement to the revised manuscript.

Figure 3B:

(1) The authors state that ATPase activity of ΔN-ClpL was "hardly affected", but from the data provided it appeared to result in an approximate 35% reduction. As discussed above, no stats are provided for this figure, but given the error bars, it is highly likely that this reduction is significant. Please perform this statistical test, and if significant, please reflect this in the written results as well as the figure. Lastly, if this reduction in ATPase activity is significant, why would this be so, and could this contribute to the reduction in aggregase activity towards luciferase and MDH observed in Figure 3A?

We applied statistical tests as suggested by the reviewer, showing that the reduction in ATPase activity of ∆N-ClpL is statistically significant. N-terminal domains of Hsp100 proteins can modulate ATPase activity as shown for the family member ClpB, functioning as auxiliary regulatory element for fine tuning of ClpB activity [7]. We speculate that the impact of the ClpL-NTD on the assembly state (stabilization of ClpL ring dimers) might affect ClpL ATPase activity. We would like to point out that other ClpL mutants (e.g. NTD mutant ClpL-Y51A; MDmutant ClpL-F354A) have a similarly reduced ATPase activity, yet exhibit substantial disaggregation activity (approx. 2-fold reduced compared to ClpL wildtype). In contrast ∆NClpL does not exhibit any disaggregation activity. This suggests that the loss of disaggregation activity is caused by a substrate binding defect but not by a partial reduction in ATPase activity. We added a comment on the reduced ATPase activity and also discuss its potential reasons in the discussion section.

(2) I think the authors' conclusion that deletion of the ClpL NTD does not contribute to structural defects of ClpL is premature given the apparent reduction in ATPase activity. Did the authors perform any biophysical analysis of ΔN-ClpL to confirm this conclusion? Thermal shift assays, Native-PAGE, or size-exclusion chromatography for aggregates would all be good assays to demonstrate that the wild-type and ΔN-ClpL have similar structural properties. Surprisingly, Figure 6 describes significant macromolecular changes associated with ΔN-ClpL such that it preferentially forms a dimer of rings. Furthermore, in Supp. Figure 6D the authors report that ΔN-ClpL appears to have an increased Tm as compared to WT- or ΔM-ClpL. The authors should reflect these observations as deletion of the ClpL NTD does appear to contribute to structural changes, though perhaps only at the macromolecular scale, i.e. dimerization of the rings.

We have characterized the oligomeric state of ∆N-ClpL by size exclusion chromatography (Figure 6 – figure supplement 1A) and negative staining electron microscopy (Figure 6C), both showing that it forms assemblies similar to ClpL wildtype. We did not observe an increased tendency of ∆N-ClpL to form aggregates and the protein remained fully soluble after several cycles of thawing and freezing. EM data reveal that ∆N-ClpL exclusively form ring dimers, suggesting that the NTDs destabilize MD-MD interactions. The stabilized interaction between two ∆N-ClpL rings can explain the increased thermal stability (Figure 6 – figure supplement 1D). We speculate that the ClpL NTDs either affect MD-MD interactions through steric hindrance or by directly contacting MDs. We have added a respective statement to the discussion section.

Figure 3C and 3D:

(1) Given the larger error in samples expressing ClpG (100) or ClpL (100) statistical analysis with p-values is required to make conclusions regarding the comparison of these samples vs. plasmid-only control. The effect of ΔN-ClpL vs. wild-type ClpL looks compelling and does appear to attenuate the ClpL-induced thermotolerance. This is nicely demonstrated in Figure 3D.

We quantified respective spot tests (new Figure 3E) and tested for statistical significance as suggested by the reviewer. We show that restoration of heat resistance is significant for the first 30 min. While we always observe rescue at later timepoints significance is lost here due to larger deviations in the number of viable cells and thus the degree of complementation.

Figure 3F:

(1) What is the role of the ClpB NTD? It appears to be dispensable for disaggregase activity, assuming that ClpB is co-incubated with KJE. A quick explanation of this domain in ClpB could be useful.

The ClpB NTD is not required for disaggregation activity, as ClpB is recruited to protein aggregates by DnaK, which interacts with the ClpB MDs. Still, two functions have been described for the ClpB NTD. First, it can bind soluble unfolded substrates such as casein [8]. This substrate binding function can increase ClpB disaggregation activity towards some aggregated model substrates (e.g. Glucose-6-phosphate dehydrogenase) [9]. However, NTD deletion usually does not decrease ClpB disaggregation activity and can even lead to an increase [7, 10, 11]. An increased disaggregation activity of ∆N-ClpB correlates with an enhanced ATPase activity, which is explained by NTDs stabilizing a repressing conformation of the ClpB MDs, which function as main regulators of ClpB ATPase activity [7]. We added a short description on the role of the ClpB NTD to the respective results section.

(2) The result of fusing the ClpL NTD to ClpB supports a role for this NTD in promoting autonomous disaggregase activity. What would you expect to observe if the fused Ln-ClpB protein was co-incubated with KJE? Would this further promote disaggregase activity, or potentially impair through competition? This experiment could potentially support the authors' hypothesis that ClpL and ClpB/KJE can compete with each other for aggregated substrates as suggested in Figure 1.

We have performed the suggested experiment using aggregated MDH as model substrate. We did not observe an inhibition of LN-ClpB disaggregation activity in presence of KJE. In contrast ClpL disaggregation activity towards aggregated MDH is inhibited upon addition of KJE due to competition for aggregate binding (Figure 1 – figure supplement 2D/F). Disaggregation activity of LN-ClpB in presence of KJE can be explained by functional cooperation between both chaperone systems, which involves interactions between aggregate-bound DnaK and the ClpB MDs of the LN-ClpB fusion construct. We prefer showing these data only in the response letter but not including them in the manuscript, as respective results distract from the main message of the LN-ClpB fusion construct: the ClpL NTD functions as autonomous aggregatetargeting unit that can be transferred to other Hsp100 family members.

Author response image 2.

LN-ClpB cooperates with DnaK in protein disaggregation. Relative MDH disaggregation activities of indicated disaggregation systems were determined. KJE: DnaK/DnaJ/GrpE. The disaggregation activity of Lm ClpL was set to 1. Statistical Analysis: Oneway ANOVA, Welch’s Test for post-hoc multiple comparisons. Significance levels: **p < 0.001. n.s.: not significant.

Figures 4E and 4F:

(1) While the effect of various NTD mutations follows a similar trend in regard to the impairment of ClpL-mediated disaggregation of luciferase and MDH, the degree of these effects does appear different. For example, patch A and C mutations reduce ClpL disaggregase activity towards luciferase (~60% / 50% reduction) vs. MDH (>90%) respectively. While these results do suggest a critical role for residues in patches A and C of ClpL, these substrate-specific differences are not discussed. Why would we expect a difference in the effect of these patch A/C ClpL mutations on different substrates?

We speculate that the aggregate structure and the presence or distributions of ClpL NTD binding sites differ between aggregated Luciferase and MDH. A difference between both aggregated model substrates was also observed when testing for an inhibitory effect of Lm KJE (and Ec KJE) on ClpL disaggregation activity (see comment above). We speculate that the mutated NTD residues make specific contributions to aggregate recognition. The severity of binding defects (and reduction of disaggregation activities) of these mutants will depend on specific features of the aggregated model substrates. We now point out that ClpL NTD patch mutants can differ in disaggregation activities depending on the aggregated model substrate used and refer to potential differences in aggregate structures.

(2) The authors suggest that the loss of disaggregation activity of selected NTD mutants could be linked to reduced binding to aggregated luciferase. While this is likely given that these mutations do not appear to affect ATPase activity (Supp. 4), it could be possible that these mutants can still bind to aggregated luciferase and some other mechanism may impair disaggregation. A pull-down assay would help to prove whether reduced binding is observed in these NTD ClpL mutants. This also needs to be confirmed for Supp. Figure 4.2H.

We have shown a strong correlation between loss of aggregate binding and disaggregation activity for several NTD mutants (Fig. 4G, Figure 4 – figure supplement 2H). We decided to perform the aggregate binding assay only with mutants that show a full but not a partial disaggregation defect as we made the experience that the centrifugation-based assay provides clear and reproducible results for loss-of-activity mutants but has limitations in revealing differences for partially affected mutants. This might be explained by the use of nonhydrolyzable ATPγS in these experiments, which strongly stabilizes substrate interactions, potentially covering partial binding defects. We agree with the reviewer that some ClpL NTD mutants might have additional effects on disaggregation activity by e.g. controlling substrate transfer to the processing pore site. We have added a respective comment to the revised manuscript.

(3) Supp. Figure 4.2H has no description in the figure legend. The Y-axes states % aggregate bound to chaperone. How was this measured? See the above comments for Figures 4E and 4F.

We apologize and added the description to the figure legend. The determination of % aggregate bound chaperone is based on the quantifications of chaperones present in the supernatant and pellet fractions after sample centrifugation. Background levels of chaperones in the pellet fractions in absence of protein aggregates were subtracted. We added this information to the materials and methods section.

Figure 6G:

The authors observed reduced disaggregase activity and ATPase activity of mutant T355C under both oxidative and reducing conditions. While this observation under oxidative conditions supports the authors' hypothesis, under reducing conditions (+DTT) we would expect the enzyme to behave similarly to wild-type ClpL unless this mutation has other effects. Can the authors please comment on this and provide an explanation or hypothesis?

The reviewer is correct, ClpL-T355C exhibit a reduced disaggregation activity (Figure 6 – figure supplement 2B). We observe a similar reduction in disaggregation activity for the ClpL MD mutant F354A, pointing to an auxiliary function of the MD in protein disaggregation. We have made a respective comment in the discussion section of the revised manuscript. How exactly ClpL MDs support protein disaggregation is currently unclear and will be subject of future analysis in the lab. We strongly believe that such analysis should be part of an independent study.

Discussion:

In the fourth feature, it is discussed that one disaggregase feature of ClpL is that it does not cooperate with the ClpP protease. While a reference is provided for the canonical ClpB, no data in this paper, nor a reference, is provided demonstrating that ClpL does not interact with ClpP. As discussed, it is highly unlikely that ClpL interacts with ClpP given that ClpL does not contain the IGL/F loops that mediate the interaction of ClpP with cochaperones, such as ClpX, but data or a reference is needed to make such a factual statement.

The absence of the IGL/F loop makes an interaction between ClpL and ClpP highly unlikely. However, the reviewer is correct, direct evidence for a ClpP-independent function of ClpL, though very likely, is not provided. We have therefore rephrased the respective statement: “Forth, novel disaggregases lack the specific IGL/F signature motif, which is essential for cooperation of other Hsp100 proteins with the peptidase ClpP. This feature is shared with the canonical ClpB disaggregase [12] suggesting that protein disaggregation is primarily linked to protein refolding.”.

References

(1) Katikaridis P, Simon B, Jenne T, Moon S, Lee C, Hennig J, et al. Structural basis of aggregate binding by the AAA+ disaggregase ClpG. J Biol Chem. 2023:105336.

(2) Glover JR, Lindquist S. Hsp104, Hsp70, and Hsp40: A novel chaperone system that rescues previously aggregated proteins. Cell. 1998;94:73-82.

(3) Krzewska J, Langer T, Liberek K. Mitochondrial Hsp78, a member of the Clp/Hsp100 family in Saccharomyces cerevisiae, cooperates with Hsp70 in protein refolding. FEBS Lett. 2001;489:92-6.

(4) Seyffer F, Kummer E, Oguchi Y, Winkler J, Kumar M, Zahn R, et al. Hsp70 proteins bind Hsp100 regulatory M domains to activate AAA+ disaggregase at aggregate surfaces. Nat Struct Mol Biol. 2012;19:1347-55.

(5) Haslberger T, Zdanowicz A, Brand I, Kirstein J, Turgay K, Mogk A, et al. Protein disaggregation by the AAA+ chaperone ClpB involves partial threading of looped polypeptide segments. Nat Struct Mol Biol. 2008;15:641-50.

(6) Theyssen H, Schuster H-P, Bukau B, Reinstein J. The second step of ATP binding to DnaK induces peptide release. J Mol Biol. 1996;263:657-70.

(7) Iljina M, Mazal H, Goloubinoff P, Riven I, Haran G. Entropic Inhibition: How the Activity of a AAA+ Machine Is Modulated by Its Substrate-Binding Domain. ACS chemical biology. 2021;16:775-85.

(8) Rosenzweig R, Farber P, Velyvis A, Rennella E, Latham MP, Kay LE. ClpB N-terminal domain plays a regulatory role in protein disaggregation. Proc Natl Acad Sci U S A. 2015;112:E6872-81.

(9) Barnett ME, Nagy M, Kedzierska S, Zolkiewski M. The amino-terminal domain of ClpB supports binding to strongly aggregated proteins. J Biol Chem. 2005;280:34940-5.

(10) Beinker P, Schlee S, Groemping Y, Seidel R, Reinstein J. The N Terminus of ClpB from Thermus thermophilus Is Not Essential for the Chaperone Activity. J Biol Chem. 2002;277:47160-6.

(11) Mogk A, Schlieker C, Strub C, Rist W, Weibezahn J, Bukau B. Roles of individual domains and conserved motifs of the AAA+ chaperone ClpB in oligomerization, ATP-hydrolysis and chaperone activity. J Biol Chem. 2003;278:15-24.

(11) Weibezahn J, Tessarz P, Schlieker C, Zahn R, Maglica Z, Lee S, et al. Thermotolerance Requires Refolding of Aggregated Proteins by Substrate Translocation through the Central Pore of ClpB. Cell. 2004;119:653-65.

Reviewer #1 (Public Review):

Summary:

This work describes the mechanism of protein disaggregation by the ClpL AAA+ protein of Listeria monocytogenes. Using several model subtrate proteins the authors first show that ClpL possesses a robust disaggregase activity that does not further require the endogenous DnaK chaperone in vitro. In addition, they found that ClpL is more thermostable than the endogenous L. monocytogenes DnaK and has the capacity to unfold tightly folded protein domains. The mechanistic basis for the robust disaggregase activity of ClpL was also dissected in vitro and in some cases, supported by in vivo data performed in chaperone-deficient E. coli strains. The data presented show that the two AAA domains, the pore-2 site and the N-terminal domain (NTD) of ClpL are critical for its disaggregase activity. Remarkably, grafting the NTD of ClpL to ClpB converted ClpB into an autonomous disaggregase, highlighting the importance of such a domain in the DnaK-independent disaggregation of proteins. The role of the ClpL NTD domain was further dissected, identifying key residues and positions necessary for aggregates recognition and disaggregation. Finally, using sets of SEC and negative staining EM experiments combined with conditional covalent linkages and disaggregation assays the authors found that ClpL shows significant structural plasticity, forming dynamic hexameric and heptameric active single rings that can further form higher assembly states via their middle domains.

Strengths:

The manuscript is well written and the experimental work well executed. It contains a robust and complete set of in vitro data that push further our knowledge of such important disaggregases. It shows the importance of the atypical ClpL N-terminal domain in the disaggregation process as well as the structural malleability of such AAA+ proteins. More generally, this work expands our knowledge of heat resistance in bacterial pathogens.

Weaknesses:

There is no specific weakness in this work, although it would have helped to have a drawing model showing how ClpL performs protein disaggregation based on their new findings. The function of the higher assembly states of ClpL remains unresolved and will need further extensive research. Similarly, it will be interesting in the future to see whether the sole function of the plasmid encoded ClpL is to cope with general protein aggregates under heat stress.

Reviewer #2 (Public Review):

The manuscript by Bohl et al. is an interesting and carefully done study on the biochemical properties and mode of action of potent autonomous AAA+ disaggregase ClpL from Listeria monocytogenes. ClpL is encoded on plasmids. It shows high thermal stability and provides Listeria monocytogenes food-pathogen substantial increase in resistance to heat. The authors show that ClpL interacts with aggregated proteins through the aromatic residues present in its N-terminal domain and subsequently unfolds proteins from aggregates translocating polypeptide chains through the central pore in its oligomeric ring structure. The structure of ClpL oligomers was also investigated in the manuscript. The results suggest that mono-ring structure and not dimer or tetramer of rings, observed in addition to mono-ring structures under EM, is an active specie of disaggregase. In the revised version additional data is presented suggesting that dimer or tetramer of ClpL rings play a protective role in cell by restricting ClpL activity.

Presented experiments are conclusive and well controlled. I think the presentation and discussion of results are better in revised version.<br /> The study's strength lies in the direct comparison of ClpL biochemical properties with autonomous ClpG disaggregase present in selected Gram-negative bacteria and well-studied E. coli system consisting of ClpB disaggregase and DnaK and its cochaperones. This puts the results in a broader context.

Reviewer #3 (Public Review):

Summary:

This manuscript details the characterization of ClpL from L. monocytogenes as a potent and autonomous AAA+ disaggregase. The authors demonstrate that ClpL has potent and DnaK-independent disaggregase activity towards a variety of aggregated model substrates, and that this disaggregase activity appears to be greater than that observed with the canonical DnaK/ClpB co-chaperone. Furthermore, LmClpL appears to have greater thermostability as compared to LmDnaK, suggesting that ClpL-expressing cells may be able to withstand more severe heat stress conditions. Interestingly, LmClpL can provide thermotolerance to E. coli that have been genetically depleted of either ClpB or in cells expressing a mutant DnaK103. The authors further characterized the mechanisms by which ClpL interacts with protein aggregates, identifying that the N-terminal domain of ClpL is essential for disaggregase function. Lastly, by EM and mutagenesis analysis the authors report that ClpL can exist in a variety of larger macromolecular complexes, including dimer or trimers of hexamers/heptamers, and they provide evidence that the N-terminal domains of ClpL prevent dimer ring formation, thus promoting an active and substrate-binding ClpL complex. Throughout this manuscript the authors compare LmClpL to ClpG, another potent and autonomous disaggregase found in gram-negative bacteria that has been reported on previously, demonstrating that these two enzymes share homologous activity and qualities. Taken together this report clearly establishes ClpL as a novel and autonomous disaggregase.

Analysis:

The work presented in this report amounts to a significant body of novel and significant work that will be of interest to protein chaperone community. Furthermore, by providing examples of how ClpL can provide in vivo thermotolerance to both E. coli and L. gasseri the authors have expanded the significance of this work and provides novel insight into potential mechanisms responsible for thermotolerance in food-borne pathogens. The figures are clearly depicted, well-labeled, and easy to understand, and the manuscript is well-written. Experimentally the work was performed to a high standard with excellent controls, aiding in the ability for the audience to understand the major findings and conclusions. Additionally, the authors have effectively and efficiently expanded on their work through the peer review process, further increasing the understandability and significance of their work. Overall, the data presented, and analysis thereof, support the authors' conclusions, and thus this study represents an important addition to our understanding of molecular chaperone biochemistry. Lastly, this study establishes new avenues for research into autonomous disaggregates, their role in in vivo thermotolerance, and the mechanisms by which AAA+ chaperones recognize and interact with substrate proteins.

Author Response:

Reviewer #1 (Public Review):

Summary:

The authors identified that genetically and pharmacological inhibition of CERS1, an enzyme implicated in ceramides biosynthesis worsen muscle fibrosis and inflammation during aging.<br /> Strengths:

The study points out an interesting issue on excluding CERS1 inhibition as a therapeutic strategy for sarcopenia. Overall, the article it's well written and clear.<br /> Weaknesses:

Many of the experiments confirmed previous published data, which also show a decline of CERS1 in ageing and the generation and characterization of a muscle specific knockout mouse line. The mechanistic insights of how the increased amount of long ceramides (cer c24) and the decreased of shorter ones (cer c18) might influence muscle mass, force production, fibrosis and inflammation in aged mice have not been addressed.

We thank the reviewer for the assessment and would like to point out that Cers1 had not previously been studied in the context of aging. Moreover, our unbiased pathway analyses in human skeletal muscle implicate CERS1 for the first time with myogenic differentiation, which we validate in cell culture systems. To improve mechanistic insights, as suggested by Reviewer #1, we performed more experiments to gain insights how Cers1 derived c18, and Cers2 derived c24 ceramide species affect myogenesis. We recently showed that knocking out Cers2 reduces c24:0/c24:1 and promotes muscle cell maturation (PMID: 37118545, Fig. 6m-r and Supplementary Fig. 5e). This suggests that the very long chain ceramides c24 might indeed be driving the effect we see upon Cers1 inhibition because we observe an accumulation of c24 ceramides upon Cers1 (c18) inhibition (Fig 2B, Fig 3B, Fig 4A, Fig S3E), which is associated with impaired muscle maturation (Fig 4B-C, Fig S3G-I, Fig S4G-I). To study whether impaired muscle cell differentiation upon Cers1 inhibition is dependent on Cers2, we knocked-down Cers1 alone, or in combination with the knockdown of Cers2. Results show that reduced muscle cell maturation mediated by Cers1KD is rescued by the simultaneous knockdown of Cers2 as shown by gene expression analyses and immunohistochemical validation and quantification. Hence, we believe that reducing Cers1 function during aging might lead to an increase in sphingosine levels as has been shown previously (PMID: 31692231). Increased sphingosine triggers cell apoptosis due to its toxicity (PMID: 12531554). Therefore, channeling accumulating sphingosine towards C24 ceramides may avoid toxicity but, as we show in this manuscript, will reduce the myogenic potential in muscle. However, if also C24 production is blocked by Cers2 inhibition, sphingosine is forced towards the production of other, potentially less toxic or myogenesis-impairing ceramides. We added these new data to the revised manuscript as new Fig 5D-E and new Fig S5G-I.

Reviewer #2 (Public Review):

Summary:

The manuscript by Wohlwend et al. investigates the implications of inhibiting ceramide synthase Cers1 on skeletal muscle function during aging. The authors propose a role for Cers1 in muscle myogenesis and aging sarcopenia. Both pharmacological and AAV-driven genetic inhibition of Cers1 in 18month-old mice lead to reduced C18 ceramides in skeletal muscle, exacerbating age-dependent features such as muscle atrophy, fibrosis, and center-nucleated fibers. Similarly, inhibition of the Cers1 orthologue in C. elegans reduces motility and causes alterations in muscle morphology.<br /> Strengths:

The study is well-designed, carefully executed, and provides highly informative and novel findings that are relevant to the field.

Weaknesses:

The following points should be addressed to support the conclusions of the manuscript.

(1) It would be essential to investigate whether P053 treatment of young mice induces age-dependent features besides muscle loss, such as muscle fibrosis or regeneration. This would help determine whether the exacerbation of age-dependent features solely depends on Cers1 inhibition or is associated with other factors related to age- dependent decline in cell function. Additionally, considering the reported role of Cers1 in whole-body adiposity, it is necessary to present data on mice body weight and fat mass in P053treated aged-mice.

We thank the reviewer to suggest that we study Cers1 inhibition in young mice. In fact, a previous study shows that muscle-specific Cers1 knockout in young mice impairs muscle function (PMID: 31692231). Similar to our observation, these authors report reduced muscle fiber size and muscle force. Therefore, we do not believe that our observed effects of Cers1 inhibition in aged mice are specific to aging, although the phenotypic consequences are accentuated in aged mice. As requested by the reviewer, we attached the mice body weights and fat mass (Author response image 1A-B). The reduced fat mass upon P053 treatment is in line with previously reported reductions in fat mass in chow diet or high fat diet fed young mice upon Cers1 inhibition (PMID: 30605666, PMID: 30131496), again suggesting that the effect of Cers1 inhibition might not be specific to aging.

Author response image 1.

(A-B) Body mass (A) and Fat mass as % of body mass (B) were measured in 22mo C57BL/6J mice intraperitoneally injected with DMSO or P053 using EchoMRI (n=7-12 per group). (C-D) Grip strengh measurements in all limbs (C) or only the forelimbs (D) in 24mo C57BL/6J mice intramuscularly injected with AAV9 particles containing scramble, or shRNA targeting Cers1 (n=8 per group). (E-F) Pax7 gene expression in P053 or AAV9 treated mice (n=6-7 per group) (E), or in mouse C2C12 muscle progenitor cells treated with 25nM scramble or Cers1 targeting shRNA (n=8 per group) (F). (G) Proliferation as measured by luciferase intensity in mouse C2C12 muscle muscle cells treated with 25nM scramble or Cers1 targeting shRNA (n=24 per group). Each column represents one biological replicate. (H) Overlayed FACS traces of Annexin-V (BB515, left) and Propidium Iodide (Cy5, right) of mouse C2C12 muscle myotubes treated with 25nM scramble or Cers1 targeting shRNA (n=3 per group). Quantification right: early apoptosis (Annexin+-PI-), late apoptosis (Annexin+-PI+), necrosis (Annexin--PI+), viability (Annexin--PI-). (I) Normalized Cers2 gene expression in mouse C2C12 muscle muscle cells treated with 25nM scramble or Cers1 targeting shRNA (n=6-7 per group). (J-K) Representative mitochondrial respiration traces of digitonin-permeablized mouse C2C12 muscle muscle cells treated DMSO or P053 (J) with quantification of basal, ATP-linked, proton leak respiration as well as spare capacity and maximal capacity linked respiration (n=4 per group). (L) Reactive oxygen production in mitochondria of mouse C2C12 muscle muscle cells treated DMSO or P053. (M) Enriched gene sets related to autophagy and mitophagy in 24mo C57BL/6J mouse muscles intramuscularly injected with AAV9 particles containing scramble, or shRNA targeting Cers1 (left), or intraperitoneally injected with DMSO or P053 (right). Color gradient indicates normalized effect size. Dot size indicates statistical significance (n=6-8 per group). (N) Representative confocal Proteostat® stainings with quantifications of DMSO and P053 treated mouse muscle cells expressing APPSWE (top) and human primary myoblasts isolated from patients with inclusion body myositis (bottom). (O) Stillness duration during a 90 seconds interval in adult day 5 C. elegans treated with DMSO or 100uM P053. (P) Lifespan of C. elegans treated with DMSO or P053. (n=144-147 per group, for method details see main manuscript page 10).

(2) As grip and exercise performance tests evaluate muscle function across several muscles, it is not evident how intramuscular AAV-mediated Cers1 inhibition solely in the gastrocnemius muscle can have a systemic effect or impact different muscles. This point requires clarification.

The grip strength measurements presented in the manuscript come from hindlimb grip strength, as pointed out in the Methods section. We measured grip strength in all four limbs, as well as only fore- (Author response image 1C-D). While forelimb strength did not change, only hindlimb grip strength was significantly different in AAV-Cers1KD compared to the scramble control AAV (Fig 3I), which is in line with the fact that we only injected the AAV in the hindlimbs. This is similar to the effect we observed with our previous data where we saw altered muscle function upon IM AAV delivery in the gastrocnemius (PMID: PMID: 34878822, PMID: 37118545). The gastrocnemius likely has the largest contribution to hindlimb grip strength given its size, and possibly even overall grip strength as suggested by a trend of reduced grip strength in all four limbs (Author response image 1C). We also suspect that the hindlimb muscles have the largest contribution to uphill running as we could also see an effect on running performance. While we carefully injected a minimal amount of AAV into gastrocnemius to avoid leakage, we cannot completely rule out that some AAV might have spread to other muscles. We added this information to the discussion of the manuscript as a potential limitation of the study.

(3) To further substantiate the role of Cers1 in myogenesis, it would be crucial to investigate the consequences of Cers1 inhibition under conditions of muscle damage, such as cardiotoxin treatment or eccentric exercise.<br /> While it would be interesting to study Cers1 in the context of muscle regeneration, and possibly mouse models of muscular dystrophy, we think such work would go beyond the scope of the current manuscript.

(4) It would be informative to determine whether the muscle defects are primarily dependent on the reduction of C18-ceramides or the compensatory increase of C24-ceramides or C24-dihydroceramides.

To improve mechanistic insights, as suggested by Reviewer #2, we performed more experiments to gain insights how Cers1 derived c18, and Cers2 derived c24 ceramide species affect myogenesis. We recently showed that knocking out Cers2 reduces c24:0/c24:1 and promotes muscle cell maturation (PMID: 37118545, Fig. 6m-r and Supplementary Fig. 5e). This suggests that the very long chain ceramides c24 might indeed be driving the effect we see upon Cers1 inhibition because we observe an accumulation of c24 ceramides upon Cers1 (c18) inhibition (Fig 2B, Fig 3B, Fig 4A, Fig S3E), which is associated with impaired muscle maturation (Fig 4B-C, Fig S3G-I, Fig S4G-I). To study whether impaired muscle cell differentiation upon Cers1 inhibition is dependent on Cers2, we knocked-down Cers1 alone, or in combination with the knockdown of Cers2. Results show that reduced muscle cell maturation mediated by Cers1KD is rescued by the simultaneous knockdown of Cers2 as shown by gene expression analyses and immunohistochemical validation and quantification. We added these data to the manuscript as new Fig 5D-E, new Fig S5G-I. These data, together with our previous results showing that Degs1 knockout reduces myogenesis (PMID: 37118545, Fig. 6s-x and Fig. 7) suggest that C24/dhC24 might contribute to the age-related impairments in myogenesis. We added the new results to the revised manuscript.

(5) Previous studies from the research group (PMID 37118545) have shown that inhibiting the de novo sphingolipid pathway by blocking SPLC1-3 with myriocin counteracts muscle loss and that C18-ceramides increase during aging. In light of the current findings, certain issues need clarification and discussion. For instance, how would myriocin treatment, which reduces Cers1 activity because of the upstream inhibition of the pathway, have a positive effect on muscle? Additionally, it is essential to explain the association between the reduction of Cers1 gene expression with aging (Fig. 1B) and the age-dependent increase in C18-ceramides (PMID 37118545).

Blocking the upstream enzyme of the ceramide pathway (SPT1) shuts down the entire pathway that is overactive in aging, and therefore seems beneficial for muscle aging. While most enzymes in the ceramide pathway that we studied so far (SPTLC1, CERS2) revealed muscle benefits in terms of myogenesis, inflammation (PMID: 35089797; PMID: 37118545) and muscle protein aggregation (PMID: 37196064), the CERS1 enzyme shows opposite effects. This is also visible in the direction of CERS1 expression compared to the other enzymes in one of our previous published studies (PMID: 37118545, Fig. 1e and Fig. 1f). In the current study, we show that Cers1 inhibition indeed exacerbates age-related myogenesis and inflammation as opposed to the inhibition of Sptlc1 or Cers2. As the reviewer points out, both C18- and C24-ceramides seem to accumulate upon muscle aging. We think this is due to an overall overactive ceramide biosynthesis pathway. Blocking C18-ceramides via Cers1 inhibition results in the accumulates C24-ceramides and worsens muscle phenotypes (see reply to question #4). On the other hand, blocking C24-ceramides via Cers2 inhibition improves muscle differentiation. These observations together with the finding that Cers1 mediated inhibition of muscle differentiation is dependent on proper Cers2 function (new Fig 5D-E, new Fig S5G-I) points towards C24-ceramides as the main culprit of reduced muscle differentiation. Hence, at least a significant part of the benefits of blocking SPTLC1 might have been related to reducing very long-chain ceramides. We believe that reduced Cers1 expression in skeletal muscle upon aging, observed by us and others (PMID: 31692231), might reflect a compensatory mechanism to make up for an overall overactive ceramide flux in aged muscles. Reducing Cers1 function during aging might lead to an increase in sphingosine levels as has been shown previously (PMID: 31692231). Increased sphingosine triggers cell apoptosis due to its toxicity (PMID: 12531554). Therefore, channeling accumulating sphingosine towards C24 ceramides may avoid toxicity but, as we show in this manuscript, will reduce the myogenic potential in muscle. However, if also C24 production is blocked by Cers2 inhibition (new Fig 5E-D, new Fig S5G-I), sphingosine is forced towards the production of other, potentially less toxic, or myogenesis-impairing ceramides. These data are now added to the revised manuscript (see page 7). Details were added to the discussion of the manuscript (see page 8).

Addressing these points will strengthen the manuscript's conclusions and provide a more comprehensive understanding of the role of Cers1 in skeletal muscle function during aging.

Reviewer #1 (Recommendations For The Authors):

The authors identified that genetical and pharmacological inhibition of CERS1, an enzyme implicated in ceramides biosynthesis worsen muscle fibrosis and inflammation during aging.

Even though many of the experiments only confirmed previous published data (ref 21, 11,37,38), which also show a decline of CERS1 in ageing and the generation and characterization of a muscle specific knockout mouse line, the study points out an interesting issue on excluding CERS1 inhibition as a therapeutic strategy for sarcopenia and opens new questions on understanding how inhibition of SPTLC1 (upstream CERS1) have beneficial effects in healthy aging (ref 15 published by the same authors).

Overall, the article it's well written and clear. However, there is a major weakness. The mechanistic insights of how the increased amount of long ceramides (c24) and the decreased of shorter ones (cer c18) might influence muscle mass, force production, fibrosis and inflammation in aged mice have not been addressed. At the present stage the manuscript is descriptive and confirmatory of CERS1 mediated function in preserving muscle mass. The authors should consider the following points:

Comments:

(1) Muscle data

(a) The effect of CERS1 inhibition on myotube formation must be better characterized. Which step of myogenesis is affected? Is stem cell renewal or MyoD replication/differentiation, or myoblast fusion or an increased cell death the major culprit of the small myotubes? Minor point: Figure S1C: show C14:00 level at 200 h; text of Fig S2A and 1F: MRF4 and Myogenin are not an early gene in myogenesis please correct, Fig S2B and 2C: changes in transcript does not mean changes in protein or myotube differentiation and therefore, authors must test myotube formation and myosin expression.

Cers1 inhibition seems to affect differentiation and myoblast fusion. To test other suggested effects we performed more experiments as delineated. Inhibiting Cers1 systemically with the pharmacological inhibitor of Cers1 (P053) or with intramuscular delivery of AAV expressing a short hairpin RNA (shRNA) against Cers1 in mice did not affect Pax7 transcript levels (Author response image 1E). Moreover, we did also not observe an effect of shRNA targeting Cers1 on Pax7 levels in mouse C2C12 muscle progenitor cells (Author response image 1F). To characterize the effect of Cers1 inhibition on muscle progenitor proliferation/renewal, we used scramble shRNA, or shRNA targeting Cers1 in C2C12 muscle progenitors and measured proliferation using CellTiter-Glo (Promega). Results showed that Cers1KD had no significant effect on cell proliferation (Author response image 1G). Next, we assayed cell death in differentiating C2C12 myotubes deficient in Cers1 using FACS Analysis of Annexin V (left) and propidium iodide (right). We found no difference in early apoptosis, late apoptosis, necrosis, or muscle cell viability, suggesting that cell death can be ruled out to explain smaller myotubes (Author response image 1H). These findings support the notion that the inhibitory effect of Cers1 knockdown on muscle maturation are primarily based on effects on myogenesis rather than on apoptosis. Our data in the manuscript also suggests that Cers1 inhibition affects myoblast fusion, as shown by reduced myonucleation upon Cers1KD (Fig S3H right, Fig S5I).

(b) The phenotype of CESR1 knockdown is milder than 0P53 treated mice (Fig S5D and Figure 3F, 3H are not significant) despite similar changes of Cer18:0, Cer24:0, Cer 24:1 concentration in muscles . Why?

Increases in very long chain ceramides were in fact larger upon P053 administration compared to AAVmediated knockdown. For example, Cer24:0 levels increased by >50% upon P053 administration, compared to 20% by AAV injections. Moreover, dhC24:1 increased by 6.5-fold vs 2.5-fold upon P053 vs AAV treatment, respectively. These differences might not only explain the slightly attenuated phenotypes in the AA- treated mice but also underlines the notion that very long chain ceramides might cause muscle deterioration. We believe inhibiting the enzymatic activity of Cers1 (P053) as compared to degrading Cers1 transcripts is a more efficient strategy to reduce ceramide levels. However, we cannot completely rule out multi-organ, systemic effects of P053 treatment beyond its direct effect on muscle. We added these details in the discussion of the revised manuscript (see page 8 of the revised manuscript).

(c) The authors talk about a possible compensation of CERS2 isoform but they never showed mRNA expression levels or CERS2 protein levels aner treatment. Is CERS2 higher expressed when CERS1 is downregulated in skeletal muscle?

We appreciate the suggestion of the reviewer. We found no change in Cers2 mRNA levels upon Cers1 inhibition in mouse C2C12 myoblasts (Author response image 1I). We would like to point out that mRNA abundance might not be the optimal measurement for enzymes due to enzymatic activities. Therefore, we think metabolite levels are a better proxy of enzymatic activity. It should also be pointed out that “compensation” might not be an accurate description as sphingoid base substrate might simply be more available upon Cers1KD and hence, more substrate might be present for Cers2 to synthesize very long chain ceramides. This “re-routing” has been previously described in the literature and hypothesized to be related to avoid toxic (dh)sphingosine accumulation (PMID: 30131496). Therefore, we changed the wording in the revised manuscript to be more precise.

(d) Force measurement of AAV CERS1 downregulated muscles could be a plus for the study (assay function of contractility)

In the current study we measured grip strength in mice, which had previously been shown to be a good proxy of muscle strength and general health (PMID: 31631989). Indeed, our results of reduced muscle grip strength are in line with previous work that shows reduced contractility in muscles of Cers1 deficient mice (PMID: 31692231).

(e) How are degradation pathways affected by the downregulation of CERS1. Is autophagy/mitophagy affected? How is mTOR and protein synthesis affected? There is a recent paper that showed that CerS1 silencing leads to a reduction in C18:0-Cer content, with a subsequent increase in the activity of the insulin pathway, and an improvement in skeletal muscle glucose uptake. Could be possible that CERS1 downregulation increases mTOR signalling and decreases autophagy pathway? Autophagic flux using colchicine in vivo would be useful to answer this hypothesis

Cers1 in skeletal muscle has indeed been linked to metabolic homeostasis (see PMID: 30605666). In line with their finding in young mice we also find reduced fat mass upon P053 treatment in aged mice (Author response image 1A-B). We also looked into mitochondrial bioenergetics upon blocking Cers1 with P053 treatment using an O2k oxygraphy (Author response image 1J-L). Results show that Cers1 inhibition in mouse muscle cells increases mitochondrial respiration, similar to what has been shown before (PMID: 30131496). However, we also found that reactive oxygen species production in mouse muscle cells is increased upon P053 treatment, suggesting the presence of dysfunctional mitochondria upon inhibiting Cers1 with P053.We next looked into the mitophagy/autophagy degradation pathways suggested by the reviewer and do not find convincing evidence supporting that Cers1 has a major impact on autophagy or mitophagy derived gene sets in mice treated with shRNA against Cers1, or the Cers1 pharmacological inhibitor P053 (Author response image 1M).

We then assessed the effect of Cers1 inhibition on transcripts levels related to the mTORC1/protein synthesis, as suggested by the reviewer. Cers1 knockdown in differentiating mouse muscle cells showed only a weak trend to reduce mTORC1 and its downstream targets (new Fig S4A). In line with this, there was no notable difference in protein synthesis in differentiating, Cers1 deficient mouse C2C12 myoblasts as assessed by L-homopropargylglycine (HPG) amino acid labeling using confocal microscopy (new Fig S4B) or FACS analyses (new Fig S4C). However, Cers1KD increased transcripts related to the myostatin-Foxo1 axis as well as the ubiquitin proteasome system (e.g. atrogin-1, MuRF1) (new Fig S4D), suggesting Cers1 inhibition increases protein degradation. We added these details to the revised manuscript on page 7. We recently implicated the ceramide pathway in regulating muscle protein homeostasis (PMID: 37196064). Therefore, we assessed the effect of Cers1 inhibition with the P053 pharmacological inhibitor on protein folding in muscle cells using the Proteostat dye that intercalates into the cross-beta spine of quaternary protein structures typically found in misfolded and aggregated proteins. Interestingly, inhibiting Cers1 further increased misfolded proteins in C2C12 mouse myoblasts expressing the Swedish mutation in APP and human myoblasts isolated from patients with inclusion body myositis (Author response imageure 1N). These findings suggest that deficient Cers1 might upregulate protein degradation to compensate for the accumulation of misfolded and aggregating proteins, which might contribute to impaired muscle function observed upon Cers1 knockdown. Further studies are needed to disentangle the underlying mechanstics.

(f) The balances of ceramides have been found to play roles in mitophagy and fission with an impact on cell fate and metabolism. Did the authors check how are mitochondria morphology, mitophagy or how dynamics of mitochondria are altered in CERS1 knockdown muscles? (fission and fusion). There is growing evidence relating mitochondrial dysfunction to the contribution of the development of fibrosis and inflammation.

Previously, CERS1 has been studied in the context of metabolism and mitochondria (for reference, please see PMID: 26739815, PMID: 29415895, PMID: 30605666, PMID: 30131496). In summary, these studies demonstrate that C18 ceramide levels are inversely related to insulin sensitivity in muscle and mitochondria, and that Cers1 inhibition improves insulin-stimulated suppression of hepatic glucose production and reduced high-fat diet induced adiposity. Moreover, improved mitochondrial respiration, citrate synthase activity and increased energy expenditure were reported upon Cers1 inhibition. Lack of Cers1 specifically in skeletal muscle was also reported to improve systemic glucose homeostasis. While these studies agree on the effect of Cers1 inhibition on fat loss, results on glucose homeostasis and insulin sensitivity differ depending on whether a pharmacologic or a genetic approach was used to inhibit Cers1. The current manuscript describes the effect of CERS1 on muscle function and myogenesis because these were the most strongly correlated pathways with CERS1 in human skeletal muscle (Fig 1C) and impact of Cers1 on these pathways is poorly studied, particularly in the context of aging. Therefore, we would like to refer to the mentioned studies investigating the effect of CERS1 on mitochondria and metabolism.

(2) C.elegans data:

(a) The authors checked maternal RNAi protocol to knockdown lagr-1 and showed alteration of muscle morphology at day 5. They also give pharmacological exposure of P053 drug at L4 stage. Furthermore, the authors also used a transgenic ortholog lagr-1 to perform the experiments. All of them were consistent showing a reduced movement. It would be important to show rescue of the muscle phenotype by overexpressing CERS1 ortholog in knockdown transgenic animals.

We used RNAi to knockdown the Cers1 orthologue, lagr-1, in C.elegans. Therefore, we do not have transgenic animals. Overexpressing lagr-1 in the RNAi treated animals would also not be possible as the RNA from the overexpression would just get degraded.

(b) The authors showed data about distance of C.elegans. It would be interesting to specify if body bends, reversals and stillness are affected in RNAi and transgenic Knockdown worms.

As suggested, we measured trashing and stillness as suggested by the reviewer and found reduced trashing (new Fig S5B) and a trend towards an increase in stillness (Author response image 1O) in P053 treated worms on day 5 of adulthood, which is the day we observed significant differences in muscle morphology and movement (Fig 4D-E, Fig S5A). These data are now included in the revised manuscript.

(c) Is there an effect on lifespan extension by knocking down CERS1?

We performed two independent lifespan experiments in C.elegans treated with the Cers1 inhibitor P053 and found reduced lifespan in both replicate experiments (for second replicate, see Author response image 1P). We added these data to the revised manuscript as new Fig 4H.

How do the authors explain the beneficial effect of sptlc1 inhibition on healthy aging muscle? Discuss more during the article if there is no possible explanation at the moment.

We believe that blocking the upstream enzyme of the ceramide pathway (SPT1) shuts down the entire pathway that is overactive in aging, and therefore is more beneficial for muscle aging. Our current work suggests that at least a significant part of Sptlc1-KD benefits might stem from blocking very long chain ceramides. While SPTLC1 and CERS2 revealed muscle benefits in terms of myogenesis, inflammation (PMID: 35089797; PMID: 37118545) and muscle protein aggregation (PMID: 37196064), the CERS1 enzyme shows opposite effects, which is also visible in Fig 1e and Fig 1f of PMID: 37118545. In the current study, we show that Cers1 inhibition indeed exacerbates aging defects in myogenesis and inflammation as opposed to the inhibition of Sptlc1 or Cers2. The fact that the effect of Cers1 on inhibiting muscle differentiation is dependent on the clearance of Cers2-derived C24-ceramides suggests that reducing very long chain ceramides might be crucial for healthy muscle aging. We added details to the discussion.

Author Response

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

Public Review:

Summary:

This paper reports how mycobacterial cAMP level is increased under stressful conditions and that the increase is important in the survival of the bacterium in animal hosts.

Strengths:

The authors show that under different stresses the response regulator PhoP represses a phosphodiesterase (PDE) that degrades cAMP specifically. Identification of a PDE specific to cAMP is significant progress in understanding Mtb pathogenesis. An increase in cAMP apparently increases bacterial survival upon infection. On the practical side, the reduction of cAMP by increasing PDE can be a means to attenuate the growth of the bacilli. The results have wider implications since PhoP is implicated in controlling diverse mycobacterial stress responses and many bacterial pathogens modulate host cell cAMP level. The results here are straightforward, internally consistent, and of both theoretical and applied interests.

We thank the reviewers for these extremely encouraging comments.

Weaknesses:

Repression of PDE promoter by binding of phosphorylated PhoP could have been shown at higher precision. The binding is now somewhere along a roughly 500 bp region. Although the regulation of PDE is shown to be by transcriptional repression only, it has been described as a homeostatic mechanism. The latter would have required a demonstration of both repression and activation by negative feedback.

We agree. We have now performed EMSA (Electrophoretic Mobility Shift Assay) experiments and included the data showing DNA binding of PhoP to the upstream regulatory region of rv0805 (rv0805up) as a supplemental figure (see Figure 2-figure supplement 1). The supplemental figure, figure caption, and the relevant results have been adjusted accordingly in the revised manuscript.

Further, as recommended by the reviewer we have now removed the term ‘homeostatic mechanism’ and rephrased it with ‘maintenance of cAMP level’ in the manuscript.

Response to Reviewers’ comments

Reviewer #1:

The authors have used homeostasis inappropriately. Homeostasis usually requires negative feedback (a clear example is the regulation of Lambda prm promoter). Here, there is no feedback from changes in PDE or cAMP level to their synthesis. Homeostasis does not belong to this paper anywhere.

As recommended by the reviewer, we have now removed “homeostasis” from the manuscript and mostly replaced it with “maintenance of cAMP level” in the revised manuscript.

The authors have frequently used adverbs at the beginning of a sentence, such as Notably (l.240, 272, 376), Importantly (l.66, 213), More importantly (l.134), Remarkably (l.264), Interestingly (l.115,301), Intriguingly (l.344), unambiguously (l.347), etc. The use of these words is generally counter-productive. The authors should scan the ms. to eliminate them as far as possible. The sentences would read more clearly and become more impactful.

Following reviewer’s recommendation, we have now eliminated most of the adverbs, mostly used at the beginning of sentences, in the revised manuscript.

Specific comments

(1) L.1: "maintenance of homeostasis" or increasing cAMP level.

As suggested by the reviewer, we have now replaced “maintenance of cAMP homeostasis” with “maintenance of cAMP level”.

(2) L.27: mechanism or reason; varying or various.

As recommended by the reviewer, we have now replaced “mechanism” with “reason” and the word “varying” is deleted while incorporating suggested changes in the abstract.

(3) L.28-29: The logic of connecting PhoP to cAMP doesn't follow well. The logic is much better in l.54, l.112-5 and l.130.

We thank the reviewer for this suggestion. We have now modified the statement within the ‘abstract’ in the revised manuscript (duplicated below):

“cAMP is one of the most widely used second messengers which impacts on a wide range of cellular responses in microbial pathogens including M. tuberculosis. Herein, we hypothesized that intra-mycobacterial cAMP level could be controlled by the phoP locus since the major regulator plays a key role in bacterial response against numerous stress conditions.”

(4) L.30: discovers or reveals (?). Also, in l.101.

As recommended by the reviewer, we have now replaced ‘discovers’ with ‘reveals’ in the Abstract and ‘uncovered’ with ‘revealed’ in the Introduction section of the manuscript.

(5) L.31: Delete "The most - - derived". It is not obvious what most fundamental means here. I suggest: We find that PhoP-dependent ---involves specific binding of the regulator---PDE gene.

As recommended by the reviewer, we have modified the statement (duplicated below): “In keeping with these results, we find specific recruitment of the regulator within the promoter region of rv0805 PDE, and absence of phoP or ectopic expression of rv0805 independently accounts for elevated PDE synthesis leading to depletion of intra-mycobacterial cAMP level.”

(6) L.36: --pathway decreases cAMP level, stress tolerance, and survival of the bacilli.

As recommended by the reviewer, we have now modified the statement (duplicated below): “Thus, genetic manipulation to inactivate PhoP-Rv0805-cAMP pathway decreases cAMP level, stress tolerance, and intracellular survival of the bacilli.

(7) L.41: 'keeps encountering" or encounters?

As suggested by the reviewer, we have replaced ‘keeps encountering’ with ‘encounters’ in the ‘Introduction’ section of the revised manuscript.

(8) L.61: responds, carries.

Our apologies for the embarrassing grammatical mistakes. We have rectified these errors in the revised manuscript.

(9) L.67: you mean burst in synthesis level, not burst of cAMP itself.

To improve clarity, we have now modified the statement in the revised manuscript (duplicated below): “Agarwal and colleagues had shown that burst in synthesis of bacterial cAMP upon infection of macrophages, improved bacterial survival by interfering with host signalling pathways (Agarwal et al., 2009)”

Reference

Agarwal N, Lamichhane G, Gupta R, Nolan S, Bishai WR (2009) Cyclic AMP intoxication of macrophages by a Mycobacterium tuberculosis adenylate cyclase. Nature 460: 98-102

(10) L.77: Change Off to Of.

We are sorry for the inaccuracy. The suggested change has been made to the text.

(11) L.83: Did not discuss "degradation" earlier.

Following reviewer’s recommendation, we have now modified the statement in the revised manuscript (duplicated below).

“Together, these results strongly suggest that a balance between cAMP synthesis by adenylate cyclases and cAMP degradation by phosphodiesterases contributes to rapid adaptive response of mycobacteria in a hostile intracellular environment (Johnson and McDonough, 2018; McDonough and Rodriguez, 2011).”

Reference

Johnson RM, McDonough KA (2018) Cyclic nucleotide signaling in Mycobacterium tuberculosis: an expanding repertoire. Pathog Dis 76 (5)

McDonough KA, Rodriguez A (2011) The myriad roles of cyclic AMP in microbial pathogens: from signal to sword. Nature reviews Microbiology 10: 27-38

(12) L.95: Isn't PhoPR a two-component signal transduction system, the terminology that is more specific than a two-protein regulatory system?

As recommended by the reviewer, we have replaced “two protein regulatory system” with more specific “two-component signal transduction system” in the revised manuscript.

(13) L.124: check-point prevents things from happening. Here the mechanism you found allows growth and survival.

We agree. As recommended by the reviewer, we have now modified the sentence in the revised manuscript (duplicated below).

“Together, the newly identified mechanism of regulation of cAMP level allows intraphagosomal survival and growth program of mycobacteria.”

(14) L.132: why not say directly-"---under normal, and NO and acid stress conditions (Fig. 1A).

As recommended by the reviewer, we have now deleted the first part of the sentence and directly stated that “we compared cAMP levels………. under normal, NO and acidic stress conditions” (duplicated below).

“We compared cAMP levels of WT and phoPR-KO (lacking both phoP and phoR), grown under normal, NO stress and acid stress conditions (Fig. 1A).”

(15) L.134: The complementation is quite variable. Also true in Fig. 2A. If no simple answer, you can say- cAMP values increased in complemented cells, although to a variable extent, for reasons unknown.

We agree with the reviewer. We have now incorporated new text in the ‘Results’ section of the revised manuscript (duplicated below):

“A higher cAMP level in the complemented strain under NO stress is possibly attributable to reproducibly higher phoP expression in the complemented mutant under specific stress conditions (Khan et al., 2022).”

(16) L.154: You rather not say "conclude" and "most likely" at the same time. How about replacing "we conclude" with suggests? In that case, no need to say "most likely". Also, in l.306-7 & l.322-3.

We thank the reviewer for these suggestions. We have now modified the statements in the revised manuscript (duplicated below).

“We suggest that lower cAMP level of the mutant is not due to its higher efficacy of cAMP secretion.”

Following reviewer’s recommendation, we have incorporated similar changes in two other places of the ‘Results’ section of the revised manuscript.

(17) L.161: introduce both the acronyms here and not in l.162.

Following reviewer’s recommendation, we have made the suggested changes.

(18) L.164: Second, (to be in line with First).

We have made the suggested change.

(19). Fig. 2C: There are no black and white bars. This is an important figure because the results appear in the abstract. The signal change from pH 7 to 4.5 is not much. An independent approach would have been desirable. If it were E. coli, I would have suggested beta-gal assay or in vivo footprints. Is a PhoP binding site recognizable in the promoter region of rv0805?

We apologize for the inaccuracy. We have corrected it in the revised manuscript. Also, we have now carried out DNA binding assays, and included the EMSA data of rv0805 upstream regulatory region binding to phosphorylated PhoP (P~PhoP) as a supplemental figure (Figure 2-figure supplement 1A-B). In this figure, we have also incorporated our results on the likely PhoP binding site within rv0805up. The new figure, figure caption and the relevant results have been adjusted accordingly in the revised manuscript.

(20) L.209: ORFs; also delete "of growth" from the sentence.

The suggested changes were made to the text.

(21) L.213: Delete Importantly and change "failed to" to 'did not' (since you did not motivate the expectation earlier, it is better to state the results in an unbiased way).

As recommended by the reviewer, both changes were included in the revised manuscript.

(22) L.217: The requirement of PhoR is a new result - why say "confirm". Change it to indicate. Also, delete "indeed" here and from L.233.

As recommended by the reviewer, both changes were included in the revised manuscript.

(23) L.224: Are the results in Fig 3-S1A under inducing conditions?

The results shown in Fig 3-S1A are not under inducing conditions of expression. For better clarity, we have modified the sentence describing Figure 3-figure supplement 1A (duplicated below).

“rv0805 ORF was cloned within the multicloning site of integrative pSTki (Parikh et al., 2013) between EcoRI and HindIII sites under the control of Pmyc1tetO promoter, and expression of rv0805 under non-inducing condition was verified by determining the mRNA level (Figure 3 - figure supplement 1A).

Reference:

Parikh et al (2013) Development of a new generation of vectors for gene expression, gene replacement, and protein-protein interaction studies in mycobacteria. Applied and environmental microbiology 79: 1718-1729

(24) L.225: ---cAMP level. Add (Fig. 3C) at the end of the next sentence.

As recommended by the reviewer, both the suggested changes were made to the revised text.

(25) L.231: Delete "Most importantly"- you didn't specify what are other less important results.

We agree. We have now deleted “most importantly” from the sentence in the revised text.

(26) L.243 & 254: Change homeostasis to level? Here you are showing mechanisms that can change cAMP level. Homeostasis here would mean how fluctuations in cAMP level are adjusted, usually requiring negative feedback.

As recommended by the reviewer, ‘homeostasis’ was replaced with ‘level’ in both places.

(27) L.256: stress response or stress? Also, in l.272

We are sorry for the inaccuracy. We have corrected these errors in the revised version of the manuscript.

(28) L.259: Change "maintenance of homeostasis" to 'repressing the rv0805 PDE gene'. It is safer to use a fact-based title. In this section, direct measurement of rv0805 mRNA, and/or cAMP levels in different genetic backgrounds seem desirable.

We agree. As recommended by the reviewer, we have modified the title of the ‘Results’ section in the revised manuscript (duplicated below).

“PhoP contributes to mycobacterial stress tolerance and intracellular survival by repressing the rv0805 PDE expression.”

Please note that direct measurements of rv0805 mRNA and cAMP levels are part of Fig. 3 and Figure 3- figure supplement 1A, respectively.

(29) Fig, 4A: White and grey symbols are not easily discriminated without zooming. Use color for phoPR-KO.

We agree. We have now indicated the phoPR-KO in blue in the revised Fig. 4.

(30) L.264: Delete remarkable or explain what is so remarkable. Aren't the results expected- the PDE level would go up in both cases. Direct measurement of PDE /cAMP levels would take the mystery out of the results.

As recommended by the reviewer, we have deleted ‘remarkably’ in the revised text. We have measured cAMP and PDE expression levels of the four strains in Fig. 3 and Figure 3-figure supplement 1.

(31) L.273: --suggesting a role of ---

We have modified this sentence in the revised version of the manuscript (duplicated below).

“A previous study had reported that phoP-deleted mutant strain was more sensitive to Cumene Hydrogen Peroxide (CHP), suggesting a role of PhoP in regulating mycobacterial stress response to oxidative stress (Walters et al., 2006).”

Reference:

Walters et al. (2006) The Mycobacterium tuberculosis PhoPR two-component system regulates genes essential for virulence and complex lipid biosynthesis. Mol Microbiol 60: 312-330

(32) L.275: Delete "transcriptome". CHP sensitivity alone doesn't speak for transcriptome.

As suggested by the reviewer, we have deleted “transcriptome”. Also, please see our response to the previous comment (above).

(33) Fig. 4D and E: % Colocalization in the Merge panels is not much different among the four strains tested (to an untrained eye). Can the results be explained to readers not used to in vivo studies?

As recommended by the reviewer, we have now incorporated new text to explain the in vivo experiment (duplicated below).

“In this assay, WT-H37Rv inhibits phagosome maturation, whereas phagosomes with phoPR-KO mature into phagolysosomes (Anil Kumar et al., 2016).”

Further, for better clarity of the results shown in Fig. 4D, we have (a) increased size of the figure to highlight the difference in the ‘merge’ panel; (b) included “white arrowheads” in the merge panels of Fig. 4D to indicate auramine labeled mycobacteria, which either have inhibited or facilitated trafficking into lysosomes, and finally (c) incorporated method used to calculate percent co-localization in greater details in the ‘Material and Methods’ section of the revised manuscript.

Reference

Anil Kumar et al. (2016) EspR-dependent ESAT-6 secretion of Mycobacterium tuberculosis requires the presence of virulence regulator PhoP. J Biol Chem. 291, 19018-19030

(34) L.275-6: Delete "next" (also in l.347) and "Note that". In this paragraph, I was expecting some explanation on how phoPR-KO and WT-Rv0805 are behaving similarly. Even if the reason is not known, it should be mentioned.

The suggested changes have been made to the text. Also, as recommended by the reviewer, we have included the following text in the revised manuscript (duplicated below):

“Together, these results reveal similar behaviour of phoPR-KO, and WT-Rv0805 by demonstrating a comparably higher susceptibility of these strains to acidic pH and oxidative stress relative to WT bacteria and indicate a link between intra-mycobacterial cAMP level and bacterial stress response. Collectively, these data suggest that at least one of the mechanisms by which PhoP contributes to global stress response is attributable to maintenance of cAMP level.”

(35) L.281: ---WT and indicate a link between cAMP level and stress response in mycobacteria. (No mention of homeostasis).

The suggested change has been made to the revised text. Please see above our response to point # 34.

(36) L.288, 290: No Thus and no clearly.

Both the suggested changes have been made to the text.

(37) L.297: Can you be more direct and state --is due to reduced cAMP level?

As recommended by the reviewer, we have now modified the sentence to make it more direct in the revised manuscript (duplicated below):

“Together, our findings facilitate an integrated view of our results, suggesting that higher susceptibility of WT-Rv0805 to stress conditions, is attributable to its reduced cAMP level.”

(38) L.307: May delete "most likely----homeostasis". cAMP is not discussed here. The same deletion is desired in l.324.

We agree. As recommended by the reviewer, we have now modified the relevant texts in the revised manuscript. These are duplicated below.

“From these results, we suggest that ectopic expression of rv0805 impacts phagosome maturation arguing in favour of a role of PhoP in influencing phagosome-lysosome fusion in macrophages.”

“Thus, we suggest that one of the reasons which accounts for an attenuated phenotype of phoPR-KO in both cellular and animal models is attributable to PhoP-dependent repression of rv0805 PDE activity, which controls mycobacterial cAMP level.”

(39) L.342: cAMP level is regulated remains---

The suggested change has been made to the revised text (duplicated below):

“Although many bacterial pathogens modulate host cell cAMP level as a common strategy, the mechanism of regulation of mycobacterial cAMP level remains unknown.”

(40) L.373: tone down "most fundamental". It is not obvious what is so profound about a stress-response system that depends on PhoP also depends on PhoR. OR justify what is most fundamental about it.

We agree. Following reviewer’s recommendation, we have modified the text in the revised manuscript (duplicated below):

“In keeping with these results, we find that PhoP-dependent rv0805 expression requires PhoR (Figs. 3A-B), the cognate kinase which activates PhoP in a signal-dependent manner (Gupta et al., 2006; Singh et al., 2023).”

References:

Gupta et al. (2006) Transcriptional autoregulation by Mycobacterium tuberculosis PhoP involves recognition of novel direct repeat sequences in the regulatory region of the promoter. FEBS Letters 580, 5328-5338.

Singh et al. (2023) Dual functioning by the PhoR sensor is a key determinant to Mycobacterium tuberculosis virulence. PLoS Genetics 19(12): e1011070.

(41) L.395: delete correspondingly (?)

The suggested change has been made to the text.

(42) L.396: Delete "appear to" and "somewhat". The uncertainty is already implied in "suggest". The evidence that ectopic expression of rv0805 is functionally equivalent to phoP deletion is quite clear in this paper and not saying that clearly is confusing.

We agree with the reviewer. The suggested changes have been made to the revised text (duplicated below):

“Thus, our results suggest that ectopic expression of rv0805 is functionally equivalent to deletion of the phoP locus.”

(43) L.401: --over-expressing bacilli, induction level of rv0805 expression was significantly different in Matange et al and our studies. The next sentence is also very wordy.

We have made changes to the text to address the reviewer’s concern. Also, the next sentence has been rewritten (duplicated below).

“Although both studies were performed with rv0805 over-expressing bacilli, the fact that important differences in the expression of PDEs, in this study (Matange et al., 2013) and in our assays - yielding significantly different levels of rv0805 expression - most likely account for this discrepancy. While we cannot rule out the possibility of cleavage of other cyclic nucleotides by Rv0805 (Keppetipola & Shuman, 2008; Shenoy et al., 2007; Shenoy et al., 2005), consistent with a previous study our results correlate rv0805 expression with intra-mycobacterial cAMP level (Agarwal et al., 2009).”

References:

Matange et al. (2013) Overexpression of the Rv0805 phosphodiesterase elicits a cAMP-independent transcriptional response. Tuberculosis (Edinb) 93: 492-500.

Keppetipola N, Shuman S (2008) A phosphate-binding histidine of binuclear metallophosphodiesterase enzymes is a determinant of 2',3'-cyclic nucleotide phosphodiesterase activity. J Biol Chem 283: 30942-30949

Shenoy et al. (2007) Structural and biochemical analysis of the Rv0805 cyclic nucleotide phosphodiesterase from Mycobacterium tuberculosis. Journal of molecular biology 365: 211-225

Shenoy et al. (2005) The Rv0805 gene from Mycobacterium tuberculosis encodes a 3',5'-cyclic nucleotide phosphodiesterase: biochemical and mutational analysis. Biochemistry 44: 15695-15704

Agarwal N, Lamichhane G, Gupta R, Nolan S, Bishai WR (2009) Cyclic AMP intoxication of macrophages by a Mycobacterium tuberculosis adenylate cyclase. Nature 460: 98-102

(44) L.409: To avoid saying "conclude" and "most likely" at the same time, can you start the sentence thus: 'We infer that Pho-----rv0805 is a---.

We agree. We have made suggested changes to the text. The modified sentence is duplicated below:

“We infer that PhoP-dependent regulation of Rv0805 is a critical regulator of intra-mycobacterial cAMP level.”

(45) L.424. Delete "According to this model". In the preceding sentence, the subject is results, not model. This whole paragraph needs to be rewritten in fewer lines. The shorter the summary statement, the greater would be its impact (less is more here). I would delete the red circles from the figure- it appears that in the repressed state, you are making more products. Replace the circles with an arrow. The legend could be "Increased cAMP level and effective stress response" and "Decreased cAMP---and reduced---.

We thank the reviewer for these suggestions. Following reviewer’s recommendations, we have made numerous changes and rewritten the paragraph in the revised manuscript (duplicated below):

“In summary, upon sensing low acidic pH as a signal PhoR activates PhoP, P~PhoP binds to rv0805 upstream regulatory region and functions as a specific repressor of Rv0805. Therefore, we observed (a) a reproducibly lower level of cAMP in phoPR-KO relative to WT-H37Rv, (b) a significantly reduced expression of rv0805 in WT-H37Rv, grown under acidic pH relative to normal conditions, and (c) comparable cAMP levels in phoPR-KO and WT-Rv0805. This is why the two strains remain ineffective to mount an appropriate stress response, most likely due to their inability to coordinate regulation of gene expression because of dysregulation of intra-mycobacterial cAMP level. However, without uncoupling regulatory control of PhoPR and rv0805 expression, we cannot confirm that dysregulation of cAMP level accounts for virulence attenuation of phoPR-KO. Given the fact that rv0805-depleted M. tuberculosis is growth attenuated in vivo (McDowell et al., 2023), paradoxically ectopic expression of rv0805 leads to dysregulated metabolic adaptation, thereby resulting in reduced stress tolerance and intracellular survival.”

Also, the suggested changes have been incorporated in Fig. 6 and the figure caption.

Reference

McDowell JR, Bai G, Lasek-Nesselquist E, Eisele LE, Wu Y, Hurteau G, Johnson R, Bai Y, Chen Y, Chan J et al (2023) Mycobacterial phosphodiesterase Rv0805 is a virulence determinant and its cyclic nucleotide hydrolytic activity is required for propionate detoxification. Mol Microbiol 119: 401-422

(46) L.458 & 500: ---was used to transform.

Following reviewer’s recommendation, the suggested changes were made to the text in the Materials and Methods section of the revised manuscript.

(47) L.460: --- antibiotics plates.

Both suggested changes were made to the text.

(48) L.466-7: --they were transferred-pH 4.5) and grown for further-

We thank the reviewer for these suggestions. The suggested changes were made to the text.

(49) L.486: ---full-length ORFs of interest were---

The suggested changes were incorporated in the revised manuscript.

(50) L.497: The RNAs were 20 nt long and complementary---

As recommended by the reviewer, we have modified the text in the revised manuscript (duplicated below).

“The RNAs were 20 nt long and complementary to the non-template strand of the target gene.”

Reviewer #2:

(1) Rephrase this sentence in the abstract: “Because growing evidence connects PhoP with varying stress response, we hypothesized that the level of 3’,5’ cAMP, one of the most widely used second messengers, was regulated by the phoP locus, linking numerous stress responses with cAMP production”.

As recommended by the reviewer, we have now rewritten the sentence. The modified text is incorporated in the revised manuscript (duplicated below):

“cAMP is one of the most widely used second messengers, which impacts on a wide range of cellular responses in microbial pathogens including M. tuberculosis. Herein, we hypothesized that intra-mycobacterial cAMP level could be controlled by the phoP locus since the major regulator plays a key role in bacterial responses against numerous stress conditions.”

Also, please see our response to specific comments #1-3 of Reviewer 1.

(2) Line 134: please describe the complementation strain features as it is mentioned for the first time (plasmid, copy number, promoter etc.) in the manuscript. Especially under NO stress what could be the authors' justification regarding the high cAMP concentration in the complementation strain?

As recommended by the reviewer, the details of construction of the complemented strain have been incorporated in the ‘Materials and Methods’ section of the revised manuscript (duplicated below):

“To complement phoPR expression, pSM607 containing a 3.6- kb DNA fragment of M. tuberculosis phoPR including 200-bp phoP promoter region, a hygromycin resistance cassette, attP site and the gene encoding phage L5 integrase, as detailed earlier (Walters et al., 2006) was used to transform phoPR mutant to integrate at the L5 attB site.”

To address the reviewer’s other concern, we have now included the following sentence in the ‘Results’ section of the revised manuscript (duplicated below):

“A higher cAMP level in the complemented strain under NO stress is possibly attributable to reproducibly higher phoP expression in the complemented mutant under specific stress condition (Khan et al., 2022).”

Reference:

Khan et al. (2022) Convergence of two global regulators to coordinate expression of essential virulence determinants of Mycobacterium tuberculosis. eLife 2022, 11:e80965.

(3) In Figure 1C, it is a bit confusing to see the numbers 1,2,3 and 4 and nothing is referred to these numbers in the figure legend so it's better to remove them.

We agree with the reviewer. We have now removed the lane numbers from the figure (Fig. 1C) in the revised manuscript.

(4) Line 852: rephrase it "insignificantly different".

The suggested change has been made to the text. The modified text is incorporated in the manuscript (duplicated below):

“Note that the difference in expression levels of rv0805 between WT and phoPR-KO was significant (p<0.01), whereas the fold difference in mRNA level between WT and the complemented mutant (Compl.) remains nonsignificant (not indicated).”

(5) Line198-200: There are no open/black bars, they all are coloured bars. Correct the same. The significance test should be done for the same gene (suppose rv0805 up) in different pH conditions. Right now, it is not revealing anything and misleading.

We apologize for the inaccuracy. We have now rectified the error. As recommended by the reviewer, Fig. 4C was modified, and the significance tests were carried out between samples involving identical promoter enrichments under different pH conditions. The modified figure, figure legend, and the relevant results have been adjusted accordingly in the revised manuscript.

(6) Line 213: Is there any difference between this complementation strain (phoPR-KO:: phoPphoR with the one used in Figure 1A, 1B, and 2A? If yes, then please describe it.

The same complemented mutant strain, which has been described in the ‘Materials and Methods’ section of the revised manuscript, was used in the experiments described in Fig. 1A, Fig.1B and Fig. 2A.

(7) Line 223: Please mention the copy number and promoter of the vector construct.

As recommended by the reviewer, we have now mentioned the promoter of the vector and incorporated new text with regard to copy number of the expression vector in the revised manuscript (duplicated below).

“Although copy number of episomal vectors with pAl5000 origin of replication (oriM) have been reported to be 3 by Southern hybridization (Ranes et al, 1990), in this case wild-type and mutant Rv0805 proteins were expressed from single-copy chromosomal integrants (Parikh et al., 2013).”

References

Ranes et al., (1990) Functional analysis of pAL5000, a plasmid from Mycobacterium fortuitum: construction of a "mini" mycobacterium-Escherichia coli shuttle vector. J Bacteriol 172: 2793-2797

Parikh et al., (2013) Development of a new generation of vectors for gene expression, gene replacement, and protein-protein interaction studies in mycobacteria. Applied and environmental microbiology 79: 1718-1729

(8) Figure 3 - Figure Supplement 1: not sure why the authors measured mRNA levels of rv1357 and rv2387? These genes were not overexpressed!

The mRNA levels of rv1357 and rv2387 were measured to show that overexpression of either the wild-type or mutant Rv0805 did not influence expression of other PDEs like Rv1357 and Rv2387. We have now mentioned it explicitly in the revised manuscript (duplicated below).

“In contrast, other PDE encoding genes (rv1357 and rv2387), under identical conditions, demonstrate comparable expression levels in WT-H37Rv and rv0805 over-expressing strains.”

(9) Line 234: Wrong interpretation it should be PDE mRNA levels in WT-Rv0805 and WT-Rv0805M.

As recommended by the reviewer, we have now modified the statement to improve clarity (duplicated below).

“The corresponding mRNA levels of PDEs (wild-type and the mutant) are over-expressed approximately 4.5-6 -fold relative to the genomic rv0805 level of WT-H37Rv (Figure 3-figure supplement 1A).”

(10) Line 237: Remove the sentence "Thus, we conclude......identical expression strategy", you have already talked about why phosphodiesterase activity is crucial for cAMP concentration and it is well understood.

Following reviewer’s recommendation, we have now removed the sentence from the revised manuscript.

(11) Figure 3E: Authors should comment on why the cAMP concentration is not significantly changed even though the mRNA level changes are drastic (~90%). How do you correlate that? Is it because of other PDEs?

We agree. As suggested by the reviewer, we have now incorporated new text in the revised manuscript (duplicated below).

“We speculate that effective knocking down of phoP or rv0805 is not truly reflected in the extent of variation of cAMP levels possibly due to the presence of numerous other mycobacterial PDEs.”

(12) Line 505,506: Is it the translation start site or the transcription start site? Because mRNA level changes are reported.

It is the translational start sites, and gene-specific small guide RNAs were designed to inhibit mRNA expression.

(13) Line 292: There is a difference between red and green bars. Authors should do statistical analysis and then comment on whether overexpression of WT and mutant pde are different or similar, to me they are different; also, explain why the WT-Rv0805 strain is different than the phoPR-KO strain in the context of cell wall metabolism.

As recommended by the reviewer, we have now included statistical significance of the data in the revised version, and modified the text accordingly in the manuscript.

Also, we included text explaining why WT-Rv0805 is different compared to phoPR-KO strain in the context of cell wall metabolism (duplicated below).

“Together, these results suggest that both strains expressing wild type or mutant PDEs share a largely similar cell-wall properties and are consistent with (a) a recent study reporting no significant effect of cAMP dysregulation on mycobacterial cell wall structure/permeability (Wong et al., 2023), and (b) role of PhoP in cell wall composition and complex lipid biosynthesis (Walters et al., 2006; Asensio et al., 2006; Goyal et al., 2011).”

References:

Wong et al. (2023) Cyclic AMP is a critical mediator of intrinsic drug resistance and fatty acid metabolism in M. tuberculosis. eLife 2023; 12: e81177

Walters et al. (2006) The Mycobacterium tuberculosis PhoPR two-component system regulates genes essential for virulence and complex lipid biosynthesis. Mol Microbiol 60: 312-330

Asensio et al. (2006) The Virulence-associated Two-component PhoP-PhoR System Controls the Biosynthesis of Polyketide-derived Lipids in Mycobacterium tuberculosis. J Biol Chem 281: 1313-1316.

Goyal et al. (2011) Phosphorylation of PhoP protein plays direct regulatory role in lipid biosynthesis of Mycobacterium tuberculosis. J Biol Chem 286: 45197-45208

(14) Line 299-303: Authors should explain how the colocalization % are calculated. Also, in the figure 4D merge panel please highlight the difference.

As suggested by the reviewer, we have now explained the methodology used to calculate percent colocalization in greater details. Also, we have modified Figure 4D to highlight the difference between samples shown in merge panel. Please see our response to comment # 33 from the Reviewer 1.

(15) General comment: There are multiple instances where writing needs to be improved.

We are sorry for the inaccuracies. We have now done thorough editing of the manuscript and made numerous corrections throughout.

eLife assessment

This important study describes how PhoP regulates cyclic-AMP production in the human pathogen Mycobacterium tuberculosis. The authors provide convincing evidence that PhoP acts as a repressor of the cyclic-AMP-specific phosphodiesterase, Rv0805, which can degrade cyclic-AMP. The work will be of interest to bacteriologists and whilst the revised manuscript has been substantively improved, there are some outstanding points that could improve clarity and presentation.

Reviewer #1 (Public Review):

Summary:

This paper provides a straightforward mechanism of how mycobacterial cAMP level is increased under stressful conditions and shows that the increase is important for the survival of the bacterium in animal hosts. The cAMP level is increased by decreasing the expression of an enzyme that degrades cAMP.

Strengths:

The paper shows that under different stresses the response regulator PhoP represses a phosphodiesterase (PDE) that degrades cAMP specifically. Identification of<br /> PhoP as a regulator of cAMP is significant progress in understanding Mtb pathogenesis, as increase in cAMP apparently increases bacterial survival upon infection. On the practical side, reduction of cAMP by increasing PDE can be a means to attenuate the growth of the bacilli. The results have wider implications since PhoP is implicated in controlling diverse mycobacterial stress responses and many bacterial pathogens modulate host cell cAMP level. The results here are straightforward, internally consistent, and of both theoretical and applied interests. The results also open considerable future work, especially how increases in cAMP level help to increase survival of the pathogen.

Weaknesses:

It is not clear whether PhoP-PDE Rv0805 is the only pathway to regulate cAMP level under stress.

Reviewer #2 (Public Review):

Summary: In the manuscript, the authors have presented new mechanistic details to show how intracellular cAMP levels are maintained linked to the phosphodiesterase enzyme which in turn is controlled by PhoP. Later, they showed the physiological relevance linked to altered cAMP concentrations.

Strengths: Well thought out experiments. The authors carefully planned the experiments well to uncover the molecular aspects of it diligently.

Weaknesses: Some fresh queries were made based on the author's previous responses and hope to get satisfactory answers this time.

Author Response

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

eLife assessment

This important study reports a novel measurement for the chemotactic response to potassium by Escherichia coli. The authors convincingly demonstrate that these bacteria exhibit an attractant response to potassium and connect this to changes in intracellular pH level. However, some experimental results are incomplete, with additional controls/alternate measurements required to support the conclusions. The work will be of interest to those studying bacterial signalling and response to environmental cues.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This paper shows that E. coli exhibits a chemotactic response to potassium by measuring both the motor response (using a bead assay) and the intracellular signaling response (CheY phosporylation level via FRET) to step changes in potassium concentration. They find increase in potassium concentration induces a considerable attractant response, with an amplitude larger than aspartate, and cells can quickly adapt (but possibly imperfectly). The authors propose that the mechanism for potassium response is through modifying intracellular pH; they find both that potassium modifies pH and other pH modifiers induce similar attractant responses. It is also shown, using Tar- and Tsr-only mutants, that these two chemoreceptors respond to potassium differently. Tsr has a standard attractant response, while Tar has a biphasic response (repellent-like then attractant-like). Finally, the authors use computer simulations to study the swimming response of cells to a periodic potassium signal secreted from a biofilm and find a phase delay that depends on the period of oscillation.

Strengths:

The finding that E. coli can sense and adapt to potassium signals and the connection to intracellular pH is quite interesting and this work should stimulate future experimental and theoretical studies regarding the microscopic mechanisms governing this response. The evidence (from both the bead assay and FRET) that potassium induces an attractant response is convincing, as is the proposed mechanism involving modification of intracellular pH.

Weaknesses:

The authors show that changes in pH impact fluorescent protein brightness and modify the FRET signal; this measurement explains the apparent imprecise adaptation they measured. However, this effect reduces confidence in the quantitative accuracy of the FRET measurements. For example, part of the potassium response curve (Fig. 4B) can be attributed to chemotactic response and part comes from the pH modifying the FRET signal. Measuring the full potassium response curve of the no-receptor mutants as a control would help quantify the true magnitude of the chemotactic response and the adaptation precision to potassium.

Response: We thank the reviewer for the suggestion. We have now measured the full potassium response curve for the no-receptor mutant (HCB1414-pVS88), as shown in Fig. S4. We characterized the pH effects on CFP and YFP channels at different concentrations of KCl, and the relationship between the ratio of the signal post- to pre-KCl addition and the KCl concentration was established for both channels, as shown in Fig. S4C. The pH-corrected signal after KCl addition for strains with receptors was obtained by dividing the original signal after KCl addition by this ratio at the specific KCl concentration. This was done for both CFP and YFP channels. The pH-corrected responses for the Tar-only and Tsr-only strains are represented by red dots in Fig. 5BC. The recalculated response curve and adaptation curve for the wild-type strain are shown in Fig. S5. The same correction was applied to Fig. 3 as well. We also re-performed the simulations using the corrected dose-response curve and replotted Fig. 6, though the simulation results did not change much.

We have now added a subsection “Revised FRET responses by correcting the pH effects on the brightness of eCFP and eYFP” at line 296 in “Results” to describe this.

The measured response may also be impacted by adaptation. For other strong attractant stimuli, the response typically shows a low plateau before it recovers (adapts). However, in the case of Potassium, the FRET signal does not have an obvious plateau following the stimuli. Do the authors have an explanation for that? One possibility is that the cells may have already partially adapted when the response reaches its minimum, which could indicate a different response and/or adaptation dynamics from that of a regular chemo-attractant? In any case, directly measuring the response to potassium in mutants without adaptation enzymes (CheR, CheB) and with the receptors in different methylation levels would shed more light on the problem.

Response: We appreciate the reviewer’s insightful questions. To observe the low plateau before adaptation, a saturating amount of attractant should be added in a stepwise manner. According to the dose-response curve we measured for potassium, a saturating amount of potassium would be close to 100 mM. In fact, there is a small segment of the low plateau in the step response to 30 mM KCl (Fig. 4C or Fig. S5A). To observe more of this low plateau, we could have used a higher concentration of KCl. However, a stimulation higher than 30 mM KCl will induce substantial physiological changes in the cell, resulting in a significant decrease in fluorescence for both channels (Fig. S7). Therefore, the range of KCl concentration that can be reliably applied in FRET measurements is limited.

The half-time of adaptation at 30 mM KCl was measured to be approximately 80 s, demonstrating a faster adaptation than 0.1 mM MeAsp, which induced a similar magnitude of response. Nevertheless, this is still significantly slower than the time required for medium exchange in the flow chamber, which takes less than 10 s to replace 99% of the medium. Thus, the effect on the measured response magnitude due to adaptation should be small (less than 10%).

We thank the reviewer for the suggestion of measuring the response to potassium in mutants without adaptation enzymes (CheR, CheB) and with the receptors in different methylation levels. However, these mutants are typically less sensitive than the wild-type, exhibiting higher values of K0.5 (Sourjik & Berg, PNAS 99:123, 2002), and thus require an even higher KCl concentration to see the low plateau. Consistent with this, we attempted to measure the response to potassium in a cheRcheB mutant (HCB1382-pVS88). As shown in Fig. R1 below, there is no response to up to 30 mM KCl, suggesting that the sensitive region of the mutant is beyond 30 mM KCl.

The relevant text was added at line 413-424.

Author response image 1.

The response of the cheRcheB mutant (HCB1382-pVS88) to different concentrations of KCl. The blue solid line denotes the original signal, while the red dots represent the pH-corrected signal. The vertical purple (green) dashed lines indicate the moment of adding (removing) 0.01 mM, 0.1 mM, 0.3 mM, 1 mM, 3 mM, 10 mM and 30 mM KCl, in chronological order.

There seems to be an inconsistency between the FRET and bead assay measurements, the CW bias shows over-adaptation, while the FRET measurement does not.

Response: We thank the reviewer for pointing this out. We have now demonstrated that the imprecise adaptation shown in the FRET assay primarily resulted from the pH-induced intensity change of the fluorescent proteins. As shown in Fig. S5A&C, the FRET signal also shows over-adaptation, similar to the bead assay, when we recalculated the response by correcting the CFP and YFP channels.

Now we clarified it at line 315.

The small hill coefficient of the potassium response curve and the biphasic response of the Tar-only strain, while both very interesting, require further explanation since these are quite different than responses to more conventional chemoattractants.

Response: We thank the reviewer for pointing this out. We have now recalculated the pH-corrected results for the dose-response curve (Fig. S5) and the biphasic response of the Tar-only strain (Fig. 5C). The new Hill coefficient is 0.880.14 (meanSD), which is close to the response to MeAsp (1.2) (ref. 46). We suspected that this Hill coefficient of slightly less than 1 resulted from the different responses of Tar and Tsr receptors to potassium.

The Tar-only strain exhibits a repellent response to stepwise addition of low concentrations of potassium less than 10 mM, and a biphasic response above (Fig. 5C). This biphasic response might result from additional pH-effects on the activity of intracellular enzymes such as CheRB and CheA, which may have a different timescale and response from the Tar receptor. We have now added the penultimate paragraph in “Discussion” to talk about the response of the Tar-only strain.

Reviewer #2 (Public Review):

Summary:

Zhang et al investigated the biophysical mechanism of potassium-mediated chemotactic behavior in E coli. Previously, it was reported by Humphries et al that the potassium waves from oscillating B subtilis biofilm attract P aeruginosa through chemotactic behavior of motile P aeruginosa cells. It was proposed that K+ waves alter PMF of P aeruginosa. However, the mechanism was this behaviour was not elusive. In this study, Zhang et al demonstrated that motile E coli cells accumulate in regions of high potassium levels. They found that this behavior is likely resulting from the chemotaxis signalling pathway, mediated by an elevation of intracellular pH. Overall, a solid body of evidence is provided to support the claims. However, the impacts of pH on the fluorescence proteins need to be better evaluated. In its current form, the evidence is insufficient to say that the fluoresce intensity ratio results from FRET. It may well be an artefact of pH change. Nevertheless, this is an important piece of work. The text is well written, with a good balance of background information to help the reader follow the questions investigated in this research work.

In my view, the effect of pH on the FRET between CheY-eYFP and CheZ-eCFP is not fully examined. The authors demonstrated in Fig. S3 that CFP intensity itself changes by KCl, likely due to pH. They showed that CFP itself is affected by pH. This result raises a question of whether the FRET data in Fig3-5 could result from the intensity changes of FPs, but not FRET. The measured dynamics may have nothing to do with the interaction between CheY and CheZ. It should be noted that CFP and YFP have different sensitivities to pH. So, the measurement is likely confounded by the change in intracellular pH. Without further experiments to evaluate the effect of pH on CFP and YFP, the data using this FRET pair is inconclusive.

Response: We thank the reviewer for pointing this out. We have now measured the full potassium response curve for the no-receptor mutant (HCB1414-pVS88), as shown in Fig. S4. We characterized the pH effects on CFP and YFP channels at different concentrations of KCl, and the relationship between the ratio of the signal post- to pre-KCl addition and the KCl concentration was established for both channels, as shown in Fig. S4C. The pH-corrected signal after KCl addition for strains with receptors was obtained by dividing the original signal after KCl addition by this ratio at the specific KCl concentration. This was done for both CFP and YFP channels. The pH-corrected responses for the Tar-only and Tsr-only strains are represented by red dots in Fig. 5BC. The recalculated response curve and adaptation curve for the wild-type strain are shown in Fig. S5. The same correction was applied to Fig. 3 as well. We also re-performed the simulations using the corrected dose-response curve and replotted Fig. 6, though the simulation results did not change much.

We have now added a subsection “Revised FRET responses by correcting the pH effects on the brightness of eCFP and eYFP” at line 296 in “Results” to describe this.

The data in Figure 1 is convincing. It would be helpful to include example videos. There is also ambiguity in the method section for this experiment. It states 100mM KCl was flown to the source channel. However, it is not clear if 100 mM KCl was prepared in water or in the potassium-depleted motility buffer. If KCl was prepared with water, there would be a gradient of other chemicals in the buffer, which confound the data.

Response: We apologize for the ambiguity. The KCl solution used in this work was prepared in the potassium-depleted motility buffer. We have now clarified this at both lines 116 and 497. We now provided an example video, Movie S1, with the relevant text added at line 123.

The authors show that the FRET data with both KCl and K2SO4, and concluded that the chemotactic response mainly resulted from potassium ions. However, this was only measured by FRET. It would be more convincing if the motility assay in Fig1 is also performed with K2SO4.

Response: We thank the reviewer for the suggestion. The aim of comparing the responses to KCl and K2SO4 was to determine the role of chloride ions in the response and to prove that the chemotactic response of E. coli to KCl comes primarily from its response to potassium ions. It is more sensitive to compare the responses to KCl and K2SO4 by using the FRET assay. In contrast, the microfluidic motility assay is less sensitive in revealing the difference in the chemotactic responses, making it difficult to determine the potential role of chloride ions.

Methods:

• Please clarify the promotes used for the constitutive expression of FliCsticky and LacI.

Response: The promoters used for the constitutive expression of LacIq and FliCsticky were the Iq promoter and the native promoter of fliC, respectively (ref. 57).

Now these have been clarified at line 471.

• Fluorescence filters and imaging conditions (exposure time, light intensity) are missing.

Response: Thank you for the suggestion. We have now added more descriptions at lines 535-546: The FRET setup was based on a Nikon Ti-E microscope equipped with a 40× 0.60 NA objective. The illumination light was provided by a 130-W mercury lamp, attenuated by a factor of 1024 with neutral density filters, and passed through an excitation bandpass filter (FF02-438/24-25, Semrock) and a dichroic mirror (FF458-Di02-25x36, Semrock). The epifluorescent emission was split into cyan and yellow channels by a second dichroic mirror (FF509-FDi01-25x36, Semrock). The signals in the two channels were then filtered by two emission bandpass filters (FF01-483/32-25 and FF01-542/32-25, Semrock) and collected by two photon-counting photomultipliers (H7421-40, Hamamatsu, Hamamatsu City, Japan), respectively. Signals from the two photomultipliers were recorded at a sampling rate of 1 Hz using a data-acquisition card installed in a computer (USB-1901(G)-1020, ADlink, New Taipei, Taiwan).

• Please clarify if the temperature was controlled in motility assays.

Response: All measurements in our work were performed at 23 ℃. It was clarified at line 496.

• L513. It is not clear how theta was selected. Was theta set to be between 0 and pi? If not, P(theta) can be negative?

Response: The θ was set to be between 0 and π. This has now been added at line 581.

• Typo in L442 (and) and L519 (Koff)

Response: Thank you. Corrected.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

(1) From the motor measurements the authors find that the CW bias over-adapts to a level larger than prestimulus, but this is not seen in the FRET measurements. What causes this inconsistency? Fig. 2D seems to rule out any change in CheY binding to the motor.

Response: We thank the reviewer for pointing this out. We have now demonstrated that the imprecise adaptation shown in the FRET assay primarily resulted from the pH-induced intensity change of the fluorescent proteins. As shown in Fig. S5A&C, the FRET signal also shows over-adaptation, similar to the bead assay, when we recalculated the response by correcting the CFP and YFP channels.

We now clarified it at line 315.

(2) It would be useful to compare the response amplitude for potassium (Fig. 3C) to a large concentration of both MeAsp and serine. This is a fairer comparison since your work shows potassium acts on both Tar and Tsr. Alternatively, testing a much larger concentration (~10^6 micromolar) at which MeAsp also binds to Tsr would also be useful.

Response: We thank the reviewer for pointing this out. We have now recalculated the response to potassium by correcting the pH-induced effects on fluorescence intensity of CFP and YFP. The response to 30 mM KCl was 1.060.10 times as large as that to 100 μM MeAsp. The aim of the comparison between the responses to potassium and MeAsp was to provide an idea of the magnitude of the chemotactic response to potassium. The stimulus of 100 μM MeAsp is already a saturating amount of attractant and induces zero-kinase activity, thus using a higher stimulus (adding serine or a larger concentration of MeAsp) is probably not needed. Moreover, a larger concentration (~10^6 micromolar) of MeAsp would also induce an osmotactic response.

(3) The fitted Hill coefficient (~0.5) to the FRET response curve is quite small and the authors suggest this indicates negative cooperativity. Do they have a proposed mechanism for negative cooperativity? Have similar coefficients been measured for other responses?

Response: We thank the reviewer for pointing this out. We have now recalculated the pH-corrected results for the dose-response curve (Fig. S5). The new Hill coefficient is 0.880.14 (meanSD), which is close to the response to MeAsp (1.2) (ref. 46). We suspect that this Hill coefficient of slightly less than 1 results from the differing responses of Tar and Tsr receptors to potassium.

(3a) The authors state a few times that the response to potassium is "very sensitive", but the low Hill coefficient indicates that the response is not very sensitive (at least compared to aspartate and serine responses).

Response: We apologize for the confusion. We described the response to potassium as “very sensitive” due to the small value of K0.5. This has now been clarified at line 236.

(3b) Since the measurements are performed in wild-type cells the response amplitude following the addition of potassium may be biased if the cell has already partially adapted. This seems to be the case since the FRET time series does not plateau after the addition of the stimulus. The accuracy of the response curve and hill coefficient would be more convincing if the experiment was repeated with a cheR cheB deficient mutant.

Response: We thank the reviewer for raising these questions. To observe the low plateau before adaptation, a saturating amount of attractant should be added in a stepwise manner. According to the dose-response curve we measured for potassium, a saturating amount of potassium would be close to 100 mM. In fact, there is a small segment of the low plateau in the step response to 30 mM KCl (Fig. 4C or Fig. S5A). To observe more of this low plateau, we could have used a higher concentration of KCl. However, a stimulation higher than 30 mM KCl will induce substantial physiological changes in the cell, resulting in a significant decrease in fluorescence for both channels (Fig. S7). Therefore, the range of KCl concentration that can be reliably applied in FRET measurements is limited.

The half-time of adaptation at 30 mM KCl was measured to be approximately 80 s, demonstrating a faster adaptation than 0.1 mM MeAsp, which induced a similar magnitude of response. Nevertheless, this is still significantly slower than the time required for medium exchange in the flow chamber, which takes less than 10 s to replace 99% of the medium. Thus, the effect on the measured response magnitude due to adaptation should be small (less than 10%).

We thank the reviewer for the suggestion of measuring the response to potassium in mutants without adaptation enzymes (CheR, CheB) and with the receptors in different methylation levels. However, these mutants are typically less sensitive than the wild-type, exhibiting higher values of K0.5 (ref. 46), and thus require an even higher KCl concentration to see the low plateau. Consistent with this, we attempted to measure the response to potassium in a cheRcheB mutant (HCB1382-pVS88). As shown in Fig. R1, there is no response to up to 30 mM KCl, suggesting that the sensitive region of the mutant is beyond 30 mM KCl.

The relevant text was added at line 413-424.

(4) The authors show that the measured imprecise adaptation can be (at least partially) attributed to pH impacting the FRET signal by changing eCFP and eYFP brightness.

(4a) Comparing Fig. 5C and D, the chemosensing and pH response time scales look similar. Therefore, does the pH effect bias the measured response amplitude (just as it biases the adapted FRET level)?

Response: We agree with the reviewer that the pH effect on CFP and YFP biases the measured response amplitude. We have now performed the measurement of dose-response curve to potassium for the no-receptor mutant (HCB1414-pVS88), as shown in Fig. S4. The pH effects on CFP and YFP were corrected. The dose-response curve and adaptation curve were recalculated and plotted in Fig. S5.

(4b) It would help to measure a full response curve (at many concentrations) for the no-receptor strain as a control. This would help distinguish, as a function of concentration, how much response can be attributed to pH impacting the FRET signal versus the true chemotactic response.

Response: We thank the reviewer for the suggestion. We have now performed the measurements for the no-receptor strain. The impact of pH on CFP and YFP has been corrected. The pH-corrected results, previously in Fig.3-5, are now presented in Fig. 3, Fig. S5 and Fig. 5, respectively.

(5) The biphasic response of Tar is strange and warrants further discussion. Do the authors have any proposed mechanisms that lead to this behavior? For the 10mM and 30mM KCl measurements there is a repellent response followed by an attractant response for both adding and removing the stimuli, why is this?

Response: We thank the reviewer for pointing this out. The Tar-only strain exhibits a repellent response to stepwise addition of low concentrations of potassium less than 10 mM, and a biphasic response above (Fig. 5C). This biphasic response might result from additional pH-effects on the activity of intracellular enzymes such as CheRB and CheA, which may have a different timescale and response from the Tar receptor. We have now added the penultimate paragraph in “Discussion” to talk about the response of the Tar-only strain.

(5a) The fact that Tar and Tsr are both attractant (after the initial repellant response in Tar) appears to be inconsistent with previous work on pH response (Ref 52, Yang and Sourjik Molecular Microbiology (2012) 86(6), 1482-1489). This study also didn't see any biphasic response.

Response: We thank the reviewer for pointing this out. The Tar-only strain shows a repellent response to stepwise addition of low concentrations of potassium, specifically less than 10 mM. This is consistent with previous observations of the response of Tar to changes in intracellular pH (refs. 44,45) and also with the work of Yang and Sourjik (new ref. 53), although the work in ref. 53 dealt with the response to external pH change, and bacteria were known to maintain a relatively stable intracellular pH when external pH changes (Chen & Berg, Biophysical Journal (2000) 78:2280-2284). Interestingly, the Tar-only strain exhibits a biphasic response to high potassium concentrations of 10 mM and above. This biphasic response might result from additional pH-effects on the activity of intracellular enzymes such as CheRB and CheA (ref. 56), which may have a different timescale and response from the Tar receptor. We have now added the penultimate paragraph in “Discussion” to talk about the response of the Tar-only strain.

(5b) The response of Tar to the removal of sodium benzoate (Fig. S2) seems to be triphasic, is there any explanation for this?

Response: We thank the reviewer for pointing this out. We have now acknowledged in the legend of Fig. S2 that this response is interesting and warrants further exploration: “The response to the removal of sodium benzoate seems to be a superposition of an attractant and a repellent response, the reason for which deserves to be further explored.”

(6) Fitting the MWC model leads to N=0.35<1. It is fine to use this as a phenomenological parameter, but can the authors comment on what might be causing such a small effective cluster size for potassium response?

Response: We thank the reviewer for pointing this out. We have now recalculated the pH-corrected results for the dose-response curve (Fig. S5). The new Hill coefficient is 0.880.14 (meanSD), which is close to the response to MeAsp (1.2) (ref. 46). We now refit the MWC model to the pH-corrected dose-response curve, obtaining N of 0.85. We think the small N is due partly to the fact that we are fitting the curve with four parameters: N, Kon, Koff, and fm, while only three features of the sigmoid does-response curve are relevant (the vertical scale, the midpoint concentration, and the slope of the sigmoid). Future experiments may determine these parameters more accurately, but they should not significantly affect the simulation results as long as the wild-type dose-response curve is accurate.

(7) The results of the modeling are closely related to Zhu et. al. Phys. Rev. Lett. 108, 128101. Is the lag time for large T related to the adaptation time?

Response: We thank the reviewer for pointing this out. We used a similar framework of modeling as Zhu et. al. The potassium response was also analogous to the chemotactic response to MeAsp. Thus, the results are closely related to Zhu et al. We have now cited Zhu et al. (Ref. 52) and noted this at line 366.

The lag time for large T is related to the adaptation time. We have now simulated the chemotaxis to potassium for large T with different adaptation time by varying the methylation rate kR. The results are shown in Fig. S8. The simulated lag time decreases with the methylation rate kR, but levels off at high values of kR. Now this has been added at line 603.

Minor issues:

• Fig. 1C: should the axis label be y?

Response: Yes, thank you. Now corrected.

• Line 519: Koff given twice, the second should be Kon.

Response: Thank you. Corrected.

• When fitting the MWC model (Eq. 3 and Fig. 6B) did you fix a particular value for m?

Response: m was treated as a fitting parameter, grouped in the parameter fm.

Reviewer #2 (Recommendations For The Authors):

Minor points: - I suggest explaining the acronyms when they first appear in the text (eg CMC, CW, CCW).

Response: Thank you. Now they have been added.

• L144. L242. "decrease" is ambiguous since membrane potential is negative. I understand the authors meant less negative (which is an increase). I suggest to avoid this expression.

Response: Thank you for the suggestion. Now they have been replaced by “The absolute value of the transmembrane electrical potential will decrease”.

• For Fig 1b - it says the shaded area is SEM in the text, but SD in the legend. Please clarify.

Response: Thank you. The annotation in the legend has now been revised as SEM.

• Fig 1C label of x axis should be "y" instead of "x" to be consistent with Fig 1A.

Response: Thank you. It has now been revised.

• In Figure 2, the number of independent experiments as well as the number of samples should be included.

Response: Thank you. The response in Fig. 2C is the average of 83 motors from 5 samples for wild-type strain (JY26-pKAF131). The response in Fig. 2D is the average of 22 motors from 4 samples for the chemotaxis-defective strain (HCB901-pBES38). They have now been added to the legend.

• Regarding the attractant or repelling action of potassium and sucrose, it would be important to have a move showing the cells' behaviours.

Response: We thank the reviewer for the suggestion. We have now provided Movie S1 to show the cells’ behavior to potassium. As shown in Fig. 3B, the chemotactic response to 60 mM sucrose is very small compared to the response to 30 mM KCl. This implies that a noticeable response to sucrose necessitates higher concentrations of stimulation. However, Jerko et al. [Rosko, J., Martinez, V. A., Poon, W. C. K. & Pilizota, T. Proc. Natl Acad. Sci. USA 114, E7969-E7976 (2017).] have shown that high concentrations of sucrose lead to a significant reduction in the speed of the flagella motor. Thus, in a motility assay for sucrose, the osmolarity-induced motility effect may overwhelm the minor repellent-like response.

eLife assessment

In this important study, the authors report a novel measurement of the Escherichia coli chemotactic response and demonstrate that these bacteria display an attractant response to potassium, which is connected to intracellular pH level. Whilst the experiments are mostly convincing, there are some confounders regards pH changes and fluorescent proteins that remain to be addressed.

Reviewer #1 (Public Review):

Summary:

This paper shows that E. coli exhibits a chemotactic response to potassium by measuring both the motor response (using a bead assay) and the intracellular signaling response (CheY phosporylation level via FRET) to step changes in potassium concentration. They find increase in potassium concentration induces a considerable attractant response, with amplitude comparable to aspartate, and cells can quickly adapt (and generally over-adapt). The authors propose that the mechanism for potassium response is through modifying intracellular pH; they find both that potassium modifies pH and other pH modifiers induce similar attractant responses. It is also shown, using Tar- and Tsr-only mutants, that these two chemoreceptors respond to potassium differently. Tsr has a standard attractant response, while Tar has a biphasic response (repellent-like then attractant-like). Finally, the authors use computer simulations to study the swimming response of cells to a periodic potassium signal secreted from a biofilm and find a phase delay that depends on the period of oscillation.

Strengths:

The finding that E. coli can sense and adapt to potassium signals and the connection to intracellular pH is quite interesting and this work should stimulate future experimental and theoretical studies regarding the microscopic mechanisms governing this response. The evidence (from both the bead assay and FRET) that potassium induces an attractant response is convincing, as is the proposed mechanism involving modification of intracellular pH. The updated manuscript controls for the impact of pH on the fluorescent protein brightness that can bias the measured FRET signal. After correction the response amplitude and sharpness (hill coefficient) are comparable to conventional chemoattractants (e.g. aspartate), indicating the general mechanisms underlying the response may be similar. The authors suggest that the biphasic response of Tar mutants may be due to pH influencing the activity of other enzymes (CheA, CheR or CheB), which will be an interesting direction for future study.

Weaknesses:

The measured response may be biased by adaptation, especially for weak potassium signals. For other attractant stimuli, the response typically shows a low plateau before it recovers (adapts). In the case of potassium, the FRET signal does not have an obvious plateau following the stimuli of small potassium concentrations, perhaps due to the faster adaptation compared to other chemoattractants. It is possible cells have already partially adapted when the response reaches its minimum, so the measured response may be a slight underestimate of the true response. Mutants without adaptation enzymes appear to be sensitive to potassium only at much larger concentrations, where the pH significantly disrupts the FRET signal; more accurate measurements would require development of new mutants and/or measurement techniques.

Reviewer #2 (Public Review):

Zhang et al investigated the biophysical mechanism of potassium-mediated chemotactic behavior in E coli. Previously, it was reported by Humphries et al that the potassium waves from oscillating B subtilis biofilm attract P aeruginosa through chemotactic behavior of motile P aeruginosa cells. It was proposed that K+ waves alter PMF of P aeruginosa. However, the mechanism was this behaviour was not elusive. In this study, Zhang et al demonstrated that motile E coli cells accumulate in regions of high potassium levels. They found that this behavior is likely resulting from the chemotaxis signalling pathway, mediated by an elevation of intracellular pH. Overall, a solid body of evidence is provided to support the claims. However, the impacts of pH on the fluorescence proteins need to be better evaluated. In its current form, the evidence is insufficient to say that the fluoresce intensity ratio results from FRET. It may well be an artefact of pH change.

The authors now carefully evaluated the impact of pH on their FRET sensor by examining the YFP and CFP fluorescence with no-receptor mutant. The authors used this data to correct the impact of pH on their FRET sensor. This is an improvement, but the mathematical operation of this correction needs clarification. This is particularly important because, looking at the data, it is not fully convincing if the correction was done properly. For instance, 3mM KCl gives 0.98 FRET signal both in Fig3 and FigS4, but there is almost no difference between blue and red lines in Fig 3. FigS4 is very informative, but it does not address the concern raised by both reviewers that FRET reporter may not be a reliable tool here due to pH change.

The authors show the FRET data with both KCl and K2SO4, concluding that the chemotactic response mainly resulted from potassium ions. However, this was only measured by FRET. It would be more convincing if the motility assay in Fig1 is also performed with K2SO4. The authors did not address this point. In light of complications associated with the use of the FRET sensor, this experiment is more important.

Author Response

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

Public Reviews:

Reviewer #1 (Public Review):

Weaknesses:

The weaknesses are the brevity of the simulations, the concomitant lack of scope of the simulations, the lack of depth in the analysis, and the incomplete relation to other relevant work.

A 1 µs simulation of CCh (Video 1, part 2) shows that m3 (ACHA) is stable, throughout. The DG comparisons, in silico versus in vitro, indicate that 200 ns simulations are sufficient to identify LA versus HA conformational populations. Figure 6-table supplement 1 shows distances. New citations have been added.

Reviewer #2 (Public Review):

Weaknesses:

After carrying out all-atom molecular dynamics, the authors revert to a model of binding using continuum Poisson-Boltzmann, surface area, and vibrational entropy. The motivations for and limitations associated with this approximate model for the thermodynamics of binding, rather than using modern atomistic MD free energy methods (that would fully incorporate configurational sampling of the protein, ligand, and solvent) could be provided. Despite this, the authors report a correlation between their free energy estimates and those inferred from the experiment. This did, however, reveal shortcomings for two of the agonists. The authors mention their trouble getting correlation to experiment for Ebt and Ebx and refer to up to 130% errors in free energy. But this is far worse than a simple proportional error, because -24 Vs -10 kcal/mol is a massive overestimation of free energy, as would be evident if the authors were to instead express results in terms of KD values (which would have an error exceeding a billion fold). The MD analysis could be improved with better measures of convergence, as well as a more careful discussion of free energy maps as a function of identified principal components, as described below. Overall, however, the study has provided useful observations and interpretations of agonist binding that will help understand pentameric ligand-gated ion channel activation.

The objective of the calculations was to identify structural populations, not to estimate binding free energies. We knew the actual LA and HA energies (for all 4 agonists) from real-world electrophysiology experiments. We conclude that the simple PBSA method worked as a tool for identification because the calculated efficiencies match those from experiments (Figure 4B, Figure 4-Source Data 1). We discuss the mismatches in absolute G in the Results and Discussion. Methods for estimating experimental binding free energies are described in a cited, eLife companion paper. The G ratio relates to agonist efficiency.

Main points:

Regarding the choice of model, some further justification of the reduced 2 subunit ECD-only model could be given. On page 5 the authors argue that, because binding free energies are independent of energy changes outside the binding pocket, they could remove the TMD and study only an ECD subunit dimer. While the assumption of distant interactions being small seems somewhat reasonable, provided conformational changes are limited and localised, how do we know the packing of TMD onto the ECD does not alter the ability of the alpha-delta interface to rearrange during weak or strong binding? They further write that "fluctuations observed at the base of the ECD were anticipated because the TMD that offers stability here was absent.". As the TMD-ECD interface is the "gating interface" that is reshaped by agonist binding, surely the TMD-ECD interface structure must affect binding. It seems a little dangerous to completely separate the agonist binding and gating infrastructure, based on some assumption of independence. Given the model was only the alpha and delta subunits and not the pentamer with TMD, I am surprised such a model was stable without some heavy restraints. The authors state that "as a further control we carried out MD simulation of a pentamer docked with ACh and found similar structural changes at the binding pocket compared to the dimer." Is this sufficient proof of the accuracy of the simplified model? How similar was the model itself with and without agonist in terms of overall RMSD and RMSD for the subunit interface and the agonist binding site, as well as the free energy of binding to each model to compare?

The statement that distant interactions are small is not an "assumption", but rather a conclusion based on data. Mutant cycle analysis of 83 pairs shows (with a few exceptions) non-additivity of free energy change prevails only with separations <~15 A (Fig.3 in Gupta et al 2017). Regardless, the adequacy of dimers and convergence by 200 ns are supported by the calculated and experimental agonist efficiencies match (Figure 4B) and the 1 ms simulation (Video 1 part 2). Apo 200ns simulation of the ECD dimer is now added (Figure 2-figure supplement 2) and the dimer interface seems to be adequate (stable).

Although the authors repeatedly state that they have good convergence with their MD, I believe the analysis could be improved to convince us. On page 8 the authors write that the RMSD of the system converged in under 200 ns of MD. However, I note that the graph is of the entire ECD dimer, not a measure for the local binding site region. An additional RMSD of local binding site would be much more telling. You could have a structural isomerisation in the site and not even notice it in the existing graph. On page 9 the authors write that the RMSF in Figure S2 showed instability mainly in loops C and F around the pocket. Given this flexibility at the alpha-delta interface, this is why collecting those regions into one group for the calculation of RMSD convergence analysis would have been useful. They then state "the final MD configuration (with CCh) was well-aligned with the CCh-bound cryo-EM desensitized structure (7QL6)... further demonstrating that the simulation had converged." That may suggest a change occurred that is in common with the global minimum seen in cryo EM, which is good, but does not prove the MD has "converged". I would also rename Figure S3 accordingly.

The description is now changed to “aligns well” with desensitized structure (7QL6.PDB)”. RMSD of not just the binding pocket but the whole ECD dimer is well aligned with first apo (m1) and with desensitized state (m3).

The authors draw conclusions about the dominant states and pathways from their PCA component free energy projections that need clarification. It is important first to show data to demonstrate that the two PCA components chosen were dominant and accounted for most of the variance. Then when mapping free energy as a function of those two PCA components, to prove that those maps have sufficient convergence to be able to interpret them. Moreover, if the free energies themselves cannot be used to measure state stability (as seems to be the case), that the limitations are carefully explained. First, was PCA done on all MD trajectories combined to find a common PC1 & PC2, or were they done separately on each simulation? If so, how similar are they? The authors write "the first two principal components (PC-1 and PC-2) that capture the most pronounced C. displacements". How much of the total variance did these two components capture? The authors write the changes mostly concern loop C and loop F, but which data proves this? e.g. A plot of PC1 and PC2 over residue number might help.

The PCA analyses have been enriched. Figure 3-Source Data 1. shows the dominance of PC1 and PC2. Because the binding energy match was sufficient to identify affinity states, we did not explore additional PCs. Residue-wise PC1 and PC2 analysis and comparison with RMSF are in Figure 2-figure supplement 2. PC1 and PC2 both correlate with fluctuations in loops C and F. Overlap analysis in different runs is shown in Figure 3-figure supplement 1. Lower variance in a particular region of the PCA landscape indicates that the system frequently visits these states, suggesting stability (a preference for these conformations).

The authors map the -kTln rho as a free energy for each simulation as a function of PC1 & PC2. It is important to reveal how well that PC1-2 space was sampled, and how those maps converged over time. The shapes of the maps and the relative depths of the wells look very different for each agonist. If the maps were sampled well and converged, the free energies themselves would tell us the stabilities of each state. Instead, the authors do not even mention this and instead talk about "variance" being the indicator of stability, stating that m3 is most stable in all cases. While I can believe 200ns could not converge a PC1-2 map and that meaningful delta G values might not be obtained from them, the issue of lack of sampling must be dealt with. On page 12 they write "Although the bottom of the well for 3 energy minima from PCA represent the most stable overall conformation of the protein, they do not convey direct information regarding agonist stability or orientation". The reasons why not must be explained; as they should do just that if the two order parameters PC1 and PC2 captured the slowest degrees of freedom for binding and sampling was sufficient. The authors write that "For all agonists and trajectories, m3 had the least variance (was most stable), again supporting convergence by 200 ns." Again the issue of actual free energy values in the maps needs to be dealt with. The probabilities expressed as -kTln rho in kcal/mol might suggest that m2 is the most stable. Instead, the authors base stability only on variance (I guess breadth of the well?), where m3 may be more localised in the chosen PC space, despite apparently having less preference during the MD (not the lowest free energy in the maps).

The motivations and justifications for the use of approximate PBSA energetics instead of atomistic MD free energies should be dealt with in the manuscript, with limitations more clearly discussed. Rather than using modern all-atom MD free energy methods for relative or absolute binding free energies, the author selects clusters from their identified states and does Poisson-Boltzmann estimates (electrostatic, vdW, surface area, vibrational entropy). I do believe the following sentence does not begin to deal with the limitations of that method: "there are limitations with regard to MM-PBSA accurately predicting absolute binding free energies (Genheden & Ryde, 2015; Hou et al., 2011) that depends on the parameterization of the ligand (Oostenbrink et al., 2004)." What are the assumptions and limitations in taking continuum electrostatics (presumably with parameters for dielectric constants and their assignments to regions after discarding solvent), surface area (with its assumptions and limitations), and of course assuming vibration of a normal mode can capture entropy. On page 30, regarding their vibrational entropy estimate, they write that the "entropy term provides insights into the disorder within the system, as well as how this disorder changes during the binding process". It is important that the extent of disorder captured by the vibrational estimate be discussed, as it is not obvious that it has captured entropy involving multiple minima on the system's true 3N-dimensional energy surface, and especially the contribution from solvent disorder in bound Vs dissociated states.

As discussed above, errors in the free energy estimates need to be more faithfully represented, as fractional errors are not meaningful. On page 21 the authors write "The match improved when free energy ratios rather than absolute values were compared." But a ratio of free energies is not a typical or expected measure of error in delta G. They also write "For ACh and CCh, there is good agreement between.Gm1 and GLA and between.Gm3 and GHA. For these agonists, in silico values overestimated experimental ones only by ~8% and ~25%. The agreement was not as good for the other 2 agonists, as calculated values overestimated experimental ones by ~45%(Ebt) and ~130% (Ebt). However, the fractional overestimation was approximately the same for GLA and GHA." See the above comment on how this may misrepresent the error. On page 21 they write, in relation to their large fractional errors, that they "do not know the origin of this factor but speculate that it could be caused by errors in ligand parameterization". However the estimates from the PBSA approach are, by design, only approximate. Both errors in parameterisation (and their likely origin) and the approximate model used, need discussion.

Again, the goal of calculating binding free energy was to identify structural correspondence to LA and HA and not to obtain absolute binding free energy values. Along with the least variance (distribution) for the principle component for m3, it also had the highest binding free energy. An association of m1 to LA and m3 to HA was done after comparing them to experimental values (efficiencies). This comparison not only validates our approach but also underscores the utility of PBSA in supplementing MD and PCA analyses with broader energetics perspectives.

Reviewer #3 (Public Review):

Weaknesses:

Although the match in simulated vs experimental energies for two ligands was very good, the calculated energies for two other ligands were significantly different than the experiment. It is unclear to what extent the choice of method for the energy calculations influenced the results. See above.

A control simulation, such as for an apo site, is lacking. Figure 2-figure supplement 2. shows the results of 200 ns MD simulations of the apo structure (n=2).

Reviewer #4 (Public Review):

Weaknesses:

Timescales (200 ns) do not capture global rearrangements of the extracellular domain, let alone gating transitions of the channel pore, though this work may provide a launching point for more extended simulations. A more general concern is the reproducibility of the simulations, and how representative states are defined. It is not clear whether replicates were included in principal component analysis or subsequent binding energy calculations, nor how simulation intervals were associated with specific states.

We are interested eventually in using MD to study the full isomerization, but these investigations are for the future and likely will involve full length pentamers and longer timescales. However, in response to this query we have in the Discussion raised this issue and offer speculations. See above, PCA has be compared between replicates (Figure 3-figure supplement 1).

Structural analysis largely focuses on snapshots, with limited direct evidence of consistency across replicates or clusters. Figure legends and tables could be clarified.

Snapshots and distance measurements (Figure 6-table supplement 1) were extracted from m1, m2 and m3 plateau regions of trajectories. Incorporated in the legend.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

This study gives interesting insights into the possible dynamics of ligand binding in ACh receptors and establishes some prerequisites for necessary and urgent further work. The broad interest in this receptor class means this work will have some reach.

Suggestions:

(1) I found the citation of relevant literature to be rather limited. In the following paper, the agonist glutamate was shown to bind in two different orientations, and also to convert. These are much longer simulations than what is presented here (nearly 50 µs), which allowed a richer view of conformational changes and ligand binding dynamics in the AMPA Receptor. Albert Lau has published similar work on NMDA, delta, and kainate receptors, including some of it in eLife. Perhaps the authors could draw some helpful comparisons with this work.

Yu A et al. (2018) Neurotransmitter Funneling Optimizes Glutamate Receptor Kinetics. Neuron

Likewise, the comparison to a similar piece of work on glycine receptors (not cited, https://pubs.acs.org/doi/10.1021/bi500815f) could be instructive. Several similar computational techniques were used, and interactions observed (in the simulations) between the agonist and the receptor were tested in the context of wet experiments. In the absence of an equivalent process in this paper (no findings were tested using an orthogonal approach, only compared against known results, from perhaps a narrow spectrum of papers), we have to view the major findings of the paper (docking in cis that leads to a ligand somersault) with some hesitancy.

The Gharpure 2019 paper is cited in the context of the delta subunit but this paper was about a3b4 neuronal nicotinic receptors. This could be tidied up. Also, the simulations from that paper could be used as an index of the stability of the HA state (if ligand orientation is being cited as transferrable, other observations could be too).

New citations have been added. It is difficult to generalize from Yu A and Yu R eta al, because in neither study was the ligand orientation associated with LA versus HA binding energy.

(2) "To start, we associated the agonist orientation in the hold end states as cis in AC-LA versus trans in AC-HA."

I think this a valid start, but one is left with the feeling that this is all we have and the validity of the starting state is not tested. What was really shown here? Is the docking reliable? What evidence can the authors summon for the ligand orientation that they use as a starting structure? In addition to docking energies, the match between PBSA and electrophysiology Gs and temporal sequence (m1-m2-m3) support the assignment.

Given that these simulations cover a circumscribed part of the binding process, I think the limitations should be acknowledged. Indeed the authors do mention a number of remaining open questions.

Paragraphs regarding 'catch' have been added to the Discussion.

(3) Results around line 90. Hypothetical structures and states that were determined from Markov analyses are discussed as if they are well understood and identified. Plausible though these are, I think the text should underline at least the source of such information. In these simulations, a further intermediate has been identified.

The model in Figure 1B was first published in 2012 and has been used and extended over the intervening years. In our lab, catch-and-hold is standard. We have published many papers (in top journals), plus reviews, regarding this scheme. We made presentations that are on Youtube. Here, at the end of the Introduction we now cite a new review article (Biophysical Journal, 2024). I am not sure what more we can do to raise awareness regarding catch and hold.

(4) The figures are dense and could be better organised. Figure 2 is key but has a muddled organization. The placement of the panel label (C) makes it look like the top row (0 ns) is part of (A). Panel B- what is shown in the oval inset (not labeled or in legend). Why not show more than one view, perhaps a sequence of time points? It is confusing to change the colour of the loops in (C). Please show the individual values in D.

Figure 2 has been redone.

(5) A lot is made of the aK145 salt bridge with aD200 and the distances - but I didn't see any measurements, or time course. This part is vague to the point of having no meaning ("bridge tightening").

We present a Table of distance measurements in the SI (Figure 6-table supplement 1).

Reviewer #2 (Recommendations For The Authors):

All main comments have been given in the above review. There are a few other minor comments below.

The 4 agonists examined were acetylcholine (ACh), carbamylcholine (CCh), epibatidine (Ebt), and epiboxidine (Ebx). Could the choices be motivated for the reader?

New in Methods: the agonists are about the same size yet represent different efficiency classes (citation to companion eLife paper). One of our (unmet) objectives was to understand the structural correlates of agonist efficiency.

The authors write that state structures generated in the MD simulation were identified by aligning free energy values with those from experiments. It would be good to explain to the reader, in the introduction, how LA and HA free energies were extracted from experiments, rather than relying on them to read older papers.

In the Introduction, we say that to get G, just measure an equilibrium constant and take the log. We think it is excessive to explain in detail in this paper how to measure the equilibrium binding constants (several methods suffice). However, we have added in Methods our basic approach: measure KLA and L2 by using electrophysiology, and compute KHA from the thermodynamic cycle using L0. We think this paper is best understood in the context of its companion, also in eLife.

In all equilibrium equations of the type A to B (e.g. on page 5), rather than using "=" signs it would be much better to use equilibrium reversible arrow symbols.

It is incorporated.

Reviewer #3 (Recommendations For The Authors):

(1) Although the match in simulated vs experimental energies for two ligands was very good, the calculated energies for Ebt and Ebx were significantly different than the experiment. Are there any alternative methods for calculating binding energies from the MD simulations that could be readily compared to?

See above. We did not use more sophisticated energy calculations because we already knew the answers. Our objective was to identify states, not to calculate energies.

(2) It would be nice to see control simulations of an apo site to ensure that the conformational changes during the MD are due to the ligands and not an artifact of the way the system is set up. I am primarily asking about this as the simulation of the isolated ECDs for the binding site interface seems like it may be unhappy without the neighboring domains that would normally surround it. On that note, was the protein constrained in any way during the MD?

Apo simulation results are presented in Figure 2-figure supplement 2. The dimer interface seems to be adequate (stable).

(3) Figure 4A-B: Should the colors for m1 and m3 be reversed?

Colors have been changed and a bar chart has been added.

Reviewer #4 (Recommendations For The Authors):

(1) Although simulations are commendably run in triplicate, it is difficult in some places to discern their consistency.

(1a) Table S1 provides important quantification of deviations in different replicates and with different agonists. Please confirm that the reported values are accurate. All values reported for the epibatidine system are identical to those reported for carbamylcholine, which seems statistically improbable. Similarly, runs 1 and 3 with epiboxidine seem identical to one another, and runs 1 and 2 with acetylcholine are nearly the same.

Figure 2-Source Data 1 has been corrected.

(1b) In reference to Figure S3, the authors comment that the simulated system (one replicate with carbamylcholine) converges within 0.5 Å RMSD of a desensitized experimental structure. This seems amazing; please specify over what atoms this deviation was calculated and with reference to what alignment. It would be interesting to know the reproducibility of this remarkable convergence in additional replicates or with other ligands; for example, Figure 5 indicates that loop C transitions to a lesser extent in the context of epibatidine than other agonists.

The comparison was for the entire dimer ECD; 0.5 Å is the result. It may be worthwhile to pursue this remarkable convergence, but not in this paper. Here, we are concerned with identifying ACLA and ACHA. Similarity between ACHA and AD structures is for a different study.

(1c) For principal-component and subsequent analyses, it appears that only one trajectory was considered for each system. Please clarify whether this is the case; if so, a rationale for the selection would be helpful, and some indication of how reproducible other replicates are expected to be.

We have added new PCA results (Results, Figure 3-figure supplement 1) that show comparable principal components in other replicates.

(2) Figure 3 shows free energy landscapes defined by principal components of fluctuation in Cα positions.

(2a) Do experimental structures (e.g. PDB IDs 6UWZ, 7QL6u) project onto any of these landscapes in informative ways?

6UWZ.pdb matches well with the apo (7QKO.pdb), comparable to m1, and 7QL6.pdb with the m3.

(2b) Please indicate the meaning of colored regions in the righthand panels.

The color panels in the top left panel indicate the colored regions in the righthand panel also, which is indicative of direction and magnitude of changes with PC1 and PC2.

(2c) Please also check the legend; do the porcupine plots really "indicate the direction and magnitude of changes between PC1 and PC2," or rather between negative and positive values of each principal component?

It indicates the direction and magnitude of changes with PC1 and PC2.

(3) It would be helpful to clarify how trajectory segments were assigned to specific minima, particularly m2 and m3.

(3a) Please verify the timeframes associated with the m2 minima, reported as "20-50 ns [with acetylcholine], 50-60 ns [with carbamylcholine], 60-100 ns [with epibatidine, and] 100-120 ns [with epiboxidine]." It seems improbable that these intervals would interleave so precisely in independent systems. Furthermore, the intervals associated with acetylcholine and epiboxidine do not appear to correspond to the m2 regions indicated in Figure S8.

Times are given in Figure 4-Source Data 1 and Figure 3-figure supplement 2. The m2 classification is based on loop displacement as well as agonist orientation. For all agonists, the selection was strictly from PCA and cluster analysis.

(3b) The text (and legend to Figure 3) indicate that 180+ ns of each trajectory was assigned to m3, which seems surprisingly consistent. However, Figure S5 indicates this minimum is more variable, appearing at 160 ns with acetylcholine but at 186 ns with carbamylcholine. Please clarify.

see above: the selection was from PCA and cluster analysis. Times are in Figure 3-figure supplement 2 and also in Figure 4-Source Data 1 (none in Fig. 3 legend).

(3c) Figures 5, 6, S6, and S7 illustrate structural features of free-energy minima in each ligand system. Please clarify what is shown, e.g. a representative snapshot, centroid, or average structure from a particular prominent cluster associated with a given minimum.

They are all representative snapshots (now in Methods). Snapshots and distance measurements (Figure 6-table supplement 1) were extracted from m1, m2 and m3 plateau regions of trajectories.

(4) Figure S4 helpfully shows the behavior of a pentameric control system; however, some elements are unclear.

(4a) The 2.5-6.5 Å jump in RMSD at ~40 ns seems abrupt; can it be clarified whether this corresponds to a transition to either m2 or m3 poses, or to another feature of e.g. alignment?

Figure 2-figure supplement 4 left bottom is just the ligand. The jump is the flip, m1 to m2.

(4b) It seems difficult to reconcile the apparently bimodal distribution of states with the proposed 3-state model. Into which RMSD peak would the m2 intermediate fall?

The simulations are only to 100 ns, where we found a complete flip of the agonist represented in the histograms. This confirmed that dimer showed similar pattern as the pentamer. In depth analysis was only done only on dimers.

(4c) The top panel is labeled "Com" with a graphical legend indicating "ACh." Does this indicate the ligand or, as described in the text legend, "the pentamer" (i.e. the receptor)? For both panels, please verify whether they are calculated on the basis of center-of-mass, heavy atoms, Cα, etc.

"Com" (for complex) has been changed to system (protein+ligand).

(5) Minor concerns:

(5a) In Figures 1 and S3, correct the PDB references (6UWX and 7QL7 are not nAChRs).

They are now corrected.

(5b) In Figure 4, do all panels represent mean {plus minus} standard deviation calculated across all cluster-frames reported in Table 1?

Yes.

Also check the graphical legend in panel A: presumably the red bars correspond to m1/LA, and the blue to m3/HA?

Corrected

(5c) In the legend to Figure S1, please clarify that panel B is reproduced from Indurthi & Auerbach 2023.

This figure has been deleted.

(5d) As indicated in Figure S2, it seems surprising that the RMSF is so apparently low at the periphery, where the subunits should contact neighbors in the extracellular domain; how might the authors account for this? Specify whether these results apply to all replicates of each system.

The redness in the periphery for all four systems indicates the magnitude of fluctuation. As we focus on the orthosteric site, we highlight the loops around the agonist binding pocket and kept other regions 75% transparent. We now include Apo simulations and the dimer appears to be stable even without an agonist present.

(5e) Within each minimum in Figure S5, three "prominent" clusters appear to be colored (by heteroatom) with carbons in cyan, pink, and yellow respectively. If this is correct, note these colors in the text legend.

Colors have been added to the legend.

(5f) In Figure S6, note in the legend that key receptor sidechains are shown as spheres, with the ligand as balls-and-sticks, and that ligand conformations in both low- and high-affinity complexes are shown in both receptor states for comparison.

This is now added in the legend.

(5g) The legend to Figure S6 also notes "The agonists are as in Fig S4," but that figure contains a single replicate of a different system; please check this reference.

This has been updated to Figure 5.

(5h) In Figure S8, the colors in the epibatidine system appear different from the others.

The colors are the same for m1, m2 and m3 in all systems including epibatidine.

(5i) In Table 1, does "n clusters" indicate the number of simulation frames included in the three prominent clusters chosen for MM-PBSA analysis? Perhaps "n frames" would be more clear.

It was a good suggestion. It has now been changed to ‘n frames’

(5j) Pg 24-ln 453 presumably should read "...that separate it from m1 and m3..."

This sentence is now changed in the discussion.

eLife assessment

This useful work provides insight into agonist binding to nicotinic acetylcholine receptors, which is the stimulus for channel activation that regulates muscle contraction at the neuromuscular junction. The authors use in silico methods to explore the transient conformational change from a low to high affinity agonist-bound conformation as occurs during channel opening, but for which structural information is lacking owing to its transient nature. The evidence supporting the main conclusion that ligands flip ~180 degrees in the binding site as it transitions from a low to high affinity bound conformation is incomplete because little support is available for the starting low affinity docked conformations, and the rather approximate methods for computing binding free energies differ significantly from experimental measures for two of the four tested ligands. Nonetheless, this work presents an intriguing possibility for the nature of a transient conformational change at the agonist binding site correlated with channel opening. If the ligand flip observed in these simulations can be reproduced or verified by other studies, then this work would stand as a significant advance in our knowledge of nicotinic receptor gating.

Reviewer #1 (Public Review):

Summary:

The authors want to understand fundamental steps in ligand binding to muscle nicotinic receptors using computational methods. Overall, although the work provides new information and support for existing models of ligand activation of this receptor type, some limitations in the methods and approach mean that the findings are not as conclusive as hoped.

Strengths:

The strengths include the number of ligands tried, and the comparison to the existing mature analysis of receptor function from the senior author's lab.

Weaknesses:

The weakness are the brevity of the simulations, the concomitant lack of scope of the simulations, the lack of depth in the analysis and the incomplete relation to other relevant work. The free energy methods use seem to lack accuracy - they are only correct for 2 out of 4 ligands.

Reviewer #2 (Public Review):

Summary:

The aim of this manuscript is to use molecular dynamics (MD) simulations to describe the conformational changes of the neurotransmitter binding site of a nicotinic receptor. The study uses a simplified model including the alpha-delta subunit interface of the extracellular domain of the channel and describes the binding of four agonists to observe conformational changes during the weak to strong affinity transition.

Strength:

The 200 ns-long simulations of this model suggest that the agonist rotates about its centre in a 'flip' motion, while loop C 'flops' to restructure the site. The changes appear to reproduced across simulations and different ligands and are thus a strong point of the study.

Weaknesses:

After carrying out all-atom molecular dynamics, the authors revert to a model of binding using continuum Poisson-Boltzmann, surface area and vibrational entropy. The motivations for and limitations associated with this approximate model for the thermodynamics of binding, rather than using modern atomistic MD free energy methods (that would fully incorporate configurational sampling of the protein, ligand and solvent) could be provided. Despite this, the authors report correlation between their free energy estimates and those inferred from experiment. This did, however, reveal shortcomings for two of the agonists. The authors mention their trouble getting correlation to experiment for Ebt and Ebx and refer to up to 130% errors in free energy. But this is far worse than a simple proportional error, because -24 Vs -10 kcal/mol is a massive overestimation of free energy, as would be evident if it the authors were to instead to express results in terms of KD values (which would have error exceeding a billion fold). The MD analysis could be improved with better measures of convergence, as well as more careful discussion of free energy maps as function of identified principal components, as described below. Overall, however, the study has provided useful observations and interpretations of agonist binding that will help understand pentameric ligand-gated ion channel activation.

Main points:

Regarding the choice of model, some further justification of the reduced 2 subunit ECD-only model could be given. On page 5 the authors argue that, because binding free energies are independent of energy changes outside the binding pocket, they could remove the TMD and study only an ECD subunit dimer. While the assumption of distant interactions being small seems somewhat reasonable, provided conformational changes are limited and localised, how do we know the packing of TMD onto the ECD does not alter the ability of the alpha-delta interface to rearrange during weak or strong binding? They further write that "fluctuations observed at the base of the ECD were anticipated because the TMD that offers stability here was absent.". As the TMD-ECD interface is the "gating interface" that is reshaped by agonist binding, surely the TMD-ECD interface structure must affect binding. It seems a little dangerous to completely separate the agonist binding and gating infrastructure, based on some assumption of independence. Given the model was only the alpha and delta subunits and not the pentamer with TMD, I am surprised such a model was stable without some heavy restraints. The authors state that "as a further control we carried out MD simulation of a pentamer docked with ACh and found similar structural changes at the binding pocket compared to the dimer." Is this sufficient proof of the accuracy of the simplified model? How similar was the model itself with and without agonist in terms of overall RMSD and RMSD for the subunit interface and the agonist binding site, as well as the free energy of binding to each model to compare?

Although the authors repeatedly state that they have good convergence with their MD, I believe the analysis could be improved to convince us. On page 8 the authors write that the RMSD of the system converged in under 200 ns of MD. However, I note that the graph is of the entire ECD dimer, not a measure for the local binding site region. An additional RMSD of local binding site would be much more telling. You could have a structural isomerisation in the site and not even notice it in the existing graph. On page 9 the authors write that the RMSF in Fig.S2 showed instability mainly in loops C and F around the pocket. Given this flexibility at the alpha-delta interface, this is why collecting those regions into one group for the calculation of RMSD convergence analysis would have been useful. They then state "the final MD configuration (with CCh) was well-aligned with the CCh-bound cryo-EM desensitized structure (7QL6)... further demonstrating that the simulation had converged." That may suggest a change occurred that is in common with the global minimum seen in cryo EM, which is good, but does not prove the MD has "converged". I would also rename Fig.S3 accordingly.

The authors draw conclusions about the dominant states and pathways from their PCA component free energy projections that need clarification. It is important first to show data to demonstrate that the two PCA components chosen were dominant and accounted for most of the variance. Then when mapping free energy as a function of those two PCA components, to prove that those maps have sufficient convergence to be able to interpret them. Moreover, that if the free energies themselves cannot be used to measure state stability (as seems to be the case), that the limitations are carefully explained. First, was PCA done on all MD trajectories combined to find a common PC1 & PC2, or were they done separately on each simulation? If so, how similar are they? The authors write "the first two principal components (PC-1 and PC-2) that capture the most pronounced C. displacements". How much of the total variance did these two components capture? The authors write the changes mostly concern loop C and loop F, but which data proves this? e.g. A plot of PC1 and PC2 over residue number might help?

The authors map the -kTln rho as a free energy for each simulation as function of PC1 & PC2. It is important to reveal how well that PC1-2 space was sampled, and how those maps converged over time. The shapes of the maps and the relative depths of the wells look very different for each agonist. If the maps were sampled well and converged, the free energies themselves would tell us the stabilities of each state. Instead, the authors do not even mention this and instead talk about "variance" being the indicator of stability, stating that m3 is most stable in all cases. While I can believe 200ns could not converge a PC1-2 map and that meaningful delta G values might not be obtained from them, the issue of lack of sampling must be dealt with. On page 12 they write "Although the bottom of the well for 3 energy minima from PCA represent the most stable overall conformation of the protein, they do not convey direct information regarding agonist stability or orientation". The reasons why not must be explained; as they should do just that if the two order parameters PC1 and PC2 captured the slowest degrees of freedom for binding and sampling was sufficient. The authors write that "For all agonists and trajectories, m3 had the least variance (was most stable), again supporting convergence by 200 ns." Again the issue of actual free energy values in the maps needs to be dealt with. The probabilities expressed as -kTln rho in kcal/mol might suggest that m2 is the most stable. Instead, the authors base stability only on variance (I guess breadth of the well?), where m3 may be more localised in the chosen PC space, despite apparently having less preference during the MD (not the lowest free energy in the maps).

The motivations and justifications for use of approximate PBSA energetics instead of atomistic MD free energies should be dealt with in the manuscript, with limitations more clearly discussed. Rather than using modern all-atom MD free energy methods for relative or absolute binding free energies, the author select clusters from their identified states and do Poisson-Boltzmann estimates (electrostatic, vdW, surface area, vibrational entropy). I do believe the following sentence does not begin to deal with the limitations in that method: "there are limitations with regard to MM-PBSA accurately predicting absolute binding free energies (Genheden & Ryde, 2015; Hou et al., 2011) that depends on parameterization of the ligand (Oostenbrink et al., 2004)." What are the assumptions and limitations in taking a continuum electrostatics (presumably with parameters for dielectric constants and their assignments to regions after discarding solvent), surface area (with its assumptions and limitations) and of course assuming vibration of a normal mode can capture entropy. On page 30, regarding their vibrational entropy estimate, they write that the "entropy term provides insights into the disorder within the system, as well as how this disorder changes during the binding process". It is important that the extent of disorder captured by the vibrational estimate be discussed, as it is not obvious that it has captured entropy involving multiple minima on the system's true 3N-dimensional energy surface, and especially the contribution from solvent disorder in bound Vs dissociated states.

As discussed above, errors in the free energy estimates need to be more faithfully represented, as fractional errors are not meaningful. On page 21 the authors write "The match improved when free energy ratios rather than absolute values were compared." But a ratio of free energies is not a typical or expected measure of error in delta G. They also write "For ACh and CCh, there is good agreement between.Gm1 and GLA and between.Gm3 and GHA. For these agonists, in silico values overestimated experimental ones only by ~8% and ~25%. The agreement was not as good for the other 2 agonists, as calculated values overestimated experimental ones by ~45%(Ebt) and ~130% (Ebt). However, the fractional overestimation was approximately the same for GLA and GHA." See above comment on how this may misrepresent the error. On page 21 they write, in relation to their large fractional errors, that they "do not know the origin of this factor but speculate that it could be caused by errors in ligand parameterization". But the estimates from the PBSA approach are, by design, only approximate. Both errors in parameterisation (and their likely origin) and the approximate model used, need discussion.

Reviewer #3 (Public Review):

Summary:

The authors use docking and molecular dynamics (MD) simulations to investigate transient conformations that are otherwise difficult to resolve experimentally. The docking and simulations suggest an interesting series of events whereby agonists initially bind to the low affinity site and then flip 180 degrees as the site contracts to its high affinity conformation. This work will be of interest to the ion channel community and to biophysical studies of pentameric ligand-gated channels.

Strengths:

I find the premise for the simulations to be good, starting with an antagonist bound structure as an estimate of the low affinity binding site conformation, then docking agonists into the site and using MD to allow the site to relax to a higher affinity conformation that is similar to structures in complex with agonists. The predictions are interesting and provide a view into what a transient conformation that is difficult to observe experimentally might be like.

Weaknesses:

A weakness is that the relevance of the initial docked low affinity orientations depend solely on in silco results, for which simulated vs experimental binding energies deviate substantially for two of the four ligands tested. This raises some doubt as to the validity of the simulations. I acknowledge that the calculated binding energies for two of the ligands were closer to experiment, and simulated efficiencies were a good representation of experimental measures, which gives some support to the relevance of the in silico observations. Regardless, some of the reviewers comments regarding the simulation methodology were not seriously addressed.

Reviewer #4 (Public Review):

Summary:

In their revised manuscript "Conformational dynamics of a nicotinic receptor neurotransmitter binding site," Singh and colleagues present molecular docking and dynamics simulations to explore the initial conformational changes associated with agonist binding in the muscle nicotinic acetylcholine receptor, in context with the extensive experimental literature on this system. Their central findings are of a consistently preferred pose for agonists upon initial association with a resting channel, followed by a dramatic rotation of the ligand and contraction of a critical loop over the binding site. Principal component analysis also suggests the formation of an intermediate complex, not yet captured in structural studies. Binding free energy estimates are consistent with the evolution of a higher-affinity complex following agonist binding, with a ligand efficiency notably similar to experimental values. Snapshot comparisons provide a structural rationale for these changes on the basis of pocket volume, hydration, and rearrangement of key residues at the subunit interface.

Strengths:

Docking results are clearly presented and remarkably consistent. Simulations are produced in triplicate with each of four different agonists, providing an informative basis for internal validation. They identify an intriguing transition in ligand pose, not well documented in experimental structures, and potentially applicable to mechanistic or even pharmacological modeling of this and related receptor systems. The paper seems a notable example of integrating quantitative structure-function analysis with systematic computational modeling and simulations, likely applicable to the wider journal audience.

Weaknesses:

The response to initial review is somewhat disappointing, declining in some places to implement suggested clarifications, and propagating apparent errors in at least one table (Fig 2-source data 1). Some legends (e.g. Fig 2-supplement 4, Fig 3, Fig 4) and figure shadings (e.g. Fig 2-supplement 2, Fig 6-supplement 2) remain unclear. Apparent convergence of agonist-docked simulations towards a desensitized state (l 184) is difficult to interpret in absence of comparative values with other states, systems, etc. In more general concerns, aside from the limited timescales (200 ns) that do not capture global rearrangements, it is not obvious that landscapes constructed on two principal components to identify endpoint and intermediate states (Fig 3) are highly robust or reproducible, nor whether they relate consistently to experimental structures.

Author Response

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

Thank you and the two reviewers for the thorough review of our manuscript. We thank you very much for the positive evaluation of our manuscript and your encouragement to continue in this fascinating topic. In this version we made minor changes in the text to address the comments and suggestion of the second reviewer and increase the clarity of the text.

Reviewer #2 Recommendation to the authors

We thank the reviewer for the sharp comments that help us improve the clarity of the paper. Below we list the changes we made to correct and revise the paper in accordance to the reviewer’s comments.

(1) Line 90. Isn't the genus Paracentrotus?

Yet it is, thank you. We corrected the typo.

(2) Figure 1 and supplementary figure 2. To this reviewer supplementary Figure 2 doesn't really help the story as written in the paragraph from line 96-110. You want to report expression of ROCK in skeletogenic cells. You do that quite well in Figure 1. Since Fig. S2 reports whole embryo expression of ROCK when only 5% of the cells in the embryo are the subject of interest here, and the Axitinib is selective, presumably for skeletogenic cells, the relative lack of effect in Fig. S2 is not surprising and again, doesn't really help the theme you wish to establish by focusing on the role of ROCK in skeletogenic cells over time. If anything, the data reported in Fig. S2 shows that perturbation of VEGF signaling has very little effect embryo-wide, while Fig. 1 shows that perturbation of VEGF signaling has a noticeable effect on ROCK expression in skeletogenic cells. If you choose to keep Fig. S2, I recommend that you indicate that embryo-wide vs skeletogenic cell difference more succinctly than given at present. It will also strengthen your paragraph in lines 110-127.

The importance of the western blot presented in Fig. S2 is to validate that the antibody recognizes a protein of the expected size. This strengthen the credibility of this commercial antibody to detect the sea urchin ROCK protein. We agree with the reviewer that the fact that the skeletogenic cells are less than 5% of the embryonic cells is important to explain why we didn’t see an affect of VEGFR inhibition in the western blot, and we changed the text to express it (lines 108-111): “Yet, this measurement was done on proteins extracted from whole embryos, of which the skeletogenic cells, where VEGFR is active, are less than 5% of the total cell mass (42). We therefore wanted to study the spatial expression of ROCK and specifically, its regulation in the skeletogenic cells.”

(3) Comparison of Fig. 2 and Fig. S3. To me the reader is confused when Fig. S3 is 33hpf as reported in the text (but not in the figure legend), and Fig. 2 shows 2 day old embryos - on the figure and figure legend but not in the text. So, the reader sees the text indicating 33hpf and looks around and the figure 2 says 2dpf. Does that mean 33hpf = 2dpf, the reader is thinking. To clarify, I suggest including the 2dpf in the text or simply drop the time in the text and report it in the two figures. Further, in the middle of the paragraph 130-143 you switch from reporting on Fig.S3 to Fig. 2, yet the reader doesn't know that. The reader is still looking at Fig. S3. The problem here is that at 33hpf the skeleton doesn't yet show the reduction or abnormalities that are shown later at 2dpf in Fig. 2. In clarifying this paragraph both the reduction in ROCK expression and the subsequent alterations in growth and patterning of the skeleton will be clear to the reader.

Thank you for raising this point. We added in the caption of Fig. S3 that the measurements were done in 33hpf. We also added in the text, that the observations of the skeletogenic phenotypes were done at 2dpf (48hpf). We made a break between the first paragraph discussing Fig. S3 and the paragraph discussing Fig. 2.

(4) The experiment with Y27632, an inhibitor of ROCK, is significantly improved in this revision. The concern earlier was the possibility that at the concentration used there might be off-target effects since other kinases are affected by higher concentrations of this selective inhibitor. The authors have modified this component of the paper and performed experiments at lower concentrations where other reports indicate the inhibitor is highly selective for ROCK, and they still demonstrate an inhibition of skeletal production. This, plus the added citations greatly increases confidence that this inhibition is selective for ROCK, thus enabling a stronger conclusion that ROCK has a role in skeletal growth and patterning.

Thank you for asking us to test this lower concentration which improved the credibility of our findings.

Line 239 - should be: indicating instead of indicting We corrected that.

(5) Line 402-403."The first step in generating the sea urchin spicules is the construction of the spicule cavity, a membrane filled with calcium carbonate and coated with F-actin (Fig. 8A)". I suggest more precise language. The way this now reads (above) is that somehow the spicule cavity is a membrane and that membrane is filled with CaCO3. And further the membrane is coated with F-actin. Isn't the spicule cavity what is filled with CaCO3? And isn't that cavity surrounded by a membrane? And the F-actin must be in the cortex of the cell since there is very little cytoplasm associated with the pseudopodial extensions that surround the spicule.

We change this sentence to: “The first step in generating the sea urchin spicules is the construction of the spicule cavity where the mineral is engulfed in a membrane coated with F-actin” (lines 403-404). Our observations show that F-actin is enriched around the spicule cavity. It could be an extension of the cell cortex, but we did not prove it, so we prefer to simply describe what we saw.

Line 405-408. Thank you for putting in this unknown. It is important to point out that while you've shown that ROCK contributes to regulation of actomyosin, it is not clear whether this is direct or indirect. You have also shown that ROCK somehow contributes to regulation of the GRN that leads to skeletogenesis. Thus, your data are consistent in showing that ROCK perturbation cripples normal skeletogenesis both via morpholino and with a selective inhibitor. Your last part of the discussion then offers speculation as to what might be affected specifically. That discussion sets the stage for digging even deeper to identify specific targets of ROCK activity.

Thank you, we agree with you that there is an exciting road ahead of us!

eLife assessment

This valuable study addresses the role of Rho-associated coiled-coil kinase (ROCK) and the cytoskeleton in the initiation and growth of the calcified endoskeleton of sea urchin embryos. Perturbation by two independent approaches (a morpholino and a selective inhibitor) provides convincing evidence that ROCK participates both in actomyosin regulation and in the gene regulatory network that controls skeletogenesis. Exciting areas of future work will be to elucidate the mechanisms by which ROCK influences gene expression and to further dissect the role of the cytoskeleton in mineralization.

Reviewer #1 (Public Review):

Using a pharmacological and knock-down approach, the authors could demonstrate that ROCK activity is required for the normal development of the larval skeleton. The presence of ROCK in the pluteus stage depends on the activity of VEGF that is responsible for the formation of the tubular syncytial sheath of the calcifying primary mesenchyme cells in which the skeleton forms. The importance of ROCK in skeleton formation was confirmed in cell culture experiments, demonstrating that ROCK inhibition leads to decreased elongation and abnormal branching of spicules. µCT analyses underline this finding demonstrating that the inhibition of ROCK mainly affects elongation of spicules while growth in girth is little affected. F-actin labeling experiments could demonstrate that ROCK inhibition interferes with the organization of the actomyosin network in the early phase of skeleton formation, while f-actin organization in the tips of the elongating spicule is unaffected by the pharmacological inhibition of ROCK. Finally, ROCK inhibition strongly affects the expression of major regulatory and calcification-related genes in the calcifying cells. Based on these findings the authors propose a model for the regulatory interaction between the skeletogenic GRN, ROCK and the f-actin system relevant for skeletogenesis.

Comments on revised version:

In their manuscript Hijaze et al. adequately addressed the majority of my previous concerns in a satisfactory manner. In particular, they validated their morpholino knock-down experiments by explaining how they determined the optimal concentrations and provided an immunohistological evidence for the reduction in ROCK protein abundance. The authors also added new antibody stainings providing evidence that ROCK and F-actin do not interact directly but likely through other kinases that modulate f-actin, and that the localization of f-actin at the spicule tips remains unaffected by the knock-down. In addition, the authors revised their discussion to not overstate their observations, and by focusing on the potential mechanisms by which ROCK may affect biomineralization (i.e. mechano sensing and exocytosis of vesicles). Here I would like to add, that f-actin mediated exocytosis does not necessarily target mineral baring vesicles but may also promote the exocytosis of matrix proteins that are essential for the normal formation of the spicules and that are an integral component of other biominerals, as well. I strongly encourage the authors to continue on this exciting research, including the development of methods to analyze the molecular mechanisms that control vesicular trafficking in mineralizing systems.

Reviewer #2 (Public Review):

This project is on the role of ROCK in skeletogenesis during sea urchin development. That skeleton is produced by a small number of cells in the embryo with signaling inputs from the ectoderm providing patterning cues. The skeleton is built from secretion of CaCO3 by the skeletogenic cells. The authors conclude that ROCK is involved in the regulation of skeletogenesis with a role both in regulating actomyosin in the process, and in the gene regulatory network (GRN) underlying the entire sequence of events.

The strength of the paper is that they show in detail how perturbations of ROCK results in abnormal actomyosin activity in the skeletogenic cells, and they show alterations both in expression of transcription factors of the GRN, and expression of genes involved in assembly of the skeletal matrix. Two different approaches lead to this conclusion: morpholino perturbations and the actions of a selective inhibitor of the kinase activity. Thus, they achieved their goal which was to test the hypothesis that ROCK is involved in the process of skeletogenesis. Those tests support the hypothesis with data that was quantitatively significant.

The discussion was transparent regarding where the analysis ended and where the next phase of work should begin. While actomyosin involvement was altered when ROCK was perturbed, it isn't known how direct or indirect the role of ROCK might be. Also, while the regulatory input to spicule initiation and growth is affected when ROCK is inhibited, it isn't clear exactly where ROCK is involved.

Reviewer #2 (Public Review):

Summary:

In this paper, the authors set out to better understand the mechanism by which the FtsZ-associated protein ZapD crosslinks FtsZ filaments to assemble a large scale cytoskeletal assembly. For this aim, they use purified proteins in solution and a combination of biochemical, biophysical experiments and cryo-EM. The most significant finding of this study is the observation of FtsZ toroids that form at equimolar concentrations of the two proteins.

Strengths:

Many experiments in this paper confirm previous knowledge about ZapD. For example, it shows that ZapD promotes the assembly of FtsZ polymers, that ZapD bundles FtsZ filaments, that ZapD forms dimers and that it reduces FtsZ's GTPase activity.

The most novel discovery is the observation of different assemblies as a function of ZapD:FtsZ ratio. In addition, using CryoEM to describe the structure of toroids and bundles, the papers provides some information about the orientation of ZapD in relation to FtsZ filaments. For example, they found that the organization of ZapD in relation to FtsZ filaments is "intrinsic heterogeneous" and that FtsZ filaments were crosslinked by ZapD dimers pointing in all directions. The authors conclude that it is this plasticity that allows for the formation of toroids and its stabilization. Unfortunately, a high-resolution structure of the protein organization was not possible.

Weaknesses:

While the data is convincing, their interpretation has some substantial weaknesses that the authors should address for the final version of this paper.

For example, as the authors are the first to describe FtsZ-ZapD toroids, a discussion why this has not been observed in previous studies would be very interesting, i.e. is it due to buffer conditions, sample preparation?

At parts of the manuscript, the authors try a bit too hard to argue for the physiological significance of these toroids. This, however, is at least very questionable, because:<br /> The typical diameter is in the range of 0.25-1.0 μm, which requires some flexibility of the filaments to be able to accommodate this. It's difficult to see how a FtsZ-ZapD toroid, which appears to be quite rigid with a narrow size distribution of 502 nm {plus minus} 55 nm could support cell division rather than stalling it at that cell diameter. which the authors say is similar to the E. coli cell.

For cell division, FtsZ filaments are recruited to the membrane surface via an interaction of FtsA or ZipA the C-terminal peptide of FtsZ. As ZapD also binds to this peptide, the question arises who wins this competition or where is ZapD when FtsZ is recruited to the membrane surface? Can such a toroidal structure of FtsZ filaments form on the membrane surface? Additional experiments would be helpful, but a more detailed discussion on how the authors think ZapD could act on membrane-bound filaments would be essential.

The authors conclude that the FtsZ filaments are dynamic, which is essential for cell division. But the evidence for dynamic FtsZ filaments within these toroids seems rather weak, as it is solely the partial reassembly after addition of GTP. As ZapD significantly slows down GTP hydrolysis, I am not sure it's obvious to make this conclusion.

On a similar note, on page 5 the authors claim that ZapD would transiently interact with FtsZ filaments. What is the evidence for this? They also say that this transient interaction could have a "mechanistic role in the functionality of FtsZ macrostructures." Could they elaborate?

The author should also improve in putting their findings into the context of existing knowledge. For example:

The authors observe a straightening of filament bundles with increasing ZapD concentration. This seems consistent with what was found for ZapA, but this is not explicitly discussed (Caldas et al 2019)

A paragraph summarizing what is known about the properties of ZapD in vivo would be essential: i.e. what has been found regarding its intracellular copy number, location and dynamics?

In the introduction, the authors write that "GTP binding and hydrolysis induce a conformational change in each monomer that modifies its binding potential, enabling them to follow a treadmilling behavior". This seems inaccurate, as shown by Wagstaff et al. 2022, the conformational change of FtsZ is not associated with the nucleotide state. In addition, they write that FtsZ polymerization depends on the GTPase activity. It would be more accurate to write that polymerization depends on GTP, and disassembly on GTPase activity.

On page 2 they also write that "the mechanism underlying bundling of FtsZ filaments is unknown". I would disagree, the underlying mechanism is very well known (see for example Schumacher, MA JBC 2017), but how this relates to the large-scale organization of FtsZ filaments was not clear.

The authors describe the toroid as a dense 3D mesh, how would this be compatible with the Z-ring and its role for cell division? I don't think this corresponds to the current model of the Z-ring (McQuillen & Xiao, 2020). Apart from the fact it's a ring, I don't think the organization of FtsZ obviously similar to the current of the Z-ring in the bacterial cell, in particular because it's not obvious how FtsZ filaments can bind ZapD and membrane anchors simultaneously.

The authors write that "most of these modulators" interact with FtsZ's CTP, but then later that ZapD is the only Zap protein that binds CTP. This seems to be inconsistent. Why not write that membrane anchors usually bind the CTP, most Zaps do not, but ZapD is the exception?

I also have some comments regarding the experiments and their analysis:

Regarding cryoET: the filaments appear like flat bands, even in the absence of ZapD, which further elongates these bands. Is this due to an anisotropic resolution? This distortion makes the conclusion that ZapD forms bi-spherical dimers unconvincing.

The authors say that the cryoET visualization provides crucial information on the length of the filaments within this toroid. How long are they? Could the authors measure it?

Regarding the dimerization mutant of ZapD: there is actually no direct confirmation that mZapD is monomeric. Did the authors try SEC MALS or AUC? Accordingly, the statement that dimerization is "essential" seems exaggerated (although likely true).

What do the authors mean that toroid formation is compatible with robust persistence length? I.e. What does robust mean? It was recently shown that FtsZ filaments are actually surprisingly flexible, which matches well the fact that the diameter of the Z-ring must continuously decrease during cell division (Dunajova et al Nature Physics 2023).

the authors claim that their observations suggest „that crosslinkers ... allows filament sliding in an organized fashion". As far as I know there is no evidence of filament sliding, as FtsZ monomers in living cells and in vitro are static.

What is the „proto-ring FtsA protein"?

The authors refer to „increasing evidence" for „alternative network remodling mechanisms that do not rely on chemical energy consumption as those in which entropic forces act through diffusible crosslinkers, similar to ZapD and FtsZ polymers." A reference should be given, I assume the authors refer to the study by Lansky et al 2015 of PRC on microtubules. However, I am not sure how the authors made the conclusion that this applies to FtsZ and ZapD, on which evidence is this assumption based?

Some inconsistencies in supplementary figure 3: The normalized absorbances in panel a do not seem to agree with the absolute absorbance shown in panel e, i.e. compare maximum intensity for ZapD = 20 µM and 5 µM in both panels.

It's not obvious to me why the structure formed by ZapD and FtsZ disassembles after some time even before GTP is exhausted, can the authors explain? As the structures disassemble, how is the "steady-state turbidity" defined? Do the structures also disassemble when they use a non-hydrolyzable analog of GTP?

Conclusion:

Despite some weaknesses in the interpretation of their findings, I think this paper will likely motivate other structural studies on large scale assemblies of FtsZ filaments and its associated proteins. A systematic comparison of the effects of ZapA, ZapC and ZapD and how their different modes of filament crosslinking can result in different filament networks will be very useful to understand their individual roles and possible synergistic behavior.

eLife assessment

The formation of the Z-ring at the time of bacterial cell division interests researchers working towards understanding cell division across all domains of life. The manuscript by Jasnin et al reports the cryoET structure of toroid assembly formation of FtsZ filaments driven by ZapD as the cross linker. The findings are important and have the potential to open a new dimension in the field, but the evidence to support these exciting claims is currently inadequate.

Reviewer #1 (Public Review):

Summary:

The major result in the manuscript is the observation of the higher order structures in a cryoET reconstruction that could be used for understanding the assembly of toroid structures. The cross-linking ability of ZapD dimers result in bending of FtsZ filaments to a constant curvature. Many such short filaments are stitched together to form a toroid like structure. The geometry of assembly of filaments - whether they form straight bundles or toroid like structures - depends on the relative concentrations of FtsZ and ZapD.

Strengths:

In addition to a clear picture of the FtsZ assembly into ring-like structures, the authors have carried out basic biochemistry and biophysical techniques to assay the GTPase activity, the kinetics of assembly, and the ZapD to FtsZ ratio.

Weaknesses:

The discussion does not provide an overall perspective that correlates the cryoET structural organisation of filaments with the biophysical data.

The crosslinking nature of ZapD is already established in the field. The work carried out is important to understand the ring assembly of FtsZ. However, the availability of the cryoET observations can be further analysed in detail to derive many measurements that will help validate the model, and obtain new insights.

Reviewer #3 (Public Review):

Summary:

The authors provide the first image analysis by cryoET of toroids assembled by FtsZ crosslinked by ZapD. Previously toroids of FtsZ alone have been imaged only in projection by negative stain EM. The authors attempt to distinguish ZapD crosslinks from the underlying FtsZ filaments. I did not find this distinction convincing, especially because it seems inconsistent with the 1:1 stoichiometry demonstrated by pelleting. I was intrigued by one image showing straight filament pairs, which may suggest a new model for how ZapD crosslinks FtsZ filaments.

Strengths:

(1) The first image analysis of FtsZ toroids by cryoET.<br /> (2) The images are accompanied by pelleting assays that convincingly establish a 1:1 stoichiometry of FtsZ:ZapD subunits.<br /> (3) Fig. 5 shows an image of a pair of FtsZ filaments crosslinked by ZapD. This seems to have higher resolution than the toroids. Importantly, it suggests a new model for the structure of FtsZ-ZapD that resolves previously unrecognized conflicts. (This is discussed below under weaknesses, because it is so far only supported by a single image.)

Weaknesses:

This paper reports a study by cryoEM of polymers and bundles assembled from FtsZ plus ZapD. Although previous studies by other labs have focused on straight bundles of filaments, the present study found toroids mixed with these straight bundles, and they focused most of their study on the toroids. In the toroids they attempt to delineate FtsZ filaments and ZapD crosslinks. A major problem here is with the stoichiometry. Their pelleting assays convincingly established a stoichiometry of 1:1, while the mass densities identified as ZapD are sparse and apparently well below the number of FtsZ (FtsZ subunits are not resolved in the reconstructions, but the continuous sheets or belts seem to have a lot more mass than the identified crosslinks.) Apart from the stoichiometry I don't find the identification of crosslinks to be convincing. It is missing an important control - cryoET of toroids assembled from pure FtsZ, without ZapD.

However, if I ignore these and jump to Fig. 5, I think there is an important discovery that resolves controversies in the present study as well as previous ones, controversies that were not even recognized. The controversy is illustrated by the Schumacher 2017 model (their Fig. 7), which is repeated in a simplified version in Fig. 1a of the present mss. That model has a two FtsZ filaments in a plane facing ZapD dimers which bridge them. In this planar model the C-terminal linker, and the ctd of FtsZ that binds ZapD facing each other and the ZapD in the middle, with. The contradiction arises because the C-terminus needs to face the membrane in order to attach and generate a bending force. The two FtsZ filaments in the planar model are facing 90{degree sign} away from the membrane. A related contradiction is that Houseman et al 2016 showed that curved FtsZ filaments have the C terminus on the outside of the curve. In a toroid the C termini should all be facing the outside. If the paired filaments had the C termini facing each other, they could not form a toroid because the two FtsZ filaments would be bending in opposite directions.

Fig. 5 of the present mss seems to resolve this by showing that the two FtsZ filaments and ZapD are not planar, but stacked. The two FtsZ filaments have their C termini facing the same direction, let's say up, toward the membrane, and ZapD binds on top, bridging the two. The spacing of the ctd binding sites on the Zap D dimer is 6.5 nm, which would fit the ~8 nm width of the paired filament complex observed in the present cryoEM (Fig S13). In the Schumacher model the width would be about 20 nm. Importantly, the stack model has the ctd of each filament facing the same direction, so the paired filaments could attach to the membrane and bend together (using ctd's not bound by ZapD). Finally, the new arrangement would also provide an easy way for the complex to extend from a pair of filaments to a sheet of three or four or more.

A problem with this new model from Fig. 5 is that it is supported by only a single example of the paired FtsZ-ZapD complex. If this is to be the basis of the interpretation, more examples should be shown. Maybe examples could be found with three or four FtsZ filaments in a sheet.

What then should be done with the toroids? I am not convinced by the identification of ZapD as "connectors." I think it is likely that the ZapD is part of the belts that I discuss below, although the relative location of ZapD in the belts is not resolved. It is likely that the resolution in the toroid reconstructions of Fig. 4, S8,9 is less than that of the isolated pf pair in Fig. 5c.

Importantly, If the authors want to pursue the location of ZapD in toroids, I suggest they need to compare their ZapD-containing toroids with toroids lacking ZapD. Popp et al 2009 have determined a variety of solution conditions that favor the assembly of toroids by FtsZ with no added protein crosslinker. It would be very interesting to investigate the structure of these toroids by the present cryoEM methods, and compare them to the FtsZ-ZapD toroids. I suspect that the belts seen in the ZapD toroids will not be found in the pure FtsZ toroids, confirming that their structure is generated by ZapD.

eLife assessment

The manuscript reports on an important comparison of a range of approved clinical inhibitors for BTK used for the treatment of chronic lymphocytic leukemia (CLL). The authors provide solid evidence for their claims, using a combination of HDX-MS and NMR spectroscopy. The novelty is that this study also seeks to evaluate resistance mutation bias. The scope of the study is highly exciting but would benefit from a clear link of the biophysical studies to the functional assays - specifically nucleotide binding.

Reviewer #1 (Public Review):

Summary:

The work by Joseph et al "Impact of the clinically approved BTK inhibitors on the conformation of full-length BTK and analysis of the development of BTK resistance mutations in chronic lymphocytic leukemia" seeks to comparatively analyze the effect of a range of covalent and noncovalent clinical BTK inhibitors upon BTK conformation. The novel aspect of this manuscript is that it seeks to evaluate the differential resistance mutations that arise distinctly from each of the inhibitors.

Strengths:

This is an exciting study that builds upon the fundamental notion of ensemble behavior in solutions for enzymes such as BTK. The HDX-MS and NMR experiments are adequately and comprehensively presented.

Weaknesses:

While I commend the novelty of the study, the absence of important controls greatly tempers my enthusiasm for this work. As stated in the abstract, there are no broad takeaways for how resistance mutation bias operated from this study, although the mechanism of action of 2 common resistance mutations is useful. How these 2 resistance mutations connect to ensemble behavior, is not obvious. This is partly because BTK does not populate just binary "open"/"closed" conformations, but there are likely multiple intermediate conformations. Each inhibitor appears to preferentially "select" conformations by the authors' own assessment (line 236) and this carries implications for the emergence of resistance mutations. The most important control that would help is to use ADP or nonhydrolyzable and ATP as a baseline to establish the "inactive" and "active" conformations. All of the HDX-MS and NMR studies use protein that has no nucleotide present. A major question that remains is whether each of the inhibitors preferentially favors/blocks ADP or ATP binding. This then means it is not equivalent to correlate functional kinase assay conditions with either HDX-MS or NMR experiments.

Reviewer #2 (Public Review):

Summary:

Previous NMR and HDX-MS studies on full-length (FL) BTK showed that the covalent BTKi, ibrutinib, causes long-range effects on the conformation of BTK consistent with disruption of the autoinhibited conformation, based on HDX deuterium uptake patterns and NMR chemical shift perturbations. This study extends the analyses to four new covalent BTKi, acalabrutinib, zanubrutinib, tirabrutinib/ONO4059, and a noncovalent ATP competitive BTKi, pirtobrutinib/LOXO405.

The results show distinct conformational changes that occur upon binding each BTKi. The findings show consistent NMR and HDX changes with covalent inhibitors, which move helix aC to an 'out' position and disrupt SH3-kinase interactions, in agreement with X-ray structures of the BTKi complexed with the BTK kinase domain. In contrast, the solution measurements show that pirtobrutinib maintains and even stabilizes the helix aC-in and autoinhibited conformation, even though the BTK:pritobrutinib crystallizes with helix aC-out. This and unexpected variations in NMR and HDX behavior between inhibitors highlight the need for solution measurements to understand drug interactions with the full-length BTK. Overall the findings present good evidence for allosteric effects by each BTKi that induce distal conformational changes which are sensitive to differences in inhibitor structure.

The study goes on to examine BTK mutants T474I and L528W, which are known to confer resistance to pirtobrutinib, zanubritinib, and tirabrutinib. T474I reduces and L528W eliminates BTK autophosphorylation at pY551, while both FL-BTK-WT and FL-BTK-L528W increase HCK autophosphorylation and PLCg phosphorylation. These show that mutants partially or completely inactivate BTK and that inactive FL-BTK can activate HCK, potentially by direct BTK-HCK interactions. But they do not explain drug resistance. However, HDX and NMR show that each mutant alters the effects of BTKi binding compared to WT. In particular, T474I alters the effects of all three inhibitors around W395 and the activation loop, while L528W alters interactions around W395 with tirabrutinib and pirtobrutinib, and does not appear to bind zanubrutinib at all. The study concludes that the mutations might block drug efficacy by reducing affinity or altering binding mode.

Strengths:

The work presents convincing evidence that BTK inhibitors alter the conformation of regions distal to their binding sites, including those involved in the SH3-kinase interface, the activation loop, and a substrate binding surface between helix aF and helix aG. The findings add to the growing understanding of allosteric effects of kinase inhibitors, and their potential regulation of interactions between kinase and binding proteins.

Weaknesses:

The interpretation of HDX, NMR, and kinase assays is confusing in some places, due to ambiguity in quantifying how much kinase is bound to the inhibitor. It would be helpful to confirm binding occupancy, in order to clarify if mutants lower the amount of BTK complexed with BTKi as implied in certain places, or if they instead alter the binding mode. In addition, the interpretation of the mutant effects might benefit from a more detailed examination of how each inhibitor occupies the ATP pocket and how substitutions of T474 and L528 with Ile and Trp respectively might change the contacts with each inhibitor.

eLife assessment

In this interesting study, Drożdżyk and colleagues analyze the ability of placental CALHM orthologs to form stable complexes, identifying that CALHM2 and CALHM4 form heterooligomeric channels. The authors then determine cryo-EM structures of heterooligomeric CALHM2 and CALHM4 that reveal a distinct arrangement in which the two orthologs can interact, but preferentially segregate in the channel. This is an important study; the data provide compelling support for the interpretations and overall, the work is clearly described.

Reviewer #1 (Public Review):

The Calcium Homeostasis Modulators (CALHM) are a family of large pore channels, of which the physiological role of CALHM1 and 3 is well understood, in particular their key role in taste sensation via the release of the neurotransmitter ATP. The activation mechanism of CALHM1 involves membrane depolarization and a decrease in extracellular Ca concentration, allowing the passage of large cellular metabolites. However, the activation mechanism and physiological roles of other family members are much less well understood. Many structures of homomeric CALHM proteins have been determined, revealing distinct oligomeric assemblies despite a common transmembrane domain topology. CALHM1 and 3 have been shown functionally to form heteromeric assemblies with properties distinct from those of homomeric CALHM1. However, the structural basis of heteromeric CALHM1 and 3 remains unexplored.

In this paper, Drozdzyk et al. present an important study on the structures of heteromeric channels composed of CALHM2 and CALHM4, extending the structural understanding of the CALHM family beyond homomeric channels. The study relies primarily on cryo-EM. Despite the inherent challenges of structural determination due to the similar structural features of CALHM2 and CALHM4, the authors innovatively use synthetic nanobodies to distinguish between the subunits. Their results show a broad distribution of different heteromeric assemblies, with CALHM4 conformation similar to its homomeric form and CALHM2 conformation influenced by its proximity to CALHM4, and provide detailed insights into the interaction between CALHM2 and CALHM4.

The manuscript is well-structured and presents clear results that support the conclusions drawn. The discovery of heteromeric CALHM channels, although currently limited to an overexpressed system, represents a significant advance in the field of large-pore channels and will certainly encourage further investigation into the physiological relevance and roles of heteromeric CALHM channels. The manuscript would benefit from further insight into the functional properties of these heteromeric channels. However, this is not a weakness as the identification of precise activation stimuli for CALHM2 and 4 is beyond the scope of this work.

A challenge noted is the wide distribution of heteromeric assemblies in the 3D classification, resulting in insufficient particles for high-resolution structure determination of each assembly. The authors choose to combine particles from assemblies with 2-4 copies of CALHM4, which reveals the interface between CALHM2 and 4 but may compromise the quality of structural details. I recommend an alternative data processing strategy. First, refine particles with 2-4 CALHM4 subunits with symmetry imposed. This is followed by symmetry expansion, signal subtraction of two adjacent subunits, and subsequent classification and refinement of the subtracted particles. This approach, while not guaranteed, can potentially provide a clearer definition of CALHM2 and CALHM4 interfaces and show whether CALHM2 subunits adopt different conformations based on their proximity to CALHM4 subunits.

Reviewer #2 (Public Review):

Summary:

The authors identified that two of the placental CALHM orthologs, CALHM2 and CALHM4 can form heterooligomeric channels that are stable following detergent solubilization. By adding fiducial markers that specifically recognize either CALHM2 or CALHM4, the authors determine a cryo-EM density map of heterooligomeric CALHM2/CALHM4 from which they can determine how the channel is assembled. Surprisingly, the two orthologs segregate into two distinct segments of the channel. This segregation enables the interfacial subunits to ease the transition between the preferred conformations of each ortholog, which are similar to the confirmation that each ortholog adopts in homooligomeric channels.

Strengths:

Through the use of fiducial markers, the authors can clearly distinguish between the CALHM2 and CALHM4 promoters in the heterooligomeric channels, strengthening their assignment of most of the promoters. The authors take appropriate caution in identifying two subunits that are likely a mix of the two orthologs in the channel.

Weaknesses:

Despite the authors' efforts, no currents could be observed that corresponded to CALHM2/CALHM4 channels and thus the functional effect of their interaction is not known.

eLife assessment

Morphological characteristics and phenotypes of mutations in key developmental genes suggest that head, trunk, and tail development are regulated by discernible modules. Gdf11 signalling plays a crucial role in orchestrating the transition from trunk to tail tissues in vertebrate embryos. This important study presents convincing evidence that Tgfbr1 acts upstream of Isl1 (a pivotal effector of Gdf11 signalling) and regulates blood vessels, the lateral plate mesoderm, and the endoderm associated with the trunk-to-tail transition. Together with the previous studies, this work identifies a key signal that acts as the pivot of the trunk-to-tail transition.

Joint Public Review:

Summary:

Previously, this group showed that Tgfbr1 regulates the reorganization of the epiblast and primitive streak into the chordo-neural hinge and tailbud during the trunk-to-tail transition. Gdf11 signaling plays a crucial role in orchestrating the transition from trunk to tail tissues in vertebrate embryos, including the reallocation of axial progenitors into the tailbud and Tgfbr1 plays a key role in mediating its signaling activity. Progenitors that contribute to the extension of the neural tube and paraxial mesoderm into the tail are located in this region. In this work, the authors show that Tgfbr1 also regulates the reorganization of the posterior primitive streak/base of allantois and the endoderm as well.

By analyzing the morphological phenotypes and marker gene expression in Tgfbr1 mutant mouse embryos, they show that it regulates the merger of somatic and splanchnic layers of the lateral plate mesoderm, the posterior streak derivative. They also present evidence suggesting that Tgfbr1 acts upstream of Isl1 (key effector of Gdf11 signaling for controlling differentiation of lateral mesoderm progenitors) and regulates the remodelling of the major blood vessels, the lateral plate mesoderm and endoderm associated with the trunk-to-tail transition. Through a detailed phenotypic analysis, the authors observed that, similarly to Isl1 mutants, the lack of Tgfbr1 in mouse embryos hinders the activation of hindlimb and external genitalia maker genes and results in a failure of lateral plate mesoderm layers to converge during tail development. As a result, they interpret that ventral lateral mesoderm, which generates the peri cloacal mesenchyme and genital tuberculum, fails to specify.

They also show defects in the morphogenesis of the dorsal aorta at the trunk/tail juncture, resulting in an aberrant embryonic/extraembryonic vascular connection. Endoderm reorganization defects following abnormal morphogenesis of the gut tube in the Tgfbr1 mutants cause failure of tailgut formation and cloacal enlargement. Thus, Tgfbr1 activity regulates the morphogenesis of the trunk/tail junction and the morphogenetic switch in all germ layers required for continuing post-anal tail development. Taken together with the previous studies, this work places Gdf11/8 - Tgfbr1 signaling at the pivot of trunk-to-tail transition and the authors speculate that critical signaling through Tgfbr1 occurs in the posterior-most part of the caudal epiblast, close to the allantois.

Strengths:

The data shown is solid with excellent embryology/developmental biology. This work demonstrates meticulous execution and is presented in a comprehensive and coherent manner. Although not completely novel, the results/conclusions add to the known function of Gdf11 signaling during the trunk-to-tail transition.

Weaknesses:

The authors rely on the expression of a small number of key regulatory genes to interpret the developmental defects. The alternative possibilities remain to be ruled out thoroughly. The manuscript is also quite descriptive and would benefit from more focused highlighting of the novelty regarding the absence of Tgfbr1 in the mouse embryo. They should also strengthen some of their conclusions with more details in the results.

eLife assessment

This study presents a valuable finding and developed ME3BP-7 as a novel microencapsulated form of 3BP targeting MCT1 overexpressing PDAC cells, demonstrating its specificity and efficacy in vitro and in PDAC mouse models with significant anti-tumor effects and improved serum stability. The evidence supporting the claims of the authors is solid; however, the study calls for additional comparative in vivo data to enhance its translational significance.

Reviewer #1 (Public Review):

Summary:<br /> In the present study, Rincon-Torroella et al. developed ME3BP-7, a microencapsulated formulation of 3BP, as an agent to target MCT1 overexpressing PDACs. They provided evidence showing the specific killing of PDAC cells with MCT1 overexpressing in vitro, along with demonstrating the safety and anti-tumor efficacy of ME3BP-7 in PDAC orthotopic mouse models.

Strengths:<br /> * Developed a novel agent.<br /> * Well-designed experiments and an organized presentation of data that support the conclusions drawn.

Weaknesses:<br /> There are some minor issues that could enhance the clarity and completeness of the study:

(1) Statistical results should be visually presented in Figure 4 and Figure S1.

(2) Given the tumor heterogeneity and the identification of focal high expression of MCT1 in Figure 7 and Figure S5B, it is suggested that the authors include the results of immunohistochemical (IHC) analysis of MCT1 expression in both control and ME3BP-7 treated tumor tissues. This addition may offer insight into whether the remaining tumors are composed of PDAC cells with negative MCT1 expression, while the cells with relatively high levels of MCT1 expression were eliminated by ME3BP-7 treatment.

(3.)The authors are encouraged to discuss the future directions for improving the efficacy of this study. For example, exploring the combination of ME3BP-7 with a glutaminase-1 inhibitor (PMID 37891897) could be a valuable avenue for further research.

Reviewer #2 (Public Review):

Summary:<br /> In the manuscript by Rincon-Torroella et al, the authors evaluated the therapeutic potential of ME3BP-7, a microencapsulated formulation of 3BP which specifically targets MCT-1 high tumor cells, in pancreatic cancer models. The authors showed that, compared to 3BP, ME3BP-7 exhibited much-enhanced stability in serum. In addition, the authors confirmed the specificity of ME3BP-7 toward MCT-1 high tumor cells and demonstrated the in vivo anti-tumor effect of ME3BP-7 in orthotopic xenograft of human PDAC cell line and PDAC PDX model.

Strengths:<br /> (1) The study convincingly demonstrated the superior stability of ME3BP-7 in serum.<br /> (2) The specificity of ME3BP-7 and 3BP toward MCT-1 high PDAC cells was clearly demonstrated with CRISPR-mediated knockout experiments.

Weaknesses:<br /> The advantage of ME3BP-7 over 3BP under an in vivo situation was not fully established.

eLife assessment

This is an important study on the biochemical and biophysical analysis of a transcriptional riboswitch, detailing how Mg2+ and guanidine regulate RNA conformations. The study provides compelling evidence, developing Position-specific labeling of RNA (PLOR) and single-molecule FRET (smFRET) microscopy that are well suited to investigate the conformational dynamics of RNA structure formation. The study would have been strengthened by using a bacterial instead of phage protein to better recapitulate the physiology of the process. The work is of broad interest to those interested in RNA functional architecture and regulation.

Reviewer #1 (Public Review):

Summary:

This work presents an in-depth characterization of the factors that influence the structural dynamics of the Clostridium botulinum guanidine-IV riboswitch (riboG). Using a single-molecule FRET, the authors demonstrate that riboG undergoes ligand and Mg2+ dependent conformational changes consistent with the dynamic formation of a kissing loop (KL) in the aptamer domain. Formation of the KL is attenuated by Mg2+ and Gua+ ligand at physiological concentrations as well as the length of the RNA. Interestingly, the KL is most stable in the context of just the aptamer domain compared to longer RNAs capable of forming the terminator stem. To attenuate transcription, binding of Gua+ and formation of the KL must occur rapidly after transcription of the aptamer domain but before transcription of the rest of the terminator stem.

Strengths:

(1) Single-molecule FRET microscopy is well suited to unveil the conformational dynamics of KL formation and the authors provide a wealth of data to examine the effect of the ligand and ions on riboswitch dynamics. The addition of complementary transcriptional readthrough assays provides further support for the author's proposed model of how the riboswitch dynamics contribute to function.

(2) The single-molecule data strongly support that the effect of Gua+ ligand and Mg2+ influence the RNA structure differently for varying lengths of the RNA. The authors also demonstrate that this is specific for Mg2+ as Na+ and K+ ions have little effect.

(3) The PLOR method utilized is clever and well adapted for both dual labeling of RNAs and examining RNA at various lengths to mimic co-transcriptional folding. Using PLOR, they demonstrate that a change in the structural dynamics and ligand binding can occur after the extension of the RNA transcript by a single nucleotide. Such a tight window of regulation has intriguing implications for kinetically controlled riboswitches.

Weaknesses:

(1) The authors use only one mutant to confirm that their FRET signal indicates the formation of the KL. Importantly, this mutation does not involve the nucleotides that are part of the KL interaction. It would be more convincing if the authors used mutations in both strands of the KL and performed compensatory mutations that restore base pairing. Experiments like this would solidify the structural interpretation of the work, particularly in the context of the full-length riboG RNA or in the co-transcriptional mimic experiments, which appear to have more conformational heterogeneity.

(2) The existence of the pre-folded state (intermediate FRET ~0.5) is not well supported in their data and could be explained by an acquisition artifact. The dwell times are very short often only a single frame indicating that there could be a very fast transition (< 0.1s) from low to high FRET that averages to a FRET efficiency of 0.5. To firmly demonstrate that this intermediate FRET state is metastable and not an artifact, the authors need to perform measurements with a faster frame rate and demonstrate that the state is still present.

(3) The PLOR method employs a non-biologically relevant polymerase (T7 RNAP) to mimic transcription elongation and folding near the elongation complex. T7 RNAP has a shorter exit channel than bacterial RNAPs and therefore, folding in the exit channel may be different between different RNAPs. Additionally, the nascent RNA may interact with bacterial RNAP differently. For these reasons, it is not clear how well the dynamics observed in the T7 ECs recapitulate riboswitch folding dynamics in bacterial ECs where they would occur in nature.

Reviewer #2 (Public Review):

Summary:

Gao et al. used single-molecule FRET and step-wise transcription methods to study the conformations of the recently reported guanidine-IV class of bacterial riboswitches that upregulate transcription in the presence of elevated guanidine. Using three riboswitch lengths, the authors analyzed the distributions and transitions between different conformers in response to different Mg2+ and guanidine concentrations. These data led to a three-state kinetic model for the structural switching of this novel class of riboswitches whose structures remain unavailable. Using the PLOR method that the authors previously invented, they further examined the conformations, ligand responses, and gene-regulatory outcomes at discrete transcript lengths along the path of vectorial transcription. These analyses uncover that the riboswitch exhibits differential sensitivity to ligand-induced conformational switching at different steps of transcription, and identify a short window where the regulatory outcome is most sensitive to ligand binding.

Strengths:

Dual internal labeling of long RNA transcripts remains technically very challenging but essential for smFRET analyses of RNA conformations. The authors should be commended for achieving very high quality and purity in their labelled RNA samples. The data are extensive, robust, thorough, and meticulously controlled. The interpretations are logical and conservative. The writing is reasonably clear and the illustrations are of high quality. The findings are significant because the paradigm uncovered here for this relatively simple riboswitch class is likely also employed in numerous other kinetically regulated riboswitches. The ability to quantitatively assess RNA conformations and ligand responses at multiple discrete points along the path towards the full transcript provides a rare and powerful glimpse into co-transcriptional RNA folding, ligand-binding, and conformational switching.

Weaknesses:

The use of T7 RNA polymerase instead of a near-cognate bacterial RNA polymerase in the termination/antitermination assays is a significant caveat. It is understandable as T7 RNA polymerase is much more robust than its bacterial counterparts, which probably will not survive the extensive washes required by the PLOR method. The major conclusions should still hold, as the RNA conformations are probed by smFRET at static, halted complexes instead of on the fly. However, potential effects of the cognate RNA polymerase cannot be discerned here, including transcriptional rates, pausing, and interactions between the nascent transcript and the RNA exit channel, if any. The authors should refrain from discussing potential effects from the DNA template or the T7 RNA polymerase, as these elements are not cognate with the riboswitch under study.

Reviewer #3 (Public Review):

Summary:

In this article, Gao et. al. uses single-molecule FRET (smFRET) and position-specific labelling of RNA (PLOR) to dissect the folding and behavioral ligand sensing of the Guanidine-IV riboswitch in the presence and absence of the ligand guanidine and the cation Mg2+. The results provided valuable information on the mechanistic aspects of the riboswitch, including the confirmation of the kissing loop present in the structure as essential for folding and riboswitch activity. Co-transcriptional investigations of the system provided key information on the ligand-sensing behavior and ligand-binding window of the riboswitch. A plausible folding model of the Guanidine-IV riboswitch was proposed as a final result. The evidence presented here sheds additional light on the mode of action of transcriptional riboswitches.

Strengths:

The investigations were very thorough, providing data that supports the conclusions. The use of smFRET and PLOR to investigate RNA folding has been shown to be a valuable tool for the understanding of folding and behavior properties of these structured RNA molecules. The co-transcriptional analysis brought important information on how the riboswitch works, including the ligand-sensing and the binding window that promotes the structural switch. The fact that investigations were done with the aptamer domain, aptamer domain + terminator/anti-terminator region, and the full-length riboswitch were essential to inform how each domain contributes to the final structural state if in the presence of the ligand and Mg2+.

Weaknesses:

The system has its own flaws when compared to physiological conditions. The RNA polymerase used (the study uses T7 RNA polymerase) is different from the bacterial RNA polymerase, not only in complexity, but also in transcriptional speed, which can directly interfere with folding and ligand-sensing. Additionally, rNTPs concentrations were much lower than physiological concentrations during transcription, likely causing a change in the polymerase transcriptional speed. These important aspects and how they could interfere with results are important to be addressed to the broad audience. Another point of consideration to be aware of is that the bulky fluorophores attached to the nucleotides can interfere with folding to some extent.

eLife assessment

The Hippo signaling pathway plays a crucial role in controlling organ size, cell proliferation, and apoptosis, though its role in endocrine pancreas development has remained unclear. In this useful work, the authors study the function of the Tead1 transcription factor, a Hippo effector, specifically in pancreatic beta cells. They provide solid evidence, using multiple different conditional knockout models to reveal Tead1's regulatory functions in insulin secretion and beta cell proliferation. However, deeper exploration of their data and incorporating findings from existing literature on this topic would provide a clearer understanding of Tead1's role in β-cell function, within or beyond the Hippo pathway.

Reviewer #1 (Public Review):

Summary:

Hippo pathway activity is required for pancreas morphogenesis, but its role in endocrine pancreas function remains elusive. The author aims to study the function of the TEAD1 gene in b-cells.

Strengths:

The authors generated TEAD1 conditional knockout animals by crossing the TEAD1f/f mice with three Cre strains (RIP-Cre, Ins1-Cre, and MIP-CreERT). In all of them, the KO animals showed progressive loss of insulin secretion with normal beta cell mass. Further characterization of the animals indicated glucose-induced insulin secretion defect and increased beta cell proliferation rate. RNA-Seq and ChIP-Seq experiments identified Pdx1, MafA, and Glut2, etc. as direct targets of TEAD1, which might be responsible for the insulin secretion defect in the animals. Of interest, the authors also uncovered the cell cycle-related gene p16 as a direct target of TEAD1. Reduction of p16 is likely to drive the beta cell proliferation in the TEAD1 knockout model. Thus, they proposed that TEAD1 is a regulator of the proliferative quiescence process in beta cells. Overall, the evidence provided by the authors is highly relevant and supports their conclusion.

Weaknesses:

(1) The authors don't explicitly mention that some results appeared in a previous publication (https://doi.org/10.1093/nar/gkac1063) from them.

(2) The authors begin their story by introducing TEAD1 as part of the Hippo pathway. They showed Taz expression data in Figure 1. Did they do any experiments to detect Taz in their TEAD1 model? Did the authors detect any expression changes in CTGF following TEAD1 knockout? I could not see this changed. The phenotype characterization data presented here contrasts with what has been shown in TAZ b-cell knockout mice (https://doi.org/10.1101/2022.05.31.494216). Based on the data presented here, Hippo is not involved, which should at least be discussed in length.

(3) Figure 1B - TAZ staining looks different in the three-month age group.

(4) TEAD ChIP-seq data doesn't look very convincing to me. It's hard to tell whether those highlighted regions in Figures 3A and 5J were signals or background noise. Although the authors also performed ChIP-qPCR in MIN6, it's unclear whether these binding events occur in vivo. The analysis of ChIP-seq dataset is limited as well. How many peaks called? What proportion of differentially expressed genes are bound by TEAD1? Was TEAD1 also detectable at NGN3 and NEUROD1 gene regions? If acquiring enough cells is not possible, the authors could try CUT&RUN or CUT&Tag to improve the data quality.

(5) The authors should perform RNA-seq or gene expression studies in MIP-CreERT to confirm, which could help narrow down the actual targets of TEAD1 as well.

(6) Figure 6 - the experiment lacks a control: Ezh2 beta cell KO. In addition to p16, Ezh2, and PRC2 have other targets in beta-cells, the authors could not rule out the contribution of those to the phenotype, so the implication of this experiment is vague.

Reviewer #2 (Public Review):

In this manuscript, Lee et al. assessed the role of Tead1 in mouse beta cells using three Cre-driver lines: Rip-Cre, Ins-Cre, and Mip-CreERT. The authors demonstrate that loss of TEAD1 during development and in mature beta cells leads to increased cell-autonomous beta cell proliferation and reduced insulin secretion. The phenotype of Tead1 knockout is not surprising, given that it is a key player in the Hippo pathway - a well-characterized pathway controlling cell proliferation. However, as the authors suggested, the phenotype observed in Tead1 might be through other non-Hippo pathway factors as well. The authors further convincingly established PDX1 and p16 as the target of Tead1 in controlling beta cell function and proliferation correspondingly. I have the following specific comments:

(1) As the authors mentioned, there are concerns over the usage of some Cre transgenic lines. Another useful control would be the naive Cre line that is not bred to floxed mutant, in addition to the floxed mice used by the authors in the manuscript here.

(2) The logic to rely on the deletion of Ezh2 to restore p16 in the Tead1 knockout mice is unclear. Ezh2 has so many more targets than p16. Why not a direct rescue experiment by overexpression of p16?

(3) The observed correlation of PDX1 and TEAD1 in expression in human islets is intriguing. But does this correlation translate to beta cell proliferation and function? Does TEAD1 knockout in human islets elicit a similar proliferation versus function response?

(4) The argument of Tead1 only controls maturation but not differentiation and that maturation function versus proliferation phenotype is independently controlled is weak. It appears that this conclusion is only based on that "many disallowed genes...were not altered in Tead1-deficient islets". Perhaps the authors can perform a formal comparison between the transcriptomic changes of Tead1 knockout and Myc overexpressing/Notch gain of function beta cells and show that these two processes are different. In addition, what are the signatures of genes that are upregulated in Tead1 knockout compared with controls?

Reviewer #1 (Public Review):

Summary:

Ewing sarcoma is an aggressive pediatric cancer driven by the EWS-FLI oncogene. Ewing sarcoma cells are addicted to this chimeric transcription factor, which represents a strong therapeutic vulnerability. Unfortunately, targeting EWS-FLI has proven to be very difficult, and a better understanding of how this chimeric transcription factor works is critical to achieving this goal. Towards this perspective, the group had previously identified a DBD-𝛼4 helix (DBD) in FLI that appears to be necessary to mediate EWS-FLI transcriptomic activity. Here, the authors used multi-omic approaches, including CUT&tag, RNAseq, and MicroC to investigate the impact of this DBD domain. Importantly, these experiments were performed in the A673 Ewing sarcoma model where endogenous EWS-FLI was silenced, and EWS-FLI-DBD proficient or deficient isoforms were re-expressed (isogenic context). They found that the DBD domain is key to mediating EWS-FLI cis activity (at msat) and to generating the formation of specific TADs. Furthermore, cells expressing DBD-deficient EWS-FLI display very poor colony-forming capacity, highlighting that targeting this domain may lead to therapeutic perspectives.

Strengths:

The group has strong expertise in Ewing sarcoma genetics and epigenetics and also in using and analyzing this model (Theisen et al., 2019; Boone et al., 2021; Showpnil et al., 2022).

They aim at better understanding how EWS-FLI mediated its oncogenic activity, which is critical to eventually identifying novel therapies against this aggressive cancer.

They use the most recent state-of-the-art omics methods to investigate transcriptome, epigenetics, and genome conformation methods. In particular, Micro-C enables achieving up to 1kb resolved 3D chromatin structures, making it possible to investigate a large number of TADs and sub-TADs structures where EWS-FLI1 mediates its oncogenic activity.

They performed all their experiments in an Ewing sarcoma genetic background (A673 cells) which circumvents bias from previously reported approaches when working in non-orthologous cell models using similar approaches.

Weaknesses:

The main weakness comes from the poor reproducibility of Micro-C data. Indeed, it appears that the distances/clustering observed between replicates are typically similar or even larger than between biological conditions. For instance, in Figure 1B, I do not see any clustering when considering DBD1, DBD2, DBD+1, DBD+2.

Lanes 80-83: "KD replicates clustered together with DBD replicate 1 on both axes and with DBD replicate 2 on the y-axis. DBD+ replicates, on the other hand, clustered away from both KD and DBD replicates. These observations suggest that the global chromatin structure of DBD replicates is more similar to KD than DBD+ replicates."

When replacing DBD replicate 1 with DBD replicate 2, their statement would not be true anymore.

Additional replicates to clarify this aspect seem absolutely necessary since those data are paving the way for the entire manuscript.

Similarly:<br /> - In Figure 1C, how would the result look when comparing DBD2/KD2/DBD+2? Same when comparing DBD 1 with KD1 and DBD+1. Would the difference go in the same direction?<br /> - Figure 1D-E. How would these plots look like when comparing each replicate to each other's? How much difference would be observed when comparing, for instance, DBD1/DBD2 ? or DBD1/DBD+1?<br /> - Figure 2: again, how would these analyses look like when performing the analysis with only DBD1/DBD+1/KD1 or DBD2/DBD+2/KD?

Another major question is the stability of EWS-FLI DBD vs EWS-FLI DBD+ proteins. Indeed, it seems that they have more FLAG (i.e., EWS-FLI) peaks in the DBD+ condition compared to the DBD condition (Figure 2B). In the WB, FLAG intensities seem also higher (2/3 replicates) in DBD+ condition compared to the DBD condition (Figure S1B).

Would it be possible that DBD+ is just more expressed or more stable than DBD? The higher stability of the re-expressed DBD+ could also partially explain their results independently of the 3D conformational change. In other words, can they exclude that DBD+ and DBD binding are not related to their respective protein stability or their global re-expression levels?

Surprisingly, WB FLI bands in DBD+ conditions are systematically (3/3 replicates) fainter than in DBD conditions (Figure S1B). How do the authors explain these opposite results between FLI and FALG in the WB?

Reviewer #2 (Public Review):

Summary:

The manuscript by Bayanjargal et al. entitled "The DBD-alpha4 helix of EWS::FLI is required for GGAA microsatellite binding that underlies genome regulation in Ewing sarcoma" reports on the critical role of a small alpha helix in the DNA binding domain (DBD) of the FLI1 portion of EWS::FLI1 that is critical for binding to repetitive stretches of GGAA-motifs, i.e. GGAA microsatellites, which serve as potent neoenhancers in Ewing sarcoma.

Strengths:

The paper is generally well-written, and easy to follow and the data presented are of high quality, well-described and underpin the conclusions of the authors. The report sheds new light on how EWS::FLI1 mechanistically binds to and activates GGAA microsatellite enhancers, which is of importance to the field.

Weaknesses:

While there are no major weaknesses in this paper, there are a few minor issues that the authors may wish to address:

(1) While the official protein symbol for the gene EWSR1 is indeed EWS, the protein symbol for the gene FLI1 is identical, i.e. FLI1. The authors nominate the fusion oncoprotein EWS::FLI1 (even in the title) but it appears more adequate to use EWS::FLI1.

(2) The used cell lines should be spelled according to their official nomenclature (e.g. A-673 instead of A673).

(3) It appears as if the vast majority of results were generated in a single Ewing sarcoma cell line (A-673) which is an atypical Ewing sarcoma cell line harboring an activating BRAF mutation and may be genomically quite unstable as compared to other Ewing sarcoma cell lines (Kasan et al. 2023 preprint at bioRxiv https://www.biorxiv.org/content/10.1101/2023.11.20.567802v1). Hence, it may be supportive for the paper to recapitulate/cross-validate a few key results in other Ewing sarcoma cell lines, e.g. by using EWS::ERG-positive cell lines. Perhaps the authors could make use of available published data.

(4) Figure 6 and Supplementary Figure 5 are very interesting but focus on two selected target genes of the fusion (FCGRT and CCND1). It would be interesting to see whether these findings also extend to common EWS::ETS transcriptional signatures that have been reported. The authors could explore their data and map established consensus EWS::ETS signatures to investigate which other hubs might be affected at relevant target genes.

(5) Table 1 is a bit hard to read. In my opinion, it is not necessary to display P-values with up to 8 decimal positions. The gene symbols should be displayed in italic font.

eLife assessment

The main finding in this paper is that EGFR can be a novel substrate of the membrane ZNRF3/RNF43 E3 ligases. This is significant as the prevailing understanding posits that the Wnt receptors Frizzled and LRP5/6 exclusively served as substrates for these ligases. Given the frequent occurrence of mutations in ZNRF3/RNF43 or compromised expression levels in human cancers, the new evidence that aberrant EGFR expression and signaling may also contribute to the tumorigenic effects of ZNRF3/RNF43 mutations in cancer is important. The conclusions of the manuscript are supported by solid data, but some aspects of the mechanism presented need to be reinforced to fully support the claims made by the authors.

Reviewer #1 (Public Review):

Summary:

In this manuscript, the authors provide strong evidence that the cell surface E3 ubiquitin ligases RNF43 and ZNRF3, which are well known for their role in regulating cell surface levels of WNT receptors encoded by FZD genes, also target EGFR for degradation. This is a newly identified function for these ubiquitin ligases beyond their role in regulating WNT signaling. Loss of RNF43/ZNRF3 expression leads to elevated EGFR levels and signaling, suggesting a potential new axis to drive tumorigenesis, whereas overexpression of RNF43 or ZNRF3 decreases EGFR levels and signaling. Furthermore, RNF43 and ZNRF3 directly interact with EGFR through their extracellular domains.

Strengths:

The data showing that RNF43 and ZNRF3 interact with EGFR and regulate its levels and activity are thorough and convincing, and the conclusions are largely supported.

Weaknesses:

While the data support that EGFR is a target for RNF43/ZNRF3, some of the authors' interpretations of the data on EGFR's role relative to WNT's roles downstream of RNF43/ZNRF3 are overstated. The authors, perhaps not intentionally, promote the effect of RNF43/ZNRF3 on EGFR while minimizing their role in WNT signaling. This is the case in most of the biological assays (cell and organoid growth and mouse tumor models). For example, the conclusion of "no substantial activation of Wnt signaling" (page 14) in the prostate cancer model is currently not supported by the data and requires further examination. In fact, examination of the data presented here indicates effects on WNT/b-catenin signaling, consistent with previous studies.<br /> Cancers in which RNF43 or ZNRF3 are deleted are often considered to be "WNT addicted", and inhibition of WNT signaling generally potently inhibits tumor growth. In particular, treatment of WNT-addicted tumors with Porcupine inhibitors leads to tumor regression. The authors should test to what extent PORCN inhibition affects tumor (and APC-min intestinal organoid) growth. If the biological effects of RNF43/ZNRF3 loss are mediated primarily or predominantly through EGFR, then PORCN inhibition should not affect tumor or organoid growth.

Reviewer #2 (Public Review):

Using proteogenomic analysis of human cancer datasets, Yu et al, found that EGFR protein levels negatively correlate with ZNFR3/RNF43 expression across multiple cancers. Interestingly, they found that CRC harbouring the frequent RNF43 G659Vfs*41 mutation exhibits higher levels of EGFR when compared to RNF43 wild-type tumors. This is highly interesting since this mutation is generally not thought to influence Frizzled levels and Wnt-bcatenin pathway activity. Using CRISPR knockouts and overexpression experiments, the authors show that EGFR levels are modulated by ZNRF3/RNF43. Supporting these findings, modulation of ZNRF3/RNF43 activity using Rspondin also leads to increased EGFR levels. Mechanistically, the authors, show that ZNRF3/RNF43 ubiquitinate EGFR and leads to degradation. Finally, the authors present functional evidence that loss of ZNRF3/RNF43 unleashes EGFR-mediated cell growth in 2D culture and organoids and promotes tumor growth in vivo.

Overall, the conclusions of the manuscript are well supported by the data presented, but some aspects of the mechanism presented need to be reinforced to fully support the claims made by the authors. Additionally, the title of the paper suggests that ZNRF3 and RNF43 loss leads to the hyperactivity of EGFR and that its signalling activity contributes to cancer initiation/progression. I don't think the authors convincingly showed this in their study.

Major points:

(1) EGFR ubiquitination. All of the experiments supporting that ZNFR3/RNF43 mediates EGFR ubiquitination are performed under overexpression conditions. A major caveat is also that none of the ubiquitination experiments are performed under denaturing conditions. Therefore, it is impossible to claim that the ubiquitin immunoreactivity observed on the western blots presented in Figure 4 corresponds to ubiquitinated-EGFR species.

Another issue is that in Figure 4A, the experiments suggest that the RNF43-dependent ubiquitination of EGFR is promoted by EGF. However, there is no control showing the ubiquitination of EGFR in the absence of EGF but under RNF43 overexpression. According to the other experiments presented in Figures 4B, 4C, and 4F, there seems to be a constitutive ubiquitination of EGFR upon overexpression. How do the authors reconcile the role of ZNRF3/RNF43 vs c-cbl ?

(2) EGFR degradation vs internalization. In Figure 3C, the authors show experiments that demonstrate that RNF43 KO increases steady-state levels of EGFR and prevents its EGF-dependent proteolysis. Using flow cytometry they then present evidence that the reduction in cell surface levels of EGFR mediated by EGF is inhibited in the absence of RNF43. The authors conclude that this is due to inhibition of EGF-induced internalization of surface EGF. However, the experiments are not designed to study internalization and rather merely examine steady-state levels of surface EGFR pre and post-treatment. These changes are an integration of many things (retrograde and anterograde transport mechanisms presumable modulated by EGF). What process(es) is/are specifically affected by ZNFR3/RNF43 ? Are these processes differently regulated by c-cbl ? If the authors are specifically interested in internalization/recycling, the use of cell surface biotinylation experiments and time courses are needed to examine the effect of EGF in the presence or absence of the E3 ligases.

(3) RNF43 G659fs*41. The authors make a point in Figure 1D that this mutant leads to elevated EGFR in cancers but do not present evidence that this mutant is ineffective in mediated ubiquitination and degradation of EGFR. As this mutant maintains its ability to promote Frizzled ubiquitination and degradation, it would be important to show side by side that it does not affect EGFR. This would perhaps imply differential mechanisms for these two substrates.

(4) "Unleashing EGFR activity". The title of the paper implies that ZNRF3/RNF43 loss leads to increased EGFR expression and hence increased activity that underlies cancer. However, I could find only one direct evidence showing that increased proliferation of the HT29 cell line mutant for RNF43 could be inhibited by the EGFR inhibitor Erlotinib. All the other evidence presented that I could find is correlative or indirect (e.g. RPPA showing increased phosphorylation of pathway members upon RNF43 KO, increased proliferation of a cell line upon ZNRF3/ RNF43 KO, decreased proliferation of a cell line upon ZNRF3/RNF43 OE in vitro or in xeno...). Importantly, the authors claim that cancer initiation/ progression in ZNRF3/RNF43 mutants may in some contexts be independent of their regulation of Wnt-bcatenin signaling and relying on EGFR activity upregulation. However, this has not been tested directly. Could the authors leverage their znrf3/RNF43 prostate cancer model to test whether EGFR inhibition could lead to reduced cancer burden whereas a Frizzled or Wnt inhibitor does not?

More broadly, if EGFR signaling were to be unleashed in cancer, then one prediction would be that these cells would be more sensitive to EGFR pathway inhibition. Could the authors provide evidence that this is the case? Perhaps using isogenic cell lines or a panel of patient-derived organoids (with known genotypes).

eLife assessment

This compelling study reports the to-date most comprehensive neurotransmitter atlas of any organism, using fluorescent knock-in reporter lines. It represents an extremely useful tool for a broad audience of scientists interested in neuronal cell type differentiation and function.

Reviewer #1 (Public Review):

Wang and colleagues conducted a study to determine the neurotransmitter identity of all neurons in C. elegans hermaphrodites and males. They used CRISPR technology to introduce fluorescent gene expression reporters into the genomic loci of NT pathway genes. This approach is expected to better reflect in vivo gene expression compared to other methods like promoter- or fosmid-based transgenes, or available scRNA datasets. The study presents several noteworthy findings, including sexual dimorphisms, patterns of NT co-transmission, neuronal classes that likely use NTs without direct synthesis, and potential identification of unconventional NTs (e.g. betaine releasing neurons). The data is well-described and critically discussed, including a comparison with alternative methods. Although many of the observations and proposals have been previously discussed by the Hobert lab, the current study is particularly valuable due to its comprehensiveness. This NT atlas is the most complete and comprehensive of any nervous system that I am aware of, making it an extremely useful tool for the community.

Reviewer #2 (Public Review):

Summary:

Together with the known anatomical connectivity of C. elegans, a neurotransmitter atlas paves the way toward a functional connectivity map. This study refines the expression patterns of key genes for neurotransmission by analyzing the expression patterns from CRISPR-knocked-in GFP reporter strains using the color-coded Neuropal strain to identify neurons. Along with data from previous scRNA sequencing and other reporter strains, examining these expression patterns enhances our understanding of neurotransmitter identity for each neuron in hermaphrodites and the male nervous system. Beyond the known neurotransmitters (GABA, Acetylcholine, Glutamate, dopamine, serotonin, tyramine, octopamine), the atlas also identifies neurons likely using betaine and suggests sets of neurons employing new unknown monoaminergic transmission, or using exclusively peptidergic transmission.

Strengths:

The use of CRISPR reporter alleles and of the Neuropal strain to assign neurotransmitter usage to each neuron is much more rigorous than previous analysis and reveals intriguing differences between scRNA seq, fosmid reporter, and CRISPR knock-in approaches. Among other mechanisms, these differences between approaches could be attributed to 3'UTR regulatory mechanisms for scRNA vs. knockin or titration of rate-limited negative regulatory mechanisms for fosmid vs. knockin. It would be interesting to discuss this and highlight the occurrences of these potential phenomena for future studies.

Weaknesses:

For GABAergic transmission, one shortcoming arises from the lack of improved expression pattern by a knockin reporter strain for the GABA recapture symporter snf-11. In its absence, it is difficult to make a final conclusion on GABA recapture vs GABA clearance for all neurons expressing the vesicular GABA transporter neurons (unc-47+) but not expressing the GAD/UNC-25 gene e.g. SIA or R2A neurons. At minima, a comparison of the scRNA seq predictions versus the snf-11 fosmid reporter strain expression pattern would help to better judge the proposed role of each neuron in GABA clearance or recycling.

Considering the complexities of different tagging approaches, like T2A-GFP and SL2-GFP cassettes, in capturing post-translational and 3'UTR regulation is important. The current formulation is simplistic. e.g. after SL2 trans-splicing the GFP RNA lacks the 5' regulatory elements, T2A-GFP self-cleavage has its own issues, and the his-44-GFP reporter protein does certainly have a different post-translational life than vesicular transporters or cytoplasmic enzymes.

Do all splicing variants of neurotransmitter-related genes translate into functional proteins? The possibility that some neurons express a non-functional splice variant, leading to his-74-GFP reporter expression without functional neurotransmitter-related protein production is not addressed. Also, one tagged splice variant of unc-25 is expected to fail to produce a GFP reporter, can this cause trouble?

Reviewer #3 (Public Review):

Summary:

In this paper, Wang et al. provide the most comprehensive description and comparison of the expression of the different genes required to synthesize, transport, and recycle the most common neurotransmitters (Glutamate, Acetylcholine, GABA, Serotonin, Dopamine, Octopamine, and Tyramine) used by hermaphrodite and male C. elegans. This paper will be a seminal reference in the field. Building and contrasting observations from previous studies using fosmid, multicopy reporters, and single-cell sequencing, they now describe CRISPR/Cas-9-engineered reporter strains that, in combination with the multicolor pan-neuronal labeling of all C. elegans neurons (NeuroPAL), allows rigorous elucidation of neurotransmitter expression patterns. These novel reporters also illuminate previously unappreciated aspects of neurotransmitter biology in C. elegans, including sexual dimorphism of expression patterns, co-transmission, and the elucidation of cell-specific pathways that might represent new forms of neurotransmission.

Strengths:

The authors set out to establish neurotransmitter identities in C. elegans males and hermaphrodites via varying techniques, including integration of previous studies, examination of expression patterns, and generation of endogenous CRISPR-labeled alleles. Their study is comprehensive, detailed, and rigorous, and achieves the aims. It is an excellent reference for the field, particularly those interested in biosynthetic pathways of neurotransmission and their distribution in vivo, in neuronal and non-neuronal cells.

Weaknesses:

No weaknesses were noted. The authors do a great job linking their characterizations with other studies and techniques, giving credence to their findings. As the authors note, there are sexually dimorphic differences across animals and varying expression patterns of enzymes. While it is unlikely there will be huge differences in the reported patterns across individual animals, it is possible that these expression patterns could vary developmentally, or based on physiological or environmental conditions. It is unclear from the study how many animals were imaged for each condition, and if the authors noted changes across individuals during development (could be further acknowledged in the discussion?)

eLife assessment

Dong et al. investigate the role of the small Ras-like GTPase Rab10 in the exocytosis of DCVs in mouse hippocampal neurons, showing that Rab10 depletion hinders DCV exocytosis independently of its effects on neurite outgrowth. Their findings are convincing and provide evidence that Rab10 depletion leads to altered ER morphology, impaired ER-based calcium buffering, and decreased ribosomal protein expression, which collectively contributes to defective DCV secretion. The study comes to the important conclusion that Rab10 is critical for DCV release by ensuring ER calcium homeostasis.

Reviewer #1 (Public Review):

Summary:

Dong et al here have studied the impact of the small Ras-like GTPase Rab10 on the exocytosis of dense core vesicles (DVC), which are important mediators of neuropeptide signaling in the brain. They use optical imaging to show that lentiviral depletion of Rab10 in mouse hippocampal neurons in culture independent of the established defects in neurite outgrowth hamper DCV exocytosis. They further demonstrate that such defects are paralleled by changes in ER morphology and defective ER-based calcium buffering as well as reduced ribosomal protein expression in Rab10-depleted neurons. Re-expression of Rab10 or supplementation of exogenous L-leucine to restore defective neuronal protein synthesis rescues impaired DCV secretion. Based on these results they propose that Rab10 regulates DCV release by maintaining ER calcium homeostasis and neuronal protein synthesis.

Strengths:

This work provides interesting and potentially important new insights into the connection between ER function and the regulated secretion of neuropeptides via DCVs. The authors combine advanced optical imaging with light and electron microscopy, biochemistry, and proteomics approaches to thoroughly assess the effects of Rab10 knockdown at the cellular level in primary neurons. The proteomic dataset provided may be valuable in facilitating future studies regarding Rab10 function. This work will thus be of interest to neuroscientists and cell biologists.

Weaknesses:

While the main conclusions of this study are comparably well supported by the data, I see three major weaknesses:

(1) For some of the data the statistical basis for analysis remains unclear. I.e. is the statistical assessment based on N= number of experiments or n = number of synapses, images, fields of view etc.? As the latter cannot be considered independent biological replicates, they should not form the basis of statistical testing.

(2) As it stands the paper reports on three partially independent phenotypic observations, the causal interrelationship of which remains unclear. Based on prior studies (e.g. Mercan et al 2013 Mol Cell Biol; Graves et al JBC 1997) it is conceivable that defective ER-based calcium signaling and the observed reduction in protein synthesis are causally related. For example, ER calcium release is known to promote pS6K1 phosphorylation, a major upstream regulator of protein synthesis and ribosome biogenesis. Conversely, L-leucine supplementation is known to trigger calcium release from ER stores via IP3Rs. Given the reported impact of Rab10 on axonal transport of autophagosomes and, possibly, lysosomes via JIP3/4 or other mediators (see e.g. Cason and Holzbaur JCB 2023) and the fact that mTORC1, the alleged target of leucine supplementation, is located on lysosomes, which in turn form membrane contacts with the ER, it seems worth analyzing whether the various phenotypes observed are linked at the level of mTORC1 signaling.

(3) The claimed lack of effect of Rab10 depletion on SV exocytosis is solely based on very strong train stimulation with 200 Aps, a condition not very well suited to analyze defects in SV fusion. The conclusion that Rab10 loss does not impact SV fusion thus seems premature.

Reviewer #2 (Public Review):

Summary:

In this paper, the authors assess the function of Rab10 in dense core vesicle (DCV) exocytosis using RNAi and cultured neurons. The author provides evidence that their knockdown (KD) is effective and provides evidence that DCV is compromised. They also perform proteomic analysis to identify potential pathways that are affected upon KD of Rab10 that may be involved in DCV release. Upon focusing on ER morphology and protein synthesis, the authors conclude that defects in protein synthesis and ER Ca2+ homeostasis contributes to the DVC release defect upon Rab10 KD. The authors claim that Rab10 is not involved in synaptic vesicle (SV) release and membrane homeostasis in mature neurons.

Strengths:

The data related to Rab10's role in DCV release seems to be strong and carried out with rigor. While the paper lacks in vivo evidence that this gene is indeed involved in DCV in a living mammalian organism, I feel the cellular studies have value. The identification of ER defect in Rab10 manipulation is not truly novel but it is a good conformation of studies performed in other systems. The finding that DCV release defect and protein synthesis defect seen upon Rab10 KD can be significantly suppressed by Leucine supplementation is also a strength of this work.

Weaknesses:

The data showing Rab10 is NOT involved in SV exocytosis seems a bit weak to me. Since the proteomic analysis revealed so many proteins that are involved in SV exo/encodytosis to be affected upon Rab10, it is a bit strange that they didn't see an obvious defect. Perhaps this could have been because of the protocol that the authors used to trigger SV release (I am not an E-phys expert but perhaps this could have been a 'sledge-hammer' manipulation that may mask any subtle defects)? Perhaps the authors can claim that DCV is more sensitive to Rab10 KD than SV, but I am not sure whether the authors should make a strong claim about Rab10 not being important for SV exocytosis.

Also, the authors mention "Rab10 does not regulate membrane homeostasis in mature neurons" but I feel this is an overstatement. Since the authors only performed KD experiments, not knock-out (KO) experiments, I believe they should not make any conclusion about it not being required, especially since there is some level of Rab10 present in their cells. If they want to make these claims, I believe the authors will need to perform conditional KO experiments, which are not performed in this study.

Finally, the authors show that protein synthesis and ER Ca2+ defects seem to contribute to the defect but they do not discuss the relationship between the two defects. If the authors treat the Rab10 KD cells with both ionomycin and Leucine, do they get a full rescue? Or is one defect upstream of the other (e.g. can they see rescue of ER morphology upon Leucine treatment)? While this is not critical for the conclusions of the paper, several additional experiments could be performed to clarify their model, especially considering there is no clear model that explains how Rab10, protein synthesis, ER homeostasis, and Ca2+ are related to DCV (but not SV) exocytosis.

Reviewer #3 (Public Review):

In the submitted manuscript, Dong and colleagues set out to dissect the role of the Rab10 small GTPase on the intracellular trafficking and exocytosis of dense core vesicles (DCVs). While the authors have already shown that Rab3 plays a central role in the exocytosis of DVC in mammalian neurons, the roles of several other Rab-members have been identified genetically, but their precise mechanism of action in mammalian neurons remains unclear. In this study, the authors use a carefully designed and thoroughly executed series of experiments, including live-cell imaging, functional calcium-imaging, proteomics, and electron microscopy, to identify that DCV secretion upon Rab10 depletion in adult neurons is primarily a result of dysregulated protein synthesis and, to a lesser extent, disrupted intracellular calcium buffering. Given that the full deletion of Rab10 has a deleterious effect on neurons and that Rab10 has a major role in axonal development, the authors cautiously employed the knock-down strategy from 7 DIV, to focus on the functional impact of Rab10 in mature neurons. The experiments in this study were meticulously conducted, incorporating essential controls and thoughtful considerations, ensuring rigorous and comprehensive results.

eLife assessment

Sisigano et al. report findings about the role of sphingolipids using lipidomics with machine learning in paclitaxel-induced peripheral neuropathy and preliminary translation of the impact of SA1P in cultured neuronal cells. This study presents a valuable finding on the increased activity of two well-studied signal transduction pathways in a subtype of breast cancer. The strength of evidence is incomplete with some support for the main claims with some limitations.

Reviewer #1 (Public Review):

Summary:

This study examines lipid profiles in cancer patients treated with the neurotoxic chemotherapy paclitaxel. Multiple methods, including machine learning as well as more conventional statistical modelling, were used to classify lipid patterns before and after paclitaxel treatment and in conjunction with neuropathy status. Lipid profiles before and after paclitaxel therapy were analysed from 31 patients. The study aimed to characterize from the lipid profile if plasma samples were collected pre-paclitaxel or post-paclitaxel and their relevance to neuropathy status. Sphingolipids including sphinganine-1-phosphate (SA1P) differed between patients with and without neuropathy. To examine the potential role of SA1P, it was applied to murine primary sensory neuron cultures, and produced calcium transients in a proportion of neurons. This response was abolished by the application of a TRPV1 antagonist. The number of neurons responding to SA1P was partially reduced by the sphingosine 1-phosphate receptor (S1PR1) modulator fingolimod.

Strengths:

The strengths of this study include the use of multiple methods to classify lipid patterns and the attempt to validate findings from the clinical cohort in a preclinical model using primary sensory neurons.

Weaknesses:

There are a number of weaknesses in the study. The small sample size is a significant limitation of the study. Out of 31 patients, only 17 patients were reported to develop neuropathy, with significant neuropathy (grade 2/3) in only 5 patients. The authors acknowledge this limitation in the results and discussion sections of the manuscript, but it limits the interpretation of the results. Also acknowledged is the limited method used to assess neuropathy.

Potentially due to this small number of patients with neuropathy, the machine learning algorithms could not distinguish between samples with and without neuropathy. Only selected univariate analyses identified differences in lipid profiles potentially related to neuropathy.

Three sphingolipid mediators including SA1P differed between patients with and without neuropathy at the end of treatment. These sphingolipids were elevated at the end of treatment in the cohort with neuropathy, relative to those without neuropathy. However, across all samples from pre to post-paclitaxel treatment, there was a significant reduction in SA1P levels. It is unclear from the data presented what the underlying mechanism for this result would be. If elevated SA1P is associated with neuropathy development, it would be expected to increase in those who develop neuropathy from pre to post-treatment time points.

Primary sensory neuron cultures were used to examine the effects of SA1P application. SA1P application produced calcium transients in a small proportion of sensory neurons. It is not clear how this experimental model assists in validating the role of SA1P in neuropathy development as there is no assessment of sensory neuron damage or other hallmarks of peripheral neuropathy. These results demonstrate that some sensory neurons respond to SA1P and that this activity is linked to TRPV1 receptors. However, further studies will be required to determine if this is mechanistically related to neuropathy.

Impact:

Taken in total, the data presented do not provide sufficient evidence to support the contention that SA1P has an important role in paclitaxel-induced peripheral neuropathy. Further, the results do not provide evidence to support the use of S1PR1 receptor antagonists as a therapeutic strategy. It is important to be careful with language use in the discussion, as the significance of the present results is overstated.

However, based on the results of previous studies, it is likely that sphingolipid metabolism plays a role in chemotherapy-induced peripheral neuropathy. Based on this existing evidence, the S1PR1 receptor antagonist fingolimod has already been examined in experimental models and clinical trials. Further work is needed to examine the links between lipid mediators and neuropathy development and identify additional strategies for intervention.

Reviewer #2 (Public Review):

Summary:

The study investigates the mechanisms underlying chemotherapy-induced peripheral neuropathy (CIPN), a notable side effect of commonly used anticancer drugs like paclitaxel. It aims to comprehend the putative mechanisms through lipidomics analysis of plasma samples from cancer patients pre and post-paclitaxel treatment, drawing inspiration from preclinical studies highlighting the role of sphingolipids. While the use of patient plasma samples stands out as a major strength, shortcomings in the result presentation undermine the study's significance. The introduction lacks a robust rationale, failing to articulate the utility of machine learning methods over conventional lipidomics analysis and the relevance of broader neuropathy in the context of the study's goal of investigating peripheral neuropathy. The failure to robustly link neuropathy to paclitaxel treatment, with only around 50% of patients developing neuropathy, mostly at Grade 1, with no or mild symptoms that require no intervention, weakens the study's impact. The presentation of results lacks clarity on sphingolipid dysregulation, leaving uncertainty regarding downregulation or upregulation. Furthermore, no clarity in validation for the machine learning-based analysis with conventional methods and an overall weakness in result representation weaken the study, despite addressing an important question in the field.

Strengths:

The study leverages patient plasma samples before and after paclitaxel treatment, enhancing the translatability of findings to patient impact. The attempt to employ machine learning (ML) methods for analyzing biological samples and classifying patient groups is commendable, pushing the biomedical sciences towards ML applications for handling complex data. The chosen topic of investigating chemotherapy-induced peripheral neuropathy (CIPN) is clinically important, offering potential benefits for cancer patients undergoing chemotherapy treatment.

Weaknesses:

The article is poorly written, hindering a clear understanding of core results. While the study's goals are apparent, the interpretation of sphingolipids, particularly SA1P, as key mediators of paclitaxel-induced neuropathy lacks robust evidence. The introduction fails to establish the significance of general neuropathy or peripheral neuropathy in anticancer drug-treated patients, and crucial details, such as the percentage of patients developing general neuropathy or peripheral neuropathy, are omitted. This omission is particularly relevant given that only around 50% of patients developed neuropathy in this study, primarily of mild Grade 1 severity with negligible symptoms, contradicting the study's assertion of CIPN as a significant side effect. The lack of clarity in distinguishing results obtained by lipidomics using machine learning methods and conventional methods adds to the confusion. The poorly written results section fails to specify SA1P's downregulation or upregulation, and the process of narrowing down to sphingolipids and SA1P is inadequately explained. Integrating a significant portion of the discussion section into the results section could enhance clarity. An explanation of the utility of machine learning in classifying patient groups over conventional methods and the citation of original research articles, rather than relying on review articles, may also add clarity to the usefulness of the study.

eLife assessment

This important modeling work demonstrates out-of-distribution generalization using a grid cell coding scheme combined with an attentional mechanism that operates over these representations (Determinantal Point Process Attention). The simulations provide compelling evidence that the model can improve generalization performance for analogies, addition, and multiplication. The paper is significant in demonstrating how neural grid codes can support human-like generalization capabilities in analogy and arithmetic tasks, which has been a challenge for prior models.

Reviewer #1 (Public Review):

This paper presents a cognitive model of out-of-distribution generalisation, where the representational basis is grid-cell codes. In particular, the authors consider the tasks of analogies, addition, and multiplication, and the out-of-distribution tests are shifting or scaling the input domain. The authors utilise grid cell codes, which are multi-scale as well as translationally invariant due to their periodicity. To allow for domain adaptation, the authors use DPP-A which is, in this context, a mechanism of adapting to input scale changes. The authors present simulations results demonstrating this model can perform out-of-distribution generalisation to input translations and re-scaling, whereas other models fail.

This paper makes the point it sets out to - that there are some underlying representational bases, like grid cells, that when combined with a domain adaptation mechanism, like DPP-A, can facilitate out-of-generalisation. I don't have any issues with the technical details.

The paper nicely demonstrates how neural codes can be transformed into a common representational space so that analogies, and presumably other useful tasks/computations, can be performed.

Author Response

Reviewer #1 (Public Review):

Weaknesses:

The manuscript needs proper editing and is not complete. Some wordings lack precision and make it difficult to follow (e.g. line 98 "we assembled a chromosome-scale genome of ..." should read instead "we assembled a chromsome-scla genome sequence of ...". Also, panel Figure 2E is missing.

We will make the suggested change of adding “sequence”. Concerning additional changes, we have carefully edited our manuscript and looked for any incomplete sections. Unfortunately, it is difficult to see what other issues are being raised here without any further information. And the example given is not helpful to ascertain what other changes may be necessary, since we cannot see any problem with the sentence “we assembled a chromosome-scale genome of” as this phrase is widely used in many similar publications.

As for panel E of figure 2, it is not missing. The panel located to the right, just below “Target Cells”.

The shortcomings of the manuscripts are not limited to the writing style, and important technical and technological information is missing or not clear enough, thereby preventing a proper evaluation of the resolution of the genomic resources provided:

• Several RNASeq libraries from different tissues have been built to help annotate the genome and identify transcribed regions. This is fine. But all along the manuscript, gene expression changes are summarized into a single panel where it is not clear at all which tissue this comes from (whole embryo or a specific tissue ?), or whether it is a cumulative expression level computed across several tissues (and how it was computed) etc. This is essential information needed for data interpretation.

No fertilised eggs or embryos have been sequenced, individual tissues derived from juvenile fish were used for the genome annotation and whole larval fish for the developmental analysis. We will specify in the figures and text that the results shown are from whole larvae, and add more detail to the material and methods section about which type of sample was analysed in which way.

• The bioinformatic processing, especially of the assemble and annotation, is very poorly described. This is also a sensitive topic, as illustrated by the numerous "assemblathon" and "annotathon" initiatives to evaluate tools and workflows. Importantly, providing configuration files and in-depth description of workflows and parameter settings is highly recommended. This can be made available through data store services and documents even benefit from DOIs. This provides others with more information to evaluate the resolution of this work. No doubt that it is well done,but especially in the field of genome assembly and annotation, high resolution is VERY cost and time-intensive. Not surprisingly, most projects are conditioned by trade-offs between cost, time, and labor. The authors should provide others with the information needed to evaluate this.

We will upload the code used to assemble and annotate this genome to a public repository or add it to the supplementary material.

The genome assembly did not use a specific workflow (e.g., nextflow), but was done with a simple command and standard parameters in IPA. Scaffolding was carried out by Phase Genomics using their standardised proprietary workflow, of which a detailed description provided by Phase Genomics can be found in the supplementary material. The annotation workflow has been described in a previous publication already, but an in-depth description can also be found in the Material and methods section, including parameters used for specific steps. The RNA-seq mapping and analysis part has also been described in the Material and Methods section, including parameters and models for DESEq2.

• Quantifications of T3 and T4 levels look fairly low and not so convincing. The work would clearly benefit from a discussion about why the signal is so low and what are the current technological limitations of these quantifications. This would really help (general) readers.

We will add a comment on this in the manuscript as suggested. Basically, the T3/T4 levels are consistent with other published work in fish. In the present manuscript for grouper we have a peak level of 1.2 ng/g (1,200 pg/g) of T4 and 0.06 ng/g (60 pg/g) of T3. This is a higher level of T4 and comparable level of T3 to what was found in convict tang (Holzer et al. 2017; Figure 2) with 30 pg/g of T4 and 100 pg/g of T3. Of course, there are also examples with higher levels, such as clownfish (Roux et al. 2023; Figure 1), with 10 ng/g (10,000 pg/g) of T4 and 2 ng/g (2,000 pg/g) of T3.

The differences could be due to different structure of fish tissues and therefore different hormone extraction efficiency, different hormone measurement protocols, different fish physiology, different fish size (e.g., the weighting of tiny grouper larvae is difficult and less precise than in convict tang). What is important is not the absolute level but the relative level, which shows the change within different larval stages of a species with identical extraction and measurement protocols. Which means our data is internally consistent and coherent with what the grouper literature says.

Holzer, Guillaume, et al. "Fish larval recruitment to reefs is a thyroid hormone-mediated metamorphosis sensitive to the pesticide chlorpyrifos." Elife 6 (2017): e27595.

Roux, Natacha, et al. "The multi-level regulation of clownfish metamorphosis by thyroid hormones." Cell Reports 42.7 (2023).

• Differential analysis highlights up to ~ 15,000 differentially expressed genes (DEG), out of a predicted 26k genes. This corresponds to more than half of all genes. ANOVA-based differential analysis relies on the simple fact that only a minority of genes are DEG. Having >50% DEG is well beyond the validity of the method. This should be addressed, or at least discussed.

As the reviewer notes, there are a large number of differentially expressed genes due to the fact that this is coming from a larval developmental transcriptome going from one day old larva to fully metamorphosed juveniles at around day 60.

While DESeq2 indeed works on an assumption that most genes are not differentially expressed, this affects normalization but not hypothesis testing (Wald-test, LRT tests or ANOVA). Normalisation in DESeq2 is fairly robust to this assumption. According to the author of DESeq2, Micheal Love, DESeq2 is using the median ratio for normalisation, and as long as the number of up and down regulated genes is relatively even, DESeq2 will be able to handle the data. As part of our general quality control for this project we consulted the MA plots, which do not show any overrepresented up or down expression patterns. Additionally see Michael Love comment on comparing different tissues, which is also applicable here when comparing vastly different larval stages (https://support.bioconductor.org/p/63630/): “For experiments where all genes increase in expression across conditions, the median ratio method will not be able to capture this difference, but this is typically not the case for a tissue comparison, as there are many "housekeeping" genes with relatively similar expression pattern across tissues.”

Reviewer #3 (Public Review):

Weaknesses:

However, the authors make substantial considerations that are not proven by experimental or functional data. In fact, this is a descriptive study that does not provide any functional evidence to support the claims made.

We agree with the reviewer that our paper lacks functional experiments but despite that, the transcriptomic data clearly show the activation of TH and corticoid pathways during two distinct periods; an early activation between D1 and D10, and a second one between D32 and juvenile stage. These data are interesting as they call for further examination of 1) the possible interaction of corticoids and TH during metamorphosis, a question that is certainly not settled yet in teleost fishes, and 2) the existence of an early larval developmental step also involving TH and corticosteroids.

Especially 2) is of interest and importance, since this early activation (unique to our knowledge in any teleost fish studied so far) raises a lot of new questions and once again will certainly be scrutinised by other groups in the years to come, therefore ensuring a good citation impact of our study. We hope that the reviewer, while disagreeing with some our statements, will recognize that our study will be stimulating at that level and that this is what scientific studies should do.

The consideration that cortisol is involved in metamorphosis in teleosts has never been shown, and the only example cited by the authors (REF 20) clearly states that cortisol alone does not induce flatfish metamorphosis. In that work, the authors clearly state that in vivo cortisol treatment had no synergistic effect with TH in inducing metamorphosis. Moreover, in Senegalensis, the sole pre-otic CRH neuron number decreases during metamorphosis, further arguing that, at least in flatfish, cortisol is not involved in flatfish metamorphosis (PMID: 25575457).

We will do our best to improve the clarity of the revised manuscript to avoid any misunderstanding about our claims. However, we would like to point out the semantic shift in the reviewer first sentence: Indeed “being involved” is not the same as “cortisol alone does not induce”. In ref 20 the authors explicitly wrote that “Cortisol further enhanced the effects of both T4 and T3, but was ineffective in the absence of thyroid hormones” and in our view this indeed corresponds to ”being involved in metamorphosis”.

We are not claiming that cortisol alone is involved in metamorphosis as the reviewer suggests, but simply that there is a possible involvement of cortisol together with TH in metamorphosis. We stand on this claim as we indeed observed an activation of corticoid pathway genes around D32, which is sufficient to say it is involved. We do agree that functional experiments will be needed to properly demonstrate the involvement of corticoids in grouper metamorphosis, but this was not possible in the current study as it would imply to set up a full grouper life cycle in lab conditions which is impossible for the scope of this manuscript.

We also mentioned in the discussion that the role of corticoids in fish larval development is still debated, and we agree that this remain a contentious issue.

We wrote that “there is contrasting evidence of communication between these two pathways [TH and corticosteroids] in teleost fish with some data suggesting a synergic and other an antagonistic relationship. In terms of synergy, an increase in cortisol level concomitantly with an increase in TH levels has been observed in flatfish (ref 19), golden sea bream (ref 100) and silver sea bream (ref 101). Cortisol was also shown to enhance in vitro the action of TH on fin ray resorption (phenomenon occurring during flatfish metamorphosis) in flounder (ref 20). TH exposure increases MR and GR genes expression in zebrafish embryo (ref 55). It has also been shown that cortisol regulates local T3 bioavailability in the juvenile sole via regulation of deiodinase 2 in an organ-specific manner (ref 56) On the antagonistic side, it has been shown that experimentally induced hyperthyroidism in common carp, decreasing cortisol levels (ref 57), whereas cortisol exposure decreases TH levels in European eel (ref 58). Given this scattered evidence, the existence of a crosstalk active during teleost metamorphosis has never been formally demonstrated. The results we obtained in grouper are clearly indicating that HPI axis and cortisol synthesis are activated (i) during early development and (ii) during metamorphosis. This may suggest that in some aspect cortisol synthesis can work in concert with TH, as has been shown in several different contexts in amphibians (ref 17).” In the revised manuscript, we will also add the interesting case of the Senegal sole mentioned by the reviewer.

In the last revision, we had also added that our results “brought a first insight into the potential role of corticoids in the metamorphosis of E. malabaricus and call for functional experiments directly testing a possible synergy” meaning that we clearly acknowledge that we are only revealing a hypothesis that remains to be tested. We later follow up with a discussion about the most novel observation and focus of our study, the increase in THs and cortisol during early development, which was unexpected and very intriguing. Again, these results suggest that there might be a link between the two, as has been shown in amphibians. This is typically the kind of results that should encourage more investigations into other fish species. Indeed, this has been pointed out by other authors and in particular by Bob Denver (probably the foremost expert on this topic) in Crespi and Denver 2012: “Elevation in HPA/I axis activity has been described prior to Metamorphosis in amphibians and fish, birth in mammals (reviewed in Crespi & Denver 2005a; Wada 2008)”. B. Denver also adds that: “Experiments in which GCs were elevated prior to metamorphosis or prior to hatching or birth (e.g. Weiss, Johnston & Moore 2007) or inhibited by treatments with GC synthesis blockers (e.g. metyrapone) or receptor antagonists (e.g. RU486, Glennemeir & Denver 2002) demonstrate that GCs play a causal role in precipitating these life-history transitions (also reviewed in Crespi & Denver 2005a; Wada 2008).” We believe the reviewer will be convinced by these elements coming from a colleague unanimously respected in the field.

Furthermore, the authors need to recognise that the transcriptomic analysis is whole-body and that HPA axis genes are upregulated, which does not mean they are involved in regulating the HPT axis. The authors do not show that in thyrotrophs, any CRH receptor is expressed or in any other HPT axis-relevant cells and that changes in these genes correlate with changes in TSH expression. An in-situ hybridisation experiment showing co-expression on thyrotrophs of HPA genes and TSH could be a good start. However, the best scenario would be conducting cortisol treatment experiments to see if this hormone affects grouper metamorphosis.

We agree that functional experiments are needed to validate our hypothesis. As the early peaks of expression levels observed for many genes were very intriguing for us, we did carry out thyroid hormones and goitrogenic treatment on young grouper larvae to test their effect on the morphological changes. Unfortunately, such experiments, already tricky on metamorphosing larvae, are even more risky on such tiny individuals just after hatching and we encountered high mortality rates. We must add that because we cannot establish a full grouper life cycle under lab conditions, we have done these experiment in the context of a commercial husbandry system in Japan, which while excellent limits the scope of possible experiments. We were thus not able to provide functional validation of our hypothesis. Such experiments will be a full project in itself, requiring setting up a rearing system suitable for both larval survival and economical constraints related to drug treatments. We were further limited by the spawning times of the grouper in the operational aquaculture farm, which are limited to a short time during each year. So even if we strongly agree with the necessity of conducting such experiments, we think that this is not in the scope of the present paper, but something future research can explore.

High TSH and Tg levels usually parallel whole-body TH levels during teleost metamorphosis. However, in this study, high Tg expression levels are only achieved at the juvenile stage, whereas high TSH is achieved at D32, and at the juvenile stage, they are already at their lowest levels.

This is exactly our point. We observe two peaks in TSH expression, one at D3 and one at D32. The peak at D3 coincides with high thyroid hormone levels on the same day, and while we have not measured TH at D32, existing literature shows that there is a peak in TH during that time (e.g., de Jesus et al., 1998). Similarly, there is a small peak of Tg at D3. Our manuscript focused more on the upregulation of these genes at D3, which has not been reported before in the literature and raised the question of the role of TH so early in the larval development, outside of the metamorphosis period.

Regarding the respective levels of TSH and Tg, we first would like to add that their respective order of appearance before metamorphosis (TSH at D32, Tg after) is consistent with what we would expect. We agree however that the strong increase of Tg and TPO expression is later than expected. We will make this clear in the revised manuscript.

It is very difficult to conclude anything with the TH and cortisol levels measurements. The authors only measured up until D10, whereas they argue that metamorphosis occurs at D32. In this way, these measurements could be more helpful if they focus on the correct developmental time. The data is irrelevant to their hypothesis.

We respectfully disagree with the reviewer, considering that 1) TH levels have already been investigated in groupers coinciding with pigmentation changes and fin rays resorption, 2) that there is also evidence in numerous fish species that TH level increase is concomitant with increase of TH related genes, and 3) that we observed in our data an increase in the expression of TH related genes as well as pigmentation changes and fin rays resorption. Based on our experience in fish metamorphosis and the literature we can say confidently that those observations indicate that metamorphosis is occurring between D32 and the juvenile stage. To reinforce our point, we plan to add a figure to the revised manuscript, which puts our data in the context of earlier studies done in grouper. This will clearly show that our inference is correct. Additionally, we would like to point out that from our experience in several fish species transcriptomic data are more robust and precise than hormone measurements.

However, as we were surprised by the activation of TH and corticoid pathway genes very early in the larval development (at D3), which is clearly outside of the metamorphosis period, we decided to measure TH and cortisol levels during this period of time to determine if whether or not there this surprising early activation was indeed corresponding to an increase in both TH and cortisol. As such observation has never been made in other teleost species (to our knowledge), and as we were wondering if gene activation was accompanied by hormonal increase, the measurements we did for TH and cortisol between D1 and D10 are relevant. We will make sure to improve the clarity of the revised version of the manuscript to avoid any confusion between the two periods we are studying: early larval development (between D1 and D10) and metamorphosis (between D32 and juvenile stage).

Moreover, as stated in the previous review, a classical sign of teleost metamorphosis is the upregulation of TSHb and Tg, which does not occur at D32 therefore, it is very hard for me to accept that this is the metamorphic stage. With the lack of TH measurements, I cannot agree with the authors. I think this has to be toned down and made clear in the manuscript that D32 might be a putative metamorphic climax but that several aspects of biology work against it. Moreover, in D10, the authors show the highest cortisol level and lowest T4 and T3 levels. These observations are irreconcilable, with cortisol enhancing or participating in TH-driven metamorphosis.

We thank the reviewer for this comment, but we think that there might be a misunderstanding here.

(1) We clearly observed an increase of TSHb (that occurs between D18 and juvenile stage) and an increase of tg from D32 which coincide with the activation of other genes involved in TH pathway (dio2, dio3, and also a strong increase of TRb). All this and put in the context of what we know from previous grouper studies, clearly supports our conclusion that TH-regulated metamorphosis is starting at around D32 in grouper. We also observed morphological changes such as fin rays resorption and pigmentation changes between D32 and juvenile stage. Such morphological changes have already been associated as corresponding to metamorphosis in groupers (De Jesus et al 1998) as they occur during TH level increase, and they also happen to be under the control of TH in grouper (De Jesus et al 1998). Based on this study but also on studies (conducted on many other teleost species) showing that the increase of TH levels is always associated with an activation of TH pathway genes and morphological and pigmentation changes we concluded that metamorphosis of E. malabaricus occurs between D32 and juvenile stage. We will improve the clarity of the manuscript to make sure that our conclusion is based on our transcriptomic and morphological data plus the available literature.

(2) We clearly observed another activation of TH related gene earlier in the development (between D1 and D10, with a surge of trhrs, tg and tpo at D3. As this activation was very unexpected for us, we decided to focus the analysis of TH levels between D1 and D10 and very interestingly we observed high level of T4 at D3 indicating that THs are instrumental very precociously in the larval development of the malabar grouper which has never been shown before. We declared line 195 that our “data reinforce the existence of two distinct periods of TH signalling activity, one early on at D3 and one late corresponding to classic metamorphosis at D32”. However, we agree that we could have been clearer and clearly explained that this early activation was very intriguing for us and that we wanted to investigate hormonal levels around that period. However, we never claimed anywhere in the manuscript that this early developmental period corresponds to metamorphosis. Something else is occurring and both TH and cortisol seem to be involved but further experiments need to be conducted to understand their role and their possible interaction.

(3) Finally, regarding the comment about cortisol enhancing or participating in TH driven metamorphosis, our data clearly showed an activation of the corticoid pathway genes around metamorphosis (between D32 and juvenile stage) suggesting a potential implication of corticoids in metamorphosis, but we agree with the reviewer that further experiment are needed to test that. We never claimed that cortisol was enhancing or participating in metamorphosis, on the contrary we are “suggesting a possible interaction between TH and corticoid pathway during metamorphosis”. And we also say that our “results brought a first insight into the potential role of corticoids in the metamorphosis of E. malabaricus and call for functional experiments directly testing a possible synergy.” Nonetheless, we agree that some parts of our manuscript can be confusing in regards of cortisol synthesis during metamorphosis as we did not measure cortisol levels between D32 and juvenile stage. We will correct this in the revised version.

Given this, the authors should quantify whole-body TH levels throughout the entire developmental window considered to determine where the peak is observed and how it correlates with the other hormonal genes/systems in the analysis.

We did not measure TH levels at later stages as it has already been measured during Epinephelus coioides metamorphosis and the morphological changes observed in this species around the TH peak corresponds to what we observed in Epinephelus malabaricus around the peak of expression of TH pathway genes (see De Jesus et al., 1998 General and Comparative Endocrinology, 112:10-16). We are planning to add a figure reconciling all these data together. However, the main focus of this manuscript is the novel observation of the existence of an early activation period observed at D3, and for which we needed TH levels to determine if they were involved in another early developmental process (not related to metamorphosis). Our hypothesis is that this early activation might be related to the growth of fin rays necessary to enhance floatability during the oceanic larval dispersal. As we may have arrived at the explanation of this hypothesis too rapidly without setting up the context well enough, we will pay attention to improve that part too.

Even though this is a solid technical paper and the data obtained is excellent, the conclusions drawn by the authors are not supported by their data, and at least hormonal levels should be present in parallel to the transcriptomic data. Furthermore, toning down some affirmations or even considering the different hypotheses available that are different from the ones suggested would be very positive.

We thank the reviewer for acknowledging the solidity of the method of our paper and the quality of the results. We agree that there were several parts where our message is unclear, which we will address in the revised version of the manuscript to make sure there is no more confusion between the two distinct periods we studied in this paper (early larval development and metamorphosis). We will also make sure that our claims about TH/corticoids interaction during both periods remain hypothetical as we cannot yet, despite trials, sustain them with functional experiment.

eLife assessment

In this work, Huerlimann and colleagues suggest an intertalk between the thyroid and corticosteroid axis in regulating grouper metamorphosis. The work provides valuable genomic resources to address the endocrine control of a life cycle transition in the Malabar grouper fish. The evidence is still incomplete and it does not fully support an interaction between the thyroid and corticosteroid axis.

Reviewer #1 (Public Review):

Summary and strength:

The authors undertook to assemble and annotate the genome sequence of the Malabar grouper fish, with the aim of providing molecular resources for fundamental and applied research. Even though this is more mainstream, the task is still daunting and labor-intensive. Currently, high-quality and fully annotated genome sequences are of strategic importance in modern biology. The authors make use of the resource to address the endocrine control of an ecologically and developmentally relevant life cycle transition, metamorphosis. As opposed to amphibian and flat fish where body plan changes, fish metamorphosis is anatomically more subtle and much less known, although it is clear that thyroid hormone (TH) signaling is a key player. The authors thus provide a repertoire of TH-relevant gene expression changes during development and across metamorphosis and correlate developmental stages with changes in gene expression. Overall, this work has a strong potential to meet its target.

Weaknesses:

The manuscript needs proper editing and is not complete. Some wordings lack precision and make it difficult to follow (e.g. line 98 "we assembled a chromosome-scale genome of ..." should read instead "we assembled a chromsome-scla genome sequence of ...". Also, panel Figure 2E is missing.

The shortcomings of the manuscripts are not limited to the writing style, and important technical and technological information is missing or not clear enough, thereby preventing a proper evaluation of the resolution of the genomic resources provided:

- Several RNASeq libraries from different tissues have been built to help annotate the genome and identify transcribed regions. This is fine. But all along the manuscript, gene expression changes are summarized into a single panel where it is not clear at all which tissue this comes from (whole embryo or a specific tissue ?), or whether it is a cumulative expression level computed across several tissues (and how it was computed) etc. This is essential information needed for data interpretation.

- The bioinformatic processing, especially of the assemble and annotation, is very poorly described. This is also a sensitive topic, as illustrated by the numerous "assemblathon" and "annotathon" initiatives to evaluate tools and workflows. Importantly, providing configuration files and in-depth description of workflows and parameter settings is highly recommended. This can be made available through data store services and documents even benefit from DOIs. This provides others with more information to evaluate the resolution of this work. No doubt that it is well done,<br /> but especially in the field of genome assembly and annotation, high resolution is VERY cost and time-intensive. Not surprisingly, most projects are conditioned by trade-offs between cost, time, and labor. The authors should provide others with the information needed to evaluate this.

- Quantifications of T3 and T4 levels look fairly low and not so convincing. The work would clearly benefit from a discussion about why the signal is so low and what are the current technological limitations of these quantifications. This would really help (general) readers.

- Differential analysis highlights up to ~ 15,000 differentially expressed genes (DEG), out of a predicted 26k genes. This corresponds to more than half of all genes. ANOVA-based differential analysis relies on the simple fact that only a minority of genes are DEG. Having >50% DEG is well beyond the validity of the method. This should be addressed, or at least discussed.

Reviewer #3 (Public Review):

Summary:

The manuscript by Huerlimann et al. entitled "The transcriptional landscape underlying metamorphosis in the Malabar grouper (Epinephelus malabaricus)." describes the transcriptional landscape of the Malabar grouper during selected metamorphic stages. The authors find evidence of dynamic regulation of HPT axis genes, TH signalling genes, and HPA and metabolic-related genes during post-natal development. Finally, the authors argue that the HPA is involved in grouper metamorphosis, given the related genes' dynamic expression during this developmental time.

Strengths:

The work is technically very good, and the methodology applied is solid.

Weaknesses:

However, the authors make substantial considerations that are not proven by experimental or functional data. In fact, this is a descriptive study that does not provide any functional evidence to support the claims made.

The consideration that cortisol is involved in metamorphosis in teleosts has never been shown, and the only example cited by the authors (REF 20) clearly states that cortisol alone does not induce flatfish metamorphosis. In that work, the authors clearly state that in vivo cortisol treatment had no synergistic effect with TH in inducing metamorphosis. Moreover, in Senegalensis, the sole pre-otic CRH neuron number decreases during metamorphosis, further arguing that, at least in flatfish, cortisol is not involved in flatfish metamorphosis (PMID: 25575457). Furthermore, the authors need to recognise that the transcriptomic analysis is whole-body and that HPA axis genes are upregulated, which does not mean they are involved in regulating the HPT axis. The authors do not show that in thyrotrophs, any CRH receptor is expressed or in any other HPT axis-relevant cells and that changes in these genes correlate with changes in TSH expression. An in-situ hybridisation experiment showing co-expression on thyrotrophs of HPA genes and TSH could be a good start. However, the best scenario would be conducting cortisol treatment experiments to see if this hormone affects grouper metamorphosis.

High TSH and Tg levels usually parallel whole-body TH levels during teleost metamorphosis. However, in this study, high Tg expression levels are only achieved at the juvenile stage, whereas high TSH is achieved at D32, and at the juvenile stage, they are already at their lowest levels.

It is very difficult to conclude anything with the TH and cortisol levels measurements. The authors only measured up until D10, whereas they argue that metamorphosis occurs at D32. In this way, these measurements could be more helpful if they focus on the correct developmental time. The data is irrelevant to their hypothesis.

Moreover, as stated in the previous review, a classical sign of teleost metamorphosis is the upregulation of TSHb and Tg, which does not occur at D32 therefore, it is very hard for me to accept that this is the metamorphic stage. With the lack of TH measurements, I cannot agree with the authors. I think this has to be toned down and made clear in the manuscript that D32 might be a putative metamorphic climax but that several aspects of biology work against it. Moreover, in D10, the authors show the highest cortisol level and lowest T4 and T3 levels. These observations are irreconcilable, with cortisol enhancing or participating in TH-driven metamorphosis.

Given this, the authors should quantify whole-body TH levels throughout the entire developmental window considered to determine where the peak is observed and how it correlates with the other hormonal genes/systems in the analysis.

Even though this is a solid technical paper and the data obtained is excellent, the conclusions drawn by the authors are not supported by their data, and at least hormonal levels should be present in parallel to the transcriptomic data. Furthermore, toning down some affirmations or even considering the different hypotheses available that are different from the ones suggested would be very positive.

Reviewer #3 (Public Review):

Summary:

This study aimed to investigate whether the development of functional connectivity (FC) is modulated by early physical growth and whether these might impact cognitive development in childhood. This question was investigated by studying a large group of infants (N=204) assessed in Gambia with fNIRS at 5 visits between 5 and 24 months of age. Given the complexity of data acquisition at these ages and following data processing, data could be analyzed for 53 to 97 infants per age group. FC was analyzed considering 6 ensembles of brain regions and thus 21 types of connections. Results suggested that: i) compared to previously studied groups, this group of Gambian infants have different FC trajectory, in particular with a change in frontal inter-hemispheric FC with age from positive to null values; ii) early physical growth, measured through weight-for-length z-scores from birth on, is associated with FC at 24 months. Some relationships were further observed between FC during the first two years and cognitive flexibility at 4-5 years of age, but results did not survive corrections for multiple comparisons.

Strengths:

The question investigated in this article is important for understanding the role of early growth and undernutrition on brain and behavioral development in infants and children. The longitudinal approach considered is highly relevant to investigate neurodevelopmental trajectories. Furthermore, this study targets a little-studied population from a low-/middle-income country, which was made possible by the use of fNIRS outside the lab environment. The collected dataset is thus impressive and it opens up a wide range of analytical possibilities.

Weaknesses:

- Analyzing such a huge amount of collected data at several ages is not an easy task to test developmental relationships between growth, FC, and behavioral capacities. In its present form, this study and the performed analyses lack clarity, unity and perhaps modeling, as it suggests that all possible associations were tested in an exploratory way without clear mechanistic hypotheses. Would it be possible to specify some hypotheses to reduce the number of tests performed? In particular, considering metrics at specific ages or changes in the metrics with age might allow us to test different hypotheses: the authors might clarify what they expect specifically for growth-FC-behaviour associations. Since some FC measures and changes might be related to one another, would it be reasonable to consider a dimensionality reduction approach (e.g., ICA) to select a few components for further correlation analyses?

- It seems that neurodevelopmental trajectories over the whole period (5-24 months) are little investigated, and considering more robust statistical analyses would be an important aspect to strengthen the results. The discussion mentions the potential use of structural equation modelling analyses, which would be a relevant way to better describe such complex data.

- Given the number of analyses performed, only describing results that survive correction for multiple comparisons is required. Unifying the correction approach (FDR / Bonferroni) is also recommended. For the association between cognitive flexibility and FC, results are not significant, and one might wonder why FC at specific ages was considered rather than the change in FC with age. One of the relevant questions of such a study would be whether early growth and later cognitive flexibility are related through FC development, but testing this would require a mediation analysis that was not performed.

- Growth is measured at different ages through different metrics. Justifying the use of weight-for-length z-scores would be welcome since weight-for-age z-scores might be a better marker of growth and possible undernutrition (this impacting potentially both weight and length). Showing the distributions of these z-scores at different ages would allow the reader to estimate the growth variability across infants.

- Regarding FC, clarifications about the long-range vs short-range connections would be welcome, as well as drawing a summary of what is expected in terms of FC "typical" trajectory, for the different brain regions and connections, as a marker of typical development. For instance, the authors suggest that an increase in long-range connectivity vs a decrease in short-range is expected based on previous fNIRS studies. However anatomical studies of white matter growth and maturation would suggest the reverse pattern (short-range connections developing mostly after birth, contrarily to long-range connections prenatally).

The authors test associations between FC and growth, but making sense of such modulation results is difficult without a clearer view of developmental changes per se (e.g., what does an early negative FC mean? Is it an increase in FC when the value gets close to 0? In particular, at 24m, it seems that most FC values are not significantly different from 0, Figure 2B). Observing positive vs negative association effects depending on age is quite puzzling. It is also questionable, for some correlation analyses with cognitive flexibility, to focus on FC that changes with age but to consider FC at a given age.

- The manuscript uses inappropriate terms "to predict", "prediction" whereas the conducted analyses are not prediction analyses but correlational.

eLife assessment

This important study details the development of brain functional connectivity in a longitudinal cohort of Gambian children assessed outside a lab setup with functional near-infrared spectroscopy (fNIRS) from age 5 to 24 months, in relation to early physical growth and cognitive flexibility capacities at 4-5 years of age. Although the evidence supporting some conclusions is solid, the relevance of the results would be improved by defining clearer hypotheses regarding the developmental changes expected for the different connections, and by discussing the unexpected findings on early negative connectivity and connectivity decreases. Considering more advanced analytical approaches would allow the authors to deal with longitudinal data and integrate mediation links, even if the study might be underpowered to link adverse conditions such as undernutrition and later cognitive development. This study will be of interest to neuroscientists, psychologists, and neuroimaging researchers working on infant development in relation to environmental factors.

Reviewer #1 (Public Review):

Summary:

Cognitive and brain development during the first two years of life is vast and determinant for later development. However, longitudinal infant studies are complicated and restricted to occidental high-income countries. This study uses fNIRS to investigate the developmental trajectories of functional connectivity networks in infants from a rural community in Gambia. In addition to resting-state data collected from 5 to 24 months, the authors collected growing measures from birth until 24 months and administrated an executive functioning task at 3 or 5 years old.

The results show left and right frontal-middle and right frontal-posterior negative connections at 5 months that increase with age (i.e., become less negative). Interestingly, contrary to previous findings in high-income countries, there was a decrease in frontal interhemispheric connectivity. Restricted growth during the first months of life was associated with stronger frontal interhemispheric connectivity and weaker right frontal-posterior connectivity at 24 months. Additionally, the study describes that some connectivity patterns related to better cognitive flexibility at pre-school age.

Strengths:

- The authors analyze data from 204 infants from a rural area of Gambia, already a big sample for most infant studies. The study might encourage more research on different underrepresented infant populations (i.e., infants not living in occidental high-income countries).

- The study shows that fNIRS is a feasible instrument to investigate cognitive development when access to fMRI is not possible or outside a lab setting.

- The fNIRS data preprocessing and analysis are well-planned, implemented, and carefully described. For example, the authors report how the choices in the parameters for the motion artifacts detection algorithm affect data rejection and show how connectivity stability varies with the length of the data segment to justify the threshold of at least 250 seconds free of artifacts for inclusion.

- The authors use proper statistical methods for analysis, considering the complexity of the dataset.

Weaknesses:

- No co-registration of the optodes is implemented. The authors checked for correct placement by looking at pictures taken during the testing session. However, head shape and size differences might affect the results, especially considering that the study involves infants from 5 months to 24 months and that the same fNIRS array was used at all ages.

- The authors regress the global signal to remove systemic physiological noise. While the authors also report the changes in connectivity without global signal regression, there are some critical differences. In particular, the apparent decrease in frontal inter-hemispheric connections is not present when global signal regression is omitted, even though it is present for deoxy-Hb. The authors use connectivity results obtained after applying global signal regression for further analysis. The choice of regressing the global signal is questionable since it has been shown to introduce anti-correlations in fMRI data (Murphy et al., 2009), and fNIRS in young infants does not seem to be highly affected by physiological noise (Emberson et al., 2016). Systemic physiological noise might change at different ages, which makes its remotion critical to investigate functional network development. However, global signal regression might also affect the data differently. The study would have benefited from having short separation channels to measure the systemic psychological component in the data.

- I believe the authors bypass a fundamental point in their framing. When discussing the results, the authors compare the developmental trajectories of the infants tested in a rural area of Gambia with the trajectories reported in previous studies on infants growing in occidental high-income countries (likely in urban contexts) and attribute the differences to adverse effects (i.e., nutritional deficits). Differences in developmental trajectories might also derive from other environmental and cultural differences that do not necessarily lead to poor cognitive development.

- While the study provides a solid description of the functional connectivity changes in the first two years of life at the group level, the evidence regarding the links between adverse situations, developmental trajectories, and later cognitive capacities is weaker. The authors find that early restricted growth predicts specific connectivity patterns at 24 months and that certain connectivity patterns at specific ages predict cognitive flexibility. However, the link between development trajectories (individual changes in connectivity) with growth and later cognitive capacities is missing. To address this question adequately, the study should have compared infants with different growing profiles or those who suffered or did not from undernutrition. However, as the authors discussed, they lacked statistical power.

Reviewer #2 (Public Review):

Summary and strengths:

The article pertains to a topic of importance, specifically early life growth faltering, a marker of undernutrition, and how it influences brain functional connectivity and cognitive development. In addition, the data collection was laborious, and data preprocessing was quite rigorous to ensure data quality, utilizing cutting-edge preprocessing methods.

Weaknesses:

However, the subsequent analysis and explanations were not very thorough, which made some results and conclusions less convincing. For example, corrections for multiple tests need to be consistently maintained; if the results do not survive multiple corrections, they should not be discussed as significant results. Additionally, alternative plans for analysis strategies could be worth exploring, e.g., using ΔFC in addition to FC at a certain age. Lastly, some analysis plans lacked a strong theoretical foundation, such as the relationship between functional connectivity (FC) between certain ROIs and the development of cognitive flexibility.

Thus, as much as I admire the advanced analysis of connectivity that was conducted and the uniqueness of longitudinal fNIRS data from these samples (even the sheer effort to collect fNIRS longitudinally in a low-income country at such a scale!), I have reservations about the importance of this paper's contribution to the field in its present form. Major revisions are needed, in my opinion, to enhance the paper's quality.

Reviewer #1 (Public Review):

Summary:

In this study, participants completed two different tasks. A perceptual choice task in which they compared the sizes of pairs of items and a value-different task in which they identified the higher value option among pairs of items with the two tasks involving the same stimuli. Based on previous fMRI research, the authors sought to determine whether the superior frontal sulcus (SFS) is involved in both perceptual and value-based decisions or just one or the other. Initial fMRI analyses were devised to isolate brain regions that were activated for both types of choices and also regions that were unique to each. Transcranial magnetic stimulation was applied to the SFS in between fMRI sessions and it was found to lead to a significant decrease in accuracy and RT on the perceptual choice task but only a decrease in RT on the value-different task. Hierarchical drift-diffusion modelling of the data indicated that the TMS had led to a lowering of decision boundaries in the perceptual task and a lower of non-decision times on the value-based task. Additional analyses show that SFS covaries with model-derived estimates of cumulative evidence and that this relationship is weakened by TMS.

Strengths:

The paper has many strengths including the rigorous multi-pronged approach of causal manipulation, fMRI and computational modelling which offers a fresh perspective on the neural drivers of decision making. Some additional strengths include the careful paradigm design which ensured that the two types of tasks were matched for their perceptual content while orthogonalizing trial-to-trial variations in choice difficulty. The paper also lays out a number of specific hypotheses at the outset regarding the behavioural outcomes that are tied to decision model parameters and are well justified.

Weaknesses:

Unless I have missed it, the SFS does not actually appear in the list of brain areas significantly activated by the perceptual and value tasks in Supplementary Tables 1 and 2. Its presence or absence from the list of significant activations is not mentioned by the authors when outlining these results in the main text. What are we to make of the fact that it is not showing significant activation in these initial analyses?

The value difference task also requires identification of the stimuli, and therefore perceptual decision-making. In light of this, the initial fMRI analyses do not seem terribly informative for the present purposes as areas that are activated for both types of tasks could conceivably be specifically supporting perceptual decision-making only. I would have thought brain areas that are playing a particular role in evidence accumulation would be best identified based on whether their BOLD response scaled with evidence strength in each condition which would make it more likely that areas particular to each type of choice can be identified. The rationale for the authors' approach could be better justified.

TMS led to reductions in RT in the value-difference as well as the perceptual choice task. DDM modelling indicated that in the case of the value task, the effect was attributable to reduced non-decision time which the authors attribute to task learning. The reasoning here is a little unclear. If task learning is the cause, then why are similar non-decision time effects not observed in the perceptual choice task? Given that the value-task actually requires perceptual decision-making, is it not possible that SFS disruption impacted the speed with which the items could be identified, hence delaying the onset of the value-comparison choice?

The sample size is relatively small. The authors state that 20 subjects is 'in the acceptable range' but it is not clear what is meant by this.

Reviewer #3 (Public Review):

Summary:

Garcia et al., investigated whether the human left superior frontal sulcus (SFS) is involved in integrating evidence for decisions across either perceptual and/or value-based decision-making. Specifically, they had 20 participants perform two decision-making tasks (with matched stimuli and motor responses) in an fMRI scanner both before and after they received continuous theta burst transcranial magnetic stimulation (TMS) of the left SFS. The stimulation thought to decrease neural activity in the targeted region, led to reduced accuracy on the perceptual decision task only. The pattern of results across both model-free and model-based (Drift diffusion model) behavioural and fMRI analyses suggests that the left SLS plays a critical role in perceptual decisions only, with no equivalent effects found for value-based decisions. The DDM-based analyses revealed that the role of the left SLS in perceptual evidence accumulation is likely to be one of decision boundary setting. Hence the authors conclude that the left SFS plays a domain-specific causal role in the accumulation of evidence for perceptual decisions. These results are likely to add importance to the literature regarding the neural correlates of decision-making.

Strengths:

The use of TMS strengthens the evidence for the left SFS playing a causal role in the evidence accumulation process. By combining TMS with fMRI and advanced computational modelling of behaviour, the authors go beyond previous correlational studies in the field and provide converging behavioural, computational, and neural evidence of the specific role that the left SFS may play.

Sophisticated and rigorous analysis approaches are used throughout.

Weaknesses:

Though the stimuli and motor responses were equalised between the perception and value-based decision tasks, reaction times (according to Figure 1) and potential difficulty (Figure 2) were not matched. Hence, differences in task difficulty might represent an alternative explanation for the effects being specific to the perception task rather than domain-specificity per se.

No within- or between-participants sham/control TMS condition was employed. This would have strengthened the inference that the apparent TMS effects on behavioural and neural measures can truly be attributed to the left SFS stimulation and not to non-specific peripheral stimulation and/or time-on-task effects.

No a priori power analysis is presented.

eLife assessment

This important study combined fMRI, TMS and computational modelling of behaviour to investigate the functional role of the left superior frontal sulcus (SFS) in both perceptual and value-based decisions. Based on sophisticated analyses, the results provide solid evidence that downregulating left SFS activity through TMS selectively alters perceptual decision accuracy but does not influence value-based decisions. The work will be of interest to cognitive neuroscientists investigating the neural correlates of decision-making and may have implications for computational psychiatry.

Reviewer #2 (Public Review):

Summary:

The authors set out to test whether a TMS-induced reduction in excitability of the left Superior Frontal Sulcus influenced evidence integration in perceptual and value-based decisions. They directly compared behaviour - including fits to a computational decision process model - and fMRI pre and post-TMS in one of each type of decision-making task. Their goal was to test domain-specific theories of the prefrontal cortex by examining whether the proposed role of the SFS in evidence integration was selective for perceptual but not value-based evidence.

Strengths:

The paper presents multiple credible sources of evidence for the role of the left SFS in perceptual decision-making, finding similar mechanisms to prior literature and a nuanced discussion of where they diverge from prior findings. The value-based and perceptual decision-making tasks were carefully matched in terms of stimulus display and motor response, making their comparison credible.

Weaknesses:<br /> More information on the task and details of the behavioural modelling would be helpful for interpreting the results. I had the following concerns:

(1) The evidence for a choice and 'accuracy' of that choice in both tasks was determined by a rating task that was done in advance of the main testing blocks (twice for each stimulus). For the perceptual decisions, this involved asking participants to quantify a size metric for the stimuli, but the veracity of these ratings was not reported, nor was the consistency of the value-based ones. It is my understanding that the size ratings were used to define the amount of perceptual evidence in a trial, rather than the true size differences, and without seeing more data the reliability of this approach is unclear. More concerning was the effect of 'evidence level' on behaviour in the value-based task (Figure 3a). While the 'proportion correct' increases monotonically with the evidence level for the perceptual decisions, for the value-based task it increases from the lowest evidence level and then appears to plateau at just above 80%. This difference in behaviour between the two tasks brings into question the validity of the DDM which is used to fit the data, which assumes that the drift rate increases linearly in proportion to the level of evidence.

(2) The paper provides very little information on the model fits (no parameter estimates, goodness of fit values or simulated behavioural predictions). The paper finds that TMS reduced the decision bound for perceptual decisions but only affected non-decision time for value-based decisions. It would aid the interpretation of this finding if the relative reliability of the fits for the two tasks was presented.

(3) Behaviourally, the perceptual task produced decreased response times and accuracy post-TMS, consistent with a reduced bound and consistent with some prior literature. Based on the results of the computational modelling, the authors conclude that RT differences in the value-based task are due to task-related learning, while those in the perceptual task are 'decision relevant'. It is not fully clear why there would be such significantly greater task-related learning in the value-based task relative to the perceptual one. And if such learning is occurring, could it potentially also tend to increase the consistency of choices, thereby counteracting any possible TMS-induced reduction of consistency?

Reviewer #1 (Public Review):

Summary:

In their manuscript, "Nicotine enhances the stemness and tumorigenicity in intestinal stem cells via Hippo-YAP/TAZ and Notch signal pathway", authors Isotani et al claimed that this study identifies a NIC-triggered pathway regulating the stemness and tumorigenicity of ISCs and suggest the use of DBZ as a potential therapeutic strategy for treating intestinal tumors. However, the presented data do not support the primary claims.

Weaknesses:

My main reservation is that the quality of the results presented in the manuscript may not fully substantiate their conclusions. For instance, in Figure 2 A and B, it is challenging to discern a healthy organoid. This is significant, as the entirety of Figure 2 and several panels in Figures 3 - 5 are based on these organoid assays. Additionally, there seems to be a discrepancy in the quality of results from the western blot, as the lanes of actin do not align with other proteins (Figure 6B).

eLife assessment

This study presents a valuable finding on a potential signaling pathway responsible for the direct effects of nicotine on intestinal stem cell growth and tumorigenesis. However, the evidence supporting the authors' claims remains incomplete. Additional analysis on how stem cells uniquely respond to nicotine could provide more definitive evidence and strengthen the study. This research will be of interest to medical biologists specializing in intestinal tumors.

Reviewer #2 (Public Review):

Summary:

The manuscript by Isotani et al characterizes the hyperproliferation of intestinal stem cells (ISCs) induced by nicotine treatment in vivo. Employing a range of small molecule inhibitors, the authors systematically investigated potential receptors and downstream pathways associated with nicotine-induced phenotypes through in vitro organoid experiments. Notably, the study specifically highlights a signaling cascade involving α7-nAChR/PKC/YAP/TAZ/Notch as a key driver of nicotine-induced stem cell hyperproliferation. Utilizing a Lgr5CreER Apcfl/fl mouse model, the authors extend their findings to propose a potential role of nicotine in stem cell tumorgenesis. The study posits that Notch signaling is essential during this process.

Strengths and Weaknesses:

One noteworthy research highlight in this study is the indication, as shown in Figure 2 and S2, that the trophic effect of nicotine on ISC expansion is independent of Paneth cells. In the Discussion section, the authors propose that this independence may be attributed to distinct expression patterns of nAChRs in different cell types. To further substantiate these findings, it is suggested that the authors perform tissue staining of various nAChRs in the small intestine and colon. This additional analysis would provide more conclusive evidence regarding how stem cells uniquely respond to nicotine. It is also recommended to present the staining of α7-nAChR from different intestinal regions. This will provide insights into the primary target sites of nicotine in the gut tract. Additionally, it is recommended that the authors consider rephrasing the conclusion in this section (lines 123-124). The current statement implies that nicotine does not affect Paneth cells, which may be inaccurate based on the suggestion in line 275 that nicotine might influence Paneth cells through α2β4-nAChR. Providing a more nuanced conclusion would better reflect the complexity of nicotine's potential impact on Paneth cells.

As shown in the same result section, the effect of nicotine on ISC organoid formation appears to be independent of CHIR99021, a Wnt activator. Despite this, the authors suggest a potential involvement of Wnt/β-catenin activation downstream of nicotine in Figure 4F. In the Lgr5CreER Apcfl/fl mouse model, it is known that APC loss results in a constitutive stabilization of β-catenin, thus the hyperproliferation of ISCs by nicotine treatment in this mouse model is likely beyond Wnt activation. Therefore, it is recommended that the authors reconsider the inclusion of Wnt/β-catenin as a crucial signaling pathway downstream of nicotine, given the experimental evidence provided in this study.

In Figure 4, the authors investigate ISC organoid formation with a pan-PKC inhibitor, revealing that PKC inhibition blocks nicotine-induced ISC expansion. It's noteworthy that PKC inhibitors have historically been used successfully to isolate and maintain stem cells by promoting self-renewal. Therefore, it is surprising to observe no effect or reversal effect on ISCs in this context. A previous study demonstrated that the loss of PKCζ leads to increased ISC activity both in vivo and in vitro (DOI: 10.1016/j.celrep.2015.01.007). Additionally, to strengthen this aspect of the study, it would be beneficial for the authors to present more evidence, possibly using different PKC inhibitors, to reproduce the observed results with Gö 6983. This could help address potential concerns or discrepancies and contribute to a more comprehensive understanding of the role of PKC in nicotine-induced ISC expansion.

An additional avenue that could enhance the clinical relevance of the study is the exploration of human datasets. Specifically, leveraging scRNA-seq datasets of the human intestinal epithelium (DOI: 10.1038/s41586-021-03852-1) could provide valuable insights. Analyzing the expression patterns of nAChRs across diverse regions and cell types in the human intestine may offer a potential clinical implication.

In summary, the results generally support the authors' conclusions that nicotine directly influences ISC growth, potentially contributing to tumorgenesis. The identification of the α7-nAChR/PKC/YAP/TAZ/Notch pathway adds significant mechanistic insight. However, certain aspects of the experimental evidence, such as the receptor expression pattern, PKC inhibition response, and the involvement of Wnt/β-catenin activation, may require further clarification and exploration, especially considering previous literature suggesting potential discrepancies.

Author Response

We provide here a provisional response to the Public Comments and main issues raised by the reviewers. We appreciate the opportunity to submit a revision and will give all of the reviewers’ comments careful consideration when modifying the manuscript.

(1) BioRxiv version history.

Reviewer 1 correctly noted that we have posted different versions of the paper on bioRxiv and that there were significant changes between the initial version and the one posted as part of the eLife preprint process. Here we provide a summary of that history.

We initially posted a bioRxiv preprint in November, 2021 (Version 1) that included the results of two experiments. In Experiment 1, we compared conditions in which the stimulation frequency was at 2 kHz, 3.5 kHz, or 5.0 kHz. In Experiment 2, we replicated the 3.5 kHz condition of Experiment 1 and included two amplitude-modulated (AM) conditions, with a 3.5 kHz carrier signal modulated at 20 Hz or 140 Hz. Relative to the sham stimulation, non-modulated kTMP at 2 kHz and 3.5 kHz resulted in an increase in cortical excitability in Experiment 1. This effect was replicated in Experiment 2.

In the original posting, we reported that there was an additional boost in excitability in the 20 Hz AM condition above that of the non-modulated condition. However, in re-examining the results, we recognized that the 20 Hz AM condition included an outlier that was pulling the group mean higher. We should have caught this outlier in the initial submission given that the resultant percent change for this individual is 3 standard deviations above the mean. Given the skew in the distribution, we also performed a log transform on the MEPs (which improves the normality and homoscedasticity of MEP distributions) and repeated the analysis. However, even here the participant’s results remained well outside the distribution. As such, we removed this participant and repeated all analyses. In this new analysis, there was no longer a significant difference between the 20 Hz AM and nonmodulated conditions in Experiment 2. Indeed, all three true stimulation conditions (nonmodulated, AM 20 Hz, AM 140 Hz) produced a similar boost in cortical excitability compared to sham. Thus, the results of Experiment 2 are consistent with those of Experiment 1, showing, in three new conditions, the efficacy of kHz stimulation on cortical excitability. But the results fail to provide evidence of an additional boost from amplitude modulation.

We posted a second bioRxiv preprint in May, 2023 (Version 2) with the corrected results for Experiment 2, along with changes throughout the manuscript given the new analyses.

Given the null results for the AM conditions, we decided to run a third experiment prior to submitting the work for publication. Here we used an alternative form of amplitude modulation (see Kasten et. al., NeuroImage 2018). In brief, we again observed a boost in cortical excitability in from non-modulated kTMP at 3.5 kHz, but no additional effect of amplitude modulation. This work is included in the third bioRrxiv preprint (Version 3), the paper that was submitted and reviewed at eLife.

(2) Statistical analysis.

Reviewer 1 raised a concern with the statistical analyses performed on aggregate data across experiments. We recognize that this is atypical and was certainly not part of an a priori plan. Here we describe our goal with the analyses and the thought process that led us to combine the data across the experiments.

Our overarching aim is to examine the effect of corticospinal excitability of different kTMP waveforms (carrier frequency and amplitude modulated frequency) matched at the same estimated cortical E-field (2 V/m). Our core comparison was of the active conditions relative to a sham condition (E-field = 0.01 V/m). We included the non-modulated 3.5 kHz condition in Experiments 2 and 3 to provide a baseline from which we could assess whether amplitude modulation produced a measurable difference from that observed with non-modulated stimulation. Thus, this non-modulated condition as well as the sham condition was repeated in all three experiments. This provided an opportunity to examine the effect of kTMP with a relatively large sample, as well as assess how well the effects replicate, and resulted in the strategy we have taken in reporting the results.

As a first step, we present the data from the 3.5 kHz non-modulated and sham conditions (including the individual participant data) for all three experiments in Figure 4. We used a linear mixed effect model to examine if there was an effect of Experiment (Exps 1, 2, 3) and observed no significant difference within each condition. Given this, we opted to pool the data for the sham and 3.5 kHz non-modulated conditions across the three experiments. Once data were pooled, we examined the effect of the carrier frequency and amplitude modulated frequency of the kTMP waveform.

(3) Carry-over effects

As suggested by Reviewer 1, we will examine in the revision if there is a carry-over effect across sessions (for the most part, 2-day intervals between sessions). For this, we will compare MEP amplitude in baseline blocks (pre-kTMP) across the four experimental sessions.

Reviewer 1 also commented that mixing the single- and paired-pulse protocols might have impacted the results. While our a priori focus was on the single-pulse results, we wanted to include multiple probes given the novelty of our stimulation method. Mixing single- and different paired-pulse protocols has been relatively common in the noninvasive brain stimulation literature (e.g., Nitsche 2005, Huang et al, 2005, López-Alonso 2014, Batsikadze et al 2013) and we are unaware of any reports suggested that mixed designs (single and paired) distort the picture compared to pure designs (single only).

(4) Sensation and Blinding

Reviewer 2 bought up concerns about the sham condition and blinding of kTMP stimulation. We do think that kTMP is nearly ideal for blinding. The amplifier does emit an audible tone (at least for individuals with normal hearing) when set to an intensity to produce a 2 V/m E-field. For this reason, the participants and the experimenter wore ear plugs. Moreover, we played a 3.5 kHz tone in all conditions, including the sham condition, which effectively masked the amplifier sound. We measured the participant’s subjective rating of annoyance, pain, and muscle twitches after each kTMP session (active and sham). Using a linear mixed effect model, we found no difference between active and sham for each of these ratings suggesting that sensation was similar for active and sham (Fig 8). This matches our experience that kHz stimulation in the range used here has no perceptible sensation induced by the coil. To blind the experimenters (and participants) we used a coding system in which the experimenter typed in a number that had been randomly paired to a stimulation condition that varied across participants in a manner unknown to the experimenter.

Reviewer 1 asked why we did not explicitly ask participants if they thought they were in an active or sham condition. This would certainly be a useful question. However, we did not want to alert them of the presence of a sham condition, preferring to simply describe the study as one testing a new method of non-invasive brain stimulation. Thus, we opted to focus on their subjective ratings of annoyance, pain, and finger twitches after kTMP stimulation for each experimental session.

eLife assessment

This important study reports the first results on the effects of subthreshold kilohertz sinusoidal transcranial magnetic stimulation (TMS) on the brain. Stimulation at fields of 2V/m and 3.5kHz enhances cortical motor excitability as measured by motor-evoked potentials elicited by single-pulse TMS. The evidence in support of this claim is compelling. This result is of importance to the field of non-invasive brain stimulation and to cognitive neuroscience as a whole.

Reviewer #1 (Public Review):

Summary:

This paper reports the first results on the effects of a novel waveform for weak transcranial magnetic stimulation, which they refer to as "perturbation" (kTMP). The waveform is sinusoidal at kHz frequency with subthreshold intensities of 2V/m, instead of the suprathreshold pulses used in conventional TMS (~100V/m). The effect reported here concerns motor-evoked potentials (MEPs) elicited on the hand with single-pulse TMS. These MEPs are considered a marker of "corotico-spinal excitability. The manuscripts report that kTMP at 3.5kHz enhances MEPs with a medium effect size, and reports independent replications of this fining on 3 separate cohorts of subjects (N=16, 15, 16). This result is important for the field of non-invasive brain stimulation. The evidence in support of this claim is compelling.

Strengths:

• This is a novel modality for non-invasive brain stimulation.

• Knowing the history in this field, is likely to lead to a large number of follow-up studies in basic and clinical research.

• The modality cases practically no sensation which makes it perfectly suitable for control conditions. Indeed, the study itself used a persuasive double-blinding procedure.

• The replication of the main result in two subsequent experiments is very compelling.

• The effect size of Cohen's d=0.5 is very promising.

• It is nice the E-fields were actually measured on a phantom, not just modeled.

Weakness:

• The within-subject design may have carry-over effects, although a 2-day gap is probably enough for washout.

• It would have been nice to assess washout by comparing the per-conditions between days. Particularly problematic are the paired-pulse effects that are done within sessions in experiments 2 and 3 which could have carried over to the main metric of interest, which was the single pulse MEP.

• Statistical analysis combining Experiments 1, 2, and 3 is a little muddled.

• Related, the biorxiv version history of this work as experiments 1-3 came together to point to diverging results, and changing analysis methods. Specifically, an earlier version of the work claims that modulated kHz sinusoids are more effective than un-modulated sinusoids, yet the current version says that no differences were detected - which seems consistent with the data presented in this version. However, it does raise concerns about analytic methods, which seem to have shifted over time.

• While sensation has been documented nicely, it does not seem like blinding has been directly assessed, by asking participants at the end which group they thought to be in.

Reviewer #2 (Public Review):

Summary:

kTMP is a novel method of stimulating the brain using electromagnetic fields. It has potential benefits over existing technology because it is safe and easy. It explores a range of brain frequencies that have not been explored in depth before (2-5kHz) and thus offers new opportunities.

Strengths:

This work relied on standard methods and was carefully and conservatively performed.

Weaknesses:

The sham condition was prepared as well as could be done, but sham is always challenging in a treatment with sound and sensation and with knowledgeable operators. New technology, also, is very exciting to subjects and it is difficult to achieve a natural experiment. These difficulties are related to the technology, however, and not to the execution of these experiments.

Author Response

Provisional Response to Public Reviews

Reviewer #1 (Public Review):

Summary:

The work by Zeng et al. comprehensively explored the differences in the effects of leaf and soil microbes on the seed germination, seedling survival, and seedling growth of an invasive forb, Ageratina adenophora, and found evidence of stronger effects of leaf microbes on Ageratina compared with soil microbes, which were negative for seed germination and seedling survival but positive for seedling growth. By further DNA sequencing and fungal strain cultivation, the authors were able to identify some of the key microbial guilds that may facilitate such negative and positive feedback.

Thank you very much for your assessment.

Strengths:

(1) The theoretic framework is well-established.

(2) Relating the direction of plant-microbe feedback to certain microbial guilds is always hard, but the authors have done a great job of identifying and interpreting such relationships.

Thank you very much for your assessment.

Weaknesses:

(1) In the G0 and G21 inoculation experiments, allelopathic effects from leaf litters had not been accounted for, while these two experiments happened to be the ones where negative feedback was detected.

We did not directly test the allelopathic effects. However, our inoculation of sterile litter or soil indicated the potential allelopathic role in germination and seedling mortality. Interestingly, such allelopathic effects are elicited by leaf litter not by soil, which include delaying germination time (see Fig. 1a) and killing some seedlings (see Fig. 1c). Nonetheless, microbial effects are significantly more adverse than allelopathic (also see Fig. 1e). We will discuss this point in the resubmitted version.

(2) The authors did not compare the fungal strains accumulated in dead seedlings to those accumulated in live seedlings to prove that the live seedlings indeed accumulated lower abundances of the strains that were identified to increase seedling mortality.

Thanks for your concerns. We have not isolated fungi from healthy seedlings to make a comparative study. However, our team work previously found that the seedling-killing Allophoma strains obtained in this study had the same ITS genes as the leaf endophyte and leaf spot pathogen Allophoma associated with mature A. adenophora individual; some seedling-killing Alternaria also occur in healthy seedlings inoculated by leaf litter. We thus assumed that these seedling-killing fungi, e.g., Allophoma and Alternaria, likely exist in A. adenophora mature individual by a lifestyle switch from endophytic to pathogenic, and these fungi can kill seedling only at very early life stage of A. adenophora.

Thus, we discussed this point as: “We did not isolate fungi from healthy seedlings in this study. However, a previous report revealed that the dominant genera in healthy seedlings inoculated with leaf litter were Didymella and Alternaria (Kai Fang et al., 2019). Based on these results, these fungal genera likely exist in A. adenophora by a lifestyle switch from endophytic to pathogenic. The virulence of these strains for seedling survival under certain conditions may play an essential role in limiting the population density of A. adenophora monocultures.” See Lines 416-435.

Here, we also will consider adding more sentences to discuss your concerns in the resubmitted version as: “It is worth to explore the dynamic of these strains along with seedling development and to determine if these strains kill seedling only at very early stage.”

(3) The data of seed germination and seedling mortality could have been analyzed in the same manner as that of seedling growth, which makes the whole result section more coherent. I don't understand why the authors had not calculated the response index (RI) for germination/mortality rate and conducted analyses on the correlation between these RIs with microbial compositions.

Thanks so much. Response index (RI) was calculated as: (variablenon-sterile–variablesterile)/variablesterile)). Because mortality rates of some sterile groups were zero values, it is impossible to calculate their RIs. Relatively, microbes rarely affect seed germination time (GT) and rate (GR) (see Fig. 1a,b). Therefore, we preferred to make a direct comparison of their difference between non-sterile and sterile treatments (see also Figure S2), and we also conducted a correlation by these values with microbial compositions rather than by RIs (see Fig. 4).

We will emphasis this point in the Materials and Methods when resubmit our revision.

(4) The language of the manuscript could be improved to increase clarity.

We will improve this in the resubmitted version.

Reviewer #2 (Public Review):

Summary:

The study provides strong evidence that leaf microbes mediate self-limitation at an early life stage. It highlights the importance of leaf microbes in population establishment and community dynamics.

Thank you very much for your assessment.

The authors conducted three experiments to test their hypothesis, elucidating the effects of leaf and soil microbial communities on the seedling growth of A. adenophora at different stages, screening potential microbial sources associated with seed germination and seedling performance, and identifying the fungus related to seedling mortality. The conclusions are justified by their results. Overall, the paper is well-structured, providing clear and comprehensive information.

Thank you very much for your assessment.

eLife assessment

This valuable study advances our understanding of how leaf and soil microbes separately affect the performance of an invasive plant, Ageratina adenophora. The conclusions regarding the roles of litter microbes in regulating the A. adenophora population are currently supported only by incomplete evidence owing to limitations in experimental design and statistical analyses, as well as uncertainties associated with the presentation. The work will be of interest to invasion biologists.

Reviewer #1 (Public Review):

Summary:

The work by Zeng et al. comprehensively explored the differences in the effects of leaf and soil microbes on the seed germination, seedling survival, and seedling growth of an invasive forb, Ageratina Adenophora, and found evidence of stronger effects of leaf microbes on Ageratina compared with soil microbes, which were negative for seed germination and seedling survival but positive for seedling growth. By further DNA sequencing and fungal strain cultivation, the authors were able to identify some of the key microbial guilds that may facilitate such negative and positive feedback.

Strengths:

(1) The theoretic framework is well-established.

(2) Relating the direction of plant-microbe feedback to certain microbial guilds is always hard, but the authors have done a great job of identifying and interpreting such relationships.

Weaknesses:

(1) In the G0 and G21 inoculation experiments, allelopathic effects from leaf litters had not been accounted for, while these two experiments happened to be the ones where negative feedback was detected.

(2) The authors did not compare the fungal strains accumulated in dead seedlings to those accumulated in live seedlings to prove that the live seedlings indeed accumulated lower abundances of the strains that were identified to increase seedling mortality.

(3) The data of seed germination and seedling mortality could have been analyzed in the same manner as that of seedling growth, which makes the whole result section more coherent. I don't understand why the authors had not calculated the response index (RI) for germination/mortality rate and conducted analyses on the correlation between these RIs with microbial compositions.

(4) The language of the manuscript could be improved to increase clarity.

Reviewer #2 (Public Review):

Summary:

The study provides strong evidence that leaf microbes mediate self-limitation at an early life stage. It highlights the importance of leaf microbes in population establishment and community dynamics.

The authors conducted three experiments to test their hypothesis, elucidating the effects of leaf and soil microbial communities on the seedling growth of A. adenophora at different stages, screening potential microbial sources associated with seed germination and seedling performance, and identifying the fungus related to seedling mortality. The conclusions are justified by their results. Overall, the paper is well-structured, providing clear and comprehensive information.

eLife assessment

This work provides a valuable characterization of neural activity in the anterior insular cortex during fear. The data were collected using behavior, single unit recording, and optogenetic control of neural activity. The study is a great starting point on the path to testing hypotheses about bidirectional control of behavior via segregated, anatomically defined output populations, since the authors recorded true neural activity, as opposed to bulk calcium flux, which is often used in this kind of experiment. However, the link between spiking, anatomy, and behavior is incomplete, and additional controls are necessary to support the claim for associative learning.

Reviewer #1 (Public Review):

The authors sought to test whether anterior insular cortex neurons increase or decrease firing during fear behavior and freezing, bi-directionally control fear via separate, anatomically defined outputs. Using a fairly simple behavior where mice were exposed to tone-shock pairings, they found roughly equal populations that do indeed either increase or decrease firing during freezing. Next, they sought to test whether these distinct populations may also have distinct outputs. Using retrograde tracers they found that the anterior insular cortex contains non-overlapping neurons which project to the mediodorsal thalamus or amygdala. Mediodorsal thalamus-projecting neurons tended to cluster in deep cortical layers while amygdala-projecting neurons were primarily in more superficial layers. Stimulation of insula-thalamus projection decreased freezing behavior, and stimulation of insula-amygdala projections increased fear behavior. Given that the neurons that increased firing were located in deep layers, that thalamus projections occurred in deep layers, and that stimulation of insula-thalamus neurons decreased freezing, the authors concluded that the increased firing neurons may be thalamus projections. Similarly, given that decreased-firing neurons tended to occur in more superficial layers, that insula-amygdala projections were primarily superficial, and that insula-amygdala stimulation increased freezing behavior, authors concluded that the decreased firing cells may be amygdala projections. The study has several strengths though also some caveats.

Strengths:

The potential link between physiological activity, anatomy, and behavior is well laid out and is an interesting question. The activity contrast between the units that increase/decrease firing during freezing is clear.

It is nice to see the recording of extracellular spiking activity, which provides a clear measure of neural output, whereas similar studies often use bulk calcium imaging, a signal that rarely matches real neural activity even when anatomy suggests it might (see London et al 2018 J Neuro - there are increased/decreased spiking striatal populations, but both D1 and D2 striatal neurons increase bulk calcium).

Weaknesses:

The link between spiking, anatomy, and behavior requires assumptions/inferences: the anatomically/genetically defined neurons which had distinct outputs and opposite behavioral effects can only be assumed the increased/decreased spiking neurons, based on the rough area of the cortical layer they were recorded.

The behavior would require more control to fully support claims about the associative nature of the fear response (see Trott et al 2022 eLife) - freezing, in this case, could just as well be nonassociative. In a similar vein, fixed intertrial intervals, though common practice in the fear literature, pose a problem for neurophysiological studies. The first is that animals learn the timing of events, and the second is that neural activity is dynamic and changes over time. Thus it is very difficult to determine whether changes in neural activity are due to learning about the tone-shock contingency, timing of the task, simply occur because of time and independently of external events, or some combination of the above.

Reviewer #2 (Public Review):

In this study, the authors aim to understand how neurons in the anterior insular cortex (insula) modulate fear behaviors. They report that the activity of a subpopulation of insula neurons is positively correlated with freezing behaviors, while the activity of another subpopulation of neurons is negatively correlated to the same freezing episodes. They then used optogenetics and showed that activation of anterior insula excitatory neurons during tones predicting a footshock increases the amount of freezing outside the tone presentation, while optogenetic inhibition had no effect. Finally, they found that two neuronal projections of the anterior insula, one to the amygdala and another to the medial thalamus, are increasing and decreasing freezing behaviors respectively. While the study contains interesting and timely findings for our understanding of the mechanisms underlying fear, some points remain to be addressed.

eLife assessment

This paper describes the structure and connectivity of brain neurons that send descending connections to motor neurons and muscle in the fruit fly nerve cord, using a synapse-resolution connectome. This valuable work provides a wealth of hypotheses and predictions for future experimentation and modelling. Using state-of-the-art methods, the authors provide solid evidence for their conclusions. Some conclusions however could be qualified to acknowledge currently unavoidable ambiguities associated with current methodologies.

Reviewer #1 (Public Review):

Summary:

Cheong et al. use a synapse-resolution wiring map of the fruit fly nerve cord to comprehensively investigate circuitry between descending neurons (DNs) from the brain and motor neurons (MNs) that enact different behaviours. These neurons were painstakingly identified, categorised, and linked to existing genetic driver lines; this allows the investigation of circuitry to be informed by the extensive literature on how flights walk, fly, and escape from looming stimuli. New motifs and hypotheses of circuit function were presented. This work will be a lasting resource for those studying nerve cord function.

Strengths:

The authors present an impressive amount of work in reconstructing and categorising the neurons in the DN to MN pathways. There is always a strong link between the circuitry identified and what is known in the literature, making this an excellent resource for those interested in connectomics analysis or experimental circuits neuroscience. Because of this, there are many testable hypotheses presented with clear predictions, which I expect will result in many follow-up publications. Most MNs were mapped to the individual muscles that they innervate by linking this connectome to pre-existing light microscopy datasets. When combined with past fly brain connectome datasets (Hemibrain, FAFB) or future ones, there is now a tantalising possibility of following neural pathways from sensory inputs to motor neurons and muscle.

Weaknesses:

As with all connectome datasets, the sample size is low, limiting statistical analyses. Readers should keep this in mind, but note that this is the current state-of-the-art. Some figures are weakened by relying too much on depictions of wiring diagrams as evidence of circuit function, similarity between neuropils, etc. without additional quantitative justification.

Reviewer #2 (Public Review):

Summary:

In Cheong et al., the authors analyze a new motor system (ventral nerve cord) connectome of Drosophila. Through proofreading, cross-referencing with another female VNC connectome, they define key features of VNC circuits with a focus on descending neurons (DNs), motor neurons (MNs), and local interneuron circuits. They define DN tracts, MNs for limb and wing control, and their nerves (although their sample suffers for a subset of MNs). They establish connectivity between DNs and MNs (minimal). They perform topological analysis of all VNC neurons including interneurons. They focus specifically on identifying core features of flight circuits (control of wings and halteres), leg control circuits with a focus on walking rather than other limbed behaviors (grooming, reaching, etc.), and intermediate circuits like those for escape (GF). They put these features in the context of what is known or has been posited about these various circuits.

Strengths:

Some strengths of the manuscript include the matching of new DN and MN types to light microscopy, including the serial homology of leg motor neurons. This is a valuable contribution that will certainly open up future lines of experimental work.

Also, the analysis of conserved connectivity patterns within each leg neuromere and interconnecting connectivity patterns between neuromeres will be incredibly valuable. The standard leg connectome is very nice.

Finally, the finding of different connectivity statistics (degrees of feedback) in different neuropils is quite interesting and will stimulate future work aimed at determining its functional significance.

Weaknesses:

First, it seems like quite a limitation that the neurotransmitter predictions were based on training data from a fairly small set of cells, none of which were DNs. It's wonderful that the authors did the experimental work to map DN neurotransmitter identity using FISH, and great that the predictions were overall decently accurate for both ACh and Glu, but unfortunate that they were not accurate for GABA. I hope there are plans to retrain the neurotransmitter predictions using all of this additional ground truth experimental data that the authors collected for DNs, in order to provide more accurate neurotransmitter type predictions across more cell types.

Second, the degradation of many motor neurons is unfortunate. Figure 5 Supplement 1 shows that roughly 50% of the leg motor neurons have significantly compromised connectivity data, whereas, for non-leg motor neurons, few seem to be compromised. If that is the correct interpretation of this figure, perhaps a sentence like this that includes some percentages (~50% of leg MNs, ~5% of other MNs) could be added to the main text so that readers can get a sense of the impact more easily.

As well, Figure 5 Supplement 1 caption says "Note that MN groups where all members of the group have reconstruction issues may not be flagged" - could the authors comment on how common they think this is based on manual inspection? If it changes the estimate of the percentage of affected leg motor neurons from 50% to 75% for example, this caveat in the current analysis would need to be addressed more directly. Comparing with FANC motor neurons could perhaps be an alternative/additional approach for estimating the number of motor neurons that are compromised.

This analysis might benefit from some sort of control for true biological variability in the number of MN synapses between left and right or across segments. I assume the authors chose the threshold of 0.7 because it seemed to do a good job of separating degraded neurons from differences in counts that could just be due to biological variability or reconstruction imperfections, but perhaps there's some way to show this more explicitly. For example, perhaps show how much variability there is in synapse counts across all homologs for one or two specific MN types that are not degraded and are reconstructed extremely well, so any variability in input counts for those neurons is likely to be biologically real. Especially because the identification of serial homologs among motor neurons is a key new contribution of this paper, a more in-depth analysis of similarities and differences in homologous leg MNs across segments could be interesting to the field if the degradation doesn't preclude it.

Fourth, the infomap communities don't seem to be so well controlled/justified. Community detection can be run on any graph - why should I believe that the VNC graph is actually composed of discrete communities? Perhaps this comes from a lack of familiarity with the infomap algorithm, but I imagine most readers will be similarly unfamiliar with it, so more work should be done to demonstrate the degree to which these communities are really communities that connect more within than across communities.

I think the length of this manuscript reduces its potential for impact, as I suspect the reality is that many people won't read through all 140 pages and 21 main figures of (overall excellent) work and analysis.

eLife assessment:

This study reports compelling data supporting the role of beta-catenin on intercellular communication occurring via extracellular vesicles, with implications for immune evasion in hepatocellular carcinoma (HCC). This fundamental insight sheds light on the underlying biology of HCC at a time when an increasing number of treatment options, including targeted small molecular inhibitors and immuno-oncologic drugs are being used to treat patients in the absence of validated predictive biomarkers. This work will be of special interest to researchers investigating the basic biology of liver function and HCC, and also to readers investigating novel pathways for therapeutic targeting of HCC.

Reviewer #1 (Public Review):

Summary:

This finding shows a connection between cancer-associated beta-catenin mutations and extracellular vesicle secretion. A link between the beta-catenin mutation and expression of trafficking and exocytosis machinery. They used a multidisciplinary approach to explore expression levels of relevant proteins and single-particle imaging to directly explore the release of extracellular vesicles. These results suggest a role of extracellular vesicles in immune evasion in liver cancer with the role needing to be further explored in other forms of cancer. I find this work to be compelling and of strong significance.

Strengths:

This paper uses multidisciplinary methods to demonstrate the compelling role of beta-catenin mutations in suppressing EV secretion in tumors. The results and imaging are extremely convincing and compelling.

Weaknesses:

There is no major weakness in this work. There are only things that left me more intrigued about this work. While the role of Rab27 was strongly examined, the hits of the VAMP proteins were not explored in detail. I was wondering if the decrease in the presence of VAMPS directly suggests the final step of membrane fusion in the exocytosis of EVs is what is being impaired. Or if it is other trafficking steps along the EV secretion pathway.

Reviewer #2 (Public Review):

Summary:

Dantzer and colleagues are investigating the pivotal role of ß-catenin, a gene that undergoes mutation in various cancer cells, and its influence on promoting the evasion of immune cells. In their initial experiments, the authors developed a HepG2 mutated ß-catenin KD model, conducting transcriptional and proteomic analyses. The results revealed that the silencing of mutated ß-catenin in HepG2 cells led to an up-regulation in the expression of exosome biogenesis genes.

Furthermore, the researchers verified that these KD cells exhibited increased production of exosomes, with the mutant form of ß-catenin concurrently decreasing the expression of SDC4 and Rab27a. Intriguingly, applying a GSK inhibitor to the cells resulted in reduced expression of SDC4 and Rab27a. Subsequent findings indicated that mutated ß-catenin actively facilitates immune escape through exosomes, and silencing exosome biogenesis correlates with a decrease in immune cell infiltration.<br /> In a crucial clinical correlation, the study demonstrated that patients with ß-catenin mutations exhibited low levels of exosome biogenesis.

Strengths:

Overall, the data robustly supports the outlined conclusions, and the study is commendably designed and executed. However, there are a few suggestions for manuscript improvement.

Weaknesses:<br /> No weaknesses were identified by this reviewer.

Reviewer #3 (Public Review):

Summary:

In this very important study by Dantzer et al., 'Emerging role of oncogenic b-catenin in exosome biogenesis as a driver of immune escape in hepatocellular carcinoma' the authors define a role for oncogenic b-catenin on exosome biology and explore the link between reduce exosome secretion and tumor immune cell evasion. Using transcriptional and proteomic analysis of hepatocellular carcinoma cells with either oncogenic or wildtype b-catenin the authors find that oncogenic b-catenin negatively regulates exosome biogenesis.

The authors can provide compelling evidence that oncogenic b-catenin in different hepatocellular carcinoma cells negatively regulates exosome biogenesis and secretion, by downregulation of, amongst others, SDC4 and RAB27A, two proteins involved in exosome biogenesis. The authors corroborate these results by inducing b-catenin activation using CHIR99021 in a hepatocarcinoma cell line with non-oncogenic bCatenin (Huh7 cells). The authors can further demonstrate convincingly that a reduction in exosome release by hepatocarcinoma spheroids leads to a reduction in immune cell infiltration into the tumor spheroid.

Strengths:

This is a very important and well-conceived study, that appeals to a readership beyond the field of hepatocarcinoma. The authors demonstrate a compelling link between oncogenic bCatenin and exosome biogenesis. Their results are convincing and with well-designed control experiments. The authors included various complementary lines of investigation to verify their findings.

Weaknesses:

One limitation of this study is that the mechanistic relationship of exosome release and how they affect immune cells remains to be elucidated. In this context, the authors conclusions rest on the assumption that hepatocarcinoma immune evasion is based exclusively on the reduced number of exosomes. However, the authors do not analyze exosome composition between exosomes of wild type and oncogenic background, which could be different.

eLife assessment

Using unbiased transcriptional profiling, the study reports a fundamental discovery of a novel hepatic lncRNA, FincoR, which regulates FXR. The convincing findings have therapeutic implications in the treatment of MASH. The authors use state-of-the-art methodology and use unbiased transcriptomic profiling and epigenetic profiling, including validation in mouse models and human samples.

Author Response

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

eLife assessment:

The authors report a novel hepatic lncRNA FincoR regulating FXR with therapeutic implications in the treatment of MASH. The findings are important and use an appropriate methodology in line with the current state-of-the-art, with convincing support for the claims.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In the article titled "Hammerhead-type FXR agonists induce an eRNA FincoR that ameliorates nonalcoholic steatohepatitis in mice," the authors explore the role of the Farnesoid X Receptor (FXR) in treating metabolic disorders like NASH. They identify a new liver-specific long non-coding RNA (lncRNA), FincoR, regulated by FXR, notably induced by agonists such as tropifexor. The study shows that FincoR plays a significant role in enhancing the efficacy of tropifexor in mitigating liver fibrosis and inflammation associated with NASH, suggesting its potential as a novel therapeutic target. The study makes a promising contribution to understanding the role of FincoR in alleviating liver fibrosis in NASH, providing initial insights into the mechanisms involved. While it offers a valuable starting point, there is potential for further exploration into the functional roles of FincoR and their specific actions in human NASH cases. Building upon the current findings to elucidate more detailed mechanistic pathways through which FincoR exerts its therapeutic effects in liver disease would elevate the research's significance and potential impact in the field.

Strengths:

This study stands out for its comprehensive and unbiased approach to investigating the role of FincoR, a liver-specific lncRNA, in the treatment of NASH. Key strengths include: 1) The application of advanced sequencing methods like GRO-seq and RNA-seq offered a comprehensive and unbiased view of the transcriptional changes induced by tropifexor, particularly highlighting the role of FincoR. 2) Utilizing a genetic mouse model of FXR KO and a FincoR liver-specific knockdown (FincoR-LKD) mouse model provided a controlled and relevant environment for studying NASH, allowing for precise assessment of tropifexor's therapeutic effects. 3) The inclusion of tropifexor, an FDAapproved FXR agonist, adds significant clinical relevance to the study. It bridges the gap between experimental research and potential therapeutic application, providing a direct pathway for translating these findings into real-world clinical benefits for NASH patients. 4) The study's rigorous experimental design, incorporating both negative and positive controls, ensured that the results were specifically attributable to the action of FincoR and tropifexor.

Weaknesses:

The study presents several notable weaknesses that could be addressed to strengthen its findings and conclusions: 1) The authors focus on FincoR, but do not extensively test other lncRNAs identified in Figure 1A. A more comprehensive approach, such as rescue experiments with these lncRNAs, would provide a better understanding of whether similar roles are played by other lncRNAs in mitigating NASH. 2) FincoR was chosen for further study primarily because it is the most upregulated lncRNA induced by GW4064. Including another GW4064-induced lncRNA as a control in functional studies would strengthen the argument for FincoR's unique role in NASH. 3) The study does not conclusively demonstrate whether FincoR is specifically expressed in hepatocytes or other liver cell types. Conducting FincoR RNA-FISH with immunofluorescent experiments or RT-PCR, using markers for different liver cell types, would clarify its expression profile. 4) Understanding the absolute copy number of FincoR is crucial. Determining whether there are sufficient copies of FincoR to function as proposed would lend more credibility to its suggested role. 5) The manuscript, although technically proficient, does not thoroughly address the relevance of these findings to human NASH. Questions like the conservation of FincoR in humans and its potential role in human NASH should be discussed.

Reviewer #2 (Public Review):

Summary:

Nonalcoholic fatty liver disease (NASH), recently renamed as metabolic dysfunctionassociated steatohepatitis (MASH) is a leading cause of liver-related death. Farnesoid X receptor (FXR) is a promising drug target for treating NASH and several drugs targeting FXR are under clinical investigation for their efficacy in treating NASH. The authors intended to address whether FXR mediates its hepatic protective effects through the regulation of lncRNAs, which would provide novel insights into the pharmacological targeting of FXR for NASH treatment. The authors went from an unbiased transcriptomics profiling to identify a novel enhancer-derived lncRNA FincoR enriched in the liver and showed that the knockdown of FincoR in a murine NASH model attenuated part of the effect of tropifexor, an FXR agonist, namely inflammation and fibrosis, but not steatosis. This study provides a framework for how one can investigate the role of noncoding genes in pharmacological intervention targeting known protein-coding genes. Given that many disease-associated genetic variants are located in the non-coding regions, this study, together with others, may provide useful information for improved and individualized treatment for metabolic disorders.

Strengths:

The study leverages both transcriptional profile and epigenetic signatures to identify the top candidate eRNA for further study. The subsequent biochemical characterization of FincoR using FXR-KO mice combined with Gro-seq and Luciferase reporter assays convincingly demonstrates this eRNA as a FXR transcriptional target sensitive to FXR agonists. The use of in vitro culture cells and the in vivo mouse model of NASH provide multi-level evaluation of the context-dependent importance of the FincoR downstream of FXR in the regulation of functions related to liver dysfunction.

Weaknesses:

As discussed, future work to dissect the mechanisms by which FincoR facilitates the action of FXR and its agonists is warranted. It would be helpful if the authors could base this on the current understanding of eRNA modes of action and the observed biochemical features of FincoR to speculate potential molecular mechanisms explaining the observed functional phenotype. It is unclear if this eRNA is conserved in humans in any way, which will provide relevance to human disease. Additionally, the eRNA knockdown was achieved by deletion of an upstream region of the eRNA transcription. A more direct approach to alter eRNA levels, e.g., overexpression of FincoR in the liver would provide important data to interpret its functional regulation.

We thank the Editor and Reviewers for their constructive comments. We believe we have addressed all of the issues (detailed below) and the revisions have greatly strengthened the manuscript.

Reviewer 1:

The study presents several notable weaknesses that could be addressed to strengthen its findings and conclusions:

(1) The authors focus on FincoR, but do not extensively test other lncRNAs identified in Figure 1A. A more comprehensive approach, such as rescue experiments with these lncRNAs, would provide a better understanding of whether similar roles are played by other lncRNAs in mitigating NASH.

(2) FincoR was chosen for further study primarily because it is the most upregulated lncRNA induced by GW4064. Including another GW4064-induced lncRNA as a control in functional studies would strengthen the argument for FincoR's unique role in NASH.

(3) The study does not conclusively demonstrate whether FincoR is specifically expressed in hepatocytes or other liver cell types. Conducting FincoR RNA-FISH with immunofluorescent experiments or RT-PCR, using markers for different liver cell types, would clarify its expression profile.

(4) Understanding the absolute copy number of FincoR is crucial. Determining whether there are sufficient copies of FincoR to function as proposed would lend more credibility to its suggested role.

Response to 1 - 4): We thank Reviewer 1 for the positive comments on the strength of our work, including the open-ended approach, the novel eRNA FincoR and its strong relevance to liver disease. We also value the constructive feedback provided by the reviewer and agree that additional studies are important to fully understand the mechanisms of FincoR and the functional significance of other FXR-induced lncRNAs. In this manuscript we report the discovery and initial characterization of FincoR, as well as its potential function in FXR action in response to hammerhead agonists, but a number of interesting questions are raised. Future experiments, as suggested by reviewer, will be needed to examine the role of other FXR-induced lncRNAs, the potential role of FincoR induction by other nuclear receptors with binding sites at FincoR, whether FincoR is expressed in liver cell types in addition to hepatocytes, and the expression abundance of FincoR. These are all excellent suggestions for future experimentation which we feel are beyond the scope of the present report. For example, generating a genetic CRISPR/Cas9 of another lncRNA is not trial as it takes a significant amount of work with murine models. Also, we did not mean to exclude if other lncRNAs induced by FXR also bear functions. Technically, rescue experiment is not possible as FincoR RNA can be potentially very long (~10 kb if estimated by RNA-seq pattern in Fig.1C), and it is not feasible now to properly express it by exogenous vectors to ensure the expression levels are similar to endogenous ones. We therefore consider that these important questions are more suitable for future work to fully address. Our belief is that a comprehensive exploration of FXR-regulated lncRNAs holds the potential to unveil novel insights crucial for the development of therapies targeting NASH and other metabolic diseases. The study of FincoR is the beginning of this area of research.

(5) The manuscript, although technically proficient, does not thoroughly address the relevance of these findings to human NASH. Questions like the conservation of FincoR in humans and its potential role in human NASH should be discussed.

Response: These are important questions. To respond to the reviewer’s comment, new experiments are presented in our final revised manuscript in which we utilized mouse models of NAFLD/NASH and cholestatic liver injury to determine FincoR’s role in these diseases. Hepatic FincoR levels were significantly increased in mice fed with high fat diet (HFD) for 12 weeks (Supplementary Figure S1A) and in mice fed a HFD with high fructose (HFHF) in drinking water for 12 weeks (Supplementary Figure S1B). Elevated hepatic FincoR levels were also observed in mice treated with α-naphthylisothiocyanate (ANIT), a chemical inducer of liver cholestasis (Supplementary Figure S1C), and in mice with bile duct ligation (BDL), a surgical method to induce cholestatic liver injury (Supplementary Figure S1D).

In terms of the human relevance, we have provided additional information and figures showing that there is sequence similarity between mouse FincoR and a human loci. FincoR sequence is moderately conserved between mice and humans as displayed in the UCSC genome browser (Supplementary Figure S1E). Annotation of these conserved human sequences revealed that they overlap with a functionally uncharacterized human lncRNA XR_007061585.1 (Supplementary Figure S1F). Further, we conducted qRT-PCR experiment from human patient’s RNA samples, which demonstrated that hepatic lncRNA XR_007061585.1 levels are elevated in patients with NAFLD and PBC, but not in severe NASH-fibrosis patients (Supplementary Figure S1G, H). These results demonstrate that hepatic levels of a potential human analog of FincoR are elevated in NAFLD and PBC patients, which is consistent with FincoR’s upregulation in mouse models of chronic liver disease with hepatic inflammation and liver injury. Whether human lncRNA XR_007061585.1 is entirely analogous to mouse FincoR in terms of functions and mechanisms, and whether the elevation of this human lncRNA has a role in liver disease progression or is an adaptive response to liver injury remains to be determined.

Reviewer #2 (Recommendations For The Authors):

(1) In the introduction Line 96, "..., while the vast majority are transcribed into ncRNAs" may not be accurate. Please refer to Pointing and Haerty Annu Rev 2022 for a related discussion.

Response: We would like to thank the reviewer for pointing out this inaccurate information in the introduction. We have changed the content in the text, “While a significant portion of the genome was initially thought to be "junk DNA", it has been established that many non-coding regions give rise to functional non-coding RNAs.”

(2) Figure 5: the authors should provide a clear illustration demonstrating the sequence targeted by the sgRNA in relation to the transcriptional and epigenetic profile (i.e., RNAseq and H3K27ac ChIP-seq data).

Response: The illustration (Figure 5-figure supplement 1A, right panel) demonstrating the sequence targeted by the sgRNA has been updated as suggested by the reviewer.

In this model, the upstream of FincoR is deleted, leading to the inhibition of FincoR transcription. Does the deleted region include FXR binding sites? If so, would the phenotype be due to the deletion of these binding sequences, rather than the decreased FincoR transcripts? Accordingly, the limitation or alternative interpretation should be discussed.

Response: The reviewer made a good point. The deleted region includes FXR binding sites so that we cannot rule out decreased binding of FXR or decreased transcription of the region per se, in addition to the decreased levels of FincoR, to bear a role in the phenotypic changes we observed. In the final revision, we have added discussion of this alternative (6th paragraph in the revised discussion section).

(3) Figure 6C, the images should be accompanied by quantification. It appears the FincoR-KD shows a visible difference as compared to Tropifexor-treated control mice, which does not match entirely what is written in the results.

Response: The quantitation of Oil Red O staining has been done as suggested by the reviewer (Figure 6C). The result is consistent with the triglyceride result showing that tropifexor treatment markedly reduced neutral lipids determined by Oil Red O staining of liver sections (Figure 6C) and liver TG levels (Figure 6D) and these beneficial effects on reducing fatty liver were not altered by FincoR.

(4) Figure 7, does AST show the same pattern as ALT? As indicated from Line 335, "tropifexor treatment reduced mRNA levels of several genes that promote fibrosis (Col1a1, Col1a2, ...)". Fig. 7D does not seem to match the description of Col1a1. Authors may need to modify the results.

Response: AST has been measured and has the same pattern as ALT. The new data have been added to Figure 7B. Col1a1 expression has been re-measured and the results have been updated in Figure 7D.

(5) Is FincoR level reduced in NASH conditions?

Response: We thank the Reviewer for this question. We now added new data to examine the levels of FincoR in mouse liver disease models and also examined levels of a potential human analog of FincoR in human liver specimens from PBC, NAFLD, and NASH patients. Please see our new data and description above in the response to comment 5 by Reviewer 1 (most data now included in the new Supplementary Figure S1).

(6) Please provide information on the conservation of FincoR (DNA and RNA) in humans. This would be important to provide the human disease relevance.

Response: As described above in the response to comment 5 of reviewer 1, a human loci shows sequence similarity to mouse FincoR and this conserved region has an annotated uncharacterized human lncRNA. We also examined the levels of this human homolog in human diseased liver samples. Our new results demonstrate that hepatic levels of a potential human analog of FincoR are elevated in NAFLD and PBC patients, which is consistent with FincoR’s upregulation in mouse models of chronic liver disease with hepatic inflammation and liver injury. Whether human lncRNA XR_007061585.1 is entirely analogous to mouse FincoR in terms of functions and mechanisms, and whether the elevation of this human lncRNA has a role in liver disease progression or is an adaptive response to liver injury remains to be determined.

(7) Several discussion points for the authors' consideration:

(7.1) human-mouse conservation as alluded to in #6;

Response: Potential human-mouse conservation is discussed with new data in the last paragraph of the Results section.

(7.2) potential molecular mechanism involved in FincoR-regulated hepatocyte function;

Response: We thank Reviewer for this comment. We have added more discussion as shown below: “RNA inside the cells usually associates with different RNA-binding proteins (RBPs). To predict those potential binding proteins of FincoR. Additional bioinformatic analysis identified proteins that potentially binding FincoR, including KHDRBS1, RBM38, YBX2 and YBX3 (Supplemental Table S5). These findings and potential functions of the binding proteins are discussed in the 5th paragraph of the discussion section in the final revised manuscript. Whether these predicted RBPs interact with FincoR and the underlying mechanisms will need to be investigated in future experimentation to understand the mechanisms involved in FincoR-regulated hepatocyte function.”

(7.3) any disease-associated SNPs in the FincoR locus.

Response: No SNPs were noted in the annotation of the human loci with sequence similarity to mouse FincoR in the NCBI genome data viewer.

(7.4) the in vitro induction of FincoR is transient but in vivo this occurs after 12 days of drug treatment. How do the authors reconcile the differential induction patterns?

Response: To clarify, the induction of FincoR after a single dose of GW4064 in vivo was transient, peaked within 1 h and then declined gradually (Figure 1-figure Supplement 1C). In the tropifexor treatment protocol (also in vivo), the mice were treated daily with tropifexor for 12 days so that the multiple doses maintained FincoR induction. The beneficial effect of tropifexor by inducing FincoR, therefore, accumulated over the 12 days.

It is worthy to note that we failed to see induction of FincoR in isolated primary mouse hepatocytes treated with GW4064 in vitro. We can only detect FincoR in primary hepatocytes isolated from GW4064-treated mice liver. This may be due to the loss of key factors mediating FincoR induction in the cultured primary hepatocytes.

Reviewer #1 (Public Review):

Summary:

In their article, the authors delve into the therapeutic potential of a newly identified liver-specific lncRNA, FincoR, regulated by the Farnesoid X Receptor (FXR) and induced by the agonist tropifexor, in treating nonalcoholic steatohepatitis (NASH). They demonstrate that FincoR significantly enhances tropifexor's effectiveness in reducing liver fibrosis and inflammation in NASH, presenting it as a promising therapeutic target. The manuscript revisions broaden the study to include both mouse and human data, showing elevated FincoR levels in various mouse models of liver disease and identifying a similar lncRNA in humans, potentially indicating a conserved therapeutic mechanism. This research offers valuable insights into FincoR's role in NASH and suggests further exploration into its functions and mechanisms in liver disease treatment.

Strengths:

This study enhances our understanding of FincoR, a liver-specific lncRNA, and its therapeutic potential in treating NASH through a multifaceted research approach. The revised manuscript further strengthens this contribution by incorporating additional experiments and human relevance, summarized as follows: 1) The use of GRO-seq and RNA-seq technologies has provided an in-depth and unbiased view of the transcriptional alterations driven by the FXR agonist tropifexor, especially emphasizing FincoR's pivotal role. 2) The research expands on the original findings by including diverse mouse models of NAFLD/NASH and cholestatic liver injury. These models demonstrate significant increases in hepatic FincoR levels across various conditions, such as diets high in fat and fructose, chemical induction of liver cholestasis with ANIT, and surgical induction via bile duct ligation. This broadened scope underscores FincoR's involvement in liver disease mechanisms beyond the initial models of FXR knockout (KO) and FincoR liver-specific knockdown (FincoR-LKD). 3) Incorporation of tropifexor, an FDA-approved FXR agonist, alongside these experimental models bridges experimental findings to potential therapeutic applications for NASH patients. 2) The manuscript revision includes promising data on the sequence similarity between mouse FincoR and a human locus, identifying a partially conserved human lncRNA (XR_007061585.1) with elevated levels in NAFLD and PBC patients. This addition enhances the study's relevance to human health. 3) The study's design, with the inclusion of both negative and positive controls and now enriched with a wider array of mouse models and human data, ensures that the observed therapeutic effects can be confidently attributed to FincoR's modulation by tropifexor.

Weaknesses:

The authors acknowledge that certain questions remain unanswered within the scope of this study on FincoR, due to feasibility and technical challenges. While it's important to note that such limitations are rooted in the practical and technical complexities, these unresolved issues might limit the study's immediate impact. The decision to focus on the discovery and initial characterization of FincoR, is strategically but not scientifically justified.

Reviewer #2 (Public Review):

Summary:

Nonalcoholic fatty liver disease (NASH), recently renamed as metabolic dysfunction-associated steatohepatitis (MASH) is a leading cause of liver-related death. Farnesoid X receptor (FXR) is a promising drug target for treating NASH and several drugs targeting FXR is under clinical investigation for its efficacy in treating NASH. The authors intended to address whether FXR mediates its hepatic protective effects through regulation of lncRNAs, which would provide novel insights into the pharmacological targeting of FXR for NASH treatment. The authors went from an unbiased transcriptomics profiling to identify a novel enhancer-derived lncRNA FincoR enriched in the liver and showed that the knockdown of FincoR in a murine NASH model attenuated part of the effect of tropifexor, an FXR agonist, namely inflammation and fibrosis, but not steatosis. This study provides a framework how one can investigate the role of noncoding genes in pharmacological intervention targeting a known protein coding genes. Given that many disease-associated genetic variants are located in the non-coding regions, this study, together with others, may provide useful information for improved and individualized treatment for metabolic disorders.

Strengths:

The study leverages both transcriptional profile and epigenetic signatures to identify the top candidate eRNA for further study. The subsequent biochemical characterization of FincoR using FXR-KO mice combined with Gro-seq and Luciferase reporter assays convincingly demonstrates this eRNA as a FXR transcriptional targets sensitive to FXR agonists. The use of in vitro culture cells and the in vivo mouse model of NASH provide multi-level evaluation of the context-dependent importance of the FincoR downstream of FXR in regulation of functions related to liver dysfunction.

Weaknesses:

Future work to dissect the detailed mechanisms by which FincoR facilitates action of FXR and its agonists is warranted. A more direct approach to alter eRNA levels, e.g., overexpression of FincoR in the liver would provide important data to interpret its functional regulation.

Author Response

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

Response to Public Reviewer Comments

We again thank the reviewers for the time and effort they clearly put into reviewing our manuscript. We have revised our manuscript to take into account the majority of their suggestions, primary among them being refinements of our model and classification approach, detailed sensitivity analysis of our model, and several new simulations. Their very constructive feedback has resulted in what we feel is a much-improved paper. In what follows, we respond to each of their points.

Reviewer #1:

COMMENT: The reviewer suggested that our control policy classification thresholds should be increased, especially if the behavioral labels are to be subsequently used to guide analyses of neural data which “is messy enough, but having trials being incorrectly labeled will make it even messier when trying to quantify differences in neural processing between strategies.”

REPLY: We appreciate the observation and agree with the suggestion. In the revised manuscript, we simplified the model (as another reviewer suggested), which allowed for better training of the classifier. This enabled an increase in the threshold to 95% to have more confidence in the identified control strategies. Figures 7 and 8 were regenerated based on the new threshold.

COMMENT: The reviewer asked if we could discuss what one might expect to observe neurally under the different control policies, and also suggested that an extension of this work could be to explore perturbation trials, which might further distinguish between the two control policies.

REPLY: It is indeed interesting to speculate what neural activity could underlie these different behavioral signatures. As this task is novel to the field, it is difficult to predict what we might observe once we examine neural activity through the lens of these control regimes. We hope this will be the topic of future studies, and one aspect worthy of investigation is how neural activity prior to the start of the movement may reflect two different control objectives. Previous work has shown that motor cortex is highly active and specific as monkeys prepare for a cued movement and that this preparatory activity can take place without an imposed delay period (Ames et al., 2014; Cisek & Kalaska, 2005; Dekleva et al., 2018; Elsayed et al., 2016; Kaufman et al., 2014; Lara et al., 2018; Perich et al., 2018; Vyas et al., 2018; Zimnik & Churchland, 2021). It seems possible that the control strategies we observed correspond to different preparatory activity in the motor cortex. We added these speculations to the discussion.

The reviewer’s suggestion to introduce perturbations to probe sensory processing is very good and was also suggested by another reviewer. We therefore conducted additional simulations in which we introduced perturbations (Supplementary Material; Figure S10). Indeed, in these model simulations the two control objectives separated more. However, testing these predictions via experiments must await future work.

COMMENT: “It seems like a mix of lambda values are presented in Figure 5 and beyond. There needs to be some sort of analysis to verify that all strategies were equally used across lambda levels. Otherwise, apparent differences between control strategies may simply reflect changes in the difficulty of the task. It would also be useful to know if there were any trends across time?”

REPLY: We appreciate and agree with the reviewer’s suggestion. We have added a complementary analysis of control objectives with respect to task difficulty, presented in the Supplementary Material (Figures S7 and S8). We demonstrate that, overall, the control objectives remain generally consistent throughout trials and difficulty levels. Therefore, it can be concluded that the difference in behavior associated with different control objectives does not depend on the trial sequence or difficulty of the task. A statement to this extent was added to the main text.

COMMENT: “Figure 2 highlights key features of performance as a function of task difficulty. …However, there is a curious difference in hand/cursor Gain for Monkey J. Any insight as to the basis for this difference?”

REPLY: The apparently different behavior of Monkey J in the hand/cursor RMS ratio could be due to subject-to-subject variability. Given that we have data from only two monkey subjects, we examined inter-individual variations between human subjects in the Supplementary Material by presenting individual hand/cursor gain data for all individual human subjects (Figure S1). As can be seen, there was indeed variability, with some subjects not exhibiting the same clear trend with task difficulty. However, on average, the RMS ratio shows a slight decrease as trials grow more difficult, as was earlier shown in Figure 2. We added a sentence about the possibility of inter-individual variations to address the difference in behavior of monkey J with reference to the supplementary material.

Reviewer #2:

(Reviewer #2's original review is with the first version of the Reviewed Preprint. Below is the authors' summary of those comments.)

COMMENT: The reviewer commends the care and effort taken to characterize control policies that may be used to perform the CST, via dual human and monkey experiments and model simulations, noting the importance of doing so as a precursor to future neural recordings or BMI experiments. But the reviewer also wondered if it is all that surprising that different subjects might choose different strategies: “... it makes sense that different subjects might choose to favor different objectives, and also that they can do so when instructed. But has this taught us something about motor control or simply that there is a natural ambiguity built into the task?”

REPLY: The redundancy in the task that allowed different solutions to achieve the task was deliberate, and the motivation for choosing this task for this study. We therefore did not regard the resulting subject-to-subject variability as a finding of our study. Rather, redundancy and inter-individual variability are features ubiquitous in all everyday actions and we explicitly wanted to examine behavior that is closer to such behavior. As commended by the reviewers, CST is a rich task that extends our research beyond the conventional highly-constrained reaching task. The goal of our study was to develop a computational account to identify and classify such differences to better leverage future neural analyses of such more complex behaviors. This choice of task has now been better motivated in the Introduction of the revised manuscript.

COMMENT: The reviewer asks about our premise that subjects may use different control objectives in different trials, and whether instead a single policy may be a more parsimonious account for the different behavioral patterns in the data, given noise and instability in the system. In support of this view, the reviewer implemented a simple fixed controller and shared their own simulations to demonstrate its ability to generate different behavioral patterns simply by changing the gain of the controller. The reviewer concludes that our data “are potentially compatible with any of these interpretations, depending on which control-style model one prefers.”

REPLY: We first address the reviewer’s concern that a simple “fixed” controller can account for the two types of behavioral patterns observed in Experiment 2 (instructed groups) by a small change in the control gain. We note that our controller is also fixed in terms of the plant, the actuator, and the sensory feedback loop; the only change we explore is in the relative weights of position vs. velocity in the Q matrix. This determines whether it is deviations in position or in velocity that predominate in the cost function. This, in turn, generates changes in the gain vector L in our model, since the optimal solution (i.e. the gains L that minimize the cost function) depends on the Q matrix as well as the dynamics of the plant (specifically, the lambda value). Hence, one could interpret the differences arising from changes in the control objective (the Q matrix) as changes in the gains of our “fixed” controller.

More importantly, while the noise and instability in the system may indeed occasionally result in distinct behavioral patterns (and we have observed such cases in our simulations as well), these factors are far from giving an alternative account for the structural differences in the behavior that we attribute to the control objective. To substantiate this point, we performed additional simulations that are provided in the Supplementary Material (Figures S4—6). These simulations show that neither a change in noise nor in the relative cost of effort can account for the two distinct types of behavior. These differences are more consistently attributed to a change in the control objective.

In addition, our approach provides a normative account of the control gains needed to simulate the observed data, as well as the control objectives that underlie those gains. As such, the two control policies in our model (Position and Velocity Control) resulted in control gains that captured the differences in the experimental groups (Experiment 2), both at the single trial and aggregate levels and across different task difficulties. Figure S9 in the Supplementary Material shows how the control gains differ between Position and Velocity Control in our model across different difficulty levels.

We agree,with the reviewer’s overall point, that there are no doubt many models that can exhibit the variability observed in our experimental data, our simulations, or the reviewer’s simulations. Our study aimed to explore in detail not only the model’s ability to generate the variable behavior observed in experimental data, but also to match experimental results in terms of performance levels, gains, lags and correlations across a wide range of lambda values, wherein the only changes in the model were the lambda value and the control objective. Without the details of the reviewer’s model, we are unable to perform a detailed analysis of that model. Even so, we are not claiming that our model is the ‘ground truth,’ only that it is certainly a reasonable model, adopted from the literature, that provides intuitive and normative explanation about the performance of humans and monkeys over a range of metrics, system dynamics, and experimental conditions.

Finally, we understand the reviewer’s concern regarding whether the trial-by-trial identification of control strategy in Figure 8 suggests that (uninstructed) subjects constantly switch control objectives between Position and Velocity. Although it is not unreasonable to imagine that individuals would intuitively try different strategies between ‘keeping the cursor still’ and ‘keeping the cursor at the center’ across trials, we agree that it is generally difficult to determine such trial-to-trial changes, especially when the behavior lies somewhere in between the two control objectives. In such cases, as we originally discussed in the manuscript, an alternative explanation could be a mixed control objective that generates behavior at the intersection of Position and Velocity Control, i.e., between the two slopes in Figure 8. We believe, however, that our modeling approach is still helpful in cases where performance is predominantly based on Position or Velocity Control. After all, the motivation for this study was to parse neural data into two classes associated with each control objective to potentially better identify structure underlying these behaviors.

We clarified these points in the main text by adding further explanation in the Discussion section.

COMMENT: The reviewer suggested additional experiments, such as perturbation trials, that might be useful to further explore the separability of control objectives. They also suggested that we temper our conclusion that our approach can reliably discriminate amongst different control policies on individual trials. Finally, the reviewer suggested that we modify our Introduction and/or Discussion to note past human/monkey research as well as investigations of minimization of velocity-error versus position-error in the smooth pursuit system.

REPLY: We have expanded our simulations to investigate the effects of perturbation on the separability of different control objectives (Figure S10 in Supplementary Materials). We demonstrated that introducing perturbations more clearly differentiated between Position and Velocity Control. These results provide a good basis for further experimental verifications of the control objectives, but we defer these for future work.

We also appreciate the additional past work that bridges human and monkey research that the reviewer highlights, including the related discussions in the eye movement literature on position versus velocity control. We have modified our Introduction and Discussion accordingly.

Reviewer #3:

COMMENT: The reviewer asked whether the observed differences in behavior might be due to some other factors besides the control policy, such as motor noise or effort cost, and suggested that we more systematically ruled out that possibility.

REPLY: We appreciate and have heeded the reviewer’s suggestion. The revised manuscript now includes additional simulations in which the control objective was fixed to either Position or Velocity Control, while other parameters were systematically varied. Specifically, we examined the influence of the relative effort cost, the sensory delay, and motor noise, on performance. The results of these sensitivity analyses are presented in the Supplementary Material, Figures S4—6. In brief, we found that changing the relative effort cost, delay, or noise levels, mainly affected the success rate in performance (as expected), but did not affect the behavioral features originally associated with control objectives. We include a statement about this result in the main text with reference to the details provided in the Supplementary Material.

COMMENT: The reviewer questioned our choice of classification features (RMS position and velocity) and wondered if other features might yield better class separation, such as the hand/cursor gain. In a similar vein, reviewer 2 suggested in their recommendations that we examine the width of the autocorrelation function as a potentially better feature.

REPLY: We note first that our choice of cursor velocity and position stems from a dynamical systems perspective, where position-velocity phase-space analysis is common. However, we also explored other features as suggested. We found that they, too, exhibited overlap between the two different control objectives, and did not provide any significant improvement in classification performance (Figures S2 and S3; Supplementary Materials). Of course, that is not to say that a more exhaustive examination of features may not find ones that yield better classification performance than those we investigated, but that is beyond the scope of our study. We refer to this consideration of alternative metrics in the discussion.

COMMENT: The reviewer notes that “It seems that the classification problem cannot be solved perfectly, at least on a single-trial level.” To address this point, the reviewer suggested that we conduct additional simulations under the two different control objectives, and quantify the misclassifications.

REPLY: We appreciate the reviewer’s suggestion, and have conducted the additional simulations as suggested, the results of which are included in the revised manuscript.

COMMENT: “The problem of inferring the control objective is framed as a dichotomy between position control and velocity control. In reality, however, it may be a continuum of possible objectives, based on the relative cost for position and velocity. How would the problem differ if the cost function is framed as estimating a parameter, rather than as a classification problem?”

REPLY: A blended control strategy, formulated as a cost function that is a weighted combination of position and velocity costs, is indeed a possibility that we briefly discussed in the original manuscript. This possibility arises particularly for individuals whose performance metrics lie somewhere between the purely Position or purely Velocity Control. While our model allows for a weighted cost function, which we will explore in future work, we felt in this initial study that it was important to first identify the behavioral features unique to each control objective.

Response to Recommendations for the Authors:

Reviewer #1 (Recommendations For The Authors):

None beyond those stated above.

Reviewer #2 (Recommendations For The Authors):

COMMENT: Line 166 states "According to equation (1), this behavior was equivalent to reducing the sum (𝑝 + 𝑥) when 𝜆 increased, so as to prevent rapid changes in cursor velocity". This doesn't seem right. In equation 1, velocity (not acceleration) depends on p+x. So a large p+x doesn't create a "rapid change in cursor velocity", but rather a rapid change in cursor position.

REPLY: The reviewer is correct and we have corrected this misworded sentence; thank you for catching that.

COMMENT: The reviewer points out the potential confusion readers may have, given our unclear use of ‘control strategy’ vs. ‘control policy’ vs. ‘control objective’. The reviewer suggests that “It would be helpful if this could be spelled out early and explicitly. 'Control strategy' seems perilously close to 'control policy', and it would be good to avoid that confusion. The authors might prefer to use the term 'cost function', which is really what is meant. Or they might prefer 'control objective', a term that they introduce as synonymous with 'control strategy'.”

REPLY: We thank the reviewer for noting this ambiguity. We have clarified the language in the Introduction to explicitly note that by strategy, we mean the objective or cost function that subjects attempt to optimize. We then use ‘control objective’ consistently and removed the term ‘policy’ from the paper to avoid confusion. We also now use Position Control and Velocity Control as the labels for our two control objectives.

COMMENT: The reviewer notes that in Figure 2B and the accompanying text in the manuscript, we need to be clearer about what is being correlated; namely, cursor and hand position.

REPLY: Thank you for pointing out this lack of clarity, which we have corrected as suggested.

COMMENT: The reviewer questions our attribution of decreasing lag with task difficulty as a consequence of subjects becoming more attentive/responsive when the task is harder, and points out that our model doesn’t include this possible influence yet the model reproduces the change in lag. The reviewer suggests that a more likely cause is due to phase lead in velocity compared to position, with velocity likely increasing with task difficulty, resulting in a phase advance in the response.

REPLY: Our attribution of the decrease in lag with task difficulty being due to attention/motivation was a recapitulation of this point made in the paper by Quick et al. [2018]. But as noted by the reviewer, this potential influence on lag is not included in our model. Accordingly, the change in lag is more likely a reflection of the phase response of the closed loop system, which does change with task difficulty since the optimal gains depend upon the plant dynamics (i.e., the value of lambda). We have, therefore, deleted the text in question.

COMMENT: “The Methods tell us rather a lot about the dynamics of the actual system, and the cost functions are also well defined. However, how they got from the cost function to the controller is not described. I was also a bit confused about the controller itself. Is the 50 ms delay assumed when deriving the controller or only when simulating it (the text seems to imply the latter, which might make sense given that it is hard to derive optimal controllers with a hard delay)? How similar (or dissimilar) are the controllers for the two objectives? Is the control policy (the matrix that multiplies state to get u) quite different, or only subtly?”

REPLY: Thanks for pointing this out. For brevity, we had omitted the details and referred readers to the original paper (Todorov, 2005). However, we now revised the manuscript to now include all the details in the Methods section. Hence, the entire section on the model is new. This also necessitated updating all data figures (Figures 3, 4, 5, 6, 7, 8) as they contain modeling results.

COMMENT: “Along similar lines, I had some minor to moderate confusions regarding the OFC model as described in the main text. Fig 3 shows a model with a state estimator, but it isn't explained how this works. …Here it isn't clear whether there is sensory noise, or a delay. The methods say a delay was included in the simulation (but perhaps not when deriving the controller?). Noise appears to have been added to u, but I'm guessing not to x or x'? The figure legend indicates that sensory feedback contains only some state variables, and that state estimation is used to estimate the rest. Presumably this uses a Kalman filter? Does it also use efference copy, as would be typical? My apologies if this was stated somewhere and I missed it. Either way, it would be good to add a bit more detail to the figure and/or figure legend.”

REPLY: As the lack of detail evidently led to some confusion, we now more clearly spell out the details of the model in the Methods, including the state estimation procedure.

COMMENT: The reviewer wondered why we chose to plot mean velocity vs. mean position as in Figure 5, noting that, “ignoring scale, all scatter plots would be identical if the vertical axis were final position (because mean velocity determines final position). So what this plot is really examining is the correlation between final position and average position. Under position control, the autocorrelation of position is short, and thus final position tends to have little to do with average position. Under velocity control, the autocorrelation of position is long, and thus final position tends to agree with average position. Given this, why not just analyze this in terms of the autocorrelation of position? This is expected to be much broader under velocity control (where they are not corrected) than under position control (where they are, and thus disappear or reverse quickly). To me, thinking of the result in terms of autocorrelation is more natural.”

REPLY: The reviewer is correct that the scatter plots in Fig. 5 would be the same (to within a scale factor of the vertical axis) had we plotted final position vs. mean position instead of mean velocity vs. mean position as we did. Our preference for mean velocity vs. mean position stems from a dynamical systems perspective, where position-velocity phase-space analysis is common. We now mention these perspectives in the revised manuscript for the benefit of the reader.

As suggested, we also investigated the width of the (temporal) autocorrelation function (acf) of cursor position for 200 simulated position control trials and 200 simulated velocity control trials, at four different lambda values (50 simulated trials per lambda). Figs. S2A and B (Supplementary Materials) show example trials and histograms of the acf width, respectively. As the reviewer surmised, velocity control trials tend to have wider acfs than position control trials. However, as with the metrics we chose to analyze, there is overlap and there is no visible benefit for the classification.

COMMENT: “I think equation ten is incorrect, but would be correct if the identity matrix were added? Also, why is the last term of B set to 1/(Tau*M). What is M? Is it mass (which above was lowercase m)? If so, mass should also be included in A (it would be needed in two places in the last column). Or if we assume m = 1, then just ignore mass everywhere, including here and equation 5. Or perhaps I'm confused, and M is something else?”

REPLY: Thanks for pointing this out. The Matrix A shown in the paper is for the continuous-time representation of the model. However, as the reviewer correctly mentioned, for the discrete-time implementation of the model, a modification (identity matrix) was added in our simulations. We have now clarified this in the Methods section of the revised manuscript. Also, as correctly pointed out, M is the mass of the hand, which depending on whether the hand acceleration (d^2 p/dt^2) or hand force (F) are taken as the state, it can be included in the A matrix. In our case, the A matrix is modified according to the state vector. Similarly, the B matrix is also modified. This is now clarified in the Methods section of the manuscript.

Reviewer #3 (Recommendations For The Authors):

COMMENT: “Equations 4-8 are written in continuous time, but Equation 9 is written in discrete time. Then Equation 10 is in discrete time. This needs to be tidied up. … I would suggest being more detailed and systematic, perhaps formulating the control problem in continuous time and then converting to discrete time.”

REPLY: Thank you for this helpful suggestion. The model section in the Methods has been expanded to provide further details of the equation of motion, the discretization process, the control law calculation and the state estimation process.

COMMENT: “It seems slightly odd for the observation to include only position and velocity of the cursor. Presumably participants can also observe the state of their own hand through proprioception (even if it were occluded). How would it affect the model predictions if the other states were observable?”

REPLY: Thanks for pointing this out. We initially included only cursor position and velocity since we felt that was the most prominent state feedback, and the system is observable in that case. Nevertheless, we revised the manuscript and repeated all simulations using a full observability matrix. Our findings and conclusions remain unchanged. With the changes in the modeling, the figures were also updated (Fig.3, 4, 5, 6, 7, 8).

COMMENT: “It seems unnecessary to include the acceleration of the cursor in the formulation of the model. …the acceleration is not even part of the observed state according to line 668… I think the model could therefore be simplified by omitting cursor acceleration from the state vector.”

REPLY: We agree. We have simplified the model, and generated new simulations and figures. Our results and conclusions were unchanged by this modification. With the changes in the modeling, the figures were also updated (Fig.3, 4, 5, 6, 7, 8).

COMMENT: “In the cost function, it's not clear why any states other than position and velocity of the cursor need to have non-zero values. …The choice to have the cost coefficient for these other states be 1 is completely arbitrary… If the point is that the contribution of these other costs should be negligible, then why not just set them to 0?”

REPLY: We agree, and have made this change in the Methods section. Our findings and conclusions were unaffected.

COMMENT: “It seems that the cost matrices were specified after transforming to discrete-time. It is possible however (and perhaps recommended) to formulate in continuous time and convert to discrete time. This can be done cleanly and quite straightforwardly using matrix exponentials. Depending on the discretization timestep, this can also naturally lead to non-zero costs for other states in the discrete-time formulation even if they were zero under continuous time. … A similar comment applies to discretization of the noise.”

REPLY: Thanks for the suggestion. We have expanded on the discretization process in our Methods section, which uses a common approximation of the matrix exponentiation method.

COMMENT: “Most of the parameters of the model seem to be chosen arbitrarily. I think this is okay as the point is to illustrate that the kinds of behaviors observed are within the scope of the model. However, it would be helpful to provide some rationale as to how the parameters were chosen. e.g. Were they taken directly from prior literature, or were they hand-tuned to approximately match observed behavior?”

REPLY: We have revised the manuscript to more clearly note that the noise parameters, as well as parameters of the mechanical system (mass, muscle force, time scale, etc) in our model were taken from previous publications (Todorov, 2005, Cluff et al. 2019). As described in the manuscript, the parameter values of the cost function (Q matrix) were obtained by tuning the parameters to achieve a similar range of success rate with the model as observed in the experimental data. This is now clarified in the Methods section.

COMMENT: “The ‘true’ cost function for this task is actually a 'well' in position space - zero cost within the screen and very high cost elsewhere. In principle, it might be possible to derive the optimal control policy for this more veridical cost function. It would be interesting to consider whether or not this model might reproduce the observed behaviors.”

REPLY: This is indeed a very interesting suggestion, but difficult to implement based on the current optimal feedback control framework. However, this is interesting to consider in future work.

Minor Comments:

COMMENT: “In Figs 4 and 5, the data points are drawn from different conditions with varying values of lambda. How did the structure of this data depend on lambda? Might it be possible to illustrate in the figure (e.g. the shade/color of each dot) what the difficulty was for each trial?”

REPLY: We performed additional analyses to show the effects of task difficulty on the choice of control objective. Overall, we found that the main behavioral characteristics of the control objective remained fairly unchanged across different task difficulties or across time. The results of this analysis are included in Fig. S7 and S8 of the Supplementary Materials.

COMMENT: “Should mention trial duration (6s) in the main narrative of the intro/results.”

REPLY: We now mention this detail when we describe the task for the first time.

COMMENT: “As an alternative to training on synthetic data (which might not match behavior that precisely, and was also presumably fitted to subject data at some level) it might be worth considering to do a cross-validation analysis, i.e. train the classifier on subsets of the data with one participant removed each time, and classify on the held-out participant.”

REPLY: This is indeed a valid point. The main reason to train the classifier based on model simulations was two-fold: first, to have confidence in the training data, as the experimental data was limited and noisy, which would result in less reliable classifications; and second, the model simulations are available for different contexts and conditions, where experimental data is not necessarily available. The latter is a more practical reason to be able to identify control objectives for any subject (who received no instructions), without having to collect training data from matching control subjects who received explicit instructions. Nonetheless, we appreciate the reviewer’s recommendation and will consider that for our future studies.

COMMENT: “line 690 - Presumably the optimal policy was calculated without factoring in any delay (this would be tricky to do), but the 50ms delay was incorporated at the time of simulation?”

REPLY: The discretization of the system equations allowed us to incorporate the delay in the system dynamics and solve for the optimal controller with the delay present. This was done simply by system augmentation (e.g., Crevecoeur et al., 2019), where the states of the system in the current time-step were augmented with the states from the 5 preceding time-steps to form the new state vector x(t)_aug =[x(t) , x(t-1) , … , x(t-d) ]. Similarly, the matrices A, B, and H from the system dynamics could be expanded accordingly to form the new dynamical system:

$$x(t+1){aug} = A{aug} * x(t){aug} + B{aug} * u$$

Then, the optimal control was implemented on the new (augmented) system dynamics.

We have revised the manuscript (Methods) to clarify this issue.

eLife assessment

This study represents a step towards integrating human and non-human primate research towards a broader understanding of the neural control of motor strategies. It could offer valuable insights into how humans and non-human primates (Rhesus monkeys) manage visuomotor tasks, such as stabilizing an unstable virtual system, potentially leading to discoveries in neural behaviour mechanisms. While the evidence is mostly solid, some results, particularly from the binary classification of control strategies for non instructed behaviour, require further validation before it could be conclusively interpreted.

Author Response

Reviewer #1 (Public Review):

Summary:

The present study's main aim is to investigate the mechanism of how VirR controls the magnitude of MEV release in Mtb. The authors used various techniques, including genetics, transcriptomics, proteomics, and ultrastructural and biochemical methods. Several observations were made to link VirR-mediated vesiculogenesis with PG metabolism, lipid metabolism, and cell wall permeability. Finally, the authors presented evidence of a direct physical interaction of VirR with the LCP proteins involved in linking PG with AG, providing clues that VirR might act as a scaffold for LCP proteins and remodel the cell wall of Mtb. Since the Mtb cell wall provides a formidable anatomical barrier for the entry of antibiotics, targeting VirR might weaken the permeability of the pathogen along with the stimulation of the immune system due to enhanced vesiculogenesis. Therefore, VirR could be an excellent drug target. Overall, the study is an essential area of TB biology.

Strengths:

The authors have done a commendable job of comprehensively examining the phenotypes associated with the VirR mutant using various techniques. Application of Cryo-EM technology confirmed increased thickness and altered arrangement of CM-L1 layer. The authors also confirmed that increased vesicle release in the mutant was not due to cell lysis, which contrasts with studies in other bacterial species.

Another strength of the manuscript is that biochemical experiments show altered permeability and PG turnover in the mutant, which fits with later experiments where authors provide evidence of a direct physical interaction of VirR with LCP proteins.

Transcriptomics and proteomics data were helpful in making connections with lipid metabolism, which the authors confirmed by analyzing the lipids and metabolites of the mutant.

Lastly, using three approaches, the authors confirm that VirR interacts with LCP proteins in Mtb via the LytR_C terminal domain.

Altogether, the work is comprehensive, experiments are designed well, and conclusions are made based on the data generated after verification using multiple complementary approaches.

Weaknesses:

The major weakness is that the mechanism of VirR-mediated EV release remains enigmatic. Most of the findings are observational and only associate enhanced vesiculogenesis observed in the VirR mutant with cell wall permeability and PG metabolism. The authors suggest that EV release occurs during cell division when PG is most fragile. However, this has yet to be tested in the manuscript - the AFM of the VirR mutant, which produces thicker PG with more pore density, displays enhanced vesiculogenesis. No evidence was presented to show that the PG of the mutant is fragile, and there are differences in cell division to explain increased vesiculogenesis. These observations, counterintuitive to the authors' hypothesis, need detailed experimental verification.

Response: We thank the reviewer for this comments. We would like to convince this reviewer about the fact that the VirR mutant is truly caring a more fragile PG. We will perfume additional experiments that would support this notion. We will determine the degree of PG release to the extracellular space and run additional mass spectrometry data on isolated PG.

Transcriptomic data only adds a little substantial. Transcriptomic data do not correlate with the proteomics data. It remains unclear how VirR deregulates transcription. TLCs of lipids are not quantitative. For example, the TLC image of PDIM is poor; quantitative estimation needs metabolic labeling of lipids with radioactive precursors. Further, change in PDIMs is likely to affect other lipids (SL-1, PAT/DAT) that share a common precursor (propionyl- CoA).

Response: We agree with the reviewer that TLC analysis is not quantitative. Additional TLCs will be run to investigate other lipids sharing common precursors. At the present time, we can not run radioactive experiments on the lab.

The connection of cholesterol with cell wall permeability is tenuous. Cholesterol will serve as a carbon source and contribute to the biosynthesis of methyl-branched lipids such as PDIM, SL-1, and PAD/DAT. Carbon sources also affect other aspects of physiology (redox, respiration, ATP), which can directly affect permeability and import/export of drugs. Authors should investigate whether restoration of the normal level of permeability and EV release is not due to the maintenance of cell wall lipid balance upon cholesterol exposure of the VirR mutant.

Response: We concur with the reviewer that cholesterol as sole carbon source is introducing many changes in Mtb cells beside permeability. Our central hypothesis regarding this data is that cholesterol will make Mtb cell membrane less fluid and this fact will make Ev release to be reduced. We will try to measure membrane fluidity in the presence and absence of cholesterol. However, permeability changes in Mtb cells can be manifested at different levels of the cell envelope. This would suggest that the increased permeability observed in the VirR mutant, could be different than that of observed upon TRZ treatment. The main point on this is that vesiculogenesis could be a general process responding to changes in permeability regardless of the cell envelope compartment affected. We need to define experiments here, but we will try to demonstrate this.

Finally, protein interaction data is based on experiments done once without statistical analysis. If the interaction between VirR and LCP protein is expected on the mycobacterial membrane, how the SPLIT_GFP system expressed in the cytoplasm is physiologically relevant. No explanation was provided as to why VirR interacts with the truncated version of LCP proteins and not with the full-length proteins.

Response: Split-GFP has been previously used with cell membrane proteins with success. However, we will repeat the experiments and perform statistics.

Reviewer #2 (Public Review):

Summary:

In this work, Vivian Salgueiro et al. have comprehensively investigated the role of VirR in the vesicle production process in Mtb using state-of-the-art omics, imaging, and several biochemical assays. From the present study, authors have drawn a positive correlation between cell membrane permeability and vasculogenesis and implicated VirR in affecting membrane permeability, thereby impacting vasculogenesis.

Strengths:

The authors have discovered a critical factor (i.e. membrane permeability) that affects vesicle production and release in Mycobacteria, which can broadly be applied to other bacteria and may be of significant interest to other scientists in the field. Through omics and multiple targeted assays such as targeted metabolomics, PG isolation, analysis of Diaminopimelic acid and glycosyl composition of the cell wall, and, importantly, molecular interactions with PG-AG ligating canonical LCP proteins, the authors have established that VirR is a central scaffold at the cell envelope remodelling process which is critical for MEV production.

Response: We thank the reviewer for this kind words.

Weaknesses:

Throughout the study, the authors have utilized a CRISPR knockout of VirR. VirR is a non-essential gene for the growth of Mtb; a null mutant of VirR would have been a better choice for the study.

Response: We thank the reviewer for bringing up this issue. Contrary to predictions, we believe that virR is an essential gene as we have tried to delete it several times with no success. We used in the study a transposon mutant and its complementing strain since they have been the base of previous studies to establish their genetic implications in vesiculogenesis in Mtb. The choice of CRISPRi was run similar experiments in a background different from transposon mutagenesis. Our data, support similar phenotypes in term of vesicle release.