Author Response

Reviewer #1 (Public Review):

The manuscript by Curtis et al. reports the interaction between CaMKII and alpha-actinin-2. The authors found that the interaction was elevated after NMDA receptor activation in dendritic spines. In addition, this study reveals NMDA receptor binding to CaMKII facilitates alpha-actinin-2 access to the CaMKII regulatory segment, indicating that the NMDA receptor is involved in this interaction. The authors identified the EF1-4 motifs mediated this interaction, and overexpression of this motif inhibited structural LTP. Moreover, biochemical measurements of affinities from various combination of protein fragments including autoinhibited CaMKII 1-315, regulatory segments of CaMKII, and the EFhand motif reveals that autoinhibited CaMKII has limited access to alpha-actinin-2. The authors also solved the structure of the interaction, supporting their finding in neurons at the molecular level. The authors claim that the interaction between CaMKII and alpha-actinin-2 is essential for structural LTP through cooperative action by the NMDA receptor and actin cytoskeleton.

Overall, the experiments are well-designed and the results are largely convincing and well-interpreted. But some aspects of the experiments need to be clarified.

1) Time resolution of the interaction analysis appears to be poor, as calcium elevation in a dendritic spine would be at milli-second order. What is the time window to interact alpha-actinin-2 with CaMKII during NMDA receptor activation or LTP?

We have performed additional time-course experiments to determine how quickly interactions between alpha-actinin-2 and CaMKII are elevated following NMDAR activation. The results of these experiments are shown in Figure 2A and Figure 2-Figure Supplement 1. We found that the change in association was established rapidly after NMDAR activation (t50% = 22±1 s, Figure 2A), which is consistent with proposed time-courses for CaMKII interactions following the induction of LTP (see Yasuda, Hayashi & Hell, Nat Rev Neuroscience, 2022, PMID 36056211). We have included additional text in the results (lines 138-147), methods (lines 609-611 & 650-652), and discussion (lines 426-427) sections explaining these experiments, and figure legends are provided for the new figures on lines 10061009 and lines 1096-1101.

2) The authors analyzed the binding of CaMKII and alpha-actinin-2 with partial fragments. It remains to be unknown whether CaMKII can form a protein complex with GluN2B and alpha-actinin-2 in a single CaMKII protomer.

The reviewer is referring to experiments shown in figure 5, in which we found that a fragment of GluN2B (1260-1492) increases pull-down of full-length CaMKIIa with a fusion of GST to the EF3-4 region of a-actinin-2. This region of GluN2B contains a CaMKII phosphorylation sequence (positions 1290-1309) that occupies the substrate binding groove of the kinase domain (Stratton et al., Cell Reports, 2023, PMID 35830796). Therefore, the most logical explanation for the results of the pulldown experiment is that GluN2B increases a-actinin-2 access to the regulatory segment by binding to the substrate binding groove of the same CaMKII protomer. Nevertheless, we discuss the difficulty of conceptualising and investigating interactions between oligomeric proteins within the PSD on lines 451461.

3) Besides synaptic localization, the effect of the interaction on the enzymatic activity of CaMKII is not known.

The Colbran laboratory has previously examined the effect of a-actinin-2 on CaMKII activity. Jalan-Sakrikar and colleagues (JBC, 2012, PMID 22427672) showed that a fragment of aactinin-2 corresponding to EF hands 3 and 4 is able to weakly activate CaMKII (~ 10 % compared to Ca2+/CaM) towards peptide substrates autocamtide-2 and GluN2B but not syntide-2 (see Figure 1B&C of this paper). An earlier study by Robison and colleagues (JBC, 2005, PMID 16172120) found that aactinin-2 antagonises Ca2+/CaM-dependent activation of unphosphorylated CaMKII towards autocamtide2, but does not affect the activity of pT286 auto-activated CaMKII (see Figure 4A of this paper). This work is referred to on lines 63-65 of the introduction.

4) Although the authors quantify the effect of the EF-hand disruptor by measuring numbers of the dendritic spine by its shape, the specificity of the EF-hand disruptor needs to be clarified.

There are two known interaction partners for the EF hand region of a-actinin-2: CaMKII and Titin (Young et al., EMBO J, 1998, PMID 9501083; Atkinson et al., Nat Struct Biol, 2001, PMID 11573089). Titin is an extremely long sarcomeric protein that is expressed in striated muscle cells but not neurons. Therefore, the effects of the disruptor are highly likely to reflect disruption of interactions to CaMKII. We also performed control experiments with EF34 L854R that does not bind CaMKII effectively (Figure 3-figure supplement 1C). We have added a sentence to clarify the specificity of the EF-hand disruptor on lines 182-184, as follows: ” Furthermore, the only known interaction partner for the EF14 region of a-actinin-2 besides CaMKII is the muscle-specific protein titin (Young et al., 1998), so any effects of EF14 in neurons are likely to reflect destabilisation of native interactions between CaMKII and a-actinin-2”.

Author Response

Reviewer #1 (Public Review):

This study uses electrophysiological techniques in vitro to address the role of the Na+ leak channel NALCN in various physiological functions in cartwheel interneurons of the dorsal cochlear nucleus. Comparing wild type and glycinergic neuron-specific knockout mice for NALCN, the authors show that these channels 1) are required for spontaneous firing, 2) are modulated by noradrenaline (NA, via alpha2 receptors) and GABA (through GABAB receptors), 3) how the modulation by NA enhances IPSCs in these neurons.

This work builds on previous results from the Trussell's lab in terms of the physiology of cartwheel cells, and from other labs in terms of the role of NALCN channels, that have been characterized in more and more brain areas somewhat recently; for this reason, this study could be of interest for researchers that work in other preparations as well. The general conclusions are strongly supported by results that are clearly and elegantly presented.

I have a few comments that, in my opinion, might help clarify some aspects of the manuscript.

1) It is mentioned throughout the manuscript, including the abstract, that the results suggest a closed apposition of NALCN channels and alpha2 and GABAB receptors. From what I understand, this conclusion comes from the fact that GABAB receptors activate GIRK channels through a membrane-delimited mechanism. Is it possible that these receptors converge on other effectors, for example adenylate cyclase (see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6374141/).

It will be of interest to test the roles of adenylyl cyclase modulation in the control of NALCN, as a complement to the studies we have presented here.

2) In Figure 2G, the neurons from NALCN KO mice appear to reach a significantly higher frequency than those from WT (figure 2E, 110 vs. 70 spikes/s). Was this higher frequency a feature of all experiments? The results mention a rundown of peak firing rate due to whole-cell dialysis, but, from what I understand, the control conditions should be similar for all experiments.

The peak firing rates in control solutions for WT and KO CWC are not statistically different.

3) Also in Figure 2, the firing patterns for neurons from WT and NALCN KO mice appear to be quite different, with spikes appearing to be generated during the hyperpolarization of the bursts in the second half of the current step for WT neurons but always during the depolarization in KO neurons. Was this always the case? If so, could NALCN channels be involved in this type of firing? Along these lines, it would be interesting to show an example of a firing pattern of neurons from WT mice in the presence of NA, which inhibits NALCN channels.

The specific pattern of spikes in CWC is quite variable from trial-to-trial or cell-to-cell, as it is dependent on multiple CaV and calcium dependent K channels subtypes, and is not dependent on the genotypes used here. The primary effects observed in the KO are in background firing and sensitivity to NA, both reflected alterations in rheobase. The firing pattern example requested was shown in the raster plot of fig 2B2.

4) It might be interesting to discuss how the hyperpolarization induced by the activation of GIRK channels and inhibition of NALCN channels could have different consequences due to their opposite effect on the input resistance.

We considered this as a point of discussion, but decided that making sense of it would depend on assumptions about the location of the channels (dendritic vs somatic, distance to AIS) that we do not have data for. For example, a dendritic increase in resistance through NALCN block, leading to a hyperpolarization of the soma, might have actions similar to a somatic hyperpolarizing conductance increase by GIRK, as far as the voltage at the AIS is concerned.

Reviewer #3 (Public Review):

The study by Ngodup and colleagues describes the contribution of sodium leak NALCN conductance on the effects of noradrenaline on cartwheel interneurons of the DCN. The manuscript is very well-written and the experiments are well-controlled. The scope of the study is of high biological relevance and recapitulates a primary finding of the Khaliq lab (Philippart et al., eLife, 2018) in ventral midbrain dopamine neurons, that Gi/o-coupled receptors inhibit NALCN current to reduce neuronal excitability. Together these studies provide unequivocable evidence for NALCN as a downstream target of these receptors. There are no major concerns. I have only minor suggestions:

Minor

1) As introduced in the introduction, NALCN is inhibited by extracellular calcium which has led to some discourse of the relevance of NALCN when recorded in 0.1 mM calcium. A strength of this study is the effect of NA on NALCN is recorded in physiological levels of calcium (1.2 mM). I suggest including the concentration of extracellular calcium in the aCSF in the Results section instead of relying on the reader to look to the Methods.

Will do.

2) It would be interesting to include the basal membrane properties of the KO compared to wildtype, including membrane resistance and resting membrane potential. From the example recording in Figure 2, one might think that the KOs have lower membrane resistance, so it is interesting that the 2 mV hyperpolarization produced similar effects on rheobase. In addition, from the example in Figure 2G, it appears that NA has an effect on firing frequency with large current injection in the KO. Is this true in grouped data and if so, is there any speculation into how this occurs?

Will do.

3) Please expand on the rationale for why GABAB and alpha2 must be physically close to NALCN. To my knowledge, the mechanism by which these receptors inhibit NALCN is not known. Must it be membrane-delimited?

Given the known membrane delimited modulation of GIRK by GABAB, and that alpha2 and GABAB receptors appear to share the same population of NALCN channels, and that alpha2 receptors do not appear to target GIRK channels, we felt the simplest explanation would be coupling through G-proteins, with spatial segregation of different receptor/channel pools providing the means for separating GIRK and NALCN effects. However, the involvement of an additional second messenger is testable.

Author Response

We wish to thank the Reviewers for the appreciation they have expressed for our work, and the constructive feedback that they offered. We agree that clarifying the interpretation of synergy and information decomposition in the context of macroscale BOLD signals and loss of consciousness will be a valuable addition to the manuscript, and so will be improving the quality of our figures, and we will endeavour to do so. Briefly, at this stage we just wish to clarify that it is not our intention to claim that Phi-R and synergy, as measured at the level of regional BOLD signals, represent a direct cause of consciousness, or are identical to it. Rather, our work is intended to use these measures similarly to the use of sample entropy and LZC for BOLD signals: as theoretically grounded macroscale indicators, whose empirical relationship to consciousness may reveal the relevant underlying phenomena. We will ensure that our updated manuscript reflects this additional nuance.

Author Response

We thank the reviewers for their very thorough and detailed comments as well as the overall positive reception of the work. Additionally, the reviewers provided excellent detailed suggestions for future work.

Specific response to Reviewer 1:

“Indeed, the major disappointment of this work is the clinical relevance that was highlighted in the Introduction but was never really studied in the end. iPSC from patients could be added to the study.”

We completely agree that it would be very exciting to use patient-derived iPSC in the platform that we describe in this manuscript. We recognize that extensive work to characterize and validate BMECS differentiated from patient-derived iPSCs would be required, including validating BBB-like properties, before retinol transport data could be collected and interpreted. This work is beyond the scope of the current manuscript. We hope that in the future the in vitro model we describe in this manuscript will be used for exactly this type of clinically relevant application.

Specific response to Reviewer 2:

“1) The authors assume that there is a significant fraction of free ROL, 20% for ROH/RBP and 7% for RBP/TTR complexes (summarized in Table 1). This implies that at the physiological concentration of ROH/RBP in the plasma of 2 uM, free ROL represents 0.4 uM. However, the concentration of free ROL is limited by its poor solubility in the aqueous phase, which is around 0.06 uM (Szuts EZ, 1991, Arch Biochem Biophys). Moreover, taking into account the large concentration of other potential nonspecific carriers for lipids, it is safe to assume that there is virtually no free ROH in the plasma. There is also an important physiological reason for the limited amount of free ROL. Its rapid and nonspecific partition into cells (also observed in this study) would work against the highly specific RBP/STRA6-dependent ROH uptake pathway, undermining its physiological function.”

The reviewer raises an important point that we considered carefully during the design of the research. As the reviewer says, Szuts (1991) reported retinol (ROH) solubility of ~0.06 µM (range of 0.03 – 0.11 µM). Szuts defined ROH solubility as ‘the amount of dissolved solute in equilibrium with its solid state…includ[ing] all its dissolved forms (monomers, multimers, and micelles)’. We are using a definition of ‘free’ ROH as ‘ROH not bound to protein’; in our work ‘free’ ROH could include retinol multimers and micelles, which likely do exist under our experimental conditions. (We did not see any evidence of solid ROH.) That said, we calculate that the concentration of free ROH (ROH not bound to protein) is ~0.14 µM when both RBP and TTR are present. In more complex biological mixtures containing other ROH carriers, the concentration of unbound ROH is expected to be lower, in agreement with the reviewer.

One key point is that the free ROH concentration depends on the experimental setup, and must be correctly accounted for. For example, in some of the literature investigating STRA6-mediated uptake and signaling in vitro, purified ROH-RBP is used as the retinol source and samples do not include TTR. In such a case, the unbound ROH concentration in an equilibrated sample is anticipated to be significantly higher than the physiological concentration. Our investigation demonstrates that unbound ROH can accumulate intracellularly; thus, failure to include TTR and/or to account for the action of unbound ROH could lead to errors in mechanistic interpretation of experimental studies on retinol transport into cells or across barriers such as the BBB.

2) “However, a question remains: would the outcome of the experiment be different if the basolateral chamber contained an ROH acceptor (retinol-binding proteins) rather than Hank's balanced salt solution, to which the partition of ROL is limited by its water solubility?”

We agree with the reviewer that it would be very interesting to determine whether retinol permeability changes in the presence of RBP and/or TTR on the basolateral side. This is a logical next step and can readily be performed in the Transwell setup. We chose not to do this for this project because we wanted to compare our setup with other in vitro models (e.g., with porcine BMECs) where no retinol-binding proteins were present basolaterally.

3) “The authors claim that transthyretin (TTR) increases BMECs permeability when compared to ROH/RBP. However, the mechanistic explanation for this phenomenon remains unclear. Do the authors imply the presence of a putative TTR receptor whose signaling could affect the efflux of ROL at the basolateral side of BMECs? TTR is an ubiquitous plasma protein. The concentration of TTR is tightly regulated and maintained between 300 - 330 mg/L. Therefore, it is questionable how TTR can serve as a signaling molecule modulating retinoid homeostasis in the brain.”

We disagree with the reviewer about the TTR concentration. Per Johnson et al (Clin Chem Lab Med 2007, 45:419-426), TTR concentration varies with age, gender, inflammation and nutritional status, with typical concentrations for adults ranging from 150-450 mg/L. We were surprised at our observations that TTR enhanced ROH permeability across BMECs and that LRAT expression increased in the presence of TTR. We do not currently have a mechanistic interpretation and agree with the reviewer that further exploration of these tantalizing observations is warranted.

“Additional technical issues that could affect the experimental outcomes: The formation of the ROH/RBP-TTR complex should be confirmed and purified using gel filtration to separate free TTR and ROH/RBP. Only fractions containing the complex should be used in the experiments. Assuming that the complex is formed with 100% efficiency is overly optimistic.”

We respectfully disagree with the reviewer regarding using gel filtration to isolate TTR/ROH/RBP complexes. Any such isolated complexes will fairly rapidly re-equilibrate so that some protein and some ROH is unbound. It is important to note that we do not assume that the complex is formed with 100% efficiency. In fact, on the contrary, we explicitly take into account the distribution of materials (free TTR, free RBP, free ROH, RBP-ROH, TTR-RBP-ROH) in any sample; values are reported in the manuscript. This issue is also relevant to the first point raised by the reviewer. We routinely validated binding of ROH to RBP by FRET and ROH-RBP to TTR by fluorescence anisotropy.

“Reloading RBP with isotopically labeled ROH requires an additional purification step. Stripping ROL from the ROH/RBP complex with organic solvent (diethyl ether) is appropriate but relatively harsh, causing partial unfolding of a fraction of RBP. Therefore, assuming that 100% of stripped RBP remains functional and can be reloaded with ROH is inaccurate. Reloading apo-RBP with a stoichiometric amount of ROH without an additional purification step (e.g., ion exchanger) leads to an excess of free ROL and/or its nonspecific association with nonfunctional RBP fractions. Measuring absorbance at 330 nm is not sufficient proof of binding since free ROH also absorbs at the same wavelength.”

We produced RBP by refolding of guanidine-denatured RBP in an excess of ROH to ensure near 100% ROH loading. High quality refolded RBP can qualitatively be determined by examination of the A330/280 absorbance ratio, which should be ~1.0. We then extract ROH to completion by diethyl ether to produce pure apo-RBP (ROH-free). We utilized this diethyl-ether stripped apo-RBP stock for all future characterizations, including binding to ROH and TTR. We found our stripped apo-RBP was a suitable replacement for serum sources in every biophysical assay performed. Reloaded ROH-RBP elutes as a single peak on ion exchange chromatography, indicating the vast majority of stripped RBP is available for ROH binding. We provide detailed information about RBP characterization in Est and Murphy, Prot. Exp. Purif. (2020), to which the interested reader is referred.

Author Response

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

Comment 1: The descriptions about body weights should be matched.

Regrettably, we did not monitor the body weights throughout the study. We have now revised the description clarifying the confusions. Importantly we evaluated the weights of the muscle (EDL and soleus) and heart tissues in 8-month-old mice (Fig. 1A).

Comment 2: Quantitative data for figures.

As stated in the manuscript, the presented images are representatives of at least three mice per genotype. However, assessing specific measurements such as cell sizes, diameters, or mitochondria sizes in histological tissue sections and electron microscopical fields is not feasible due to practical limitations. Unfortunately, we do not have access to specialized software for such analyses. While semi-quantification of Western blot bands is possible, implementing this for all Western blots in the manuscript would result in a substantial increase in the number of bar graphics. Below are Western blots from additional two pairs of mice used in all figures.

Comment 3: Confusions about “total mitochondrial content”.

The mitochondria content in cells was assessed by quantitatively comparing the DNA level of the mitochondrial gene cytochrome B to that of the nuclear gene 18S using quantitative PCR. This method is commonly used to determine the relative number of mitochondria in cells. However, we have revised and provided a clearer description in the figure legend to avoid any potential confusion.

Comment 4: Suggestions on further analyses of PGC1-alpha and TFAM. LC3-I and -II.

We evaluated LC3-I/II levels in PTPMT1 knockout muscles, and our findings did not indicate any signs of increased autophagic activity (Supplementary Figure S3). We will examine PGC1-alph and TFAM levels in our future studies. It is worth noting that in our previous RNA-seq analyses of PTPMT1 knockout hematopoietic cells, we did not observe any significant alterations in the expression levels of these two genes.

Comment 5: Description on fibrotic lesions.

Quantifying fibrotic areas poses a significant challenge. Therefore, we were only able to describe this finding.

Comment 6: Fig 6 is not well organized and aligned.

In response to your suggestion, we have reorganized this figure accordingly. Panels C, D, and E display mitochondrial OCR data derived from three biological replicates/genotype. We feel that these changes are sufficient to demonstrate the differences in substrate utilization between PTPMT1 knockout and control mitochondria.

Comment 7: Descriptions on glucose oxidation and glycolysis in different types of muscle fibers are confusing

We have followed the suggestions and revised the descriptions accordingly.

Comment 8: A discussion about lactate utilization in cardiomyocytes would be helpful.

Following this suggestion, we have now added a brief discussion.

Comment 9: “Cropped” images were used in Fig 10.

The images shown in Fig. 10 were not cropped images. In order to efficiently use the tissue and mitochondrial lysates, the Western blot membranes were intentionally cut into smaller fragments based on the molecular weights of the proteins to be detected. These smaller membrane sections were then employed for individual Western blotting purposes.

Minor comment 1: The order of Fig 1 panels should be reorganized.

Following this suggestion, we have now reorganized this figure.

Minor comment 2: Suggestion for an Echocardiograph result table.

These analyses were carried out by trained personnel at the Emory Animal Physiology Core. The data presented in our manuscript was provided by them. It is important to note that no additional parameters were measured beyond the data provided by the Core.

Minor comment 3: Is ROS production increased in PTPMT1 knockout muscle cells?

Yes, PTPMT1 knockout tissues showed elevated overall cellular ROS levels even at 3 months (Figure 6I).

Minor comment 4: Typo in S10 legend.

The typo has been corrected.

---

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

Comment 1: The effects of PTPMT1 on the skeletal muscle and heart might be an embryonic defect. They might be mediated by significantly reduced mTOR signaling

We acknowledge the valid point made by this reviewer. While both CKMM-Cre and Myh6Cre express Cre during the embryonic stage, we did not observe any developmental defects in skeletal muscle-specific (PTPMT1fl/fl/CKMM-Cre) or heart-specific (PTPMT1fl/fl/Myh6-Cre) knockout mice. These knockout mice appeared indistinguishable from their WT littermates until the age of 3-4 months.

Morphologically, the skeletal muscle and heart dissected from these mice showed no abnormalities. Additionally, mitochondria isolated from these tissues did not exhibit any morphological/structural defects. Undoubtedly, the late-onset phenotypes observed in the knockout mice over time was attributed to the metabolic defects arising from the loss of PTPMT1 in the embryos. Although PTPMT1 knockout muscle cells and cardiomyocytes initially maintained energy homeostasis through enhanced fatty acid and glutamate oxidation, along with metabolic adaptations or activation of alternative energy-producing pathways in the first few months, they eventually encountered substantial energy deficits. This was attributed to the subsequent occurrence of oxidative stress and mitochondrial damage. In response to this valuable feedback, we have included a brief discussion in the manuscript's discussion section to address this point.

As mentioned in the manuscript, the late-onset phenotypes observed in our study were likely a result of subsequent damages induced by prolonged metabolic substrate shift and lipid accumulation within the cells. We agree with the reviewer that decreased mTOR activities may also contribute to these late effects, and have included a brief discussion in the discussion section.

Comment 2: Why are the effects of the loss of PTPMT1 similar in the skeletal muscle and heart.

The depletion of PTPMT1 yields similar effects in both tissue types; however, the manifestations occur earlier in the skeletal muscle. Although mitochondria in the skeletal muscle and heart have distinct preferences for energy sources, prolonged forced utilization of fatty acids caused by PTPMT1 depletion eventually leads to lipid accumulation and cellular damage (lipotoxicity) in both tissue types. This phenomenon underscores the importance of maintaining a balance in substrate utilization to prevent adverse effects on cellular health in the skeletal muscle and heart.

Comment 3: AMPK is activated in PTPMT1 knockout cardiomyocytes; this should have cardioprotective effects.

AMPK can be activated through various mechanisms. In our study, AMPK activation occurs in response to energetic stress in late-stage PTPMT1 knockout tissues that displayed significantly reduced ATP levels, aligning with its role as a bioenergetic stress sensor. It is possible that AMPK activation alone was insufficient to overcome the secondary damages induced by the prolonged metabolic switch from carbohydrate metabolism to fatty acid metabolism.

Comment 4: Knockout skeletal muscles and hearts had lipid accumulation; why were knockout mice smaller than controls? Are there any changes in white fat, core temperature or browning of fat? Rescue experiments should be considered to prove that lipid accumulation is the cause of death in the knockout mice.

We believe that the lipid accumulation observed in muscle cells and cardiomyocytes of the knockout mice does not necessarily imply that these tissue-specific knockout mice would be heavier or have increased body fat. We appreciate the suggestions regarding energy expenditure tests and rescue experiments. We will certainly consider incorporating these experiments into our future study.

As stated in the manuscript, we did not observe any morphological changes in white or brown fat tissues in the adipocyte-specific PTPMT1 knockout mice. Furthermore, we assessed body temperature and its response to a cold environment (4°C), and no differences were detected between the knockout mice and the control mice.

Comment 5: Are there sex differences in muscle and heart phenotypes in the tissue specific knockout mice?

We did not observe significant differences in phenotypes between male and female knockout mice.

Comment 6: What happens to UCP2 activity in PTPMT1 deleted cells and what is its function in mediating AMPK and/mTOR regulation.

Currently, there is a lack of direct methods available to measure UCP2 activity. The relationship between UCP2 and the regulation of AMPK and mTOR has not been extensively investigated.

Comment 7: What is the effect of PTPMT1 deletion on cardiolipin synthesis?

PTPMT1 has been implicated in both facilitating mitochondrial utilization of pyruvate and participating in the synthesis of cardiolipin. To investigate the impact of PTPMT1 knockout on cardiolipin levels, we plan to establish a mass spectrometry assay for the quantitative analysis of cardiolipin in knockout mitochondria. Completing these experiments might require a considerable amount of time. Nonetheless, we extensively addressed this point in the discussion section.

Minor concerns:

Comment 8: The title needs more specificity.

As suggested, we have revised the title to "Loss of PTPMT1 restricts mitochondrial utilization of carbohydrates and induces muscle atrophy and heart failure in tissue-specific knockout mice".

Comment 9: Heart and skeletal muscle weights in Fig 1A should be normalized against tibia length.

Unfortunately, we did not perform normalization in this study. However, we appreciate the suggestion and will incorporate it into our future studies. It is important to note that the lengths of tibias in the knockout mice were only marginally shorter.

Comment 10: Low magnification and longitudinal section of the muscle should be shown in Fig 1B and 2A.

The histological images provide supporting evidence for the conclusion, despite not being optimal in quality. We acknowledge the suggested improvements and assure you that we will integrate them into our future studies. It is crucial to emphasize that each conclusion in this study was derived from multiple experimental designs, rather than solely relying on morphological changes.

Comment 11: Fig 1F is mislabeled as 1G.

We have conducted a thorough review and can confidently confirm that the labeling is correct.

Comment 12: Fig 2F and 6B should be quantified.

As indicated in the manuscript, the images presented are representatives of at least three mice per genotype. While semi-quantification of Western blot bands is possible, implementing this for all Western blots in the manuscript would result in a substantial increase in the number of bar graphics. Below are Western blot images from additional two pairs of mice included in Fig. 2F and Fig. 6B. Furthermore, Western blot images from two additional pairs of mice in other figures are also provided below.

Author response image 1

Western blotting data from additional two pairs of mice in Fig. 2F.

Author response image 2

Western blotting data from additional two pairs of mice in Fig. 6B.

Author response image 3

Western blotting data from additional two pairs of mice in Supplementary Fig. 2G.

Author response image 4

Western blotting data from additional two pairs of mice in Supplementary Fig. 3A.

Author response image 5

Western blotting data from additional two pairs of mice in Supplementary Fig. 3C.

Author response image 6

Western blotting data from additional two pairs of mice in Supplementary Fig. 3D.

Author response image 7

Western blotting data from additional two pairs of mice in Supplementary Fig. 4F.

Author response image 8

Western blotting data from additional two pairs of mice in

Author response image 9

Western blotting data from additional two pairs of mice in Supplementary Fig. 7C.

Comment 13: Knockout mice should be placed on HFD or keto diet to test for the effects of PTPMT1 depletion.

We appreciate this thoughtful suggestion. We will certainly incorporate this suggestion into our future studies, expanding beyond the scope of the current initial report.

Comment 14: Suggestions on Fig 4A.

Please see our response to Comment 10.

Comment 15: Suggestions for improving echocardiographs.

These analyses were conducted by trained personnel at the Emory Animal Physiology Core. The data presented in our manuscript was provided by them. We appreciate bringing the issues to our attention, and we will inform them accordingly.

Comment 16: Comment on Fig 5B.

The tissues were sectioned at comparable, if not identical, levels. WT and PTPMT1 knockout heart sections look dramatically different because of the dilated myopathy observed in the knockout hearts.

Comment 17: Comment on Fig 5C.

We believe the cell death occurred predominantly in cardiomyocytes.

Author Response

We thank the reviewers for their careful reading of our manuscript and for their constructive and positive comments. We will revise the manuscript to address their key points. Here, we address the reviewer’s scepticism of sleep-learning being mediated by the episodic memory system. We agree that the reported unconscious learning of novel verbal associations during sleep may not match textbook definitions of episodic memory. However, the traditional definitions of episodic memory have long been criticized (e.g, Henke, 2010; Hannula, Minor, Slabbekoorn, 2023; Shohamy & Turk-Browne, 2013; Dew & Cabeza, 2011; Reder et al, 2009). We stand by our claim that sleep-learning was of episodic nature. Here, we provide arguments for this claim:

In the introduction and the discussion, we are reporting that we use a computational definition of episodic memory (Cohen & Eichenbaum, 1993; Henke, 2010; O’Reilly et al., 2014; O’Reilly & Rudy, 2000), and not the traditional definition of episodic memory that ties episodic memory to wakefulness and conscious awareness (Gabrieli, 1998; Moscovitch, 2008; Schacter, 1998; Squire & Dede, 2015; Tulving, 2002). Consciousness and wakefulness are no properties of episodic memory according to the computational definition of episodic memory. Instead, the core computational features of episodic memory according to the computational definition are 1) rapid learning, 2) association formation, and 3) a compositional and flexible representation of the associations in long-term memory. We designed the retrieval task in the current study to assess only the retention of sleep-formed flexibly and compositionally stored word-word associations. Reviewer 3 suggests that sound-sound associations may have been formed during sleep and may have been reactivated at test resulting in the translation of the sound pattern of the translation word to the meaning of the translation word and further to the correct superordinate semantic category of the translation word. Although these processing steps during sleep and during the wake retrieval are possible, the rapid sound-sound associative encoding, long-term storage, and the flexible sound retrieval would still require hippocampal processing and hence computations in the episodic memory system. The interpretation in terms of associative auditory learning with a double semantic translation at wake testing is laborious and inefficient and hence a less parsimonious interpretation of sleep-learning than conceptual associative encoding during sleep. Our view resonates the findings by Andrillon et al. (2017) that mere auditory perceptual learning during slow-wave sleep was not possible at all or led to suppressive memory traces that could not be retrieved following awakening.

Importantly, Züst et al. (Current Biology, 2019) had also presented pseudowords and translation words for paired-associative word encoding during slow-wave sleep. Retrieval testing was performed in the waking state following sleep by use of a cued-recall task, as in the current study. During retrieval testing, Züst et al. recorded brain blood oxygenation using functional magnetic resonance imaging. Importantly, the hippocampus was activated during correctly, but not during incorrectly retrieved memories that had been formed during sleep. Crucially, activation resulting from this contrast within the posterior and anterior hippocampus and within lexical-semantic storage sites in the left temporal pole correlated between participants with retrieval performance (Züst et al., 2019). These correlation results demonstrate that those participants, who learned the vocabulary best during slow-wave sleep activated the hippocampus and lexical-semantic storage sites the most during wake retrieval testing. Because the learning and retrieval tasks in the current study were similar to Züst et al. (2019), the hippocampus was likely mediating the retrieval of the sleep-formed associations in the current study. We have also measured the brain oxygenation using functional magnetic resonance imaging in five persons while they learned pairs of pseudowords and translation words during slow-wave sleep and found the hippocampus activated (besides language areas) in all persons (unpublished).

For these reasons, we believe that vocabulary presentations during sleep had triggered a hippocampus-mediated rapid conceptual-associative encoding process that provided for flexible representations of combinations of pseudowords and translation words in episodic memory.

Author Response

We thank the reviewers for their insightful reviews of our work, including both its strengths and limitations. Below we present minor corrections to the preprint and responses to the main points brought up by each reviewer.

Erratum:

• Line 330 refers to Fig. 7F (instead of 7D).

• Line 331 refers to Fig. 7G (instead of 7E).

Reviewer #1 (Public Review):

The experimental design presented cannot clearly show that the effect of passive exposure was due to the specific exposure to task-relevant stimuli since there is no control group exposed to irrelevant stimuli.

We acknowledge the possibility that exposure to task-irrelevant stimuli could result in improvements in learning. Testing this possibility would be a worthwhile goal of future experiments, but it is outside the scope of our current study. We have been careful in our paper to only draw conclusions about the effects of exposure to task-relevant stimuli compared to no exposure. We will also add a discussion of this point and references to the literature pointed out by the reviewer to the final version of our manuscript.

The conclusion that "passive exposure influences responses to sounds not used during training" (line 147) does not seem fully supported by the authors' analysis. The authors show that there is an increase in accuracy for intermediate sweep speeds despite the fact that this is the first time the animals encounter them in the active session. However, it seems impossible to exclude that this effect is not simply due to the increased accuracy of the extreme sounds that the animals had been trained on.

The conclusion that the reviewer quotes from our paper is drawn from Figure 3, in which we show that mice exhibit an improvement on non-extreme stimuli after training on extreme stimuli. Panel 3D illustrates that the observed improvements are not just changes in psychometric performance driven by the extreme sounds. In the context of this result, the conclusion relates to generalization in performance on task-relevant stimuli that are closely related to the training stimuli. In our view, it was not entirely obvious a priori that this result would have to occur, since it is possible that performance could improve at the extremes without improving at the intermediate stimuli.

In the modelling section, the authors adjusted the hyper-parameters to maximize the difference between pure active and passive/active learning. This makes a comparison of learning rates between models somewhat confusing.

We apologize for the confusion. None of our conclusions are based on comparisons of learning speed between models, but perhaps this was not pointed out sufficiently clearly. The relevant comparisons between conditions for each specific model are made using the same hyperparameters. We will clarify this in the updated version of our manuscript.

The description of the sound does not state whether when reducing the slope of the sweeps the center or the onset frequency of the sounds is preserved.

Frequency modulated sounds of different FM slopes were generated such that the center frequency was always the same. This will be clarified in the updated version of our manuscript.

Reviewer #2 (Public Review):

One limitation here is that the presented analysis is somewhat simplistic, does not include any detailed psychometric analysis (bias, lapse rates etc), and primarily focuses on learning speed.

In our analyses of trials that included extreme and intermediate stimuli, we investigated some metrics of the type that the reviewer suggests here. However, since such additional psychometric analyses generally led to null results and would in any case be somewhat tangential to our main results, which are about learning speed and responses to sounds not included during training, we did not include these in our manuscript. A limitation of our study is that the available data does not allow for an analysis of psychometrics during the initial learning stages, since only the extreme stimuli were presented during the task.

Reviewer #3 (Public Review):

The first [major weakness] is that even Model 5 differs from their data. For example, the A+P (passive interleaved condition) learning curve in Figure 7 seems to be non-monotonic, and has some sort of complex eigenvalue in its decay to the steady state performance as trials increase. This wasn't present in their experimental data (Figure 2D), and implies a subtle but important difference. There also appear to be differences in how quickly the initial learning (during early trials) occurs for the A+P and A:P conditions. While both A+P and A:P conditions learn faster than A only in M5, A+P and A:P seem to learn in different ways, which isn't supported in their data.

The reviewer is correct that there are subtle differences between the two learning curves produced by Model 5. Due to noise in the experimental data, however, it is possible that such subtle distinctions also appear in the learning curves of the mice. Further, the slight overshoot of the learning curve that the reviewer mentions is not constrained by the experimental data due to the fact that different mice reach asymptotic performance at different times, and many of them have not even reached asymptotic performance by the end of the training period.

However, even if there are minor discrepancies between the learning curves produced by the final version of the model and by the mice, we do not see this as being especially surprising or problematic. As in any model, there are a large number of potentially important features that are not included in any of our models–for example, realistic spectrotemporal neural responses, nonlinearity in neural activations, heterogeneity across mice, and many others. The aim of our modeling was to choose a space of possible models (which is inevitably restricted) and show which model version within that space best captures our experimental observations. Expanding the space of possible models that we considered to capture further nuances in the data will be a task for future work.

The second major weakness is that the authors also don't generate any predictions with M5. Can they test this model of learning somehow in follow-up behavioural experiments in mice? ... Without follow-up experiments to test their mechanism of why passive exposure helps in a schedule-independent way, the impact of this paper will be limited.

Although testing behavioral predictions from our models was beyond the scope of the current study, we do generate specific predictions with M5 (specifically, about neural representations). Our model produces predictions about neural representations and the ways in which they evolve through learning, and we hope to test these predictions in future work.

I believe the authors need to place this work in the context of a large amount of existing literature on passive (unsupervised) and active (supervised) learning interactions. This field is broad both experimentally and computationally. For example, there is an entire sub-field of machine learning, called semi-supervised learning that is not mentioned at all in this work.

We thank the reviewer for pointing this out. The updated version of our manuscript will include a discussion on how our results fit in with this literature.

Author Response

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

First, the authors would like to thank the reviewers and editors for their thoughtful comments. The comments were used to guide our revision, which is substantially improved over our initial submission. We have addressed all comments in our responses below, through a combination of clarification, new analyses and new experimental data.

Reviewer #1 (Public Review):

In this manuscript, the authors identified and characterized the five C-terminus repeats and a 14aa acidic tail of the mouse Dux protein. They found that repeat 3&5, but not other repeats, contribute to transcriptional activation when combined with the 14aa tail. Importantly, they were able to narrow done to a 6 aa region that can distinguish "active" repeats from "inactive" repeats. Using proximal labeling proteomics, the authors identified candidate proteins that are implicated in Dux-mediated gene activation. They were able to showcase that the C-terminal repeat 3 binds to some proteins, including Smarcc1, a component of SWI/SNF (BAF) complex. In addition, by overexpressing different Dux variants, the authors characterized how repeats in different combinations, with or without the 14aa tail, contribute to Dux binding, H3K9ac, chromatin accessibility, and transcription. In general, the data is of high quality and convincing. The identification of the functionally important two C-terminal repeats and the 6 aa tail is enlightening. The work shined light on the mechanism of DUX function.

A few major comments that the authors may want to address to further improve the work:

We thank the reviewer for their efforts and constructive comments, which have guided our revisions.

1) The summary table for the Dux domain construct characteristics in Fig. 6a could be more accurate. For example, C3+14 clearly showed moderate weaker Dux binding and H3K9ac enrichment in Fig 3c and 3e. However, this is not illustrated in Fig. 6a. The authors may consider applying statistical tests to more precisely determine how the different Dux constructs contribute to DNA binding (Fig. 3c), H3K9ac enrichment (Fig. 3e), Smarcc1 binding (Fig. 5e), and ATAC-seq signal (Fig. 5f).

We thank the reviewer for this comment, and agree that there were some modest differences in construct characteristics that were not captured in the Summary Table (6a). To better reflect the differences between constructs, we added additional dynamic range to our depiction/scoring, and believe that the new scoring system provides sufficient qualitative range to capture the difference without imposing a statistical approach.

2) Another concern is that exogenous overexpressed Dux was used throughout the experiments. The authors may consider validating some of the protein-protein interactions using spontaneous or induced 2CLCs (where Dux is expressed).

We agree that it would be helpful to determine endogenous DUX interaction with our BioID candidates. Here, we attempted co-IPs for endogenous DUX protein with the DUX antibody and were unsuccessful, which indicated that the DUX antibody is useful for detection but not efficient in the primary IP. This is why we utilized the mCherry tag for DUX IP experiments, which worked exceptionally well.

3) It could be technically challenging, but the authors may consider to validate Dux and Smarcc1 interaction in a biologically more relevant context such as mouse 2-cell embryos where both proteins are expressed. Whether Smarcc1 binding will be dramatically reduced at 4-cell embryos due to loss of Dux expression?

While we agree that it would be interesting to validate the in vivo interaction of DUX and SMARCC1 in the early embryo, it is not technically feasible for us to conduct the experiment, as the IP would require thousands of two-cell embryos, and we have the issue of poor co-IP quality with the DUX antibody.

Reviewer #2 (Public Review):

In this manuscript, Smith et al. delineated novel mechanistic insights into the structure-function relationships of the C-terminal repeat domains within the mouse DUX protein. Specifically, they identified and characterised the transcriptionally active repeat domains, and narrowed down to a critical 6aa region that is required for interacting with key transcription and chromatin regulators. The authors further showed how the DUX active repeats collaborate with the C-terminal acidic tail to facilitate chromatin opening and transcriptional activation at DUX genomic targets.

Although this study attempts to provide mechanistic insights into how DUX4 works, the authors will need to perform a number of additional experiments and controls to bolster their claims, as well as provide detailed analyses and clarifications.

We thank this reviewer for their constructive comments, and have conducted several new analyses, additional experiments and clarifications – which have strengthened the manuscript in several locations. Highlights include a statistical approach to the similarity of mouse repeats to themselves and to orthologs (Figure S1d) and clarified interpretations, a wider dynamic range to better reflect changes in DUX construct behaviors (Figure 6a), and additional data on construct behavior, including ‘inactive’ constructs (e.g C1+14aa in Figure 1a,d, new ATAC-seq in Figure S1g), and active constructs such as C3+C5+14aa and C3+C514aa (in Figure S1b).

Reviewer #3 (Public Review):

Dux (or DUX4 in human) is a master transcription factor regulating early embryonic gene activation and has garnered much attention also for its involvement in reprogramming pluripotent embryonic stem cells to totipotent "2C-like" cells. The presented work starts with the recognition that DUX contains five conserved c. 100-amino acid carboxy-terminal repeats (called C1-C5) in the murine protein but not in that of other mammals (e.g. human DUX4). Using state-of-the-art techniques and cell models (BioID, Cut&Tag; rescue experiments and functional reporter assays in ESCs), the authors dissect the activity of each repeat, concluding that repeats C3 and C5 possess the strongest transactivation potential in synergy with a short C-terminal 14 AA acidic motif. In agreement with these findings, the authors find that full-length and active (C3) repeat containing Dux leads to increased chromatin accessibility and active histone mark (H3K9Ac) signals at genomic Dux binding sites. A further significant conclusion of this mutational analysis is the proposal that the weakly activating repeats C2 and C4 may function as attenuators of C3+C5-driven activity.

By next pulling down and identifying proteins bound to Dux (or its repeat-deleted derivatives) using BioID-LC/MS/MS, the authors find a significant number of interactors, notably chromatin remodellers (SMARCC1), a histone chaperone (CHAF1A/p150) and transcription factors previously (ZSCAN4D) implicated in embryonic gene activation.

The experiments are of high quality, with appropriate controls, thus providing a rich compendium of Dux interactors for future study. Indeed, a number of these (SMARCC1, SMCHD1, ZSCAN4) make biological sense, both for embryonic genome activation and for FSHD (SMCHD1).

A critical question raised by this study, however, concerns the function of the Dux repeats, apparently unique to mice. While it is possible, as the authors propose, that the weak activating C1, C2 C4 repeats may exert an attenuating function on activation (and thus may have been selected for under an "adaptationist" paradigm), it is also possible that they are simply the result of Jacobian evolutionary bricolage (tinkering) that happens to work in mice. The finding that Dux itself is not essential, in fact appears to be redundant (or cooperates with) the OBOX4 factor, in addition to the absence of these repeats in the DUX protein of all other mammals (as pointed out by the authors), might indeed argue for the second, perhaps less attractive possibility.

In summary, while the present work provides a valuable resource for future study of Dux and its interactors, it fails, however, to tell a compelling story that could link the obtained data together.

We appreciated the reviewer’s views regarding the high quality of the work and our generation of an important dataset of DUX interactors. We also appreciate the comments provided to improve the work, and have performed and included in the revised version a set of clarifications, additional analyses and additional experiments that have served to reinforce our main points and provide additional mechanistic links. We also agree that more remains to be done to understand the function and evolution of repeats C1, C2 and C4.

Reviewer #1 (Recommendations For The Authors):

1) For immuno-blots, authors may indicate the expected bands to help readers better understand the results.

Agreed, and we have included the predicted molecular weight of proteins in the Figure Legends. We note that our work shows that the C-terminal domains confer anomalous migration in SDS-PAGE.

2) Fig. 5b, a blot missing for the mCherry group?

Figure 5b is a volcano blot, so we believe the reviewer is referring to Figure 5d, which is a coimmunoprecipitation experiment between SMARCC1 and mCherry-tagged DUX constructs. However, we are unsure of the comment as an anti mCherry sample is present in that panel.

3) Line 99-100, Fig. S1d, it seems that repeat2, but not repeat3, is more similar to human DUX4 C-terminal region.

This comment and one by another reviewer have prompted us to re-examine the similarities of the DUX repeats, and we have new analyses (Figure S1d) and an alternative framing in the manuscript as a result. We have expanded on this in our response to Reviewer #2, point #1 – and direct the reviewer there for our expanded treatment.

4) There are a few references are misplaced. For example, line 48, the studies that reported the role of Dux in inducing 2CLCs should be from Hendrickson et al., 2017, De Iaco et al., 2017, and Whiddon et al., 2017. The authors may want to double check all references.

Thanks for pointing these out. These issues have been corrected in the manuscript.

5) In the materials & methods section, a few potential errors are noticed. For example, concentrations of PD0325901 and CHIR99021 in mESC medium appear ~1000-fold higher than standards.

Thanks – corrected.

Reviewer #2 (Recommendations For The Authors):

Major Points

1) Line 99 - The authors claimed that the "human DUX4 C-terminal region is most similar to the 3rd repeat of mouse DUX", but based on Supp. Fig. 1d, the human DUX4 C-term should be most similar to the 2nd repeat of mouse DUX. If this is indeed the case, it will undermine the rest of this study, since the authors claim that the 3rd repeat is transcriptionally active, whereas the 2nd repeat is transcriptionally inactive, and the bulk of this study largely focused on how the active repeats, not the inactive repeats, are critical in recruiting key transcriptional and chromatin regulators to induce the embryonic gene expression program.

We thank the reviewer for their comments here. Since submission,and as mentioned above for reviewer #1 we have revisited the issue of similarity of the DUX4 C-terminal region to the mouse C-terminal repeats, with a BLAST-based approach that is more rigorous and informed by statistics – which is in Author response table 1 and now in the manuscript as Figure S1d, and has affected our interpretation. Our prior work involved a simple % identity comparison table and we now appreciate that some of the similarity analyses did not meet statistical significance, and therefore we are unable to draw certain conclusions. We make the appropriate modifications in the text. For example, we no longer state that the DUX4 C-terminus appears to be most similar to mouse repeats 3 and 5. This does not affect the main conclusions of the paper regarding interactions of the C-terminus with chromatin-related proteins, only our speculation on which repeat might have represented the original single repeat in the mouse – an issue we think of some interest, but did not rise to the level of mentioning in the original or current abstract.

Author response table 1.

Parameters: PAM250 matrix. Gap costs of existence: 15 and extension: 3. Numbers represent e-value of each pairwise comparison

*No significant similarities found (>0.05).

2) In Supp Fig 1d, it seems that the rat DUX4 C-terminal region is most similar to the 4th repeat of mouse DUX, which according to the author is supposedly transcriptionally inactive. This weakens the authors justification that the 3rd or 5th repeat is likely the "parental repeat for the other four", and further echoes my concern in point 1 where the human DUX4 C-term is most similar to the 2nd (inactive) repeat of mouse DUX.

The reviewer’s point is well taken and is addressed in point #1 above.

3) In Fig. 1d, the authors showed that DUX4-containing C3 and C5, but lacking acidic tail, can promote MERVL::GFP expression, albeit to a slightly lower extent compared to FL. However, in Fig. 2b, C3 or C5 alone (lacking acidic tail) completely failed to promote MERVL::GFP expression. However, in the presence of the acidic tail, both versions were able to promote MERVL::GFP expression, similar to that of FL. The latter would suggest that it is the acidic tail that is crucial for MERVL::GFP expression, and this does not quite agree with Fig 1b, where C12345 (lacking acidic tail) was able to promote MERVL::GFP expression. Although C12345 did not activate MERVL to a similar level as FL, it is clearly proficient, compared to C3 or C5 alone (lacking acidic tail) where there is no increase in MERVL at all. Additional constructs will be helpful to clarify these points. For example, 'C3+C5 minus acidic tail' and 'HD1+HD2+acidic tail only' constructs.

We agree that constructs such as those mentioned would add to the work. First, we have done the additional construct HD1+HD2+14aa tail, which is presented as ΔC12345+14aa in Figure 2a and in S2a. Additionally, we performed experiments on the requested C3+C5+14aa and C3+C5Δ14aa (see samples 6 and 7 in Author response image 1, which are now included in Supplemental Figure 2b). The results reinforce our hypothesis of an additive effect toward DUX target gene activation by increasing C-terminal repeats and including the 14aa tail.

Author response image1.

4) Related to the above, the flow cytometry data for the MERVL::GFP reporter as presented in Figures 1 and 2, as well as in Supp. Fig. 2, show a considerably large difference in the %GFP|mCherry for the FL construct, ranging from ~6-26%. This makes it difficult to convince the reader which of the different DUX domain constructs cannot or can partially induce GFP|mCherry signal when compared to FL, and hence it is tough to definitively ascertain the exact contribution of each of the 5 C-terminal repeats with high confidence, as it appears that there exists a significant amount of variability in this MERVL::GFP reporter system. The authors need to address this issue since this is their primary method to elucidate the transcriptional activity of each of the mouse DUX repeat domains.

We note that with the Dux-/- cell lines we used throughout the timeline of the study, the percent of %GFP|mCherry expression progressively and slowly decreased – possibly due to slow/modest epigenetic silencing of the reporter. However, we always used the full-length DUX construct to establish the dynamic range. We emphasize that the relative differences between constructs over multiple cell line replicates remained relatively consistent. However, we elected to show absolute values in each experiment, rather than simply normalizing the full-length to 100% and showing relative.

5) Lines 140-142 - The authors claimed that the functional difference between the transcriptionally active and inactive repeats could be narrowed down to a "6aa region which is conserved between repeats C3 and C5, but not conserved in C1, C2 and C4". Assuming the 6aa sequence is DPLELF, why does C1C3a elicit almost twice the intensity of GFP|mCherry signal compared to C3C1c, despite both constructs having the exact same 6aa sequence?

Indeed, C1C3a and C3C1c both containing the ‘active’ DPL sequence but having different relative levels of %GFP|mCherry. This is consistent with these sequences having a positive role in DUX target gene regulation – but likely in combination with other other regions which potentiate its affect, possibly through interacting proteins or post-translational modifications.

Why does DPLEPL (the intermediate C3C1b construct) induce a similar extent of GFP|mCherry signal as the FL construct, even though the former includes 3aa from a transcriptionally inactive repeat? In contrast, GSLELF (the other intermediate C1C3b construct) that also includes 3aa from a transcriptionally inactive repeat is almost completely deficient in inducing any GFP|mCherry signal. Why is that so? Is DPL the most crucial sequence? It will be important to mutate these 3 (or the above 6) residues on FL DUX4 to examine if its transcriptional activity is abolished.

These are interesting points. DPL does appear to be the most important region in the mouse DUX repeats. However, DPL is not shared in the C-terminus of human DUX4. Notably, the DUX4 C-terminus is sufficient to activate the mouse MERVL::GFP reporter when cloned to mouse homeodomains (see Author response image 2, second sample) and other DUX target genes (initially published in Whiddon et al. 2017). One clear possibility is that the DPL region is helping to coordinate the additive effects of multiple DUX repeats, which only exist in the mouse protein.

Author response image 2.

6) Line 154 - The intermediate DUX domain construct C1C3b occupied a different position on the PCA plot from the C1C3c construct that does not contain any of the critical 6aa sequence, as shown in Fig. 2e. However, both these constructs appear to be similarly deficient in inducing any GFP|mCherry signal, as seen in Fig. 2c. Why is that so?

The PCA plot assesses the impact on the whole transcriptome and not just the MERVL::GFP reporter, suggesting the 3aa region has transcriptional effects on the genome beyond what is detected in the MERVL::GFP reporter.

7) To strengthen the claim that "Chromatin alterations at DUX bindings sites require a transcriptionally active DUX repeat", the authors should also perform CUT&Tag for constructs containing transcriptionally inactive DUX repeats (e.g. C1+14aa), and show that such constructs fail to occupy DUX binding sites, as well as are deficient in H3K9ac accumulation.

This is a good comment. We elected to control this with constructs containing or lacking an active repeat. Although we have not pursued this by CUT&TAG, we have examined the impact of DUX constructs with inactive repeats (including the requested C1+14aa, new Figure S1g) by ATAC-seq (see #12, ATAC-seq section, below), and observe no chromatin opening, suggesting that the lack of transcriptional activity is rooted in the inability to open chromatin.

8) It would be good if the authors could also include CUT&Tag data for some of the C1C3 chimeric constructs that were used in Fig. 2, since the authors argued that the minimal 6aa region is sufficient to activate many of the DUX target genes. This would also strengthen the authors’ case that the transcriptionally active, not inactive, repeats are critical for binding at DUX binding sites and ensuring H3K9ac occupancy.

We agree that these would be helpful, and have examined the inactive repeats in transcription and ATAC-seq formats during revision (new data in Figures 1d and S1g), but not yet the CUT&TAG format.

9) Line 213 - "SMARCA4" should have been "SMARCA5"? Based on Fig. 4d, SMARCA5 is picked up in the BirA*-DUX interactome, not SMARCA4.

Thanks – corrected.

10) Lines 250-252 - The authors compared the active BirA-C3 against the inactive BirA-C1 to elucidate the interactome of the transcriptionally active C3 repeat, as illustrated in Fig. 5c. They found 12 proteins more enriched in C1 and 154 proteins in C3. This information should be presented clearly as a separate tab in Supp Table 2. What are the proteins common to both constructs, i.e. enriched to a similar extent? Do they include chromatin remodellers too? Although the authors sought to identify differential interactors between the 2 constructs, it is also meaningful to perform 2 separate comparisons - active BirA-C3 against BirA alone control, and inactive BirA-C1 against BirA alone control - like in Fig. 4d, so as to more accurately define whether the active C3 repeat, and not the inactive C1 repeat, interacts with proteins involved in chromatin remodeling.

We thank the reviewer for this comment, and we have modified the manuscript by adding a second sheet in Supplementary Table 2 including the results for enriched proteins in BirA-C1 vs. C3. Additionally, due to limitations of annotation between BirA alone and BirA*-C3 being sequenced in different mass spectrometry experiments, it is difficult to quantitatively compare the two datasets with pairwise comparisons.

11) Fig 5d: The authors mentioned in the legend that endogenous IP was performed for SMARCC1. However, in line 266, they stated Flag-tagged SMARCC1. Is SMARCC1 overexpressed? The reciprocal IP should also be presented. More importantly, C1 constructs (e.g. C1+14aa and C1Δ14aa) should also be included.

To clarify, Figure 4e used exogenously overexpressed FLAG-SMARCC1 in HEK-293T cells to confirm the results of the full-length DUX BioID experiment. Figure 5d was performed with overexpressed DUX construct, but involved endogenous SMARCC1 in mESCs. This has now been made clearer in the revised manuscript.

12) For both the SMARCC1 CUT&Tag and ATAC-seq experiments shown in Figures 5e and 5f respectively, the authors need to include DUX derivatives that contain transcriptionally inactive repeats with and without the 14aa acidic tail, i.e. C1+14aa and C1Δ14aa, and show that these constructs prevent the binding/recruitment of SMARCC1 to DUX genomic targets, and correspondingly display a decrease in chromatin accessibility. Only then can they assert the requirement of the transcriptionally active repeat domains for proper DUX protein interaction, occupancy and target activation.

We agree that examination of an inactive repeat in certain approaches would improve the manuscript. Importantly, we have now included C1+14 in our ATAC-seq experiments, and in Author response image 3 two individual replicates, which constitute a new Figure S1g. Compared to the transcriptionally active DUX constructs, which see opening at DUX binding sites, we do not see chromatin opening at DUX binding sites with transcriptionally inactive C1+14.

Author response image 3.

13) To prove that DUX-interactors are important for embryonic gene expression, it will be important to perform loss of function studies. For instance, will the knockdown/knockout of SMARCC1 in cells expressing the active DUX repeat(s) lead to a loss of DUX target gene occupancy and activation?

We agree that it would be interesting to better understand SMARCC1 cooperation with DUX function in the embryo, but we believe this is beyond the scope of this paper.

Minor Points

1) Lines 124-126 - What is the reason/rationale for why the authors used one linker (GGGGS2) for constructs with a single internal deletion, but 2 different linkers (GGGGS2 and GAGAS2) for constructs with 2 internal deletions?

With Gibson cloning, there are homology overhang arms for each PCR amplicon that are required to be specific for each overlap. Additionally, each PCR amplicon needs to be specific enough from one another so that all inserts (up to 5 in this manuscript) are included and oriented in the right order. The linker sequences were included in the homology arm overlaps, so the nucleotide sequences for each linker needed to be specific enough to include all inserts. This is a general rule to Gibson cloning. Additionally, both GGGGS2 and GAGAS2 are common linker sequences used in molecular biology and the amino acids structures are similar to one another, suggesting there is no functional difference between linkers.

2) Line 704 - 705: In the figure legend, the authors stated that 'Constructs with a single black line have the linker GGGGS2 and constructs with two black lines have linkers with GGGGS2 and GAGAS2, respectively.'. This was not obvious in the figures.

Constructs used for flow and genomics experiments that are depicted in Figure 2, Supplementary Figure 2, Figure 3, Figure 4, and Figure 5 have depicted black lines where deletions are present. Where these deletions are present, there are linkers in order to preserve spacing and mobility for the protein.

3) Line 160 - Clusters #1 and #2 are likely written in the wrong order. It should have been "activating the majority of DUX targets in cluster #2, not cluster #1" and "failed to activate those in cluster #1, not cluster #2", based on the RNA-seq heatmap in Fig. 2f.

We thank the reviewer for this comment, and the error has been corrected in the manuscript.

4) Line 188 - Delete the word "of" in the following sentence fragment: "DUX binding sites correlating with the of transcriptional".

Thanks – corrected.

5) Line 191 - Delete the word "aids" in the following sentence fragment: "important for conferring H3K9ac aids at bound".

Thanks – corrected.

6) Line 711 - "C1-C3 a,b,d" should be "C1-C3 a,b,c".

Thanks – corrected.

7) Lines 711-712 - The colors "pink to blue" and "blue to pink" are likely written in the wrong order. Based on Fig. 2c, the blue to pink bar graphs should represent C1-C3 a,b,c in that order, and likewise the pink to blue bar graphs should represent C3-C1 a,b,c in that order.

Thanks – corrected.

8) There is an overload of data presented in Fig. 2c, such that it is difficult to follow which part of the figure represents each data segment as written in the figure legend. It is recommended that the data presented here is split into 2 sub-figures.

Figure 2c has a supporting figure in Supplementary Figure 2b. While there is both a graphical depiction of the constructions and the data both in the main panel of Figure 2C, we have depicted it as so to be as clear as possible for the reader to interpret the complexity and presentence of amino acids in each of the constructs.

9) Line 717 - "following" is misspelt.

Thanks – corrected.

10) Lines 720-721 - "(Top)" and "(Bottom)" should be replaced with "(Left)" and "(Right)", as the 2 bar graphs presented in Fig. 2d are placed side by side to each other, not on the top and bottom.

Thanks – corrected.

11) Lines 725 and 839 - "Principle" is misspelt. It should be "Principal".

Thanks – corrected.

12) In Figures 3d and 3e, the sample labeled "C3+14_1" should be re-labeled to "C3+14", in accordance with the other sub-figures. Additionally, for the sake of consistency, "aa" should be appended to the relevant constructs, e.g. "C3+14aa" and "C3Δ14aa".

Thanks – corrected.

13) Line 773 - Were the DUX domain constructs over-expressed for 12hr (as written in the figure legend) or 18hr (as labeled in Fig. 5d)?

Thanks – corrected.

14) Related to minor point 19 above, is there a reason/rationale for why some of the experiments used 12hr over-expression of DUX domain constructs (e.g. for CUT&TAG in Fig. 3), whereas in other experiments 18hr over-expression was chosen instead (e.g. flow cytometry for MERVL::GFP reporter in Figures 1 and 2, and co-IP validations of BirA*-DUX interactions in Fig. 4)?

Thanks for the opportunity to explain. In this work, experiments that reported on proteins that are translated following DUX gene activation (e.g. MERVL:GFP via flow) were done at 18hr to allow for enough time for transcription and translation of GFP (or other DUX target genes). For experiments that report on the impact of DUX on chromatin and transcription, such as RNA-seq, CUT&Tag, and ATAC-seq, we induced DUX domain constructs for 12 hours.

15) Line 804 - "ΔHDs" is missing between "C2345+14aa" and "ΔHD1".

Thanks – corrected.

16) In Fig. 5c, "Chromatin remodelers" is misspelt.

Thanks – corrected.

17) There is no reference in the manuscript to the proposed model that is presented in Fig. 6b.

Thanks – corrected.

Reviewer #3 (Recommendations For The Authors):

Given the uncertainty of the function of the Dux peptide repeats in mice, could it not also be possible that the underlying repeated nature of the (coding) DNA? That is, could these DNA repeats exert a regulatory function on Dux transcription itself (also given the dire consequences of misregulated DUX4 expression as seen in FSHD, for example).

Yes, it remains possible that the internal coding repeats within Dux are playing a role in locus regulation, and might be interesting to examine. However, we consider this question as being outside the scope of the current paper.

Finally, it would be interesting to know whether these repeats are, in fact, present in all mouse species. Already no longer present in rat, do they exist, or not, in more "distant" mice, e.g. M. caroli?

Determining whether all mouse strains contain C-terminal repeats in DUX is a question we also considered. However, Dux and its orthologs are present in long and very complex repeat arrays that are not present in the sequencing data or annotation in other mouse strains. Therefore, we are not unable to answer this question from existing sequencing data. Answering would require a considerable genome sequencing and bioinformatics effort, or alternatively a considerable effort aimed at cloning ortholog cDNAs from 2-cell embryos.

Minor points:

line 169: here it seems, in fact, that the 'inactive' C2, C4 repeats are more similar to each other (my calculation: 91 and 96% identity at the protein and DNA level, respectively) than the active C3 and C5 repeats (82 and 89% identity, resp.), the outlier being C1.

Thanks for this comment, which was mentioned by other reviewers as well and has been addressed through new statistical analyses and interpretation (see new Figure S1d).

line 191: I'm not sure this sentence parses correctly ("...14AA tail is important for conferring H3K9Ac aids at bound sites...")

We thank the reviewer for this comment, and we have corrected the sentence in the manuscript.

Author Response

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

Reviewer #1 (Public Review):

This study presents an important finding on human m6A methyltransferase complex (including METTL3, METTL14 and WTAP). The evidence supporting the claims of the authors is convincing, although the model and assays need to be further modified. The work will be of interest to biologists working on RNA epigenetics and cancer biology.

In mammals, a large methyltransferase complex (including METTL3, METTL14 and WTAP) deposits m6A across the transcriptome, and METTL3 serves as its catalytic core component. In this manuscript, the authors identified two cleaved forms of METTL3 and described the function of METTL3a (residues 239-580) in breast tumorigenesis. METTL3a mediates the assembly of METTL3-METTL14-WTAP complex, the global m6A deposition and breast cancer progression. Furthermore, the METTL3a-mTOR axis was uncovered to mediate the METTL3 cleavage, providing potential therapeutic target for breast cancer. This study is properly performed and the findings are very interesting; however, some problems with the model and assays need to be modified. It is widely known that METTL3 and METTL14 form a stable heterodimer with the stoichiometric ratio of 1:1 (Wang X et al. Nature 534, 575-578 (2016), Su S et al. Cell Res 32(11), 982994 (2022), Yan X et al. Cell Res 32(12), 1124-1127 (2022)), the numbers of METTL3 and METTL14 in the model of Fig 7P are not equivalent and need to be modified.

We thank for reviewer’s good suggestion. We have modified the model in Fig. 7P.

Reviewer #2 (Public Review):

In this study, Yan et al. report that a cleaved form of METTL3 (termed METTL3a) plays an essential role in regulating the assembly of the METTL3-METTL14-WTAP complex. Depletion of METTL3a leads to reduced m6A level on TMEM127, an mTOR repressor, and subsequently decreased breast cancer cell proliferation. Mechanistically, METTL3a is generated via 26S proteasome in an mTOR-dependent manner.

The manuscript follows a smooth, logical flow from one result to the next, and most of the results are clearly presented. Specifically, the molecular interaction assays are welldesigned. If true, this model represents a significant addition to the current understanding of m6A-methyltransferase complex formation.

A few minor issues detailed below should be addressed to make the paper even more robust. The specific comments are contained below.

1) The existence of METTL3a and METTL3b.<br /> In this study, the author found the cleaved form of METTL3 in breast cancer patient tissues and breast cancer cell lines. Is it a specific event that only occurs in breast cancer? The author may examine the METTL3a in other cell lines if it is a common rule.

We thank reviewer for point this out. We discovered the cleaved form of METTL3 in breast cancer, and we further examined this cleaved METTL3 in other cell lines such as lung cancer cell lines, renal cancer cell lines, HCT116 and MEF (new Supplementary Figures 1A-1C), these data suggest that it is a common rule. Therefore, we speculate that METTL3a may be ubiquitiously expressed. We have added this part in the revised manuscript, please see Line 118-120.

2) Generation of METTL3a and METTL3b.

1) Figure 1 shows that METTL3a and METTL3b were generated from the C-terminal of full-length METTL3. Because the sequence of METTL3a is involved in the sequences of METTL3b, can METTL3b be further cleaved to produce METTL3a?

Although the sequence of METTL3a is involved in the sequences of METTL3b, overexpression of METTL3b in T47D, MDA-MB-231 and 293T cells did not show METTL3a expression in these cells (please see Figures 3A, 3C, 3G), suggesting that METTL3b can not be further cleaved to produce METTL3a, and the METTL3 cleavage may require its N-terminal region. We have added this in the discussion, please see Line 358 to 360.

2) Based on current data, the generation of METTL3a and METTL3b are separated. Are there any factors that affect the cleavage ratio between METTL3a and METTL3b?

We thank for reviewer’s excellent question. In this study, we show that both METTL3a and METTLb are produced through proteasomal cleavage, and both of them are positively regulated by the mTOR pathway. On the other hand, we indeed observed the differential cleavage ratios between METTL3a and METTL3b across different cell lines. For example, METTL3a/METTLb ratio was greater than 1 in MDA-MB-231 cells (see Figure 7C), less than 1 in T47D and 293T cell lines (see Figure 7A and 7B), and equal to 1 in MEF cells (see Figure 7O). Based on these results, we speculate that there may be some factors that control the cleavage ratio between METTL3a and METTL3b, which warrants further investigation. We have added this in the discussion, please see Line 374 to 379.

3) In Figure 2G, the author shows the result that incubation of the Δ198+Δ238 METTL3 protein with T47D cell lysates cannot produce the METTL3a and METTL3b variants. The author may also show the results that Δ198 METTL3 protein or Δ238 METTL3 protein incubates with T47D cell lysates, respectively.

Following the reviewer’s suggestion, we had performed in vitro cleavage assays by incubation of METTL3-Δ238 or METTL3-Δ198 with T47D cell lysates, and had incorporated this result in the revised manuscript. Please see our new Supplementary Figure 3A.

4) As well as many results published in previous studies, the in vitro methylation assay shows that WT METTL3 is capable of methylating RNA probe (figure 2H). The main point of this study is that METTL3a is required for the METTL3-METTL14 assembly. However, the absence of METTL3a in the in vitro system did not inhibit METTL3METTL14 methylation activity. Moreover, the presence of METTL3a even resulted in a weak m6A level.

The main point of this study is that METTL3a is required for the METTL3WTAP interaction, but dispensable for the METTL3-METTL14 assembly (see Figure 4A-4B). In this in vitro methylation assays, METTL3 and METTL14 is capable of methylating RNA probe in the absent of WTAP. In this condition, we found that METTL3 WT as well as its different variants (METTL3-Δ238, METTL3-Δ198, METTL3b and METTL3a) except the catalytically dead mutant METTL3 APPA showed methylation activity in vitro.

5) In Figure 4A, the author suggests that WTAP cannot be immunoprecipitated with METTL3a and 3b because WTAP interacted with the N-terminal of METTL3. If this assay is performed in WT cells, the endogenous full-length METTL3 may help to form the complex. In this case, WTAP is supposed to be co-immunoprecipitated.

We thank reviewer for point this out. METTL3 interacts with WTAP through its N-terminal (1-33aa) (1). Consistently, we find that the two cleaved forms METTL3a and METTL3b which lack the N-terminal region are not able to bind with WTAP. In Figure 4A, we overexpressed METTL3 WT as well as its different variants METTL3-Δ238, METTL3-Δ198, METTL3b and METTL3a respectively in WT cells, and compared the binding ability with WTAP or METTL14 across these overexpressed METTL3 variants. We acknowledge that the exogenous METTL3a and METTL3b interact with endogenous full-length METTL3, and the endogenous full-length METTL3 may help them to form the complex with WTAP. But in fact, the exogenous expression levels of METTL3a and METTL3b are much higher than that of endogenous full-length METTL3 (see Figure 3A and 3C). In this case, METTL3a or METTL3b predominantly interacts with itself, METTL3, METTL14 or other potential interacting proteins through its C-terminal region, this may greatly dilute the condition for the interaction between WTAP and endogenous full-length METTL3. Moreover, in Figure 4A, the comparison is among overexpressed METTL3 variants, the week indirect interactions through much lower expression levels of endogenous protein are probably not comparable to those direct interactions between overexpressed METTL3 variants and WTAP.

Reference:

1) Schöller, E., Weichmann, F., Treiber, T., Ringle, S., Treiber, N., Flatley, A., Feederle, R., Bruckmann, A., and Meister, G. (2018). Interactions, localization, and phosphorylation of the m6A generating METTL3–METTL14–WTAP complex. Rna 24, 499-512

Reviewer #1 (Recommendations For The Authors):

Major points:

1) It is widely known that METTL3 and METTL14 form a stable heterodimer with the stoichiometric ratio of 1:1 (Wang X et al. Nature 534, 575-578 (2016), Su S et al. Cell Res 32(11), 982-994 (2022), Yan X et al. Cell Res 32(12), 1124-1127 (2022)), the numbers of METTL3 and METTL14 in the model of Fig 7P are not equivalent and need to be modified.

We thank for reviewer’s good suggestion. We have modified the model in Fig. 7P.

2) The in vitro methylation activity was detected by the m6A antibody, which has limited linear range. The MTase-Glo{trade mark, serif} Methyltransferase Assay is a SAMdependent enzyme assay with wide applications (Please refer to the references below).

Could this assay be performed by authors?

Wilkinson AW et al. Nature 565(7739), 372-376 (2019).

Yu D et al. Nucleic Acids Res 49(20),11629-11642 (2021).

Yan X et al. Cell Res 32(12), 1124-1127 (2022).

Chen J et al. Nat Commun 13(1), 3257 (2022).

Thanks for reviewer’s good suggestion. We had performed the in vitro methylation assay by using MTase-Glo kit, and the data is consistent with the dot blot results. Please see the new Figure 2H-J.

3) When expressed alone in mammalian cell lines, METTL14 is unstable and is easily contaminated with endogenous METTL3 during purification (Yang W et al. Nat Cell Biol 16(2), p.191-8 (2014), Fig 1e). In Fig 2I, Co-expressing METTL3 and METTL14 maybe a good choice.

We thank for reviewer’s good suggestion. In fact, we co-expressed METTL3 and METTL14 in this in vitro methylation assay in Fig 2I (new Figure 2J in the revised version), METTL3-Flag or its mutant with Flag tag and METTL14-Flag were co-transfected into 293T cells, and co-purified by using Flag M2 magnetic beads from the cell lysates. We have added these details in the indicated method section, please see Line 574-585.

Other minor points:

1) In Fig 5D, the protein domain information of METTL3 and relevant references need to be added (Su S et al. Cell Res 32(11), 982-994 (2022), Fig 6g; Yan X et al. Cell Res 32(12), 1124-1127 (2022), Fig 1a).

We have added these references in the revised manuscript.

2) In Fig 5, would METTL3b contribute to the METTL3-METTL3 interaction?

Our data showed that METTL3a but not METTL3b is responsible for the METTL3-WTAP interaction, breast cancer cell proliferation and the m6A modification. Then, we investigated the mechanism of how METTL3a regulates the METTL3-WTAP interaction, and found that METTL3a is essential for METTL3-METTL3 interaction, which is a prerequisite step for WTAP recruitment in MTC complex. In this case, we speculate that METTL3b is not required for the METTL3-METTL3 interaction. Indeed, through Co-IP assays,we found that METTL3b has no effect on the METTL3-METTL3 interaction (new supplementary Figure 4D), which is consistent with our above data showing that METTL3b is dispensable for the METTL3-WTAP interaction. We have added this comment in Page 6, Line 226 to 228.

3) In Fig 3F, the color in the legend and figure is inconsistent.

We have corrected the inconsistent color in the revised manuscript.

Reviewer #2 (Recommendations For The Authors):

1) In Figure 5D, the construction details of METTL3-HA and Flag should have been included in the method section. Are these tag sequences in the N-terminal of METTL3 protein?

These tags are all in the C-terminal of METTL3. We have added the construction details of these plasmids in the method section. Please see Line 434.

2) In Figure 7A, the labels of the inhibitors are overlapped with the figures.

We have corrected the labels of the inhibitors in Figure 7A in the revised manuscript.

Author Response

We thank the reviewers and editors for their thoughtful evaluation of our preprint. We felt that the reviews were fair and that addressing them will improve the rigor and clarity of our presentation. We are working to address all of the comments, with intent to submit a revised manuscript in the near future.

Author Response

Reviewer #1 (Public Review):

This cross-sectional study examined the results of a survey about cancer treatment disruption during June-August 2020 in 82 counties located in Missouri and Illinois in the U.S. The main outcome was disruption in cancer care. Authors reported that higher education, being a female, experiencing more discrimination in healthcare settings, and having scheduled a telehealth appointment were associated with higher odds of care disruption. Lack of a research focus, lack of following any conceptual framework, the cross-sectional nature of the study, and the small sample size were the noted shortcomings of the manuscript.

We thank Reviewer 1 for their comments. We agree that it is important to understand COVID-related care disruptions using causal methods. However, this manuscript aimed to examine the local impact of COVID care disruptions. We focused on the Siteman Cancer Center’s (SCC) catchment area because the co-author team includes the SCC’s Associate Director of Community Outreach and Engagement (COE) program, the SCC Associate Director for Diversity, Equity, and Inclusion, multiple members of the SCC COE leadership team. Thus, we are uniquely positioned to mobilize and identify outreach opportunities and/or programs that address any gaps we discover. Moreover, this focus on our catchment area and the motivation for this survey aligns with the National Cancer Institute’s priorities of population health assessments to characterize cancer-relevant knowledge, attitudes, beliefs, and behaviors across cancer center catchment areas. While this is a crosssectional study, this snapshot of care disruption will be helpful in planning local outreach strategies. Lastly, our catchment area is challenged with multiple cancer disparities patterned by social identities. Therefore, our analysis was guided by the theory that social identities related to race, ethnicity, class, and gender shape access to healthcare and disease processes and are the fundamental drivers of health. Thus, we included variables that impact health and are patterned by these social factors.

Reviewer #2 (Public Review):

Dr. Kia Davis and colleagues present a thoughtful analysis of disruptions to cancer care during COVID-19 in the article, "Understanding disruptions in cancer care to reduce increased cancer burden: a cross-sectional study." The article is based on an online survey of 680 residents in the Siteman Cancer Center catchment area in Summer 2020. The authors aim to characterize demographic differences in cancer care disruptions. Information about the causes and distribution of care disruption can help reduce the impacts of COVID-19 and guide the recovery of programs and services. The article provides a clear and detailed assessment of factors associated with care disruption and return to care during the first six months of the pandemic.

A strength of the study is the focus on the catchment area of the cancer center during a period of dramatic change. The results would provide timely and actionable data to address emerging barriers to care and associated social or contextual factors. This information helps the Community Outreach and Engagement efforts to be responsive to community priorities despite rapidly evolving circumstances.

The analysis would benefit from greater detail in three areas. First, it would be helpful to have more information about how the outcome measures were originally developed or tested. Second, for the regression analysis, it would be helpful to show the demographic characteristics of the two strata to better understand the sample composition. Third, the authors should demonstrate that the data do not violate the assumptions for conducting logistic regression to improve confidence in the findings.

COVID-19 affected all aspects of the cancer continuum. The study reports factors associated with postponing or canceling cancer-related appointments during the pandemic. It will be of great interest to researchers and practitioners in cancer prevention and control.

We thank Reviewer 2 for their thoughtful critique of our work. Their suggestions have strengthened our manuscript. Since our article was submitted, the questionnaire where we derived our outcome measure has been published. The questions were drawn from validated measures assessing the impact of pandemics such as H1N1, and major life disruptions such as natural disasters. This language was updated in the manuscript as were the references. Moreover, we added a supplemental Table 2 to show the demographic characteristics by race strata. Finally, we tested and can confirm that the analysis does not validate the assumptions of logistic regression. We believe that our results will aid in the understanding of how COVID impacted cancer care in our catchment area so that we can better mobilize resources. While we understand this is a cross-sectional study with the potential for unmeasured confounding, we believe this snapshot of cancer care during the pandemic will also be of interest to researchers, clinicians, and other practitioners in cancer prevention and control in locations like ours.

Author Response

Reviewer #3 (Public Review):

In this manuscript, Castano et al generate and test a small molecule inhibitor of CDKL5, an Xlinked kinase whose loss-of-function is the cause of a severe neurodevelopmental disorder. Since the current knowledge of CDKL5 functions mainly rely on genetic models it is still unclear which effects are caused directly by CDKL5 loss and which can be ascribed to indirect effects. A specific inhibitor would therefore be an important tool for the field.

Castano and colleagues therefore tested a panel of twenty kinase inhibitors for their capacity to block phosphorylation of a EB2, a bona fide CDKL5 substrate, in rat neurons. Among the three that could inhibit EB2 phosphorylation at low concentrations, one was found to inhibit CDKL5 while not affecting GSK3 kinases, which share significant homology to CDKL5. Considering that genetic studies have previously linked CDKL5 to excitatory synaptic transmission, acute hippocampal slices were exploited to test the consequences of CDKL5 inhibition. While CDKL5 loss in the past was found to affect both AMPA- and NMDA-Rs, the small molecule-based inhibition affected only AMPA-R responses at the post-synaptic level. Since pharmacokinetic analyses showed that the inhibitor has a low capacity for brain penetration the molecule remains limited for testing the acute inhibition of CDKL5 in vitro and ex vivo. Such a tool represents an important aspect in the CDKL5 field and the findings suggesting a direct role of CDKL5 in regulating AMPA-R functions are interesting. However, the manuscript could be improved to render it more readable.

Thank you for this positive feedback and we hope that our adjustments improve the readability.

The description of the binding and orthogonal assays, which are the basis for the selection of the small molecule inhibitor, is not straightforward to understand for non-expert readers and could be improved.

We have added additional text to the Methods and Results to better explain the assays.

While the in vitro and ex vivo assays are well presented, it is not clear why the myelin basic protein is used as a substrate for CDKL5 in the in vitro kinase assays. Does this protein contain a CDKL5 consensus site?

To execute the in vitro kinase assays, myelin basic protein (Active Motif, 31314) was employed as a substrate for recombinant CDKL5. Myelin basic protein is used as a substrate for multiple kinases, both serine/threonine and tyrosine kinases, to enable in vitro kinase assays due to the presence of multiple sites for phosphorylation. As such, we and others have used this protein as a kinase substrate for evaluating kinase activity[2, 4]. MBP does not contain a CDKL5 consensus site of RPXS/T*, and as such could be considered a less than ideal substrate to study CDKL5 activity, however for in vitro kinase assays MBP is still suitable as it can be phosphorylated by CDKL5. In addition, CDKL5 is known to phosphorylate substrates that do not contain a consensus motif[3].

Author Response

Reviewer #1 (Public Review):

This study demonstrates that a hybrid measurement method increases 3 fold the resolution of mouse USV localization. This increased resolution enables to revise previous occurrence frequency measures for female vocalizations and establishes the existence of vocal dominance in triadic interactions. The method is well described and its efficiency is carefully quantified. A limitation of the study is the absence of ground truth data, which may have been generated eventually with miniaturized loudspeakers in mouse puppets. However, a careful error estimation partially compensates for the absence of these likely challenging calibrations. In addition, the conclusions take into account this uncertainty. The gain in accuracy with respect to previous methods is clear and the impact of localisation accuracy on biological conclusions about vocalisation behavior is clearly exemplified. This study demonstrates the impact of the new method for understanding vocal interactions in the mouse model, which should be of tremendous interest for the growing community studying social interactions in mice.

We have performed the requested, additional ground estimate using a movable miniature speaker, for more details see point 2 of Reviewer 2, and the new supplementary figure.

Reviewer #2 (Public Review):

Past systems for identifying and tracking rodent vocalizations have relied on triangulating positions using only a few high-quality ultrasonic microphones. There are also large arrays of less sensitive microphones, called acoustic cameras that don't capture the detail of the sounds, but do more accurately locate the sound in 3D space. Therefore the key innovation here is that the authors combine these two technologies by primarily using the acoustic camera to accurately find the emitter of each vocalization, and matching it to the highresolution audio and video recordings. They show that this strategy (HyVL) is more accurate than other methods for identifying vocalizing mice and also has greater spatial precision. They go on to use this setup to make some novel and interesting observations. The technology and the study are timely, important, and have the potential to be very useful. As machine learning approaches to behavior become more widespread in use, it is easy to imagine this being incorporated and lowering entry costs for more investigators to begin looking at rodent vocalizations. I have a few comments.

1) What is the relationship of the current manuscript to this: https://www.biorxiv.org/content/10.1101/2021.10.22.464496v1 which has a number of very similar figures and presents a SLIM-only method that reportedly has lower precision than the current HyVL approach. Is this superseded by the submitted paper?

The referred manuscript (now published in Scientific Reports) is indeed related to the current work: The currently presented system is based on the integration between SLIM (based on 4 high quality microphones) and Beamforming (based on the 64-channel microphone array). The accuracy of SLIM is generally lower than that of HyVL, but it makes essential contributions to the overall accuracy of HyVL through the integration of the complementary strengths of the two methods/microphone arrays (see Fig. 3A, L-shape of errors). To our knowledge, SLIM was the previously most accurate technique (based on 4 microphones, see comparison in the Discussion), but HyVL exceeds this by a substantial margin. Some figures appear similar mostly due to related code in the underlying analysis pipeline and visualization scripts (e.g. the half-disc densities). However, the set of dyadic and triadic recordings was collected specifically for the present study, and all top-level analyses were performed separately. The single mouse (C57Bl/6 WT) ground truth dataset is shared between the two studies, where in the SLIM paper only the USM4/SLIM part was evaluated (leading to a correspondingly lower, single animal accuracy).

We felt that the level of detail above would probably impede the reading of the manuscript, and we have therefore added a subset of the above clarifications to the methods and the first time the other study is mentioned.

2) Can the authors provide any data showing the accuracy of their system in localizing sounds emitted from speakers as a function of position and amplitude? I am imagining that it would be relatively easy to place multiple speakers around the arena as ground truth emitting devices to quantify the capabilities of the system.

Ground truth data is critical for any meaningful comparison. First, we would like to highlight that we already provided ground truth data in the previous version of the manuscript: In Fig. 3C. we analyzed vocalization data from trials with (1) just a single mouse as well as (2) vocalization at times when all mice were far apart in relation to the accuracy of HyVL (>100 mm, i.e. >25x the accuracy of HyVL) where the chances of erroneous assignment are negligible. We think that these tests are the most relevant, as they are conducted with the relevant sounds, at their actual intensity, spectral profile and emitter acoustics.

In addition, we have now conducted a series of tests with sounds produced by a miniature speaker placed in 25 different locations to demonstrate the lower-bound of accuracy achievable with the system. The tests indicate an accuracy of MAE < 1mm under these ideal conditions, i.e. without the absorption of the mouse bodies, varying direction of emission of the mouse snout, varying intensity, varying spectral content, duration, etc. Exploring the dependence on all these parameters is in itself interesting, but requires a detailed study in itself. The detailed experimental conditions and results are now provided in Supplementary Fig. 4, including a quantification of the dependence on amplitude.

3) How is the system's performance affected by overlapping vocalizations? It might be useful to compare the accuracy of caller identification for periods where only one animal is calling at a time vs. periods where multiple animals are simultaneously calling.

This is an excellent question. Our current code for detecting vocalizations cannot automatically determine if one or multiple vocalizations are concurrently present. We have therefore manually checked all vocalizations for overlapping instances, including those in triadic recordings with two males, where this would be expected to occur most frequently.

We considered vocalizations to be overlapping if the overlapping constituent timefrequency traces did not form a harmonic stack. Overall, overlaps were surprisingly rare. We did find a couple of cases (<0.1%) where our detection algorithm produced a longer vocalization interval that contained multiple, differently shaped vocalization traces that, when re-analyzed in shortened time-frequency bins with beamforming, belonged to two different males. Note here that beamforming is separately performed from the onset to the end of each vocalization, so the cumulative heatmap can change depending on these onset and end times, which are normally determined by our detection algorithm.

However, although the identity of the assigned vocalizer could shift in these very rare cases depending on which time bin was re-analyzed, the system’s localization performance remained in principle unaffected: as mentioned above, shorter time bins on non-overlapping parts correctly show the origin of the vocalizations in this case, and therefore a solution to this issue could be a USV detection algorithm that is able to detect the overlap based on the spectral shapes and parses them apart. During the beamforming each vocalization can then be separately localized, by restricting the beamforming to the corresponding time and frequency range. Further, the analysis could be refined so that multiple salient peaks can be detected in the soundfield estimate. This would, however, substantially change the analysis approach, i.e. rather than a single estimate per USV, a sequence of soundfield estimates should be computed and later fused again. Since such a procedure uses less data per single estimate, it also increases the possibility of false positives, which in the current situation with very few overlaps in time, would likely reduce the overall accuracy of the system, we decided to not modify the algorithm in this direction, but we agree that ideally a joint approach - combining separation on the spectrogram and soundfield level - should be pursued. For the present data, if a time window was analyzed such that the intensity map of the sound field contains multiple hotspots of an approximately equal magnitude, the USV would likely remain unassigned, because the within soundfield uncertainty would be higher than for a single peak, and this would reduce the MPI. However, given the rarity of these cases in our dataset, we do not think that their exclusion would change the results appreciably. This information was added as a paragraph to the Discussion.

It is worth noting that HyVL is very robust: There were a number of cases (<5%) where environmental dampening in combination with harmonic stacking produced interesting timefrequency traces in some of the USM4 microphones, but our system did not have any issue spatially localizing this - what seems like a - smeared vocalization trace. We provide a few examples of this kind in a short video (see Rebuttal Video 2 and the legend at the bottom of this document), where the overlap is also reflected in the intensity map of the sound field, overlaid onto the platform.

4) Can the authors comment on how sound shadows cast by animals standing between the caller and a USM4 affect either the accuracy of identification or the fidelity of the vocal recording?

An important point to raise. Sound scattering and dampening caused by the conspecifics of the vocalizing animal can impede the accuracy of any sound localization system, but can unfortunately not be avoided in a social setting. To address this issue, we raised all USM4 microphones by ~12 cm above the interaction platform to minimize the instances of sound blocked by the mice. Further, the Cam64 device should largely be unaffected by sound shadows as it is centrally located above the platform. We have added a modified version of the above comment to the discussion under the heading "Current limitations and future improvements of the presented system".

5) I'm a bit confused about how the algorithm uses the information from the video camera. Reading through the methods, it seems like they primarily calculate competing location estimates by the two types of microphone data and then make sure that a mouse is in close proximity to one location, discarding the call if there isn't. Why did the authors choose this procedure rather than use the tracked position of the snouts as constrained candidate locations and use the microphone data to arbitrate between them? Do they think that their tracking data are not reliable or accurate enough?

Thanks for this important suggestion, which we have actually grappled with a lot during the analysis. First of all, the visual tracking data, in particular the manual data, is in our opinion (based on human visual identification) near perfect (within the limits of the video resolution, pixel resolution = 0.8 mm), i.e. on the order of 1-2 mm, and is therefore not the source of any unattributable vocalizations. If we understand the reviewer correctly, then we indeed perform the attribution as he indicates based on the tracked snouts of all mice, specifically by measuring the MPI's of both acoustic location estimates for all mice and then choosing the most reliable one. Specifically, the attributions can be grouped into 3 cases: (i) Estimated origin close to one snout, and snouts rather far apart, (ii) Estimated origin close to one snout and snouts close, and (iii) estimated origin not close to either snout. (i) is easy to address, (ii) is appropriately handled by the mouse probability index, but (iii) is tricky. Since the vocalization has to come from one of the mice, this already indicates that the localization is not working well in this case. Therefore we found it prudent (similar to Neunuebel et al. 2015) to not assign in these cases. Interestingly the MPI is not useful in these cases, as due to the exponential dependence of the normal density on distance, for example a case with a distance of 50 mm to one snout and 60 mm to another snout could lead to an MPI close to 1, which is likely not trustable. We have described this in the Methods as follows:

"This distance threshold mainly serves to compensate for a deficiency of the 𝑀𝑃𝐼: if all mice are far from the estimate, all 𝑃𝑘 are extremely small, however, the 𝑀𝑃𝐼𝑘 will often exceed 0.95."<br /> Due to the inherent limit for localizing very quiet, short USVs by any system, we think this kind of selection (introduced originally by Neunuebel et al 2015) is a valuable and necessary step in the processing to avoid confusions (which are of course already substantially reduced through HyVL here).

6) I guess the authors have code that we can run, but I couldn't access it. The manuscript describes the algorithms and equations that are used to calculate the location, but this doesn't really give me a feel for how it works. If you want to have the broadest impact possible, I think you would do well to make the code user-friendly (maybe it is, I don't know). In pursuit of that goal, I would suggest that the authors devote some of the paper to a guided example of how to use it.

While the code was made available to the reviewers via the link at the beginning of the manuscript (p2, before abstract), we completely agree that this method of distribution is not very accessible. We have therefore created a publicly available GitHub repository (https://github.com/benglitz/HyVL) which hosts the code and details its use on the basis of a sample data set (which is available to the reviewers in the repository link, and later to the public under https://doi.org/10.34973/7kgc-ta72). While we do provide a sample video and analysis workflow there, our data analysis pipeline is quite integrated and other labs will likely use different pipelines. We have therefore tried to make the core functions independent of our pipeline and thus easy to integrate by others into their analysis pipelines.

Reviewer #3 (Public Review):

The present manuscript describes a new method to identify the emitter of ultrasonic vocalisations during social interactions between 2 or 3 mice. The method combines two technologies (an "acoustic camera" and a set of four microphones) and succeeds in increasing the spatial precision and the attribution of USV emission to one of the mice. The manuscript describes the characteristics and advantages of each method and the advantages of using both to optimize the identification of USV emitter. The authors used the method to confirm that females are also vocalising during male-female interactions and that females emit USV mostly during nose-nose contact while this was not the case for males. Interestingly, the authors identified that the vocal behaviour of two competing males was strongly asymmetric when facing a female. This was not the case for two females facing one male.

The method is really promising since the identification of the emitter of USVs during mouse social interactions is a necessary step to speed up our understanding of this communication modality. The increase in spatial precision and in the proportion of attributed vocalisations is non-negligible and will be of great utility in the future.

We would like to thank the reviewer for this positive perspective on the future utility of our system.

Generally, the statistical analyses should be adjusted. Indeed, the statistical analyses do not consider the fact that the same individuals were recorded several times (if we understood well the methods). Each point was considered independent (in non-parametric Wilcoxon tests), while this is not the case given the repetitions with the same individuals (the number of repeated encounters per individual should be given in the methods section, by the way). We strongly recommend revising the statistical analyses of the results in Figures 4 and 5. In addition, it could be interesting to check whether the vocal behaviour is stable within each individual (i.e., a male that is vocalising frequently in one situation vocalises always frequently in other situations).

We generally agree with this suggestion: In order to properly conduct the analysis for individuals as you suggest, a balanced dataset should be used. We had initially collected such a balanced dataset, which was previously not detailed in the manuscript, as the focus was on USV localization/attribution and hence only the recordings containing USVs were analyzed (detailed now in the beginning of Results and Methods). However, overall, the probability of a recording containing vocalizations at all is low: in our balanced set only 23/112 recordings contained vocalizations. We therefore had collected additional recordings with the best vocalizers which created the previously analyzed set of 83 recordings containing USVs recorded with all microphones. This dataset is therefore dominated by recordings from mice that are active vocalizers. While this does not raise any issue for the estimation of the accuracy of the method (Figure 3) or the female vocalizations (Figure 4, because recordings were always randomized across female mice), it precludes an encompassing analysis of individual differences in Figure 5, i.e. the dyadic-triadic comparison. In the new Figure 5, we address the reviewer's question for the dyadic recordings, finding that the current set of recordings does not provide sufficient evidence that individual male mice had significantly different vocalization rates. We would, however, like to point out that this is likely a consequence of the n=4 recordings that are compared here. For the female mice, we also did not find differences in vocalization rates, which is based on n=14 recordings and thus a more reliable result (p=0.16, 1-way ANOVA with factor individual).

For the triadic recordings, however, due to a limitation in the experiment execution, we unfortunately do not have the complete information available on an experiment level for the triadic recordings, i.e. the video stream was accidentally started after all mice were placed in the platform, and since the same sex animals are visually not separable (while the female mice are separable from the males, based on a slightly shaved region on their head), we cannot completely assess this question in triadic recordings based on the available data. When including the triadic recordings in addition and assuming a single vocalizer (combining all male USVs, see below for why the males could not be assigned in the triadic condition) the male individual comparison can be approximately performed with n=8 recordings, and then the dependence on individual becomes borderline significant (p=0.028, 2-way ANOVA with factors individual and condition).

For the comparison of vocalization rates in the previous Figure 5 that the reviewer was referring to, we cannot perform a rigorous analysis on the individual level, due to the lack of balance. While we thus agree that differences between individual mice can contribute to the differences observed, we do not think that this would change the conclusion that one of the mice dominates the vocal emissions. If the reviewers agree, we would thus leave Figures 6 (old Fig. 5) and new Figure 7 (behavioral confirmation of dominant/subordinate division) as part of the manuscript, with a clear cautioning about the possible contribution of individual differences to the observed differences. If the reviewers find it inappropriate to leave the results based on the unbalanced dataset in, all results after figure 5 could also be excluded (although we would find this unfortunate, given the additional time and effort we have invested in these).

It is not easy to understand the rationale behind testing animals in pairs and in triads from the beginning of the manuscript. The authors should better introduce this aspect in the manuscript, especially given the fact that biological results deal with this aspect in Figure 5. The authors might strengthen the parts of the biological results extracted from their new method.

Thank you for pointing out the need for clarification regarding the rationale behind testing animals in pairs and in triads. It is because courtship interactions are particularly vocal and social, that they are of interest to many fields, e.g. neurodevelopmental disorders.3,4 Due to the natural competitiveness between mice during courtship interactions, high accuracy is particularly beneficial in this regard because it allows disentangling USVs at close distances. We adapted the introduction to better reflect this reasoning and included an extra paragraph in the introduction and also where the biological results from old Fig. 5 / new Fig. 6 are summarized.

More specifically, the fact that one male takes over the vocal behaviour within a triad is of high interest. Nevertheless, some behavioural data would be needed to strengthen these findings.

We agree that this is an interesting finding and also agree that some additional behavioral analysis is useful to complement it. In order to arrive at this analysis, we performed all-frame, 3-animal tracking on the 14 triadic recordings with two males. This required switching to skeleton tracking with SLEAP5 in addition to manual post-processing to ensure that no identity switches occur. In each recording the dominant male was then defined as the one that emitted more vocalizations, and then the vocalization-independent spatial interaction histogram was computed, similar to the ones in Fig.4, but now separating between the dominant and the subordinate males (see new Figure 7). The results are consistent with the most typical location of vocalization of the male, in proximity to the female abdomen: The dominant male's spatial interaction histogram (Fig. 7A) was more clearly peaked in the location of the female abdomen very close to the male's snout, in comparison with the subordinate male's histogram (Fig. 7B), which shows up very clearly in the difference between the normalized histograms (Fig. 7C). Significance analysis was performed using 100x bootstrapping on the relative spatial positions to estimate p=0.99 confidence bounds around the histograms of the dominant and subordinate respectively. Significance at a level of p<0.01 highlights multiple relative spatial positions (Fig. 7D), including the one proximal to the snout which has the largest absolute difference (Fig. 7C). Note, that these analyses were conducted on the basis of the non-balanced dataset which contained enough vocalizations to assess the dominant male based on the vocalization rates and thus individual traits of certain animals remain as a possible confound.

A small proportion of USVs was not assigned. The authors did not discuss the potential reason for this failure (Were the USVs too soft? Did they include specific acoustic characteristics that render them difficult to localise?). These points could be of interest when testing other mouse strains or other species.

Good point, we agree that it is interesting to know the reasons for failure. As so often, there is not a single property that makes localization hard, but multiple factors contribute. In the SLIM paper, we already identified duration and intensity as important contributors (Fig. 3E/F), and in the speaker test (see new Supplementary Fig. 4) we again demonstrated the influence of intensity. In addition, frequency bandwidth and acoustic occlusion are two other main contributors that each influence the availability of the information/signal-to-noise ratio at the microphones:

• Frequency bandwidth: In signals that are very narrowband, there are more opportunities for phase ambiguity, in particular for very high-frequency signals. These are avoided/reduced for more wideband signals.

• Acoustic occlusion: As ultrasonic sounds can be quite directional, if an animal is vocalizing away from a microphone, which in addition would put its body in the way of the sounds to the microphone, then this can reduce the intensity at the microphone to a level where the information is insufficient to utilize information from this microphone. This mostly influences the 4 microphones surrounding the platform, while the Cam64 overhead will likely not be affected by acoustic occlusion in the plain.

We have added a brief version of this explanation to the discussion under the heading: "Current limitations and future improvements of the presented system"

Author Response

Reviewer #1 (Public Review):

In this manuscript, Marmor and colleagues reanalyze a previously published dataset of chronic widefield Ca2+ imaging from the dorsal cortex of mice as they learn a go/no-go somatosensory discrimination task. Comparing hit trials that have a distinct history (i.e. are preceded by distinct trial types), the authors find that hit trials preceded by correct rejections of the nontarget stimulus are associated with larger subsequent neural responses than trials precede by other hits, across the cortex. The authors analyze the time course over which this effect emerges in the barrel cortex (BC) and the rostrolateral visual area (RL), and find that its magnitude increases as the animals become expert task performers. Although the findings are potentially interesting, I, unfortunately, believe that there are important methodological concerns that could put them into question. I also disagree with the rationale that singles out BC and RL as being especially important for the emergence of trial history effects on neural responses during decision-making. I detail these points below .

1) The authors did not perform correction for hemodynamic contamination of GCaMP fluorescence. In widefield imaging, blood vessels divisively decrease neural signals because they absorb green-wavelength photons, which could lead to crucial confounds in the interpretation of the main results because of neurovascular coupling, which lags neural activity by seconds. For example, if a reward response from the previous trial is associated with a lagged hemodynamic contamination that artificially decreases the signal in the following trial, one could get artificially higher activity in trials that were not preceded by a reward (i.e. CR), which is what the authors observed. Ideally, the experiments would be repeated with proper hemodynamic correction, but at the very least the authors should try to address this with control analyses.

Done. We basically redone the experiment with proper hemodynamic correction and maintained trial history results. Please see point 1 above for more details (Figures S4 and S5). In addition to hemodynamic controls, we also present novel two-photon single cell data with similar results in Figure S6. We also added a dedicated section for this in the Methods section (pg. 12).

For example, what is the time course of reward-related responses in BC and elsewhere?

In general, and specifically in BC, reward related responses return to baseline up to 5 seconds after the start of the reward period and at least 5 seconds before the stimulus presentation of the next trial. In the novel experiments we even extended the baseline period by an additional 2 seconds just in case. Trial history information was still present with an extended inter-trial interval.

The text now reads (pg. 4): "We further report that responses during the reward period in cortex and specifically in BC went back to baseline 4-5 seconds after the start of the reward period and 6-8 seconds before the presentation of the next stimulus (total inter-trial interval ranged between 10-12 seconds)."

Do hemodynamics artifacts have a trial-by-trial correlation with the subsequent trial history effect?

We have now done the proper hemodynamic control (Figure 2) and we did not find a strong effect of hemodynamic responses on trial history information.

What is the learning time course of reward responses?

Responses during the reward period as a function of learning were not significantly modulated. We further show the whole learning profile for BC response during the reward period in Author response image 1.

Author response image 1.

Response in BC averaged during the reward period (2-4 sec after texture stop) as a function of learning for each mouse separately.

The text now reads (pg. 4): "In addition, responses in BC during the reward period were not consistently modulated as a function of learning (p>0.05; Wilcoxon signed-rank test between naïve and expert, BC response averaged during the reward period, 2-4 seconds after stimulus onset; n=7 mice). Taken together, we find that direct responses from the reward period do not effect history-related responses during the next trial."

Note that I don't believe the FA-Hit condition analysis that the authors have already presented provides adequate control, as punishment responses are also pervasive in the cortex and therefore suffer from the same interpretational caveat. Unfortunately, I believe this is a serious methodological issue given the above. However, I will proceed to take the reported results at face value .

We hope that our additional control analysis regarding the hemodynamic controls are satisfactory.

2) The statistics used to assess the effect of trial history over learning are inadequate (e.g., Fig 2b). The existence of a significant effect in one condition (e.g., CR-Hit vs. Hit-Hit in expert) but not in another (e.g., same comparison in naive) does not imply that these two conditions are different. This needs to be tested directly. Moreover, the present analysis does not account for the fact that measures across learning stages are taken from the same animals. Thus, the appropriate analysis for these cases would be to first use a two-way ANOVA with repeated measures with factors of trial history and learning stage (or equivalent non-parametric test) and then derive conclusions based on post hoc pairwise tests, corrected for multiple comparisons .

Done. We performed 2 way ANOVA as suggested and found significant history and learning effects along with a significant interaction effect for BC.

The text now reads (pg. 4): "This difference was significant during the stim period in learning and expert phases across mice (Fig. 2b; 2-way ANOVA with repeated measures; DF (1-6) F=51 p<0.001, DF (2-12) F=18 p<0.001, DF(2-12) F=5 p<0.05 for trial history, learning and the interaction between trial history and learning; Post hoc Tukey analysis p<0.05 for trial history in learning and expert phases; p>0.05 in the naïve phase)."

3) I am not convinced that BC and RL are especially important for trial-history-dependent effects. Figures 4 and 5 suggest that this modulation is present across the cortex, and in fact, the difference between CR-Hit and Hit-Hit in some learning stages appears stronger in other areas. BC and RL do have the highest absolute activity during the epochs in Figs 4 and 5, but I would argue that this is likely due to other aspects of the task (e.g., touch) and therefore is not necessarily relevant to the issue of trial history .

Done. First, we would like to point out that RL during the pre period displays the largest difference between the CR-Hit and Hit-Hit conditions (Fig. 5c bottom). Second, we now show difference maps (i.e., activity in CR-Hit minus Hit-Hit) which clearly show a positive activity patch in BC during the stim period for 5 out of the 7 mice (Fig. S10a). Example maps also highlight RL during the pre period (Fig. S10b). We note that activity patches somewhat spread over to other areas and also slightly vary across mice. This is why the grand average may slightly average out trial history information. Taken together, we strongly feel that during the pre period, trial history information emerges in RL (and adjacent posterior association areas) which shift towards BC during the stim period

Nevertheless, we agree with the reviewer that other areas (that do not necessarily display high activity) may encode trial history information and we now clearly report this in the text (pg. 5): "We note that other areas, e.g., different association areas, also encoded historydependent information especially during learning and expert phases. In addition, we present activity difference maps between CR-Hit and Hit-Hit conditions during the stim period (Fig. S10a). These maps clearly show the highest trial history information (i.e., difference in activity) in BC. Taken together, these results indicate that BC encodes history-dependent information that emerges during the stim period and just after learning. "

And also in (pg. 6): " In addition, we present activity difference maps between CR-Hit and HitHit conditions during the pre period (Fig. S10b). These maps localize trial history information to RL which also spreads to other adjacent association areas. Moreover, activity patches slightly vary across the different mice which may affect the grand average (averaged across mice) of each area."

4) Because of similar arguments to the above, and because this was not directly assessed, I do not believe the conclusion that history information emerges in RL and is transferred to BC is warranted. For instance, there is no direct comparison between areas, but inspection of the ROC plots in Fig 6b suggests that history information emerges concomitantly across cortical areas. I suggest directly comparing the time course between these and other areas

Done. We now add example history AUC maps and quantify history AUC for all 25 areas during the pre and stim periods. During the pre period (Fig. 6), AUC values are concentrated around the RL (and other PPC areas), whereas during the stim periods AUC values shift to BC. Again, due to the inter-mouse variability, these differences are slightly averaged out which also makes it tough to have strong statistical test (with only 7 mice).

The text now reads (pg. 7): "We next calculated the history AUC for each pixel during either the pre or stim period. The history AUC maps during the pre period display AUC values around the RL areas (Fig. 6f). In contrast, the history AUC maps during the stim period display AUC values mostly in BC (Fig. 6g). Quantified across 25 areas and averaged across mice, RL displays the highest history AUC during the pre period, whereas BC displays the highest history AUC values during the stim period (Fig. 6h). We note that other cortical areas such as other association areas also display high history AUC values. Taken together, we find that trial history emerges in RL before the texture arrives and then shifts to BC during stimulus presentation. "

5) How much is task performance itself modulated by trial history? How does this change over the course of learning? These behavioral analyses would greatly help interpret the neural findings and how this trial history might be used behaviorally .

Done, we have now calculated the dprime for Hit-Hit and CR-Hit trials separately. We find no significant differences between conditions both within and across mice (see Fig. S2 below).

The text now reads pg. 3): "We note that learning curves that are calculated separately for each pair (i.e., either a preceding Hit or CR trial) were not significantly different (Fig. S2)."

Reviewer #2 (Public Review):

Marmor et al. mine a previously published dataset to examine whether recent reward/stimulus history influences responses in sensory (and other) cortices. Bulk L2/3 calcium activity is imaged across all of the dorsal cortex in transgenic mice trained to discriminate between two textures in a go/no-go behavior. The authors primarily focus on comparing responses to a specific stimulus given that the preceding trial was or was not rewarded. There are clear differences in activity during stimulus presentation in the barrel cortex along with other areas, as well as differences even before the second stimulus is presented. These differences only emerge after task learning. The data are of high quality and the paper is clear and easy to follow. My only major criticism is that I am not completely convinced that the observed difference in response is not due to differences in movement by the animal on the two trial types. That said, the demonstration of differences in sensory cortices is relatively novel, as most of the existing literature on trial history effect demonstrates such differences only in higher-order areas .

Major :

1a) The claim that body movements do not account for the results is in my view the greatest weakness of the paper - if the difference in response simply reflects a difference in movement, perhaps due to "excitement" in anticipation of reward after not receiving one on CR-H vs. HH trials, then this should show up in movement analysis. The authors do a little bit of this, but to me, more is needed .  

Done. We have now extensively and carefully analyzed body and whisker movements for CRHit and Hit-Hit conditions. First, In the figure below we decomposed body movements into 22 different body parts using DeepLabCut. In short, we find no significant difference between CRHit and Hit-Hit conditions in each body part separately (Fig. S7 below). This was true for the naïve, learning and expert phases. Please see additional analyses in the points below.

This is now reported in the text (pg. 4): “In addition, we performed a more detailed body and whisker analysis, e.g., decomposing the movement to different body parts and obtaining single whisker dynamics. These analyses did not find significant differences in movement parameters between CR-Hit and Hit-Hit conditions (Fig. s7 and s8).”

First, given the small sample size and use of non-parametric tests, you will only get p<.05 if at least 6 of the 7 mice perform in the same way. So getting p>.05 is not surprising even if there is an underlying effect. This makes it especially important to do analyses that are likely to reveal any differences; using whisker angle and overall body movement, which is poorly explained, is in my opinion insufficient. An alternative approach would be to compare movements within animals; small as the dataset is, it is feasible to do an animal-by-animal analysis, and then one could leverage the large trial count to get much greater statistical power, foregoing summary analyses that pool over only n=7 .

We agree with this point and are have now dramatically improved our statistical analysis.

1) We now perform within mouse statistics for responses in BC during naïve, learning and expert (see Fig. S4 below). In short, we find statistical significance for 7 out of 7 mice during the expert phase, 6 out of 7 mice in the learning phase and 0 out of 7 in the naive phase. For RL during the pre period we find significant difference in 5 out of 7 expert mice.

This is now reported in the text (pg. 4): "In addition, a statistical comparison between CR-Hit and Hit-Hit responses within each mouse separately maintained significance for expert (7/7 mice Mann-Whitney U-test p<0.05) and learning (6/7 mice) but not for naïve (0/7 mice. Fig. S3)."

And also in (pg. 5): "In addition, a statistical comparison between CR-Hit and Hit-Hit responses in RL within each mouse separately maintained significance for expert (5/7 mice; MannWhitney U-test p<0.05)."

2) We would like to point out that we have now added 3 additional mice (with hemodynamics control) and performed within mouse statistics in BC and RL (Fig. S5), adding to our initial observations.

3) In terms of body movements, we now performed within mice statistics and compared body movements between CR-Hit and Hit-Hit conditions. In general, most mice did not show a significant difference in body movements or whisker envelope.

This is now reported in the text (pg. 4): "A within mouse statistical comparison between body or whisker parameters in CR-Hit and Hit-Hit maintained a non-significant difference in expert (1/7 mice displayed a significant difference; Mann-Whitney U-test p>0.05), learning (2/7 mice) and naïve (0/7 mice)."

And also in (pg. 4): "Body movements and whisker parameters did not significantly differ between CR-Hit and Hit-Hit conditions during the pre-period (Similar to the stim period. Across and within mice. P>0.05; Mann-Whitney U-test)."

In summary, we have now substantially improved our statistical analysis and further decomposed the body movements, maintaining the trial history results.

The authors only consider a simple parametrization of movement (correlation across successive frames), and given the high variability in movement across animals, it is likely that different mice adopt different movements during the task, perhaps altering movement in specific ways. Aggregating movement across different body parts after an analysis where body parts are treated separately seems like an odd choice - perhaps it is fine, but again, supporting evidence for this is needed. As it stands, it is not clear if real differences were averaged out by combining all body parts, or what averaging actually entails .

Please see the above point where we decomposed body movements (Fig. S7 and Methods section in Pg. 14).

If at all possible, I would recommend examining curvature and not just the whisker angle, since the angle being the same is not too surprising given that the stimulus is in the same place. If the animal is pressing more vigorously on CR-H trials, this should result in larger curvature changes .

Done. We now decompose whisker dynamics (i.e., curvature) using DeepLabCut (Fig. S8 see below). In general, we find no significant differences in whisker parameters between Hit-Hit and CR-Hit conditions.

This is now reported in the text (pg. 4): "In addition, we performed a more detailed body and whisker analysis, e.g., decomposing the movement to different body parts. This analysis did not find significant differences between CR-Hit and Hit-Hit conditions (Fig. S7 and S8)."

Finally, the authors presumably have access to lick data. Are reaction times shorter on CR-H trials? Is lick count or lick frequency shorter?

Done. We now calculated lick reaction time and lick rate and find a significant difference for the lick reaction time but not in lick rate. We show a figure below for the reviewer and report this in the text

The text now reads (pg. 3): "In addition, the lick reaction time (but not the lick rate) between Hit-Hit and CR-Hit were significantly different (p<0.05; Wilcoxon signed-rank test) ,maybe indicating a more considered response after a previous stop signal."

If movement differs across trial types, it is entirely plausible that at least barrel cortex activity differences reflect differences in sensory input due to differences in whisker position/posture/etc. This would mitigate the novelty of the present results .

As detailed above, have now meticulously analyzed the whisker parameter differences between both conditions and did not find any significant differences.

1b) Given the importance of this control to the story, both whisker and body movement tracking frames should be explicitly shown either in the primary paper or as a supplement. Moreover, in the methods, please elaborate on how both whisker and body tracking were performed .

Done. Please see Figs. S7 and S8 for tracking frames. This is now detailed in the above points and also the revised relevant methods section

2) .Did streak length impact the response? For instance, in Fig. 1f "Learning", there is a 6-trial "no-go" streak; if the data are there, it would be useful to plot CR-H responses as a function of preceding unrewarded trials.

Done. We have now calculated response in CR-Hit as a function of the number of preceding CRs. In general, we obtain inconsistent results across mice that may be due to the small number of trials that have more than one preceding CR. Nevertheless, some mice have a trend, sometimes significant, in which CR-Hit responses are higher for longer CR preceding streaks. This is especially true during the learning phase. We have decided not to include this in the manuscript and present this figure only to the reviewer.

Author Response

Reviewer #1 (Public Review):

The central claim that the R400Q mutation causes cardiomyopathy in humans require(s) additional support.

We regret that the reviewer interpreted our conclusions as described. Because of the extreme rarity of the MFN2 R400Q mutation our clinical data are unavoidably limited and therefore insufficient to support a conclusion that it causes cardiomyopathy “in humans”. Importantly, this is a claim that we did not make and do not believe to be the case. Our data establish that the MFN2 R400Q mutation is sufficient to cause lethal cardiomyopathy in some mice (Q/Q400a; Figure 4) and predisposes to doxorubicin-induced cardiomyopathy in the survivors (Q/Q400n; new data, Figure 7). Based on the clinical association we propose that R400Q may act as a genetic risk modifier in human cardiomyopathy.

To avoid further confusion we modified the manuscript title to “A human mitofusin 2 mutation can cause mitophagic cardiomyopathy” and provide a more detailed discussion of the implications and limitations of our study on page 11).

First, the claim of an association between the R400Q variant (identified in three individuals) and cardiomyopathy has some limitations based on the data presented. The initial association is suggested by comparing the frequency of the mutation in three small cohorts to that in a large database gnomAD, which aggregates whole exome and whole genome data from many other studies including those from specific disease populations. Having a matched control population is critical in these association studies.

We have added genotyping data from the matched non-affected control population (n=861) of the Cincinnati Heart study to our analyses (page 4). The conclusions did not change.

For instance, according to gnomAD the MFN2 Q400P variant, while not observed in those of European ancestry, has a 10-fold higher frequency in the African/African American and South Asian populations (0.0004004 and 0.0003266, respectively). If the authors data in table one is compared to the gnomAD African/African American population the p-value drops to 0.029262, which would not likely survive correction for multiple comparison (e.g., Bonferroni).

Thank you for raising the important issue of racial differences in mutant allele prevalence and its association with cardiomyopathy. Sample size for this type of sub-group analysis is limited, but we are able to provide African-derived population allele frequency comparisons for both the gnomAD population and our own non-affected control group.

As now described on page 4, and just as with the gnomAD population we did not observe MFN2 R400Q in any Caucasian individuals, either cardiomyopathy or control. Its (heterozygous only) prevalence in African American cardiomyopathy is 3/674. Thus, the R400Q minor allele frequency of 3/1,345 in AA cardiomyopathy compares to 10/24,962 in African gnomAD, reflecting a statistically significant increase in this specific population group (p=0.003308; Chi2 statistic 8.6293). Moreover, all African American non-affected controls in the case-control cohort were wild-type for MFN2 (0/452 minor alleles).

(The source and characteristics of the subjects used by the authors in Table 1 is not clear from the methods.)

The details of our study cohorts were inadvertently omitted during manuscript preparation. As now reported on pages 3 and 4, the Cincinnati Heart Study is a case-control study consisting of 1,745 cardiomyopathy (1,117 Caucasian and 628 African American) subjects and 861 non-affected controls (625 Caucasian and 236 African American) (Liggett et al Nat Med 2008; Matkovich et al JCI 2010; Cappola et al PNAS 2011). The Houston hypertrophic cardiomyopathy cohort [which has been screened by linkage analysis, candidate gene sequencing or clinical genetic testing) included 286 subjects (240 Caucasians and 46 African Americans) (Osio A et al Circ Res 2007; Li L et al Circ Res 2017).

Relatedly, evaluation in a knock-in mouse model is offered as a way of bolstering the claim for an association with cardiomyopathy. Some caution should be offered here. Certain mutations have caused a cardiomyopathy in mice when knocked in have not been observed in humans with the same mutation. A recent example is the p.S59L variant in the mitochondrial protein CHCHD10, which causes cardiomyopathy in mice but not in humans (PMID: 30874923). While phenocopy is suggestive there are differences in humans and mice, which makes the correlation imperfect.

We understand that a mouse is not a man, and as noted above we view the in vitro data in multiple cell systems and the in vivo data in knock-in mice as supportive for, not proof of, the concept that MFN2 R400Q can be a genetic cardiomyopathy risk modifier. As indicated in the following responses, we have further strengthened the case by including results from 2 additional, previously undescribed human MFN2 mutation knock-in mice.

Additionally, the argument that the Mfn2 R400Q variant causes a dominant cardiomyopathy in humans would be better supported by observing of a cardiomyopathy in the heterozygous Mfn2 R400Q mice and not just in the homozygous Mfn2 R400Q mice.

We are intrigued that in the previous comment the reviewer warns that murine phenocopies are not 100% predictive of human disease, and in the next sentence he/she requests that we show that the gene dose-phenotype response is the same in mice and humans. And, we again wish to note that we never argued that MFN2 R400Q “causes a dominant cardiomyopathy in humans.” Nevertheless, we understand the underlying concerns and in the revised manuscript we present data from new doxorubicin challenge experiments comparing cardiomyopathy development and myocardial mitophagy in WT, heterozygous, and surviving (Q/Q400n) homozygous Mfn2 R400Q KI mice (new Figure 7, panels E-G). Homozygous, but not heterozygous, R400Q mice exhibited an amplified cardiomyopathic response (greater LV dilatation, reduced LV ejection performance, exaggerated LV hypertrophy) and an impaired myocardial mitophagic response to doxorubicin. These in vivo data recapitulate new in vitro results in H9c2 rat cardiomyoblasts expressing MFN2 R400Q, which exhibited enhanced cytotoxicity (cell death and TUNEL labelling) to doxorubicin associated with reduced reactive mitophagy (Parkin aggregation and mitolysosome formation) (new Figure 7, panels A-D). Thus, under the limited conditions we have explored to date we do not observe cardiomyopathy development in heterozygous Mfn2 R400Q KI mice. However, we have expanded the association between R400Q, mitophagy and cardiomyopathy thereby providing the desired additional support for our argument that it can be a cardiomyopathy risk modifier.

Relatedly, it is not clear what the studies in the KI mouse prove over what was already known. Mfn2 function is known to be essential during the neonatal period and the authors have previously shown that the Mfn2 R400Q disrupts the ability of Mfn2 to mediate mitochondrial fusion, which is its core function. The results in the KI mouse seem consistent with those two observations, but it's not clear how they allow further conclusions to be drawn.

We strenuously disagree with the underlying proposition of this comment, which is that “mitochondrial fusion (is the) core function” of mitofusins. We also believe that our previous work, alluded to but not specified, is mischaracterized.

Our seminal study defining an essential role for Mfn2 for perinatal cardiac development (Gong et al Science 2015) reported that an engineered MFN2 mutation that was fully functional for mitochondrial fusion, but incapable of binding Parkin (MFN2 AA), caused perinatal cardiomyopathy when expressed as a transgene. By contrast, another engineered MFN2 mutant transgene that potently suppressed mitochondrial fusion, but constitutively bound Parkin (MFN2 EE) had no adverse effects on the heart.

Our initial description of MFN2 R400Q and observation that it exhibited impaired fusogenicity (Eschenbacher et al PLoS One 2012) reported results of in vitro studies and transgene overexpression in Drosophila. Importantly, a role for MFN2 in mitophagy was unknown at that time and so was not explored.

A major point both of this manuscript and our work over the last decade on mitofusin proteins has been that their biological importance extends far beyond mitochondrial fusion. As introduced/discussed throughout our manuscript, MFN2 plays important roles in mitophagy and mitochondrial motility. Because this central point seems to have been overlooked, we have gone to great lengths in the revised manuscript to unambiguously show that impaired mitochondrial fusion is not the critical functional aspect that determines disease phenotypes caused by Mfn2 mutations. To accomplish this we’ve re-structured the experiments so that R400Q is compared at every level to two other natural MFN2 mutations linked to a human disease, the peripheral neuropathy CMT2A. These comparators are MFN2 T105M in the GTPase domain and MFN2 M376A/V in the same HR1 domain as MFN2 R400Q. Each of these human MFN2 mutations is fusion-impaired, but the current studies reveal that that their spectrum of dysfunction differs in other ways as summarized in Author response table 1:

Author response table 1.

We understand that it sounds counterintuitive for a mutation in a “mitofusin” protein to evoke cardiac disease independent of its appellative function, mitochondrial fusion. But the KI mouse data clearly relate the occurrence of cardiomyopathy in R400Q mice to the unique mitophagy defect provoked in vitro and in vivo by this mutation. We hope the reviewer will agree that the KI models provide fresh scientific insight.

Additionally, the authors conclude that the effect of R400Q on the transcriptome and metabolome in a subset of animals cannot be explained by its effect on OXPHOS (based on the findings in Figure 4H). However, an alternative explanation is that the R400Q is a loss of function variant but does not act in a dominant negative fashion. According to this view, mice homozygous for R400Q (and have no wildtype copies of Mfn2) lack Mfn2 function and consequently have an OXPHOS defect giving rise to the observed transcriptomic and metabolomic changes. But in the rat heart cell line with endogenous rat Mfn2, exogenous of the MFN2 R400Q has no effect as it is loss of function and is not dominant negative.

Our results in the original submission, which are retained in Figures 1D and 1E and Figure 1 Figure Supplement 1 of the revision, exclude the possibility that R400Q is a functional null mutant for, but not a dominant suppressor of, mitochondrial fusion. We have added additional data for M376A in the revision, but the original results are retained in the main figure panels and a new supplemental figure:

Figure 1D reports results of mitochondrial elongation studies (the morphological surrogate for mitochondrial fusion) performed in Mfn1/Mfn2 double knock-out (DKO) MEFs. The baseline mitochondrial aspect ratio in DKO cells infected with control (b-gal containing) virus is ~2 (white bar), and increases to ~6 (i.e. ~normal) by forced expression of WT MFN2 (black bar). By contrast, aspect ratio in DKO MEFs expressing MFN2 mutants T105M (green bar), M376A and R400Q (red bars in main figure), R94Q and K109A (green bars in the supplemental figure) is only 3-4. For these results the reviewer’s and our interpretation agree: all of the MFN2 mutants studied are non-functional as mitochondrial fusion proteins.

Importantly, Figure 1E (left panel) reports the results of parallel mitochondrial elongation studies performed in WT MEFs, i.e. in the presence of normal endogenous Mfn1 and Mfn2. Here, baseline mitochondrial aspect ratio is already normal (~6, white bar), and increases modestly to ~8 when WT MFN2 is expressed (black bar). By comparison, aspect ratio is reduced below baseline by expression of four of the five MFN2 mutants, including MFN2 R400Q (main figure and accompanying supplemental figure; green and red bars). Only MFN2 M376A failed to suppress mitochondrial fusion promoted by endogenous Mfns 1 and 2. Thus, MFN2 R400Q dominantly suppresses mitochondrial fusion. We have stressed this point in the text on page 5, first complete paragraph.

Additionally, as the authors have shown MFN2 R400Q loses its ability to promote mitochondrial fusion, and this is the central function of MFN2, it is not clear why this can't be the explanation for the mouse phenotype rather than the mitophagy mechanism the authors propose.

Please see our response #7 above beginning “We strenuously disagree...”

Finally, it is asserted that the MFN2 R400Q variant disrupts Parkin activation, by interfering with MFN2 acting a receptor for Parkin. The support for this in cell culture however is limited. Additionally, there is no assessment of mitophagy in the hearts of the KI mouse model.

The reviewer may have overlooked the studies reported in original Figure 5, in which Parkin localization to cultured cardiomyoblast mitochondria is linked both to mitochondrial autophagy (LC3-mitochondria overlay) and to formation of mito-lysosomes (MitoQC staining). These results have been retained and expanded to include MFN2 M376A in Figure 6 B-E and Figure 6 Figure Supplement 1 of the revised manuscript. Additionally, selective impairment of Parkin recruitment to mitochondria was shown in mitofusin null MEFs in current Figure 3C and Figure 3 Figure Supplement 1, panels B and C.

The in vitro and in vivo doxorubicin studies performed for the revision further strengthen the mechanistic link between cardiomyocyte toxicity, reduced parkin recruitment and impaired mitophagy in MFN2 R400Q expressing cardiac cells: MFN2 R400Q-amplified doxorubicin-induced H9c2 cell death is associated with reduced Parkin aggregation and mitolysosome formation in vitro, and the exaggerated doxorubicin-induced cardiomyopathic response in MFN2 Q/Q400 mice was associated with reduced cardiomyocyte mitophagy in vivo, measured with adenoviral Mito-QC (new Figure 7).

Reviewer #2 (Public Review):

In this manuscript, Franco et al show that the mitofusin 2 mutation MFN2 Q400 impaires mitochondrial fusion with normal GTPase activity. MFN2 Q400 fails to recruit Parkin and further disrupts Parkin-mediated mitophagy in cultured cardiac cells. They also generated MFN2 Q400 knock-in mice to show the development of lethal perinatal cardiomyopathy, which had an impairment in multiple metabolic pathways.

The major strength of this manuscript is the in vitro study that provides a thorough understanding in the characteristics of the MFN2 Q400 mutant in function of MFN2, and the effect on mitochondrial function. However, the in vivo MFN2 Q/Q400 knock-in mice are more troubling given the split phenotype of MFN2 Q/Q400a vs MFN2 Q/Q400n subtypes. Their main findings towards impaired metabolism in mutant hearts fail to distinguish between the two subtypes.

Thanks for the comments. We do not fully understand the statement that “impaired metabolism in mutant hearts fails to distinguish between the two (in vivo) subtypes.” The data in current Figure 5 and its accompanying figure supplements show that impaired metabolism measured both as metabolomic and transcriptomic changes in the subtypes (orange Q400n vs red Q400a in Figure 5 panels A and D) are reflected in the histopathological analyses. Moreover, newly presented data on ROS-modifying pathways (Figure 5C) suggest that a central difference between Mfn2 Q/Q400 hearts that can compensate for the underlying impairment in mitophagic quality control (Q400n) vs those that cannot (Q400a) is the capacity to manage downstream ROS effects of metabolic derangements and mitochondrial uncoupling. Additional support for this idea is provided in the newly performed doxorubicin challenge experiments (Figure 7), demonstrating that mitochondrial ROS levels are in fact increased at baseline in adult Q400n mice.

While the data support the conclusion that MFN2 Q400 causes cardiomyopathy, several experiments are needed to further understand mechanism.

We thank the reviewer for agreeing with our conclusion that MFN2 Q400 can cause cardiomyopathy, which was the major issue raised by R1. As detailed below we have performed a great deal of additional experimentation, including on two completely novel MFN2 mutant knock-in mouse models, to validate the underlying mechanism.

This manuscript will likely impact the field of MFN2 mutation-related diseases and show how MFN2 mutation leads to perinatal cardiomyopathy in support of previous literature.

Thank you again. We think our findings have relevance beyond the field of MFN2 mutant-related disease as they provide the first evidence (to our knowledge) that a naturally occurring primary defect in mitophagy can manifest as myocardial disease.

Author Response

Reviewer #2 (Public Review):

This manuscript reports on an important study that aims to identify symptom trajectories for the early detection of pancreatic cancer. The study's findings are based on the analysis of two complementary data sources: structured data obtained from the Danish National Patient Registry and unstructured information extracted from the free-text sections of patient notes. The researchers successfully identified various symptoms and disease trajectories that are strongly associated with pancreatic cancer, with compelling evidence from both data sources. Additionally, the study provides a detailed comparison and contrast of the results obtained from each data source, adding valuable insights into the strengths and limitations of each method.

Strengths:

The work is well motivated by the urgent need for early detection of pancreatic cancer, which is often difficult due to the lack of effective (computational) methods. The manuscript is generally well-written and includes relevant studies, providing a comprehensive overview of the current state of the field.

One of the unique contributions of this work is its use of both structured registry data and unstructured clinical notes to leverage complementary information. This approach enables a more nuanced and comprehensive understanding of the disease symptom trajectories, which is critical for improving early disease diagnosis and prognosis.

The methodology employed in this study is sound and robust, and the authors have candidly discussed its limitations. The results are significant and highlight previously unknown insights into symptom disease trajectories, which have important implications for the management of pancreatic cancer.

Overall, this is a well-designed and executed study that makes an important contribution to the field of cancer/informatics research, and it should be of great interest to both researchers and clinicians.

Weaknesses:

To complement the results in Figure 1, I'd also suggest that the authors compile a list of the most common (known) symptoms of pancreatic cancer as a reference. In other words, not only can you compare results found from the two sources but also compare them with existing knowledge. This is something you discussed partly in lines 245 but including this early as part of the results in Figure 1 would be more informative.

We agree that this would be informative to include into the Venn diagram. Hence, we have created a list of the most established and well-known symptoms of pancreatic cancer (Supplementary table S1) and converted these to the comparable ICD-10 level that we also use for the text mining and registry counts in Fig. 1. We have included the Venn diagram as Supplementary Figure S1.

In terms of the text mining evaluation results, providing information on recall errors would be beneficial to better understand the performance of the method. Additionally, line 144 mentions 53 words, but it is still not clear to me what these words refer to. Could you please clarify this point or provide more context?

We have added sensitivity/recall measures on the text mining procedure and furthermore added two references in the Discussion of the Tagcorpus program which was used for text mining the clinical notes. These references also mention similar sensitivities for the studies. The 53 words are false positives and we have clarified why these have been captured as false positives by the Tagcorpus (negations).

The disparities between Figure 2A and 2B are noteworthy, from very different initial symptoms to the proportion of short median survival dates (<=90 days), with much more pronounced differences than those observed in Figure 1 comparing two data sources. The highlighted trajectories are almost completely different. Should this be expected? I was hoping to see at least some overlap between the two results.

After updating the case population (via the cancer registry) and showing only symptoms trajectories in this revised version, we can clearly see that the trajectories are more similar. This gives an indication that the methods pick up on similar pancreatic-cancer symptoms, but there are also differences that show how each data type can complement the other, such as the text-mined trajectories being able to pick up longer symptom trajectories prior to the cancer.

All trajectories shown in Figure 2 include three symptoms. Is this by design? Could there be meaningful trajectories with different numbers of symptoms (e.g. 4 or more)?

We agree and have added the significant length 4 trajectories (for the registry data) as supplementary figure S2. No trajectories with length 5 or higher were found in the registry-based analysis. No length 4 (or higher) trajectories were found for the text-mined patients (presumably due to the data set size).

Considering those patients with both clinical notes and registry data, it may be beneficial to merge their symptoms to generate more informative trajectories.

This could be interesting but is out of scope for this paper. Here we would like to stress the proof-of-concept that the two data types can complement each other. The next steps would be to generate these multimodal trajectories to for example test if they are predictive of pancreatic cancer. Nonetheless, we acknowledge the significance of this perspective and have incorporated it into the Discussion section of the manuscript.

Given that results from two sources are being compared in Figures 1 and 2, have you considered calculating the top 20 most significant symptoms from the registry data as well?

We have done this and added them to Supplementary figure S3.

While there is a discussion related to cardiovascular diseases, I noticed no mention of cataracts or gonarthrosis, which were found to be prevalent among patients with short survival in Figure 2.

Since we now only include symptoms trajectories in the Results, we have chosen to not include these results in the Discussion for the final version of the manuscript. However, the diagnosis-wide trajectories are included in the Supplementary figure S2. Cataract and gonarthrosis have still been found significant in the results even though they are not shown in the Supplementary figure due to its visualization threshold of min. 400 patients per trajectory.

Ultimately, the goal of this research is to improve the early detection and prognosis of pancreatic cancer, thus it is important to discuss how the findings of this work could be applied in practice towards this goal (e.g. used by disease prediction algorithms?)

We agree that this is very important and have added a small section on this in the Discussion. We have also cited a recent publication using deep learning algorithms to predict pancreatic cancer based solely on registry data (Placido et al. 2023).

Author Response

Reviewer #1 (Public Review):

In general, in the discussion, I miss two of the main points that led to suspend screening programs in most countries during the pandemic:

1) protecting women from the risk of infection linked to attending a clinic during pandemic when health facilities were mostly attended by symptomatic people seeking care for Covid-19;

We agree. We have added this to the background and Discussion section (page 3, lines 76-78 & page 9, lines 296-299).

2) the of health professionals because they were mostly involved in covid related activities: lack of radiologists (addressed to the emergency department to assure diagnoses of pneumonia), lack of anesthesiologists (due to the expansion of intensive care), thus risking not having timely surgical treatment; lack of screening organization personal for invitations and phone calls (working on contact tracing).

We agree. We have added this to the background and Discussion section (page 3, lines 76-78 & page 9, lines 296-299).

Lacking the rationale for suspending screening, it is not clear to the reader how the Danish program afforded these issues and was able to maintain open the program.

We have elaborated on this in the Discussion section (page 296-299), arguing that Denmark may have partly decreased the issue of staff shortage due to e.g., a lower burden of COVID-19, use of laymen and medical student for testing and vaccinations and a high vaccine coverage.

Author Response

Reviewer #1 (Public Review):

Hoang, Tsutsumi and colleagues use 2-photon calcium imaging to study the activity of Purkinje cells during a Go/No-go task and related this activity to their location in Aldolase-C bands. Tensor component analysis revealed that a substantial part of the calcium responses can be linked to four functional components. The manuscript addresses an important question with an elegant technical approach and careful analysis. There are a few points that I think could be addressed to further improve the quality of the manuscript.

1) The authors should be careful not to overstate the goal and results. For instance, in the abstract it is stated that dynamical functional organization is necessary for dimension reduction. However, the statement that the 4 TCs together account for about half of the variance (line 220) indicates that dimensionality may not be reduced that much. I would suggest revising the first and last sentence of the abstract accordingly.

Dynamic functional organization of TC1 and TC2 by synchronization is the major finding of this study and we believe that it is one of the most efficient mechanisms of dimension reduction, given the unique anatomy of the cerebellum. In the revised manuscript, we added a supplemental result showing that the dimensionality of TC1 and TC2 neurons decreased and increased, respectively, in accordance with bi-directional changes in their synchronization (Figure 3 – figure supplement 1DE). Dimension reduction was further confirmed by conventional PCA (Figure 6 – figure supplement 1). However, we agree that the statement that the cerebellum reduces dimensions by self-organization of components is speculative, and we revised the abstract accordingly.

At the end of the introduction, the authors refer to "the first evidence supporting the two major theories of cerebellar function" but which two theories is referred to and how this manuscript support them is not very obvious. Similarly, they state that "This study unveiled the secret of cerebellar functional architecture", which I would consider to be an unnecessary overstatement of the impact of the work described.

In the revised Introduction, we explicitly stated that TC1 and TC2 are related to timing control and cognitive error learning, respectively, with some indirect causal evidence. We also revised the last paragraph of the Introduction to emphasize that this study provides the first evidence to support the view that distinct cerebellar components may serve divergent cerebellar functions in a single task. The statement "This study unveiled the secret of cerebellar functional architecture" was removed.

In the title, the authors use the word modular. In the consensus paper on cerebellar modules (Apps et al., 2018) an attempt is made to unify the terms used to describe cerebellar anatomical structures. Here "module" is used for the longitudinal zone of interconnected PCs, CN neurons and olivary neurons. As the authors only studied PC activity (and indirectly the IO), I would suggest using band, stripe or subpopulation instead.

Because we used TCA to identify functional components underlying the Go/No-go data, we changed the word “module” to “component” in the title.

Finally, the term "CF firing" or "CF activity" is used when referring to the recorded signals. However, the authors measure postsynaptic calcium responses that are indeed likely driven by CF inputs, but could also be influenced by PF inputs. At the very least, because Purkinje cells and not climbing fibers are being imaged, "complex spike" should be used instead. It would be more accurate still to use the more general "calcium response" and make less of an assumption about the origin of the calcium response.

In this study, CF-dependent dendritic Ca2+ signals in adjacent AldC compartments were recorded by the two-photon imaging. The HA_time algorithm (Hoang et al. 2020) was then applied to extract spike timings from the recorded signals. In the revised manuscript, we used the terms “calcium responses” and “complex spikes” when referring to the recorded Ca2+ signals and the estimated spikes, respectively.

2) For some figure panels and statements in the manuscript error bars or confidence intervals and statistics are missing. This is the case for, for example, the changes in fraction correct, lick latency, fraction incorrect, etc. (Fig 1B, 2E-F, TC levels in 3, 4D-E and 5A-C). Including these is particularly relevant in Fig 4E as this is a key result, mentioned also in the abstract. Please indicate clearly if these plots are cumulative for all mice or per mouse and averaged. I advise the authors to statistically support the claim that the changes are significant and in opposite direction as this element of the study is referred to in the abstract and discussion (summary).

We added the error bars / confidence intervals to the related figures. Most importantly, we added histograms of synchrony strength for TC1/TC2 neurons (Figure 4E) and conducted statistical tests to strengthen the claim of bi-directional changes in synchronization of TC1/TC2.

3) Data presentation sometimes does not do the work justice. For example, the data in Figure 6 are very interesting, but hard to read because of the design of the figure. It is clear how the components are mostly confined to Aldolase-C domains, but within the domains the distribution is not clear. I would advise to also more clearly indicate what the locations of the colors within the bands refers to. The spatial distribution of the selected top 300 cells for each TC could be added.

We added pie-chart plots for the fraction of TC1-4 neurons in each Ald-C zone and learning stage. We also indicated in the figure legend that the location of a single-color bar referred to the geographic distance of the corresponding neuron relative to Ald-C boundaries. We included spatial distribution of the selected neurons in Figure 4 – figure supplement 1D.

Author Response

Reviewer #1 (Public Review):

The authors investigate the mechanistic underpinning of paradoxical activation (PA) of RAF by small molecule kinase inhibitors using mathematical modeling. The main novelty of the study is the consideration of RAF conformational autoinhibition by its N-terminal regulatory domains as a new determinant of PA. This mechanism has not been explicitly considered in previous theoretical studies, which are based on two other mechanisms: drug-induced RAF oligomerization into active dimers (dimer potentiation DP) and negative cooperativity (NC) of inhibitor binding by a second monomer in the inhibitor-induced RAF kinase dimerization. An important discovery of this study is that conformational autoinhibition is a critical determinant of PA and that in some cases, it can contribute to PA in the absence of DP and NC. Another novelty is the consideration of RAF interaction with 14-3-3 proteins, as a determinant of PA. The 14-3-3 dimeric scaffolds play an important role in the regulation of both autoinhibited and active states of RAF and thus understanding how their interaction with RAF influences PA by RAF inhibitors is important. Using mathematical modeling the authors show that 14-3-3 binding does indeed enhance PA in response to a spectrum of RAF inhibitors.

We thank Reviewer #1 for reviewing our manuscript, and we agree with this summary.

Strengths

The overall strength of this study is that it increases the mechanistic understanding of how PA of RAF originates in response to its inhibitors. Consideration of the effect that the inhibitors play in breaking the autoinhibited conformation has been overlooked by previous mathematical analyses of PA, and this study bridges this gap. By doing so, the authors discover that breaking that autoinhibited state is in fact the biggest contribution to PAB by RAF inhibitors. In my opinion, this is the most impactful finding of this study, which additionally speaks to how important are the autoinhibitory mechanisms for constraining basal RAF signaling in cells. The presented analysis also shows that consideration of conformational autoinhibition can explain PA by all different types of RAF inhibitors (1, 1.5, and 2), which until now has been difficult to reconcile.

Another important contribution of this study is the investigation of how the 14-3-3 scaffold proteins can further contribute to PA. This is exciting, especially in light of recent elegant structural studies that unveiled complex regulation of RAF by 14-3-3 (which are both important for RAF inhibition and stabilization of the active dimers). The authors dissect these opposing roles of 14-3-3 in their model and show the autoinhibitory interaction with 14-3-3, but not the activating one, significantly increases the PA response. Their findings that an increase in the 143-3 levels amplifies PA is very interesting and somewhat provocative as it is unclear how much 14-3-3 levels in cells can oscillate. To this end, the authors show that elevated 14-3-3 levels are observed with increased time of RAF inhibitor treatment, which might point to a new mechanism of resistance to RAF inhibitors.

We thank reviewer #1 for the enthusiastic review and for highlighting the value of bringing conformational autoinhibition into the study and understanding of paradoxical activation. We also appreciate the positive consideration of the 14-3-3 section of the manuscript and the helpful suggestions later in the review. In this revision, we have taken the offered option of removing all of the 14-3-3 theoretical and experimental work. We plan to expand the 14-3-3 work in our ongoing work, in accordance with the thoughtful input from reviewers #1, #2, and #3 on this topic. Thank you.

Weaknesses

The main weakness of the study is the limited experimental analysis conducted to test the predictions that arise from the mathematical models. While some of these predictions might be challenging to test, the one which is tested is not tested rigorously. The experiments focus on 14-3-3-based regulation and are conducted in cells by observing the effect of 14-3-3 overexpression on the inhibition of RAF signaling by its different kinase inhibitors. While the authors acknowledge that too, 14-3-3 overexpression will have a multifaceted effect on signaling as these scaffold proteins participate in the regulation of almost all signaling events. Thus, the proposed experiments are not sufficient to conclude that the observed effects are in fact a result of 14-3-3/RAF interaction.

The authors consider conformational autoinhibition and 14-3-3 stabilization of autoinhibited RAF as two different mechanisms. While it is not a weakness, I am curious how accurate is the consideration of the autoinhibited state of RAF in the absence of 14-3-3. Is it known how the proportion of RAF in cells in its inactive state exists while not bound to 14-3-3?

We thank Reviewer #1 for this input on how we can significantly improve the 14-3-3 section of the manuscript. We have removed the 14-3-3 sections due to the consensus input of all three reviewers and the presented option of focusing on the theoretical results of how conformational autoinhibition influences PA. We do plan to continue this research program on beyond this manuscript, and we therefore very much appreciate these insights into which aspects should be supported with additional experiments and the challenges that follow from the pleiotropic activities of 14-3-3 proteins. The suggestion of quantifying the ratio of autoinhibited to non-autoinhibited forms of RAF when 14-3-3 proteins are present and absent is an experiment we plan to pursue in our future work. It will require us to learn new methods and/or to form new collaborations, and we therefore appreciate the consensus opinion that this would be outside of our current expertise and outside of the scope of the focused manuscript on modeling the impact of conformational autoinhibition on PA.

Reviewer #2 (Public Review):

In this study, the authors set out to investigate factors that have been neglected in existing mathematical models for the paradoxical activation (PA) of RAF by pharmacological inhibitors. The PA phenomenon is well known and is thought to be an important factor in limiting the effectiveness of RAF inhibitors. The authors primarily use mathematical models, first to examine the importance of conformational autoinhibition of RAF monomers, and later to investigate the potential role played by binding of 14-3-3 proteins to either autoinhibited monomers or active dimers. The authors develop several model variants containing different candidate mechanisms and generate analytical solutions that demonstrate under which parameter conditions PA may occur within these models. The use of analytical solutions is a strong point of the paper, as it allows evaluation of the models independently of specific parameter values. This analysis suggests that conformational autoinhibition is a very strong contributor to paradoxical activation, as models that include this mechanism show substantially larger concentration ranges under which RAF is activated by inhibitors. Fitting the parameters of the model to a published dataset on multiple inhibitors further suggests that conformational activation is important, as models containing this mechanism can fit the dataset with significantly lower error. Another interesting observation is that the different types of RAF inhibitors (1, 1.5, 2) fit the data with parameter values that are reasonably similar within each type. A moderate weakness in this analysis is that all of these observations provide indirect evidence for the importance of conformational autoinhibition. A direct test of whether PA is reduced when conformational autoinhibition is removed would be more compelling, but such a test could be difficult to set up experimentally.

We thank Reviewer #2 for reviewing our manuscript, and we agree with this summary. We agree that an experimental test where conformational autoinhibition is removed from the system would a very compelling experiment, but that it would be difficult to set up experimentally. We appreciate the option to focus on the theoretical advance in our revision, and we will be working toward such an experiment.

The authors then perform an analysis of how 14-3-3 binding to either autoinhibited monomers or active dimers might enhance PA. A new model is constructed that contains these binding events in the context of conformational activation, but without negative cooperativity or dimer potentiation included, for the sake of limiting complexity. These models implicate monomer binding, but not dimer binding as a contributor to PA. They follow up on this model result by overexpressing 14-3-3 proteins in two RAS-mutant cell lines, which leads to both higher baseline ERK phosphorylation and to a wider range of inhibitor-induced PA, as predicted by the model. A cell-based RAF dimerization assay also shows higher dimerization effects when 14-3-3 plasmids are transfected as well. This experimental evidence provides strong support for the model, although one drawback, which is noted by the authors in the discussion, is that 14-3-3 overexpression could potentially exert effects on RAF activity through pleiotropic effects other than the binding actions included in the model.

We thank Reviewer #2 for the input on the 14-3-3 section of the manuscript. Although it has been removed from the revision, all of the comments from the review will be helpful for our ongoing work.

Overall, this study makes a strong contribution to understanding the paradoxical effects of RAF inhibitors on the RAS/ERK signaling pathway, which remains a significant problem in the use of targeted inhibitors for cancer. Demonstrating that both conformational activation and 14-3-3 binding strongly contribute to the PA effect is an important step forward, as it establishes that these mechanisms should not be overlooked when designing strategies to use Raf inhibitors.

We appreciate the thoughtful review and helpful comments to improve the manuscript.

Reviewer #3 (Public Review):

The authors describe a mathematical and computational modeling study of RAF paradoxical activation (PA), a phenomenon in which RAF inhibitors exhibit a bell-shaped dose-response curve of Erk phosphorylation - activating signaling through wild-type RAF at low drug concentrations before inhibiting it at higher concentrations. They explore three distinct mechanisms that may contribute to PA - conformational autoinhibition, negative cooperativity, and drug-induced dimerization - and conclude that all three are required to best fit published data that show the PA phenomenon. They explore the effect of 14-3-3 binding to RAF both computationally and experimentally and reach the conclusion that 14-3-3 can potentiate the PA phenomenon via stabilization of the autoinhibited conformation.

We thank Reviewer #3 for reviewing our manuscript, and for the helpful comments in the review.

Strengths:

One key finding will be quite valuable to the field - that paradoxical activation can arise in the absence of negative cooperativity and without any effect of the inhibitor on the propensity of RAF to dimerize, provided that there exists a "conformationally autoinhibited" state that cannot dimerize and cannot bind inhibitor. This finding is important because negative cooperativity and dimer-induction have been a major focus - arguably the main focus - of prior studies of the phenomenon and also a source of considerable confusion. Inhibitors with very different chemical structures and binding properties - type 1.5 inhibitors that are dimer-breakers (and may or may not exhibit negative cooperativity) and type I and II inhibitors that can promote dimers (and almost certainly do not exhibit negative cooperativity) can nevertheless both exhibit PA. Thus the authors' modeling provides a unifying explanation - it is not dimerinduction or negative cooperativity that is at the root of PA, rather it is that there exists an autoinhibited state that can neither bind inhibitor nor dimerize. The authors further show that negative cooperativity and dimer-induction can act in concert with "conformational autoinhibition" to modify the PA response in a drug-specific manner.

We thank Reviewer #3 for highlighting these strengths and their value to the field. In the focused paper, we have updated our discussion of the fits and of the model to highlight these points better.

Weaknesses:

Unfortunately, the authors don't really explain in a straightforward manner what is going on with the conformational autoinhibition model (Figure 2A). One has to read carefully and all the way to section 3 of appendix 1 to piece it together. In short, what the math shows is that at least for certain ranges of parameter values, the presence of an inhibitor can increase the concentration of dimers, even when it does not change the equilibrium constant for dimer formation, and some of those dimers will have an active, drug-free protomer. This is because the inhibitor effectively traps open monomers, which can then capture drug-free open monomers to form active dimers (active in one subunit, inactive and drug-bound in the other). As inhibitor concentration increases, the pool of autoinhibited RAF is diminished, and eventually, it is shifted completely to fully inhibited dimers. But at low concentrations of inhibitor, there is a net increase in dimerized (active) but drug-free protomers (see figure on page 27 of the appendix). Voila, paradoxical activation, with no need to invoke negative cooperativity.

We apologize for the confusion, and agree that the description/walk through in the appendix should be featured more prominently in the manuscript. To this end, we have added a section to the main manuscript (titled “Paradoxical activation reflects a shifting balance of signaling complexes”) that includes the content that was previously in the appendix, and we have added a supplementary figure (Figure 2 – figure supplement 2) which includes the figures from the appendix. Thank you for your thorough review and working through the appendix, and we appreciate this suggestion.

Considering the potential for confusion around what is meant by "drug-induced dimerization" as an effect distinct from the effect of the drug in promoting RAF dimerization in their conformational autoinhibition model, it would have been helpful for the authors to explicitly address the distinction (drug-induced dimerization alters the equilibrium constant for dimerization; this is not a feature of the conformational autoinhibition model).

Thank you for this suggestion. We have clarified our text by rewriting it to read: … some RAF inhibitors have been shown to result in an increased level of RAF dimerization (Hatzivassiliou et al., 2010; Jin et al, 2017; Karoulia et al., 2016; Lavoie et al, 2013). This druginduced dimer potentiation is commonly thought of as manifesting in a higher affinity between RAF protomers when one (or both) are bound to a RAF inhibitor (Kholodenko, 2015).

Also, I am confused by Figure 3C. The figure shows, and the authors state in the text, that for type II inhibitors an f > ~1 indicates a propensity to break dimers. But type 1.5 inhibitors should break dimers, and Type I and II inhibitors should promote dimers (at least some Type I and II drugs have been shown to promote kinase dimers). Seems that the predictions of the model are inconsistent with experimental data, at least for some inhibitors.

We agree that discussing the fits, relating them to experimental data and current thinking in the field, is important. We have therefore significantly extended our discussion of the fits in Figure 3C in the Discussion of the text. The new text reads:

It has previously been difficult to reconcile PA for Type I.5 inhibitors, which are sometimes thought of as dimer breakers because they position the alpha-C helix in the “out” position (in contrast to Type I and Type II inhibitors). Studies with recombinant protein and analytic ultracentrifugation clearly found type I.5 inhibitors to predominantly be in the monomeric form (Lavoie et al., 2013). Within-cell assays have similarly found type I.5 inhibitors to promote dimerization less than other Type I and Type II RAF inhibitors (Hatzivassiliou et al., 2010; Peng et al., 2015; Thevakumaran et al, 2015), however, RAF inhibitors still appeared to promote some dimerization in those in-cell assays. 14-3-3 binding proteins, which can help stabilize RAF dimers, may help explain this discrepancy (Kondo et al., 2019; Liau et al, 2020; Park et al., 2019). For example, by promoting the non-autoinhibited form, a type I.5 inhibitorbound RAF monomer is more dimerization capable than an autoinhibited (and non-inhibitor bound) RAF monomer, and even if the affinity is reduced compared to a non-autoinhibited and non-inhibitor bound RAF monomer, 14-3-3 proteins may be able to bind and overcome the effect. As our model does not explicitly include 14-3-3 proteins, this effect may contribute to our parameter estimation process finding an elevated binding affinity for type I.5 bound RAF monomers.

Although negative cooperativity has been difficult to precisely measure experimentally, it has widely been assumed to be present to help explain the paradoxical activation caused by Type I.5 inhibitors that do not promote dimerization as strongly as other RAF inhibitors. Our best fit parameters did tend to have g values that were larger than 1, indicating that the model fit best when there was some negative cooperativity. This could suggest that negative cooperativity is more abundant than widely believed. Alternatively, the model without negative cooperativity was able to fit the data nearly as well as the full model that included negative cooperativity (i.e., Figure 3D). This may suggest that other processes not included in the model may be modulating paradoxical activation and the g parameter, as the only other term the model, is contributing to the models ability to account for these otherwise not included effects.

We found parameter sets that reproduced available, published, data in order to test our model and investigate the potential for it to help illuminate aspects of PA. The best fit parameter sets further support a role for conformational autoinhibition and its modulation by RAF inhibitors in PA. However, it is also important not to read too deeply into the fits. For example, the data for the type II inhibitors AZ-628, LY3009120, and TAK-632 had small total fold-change PA magnitudes, and our fits for them have even less PA. We anticipate that the model-fitting approach would converge to increasingly accurate estimates for the parameters as the set of data being fit to expands. Additionally, quantitative experimental measurements of the parameters being fit should also cascade to impact other parameters and result in better estimates (Gutenkunst et al, 2007).

A large part of the paper deals with the effect of 14-3-3 binding. In my view, this part of the manuscript is not particularly helpful. There is no evidence (that I am aware of) that 14-3-3 concentrations vary significantly, or that their variation affects RAF activity/signaling. Considering their abundance relative to RAF, and relatively high affinity for RAF, it is likely that both autoinhibited and active RAF are saturated with 14-3-3. (RAF that is not 14-3-3-bound is likely mostly bound to chaperones and not active). That said, the authors' conclusion (based on modeling) that 14-3-3 can increase the extent of paradoxical activation by stabilizing the autoinhibited state seems sensible, but hard to reconcile with their experimental result where they find increased basal signaling with 14-3-3 over-expression. It is also difficult to understand how increased 14-3-3 binding to RAF could lead to active RAF dimers that are not inhibited at 10-100 uM concentrations of potent RAF dimer inhibitors like LY3009120 (Fig. 5C). It seems more likely that 14-3-3 overexpression is promoting Erk phosphorylation in a manner that is (at least partially) Raf-independent. To their credit, the authors acknowledge this concern.

We thank Reviewer #3 for the helpful critique of the section on 14-3-3. Although we have cut this section as part of the consensus review and suggestions for how to proceed with the revision, these points are very helpful for us as we consider how to interpret the modeling and experimental results of this section, how it fits into what is known, and what we should investigate next. Thank you.

Finally, one comment regarding the presentation. The authors discuss conformational inhibition and 14-3-3 binding as if they are promoting and/or inducing paradoxical activation. This is pervasive in the paper, including in the title, and is distracting and potentially will mislead some readers. Obviously, it is RAF inhibitor that induces or promotes paradoxical activation. Conformational autoinhibition - mediated by 14-3-3 - is a feature of the system that makes paradoxical activation possible.

We completely agree. We have rephrased to avoid this interpretation and we apologize for not recognizing it previously. Thank you for catching this and noting it for us to fix. As examples of the revisions to address this point, the last sentence of our abstract now reads:

Overall, this work establishes conformational autoinhibition as a robust mechanism for RAFinhibitor driven PA based solely on equilibrium dynamics of canonical interactions that comprise RAF signaling and inhibition.

And as another example, the third to last sentence in our Introduction now reads:

Our modeling reveals that, under certain conditions, RAF autoinhibitory conformational changes and their modulation by RAF inhibitor binding can be sufficient to drive PA.

Lastly, we have a last paragraph in the discussion that summarizes and hypothesizes to generalization:

\Our analysis was motivated by RAF inhibitors and PA in RAS mutant cells treated with a RAF inhibitor. Our model, however, is generalizable to other systems that share the modeled features. We anticipate that PA will be observed for other proteins (a) that have a dynamic-equilibrium of conformations, (b) where not all conformations can dimerize, and (c) where drug binding the protein stabilizes one or more of the conformations that can dimerize. As dimerization and conformational autoinhibition are both common features for kinase regulation (Huse & Kuriyan, 2002; Lavoie et al, 2014), it seems reasonably to hypothesize PA will be observed for more kinases through modulation of the conformation and dimerization dynamic-equilibrium. Thank you for suggesting these changes.

Author Response

Reviewer #1 (Public Review):

This manuscript reports a study to investigate the reporting practices in three top cardiovascular research journals for articles published in 2019. The study was preregistered, which makes the intent and methodology transparent, and the authors also make their materials, data, and code open. While the preregistration and sample strategy is a strength, it suffers from a higher than expected number of non-empirical articles decreasing the sample size and thus inference that can be drawn. The author's focus was mainly on transparency of reporting and not on the actual reproducibility or replicability of the articles; however, the accessibility of data, code, materials, and methods is a prerequisite. While the authors were still able to draw inferences to their main objectives, they could not perform some of their proposed analyses because of a small sample size (due partly to the less than half empirical articles in their sample as well as the low number of papers with accessible information to code). One of the descriptive analyses they performed, the country level scores (Figure 6), in particular suffers from the small sample size and while the authors state indicates this in their manuscript I do not think it would be reasonable to include as it has the potential to be misinterpreted since so many are based on an n=1. Overall, I found the authors presentation and discussion clear and concise; however, a lack of a more in-depth discussion is an area to improve the current manuscript. The manuscript outlines opportunities for researchers, journals, funders, and institutions to improve the way cardiovascular research is reported to enable discovery, reuse, and reproducibility.

We appreciate the reviewer’s recognition of our pre-registration, methodology, and resource sharing and also their feedback regarding the small sample size of empirical research articles and need for a more in-depth discussion of the impacts of our study. We have now increased the number of empirical studies to a total of 393 out of 639 articles screened. We also agree that our study focuses more on transparency than reproducibility and replicability, and we have changed our title to reflect this. While the sample size of empirical papers has increased, a comparison of accessibility scores across countries continued to suffer from small sample size and we have removed it based on the recommendation of the reviewers. We have updated the Materials and Methods section to reflect our updated analyses, as well as included additional paragraphs on Limitations and Future Work in our Discussion to acknowledge future improvements that could be made to the accessibility score used in our study.

Reviewer #2 (Public Review):

This is a descriptive paper in the field of metascience, which documents levels of accessibility and reproducible research practices in the field of cardiovascular science. As such, it does not make a theoretical contribution, but it argues, first, that there is a problem for this field, and second, it provides a baseline against which the impact of future initiatives to improve reproducibility can be assessed. The study was pre-registered and the methods and data are clearly documented. This kind of study is extremely labour-intensive and represents a great deal of work.

I have a major concern about the analysis. It is stated that to be fully reproducible, publications must include sufficient resources (materials, methods, data and analysis scripts). But how about cases where materials are not required to reproduce the work? In line 128-129 it is noted that the materials criterion was omitted for meta-analyses, but what about other types of study where materials may be either described adequately in the text, readily available (eg published questionnaires), or impossible to share (e.g. experimental animals).

To see how valid these concerns might be, I looked at the first 4 papers in the deposited 'EmpricalResearchOnly.csv' file. Two had been coded as 'No Materials availability statement' and for two the value was blank.

Study 1 used registry data and was coded as missing a Materials statement. The only materials that I could think might be useful to have might be 'standardized case report forms' that were referred to. But the authors did note that the Registry methods were fully documented elsewhere (I am not sure if that is the case).

Study 2 was a short surgical case report - for this one the Materials field was left blank by the coder.

Study 3 was a meta-analysis; the Materials field was blank by the coder

Study 4 was again coded as lacking a Material statement. It presented a model predicting outcome for cardiac arrhythmias. The definitions of the predictor variables were provided in supplementary materials. I am not clear what other materials might be needed.

These four cases suggest to me that it is rather misleading to treat lack of a Materials statement as contributing to an index of irreproducibility. Certainly, there are many studies where this is the case, but it will vary from study to study depending on the nature of the research. Indeed, this may also be true for other components of the irreproducibility index: for instance, in a case study, there may be no analysis script because no statistical analysis was done. And in some papers, the raw data may all be present in the text already - that may be less common, but it is likely to be so for case studies, for instance.

A related point concerns the criteria for selecting papers for screening: it was surprising that the requirement for studies to have empirical data was not imposed at the outset: it should be possible to screen these out early on by specifying 'publication type'; instead, they were included and that means that the numbers used for the actual analysis are well below 400. The large number of non-empirical papers is not of particular relevance for the research questions considered here. In the Discussion, the authors expressed surprise at the large number of non-empirical papers they found; I felt it would have been reasonable for them to depart from their pre registered plan on discovering this, and to review further papers to bring the number up to 400, restricting consideration to empirical papers only - also excluding case reports, which pose their own problems in this kind of analysis.

A more minor point is that some of the analyses could be dropped. The analysis of authorship by country had too few cases for many countries to allow for sensible analysis.

Overall, my concern is that the analysis presented here may create a backlash against metascientific analyses like this because it appears unfair on authors to use a metric based on criteria that may not apply to their study. I am strongly in favour of open, reproducible science, and agree it is important to document the state of the science for different disciplines. But what this study demonstrates to me is that if you are going to evaluate papers as to whether they include things like materials/data/ availability statements, then you need to have a N/A option. Unfortunately, I suspect it may not be possible to rely on authors' self-evaluation of N/A and that means that metascientists doing an evaluation would need to read enough of the paper to judge whether such a statement should apply.

We thank the reviewer for the time taken to review our paper, the appreciation of the work we conducted, and for the suggestions for improving our research methods. To address the initial concern about our analytical approach, the definition for fully reproducible publications that we used was only applicable to research that utilized empirical research methods. We recognize that publications such as editorials and reviews are not inherently reproducible experimental studies; thus, such papers were not provided with an accessibility score, were only screened for the components such as funding and conflict of interest information, and were only compared amongst each other. Additionally, articles such as meta-analyses and systematic reviews that do not include materials had adjusted accessibility scores. We expanded our Methods and Discussion section to further explain our screening process and our assumption that all empirical research articles contain methods, data, and analysis scripts and to acknowledge the limitations of our approach. We also agree that screening more empirical research articles is more in line with the intent of our pre-registration and we expanded the number of empirical research articles screened to 393. We also agree with the reviewer that the analysis by country should be excluded because of the small sample size for most countries, and we have adjusted the manuscript accordingly.

Author Response

We thank the reviewers for their insightful comments, which raise several important points regarding our study. As the reviewers have recognised, we introduced a number of simplifications in order to perform this complex optimisation problem, such as by restricting the analysis to a single intervention (insecticide-treated nets) and modelling countries at a national level. Despite their clear relevance to the study, computationally it would not have been feasible to run the multitude of scenarios suggested by reviewer 1, which we recognise as a limitation. As such we agree with the assessment that this study primarily represents a thought experiment to assess whether current policies are aligned with an optimal allocation strategy or whether there might be a need to consider alternative strategies. The findings are relevant primarily to global funders and should not be used to inform individual country allocation decisions. This perspective also underlies our decision to start the analysis from a baseline of year 2000 as opposed to modelling the current 2023 malaria situation: the largest international donor (the Global Fund) also uses baseline malaria levels in the period 2000-2004 as the basis of their allocation calculations (The Global Fund, Description of the 2020-2022 Allocation Methodology, December 2019). A simplified version of this method is represented by our “proportional allocation” strategy. We will further address these points in a revised manuscript and detailed responses to the reviewer comments.

Author Response

Reviewer #2 (Public Review):

Machold and colleagues develop and describe an intersectional genetic mouse (Id2Cre:Dlx5/6FlpE) that allows for the targeting of a cortical interneuron subpopulation predominantly consisting of the neurogliaform cell subtype (NGFCs). The strategy is a modification of that previously published by the authors (Id2cre:Nkx2-1Flpo; Valero et al., 2021) in which a subset of deep layer 6 NGFCs with distinct embryonic origins were targeted. Conversely, using the NDNF transgenic mouse lines previous studies, including thosefrom the Rudy laboratory, have clearly shown the prevalence of NGFCs in the outermost cortical Layer 1 region. Thus, the Id2Cre:Dlx5/6FlpE mouse poses an advantage over these previous approaches permitting the targeting of NGFCs in Layers 2-5. NGFCs in these regions have been hitherto difficult to study in an expedited manner.

The manuscript is of the resource/toolbox type and the authors are thorough in their description of the distribution and molecular characteristics of the ID2 neurons labelled by this intersectional approach. Furthermore, the authors perform a series of in vivo experiments. These entail the identification of NGFCs, the assessment of their influence on other neuronal populations, and the ability to delineate their activity during various network and behavioral states. Indeed, the authors reveal an activity pattern that is unique to NGFCs across epochs of specific network states. Therefore, this clearly demonstrates the applicability of the ID2Cre:Dlx5/6Flpe mouse to study the role of L2-5 NGFCs in a whole brain setting and these in vivo experiments constitute a major strength of the current study.

However, as with many transgenic mice, they are not always perfect, and the authors are very transparent regarding the additional, albeit a relatively smaller number of reported non-NGFCs particularly those of the CCK IN subtype. Indeed, clear morpho- functional divergence is revealed by the authors between these ID2 IN subpopulations. Furthermore, it is possible that this variability may differ across varying cortical regions. Thus, careful consideration of this caveat is necessary when using this mouse for future in vitro and in vivo studies. Related to this matter is a concern regarding the framing of the manuscript. The authors term the ID2 mixed population as the "4th group" since they do not express PV, SST, and VIP. One could argue this is a matter of semantics but to combine IN types that display distinct morphological and physiological properties into a single "group" based on one molecular feature is not consistent with that proposed by the widely accepted Petilla terminology (Ascoli et al., 2008).

We agree that the definition of “group” here for INs delineated by the molecular markers PV, SST, VIP and Id2 is oversimplified, but in practice, the use of the corresponding genetic tools (e.g., Pvalb-Cre, Sst-Cre etc.) has resulted in widespread adoption of this marker-based organization of IN diversity. For example, PV+ INs targeted with PV-Cre encompass both basket cells and chandelier cells that while sharing some electrophysiological properties (e.g., fast-spiking behavior) are completely distinct morphologically, and innervate different subcellular compartments (soma vs. axon initial segment). The same is true for SST INs, in that there appear to be at least three main subtypes – Martinotti, non-Martinotti, and long range projecting – each with distinct axonal projections and electrophysiology. Thus, while the molecular targeting approaches developed to date have greatly facilitated functional studies of IN subtypes, they have prioritized marker expression over the other aspects of IN diversity outlined in the Petilla framework.

Of interest to many who investigate cortical INs is the ability to genetically target specific subtypes during development. To this end, a potential and welcome addition to the manuscript would be an analysis (perhaps restricted to distribution/molecular characterization) highlighting whether the Id2cre:Dlx5/6Flpe strategy allows genetic access to layer 2-5 NGFCs during postnatal development following maternal tamoxifen administration.

We agree that a method to target NGFC at early postnatal ages would be useful; however, the expression of Id2 is dynamic during development, and is robust in ventricular zone progenitors at embryonic stages (Neuman et al., 1993 Dev. Biol. PMID 8224536) so maternal tamoxifen administration is likely to result in nonspecific labeling. Furthermore, we found that multiple doses of tamoxifen were necessary to achieve decent labeling of the Id2 IN population in adult animals, a protocol that would be difficult to perform in pregnant dams or early postnatal animals due to pup lethality.

Regardless, the experiments in the current study are, in general, well performed and clearly presented with the authors' conclusions supported by the results. Thus, it is clear that further refinements to genetic strategies are obviously required to exclusively target NGFCs throughout the cortical depth. Nevertheless, in the interim, the approach described in this current manuscript will be of use to the neuroscience community and help to further unravel the physiological role of this relatively understudied neuronal subtype.

Author Response

Reviewer #3 (Public Review):

Because of the position of pigeon embryos in eggs, light exposure will only stimulate the right eye, leading to lateralisation of brain responses and behaviour. Lorenzi and colleagues injected manganese chloride into pigeon eggs, to assess neuronal activation in the embryonic brain. While the eggs were placed in the light or dark, manganese ions accumulated in neurons that were activated (in cell bodies and axons), which was then visualized with MRI of the embryos before hatching. The authors report lateralisation of neuronal activity in three brain regions, which could potentially be important for our understanding of experience-dependent development of lateralised neural activation.

The tectofugal pathway in pigeons projects from the retina to the optical tectum, then to the nucleus rotundus in the thalamus, and then to the entopallium. The thalamofugal pathway projects from the retina to the GLd in the thalamus, and then to the wulst in the hyperpallium. The two pathways involve different thalamic nuclei (e.g., Deng 2006). In the methods and throughout the manuscript it should be specified which thalamic region is used as ROI.

Here we refer to the Gld in the thalamofugal visual pathway, we did not estimate activity in the n. rotundus. We have now clarified this point in the revised MS (ll. 54, 80, 86).

This manuscript only describes neural activity, but the MEMRI technique should also be used to assess the effect of experimental manipulations on axonal connectivity. It is important to learn about the asymmetry of contralateral projections in the light vs dark groups for answering the research question.

Here we used systemic administration of Mn through the CAM. The Blood Brain Barrier at this embryonic stage is not completely developed and its permeability to ions and small molecules is way higher in embryo than in later stages of development (Engelhardt, B. (2003). Development of the blood-brain barrier. Cell and tissue research, 314(1), 119-129.). Other studies involving direct, local injection in selected brain regions are more apt to investigate connectivity, but this is not the protocol used here. We appreciate the reviewer’s suggestion, and this will be the object of future experiments. However, we would like to disseminate the current protocol and the results it led to at an early stage to enable and encourage its use by other researchers in the field.

There is an overinterpretation of post-hoc statistics that are reported without correction for multiple testing. The wulst light group lateralization is probably not actually different from zero (uncorrected p=0.04).

We considered the reviewer's observation regarding the need for improvements in the statistical methods. In response, we have made amendments to the relevant section of the manuscript, explicitly stating that significant findings were obtained using a two-way ANOVA. For comparisons between conditions within specific brain regions, we conducted two-sample t-tests, and the results were corrected for Type I errors using the false discovery rate (FDR) method. Post-hoc one-sample t-tests were employed to assess lateralization across brain regions and conditions, and the corresponding p-values were reported without correction for multiple comparisons (as explicitly reported in the text, to avoid any confusion).

The first line in the discussion states that there is thalamofugal lateralization, but no lateralization in the tectofugal pathway. To my understanding, previous literature reported it the other way around: in altricial pigeons, light exposure in the egg mainly affected the tectofugal pathway (Deng & Rogers 2002), while the thalamofugal pathway in pigeons was not lateralized (Strockens et al., 2013). The manuscript should compare the current findings with the literature and discuss differences.

We are aware of the substantial differences in brain lateralization of the two visual pathways between pigeons and chicks after embryonic light exposure. However, in the present work we employed chick embryos (Gallus gallus domesticus), and the space limitations of a Brief Communication do not allow for an in-depth discussion of these differences between avian species.

Moreover, the tectum is the only region shown here from the tectofugal pathway. However, lateralization of contralateral connections is expected from tectum to the nucleus rotundus in the thalamus, and thus lateralization of activation may only arise in downstream brain regions from the optical tectum. Therefore, the conclusion that there is no lateralization in the tectofugal pathway is not supported by the data.

In conclusion, I think it is interesting and worthwhile that the authors assessed neural activity in response to visual stimulation in the embryo prior to hatching, but multiple methodological weaknesses and unclarities should be addressed.

The ROI that we here named Thalamus does not include the nucleus rotundus, but is referring to the nucleus geniculatus lateralis (Gld). We have now clarified this point in the revised MS (ll. 54, 80, 86), and we now refer only to the tectum, without generalizing to the entire tectofugal pathway, which will be the subject of future investigations.

Author Response

Reviewer #3 (Public Review):

This manuscript proposes to tackle a very interesting and methodologically challenging topic: the mechanistic underpinnings of neural specialization in the infant brain. The authors presented 4- to 7-month-old infants with social and non-social stimuli while their neural, hemodynamic, and metabolic activity was monitored, and they report a complex pattern of relationships between neural and metabolic or hemodynamic responses during social processing on the one hand, and during non-social processing on the other hand.

The approach described in this manuscript is very interesting and the combined use of EEG and bNIRS data appears very promising. However, there is some confusion between the initial aims of the study, and the analyses performed, which jeopardizes the clarity and the impact of this manuscript. Besides, the predictions of the authors are often underspecified which complexifies the interpretation of the results.

Based on its abstract, the goal of this work is to "combine simultaneous measures of coordinated neural activity metabolic rate and oxygenated blood supply to measure emerging specialization in the infant brain". The introduction nicely elaborates on the "interactive specialization theory" and the potential role of the interplay between brain energy consumption and neural activity in the emergence of functionally specialized brain regions during development. The authors present a novel multimodal approach, with potentially important implications for the study of brain specialization as a function of experience or maturation. Yet the experimental procedure presented in this manuscript only assesses specialized brain activity in response to social processing in 4- to 7-month-old infants, using multimodal neuroimaging.

Indeed, the authors presented 4- to 7-month-old infants with social and non-social stimuli while their neural, hemodynamic, and metabolic activity was monitored. The authors report significant differences between the two conditions in terms of neural activity in the delta, alpha, beta, and gamma bands; as well as in the pattern of hemodynamic to metabolic coupling. Using a GLM approach, the authors report on fNIRS channels and EEG sensors showing significant relationships between the evoked neural activity in the beta and gamma frequency bands, and each of the bNIRS signals (HbO, HbR & CCO), in the social and in the non-social conditions. The authors identify a particular fNIRS channel overlaying posterior STS, showing a positive relationship between Pz EEG beta activity and HbO, as well as CCO, together with a negative relationship between that same neural activity and HbR, in the social condition. This pattern of activity was not observed in the non-social condition.

Overall, these results indicate differential neural responses to social and non-social stimuli, coupled metabolic and hemodynamic activity in response to social as well as nonsocial stimuli.

These results additionally indicate coordinated metabolic, hemodynamic, and neural responses in brain regions selective for social processing, but it does not allow us to conclude that this coordinated activity is actually related to the functional specialization process (e.g. last sentence of the abstract).

We would like to thank the reviewer for their detailed comments. Based on their suggestions, we have made several changes to the manuscript. This study was the first to combine EEG and broadband NIRS and therefore served as a proof of principle study. At the onset of this work, there were many elements to develop such as the technical aspect of simultaneous bNIRS – EEG measurements as well as the methodology to combine the signals from both techniques with such different time resolutions. Therefore, we focused on one age group of infants rather than performing a study involving multiple age groups. The 4-to-7-month-old age group has been studied extensively using fNIRS, particularly to look at social brain development using similar stimuli as those used in the present study. Previous studies have demonstrated that social selectivity can be detected at 4 – 8 months of age (Grossmann et al., 2010; Lloyd-Fox et al., 2012, 2013, 2017). As this was a proof of principle study, we wanted to ensure that we were able to replicate results from previous studies with this new methodology. We therefore used one age group of 4-to-7-months. This has also been added to the introduction of the manuscript to provide clearer reasoning for using this age group.

The reviewer is correct that the current study does not provide direct evidence of developmental change in functional specialisation or the hypothesised interactive process through which functional specialisation may occur. Rather, we are measuring the status of functional specialisation (the idea that different areas in the brain are specialised for different functions) at the age we study, by testing whether the signals we observe are selective to social but not non-social stimuli. We have reframed the abstract and introduction of the manuscript to ensure this is clear, and we additionally now focus more on the methodology developed to answer such questions. Future studies can leverage our methodology to study different age groups to establish how the relationships between neural and vascular/metabolic signals changes over developmental time, which may provide greater insight into the specialisation process.

Grossmann, T., Oberecker, R., Koch, S. P., & Friederici, A. D. (2010). The Developmental Origins of Voice Processing in the Human Brain. Neuron, 65(6), 852–858. https://doi.org/https://doi.org/10.1016/j.neuron.2010.03.001

Lloyd-Fox, S., Begus, K., Halliday, D., Pirazzoli, L., Blasi, A., Papademetriou, M., Darboe, M. K., Prentice, A. M., Johnson, M. H., Moore, S. E., & Elwell, C. E. (2017). Cortical specialisation to social stimuli from the first days to the second year of life: A rural Gambian cohort. Developmental Cognitive Neuroscience, 25, 92–104. https://doi.org/10.1016/j.dcn.2016.11.005

Lloyd-Fox, S., Blasi, A., Elwell, C. E., Charman, T., Murphy, D., & Johnson, M. H. (2013). Reduced neural sensitivity to social stimuli in infants at risk for autism. Proceedings of the Royal Society B: Biological Sciences, 280(1758), 20123026. https://doi.org/10.1098/rspb.2012.3026

Lloyd-Fox, S., Blasi, A., Mercure, E., Elwell, C. E., & Johnson, M. H. (2012). The emergence of cerebral specialization for the human voice over the first months of life. Social Neuroscience, 7(3), 317–330. https://doi.org/10.1080/17470919.2011.614696

Another weakness of this manuscript relates to the unclear or underspecified motivations behind some of the performed analyses. For example, the authors contrast brain responses to social vs. baseline, non-social vs. baseline, and social vs. non-social. For clarity in the manuscript, the authors should specify the motivation behind each of these contrasts and their predictions.

We thank the reviewer for their suggestion. We have added the predictions for each of the analyses in the introduction section, lines 436 – 527. We have removed the “social minus non-social” comparison for the EEG topographical maps from Figure 2 as there was no value added by including this comparison.

Another example is in the analysis of the hemodynamic and metabolic coupling analysis, here the authors analyze only the social vs. baseline and non-social vs. baseline contrast, and they do not analyze the social vs non-social contrast. It would be useful for the reader to understand why only these two contrasts are performed and not the social vs. non-social, and what are the predictions of the authors.

We have now added this into the manuscript and the results can be seen in Figure 3c. We have clarified our predictions both at the end of the introduction (lines 436 - 527) and at the beginning of the discussion (lines 685 – 755).

The following has been added to the introduction:

For EEG, we expected an increase in neural activity in response to the social condition and a decrease in neural activity in response to the non-social condition. Based on previous work, this was expected to be strongest in the theta frequency band [3]. Moreover, for the combined bNIRS-EEG analyses, we hypothesised differentiated haemodynamic/metabolic coupling with neural activity for the social and non-social stimulus conditions. We performed two types of statistical tests: a) individual comparisons of the social and non-social conditions and b) comparison of the social condition versus the non-social condition. The individual condition tests were performed to show the scale and spatial location/sensitivity of the coupling between haemodynamics/metabolism and neural activity for each condition. Meanwhile, the social versus non-social comparison was performed to show where there was a significant difference in the coupling between the two conditions. With comparison (a) we aimed to identify regions involved in the processing of social and non-social stimuli by identifying the regions where the coupling was significant. With comparison (b) we aimed to identify regions where coupling was significantly different between conditions. We predicted that for the individual comparison of the social condition, we would observe positive associations between bNIRS and EEG measures, i.e. coordinated increases in haemodynamics/metabolism and neural oscillatory activity in the beta and gamma frequency bands (based on previous combined EEG – fMRI studies [16], [18]–[21], [23], [30]) which would be localised to core social brain regions. We hypothesised that for the non-social condition, over the same brain regions, positive associations would be observed between bNIRS and EEG measures, but they would be coordinated decreases in haemodynamics/metabolism and oscillatory activity. We also expected coordinated increases in haemodynamics/metabolism and oscillatory activity localised to the parietal brain region. These predictions are based on our previous work [29] where we demonstrated that stronger coupling between haemodynamics and metabolism was observed in the temporo-parietal regions for the social condition and in parietal region for the non-social condition which is known to play an important role in object processing [31], [32]. For the social versus the non-social contrast, we predicted that haemodynamic activity and metabolism would be coupled with neuronal oscillatory activity more strongly for the social stimuli in comparison to the non-social stimuli, with significant differences being observed in the temporo-parietal regions.

The following has been added to the discussion:

As a proof of principle, we examined the relationship between these measures to identify regional selectivity to social versus non-social stimuli. To first demonstrate the scale and spatial sensitivity of the coupling between haemodynamic/metabolic activity and neuronal oscillatory activity, comparisons were performed individually for the social and non-social conditions. For this, we predicted coordinated increases in haemodynamics/metabolism and neural activity in the beta and gamma frequency band. We predicted that for the social condition this would be localised to the core social brain regions (temporo-parietal region) while for the non-social condition, we expected the coupling to be localised to parietal regions, known to be involved in object processing [31], [32]. We additionally expected coordinated decreases in haemodynamic/metabolic activity and neural activity over the temporo-parietal region for the non-social condition, in accordance with our previous work [29]. Next, to demonstrate differential coupling for social and non-social stimuli, we performed a comparison of the social condition versus the non-social condition. For this, we hypothesised that in the beta and gamma frequency bands, there would be stronger coupling between haemodynamics/metabolism and neural activity for the social condition over the temporo-parietal region.

Finally, the core result of this work derives from the final GLM analysis which relates EEG activity to hemodynamic or metabolic responses. This analysis implies the inspection of interactions between 3 neuroimaging modalities, with 4 EEG measures, 2 hemodynamic measures, and 1 metabolic measure, which represents a very rich and relatively complex analytic approach. Unfortunately, the predictions are not clearly specified, which makes results interpretation difficult.

We appreciate that the methods are complex, and the hypotheses should be stated more clearly. The hypotheses have now been explicitly stated both at the end of the introduction (lines 436 - 527) and at the beginning of the discussion (lines 685 – 755).

Based on the results (L160-162) and discussion (L233-235) sections, it appears that the authors aim at identifying brain regions showing a precise pattern of activity, with a positive relationship between EEG activity and HbO/CCO responses together with a concurrent negative relationship between EEG and HbR responses in response to social events, but not in response to non-social events. Importantly, the social vs. non-social contrast seems crucial to assess the selectivity of the response. Yet, the authors analyze the 3 chromophores separately, and they do not contrast the two conditions (figure 3). As a result, the authors are limited to reporting a descriptive pattern of relationships between EEG and HbO/HbR/CCO activations for the social condition. And another one for the non-social condition. Overall, the authors conclude that channel 14, overlaying the right TPJ, shows the expected pattern of activity, specifically in response to social stimuli. Yet, this statement is only supported by visual inspection/comparison of the results between the social vs baseline and non-social vs baseline conditions. The authors do not assess analytically the differential patterns of activations between the two conditions. Instead, a GLM including all 3 chromophores and contrasting the two experimental conditions would allow us to directly test the predicted pattern of activity, and the selectivity of the activity for social stimuli.

As per the reviewer’s comment, we have now included the comparison of the social and non-social conditions, shown in Figure 3c. The results from this comparison showed that haemodynamics and metabolic activity at channels 11 and 14 (located spatially close to one another) had a significantly greater association to EEG electrode “Pz” for the social condition, in comparison to the non-social condition for the beta and gamma bands. These results support/indicate the selectivity of the response to the social condition, analytically.

We have kept the results showing the individual comparison of the social and non-social conditions. The individual condition tests were performed to show the scale and spatial location/sensitivity of the coupling between haemodynamics/metabolism and neural activity for each condition. Meanwhile, the social versus non-social comparison was performed to show where there was a significant difference in the coupling between the two conditions. With comparison (a) we aimed to identify regions involved in the processing of social and non-social stimuli by identifying the regions where the coupling was significant. With comparison (b) we aimed to identify regions where coupling was significantly different between conditions. The following has been added on line 533 – 541 to explain the reasoning behind the comparisons performed.

We performed two types of statistical tests: a) individual comparisons of the social and non-social conditions and b) comparison of the social condition versus the non-social condition. The individual condition tests were performed to show the scale and spatial location/sensitivity of the coupling between haemodynamics/metabolism and neural activity for each condition. Meanwhile, the social versus non-social comparison was performed to show where there was a significant difference in the coupling between the two conditions. With comparison (a) we aimed to identify regions involved in the processing of social and non-social stimuli by identifying the regions where the coupling was significant. With comparison (b) we aimed to identify regions where coupling was significantly different between conditions.

As our interest was in looking at the selectivity of the response and not comparing the chromophores, we did not perform a comparison between chromophores.

Author Response

Reviewer 2 (Public Review):

1) My major criticism of the study is that the authors argue for CD8+ Trm activity as a key mechanism for OLP pathogenesis but have presented mostly descriptive datasets. The data strongly argue for CD8+ Trm cells as a defining feature of erosive OLP, but there is no data to support their involvement in disease pathogenesis. The authors note the lack of a mouse model for OLP which represents a significant technical barrier to interrogating the role of CD8+ Trm cells in OLP pathogenesis.

Thank you for bringing this to our attention, and please accept our apologies for any confusion caused by our previous article. The pathogenesis of OLP is responsible for the immune disease caused by multiple factors, but there is no corresponding animal model at present, which has obvious limitations on the research. Therefore, we focus on the research on the reasons for the change of the clinical state of the disease. Our study found that CD8+ TRM cells play an important role in the changes observed in the local presentation of OLP, specifically erosions. However, it is important to note that they are not the primary driver of the disease. In addition, we use cohort studies combined with transcriptome data to increase the strength of evidence for causal effects. We have revised and emphasized this point in the updated text.

The modified description in introduction is as follows:

Notably, EOLP has a significantly higher risk of malignant transformation than non-erosive oral lichen planus (NEOLP) (Danielsson et al., 2013). To reduce the psychological and economic burden of OLP patients, improve their quality of life, and decrease the risk of cancer, it is crucial to maintain the disease in a relatively stable non-erosive stage for as long as possible. However, clinical experience suggests that OLP often exhibits a prolonged and recurrent disease course, with alternating periods of non-erosive and erosive lesions. Despite this, the underlying causes and mechanisms of lesion type switching remain unclear (Husein-ElAhmed and Steinhoff, 2022). (Page 4, lines 13-21)

2) Another criticism is the lack of strong findings in the analysis of CD8+ Trm cells isolated from non-erosive and erosive OLP tissues. The authors note increases in CD8+ Trm cell recovery, however, they only observe minor changes in CD8+ Trm activity upon restimulation. Analyzing the activation status or proliferative capacity of CD8+ Trm cells from non-erosive and erosive OLP could be informative and more robust measures of functional changes.

We appreciate your suggestion to test the activation status and proliferation of sorted CD8+ Trm cells to further investigate the differences between the two groups. However, due to the limited amount of tissue available for our study, it was so hard to obtain sufficient numbers of CD8+ Trm cells for these experiments. Additionally, there is a lack of established methods for in vitro culture of CD8+ Trm cells, which further limited our options for functional studies.

To investigate the function of CD8+ Trm cells in the two tissue groups, we instead measured inflammatory factors in the supernatant of CD8+ Trm cells after in vitro stimulation. This allowed us to indirectly assess the activity of CD8+ Trm cells in non-erosive and erosive OLP. We used ELISA assay to measure the levels of several inflammatory cytokines, which are known to be produced by activated T cells, including CD8+ Trm cells.

We acknowledge that this method has limitations and is an indirect measure of CD8+ Trm cell function. However, we believe that our approach provides useful information on the potential role of CD8+ Trm cells in oral lichen planus and represents a valuable contribution to the field.

3) A minor criticism is the formatting of the data presented in Figure 4. The authors should clearly label each marker used in the flow cytometry experiments as well as clearly labeling y-axes for graphs 4H and 4I.

Thank you for your valuable comments, I have modified the flow cytometry diagram accordingly and labeled each step of the gating strategy, also modified the other two diagrams. And 4H and 4I figure numbers changed to 4G and 4H.

Author Response

Reviewer #1 (Public Review):

This paper investigates whether bistable rhodopsins can be used to manipulate GPCR signalling in zebrafish. As a first step, the authors compared the performance of bistable rhodopsins fused with a flag tag or with a fluorescent protein tag (TagCFP). Constructs were compared by expressing in HEK cells followed by calcium imaging with aequorin or cAMP monitoring with GloSensor. This showed that the protein with a smaller flag tag performed better. Then, a series of transgenic zebrafish lines were made, in which tagged rhodopsins were expressed in reticulospinal neurons or cardiomyocytes.

The data indicate that bistable rhodopsin can be used to manipulate Gq and Gi/o signalling in zebrafish. The Gq-coupled SpiRh1 was effective in manipulating reticulospinal neurons, as indicated by analysis of tail movements and calcium imaging of the neurons. Gi/o signalling could be manipulated by Opn3 from mosquitoes, TMT from pufferfish, and parapinopsin from lamprey, as shown by their effects on the heartbeat. Lamprey parapinopsin has the interesting property that it can be turned on and off by different wavelengths of light, and this was used to stop and restart the heart. Finally, the authors show that the cardiac effects are mediated by an inward-rectifier K+ channel, through the use of pharmacological inhibitors.

A strength of this paper is the testing of a range of bistable rhodopsins, with a total of 10 proteins tested. This provides a good resource for future experiments. A weakness is the failure to show that some experiments involved repeated sampling of the same animal. Figure 3 gives the impression that there are 48 independent datapoints. However, there are 8 animals, with 6 datapoints coming from each. Similarly, Figure 4 shows the data from 6 trials of 4 animals, not 24 independent animals. Repeated sampling should be reflected in the data presentation, and in the statistical analysis. Was there an effect of trial number, which is suggested in Figure 6?

In response to the reviewer’s comments, we modified the graph to show the average data for individual animals in Figure 3A-E, Figure 3-supplement 2, Figure 4D-F, H, and Figure 4-supplement 2B. We also showed the effect of trial number (difference between trials 1 and 6) in Figure 3-supplement 1 and Figure 4-supplement 1. In addition, we also showed all data as source data. We believe that more accurate statistical analyses were conducted using data from each individual animal.

Delta F/F refers to relative change, which should be (F-F0)/F0. This should be zero when t = 0. The values in Figure 3E, and 3F are ~ 1 when t = 0, however. Are these figures showing F/F0?

The reviewer is correct. It is indeed F-F0/F0 (ΔF/F0). In Figure 3F (3E in the original manuscript), t=0 was the time when 470-495 nm light (for both stimulation of SpiRh1 and detection of GCaMP6s fluorescence) started to be applied. In the experiment in Figure 3G (3F in the original manuscript), 405 nm light was applied to activate SpiRh1[S186F] for 2 s and then 470-495 nm light was applied to detect GCaMP6s fluorescence. In other words, t=0 is the time when 405 nm light started to be applied.

The authors' conclusions that the bistable rhodopsins are useful tools in the zebrafish system appear largely justified. This is consistent with findings from other organisms, including mouse (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8097317/, https://www.sciencedirect.com/science/article/pii/S0896627321001616). The tools here are likely to find broad use by scientists who use the zebrafish as the experimental system for a variety of different areas.

For the studies on LamPP and MosOpn3, we cited the references mentioned by the reviewer. We believe that our study substantiates that LampPP and MosOpn3, as well as other bistable rhodopsins, are valuable tools for zebrafish research, as pointed out by the reviewer.

Reviewer #2 (Public Review):

The presented study aims at deciphering the physiological function of GPCR signaling in excitable cells. To this end, the authors developed transgenic zebrafish models expressing a selection of Gq- and Gi/o-coupled bistable rhodopsins in either reticulospinal neurons or cardiomyocytes and elucidated behavioral responses (tail movements) or physiological responses (heartbeat) as well as intracellular Ca2+ dynamics following optical stimulation of rhodopsins.

One of the major strengths of the presented study is the functional comparison of five Gq- and five Gi/o-coupled rhodopsins in two major classes of excitable cells, however; the selection of rhodopsins tested remains elusive. More importantly, it is not obvious why some of the effects of rhodopsin activation were assessed in both neurons and cardiomyocytes, while others were only tested in one of the two systems without further explanation. The main chosen experimental readouts (swimming/tail bending or cardiac contractions) have limited informative value regarding GPCR signaling, as they will only report the peak of the iceberg, namely whether movements are elicited or heartbeats inhibited. No analysis on subtle changes in heart rate and contraction force was included, but such modulation of cardiac activity (e.g. positive or negative chronotropic, inotropic, dromotropic, bathmotropic, and/or lusitropic responses) would represent better the physiological modulation of the heart via GPCR and down-stream signaling events. In line, the presented data only represents behavior at one light intensity tested, whereas a light titration of observed effects could provide more meaningful insight into both rhodopsin responses and signaling mechanisms. Also, the potential promiscuity of G protein activation of selected receptors has not been addressed, neither experimentally nor in the discussion part. As a result of the above-mentioned limitations, it is difficult to follow the logic of the study and especially to interconnect the data obtained in reticulospinal neurons (where activation of jumping spider rhodopsin elicited tail bending) to myocyte data (where three Gi-coupled rhodopsins suppressed cardiac activity). Moreover, as such, the study does not provide explanations on why a certain tool might evoke an effect in one system or the other, or not, which could be the main deliverable of such a comparative analysis.

We are grateful for helpful and insightful comments from the reviewer. We believe that the presentation of experimental findings in the original manuscript may have led to a misunderstanding. We examined the effects of Gq and Gi/o-coupled bistable rhodopsins on both reticulospinal V2a neurons and cardiomyocytes. We observed noticeable effects of Gq rhodopsins on reticulospinal V2a neurons, but no significant effects on cardiomyocytes. Similarly, we found effects of Gi/o-coupled rhodopsins on cardiomyocytes, but no significant effects on reticulospinal V2a neurons. These discrepancies could be attributed to differences in the target cells and experimental conditions, suggesting the need for further optimization. We described the data on page 13, lines 16-22 and page 16, lines 9-10 in the Result section and Table 1, and discussed the relationship between the activity of bistable rhodopsins and their effects on target cells on page 21, lines 6-15 and page 24, line 19-page 25, line 2 in the Discussion section of the revised manuscript.

In order to clarify the function of Gi/o-coupled rhodopsins on the heart in more detail, we conducted experiments in which we activated cardiomyocytes expressing bistable rhodopsins at various light intensities to observe the effects on heartbeats. We analyzed cardiac arrest rate, latency to cardiac arrest, and time to resumption of heartbeat. The results of these experiments are shown in Figure 4 and Figure 4-supplement 2, 3 in the revised manuscript. We described the data on page 15, line 16-page 16, line 1 in the revised manuscript, as follows.

To analyze the photosensitivity of Gi/o-coupled rhodopsins, we applied light of various intensities for 1 s and examine their effect on HBs (Figure 4-supplement 2). Cardiac arrest was induced and sustained for over 20 s after stimulation of MosOpn3 with 0.05 mW/mm2 light for 1 s. Photoactivation of PufTMT and LamPP at lower light intensities (0.2 or 0.05 mW/mm2) resulted in cardiac arrest, but faster HB recovery than stimulation with 0.5 mW/mm2 light (Figure 4-supplement 2). The data indicate that the ability of MosOpn3 to suppress HBs is more photosensitive than PufTMT and LamPP in the zebrafish heart. We further examined atrial-ventricular (AV) conductivity by measuring the time difference between atrial and ventricular contractions before and after light stimulation when HBs had slightly recovered. There was no significant difference in AV conductivity before and after light stimulation (Figure 4-supplement 3).

We performed experiments to the best of our ability with current technology regarding cardiac function. However, we hope that the reviewer is willing to acknowledge that there are certain limitations in conducting a detailed analysis of the zebrafish larval heart, since many experimental techniques, such as electrophysiological analysis, have not yet been fully or effectively established for this animal model.

While the presented data is interesting, the graphical presentation and description of the data are insufficient. Most importantly, the current version of the text does not include a quantitative description of effects and statistical analyses (which are found in the figures and legends!). The lack of quantitative description also extends to both the introduction and discussion, which remain general without a specific dissection of observed effects.

We have described quantitative data in the Result section.

One major concern is the selective citation of own work. While single statements in both the introduction and discussion are supported by up to ten own papers, recent studies using rhodopsins for dissecting GPCR signaling in neurons are not sufficiently discussed and new data is not compared to published results by other teams. Moreover, relevant papers on cardiomyocytes (e.g. PMID: 35579776, 35365606, 34987414, 30894542) are not cited at all, despite the use of similar rhodopsins and/or optogenetic activation of the same signaling pathways. Taking into account these published studies may help to better understand the observed responses.

We apologize for not citing important relevant papers in the original manuscript. We have now cited all four papers (Dai et la., 2022; Wagdi et al., 2022; Cokic et al., 2021; Makowka et al., 2019) mentioned by the reviewer, as well as a new paper describing the use of MosOpn3 and LamPP in C. elegans neurons (Koyanagi et al., 2022) in the Introduction section. We also discussed the differences between our findings and previously published data in the Discussion section.

Additional comment: Data were obtained from larvae zebrafish. It would be useful to include a discussion on how GPCR signaling might be different in adult fish compared to larvae, and how to test whether the observed effects are more generally applicable.

We discussed the differences between the hearts of zebrafish larvae and adults, and the differences in GPCR signaling, on page 27, lines 10-16, as follows. In this study, we used zebrafish larvae to study the role of GPCR signaling in cardiac function, and there are differences in heart structure and function between larvae and adult zebrafish. As a zebrafish grows, blood pressure increases and the heart becomes more complex with the development of valves and ventricular trabeculae. Therefore, GPCR signaling, which regulates heart structure and function, may differ between juvenile and adult fish. Optogenetic manipulation of the heart’s function in adult zebrafish using bistable opsins should clarify this issue.

Author Response

Reviewer #1 (Public Review):

This paper aims to test whether a series of light activated ion channels (GtCCRT4, KnChR) and enzymes that regulate second messengers (BeGC1, bPac, OaPac) can be used to manipulate cells in the zebrafish.

Among the strengths of the paper are the use of several independent methods to test whether the tools are functional - e.g. electrophysiology of mammalian cells for GtCCR4, calcium and cAMP imaging in zebrafish cells in vivo, behaviour tests (tail movement) and monitoring of heart beat. Multiple transgenic lines were established, to select for lines with optimal expression levels. Experiments are carried out in two cell types - reticulospinal neurons in the hindbrain and cardiomyocytes.

The authors have largely achieved their aim of determining whether the rhodopsins can be used in zebrafish. They demonstrate that the cation channel KnChR is particularly sensitive in triggering depolarization of the reticulospinal neurons, as indicated by tail movement. They show that the photoactivatable adenylyl cyclase bPAC and cation channels have an effect on heartbeat. Two other photoactivatable enzymes OaPAC and BeGC1 have no effect on heartbeat, although it is not evident whether this is due to lack of effect on cAMP and cGMP levels.

The abstract sets out to investigate the role of second messengers, emphasizing the need for specificity. However, KnChR is not specific for Na+. As noted by Tashiro et al, the channel can also conduct H+, Ca2+ and Mg2+. The knowledge gap that is being addressed by the manuscript thus needs to be reframed. The concluding statement of the abstract, that the tools tested here can be used to investigate second messengers, is not accurate given the broad conductance of KnChR.

We agree with the reviewer. We changed the title to “Optogenetic manipulation of neuronal and cardiomyocyte functions in zebrafish using microbial rhodopsins and adenylyl cyclases” and revised the abstract and introduction, accordingly. The last sentence of the abstract was modified to “These data suggest that these optogenetic tools can be used to reveal the function and regulation of zebrafish neurons and cardiomyocytes.”

The tools described here have been tested previously in other species, either in cultured mammalian cells (GtCCR4, KnChR, OaPAC) or in vivo (bPAC and BeGC1). The current work thus does not introduce novel tools, but provides evidence that some of these tools can be used in zebrafish. Overall, the lines characterized here will be of use to scientists using zebrafish as the experimental system in a variety of areas.

We appreciated the positive comments from the reviewer. It was worthwhile generating and analyzing so many transgenic zebrafish.

Reviewer #2 (Public Review):

Optogenetic proteins are important tools for circuit neuroscience. The authors characterize five proteins, GtCCR4, KnCHR2, BeGC1, bPAC, and OaPAC with respect to their ability to suppress normal cell excitability and compare the results to those for the more established GtACR1 and CrChR2[T159]. The study makes use of expression in the zebrafish heart and hindbrain, as well as in a cell line. Electrophysiology in the cell line demonstrates that GtCCR photo-activation induces similar currents as CrChR2 activation and shows less signs of desensitization. Using a transgenic vsx2:Gal4 zebrafish line, immunohistochemistry shows that the tools are expressed. When activated, they triggered the expected behavioral responses (swimming) at short latency (<4s). This was true even for the three tools that are guanylyl or adenylyl cyclases (BeGC1, bPAC, OaPAC) and thus affect cell excitability only indirectly. At the tested light intensity, the Klebsormidium nitens channelrhodopsin (KnChR) had the shortest latency (<0.5 s) and highest (100%) probabilities of inducing locomotion. When expressing the tools in the zebrafish heart, brief illumination (100 ms) induces brief (100 ms - 1500 ms) suppression of the heartbeat. Notably, also tools that evoke depolarization induce heartbeat suppression. Heartbeat movies and calcium imaging demonstrate that this is caused by prolonged cardiomyocyte contraction. The optogenetic guanylyl and adenylyl cyclases were not effective in perturbing zebrafish heartbeat (except for bPAC over longer time scales).

Given the large number of optogenetic proteins available to date and the challenge of employing them in well-controlled neuroscience experiments, this study presents an important contribution for neuroscientists performing optogenetic research in animal models. Two light-gated cation channels, GtCCR4 and KnChR, are tested for the first time in vivo. The evidence supporting the claims regarding heartbeat and induced swimming behavior is solid. Since GtCCR4 is more Na+-selective than other channelrhodopsins, it should allow better control of experimental variables and is a valuable addition to the optogenetic tool box. The created transgenic zebrafish lines will be useful for the zebrafish neuroscience community.

The expression in zebrafish was compared using immunohistochemical staining (of a single Gal4 driver line). From this experiment alone, it is difficult to judge the expression level, the in vivo visibility of the fluorescence under the microscope, and the proportion of target cells that do express the optogenetic gene of interest.

The evidence for optogenetically induced alteration of swimming behavior is compelling. However, the associated neuronal responses and their dependence on different light intensity levels remain uncharacterized. Therefore, if anyone plans to use these tools to investigate a neural circuit in the future, the needed light levels and the specificity of the manipulation would still need to be determined.

We stimulated neuronal ND7/23 cells, reticulospinal V2a neurons or cardiomyocytes expressing microbial optogenetic tools at various light intensities and examined their effects on neuronal activities and behaviors (tail movements and cardiac arrest). These data are shown in revised Figure 1, Figure 1-supplement 1, Figure 3, Figure 3-supplements 2, 3, Figure 5, and Figure 5-supplements 1, 2. We described the data on page 12, line-page 13, line 1 and page 14, lines 10-13 in the revised manuscript.

For the optogenetic guanylyl and adenylyl cyclases, which clearly were able to alter behavioral responses, the signaling and circuit mechanisms that lead to neuronal depolarization remain unknown, but possible activation pathways are discussed.

Reviewer #3 (Public Review):

In this study, the authors set out to test several new optogenetic tools in zebrafish. They motivate the study by citing differences in ion selectivity of channelrhodopsins and the potential utility of photoactivatable anenylyl and guanylyl cyclases to control cell functions. Although the study provides some useful new information about the utility of these tools in zebrafish, the characterization is limited and there are serious caveats around interpretation of behavioral responses.

The latency of behavioral responses is often extremely long and there is a lack of control data from opsin negative animals, raising serious doubts as to whether these responses are optogenetically mediated.

In other words, many of these responses may not result from optogenetic activation of V2a cells, but instead arise from indirect effects such as visual stimulation of the animal. Previous zebrafish studies have shown swimming responses in opsin-negative control animals at latencies above ~100 ms and used a 50 ms cut-off for optogenetically evoked swims. One can see evidence suggestive of this issue in the authors' data: latency data for GtCCR4 appears bimodal with a cluster of short latency swims and a second spread at latencies >2s; this could be a mix of fast optogenetic and slow artifactual responses. As the authors have already tested opsin negative control animals, they should examine the latency distribution of these responses. The long latency is even more striking in the case of BeGC1, pPAC and OaPAC where in all cases mean latency exceeds 2 seconds. No short latency responses are apparent and the delay is too long to be solely a result of second messenger action (e.g. activation of cyclic nucleotide gated ion channels). In any case, no explanation is provided.

We understand the reviewer’s concern that the responses were too slow. However, the neurons responded after accumulation of cAMP or cGMP, which bind and activate CNG in the neurons. Similar delayed responses were observed when G protein-coupled bistable rhodpsins were activated in reticulospinal V2a neurons (please see the accompanying manuscript).

We compared the latency of zebrafish larvae expressing each tool with those not expressing the tool. The data are shown in Figure 3, Figure 3-supplement 1, Figure 5, Figure 6, Figure 7, and Figure 7-supplement 1. Statistically, we considered responses within 8 s after the start of light stimulation as positive, and significant differences in responses were observed depending on the presence or absence of tool expression, suggesting that tail movements were induced by tool activation. In the absence of tool expression, spontaneous movements were occasionally observed, but they did not often occur within 8 s. We have described the data on page 15, line 20-page 16, line 4 in the revised manuscript.

Although this study is motivated by the need to precisely control the flux of specific ions and modulate specific second messenger pathways, there is almost no characterisation of these processes in zebrafish cells. As such, the degree to which these tools are useful to "precisely control second messengers in vivo" is unclear and the lack of mechanistic data also leaves open questions about unexpected aspects of behavioral results (e.g. the long latency of presumed cyclic-nucleotide induced behavior, above).

We believe that the description "controlling second messengers" was misleading. Since Reviewer #3 has taken issue with this aspect, we note that this paper does not provide a detailed analysis of second(ary) messengers. We have restructured the entire manuscript to focus on optogenetic regulation of zebrafish neurons and cardiomyocytes rather than on "control messenger regulation".

Finally, there is little comparison with other commonly used optogenetic actuators. CrChR2[T159C] is used as the only control but more recent tools (e.g. CoChR, Chrmine, ChroME) are not considered. Thus, beyond showing that the new tools have behavioral effects in zebrafish, the usefulness of this report for researchers wanting to compare and select between tools is limited.

We examined the activity of CoChR and ChrimsonR in neuronal ND7/23 cells. In addition, we generated transgenic zebrafish expressing CoChR or ChrimsonR, and examined their activity in V2a neurons and cardiomyocytes. We thereby compared the activity of GtACR4, KnChR, and CrChR2[T159C] with that of CoChR and ChrimsonR. The data are shown in Figure 1, Figure 1-supplement 1, Figure 2, Figure 3, Figure 3-supplement 3, and Figure 5-supplements 1, 2. We described the data for CoChR and ChrimsonR in the relevant part of the Result section (pages 8-14) and discussed a comparison on page 18, lines 2-16 in the revised manuscript.

We found that KnChR was a more potent optogenetic tool than GtCCR4, CrChR2, and ChrimsonR in zebrafish reticulospinal V2a neurons. Optogenetic activity of KnChR was comparable to that of CoChR in both reticulospinal V2a neurons and cardiomyocytes (Figures 1, 3, 5). Truncation of KnChR prolonged the channel open lifetime by more than 10-fold (Tashiro et al. , 2021) (Figure 1). KnChR conducts various monovalent and bivalent cations, including H+, Na+, and Ca2+, while KnChR has a higher permeability to Na+ and Ca2+ and a higher permeability ratio of Ca2+ to Na+ than CrChR2 (Tashiro et al. , 2021). These properties may contribute to the high photo-inducible activity of KnChR. Activation of KnChR may induce influx of more cations with a longer channel open time than CrChR2 and ChrimsonR, leading to stronger cell depolarization. Optogenetic activity of KnChR was comparable to that of GtCCR4 in cultured cells, but higher than GtCCR4 in zebrafish reticulospinal V2a neurons and cardiomyocytes. While the exact reason is unclear, it is possible that the expression of functional KnChR protein may be high in zebrafish cells.

Author Response

Reviewer #2 (Public Review):

Dipeptide repeat (DPR) proteins produced from both sense GGGGCC (poly-GA, poly-GP and poly-GR) and antisense CCCCGG (poly-PR, poly-PG, poly-PA) repeat RNAs are found C9ORF72-linked ALS/FTD and contribute to neurodegeneration. The translation of the repeat RNA can initiate without the AUG start codon, a process known as repeat associated non-AUG (RAN) translation. In this manuscript, the authors used luciferase reporter construct to show that the translation of PR and PG from the CCCCGG repeats initiated from in-frame AUG in the C9 sequences before the repeats. After mutating candidate AUG codons, the translation can initiate from other AUG, so there is redundancy. But if mutating all the in-frame AUG codons, the luciferase was dramatically reduced, supporting the translation initiated at the AUG start codon. The translation initiation factor eIF2D has been shown to be important for CUG start codon-dependent poly-GA translation from GGGGCC repeats. Here it is shown that eIF2D is not required for poly-PG and poly-PR translation from CCCCGG repeats using both reporter and patient iPS-neurons. The data using luciferase reporter to study antisense repeat translation is solid, the translation initiates from AUG start codon as there are AUG in frame with PG and PR in the constructs containing the antisense sequences.

We thank the reviewer for the constructive feedback.

On the other hand, as the reporter construct includes the sequences containing the AUG codon, it is not surprising that AUG was used. This is canonical translation.

We completely agree. In the revised Introduction, we now point out that, before our study, it was not clear which mode of translation (RAN vs AUG canonical) is employed for DPR synthesis.

Also, in the revised Discussion (lines 251-257)), we mention the following: “Hence, our findings together with these previous studies suggest that DPR synthesis may involve at least three different modes of translation: (a) near-cognate start codon (e.g., CUG, AGG) dependent-translation for poly-GA and poly-GR from sense GGGGCC transcripts, (b) canonical AUG-dependent translation for poly-PR and poly-PG synthesis from antisense CCCCGG transcripts, and (c) DPR synthesis may also occur through RAN translation mechanisms that solely utilize the repeat. It is conceivable that all three modes of translation may occur simultaneously in disease, and that the use of non-canonical and canonical initiation codons may be the primary contributors of DPR production ”.

The 1,000bp intronic sequence included in our antisense 35xCCCCGG constructs (Figure 1A) is the authentic human intronic sequence. We agree that it does contain multiple putative initiation codons, and this was our motivation for conducting systematic mutagenesis of all these codons. To narrow down the list of putative initiation codons, we used our recently developed machine-learning algorithm for initiation codon prediction (PMID: 35648796). We found a CUG and an AUG in poly-PR frame; a CUG and three AUGs in the poly-PG frame), all of which had a good Kozak sequence (as mentioned in Results). Systematic mutagenesis of these codons (single and multiple codon mutations were generated) revealed that an AUG at -273bp is necessary for poly-PR synthesis (Figure 2). Of note, poly-PR is one of the most toxic DPRs, for which an initiation codon had not been previously identified in the literature.

Additionally, the AUG-initiated translation of antisense repeats has been reported previously. Therefore, the novelty is limited.

We agree that an AUG initiation codon was previously described for poly-PG (Boivin et al., EMBO J, 2020, PMID: 31930538). However, our findings significantly extend this observation because redundancy at the level of AUG initiation codon usage was not reported in that study.

We believe our study significantly contributes to the field of C9ORF72 ALS/FTD in the following way:

(i) We identified for the first time an AUG (at -273nt) necessary for synthesis of poly-PR, one of the most toxic DPRs.

(ii) We propose the concept of initiation codon redundancy for poly-PG, which may apply to other DPRs in C9ORF72 ALS/FTD, as well as in other neurological disorders caused by nucleotide repeat expansion mutations.

(iii) Our findings merged with those of previous studies suggest that DPR synthesis may involve at least three different modes of translation: (a) near-cognate start codon (e.g., CUG, AGG) dependent-translation for poly-GA and poly-GR from sense GGGGCC transcripts, (b) canonical AUG-dependent translation for poly-PR and poly-PG synthesis from antisense CCCCGG transcripts, and (c) DPR synthesis may also occur through RAN translation mechanisms that solely utilize the repeat. It is conceivable that all three modes of translation may occur simultaneously in disease, and the use of non-canonical and canonical initiation codons may be the primary contributor of DPR production”.

(iv) We found that the non-canonical translation initiation factor eIF2D is mainly responsible for poly-GA (sense DPR) production without affecting anti-sense DPRs. Hence, we propose a model where DPR translation occurs in a “piecemeal manner”, i.e., a distinct machinery of translation initiation factors may be needed for the synthesis of each DPR.

In the revised manuscript, we now better highlight these key contributions.

How the antisense DPRs are translated endogenously, AUG-canonical translation or RAN translation, depends on whether the AUG is included in the antisense RNA in patients and where the transcription of the antisense starts, upstream or downstream of the AUG start codons. However, this is not considered in the manuscript.

Thank you for this important point. Zu et al., (PNAS, 2013) observed antisense DPR aggregation in brain samples of C9ORF72 ALS/FTD patients. In the same study, the authors conducted 5’ Rapid Amplification of cDNA Ends (RACE). Although this analysis did not identify the exact transcription start site for the antisense CCCCGG RNA, it did show that the region that includes the AUG codons, which we found to be important for poly-PR or poly-PG, is included in the antisense RNA from human C9ORF72 ALS/FTD samples. In page E4969, Zu et al write: “RACE analysis of FCX samples showed intron 1b antisense transcripts begin at varying sites 251–455 bp upstream of the G2C4 repeat”. The same study also detected antisense RNA foci in brain samples of C9ORF72 ALS/FTD patients.

The exact transcription start site for the antisense (and sense) transcript remains unknown. In the near future, we plan RACE experiments to identify it and share these finding with the community in a separate manuscript.

We have modified the Results (lines 133-136) to: “These results strongly suggest that AUG at -273 bp is the start codon for translation of poly-PR, one of the most toxic DPRs in C9ORF72 ALS/FTD. This AUG is predicted to be included in the endogenous antisense CCCCGG transcript based on 5’ Rapid Amplification of cDNA Ends (RACE) analysis on brain samples of C9ORF72 ALS/FTD patients14.”

Author Response

Reviewer #1 (Public Review):

1) While the current dataset aims to demonstrate a "correlation" between grid cell encoding and task performance, the other variables that could confound this correlation should be carefully examined.

(1) The exact breakdown of the fraction of beaconed/non-beaconed/probe trials is never shown. if the session makeup has a significant effect on the coding scheme or other results, this variable should be accounted for.

(2) The manuscript did not provide information about whether individual mice experienced sessions with different combinations of the three trial types, and whether they show different preferences in position or distance encoding even in comparable sessions. This leads to the question of whether different behaviour and activity encoding were dominated by experimental or natural differences between individual mice. Presenting the data per mouse will be helpful.

(3) Related to the above point, in Figure 5, the mice appeared to behave worse in probe trials than non-beaconed trials. If the mouse did not know if a trial is a probe or a non-beacon trial, they should behave equivalently until the reward location and thus should stop an equal amount. If this difference is because multiple probe trials are placed consecutively, did the mouse learn that it will not get a reward and then stop trying to get rewards? Did this affect switching between position and distance coding?

(4) It is not shown how the behaviours (e.g., running speed away from the reward zone, licking for reward) in beaconed/non-beaconed/probe trials were different and whether the difference in behaviours led to the different encoding schemes.

We appreciate these suggestions and will add all of the requested analyses in a revised manuscript. We note here that while the proportion of trial types differed between sessions, in all sessions trial types were varied in a repeating sequence, so blocks of behaviour where grid firing is anchored (or not anchored) to the track coordinates can not be explained as a consequence of a particular trial type. We will make this clearer in a revised manuscript.

2) Regarding the behaviour and activity encoding on a trial-by-trial basis, did the behavioural change occur first, or did the encoding switch occur first, or did they happen within the same trial? This analysis will potentially determine whether the encoding is causal for the behaviour, or the other way around.

We agree this is an important point and the corresponding analyses will be reported in a revised manuscript.

3) The author determined that the grid cell coding schemes were limited to distance encoding and position encoding. However, there could be other schemes, such as switching between different position encodings (with clear spatial fields but at different locations), as indicated by Low et. al., 2021, and switching between different distant encodings (with different distance periods). If these other schemes indeed existed in the data, they might contribute to the variation of the behaviours.

We did not observe switching between coding schemes of the same type within our dataset and so did not document this. We agree it is important to do so and will provide additional analyses in the revised manuscript

4) The percentage of neurons categorised in each coding scheme was similar between non-grid and grid cells. This implies that non-grid cells might switch coding schemes in sync with grid cells, which would mean the whole MEC network was switching between distance and position coding. This raises the question of whether the grid cell coding scheme was important per se, or just the MEC network coding scheme.

We appreciate the suggestion and very much agree that looking at cells outside of just grid cells is important in determining which cells are functionally relevant in spatial behaviours. We will provide additional analyses in a revised manuscript.

5) In Figure 2 there are several cell examples that are categorised as distance or position coding but have a high fraction of the other coding scheme on a per-trial basis. Given this variation, the full session data in F should be interpreted carefully, since this included all cells and not just "stable" coding cells. It will be cleaner to show the activity comparison only between the stable cells.

We agree that showing stable examples before introducing examples that switch on a per-trial basis will be helpful. We will amend this in a revised manuscript.

6) The manuscript is not well written. Throughout the manuscript, there are many unexplained concepts (especially in the introduction) and methods, mis-referenced figures, and unclear labels.

We appreciate the feedback and will work to address the concerns in a revised manuscript.

Reviewer #2 (Public Review):

This study is very timely as there is a pressing need to identify/delimitate the contribution of grid cells to spatial behaviors. More studies in which grid cell activity can be associated with navigational abilities are needed. The link proposed by Clark and Nolan between "virtual position" coding by grid cells and navigational performance is a significant step toward better understanding how grid cell activity might support behavior. It should be noted that the study by Clark and Nolan is correlative. Therefore, the effect of selective manipulations of grid cell activity on the virtual task will be needed to evaluate whether the activity of grid cells is causally linked to the behavioral performance on this task. In a previous study by the same research group, it was shown that inactivating the synaptic output of stellate cells of the medial entorhinal cortex affected mice's performance of the same virtual task (Tennant et al., 2018). Although this manipulation likely affects non-grid cells, it is still one of the most selective manipulations of grid cells that are currently available.

We appreciate this additional context provided here. In our view, it is critical to narrow down the space of possible behaviours that grid cells might contribute to. As the reviewer notes, our previous work provided evidence that speaks to this question by targeting genetic manipulations (Tennat et al., 2018), but while this approach was specific to stellate cells it does not discriminate grid from non-grid cells and so does not tell us specifically about roles for grid cells. As far as we are aware there is currently no manipulation that will do this. In the experiments here, we take a complementary approach, leveraging the variability inherent in behaviour and the fact that in our location memory task animals will perform many trials in a session. By showing that spatially anchored grid firing does not predict behavioural success on cued trials, but does predict success on trials that are solved by path integration, we substantially narrow the space of behaviours that grid cells could contribute to. Importantly, stellate cells appear necessary for both cued and uncued behaviour in the task (Tennant et al., 2018), suggesting that their roles are more general than the grid cell population, which is likely to be only a subset of stellate cells. We will more carefully address this point in a revised manuscript.

When interpreting the "position" and "distance" firing mode of grid cells, it is important to appreciate that the "position" code likely involves estimating distance. The visual cues on the virtual track appear to provide mainly optic flow to the animal. Thus, the animal has to estimate its position on the virtual track by estimating the distance run from the beginning of the track (or any other point in the virtual world).

We agree this terminology has the potential for causing confusion. A simpler descriptive definition would be track-anchored and track-independent rather than position and distance coding. We will consider this and other alternatives for a revised manuscript.

Reviewer #3 (Public Review):

This study addresses the major question of 'whether and when grid cells contribute to behaviour'. There is no doubt that this is a very important question. My major concern is that I'm not convinced that this study gives a significant contribution to this question, although this study is well-performed and potentially interesting. This is mainly due to the fact that the relation between grid cell properties and behaviour is exclusively correlative and entirely based on single cell activity, although the introduction mentions quite often the grid cell network properties and dynamics. In general, this study gives the impression that grid cells exclusively support the cognitive processes involved in this task. This problem is in part related to the text. However, it would be interesting to look at the population level (even beyond grid cells) to test whether at the network level, the link between behavioural performance and neural activity is more straightforward compared to the single-cell level.

We appreciate the feedback and suggestions. As we note in our response to Reviewer #2, there is currently no method for selective manipulation of grid cells, while testing correlation is a critical step on the path to establishing causation. Our study contributes by reducing the space of possible functions of grid cells to exclude behaviours in which local cues are available, while providing evidence for a clear relationship between anchoring of grid cells and successful outcomes when path integration is used for localisation. We’re unclear here about what the reviewer means by ‘more straightforward’ as the relationships we establish do not appear overly complicated, and as strong relationships between activity of single grid cells and populations of grid cells are already well established (Gardner et al., 2021; Waaga et al., 2021; Yoon et al., 2013).

The authors used a statistical method based on the computation of the frequency spectrum of the spatial periodicity of the neural firing to classify grid cells as 'position-coding' (with fields anchored to the virtual track) and 'distance-coding' (with fields repeating at regular intervals across trials). This is an interesting approach that has nonetheless the default to be based exclusively on autocorrelograms. It would be interesting to compare with a different method based on the similarities between raw maps.

We’re not sure we understand the point here. The manuscript provides analyses comparing rate maps for activity periods in which grid cells are / are not anchored to the task environment (e.g. Figure 2A-C, Figure 3B-E); when grid cells are anchored the rate maps are clearly spatial, when they are not anchored we show that spatial information (in the track reference frame) is very substantially reduced.

Beyond this minor point, cell categorization is performed using all trial types. Each trial type (i.e. beacon or non-beacon) is supposed to force mice to use different strategies and should induce different spatial representations within the entorhinal-hippocampal circuit (and not only in the grid cell system). In that context, since all trials are mixed, it is difficult to extrapolate general information.

Again, we’re not sure we understand the point. We appreciate this likely reflects a lack of clarity on our part in the writing of the manuscript. As noted in our response to Reviewer #1, we will include additional details about the organisation of trials and relationships between trials, behavioural outcomes and neural codes observed. We should note here that mice are not ‘forced’ to adopt any particular strategy. Rather, on uncued trials a path integration strategy is the most efficient way to solve the task. Mice could instead use a less efficient strategy of stopping at short intervals and still obtain rewards, although the behavioural evidence suggests they do not choose to do this after learning the task.

On page 5 the authors state that 'Since only position representations should reliably predict the reward location, ..., we reasoned that the presence of positional coding could be used to assess whether grid firing contributes to the ongoing behaviour'. I do not agree with this statement. First of all, position coding should be more informative only in a cue-guided trial. Second, distance coding could be as informative as position coding since at the network level may provide information relevant to the task (such as distance from the reward).

Again, this point perhaps reflects a lack of clarity on our part in writing the manuscript. When grid cells are anchored to the track reference frame (position encoding in the manuscript), then the location of the rate peaks in grid firing is reliable from trial to trial. This is the case whether or not the trial is cued. When grid cells are independent of the track reference frame (distance encoding in the manuscript, but we now appreciate this is a poor choice of words), then the location of the firing rate peaks vary from trial to trial; thus position can not be read out directly from trial to trial. In principle, when grid cells are not anchored to the track the mouse could read out track position by storing the grid network configuration at the start of each trial and then subtracting this from readouts of distance as mice move along the track. If mice do use this computation we would expect them to do so equally well on cued and uncued trials, whereas our results clearly show a dissociation between trial types in the relationship between grid firing and behavioural outcome. We will highlight this possibility in a revised manuscript.

Third, position-coding is interpreted as more relevant because it predominates in correct trials. However, this does not imply that this coding scheme is indeed used to perform correct trials.

As we note above, our analyses reduce the space of behaviours to which grid cells might contribute, by providing evidence that anchoring of grid firing is associated with successful outcomes specifically when mice adopt a path integration strategy. We agree that alternative models remain plausible, for example perhaps the behaviourally relevant computations are implemented elsewhere in the brain with grid anchoring to the track as an indirect consequence. Nevertheless, the space of alternative models is substantially reduced given our experiments and analyses, while our approach complements tests of grid-behaviour functions that rely on manipulations which leave open alternative explanations based on off target effects. We expect that inclusion in a revised manuscript of the further analyses suggested above should provide further tests of the grid-behaviour relationship.

It could be more informative to push forward the correlative analysis by looking at whether behavioural performance can be predicted by the coding scheme on a trial-by-trial basis.

Figure 5E shows the recommended analysis.

Author Response

eLife assessment

This useful study emphasizes some previously ignored aspects of synaptic communication between Purkinje neurons and their targets in the cerebellar nuclei. Reviewers felt that some aspects of the evidence were solid but that others were incomplete.

We think this is an extensive and complete study. The major issue that the reviewers raised is about the usage of high chloride internals in our recordings. We feel that this single issue does not really match the statement “others were incomplete”, which suggests that this study is incomplete in some way. Please note that in our complete revision we will respond to the issue of chloride by pointing out: (1) the advantages of using high chloride internals to determine the distribution of input sizes, (2) the challenges of estimating the relationship between input sizes for different chloride internals, (3) the previous studies that have established the relationship between input sizes and chloride levels at other synapses, and (4) additional simulations will be provided indicating that subtle changes in the input sizes would have minor quantitative effects on the influences of individual inputs, but would not affect the main conclusions of the paper.

Reviewer #1 (Public Review):

This manuscript explores physiological properties of Purkinje-to-nuclear synapses. The report provides largely incremental advances over what has already been discovered about this synaptic relationship. The main findings, as articulated by the authors, are that Purkinje-to-nuclear synaptic strength is variable, with a few very strong inputs to the cerebellar nuclei. They show that single inputs effectively inhibit nuclear firing and that the diversity of synaptic strength influences nuclear neuron responsivity to inputs by enhancing synaptic variance. In addition, while not necessarily surprising, it's nice to see that stronger inputs would have a stronger influence on a postsynaptic cell, both in terms of rates and temporal coding transfer. Overall, as it stands, the manuscript is not very scholarly, overstates the novelty of findings, and frames a straw-man. That said, buried in here are some potentially interesting observations.

This review provides us with an opportunity to more clearly summarize what is new in our findings. Our study builds upon Person and Raman (2012) and other studies, and makes a number of important advances. (1) We provide a much more extensive characterization of input sizes (n=157) than previous studies, and show that the distribution of input sizes is skewed, with the largest inputs almost 100 times larger than the smallest inputs. This distribution is clearly different from that of Person and Raman (2012), where the estimation of unitary PC input sizes was based on small sample sizes from a broad range of age (n=30, P13-29 animals). The high Cl- concentration internal we used in our recordings provides us with superior stability and sensitivity in detecting such variability in input size. (2) We show for the first time that the distribution of input sizes becomes more skewed in juvenile animals than in young animals, suggesting that PC-CbN synapses are modified by plasticity mechanisms during development. (3) Our dynamic clamp approach is based on the skewed distribution of input sizes we observed, and the Purkinje cell firing patterns we recorded in vivo, whereas Person and Raman (2012) primarily focused their dynamic clamp studies on 40 uniform sized inputs (even though they recognized that there are also somewhat larger inputs), with their firing interspike intervals drawn from Gaussian distributions (which lack refractory periods and do not represent realistic PCs firing patterns). We also complement our dynamic clamp studies with simulations using an integrate-and-fire model that does a good job of replicating our dynamic clamp studies. This allowed us to more thoroughly explore the effects of different size input that would not be practical with dynamic clamp studies. (4) We show that individual PC inputs powerfully regulate the rate and timing of CbN neuron firing, without requiring a high degree of PC synchrony. (5) We further show that timing control by PCs leads to strong inhibition of CbN firing and, surprisingly, a brief elevation prior to the inhibition. This result from the refractory period of PCs, which generate a disinhibition period prior to the inhibition, and is shaped by the firing statistics of PC inputs. If such an elevation prior to inhibition was observed in vivo, it could be misinterpreted as excitation of CbN neurons by other inputs (e.g., mossy fiber collaterals) preceding the PC inputs. (6) We show that the total inhibitory conductance and the coefficient of variation (CV) of this conductance are both important factors in controlling the firing rate of CbN neurons. Having variable input sizes or synchronized inputs all lead to higher CV of the inhibitory conductance and therefore higher firing rates. (7) We show that all different-sized PC inputs transmit a robust rate code that simply depend on their sizes. (8) Our study helps to resolve a long-standing controversy in the field. Some thought that PC synchrony is an effective way of controlling CbN neuron firing, while others doubted the physiological relevance of PC synchrony. Here we show that a single large input is functionally equivalent to many small, perfectly synchronized inputs, which can influence the rate and timing of CbN firing as previously proposed (Person and Raman, 2012a), but without requiring a high degree of PC synchrony. We also suggest that a high degree of synchrony is not a prerequisite for an appreciable influence, because synchronizing a few large inputs can have large effects on CbN neuron firing. We strived to be fair and thorough, and we think that the study is scholarly. Prior to the initial submission, we sought advice from experts in the field, Indira Raman and Nicolas Brunel, and their input was very helpful in this regard. We will revise the manuscript to more clearly articulate what has been done previously, and what aspects of our study are new.

Reviewer #2 (Public Review):

In this manuscript, the authors address how cerebellar Purkinje cells (PC) control the firing of nuclear cells (CbN), the output stage of the cerebellar. They used patch-clamp recordings in acute cerebellar slices, and combined dynamic clamp with simulations of nuclear cell firing rate.

This article addresses one of the most fundamental unresolved question of the cerebellar physiology: how inhibitory PCs control the output stage of the cerebellum?

They first described a developmental evolution of the that PC-CbN synapses. Inhibitory synaptic weights become highly variable after three weeks of age, with a group of very large PC inputs. They used dynamic clamp to examine the influence of these variable inputs on CbN firing rate. They demonstrate that while all input size affect CbN discharge, larger ones can stop them for a few milliseconds. Using a distribution of variable input size, they showed that increasing the variability of PC inputs favor CbN discharge, while increasing the magnitude of a constant inhibitory conductance decrease their firing rate. By varying the frequency of PC inputs, they suggest that CbNs faithfully transmit rate code, but larger inputs are more effective to decrease their firing rate. Finally, addressing how synchrony of variable PC inputs influence CbN discharge, dynamic clamp studies and simulations showed that input synchronization enhance firing, but driven by the total charge of the inhibitory input.

The keystone observations that PC inputs are highly variable is very interesting and convincing and open new questions about PC-CbN plasticity. More importantly the combination of dynamic clamp and simulations is a real strength of the study, allowing the authors to test many combinations of inputs in real cells and extrapolating their hypotheses in silico. Weaknesses result from the assumptions made on the construction of the distribution of inputs and the many different conditions explored. The organization of the article could be difficult to read for a non-specialist of cerebellar physiology.

We thank the reviewer for their kind comments. We will revise the manuscript to clarify the assumptions made to construct the distribution of input sizes. We will do our best to revise the manuscript to make it easier for a non-specialist to read.

Author Response

We thank the editors and the reviewers for their comments. In response, we plan to revise the manuscript in order to provide the details requested and include additional bioinformatic analysis of the data, along the lines suggested by the reviewers. We will also take into account individual variations among the subjects investigated in this study, and discuss the extent to which factors other than age might contribute to the results. And we will expand the discussion to consider how our results may apply to other cells/tissues and how they relate to other findings in the field.

Author Response

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

We will make some minor changes to address the issues in the revised manuscript during preparation of the Version of Record.

1) Acknowledge the previous discovery that COUPTFII expression is confined to the ventral hippocampus in early human fetal forebrain (doi: 10.1093/cercor/bhx185).

We agree. We will incorporate the previous discovery that COUPTFII expression is confined to the ventral hippocampus in early human fetal forebrain (doi: 10.1093/cercor/bhx185) in the discussion section of "COUP-TFII governs the distinct characteristics of the ventral hippocampus".

2) Give some consideration to this observation from my original review "Abnormalities in the trisynaptic circuit. No studies of actual synapses, either physiological or morphological, were carried out. I wonder to what extent these immunohistochemical studies just further reflect the abnormalities in hippocampal morphology presented earlier in the manuscript without specifically telling us about synaptic circuits? Although the immunohistochemical preparations are beautiful, they are inadequate on their own in telling us much about what sort of synaptic circuitry exists in the transgenic animals".

Our data in Figure 4 show clearly that at the neural circuit level, compared with the corresponding control, the trisynaptic circuit is abnormal in all three models; therefore, in the discussion section of "COUP-TF genes are imperative for the formation of the trisynaptic circuit", we will add the following sentence, "We would like to investigate what sort of synaptic circuitry is compromised either physiologically or morphologically in the trisynaptic circuit of individual animal model in detail in the future studies.

In addition, we will correct a reference related to the COUP-TFII gene and congenital heart defects.

The reference of "High, F. A., Bhayani, P., Wilson, J. M., Bult, C. J., Donahoe, P. K., & Longoni, M. (2016). De novo frameshift mutation in COUP-TFII (NR2F2) in human congenital diaphragmatic hernia. Am J Med Genet A, 170(9), 2457-2461. doi:10.1002/ajmg.a.37830" was replaced with "Al Turki, S., Manickaraj, A. K., Mercer, C. L., Gerety, S. S., Hitz, M. P., Lindsay, S., . . . Hurles, M. E. (2014). Rare variants in NR2F2 cause congenital heart defects in humans. Am J Hum Genet, 94(4), 574-585. doi:10.1016/j.ajhg.2014.03.007".

—————

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

Reviewer #1(Recommendations For The Authors):

1) Better presentation of the western blot results

We agree with the reviewer. Based on the suggestion, new information about the western blot results has been added in the revised Figure 1Ap. We added a dash to each western blot image to indicate the target band of COUP-TFI (46 KDa), COUP-TFII (45 KDa), and GAPDH (37 KDa), respectively. There were two bands in the blot of COUP-TFII, with the upper band corresponding to mouse IgG at 50 KDa, and the bottom band corresponding to COUP-TFII protein at 45 KDa. Therefore, only the lower bands of COUP-TFII are used for the quantitative analysis. The expression of COUP-TFII in the ventral hippocampus is clearly higher than that in the dorsal hippocampus.

2) Full presentation of the Immunohistochemistry and qPCR results for at E11.5 and E14.5 in double knockdown mice.

Thanks for the suggestion. Based on the suggestion, we added immunofluorescent data in the double knockout mice at E11.5 in the Figure 5Ba-h. Meanwhile, given that it takes time to prepare animal samples at E14.5 for RT-qPCR assays, we performed immunofluorescent assays at both E13.5 and E14.5 to make sure that the changes of Lhx5 and Lhx2 expression in the hippocampal regions between the control and mutant mice were consistent. As shown in the new Figure 5B, consistent with the downregulated expression of Lhx5 transcripts in the double mutant, the expression of the Lhx5 protein was reduced in the CH in the double mutants at E11.5; moreover, the numbers of Lhx5-positive Cajal-Retzius cells decreased in the double mutant embryos at E11.5, E13.5 and E14.5 (Figure 5Ba-d, a’-d’, a’’-d’’, i-l, i’-l’, q-t, q’-t’). Consistent with RT-qPCR data, the expression of Lhx2 was comparable between the control and double-mutant mice at E11.5 (Figure 5Be-h, e’-h’). Interestingly, the expression of the Lhx2 protein was increased in the hippocampal primordium in the COUP-TF double-mutant mice at E13.5 and E14.5 (Figure 5Bm-p, m’-p’, u-x, u’-x’). Please find the altered descriptions in the Page 15, lines 347-351, 353-358 and Page 21, lines 500-503 in the revised manuscript.

3) Minor corrections. Lines 159-162, prospected not quite the right word. I would suggest "an ectopic CA-like region was observed medially in the temporal hippocampus in the COUP1TFII mutant, where the prospective posterior part of the medial amygdaloid nucleus was situated, (MeP), indicated by the star (Figure 1Ba-f). The presence of the ectopic CA-like region in the ventral but not dorsal hippocampus of the mutant was further confirmed by the presence of the prospective MeP and amygdalohippocampal area (AHi) in sagittal sections, as indicated by the star. See also line 251. Line437/438 I would suggest "... most important breakthroughs in understanding the role of the hippocampus in memory."

Thanks for the suggestion. We made the changes based on the suggestion. Please find the amendments in Page 8, lines 178-181; Page 12, lines 270, 276; Page 14, line 318; Page 19, lines 451; Page 20, lines 461-462 in the revised manuscript.

Reviewer #2 (Recommendations For The Authors):

1) It is also important to point out that the immunofluorescence data in Figure 5B is contrary to what is known for Lhx5 (it's not expressed in the neocortical and hippocampal vz) and Lhx2 (it's not expressed in the choroid plexus). Authors should explain how their conclusions could align more clearly, and consider the possibility that their results are due to a possible artifact of image setting issues or worse, antibody specificity issues.

Very good point. Based on the comments and suggestions, we first tested another Lhx5 antibody, R&D, Cat # AF6290, in the immunofluorescence assays. Indeed, there was something wrong with the previous Lhx5 antibody, Millipore, Cat # AB5762. With the new Lhx5 antibody, consistent with the reported in situ data, the expression of Lhx5 was detected specifically in the CH at E11.5, and in the Cajal-Retzius cells in the marginal zone of the telencephalon. The same Lhx2 antibody, Santa Cruz, Cat # sc-19344, which has been used successfully in one of our previous studies (Tang et al., Development, 2012) (PMID: 22492355), was used in the present study. We believe that the observations at the MP and DP of the samples are really associated with the expression of Lhx2 protein. We performed new immunofluorescence assays with the new Lhx5 antibody and confirmed with the Lhx2 antibody. As shown in new Figure 5B, consistent with the downregulated expression of Lhx5 transcripts in the double mutant, the expression of the Lhx5 protein was reduced in the CH in the double mutants at E11.5; moreover, the numbers of Lhx5-positive Cajal-Retzius cells decreased in the double mutant embryos at E11.5, E13.5 and E14.5 (Figure 5Ba-d, a’-d’, a’’-d’’, i-l, i’-l’, q-t, q’-t’). Consistent with RT-qPCR data, the expression of Lhx2 was comparable between the control and double-mutant mice at E11.5 (Figure 5Be-h, e’-h’). Interestingly, the expression of the Lhx2 protein was increased in the hippocampal primordium in the COUP-TF double-mutant mice at E13.5 and E14.5 (Figure 5Bm-p, m’-p’, u-x, u’-x’). Please find the changed descriptions in Page 15, lines 347-351, 353-358 and Page 21, lines 500-503 in the revised manuscript.

The reference:

Tang, K., Rubenstein, J. L., Tsai, S. Y., & Tsai, M. J. (2012). COUP-TFII controls amygdala patterning by regulating neuropilin expression. Development, 139(9), 1630-1639. doi:10.1242/dev.075564

2) The expression domain of RxCre remains poorly explained, and the early expression of COUPTFI and II (E10.5-E12.5) could be considered major weaknesses of the paper.

Thanks for the suggestion. The generation of RXCre was reported by Swindell et al., Genesis, 2006 (PMID: 16850473). Given that the activation of the LacZ expression serves as an indicator for the deletion of the COUP-TFII gene (Tang et al., Development, 2012) (PMID: 22492355), we performed the immunofluorescent data with antibodies against COUP-TFII and LacZ on the sagittal sections of RXCre/+; COUP-TFIIF/+ heterozygous mutant and RXCre/+; COUP-TFIIF/F homozygous mice at E11.5. As shown in the new Figure 1—figure supplement 1Da-f, COUP-TFII was readily detected at the hippocampal primordium of the heterozygous mutant embryo at E11.5 (Figure 1—figure supplement 1Da, c, g); in contrast, the expression of COUP-TFII significantly decreased in the homozygous mutant (Figure 1—figure supplement 1Dd, f, j). In addition, compared with the heterozygous mutant embryo, the LacZ signals increased distinctly in the hippocampal primordium of the homozygous mutant embryo at E11.5 (Figure 1—figure supplement 1Db-c, e-f, h, k), suggesting that RXCre recombinase can efficiently excise the COUP-TFII gene in the hippocampal primordium as early as E11.5. Please find the corresponding changes in Page 7, lines 149-159 and Page 8, lines 160-164 in the revised manuscript.

Meanwhile, we also added the early expression of COUP-TFI and -TFII at E10.5 and E11.5 in new Figure 1—figure supplement 1Aa-d. At embryonic days 10.5 (E10.5), COUP-TFI was detected in the dorsal pallium (DP) laterally and COUP-TFII was expressed in the MP and CH medially (Figure 1—figure supplement 1Aa, b). At E11.5, the expression of COUP-TFII remained in the hippocampal primordium, including MP and CH (Figure 1—figure supplement 1Ac, d). Please find the corresponding changes in Page 6, lines 129-132 and Page 9, lines 202-203 in the revised manuscript.

The references:

Swindell, E. C., Bailey, T. J., Loosli, F., Liu, C., Amaya-Manzanares, F., Mahon, K. A., . . . Jamrich, M. (2006). Rx-Cre, a tool for inactivation of gene expression in the developing retina. Genesis, 44(8), 361-363. doi:10.1002/dvg.20225

Tang, K., Rubenstein, J. L., Tsai, S. Y., & Tsai, M. J. (2012). COUP-TFII controls amygdala patterning by regulating neuropilin expression. Development, 139(9), 1630-1639. doi:10.1242/dev.075564

Reviewer #3 (Recommendations For The Authors):

1) Regarding the RxCre line, I was also confused about its spatiotemporal expression, as this line is not a commonly used Cre line and no detailed description is provided in the manuscript. Searching this line shows a previous paper by the authors (PMID: 22492355) in which they tested the RxCre recombinase activity. At E12.5, RxCre induced high LacZ expression in the ventral telencephalon but much less in the dorsal telencephalon. But they did not check later stage. Therefore, it's hard to explain the defective dorsal hippocampus in RxCre, CFI CKO. They should check later stage.

The generation of RXCre was reported by Swindell et al., Genesis, 2006 (PMID: 16850473), which reveals high Cre recombinase activity of RXCre in the eye and ventral telencephalon. Given that the activation of the LacZ expression serves as an indicator for the deletion of COUP-TFII gene, Tang et al., Development, 2012 (PMID: 22492355), we performed the immunofluorescent data with antibodies against COUP-TFII and LacZ on the sagittal sections of RXCre/+; COUP-TFIIF/+ heterozygous mutant and RXCre/+; COUP-TFIIF/F homozygous mice at E11.5. As shown in new Figure 1—figure supplement 1D, compared with the heterozygous mutant embryo, the expression of COUP-TFII was significantly decreased in the homozygous mutant; in addition, the LacZ signals evidently increased in the hippocampal primordium of the homozygous mutant embryo at E11.5, suggesting that RXCre recombinase can efficiently excise the target gene in the hippocampal primordium as early as E11.5. The expression of COUP-TFI is barely detectable in the early developing hippocampal primordium including MP at E10.5, E11.5 and E12.5. The expression of COUP-TFI is high in the MP of the control (Figure 1Cj, l); in contrast, the COUP-TFI expression is barely detectable in the MP of the homozygous double mutant at E14.5, indicating that RXCre can efficiently delete the COUP-TFI gene in the hippocampal primordium at E14.5. The loss of the COUP-TFI gene in the MP as early as E14.5 by RXCre initiates the defective dorsal hippocampus in RXCre/+; COUP-TFIF/F knockout mice.

2) Authors should check and review extensively for improvements to the use of English.

We carefully checked and made changes throughout the manuscript accordingly. For example, “imperative” was used 6 times in the previous manuscript, lines 20, 255, 486, 499, 522, 553; “imperative” was used only once in Page 22, line 522 in the revised manuscript.

3) Please correct the manuscript; 1-month-old mice are not adult mice.

Thanks for the suggestion. Based on the suggestion, we have corrected related words and sentences in the manuscript. Please find the amendments in the revised manuscript (Page 7, line 146; Page 9, lines 203-204; Page 10, line 213; Page 13, lines 299-300; Page 17, line 406; Page 20, line 476).

4) Additional ref should be added at line 93 on page 5.

Based on the suggestion, we added some new references (Bertacchi et al., EMBO J, 2020) (PMID: 32572460); (Del Pino et al., Cereb Cortex, 2020) (PMID: 32484994); (J. Feng et al., Sci Adv, 2021) (PMID: 34215582) at line 96 on page 5.

The references:

Bertacchi, M., Romano, A. L., Loubat, A., Tran Mau-Them, F., Willems, M., Faivre, L., . . . Studer, M. (2020). NR2F1 regulates regional progenitor dynamics in the mouse neocortex and cortical gyrification in BBSOAS patients. Embo j, 39(13), e104163. doi:10.15252/embj.2019104163

Del Pino, I., Tocco, C., Magrinelli, E., Marcantoni, A., Ferraguto, C., Tomagra, G., . . . Studer, M. (2020). COUP-TFI/Nr2f1 Orchestrates Intrinsic Neuronal Activity during Development of the Somatosensory Cortex. Cereb Cortex, 30(11), 5667-5685. doi:10.1093/cercor/bhaa137

Feng, J., Hsu, W. H., Patterson, D., Tseng, C. S., Hsing, H. W., Zhuang, Z. H., . . . Chou, S. J. (2021). COUP-TFI specifies the medial entorhinal cortex identity and induces differential cell adhesion to determine the integrity of its boundary with neocortex. Sci Adv, 7(27). doi:10.1126/sciadv.abf6808

5) I am confused why the authors analyzed 1-month-old mice in some instances but 3-month-old mice in others.

The RXCre/+; COUP-TFIF/F; COUP-TFIIF/F double mutant mice barely survived beyond postnatal 3 weeks. To make our findings consistent and comparable, we mainly prepared figures with observations on about 1-month-old mice in the RXCre related single or/and double gene mutant mouse models. In the study of the Emx1Cre related COUP-TFI mouse model, due to behavioral tests such as the Morris water maze test, experiments were performed with the adult experimental animal about postnatal 3 months. In order to be consistent with the stage of the mice for the behavioral tests, we only displayed morphological data with observations on the control and Emx1Cre/+; COUP-TFIF/F mutant mice at about postnatal 3-month.

Author Response

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

We thank the reviewers for their comments. We have now addressed all the comments in a revised version of the manuscript, which we believe has strengthened our paper.

1) Introduction LINE 60: the authors cite Funato et al 2016 as the paper first describing a role for SIk3 in sleep regulation. In fact, the role for this kinase was first identified nearly a decade earlier in C. elegans (Van der Linden et al, Genetics 2008 PMID 18832350).

Thank you for pointing us to this reference. Van der Linden et al. demonstrated that the C. elegans homolog of Sik3 (KIN-29) regulates satiety quiescence, in which worms stop moving following feeding on high quality food. However, as pointed out in Trojanowski and Raizen “Call it Worm Sleep” (2016), not all of the behavioral criteria for sleep has been applied to C. elegans satiety quiescence, and we cannot find any references that unequivocally demonstrate satiety quiescence is a sleep state. As McClanahan et al., (2020) show, quiescent states following mild sensory arousal do not fulfill the sleep criteria of changes in arousal threshold and homeostatic regulation, so not all quiescent states in C. elegans are sleep. Then again Grubbs et al, 2020 does demonstrate that KIN29 regulates both developmentally timed and stress induced sleep states in worms, suggesting that the observations in Van der Linden were ahead of its time and these behavioral states are possibly inter-related. We believe, though, that our line “the roles of… SIK3 kinase in modulating sleep homeostasis in mice (Funato et al. 2016) were identified in genetic screens” remains accurate.

2) Introduction LINE 71: remove the word "known" from "...while some known human sleep/wake regulators, such as the...")

Good idea. Done.

3) I was confused regarding Supplemental data 1 describing the genes they targeted with their forward genetic screen. Am I understanding correctly from the "Summary stats" tab that 702 fish lines with virus insertions were screened behaviorally? In Figure S1, it looks like about 60 are shown in the histograms but in the text (in the Discussion) they say 25 were screened. Were all the genes listed under the Excel tabs (GPCRs, channels, etc) tested? Or was just a subset tested? Where are the sleep data for these lines? Negative results may be relevant to their manuscript since they listed (tested??) a number of ion channel genes under tab "channels" which appear to NOT have a sleep phenotype.

We apologize for the confusion on these points. As highlighted in the legend to Supplementary Figure S1, we had planned a screening strategy with the following pipeline: Candidate mammalian gene → Zebrafish ortholog → ID viral insertion from “Zenemark” library → grow viral insertion lines from frozen sperm→ phenotype F3 heterozygous and homozygous mutant generation. Unfortunately, the company, Znomics, which held the Zenemark library, could not reliably reconstitute the correct live fish from the sperm library, and of the 702 lines we planned to screen, we could only screen 26 (25 was a typo) lines. We treated heterozygous and homozygous animals for each line independently, for a total of 52 screened lines in the histograms.

To make this clearer, we have edited the main text as follows (lines 104-105): “For screening, we identified zebrafish sperm samples from the Zenemark collection (Varshney et al., 2013) that harboured viral insertions in genes of interest and used these samples for in vitro fertilization and the establishment of F2 families, which we were able to obtain for 26 lines.” And lines 111-112: “While most screened heterozygous and homozygous lines had minimal effects on sleep-wake behavioural parameters (Figure S1B-S1C),”

We believe it is important to include the full set of Supplementary Data 1, even though the vast majority of these candidate lines were not tested.

4) Results LINE 117: remove the word "prominent", which is subjective, from the sentence "...showed a prominent decrease in sleep during the..."

Good point. Done.

5) LINES 185-186: did you see any circadian variation in your dmist:GFP protein abundance or localization? Protein trafficking has been described as a mechanism of circadian regulation of excitability.

For practical reasons, we imaged the membrane localization of Dmist:GFP in plasmidinjected embryos at 90% epiboly, which is about 9 hours after fertilization and when the cells remain large and in a relatively flat epithelium. Thus, we could not follow circadian fluctuations in abundance or localization. For circadian studies, we believe the best method will be to raise an antibody that recognizes Dmist.

6) LINE 203: does the GFP-tagged Dmist rescue the loss-of-function phenotype? This is relevant to Figure 2E. it is also relevant to the issue of structure-function. If it rescues, then the C-terminus may not be essential to protein function.

As noted, for practical reasons, we observed Dmist-GFP only transiently at early stages of development, expressed using a strong, ubiquitous promoter. A rescue experiment is a good idea for future experiments, where we carefully control the expression of Dmist in neurons.

7) LINE 220: explain what you mean by "...consistent with nonsense-mediated decay." and/or give a reference.

In zebrafish and other species including humans, mutant transcripts that have premature stop codons often undergo “nonsense mediated decay”, whereby the expression levels are largely reduced (Wittkopp et al., 2009). In the zebrafish community, this is often used as secondary evidence of a loss of function mutation, as relatively few antibodies are available to directly observe zebrafish proteins. We have added a reference that describes this phenomenon (Wittkopp et al., 2009).

8) LINE 225: define "LME model"

Now reads: “Linear mixed effects (LME).”

9) LINES 227-229: could the vir/vir phenotype be explained by specific effects on protein structure? could vir/vir be a gain-of-function allele?

We can’t rule this out formally, and vir/+ animals do show some sleep phenotypes, albeit weaker than those of vir/vir animals (Figure 1G). However, it is not uncommon for heterozygous mutants to show significant phenotypes that are weaker than those of their homozygous mutant siblings, and the strong suppression of dmist expression by the viral insertion (which is located in the dmist intron) is more consistent with a hypomorphic loss-of-function phenotype for the vir allele.

10) LINES 229-230: I don't quite follow the argument for pursuing further studies only of i8/i8. i8/i8 seems to also be a hypomorphic allele based on your qPCR data.

First, the dmist viral line was generated by an insertional mutagenesis method followed by sequencing, and each line has multiple other inserts in a background that does not match the background of the other animals reported in this paper. Second, the dmist vir allele is an insertion in the intron, leading to reduced, but not complete loss of expression. In contrast, the i8 allele was generated on the same background strain as our other existing and newly reported lines. Moreover, our i8 line is likely a loss-of-function allele and not a hypomorph. Yes, dmist expression is reduced in the i8 allele; however, this is likely due to nonsense mediated decay of dmist mRNA. The mutation introduces a frameshift in the dmist coding sequence, and as a result the amino acid sequence of the protein is altered after the N-terminal signal sequence.

11) LINES 241-243: grammar.

Fixed

12) LINE 245: define "JackHMMR iterative search"

We’ve added the phrase: “and seeding a hidden Markov model iterative search (JackHMMR)”

13) LINE 246 is missing the word "we" prior to "...found distant homology between..."

Added

14) LINE 301: show data demonstrating deviation from Mendelian ratios. Also, comment on meaning of such data (embryonic lethality??).

We have added this data in the line (301):

“atp1a3b mutant larvae were not obtained at Mendelian ratios (55 wild type [52.5 expected], 142 [105] atp1a3b+/-, 13 [52.5] atp1a3b-/-; p<0.0001, Chi-squared) suggesting some impact on early stages of development leading to lethality.”

15) Discussion LINES 362-372: This paragraph seems to be of only tangential relevance to the paper. Consider removing.

Our screening strategy was a large-scale reverse genetic screen, but the number of lines was limited by the technical issues described above. We think it is important to mention that the strategy, if employed today, could benefit from newer technologies.

16) Discussion. Another model is that Dmist and NaK pump have a developmental effect. Arguing against this developmental model is the Oubain expt.

This is an important point. We’ve added the line (454:457): “We also cannot exclude a role for Dmist and the Na+/K+ pump in developmental events that impact sleep, although our observation that ouabain treatment, which inhibits the pump acutely after early development is complete, also impacts sleep, argues against a developmental role.”

17) FIGURE 1G: Are these significance cut offs corrected for multiple comparisons?

Yes, all the data is corrected for multiple comparisons.

18) performing neuronal activity measures, either via neural activity imaging or phospho-ERK labeling in different mutants at day or night conditions, to determine whether baseline neuronal activity brain-wide or in specific brain regions are altered.

These are excellent experiments that we plan to perform in the future.

19) Please check all Figure numbers for accuracy.

We have double checked these.

20) The authors emphasize the role of increased cellular sodium, but equally plausibly, the phenotypes could be due to decreased cellular potassium. The potassium channel shaker has been previously identified as a critical sleep regulator in Drosophila.

We completely agree. We would like to highlight that we did devote an entire paragraph to the possibility of changes in extracellular potassium in the discussion: “A third possibility is that Dmist and the Na+,K+-ATPase regulate sleep not by modulation of neuronal activity per se but rather via modulation of extracellular ion concentrations. Recent work has demonstrated that interstitial ions fluctuate across the sleep/wake cycle in mice. For example, extracellular K+ is high during wakefulness, and cerebrospinal fluid containing the ion concentrations found during wakefulness directly applied to the brain can locally shift neuronal activity into wake-like states (Ding et al., 2016). Given that the Na+,K+-ATPase actively exchanges Na+ ions for K+ , the high intracellular Na+ levels we observe in atp1a3a and dmist mutants is likely accompanied by high extracellular K+. Although we can only speculate at this time, a model in which extracellular ions that accumulate during wakefulness and then directly signal onto sleep-regulatory neurons could provide a direct link between Na+,K+ ATPase activity, neuronal firing, and sleep homeostasis. Such a model could also explain why disruption of fxyd1 in non-neuronal cells also leads to a reduction in night-time sleep.”

We also agree that Shaker may be an important component of this sleep regulatory mechanism. Indeed, we previously showed that another potassium channel in zebrafish regulates sleep (Rihel et al., 2010).

We have emphasized sodium homeostasis in our title and paper only because we were able to directly observe intracellular sodium levels, so we are confident that these have been altered in our mutants. We can only presume that potassium levels have also been altered, but we could not directly observe this.

21) The similar phenotype between dmist and Fxyd1 in sleep reduction yet very different expression patterns, with dmist being mostly neuronal while fxyd1 being mostly non-neuronal, raise many possible questions: 1) are the sleep phenotypes due to neuronal Na/K imbalance? Or 2) Are the sleep phenotypes due to extracellular Na/K imbalance? Or 3) both? Some feasible experiments may help achieve a better mechanistic understanding of the observed sleep defects.

Yes, we think these are excellent studies for future work. As noted in the previous point (20), we did discuss the possibility that changes to extracellular potassium might be a parsimonious explanation for the similar phenotypes of fxyd1 and dmist mutants.

Future experiment suggestions (not required)

1) Perform a double mutant analysis of fxyd1 and atp1a3a, to determine whether an epistatic relationship similar to that of dmist and atp1a3a is observed in the case of fxyd1 and atp1a3a.

This is a great experiment that we will do in the future. Unfortunately, the fxyd1 mutant had been sperm frozen during the COVID-19 pandemic, so we cannot do this experiment at this time.

2) Given the differences in the sleep phenotypes between vir/vir and i8/i8 mutants, would be informative to see the phenotype of the vir/i8 trans-heterozygote.

This is also a good experiment to perform in the future. Since obtaining the cleaner i8 allele, the dmistvir/vir lines were sperm frozen.

Author Response

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

Reviewer #1 (Public Review):

In this manuscript, the authors investigated the role of Elg1 in the regulation of telomere length. The main role of the Elg1/RLC complex is to unload the processivity factor PCNA, mainly after completion of synthesis of the Okazaki fragment in the lagging strand. They found that Elg1 physically interacts with the CST (Cdc13-Stn1- Ten1) and propose that Elg1 negatively regulates telomere length by mediating the interaction between Cdc13 and Stn1 in a pathway involving SUMOylation of both PCNA and Cdc13. Accumulation of SUMOylated PCNA upon deletion of ELG1 or overexpression of RAD30 leads to elongated telomeres. On the other hand, the interaction of Elg1 with Sten1 is SIM-dependent and occurs concurrently with telomere replication in late S phase. In contrast Elg1-Cdc13 interaction is mediated by PCNA-SUMO, is independent on the SIM of Elg1 but still dependent on Cdc13 SUMOylation. The authors present a model containing two main messages 1) PCNA- SUMO acts as a positive signal for telomerase activation 2) Elg1 promotes Cdc13/Stn1 interaction at the expense of Cdc13/Est1 interaction thus terminating telomerase action.

The manuscript contains a large amount of data that make a major inroad on a new type of link between telomere replication and regulation of the telomerase. Nevertheless, the detailed choreography of the events as well as the role of PCNA- SUMO remain elusive and the data do not fully explain the role of the Stn1/Elg1 interaction. The data presented do not sufficiently support the claim that SUMO- PCNA is a positive signal for telomerase activation.

We thank the reviewer for her/his review efforts and opinion. We have re-submitted a new version of the manuscript in which we clarify some of the criticisms presented. In a point-by-point letter we respond to all the specific queries.

Reviewer #2 (Public Review):

This paper purports to unveil a mechanism controlling telomere length through SUMO modifications controlling interactions between PCNA unloader Elg1 and the CST complex that functions at telomeres. This is an extremely interesting mechanism to understand, and this paper indeed reveals some interesting genetic results, leading to a compelling model, with potential impact on the field. The conclusions are largely supported by experiments examining protein-protein interactions at low resolution and ambiguous regarding directness of interactions like co-IP and yeast two-hybrid (Y2H) combined with genetics. However, some results appear contradictory and there's a lack of rigor in the experimental data needed to support claims. There is significant room for improvement and this work could certainly attain the quality needed to support the claims. The current version needs substantial revision and lacks the necessary experimental detail. Stronger support for the claims would add detail to help distinguish competing models.

We thank the reviewer for her/his positive opinion. We have re-submitted a new version of the manuscript in which we clarify some of the criticisms presented by thereferees, and added all the missing experimental details. In a point-by-point letter we respond to all the specific queries.

Reviewer #3 (Public Review):

This paper reveals interesting physical connections between Elg1 and CST proteins that suggest a model where Elg1-mediated PCNA unloading is linked to regulation of telomere length extension via Stn1, Cdc13, and presumably Ten1 proteins. Some of these interactions appear to be modulated by sumolyation and connected with Elg1's PCNA unloading activity. The strength of the paper is in the observations of new interactions between CST, Elg1, and PCNA. These interactions should be of interest to a broad audience interested in telomeres and DNA replication.

We thank the reviewer for her/his positive opinion. We have re-submitted a new version of the manuscript in which we clarify some of the criticisms presented. In a point-by-point letter we respond to all the specific queries.

What is not well demonstrated from the paper is the functional significance of the interactions described. The model presented by the authors is one interpretation of the data shown, and proposes that the role of sumolyation is temporally regulate the Elg1, PCNA and CST interactions at telomeres. This model makes some assumptions that are not demonstrated by this work (such as Stn1 sumolyation, as noted) and are left for future testing. Alternative models that envision sumolyation as a key in promoting spatial localization could also be proposed based on the data here (as mentioned in the discussion), in addition to or instead of a role for sumolyation in enforcing a series of switches governing a tightly sequenced series of interactions and events at telomeres. Critically, the telomere length data from the paper indicates that the proposed model depicts interactions that are not necessary for telomerase activation or inhibition, as telomeres in pol30-RR strains are normal length and telomeres in elg1∆ strains are not nearly as elongated as in stn1 strains. One possibility mentioned in the paper is the PCNAS and Elg1 interactions are contributing to the negative regulation of telomerase under certain conditions that are not defined in this work. Could it also be possible that the role of these interactions is not primarily directed toward modulating telomerase activity? It will be of interest to learn more about how these interactions and regulation by Sumo function intersect with regulation of telomere extension.

We present compelling evidence for a role of SUMOylated PCNA in telomere length regulation. Figure 1 shows that this modification is both necessary and sufficient to elongate the telomeres, indicating that PCNA SUMOylation plays a positive role in telomere elongation. The model we present is consistent with all our results. There are, of course, possible alternative models, but they usually fail to explain some of the results. We agree that the fact that pol30-RR presents normal-sized telomeres implies that SUMO-PCNA is not required for telomerase to solve the "end replication problem", but rather is needed for "sustained" activity of telomerase. Since elongated telomeres (by absence of Elg1 or by over-expression of SUMO-PCNA) was the phenotype monitored, this may require sustained telomerase activity. Similar results were seen in the past for Rnr1 (Maicher et al., 2017), and this mode depends on Mec1, rather than Tel1 (Harari and Kupiec, 2018). Telomere length regulation is complex, and we may not yet understand the whole picture. It appears that for normal “end replication problem” solution, very little telomerase activity may be needed, and spontaneous interactions at a low level may suffice. Future work may find the conditions at which telomerase switches from "end replication problem" to "sustained" activity. We have added further explanations on this subject to the Discussion section.

We suspect, but could not prove, a role for Stn1 SUMOylation in the interactions. SUMOylation is usually transient, and notoriously hard to detect, and despite the fact that many telomeric proteins are SUMOylated, Stn1 SUMOylation could not be shown directly by us and others (Hang et al, 2011).

Reviewer #1 (Recommendations For The Authors):

Suggestions for improved or additional experiments, data or analyses.

• My main concern is the claim that SUMOylated PCNA acts as a positive signal for telomerase activation. Yet the pol30-RR mutant has no impact on telomere length. The explanation of the authors is not entirely convincing.

We are aware that the regulation of telomere length is complex, and we may not fully understand it yet. Just consider the fact that ~500 genes participate in determining the final telomere length of a yeast (Askree et al., 2004). Since mutation in EACH of these genes has a phenotype, the implication is that the joint action of 500 players determines the outcome (a dialogue of 500 participants). Having said this, we clearly show in figure 1 that mutations that prevent PCNA SUMOylation prevent telomere length elongation in cells lacking Elg1, and overexpressing SUMOylated PCNA is enough to elongate the telomeres. Thus, SUMOylation of PCNA does act as a positive signal for elongation.

However, it appears that to fulfill the minimal requirement of dealing with the "end- replication problem", PCNA SUMOylation is not required, and only a "sustained activity" mode requires the S-PCNA signal (as we have also shown, surprisingly, for RNR1, Maicher et al. 2017). This sustained activity mode depends on Mec1, rather than Tel1 (Harari and Kupiec, 2018). Since elongated telomeres (by absence of Elg1 or by over-expression of SUMO-PCNA) was the phenotype monitored, this may require sustained telomerase activity. Telomere length regulation is complex, and we may not yet understand the whole picture. It appears that for normal “end replication problem” solution, very little telomerase activity may be needed, and spontaneous interactions at a low level may suffice (for example, unmodified PCNA may promote telomerase activity at a lower level than that of SUMO-PCNA. Future work may find the conditions at which telomerase switches from "end replication problem" to "sustained" activity.

We have added further explanations on this subject to the Discussion section.

• The model is entitled « Elg1 negatively regulates the telomere length by forming an interaction with the CST complex ». Nevertheless, expression of PCNA-RR completely reversed the long telomere phenotype of elg1∆ cells. Thus it appears that although the interaction between Stn1 and Cdc13 is reduced in the absence of Elg1, Elg1/Stn1 interaction is not instrumental in the formation of the CST complex and thus in the termination of telomerase activity. Does the elg1∆SIM mutant that does not interact with Stn1 impact telomere length?

• In the model part (lane 318), it is argued that the complex Elg1-Stn1 unloads SUMOylated PCNA. Elg1-Stn1 interaction depends on the SIM of Elg1. This SIM is however not required for Elg1's function in genome-wide SUMO-PCNA unloading, is it required specifically at telomeres?

The interactions between Elg1 and SUMOylated PCNA are carried out through both the SIM and the Threonines 386 and 387 (Shemesh et al, 2017). Consistently, the single elg1-SIM mutant has telomeres of normal length, and its effects on telomere length can only be seen when combined with mutations in the Threonines (elg1- TT386/7AA or elg1-TT386/7DD). Although the unloading of SUMOylated PCNA by Elg1 is important, the gene is not essential, and PCNA is either eventually unloaded by RFC, or spontaneously dis-assembles. This explains why the telomere length does not reach the same length in the absence of Elg1 as in the absence of, say, Stn1.

• The model suggests that Elg1 promotes the interaction between Cdc13 and Stn1. This is based on the data presented in Figure 5 E and F. This is an important result. Because the experiment has been done on cells synchronized in S phase and the Elg1/Stn1 interaction occurs specifically at the end of S-phase, the FACS profile should be shown or a control provided to show that the two conditions are comparable.

The FACS profile for this experiment is shown in Figure 5C.

• Does the interaction between Cdc13 and Pol30 depend on the SUMOyaltion of POL30 ?

Yes. We have added this as new Figure S2, and presented the results together with Figure 3 (Figure 3 is already too crowded).

Others points :

• Fig 1 : it should be mentioned in the Materials and Methods or in the figure legend how the average telomere lengths (horizontal bar) were calculated from the teloblot, as the position of the bar is not always intuitive

We estimate telomere length by using TelQuant (Rubinstein et al., 2014). We have added this to the Methods section.

-Fig 2 : Owing to the large span of telomere length in the stn1 mutants, the epistatic relationship between elg1∆ and stn1 mutants is poorly illustrated by the teloblot.

We repeated this experiment several times, and stn1 mutants consistently gave a very spread telomere length. In ALL the blots, however, the double mutants elg1 stn1 showed a telomere length similar to that of the single stn1 mutant, and never longer.

• It is mentioned that other mutants in the collection showed epistasis. Are any of these mutants related to telomere replication or the proposed model?

Since we used the collection of non-essential mutants (so far), it was quite devoid of genes involved in DNA replication, which are mostly essential. An exception was siz1, which showed epistasis with elg1Δ.

• The section entitled « Elg1's functional activity is essential for its interaction with Cdc13 » (lane 205) is difficult to follow. The hierarchy between the different mutants of Elg1 on their capacity to unload PCNA is not totally in agreement with the data published in Itzkovich et al 2023 and Shemesh et al. 2017. In particular it appears to me from these papers that elg1-WalkerA 238 (KK343/4AA) mutant did not show a defect in contrast to elg1-WalkerA 238(KK343/4DD).

We are sorry for the typo in the results. We used the elg1-WalkerA (KK343/4DD) allele, which has a normal SIM but no activity. In a nutshell, we used mutants that either did or did not show unloading activity and/or SIM. The results clearly show that you need to unload PCNA in order for the N-ter of Elg1 to interact with Cdc13.

• Are the synchronization done at 30{degree sign}C ?

Yes. We have added the information to the Methods section.

• ChIP experiments are not described in the Materials and Methods

We apologize for this. They are now described.

• In the figure 6, the PCNA rings are curiously placed at the beginning of the Okasaki fragments.

We thank the referee for noticing, we have corrected the figure.

Reviewer #2 (Recommendations For The Authors):

This paper purports to unveil a mechanism controlling telomere length through SUMO modifications controlling interactions between PCNA unloader Elg1 and the CST complex that functions at telomeres. This is an extremely interesting mechanism to understand, and this paper indeed reveals some interesting genetic results, leading to a compelling model, with potential impact on the field. The conclusions are largely supported by experiments examining protein-protein interactions at low resolution and ambiguous regarding directness of interactions like co-IP and yeast two-hybrid (Y2H) combined with genetics. However, some results appear contradictory and there's a lack of rigor in the experimental data needed to support claims. There is significant room for improvement and this work could certainly attain the quality needed to support the claims. The current version needs substantial revision and lacks necessary experimental detail. Stronger support for the claims would add detail to help distinguish competing models.

Specific comments:

Insufficient technical detail: I could find no explanation of how overexpression was achieved. No description of how teloChIP is performed, either for the PCNA IP or how the sequence analysis is performed. Too limited details on growth like exact temperatures for the cell cycle time course.

We have significantly expanded the Methods section to include all the technical information.

Please do not bold and underline text for emphasis-EVER

We have removed those from the text.

Lines 130-132: they have not shown "accumulation of SUMOylated PCNA" anywhere; this is an inference.

We have modified the text, it says: ”show that SUMOylated PCNA, and not unmodified or ubiquitinated PCNA, is both necessary and sufficient for telomere elongation in the presence or in the absence of Elg1.”

Fig 2A Can authors show any other very long-telomere mutant like stn1 that does show enhancement in combination with elg1∆ to show feasibility of such phenotype?

We don't think it is appropriate for the paper, but we have systematically created double mutants with elg1Δ and found many additive and even synergistic interactions. Here is an example. in Author response image 1, taken from the PhD thesis of Taly Ben-Shitrit, a PhD student in the lab.

Author response image 1.

What about cdc13 or ten1? Epistatic?

We did not test telomere length in combination with Ten1. Combining elg1 with cdc13-50 resulted in synergistic elongation. Given the complex genetic relationship between Stn1/Ten1 and Cdc13, it is hard to interpret this result.

Seems tenuous to use Y2H to decipher protein-protein interactions occurring out of context (i.e., not at telomere but at reporter gene promoter)

Y2H is a great method to detect interactions, even if they are transient. Whenever possible, we confirm our findings using co-IP or telo-ChIP.

Lines 268-270: It would be more accurate to state "can be" instead of "becomes" or "is" as they have not shown that SUMOylation or PCNA unloading have occurred.

We agree, and have changed the text.

Cdc13snm protein level?

Unfortunately our Western blot is not presentable, but the level of Cdc13snm was similar to that of the wt Cdc13, and this result has been already published by Hang et al., 2011.

Fig S3A: If SUMOylated Cdc13 mediates the Stn1-Elg1 interaction, why is Stn1-Elg1 interaction maintained in cdc13snm strain? This result seems to directly contradict the premise and overall conclusion of this section that Cdc13-SUMO mediates the (Y2H) interaction of Elg1 and Stn1.

According to our model, the interaction between Stn1 and Elg1 takes place upstream, and only then this complex interacts with SUMOylated Cdc13. Hence, if Cdc13 cannot be SUMOylated, the interaction Elg1-Stn1 is not lost, although Stn1 fails to interact with Cdc13, leading to a telomeric phenotype.

Line 279: which data establishes Stn1-Elg1 interaction as direct? Fig 2B co-Ip indicates physical but not necessarily direct interaction, but later the authors suggest that the interaction requires a SUMOylated intermediary, and Y2H in Fig. S3B doesn't demonstrate direct interaction.

We have changed the text, taking out the word "direct".

Co-Ip shows that interaction of Elg1 with Stn1 occurs mainly during later Sphase and with an overall delay compared to initial Elg1-Pol3 interaction.Co-IP Interaction between Cdc13 and Stn1 is reduced in the absence of Elg1

The subsection title: "The interaction of Elg1 with Stn1 takes place at telomeres only at late S-phase" is not well supported by the data. I agree the data are consistent with the idea of the interactions occurring at telomeres but there's no direct evidence of this.

We have changed the subsection title. It now reads: " The interaction of Elg1 with Stn1 takes place only at late S-phase"

Model: Is unloading happening at the fork? Doesn't PCNA unloading have to follow its loading which occurred behind the fork particularly on the lagging strand? Model now suggest that Stn1 itself is SUMOylated.

Yes, according to the model Elg1 moves with the fork, unloading PCNA from the lagging strand. Once Elg1 reaches the telomeres, it interacts with Stn1 (Figure 5). This interaction requires SUMOylation of Stn1 or of some other protein, which is not PCNA (Figure 3D) nor Cdc13 (Figure S3A) and could be Stn1 itself or another telomeric protein (Hang et al., 2011)

Title is rather vague.

We think it summarizes what we present in the paper.

Abstract:

"We report that SUMOylated PCNA acts as a signal that positively regulates telomerase activity."

I don't think this is supported or a good description of what they find

Figure 1B clearly shows that SUMO-PCNA is both necessary and sufficient for telomere elongation.

"and dissected the mechanism by which Elg1 and Stn1 negatively regulates telomere elongation, coordinated by SUMO."

Again, I don't think this is sufficiently supported and the model invokes SUMOylation events not demonstrated like Stn1, which might be a significant step forward.

On the positive side, their model makes several predictions that they could test much more directly and rigorously: for example, examining the impact of the relevant mutations in the recruitment of proteins to the telomere.

We have dissected the mechanism, and future work will be devoted to examining the impact of the relevant mutations in the recruitment of proteins to the telomere.

Reviewer #3 (Recommendations For The Authors):

Comments:

1) The telomere length analysis data presented here is consistent with an interpretation that Stn1 and Elg1 play roles in a similar telomere maintenance pathway because the telomere restriction fragment pattern in the double mutants are not longer than the stn1 single mutants. No comment is made with respect to the yellow bars in Figure 2 that presumably measure telomere length appearing to be slightly shorter than in the stn1 single mutants. It may be interesting and informative if the double mutants do in fact have some phenotype distinct from the single stn1 mutants. Is there an impact on viability in the double mutant?

Given the variable telomeric phenotype of the single stn1 mutants, slight variations in the measurement of the median telomere size are expected. The difference observed is not likely to be significant. What is important is that the double mutants with elg1 do not show longer telomeres. In terms of fitness, the stn1 mutants grow slightly slowly, but the elg1 mutation does not slow them down further.

2) It is somewhat surprising that no additional telomere length analysis is included that actually tests the proposed model, including whether this path could be operational only under certain conditions. Maybe this is a topic of the next paper?

Indeed, future work will explore the conditions under which PCNA SUMOylation is essential, and those under which is only needed.

3) Were the error bars in Figure 5F determined only from the experiment in E? Does this represent error in measuring the data from one biological replicate? The type of error should be made clear to avoid readers assuming the data represents measurements from more than one sample in more than one experiment. The data would be stronger if it represented measurements from multiple experiments.

The graph was made with data from three biological replicates. We show the best blot in Figure 5E. We have now stressed this in the Figure Legend.

4) Why was only one two hybrid reporter shown? Having the multiple reporters can give confidence in interactions. (Not a big deal here given the nice co-IP data.)

We thought that it is enough to show one reporter, as the results with a different reporter (B-gal assay) led to the same conclusions. since this did not add information and made the paper too lengthy (and boring), we took them out. In any case all data was verified by co-IP.

5) Line 414 - what are the 32P-radio labeled PCR fragments? Are these solely comprised of TG1-3 repeats of some length? A bit more detail in this aspect of the method could be helpful.

We have added an explanation on the probe in the Methods section.

6) Line 432-433 - which anti-HA or anti-My antibodies are these? (very minor detail)

We have added the details.

Author Response:

We would like to thank the eLife reviewers for the considerable time and effort they have invested to review these manuscripts. We have also benefited from a previous round of review of the manuscript describing the proposed burial features, which underwent two rounds of revisions in a high-impact journal over a period of approximately 8 months during 2022 and early 2023. Both sets of reviews have reflected mixed responses to the evidence we have presented, with one reviewer recommending acceptance with minor editorial revisions, two recommending acceptance with minor revisions and the fourth recommending rejection based upon similar arguments to those reflected by some of the reviewers in this current round of reviews in eLife. Ultimately the managing editor of this first journal took the decision that the review process could not be completed in a timely manner and rejected the manuscript although the submission here reflected our consideration of these reviewers suggestions.

We have chosen in this initial response to the eLife reviews to include some references to the previous anonymous reviews in order to illustrate differences of opinion and differences in revision suggestions within the review process. Our goal is to offer maximal insight into our decision-making process and to acknowledge the considerable time and effort put into the assessment of these manuscripts by reviewers (for eLife and in the case of the earlier review process). We hope that this approach will assist the readers, and reviewers, of our manuscripts in understanding why we are proceeding with certain decisions during the revision process.

This is a new process for us and the reviewers, and one way in which it significantly differs from more traditional review is that both the reviews and our reply will be public well in advance of our revisions to the manuscript. Indeed, considering the scope of the reviews, some of those revisions may take considerable time, although many can be accomplished fairly easily. Thus, we are not in a position to say that we have solved every issue raised by the reviewers. Instead, we will examine what appear to be the key critical issues raised regarding the data and the analyses and how we propose to address these as we revise the papers. We will also address several philosophical and ethical issues raised by the reviews and our proposal for dealing with these. More specific editorial and citational recommendations will be dealt with on a case-by-case basis, and we do not address these point-by-point in this reply. Please note, this response to the reviewers is not the revision of the manuscript and is only the initial opinion of the corresponding authors with some guidance from the larger group of authors of all three papers. Our final submitted revision will reflect the input of all authors included on those submissions.

We took the decision to submit three separate papers consciously. The two different categories of evidence, burials and engravings, involve different kinds of analysis and different (although overlapping) teams of researchers, and we recognized that each deserved their own presentation and assessment. Meanwhile, together they inform the context of H. naledi in a way that requires some synthetic discussion, in which both kinds of evidence are relevant, leading to a third paper. But the mutual relevance of these different kinds of evidence and their review by a common set of reviewers naturally raises cross-cutting issues, and the reviewers have cross-referenced the three articles. This has sometimes led to suggestions about one manuscript based on the contents of another. Considering the situation, we accepted the recommendation that it would be clearer to consider all three articles in a single reply. Thus, while each of the three papers will proceed separately during the revision process, it will be necessary to highlight across all three papers occasionally in our responses.

Scientific Issues:

In reading the reviews, we feel there are 9 critical points/assertions raised by one or more of the reviewers that present a problem for, or challenge to, our hypothesis that the observed evidence (bone accumulations and engravings) described in the Dinaledi subsystem are of intentional naledigenic origin. These are:

1. The evidence presented does not demonstrate a clear interruption of the floor sediments, thus failing to demonstrate excavated holes.

2. The sediments infilling the holes where the skeletal remains are found have not been demonstrated to originate from the disruption of the floor sediments and thus could be part of a natural geological process (e.g. water movement, slumping) or carnivore accumulations.

3. Previous geological interpretations by our research group have given alternative geological explanations for formation of the bony accumulations that contradict the present evidence presented here and result in alternative origins hypotheses.

4. Burial cannot be effectively assessed without complete excavation of the features and site.

5. The skeletal remains as presented do not conform clearly to typical body arrangement/positions associated with human (Homo sapiens) burials.

6. There is no evidence of grave goods or lithic scatters that are typically associated with human burials.

7. Humans may have been involved with the creation of either the Homo naledi bone accumulations, the engravings, or both.

8. Without a date of the engravings, the null hypothesis should be the engravings were created by Homo sapiens.

9. The null hypothesis for explanation of the skeletal remains in this situation should be “natural accumulation”.

Our analysis of the Dinaledi Feature 1 leads us to accept that the laminated orange-red mudstone (LORM) sedimentary layer is interrupted, indicating a non-natural intervention, and that the hole created by the interruption was then filled by both a fleshed body (and perhaps parts of other bodies) which were then covered by sediment that originated from the hole that was dug. We recognize that the four eLife reviewers are not convinced that our presentation is sufficient to establish this. Interestingly, this was not the universal opinion of earlier reviewers of the initial manuscript several of whom felt we had adequately supported this hypothesis. The lack of clarity in this current version of the burial manuscript is our responsibility. In the upcoming revision of this paper to be submitted, we will take the reviewers’ critiques to heart and add additional figures that illustrate better the disruption of the LORM and clarify the sedimentological data showing the material covering the skeletal remains in the hole are the disrupted sediments excavated from the same hole. We are proposing to isolate this most critical evidence for burial into a separate section in the revised submission based on the reviewers’ comments. The fact that the LORM layer is disrupted, a fleshed body was placed in the hole created by this disruption, and the body (and perhaps parts of other bodies) was/were then covered by the same sediments from the hole is the central feature of our hypothesis that the bone accumulations observed reflect a burial and not a natural process.

The possibility of fluvial transport or involvement in the subsystem is a topic that we have addressed extensively in past work, and it is clear from these reviews that we must enhance our current manuscript to discuss this issue at greater length. Our previous work (Dirks et al. 2015; Dirks et al. 2017) emphasized that fluvial transport of whole bodies into the subsystem was precluded by several lines of sedimentological evidence. We excavated a rich accumulation of skeletal remains, including articulated limbs and other elements in subvertical orientations inconsistent with slow sedimentary infill, which were difficult to explain without positing either a large and dense pile of bodies and/or sediment movement. We encountered fractured chunks of laminated orange-red mudstone (LORM) in random orientations within our excavation area, within and among skeletal remains, which directly refuted that the remains were inundated with water at the time of burial, and this limited the possibility of fluvial transport. Water flow sufficient to displace bodies or complete skeletal evidence would also transport large and course sediment, which is absent from the subsystem, and would sort the commingled skeletal material that we found by size, which we do not observe. But our excavation only covered less than a square meter at very limited depth, and this was the limit to our knowledge of subsurface sediment. We thus were left with uncertainty that led us to suggest the possibility of sediment slumping or movement into subsurface drains, although these were not observed near our excavation. Our current work expands our knowledge of the subsurface and presents an alternative explanation for the disposition of skeletal remains from our earlier excavation. But we acknowledge that this new explanation is vulnerable to our own previous published proposals, and we must do a better job of explaining how the new information addresses our previous suggestions. By not clearly creating a section where we explained how these previous hypotheses were now nullified by new evidence, we clearly confused the reviewers with our own previous work. We will revise the manuscript by enhancing the review of the significant geological evidence demonstrating that there is no significant fluvial action in the system and making it clear how the burial hypothesis provides a clearer explanation for the situation of skeletal remains from our previous excavation work.

One of the central issues raised by reviewers has been a perceived need to excavate these features completely, totally exhuming all skeletal remains from them. Reviewers have written that it is necessary to identify every skeletal element that is present and account for any missing elements. On this point, we have both ethical and scientific differences from these reviewers. We express our ethical concerns first. Many of the best-preserved possible burials ever discovered by archaeologists were subjected to total excavation and exhumation. Cases like La Chapelle-aux-Saints, La Ferrassie, and Skhūl were fully excavated at a time when data recording and excavation methods did not include the range of spatial and geomorphological approaches that later became routine. The judgment of early investigators that these situations were intentional burials was challenged by later workers, and the kind of information that might enable better tests had been irrevocably lost (Gargett 1999; Dibble et al. 2015; Rendu et al. 2014).

Later, improved excavation standards have not sufficed to remove uncertainty or debate about possible burials. For example, it was long presumed that well-preserved remains of young children were by themselves diagnostic of intentional burial, such as those from Dederiyeh, Border Cave, or Roc de Marsal. Such cases were also fully excavated, with adequate documentation of the positioning of skeletal remains and their surrounding stratigraphic situation, but such cases were later challenged on several bases and the complete exhumation of material has confused or precluded testing of new hypotheses (e.g. Gargett 1999). The case of Roc de Marsal is one in which data from the initial excavation combined with data from the initial excavation combined with re-excavation and geoarchaeological analysis led to a naturalistic interpretation of the skeletal material (Sandgathe et al. 2011; Goldberg et al. 2017). But even in this case, the researchers erred in their interpretation of the skeleton’s situation due to a lack of identification of parts of the infant’s skeleton (Gómez-Olivencia and García-Martinez 2019). That is to say, it is not only the burial hypothesis but other hypotheses that suffer from complete excavation. Researchers concerned with preserving all possible information have sometimes taken extraordinary measures to remove and study possible burials at high-resolution in the laboratory. Such was the case of the Shanidar IV burial removed from the site and transported in plaster jacket by Solecki, which led to the disruption and loss of internal stratigraphic information (Pomeroy et al. 2020). Arguably, the current state of the art is full excavation with partial preparation, such as that undertaken at Panga ya Saidi (Martinón-Torres et al. 2021). But again, any future attempt to reinterpret or test the hypothesis of burial must rely on the adequacy of documentation as the original context has been removed.

In our decision to leave material in place as much as possible, we are expanding upon standard practice to leave witness sections and unexcavated areas for future research. The situation is novel, representing possible burials by a nonhuman species, and that makes it doubly important in our opinion to be conservative in not fully exhuming the skeletal material from its context. We anticipate that many other researchers, including future investigators, will suggest additional methods to further test the hypothesis of burial, something that would be impossible if we had excavated the features in their entirety prior to publishing a description of our work. We believe strongly that our ethical responsibility is to publish the work and the most likely interpretation while leaving as much evidence in place as possible to enable further testing and replication. We welcome the suggestions of additional methods/analyses to test the H. naledi burial hypothesis.

This being said, we also observe that total exhumation would not resolve the concerns raised by the reviewers. The recommendation of total exhumation is in pursuit of a full account of all skeletal material present and its preservation and spatial situation, in order to demonstrate that they conform to body positions comparable to human burials. As has been highlighted in forensic casework, the excavation of an inhumation feature does not necessarily provide an accurate spatial or anatomical manifest of the stratigraphical relationships between the body, encapsulating matrix, and any cut present due to preservational, taphonomic and operational factors (Dirkmaat and Cabo, 2016; Hunter, 2014). In particular, in cases where skeletal elements are highly fragmented, friable, or degraded (such as through bioerosion) then complete excavation—even under controlled laboratory conditions—may destroy bone and severely limit skeletal identification (Henderson, 1997; Hochrein, 2002; Owsley and Compton, 1997), particularly in elements where the ratio of trabecular to cortical bone is high (Darwent and Lyman, 2002; Lyman, 1994). As such, non-invasive methods of 3D and 4D modelling (preservation in situ) are often considered preferable to complete necropsy or excavation (preservation by record) where appropriate (Bolliger and Thali, 2009; Dell’Unto and Landeschi, 2022; Randolph-Quinney et al., 2018; Silver, 2016). 

The test of burial is not primarily positional, but taphonomic and geological. The position and number of bones can elaborate on process-driven questions of decay and destruction in the burial environment, or post-mortem modification, but are not singularly indicative of whether the remains were intentionally buried – the post-mortem narrative of all the processes affecting the cadaveric island is required (Knüsel and Robb, 2016). In previous cases, researchers have disputed or accepted the hypothesis of intentional hominin burial based upon assumptions about how modern humans or Neandertals would have positioned bodies, with the idea that some positions reflect ritual intent while others do not. But applying such assumptions is unjustifiable, particularly for a species like H. naledi, whose culture may have differed fundamentally from our own. Our work acknowledges that the present evidence does not enable a full reconstruction of the burial positions, but it does show that fleshed remains were encased in sediment prior to decomposition of soft tissue, and that subsequent spatial changes can be most parsimoniously explained by natural decomposition within sedimentary matrix contained within a burial feature (after Green, 2022; Mickleburgh and Wescott, 2018; Mickleburgh et al., 2022). If the argument is that extraordinary claims require extraordinary evidence, we feel that the evidence documents excavation and interment (and will do so more clearly in the revision) and the fact of the remains do not match a “typical” human burial in body positioning is not in itself evidence that these are not H. naledi burials.

We feel that the reviewers (in keeping with many palaeoanthropologists) have a clear idea of what they “think” a burial should look like in an idealised sense, but this platonic ideal of burial form is not matched by the extensive literature in archaeothanatology, funerary archaeology and forensic science which indicates enormous variability in the activity, morphology and post-mortem system experienced by the human body in cases of interment and body disposal (e.g. Aspöck, 2008; Boulestin and Duday, 2005 and 2006; Connelly et al., 2005; Channing and Randolph-Quinney, 2006; Cherryson, 2008; Donnelly et al., 1995; Finley, 2000; Hunter, 2014; Parker Pearson, 1999; Randolph-Quinney, 2013). Decades of experience in the identification, recovery and interpretation of clandestine, deviant, and non-formal burials indicates the platonic ideal is rare, and in many contexts, the exception (Cherryson, 2008; Parker Pearson, 1999). This variability is particularly relevant to morphological traits in burial context, such as the informal nature of the grave cut in plan and section, shallow burial depth, and initial disposition of body (placement) during the early post-mortem period. These might run counter to the expectations of reviewers or others referencing the fossil hominin record, but are well accepted within the communities of researchers investigating Holocene archaeological sites and forensic contexts.

It is encouraging to see reviewers beginning to incorporate the extensive (often experimentally derived) literature from archaeothanatology and forensic taphonomy in their deliberations, and we will be taking these comments on board going forward. In particular, we acknowledge reviewers’ comments and the need to construct a more detailed post-mortem narrative, accounting for joint disarticulation (labile versus persistent joints etc), displacement, and final disposition of elements within the burial space. As such we will incorporate the hierarchy of decomposition (rank order disarticulation), associations between regions of anatomical association, areas of disassociation, and the voids produced during decomposition (after Mickleburgh and Wescott, 2018; Mickleburgh et al., 2022) into our narrative. In doing so we acknowledge the tensions between the inductive archaeolothanatological narrative-driven approach (e.g. Duday, 2005 & 2009) versus robust decomposition data derived from human forensic taphonomic experimentation recently articulated by Schotsmans and colleagues (2022) - noting that we will highlight comparative data based on forensic experimental casework and actualistic modelling over inductive intuitive approaches which come with significant evidential shortcomings (Bristow et al. 2011).

Finally, from a taphonomic perspective it is worth pointing out to reviewers that we have already addressed the issue of lack of taphonomic evidence for carnivore involvement in the formation of the Dinaledi assemblage (Dirks, et al., 2016). Absence of any carnivore-induced bone surface modifications, patterns of skeletal part representation, and a total absence of any carnivore remains found within the Dinaledi chamber (following Kuhn and colleagues, 2010) lead us to reject carnivores as possible vectors of body accumulation within the Dinaledi Chamber and Hill Antechamber.

Reviewers suggest that without a date derived from geochronological methods, the engravings cannot be associated with H. naledi, and that it is possible (or probable) that the engravings were done in the recent past by H. sapiens. This suggestion neglects the context of the site. We have previously documented the structure and extremely limited accessibility of the Dinaledi subsystem. This subsystem was not recorded on maps of the documented Rising Star Cave system prior to our work and its discovery by our teams. Furthermore, there is no evidence of prehistoric human activity in the areas of the cave related to possible subterranean entrances There is no evidence that humans in the past typically ventured into such extreme spaces like those of Rising Star. It is clear from the presence of the remains of many individuals that H. naledi ventured into these spaces again and again. It is likely that H. naledi moved through these spaces more easily than humans do based on their physique. We show that the engravings overlay each other suggesting multiple engraving events.  These engravings took time and effort and the only evidence for use of the Dinaledi subsystem by any hominin is by H. naledi. The context leads to the null hypothesis that H. naledi made the marks. In our revision, we will elaborate on this argument to clarify the evidence for our stance on this hypothesis. Several reviewers took issue with the title of the engraving paper as we did not insert a qualifier in front of the suggested date range for the engravings. We deliberately left out qualifying language so that the title took the form of a testable hypothesis rather than a weak assertation. Should future work find the engravings were not produced within this time range, then we will restate this hypothesis.

Finally, with regards to the engravings we have chosen to report them because they exist. Not reporting the presence of engraved marks on the walls of a cave above hypothesized burials would be tantamount to leaving relevant evidence out of the description of an archeological context. We recognize and state in our manuscript that these markings require substantial further study, including attempts at geochronological dating. But the current evidence is clearly relevant to the archaeological context of the subsystem. We take a similar stance with reporting the presence of the tool shaped artefact near the hand of the H. naledi skeleton in the Hill Antechamber. It is evident that this object requires further study, as we stated in our manuscript, but again omitting it from our study would be leaving out relevant evidence.

Some have suggested that the null hypothesis should be that all of these observed circumstances are of natural origin. Our team took this approach in our early investigation of the Dinaledi subsystem (Dirks et al. 2015). We adopted the null hypothesis that the geological processes involved in the accumulation of H. naledi skeletal remains were “natural” (e.g., non-naledigenic involvement), and we were able to reject many alternative explanations for the assemblage, including carnivore accumulation, “death trap” accumulation, and fluvial transport of bodies or bones (Dirks et al. 2015). This led us to the hypothesis that H. naledi were involved in bringing the bodies into the spaces where they were found. But we did not hypothesize their involvement in the formation of the deposit itself beyond bringing the bodies to the location.

This approach seems conservative. It followed the traditional view that small-brained hominins do not engage in cultural practices. But we recognize in hindsight that this null hypothesis approach did harm to our analyses. It impeded us from recognizing within our initial excavations of the puzzle box area and other excavations between 2014 – 2017 that we might be encountering remains that were intrusive in the sedimentary floor of the chamber. If we had approached the accumulation of a large number of hominins from the perspective of the null hypothesis being that the situation was likely cultural, we perhaps would have collected evidence in a slightly different manner. We certainly note that if the Dinaledi system had been full of the remains of modern humans, there would have been little doubt that the null hypothesis would have been that this was a cultural space and not a “natural space”.  We therefore respectfully disagree with the reviewers who continue to support the idea that we should approach hominin excavations with the null hypothesis that they will be natural (specifically non-cultural) in origins. If excavations continue with this mindset we believe that potential cultural evidence is almost certain to be lost.

There has been a gradient across paleoanthropological excavations, archaeological work, and forensic investigation, with increasing precision of context. The reality is that the recording precision and frame of approach is typically different in most paleontological excavations than in those related to contemporary human remains. If anything comes from the present discussion of whether the Dinaledi system is a burial site for H. naledi or not, we hope that by taking seriously the possibility of deep cultural dynamics of hominins, we will encourage other teams to meet the highest standards of excavation in order to preserve potential cultural evidence. Given H. naledi’s cranial capacity we suggest that even very early hominin skeletal assemblages should be re-examined, if there is sufficient evidence or records available.  These would include examples such as the A.L. 333 Au. afarensis site (the so called First Family site in Hadar Ethiopia), the Dikika infant skeleton, WT 15000 (Turkana Boy) and even A.L. 288 (Lucy) as such unusual taphonomic situations where skeletons are preserved cannot be simply explained away as “natural” in origin, based solely on the cranial capacity and assumed lack of cognitive and cultural complexity of the hominins as emphasized by us in Fuentes et al. (2023). We are not the first to observe that some very early hominin situations may represent early mortuary activity (Pettitt 2013), but we would advocate a step further. We suggest it may be damaging to take “natural accumulation” as the standard null hypothesis for hominin paleoanthropology, and that it is more conservative in practice to engage remains with the null hypothesis of possible cultural formation.

We are deeply grateful for the time and effort all of the 8 reviewers (across three reviews) have taken with this work.  We also acknowledge the anonymous reviewers from previous submissions who’s opinions and comments will have made the final iterations of these manuscripts better for their efforts. As this process is rather public and includes commentary outside of the eLife forum, we ask that the efforts of all 37 authors and 8 reviewers involved be respected and that the discourse remain professional in all venues as we study this fascinating and quite complex occurrence. We appreciate also the efforts of members of the public who have engaged with this relatively new process where preprints are posted prior to the reviews allowing comments and interactions from colleagues and the public who are normally not part of the internal peer review process.  We believe these interactions will make for better final papers. We feel we have met the standards of demonstrating burials in H. naledi and that the engraving are most likely associated with H. naledi. However, given the reviews we see many areas where our clarity and context, and analyses, were less strong than they can be. With the clarifications and additions taken on board through these review processes the final papers will be stronger and clearer. We, recognize that this is an ongoing process of scientific investigation and further work will allow continued, and possibly better, evaluation of these hypothesis and others.

Lee R Berger, Agustín Fuentes, John Hawks, Tebogo Makhubela

Works cited:

• Aspöck, E. (2008). What Actually is a ‘Deviant Burial’?: Comparing German-Language and Anglophone Research on ‘Deviant Burials.’ In E. M. Murphy (Ed.). Deviant Burial in the Archaeological Record. Oxford: Oxbow Books.  pp 17–34.

• Bolliger, S.A. & Thali, M.J. (2009). Thanatology. In S.A. Bolliger and M.J. Thali (eds) Virtopsy Approach:  3D Optical and Radiological Scanning and Reconstruction in Forensic Medicine. Boca Raton: CRC Press. pp 187-218.

• Boulestin, B. & Duday, H. (2005). Ethnologie et archéologie de la mort: de l’illusion des références à l’emploi d’un vocabulaire. In: C. Mordant and G. Depierre (eds) Les Pratiques Funéraires à l’Âge du Bronze en France. Actes de la table ronde de Sens-en-Bourgogne. Paris: Éditions du Comité des Travaux Historiques et Scientifiques. pp. 17–30.

• Boulestin, B. & Duday, H. (2006). Ethnology and archaeology of death: from the illusion of references to the use of a terminology. Archaeologia Polona 44: 149–169.

• Bristow, J., Simms, Z. & Randolph-Quinney, P.S. Taphonomy. In S. Black and E. Ferguson (eds.) Forensic Anthropology 2000-2010. Boca Raton, FL: CRC Press. pp 279-318.

• Channing, J. & Randolph-Quinney, P.S. (2006). Death, decay and reconstruction: the archaeology of Ballykilmore Cemetery, County Westmeath. In J. O’Sullivan and M. Stanley (eds.) Settlement, Industry and Ritual: Archaeology. National Roads Authority Monograph Series No. 3. Dublin: NRA/Four Courts Press. pp 113-126.

• Cherryson, A. K. (2008). Normal, Deviant and Atypical: Burial Variation in Late Saxon Wessex, c. AD 700–1100. In E. M. Murphy (Ed.). Deviant Burial in the Archaeological Record. Oxford: Oxbow Books. pp 115–130.

• Connolly, M., F. Coyne & L. G. Lynch (2005). Underworld : Death and Burial in Cloghermore Cave, Co. Kerry. Bray, Co. Wicklow: Wordwell.

• Darwent, C. M. & R. L. Lyman (2002). Detecting  the postburial fragmentation of carpals, tarsals and phalanges. In M. H. Sorg and W. D. Haglund (eds). Advances in Forensic Taphonomy: Method, Theory and Archeological Perspectives. Boca Raton, FL, CRC Press. pp 355-378.

• d’Errico, F., & Backwell, L. (2016). Earliest evidence of personal ornaments associated with burial: The Conus shells from Border Cave. Journal of Human Evolution, 93, 91–108.

• De Villiers. H. (1973). Human skeletal remains from Border Cave, Ingwavuma District, KwaZulu, South Africa. Annals of the Transvaal Museum, 28(13), 229–246.

• Dell’Unto, N. and Landeschi, G. (2022). Archaeological 3D GIS. London: Routledge.

• Dibble, H. L., Aldeias, V., Goldberg, P., McPherron, S. P., Sandgathe, D., & Steele, T. E. (2015). A critical look at evidence from La Chapelle-aux-Saints supporting an intentional Neandertal burial. Journal of Archaeological Science, 53, 649–657.

• Dirkmaat, D. C., & Cabo, L. L. (2016). Forensic archaeology and forensic taphonomy: basic considerations on how to properly process and interpret the outdoor forensic scene_. Academic Forensic Pathology_ 6, 439–454.

• Dirks, P. H., Berger, L. R., Roberts, E. M., Kramers, J. D., Hawks, J., Randolph-Quinney, P. S., Elliott, M., Musiba, C. M., Churchill, S. E., de Ruiter, D. J., Schmid, P., Backwell, L. R., Belyanin, G. A., Boshoff, P., Hunter, K. L., Feuerriegel, E. M., Gurtov, A., Harrison, J. du G., Hunter, R., … Tucker, S. (2015). Geological and taphonomic context for the new hominin species Homo naledi from the Dinaledi Chamber, South Africa. ELife, 4, e09561.

• Dirks, P.H.G.M., Berger, L.R., Hawks, J., Randolph-Quinney, P.S., Backwell, L.R., and Roberts, E.M. (2016). Comment on “Deliberate body disposal by hominins in the Dinaledi Chamber, Cradle of Humankind, South Africa?” [J. Hum. Evol. 96 (2016) 145-148]. Journal of Human Evolution 96:  149-153.

• Dirks, P. H., Roberts, E. M., Hilbert-Wolf, H., Kramers, J. D., Hawks, J., Dosseto, A., Duval, M., Elliott, M., Evans, M., Grün, R., Hellstrom, J., Herries, A. I., Joannes-Boyau, R., Makhubela, T. V., Placzek, C. J., Robbins, J., Spandler, C., Wiersma, J., Woodhead, J., & Berger, L. R. (2017). The age of Homo naledi and associated sediments in the Rising Star Cave, South Africa. ELife, 6, e24231.

• Donnelly, S., C. Donnelly & E. Murphy (1999). The forgotten dead: The cíllíní and disused burial grounds of Ballintoy, County Antrim. Ulster Journal of Archaeology 58, 109-113.

• Duday, H. (2005). L’archéothanatologie ou l’archéologie de la mort. In: O. Dutour, J.-J. Hublin and B. Vandermeersch (eds) Objets et Méthodes en Paléoanthropologie. Paris: Comité des Travaux Historiques et Scientifiques. pp. 153–215.

• Duday, H. (2009). Archaeology of the Dead: Lectures in Archaeothanatology. Oxford: Oxbow Books.

• Finley, N. (2000). Outside of life: Traditions of infant burial in Ireland from cillin to cist.  World Archaeology 31, 407-422.

• Gargett, R. H. (1999). Middle Palaeolithic burial is not a dead issue: The view from Qafzeh, Saint-Césaire, Kebara, Amud, and Dederiyeh. Journal of Human Evolution, 37(1), 27–90.

• Goldberg, P., Aldeias, V., Dibble, H., McPherron, S., Sandgathe, D., & Turq, A. (2017). Testing the Roc de Marsal Neandertal “Burial” with Geoarchaeology. Archaeological and Anthropological Sciences, 9(6), 1005–1015.

• Gómez-Olivencia, A., & García-Martínez, D. (2019). New postcranial remains from the Roc de Marsal Neandertal child. PALEO. Revue d’archéologie Préhistorique, 30–1, 30–1.

• Green, E.C. (2022). An archaeothanatological approach to the identification of late Anglo-Saxon burials in wooden containers. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 436-455.

• Henderson, J. (1987). Factors determining the state of preservation of human remains. In A. Boddington, A. Garland and R. Janaway (eds). Death, Decay and Reconstruction: Approaches to Archaeology and Forensic Science. Manchester: Manchester University Press. pp 43-54.

• Hunter, J. R. (2014). Human remains recovery: archaeological and forensic perspectives. In C. Smith (ed). Encyclopedia of Global Archaeology. New York: Springer New York. pp 3549-3556.

• Hochrein, M. (2002). An Autopsy of the Grave: Recognizing, Collecting and Preserving Forensic Geotaphonomic Evidence. In M. H. Sorg and W. D. Haglund (eds). Advances in Forensic Taphonomy: Method, Theory and Archeological Perspectives. Boca Raton, FL, CRC Press: 45-70.

• Knüsel, C.K. & Robb, J. (2016). Funerary taphonomy: An overview of goals and methods. Journal of Archaeological Science: Reports 10, 655-673.

• Kuhn, B.F., Berger, L.R. & Skinner, J.D. (2010). Examining criteria for identifying and differentiating fossil faunal assemblages accumulated by hyenas and hominins using extant hyenid accumulations. International Journal of Osteoarchaeology 20, 15-35.

• Lyman, R. (1994). Vertebrate Taphonomy. Cambridge, Cambridge University Press.

• Martinón-Torres, M., d’Errico, F., Santos, E., Álvaro Gallo, A., Amano, N., Archer, W., Armitage, S. J., Arsuaga, J. L., Bermúdez de Castro, J. M., Blinkhorn, J., Crowther, A., Douka, K., Dubernet, S., Faulkner, P., Fernández-Colón, P., Kourampas, N., González García, J., Larreina, D., Le Bourdonnec, F.-X., … Petraglia, M. D. (2021). Earliest known human burial in Africa. Nature, 593(7857), 7857.

• Mickleburgh, H.L & Wescott, D.J. (2018). Controlled experimental observations on joint disarticulation and bone displacement of a human body in an open pit: implications for funerary archaeology. Journal of Archaeological Science: Reports 20: 158-167.

• Mickleburgh, H.L., Wescott, D.J., Gluschitz, S. & Klinkenberg, V.M. (2022). Exploring the use of actualistic forensic taphonomy in the study of (forensic) archaeological human burials: An actualistic experimental research programme at the Forensic Anthropology Center at Texas State University (FACTS), San Marcos, Texas. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 542-562.

• Owsley, D. & B. Compton (1997). Preservation in late 19th Century iron coffin burials. In W. Haglund and M. Sorg (eds). Forensic Taphonomy: The Postmortem Fate of Human Remains. Boca Raton, FL, CRC Press: 511-526.

• Parker Pearson, M. (1999). The Archaeology of Death and Burial. College Station: Texas A&M University Press.

• Pettitt, P. (2013). The Palaeolithic Origins of Human Burial. Routledge.

• Pomeroy, E., Bennett, P., Hunt, C. O., Reynolds, T., Farr, L., Frouin, M., Holman, J., Lane, R., French, C., & Barker, G. (2020). New Neanderthal remains associated with the ‘flower burial’ at Shanidar Cave. Antiquity, 94(373), 11–26.

• Randolph-Quinney, P.S. (2013). From the cradle to the grave: the bioarchaeology of Clonfad 3 and Ballykilmore 6. In N. Brady, P. Stevens and J. Channing (eds.). Settlement and Community in the Fir Tulach Kingdom. Dublin: National Roads Authority Press. pp A2.1-48.

• Randolph-Quinney, P.S., Haines, S. and Kruger, A. (2018). The use of three-dimensional scanning and surface capture methods in recording forensic taphonomic traces: issues of technology, visualisation, and validation. In: W.J. M. Groen and P. M. Barone (eds). Multidisciplinary Approaches to Forensic Archaeology. Berlin: Springer International Publishing, pp. 115-130.

• Rendu, W., Beauval, C., Crevecoeur, I., Bayle, P., Balzeau, A., Bismuth, T., Bourguignon, L., Delfour, G., Faivre, J.-P., Lacrampe-Cuyaubère, F., Tavormina, C., Todisco, D., Turq, A., & Maureille, B. (2014). Evidence supporting an intentional Neandertal burial at La Chapelle-aux-Saints. Proceedings of the National Academy of Sciences, 111(1), 81–86.

• Sandgathe, D. M., Dibble, H. L., Goldberg, P., & McPherron, S. P. (2011). The Roc de Marsal Neandertal child: A reassessment of its status as a deliberate burial. Journal of Human Evolution, 61(3), 243–253.

• Silver, M. (2016). Conservation Techniques in Cultural Heritage. In E. Stylianidis and F. Remondino (eds) 3D Recording, Documentation and Management of Cultural Heritage. Dunbeath: Whittles Publishing. pp 15-106.

• Schotsmans, E.M.J., Georges-Zimmermann, P., Ueland, M. and Dent, B.B. (2022). From flesh to bone: Building bridges between taphonomy, archaeothanatology and forensic science for a better understanding of mortuary practices. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 501-541.

Author Response:

We would like to thank the eLife reviewers for the considerable time and effort they have invested to review these manuscripts. We have also benefited from a previous round of review of the manuscript describing the proposed burial features, which underwent two rounds of revisions in a high-impact journal over a period of approximately 8 months during 2022 and early 2023. Both sets of reviews have reflected mixed responses to the evidence we have presented, with one reviewer recommending acceptance with minor editorial revisions, two recommending acceptance with minor revisions and the fourth recommending rejection based upon similar arguments to those reflected by some of the reviewers in this current round of reviews in eLife. Ultimately the managing editor of this first journal took the decision that the review process could not be completed in a timely manner and rejected the manuscript although the submission here reflected our consideration of these reviewers suggestions.

We have chosen in this initial response to the eLife reviews to include some references to the previous anonymous reviews in order to illustrate differences of opinion and differences in revision suggestions within the review process. Our goal is to offer maximal insight into our decision-making process and to acknowledge the considerable time and effort put into the assessment of these manuscripts by reviewers (for eLife and in the case of the earlier review process). We hope that this approach will assist the readers, and reviewers, of our manuscripts in understanding why we are proceeding with certain decisions during the revision process.

This is a new process for us and the reviewers, and one way in which it significantly differs from more traditional review is that both the reviews and our reply will be public well in advance of our revisions to the manuscript. Indeed, considering the scope of the reviews, some of those revisions may take considerable time, although many can be accomplished fairly easily. Thus, we are not in a position to say that we have solved every issue raised by the reviewers. Instead, we will examine what appear to be the key critical issues raised regarding the data and the analyses and how we propose to address these as we revise the papers. We will also address several philosophical and ethical issues raised by the reviews and our proposal for dealing with these. More specific editorial and citational recommendations will be dealt with on a case-by-case basis, and we do not address these point-by-point in this reply. Please note, this response to the reviewers is not the revision of the manuscript and is only the initial opinion of the corresponding authors with some guidance from the larger group of authors of all three papers. Our final submitted revision will reflect the input of all authors included on those submissions.

We took the decision to submit three separate papers consciously. The two different categories of evidence, burials and engravings, involve different kinds of analysis and different (although overlapping) teams of researchers, and we recognized that each deserved their own presentation and assessment. Meanwhile, together they inform the context of H. naledi in a way that requires some synthetic discussion, in which both kinds of evidence are relevant, leading to a third paper. But the mutual relevance of these different kinds of evidence and their review by a common set of reviewers naturally raises cross-cutting issues, and the reviewers have cross-referenced the three articles. This has sometimes led to suggestions about one manuscript based on the contents of another. Considering the situation, we accepted the recommendation that it would be clearer to consider all three articles in a single reply. Thus, while each of the three papers will proceed separately during the revision process, it will be necessary to highlight across all three papers occasionally in our responses.

Scientific Issues:

In reading the reviews, we feel there are 9 critical points/assertions raised by one or more of the reviewers that present a problem for, or challenge to, our hypothesis that the observed evidence (bone accumulations and engravings) described in the Dinaledi subsystem are of intentional naledigenic origin. These are:

1. The evidence presented does not demonstrate a clear interruption of the floor sediments, thus failing to demonstrate excavated holes.

2. The sediments infilling the holes where the skeletal remains are found have not been demonstrated to originate from the disruption of the floor sediments and thus could be part of a natural geological process (e.g. water movement, slumping) or carnivore accumulations.

3. Previous geological interpretations by our research group have given alternative geological explanations for formation of the bony accumulations that contradict the present evidence presented here and result in alternative origins hypotheses.

4. Burial cannot be effectively assessed without complete excavation of the features and site.

5. The skeletal remains as presented do not conform clearly to typical body arrangement/positions associated with human (Homo sapiens) burials.

6. There is no evidence of grave goods or lithic scatters that are typically associated with human burials.

7. Humans may have been involved with the creation of either the Homo naledi bone accumulations, the engravings, or both.

8. Without a date of the engravings, the null hypothesis should be the engravings were created by Homo sapiens.

9. The null hypothesis for explanation of the skeletal remains in this situation should be “natural accumulation”.

Our analysis of the Dinaledi Feature 1 leads us to accept that the laminated orange-red mudstone (LORM) sedimentary layer is interrupted, indicating a non-natural intervention, and that the hole created by the interruption was then filled by both a fleshed body (and perhaps parts of other bodies) which were then covered by sediment that originated from the hole that was dug. We recognize that the four eLife reviewers are not convinced that our presentation is sufficient to establish this. Interestingly, this was not the universal opinion of earlier reviewers of the initial manuscript several of whom felt we had adequately supported this hypothesis. The lack of clarity in this current version of the burial manuscript is our responsibility. In the upcoming revision of this paper to be submitted, we will take the reviewers’ critiques to heart and add additional figures that illustrate better the disruption of the LORM and clarify the sedimentological data showing the material covering the skeletal remains in the hole are the disrupted sediments excavated from the same hole. We are proposing to isolate this most critical evidence for burial into a separate section in the revised submission based on the reviewers’ comments. The fact that the LORM layer is disrupted, a fleshed body was placed in the hole created by this disruption, and the body (and perhaps parts of other bodies) was/were then covered by the same sediments from the hole is the central feature of our hypothesis that the bone accumulations observed reflect a burial and not a natural process.

The possibility of fluvial transport or involvement in the subsystem is a topic that we have addressed extensively in past work, and it is clear from these reviews that we must enhance our current manuscript to discuss this issue at greater length. Our previous work (Dirks et al. 2015; Dirks et al. 2017) emphasized that fluvial transport of whole bodies into the subsystem was precluded by several lines of sedimentological evidence. We excavated a rich accumulation of skeletal remains, including articulated limbs and other elements in subvertical orientations inconsistent with slow sedimentary infill, which were difficult to explain without positing either a large and dense pile of bodies and/or sediment movement. We encountered fractured chunks of laminated orange-red mudstone (LORM) in random orientations within our excavation area, within and among skeletal remains, which directly refuted that the remains were inundated with water at the time of burial, and this limited the possibility of fluvial transport. Water flow sufficient to displace bodies or complete skeletal evidence would also transport large and course sediment, which is absent from the subsystem, and would sort the commingled skeletal material that we found by size, which we do not observe. But our excavation only covered less than a square meter at very limited depth, and this was the limit to our knowledge of subsurface sediment. We thus were left with uncertainty that led us to suggest the possibility of sediment slumping or movement into subsurface drains, although these were not observed near our excavation. Our current work expands our knowledge of the subsurface and presents an alternative explanation for the disposition of skeletal remains from our earlier excavation. But we acknowledge that this new explanation is vulnerable to our own previous published proposals, and we must do a better job of explaining how the new information addresses our previous suggestions. By not clearly creating a section where we explained how these previous hypotheses were now nullified by new evidence, we clearly confused the reviewers with our own previous work. We will revise the manuscript by enhancing the review of the significant geological evidence demonstrating that there is no significant fluvial action in the system and making it clear how the burial hypothesis provides a clearer explanation for the situation of skeletal remains from our previous excavation work.

One of the central issues raised by reviewers has been a perceived need to excavate these features completely, totally exhuming all skeletal remains from them. Reviewers have written that it is necessary to identify every skeletal element that is present and account for any missing elements. On this point, we have both ethical and scientific differences from these reviewers. We express our ethical concerns first. Many of the best-preserved possible burials ever discovered by archaeologists were subjected to total excavation and exhumation. Cases like La Chapelle-aux-Saints, La Ferrassie, and Skhūl were fully excavated at a time when data recording and excavation methods did not include the range of spatial and geomorphological approaches that later became routine. The judgment of early investigators that these situations were intentional burials was challenged by later workers, and the kind of information that might enable better tests had been irrevocably lost (Gargett 1999; Dibble et al. 2015; Rendu et al. 2014).

Later, improved excavation standards have not sufficed to remove uncertainty or debate about possible burials. For example, it was long presumed that well-preserved remains of young children were by themselves diagnostic of intentional burial, such as those from Dederiyeh, Border Cave, or Roc de Marsal. Such cases were also fully excavated, with adequate documentation of the positioning of skeletal remains and their surrounding stratigraphic situation, but such cases were later challenged on several bases and the complete exhumation of material has confused or precluded testing of new hypotheses (e.g. Gargett 1999). The case of Roc de Marsal is one in which data from the initial excavation combined with data from the initial excavation combined with re-excavation and geoarchaeological analysis led to a naturalistic interpretation of the skeletal material (Sandgathe et al. 2011; Goldberg et al. 2017). But even in this case, the researchers erred in their interpretation of the skeleton’s situation due to a lack of identification of parts of the infant’s skeleton (Gómez-Olivencia and García-Martinez 2019). That is to say, it is not only the burial hypothesis but other hypotheses that suffer from complete excavation. Researchers concerned with preserving all possible information have sometimes taken extraordinary measures to remove and study possible burials at high-resolution in the laboratory. Such was the case of the Shanidar IV burial removed from the site and transported in plaster jacket by Solecki, which led to the disruption and loss of internal stratigraphic information (Pomeroy et al. 2020). Arguably, the current state of the art is full excavation with partial preparation, such as that undertaken at Panga ya Saidi (Martinón-Torres et al. 2021). But again, any future attempt to reinterpret or test the hypothesis of burial must rely on the adequacy of documentation as the original context has been removed.

In our decision to leave material in place as much as possible, we are expanding upon standard practice to leave witness sections and unexcavated areas for future research. The situation is novel, representing possible burials by a nonhuman species, and that makes it doubly important in our opinion to be conservative in not fully exhuming the skeletal material from its context. We anticipate that many other researchers, including future investigators, will suggest additional methods to further test the hypothesis of burial, something that would be impossible if we had excavated the features in their entirety prior to publishing a description of our work. We believe strongly that our ethical responsibility is to publish the work and the most likely interpretation while leaving as much evidence in place as possible to enable further testing and replication. We welcome the suggestions of additional methods/analyses to test the H. naledi burial hypothesis.

This being said, we also observe that total exhumation would not resolve the concerns raised by the reviewers. The recommendation of total exhumation is in pursuit of a full account of all skeletal material present and its preservation and spatial situation, in order to demonstrate that they conform to body positions comparable to human burials. As has been highlighted in forensic casework, the excavation of an inhumation feature does not necessarily provide an accurate spatial or anatomical manifest of the stratigraphical relationships between the body, encapsulating matrix, and any cut present due to preservational, taphonomic and operational factors (Dirkmaat and Cabo, 2016; Hunter, 2014). In particular, in cases where skeletal elements are highly fragmented, friable, or degraded (such as through bioerosion) then complete excavation—even under controlled laboratory conditions—may destroy bone and severely limit skeletal identification (Henderson, 1997; Hochrein, 2002; Owsley and Compton, 1997), particularly in elements where the ratio of trabecular to cortical bone is high (Darwent and Lyman, 2002; Lyman, 1994). As such, non-invasive methods of 3D and 4D modelling (preservation in situ) are often considered preferable to complete necropsy or excavation (preservation by record) where appropriate (Bolliger and Thali, 2009; Dell’Unto and Landeschi, 2022; Randolph-Quinney et al., 2018; Silver, 2016). 

The test of burial is not primarily positional, but taphonomic and geological. The position and number of bones can elaborate on process-driven questions of decay and destruction in the burial environment, or post-mortem modification, but are not singularly indicative of whether the remains were intentionally buried – the post-mortem narrative of all the processes affecting the cadaveric island is required (Knüsel and Robb, 2016). In previous cases, researchers have disputed or accepted the hypothesis of intentional hominin burial based upon assumptions about how modern humans or Neandertals would have positioned bodies, with the idea that some positions reflect ritual intent while others do not. But applying such assumptions is unjustifiable, particularly for a species like H. naledi, whose culture may have differed fundamentally from our own. Our work acknowledges that the present evidence does not enable a full reconstruction of the burial positions, but it does show that fleshed remains were encased in sediment prior to decomposition of soft tissue, and that subsequent spatial changes can be most parsimoniously explained by natural decomposition within sedimentary matrix contained within a burial feature (after Green, 2022; Mickleburgh and Wescott, 2018; Mickleburgh et al., 2022). If the argument is that extraordinary claims require extraordinary evidence, we feel that the evidence documents excavation and interment (and will do so more clearly in the revision) and the fact of the remains do not match a “typical” human burial in body positioning is not in itself evidence that these are not H. naledi burials.

We feel that the reviewers (in keeping with many palaeoanthropologists) have a clear idea of what they “think” a burial should look like in an idealised sense, but this platonic ideal of burial form is not matched by the extensive literature in archaeothanatology, funerary archaeology and forensic science which indicates enormous variability in the activity, morphology and post-mortem system experienced by the human body in cases of interment and body disposal (e.g. Aspöck, 2008; Boulestin and Duday, 2005 and 2006; Connelly et al., 2005; Channing and Randolph-Quinney, 2006; Cherryson, 2008; Donnelly et al., 1995; Finley, 2000; Hunter, 2014; Parker Pearson, 1999; Randolph-Quinney, 2013). Decades of experience in the identification, recovery and interpretation of clandestine, deviant, and non-formal burials indicates the platonic ideal is rare, and in many contexts, the exception (Cherryson, 2008; Parker Pearson, 1999). This variability is particularly relevant to morphological traits in burial context, such as the informal nature of the grave cut in plan and section, shallow burial depth, and initial disposition of body (placement) during the early post-mortem period. These might run counter to the expectations of reviewers or others referencing the fossil hominin record, but are well accepted within the communities of researchers investigating Holocene archaeological sites and forensic contexts.

It is encouraging to see reviewers beginning to incorporate the extensive (often experimentally derived) literature from archaeothanatology and forensic taphonomy in their deliberations, and we will be taking these comments on board going forward. In particular, we acknowledge reviewers’ comments and the need to construct a more detailed post-mortem narrative, accounting for joint disarticulation (labile versus persistent joints etc), displacement, and final disposition of elements within the burial space. As such we will incorporate the hierarchy of decomposition (rank order disarticulation), associations between regions of anatomical association, areas of disassociation, and the voids produced during decomposition (after Mickleburgh and Wescott, 2018; Mickleburgh et al., 2022) into our narrative. In doing so we acknowledge the tensions between the inductive archaeolothanatological narrative-driven approach (e.g. Duday, 2005 & 2009) versus robust decomposition data derived from human forensic taphonomic experimentation recently articulated by Schotsmans and colleagues (2022) - noting that we will highlight comparative data based on forensic experimental casework and actualistic modelling over inductive intuitive approaches which come with significant evidential shortcomings (Bristow et al. 2011).

Finally, from a taphonomic perspective it is worth pointing out to reviewers that we have already addressed the issue of lack of taphonomic evidence for carnivore involvement in the formation of the Dinaledi assemblage (Dirks, et al., 2016). Absence of any carnivore-induced bone surface modifications, patterns of skeletal part representation, and a total absence of any carnivore remains found within the Dinaledi chamber (following Kuhn and colleagues, 2010) lead us to reject carnivores as possible vectors of body accumulation within the Dinaledi Chamber and Hill Antechamber.

Reviewers suggest that without a date derived from geochronological methods, the engravings cannot be associated with H. naledi, and that it is possible (or probable) that the engravings were done in the recent past by H. sapiens. This suggestion neglects the context of the site. We have previously documented the structure and extremely limited accessibility of the Dinaledi subsystem. This subsystem was not recorded on maps of the documented Rising Star Cave system prior to our work and its discovery by our teams. Furthermore, there is no evidence of prehistoric human activity in the areas of the cave related to possible subterranean entrances There is no evidence that humans in the past typically ventured into such extreme spaces like those of Rising Star. It is clear from the presence of the remains of many individuals that H. naledi ventured into these spaces again and again. It is likely that H. naledi moved through these spaces more easily than humans do based on their physique. We show that the engravings overlay each other suggesting multiple engraving events.  These engravings took time and effort and the only evidence for use of the Dinaledi subsystem by any hominin is by H. naledi. The context leads to the null hypothesis that H. naledi made the marks. In our revision, we will elaborate on this argument to clarify the evidence for our stance on this hypothesis. Several reviewers took issue with the title of the engraving paper as we did not insert a qualifier in front of the suggested date range for the engravings. We deliberately left out qualifying language so that the title took the form of a testable hypothesis rather than a weak assertation. Should future work find the engravings were not produced within this time range, then we will restate this hypothesis.

Finally, with regards to the engravings we have chosen to report them because they exist. Not reporting the presence of engraved marks on the walls of a cave above hypothesized burials would be tantamount to leaving relevant evidence out of the description of an archeological context. We recognize and state in our manuscript that these markings require substantial further study, including attempts at geochronological dating. But the current evidence is clearly relevant to the archaeological context of the subsystem. We take a similar stance with reporting the presence of the tool shaped artefact near the hand of the H. naledi skeleton in the Hill Antechamber. It is evident that this object requires further study, as we stated in our manuscript, but again omitting it from our study would be leaving out relevant evidence.

Some have suggested that the null hypothesis should be that all of these observed circumstances are of natural origin. Our team took this approach in our early investigation of the Dinaledi subsystem (Dirks et al. 2015). We adopted the null hypothesis that the geological processes involved in the accumulation of H. naledi skeletal remains were “natural” (e.g., non-naledigenic involvement), and we were able to reject many alternative explanations for the assemblage, including carnivore accumulation, “death trap” accumulation, and fluvial transport of bodies or bones (Dirks et al. 2015). This led us to the hypothesis that H. naledi were involved in bringing the bodies into the spaces where they were found. But we did not hypothesize their involvement in the formation of the deposit itself beyond bringing the bodies to the location.

This approach seems conservative. It followed the traditional view that small-brained hominins do not engage in cultural practices. But we recognize in hindsight that this null hypothesis approach did harm to our analyses. It impeded us from recognizing within our initial excavations of the puzzle box area and other excavations between 2014 – 2017 that we might be encountering remains that were intrusive in the sedimentary floor of the chamber. If we had approached the accumulation of a large number of hominins from the perspective of the null hypothesis being that the situation was likely cultural, we perhaps would have collected evidence in a slightly different manner. We certainly note that if the Dinaledi system had been full of the remains of modern humans, there would have been little doubt that the null hypothesis would have been that this was a cultural space and not a “natural space”.  We therefore respectfully disagree with the reviewers who continue to support the idea that we should approach hominin excavations with the null hypothesis that they will be natural (specifically non-cultural) in origins. If excavations continue with this mindset we believe that potential cultural evidence is almost certain to be lost.

There has been a gradient across paleoanthropological excavations, archaeological work, and forensic investigation, with increasing precision of context. The reality is that the recording precision and frame of approach is typically different in most paleontological excavations than in those related to contemporary human remains. If anything comes from the present discussion of whether the Dinaledi system is a burial site for H. naledi or not, we hope that by taking seriously the possibility of deep cultural dynamics of hominins, we will encourage other teams to meet the highest standards of excavation in order to preserve potential cultural evidence. Given H. naledi’s cranial capacity we suggest that even very early hominin skeletal assemblages should be re-examined, if there is sufficient evidence or records available.  These would include examples such as the A.L. 333 Au. afarensis site (the so called First Family site in Hadar Ethiopia), the Dikika infant skeleton, WT 15000 (Turkana Boy) and even A.L. 288 (Lucy) as such unusual taphonomic situations where skeletons are preserved cannot be simply explained away as “natural” in origin, based solely on the cranial capacity and assumed lack of cognitive and cultural complexity of the hominins as emphasized by us in Fuentes et al. (2023). We are not the first to observe that some very early hominin situations may represent early mortuary activity (Pettitt 2013), but we would advocate a step further. We suggest it may be damaging to take “natural accumulation” as the standard null hypothesis for hominin paleoanthropology, and that it is more conservative in practice to engage remains with the null hypothesis of possible cultural formation.

We are deeply grateful for the time and effort all of the 8 reviewers (across three reviews) have taken with this work.  We also acknowledge the anonymous reviewers from previous submissions who’s opinions and comments will have made the final iterations of these manuscripts better for their efforts. As this process is rather public and includes commentary outside of the eLife forum, we ask that the efforts of all 37 authors and 8 reviewers involved be respected and that the discourse remain professional in all venues as we study this fascinating and quite complex occurrence. We appreciate also the efforts of members of the public who have engaged with this relatively new process where preprints are posted prior to the reviews allowing comments and interactions from colleagues and the public who are normally not part of the internal peer review process.  We believe these interactions will make for better final papers. We feel we have met the standards of demonstrating burials in H. naledi and that the engraving are most likely associated with H. naledi. However, given the reviews we see many areas where our clarity and context, and analyses, were less strong than they can be. With the clarifications and additions taken on board through these review processes the final papers will be stronger and clearer. We, recognize that this is an ongoing process of scientific investigation and further work will allow continued, and possibly better, evaluation of these hypothesis and others.

Lee R Berger, Agustín Fuentes, John Hawks, Tebogo Makhubela

Works cited:

• Aspöck, E. (2008). What Actually is a ‘Deviant Burial’?: Comparing German-Language and Anglophone Research on ‘Deviant Burials.’ In E. M. Murphy (Ed.). Deviant Burial in the Archaeological Record. Oxford: Oxbow Books.  pp 17–34.

• Bolliger, S.A. & Thali, M.J. (2009). Thanatology. In S.A. Bolliger and M.J. Thali (eds) Virtopsy Approach:  3D Optical and Radiological Scanning and Reconstruction in Forensic Medicine. Boca Raton: CRC Press. pp 187-218.

• Boulestin, B. & Duday, H. (2005). Ethnologie et archéologie de la mort: de l’illusion des références à l’emploi d’un vocabulaire. In: C. Mordant and G. Depierre (eds) Les Pratiques Funéraires à l’Âge du Bronze en France. Actes de la table ronde de Sens-en-Bourgogne. Paris: Éditions du Comité des Travaux Historiques et Scientifiques. pp. 17–30.

• Boulestin, B. & Duday, H. (2006). Ethnology and archaeology of death: from the illusion of references to the use of a terminology. Archaeologia Polona 44: 149–169.

• Bristow, J., Simms, Z. & Randolph-Quinney, P.S. Taphonomy. In S. Black and E. Ferguson (eds.) Forensic Anthropology 2000-2010. Boca Raton, FL: CRC Press. pp 279-318.

• Channing, J. & Randolph-Quinney, P.S. (2006). Death, decay and reconstruction: the archaeology of Ballykilmore Cemetery, County Westmeath. In J. O’Sullivan and M. Stanley (eds.) Settlement, Industry and Ritual: Archaeology. National Roads Authority Monograph Series No. 3. Dublin: NRA/Four Courts Press. pp 113-126.

• Cherryson, A. K. (2008). Normal, Deviant and Atypical: Burial Variation in Late Saxon Wessex, c. AD 700–1100. In E. M. Murphy (Ed.). Deviant Burial in the Archaeological Record. Oxford: Oxbow Books. pp 115–130.

• Connolly, M., F. Coyne & L. G. Lynch (2005). Underworld : Death and Burial in Cloghermore Cave, Co. Kerry. Bray, Co. Wicklow: Wordwell.

• Darwent, C. M. & R. L. Lyman (2002). Detecting  the postburial fragmentation of carpals, tarsals and phalanges. In M. H. Sorg and W. D. Haglund (eds). Advances in Forensic Taphonomy: Method, Theory and Archeological Perspectives. Boca Raton, FL, CRC Press. pp 355-378.

• d’Errico, F., & Backwell, L. (2016). Earliest evidence of personal ornaments associated with burial: The Conus shells from Border Cave. Journal of Human Evolution, 93, 91–108.

• De Villiers. H. (1973). Human skeletal remains from Border Cave, Ingwavuma District, KwaZulu, South Africa. Annals of the Transvaal Museum, 28(13), 229–246.

• Dell’Unto, N. and Landeschi, G. (2022). Archaeological 3D GIS. London: Routledge.

• Dibble, H. L., Aldeias, V., Goldberg, P., McPherron, S. P., Sandgathe, D., & Steele, T. E. (2015). A critical look at evidence from La Chapelle-aux-Saints supporting an intentional Neandertal burial. Journal of Archaeological Science, 53, 649–657.

• Dirkmaat, D. C., & Cabo, L. L. (2016). Forensic archaeology and forensic taphonomy: basic considerations on how to properly process and interpret the outdoor forensic scene_. Academic Forensic Pathology_ 6, 439–454.

• Dirks, P. H., Berger, L. R., Roberts, E. M., Kramers, J. D., Hawks, J., Randolph-Quinney, P. S., Elliott, M., Musiba, C. M., Churchill, S. E., de Ruiter, D. J., Schmid, P., Backwell, L. R., Belyanin, G. A., Boshoff, P., Hunter, K. L., Feuerriegel, E. M., Gurtov, A., Harrison, J. du G., Hunter, R., … Tucker, S. (2015). Geological and taphonomic context for the new hominin species Homo naledi from the Dinaledi Chamber, South Africa. ELife, 4, e09561.

• Dirks, P.H.G.M., Berger, L.R., Hawks, J., Randolph-Quinney, P.S., Backwell, L.R., and Roberts, E.M. (2016). Comment on “Deliberate body disposal by hominins in the Dinaledi Chamber, Cradle of Humankind, South Africa?” [J. Hum. Evol. 96 (2016) 145-148]. Journal of Human Evolution 96:  149-153.

• Dirks, P. H., Roberts, E. M., Hilbert-Wolf, H., Kramers, J. D., Hawks, J., Dosseto, A., Duval, M., Elliott, M., Evans, M., Grün, R., Hellstrom, J., Herries, A. I., Joannes-Boyau, R., Makhubela, T. V., Placzek, C. J., Robbins, J., Spandler, C., Wiersma, J., Woodhead, J., & Berger, L. R. (2017). The age of Homo naledi and associated sediments in the Rising Star Cave, South Africa. ELife, 6, e24231.

• Donnelly, S., C. Donnelly & E. Murphy (1999). The forgotten dead: The cíllíní and disused burial grounds of Ballintoy, County Antrim. Ulster Journal of Archaeology 58, 109-113.

• Duday, H. (2005). L’archéothanatologie ou l’archéologie de la mort. In: O. Dutour, J.-J. Hublin and B. Vandermeersch (eds) Objets et Méthodes en Paléoanthropologie. Paris: Comité des Travaux Historiques et Scientifiques. pp. 153–215.

• Duday, H. (2009). Archaeology of the Dead: Lectures in Archaeothanatology. Oxford: Oxbow Books.

• Finley, N. (2000). Outside of life: Traditions of infant burial in Ireland from cillin to cist.  World Archaeology 31, 407-422.

• Gargett, R. H. (1999). Middle Palaeolithic burial is not a dead issue: The view from Qafzeh, Saint-Césaire, Kebara, Amud, and Dederiyeh. Journal of Human Evolution, 37(1), 27–90.

• Goldberg, P., Aldeias, V., Dibble, H., McPherron, S., Sandgathe, D., & Turq, A. (2017). Testing the Roc de Marsal Neandertal “Burial” with Geoarchaeology. Archaeological and Anthropological Sciences, 9(6), 1005–1015.

• Gómez-Olivencia, A., & García-Martínez, D. (2019). New postcranial remains from the Roc de Marsal Neandertal child. PALEO. Revue d’archéologie Préhistorique, 30–1, 30–1.

• Green, E.C. (2022). An archaeothanatological approach to the identification of late Anglo-Saxon burials in wooden containers. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 436-455.

• Henderson, J. (1987). Factors determining the state of preservation of human remains. In A. Boddington, A. Garland and R. Janaway (eds). Death, Decay and Reconstruction: Approaches to Archaeology and Forensic Science. Manchester: Manchester University Press. pp 43-54.

• Hunter, J. R. (2014). Human remains recovery: archaeological and forensic perspectives. In C. Smith (ed). Encyclopedia of Global Archaeology. New York: Springer New York. pp 3549-3556.

• Hochrein, M. (2002). An Autopsy of the Grave: Recognizing, Collecting and Preserving Forensic Geotaphonomic Evidence. In M. H. Sorg and W. D. Haglund (eds). Advances in Forensic Taphonomy: Method, Theory and Archeological Perspectives. Boca Raton, FL, CRC Press: 45-70.

• Knüsel, C.K. & Robb, J. (2016). Funerary taphonomy: An overview of goals and methods. Journal of Archaeological Science: Reports 10, 655-673.

• Kuhn, B.F., Berger, L.R. & Skinner, J.D. (2010). Examining criteria for identifying and differentiating fossil faunal assemblages accumulated by hyenas and hominins using extant hyenid accumulations. International Journal of Osteoarchaeology 20, 15-35.

• Lyman, R. (1994). Vertebrate Taphonomy. Cambridge, Cambridge University Press.

• Martinón-Torres, M., d’Errico, F., Santos, E., Álvaro Gallo, A., Amano, N., Archer, W., Armitage, S. J., Arsuaga, J. L., Bermúdez de Castro, J. M., Blinkhorn, J., Crowther, A., Douka, K., Dubernet, S., Faulkner, P., Fernández-Colón, P., Kourampas, N., González García, J., Larreina, D., Le Bourdonnec, F.-X., … Petraglia, M. D. (2021). Earliest known human burial in Africa. Nature, 593(7857), 7857.

• Mickleburgh, H.L & Wescott, D.J. (2018). Controlled experimental observations on joint disarticulation and bone displacement of a human body in an open pit: implications for funerary archaeology. Journal of Archaeological Science: Reports 20: 158-167.

• Mickleburgh, H.L., Wescott, D.J., Gluschitz, S. & Klinkenberg, V.M. (2022). Exploring the use of actualistic forensic taphonomy in the study of (forensic) archaeological human burials: An actualistic experimental research programme at the Forensic Anthropology Center at Texas State University (FACTS), San Marcos, Texas. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 542-562.

• Owsley, D. & B. Compton (1997). Preservation in late 19th Century iron coffin burials. In W. Haglund and M. Sorg (eds). Forensic Taphonomy: The Postmortem Fate of Human Remains. Boca Raton, FL, CRC Press: 511-526.

• Parker Pearson, M. (1999). The Archaeology of Death and Burial. College Station: Texas A&M University Press.

• Pettitt, P. (2013). The Palaeolithic Origins of Human Burial. Routledge.

• Pomeroy, E., Bennett, P., Hunt, C. O., Reynolds, T., Farr, L., Frouin, M., Holman, J., Lane, R., French, C., & Barker, G. (2020). New Neanderthal remains associated with the ‘flower burial’ at Shanidar Cave. Antiquity, 94(373), 11–26.

• Randolph-Quinney, P.S. (2013). From the cradle to the grave: the bioarchaeology of Clonfad 3 and Ballykilmore 6. In N. Brady, P. Stevens and J. Channing (eds.). Settlement and Community in the Fir Tulach Kingdom. Dublin: National Roads Authority Press. pp A2.1-48.

• Randolph-Quinney, P.S., Haines, S. and Kruger, A. (2018). The use of three-dimensional scanning and surface capture methods in recording forensic taphonomic traces: issues of technology, visualisation, and validation. In: W.J. M. Groen and P. M. Barone (eds). Multidisciplinary Approaches to Forensic Archaeology. Berlin: Springer International Publishing, pp. 115-130.

• Rendu, W., Beauval, C., Crevecoeur, I., Bayle, P., Balzeau, A., Bismuth, T., Bourguignon, L., Delfour, G., Faivre, J.-P., Lacrampe-Cuyaubère, F., Tavormina, C., Todisco, D., Turq, A., & Maureille, B. (2014). Evidence supporting an intentional Neandertal burial at La Chapelle-aux-Saints. Proceedings of the National Academy of Sciences, 111(1), 81–86.

• Sandgathe, D. M., Dibble, H. L., Goldberg, P., & McPherron, S. P. (2011). The Roc de Marsal Neandertal child: A reassessment of its status as a deliberate burial. Journal of Human Evolution, 61(3), 243–253.

• Silver, M. (2016). Conservation Techniques in Cultural Heritage. In E. Stylianidis and F. Remondino (eds) 3D Recording, Documentation and Management of Cultural Heritage. Dunbeath: Whittles Publishing. pp 15-106.

• Schotsmans, E.M.J., Georges-Zimmermann, P., Ueland, M. and Dent, B.B. (2022). From flesh to bone: Building bridges between taphonomy, archaeothanatology and forensic science for a better understanding of mortuary practices. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 501-541.

Author Response:

We would like to thank the eLife reviewers for the considerable time and effort they have invested to review these manuscripts. We have also benefited from a previous round of review of the manuscript describing the proposed burial features, which underwent two rounds of revisions in a high-impact journal over a period of approximately 8 months during 2022 and early 2023. Both sets of reviews have reflected mixed responses to the evidence we have presented, with one reviewer recommending acceptance with minor editorial revisions, two recommending acceptance with minor revisions and the fourth recommending rejection based upon similar arguments to those reflected by some of the reviewers in this current round of reviews in eLife. Ultimately the managing editor of this first journal took the decision that the review process could not be completed in a timely manner and rejected the manuscript although the submission here reflected our consideration of these reviewers suggestions.

We have chosen in this initial response to the eLife reviews to include some references to the previous anonymous reviews in order to illustrate differences of opinion and differences in revision suggestions within the review process. Our goal is to offer maximal insight into our decision-making process and to acknowledge the considerable time and effort put into the assessment of these manuscripts by reviewers (for eLife and in the case of the earlier review process). We hope that this approach will assist the readers, and reviewers, of our manuscripts in understanding why we are proceeding with certain decisions during the revision process.

This is a new process for us and the reviewers, and one way in which it significantly differs from more traditional review is that both the reviews and our reply will be public well in advance of our revisions to the manuscript. Indeed, considering the scope of the reviews, some of those revisions may take considerable time, although many can be accomplished fairly easily. Thus, we are not in a position to say that we have solved every issue raised by the reviewers. Instead, we will examine what appear to be the key critical issues raised regarding the data and the analyses and how we propose to address these as we revise the papers. We will also address several philosophical and ethical issues raised by the reviews and our proposal for dealing with these. More specific editorial and citational recommendations will be dealt with on a case-by-case basis, and we do not address these point-by-point in this reply. Please note, this response to the reviewers is not the revision of the manuscript and is only the initial opinion of the corresponding authors with some guidance from the larger group of authors of all three papers. Our final submitted revision will reflect the input of all authors included on those submissions.

We took the decision to submit three separate papers consciously. The two different categories of evidence, burials and engravings, involve different kinds of analysis and different (although overlapping) teams of researchers, and we recognized that each deserved their own presentation and assessment. Meanwhile, together they inform the context of H. naledi in a way that requires some synthetic discussion, in which both kinds of evidence are relevant, leading to a third paper. But the mutual relevance of these different kinds of evidence and their review by a common set of reviewers naturally raises cross-cutting issues, and the reviewers have cross-referenced the three articles. This has sometimes led to suggestions about one manuscript based on the contents of another. Considering the situation, we accepted the recommendation that it would be clearer to consider all three articles in a single reply. Thus, while each of the three papers will proceed separately during the revision process, it will be necessary to highlight across all three papers occasionally in our responses.

Scientific Issues:

In reading the reviews, we feel there are 9 critical points/assertions raised by one or more of the reviewers that present a problem for, or challenge to, our hypothesis that the observed evidence (bone accumulations and engravings) described in the Dinaledi subsystem are of intentional naledigenic origin. These are:

1. The evidence presented does not demonstrate a clear interruption of the floor sediments, thus failing to demonstrate excavated holes.

2. The sediments infilling the holes where the skeletal remains are found have not been demonstrated to originate from the disruption of the floor sediments and thus could be part of a natural geological process (e.g. water movement, slumping) or carnivore accumulations.

3. Previous geological interpretations by our research group have given alternative geological explanations for formation of the bony accumulations that contradict the present evidence presented here and result in alternative origins hypotheses.

4. Burial cannot be effectively assessed without complete excavation of the features and site.

5. The skeletal remains as presented do not conform clearly to typical body arrangement/positions associated with human (Homo sapiens) burials.

6. There is no evidence of grave goods or lithic scatters that are typically associated with human burials.

7. Humans may have been involved with the creation of either the Homo naledi bone accumulations, the engravings, or both.

8. Without a date of the engravings, the null hypothesis should be the engravings were created by Homo sapiens.

9. The null hypothesis for explanation of the skeletal remains in this situation should be “natural accumulation”.

Our analysis of the Dinaledi Feature 1 leads us to accept that the laminated orange-red mudstone (LORM) sedimentary layer is interrupted, indicating a non-natural intervention, and that the hole created by the interruption was then filled by both a fleshed body (and perhaps parts of other bodies) which were then covered by sediment that originated from the hole that was dug. We recognize that the four eLife reviewers are not convinced that our presentation is sufficient to establish this. Interestingly, this was not the universal opinion of earlier reviewers of the initial manuscript several of whom felt we had adequately supported this hypothesis. The lack of clarity in this current version of the burial manuscript is our responsibility. In the upcoming revision of this paper to be submitted, we will take the reviewers’ critiques to heart and add additional figures that illustrate better the disruption of the LORM and clarify the sedimentological data showing the material covering the skeletal remains in the hole are the disrupted sediments excavated from the same hole. We are proposing to isolate this most critical evidence for burial into a separate section in the revised submission based on the reviewers’ comments. The fact that the LORM layer is disrupted, a fleshed body was placed in the hole created by this disruption, and the body (and perhaps parts of other bodies) was/were then covered by the same sediments from the hole is the central feature of our hypothesis that the bone accumulations observed reflect a burial and not a natural process.

The possibility of fluvial transport or involvement in the subsystem is a topic that we have addressed extensively in past work, and it is clear from these reviews that we must enhance our current manuscript to discuss this issue at greater length. Our previous work (Dirks et al. 2015; Dirks et al. 2017) emphasized that fluvial transport of whole bodies into the subsystem was precluded by several lines of sedimentological evidence. We excavated a rich accumulation of skeletal remains, including articulated limbs and other elements in subvertical orientations inconsistent with slow sedimentary infill, which were difficult to explain without positing either a large and dense pile of bodies and/or sediment movement. We encountered fractured chunks of laminated orange-red mudstone (LORM) in random orientations within our excavation area, within and among skeletal remains, which directly refuted that the remains were inundated with water at the time of burial, and this limited the possibility of fluvial transport. Water flow sufficient to displace bodies or complete skeletal evidence would also transport large and course sediment, which is absent from the subsystem, and would sort the commingled skeletal material that we found by size, which we do not observe. But our excavation only covered less than a square meter at very limited depth, and this was the limit to our knowledge of subsurface sediment. We thus were left with uncertainty that led us to suggest the possibility of sediment slumping or movement into subsurface drains, although these were not observed near our excavation. Our current work expands our knowledge of the subsurface and presents an alternative explanation for the disposition of skeletal remains from our earlier excavation. But we acknowledge that this new explanation is vulnerable to our own previous published proposals, and we must do a better job of explaining how the new information addresses our previous suggestions. By not clearly creating a section where we explained how these previous hypotheses were now nullified by new evidence, we clearly confused the reviewers with our own previous work. We will revise the manuscript by enhancing the review of the significant geological evidence demonstrating that there is no significant fluvial action in the system and making it clear how the burial hypothesis provides a clearer explanation for the situation of skeletal remains from our previous excavation work.

One of the central issues raised by reviewers has been a perceived need to excavate these features completely, totally exhuming all skeletal remains from them. Reviewers have written that it is necessary to identify every skeletal element that is present and account for any missing elements. On this point, we have both ethical and scientific differences from these reviewers. We express our ethical concerns first. Many of the best-preserved possible burials ever discovered by archaeologists were subjected to total excavation and exhumation. Cases like La Chapelle-aux-Saints, La Ferrassie, and Skhūl were fully excavated at a time when data recording and excavation methods did not include the range of spatial and geomorphological approaches that later became routine. The judgment of early investigators that these situations were intentional burials was challenged by later workers, and the kind of information that might enable better tests had been irrevocably lost (Gargett 1999; Dibble et al. 2015; Rendu et al. 2014).

Later, improved excavation standards have not sufficed to remove uncertainty or debate about possible burials. For example, it was long presumed that well-preserved remains of young children were by themselves diagnostic of intentional burial, such as those from Dederiyeh, Border Cave, or Roc de Marsal. Such cases were also fully excavated, with adequate documentation of the positioning of skeletal remains and their surrounding stratigraphic situation, but such cases were later challenged on several bases and the complete exhumation of material has confused or precluded testing of new hypotheses (e.g. Gargett 1999). The case of Roc de Marsal is one in which data from the initial excavation combined with data from the initial excavation combined with re-excavation and geoarchaeological analysis led to a naturalistic interpretation of the skeletal material (Sandgathe et al. 2011; Goldberg et al. 2017). But even in this case, the researchers erred in their interpretation of the skeleton’s situation due to a lack of identification of parts of the infant’s skeleton (Gómez-Olivencia and García-Martinez 2019). That is to say, it is not only the burial hypothesis but other hypotheses that suffer from complete excavation. Researchers concerned with preserving all possible information have sometimes taken extraordinary measures to remove and study possible burials at high-resolution in the laboratory. Such was the case of the Shanidar IV burial removed from the site and transported in plaster jacket by Solecki, which led to the disruption and loss of internal stratigraphic information (Pomeroy et al. 2020). Arguably, the current state of the art is full excavation with partial preparation, such as that undertaken at Panga ya Saidi (Martinón-Torres et al. 2021). But again, any future attempt to reinterpret or test the hypothesis of burial must rely on the adequacy of documentation as the original context has been removed.

In our decision to leave material in place as much as possible, we are expanding upon standard practice to leave witness sections and unexcavated areas for future research. The situation is novel, representing possible burials by a nonhuman species, and that makes it doubly important in our opinion to be conservative in not fully exhuming the skeletal material from its context. We anticipate that many other researchers, including future investigators, will suggest additional methods to further test the hypothesis of burial, something that would be impossible if we had excavated the features in their entirety prior to publishing a description of our work. We believe strongly that our ethical responsibility is to publish the work and the most likely interpretation while leaving as much evidence in place as possible to enable further testing and replication. We welcome the suggestions of additional methods/analyses to test the H. naledi burial hypothesis.

This being said, we also observe that total exhumation would not resolve the concerns raised by the reviewers. The recommendation of total exhumation is in pursuit of a full account of all skeletal material present and its preservation and spatial situation, in order to demonstrate that they conform to body positions comparable to human burials. As has been highlighted in forensic casework, the excavation of an inhumation feature does not necessarily provide an accurate spatial or anatomical manifest of the stratigraphical relationships between the body, encapsulating matrix, and any cut present due to preservational, taphonomic and operational factors (Dirkmaat and Cabo, 2016; Hunter, 2014). In particular, in cases where skeletal elements are highly fragmented, friable, or degraded (such as through bioerosion) then complete excavation—even under controlled laboratory conditions—may destroy bone and severely limit skeletal identification (Henderson, 1997; Hochrein, 2002; Owsley and Compton, 1997), particularly in elements where the ratio of trabecular to cortical bone is high (Darwent and Lyman, 2002; Lyman, 1994). As such, non-invasive methods of 3D and 4D modelling (preservation in situ) are often considered preferable to complete necropsy or excavation (preservation by record) where appropriate (Bolliger and Thali, 2009; Dell’Unto and Landeschi, 2022; Randolph-Quinney et al., 2018; Silver, 2016). 

The test of burial is not primarily positional, but taphonomic and geological. The position and number of bones can elaborate on process-driven questions of decay and destruction in the burial environment, or post-mortem modification, but are not singularly indicative of whether the remains were intentionally buried – the post-mortem narrative of all the processes affecting the cadaveric island is required (Knüsel and Robb, 2016). In previous cases, researchers have disputed or accepted the hypothesis of intentional hominin burial based upon assumptions about how modern humans or Neandertals would have positioned bodies, with the idea that some positions reflect ritual intent while others do not. But applying such assumptions is unjustifiable, particularly for a species like H. naledi, whose culture may have differed fundamentally from our own. Our work acknowledges that the present evidence does not enable a full reconstruction of the burial positions, but it does show that fleshed remains were encased in sediment prior to decomposition of soft tissue, and that subsequent spatial changes can be most parsimoniously explained by natural decomposition within sedimentary matrix contained within a burial feature (after Green, 2022; Mickleburgh and Wescott, 2018; Mickleburgh et al., 2022). If the argument is that extraordinary claims require extraordinary evidence, we feel that the evidence documents excavation and interment (and will do so more clearly in the revision) and the fact of the remains do not match a “typical” human burial in body positioning is not in itself evidence that these are not H. naledi burials.

We feel that the reviewers (in keeping with many palaeoanthropologists) have a clear idea of what they “think” a burial should look like in an idealised sense, but this platonic ideal of burial form is not matched by the extensive literature in archaeothanatology, funerary archaeology and forensic science which indicates enormous variability in the activity, morphology and post-mortem system experienced by the human body in cases of interment and body disposal (e.g. Aspöck, 2008; Boulestin and Duday, 2005 and 2006; Connelly et al., 2005; Channing and Randolph-Quinney, 2006; Cherryson, 2008; Donnelly et al., 1995; Finley, 2000; Hunter, 2014; Parker Pearson, 1999; Randolph-Quinney, 2013). Decades of experience in the identification, recovery and interpretation of clandestine, deviant, and non-formal burials indicates the platonic ideal is rare, and in many contexts, the exception (Cherryson, 2008; Parker Pearson, 1999). This variability is particularly relevant to morphological traits in burial context, such as the informal nature of the grave cut in plan and section, shallow burial depth, and initial disposition of body (placement) during the early post-mortem period. These might run counter to the expectations of reviewers or others referencing the fossil hominin record, but are well accepted within the communities of researchers investigating Holocene archaeological sites and forensic contexts.

It is encouraging to see reviewers beginning to incorporate the extensive (often experimentally derived) literature from archaeothanatology and forensic taphonomy in their deliberations, and we will be taking these comments on board going forward. In particular, we acknowledge reviewers’ comments and the need to construct a more detailed post-mortem narrative, accounting for joint disarticulation (labile versus persistent joints etc), displacement, and final disposition of elements within the burial space. As such we will incorporate the hierarchy of decomposition (rank order disarticulation), associations between regions of anatomical association, areas of disassociation, and the voids produced during decomposition (after Mickleburgh and Wescott, 2018; Mickleburgh et al., 2022) into our narrative. In doing so we acknowledge the tensions between the inductive archaeolothanatological narrative-driven approach (e.g. Duday, 2005 & 2009) versus robust decomposition data derived from human forensic taphonomic experimentation recently articulated by Schotsmans and colleagues (2022) - noting that we will highlight comparative data based on forensic experimental casework and actualistic modelling over inductive intuitive approaches which come with significant evidential shortcomings (Bristow et al. 2011).

Finally, from a taphonomic perspective it is worth pointing out to reviewers that we have already addressed the issue of lack of taphonomic evidence for carnivore involvement in the formation of the Dinaledi assemblage (Dirks, et al., 2016). Absence of any carnivore-induced bone surface modifications, patterns of skeletal part representation, and a total absence of any carnivore remains found within the Dinaledi chamber (following Kuhn and colleagues, 2010) lead us to reject carnivores as possible vectors of body accumulation within the Dinaledi Chamber and Hill Antechamber.

Reviewers suggest that without a date derived from geochronological methods, the engravings cannot be associated with H. naledi, and that it is possible (or probable) that the engravings were done in the recent past by H. sapiens. This suggestion neglects the context of the site. We have previously documented the structure and extremely limited accessibility of the Dinaledi subsystem. This subsystem was not recorded on maps of the documented Rising Star Cave system prior to our work and its discovery by our teams. Furthermore, there is no evidence of prehistoric human activity in the areas of the cave related to possible subterranean entrances There is no evidence that humans in the past typically ventured into such extreme spaces like those of Rising Star. It is clear from the presence of the remains of many individuals that H. naledi ventured into these spaces again and again. It is likely that H. naledi moved through these spaces more easily than humans do based on their physique. We show that the engravings overlay each other suggesting multiple engraving events.  These engravings took time and effort and the only evidence for use of the Dinaledi subsystem by any hominin is by H. naledi. The context leads to the null hypothesis that H. naledi made the marks. In our revision, we will elaborate on this argument to clarify the evidence for our stance on this hypothesis. Several reviewers took issue with the title of the engraving paper as we did not insert a qualifier in front of the suggested date range for the engravings. We deliberately left out qualifying language so that the title took the form of a testable hypothesis rather than a weak assertation. Should future work find the engravings were not produced within this time range, then we will restate this hypothesis.

Finally, with regards to the engravings we have chosen to report them because they exist. Not reporting the presence of engraved marks on the walls of a cave above hypothesized burials would be tantamount to leaving relevant evidence out of the description of an archeological context. We recognize and state in our manuscript that these markings require substantial further study, including attempts at geochronological dating. But the current evidence is clearly relevant to the archaeological context of the subsystem. We take a similar stance with reporting the presence of the tool shaped artefact near the hand of the H. naledi skeleton in the Hill Antechamber. It is evident that this object requires further study, as we stated in our manuscript, but again omitting it from our study would be leaving out relevant evidence.

Some have suggested that the null hypothesis should be that all of these observed circumstances are of natural origin. Our team took this approach in our early investigation of the Dinaledi subsystem (Dirks et al. 2015). We adopted the null hypothesis that the geological processes involved in the accumulation of H. naledi skeletal remains were “natural” (e.g., non-naledigenic involvement), and we were able to reject many alternative explanations for the assemblage, including carnivore accumulation, “death trap” accumulation, and fluvial transport of bodies or bones (Dirks et al. 2015). This led us to the hypothesis that H. naledi were involved in bringing the bodies into the spaces where they were found. But we did not hypothesize their involvement in the formation of the deposit itself beyond bringing the bodies to the location.

This approach seems conservative. It followed the traditional view that small-brained hominins do not engage in cultural practices. But we recognize in hindsight that this null hypothesis approach did harm to our analyses. It impeded us from recognizing within our initial excavations of the puzzle box area and other excavations between 2014 – 2017 that we might be encountering remains that were intrusive in the sedimentary floor of the chamber. If we had approached the accumulation of a large number of hominins from the perspective of the null hypothesis being that the situation was likely cultural, we perhaps would have collected evidence in a slightly different manner. We certainly note that if the Dinaledi system had been full of the remains of modern humans, there would have been little doubt that the null hypothesis would have been that this was a cultural space and not a “natural space”.  We therefore respectfully disagree with the reviewers who continue to support the idea that we should approach hominin excavations with the null hypothesis that they will be natural (specifically non-cultural) in origins. If excavations continue with this mindset we believe that potential cultural evidence is almost certain to be lost.

There has been a gradient across paleoanthropological excavations, archaeological work, and forensic investigation, with increasing precision of context. The reality is that the recording precision and frame of approach is typically different in most paleontological excavations than in those related to contemporary human remains. If anything comes from the present discussion of whether the Dinaledi system is a burial site for H. naledi or not, we hope that by taking seriously the possibility of deep cultural dynamics of hominins, we will encourage other teams to meet the highest standards of excavation in order to preserve potential cultural evidence. Given H. naledi’s cranial capacity we suggest that even very early hominin skeletal assemblages should be re-examined, if there is sufficient evidence or records available.  These would include examples such as the A.L. 333 Au. afarensis site (the so called First Family site in Hadar Ethiopia), the Dikika infant skeleton, WT 15000 (Turkana Boy) and even A.L. 288 (Lucy) as such unusual taphonomic situations where skeletons are preserved cannot be simply explained away as “natural” in origin, based solely on the cranial capacity and assumed lack of cognitive and cultural complexity of the hominins as emphasized by us in Fuentes et al. (2023). We are not the first to observe that some very early hominin situations may represent early mortuary activity (Pettitt 2013), but we would advocate a step further. We suggest it may be damaging to take “natural accumulation” as the standard null hypothesis for hominin paleoanthropology, and that it is more conservative in practice to engage remains with the null hypothesis of possible cultural formation.

We are deeply grateful for the time and effort all of the 8 reviewers (across three reviews) have taken with this work.  We also acknowledge the anonymous reviewers from previous submissions who’s opinions and comments will have made the final iterations of these manuscripts better for their efforts. As this process is rather public and includes commentary outside of the eLife forum, we ask that the efforts of all 37 authors and 8 reviewers involved be respected and that the discourse remain professional in all venues as we study this fascinating and quite complex occurrence. We appreciate also the efforts of members of the public who have engaged with this relatively new process where preprints are posted prior to the reviews allowing comments and interactions from colleagues and the public who are normally not part of the internal peer review process.  We believe these interactions will make for better final papers. We feel we have met the standards of demonstrating burials in H. naledi and that the engraving are most likely associated with H. naledi. However, given the reviews we see many areas where our clarity and context, and analyses, were less strong than they can be. With the clarifications and additions taken on board through these review processes the final papers will be stronger and clearer. We, recognize that this is an ongoing process of scientific investigation and further work will allow continued, and possibly better, evaluation of these hypothesis and others.

Lee R Berger, Agustín Fuentes, John Hawks, Tebogo Makhubela

Works cited:

• Aspöck, E. (2008). What Actually is a ‘Deviant Burial’?: Comparing German-Language and Anglophone Research on ‘Deviant Burials.’ In E. M. Murphy (Ed.). Deviant Burial in the Archaeological Record. Oxford: Oxbow Books.  pp 17–34.

• Bolliger, S.A. & Thali, M.J. (2009). Thanatology. In S.A. Bolliger and M.J. Thali (eds) Virtopsy Approach:  3D Optical and Radiological Scanning and Reconstruction in Forensic Medicine. Boca Raton: CRC Press. pp 187-218.

• Boulestin, B. & Duday, H. (2005). Ethnologie et archéologie de la mort: de l’illusion des références à l’emploi d’un vocabulaire. In: C. Mordant and G. Depierre (eds) Les Pratiques Funéraires à l’Âge du Bronze en France. Actes de la table ronde de Sens-en-Bourgogne. Paris: Éditions du Comité des Travaux Historiques et Scientifiques. pp. 17–30.

• Boulestin, B. & Duday, H. (2006). Ethnology and archaeology of death: from the illusion of references to the use of a terminology. Archaeologia Polona 44: 149–169.

• Bristow, J., Simms, Z. & Randolph-Quinney, P.S. Taphonomy. In S. Black and E. Ferguson (eds.) Forensic Anthropology 2000-2010. Boca Raton, FL: CRC Press. pp 279-318.

• Channing, J. & Randolph-Quinney, P.S. (2006). Death, decay and reconstruction: the archaeology of Ballykilmore Cemetery, County Westmeath. In J. O’Sullivan and M. Stanley (eds.) Settlement, Industry and Ritual: Archaeology. National Roads Authority Monograph Series No. 3. Dublin: NRA/Four Courts Press. pp 113-126.

• Cherryson, A. K. (2008). Normal, Deviant and Atypical: Burial Variation in Late Saxon Wessex, c. AD 700–1100. In E. M. Murphy (Ed.). Deviant Burial in the Archaeological Record. Oxford: Oxbow Books. pp 115–130.

• Connolly, M., F. Coyne & L. G. Lynch (2005). Underworld : Death and Burial in Cloghermore Cave, Co. Kerry. Bray, Co. Wicklow: Wordwell.

• Darwent, C. M. & R. L. Lyman (2002). Detecting  the postburial fragmentation of carpals, tarsals and phalanges. In M. H. Sorg and W. D. Haglund (eds). Advances in Forensic Taphonomy: Method, Theory and Archeological Perspectives. Boca Raton, FL, CRC Press. pp 355-378.

• d’Errico, F., & Backwell, L. (2016). Earliest evidence of personal ornaments associated with burial: The Conus shells from Border Cave. Journal of Human Evolution, 93, 91–108.

• De Villiers. H. (1973). Human skeletal remains from Border Cave, Ingwavuma District, KwaZulu, South Africa. Annals of the Transvaal Museum, 28(13), 229–246.

• Dell’Unto, N. and Landeschi, G. (2022). Archaeological 3D GIS. London: Routledge.

• Dibble, H. L., Aldeias, V., Goldberg, P., McPherron, S. P., Sandgathe, D., & Steele, T. E. (2015). A critical look at evidence from La Chapelle-aux-Saints supporting an intentional Neandertal burial. Journal of Archaeological Science, 53, 649–657.

• Dirkmaat, D. C., & Cabo, L. L. (2016). Forensic archaeology and forensic taphonomy: basic considerations on how to properly process and interpret the outdoor forensic scene_. Academic Forensic Pathology_ 6, 439–454.

• Dirks, P. H., Berger, L. R., Roberts, E. M., Kramers, J. D., Hawks, J., Randolph-Quinney, P. S., Elliott, M., Musiba, C. M., Churchill, S. E., de Ruiter, D. J., Schmid, P., Backwell, L. R., Belyanin, G. A., Boshoff, P., Hunter, K. L., Feuerriegel, E. M., Gurtov, A., Harrison, J. du G., Hunter, R., … Tucker, S. (2015). Geological and taphonomic context for the new hominin species Homo naledi from the Dinaledi Chamber, South Africa. ELife, 4, e09561.

• Dirks, P.H.G.M., Berger, L.R., Hawks, J., Randolph-Quinney, P.S., Backwell, L.R., and Roberts, E.M. (2016). Comment on “Deliberate body disposal by hominins in the Dinaledi Chamber, Cradle of Humankind, South Africa?” [J. Hum. Evol. 96 (2016) 145-148]. Journal of Human Evolution 96:  149-153.

• Dirks, P. H., Roberts, E. M., Hilbert-Wolf, H., Kramers, J. D., Hawks, J., Dosseto, A., Duval, M., Elliott, M., Evans, M., Grün, R., Hellstrom, J., Herries, A. I., Joannes-Boyau, R., Makhubela, T. V., Placzek, C. J., Robbins, J., Spandler, C., Wiersma, J., Woodhead, J., & Berger, L. R. (2017). The age of Homo naledi and associated sediments in the Rising Star Cave, South Africa. ELife, 6, e24231.

• Donnelly, S., C. Donnelly & E. Murphy (1999). The forgotten dead: The cíllíní and disused burial grounds of Ballintoy, County Antrim. Ulster Journal of Archaeology 58, 109-113.

• Duday, H. (2005). L’archéothanatologie ou l’archéologie de la mort. In: O. Dutour, J.-J. Hublin and B. Vandermeersch (eds) Objets et Méthodes en Paléoanthropologie. Paris: Comité des Travaux Historiques et Scientifiques. pp. 153–215.

• Duday, H. (2009). Archaeology of the Dead: Lectures in Archaeothanatology. Oxford: Oxbow Books.

• Finley, N. (2000). Outside of life: Traditions of infant burial in Ireland from cillin to cist.  World Archaeology 31, 407-422.

• Gargett, R. H. (1999). Middle Palaeolithic burial is not a dead issue: The view from Qafzeh, Saint-Césaire, Kebara, Amud, and Dederiyeh. Journal of Human Evolution, 37(1), 27–90.

• Goldberg, P., Aldeias, V., Dibble, H., McPherron, S., Sandgathe, D., & Turq, A. (2017). Testing the Roc de Marsal Neandertal “Burial” with Geoarchaeology. Archaeological and Anthropological Sciences, 9(6), 1005–1015.

• Gómez-Olivencia, A., & García-Martínez, D. (2019). New postcranial remains from the Roc de Marsal Neandertal child. PALEO. Revue d’archéologie Préhistorique, 30–1, 30–1.

• Green, E.C. (2022). An archaeothanatological approach to the identification of late Anglo-Saxon burials in wooden containers. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 436-455.

• Henderson, J. (1987). Factors determining the state of preservation of human remains. In A. Boddington, A. Garland and R. Janaway (eds). Death, Decay and Reconstruction: Approaches to Archaeology and Forensic Science. Manchester: Manchester University Press. pp 43-54.

• Hunter, J. R. (2014). Human remains recovery: archaeological and forensic perspectives. In C. Smith (ed). Encyclopedia of Global Archaeology. New York: Springer New York. pp 3549-3556.

• Hochrein, M. (2002). An Autopsy of the Grave: Recognizing, Collecting and Preserving Forensic Geotaphonomic Evidence. In M. H. Sorg and W. D. Haglund (eds). Advances in Forensic Taphonomy: Method, Theory and Archeological Perspectives. Boca Raton, FL, CRC Press: 45-70.

• Knüsel, C.K. & Robb, J. (2016). Funerary taphonomy: An overview of goals and methods. Journal of Archaeological Science: Reports 10, 655-673.

• Kuhn, B.F., Berger, L.R. & Skinner, J.D. (2010). Examining criteria for identifying and differentiating fossil faunal assemblages accumulated by hyenas and hominins using extant hyenid accumulations. International Journal of Osteoarchaeology 20, 15-35.

• Lyman, R. (1994). Vertebrate Taphonomy. Cambridge, Cambridge University Press.

• Martinón-Torres, M., d’Errico, F., Santos, E., Álvaro Gallo, A., Amano, N., Archer, W., Armitage, S. J., Arsuaga, J. L., Bermúdez de Castro, J. M., Blinkhorn, J., Crowther, A., Douka, K., Dubernet, S., Faulkner, P., Fernández-Colón, P., Kourampas, N., González García, J., Larreina, D., Le Bourdonnec, F.-X., … Petraglia, M. D. (2021). Earliest known human burial in Africa. Nature, 593(7857), 7857.

• Mickleburgh, H.L & Wescott, D.J. (2018). Controlled experimental observations on joint disarticulation and bone displacement of a human body in an open pit: implications for funerary archaeology. Journal of Archaeological Science: Reports 20: 158-167.

• Mickleburgh, H.L., Wescott, D.J., Gluschitz, S. & Klinkenberg, V.M. (2022). Exploring the use of actualistic forensic taphonomy in the study of (forensic) archaeological human burials: An actualistic experimental research programme at the Forensic Anthropology Center at Texas State University (FACTS), San Marcos, Texas. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 542-562.

• Owsley, D. & B. Compton (1997). Preservation in late 19th Century iron coffin burials. In W. Haglund and M. Sorg (eds). Forensic Taphonomy: The Postmortem Fate of Human Remains. Boca Raton, FL, CRC Press: 511-526.

• Parker Pearson, M. (1999). The Archaeology of Death and Burial. College Station: Texas A&M University Press.

• Pettitt, P. (2013). The Palaeolithic Origins of Human Burial. Routledge.

• Pomeroy, E., Bennett, P., Hunt, C. O., Reynolds, T., Farr, L., Frouin, M., Holman, J., Lane, R., French, C., & Barker, G. (2020). New Neanderthal remains associated with the ‘flower burial’ at Shanidar Cave. Antiquity, 94(373), 11–26.

• Randolph-Quinney, P.S. (2013). From the cradle to the grave: the bioarchaeology of Clonfad 3 and Ballykilmore 6. In N. Brady, P. Stevens and J. Channing (eds.). Settlement and Community in the Fir Tulach Kingdom. Dublin: National Roads Authority Press. pp A2.1-48.

• Randolph-Quinney, P.S., Haines, S. and Kruger, A. (2018). The use of three-dimensional scanning and surface capture methods in recording forensic taphonomic traces: issues of technology, visualisation, and validation. In: W.J. M. Groen and P. M. Barone (eds). Multidisciplinary Approaches to Forensic Archaeology. Berlin: Springer International Publishing, pp. 115-130.

• Rendu, W., Beauval, C., Crevecoeur, I., Bayle, P., Balzeau, A., Bismuth, T., Bourguignon, L., Delfour, G., Faivre, J.-P., Lacrampe-Cuyaubère, F., Tavormina, C., Todisco, D., Turq, A., & Maureille, B. (2014). Evidence supporting an intentional Neandertal burial at La Chapelle-aux-Saints. Proceedings of the National Academy of Sciences, 111(1), 81–86.

• Sandgathe, D. M., Dibble, H. L., Goldberg, P., & McPherron, S. P. (2011). The Roc de Marsal Neandertal child: A reassessment of its status as a deliberate burial. Journal of Human Evolution, 61(3), 243–253.

• Silver, M. (2016). Conservation Techniques in Cultural Heritage. In E. Stylianidis and F. Remondino (eds) 3D Recording, Documentation and Management of Cultural Heritage. Dunbeath: Whittles Publishing. pp 15-106.

• Schotsmans, E.M.J., Georges-Zimmermann, P., Ueland, M. and Dent, B.B. (2022). From flesh to bone: Building bridges between taphonomy, archaeothanatology and forensic science for a better understanding of mortuary practices. In C.J. Knüsel and E.M.J. Schotsmans (eds.) The Routledge Handbook of Archaeothanatology. London: Routledge. pp 501-541.

Author Response:

We thank eLife and the reviewer for the nice summary of our manuscript. We largely agree with the summary and review, and just add a few small points.

First, the review asks about the reproducibility of our findings, and suggests that they are only from a single experiment. In fact, our manuscript reports data from two independent single-cell experiments: one performed at low multiplicity of infection (MOI), and another at higher MOI. The broad trends, including the lack of strong correlations between viral mRNA transcription and progeny production, are consistent across both experiments.

Second, the reviewer asks about what happens when two different virions bearing the same viral barcode infect two different cells, given that we estimate 4-8% of barcodes to be shared between multiple infecting virions. When two cells are infected by different virions with the same barcode, this breaks the one-to-one link between transcription in that cell and progeny in the supernatant, since it is not possible to determine which cell contributed the progeny with that barcode. This means that between 4-8% of the points on our correlation plots could be affected by this factor, meaning that a few outliers should be expected. Another scenario, where a single cell is infected by two barcodes, is not problematic for our method because we can simply sum the progeny output for both barcodes from that cell.

Finally, the reviewer notes that some cells appear to produce progeny virions despite failing to express one or more viral genes. Such cells can be explained in one of two ways. First, as noted immediately above, we expect a small fraction (4-8%) of the points to be erroneous due to a lack of a guaranteed one-to-one link between cell and progeny for non-unique barcodes. Second, in some cases the missing viral gene could be a technical artifact caused by a stochastic failure to capture modestly expressed transcripts from the gene; this phenomenon, known as gene dropout, occurs at a fairly high rate in single-cell experiments (see Qiu Nature Communications 2020 for a detailed discussion). Genes that are expressed at lower levels, like the Influenza virus polymerase genes, are more likely to be missed during single-cell RNA sequencing. The absent viral genes in each infected cell can be explored in detail using the interactive plots at https://jbloomlab.github.io/barcoded_flu_pdmH1N1/

Author Response:

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

Major Revisions:

1) Although we appreciate this work was carried out independently, it would improve this paper if this structure presented here was compared to the recently published structure of Cx43 (Nat Commun 14, 931 (2023)) with the conclusions including added in the discussion.

We encourage the readers to read both our study on Cx43 and the one mentioned by the reviewer. However, we believe the optimal format for such a comparison is going to be a more comprehensive review article, which is outside the scope of our study.

2) Please elaborate on the lipid-binding pockets observed for lipid 1, lipid 2, and the N-lipid/PGL. For example, what are the residues involved in these lipid-protein interactions? Are these residues conserved in other connexin isoforms? Do these lipid-binding pockets match with previous structures, including the recent Cx43 structure? Please clarify what lipid sites are ambiguous due to insufficient resolution.

Within the scope of our study, we have shown that some of the disease-linked residues are located in close proximity to the lipid sites (Fig. 4b). This suggests a possible role of the lipid sites in diseases associated with Cx43 mutations (and possibly with the mutations in other connexins, as the structures of other connexin channels also feature bound lipids inside the pore region). We feel that a more in-depth comparison will require a careful study, beyond the analysis that we have performed here, and for this reason we would like to reserve such a detailed comparison for our future work (possibly a comprehensive review article on connexin structure and function).

3) The NT domain and TM2 segments are referred to as the gate region. If there is no strong evidence to support this claim then please use "putative" gate region.

We have updated the text accordingly, referring to this region as a putative gate region where appropriate.

4) It is mentioned that there is a reorientation of extracellular loops 1 and 2 after Gap junction formation. Based on their structures, I wonder how this rearrangement alters the channel conduction pathway. For example, Do the electrostatic surface and hydrophobic properties change? Please consider adding further details as this information could be useful to understand why some properties of hemichannels differ from intercellular GJ channels.

We have updated the Fig. 5 with an illustration of the Cx43 HC surface coloured according to electrostatic potential (to match the same representation of the Cx43 GJC). It is obvious that the rearrangement of the extracellular loops 1 and 2 do not dramatically alter the electrostatic properties of the HC relative to the GJC. A more obvious difference is in the local environment of the ECLs: it is radically different in a “free” HC (exposed to the solvent or to the extracellular space of a cell), compared to the ECL environment in a connexon within a GJC (which is sealed by a docked connexon from the opposite membrane).

5) Related to the previous point, the pore profile shown in Figure 5C shows that there is a constriction site in the extracellular part with the same diameter as the observed constriction caused by the NT domain. This constriction point seems to be associated with the high energies calculated for Cl-. Please clarify if this constriction is produced by the formation of the GJC or is also present in HC?

This is the same constriction zone, and the Cl- barriers are further down the channel axis where the electrostatic potential of the protein is negative. We have included a similar calculation for the HC simulation in Fig. 5 (revised Fig. 5f).

6) Related to the MD simulations shown in Figure 5d: if the voltage is applied across the whole GJC, the free energy under voltage should not be symmetric. Please clarify.

The symmetry observed in the free energies is due to the fact that the ions enter and exit from the same hemichannel. Only at very high voltages we observe some rare full GJC permeation events, slightly unbalancing the free energy at 500 mV.

7) The scheme in Figure 6 many needs further editing. The authors propose a putative closed state in which lipids are bound next to the NT, but we suggest it should be made clearer in the figure that this is a putative model, since there is no functional evidence supporting the role of these lipids in the gating/permeation properties of Cx43. Also, please clarify what is meant by a "semi-permeable gate" - a channel that only permeates ions but not molecules?

We have updated the legend of the figure 6, to clearly reflect that this is a putative model. The “semi-permeable” state of the channel is something that was suggested previously by the authors of the Cx31.3 study, and we refer to that structure in the figure.

Minor comments:

1) In the result section there are some statements that currently lack solid experimental support. Please consider editing or moving this text to the discussion section only. A good example of this is the Diseaselinked mutation section, specifically lines 199-206. In another example: in lines, 237-238 authors state that NT can move laterally and vertically, but this idea still requires experimental validation.

We feel that the original formulations of these portions of the text are appropriate. Disrupting them would interrupt the flow of the manuscript, and we prefer to stay with the original text in this case.

2) Line 283. "With these structures in mind, we can now establish the existence of several structurally defined gating substates of the connexin channels". Please, tone down this statement. Replace "establish" with "propose" or another more appropriate word.

We have updated the text as suggested ("propose” instead of “establish”)

3) Line 313-314. " The presence of such molecules could have important implications for HC or GJC assembly, substrate permeation, and molecular gating". Currently, this entire statement does not have any support. Is there any paper that authors can discuss to suggest with some basis that lipids might have a role in assembly, permeation or gating?

We feel that this statement is sufficiently careful, conveying a thought that the presence of such molecules could have important implications for various HC- or GJC-related processes. It is not a particularly strong claim and seems to be appropriate in this context.

4) It seems that the structure shown in panels A and C in Figure 2 are shown in opposite directions, which makes the figure confusing. If needed, please rotate the structure in panel A to show the cytosolic part of the protein as panel C. Also, in the same figure, panels G and F are wrongly labeled. Please correct.

For Fig. 2a, the angle is very different from anything else we show in the figure, so we would rather keep this as it is now. We have corrected the labelling for Fig. 2g-h.

5) Check spelling mistakes in the legend of Extended data Fig.2, Extended data Fig.9, and line 243.

We are grateful to the reviewers for pointing out the typos, which have now been corrected.

6) The colors for G-L isoforms are not specified in Extended Data Fig.10. Please correct this.

We updated the figure, removing the PGL label (the correct label is “lipid-N”).

7) It is not clear what is the difference between PGL and the N-lipid density. Does PGL refers to the lipid-like density observed in nanodiscs, as indicated in Extended Fig. 4 and 10?. Please clarify this issue in the manuscript.

The labeling has been corrected in like with the revised version of the manuscript (this density element is now referred to as the “lipid-N”).

8) Page 7 line 234-235 "The pore opening has a solvent-accessible radius of ~6Å (Figure 5c) very close to the effective hydrated radius of K+ (~6.6 Å) and Cl- (~7.2 Å). This makes it the most narrow pore opening...", it should be diameter, not radius.

We have added a calculation for the HC (new Fig. 5f) and corrected the text as follows (line 234):

“The pore opening observed in our cryo-EM structures has a solvent-accessible radius of ~3 Å (Figure 2b). This makes it the most narrow pore opening observed for a connexin channel to date (a comparison of the pore openings in the cryo-EM structures of connexin channels is shown in Extended Data Fig. 12). However, the average solvent-accessible radius of the pore during molecular dynamics was ~6 Å (Figure 5c); note that the effective hydrated radius of K+ and Cl- is ~3.3 Å and ~3.6 Å, respectively.”

And line 277:

“The average pore radius during the simulations was consistent with that observed in the cryo-EM structure (Fig. 5f).”

Author Response

Reviewer #2 (Public Review):

The manuscript by Ma et al, "Two RNA-binding proteins mediate the sorting of miR223 from mitochondria into exosomes" examines the contribution of two RNA-binding proteins on the exosomal loading of miR223. The authors conclude that YBX1 and YBAP1 work in tandem to traffic and load miR223 into the exosome. The manuscript is interesting and potentially impactful. It proposes the following scenario regarding the exosomal loading of miR223: (1) YBAP1 sequesters miR223 in the mitochondria, (2) YBAP1 then transfers miR223 to YBX1, and (3) YBX1 then delivers miR223 into the early endosome for eventual secretion within an exosome. While the authors propose plausible explanations for this phenomenon, they do not specifically test them and no mechanism by which miR223 is shuttled between YBAP1 and YBX1, and the exosome is shown. Thus, the paper is missing critical mechanistic experiments that could have readily tested the speculative conclusions that it makes.

Comments:

1) The major limitation of this paper is that it fails to explore the mechanism of any of the major changes it describes. For example, the authors propose that miR223 shuttles from mitochondrially localized YBAP1 to P-body-associated YBX1 to the exosome. This needs to be tested directly and could be easily addressed by showing a transfer of miR223 from YBAP1 to YBX1 to the exosome.

Testing this idea using fluorescently labeled miR223 would indeed be an ideal experiment. However, miRNA imaging presents challenges. As reviewer 1 pointed out, and we have now confirmed, the atto-647 dye itself localizes to mitochondria. We will continue our efforts to identify a suitable fluorescent label for miR223in order to be in a position to evaluate the temporal relationship between mitochondrial and endosomal miR223.

2) If YBAP1 retains miR223 in mitochondria, what is the trigger for YBAP1 to release it and pass it off to YBX1? The authors speculate in their discussion that sequestration of mito-miR223 plays a "role in some structural or regulatory process, perhaps essential for mitochondrial homeostasis, controlled by the selective extraction of unwanted miRNA into RNA granules and further by secretion in exosomes...". This is readily testable by altering mitochondria dynamics and/or integrity.

A previous study has reported that YBAP1 can be released from mitochondria to the cytosol during HSV-1 infection (Song et al., 2021). However, due to restrictions, we are unable to conduct experiments using HSV to verify this condition. We attempted to induce mitochondrial stress by using different concentrations of CCCP, but we did not observe the release of YBAP1 from mitochondria after CCCP treatment. We speculate that not all mitochondrial stress conditions can trigger YBAP1 release. Investigating the mechanism of mito-miR223 release from mitochondria is one of our interests that we aim to explore in future studies.

3) Much of the miRNA RT-PCR analysis is presented as a ratio of exosomal/cellular. This particular analysis assumes that cellular miRNA is unaffected by treatments. For example, Figure 1a shows that the presence of exosomal miR223 is significantly reduced when YBX1 is knocked out. This analysis does not consider the possibility that YBX1-KO alters (up or down-regulates) intracellular miR223 levels. Should that be the case, the ratiometric analysis is greatly skewed by intracellular miRNA changes. It would be better to not only show the intracellular levels of the miRs but also normalize the miRNA levels to the total amount of RNA isolated or an irrelevant/unchanged miRNA.

Our previous publications demonstrated that miR223 levels are increased in YBX1-KO cells and decreased in exosomes derived from YBX1 KO cells. However, no significant changes were observed in miR190 levels (Liu et al., 2021; Shurtleff et al., 2016). The repeated data has been included in Figure 1a.

For the analysis of other miRNAs by RT-PCR, we assessed changes in intracellular and exosomal miRNA levels in the corresponding figures. In the qPCR analysis, miRNA levels were normalized to the total amount of RNA.

4) In figure 1, the authors show that in YBX1-KO cells, miR223 levels are decreased in the exosome. They further suggest this is because YBX1 binds with high affinity to miR223. This binding is compared to miR190 which the authors state is not enriched in the exosome. However, no data showing that miR190 is not present in the exosome is shown. A figure showing the amount of cellular and exosomal miR223 and 190 should be shown together on the same graph.

In previous publications we demonstrated that miR190 is not localized in exosomes and not significantly changed in YBX1 knockout (KO) cells and exosomes derived from YBX1 KO cells (Liu et al., 2021; Shurtleff et al., 2016). The repeated data has been included in Figure 1a.

5) Figure 2 Supplement 1 - As to determine the nucleotides responsible for interacting with YBX1, the authors made several mutations within the miR223 sequence. However, no explanation is given regarding the mutant sequences used or what the ratios mean. Mutant sequences need to be included. How do the authors conclude that UCAGU is important when the locations of the mutations are unclear? Also, the interpretation of this data would benefit from a binding affinity curve as shown in Fig 2C.

The ratio is of labeled miR223/unlabeled miR223 (wt and mutant). All mutant sequences of miR223 have been included in Figure 2 supplement 1.

6) While the binding of miR223mut to YBX1 is reduced, there is still significant binding. Does this mean that the 5nt binding motif is not exact? Do the authors know if there are multiple nucleotide possibilities at these positions that could facilitate binding? Perhaps confirming binding "in vivo" via RIP assay would further solidify the UCAGU motif as critical for binding to YBX1.

The binding affinity of miR223mut with YBX1 is reduced approximately 27-fold compared to miR223. We speculate that the secondary structure of miR223 may contribute to the interaction with YBX1.

Our EMSA data, in vitro packaging data, and exosome analysis reinforce the conclusion that UCAGU is critical for YBX1 binding. These findings suggest that the presence of the UCAGU motif in miR223 is crucial for its interaction with YBX1 and subsequent sorting into exosomes.

7) Figures 2g, h - It would be nice to show that miR190mut also packages in the cell-free system. This would confirm that the sequence is responsible. Also, to confirm that the sorting of miR223 is YBX1-dependent, a cell-free reaction using cytosol and membranes from YBX1 KO cells is needed.

Although we have not performed the suggested experiment, we purified exosomes from cells overexpressing miR190sort and observed an increase in the enrichment of miR190sort in exosomes compared to miR190. This finding confirmed that the UCAGU motif facilitates miRNA sorting into exosomes.

Regarding the in vitro packaging assay, our previously published paper demonstrated that cytosol from YBX1 knockout (KO) cells significantly reduces the protection of miR223 from RNase digestion. We concluded that the sorting of miR223 into exosomes is dependent on YBX1 (Shurtleff et al., 2016).

8) In Figure 3a, the authors show that miR223 is mitochondrially localized. Does the sequence of miR223 (WT or Mut) matter for localization? Does it matter for shuttling between YBAP1 and YBX1?

The localization of miR223mut has not been tested in our current study. We plan to conduct these experiments in the future.

9) Supplement 3c - Is it strange that miR190 is not localized to any particular compartment? Is miR190 present ubiquitously and equally among all intracellular compartments?

Most mature miRNAs are predominantly localized in the cytoplasm. Although there is no specific subcellular localization reported for miR190 in the literature, our experimental findings indicate a relatively high expression of miR190 in 293T cells. It is likely that most of miR190 is localized in the cytosol. However, it is also possible that a small fraction of miR190 may associate with a membrane, which could explain its distribution in various subcellular structures. Importantly, we did not observe enrichment of miR190 in the mitochondria or exosomes.

10) Figure 3h - Why would the miR223 levels increase if you remove mitochondria? Does CCCP also cause miR223 upregulation? I would have thought miR223 would just be mis-localized to the cytosol.

We report that the levels of cytoplasmic miR223 increase following the removal of mitochondria using CCCP treatment. While we cannot rule out the possibility that upregulation of miR223 is directly caused by CCCP treatment, we suggest that miR223 becomes mis-localized to the cytosol upon mitochondrial removal. Our data suggests that mitochondria contribute to the secretion of miR223 into exosomes. When mitochondria are removed by mitophagy, cytosolic miR223 is not efficiently secreted, which provides an alternative explanation for the observed increase in miR223 level after mitochondrial removal.

11) Figure 3i - What is the meaning of "Urd" in the figure label? This isn't mentioned anywhere.

“Urd” represents Uridine. Uridine is now spelled out in figure 3i. The absence of mitochondria can impact the function of the mitochondrial enzyme dihydroorotate dehydrogenase, which plays a role in pyrimidine synthesis. To address this issue, one approach is to supplement the cell culture medium with Urd. A previous study demonstrated that primary fibroblasts showed positive responses when Urd was added to the cell culture medium, resulting in improved cell viability for extended periods of time (Correia-Melo et al., 2017).

12) Figure 3j - The data is presented as a ratio of EV/cell. Again, this inaccurately represents the amount of miR223 in the EV. This issue is apparent when looking at Figures 3h and 3j. In 3h, CCCP causes an upregulation of intracellular miR223. As such, the presumed decrease in EV miR233 after CCCP (3j) could be an artifact due to increased levels of intracellular miR223. Both intracellular and EV levels of miRs need to be shown.

Both the intracellular and exosomal levels of miR223 have been included in Figure 3j.

13) In Figure 4, the authors show that when overexpressed, YBX1 will pulldown YBAP1. Can the authors comment as to why none of the earlier purifications show this finding (Figure 1 for example)? Even more curious is that when YBAP1 is purified, YBX1 does not co-purify (Figure 4 supplement 1a, b).

In Figure 4a-b, human YBX1 fused with a Strep II tag was purified from 293T cells using Strep-Tactin® Sepharose® resin in a one-step purification process. Our data has shown that YBAP1 is expressed in 293T cells.

In Figure 1 and Figure 4 Supplement 1a, human YBX1 or YBAP1 fused with His and MBP tags were purified from insect cells using a three-step purification process involving Ni-NTA His-Pur resin, amylose resin, and Superdex-200 gel filtration chromatography.

One possibility is that human YBX1 or YBAP1 may not interact well with insect YBAP1 or YBX1, which could result in separate tagged forms of YBX1 or YBAP1 isolated from insect cells.

Another possibility is that the expression levels of insect YBAP1 and YBX1 may be too low. Consequently, tagged forms YBX1 or YBAP1 expressed in insect cells may copurify with partners not readily detected by Coomassie blue stain. However, in Figure 4 Supplement 1b, human YBX1 fused with His and MBP tags was co-expressed with non-tagged human YBAP1, and both bands of YBX1 and YBAP1 were visible on the Coomassie blue gel after purification using Ni-NTA His-Pur resin, amylose resin, and Superdex-200 gel filtration chromatography.

14) Figure 4f, g - The text associated with these figures is very confusing, as is the labeling for the input. Also, what is "miR223 Fold change" in this regard? Seeing as your IgG should not have IP'd anything, normalizing to IgG can amplify noise. As such, RIP assays are typically presented as % input or fold enrichment.

The RIP assay results have been calculated and presented as a % input in Figure 4g.

15) Figure 4h - The authors show binding between miR223 and YBAP1 however it is not clear how significant this binding is. There is more than a 30-fold difference in binding affinity between miR223 and YBX1 than between miR223 and YBAP1. Even more, when comparing the EMSAs and fraction bound from figures 1 and 2 to those of Figure 4h, the binding between miR223 and YBAP1 more closely resembles that of miR190 and YBX1, which the authors state is a non-binder of YBX1. The authors will need to reconcile these discrepancies.

We agree that the binding of YBAP and YBX1 differ quite significantly in the affinity of their interaction with miR223. It is difficult to draw conclusions from a comparison of the affinities of YBX1 for miR190 and YBAP1 for miR223. Nonetheless, a quantitative difference in the interaction of YBAP1 with miR223 and miR190 is apparent (Fig. 4 h, I, j) and we observed no enrichment miR190 in isolated mitochondria (Fig. 3 supplement 1a) whereas YBAP1 selectively IP’d miR223 from isolated mitochondria (Fig. 4 f and g).

16) Can the authors present the Kd values for EMSA data?

The Kd values for the EMSA data have been added to the respective figures.

17) Figure 5 - Does YBAP1-KO affect mitochondrial protein integrity or numbers?

We generated stable cell lines expressing 3xHA-GFP-OMP25 in both 293T WT and YBAP1-KO cells, but we did not observe any alterations in mitochondrial morphology (Author response image 1).

Author response image 1.

Additionally, we performed a comparison of different mitochondrial markers using immunoblot in 293T WT cells and YBAP1-KO cells and did not observe any changes in these markers (data has been included in Figure 5b.).

18) Figure 6a - Are the authors using YBAP1 as their mitochondrial marker? Please include TOM20 and/or 22.

In Figure 4c and 4e, our data clearly demonstrate that the majority of YBAP1 is localized in the mitochondria.

To further validate this localization, we performed immunofluorescence staining using antibodies against endogenous Tom20 and YBX1. The immunofluorescence images document YBX1 associated with mitochondria (Author response image 2 and new Fig 6a.).

Author response image 2.

19) Figure 6b - Rab5 is an early endosome marker and may not fully represent the organelles that become MVBs. Co-localization at this point does not suggest that associating proteins will be present in the exosome, and it is possible that the authors are looking at the precursor of a recycling endosome. Even more, exosome loading does not occur at the early endosome, but instead at the MVB. Perhaps looking at markers of the late endosome such as Rab7 or ideally markers of the MVB such as M6P or CD63 would help draw an association between YBX1, YBAP1, and the exosome. Also, If the authors want to make the claim that interactions at the early endosome leads to secretion as an exosome, the authors should show that isolated EVs from Rab5Q79L-expressing cells contain miR223.

We have previously used overexpressed Rab5(Q79L) to monitor the localization of exosomal content, specifically CD63 and YBX1, in enlarged endosomes (Liu et al. 2021, Fig. 4A, B). These endosomes exhibit a mixture of early and late endocytic markers, including CD63. (Wegner et al., 2010). Hence, the presence of Rab5(Q79L)-positive enlarged endosomes does not solely indicate early endosomes.

20) The mentioning of P-bodies is interesting but at no time is an association addressed. This is therefore an overly speculative conclusion. Either show an association or leave this out of the manuscript.

In a previous paper we demonstrated that YBX1 puncta colocalize with P-body markers EDC4, Dcp1 and DDX6 (Liu et al., 2021).

21) In lines 55-58, the authors make the comment "However, many of these studies used sedimentation at ~100,000 g to collect EVs, which may also collect RNP particles not enclosed within membranes which complicates the interpretation of these data." Do RNPs not dissolve when secreted? Can the authors give a reference for this statement?

In a previous paper, we demonstrated that the RNP Ago2 does not dissolve in the conditioned medium and is not in vesicles but sediments to the bottom of a density gradient (Temoche-Diaz et al., 2019).

Author Response

Reviewer #1 (Public Review):

In this study, Shin and colleagues investigate the role of the posttranslational modification of the DNA methyltransferase by covalent linkage of the N-Acetylglucosamine (O-GlcNAc).

The authors present compelling evidence showing that a prolonged high fat/sucrose diet causes global protein O-GlcNAcylation in the liver and DNMT1 is among the proteins that increase their O-GlcNAc level. This result is significant because of the paucity of in vivo data addressing the interplay between metabolism and protein O-GlcNAcylation. The paper also shows that DNMT1's O-GlcNAcylation level correlated to the extracellular glucose levels in other cell types.

Using mass spectrometry, the authors identify S878 as the main site for O-GlcNAcylation. It is noteworthy that the mapping was performed with hyper-O-GlcNAcylated cells and may be different in a physiological situation. To investigate how O-GlcNAcylation of S878 of DNMT1 impacts its activity and ultimately DNA methylation patterns, Shin and colleagues mostly use a cellular model of hyper O-GlcNAcylation induced by the combination of high glucose and a chemical inhibitor of OGA (the only enzyme responsible for O-GlcNAc removal). The data shows that increased O-GlcNAcylation resulting from the combination of high glucose and OGA inhibition causes a reduction of DNMT1 activity and local loss of DNA methylation specifically at partially methylated domains.

This study brings completely new knowledge on the regulatory function of glycosylation of DNMT1 and its impact on its methyl-transferase activity and downstream genomic methylation. Furthermore, the manuscript introduces new data on the interplay between cellular metabolism and O-GlcNAcylation on DNMT1 and other proteins. The experiments are well-controlled, and their interpretation is sound. This study should be of special interest to the fields of fundamental and environmental epigenetics, as well as metabolism.

The main limitation of the study is the convolution of the functional experiments where the perturbation is a combination of high glucose and chemical inhibition of OGA. The relative contribution of the two variables is partially addressed in Figure 3-figure supplement 1B which shows that high glucose increases DNMT1 activity (Hep3B cells) while Figure 3D shows that high glucose when combined with OGA inhibitor decreases DNMT1 activity (Hep3B cells). As discussed, the data suggest that high-glucose and OGA inhibition may have an antagonistic effect on DNMT1 activity. An experiment of treatment of the cells with the OGA inhibitor in physiological glucose conditions would address this gap of knowledge.

We thank the reviewer for the suggestion. The physiological glucose levels are between 5 to 7 mM, and 25mM is in hyperglycemic range, which corresponds to severe diabetes. The new Figure 1A shows TMG treatment with physiological glucose conditions. We have included new WB data of 5mM glucose, 5mM glucose + TMG, 25mM glucose, and 25mM glucose + TMG (Figure 1A).

To understand the impact of the environment (in this study: extracellular glucose level) on the epigenome, one should keep in mind the variation of cytosine methylation patterns between individuals and over time. A recent large-scale profiling of DNA methylation of 137 individuals shows a near absence of individual variation between replicates of the same cell type, suggesting that genomic methylation patterns are largely insensitive to the environment (https://doi.org/10.1038/s41586-022-05580-6).

Comparative methylomes of healthy and diabetic individuals are needed to examine the medical significance of the findings presented here. It is possible that the modulation of DNMT1 activity by O-GlcNAc modification is relevant for a specific cell type or developmental stage that remains to be discovered.

We thank the reviewer for the suggestion. While the present study is focused on the functional impact of glucose concentrations on O-GlcNAcylation of DNMT1, the extension of this work to diabetic individuals is a goal for a follow up project.

Reviewer #2 (Public Review):

I've read the manuscript by Shin et al with great interest. The authors describe the identification of O-GlcNAcylation of DNMT1 and the impact this modification has on the maintenance activity of DNMT1 genome-wide and that modification of S878 leads to enzyme inhibition. The manuscript is written in a clear and understandable way making it easy for the reader to understand the logic as well as the steps of the experimental approach.

The authors identify O-GlcNAcylation of DNMT1 in a number of different cell lines by combining inhibition studies and WB and further on they identify the modification sites with LC/MS, predictions, and mutational studies. I really like the experimental approach, which while being straightforward (albeit technically challenging), is powerful and well-controlled in this case to unequivocally prove the modification of DNMT1 and identify the site. However, mutation of the two identified modification sites does not remove all the O-GlcNAcylation signal associated with DNMT1, thus possibly not all the possible sites were identified. While this is not a criticism of this manuscript, it would be interesting to know what other sites are modified and the enzymatic/biological effects associated.

We completely agree with the reviewer. As the O-GlcNAc band was also detected in double mutated DNMT1 (Figure 2D), it is expected that undetected O-GlcNAcylated sites will exist. This is a limitation of current MS analysis and is known to be difficult to detect in the case of modified sites located at both 5’- and 3’- ends of the protein or around the site cut by endoprotease such as trypsin. In follow up work we plan to detect more diverse O-GlcNAc modified sites using more types of endoproteases and observe changes in the phenotype of various cells accordingly.

Also, the authors isolate the modified DNMT1 from cells using immunoprecipitation, which is indeed useful to study the changes in catalytic activity but does not provide any information if the cellular localisation of modified DNMT1 changes.

We apologize for this oversight. We have added a DNMT1 localization assay via immunofluorescence (IF) in the revised manuscript (Figure 3—figure supplement 3). We found no difference in DNMT1 localization between wild type and S878A mutants.

Subsequently, the authors checked the impact of high glucose diet on the genome-wide DNA methylation patterns. The observed effects (Fig 4A) are very strong, almost as strong as observed with Aza treatment and therefore I wonder if LINE/IAP or other elements are getting activated (as observed with genome-wide demethylation with Aza).

We thank the reviewer for the suggestion. Changes in methylation of LINE-1 by hyperglycemia condition are displayed in Figure 4—figure supplement 4. In the case of LINE-1, DNA methylation is lost globally in hyperglycemia conditions. While beyond the scope of this study, a more thorough examination of the impact of the observed loss of methylation under high glucose conditions is of interest.

Do the authors see any changes in cell phenotype, slower/faster proliferation, or increased apoptosis due to the activation of mobile elements (not only ROS)?

This is also a very interesting idea. We plan on further investigating this as part of a follow up study.

Another point is that the S878A mutant seems not to be able to fully maintain the DNA methylation (Fig 4A). Does O-GlcNAcylation recruit any additional interactors? Given that the authors immunoprecipitated DNMT1 and use it for activity assay, it is possible, that the modification attracts an additional protein factor that could in turn inhibit DNMT1 activity (as observed). Therefore, the observed kinetic effect could be indirect, while still interesting and important, the mechanism of inhibition would be different.

We thank the reviewer for the great suggestions. According to Figure 4A, in the case of mutated DNMT1, a slight methylation loss appears to occur in both conditions. There could be for a number of reasons. It may be due to interacting proteins or it may be caused by some damage of DNMT1 itself. A further investigation of this is planned as a follow up project.

DNA methylation clock can be used to estimate the biological age of a tissue/cells. While not directly in the line of the manuscript, I was wondering if the DNA methylation changes in the high glucose diet would affect the methylation sites used for the DNAme clock. Meaning, would the cells/tissue epigenetically age faster when in high glucose media, and if the Ala mutant could provide resistance to that?

We thank the reviewer for the interesting suggestion. We believe this is beyond the scope of this manuscript, but we'll consider this with interest in the future.

In discussion, the authors write that this is the first investigation of O-GlcNAcylation in relation to DNA methylation, while this is true for DNMTs, TET enzymes, that oxidise 5mC and trigger active DNA demethylation have been shown before to also be modified.

We have toned down the language throughout the revised manuscript. This is the first investigation into the maintenance of DNA methylation. Although there is a great deal of evidence regarding the important regulatory role of O-GlcNAcylation in gene regulation, a direct link with maintenance of DNA methylation has not previously been established.

A nice and rigorous study, with important observations and connections to biological effects. It would be nice to prove that the effects are direct and not associated with other factors that could be recruited by the modification and impact the activity of DNMT1. I find it a bit surprising that phosphorylation of the target serine does not impact DNMT1 activity as well.

We thank the reviewer for the positive comments and agree that there are many interesting avenues to follow up on this.

Reviewer #3 (Public Review):

The authors investigate the potential effect of OGlcNacylation on the activity of the DNA methyltransferase DNMT1.

Some results that are convincingly obtained include:

• There is more overall OGlcNacylation when Glucose concentration in the culture medium or the feed is high;

• DNMT1 is OGlcNacylated, and more so in high glucose or on rich chow;

• The position S878 can be OGlcNacylated;

• The activity of transfected DNMT1 is decreased in high glucose conditions. This effect is lessened when S878 is mutated to A or D.

Some results that are suggested but not fully backed by experimental data include:

• This process happens to the endogenous protein under physiologically relevant conditions;

We agree that we could not completely rule out endogenous DNMT1 in our experiments. We have adjusted the language in the revised manuscript to acknowledge this. However, we confirmed the change in activity of recombinant DNMT1 (Figure 3D), and also demonstrated the change in activity under physiological conditions (normal physiological glucose level vs hyperglycemic range) in Figure 3—figure supplement 1B. This is a result that directly shows that the activity of DNMT1 changes under physiological conditions. In addition, DNA hypomethylation due to high glucose has been previously reported, already (Kandilya et al., 2020; Lan et al., 2016). Our results suggest a possible mechanism for this.

Kandilya, D., Shyamasundar, S., Singh, D.K., Banik, A., Hande, M.P., Stunkel, W., Chong, Y.S., and Dheen, S.T. (2020). High glucose alters the DNA methylation pattern of neurodevelopment associated genes in human neural progenitor cells in vitro. Sci Rep 10, 15676.

Lan, C.C., Huang, S.M., Wu, C.S., Wu, C.H., and Chen, G.S. (2016). High-glucose environment increased thrombospondin-1 expression in keratinocytes via DNA hypomethylation. Transl Res 169, 91-101 e101-103.

• This process is responsible for changes in DNA methylation, leading to changes in gene expression, leading to increased ROS and increased apoptosis.

We confirmed that ROS levels increased under high glucose conditions through DCFH-DA fluorescence experiments (Figure 5A). In addition, γH2A.X fluorescence experiments showed that DNA damage was increased under high glucose conditions (Fig. 5B). On the other hand, in the case of the S878A mutant, DNA damage was reduced under hyperglycemic conditions compared to wild type DNMT1 despite an increase in ROS levels (Fig. 5B). Moreover, we verified that the DNA damage did not come from oxidative stress through 8-OHdG analysis (Figure 5—figure supplement 4). Therefore, DNA oxidative stress is suppressed by DNMT1 due to the increase of ROS under high glucose conditions. However, the reduction of DNA methylation by O-GlcNAcylation of DNMT1 induces apoptosis due to oxidative stress.

Studying the connection between cellular metabolism and epigenetic phenomena is interesting. However, I feel that the article falls short of its aims because of the limits of the experimental system, some missing controls, and some data overinterpretation.

We hope the reviewer finds our revised manuscript more suitable.

Author Response

Reviewer #1 (Public Review):

Overall, this manuscript exposes key gaps in patient care resulting from the pandemic, as well as the challenges and unmet needs felt by healthcare workers in cervical cancer screening. The authors’ findings on the struggles while regaining screening volume across the nation in a sustainable way, demonstrate that pre-existing weaknesses in the cancer control system were exacerbated by the pandemic and are integral to amend. The authors were able to identify these gaps in care and work environments through their synthesis of qualitative interviews. I applaud the use of such mixed methods, which emphasizes the complementary need for both quantitative and qualitative data. What could be better strengthened in the manuscript is the authors’ justification for statistical analyses within the context of the research question, and reporting of survey administration and management.

The authors thank the reviewer for a thorough assessment of the manuscript. We have addressed the reviewer’s concerns regarding justification of statistical analyses in the Data Analysis, Quantitative survey data section, and reporting of survey administration and management in the Results, Quantitative survey data section.

Reviewer #2 (Public Review):

Fuzzell et al. conducted a mixed-method study looking into the possible impact of COVID-19 on clinician perceptions of cervical cancer screening. The authors examined how the pandemic-related staffing changes might have affected the screening and abnormal results follow-up during the period October 2021 through July 2022.

They found that 80% of the clinicians experienced decreased screening during the start of the pandemic and that ≈67% reported a return to pre-pandemic levels. The general barriers for not returning to pre-pandemic levels were staffing shortages and problems with structural systems for tracking overdue patients and those in need of follow-up after abnormal screening tests.

Strengths:

There is a high focus on the consequences and the need for action to prevent the ongoing impact of COVID-19 on cervical cancer screening. Some of the actions mentioned by the authors could be the use of HPV self-sampling kits, and it is interesting to be provided knowledge on the clinicians' views on HPV self-sampling. Both are of high interest to the general population in the US. Throughout the discussion, the authors and their claims are supported by other studies.

Weaknesses:

The lack of a National representative sample, where 63% of the responding clinicians were practicing in the Northeast, affects the possibility of generalization of the results found in the study. The overrepresentation of white females is not addressed in the discussion. This composition could have affected the results, especially when the authors report a need to look at higher salaries and better childcare to maintain adequate staffing.

The conclusions are mostly supported by the data, however, some aspects of the data analysis need to be clarified.

We thank the reviewer for their constructive feedback. Despite our best efforts, we were unable to recruit a sample more representative of all US regions. We note this limitation in the discussion: “Notwithstanding efforts to achieve a regionally diverse sample, 63% of responding clinicians were practicing in the Northeast at the time of their participation. Given that COVID-19 policies varied widely by state, this regional imbalance may limit the generalizability of our results. Despite the oversample of clinicians in the Northeast, region was not a significant predictor of either outcome.” Also, we acknowledge the high enrollment of White women in our provider sample and now address this point in the discussion: “Similarly, our sample was 85% female and 70% White. Although ideally we would have included a sample that was more diverse with respect to race and gender, these characteristics are not disparate from the majority of clinicians who perform cervical cancer screening (e.g., race: Women’s Health NPs [77% White], active Ob/Gyns [67% White], all active physicians [64% White]; gender: all NPs [92% female], Ob/Gyns [64% female], all active physicians [37% female]).” Data describing these characteristics are reported in the Association of American Medical Colleges (AAMC) 2022 Physician Specialty Data Report and Executive Summary, the 2018 NPWH Women’s Health Nurse Practitioner Workforce Demographics and Compensation Survey: Highlights Report, and a published paper describing the characteristics of nurse practitioners in the US, which are cited in text.

Reviewer #3 (Public Review):

This US study presents findings from an online survey and in-person interviews of healthcare providers regarding themes associated with cervical screening in federally qualified health centres (FQHCs). The study provides insights during the post-acute phase of the pandemic into a range of areas, including perceived changes in the provision of cervical cancer screening services and the impact of the pandemic, staffing and systems barriers to cervical cancer screening, strategies for tracking missed screens and catch-ups, follow-up of abnormal screening results, as well as attitudes towards HPV self-sampling. Results indicate persisting pandemic-related impacts on patient engagement and staffing, as well as system barriers to effective screening, catch-up of missed screens and follow-ups. Taken together, these issues may lead to increases in cervical cancer in the long-term in populations serviced by these centres, if measures are not taken to adequately support them. Participants were recruited from various regions in the US, however, the study was not conducted using a nationally-representative sample. Although highlighted issues are informative, findings cannot be generalised and larger studies are warranted in the future to monitor cervical screening provision and outcomes in FQHCs.

We thank the reviewer for their thorough assessment of the manuscript. In the discussion, we have made sure to note the non-nationally representative sample and need for continued monitoring of cervical cancer screening and related outcomes in underserved settings and communities.

Author Response

Reviewer #2 (Public review):

1) The systematic review includes data from some studies where PCOS is self-reported. While self-reported PCOS information has been found to be largely sensitive and specific, it would be of interest to know if prevalence ratios of mental health-related were impacted by self-reporting.

Thank you for your insightful comment regarding the potential impact of self-reporting on the prevalence ratios of mental health-related outcomes in women with PCOS. We agree that this is an important factor to consider.

In response, we have revisited all the studies included in our review. We have updated Supplemental Tables 2-4 to provide greater transparency and understanding. These revised tables now include a new column specifying the mental health assessment method used in each study. This update should allow for a more nuanced interpretation of the results, taking into account the potential impact of self-reporting.

Furthermore, we conducted a sensitivity analysis by rerunning the meta-analysis to discern the potential influence of self-reported PCOS on our results, excluding the studies that relied solely on self-reported PCOS diagnosis. After we excluded studies where PCOS was self-reported, the point estimate for anxiety was similar whereas point estimates for depression and eating disorder were slightly higher but none of the estimates were different beyond chance compared to the original analysis. We believe these steps significantly strengthen the clarity and robustness of our findings (Line 314; Supplemental Tables 7 and 8).

2) Likewise, the screening vs self-reported nature of the mental health disorders is not clear from the information included in the characteristics table.

We have modified our Supplemental Tables 2-5 to include a column detailing the method of ‘Mental Health Assessment’. We should note that the majority of the studies directly assessed mental health using a variety of validated questionnaires. We have also included in the Discussion a section emphasizing that some of the studies included in the review relied on self-reported PCOS diagnosis and its potential impact. We also highlighted that while self-reported information is generally reliable, it is subject to potential bias that could impact the prevalence ratios of mental health-related conditions (Line 460).

3) Calculated prevalence ratios were compared with prevalence values for the general population to determine the excess prevalence. However, the source of these general population statistics (i.e., whether these figures come from the control data in the included studies or other sources) is not clear.

Thank you for raising this important point. We have now clarified in our Methods section that the general population statistics used for determining excess prevalence were derived from the control data in the included studies. We hope this provides the necessary transparency for our approach in calculating and interpreting the prevalence ratios (Line 210).

4) The estimated costs for anxiety-, depression- and eating disorder-related care are accessed in published papers and used to calculate the excess costs. Conclusions would be strengthened by a defence of these figures, particularly for anxiety where the source paper is from 1999.

Thank you for your insightful comment. We agree that providing a justification for our choice of cost estimates, especially for the anxiety care cost from a 1999 study, would strengthen our conclusions. The 1999 source was selected because it is a seminal study that offers a comprehensive breakdown of anxiety-related care costs. Despite its age, this paper is often cited in contemporary research due to its rigorous methodology and the granularity of its cost analysis. Adjusted for inflation, its findings still provide an insightful comparison point for current data. To ensure that these figures accurately represent present-day costs, we have adjusted them for inflation using the medical care inflation calculator. Our choice of these specific studies was based on their rigorous methodology, the detailed breakdown of costs, and their relevance to our targeted age groups. The aforementioned adjustments and justifications ensure that these figures aptly represent the present-day costs of treating these conditions.

Similarly, the 2021 papers on depression and eating disorders present comprehensive and up-to-date analyses of the economic burdens associated with these conditions. These papers were selected for their rigorous methodologies, comprehensive cost breakdowns, and alignment with our age-specific focus. The Greenberg et al. (2021) paper, for example, is an authoritative source that provides detailed analysis on the economic burden of adults with major depressive disorder. Likewise, the paper by Streatfeild et al. (2021) offers a meticulous investigation into the socio-economic cost of eating disorders in the U.S., making it an apt choice for our study. We recognize the necessity of providing a robust justification for our choice of these particular papers, and we have endeavored to do so in our Methods section, thus reinforcing the transparency of our approach. We have clarified this in our Methods section to make our approach more transparent to readers (Line 225).

5) An inflation tool is used to adjust the figure, but this does not take into account changes in treatment or practice since this estimate was made. The accuracy of these estimated figures is central to the final conclusions.

Thank you for your valuable comment. We do note that the inflation figures used are a healthcare-specific inflation factor, as healthcare inflation differs from general consumer inflation. However, we agree that the inflation-adjusted figures do not necessarily account for changes in treatment practices since the original estimate was made, assuming these changes would alter the cost of care. We have added a discussion of this limitation in our manuscript and proposed future studies to validate these estimates using more recent data (Line 473).

Author Response

Reviewer #1 (Public Review):

GSK3 is a multi-tasking kinase that recognises primed (i.e. phosphorylated) substrates. One of the mechanisms by which the activity of GSK3 can be regulated is through N-terminal (pSer9) phosphorylation. In this case, the phosphorylated N-terminus turns into a pseudo-substrate that occupies the substrate binding pocket and thus inhibits the activity of GSK3 towards its real substrates.

One outstanding question is how this autoinhibitory mechanism can affect some, but not all signaling pathways that GSK3 is involved in. One example is WNT/CTNNB1 signaling. Here, GSK3 plays a central role in the turnover of CTNNB1 in the absence of WNT, but this pool of GSK3 is not affected by pSer9 phosphorylation.

Gavagan et al. address this question using an in vitro approach with purified proteins. They identify a role for AXIN1 in protecting the "WNT signaling pool" of GSK3 from the auto- inhibition that occurs upon pSer9 phosphorylation.

Specifically, they show that i) GSK3-pSer9 is less capable of binding and phosphorylating primed CTNNB1 - thus suggesting that GSK3-pSer9 does not contribute to WNT signaling, ii) in the presence of AXIN1, GSK3-pSer9 becomes more capable of binding and phosphorylating CTNNB1 - suggesting that Axin can promote binding of GSK3 and CTNNB1 even when the primed binding pocket on GSK3 is blocked initially, iii) AXIN1 specifically prevents the PKA mediated phosphorylation of GSK3B on pSer9 - while leaving the phosphorylation of other PKA substrates unaffected.

Strengths:

• The authors use an in vitro system in which they can reconstitute different interactions and reactions using purified proteins, thus allowing them to zoom in on specific biochemical events in isolation.

• The authors measure the phosphorylation of primed substrates (pSer45-CTNNB1 or WNT- independent substrates) and quantify specific kinetic parameters (kcat, KM, and kcat/KM) - of wildtype non-phosphorylated GSK3B, pSer9GSK3B, or the non-phosphorylatable S9A-GSK3B, either in the presence or absence of AXIN1 (or an AXIN1 fragment).

• The experiments appear to be well-controlled and the results appear to be interpreted correctly.

Weaknesses:

• Key experiments (e.g. Figures 2 and 3) are described as being performed as n=3 technical replicates rather than independent/biological replicates.

We suggest that the replicates described in our work can properly be described as biological replicates, and we have updated the manuscript accordingly. We apologize for the confusion and elaborate on our reasoning below.

Each replicate reported for our in vitro kinetic assays is an independent reaction prepared in a separate reaction vessel, and replicates were analyzed on separate gels. Thus, each reaction is a distinct biological sample and should have been described as a biological replicate. A technical replicate would have been repeat measurements of the same timepoint from a single reaction.

Our original description as technical replicates was based on the notion that each replicate came from the same protein purification (biological sample). However, an analogy to cell culture experiments can illustrate why our initial reasoning was incorrect. In a cell culture experiment, cells from the same initial source are typically split into independent wells for biological replicates. Similarly, our proteins come from the same initial source but are split into independent reaction vessels for biological replicates.

The critical point is that, regardless of the precise terminology, our replicates capture the variability between independent experiments.

• The validation in a biologically relevant setting (i.e. a cellular context) is limited to Figure 4C, which shows that over-expression of AXIN1 reduces the total levels of pSer9-GSK3.

The biochemical experiments presented in our work address a critical gap in the signaling field and, together with the in vivo validation in Figure 4C, establish a model that was previously speculative. We suggest that further in vivo experiments are beyond the scope of the current manuscript.

The authors convincingly show that AXIN1 can play a role in shielding GSK3 from auto- inhibition. As it stands, the impact of this work on the field of WNT/CTNNB1 signaling is likely to remain limited. This is mainly due to the reason that the mechanism by which AXIN1 shields the WNT/CTNNB1 signaling pool of GSK3 from pSer9 inhibition remains unresolved. Based on the fact that a mini AXIN1 (i.e. an AXIN1 fragment) behaves the same as WT AXIN1, the authors conclude that AXIN1 likely causes allosteric changes on GSK3 but is less likely to block PKA from binding. They cannot conclusively show this, however, as they do not have evidence in favour of one or the other explanation.

We thank the reviewer for this important comment which details the central concern raised in the review process. To address this point, we have collected additional biochemical data that conclusively shows that the Axin shielding effect is allosteric and not a steric block. We demonstrated that a minimal, 27 amino acid Axin peptide produces the same GSK3β shielding behavior as full length Axin and miniAxin. The minimal Axin peptide does not sterically occlude the GSK3β phosphorylation site. This data is included in a revised Fig 4A and described on lines 115-120 of the revised manuscript.

However, this study does offer more insight into the compartmentalisation of GSK3 and the quantitative parameters may be used in computational models describing the different cellular activities of GSK3.

This work also has conceptual significance: Scaffold proteins are known to promote signal transduction by bringing proteins together (often: kinases and substrates). Here, Gavagan et al. show that AXIN1 also plays a second role, namely in protecting one of its binding kinases (GSK3) from inhibitory signals. This could potentially hold for other scaffolding proteins as well.

Reviewer #2 (Public Review):

Gavagan et al. investigated the role of the scaffolding protein, Axin, in the cross-pathway inhibition of GSK3b. The authors utilize reconstituted Axin, b-catenin, GSK3b, and protein kinase A to test 2 models. In the first model, the formation of the complex consisting of Axin, b-catenin, and GSK3b overcomes inhibitory phosphorylation of serine 9 of GSK3b. In the second model, the binding of Axin to GSK3b inhibits serine 9 phosphorylation through allosteric effects. Previous literature has established that the phosphorylation of serine 9 of GSK3b inhibits its kinase activity. To provide a quantitative measure of inhibition, the authors determine the binding affinity and catalytic efficiency of GSK3b in comparison to GSK3b phosphoS9 towards b-catenin. Interestingly, the data demonstrate a 200-fold decrease in Kcat/Km and 7 fold increase in Km. It is unclear why serine 9 mutation to alanine increases the rate of B-catenin phosphorylation more than the GSK unphosphorylated protein in figure S10.

We thank the reviewer for catching this inconsistency. In the Michaelis-Menten plots presented in the main text (Figure 2 & Figure 3D), rates for unphosphorylated GSK3β and GSK3β_S9A are indistinguishable. These plots were used to determine the kinetic parameters reported in Table S1 (now Supplementary file 1a). The purpose of Figure S10 (now Figure 2-figure supplement 8) was to confirm that these reactions were first order (linear) in enzyme concentration, but the reviewer is correct to flag the inconsistency in absolute rates. In Figure S10A (now Figure 2-figure supplement 8A), the rates for unphosphorylated GSK3β were ~2-3-fold lower than expected.

We have reanalyzed the original frozen reaction timepoints on new western blots. The results were identical for unphosphorylated GSK3β and GSK3β_S9A, resolving the apparent discrepancy. Upon review of the original western blot images, we noted that they were relatively noisy, potentially indicating incomplete blot transfer or an antibody going bad. Because we were able to reanalyze the original samples and obtained internally consistent results, we suggest that the updated data should replace the original data. The updated data are included in a revised Figure S10A (now Figure 2-figure supplement 8A).

Next, the authors tested if the addition of Axin could overcome this inhibition. Although the addition of Axin decreases the Km, thereby producing a 20-fold increase in catalytic efficiency, the addition of Axin does not rescue the catalytic turnover of the phosphorylated GSK3b. Hence, the authors propose that Axin does not rescue the kinase activity of GSK3b from the inhibitory effects of serine 9 phosphorylation.

Next, the authors test if Axin protects GSK3b from phosphorylation by the upstream kinase PKA. Excitingly, the data show a decrease in binding affinity and catalytic efficiency of PKA with GSK3b phosphoS9 in comparison to GSK3b. The binding of Axin inhibits GSK3b serine 9 phosphorylation by PKA but does not inhibit the phosphorylation of other PKA substrates such as Creb. The authors demonstrate that a fragment of Axin, residues 384-518, behaves similarly to the full-length Axin to shield GSK3b from phosphorylation. However, it is unclear how this fragment may bind in the destruction complex and if Axin has allosteric effects on GSK3b.

Author Response

Reviewer #1 (Public Review):

Various parts of the premotor cortex have been implicated in choices underlying decisionmaking tasks. Further, norepinephrine has been implicated in modulating behavior during various decision-making tasks. Less work has been done on how noradrenergic modulation would affect M2 activity to alter decision-making, nor is it clear whether noradrenergic modulation effects on activity would differ between the male and female sexes.

This manuscript addresses some of these questions.

• In particular, clear sex differences in task engagement are seen.

• May also show some interesting differences and distributions of β2 adrenergic receptors in M2 between males and females.

We thank the reviewer for their summary of our findings and thoughtful critique of our manuscript. In our revised manuscript we have taken measures to address the reviewer’s comments in line (blue edits in text and revised figures) with direct responses outlined below. We believe these revisions improve the scientific rigor of our findings and provide relevant context for our studies. We hope that they have sufficiently addressed the reviewer’s concerns.

Less clear is the specificity of systemic antagonism of β adrenergic receptors on the changes in M2 activity reported. As propranolol was given systemically, changes in M2 firing rates could also be due to broader circuit (indirect) activity changes. As it was not given locally, nor were local receptor populations manipulated, one is unable to make the conclusion that changes in neural activity are due to the direct effects of adrenergic receptors within M2 populations.

We agree that propranolol driven changes in anterior M2 activity may arise via multiple mechanisms, including direct action on the adrenoreceptors within M2, and indirect action via other regions that project to M2. Although locally activating inhibitory interneurons within M2 is sufficient to disrupt cueguided action plans and behavior in a 2AFC task (Inagaki et al., 2018), our noradrenergic manipulation was not restricted to M2. We have clarified our conclusions and provided additional discussion to highlight that propranolol actions were multifaceted and that direct actions in M2 are likely working in concert with propranolol mediated actions in other regions.

Also not clear, is the contribution of M2 to this task, and whether the changes in M2 activity patterns observed are directly responsible for the behavioral disruptions measured.

We have revised our introduction and discussion to more clearly outline the critical role of cue-guided action plans in M2 for successful behavior in 2AFC tasks. Suppression of cue-guided activity in M2 results in behavioral performance at near chance levels, similar to what we saw in females after propranolol (Guo et al., 2017; Inagaki et al., 2018; Li et al., 2016). Furthermore, targeted photostimulation of action plan encoding neurons in M2 is sufficient to drive behavioral responses (Daie et al., 2021). In our investigations it is plausible to expect propranolol related disruptions in other cognitive, sensory or motor regions. Based on the strong foundational evidence for M2 activity in 2AFC, the propranolol driven changes in anterior M2 in females, whether direct or indirectly mediated, are likely sufficient to drive behavioral disruptions in accuracy and/or trial completion.

Reviewer #2 (Public Review):

This paper by Rodbarg et al describes an interesting study on the role of beta noradrenergic receptors in action-related activity in the premotor cortex of behaving rats. This work is precious because even if the action of neuromodulatory systems in the cortex is thought to be critical for cognition, there is very little data to actually substantiate the theories. The study is well conducted and the paper is well written. I think, however, that the paper could benefit from several modifications since I can see 3 major issues:

We thank the reviewer for their generous comments on the potential impact of our manuscript as well as their suggestions to improve this work. Below we outline responses to specific comments raised by the reviewer in addition to adresing them in the revised manuscript. We hope these responses sufficiently address the reviewer’s concerns.

Both from a theoretical and from a practical point of view, the emphasis on 'cue-related' activity and the potential influence of NA on sensory processing is problematic. First, recent studies in rodents and primates have clearly demonstrated that LC activation is more closely related to actions than to stimulus processing (see Poe et al, 2020 for review).

Indeed during optimal performance the peaks of LC activity are larger when PETH are aligned to action initiation rather than the cue itself (Clayton et al., 2004). This alignment resolves variability in decision processing times and omitted cues. Although LC responses align with action they are evoked by, and occur after, cue presentation with LC responses to visual cues occurring ~ 60ms after presentation (Aston-Jones & Bloom, 1981). The same behavioral action without preceding task relevant cues does not evoke an LC response (Rajkowski et al., 2004)

In our current study cues initiate activity in anterior M2, this is our primary interest and where our electrodes are placed. The window between cue delivery and action completion hones in on our goal of investigating the role for β noradrenergic signaling in target cortical processing, rather than LC explicitly. In both NHP and rodents NE signaling (and evoked LC) promotes sustained cortical representations between cue onset and actions across cortical regions (dlPFC, S1) (Ramos & Arnsten, 2007; Vazey et al., 2018; Wang et al., 2007). In the current study we aligned neural data to either cue presentation (Figure 3) or action (lever press; Figure 4). Both presentations support a critical role for β adrenoreceptor signaling in suppressing irrelevant information, resolving and maintaining action plans. A unique feature of aligning the data to cue onset is that it allows us to see how the neural activity changes not only on completed trials (that end with a lever press) but also on omitted trials (which strongly increase after propranolol). We propose the reason we are seeing large increases in omitted trials is because β adrenoreceptor blockade either directly or indirectly prevents anterior M2 from resolving an action plan.

Second, the analysis of neural activity around cue onset should be examined with spikes aligned on the action, since M2 is a motor region and raster plots suggest that activity is strongly related to action (I'll be more specific below).

We agree that M2 shows important action plan activity which we highlight throughout the manuscript. In cued tasks, M2 neurons have been shown to represent action plans starting at cue onset that continues up to behavioral execution. Neural data was examined and results presented aligned to cue onset (illustrated in Figure 3) and aligned to action - lever press (illustrated in Figure 4). The impact of propranolol in diminishing action plan selection was similar in both action, and cue-aligned analyses.

The distinction between neural activity and behavior or cognition is not always clear. I understand that spike count can be related to motor preparation or decision, but it should not be taken for granted that neuronal activity is action planning. The analysis should be clarified and the relation between neural activity, behavior, and potential hidden cognitive operations should be explicated more clearly.

We have worked to clarify in our revised introduction, results and discussion the specifics of the known roles of neural activity in M2 in both action planning and decision making. We further expand that the neuronal activity in our study may reflect potential changes in cognitive processing and thus alter resultant behavioral outcomes.

The sex difference is interesting, but at the moment it seems anecdotal. From a theoretical point of view, is there any ecological/ biological reason for a sex dependency of noradrenergic modulation of the cortex? Is there any background literature on sex differences in motor functions in rats, or in terms of NA action? If not, why does it matter (how does it change the way we should interpret the data?) From a practical point of view, is there a functional sex difference in absence of treatment, or is it that the drug has a distinct effect on males vs females? This has very distinct consequences, I think.

We did not find overt differences in behavior in the absence of treatment. Only when noradrenergic function was challenged using propranolol did we identify functional sex differences. We agree that this has very distinct consequences – specifically it supports sex differences that can be revealed by perturbations of normal function. These functional sex differences may be a result of differences in the anatomy of central noradrenergic systems, a hypothesis further supported by our mRNA expression findings and existing literature on LC anatomy across species (Bangasser et al., 2011, 2016; Luque et al., 1992; Mulvey et al., 2018; Ohm et al., 1997; Pinos et al., 2001). Collectively these results have potential ramifications for understanding sex differences in disease prevalence and targeted treatments.

Background literature supports some innate sex differences in motor function and executive function in rodents and humans. Of particular relevance to our investigation is an established difference in behavioral strategy with females being more risk averse than males (Grissom & Reyes, 2019). Ethologically risk adverse strategies may support parental care roles, and increased inhibitory mechanisms may be selected for in females. Although this strategy was not directly tested in our study, the large increase in omissions after propranolol seen in females is in line with avoiding risk (incorrect choices) during uncertainty (disrupted neural signaling). As with other executive functions, the utilization of norepinephrine within the cortex along with other neuromodulators, and local microcircuit interactions would all contribute to promoting risk averse behavior.

These issues could be clarified both in the introduction and in the discussion, but the authors might have a different view on what is theoretically relevant here. In the result section, however, I think that both the lack of specificity in the description of behavior and cognitive operation and the confusion between 'sensory' and 'motor' functions make it very difficult to figure out what is going on in these experiments, both at a behavioral and at a neurophysiological level. First, the description of the behavior in the task is clearly not sufficient, which makes the interpretation of the measures very difficult.

We have made an effort to better specify the task and relevant behavioral operations in both the methods and results and have included a clearer task schematic (Figure 1A). We agree that the confusion between ‘sensory’ and ‘motor’ functions may make it more difficult to understand the findings in this study. Anterior M2 plays a unique role in representing motor/action plans that can be informed by sensory information. This integrative function creates difficulty in parsing the neural activity of anterior M2 as strictly motor, sensory or cognitive. In attempts to improve clarity we have expanded and highlighted relevant information on the known roles of M2 in the introduction and discussion.

One possible interpretation of the effects of the drug is a decrease in motivation, for instance, due to a decrease in reward sensitivity or an increase in sensitivity to effort. But there are others. More importantly, none of these measures can be used to tease apart action preparation from action execution, even though the study is supposed to be about the former.

Neural activity during action planning, prior to action execution is known to be an essential function of M2 (Barthas & Kwan, 2017; Gremel & Costa, 2013; Guo et al., 2017; Inagaki et al., 2018, 2022; Li et al., 2016; Siniscalchi et al., 2016; Sul et al., 2011; Wei et al., 2019) for optimal performance in 2AFC tasks. In all, we found that the representation/separation of opposing action plans (a well validated function of M2) prior to responses (lever press) is degraded after propranolol, especially in females. We have provided additional emphasis on these foundational studies throughout our revised manuscript.

To minimize impact of motivational factors, effort and reward size remain consistent within our task, and all trials require a random initiation hold prior to cue delivery. As described in our general response to the editor above (Figure 1, above), we investigated whether motivational changes may be reflected in our M2 recordings. PETHs from the first and last 10 trials within saline sessions did not identify potential motivation related differences in anterior M2 activity. Similarly, across propranolol sessions the neural activity was consistent between early and late trials. We used early and late trials as there was a mild decrease in trial rate during saline sessions in both males and females, potentially indicative of motivation/reward sensitivity changes during these sessions. M2 neural responses consistently separate action plans (after saline) or failed to separate action plans (propranolol sessions).

Also, but this is less critical: In Figures 2C and D, it looks like there is a bimodal distribution for the effect of propranolol in females. Is there something similar in the neuronal effects of the drug? And in the distribution of receptors? Can it be accounted for by hormonal cycles/ anything else?

Although there is some clustering in behavioral outcomes all data passed normality assumption as appropriate. Propranolol treatments were not synchronized to hormonal cycles, and the data likely include animals at various hormonal stages. Similar clustering was not apparent in neuronal effects of propranolol, although propranolol increased variability in many measures.

In a pilot experiment we did not see any difference in baseline performance on our 2AFC task across the hormonal cycle (diestrous, proestrous, estrous or metestrous) of females in any measure including accuracy (F(3,33)=0.59, p=0.63, one-way ANOVA) and omissions (F(3,33)=0.51, p=0.68).

The description of neural activity is also very superficial. In general, it is not clear how spike count measures have been extracted. For example, legend and figure C are not clear, is the (long) period of cue presentation included in the 'decision time'?? "Cues were presented at a variable interval 200-700ms after initiation and until animals left the well, 'Well Exit'. The time from cue onset to well exit was identified as the decision time (yellow)." Yet on the figure only the period after cue presentation is in yellow. This is critical because, given the duration of the cue, the animals are probably capable of deciding (to exit the well) before the cue turns off. Indeed, as shown in fig 2D, the animals can decide within about 500 ms. So to what extent is the 'cue response' actually a 'decision response'?

We have clarified the task and spike count measurements in methods and added a revised task schematic. It is correct that the cues are available throughout the decision time (for up to 5 seconds or until well exit), and an action plan is generated before well exit/cues turn off as reflected by the separation of neural action plans (Fig 3, saline). Anterior M2 neurons maintain action plan representation from cue onset until the lever press under normal conditions (Fig 4, saline). These action plans encapsulate “cue responses” and “decision responses”. We have aligned neural data to discrete timestamps at either end of the window in which M2 processing is known to be critical, specifically between cues and actions (lever press) and focus on neural activity relative to those points. We refer to this activity throughout the manuscript as an ‘action plan’ as action planning functions of M2 activity have been well established in prior studies.

When looking at figure 3A, there is clearly a pattern on the raster, a line going from top left to bottom right. If the trials are sorted chronologically, something is happening over time. If, as I suspect, trials are sorted by ascending response time, this raster is showing that what authors are calling a 'response to cues' is actually a response around action. Basically, if propranolol slows down reaction time, the spikes will be delayed from cue onset only because they remain locked to the action. Then the whole analysis and interpretation need to be reconsidered. But it might be for the best: as I mentioned earlier, recent work on LC activity has clearly emphasized its influence on motor rather than sensory processing (Poe et al, 2020).

Figure 3A is a single neuron example, and data analyses focus on population-wide activity. Neural data is presented both aligned to cues, for all trials in which a cue was received, and aligned to lever press (action), for all trials on which a lever press occurred. In both cases, aligned to cue or aligned to action, the impact of propranolol is the same. β adrenoreceptor blockade reduces the separation of action plans in M2, severely so in females. However, a major finding is that females receive a cue but omit a large number of trials after propranolol, for this outcome the action does not occur. We propose this is due to the lack of action plan separation in anterior M2 (either directly or indirectly). When no behavioral response occurs, these trials cannot be aligned to action, yet we are still interested in the neural activity during the critical window between cue delivery and actions. We are not assigning this neural activity to sensory processing but using this discrete sensory event within our trials (cue) to align the data as there is substantial evidence that action plans in M2 arise after cue presentation in tasks such as ours where performance is guided by external cues.

Fig 2D-F: it is hard to believe that the increase in firing rate induced by propranolol in females is not significant. Presumably, because the range of the median firing rate is so high in the first place, distribution (2E) really indicates an increase in firing. Maybe some other test? e.g paired t.test, or standardized values (z.score) to get rid of variability in firing across neurons?

We agree that the session wide firing rate appears rightward shifted in females after propranolol. As our recordings were taken on different days, several days apart we cannot assume they are the same neurons for paired analyses. In our revised manuscript we evaluated these distributions using a MannWhitney test to increase power and decrease the impact of variability within the population. Previously we had used a Kolmogorov-Smirnov test. Using our new analysis, we can confirm that the propranolol significantly increases session wide firing rates in anterior M2 of females (p=0.027) but not males. This finding increases evidence for direct actions of propranolol within M2 and supports our hypothesis that propranolol leads to local disinhibition by reducing β noradrenergic signaling in interneurons and that without this noradrenergic tone anterior M2 is less efficient at suppressing irrelevant action plans.

Along those lines, would it be worth looking for effects on specific populations (interneurons) which are sometimes characterized by thinner spikes and higher mean firing rates? Given the distribution of beta receptors RNA on interneurons, one would actually expect an effect of propranolol on the firing rate irrespective of task events. Or what is it that prevents the influence of propranolol on interneurons from changing the firing rate? In any case, one of the strengths of this study is the localization of beta receptors on specific neuronal populations in the cortex, so I think that the authors should really try to build on it and find something related to the neurophysiological effects. Otherwise, one cannot exclude the possibility that the behavioral effects are not related to the influence of the drug on these receptors in that region.

Data were collected using stainless steel electrode arrays and our sample population of task related neurons is likely biased to pyramidal neurons, with a small number of fast spiking interneurons. We used validated spike waveform parameters of interneurons in premotor cortex (peak-to-trough ratio and duration; Giordano et al., 2023) in an attempt to isolate putative interneurons and found only a very small number of these cells in our recordings (n=5-7 per group). This population is too small to make any inferences about specific impacts. We have focused on the collective population activity of M2 as this is most strongly related to optimal action planning.

You are correct that from the given findings we cannot conclusively show that the results found here are a result of propranolol acting solely within anterior M2. We have made sure to clarify throughout our revised manuscript that the behavioral and physiological changes we identified are a result of collective direct and indirect actions of propranolol.

The conclusion that neuronal discrimination decreases because the proportion of neurons showing no effect increases is confusing (negative results, basically). It would be clearer if they were reporting the number of neurons that do show an effect, and presumably that this number shows a significant decrease.

The reviewer is correct that the number of neurons that do show an effect (task related activity) does significantly decrease with propranolol (from n=70 to 27 in females and n=71 to 48 in males). These n are now given adjacent to the proportions rather than at the end of the paragraph. Proportions were used for statistical analysis due to an overall decrease in the total number of units after propranolol. All PETH presented are from neurons that show some task related activity, these PETH confirm that neural activity no longer effectively discriminates/separates action plans in M2.

Figs 3F-I: a good proportion of neurons (at least 20%) show a significant encoding before cue onset. How is it possible? This raises the issue of noise level/ null hypothesis for this kind of repeated analysis. How did the author correct for multiple comparison issues?

In response to reviews, we have altered the manner in which we identify the significantly modulated neurons to increase rigor and no longer include these figures or analyses. The proportion of neurons showing action plan encoding prior to cue onset was likely an artifact of how the data was analyzed and an insufficient correction for multiple comparisons, allowing inclusion of internally generated action plans in some neurons.

The description of the action-related activity is globally confusing. Again, how can the authors discriminate between activity related to planning vs action itself? What is significant and what is not, in males vs females? What is being measured here? For example, a very unclear statement on line 238: "Propranolol primarily disrupted active inhibition of irrelevant action selection in M2 activity, reducing the ability to maintain action plan representation in M2, delaying lever press responses (Figure 4L, 4M)." What is 'active inhibition? What is an irrelevant action plan? What is selection? All of that should be defined using objective behavioral criteria and tested formally.

We have changed our wording to clarify what we are describing and why we have chosen the words we have, and to ensure consistency and objectivity throughout the manuscript. Much of the wording we have used – for example action planning or action plan selection, are the words used in the literature to describe M2 neural activity. We call the activity in M2 action planning (either externally/cue guided or internally guided) because that is what has been previously demonstrated. In our task design and analysis we are tracking cue guided actions, as opposed to internally guided.

We also separate the electrophysiology data as preferred and nonpreferred because the literature has shown individual M2 neurons show specific directional tuning as noted in our results, using the term ‘preferred’ encapsulates that tuning regardless of left/right direction. An example M2 neuron that increases activity for left cues and responses (preferred direction), will show active inhibition (low/negative z scores) on trials with right cues and responses (nonpreferred), other neurons would show the inverse relationship with direction.

A primary impact of propranolol was the loss of negative z-scores for nonpreferred trials ie neurons with a left preference that are usually inhibited on right trials were still firing and vice-versa. After propranolol neurons continue to fire for an irrelevant action plan (for the opposite direction), and the resulting population activity is not significantly different for opposing cues/responses. Behavioral responses normally occur after opposing action plans have significantly separated in M2, collapsing action plans by preventing relevant signaling (Guo et al., 2017; Inagaki et al., 2018; Li et al., 2016) or facilitating irrelevant signaling as we see here with propranolol leads impairments in 2AFC performance.

Also, the description of the classifier analysis should be more thorough. Referencing the toolbox is not sufficient to understand what has been done.

We have added additional explanation in both the methods and description of the results to clarify the functions of the neural decoding box and how we are using it to evaluate information encoding within M2. We have provided detail on how the algorithm was trained, how shuffled data was generated and how we determined significance of decoding accuracy.

Measuring Beta adrenoceptors is a great idea, and the results are interesting, especially the difference between neuron types. But again, how does that fit with neurophysiological results? Note, that since this is RNA measures, it should not be phrased as 'receptors' but 'receptors RNA' throughout. One possible interpretation of these anatomical results that cannot be reconciled with physiology is that protein expression at the membrane shows a distinct pattern.

We have changed the references to β receptor expression to β receptor mRNA expression throughout the manuscript. Although mRNA provides a valuable proxy for adrenoreceptor production, as noted by the reviewer protein expression at the membrane may differ. Reliable antibodies that allow quantitative analysis of membrane bound adrenoreceoptors in situ with co-labeling of specific cell types are limited. The goal of assessing mRNA expression within M2 was to determine if the functional sex differences we identified in M2 neurophysiology when manipulating β adrenoreceptor function could be mediated by basal differences in adrenoreceptors. The causal impact of differential mRNA expression in anterior M2 was not directly tested but our findings provide preliminary evidence that adrenoreceptor regulation may differ across sexes. Our results provide a plausible avenue for differential sensitivity to β adrenoreceptor manipulation across sexes, that may also be found in other brain regions.

In conclusion, I think that this is a very interesting study and that the results are potentially relevant for a wide audience. But the paper would clearly benefit from revisions. If the authors could clearly identify a significant relationship between the action of NA on beta receptors on specific cortical neurons, at a physiological and behavioral level, that would be a seminal study. At the moment, the evidence is not convincing enough but the data suggest that it is the case.

We thank the reviewer for the kind remarks. We have undertaken a number of new analyses, refined existing analysis and clarified our claims in the manuscript to improve rigor. Collectively our data reflect that the behavioral and neural deficits after systemic propranolol are likely due to both direct and indirect actions on M2. We believe this work is compelling and that it will inform future work investigating potential sex differences in central noradrenergic anatomy and functional sex differences after perturbations of noradrenergic signaling.

Author Response

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

Reviewer #1 (Public Review):

(1) What's the rationale of trypsinizing the tissue prior to mitochondrial isolation? This is not standard for subsequent proteomics analysis. This step will inevitably cause protein loss, especially for the post mitochondrial fractions (PMF). Treating samples with 0.01ug/uL trypsin for 37oC 30 min is sufficient to partially digest a substantial portion of the proteome. If samples from different subjects were not of the same weight, then this partial digestion step may introduce artificial variability as variable proportions of proteins from different subjects would be lost during this step. In addition, the mitochondrial protein enrichment in the mito fraction, despite statistically significant, does not look striking (Figure 1E, ~30% mitochondrial proteins in the mito fraction). As a comparison, Williams et al., MCP 2018 seem to have obtained high mitochondrial protein content in the mito fraction without trpsinizing the frozen quadriceps using a similar SWATH-MS-based approach.

Trypsinisation of the tissue prior to mitochondrial isolation is based on previous work and a Nature Protocol (1, 2) which isolated mitochondria for skeletal muscle. The rationale is that it aids in mechanical homogenisation from highly fibrous tissues such as quadriceps muscle by digesting extracellular matrix proteins. The trypsin/protein ratio used to aid in this process is at least 400 times lower than the amount of trypsin used for formal proteomic tryptic digestion. Three pieces of evidence suggest this step has negligible effect on downstream proteomic analysis. First, because the trypsinisation buffer is detergent free, trypsin will only affect extracellular or exposed membrane proteins. Filtering our PMF dataset for proteins with ‘extracellular matrix’ gene ontology identifies at least 90 unique extracellular matrix proteins indicating good retention of proteins susceptible to partial digestion. Second, the trypsin dose used is 50 times lower than the concentration used for passaging cultured cells, which retain viability after trypsinisation. Third, and contrary to the point raised by the reviewer, we observe less missingness in PMF samples compared to mitochondrial samples. We thank the reviewer for bringing the Williams et al. 2018 MCP paper to our attention. We note that mitochondrial enrichment between the two papers is comparable (~2- fold). To improve clarity line 408 now reads: “Whole quadriceps muscle samples were prepared as previously described with modification (99, 100). First, tissue was snap frozen with liquid nitrogen…” and line 95 reads: “Mitochondrial proteins were defined based on their presence in MitoCarta 3.0 (24) and consistent with previous work (25) were approximately two-fold enriched in the mitochondrial fraction relative to the PMF (Fig 1E).”

(2) The authors mentioned that the proteomics data were Log2 transformed and median- normalized. Would it be possible to provide a bit more details on this? Were the subjects randomized?

Samples were randomised prior to sample processing and mass spectrometry analysis. Because of possible variation in total protein content, it is critical to normalise protein intensities between samples. Median normalisation adjusts the samples so that they have the same median, thereby accounting for technical variation. Log2 normalisation helps to achieve normal distributions, critical for many downstream statistical tests. Line 471 now reads: “…to achieve normal distributions and account for technical variation in total protein.”

(3) In Figure 1D, what were the numbers of mice the authors used for the CV comparisons in each group? Were they of similar age and sex? Were the differences in CV values statistically significant?

The mitochondrial and PMF proteomes originated from the same quadriceps sample from the same mouse, and thus the age and sex are the same across both proteomes. After quality control, we had mitochondrial proteomes for 194 mice and PMF proteomes for 215 mice. The overall CV in the mitochondrial fraction was significantly greater than in the PMF, however whether the source of this variation is biological, or the result of mitochondrial isolation is unclear and as such we have avoided making a statement within the body of the manuscript. We have now more clearly described the nature of the samples in the revised manuscript and added sample sizes to figure 1F.

(4) The authors stated in lines 155-157 that proteins negatively associated with the Matsuda index were further filtered by presence of their cis-pQTLs. Perhaps more explanations would be needed to justify this filtering criterion? Having a cis-pQTL would mean the protein abundance variation is explained by the variation in its coding gene, this however conceptually would not be relevant to its association with the Matsuda index. With the data that the authors have in hand, would it not be natural to align the Matsuda index QTL with the pQTLs (cis and trans if available), and/or to perform mediation analysis to examine causal relationships with statistical significance?

The rationale for filtering by cis-pQTL was not to study the genetics of either Matsuda or associated proteins but rather to identify proteins that were more likely to be causally associated with Matsuda Index as opposed to adaptively associated. To clarify this line 165 now reads: “Filtering based on cis-pQTL presence was based on the rationale that if genetic variation can explain protein abundance differences between mice, then we can be confident that phenotype (Matsuda Index) is not driving the observed differences and therefore the protein-phenotype associations are likely causal. Importantly, this assumption can only be made for cis-acting pQTLs.” Previous work by Matthew et al. (see https://qtlviewer.jax.org/) has demonstrated that cis-pQTL have markedly higher LOD scores than trans-pQTLs, and our own unpublished work suggests that trans-pQTLs do not reproduce well between datasets. The reviewer rightfully suggests aligning protein QTL with those for Matsuda. This is our long-term goal but to identify genome wide significant peaks associated with altered Matsuda will require many more mice than studied here.

(5) It seems a bit odd that the first half of the paper focused extensively on the authors' discoveries in the mitochondrial proteome, and how proteins involved in mitochondrial processes (such as complex I) were associated with Matsuda Index, but the final fingerprint list of insulin resistance, which contained 76 proteins, only had 7 mitochondrial proteins. Was this because many mitochondrial proteins were filtered out due to no cis-pQTL presenting?

There are three reasons our fingerprint is lacking mitochondrial proteins: 1) there are more non-mitochondrial than mitochondrial proteins in the muscle proteome; 2) we focussed on negatively associated proteins, and as demonstrated in figure 2c, the mitochondrial proteome is enriched for positively associated proteins; 3) as implied by the reviewer, we filtered for pQTL presence, further reducing the number of mitochondrial proteins in our fingerprint. To improve clarity, line 170 now reads: “Low mitochondrial representation in the fingerprint is the result of selecting negatively associating proteins, and as seen (Figure 2C) previously, the mitochondrial proteome is enriched for positive contributors to insulin resistance.”

(6) The authors found that thiostrepton-induced insulin resistance reversal effects were not through insulin signalling. It activated glycolysis but the mechanism of action was not clear. What are the proteins in the fingerprint list that led to identification of thiostrepton on CMAP?

Is thiostrepton able to bind or change the expression of these proteins? Since thiostrepton was identified by searching the insulin resistance fingerprint protein list against CMAP, it would be rational to think that it exerts the biological effects by directly or indirectly acting on these protein targets.

This is indeed the implication of our data. Because of the timescales involved it is unlikely that thiostrepton is changing fingerprint protein levels but could be binding to and inhibiting them. Searching the CMAP thiostrepton signature reveals ARHGDIB and NAGK as the fingerprint proteins with the most positive and negative fold-changes respectively perhaps suggesting they play a role in thiostrepton’s mechanism of action. Experiments are underway to test this hypothesis however these are beyond the scope of the current paper.

Reviewer #2 (Public Review):

Line 105: The observation that variance in respiratory proteins is stable while lipid pathways is variable is quite interesting. Is this due to lower overall levels of lipid metabolism enzymes (ex. do these differ substantially from similar pathways ranked from high-low abundance?).

The relationship between coefficient of variation (CV) and relative abundance of proteins is important to consider. To address this, we have now also performed GSEA on proteins ranked from high to low relative abundance. These comparisons have been added to supplementary figure 1 and line 110 now reads: “As a control experiment, we also performed enrichment analysis on proteins ranked by LFQ relative abundance. High CV pathways (enriched for high CV proteins) tended to be lower in relative abundance (enriched for low relative abundance proteins) (Supplementary Fig 1a, b). However, many high variability pathways, lipid metabolism for example, were not enriched in either direction based on relative abundance suggesting differences in relative abundance do not fully explain pathway variability differences.”

Line 154: the 664 associations are impressive and potentially informative. It would be valuable to know which of these co-map to the same locus - either to distinguish linkage in a 2mb window or identify any cis-proteins which directly exert effects in trans-

To assess this, we have analysed pQTL position relative to gene position to generate a ‘hotspot’ plot. We have also generated a histogram of this pQTL density (in a 2 Mbp window) and added these figures to figure 3. We did not detect any obvious pQTL hotspots, and the distribution of pQTLs across the genome appears fairly uniform. Line 159 now reads: “These were distributed across the genome and were predominately cis acting (Figure 3A)...”

Line 194: Cross-platform validation of the CMAP fingerprint results is an admirable set of validations. It might be good to know general parameters like how many compounds were shared/unique for each platform. Also the concordance between ranking scores for significant and shared compounds.

The Connectivity Map (CMap) query included 5163 compounds, the Prestwick library included 1120, and the overlap was 420. We have added these comparisons to supplementary figure 2. Supplementary figure 2 now also contains a comparison of CMap scores between overlapping compounds (found in CMap and the Prestwick library) against all significant compounds identified by CMap (supplementary figure 2b). Interestingly, compounds present in both platforms scored higher on average, suggesting the Prestwick library captures a significant proportion of highly scoring CMap candidates. Line 206 now reads: “In total, 420 compounds were found across both platforms, and these consensus compounds captured a significant proportion of highly scoring CMap compounds (Supplementary Figure 2A, B).”

Line 319: Another consideration in the molecular fingerprint is how unique these are for muscle. While studies evaluating gene expression have shown that many cis-eQTLs are shared across tissues, to my knowledge, this hasn't been performed systematically for pQTLs. Therefore, consider adding a point to the discussion pointing out that some of the proteins might be conserved pQTLs whereas others which would be more relevant here present unique druggable targets in muscle.

To examine tissue specificity, we determined whether our skeletal muscle fingerprint proteins were detected and contained a pQTL in two metabolically important tissues, liver and adipose. Despite detecting almost all the fingerprint proteins in both adipose and liver tissue, they were depleted for pQTL compared to skeletal muscle. These data have now been added to figure 3c. Line 172 now reads: “To assess the tissue specificity of our fingerprint we searched for the same proteins in metabolically important adipose and liver tissues. Despite detecting 94% and 82% of muscle fingerprint proteins across each tissue respectively, both adipose and liver were depleted for pQTL presence (Figure 3C) suggesting that regulation of our fingerprint protein abundance is specific to skeletal muscle.”

Line 332: These are fascinating observations. 1, that in general insulin signaling and ampk were not themselves shown as top-ranked enrichments with matsuda and that this was sufficient to alter glucose metabolism without changes in these pathways. While further characterization of this signaling mechanism is beyond the scope of this study, it would be good to speculate as to additional signaling pathways that are relevant beyond ROS (ex. CNYP2 and others)

We have now added further discussion to the manuscript to address this point., Line 347 now reads: “Aside from glycolysis, other pathways may be involved in enhancing insulin sensitivity. For example, the negatively associated protein ARHGDIA (Figure 2F) is a potent negative regulator of insulin sensitivity, and our fingerprint of insulin resistance contained its homologue ARHGDIB. Both ARHGDIA and ARHGDIB have been reported to inhibit the insulin action regulator RAC1 thus lowering GLUT4 translocation and glucose uptake. Further investigations may uncover a role for thiostrepton in modulating the RAC1 signalling pathway via ARHGDIB.”

Line: 314: Remove the statement: "While this approach is less powerful than QTL co- localisation for identifying causal drivers,", as I don't believe that this has been demonstrated. Clearly, the authors provide a sufficient framework to pinpoint causality and produce an actionable set of proteins.

We have edited line 314, which now reads: “Moreover, our approach has the major advantage that it requires far fewer mice to obtain meaningful outcomes (222 mice in this study) compared to that required for genetic mapping of complex traits like Matsuda Index.”

Line 346: I would highlight one more appeal of the approach adopted by the authors. Given that these compound libraries were prioritized from patterns of diverse genetics, these observations are inherently more-likely to operate robustly across target backgrounds.

This point is further supported by our thiostrepton results in both C57BL6/j and BXH9 mice. Line 317 now reads: “Furthermore, because we have used genetically diverse datasets (DOz mice and multiple cell lines in Connectivity Map) our findings are likely robust across diverse target backgrounds.”

Line 434: I might have missed but can't seem to find where the muscle data are available to researchers. Given the importance and novelty of these studies, it will be important to provide some way to access the proteomic data.

These data are now available via the ProteomeXchange Consortium. Line 465 now reads: “The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (104) partner repository with the dataset identifier PXD042277.”

1. Frezza C, Cipolat S, Scorrano L. Organelle isolation: functional mitochondria from mouse liver, muscle and cultured filroblasts. Nat Protoc. 2007;2(2):287-95.

2. Acin-Perez R, Benador IY, Petcherski A, Veliova M, Benavides GA, Lagarrigue S, et al. A novel approach to measure mitochondrial respiration in frozen biological samples. The EMBO Journal. 2020;39(13):e104073.

3. Chick JM, Munger SC, Simecek P, Huttlin EL, Choi K, Gatti DM, et al. Defining the consequences of genetic variation on a proteome-wide scale. Nature. 2016;534(7608):500- 5.

4. Gatti DM, Svenson KL, Shabalin A, Wu L-Y, Valdar W, Simecek P, et al. Quantitative Trait Locus Mapping Methods for Diversity Outbred Mice. G3 Genes|Genomes|Genetics. 2014;4(9):1623-33.

Author Response

Reviewer #1 (Public Review):

In this study, the authors set out to investigate spatial RNA processing events, specifically alternative splicing and 3' UTR usage, in mouse brain and kidney tissues using ReadZS and SpliZ methodologies on spatial transcriptomics data. The research contributes to understanding tissue-specific gene expression regulation from a spatial perspective. The study introduces a novel approach for analyzing spatial transcriptomics data, allowing for the identification of RNA processing and regulation patterns directly from 10X Visium data. The authors present convincing evidence supporting the identification of novel RNA processing patterns using their methodology, which holds significant implications for researchers in the field of spatial transcriptomics and the study of alternative splicing and 3' UTR usage.

Thank you for this thorough overview of our work.

The conclusions of the study are mostly well-supported by the data; however, certain aspects could be improved to strengthen the findings.

1) The conclusions of this study would be strengthened by conducting a more extensive tissue sample analysis and including biological replicates. Additionally, appropriate batch effect corrections should be applied when dealing with biological replicates.

We agree that including biological replicates would strengthen our findings. We will include biological replicates of the mouse brain tissues in the revision.

2) The 3' UTR usage and alternative splicing should be compared among clearly labeled clusters for a more comprehensive analysis.

We understand that it can be difficult to see how the SpliZ quantiles map spatially onto the tissue images. For the splicing of Gng13, Myl6, and Rps24, we will include box plots broken down by spatial quadrant in the revision. However, this does result in an oversimplification of the spatial patterns found in the tissue slices, which make the plots less informative than the quantile plots to our view.

3) The authors should clarify their rationale for choosing ReadZS and SpliZ approaches and provide comparisons with other methods to demonstrate the advantages and potential limitations of their chosen methodologies.

Thank you for pointing out the lack of sufficient discussion of ReadZS and SpliZ in the manuscript. The ReadZS and SpliZ were chosen for this analysis because both of these methods provide an individual score for each cell-gene pair, which is easily adapted to providing a score for each spot-gene pair. Due to the sparsity and 3’ bias of Visium data, approaches designed to analyze RNA processing in full-length sequencing analysis are not applicable. The SpliZ and ReadZS are two of the limited number of tools available that are designed for the analysis of RNA processing in droplet-based data. Other available tools tend to rely on aggregating data across multiple cells using a method called pseudo-bulking (Li et al., 2021; Patrick et al., 2020). It is not clear how this could be used for spatial transcriptomics data without potentially obscuring subtle spatial patterns in the data. Others are based on PSI measurements, which are vulnerable to artifacts due to sparsity (Buen Abad Najar et al., 2020; Olivieri et al., 2022; Wen et al., 2022). The tradeoff between pseudo-bulking and a single score per spot-gene pair means that the ReadZS and SpliZ do not have the power to detect changes for genes with very low read counts. We will add text in the revision to clarify this point.

Reviewer #2 (Public Review):

The authors applied existing ReadZS and the SpliZ methods, previously developed to analyze RNA process in scRNA-seq data, to Visium data to study spatial splicing and RNA processing events in tissues by Moran's I. The authors showed several example genes in mouse brain and kidney, whose processing are spatially regulated, such as Rps24, Myl6, Gng13.

Thank you for this thorough overview of our work.

The paper touches on an important question in RNA biology about how RNA processing is regulated spatially. Both experimental and computational challenges remain to address it. Despite some potentially interesting findings, most claims remain to be validated by orthogonal methods such as RNA FISH and simulations.

We appreciate that the reviewer finds the question important, and that the findings are potentially interesting. In the revision we will include biological replicates for our findings in the mouse brain. Unfortunately, experimental validation is outside of our budget for this project. It is unclear what further simulations could validate the biological discoveries in this manuscript: permutations were used to calculate the p value of each discovery, and the false positive and negative rates of the SpliZ have been assessed through simulation (Olivieri et al., 2022).

In addition, the percentage of spatial processing events (splicing in 0.8-2.2% of detected genes, i.e. 8-17 genes and RNA processing in 1.1-5.5% of detected genomic windows, i.e. 57-161 windows) discovered is low. Does it suggest that most of RNA processing events were not spatially regulated across the tissue? Or does it question the assumption of treating spatial transcriptomics data similar to scRNA-seq data?

We agree that the question of the prevalence of spatial RNA processing regulation is critical. Rather than the two options proposed here, we believe that the sparsity of the data limits our ability to call more of these events. In the revision, we will provide a supplemental figure showing the relationship between read depth and p value for each gene to quantify how the fraction of observed regulation changes with sequencing depth. It is worth noting that as these technologies improve, we expect the sequencing depth of spatial technologies to increase which would likely result in more discoveries.

The unique features for ST data, such as mixture of neighboring cells, different capture biases and much smaller number of spots (pseudo cells here), may have significant effects on the power of scRNA-seq based methods, but it is not discussed in the manuscript. The lack of careful evaluation and low discovery rates could limit application of the approach to other tissues and subcellular data.

We appreciate the concern that technical differences between scRNA-seq data and spatial transcriptomics data could affect our results. We agree that this point could be addressed more thoroughly in the text. None of the specificities of spatial transcriptomics data invalidate the assumptions of the SpliZ or ReadZS. The method we use to identify genes with significant spatial regulation of RNA processing was specifically created to be used for Visium data. It takes into account mixture of RNAs in neighboring cells by randomly sampling scores of neighboring cells, rather than randomization of the location of the spots themselves, which does indeed result in a high false positive rate (see “Permutations for Moran’s I” in the Methods). We do note that there is a limit to the power of this kind of analysis based on the number of spots and the read depth, which we will quantify in a plot in the revision.

Author Response:

We thank Reviewer #1 for their positive assessment of our work.

Reviewer #2 (Public Review):

[…] Although these results confirm what we already know about processes involved in the meninges in MS and its models and gradients of pathology in sub-pial regions, this is the first to use spatial transcriptomics to demonstrate such gradients at a molecular level in an animal model that demonstrates lymphoid like tissue development in the meninges and associated grey matter pathology. The mouse EAE model being used here does reproduce many, although not all, of the pathological features of MS and the ability to look at longer time points has been exploited well. However, this particular spatial transcriptomics technique cannot resolve at a cellular level and therefore there is a lot of overlap between gene expression signatures in the meninges and the underlying grey matter parenchyma.

We appreciate the reviewer’s concise summary and comments on our manuscript. We agree that the Visium spatial sequencing technology we applied is limited in its resolution and cannot precisely distinguish individual cells or anatomic regions. For that reason, there is undoubtedly some overlap between gene expression signatures in the meninges and underlying parenchyma, particularly in spots on the borders of the meningeal inflammation clusters. However, we believe that the majority of meningeal inflammation (“cluster 11”) spots are indeed in the meninges and represent the spatial transcriptome of that niche. To support this, in the revised manuscript we will provide H&E images with the UMAP clusters overlayed to demonstrate the anatomic borders that correlate with the clusters.

The short nature of this report means that the results are presented and discussed in a vague way, without enough molecular detail to reveal much information about molecular pathogenetic mechanisms.

We thank the reviewer for this comment. The goal of this work is to transcriptomically characterize the spatial relationship between areas of meningeal inflammation and the underlying parenchyma. While we agree that mechanistic studies are needed to further evaluate the role of presented signaling pathways, those experiments are beyond the scope of this brief report.

The trajectory analysis is a good way to explore gradients within the tissues and the authors are to be applauded for using this approach. However, the trajectory analysis does not tell us much if you only choose 2 genes that you think might be involved in the pathogenetic processes going on in the grey matter. It might be more useful to choose some genes involved in pathogenetic processes that we already know are involved in the tissue damage in the underlying grey matter in MS, for which there is already a lot of literature, or genes that respond to molecules we know are increased in MS CSF, although the animal models may be very different. Why were C3 and B2m chosen here?

We appreciate the reviewer’s points here. C3 and B2m were chosen as examples of genes that have differential fit to the gradient descending pattern to assist the reader in interpreting subsequent gene set trajectory analysis. However, we agree that there are many other genes of interest and will expand the number of genes displayed in our revised manuscript. 

Strengths: <br /> - The mouse model does exhibit many of the features of the compartmentalized immune response seen in MS, including the presence of meningeal immune cell infiltrates in the central sulcus and over the surface of the cortex, with the presence of FDC's HEVs PNAd+ vessels and CXCL13 expression, indicating the formation of lymphoid like cell aggregates. In addition, disruption of the glia limitans is seen, as in MS. Increased microglial reactivity is also present at the pial surface. <br /> - Spatial transcriptomics is the best approach to studying gradients in gene expression in both white matter and grey matter and their relationship between compartments. <br /> - It would be useful to have more discussion of how the upregulated pathways in the two .compartments fit with what we know about the cellular changes occurring in both, for which presumably there is prior information from the group's previous publications.

Limitations: <br /> - EAE in the mouse is not MS and may be far removed when one considers molecular mechanisms, especially as MS is not a simple anti-myelin protein autoimmune condition. Therefore, this study could be following gene trajectories that do not exist in MS. This needs a significant amount of discussion in the manuscript if the authors suggest that it is mimicking MS. <br /> - The model does not have the cortical subpial demyelination typical of MS and it is unknown whether neuronal loss occurs in this model, which is the main feature of cytokine-mediated neurodegeneration in MS. If it does not then a whole set of genes will be missing that are involved in the neuronal response to inflammatory stimuli that may be cytotoxic. <br /> - Visium technology does not get down to single cell level and does not appear to allow resolution of the border between the meninges and the underlying grey matter. <br /> - Neuronal loss in the MS cortex is independent of demyelination and therefore not related to remyelination failure. There does not appear to be any cortical grey matter demyelination in these animals, so it is difficult to relate any of the gene changes seen here to demyelination. <br /> - No mention of how the ascending and descending patterns of gene expression may be due to the gradient of microglial activation that underlies meningeal inflammation, which is a big omission.

We thank the reviewer for their insightful comments on the strengths and limitations of our study. Regarding the SJL EAE model we use in this paper, it certainly is not a perfect model of meningeal inflammation in MS, indeed we believe that no such animal model exists, but it does recapitulate several key features of human disease as described by the reviewer. Spatial transcriptomics of cortical grey matter lesions and overlying meninges of samples derived from patients with MS would be ideal, though access to this tissue is highly limited. In the revised manuscript we will include more detailed discussion of the limitations in applying these findings to MS. However, in addition to potential implications for MS research, our data contribute more generally to understanding of meningeal inflammation and penetrance of inflammation into brain tissue.

We acknowledge that sub-pial neuronal loss has not been assessed in SJL EAE, and if present it would increase the relevance of this model to neurodegeneration. We are currently working to assess this.

We agree with the reviewer that Visium technology is limited in its ability to discriminate individual cells, as discussed above (2.2).

We agree that gene expression by activated microglia is likely a major driver of the transcriptomic changes observed in the parenchyma, and thank the reviewer for highlighting this. We will add discussion of this to our revised manuscript, and intend to generate additional data regarding the contribution of subpial microglial activation to the measured transcriptomic changes.

Finally, we thank Reviewer #3 for their assessment of our work.

Author Response

eLife assessment:

Trypanosoma brucei evades mammalian humoral immunity through the expression of different variant surface glycoprotein genes. In this fundamental paper, the authors extend previous observations that TbRAP1 both interacts with PIP5pase and binds PI(3,4,5)P3, indicating a role for PI(3,4,5)P3 binding and suggesting that antigen switching is signal dependent. While much of the evidence is compelling, one reviewer suggested that the work would benefit from further controls.

We appreciate the evaluation of the work and agree that the findings substantially advance our understanding of antigenic variation. A detailed response to the public review is included below, which addresses and clarifies the issues raised by the reviewers, including those concerning controls. We also want to highlight the comment by Reviewer #3 “The methods used in the study are rigorous and well-controlled…. their results support the conclusions made in the manuscript.”. We hope this and our comments will help address the issue of controls in this eLife statement.

Reviewer #1 (Public Review):

Trypanosoma brucei undergoes antigenic variation to evade the mammalian host’s immune response. To achieve this, T. brucei regularly expresses different VSGs as its major surface antigen. VSG expression sites are exclusively subtelomeric, and VSG transcription by RNA polymerase I is strictly monoallelic. It has been shown that T. brucei RAP1, a telomeric protein, and the phosphoinositol pathway are essential for VSG monoallelic expression. In previous studies, Cestari et al. (ref. 24) have shown that PIP5pase interacts with RAP1 and that RAP1 binds PI(3,4,5)P3. RNAseq and ChIPseq analyses have been performed previously in PIP5pase conditional knockout cells, too (ref. 24). In the current study, Touray et al. did similar analyses except that catalytic dead PIP5pase mutant was used and the DNA and PI(3,4,5)P3 binding activities of RAP1 fragments were examined. Specifically, the authors examined the transcriptome profile and did RAP1 ChIPseq in PIP5pase catalytic dead mutant. The authors also expressed several C-terminal His6-tagged RAP1 recombinant proteins (full-length, aa1-300, aa301-560, and aa 561-855). These fragments’ DNA binding activities were examined by EMSA analysis and their phosphoinositides binding activities were examined by affinity pulldown of biotin-conjugated phosphoinositides. As a result, the authors confirmed that VSG silencing (both BES-linked and MES-linked VSGs) depends on PIP5pase catalytic activity, but the overall knowledge improvement is incremental. The most convincing data come from the phosphoinositide binding assay as it clearly shows that N-terminus of RAP1 binds PI(3,4,5)P3 but not PI(4,5)P2, although this is only assayed in vitro, while the in vivo binding of full-length RAP1 to PI(3,4,5)P3 has been previously published by Cestari et al (ref. 24) already. Considering that many phosphoinositides exert their regulatory role by modulating the subcellular localization of their bound proteins, it is reasonable to hypothesize that binding to PI(3,4,5)P3 can remove RAP1 from the chromatin. However, no convincing data have been shown to support the author’s hypothesis that this regulation is through an “allosteric switch”. Therefore, the title should be revised.

We appreciate the reviewer’s detailed evaluation of our work. There are a few general comments that we would like to clarify. We will break them into three points. All data included here are new and were not previously published.

i) “RNAseq and ChIPseq analyses have been performed previously …(ref. 24).” Reference 24 is Cestari et al. 2019, Mol Cell Biol. We, or others, have not published ChIP-seq of RAP1 in T. brucei. Previous work showed ChIP-qPCR, which analyses specific loci. The ChIP-seq shows genome-wide binding sites of RAP1, and new findings are shown here, including binding sites in the BES, MESs, and other genome loci such as centromeres. We also identified DNA sequence bias defining RAP1 binding sites (Fig 2A). We also show by ChIP-seq how RAP1-binding to these loci changes upon expression of catalytic inactive PIP5Pase. As for the RNA-seq, this is also the first time we show RNA-seq of T. brucei expressing catalytic inactive PIP5Pase, which establishes that the regulation of VSG silencing and switching is dependent on PIP5Pase enzyme catalysis, i.e., PI(3,4,5)P3 dephosphorylation. To improve clarity in the manuscript, we edited page 4, line 122, as follows: “We showed that RAP1 binds telomeric or 70 bp repeats (24), but it is unknown if it binds to other ES sequences or genomic loci.”

ii) “The in vivo binding of full-length RAP1 to PI(3,4,5)P3 has been previously published by Cestari et al. (ref. 24) already.”. We published in reference 24 that RAP1-HA can bind agarose beads-conjugated synthetic PI(3,4,5)P3. Here, we were able to measure T. brucei endogenous PI(3,4,5)P3 associated with RAP1-HA (Fig 4F). Moreover, we showed that the endogenous RAP1-HA and PI(3,4,5)P3 binding is about 100-fold higher when PIP5Pase is catalytic inactive than WT PIP5Pase. The data establish that in vivo endogenous PI(3,4,5)P3 binds to RAP1-HA and how the binding changes in cells expressing mutant PIP5Pase; this data is new and relevant to our conclusions.

iii) “no convincing data have been shown to support the author’s hypothesis that this regulation is through an “allosteric switch””. We show here in vitro and in vivo data supporting the conclusion. We show that PI(3,4,5)P3 binds to the N-terminus of rRAP1-His with a calculated Kd of about 20 µM (Fig 4B-E, Table 1). In contrast, we show by EMSA and binding kinetics by microscale thermophoresis that rRAP1-His binds to 70 bp and telomeric repeats via protein regions encompassing the Myb (central) or Myb-L domains (C-terminal) but not the N-terminus containing the VHP domain (Fig 3C-G, and Fig S5). Using microscale thermophoresis, we also show that rRAP1-His binds to 70 bp and telomeric repeats with Kd of 10 and 24 nM, respectively (Fig 3 and Table 1). Notably, we show that 30 µM of PI(3,4,5)P3, but not PI(4,5,)P2 – used as a control – disrupts rRAP1-His binding to 70 bp and telomeric repeats, changing Kds to about 188 and 155 nM, respectively (Fig 5A-C). We also show that PI(3,4,5)P3 does not disrupt the binding of rRAP1-His fragments (Myb or MybL) without the N-terminus domain (Fig S5), implying binding of PI(3,4,5)P3 to RAP1 N-terminus is required for displacement of RAP1 DNA binding domains (Myb and MybL) from telomeric and 70 bp repeats, and that PI(3,4,5)P3 is not competing for Myb or Myb-L binding to DNA. Moreover, we show that RAP1-HA binding to 70 bp and telomeric repeats in vivo is displaced in T. brucei cells expressing catalytic inactive PIP5Pase (Fig 5D-G), which we show results in RAP1-HA binding about 100-fold more endogenous PI(3,4,5)P3 than in T. brucei expressing WT PIP5Pase (Fig 4F). The in vivo data agrees with the in vitro data. The data show a typical allosteric regulator system, in which binding of a ligand to one site of the protein, here PI(3,4,5)P3 binding to RAP1 N-terminus, affects other domains (RAP1 Myb and Myb-L domains) binding to DNA. To improve the clarity of the title, we will change it in the revised version to imply a direct role of PI(3,4,5)P3 regulation of RAP1 in the process. This will provide more specific information to the readers and addresses the concern of the reviewer related to the “allosteric switch”. The new title will be: PI(3,4,5)P3 allosteric regulation of RAP1 controls antigenic switching in trypanosomes

There are serious concerns about many conclusions made by Touray et al., according to their experimental approaches:

1) The authors have been studying RAP1’s chromatin association pattern by ChIPseq in cells expressing a C-terminal HA tagged RAP1. According to data from tryptag.org, RAP1 with an N-terminal or a C-terminal tag does not seem to have identical subcellular localization patterns, suggesting that adding tags at different positions of RAP1 may affect its function. It is therefore essential to validate that the C-terminally HA-tagged RAP1 still has its essential functions. However, this data is not available in the current study. RAP1 is essential. If RAP1-HA still retains its essential functions, cells carrying one RAP1-HA allele and one deleted allele are expected to grow the same as WT cells. In addition, these cells should have the WT VSG expression pattern, and RAP1-HA should still interact with TRF. Without these validations, it is impossible to judge whether the ChIPseq data obtained on RAP1-HA reflect the true chromatin association profile of RAP1.

Tryptag data show both N- and C-terminus RAP1 with nuclear localization in procyclic forms, although there are differences in signal intensities in the images (http://tryptag.org/?id=Tb927.11.370). It is important to note that Tryptag data is from procyclic forms, and DNA constructs are not validated for their integration in the correct locus. As for the RAP1-HA localization in bloodstream forms, we demonstrated that C-terminally HA-tagged RAP1 co-localizes with telomeres by a combination of immunofluorescence and fluorescence in situ hybridization (Cestari and Stuart, 2015, PNAS), and RAP1-HA co-immunoprecipitate telomeric and 70 bp repeats (Cestari et al. 2019 Mol Cell Biol). We also showed by immunoprecipitation and mass spectrometry that HA-tagged RAP1 interacts with nuclear and telomeric proteins, including PIP5Pase (Cestari et al. 2019). Others have also tagged T. brucei RAP1 in bloodstream forms with HA without disrupting its nuclear localization (Yang et al. 2009, Cell; Afrin et al. 2020, Science Advances). As for the experiment suggested by the reviewer, there is no guarantee that cells lacking one allele of RAP1 will behave as wildtype, i.e., normal growth and repression of VSGs genes. Also, less than 90% of T. brucei TRF was reported to interact with RAP1 (Yang et al. 2009, Cell), which might be indirect via their binding to telomeric DNA repeats rather than direct protein-protein interactions.

2) Touray et al. expressed and purified His6-tagged recombinant RAP1 fragments from E. coli and used these recombinant proteins for EMSA analysis: The His6 tag has been used for purifying various recombinant proteins. It is most likely that the His6 tag itself does not convey any DNA binding activities. However, using His6-tagged RAP1 fragments for EMSA analysis has a serious concern. It has been shown that His6-tagged human RAP1 protein can bind dsDNA, but hRAP1 without the His6 tag does not. It is possible that RAP1 proteins in combination with the His6 tag can exhibit certain unnatural DNA binding activities. To be rigorous, the authors need to remove the His6 tag from their recombinant proteins before the in vitro DNA binding analyses are performed. This is a standard procedure for many in vitro assays using recombinant proteins.

We show in Fig 3C-G that His-tagged full-length rRAP1 does not bind to scrambled telomeric dsDNA sequences, which indicates that His-tagged rRAP1 does not bind unspecifically to DNA. Moreover, in Fig 3G, we show that His-tagged rRAP11-300 also does not bind to 70 bp or telomeric repeats. In contrast, full-length His-tagged rRAP1, rRAP1301-560, or rRAP1561-855 bind to 70 bp or telomeric repeats (Fig 3C-G). Since all proteins were His-tagged, the His tag cannot be responsible for the DNA binding.

As for the statement that human rRAP1-His has unspecific DNA binding properties, we could not find a reference to this statement; we cannot compare it without knowing the details of the experiment. Biochemical assays can result in unspecific binding depending on binding/buffer conditions. Also, humans and T. brucei RAP1 share only 15% of amino acid identity; unspecific binding to DNA could be specific to human RAP1.

3) It is unclear why Nanopore sequencing was used for RNAseq and ChIPseq experiments. The greatest benefit of Nanopore sequencing is that it can sequence long reads, which usually helps with mapping, particularly at genome loci with repetitive sequences. This seems beneficial for RAP1 ChIPseq analysis as RAP1 is expected to bind telomere repeats. However, for ChIPseq, the chromatin needs to be fragmented. Larger DNA fragments from ChIPseq experiments will decrease the accuracy of the final calculated binding sites. Therefore, ChIPseq experiments are not supposed to have long reads to start with, so Nanopore sequencing does not seem to bring any advantage. In addition, compared to Illumina sequencing, Nanopore sequencing usually yields smaller numbers of reads, and the sequencing accuracy rate is lower. The Nanopore sequencing accuracy may be a serious concern in the current study. All telomeres have the perfect TTAGGG repeats, all VSG genes have a very similar 3’ UTR, and all 70 bp repeats have very similar sequences. In fact, the active and silent ESs have 90% sequence identity. Are sequence reads accurately mapped to different ESs? How is the sequencing and mapping quality controlled? Furthermore, it is unclear whether the read depth for RNAseq is deep enough.

The mean sequence length for the ChIP-seq was about 500 bp (see Table S3), which helps to align reads to ESs and distinguish the different ESs, and it is a reasonable size range to define RAP1 binding sites. Although sequencing depths are usually higher in Illumina than in nanopore (all depending on the amount of sequencing), most Illumina short reads map to multiple genomic sequences, making it difficult to distinguish ESs. This is particularly important for RAP1 because it binds to repeats such as 70 bp and telomeric repeats. Mapping short reads to those regions would be virtually impossible; hence, our choice of nanopore sequencing. For RNA-seq, the ~500 bp read length help sequence alignment to the subtelomeric regions containing many VSG genes. The nanopore reads obtained here had an average sequencing score 12 (i.e., base call accuracy of 94%). Filtering reads with MAPQ ≥ 20 (99% probability of correct alignment) helped us to distinguish RAP1 binding to specific ESs, including silent vs active ES (ChIP-seq) or VSG sequences (RNA-seq). The details of the analysis and sequencing metrics (i.e., sequencing depth and read length) were described in the Methods section “Computational analysis of RNA-seq and ChIP-seq” and Table S3, respectively.

4) Many statements in the discussion section are speculations without any solid evidence. For example, lines 218 - 219 “likely due to RAP1 conformational changes”, no data have been shown to support this at all. In lines 224-226, the authors acknowledged that more experiments are necessary to validate their observations, so it is important for the authors to first validate their findings before they draw any solid conclusions. Importantly, RAP1 has been shown to help compact telomeric and subtelomeric chromatin a long time ago by Pandya et al. (2013. NAR 41:7673), who actually examined the chromatin structure by MNase digestion and FAIRE. The authors should acknowledge previous findings. In addition, the authors need to revise the discussion to clearly indicate what they “speculate” rather than make statements as if it is a solid conclusion.

The statement “likely due to RAP1 conformational changes” in lines 218-219 (page 6) is part of the Discussion. We did not make a strong statement but discussed a possibility. We believe that it is beneficial to the reader to have the data discussed, and we do not feel this point is overly speculative.

For lines 224-226 (page 6), the statement refers to the finding of RAP1 binding to centromeric regions by ChIP-seq, which is a new finding but not the focus of this work. Hence, future studies are necessary for this finding, and we believe it is appropriate in the Discussion to be upfront and highlight this point to the readers. However, for the RAP1 binding to telomeric ES sites, e.g., 70 bp repeats and telomeric repeats (the focus of this work), we validated the binding by EMSA and by performing binding kinetics using microscale thermophoresis.

We did not include Pandya et al. 2013 NAR because the authors demonstrated RAP1 compaction of chromatin to occur in procyclic forms only. Pandya et al. stated in their abstract: “no significant chromatin structure changes were detected on depletion of TbRAP1 in BF cells”. Hence, the suggested reference is not relevant to the context of our conclusions in bloodstream forms. Nevertheless, we have reviewed the Discussion to avoid broad speculations in the revised version of the manuscript.

There are also minor concerns:

1) In the PIP5Pase conditional knockout system, the WT or mutant PIP5Pase with a V5 tag is constitutively expressed from the tubulin array. What’s the relative expression level of this allele and the endogenous PIP5Pase? Without a clear knowledge of the mutant expression level, it is hard to conclude whether the mutant has any dominant negative effects or whether the mutant phenotype is simply due to a lower than WT PIP5pase expression level.

The relative mRNA levels of the exclusive expression of PIP5Pase Mut compared to the WT is available in the Data S1, RNA-seq. The Mut allele’s relative expression level is 0.85-fold to the WT allele (both from tubulin loci). We also showed by Western blot the WT and Mut PIP5Pase protein expression (Cestari et al. 2019, Mol Cell Biol). Concerning PIP5Pase endogenous alleles, we compared RNA-seq reads counts per million from the conditional null PIP5Pase cells exclusively expressing WT or the Mut PIP5Pase alleles (Data S1, this work) to our previous RNA-seq of single-marker 427 strain (Cestari et al. 2019, Mol Cell Biol). We used the single-maker 427 because the conditional null cells were generated in this strain background. The PIP5Pase WT and Mut mRNAs expressed from tubulin loci are 1.6 and 1.3-fold the endogenous PIP5Pase levels in single-marker 427, respectively. We include a statement in the Methods, page 7, lines 265-268: “The WT or Mut PIP5Pase mRNAs exclusively expressed from tubulin loci are 1.6 and 1.3-fold the WT PIP5Pase mRNA levels expressed from endogenous alleles in the single marker 427 strain. The fold-changes were calculated from RNA-seq reads counts per million from this work (WT and Mut PIP5Pase, Data S1) and our previous RNA-seq from single marker 427 strain (24).”

2) In EMSA analysis, what are the concentrations of the protein and the probe used in each reaction? The amount of protein used in the binding assay appears to be very high, and this can contribute to the observation that many complexes are stuck in the well. Better quality EMSA data need to be shown to support the authors’ claims.

All concentrations were provided in the Methods section. See page 9 Electrophoretic mobility shift assays: “100 nM of annealed DNA were mixed with 1 μg of recombinant protein…”. For microscale thermophoresis, also see page 9, Microscale thermophoresis binding kinetics: “1 μM rRAP1 was diluted in 16 two-fold serial dilutions in 250 mM HEPES pH 7.4, 25 mM MgCl2, 500 mM NaCl, and 0.25% (v/v) N P-40 and incubated with 20 nM telomeric or 70 bp repeats…”. Note that two different biochemical approaches, EMSA and microscale thermophoresis, were used to assess rRAP1-His binding to DNA. Both show similar results (Fig 3 and 5, and Fig S5; microscale thermophoresis shows the binding kinetics, data available in Table 1). The EMSA images clearly show the binding of RAP1 to 70 bp or telomeric repeats but not to scramble telomeric repeat DNA.

Reviewer #2 (Public Review):

This manuscript by Touray, et al. provides a significant new twist to our understanding of how antigenic variation may be regulated in T. brucei. Key aspects of antigenic variation are the mutually exclusive expression of a single antigen per cell and the periodic switching from expression of one antigen isoform to another. In this manuscript, the authors show, as they have previously shown, that depletion of the nuclear phosphatidylinositol 5-phosphatase (PIP5Pase) results in a loss of mutually exclusive VSG expression. Furthermore, using ChIP-seq, the authors show that the repressor/activator protein 1 (RAP1) binds to regions upstream and downstream of VSG genes located in transcriptionally repressed expression sites and that this binding is lost in the absence of a functional PIP5Pase. Importantly, the authors decided to further investigate this link between PIP5Pase and RAP1, a protein that has previously been implicated in antigenic variation in T. brucei, and found that inactivation of PIP5Pase results in the accumulation of PI(3,4,5)P3 bound to the RAP1 N-terminus and that this binding impairs the ability of RAP1 to bind DNA. Based on these observations, the authors suggest that the levels of PI(3,4,5)P3 may determine the cellular function of RAP1, either by binding upstream of VSG genes and repressing their function, or by not binding DNA and allowing the simultaneous expression of multiple VSG genes in a single parasite.

While I find most of the data presented in this manuscript compelling, there are aspects of Figure 1 that are not clear to me. Based on Figure 1F, the authors claim that transient inactivation of PIP5Pase results in a switch from the expression of one VSG isoform to another. However, I am not exactly sure what the authors are showing in this panel, nor do the data in Figure 1F seem to be consistent with those shown in Figure 1C. Based on Figure 1F, a transient inactivation of PIP5Pase appears to result in an almost exclusive switch to a VSG located in BES12. However, based on Figure 1E, the VSG transcripts most commonly found after a transient inactivation of PIP5Pase are those from the previously active VSG (BES1) and VSGs located on chr 1 and 6 (I believe). The small font and the low resolution make it impossible to infer the location of the expressed VSG genes, nor to confirm that ALL VSG genes located in expression sites are activated, as the authors claim. Also, I was not able to access the raw ChIP-seq and RNA-seq reads. Thus, could not evaluate the quality of the sequencing data.

We appreciate the reviewer’s comments and evaluation of our work. Fig 1E shows VSG-seq of a population after transient (24h) exclusive expression of the PIP5Pase mutant, followed by re-expression of the WT PIP5Pase allele for 60 hours (multiple VSGs are detected). As a control, it also shows VSG-seq in cells continuously expressing WT PIP5Pase (mostly VSG2, BES1 is detected). Fig 1F and Fig S1 show the sequencing of VSGs expressed by clones isolated (5-6 days of growth) after a temporary knockdown (24h) of PIP5Pase (tet -), followed by its re-expression. For comparison, no knockdown (tet +) was included. Fig 1F shows potential switchers in the population, the Fig 1E confirms VSG switching in clones.

To clarify the difference between Fig 1E and 1F, we edited the manuscript on page 3, lines 103-110: “To verify PIP5Pase role in VSG switching, we knocked down PIP5Pase for 24h (Tet -), then restored its expression (Tet +) and isolated clones by limiting dilution and growth for 5-6 days. Analysis of isolated clones after temporary PIP5Pase knockdown (Tet -/+) confirmed VSG switching in 93 out of 94 (99%) of the analyzed clones (Fig 1F, Fig S1). The cells switched to express VSGs from silent ESs or subtelomeric regions, indicating switching by transcription or recombination mechanisms. Moreover, no switching was detected in 118 isolated clones from cells continuously expressing WT PIP5Pase (Tet +, Fig 1F).”. We also edited Fig 1F to indicate temporary knockdown (Tet -/+) vs no knockdown (Tet -). The modifications will be available in the resubmitted version of the manuscript.

We agree that the heat map is difficult to read due to the amount of information. We will include in the revised version of the manuscript a table with the data in the supplementary information; the reader will be able to evaluate the data in detail.

A preference for switching to specific ESs has been observed in T. brucei (Morrison et al. 2005, Int J Parasitol; Cestari and Stuart, 2015, PNAS), which may explain several clones switching to BES12. Many potential switchers were detected in the VSG-seq (Fig 1F, the whole cell population is over 107 parasites), but not all potential switchers were detected in the clonal analysis because we analyzed 212 clones total, a fraction of the over 107 cells analyzed by VSG-seq (Fig 1E). Also, it is possible that not all potential switchers are viable. However, the point of the clonal analysis is to validate the VSG switching after genetic perturbation of PIP5Pase.

Fig 1C shows examples of ES derepression by RNA-seq after 24h exclusive expression of the mutant compared to WT PIP5Pase. The RNA-seq shows that all ESs are derepressed (Fig 1B). This can be visualized in the volcano plot (Fig 1B, BES and MES VSGs are labelled) and on the spreadsheet Data S1. Although all ESs are derepressed after PIP5Pase mutant expression, not all ESs are selected during switching, as observed in Fig 1E-F. This agrees with our previous observations in switching assays with proteins that control VSG switching (Cestari and Stuart, 2015, PNAS).

As for metrics of sequencing and raw sequencing data. See Methods section, page 13, lines 483-485: “Sequencing information is available in Table S3 and fastq data is available in the Sequence Read Archive (SRA) with the BioProject identification PRJNA934938.” Table S3 has a summary of sequencing data. Metrics information such as sequencing quality and analysis can be found in the Methods section “Computational analysis of RNA-seq and ChIP-seq”. The latter includes information about nanopore reads, i.e., mean Q-score of 12.

Reviewer #3 (Public Review):

In this manuscript, Touray et al investigate the mechanisms by which PIP5Pase and RAP1 control VSG expression in T. brucei and demonstrate an important role for this enzyme in a signalling pathway that likely plays a role in antigenic variation in T. brucei.

The methods used in the study are rigorous and well-controlled. The authors convincingly demonstrate that RAP1 binds to PI(3,4,5)P3 through its N-terminus and that this binding regulates RAP1 binding to VSG expression sites, which in turn regulates VSG silencing. Overall their results support the conclusions made in the manuscript.

There are a few small caveats that are worth noting. First, the analysis of VSG derepression and switching in Figure 1 relies on a genome that does not contain minichromosomal (MC) VSG sequences. This means that MC VSGs could theoretically be misassigned as coming from another genomic location in the absence of an MC reference. As the origin of the VSGs in these clones isn’t a major point in the paper, I do not think this is a major concern, but I would not over-interpret the particular details of switching outcomes in these experiments.

The authors state that “our data imply that antigenic variation is not exclusively stochastic.” I am not sure this is true. While I also favor the idea that switching is not exclusively stochastic, evidence for a signaling pathway does not necessarily imply that antigenic variation is not stochastic. This pathway could be important solely for lifecycle-related control of VSG expression, rather than antigenic variation during infection. Nevertheless, these data are critical for establishing a potential pathway that could control antigenic variation and thus represent a fundamental discovery.

Another aspect of this work that is perhaps important, but not discussed much by the authors, is the fact that signalling is extremely poorly understood in T. brucei. In Figure 1B, the RNA-seq data show many genes upregulated after expression of the Mut PIP5Pase (not just VSGs). The authors rightly avoid claiming that this pathway is exclusive to VSGs, but I wonder if these data could provide insight into the other biological processes that might be controlled by this signaling pathway in T. brucei.

Overall, this is an excellent study that represents an important step forward in understanding how antigenic variation is controlled in T. brucei. The possibility that this process could be controlled via a signalling pathway has been speculated for a long time, and this study provides the first mechanistic evidence for that possibility.

We thank the reviewer for the evaluation of our work. We agree that it is difficult to ensure the origin of all VSG genes not having minichromosome sequences; hence we did not emphasize this point in the manuscript. We used the 427-2018 reference genome assembled by PacBio and Hi-C (Muller et al. 2018, Nature), which we believe is the best assembly for the 427 strain, especially related to the VSG genes.

We also agree that having signaling controlling switching in vitro does not mean the switching necessarily occurs by signaling in vivo. Nevertheless, stochastic switching is an accepted model; but it has not been proved, whereas we provide molecular evidence that signaling can cause switching. To express this reviewer’s suggestion, we edited the Discussion, page 7, line 250: from “our data imply that antigenic variation is not exclusively stochastic” to “our data suggest that antigenic variation is not exclusively stochastic”.

Most of the RNA-seq data were VSGs genes/pseudogenes. Other genes upregulated included retrotransposons and DNA/RNA processing enzymes such as endonucleases and polymerases. We included in the Results, page 3, line 100: “Other genes upregulated include primarily retrotransposons, endonucleases, and polymerase proteins.”.

Author Response

Reviewer #2 (Public Review):

Associative learning assigns valence to sensory cues paired with reward or punishment. Brain regions such as the amygdala in mammals and the mushroom body in insects have been identified as primary sites where valence assignment takes place. However, little is known about the neural mechanisms that translate valence-specific activity in these brain regions into appropriate behavioral actions. This study identifies a small set of upwind neurons (UpWiNs) in the Drosophila brain that receive direct inputs from two mushroom body output neurons (MBONs) representing opposite valences. Through a series of behavioral, imaging, and electrophysiological experiments, the authors show that UpWiNs are differentially regulated by the two MBONs, i.e., inhibited by the glutamatergic MBON-α1(encoding negative valence) while activated by the cholinergic MBON-α3 (encoding positive valence). They also show that UpWiNs control the wind-directed behavior of flies. Activation of UpWiNs is sufficient to drive flies to orient and move upwind, and inhibition of UpWiNs reduces flies' upwind movement toward the source of reward-predicting odors (CS+). These results, together with existing knowledge about the function of the mushroom body in memory processing, suggest an appealing model in which reward learning decreases and increases the responses of MBON-α1 and MBON-α3 to the CS+ odor, respectively, and these changes cause UpWiNs to respond more strongly to the CS+ odor and drive upwind locomotion. Interestingly, in the final part of the results, the authors reveal a wind-independent function of UpWiNs: increasing the probability that flies will revisit the site where UpWiNs were activated. Thus, UpWiNs guide learned reward-seeking behavior with and without airflow. Although the mushroom body has been extensively studied for its role in learning and memory, the downstream neural circuits that read the information from the mushroom body to guide memory-driven behaviors remain poorly characterized. This study provides an important piece of the puzzle for this knowledge gap.

Strength

1) Memory studies have predominantly relied on binary choice (go or no-go) assays as measures of memory performance. While these assays are convenient and efficient, they fall short of providing a comprehensive understanding of underlying behavioral structures. In an effort to overcome this limitation, the current study used video recording and tracking software to delve deeper into memory-guided behavior. This innovative approach allowed the authors to uncover novel neurons and examine their contribution to behavior with a level of detail not possible with binary choice assays.

2) This study used electron microscopy-based Drosophila hemibrain connectome data to reveal the synaptic connection between UpWiNs and MBON-α1 and MBON-α3. Using this method, the study shows that a single UpWiN receives direct input from both MBON-α1 and MBON- α3, which is confirmed by a functional imaging experiment. The connectome dataset also reveals several neurons downstream of UpWiNs, opening avenues for further research into the neural mechanisms linking memory and behavior.

Weakness

1) The authors repeatedly state in the manuscript that MBON-α1 and MBON-α3 convey appetitive or aversive memories, respectively. This assertion may not be entirely accurate. Evidence from sugar reward conditioning experiments suggests that MBON-α3 is potentiated and required for sugar reward memory retrieval. Therefore, the compartmentalization for appetitive and aversive memories appears not as obvious at the level of MBONs.

What we intended was that activation of DANs in these compartments can induce aversive and appetitive memories, respectively, when paired with odors, and that these are the sole output pathway from these compartments to read out the memories in these compartments. As we previously proposed (Aso et al., 2014a eLife), these MBONs can integrate inputs from MBONs of other compartments and their activity can reflect appetitive memory stored as synaptic plasticity in other compartments. Since DANs in the α3 compartment respond to heat, bitter and electric shock but not sugar, the observation that MBON-α3 acquires an enhanced CS+ odor response after appetitive conditioning is presumably due to these intercompartmental connections rather than plasticity of KC-MBON synapses in the α3 compartment. In any case, the fact that excitatory activity of MBON-α1 and MBON-α3 conveys opposite valence of memory still holds true since appetitive conditioning induces depression and potentiation of odor responses, respectively.

To clarify this point, we now cited related literature in the following sentence in the final paragraph of Introduction: “UpWiNs receive inputs from several types of lateral horn neurons and integrate inhibitory and excitatory inputs from MBON-α1 and MBON-α3, which are the output neurons of MB compartments that store long-lasting appetitive or aversive memories, respectively (Aso and Rubin, 2016; Ichinose et al., 2015; Jacob and Waddell, 2022a; Pai et al., 2013; Yamagata et al., 2015).”

2) This study did not conclusively establish the importance of the MBON-α1/α3 to UpWiN pathways in memory-driven behavior. In the experiments shown in Figure 5, flies were trained to associate the activation of reward-related DANs with a specific odor (CS+). After conditioning, UpWiNs were observed to show enhanced responses to the CS+ odor. However, the results should be interpreted with caution because the driver line used to activate DANs (R58E02-LexAp65) labels not only DANs projecting to the MBON-α1 compartment, but all DANs in the protocerebral anterior medial (PAM) cluster. Thus, it remains unclear to what extent the observed enhanced responses are influenced by changes in inhibitory inputs from MBON-α1. While UpWiNs have been shown to play a critical role in the expression of sugar reward memory (Figure 7), it should be noted that UpWiNs receive inputs from multiple upstream neurons, making it difficult to accurately assess the contribution of MBON-α1/α3 to UpWiN pathways in UpWiN recruitment. Further research is needed to fully address this issue.

We totally agree with this point and added a sentence to explain an alternative mechanism. “This enhancement of CS+ response can be most easily explained as an outcome of disinhibition from MBON-α1 whose output had been decreased by memory formation; MBON-α1 is inhibitory to UpWiNs (Figure 4B) and MBON-α1 response to the CS+ is reduced following the same training protocol (Yamada et al. 2023). In addition to such a mechanism, plasticity in the β1 compartment may contribute to the enhanced CS+ response in UpWiNs because the driver R58E02 contains DANs in the β1 and glutamatergic MBON from the β1 directly synapse on the dendrites of MBON-α1 and MBON-α3. “

3) UpWind neurons (UpWiNs) were so named because their activation promotes upwind locomotion. However, when activated in the absence of airflow, flies show increased locomotor speed and an increased probability of revisiting the same location (Figure 7 and Figure 7-figure supplement 1). The revisiting behavior can be observed during the activation of UpWiNs, which is distinct from the local search behavior that typically begins after a reward stimulus is turned off (e.g., Gr64f-GAL4 results in Figure 7-figure supplement 1).

Return probability was calculated within a 15-s time window. High return probability during LED ON period (10-20s) in Figure 7-figure supplement 1 does not necessarily mean that flies returned during LED ON period. If a fly is at the position A when t=10s, to be counted as “returned”, it needs to move more than 10mm away from A and move back to the position less than 3mm distance from A by t=25s. In the case of sugar sensory neuron activation with Gr64f-GAL4, the peak of return probability is shifted toward a later time point because flies stop and extend proboscis during activation period.

Because revisiting a location can also be a consequence of repeated turns, it seems more accurate to describe UpWiNs as controlling the speed and likelihood of turns and promoting upwind movement by integrating with neurons that sense the direction of airflow.

The return probability plotted in Figure 7E is probability of return to the position at the end of LED period within 15s post LED period when angular speed of SS33917>CsChrimson and SS33918>CsChrimson flies are identical to empty-split-GAL4>CsChrimson control flies (Figure 7-figure supplement 1). Thus, revisiting behavior cannot be explained by a simple increase in turing probability.

Although functions of UpWiNs are not limited to promotion of wind-directed walking, we still think that the “UpWind Neurons” is a practical name for broad readers and oral communications at the current stage of investigations, because EM neuron IDs and names (SMP348, SMP353, SMP354, SLP399 and SLP400) are too lengthy and do not contain any functional information. We initially defined a set of 11 neurons labeled by SS33197 split-GAL4 as “UpWind Neurons (UpWiNs)” based on initial optogenetic screening (Figure 2A). We found other driver lines for mushroom body interneuron cell types that can promote release of dopamine and more robust returning phenotype (e.g. SS49755), but SS33917 remained to be the champion driver line for upwind locomotion phenotype.

Reviewer #3 (Public Review):

Aso et al. provide insight into how learned valences are transformed into concrete memory-driven actions, using a diverse set of proven techniques.

Here the authors use a four-armed arena to evaluate flies' preference for a reward-predicting odor and measure upwind locomotion. This behavioral paradigm was combined with the photoactivation of different memory-eliciting neurons, revealing that appetitive memories stored in different compartments of the mushroom bodies (center of olfactory memory) induce different levels of upwind locomotion. The authors then proceed to a non-exhaustive optogenetic screen of the neurons located downstream of the output neurons of the mushroom bodies (MBONs) and identify a group of 8-11 Cholinergic neurons promoting significant changes in upwind locomotion, the UpWins. By combining confocal immunolabelling of these neurons with electron microscope images, they manage to establish the UpWins' connectome within themselves and with the MBONs. Then, using two in vivo cell recording techniques, electrophysiology, and calcium imaging, they define that UpWins integrate both inhibitory and excitatory synaptic inputs from the MBONs encoding appetitive and aversive memory, respectively. In addition, they show that the UpWins' response to a reward-predicting odor is increased after appetitive training. On a behavioral level, the authors establish that the UpWins respond to wind direction only and are not involved in lower-level motor parameters, such as turning direction and acceleration. Finally, they demonstrate that the UpWins' activity is necessary for long-term appetitive memory retrieval, and even suggest a broader role for the UpWins in olfactory navigation, as their photoactivation increases the probability of revisiting behavior. In the end, the authors state that they provide new insights into how memory is translated into concrete behavior, which is fully supported by their data. Altogether, the authors present a pretty complete study that provides very interesting and reliable data, and that opens a new field of investigation into memory-driven behaviors.

Strengths of the study:

• To support their conclusions, the authors provide detailed data from different levels of analysis (behavioral, cellular, and molecular), using multiple sophisticated techniques.

• The measurement of multiple parameters in the behavioral analysis supports the strong changes in upwind locomotion. In addition, taken individually these parameters provide precise insights into how upwind locomotion changes, and allow the authors to more precisely define the role of the UpWins.

• The authors use split-Gal4 drivers instead of Gal4, allowing them to better refine neuron labelling.

The authors discussed and investigated all possible biases, making their data very reliable. For example, they demonstrated that the phenotypes observed in the behavioral assay were wind-directed behaviors and could not be explained by bias avoidance of the arena's center area.

Limitations of the study:

• In the absence of more precise drivers, the UpWins' labelling lacks precision. For example, there is no way to know exactly which UpWin is responding in the electrophysiological experiment presented in Figure 4.

We have ongoing efforts to generate split-GAL4 and split-LexA driver lines for specific subsets of UpWiN neurons, but the data using those lines are not ready for this manuscript. However, we would like to point out that historically, identification of a group of neurons with striking phenotype has been foundational to promote follow-up studies. A good example is P1 neurons for courtship behavior.

• The screening of neurons located downstream of the MBONs is not exhaustive, meaning that other groups of neurons might be involved in memory-driven upwind locomotion. Although, it does not diminish the authors' conclusions.

The UpWiNs is certainly not the only one cell type for mediating memory-driven upwind locomotion, since our and other groups’ studies (e.g. Matheson et al., 2022; PMCID: PMC9360402) identified a collection of cell types that can promote upwind locomotion upon optogenetic activation.

In 2021, we released images and driver lines of a larger collection of split-GAL4 driver lines at https://splitgal4.janelia.org. We are preparing a manuscript to provide anatomical descriptions of these lines. This collection of new drivers will help elucidate more comprehensive views of circuits for memory-driven actions.

• All data were obtained with walking flies. So far, there have been no experiments on flying flies.

This is an intriguing question and we mentioned in Discussion that “Our study was limited to walking behaviors, and the role of UpWiNs in flight behaviors remains to be investigated.”

Author Response

Reviewer #1 (Public Review):

The authors present a PyTorch-based simulator for prosthetic vision. The model takes in the anatomical location of a visual cortical prostheses as well as a series of electrical stimuli to be applied to each electrode, and outputs the resulting phosphenes. To demonstrate the usefulness of the simulator, the paper reproduces psychometric curves from the literature and uses the simulator in the loop to learn optimized stimuli.

One of the major strengths of the paper is its modeling work - the authors make good use of existing knowledge about retinotopic maps and psychometric curves that describe phosphene appearance in response to single-electrode stimulation. Using PyTorch as a backbone is another strength, as it allows for GPU integration and seamless integration with common deep learning models. This work is likely to be impactful for the field of sight restoration.

1) However, one of the major weaknesses of the paper is its model validation - while some results seem to be presented for data the model was fit on (as opposed to held-out test data), other results lack quantitative metrics and a comparison to a baseline ("null hypothesis") model. On the one hand, it appears that the data presented in Figs. 3-5 was used to fit some of the open parameters of the model, as mentioned in Subsection G of the Methods. Hence it is misleading to present these as model "predictions", which are typically presented for held-out test data to demonstrate a model's ability to generalize. Instead, this is more of a descriptive model than a predictive one, and its ability to generalize to new patients remains yet to be demonstrated.

We agree that the original presentation of the model fits might give rise to unwanted confusion. In the revision, we have adapted the fit of the thresholding mechanism to include a 3-fold cross validation, where part of the data was excluded during the fitting, and used as test sets to calculate the model’s performance. The results of the cross- validation are now presented in panel D of Figure 3. The fitting of the brightness and temporal dynamics parameters using cross-validation was not feasible due to the limited amount of quantitative data describing temporal dynamics and phosphene size and brightness for intracortical electrodes. To avoid confusion, we have adapted the corresponding text and figure captions to specify that we are using a fit as description of the data.

We note that the goal of the simulator is not to provide a single set of parameters that describes precise phosphene perception for all patients but that it could also be used to capture variability among patients. Indeed, the model can be tailored to new patients based on a small data set. Figure 3-figure supplement 1 exemplifies how our simulator can be tailored to several data sets collected from patients with surface electrodes. Future clinical experiments might be used to verify how well the simulator can be tailored to the data of other patients.

Specifically, we have made the following changes to the manuscript:

• Caption Figure 2: the fitted peak brightness levels reproduced by our model

• Caption Figure 3: The model's probability of phosphene perception is visualized as a function of charge per phase

• Caption Figure 3: Predicted probabilities in panel (d) are the results of a 3-fold cross- validation on held-out test data.

• Line 250: we included biologically inspired methods to model the perceptual effects of different stimulation parameters

• Line 271: Each frame, the simulator maps electrical stimulation parameters (stimulation current, pulse width and frequency) to an estimated phosphene perception

• Lines 335-336: such that 95% of the Gaussian falls within the fitted phosphene size.

• Line 469-470: Figure 4 displays the simulator's fit on the temporal dynamics found in a previous published study by Schmidt et al. (1996).

• Lines 922-925: Notably, the trade-off between model complexity and accurate psychophysical fits or predictions is a recurrent theme in the validation of the components implemented in our simulator.

2) On the other hand, the results presented in Fig. 8 as part of the end-to-end learning process are not accompanied by any sorts of quantitative metrics or comparison to a baseline model.

We now realize that the presentation of the end-to-end results might have given the impression that we present novel image processing strategies. However, the development of a novel image processing strategy is outside the scope of the study. Instead, The study aims to provide an improved simulation which can be used for more realistic assessment of different stimulation protocols. The simulator needs to fit experimental data, and it should run fast (so it can be used in behavioral experiments). Importantly, as demonstrated in our end-to-end experiments, the model can be used in differentiable programming pipelines (so it can be used in computational optimization experiments), which is a valuable contribution in itself because it lends itself to many machine learning approaches which can improve the realism of the simulation.

We have rephrased our study aims in the discussion to improve clarity.

• Lines 275-279: In the sections below, we discuss the different components of the simulator model, followed by a description of some showcase experiments that assess the ability to fit recent clinical data and the practical usability of our simulator in simulation experiments

• Lines 810-814: Computational optimization approaches can also aid in the development of safe stimulation protocols, because they allow a faster exploration of the large parameter space and enable task-driven optimization of image processing strategies (Granley et al., 2022; Fauvel et al., 2022; White et al., 2019; Küçükoglü et al. 2022; de Ruyter van Steveninck et al., 2022; Ghaffari et al., 2021).

• Lines 814-819: Ultimately, the development of task-relevant scene-processing algorithms will likely benefit both from computational optimization experiments as well as exploratory SPV studies with human observers. With the presented simulator we aim to contribute a flexible toolkit for such experiments.

• Lines 842-853: Eventually, the functional quality of the artificial vision will not only depend on the correspondence between the visual environment and the phosphene encoding, but also on the implant recipient's ability to extract that information into a usable percept. The functional quality of end-to-end generated phosphene encodings in daily life tasks will need to be evaluated in future experiments. Regardless of the implementation, it will always be important to include human observers (both sighted experimental subjects and actual prosthetic implant users in the optimization cycle to ensure subjective interpretability for the end user (Fauvel et al., 2022; Beyeler & Sanchez-Garcia, 2022).

3) The results seem to assume that all phosphenes are small Gaussian blobs, and that these phosphenes combine linearly when multiple electrodes are stimulated. Both assumptions are frequently challenged by the field. For all these reasons, it is challenging to assess the potential and practical utility of this approach as well as get a sense of its limitations.

The reviewer raises a valid point and a similar point was raised by a different reviewer (our response is duplicated). As pointed out in the discussion, many aspects about multi- electrode phosphene perception are still unclear. On the one hand, the literature is in agreement that there is some degree of predictability: some papers explicitly state that phosphenes produced by multiple patterns are generally additive (Dobelle & Mladejovsky, 1974), that the locations are predictable (Bosking et al., 2018) and that multi-electrode stimulation can be used to generate complex, interpretable patterns of phosphenes (Chen et al., 2020, Fernandez et al., 2021). On the other hand, however, in some cases, the stimulation of multiple electrodes is reported to lead to brighter phosphenes (Fernandez et al., 2021), fused or displaced phosphenes (Schmidt et al., 1996, Bak et al., 1990) or unpredicted phosphene patterns (Fernández et al., 2021). It is likely that the probability of these interference patterns decreases when the distance between the stimulated electrodes increases. An empirical finding is that the critical distance for intracortical stimulation is approximately 1 mm (Ghose & Maunsell, 2012).

We note that our simulator is not restricted to the simulation of linearly combined Gaussian blobs. Some irregularities, such as elongated phosphene shapes were already supported in the previous version of our software. Furthermore, we added a supplementary figure that displays a possible approach to simulate some of the more complex electrode interactions that are reported in the literature, with only minor adaptations to the code. Our study thereby aims to present a flexible simulation toolkit that can be adapted to the needs of the user.

Adjustments:

• Added Figure 1-figure supplement 3 on irregular phosphene percepts.

• Lines 957-970: Furthermore, in contrast to the assumptions of our model, interactions between simultaneous stimulation of multiple electrodes can have an effect on the phosphene size and sometimes lead to unexpected percepts (Fernandez et al., 2021, Dobelle & Mladejovsky 1974, Bak et al., 1990). Although our software supports basic exploratory experimentation of non-linear interactions (see Figure 1-figure supplement 3), by default, our simulator assumes independence between electrodes. Multi- phosphene percepts are modeled using linear summation of the independent percepts. These assumptions seem to hold for intracortical electrodes separated by more than 1 mm (Ghose & Maunsell, 2012), but may underestimate the complexities observed when electrodes are nearer. Further clinical and theoretical modeling work could help to improve our understanding of these non-linear dynamics.

4) Another weakness of the paper is the term "biologically plausible", which appears throughout the manuscript but is not clearly defined. In its current form, it is not clear what makes this simulator "biologically plausible" - it certainly contains a retinotopic map and is fit on psychophysical data, but it does not seem to contain any other "biological" detail.

We thank the reviewer for the remark. We improved our description of what makes the simulator “biologically plausible” in the introduction (line 78): ‘‘Biological plausibility, in our work's context, points to the simulation's ability to capture essential biological features of the visual system in a manner consistent with empirical findings: our simulator integrates quantitative findings and models from the literature on cortical stimulation in V1 [...]”. In addition, we mention in the discussion (lines 611 - 621): “The aim of this study is to present a biologically plausible phosphene simulator, which takes realistic ranges of stimulation parameters, and generates a phenomenologically accurate representation of phosphene vision using differentiable functions. In order to achieve this, we have modeled and incorporated an extensive body of work regarding the psychophysics of phosphene perception. From the results presented in section H, we observe that our simulator is able to produce phosphene percepts that match the descriptions of phosphene vision that were gathered in basic and clinical visual neuroprosthetics studies over the past decades.”

5) In fact, for the most part the paper seems to ignore the fact that implanting a prosthesis in one cerebral hemisphere will produce phosphenes that are restricted to one half of the visual field. Yet Figures 6 and 8 present phosphenes that seemingly appear in both hemifields. I do not find this very "biologically plausible".

We agree with the reviewer that contemporary experiments with implantable electrodes usually test electrodes in a single hemisphere. However, future clinically useful approaches should use bilaterally implanted electrode arrays. Our simulator can either present phosphene locations in either one or both hemifields.

We have made the following textual changes:

• Fig. 1 caption: Example renderings after initializing the simulator with four 10 × 10 electrode arrays (indicated with roman numerals) placed in the right hemisphere (electrode spacing: 4 mm, in correspondence with the commonly used 'Utah array' (Maynard et al., 1997)).

• Line 518-525: The simulator is initialized with 1000 possible phosphenes in both hemifields, covering a field of view of 16 degrees of visual angle. Note that the simulated electrode density and placement differs from current prototype implants and the simulation can be considered to be an ambitious scenario from a surgical point of view, given the folding of the visual cortex and the part of the retinotopic map in V1 that is buried in the calcarine sulcus. Line 546-547: with the same phosphene coverage as the previously described experiment

Reviewer #2 (Public Review):

Van der Grinten and De Ruyter van Steveninck et al. present a design for simulating cortical- visual-prosthesis phosphenes that emphasizes features important for optimizing the use of such prostheses. The characteristics of simulated individual phosphenes were shown to agree well with data published from the use of cortical visual prostheses in humans. By ensuring that functions used to generate the simulations were differentiable, the authors permitted and demonstrated integration of the simulations into deep-learning algorithms. In concept, such algorithms could thereby identify parameters for translating images or videos into stimulation sequences that would be most effective for artificial vision. There are, however, limitations to the simulation that will limit its applicability to current prostheses.

The verification of how phosphenes are simulated for individual electrodes is very compelling. Visual-prosthesis simulations often do ignore the physiologic foundation underlying the generation of phosphenes. The authors' simulation takes into account how stimulation parameters contribute to phosphene appearance and show how that relationship can fit data from actual implanted volunteers. This provides an excellent foundation for determining optimal stimulation parameters with reasonable confidence in how parameter selections will affect individual-electrode phosphenes.

We thank the reviewer for these supportive comments.

Issues with the applicability and reliability of the simulation are detailed below:

1) The utility of this simulation design, as described, unfortunately breaks down beyond the scope of individual electrodes. To model the simultaneous activation of multiple electrodes, the authors' design linearly adds individual-electrode phosphenes together. This produces relatively clean collections of dots that one could think of as pixels in a crude digital display. Modeling phosphenes in such a way assumes that each electrode and the network it activates operate independently of other electrodes and their neuronal targets. Unfortunately, as the authors acknowledge and as noted in the studies they used to fit and verify individual-electrode phosphene characteristics, simultaneous stimulation of multiple electrodes often obscures features of individual-electrode phosphenes and can produce unexpected phosphene patterns. This simulation does not reflect these nonlinearities in how electrode activations combine. Nonlinearities in electrode combinations can be as subtle the phosphenes becoming brighter while still remaining distinct, or as problematic as generating only a single small phosphene that is indistinguishable from the activation of a subset of the electrodes activated, or that of a single electrode.

If a visual prosthesis happens to generate some phosphenes that can be elicited independently, a simulator of this type could perhaps be used by processing stimulation from independent groups of electrodes and adding their phosphenes together in the visual field.

The reviewer raises a valid point and a similar point was raised by a different reviewer (our response is duplicated). As pointed out in the discussion, many aspects about multi- electrode phosphene perception are still unclear. On the one hand, the literature is in agreement that there is some degree of predictability: some papers explicitly state that phosphenes produced by multiple patterns are generally additive (Dobelle & Mladejovsky, 1974), that the locations are predictable (Bosking et al., 2018) and that multi-electrode stimulation can be used to generate complex, interpretable patterns of phosphenes (Chen et al., 2020, Fernandez et al., 2021). On the other hand, however, in some cases, the stimulation of multiple electrodes is reported to lead to brighter phosphenes (Fernandez et al., 2021), fused or displaced phosphenes (Schmidt et al., 1996, Bak et al., 1990) or unpredicted phosphene patterns (Fernández et al., 2021). It is likely that the probability of these interference patterns decreases when the distance between the stimulated electrodes increases. An empirical finding is that the critical distance for intracortical stimulation is approximately 1 mm (Ghose & Maunsell, 2012).

We note that our simulator is not restricted to the simulation of linearly combined Gaussian blobs. Some irregularities, such as elongated phosphene shapes were already supported in the previous version of our software. Furthermore, we added a supplementary figure that displays a possible approach to simulate some of the more complex electrode interactions that are reported in the literature, with only minor adaptations to the code. Our study thereby aims to present a flexible simulation toolkit that can be adapted to the needs of the user.

Adjustments:

• Lines 957-970: Furthermore, in contrast to the assumptions of our model, interactions between simultaneous stimulation of multiple electrodes can have an effect on the phosphene size and sometimes lead to unexpected percepts (Fernandez et al., 2021, Dobelle & Mladejovsky 1974, Bak et al., 1990). Although our software supports basic exploratory experimentation of non-linear interactions (see Figure 1-figure supplement 3), by default, our simulator assumes independence between electrodes. Multi- phosphene percepts are modeled using linear summation of the independent percepts. These assumptions seem to hold for intracortical electrodes separated by more than 1 mm (Ghose & Maunsell, 2012), but may underestimate the complexities observed when electrodes are nearer. Further clinical and theoretical modeling work could help to improve our understanding of these non-linear dynamics.

• Added Figure 1-figure supplement 3 on irregular phosphene percepts.

2) Verification of how the simulation renders individual phosphenes based on stimulation parameters is an important step in confirming agreement between the simulation and the function of implanted devices. That verification was well demonstrated. The end use a visual-prosthesis simulation, however, would likely not be optimizing just the appearance of phosphenes, but predicting and optimizing functional performance in visual tasks. Investigating whether this simulator can suggest visual-task performance, either with sighted volunteers or a decoder model, that is similar to published task performance from visual-prosthesis implantees would be a necessary step for true validation.

We agree with the reviewer that it will be vital to investigate the utility of the simulator in tasks. However, the literature on the performance of users of a cortical prosthesis in visually-guided tasks is scarce, making it difficult to compare task performance between simulated versus real prosthetic vision.

Secondly, the main objective of the current study is to propose a simulator that emulates the sensory / perceptual experience, i.e. the low-level perceptual correspondence. Once more behavioral data from prosthetic users become available, studies can use the simulator to make these comparisons.

Regarding the comparison to simulated prosthetic vision in sighted volunteers, there are some fundamental limitations. For instance, sighted subjects are exposed for a shorter duration to the (simulated) artificial percept and lack the experience and training that prosthesis users get. Furthermore, sighted subjects may be unfamiliar with compensation strategies that blind individuals have developed. It will therefore be important to conduct clinical experiments.

To convey more clearly that our experiments are performed to verify the practical usability in future behavioral experiments, we have incorporated the following textual adjustments:

• Lines 275-279: In the sections below, we discuss the different components of the simulator model, followed by a description of some showcase experiments that assess the ability to fit recent clinical data and the practical usability of our simulator in simulation experiments.

• Lines 842-853: Eventually, the functional quality of the artificial vision will not only depend on the correspondence between the visual environment and the phosphene encoding, but also on the implant recipient's ability to extract that information into a usable percept. The functional quality of end-to-end generated phosphene encodings in daily life tasks will need to be evaluated in future experiments. Regardless of the implementation, it will always be important to include human observers (both sighted experimental subjects and actual prosthetic implant users in the optimization cycle to ensure subjective interpretability for the end (Fauvel et al., 2022; Beyeler & Sanchez- Garcia, 2022).

3) A feature of this simulation is being able to convert stimulation of V1 to phosphenes in the visual field. If used, this feature would likely only be able to simulate a subset of phosphenes generated by a prosthesis. Much of V1 is buried within the calcarine sulcus, and electrode placement within the calcarine sulcus is not currently feasible. As a result, stimulation of visual cortex typically involves combinations of the limited portions of V1 that lie outside the sulcus and higher visual areas, such as V2.

We agree that some areas (most notably the calcarine sulcus) are difficult to access in a surgical implantation procedure. A realistic simulation of state-of-the-art cortical stimulation should only partially cover the visual field with phosphenes. However, it may be predicted that some of these challenges will be addressed by new technologies. We chose to make the simulator as generally applicable as possible and users of the simulator can decide which phosphene locations are simulated. To demonstrate that our simulator can be flexibly initialized to simulate specific implantation locations using third- party software, we have now added a supplementary figure (Figure 1-figure supplement 1) that displays a demonstration of an electrode grid placement on a 3D brain model, generating the phosphene locations from receptive field maps. However, the simulator is general and can also be used to guide future strategies that aim to e.g. cover the entire field with electrodes, compare performance between upper and lower hemifields etc.

Reviewer #3 (Public Review):

The authors are presenting a new simulation for artificial vision that incorporates many recent advances in our understanding of the neural response to electrical stimulation, specifically within the field of visual prosthetics. The authors succeed in integrating multiple results from other researchers on aspects of V1 response to electrical stimulation to create a system that more accurately models V1 activation in a visual prosthesis than other simulators. The authors then attempt to demonstrate the value of such a system by adding a decoding stage and using machine-learning techniques to optimize the system to various configurations.

1) While there is merit to being able to apply various constraints (such as maximum current levels) and have the system attempt to find a solution that maximizes recoverable information, the interpretability of such encodings to a hypothetical recipient of such a system is not addressed. The authors demonstrate that they are able to recapitulate various standard encodings through this automated mechanism, but the advantages to using it as opposed to mechanisms that directly detect and encode, e.g., edges, are insufficiently justified.

We thank the reviewer for this constructive remark. Our simulator is designed for more realistic assessment of different stimulation protocols in behavioral experiments or in computational optimization experiments. The presented end-to-end experiments are a demonstration of the practical usability of our simulator in computational experiments, building on a previously existing line of research. In fact, our simulator is compatible with any arbitrary encoding strategy.

As our paper is focused on the development of a novel tool for this existing line of research, we do not aim to make claims about the functional quality of end-to-end encoders compared to alternative encoding methods (such as edge detection). That said, we agree with the reviewer that it is useful to discuss the benefits of end-to-end optimization compared to e.g. edge detection will be useful.

We have incorporated several textual changes to give a more nuanced overview and to acknowledge that many benefits remain to be tested. Furthermore, we have restated our study aims more clearly in the discussion to clarify the distinction between the goals of the current paper and the various encoding strategies that remain to be tested.

• Lines 275-279: In the sections below, we discuss the different components of the simulator model, followed by a description of some showcase experiments that assess the ability to fit recent clinical data and the practical usability of our simulator in simulation experiments

• Lines 810-814: Computational optimization approaches can also aid in the development of safe stimulation protocols, because they allow a faster exploration of the large parameter space and enable task-driven optimization of image processing strategies (Granley et al., 2022; Fauvel et al., 2022; White et al., 2019; Küçükoglü et al. 2022; de Ruyter van Steveninck, Güçlü et al., 2022; Ghaffari et al., 2021).

• Lines 842-853: Eventually, the functional quality of the artificial vision will not only depend on the correspondence between the visual environment and the phosphene encoding, but also on the implant recipient's ability to extract that information into a usable percept. The functional quality of end-to-end generated phosphene encodings in daily life tasks will need to be evaluated in future experiments. Regardless of the implementation, it will always be important to include human observers (both sighted experimental subjects and actual prosthetic implant users in the optimization cycle to ensure subjective interpretability for the end user (Fauvel et al., 2022; Beyeler & Sanchez-Garcia, 2022).

2) The authors make a few mistakes in their interpretation of biological mechanisms, and the introduction lacks appropriate depth of review of existing literature, giving the reader the mistaken impression that this is simulator is the only attempt ever made at biologically plausible simulation, rather than merely the most recent refinement that builds on decades of work across the field.

We thank the reviewer for this insight. We have improved the coverage of the previous literature to give credit where credit is due, and to address the long history of simulated phosphene vision.

Textual changes:

• Lines 64-70: Although the aforementioned SPV literature has provided us with major fundamental insights, the perceptual realism of electrically generated phosphenes and some aspects of the biological plausibility of the simulations can be further improved and by integrating existing knowledge of phosphene vision and its underlying physiology.

• Lines 164-190: The aforementioned studies used varying degrees of simplification of phosphene vision in their simulations. For instance, many included equally-sized phosphenes that were uniformly distributed over the visual field (informally referred to as the ‘scoreboard model’). Furthermore, most studies assumed either full control over phosphene brightness or used binary levels of brightness (e.g. 'on' / 'off'), but did not provide a description of the associated electrical stimulation parameters. Several studies have explicitly made steps towards more realistic phosphene simulations, by taking into account cortical magnification or using visuotopic maps (Fehervari et al., 2010;, Li et al., 2013; Srivastava et al., 2009; Paraskevoudi et al., 2021), simulating noise and electrode dropout (Dagnelie et al., 2007), or using varying levels of brightness (Vergnieux et al., 2017; Sanchez-Garcia et al., 2022; Parikh et al., 2013). However, no phosphene simulations have modeled temporal dynamics or provided a description of the parameters used for electrical stimulation. Some recent studies developed descriptive models of the phosphene size or brightness as a function of the stimulation parameters (Winawer et al., 2016; Bosking et al., 2017). Another very recent study has developed a deep-learning based model for predicting a realistic phosphene percept for single stimulating electrodes (Granley et al., 2022). These studies have made important contributions to improve our understanding of the effects of different stimulation parameters. The present work builds on these previous insights to provide a full simulation model that can be used for the functional evaluation of cortical visual prosthetic systems.

• Lines 137-140: Due to the cortical magnification (the foveal information is represented by a relatively large surface area in the visual cortex as a result of variation of retinal RF size) the size of the phosphene increases with its eccentricity (Winawer & Parvizi, 2016, Bosking et al., 2017).

• Lines 883-893: Even after loss of vision, the brain integrates eye movements for the localization of visual stimuli (Reuschel et al., 2012), and in cortical prostheses the position of the artificially induced percept will shift along with eye movements (Brindley & Lewin, 1968, Schmidt et al., 1996). Therefore, in prostheses with a head-mounted camera, misalignment between the camera orientation and the pupillary axes can induce localization problems (Caspi et al., 2018; Paraskevoudi & Pezaris, 2019; Sabbah et al., 2014; Schmidt et al., 1996). Previous SPV studies have demonstrated that eye-tracking can be implemented to simulate the gaze-coupled perception of phosphenes (Cha et al., 1992; Sommerhalder et al., 2004; Dagnelie et al., 2006; McIntosh et al., 2013, Paraskevoudi & Pezaris, 2021; Rassia & Pezaris 2018, Titchener et al., 2018, Srivastava et al., 2009)

3) The authors have importantly not included gaze position compensation which adds more complexity than the authors suggest it would, and also means the simulator lacks a basic, fundamental feature that strongly limits its utility.

We agree with the reviewer that the inclusion of gaze position to simulate gaze-centered phosphene locations is an important requirement for a realistic simulation. We have made several textual adjustments to section M1 to improve the clarity of the explanation and we have added several references to address the simulation literature that took eye movements into account.

In addition, we included a link to some demonstration videos in which we illustrate that the simulator can be used for gaze-centered phosphene simulation. The simulation models the phosphene locations based on the gaze direction, and updates the input with changes in the gaze direction. The stimulation pattern is chosen to encode the visual environment at the location where the gaze is directed. Gaze contingent processing has been implemented in prior simulation studies (for instance: Paraskevoudi et al., 2021; Rassia et al., 2018; Titchener et al., 2018) and even in the clinical setting with users of the Argus II implant (Caspi et al., 2018). From a modeling perspective, it is relatively straightforward to simulate gaze-centered phosphene locations and gaze contingent image processing (our code will be made publicly available). At the same time, however, seen from a clinical and hardware engineering perspective, the implementation of eye-tracking in a prosthetic system for blind individuals might come with additional complexities. This is now acknowledged explicitly in the manuscript.

Textual adjustment:

Lines 883-910: Even after loss of vision, the brain integrates eye movements for the localization of visual stimuli (Reuschel et al., 2012), and in cortical prostheses the position of the artificially induced percept will shift along with eye movements (Brindley & Lewin, 1968, Schmidt et al., 1996). Therefore, in prostheses with a head-mounted camera, misalignment between the camera orientation and the pupillary axes can induce localization problems (Caspi et al., 2018; Paraskevoudi & Pezaris, 2019; Sabbah et al., 2014; Schmidt et al., 1996). Previous SPV studies have demonstrated that eye-tracking can be implemented to simulate the gaze-coupled perception of phosphenes (Cha et al., 1992; Sommerhalder et al., 2004; Dagnelie et al., 2006, McIntosh et al., 2013; Paraskevoudi et al., 2021; Rassia et al., 2018; Titchener et al., 2018; Srivastava et al., 2009). Note that some of the cited studies implemented a simulation condition where not only the simulated phosphene locations, but also the stimulation protocol depended on the gaze direction. More specifically, instead of representing the head-centered camera input, the stimulation pattern was chosen to encode the external environment at the location where the gaze was directed. While further research is required, there is some preliminary evidence that such a gaze-contingent image processing can improve the functional and subjective quality of prosthetic vision (Caspi et al., 2018; Paraskevoudi et al., 2021; Rassia et al., 2018; Titchener et al., 2018). Some example videos of gaze-contingent simulated prosthetic vision can be retrieved from our repository (https://github.com/neuralcodinglab/dynaphos/blob/main/examples/). Note that an eye-tracker will be required to produce gaze-contingent image processing in visual prostheses and there might be unforeseen complexities in the clinical implementation thereof. The study of oculomotor behavior in blind individuals (with or without a visual prosthesis) is still an ongoing line of research (Caspi et al.,2018; Kwon et al., 2013; Sabbah et al., 2014; Hafed et al., 2016).

4) Finally, the computational capacity required to run the described system is substantial and is not one that would plausibly be used as part of an actual device, suggesting that there may be difficulties with converting results from this simulator to an implantable system.

The software runs in real time with affordable, consumer-grade hardware. In Author response image 1 we present the results of performance testing with a 2016 model MSI GeForce GTX 1080 (priced around €600).

Author response image 1.

Note that the GPU is used only for the computation and rendering of the phosphene representations from given electrode stimulation patterns, which will never be part of any prosthetic device. The choice of encoder to generate the stimulation patterns will determine the required processing capacity that needs to be included in the prosthetic system, which is unrelated to the simulator’s requirements.

The following addition was made to the text:

• Lines 488-492: Notably, even on a consumer-grade GPU (e.g. a 2016 model GeForce GTX 1080) the simulator still reaches real-time processing speeds (>100 fps) for simulations with 1000 phosphenes at 256x256 resolution.

5) With all of that said, the results do represent an advance, and one that could have wider impact if the authors were to reduce the computational requirements, and add gaze correction.

We appreciate the kind compliment from the reviewer and sincerely hope that our revised manuscript meets their expectations. Their feedback has been critical to reshape and improve this work.

Author Response

Review #1 Public Review:

This is an interesting study which attempts to assess the effect of the pandemic on diagnoses of pancreatic cancer. The authors have used a large national database to evaluate this, however, it should be noted that this database only captures 40% of the population in England. The authors have looked at specific parameters including Body Mass Index (BMI) as well as markers of diabetes and liver function. Only BMI had a difference in the frequency of measurements during the pandemic, presumably due to reduced face-to-face visits to allow weight and height to be captured.

Interestingly the authors noticed a reduction in surgery for pancreatic cancer by 25%, yet reported that there were no differences in the frequency of death within 6 months following the diagnosis of pancreatic cancer. The reduction in surgery is likely related at least in part to the loss of operating lists due to pandemic restrictions, however, this paper is not equipped to address another important possibility behind this, which is that pancreatic cancers were presenting too late for surgical intervention. It is not sufficient to comment that pancreatic cancer treatment was not affected by the pandemic based on the data presented on deaths within 6 months of the diagnosis of pancreatic cancer alone, as the median survival of patients diagnosed with pancreatic cancer within the pandemic has not been captured and compared to that of patients diagnosed in the preceding 5 years.

Therefore while the study can conclude no difference in pancreatic cancer diagnoses before and during the pandemic, more work needs to be done to truly assess if the pandemic had any effect on the outcomes from pancreatic cancer for patients diagnosed within this timeframe.

Thank you for taking time to undertake the review and for all the constructive comments. This study was designed to assess the effect of the pandemic on pancreatic cancer services in England. We focused on the quantity of healthcare.

We acknowledge and understand the comments by the reviewer with regards to the limitations of this study in relation to the effect of the COVID-19 pandemic on diagnosis and survival. We did not assess the effect of the pandemic on the staging information and survival length.

Author Response

Reviewer #1 (Public Review):

This research aimed to discern the pattern of methylation changes that occur during aging, distinguishing between a unified specific mechanism and stochastic changes. To date, no unified hypothesis exists to guide our understanding of the changes in chromatin geography observed during the aging of cells. This work analysed six different types of purified blood-borne white blood cells allowing comparison across different immune cell subsets to determine if similar patterns occurred in all cell populations. Intriguingly, each subset exhibited its own distinct differential methylation rather than a single program. However, a core set of gene changes close to age-associated CpGs was identified suggesting that a central program existed, but that individual cell type function and metabolism shaped the overall chromatin landscape for the population. These findings establish a new framework for considering the aging process and open new questions about how the individual clocks of different populations might be regulated. While circulating cells are readily accessible for evaluation in humans, the majority of immune cells that regulate immune homeostasis are found within the tissues of the body. Whether these cells exhibit a similar profile to circulating cells or are rather shaped by their tissue or organ-specific ecosystem remains to be determined. In this setting, these tissue-resident cells are exposed to very different oxygen tensions and metabolic substrates. Furthermore, genes identified have been associated with aging, they concurrently appear to be associated with inflammation, thus it is not clear whether aging and low-grade inflammation are inherently linked, or whether these two pathways can be segregated. Thus a number of questions remain warranting further investigation.

The reviewer makes a very good point regarding different tissue resident cells being exposed to different oxygen and metabolic stress. In the reviewed manuscript we have Arid3a coming up as one of the transcription factors with motifs in and around probes hypermethylated with age in monocytes. Arid3a is known to target inflammatory genes but future research is warranted to implicate the link between aging and low-grade inflammation. To address the comment about connection between aging and low-grade inflammation, in the revised manuscript, we have incorporated new analysis by looking into SomaScan array derived protein levels of seven cytokines from the same cohort of donors. We tested the hypothesis that part of the age-associated changes in DNA methylation are connected with the well-known age-related proinflammatory state. We have now added the details in the Results and Methods sections. Briefly, we run two regression models (CpGi~age+sex and CpGi~age+sex+analytej, where i is each CpG probe from EPIC array and j is each of the seven cytokines). We find that change in DNA methylation levels in nearly 70009000 CpG sites in CD4 cells and 124 CpG sites in B cells that were originally age-associated, also are associated with increasing levels of TNFRSF1A, TNFRSF1B and TNF-alpha levels thereby indicating a link between DNA methylation change and aging as well as inflammatory cytokines levels.

Author Response

Reviewer #1 (Public Review):

The authors convincingly show in this study the effects of the fas5 gene on changes in the CHC profile and the importance of these changes toward sexual attractiveness.

The main strength of this study lies in its holistic approach (from genes to behaviour) showing a full and convincing picture of the stated conclusions. The authors succeeded in putting a very interdisciplinary set of experiments together to support the main claims of this manuscript.

We appreciate the kind comments from the reviewer.

The main weakness stems from the lack of transparency behind the statistical analyses conducted in the study. Detailed statistical results are never mentioned in the text, nor is it always clear what was compared to what. I also believe that some tests that were conducted are not adequate for the given data. I am therefore unable to properly assess the significance of the results from the presented information. Nevertheless, the graphical representations are convincing enough for me to believe that a revision of the statistics would not significantly affect the main conclusions of this manuscript.

We apologize for neglecting a detailed description of statistical tests that were performed. We wrote additional paragraphs in the method part specifically explaining the statistical analyses (line 435-445; 489-502; 559-561; 586-591).

The second major problem I had with the study was how it brushes over the somewhat contradicting results they found in males (Fig S2). These are only mentioned twice in the main text and in both cases as being "similarly affected", even though their own stats seem to indicate otherwise for many of the analysed compound groups. This also should affect the main conclusion concerning the effects of fas5 genes in the discussion, a more careful wording when interpreting the results is therefore necessary.

Thank you for pointing this out. Though our focus clearly lay on the female CHC profiles as a function in sexual signaling has only been described thus far for them, we now elaborated the result and discussion for the fas5 RNAi male part (line 167-178; 258-268).

Reviewer #2 (Public Review):

Insects have long been known to use cuticular hydrocarbons for communication. While the general pathways for hydrocarbon synthesis have been worked out, their specificity and in particular the specificity of the different enzymes involved is surprisingly little understood. Here, the authors convincingly demonstrate that a single fatty acid synthase gene is responsible for a shift in the positions of methyl groups across the entire alkane spectrum of a wasp, and that the wasps males recognize females specifically based on these methyl group positions. The strength of the study is the combination of gene expression manipulations with behavioural observations evaluating the effect of the associated changes in the cuticular hydrocarbon profiles. The authors make sure that the behavioural effect is indeed due to the chemical changes by not only testing life animals, but also dead animals and corpses with manipulated cuticular hydrocarbons.

I find the evidence that the hydrocarbon changes do not affect survival and desiccation resistance less convincing (due to the limited set of conditions and relatively small sample size), but the data presented are certainly congruent with the idea that the methyl alkane changes do not have large effects on desiccation.

We appreciate the kind comments from the reviewer.

Reviewer #3 (Public Review):

In this manuscript, the authors are aiming to demonstrate that a fatty-acyl synthase gene (fas5) is involved in the composition of the blend of surface hydrocarbons of a parasitoid wasp and that it affects the sexual attractiveness of females for males. Overall, the manuscript reads very well, it is very streamlined, and the authors' claims are mostly supported by their experiments and observations.

We appreciate the kind comments from the reviewer.

However, I find that some experiments, information and/or discussion are absent to assess how the effects they observe are, at least in part, not due to other factors than fas5 and the methyl-branched (MB) alkanes. I'm also wondering if what the authors observe is only a change in the sexual attractiveness of females and not related to species recognition as well.

We appreciate the interesting point that the reviewer raises in sexual attractiveness and species recognition and now expand upon this potential aspect in the discussion (lines 327-330). However, in this manuscript, we very much focused on the effect of fas5 knockdown on the conveyance of female sexual attractiveness in a single species (Nasonia vitripennis). Therefore, we argue that species recognition constitutes a different communication modality here, and we currently cannot infer whether and how species recognition is exactly encoded in Nasonia CHC profiles despite some circumstantial evidence for species-specificity (Buellesbach et al. 2013; Mair et al. 2017). Thus, we would like to refrain from any further speculation on species recognition before this can be unambiguously demonstrated, and remain within the mechanism of sexual attractiveness within a single species which we clearly show is mediated by the female MB-alkane fraction governed by the fatty acid synthase genes. We however still consider potential alternative explanations (e.g., n-alkenes acting as a deterrent of homosexual mating attempts).

The authors explore the function of cuticular hydrocarbons (CHCs) and a fatty-acyl synthase in Nasonia vitripennis, a parasitic wasp. Using RNAi, they successfully knockdown the expression of the fas5 gene in wasps. The authors do not justify their choice of fatty-acyl synthase candidate gene. It would have been interesting to know if that is one of many genes they studied or if there was some evidence that drove them to focus their interest in fas5.

In a previous study, 5 fas candidate genes orthologous to Drosophila melanogaster fas genes were identified and mapped in the genome of Nasonia vitripennis (Buellesbach et al. 2022). We actually investigated the effects of all of these fas genes on CHC variation, but only fas5 led to such a striking, traceable pattern shift. We are currently preparing another manuscript discussing the effects of the other fas genes, but decided to focus exclusively on fas5 here, due to its significance for revealing how sexual attractiveness can be encoded and conveyed in complex chemical profiles, maintained and governed by a surprisingly simple genetic basis.

The authors observe large changes in the cuticular hydrocarbons (CHC) profile of male and females. These changes are mostly a reduction of some MB alkanes and an increase in others as well as an increase of n-alkene in fas5 knockdown females. For males fas5 knockdowns, the overall quantity of CHC is increased and consequently, multiple types of compounds are increased compared to wild-type, with only one compound appearing to decrease compared to wild-type. Insects are known to rely on ratios of compounds in blends to recognize odors. Authors address this by showing a plot of the relative ratios, but it seems to me that they do show statistical tests of those changes in the proportions of the different types of compounds. In the results section, the authors give percentages while referring to figures showing the absolute amount of CHCs. They should also test if the ratios are significantly different or not between experimental conditions. Similar data should be displayed for the males as well.

We appreciate your suggestions. We kindly refer you to our response to reviewer 1, where we addressed the statistical tests. Specifically, we generated separate subplots to display the proportions of different compound classes and performed statistical tests to compare these proportions between different treatments for both males and females. Additionally, we have revised the results section to replace relative abundances with absolute quantity, as depicted in Figure 2C-G.

Furthermore, the authors didn't use an internal standard to measure the quantity of CHCs in the extracts, which, to me, is the gold standard in the field. If I understood correctly, the authors check the abundance measured for known quantities of n-alkanes. I'm sure this method is fine, but I would have liked to be reassured that the quantities measured through this method are good by either testing some samples with an internal standard, or referring to work that demonstrates that this method is always accurate to assess the quantities of CHC in extracts of known volumes.

We actually did include 7,5 ng/μl dodecane (C12) as an “internal” standard in the hexane resuspensions of all of our processed samples (line 456, Materials and Methods). This was primarily done to allow for visually inspecting and comparing the congruence of all chromatograms in the subsequent data analysis and immediately detect any variation from sample preparation, injection process and instrument fluctuation. In our study, we have a very elaborate and standardized CHC extraction method that the volume of solvent and duration for extraction are strictly controlled to minimize the variation from sample preparation steps. Furthermore, we calibrated each individual CHC compound quantity with a dilution series of external standards (C21-C40) of known concentration. By constructing a calibration curve based on this dilution series, we achieved the most accurate compound quantification, also taking into account and counteracting the generally diminishing quantities of compounds with higher chain lengths.

The authors provide a sensible control for their RNAi experiments: targeting an unrelated gene, absent in N. vitripennis (the GFP). This allows us to see if the injection of RNAi might affect CHC profiles, which it appears to do in some cases in males, but not in females. The authors also show to the reader that their RNAi experiments do reduce the expression of the target gene. However, one of the caveats of their experiments, is that the authors don't provide evidence or information to allow the (non-expert) reader to assess whether the fas5 RNAi experiments did affect the expression of other fatty-acyl synthase genes. I'm not an expert in RNAi, so maybe this suggestion is not relevant, but it should, at least, be addressed somewhere in the manuscript that such off-target effects are very unlikely or impossible, in that case, or more generally.

We acknowledge the reviewer’s concern about potential off-target effect of the fas5 knockdown. We actually did check initially for off-target effects on the other four previously published fas genes in N. vitripennis (Lammers et al. 2019; Buellesbach et al. 2022) and did not find any effects on their respective expressions. We now include these results as supplementary data (Figure 2-figure supplement 1). However, as mentioned in the cover letter to the editor, we discovered a previously uncharacterized fas gene in the most recent N. vitripennis genome assembly (NC_045761.1), fas6, most likely constituting a tandem gene duplication of fas5. These two genes turned out to have such high sequence similarity (> 90 %, Figure 2-figure supplement 2) that both were simultaneously downregulated by our fas5 dsRNAi construct, which we confirmed with qPCR and now incorporated into our manuscript (Fig. 2H). Therefore, we now explicitly mention that the knockdown affects both genes, and either one or both could have the observed phenotypic effects. Recognizing this RNAi off-target effect, we have now also incorporated a discussion of this issue in the appropriate section of the manuscript (line 364-377), as well as the potential off-target effects of our GFP dsRNAi controls (line 262-274).

The authors observe that the modified CHCs profiles of RNAi females reduce courtship and copulation attempts, but not antennation, by males toward live and (dead) dummy females. They show that the MB alkanes of the CHC profile are sufficient to elicit sexual behaviors from males towards dummy females and that the same fraction from extracts of fas5 knockdown females does so significantly less. From the previous data, it seems that dummy females with fas5 female's MB alkanes profile elicit more antennation than CHC-cleared dummy females, but the authors do not display data for this type of target on the figure for MB alkane behavioral experiments.

Actually similar proportions of males performed antennation behavior towards female dummies with MB alkane fraction of fas5 RNAi females and CHC-cleared female dummies (55% and 50%, respectively, see Author response image 1 for the corresponding parts of the sub-figures 3 E and 4 D). We did not deem it necessary to show the same data on CHC-cleared female dummies in Figure 3 as well.

Author response image 1.

Unfortunately, the authors don't present experiments testing the effect of the non-MB alkanes fractions of the CHC extracts on male behavior toward females. As such, they are not able to (and didn't) conclude that the MB-alkane is necessary to trigger the sexual behaviors of males. I believe testing this would have significantly enhanced the significance of this work. I would also have found it interesting for the authors to comment on whether they observe aggressive behavior of males towards females (live or dead) and/or whether such behavior is expected or not in inter-individual interactions in parasitoids wasps.

In our experiment, we focus on the function of the MB-alkane fraction in female CHC profiles, and we comprehensibly demonstrate in figure 4 that the MB-alkane fraction from WT females alone is sufficient to trigger mating behavior coherent with that on alive and untreated female dummies. Therefore, we do not completely understand the reviewer’s concern about us not being ” able to (and didn't) conclude that the MB-alkane is necessary to trigger the sexual behaviors of males”. We appreciate the suggestion from the reviewer of testing the non-MB alkanes (n-alkanes and n-alkenes). However, due to the experimental procedure of separating the CHC compound class fractions through elution with molecular sieves, it was not possible for us to retrieve either the whole n-alkane or n-alkene fraction remaining bound to the sieves after separation). The role of n-alkenes in N. vitripennis is however considered in the discussion, as a deterrent for homosexual interactions between males (Wang et al. 2022a). Moreover, we did not observe aggressive behavior of males towards live or dead females.

CHCs are used by insects to signal and/or recognize various traits of targets of interest, including species or groups of origin, fertility, etc. The authors claim that their experiments show the sexual attractiveness of females can be encoded in the specific ratio of MB alkanes. While I understand how they come to this conclusion, I am somewhat concerned. The authors very quickly discuss their results in light of the literature about the role of CHCs (and notably MB alkanes) in various recognition behaviors in Hymenoptera, including conspecific recognition. Previous work (cited by the authors) has shown that males recognize males from females using an alkene (Z9C31). As such, it remains possible that the "sexual attractiveness" of N. vitripennis females for males relies on them not being males and being from the right species as well. The authors do not address the question of whether the CHCs (and the MB alkanes in particular) of females signal their sex or their species. While I acknowledge that responding to this question is beyond the scope of this work, I also strongly believe that it should be discussed in the manuscript. Otherwise, non-specialist readers would not be able to understand what I believe is one of the points that could temper the conclusions from this work.

We acknowledge the reviewer’s insight about the MB alkanes in signaling sex or species in N. vitripennis, and now include this aspect in our revised discussion (line 324-330). Moreover, we clearly demonstrate that n-alkenes have been reduced to minute trace components after our compound class separation, and the males still do not display courtship and copulation behaviors similar to WT females, thus strongly indicating that the n-alkenes do not play a role when relying solely on the changed MB-alkane patterns, further strengthening our main argument.

References

Benjamini, Y. and D. Yekutieli. 2001. The control of the false discovery rate in multiple testing under dependency. Ann. Stat. 29:1165-1188.

Buellesbach, J., J. Gadau, L. W. Beukeboom, F. Echinger, R. Raychoudhury, J. H. Werren, and T. Schmitt. 2013. Cuticular hydrocarbon divergence in the jewel wasp Nasonia: Evolutionary shifts in chemical communication channels? J. Evol. Biol. 26:2467-2478.

Buellesbach, J., C. Greim, and T. Schmitt. 2014. Asymmetric interspecific mating behavior reflects incomplete prezygotic isolation in the jewel wasp genus Nasonia. Ethology 120:834-843.

Buellesbach, J., H. Holze, L. Schrader, J. Liebig, T. Schmitt, J. Gadau, and O. Niehuis. 2022. Genetic and genomic architecture of species-specific cuticular hydrocarbon variation in parasitoid wasps. Proc. R. Soc. B 289:20220336.

Engl, T., N. Eberl, C. Gorse, T. Krüger, T. H. P. Schmidt, R. Plarre, C. Adler, and M. Kaltenpoth. 2018. Ancient symbiosis confers desiccation resistance to stored grain pest beetles. Mol. Ecol. 27:2095-2108.

Ferveur, J. F., J. Cortot, K. Rihani, M. Cobb, and C. Everaerts. 2018. Desiccation resistance: effect of cuticular hydrocarbons and water content in Drosophila melanogaster adults. Peerj 6.

Lammers, M., K. Kraaijeveld, J. Mariën, and J. Ellers. 2019. Gene expression changes associated with the evolutionary loss of a metabolic trait: lack of lipogenesis in parasitoids. BMC Genom. 20:309.

Mair, M. M., V. Kmezic, S. Huber, B. A. Pannebakker, and J. Ruther. 2017. The chemical basis of mate recognition in two parasitoid wasp species of the genus Nasonia. Entomol. Exp. Appl. 164:1-15.

Wang, Y., W. Sun, S. Fleischmann, J. G. Millar, J. Ruther, and E. C. Verhulst. 2022a. Silencing Doublesex expression triggers three-level pheromonal feminization in Nasonia vitripennis males. Proc. R. Soc. B 289:20212002.

Wang, Z., J. P. Receveur, J. Pu, H. Cong, C. Richards, M. Liang, and H. Chung. 2022b. Desiccation resistance differences in Drosophila species can be largely explained by variations in cuticular hydrocarbons. eLife 11:e80859.

Author Response

Reviewer #1 (Public Review):

The work described herein would have an impact on the field in multiple ways. Firstly, it demonstrates a novel metabolic role for MSH in the regulation of hepatic cholesterol metabolism. This may prove to be a viable therapeutic strategy for the treatment of dyslipidemia. Furthermore, the authors demonstrate an alternative signaling cascade elicited by MSH independent of cAMP, but rather relying on AMPK. This novel interaction between AMPK and MC1R could have more widespread implications beyond the control of hepatic cholesterol metabolism.

For the most part, the conclusions offered by the authors are supported by the data that is presented. There are, however, a number of concerns in the current version of this manuscript detailed below.

We thank the reviewer for the encouraging and insightful comments, and we are pleased to read that the manuscript has raised considerable interest.

1) The authors demonstrate the expression of MC1R in hepatocytes through IHC staining and western blot analysis. Furthermore, the authors show an alteration in systemic bile acid homeostasis in MC1R KO mice. However, no mention of MC1R expression or function in cholangiocytes is discussed. This is important to assess both experimentally and within the discussion given the profound role of the biliary epithelium in modulating bile acid homeostasis. Furthermore, in figure 1 the authors validate the MC1R knockdown only through mRNA expression. Given panels A and C of figure 1 shows there is clearly a functional antibody for MC1R, validation of protein knockdown is needed.

The reviewer raises an important point, which we addressed by performing immunofluorescence staining using an antibody against the cholangiocyte marker cytokeratin 19 (CK-19). These colocalization studies demonstrate the presence of MC1-R in CK19-positive cholangiocytes (Figure 1-figure supplement 1). Furthermore, we have now added a discussion on the possible role of MC1-R in modulating bile acid homestasis in cholangiocytes (page 12, lines 456-462).<br /> We also quantified MC1-R protein expression by Western blotting in the liver of LMc1r-/- mice. MC1-R protein level was significantly reduced in L-Mc1r-/- mice compared to L-Mc1+/- mice (Figure 2-figure supplement 2).

2) Figure 2 demonstrates a steatotic effect of MC1R knockdown in hepatocytes. The authors attempt to provide mechanistic insight into this phenomenon through assessing the mRNA expression of genes involved in cholesterol and fatty acid synthesis. The data provided is modest at the gene level and no protein validation was provided to demonstrate functional alterations of these proteins in MC1R KO mice. Key proteins proposed such as SREBP2 and HMGCR need to be validated via a western blot of IHC analysis.

As requested by the reviewer, we quantified the expression of key proteins in the liver of L-Mc1r-/- mice by Western blotting. We observed that the protein levels of HMGCR and DHCR7 as well as the ratio between the mature and precursor forms of SREBP2 were reduced in L-Mc1r-/- mice (Figure 2F-H, page 6/lines 182-191 & page 10-11/lines 390-401). This is likely a result of the feedback regulation, whereby cholesterol accumulation suppresses the cleavage of SREBP2 and leads to a consequent downregulation of the key cholesterol synthesis enzymes such as HMGCR and DHCR7 (Brown S & Goldstein JL, Cell. 1997 May 2;89(3):331-40).

We discussed in the original submission (page 11) as follows: ‘In the presence of excess cellular cholesterol, transcriptional induction and posttranslational activation of SREBP-2 should be attenuated, which in turn downregulates Hmgcr and Dhcr7 and reduces cholesterol synthesis as a counterregulatory mechanism. Therefore, given the increase in hepatic cholesterol content, it was unexpected that Srebp2 expression was upregulated in the liver of L-Mc1r-/- mice’. The finding of reduced SREBP2/HMGCR protein expression is thus more logical, but admittedly, it is discordant with increased Srebp2/Hmgcr mRNA expression (as reported in the original submission), which might be a compensatory response to suppressed SREBP2 cleavage. Taking into account that activation of MC1-R did not affect the protein expression of HMGCR or DHCR7 in HepG2 cells, it is plausible that hepatic cholesterol accumulation in L-Mc1r-/- mice is driven by a defect in bile acid metabolism, rather than by a direct effect of MC1-R signaling on cholesterol synthesis. To avoid unnecessary confusion, we decided to omit the qPCR data and related text parts from the manuscript and report the protein expression data instead.

4) The authors suggest the involvement of AMPK in mediating the cholesterol-lowering effects of MSH. However, MSH is still able to lower free cholesterol levels even in the presence of an AMPK inhibitor. This suggests that MSH does not in fact rely on the activation of AMPK to elicit these cholesterol-lowering effects. The authors' conclusions are stronger than the actual data support. Furthermore, the authors claim LD211 phenocopies the effects of MSH in the presence of an AMPK inhibitor. However, the authors only measured the phosphorylation of Akt as their outcome. This begs the question, does LD211 still lower total cholesterol in the presence of AMPK inhibitors? This experiment is essential to conclude whether or not LD211 phenocopies the effects of MSH.

The reviewer may have missed that we postulate in the manuscript that ‘MC1-R activation engages multiple signaling mechanisms to regulate cholesterol metabolism in HepG2 cells’ (manuscript page 8, lines 310-311 & page 13, lines 498508), since low concentration of a-MSH was still able to lower free cholesterol level in the presence of the AMPK inhibitor dorsomorphin. We have been careful not to claim that the effects of a-MSH are solely dependent on AMPK phosphorylation. Likewise, we have not claimed in the original submission that LD211 phenocopies the effects of MSH in the presence of an AMPK inhibitor. However, as suggested by the reviewer, we performed new experiments to investigate the effects of LD211 on cellular cholesterol levels in the absence and presence of dorsomorphin. We found that AMPK inhibition with dorsomorphin completely abolished the cholesterollowering effect of LD211 (Figure 7-figure supplement 2), which might indicate that this synthetic agonist has a stronger signaling bias toward the AMPK pathway compared to α-MSH.

5) The authors initiate the project by showing high-fat diet disrupts the expression of MC1R. However, all of the subsequent experiments in hepatic MC1R KO mice are performed under normal chow. This begs the question of what is the phenotype of the hepatic MC1R KO mice fed a high-fat diet. Does KO of MC1R in the liver exacerbate HFD-induced obesity, glucose intolerance, and dyslipidemia? Inversely, can WT mice challenged with an HFD be rescued metabolically by treatment with either MSH or LD211? Providing data along these lines of investigation will provide physiological/clinical relevance to their findings.

As suggested by the reviewer, we phenotyped the hepatic MC1R KO (LMc1r-/-) mice after feeding them a cholesterol- and fat-rich Western diet for 12 weeks (RD Western Diet, D12079B, Research Diets Inc, NJ, USA). This was exactly the same dietary regimen (product and duration) that was used to study the changes in hepatic MC1-R expression in wild-type C57Bl mice (Figure 1B&C). We observed that 12-week Western diet feeding induced a significant gain in body weight and total fat mass as well as an increase in plasma and hepatic cholesterol and TG levels (Figure 2-figure supplement 2). L-Mc1r-/- mice did not show a difference in body weight gain, but the weight gain was attributable to enhanced gain in fat mass and a blunted increase in lean mass compared to control Mc1rfl/fl mice (Figure 2-figure supplement 2A, D & E). Furthermore, liver weight and plasma cholesterol and TG concentrations were unchanged in HFD-fed L-Mc1r-/- mice (Figure 2-figure supplement 2B, C, F & G). Importantly, recapitulating the phenotype observed in chow-fed mice, hepatic cholesterol and TG content was significantly increased in LMc1r-/- mice after a HFD challenge (Figure 2-figure supplement 2H & I). Taken together, it appears that the phenotype of HFD-fed L-Mc1r-/- mice was slightly diluted compared to the phenotype observed in chow-fed L-Mc1r-/- mice. This phenotypic difference might relate to the finding that Western diet feeding reduced the hepatic expression of MC1-R, thus limiting the incremental effect of genetically induced MC1-R deficiency on hypercholesterolemia and hepatic lipid accumulation.

We have previously studied the effects of pharmacological MC1-R activation in Western diet-fed mice and observed that chronic treatment with a selective MC1-R agonist reduced plasma cholesterol level and upregulated hepatic Ldlr expression without affecting body weight gain (Rinne P et al, Circulation. 2017 Jul 4;136(1):8397.). These findings are also discussed on manuscript page 12, lines 475-478. Although the selective MC1-R agonist was different in that particular study, it is expected that LD211 would also elicit a similar cholesterol-lowering effect in Western diet-fed mice. Chronic treatment with a-MSH, on the other hand, would likely produce wide-ranging metabolic effects. In addition to MC1-R activation in hepatocytes and its consequent effect on liver cholesterol metabolism, a-MSH would affect feeding, energy expenditure and cholesterol metabolism via MC4-R activation in the central nervous system as well as fatty acid and glucose metabolism via MC5-R activation in the skeletal muscle. Therefore, the phenotype associated with a-MSH treatment would be complex and mediated by multiple mechanisms and MC-R subtypes, thus making it difficult to interpret the exact contribution of hepatic MC1-R signaling to the observed phenotype.

Reviewer #2 (Public Review):

Keshav Thapa et al. investigated the role of melanocortin 1 receptor (MC1-R) in cholesterol and bile acid metabolism in the liver. First, they observed that MC1-R is present in the mouse liver and that its expression is reduced in response to a cholesterolrich diet. To determine the role of MC1-R in the liver, they generated hepatocyte-specific MC1-R KO mice (L-Mc1r-/-). These animals exhibited a significant increase in liver weight, lipid accumulation, triglycerides and cholesterol levels, and fibrosis in comparison with control mice. By performing liquid chromatography-mass spectrometry, the authors also found that L-Mc1r-/- mice also have fewer bile acids in the plasma and faeces, but not in the liver. In accordance with these findings, mRNA/protein expression of different genes involved in these processes were altered in L-Mc1r-/- animals.

Secondly, in an attempt to evaluate the underlying mechanisms, they measured the expression of MC1-R in HepG2 cells under different treatments (i.e., palmitic acid, LDL, and atorvastatin). Moreover, they stimulated these cells with the endogenous MC1-R agonist - MSH, where they show that this molecule decreases the free cholesterol content, whereas increasing LDL and HDL uptake, as well as recapitulates some previously observed phenotypes in the proportions of bile acids. These effects were also encountered when using a selective agonist for MC1-R (i.e., LD211), further supporting the specific role of MC1-R. Finally, some experiments indicated that -MSH evokes not one single, but multiple intracellular signalling cascades for which MC1-R activation effects might take place.

Overall, this work provides novel and interesting findings on the role of MC1-R in cholesterol and bile acid metabolism in the liver, which undoubtedly will have some crucial implications for future research. Nevertheless, some experimental details should be better explained for the correct interpretation of the data. Besides, discrepant results exist regarding the molecular mechanisms behind MC1-R action that requires additional experimentation to support the conclusions drawn.

We thank the reviewer for the encouraging and insightful comments, and we are pleased to read that the manuscript has raised considerable interest.

Author Response

Reviewer #1 (Public Review):

The authors aim to understand the role of clonal heterogeneity of tumors in immunogenicity of clonally expressed antigens. This is a significant problem with many basic as well as translational implications.

The strength of the manuscript lies in the novel demonstration that a poorly immunogenic tumor antigen, when paired with a stronger tumor antigen, begins to elicit significant immune response. The weakness lies in the fact that the actual mechanism of the key demonstration is never shown. There is a lot of speculation and tangential experimentation, but little actual evidence of a mechanism.

By making the key observation (mentioned in the strength section in the previous paragraph), the authors did achieve their objective albeit very partially. Their observation is based on excellent experimental tools and design. This study will stimulate further experiments in this important field.

Their key observation is somewhat reminiscent of the practice of conjugating small "non-immunogenic" antigens (such as some carbohydrates) to large protein carriers (such as serum albumin) in order to elicit strong antibody response to the weaker antigen. It is interesting to contemplate if the underlying mechanisms have any commonality.

We thank the reviewer for their consideration of our work and their constructive feedback. We concur that our study has limitations and further work will be necessary to fully deconstruct the mechanism leading to the observed phenotype. We have revised the text to better reflect the aim and scope of our study. However, the goal of our work was to establish a trackable model that would allow us to model different, albeit limited, degrees of antigen expression patterns reflecting what is observed in patients with different levels of ITH. Our key observation reproduces what is observed clinically, adding strength to the model. Next, we wanted to study what was different about the induced immune responses to develop strategies to better treat tumors with heterogeneous NeoAg expression patterns that currently do not respond to checkpoint blockade therapy. Studying KP-HetHigh and KP-HetLow tumors revealed that tumor debris-carrying cDC1 draining from KP-HetLow tumors phagocytosed both NeoAgs. This population of cDC1, carrying both NeoAgs, had a more stimulatory phenotype compared to cDC1 without tumor debris or cDC1 that had engulfed only one NeoAg. We were able to develop a targeted therapy including CD40 agonism based on our key observations: KP-HetLow had a more robust response towards the weaker NeoAg which was associated with more stimulatory cDC1 presenting both NeoAgs compared to KP-HetHigh tumors. The stronger immune response increased responsiveness to CBT.

The reviewer makes an interesting point about conjugate vaccines, which canonically elicit greater responses because they engage multiple immune cells, namely T cells with B cells, resulting in stronger antibody responses. The prevalence of tumor debris-carrying cDC1 with both neoantigens in KP-HetLow does make us consider that this population of cDC1 may be engaging multiple immune populations, i.e., different neoantigen-specific T cells. We suggest this as a possible mechanism for greater Aatf responses, but further work is necessary to determine if the same cDC1 can directly interact with both neoantigen-specific T cells.

Reviewer #2 (Public Review):

There are data to suggest that intratumour mutational heterogeneity (ITH; the proportion of all mutations that are found only within cancer subclones) is associated with worse therapeutic outcomes. Specifically, patients with more mutations (and thus neoantigens) mostly expressed by subclones (high ITH) have poorer responses to checkpoint immunotherapy. The authors set out to explore the mechanisms underlying this by studying 2 dimensions of neoantigen biology: firstly, distribution (clonal vs subclonal) and secondly, immunogenicity (weak vs strong binding to MHC class I). Using a panel of lung cancer cell lines modified to express individual or dual neoantigens in order to model clonal and subclonal expression, elegant studies show that clonal co-expression with a "strong" neoantigen can boost the immunogenicity of a "weak" neoantigen and result in tumour control. Mechanistically, this is related to engulfment of both neoantigens by cross presenting type 1 conventional dendritic cells and the associated enhanced activation state of this cell type. This is an interesting and potentially important finding that may be related to mechanisms of epitope spreading as immune responses diverge from targeting more to less immunogenic epitopes. Overall, the study is thought-provoking, informative in relation to how neoantigen immunogenicity is shaped and may have practical relevance.

We greatly appreciate the constructive comments from the reviewer and their insightful comments and questions on our work. We have edited the text in response to their feedback. We believe these changes have made the writing clearer and more effectively communicates the scope of our study and our results to the reader.

Author Response:

We would like to thank the Editors and Reviewers for their positive evaluations, constructive comments, and for the opportunity to revise our manuscript. We feel that the comments and suggestions will further improve our manuscript.

In the updated manuscript we aim to incorporate all suggested changes and considerations provided by the Reviewers. In particular, we will provide further information on the quality-control ratings per subfield, as suggested by Reviewer 1. Moreover, we will evaluate whether the training-related changes were specific to CA1-3, rather than just showing significant alterations in CA1-3 and not in the other subfields. Last, as suggested by Reviewer 2, we will additionally test for multivariate associations between hippocampal subfield structure and function, to further evaluate the specificity of hippocampal subfield change as a function of training and cortisol.

Author Response

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

Reviewer #1 (Recommendations For The Authors):

This study is well presented and contains all the necessary experiments to support their claims. They made the interesting finding of an additional factor Dyn2. However, it is unclear whether it is present in the human complex. Hence, it would be interesting to see whether Dyn2 co-purifies when expressed with the other complex components in insect cells. Also, purification of a tagged complex from yeast would have indicated whether Dyn2 is part of the complex and whether other factors, like RBM15 or Hakai, present in humans are also present in yeast.

We agree that Dyn2 subunit is an exciting new finding that is worth further investigation. The IP-MS experiments suggest that Dyn2 is subunit of the complex and that the Dyn2 interaction is mediated via Slz1. We also noticed a reduction in m6A levels (50%) in the dyn2 deletion mutant. What the function of Dyn2 is and whether it is conserved remains to be determined.

Our IP-MS experiments with Mum2 identified the complex as described in the manuscript, however we did not find evidence of orthologs of RBM15 and Hakai. More follow up work is needed using in vivo and in vitro assays are needed to determine how m6A by the yeast MTC is regulated.

P3 top: Although m6A is the most abundant internal methylation variant, it is far below the methylation levels of cap-adjacent nucleotides in mammalian mRNAs (PMID: 35970556 ).

We have added the word “internal” to the first sentence of the introduction.

A list of author contributions is missing.

We have added this in the revised version.

Reviewer #2 (Recommendations For The Authors):

Most of the conclusions of this paper are well supported by data, and the text is clearly written and easy to read. Here are my suggestions and comments:

1) In Fig.2, why not use LC-MS to measure m6A levels in Ygl036w, Dyn2, Pab1, Npl3 mutants, as in Fig.1?

For measuring m6A levels, we use combination of LC-MS and m6A ELISA and m6A-seq2 throughout the manuscript. We used ELISA in the Fig2 because we had established this assay in the lab (Ensinck et al, RNA Journal, 2023). M6A-ELISA technique was more accessible and easier to execute compared to LC-MS. Additionally our collaborator for the LC-MS moved his lab to another country, which made it impractical to continue the use of LC-MS.

2) The protein purification experiment described in Fig. 4D is informative. Can they include Dyn2 in the expression system as well?

Thank you for the suggestion. Dyn2 was not the focus of the manuscript as Dyn2 has, at best, only a minor role in m6A deposition in vivo. We are also currently aiming to dissect how Dyn2 regulates m6A and the yeast MTC in follow up work. Hence we decided not to add more experiments on Dyn2 to the current manuscript.

3) Among the MTC components identified in this study, Dyn2 is a new and interesting subunit. It was shown that in C. elegans Dlc1 is involved in stabilizing the m6A writer Mett10. I wonder if yeast has a homolog of C. elegans Mett10?

As far as we know, there is no ortholog identified of Mett10 (METTL16 in mammals) in budding yeast.

4) The authors have emphasized "the m6A dependent and independent functions"; however, this is only based on previous observations. Is it possible that the less severe phenotype associated with ime4 catalytic mutant is due to residual catalytic activity? I think the data presented in Fig. 5 tell us that Ime4 and other MTC subunits have no additional moonlighting function. It is not entirely clear to me what "the m6A-independent function" is.

The observation that the yeast MTC complex has m6A dependent and independent function is based on the previous observations and the current work. In Agarwala et al 2012 PLOS Genetics, it was shown that mum2 and ime4 deletion mutants have more severe phenotype than slz1 deletion mutant or the catalytically inactive mutant of Ime4. We confirmed these observations in the revised manuscript (see Figure S5A and S5B). In this work, we showed that kar4 and vir1 deletion mutants have comparable delay in the onset of meiosis as mum2 and ime4 deletion mutants. Also, the MTC remains intact with absence of Slz1, but falls apart in ime4D, mum2D, vir1D or showed strongly reduced RNA binding (kar4 deletion mutant). Based on this we conclude that an m6A independent function of the MTC exists.

We have included data demonstrating that the catalytically inactive mutant has no residual m6A and a milder meiotic phenotype compared to the ime4 deletion mutant (see Figure S5A and S5B).

5) In Mum2-TEV-ProA IP (1B) and Kar4-TEV-ProA IP (S1A), Slz1 was not significantly enriched; however, in the repeated Mum2-TEV-ProA IP with/without RNAse (S1B, 4C), Slz1 was strongly enriched. Why are the Slz1 results so variable?

This is an astute observation, for which we do not have a definitive answer. One possibility is that Slz1 is the only subunit that is induced during meiosis. It is possible that induction of Slz1 varied between the different IP-MS experiments, hence leading to variability in its association with the MTC complex.

6) The last paragraph on page 11, "Collectively...", and the first paragraph on page 12, "Collectively...", seem redundant.

We have removed the duplicated paragraph in the revised manuscript.

Author Response

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

Reviewer #1 (Public Review):

MCM8 and MCM9 are paralogues of the eukaryotic MCM2-7 proteins. MCM2-7 form a heterohexameric complex to function as a replicative helicase while MCM8-9 form another hexameric helicase complex that may function in homologous recombination-mediated longtract gene conversion and/or break-induced replication. MCM2-7 complex is loaded during the low Cdk period by ORC, CDC6, and Cdt1, when the origin DNA may intrude into the central channel via the MCM2-MCM5 entry "gate". In the S phase, MCM2-7 complex is activated as CMG helicase with the help of CDC45 and GINS complex. On the other hand, it still remains unclear how MCM8-9 complex is loaded onto DNA and then activated.

In this study, the authors first investigated the cryo-EM structure of chicken MCM8-9 (gMCM89) complex. Based on the data obtained, they suggest that the observed gMCM8-9 structure might represent the structure of a loading state with possible DNA entry "gate". The authors further investigated the cryo-EM structure of human MCM8-9 (hMCM8-9) complex in the presence of the activator protein, HROB, and compared the structure with that obtained without HROB1, which the authors published previously. As a result, they suggest that MCM8-9 complex may change the conformation upon HROB binding, leading to helicase activation. Furthermore, based on the structural analyses, they identified some important residues and motifs in MCM8-9 complex, mutations of which actually impaired the MCM8-9 activity in vitro and in vivo.

Overall, the data presented would support the authors' conclusions and would be of wide interest for those working in the fields of DNA replication and repair. One caveat is that most of the structural data are shown only as ribbon model without showing the density map data obtained by cryo-EM, which makes accurate evaluation of the data somewhat difficult.

We thank the reviewer for the positive comments on our work. For evaluating all the structural data, in our revised manuscript, we have presented the density maps of the cryo-EM structures of the gMCM8/9 complex in supplementary figure S5 and S6. In addition, the 3D cryo-EM map of the gMCM8/9 complex and the hMCM8/9 NTD ring have been deposited to the EMDB database with accession number EMD-32346 and EMD-33989, respectively. The corresponding atomic models have been deposited at the RSCB PDB under the accession code 7W7P and 7YOX, respectively. All these data have been released in May 2023.

Reviewer #2 (Public Review):

MCM8 and MCM9 together form a hexameric DNA helicase that is involved in homologous recombination (HR) for repairing DNA double-strand breaks. The authors have previously reported on the winged-helix structure of the MCM8 (Zeng et al. BBRC, 2020) and the Nterminal structure of MCM8/9 hexametric complex (MCM8/9-NTD) (Li et al. Structure, 2021). This manuscript reports the structure of a near-complete MCM8/9 complex and the conformational change of MCM8/9-NTD in the presence of its binding protein, HROB, as well as the residues important for its helicase activity.

The presented data might potentially explain how MCM8/9 works as a helicase. However, additional studies are required to conclude this point because the presented MCM8/9 structure is not a DNA-bound form and HROB is not visible in the presented structural data. Taking into these accounts, this work will be of interest to biologists studying DNA transactions.

A strength of this paper is that the authors revealed the near-complete MCM8/9 structure with 3.66A and 5.21A for the NTD and CTD, respectively (Figure 1). Additionally, the authors discovered a conformational change in the MCM8/9-NTD when HROB was included (Figure 4) and a flexible nature of MCM8/9-CTD (Figure S6 and Movie 1).

The biochemical data that demonstrate the significance of the Ob-hp motif and the N-C linker for DNA helicase activity require careful interpretation (Figures 5 and 6). To support the conclusion, the authors should show that the mutant proteins form the hexamer without problems. Otherwise, it is conceivable that the mutant proteins are flawed in complex formation. If that is the case, the authors cannot conclude that these motifs are vital for the helicase function.

A weakness of this paper is that the authors have already reported the structure of MCM8/9NTD utilizing human proteins (Li et al. Structure, 2021). Although they succeeded in revealing the high-resolution structure of MCM8/9-NTD with the chicken proteins in this study, the two structures are extremely comparable (Figure S2), and the interaction surfaces seem to be the same (Figure 2).

Another weakness of this paper is that the presented data cannot fully elucidate the mechanistic insights into how MCM8/9 functions as a helicase for two reasons. 1) The presented structures solely depict DNA unbound forms. It is critical to reveal the structure of a DNA-bound form. 2) The MCM8/9 activator, HROB, is not visible in the structural data. Even though HROB caused a conformational change in MCM8/9-NTD, it is critical to visualize the structure of an MCM8/9HROB complex.

We appreciate the reviewer’s comments on our work. Regarding the first weakness mentioned above, the previously reported cryo-EM structure of hMCM8/9 NTD ring was achieved with a resolution of 6.6 Å. At this level of resolution, we were only able to observe the overall shape of the structure and a partial representation of the protein's secondary structure. It is hard for us to discern any specific details regarding the interaction interface between MCM8 and MCM9. In this study, we solved the structure of gMCM8/9 NTD ring with a resolution of 3.67 Å. We believe that the higher resolution of gMCM8/9 NTD structure provides a significant advantage in analyzing the interaction surface between MCM8 and MCM9. This improved resolution has enabled us to gain valuable insights into the assembly mechanism of the MCM8/9 hexamer, representing a significant step forward in our understanding of the MCM8/9 helicase complex. In response to the second weakness raised by the reviewer, we fully agree with the reviewer that high-resolution structures of the MCM8/9 complex with DNA or HROB are necessary to elucidate the mechanism of this helicase complex. We are actively working towards obtaining these complex structures using cryo-EM and X-ray crystal diffraction.

Moreover, we would like to address the reviewer's concern regarding the mutant proteins used in the in vitro helicase assays. We have conducted additional experiments to confirm that these mutant proteins do not impair the formation of the MCM8/9 hexamer. Specifically, we performed size exclusion chromatography (SEC) analyses of the wild-type (WT) MCM8/9 complex, as well as MCM8 and MCM9 mutant proteins (Author response image 1). The results demonstrated that all the proteins behaved consistently and displayed similar SEC profiles during the purification process. Notably, the N-C linker deletion mutant (hMCM8_Δ369-377+MCM9_Δ283-287) combining the MCM8 and MCM9 N-C linker deletions also behaved similarly with WT MCM8/9 (Author response image 2). These findings strongly suggest that the mutations in the OB-hps regions and the N-C linkers do not disrupt the hexamer formation of the MCM8/9 complex. Author response image 1 and Author response image 2 have been included into the supplementary figure S8 and S11, respectively.

Author response image 1.

SEC profiles of WT and OB-hps mutants of MCM8/9 complex.

Author response image 2.

SEC profiles of WT and N-C linker mutant of MCM8/9 complex.

Reviewer #1 (Recommendations For The Authors):

I would like to provide some suggestions to improve the manuscript.

1) Throughout the manuscript, more density map data obtained by the cryo-EM should be shown for accurate evaluation of the data. For example, in Figure 1C, the authors state that inner channel of the gMCM8-9 hexamer is ~28 angstrom, apparently based on the ribbon model. This is not appropriate because the space upon ribbon model is not same as that upon the density map. For Figure 1B, they state that "The domain structures of gMCM8-9 fit well into their electron map". If so, please show the actual docking data. Also for Figure 2, the docking presentation between the side chains in the ribbon model and the density map should be shown.

We sincerely appreciate the reviewer for the constructive suggestions. In addition to releasing our structural data in the EMDB and PDB, we have also followed the reviewer’s suggestions to included more density map data in the supplementary material. In fact, when calculating the dimeter of the inner channel of the MCM8/9 hexamer, we also measured that upon the density map (Author response image 3. A and B), which is consistent with our report in our manuscript. To further evaluate the structure of MCM8/9, we have included additional docking structures based on the density map (Author response image 3. C-F). Moreover, for Figure 2, more docking presentation are provided and the key residues involved in the hydrophobic interactions were highlighted in a bold manner (Author response image 4). Author response image 3 and Author response image 4 have been included into the supplementary figure S5 and S6, respectively.

Author response image 3.

The cryo-EM structure of gMCM8/9. (A and B) Reconstructed cryo-EM map of gMCM8/9. The diameter of the inner channel of MCM8/9 was measured at ~28 Å. (C-F) Representative regions of the cryo-EM structure of gMCM8/9 NTD are shown based on their density map. C, chain A (MCM9); D, chain B (MCM8); E, chain C (MCM9); F, chain D (MCM8).

Author response image 4.

Representative regions of the cryo-EM structure of gMCM8/9 NTD. (A and B), the region mediated hydrophobic interaction in figure 2B. A (MCM8), B (MCM9). (C and D), the region mediated hydrophobic interaction in figure 2C. C (MCM8), D (MCM9). The key residues were in bold.

2) Figures 4, 5, and 6: For helicase assay, more detailed experimental conditions (e.g. concentrations of DNA substrates and proteins used) should be presented. In addition, it should be described how Flag-hMCM8-9 complex (Figure 4C) was purified.

We sincerely appreciate the constructive suggestion provided by the reviewer. In the revised manuscript, we have included more experimental details in the helicase assays, including the concentrations of DNA substrates and proteins. The following paragraph describes the updated experimental procedure and also provided in the revise version of the manuscript.

Helicase assays: To prepare the substrate, the oligonucleotide (5'(dT)40GTTTTCCCAGTCACGACG-TTGTAAAACGACGGCCAGTGCC-3') containing a 40 nt region complementary to the M13mp18(+) stand and a 40 nt oligo-dT at the 5′ end was labeled at the 3′ terminus with [α-32P] dCTP (Perkin Elmer) and annealed to the single-stranded DNA M13mp18 (24). 0.1 nM (in molecules) DNA substrates were respectively mixed with 5 µg recombinant MCM8/9 complex and its mutants as indicated within each 15 µl volume reaction in the helicase buffer (25 mM HEPES, pH 7.5, 1 mM magnesium acetate, 25 mM sodium acetate, pH 5.2, 4 mM ATP, 0.1 mg/ml BSA, 1 mM DTT). 2.5 µg HROB was used as an activator. To avoid re-annealing, the reaction was supplemented with a 100-fold unlabeled oligonucleotide. The reactions were then incubated at 37 °C for 60 min and stopped by adding 1 µl of stop buffer (0.4% SDS, 30 mM EDTA, and 6% glycerol) and 1µl of proteinase K (20 mg/ml, Sigma) into the reaction for another 10 min incubation at 37 °C. The products were separated by 15% polyacrylamide gel electrophoresis in 1× TBE buffer and analyzed by the Amersham typhoon (Cytiva).

In addition, to describe the expression of Flag-hMCM8/9 complex in Figure 4C, we have included the Pull-Down Assay in the “Material and Methods” section. The description is as follow: The HEK293T cells transfected with Flag-hMCM8/9-FL or Flag-hMCM8/9-NTD were cultured overnight and washed twice with cold phosphate-buffered saline (PBS). Cell pellets were resuspended with lysis buffer (20 mM Tris, pH7.5, 150 mM NaCl, 5mM EDTA, 0.5% NP-40, 10% glycerol, protease inhibitor cocktail (Roche, 04693132001)). After incubation for 45 min at 4°C with gentle agitation, the whole-cell lysates were collected by centrifugation (12,000 × g for 15 min, at 4 °C). GST beads coupled with 2 μg GST-HROB or GST alone were then incubated with an equal volume of above HEK293T cell lysates at 4°C for 4h. The beads were washed four times with lysis buffer. Proteins bound to the beads were separated by SDS–PAGE and subsequently immunoblotted with anti-Flag antibody (Cytiva).

3) Figure 3C: This is just an assumed model. Please clearly state it in the manuscript.

We appreciate the reviewer’s comment. We guess the reviewer is referring to Figure 5C. As Figure 3C depicts the top view of the gMCM8/9 hexamer structurally aligned with the MCM2-7 double hexamer (wheat) by aligning their respective C-tier ring. On the other hand, Figure 5C represents an assumed model where we docked a forked DNA fragment into the central channel of the gMCM8/9 hexamer. To address this assumed model, we have made the following clarification in the revised manuscript: “We artificially docked a forked DNA into the central channel to generate a gMCM8/9-DNA model and found that the OB-hps of gMCM8 are capable to closely contact with it and insert their highly positively charged terminal loops into the major or minor grooves of the DNA strand, implying that they could be involved in substrate DNA processing and/or unwinding (Figure 5C)”.

4) Figure S1, C and D: The coloring of the gMCM8-9 CTD appears to show higher resolution than the NTD. May this be mispresentation?

We appreciate the reviewer's valuable feedback, and we have thoroughly re-evaluated Figure S1C and D. At the beginning, the local resolution distributions of the gMCM8/9 NTD and gMCM8/9 CTD were calculated using CryoSPARC. Upon re-examination, we found that the density maps of the gMCM8/9 CTD may be lower than 3.66 Å, because the density map of the gMCM8/9 CTD does not reveal more structural details than what is observed in the gMCM8/9 NTD. Thus, although the map shown in Figure S1D may appear to show a greater distribution of high-resolution regions., we would like to clarify that this discrepancy could be attributed to an optical illusion. We thank the reviewer for bringing this to our attention.

5) Figure S9: Is the "mean resolution" 5.21 angstrom identical to the Gold standard FSC? If not, please estimate the resolution using FSC, like other maps in this paper.

We thank the reviewer for the constructive suggestion. In response to this feedback, we would like to clarify the resolution estimation process for the gMCM8/9 CTD. Initially, we calculated the resolution of the gMCM8/9 CTD using the gold standard Fourier shell correlation (FSC) method, which yielded a resolution of 3.66 Å. However, upon further analysis, we identified an issue with the GSFSC Resolution curves, which led to an overestimation of the resolution based on the density map of the gMCM8/9 CTD. To ensure a more reliable and accurate estimation, we employed the Phenix software package to calculate the mean resolution during the refinement process of the gMCM8/9 CTD structure. The calculated mean resolution was determined to be 5.21 Å, which aligns more reasonably with the characteristics of the density map. To address any potential misunderstandings and provide clarity, we have explicitly labeled and described the evaluation process for this mean resolution in the "Single particle data processing" section of the Materials and Methods.

Minor points:

1) Throughout the manuscript, there are several typographical and grammatical errors, which should be corrected. For example, in "Introduction", "GNIS complex" should be "GINS complex".

We thank the reviewer for pointing out the typographical and grammatical errors. We have corrected the grammar errors and polished our manuscript with the help of native speakers.

Reviewer #2 (Recommendations For The Authors):

1) "During HR repair, MCM8/9 was rapidly recruited to the DNA damage sites and colocalized with the recombinase Rad51 (21). It also interacted with the nuclease complex MRN (MRE11RAD50-NBS1) and was required for DNA resection at DSBs to facilitate the HR repair (Introduction)."

There is a debate about whether MCM8/9-HROB colocalizes with RAD51 and whether it works upstream or downstream of RAD51 (Park et al. MCB, 2013; Lee et al. Nat Commun., 2015; Lutzmann et al. Mol Cell, 2012; Nishimura et al. Mol Cell, 2012; Natsume et al. G&D, 2017; Hustedt et al. G&D, 2019; Huang et al. Nat Commun., 2020).

We completely agree with the reviewer that previous studies have reported contradictory results regarding to the function of MCM8/9 in homologous recombination. Based on the structure information of MCM8/9, now we do not have direct evidence to resolve the ongoing debate. Nonetheless, based on our findings, we speculate that the MCM8/9 complex is likely involved in multiple steps within the process of homologous recombination. The structural insights provided by our study serve as a foundation for further investigations and may contribute to a better understanding of the complex and multifaceted roles of MCM8/9 in homologous recombination repair.

2) I noted that the BioRxiv version 1 (https://www.biorxiv.org/content/10.1101/2022.01.26.477944v1?versioned=true) contains a near-complete MCM8/9 with human protein based on the crystal analysis. Because its structure is comparable to chicken MCM8/9 revealed by cryo-EM, I highly suggest including this data in the manuscript.

We would like to thank the reviewer for this suggestion. The resolution of the hMCM8/9 crystal structure presented in our previous BioRxiv version is 6.6 Å, which is a little low. Moreover, it cannot provide more information than the present cryo-EM structures of MCM8/9. We are dedicated to optimizing the crystal quality and implementing strategies to enhance the resolution of the structure. We hope to present an improved crystal structure of hMCM8/9 in our forthcoming article.

Author Response

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

We thank the reviewers for their insightful comments. The main issue raised by the reviewers was that because E6AP depletion reduced checkpoint signaling vis MASTL upregulation, this pathway is likely to be involved also in DNA damage checkpoint activation, in addition to checkpoint recovery. Hence, the proposed “timer”-like model is not fully supported. However, it is important to note that, the expression level of MASTL is not upregulated during the activation stage of the DNA damage checkpoint (unless E6AP is depleted). DNA damage signaling, via ATM-dependent E6AP phosphorylation, caused MASTL accumulation over time. This ultimately shifts the balance toward checkpoint recovery and cell cycle re-entry. As such, the role of MASTL (and E6AP-depletion) in suppressing DNA damage checkpoint is in harmony with the proposed role of MASTL upregulation in promoting checkpoint recovery. We have made additional clarifications about this point in the revised manuscript.

We have also addressed other concerns raised by the reviewers, as explained in the point-to-point responses below. With the addition of new modifications and data, we believe the revised manuscript is complete and conclusive.

Reviewer #1 (Public Review):

In principle a very interesting story, in which the authors attempt to shed light on an intriguing and poorly understood phenomenon; the link between damage repair and cell cycle re-entry once a cell has suffered from DNA damage. The issue is highly relevant to our understanding of how genome stability is maintained or compromised when our genome is damaged. The authors present the intriguing conclusion that this is based on a timer, implying that the outcome of a damaging insult is somewhat of a lottery; if a cell can fix the damage within the allocated time provided by the "timer" it will maintain stability, if not then stability is compromised. If this conclusion can be supported by solid data, the paper would make a very important contribution to the field.

However, the story in its present form suffers from a number of major gaps that will need to be addressed before we can conclude that MASTL is the "timer" that is proposed here. The primary concern being that altered MASTL regulation seems to be doing much more than simply acting as a timer in control of recovery after DNA damage. There is data presented to suggest that MASTL directly controls checkpoint activation, which is very different from acting as a timer. The authors conclude on page 8 "E6AP promoted DNA damage checkpoint signaling by counteracting MASTL", but in the abstract the conclusion is "E6AP depletion promoted cell cycle recovery from the DNA damage checkpoint, in a MASTL-dependent manner". These 2 conclusions are definitely not in alignment. Do E6AP/MASTL control checkpoint signaling or do they control recovery, which is it?<br /> Also, there is data presented that suggest that MASTL does more than just controlling mitotic entry after DNA damage, while the conclusions of the paper are entirely based on the assumption that MASTL merely acts as a driver of mitotic entry, with E6AP in control of its levels. This issue will need to be resolved.

We thank the reviewer for his/her insightful comments. The main issue raised by the reviewers was that because E6AP depletion reduced checkpoint signaling vis MASTL upregulation, this pathway is likely to be involved also in DNA damage checkpoint activation, in addition to checkpoint recovery. Hence, the proposed “timer”-like model is not fully supported. However, it is important to note that, the expression level of MASTL is not upregulated during the activation stage of the DNA damage checkpoint (unless E6AP is depleted). DNA damage signaling, via ATM-dependent E6AP phosphorylation, caused MASTL accumulation over time. This ultimately shifts the balance toward checkpoint recovery and cell cycle re-entry. As such, the role of MASTL (and E6AP-depletion) in suppressing DNA damage checkpoint is in harmony with the proposed role of MASTL upregulation in promoting checkpoint recovery. We have made additional clarifications about this point in the revised manuscript.

As suggested by the reviewer, we have rephrased the statement in abstract to “E6AP depletion reduced DNA damage signaling, and promoted cell cycle recovery from the DNA damage checkpoint, in a MASTLdependent manner”.

As a mitotic kinase, MASTL promotes mitotic entry and progression. This is well in line with our findings that DNA damage-induced MASTL upregulation promotes cell cycle re-entry into mitosis. MASTL upregulation could also inhibit DNA damage signaling. This manner of feedback, inhibitory, modulation of DNA damage signaling by mitotic kinases (e.g., PLK1, CDK) has been implicated in previous studies (reviewed in Cell & Bioscience volume 3, Article number: 20 (2013)). In the revised manuscript, we have included more discussions about this aspect of checkpoint regulation.

Finally, the authors have shown some very compelling data on the phosphorylation of E6AP by ATM/ATR, and its role in the DNA damage response. But the time resolution of these effects in relation to arrest and recovery have not been addressed.

Detailed time point information is now added in the figure legends for E6AP phosphorylation data. We were able to observe this event during early stages (e.g., 1 hr, or 2-4 hr) of the DNA damage response, prior to significant MASTL protein accumulation.

Reviewer #2 (Public Review):

This is an interesting study from Admin Peng's laboratory that builds on previous work by the PI implicating Greatwall Kinase (the mammalian gene is called MASTL) in checkpoint recovery.

The main claims of this study are:

1) Greatwall stability is regulated by the E6-AP ubiquitin ligase and this is inhibited following DNA damage in an ATM dependent manner.

2) Greatwall directly interacts with E6-AP and this interaction is suppressed by ATM dependent phosphorylation of E6-AP on S218

3) E6-AP mediates Greatwall stability directly via ubiqitylation

4) E6-AP knock out cells show reduced ATM/ATR activation and quicker checkpoint recovery following ETO and HU treatment

5) Greatwall mediated checkpoint recovery via increased phosphorylation of Cdk substrates

In my opinion, there are several interesting findings presented here but the overall model for a role of the E6-AP -Greatwall axis is not fully supported by the current data and will require further work. Moreover, there are a number of technical issues making it difficult to assess and interpret the presented data.

Major points:

1) The notion that Greatwall is indeed required for checkpoint recovery hinges on two experiments shown in Figures 5A and B where Greatwall depletion blocks the accumulation of HELA cells in mitosis following recovery from ETO treatment and in G2/M following release from HU. An alternative possibility to the direct involvement of Greatwall in checkpoint recovery could be that Greatwall in HeLA cells is required for S-phase progression (as for example Charrasse et al. suggested). A simple control would be to monitor the accumulation of mitotic cells by microscopy or FACS following Greatwall depletion without any further checkpoint activation.

We thank the reviewer for his/her insightful comments.

Charrasse et al. showed ENSA knockout prolonged, but not stopped the progression of S-phase. In our experiments, MASTL (partial) knockdown did not significantly impact HeLa cells proliferation in the absence of DNA damage (Fig. 5, supplemental 1A). The reported role of MASTL in checkpoint recovery was consistently seen in response to various drugs, including etoposide which typically induces G2 arrest. Thus, we do not believe a prolonged S-phase accounts for the checkpoint recovery phenotype.

2) The changes in protein levels of Greatwall and the effects of E6-AP on Greatwall stability are rather subtle and depend mostly on a qualitative assessment of western blots. Where quantifications have been made (Figures 2D and 4F) the loading control and the starting conditions for Greatwall (0 timepoints in the right panel) appear saturated making precise quantification impossible. I would argue that the authors should at least quantify the immuno-blots that led them to conclude on changes in Greatwall levels and make sure that the exposure times used are in the dynamic range of the camera (or film). A more precise experiment would be to use the exogenously expressed CFP-Greatwall that is described in Figure 6 and measure the acute changes in protein levels using quantitative fluorescence microscopy in live cells. This is, in my opinion, a lot more trustworthy than quantitative immuno-blots.

I also note here that most experiments linking Greatwall levels to E6-AP were done using siRNA, while the E6-AP ko cells would be a more reliable background for these experiments, especially with reconstituted controls.

DNA damage-induced MASTL upregulation was observed in various cell lines and after different treatments. To further strengthen this point, as suggested by the reviewer, we have included quantification of fluorescent measurements (Fig. 2, supplemental 1 A-C). Quantification of immunoblots for MASTL upregulation was also added in Fig. 1, supplemental 1E. The effects of E6AP depletion were consistently shown for both siRNA and stable KO.

3) This study has no data linking the effects of Greatwall to its canonical target PP2A:B55. The model shown in Figure 9 is therefore highly speculative. The possibility that Greatwall could act independently of PP2A:B55 should at least be considered in the discussion given the lack of experimental evidence.

The role of MASTL in promoting cell cycle progression via suppressing PP2A/B55 has been well established. As suggested by the reviewer, we have included discussions to acknowledge that “The role of MASTL upregulation in promoting checkpoint recovery and cell cycle progression can be attributed to inhibition of PP2A/B55, although the potential involvement of additional mechanisms is not excluded”.

4) The major effect of E6-AP depletion on the checkpoint appears to be a striking reduction in ATM/ATR activation, suggesting that this ubiquitin ligase is involved in checkpoint activation rather than recovery. It is not clear if this phenotype is dependent on Greatwall. If so it would be hard to reconcile with the default model that E6-AP acts via the destabilisation of Greatwall. In the permanent absence of E6-AP, increased Greatwall levels should inactivate B55:PP2A. How would this lead to a decrease in ATM/ATR activation? This is unlikely, and indeed Figure 5E shows that the reduction of MASTL in parallel to E6-AP does not result in elevated levels of ATR/ATM activation. Conversely, the S215A E6-AP mutant does have a strong rescue impact on ATR/ATM (Figure 8D).

We do not propose that PP2A/B55 directly dephosphorylates ATM/ATR-mediated phosphorylation. In fact, PP2A/B55 dephosphorylates and inactivates mitotic kinases and substrates which can feedback inhibit DNA damage checkpoint signaling (as previously shown for PLK1 and CDK). We included in a discussion about this point in the revised manuscript.<br /> The point regarding checkpoint activation vs recovery is addressed below (point 5).

5) In summary, I do not think that the presented experiments clearly dissect the involvement of E6-AP and Greatwall in checkpoint activation and recovery. E6-AP depletion has a strong effect on checkpoint activation while Greatwall depletion is likely to have various checkpoint-independent effects on cell cycle progression.

It is important to note that, the expression level of MASTL is not upregulated during the activation stage of the DNA damage checkpoint (unless E6AP is depleted). DNA damage signaling, via ATM-dependent E6AP phosphorylation, caused MASTL accumulation over time. This ultimately shifts the balance toward checkpoint recovery and cell cycle re-entry. As such, the role of MASTL (and E6APdepletion) in suppressing DNA damage checkpoint is in harmony with the proposed role of MASTL upregulation in promoting checkpoint recovery. We have made additional clarifications about this point in the revised manuscript.

Reviewer #3 (Public Review):

In this manuscript, Li et al. describe the contribution of the ATM-E6AP-MASTL pathway in recovery from DNA damage. Different types of DNA damage trigger an increase in protein levels of mitotic kinase MASTL, also called Greatwall, caused by increased protein stability. The authors identify E3 ligase E6AP to regulate MASTL protein levels. Depletion or knockout of E6AP increases MASTL protein levels, whereas overexpression of E6AP leads to lower MASTL levels. E6AP and MASTL were suggested to interact in conditions without damage and this interaction is abrogated after DNA damage. E6AP was shown to be phosphorylated upon DNA damage on Ser218 and a phosphomimicking mutant does not interact with MASTL. Stabilization of MASTL was hypothesized to be important for recovery of the cell cycle/mitosis after DNA damage.

The identification of this novel pathway involving ATM and E6AP in MASTL regulation in the DNA damage response is interesting. However, is surprising that authors state that not a lot is known about DNA damage recovery while Greatwall and MASTL have been described to be involved in DNA damage (checkpoint) recovery. In addition, PP2A, a phosphatase downstream of MASTL is a known mediator of checkpoint recovery, in addition to other proteins like Plk1 and Claspin. Although some of the publications regarding these known mediators of DNA damage recovery are mentioned, the discussion regarding the relationship to the data in this manuscript are very limited.

We thank the reviewer for his/her insightful comments. As suggested, the previously reported role of PLK1 and other cell cycle kinases in DNA damage checkpoint recovery is discussed in more details in the revised manuscript. As for PP2A/B55, we do not think it promotes checkpoint recovery, e.g., by dephosphorylating ATM/ATR or their substrates. Instead, this phosphatase dephosphorylates cell cycle kinases or their substrates, such as CDK1 or PLK1.

The regulation of MASTL stability by E6AP is novel, although the data regarding this regulation and the interaction are not entirely convincing. In addition, several experiments presented in this paper suggest that E6AP is (additionally) involved in checkpoint signalling/activation, whereas the activation of the G2 DNA damage checkpoint was described to be independent of MASTL. Has E6AP multiple functions in the DNA damage response or is ATM-E6AP-MASTL regulation not as straightforward as presented here?

Altogether, in my opinion, not all conclusions of the manuscript are fully supported by the data.

We showed that E6AP depletion reduced checkpoint signaling vis MASTL upregulation, so this pathway is likely to be involved also in DNA damage checkpoint activation, in addition to checkpoint recovery. However, it is important to note that, the expression level of MASTL is not upregulated during the activation stage of the DNA damage checkpoint (unless E6AP is depleted). DNA damage signaling, via ATM-dependent E6AP phosphorylation, caused MASTL accumulation over time. This ultimately shifts the balance toward checkpoint recovery and cell cycle re-entry. As such, the role of MASTL (and E6APdepletion) in suppressing DNA damage checkpoint is in harmony with the proposed role of MASTL upregulation in promoting checkpoint recovery. We have made additional clarifications about this point in the revised manuscript.

Reviewer #1 (Recommendations For The Authors):

In principle a very interesting story, that attempts to shed light on an intriguing and poorly understood phenomenon; the link between damage repair and cell cycle re-entry once a cell has suffered from DNA damage. The issue is highly relevant to our understanding of how genome stability is maintained or compromised when our genome is damaged. The authors present the intriguing conclusion that this is based on a timer, implying that the outcome of a damaging insult is somewhat of a lottery; if a cell can fix the damage within the allocated time it will maintain stability, if not then stability is compromised. However, the story in its present form suffers from a number of major gaps that will need to be addressed

Major point:

My primary concern regarding the main conclusion is that altered MASTL regulation seems to be doing much more than simply promoting more rapid recovery after DNA damage. This concern comes from the following gaps that I noted whilst reading the paper:

• Knock out of E6AP, is leading to a dramatic inhibition of ATM/ATR activation after damage (Fig.5C,D,E), this is (partially) rescued by co-depletion of MASTL (Fig5E). The authors will have to show that the primary effect of altered MASTL regulation is improved recovery, rather than reduced checkpoint activation. In other words, is initial checkpoint activation in cells that have lost E6AP normal, or do these cells fail to mount a proper checkpoint response? If the latter is true, that could completely alter the take home-message of this paper, because it could mean that E6AP/MASTL do not act as a "timer", but as a "tuner" to set checkpoint strength at the start of the DNA damage response. The authors themselves conclude on page 8 "E6AP promoted DNA damage checkpoint signaling by counteracting MASTL", but in the abstract the conclusion is "E6AP depletion promoted cell cycle recovery from the DNA damage checkpoint, in a MASTL-dependent manner". These 2 conclusions are definitely not in alignment, do E6AP/MASTL control checkpoint signaling or do they control recovery?

The expression level of MASTL is not upregulated during the activation stage of the DNA damage checkpoint (unless E6AP is depleted). DNA damage signaling, via ATM-dependent E6AP phosphorylation, caused MASTL accumulation over time. This ultimately shifts the balance toward checkpoint recovery and cell cycle re-entry. As such, the role of MASTL (and E6AP-depletion) in suppressing DNA damage checkpoint is in harmony with the proposed role of MASTL upregulation in promoting checkpoint recovery. We have made additional clarifications about this point in the revised manuscript. We have also made clarification to the statement indicated by the reviewer.

• MASTL KD has a rather unexpected effect on cell cycle progression after HU synchronization (Fig.5B). It seems that the MASTL KD cells fail to exit from the HU-imposed G1/S arrest, an effect that is not rescued in the E6AP knock-outs. Inversely, E6AP knock-outs seem to more readily exit from the HU-imposed arrest, an effect that is completely lost after knock-down of MASTL. How do the authors interpret these results? Their conclusions are entirely based on a role for MASTL as a driver of mitotic entry, with E6AP in control of its levels, but this experiment suggests that MASTL and E6AP are controlling very different aspects of cell cycle control in their system.

As the reviewer pointed out, our data in checkpoint signaling and cell cycle progression suggested that MASTL upregulation could also inhibit DNA damage signaling, in addition to promoting cell cycle progression. This manner of feedback, inhibitory, modulation of DNA damage signaling by mitotic kinases (e.g., PLK1, CDK) has been implicated in previous studies (reviewed in Cell & Bioscience volume 3, Article number: 20 (2013)). In the revised manuscript, we have included discussions about this aspect of checkpoint regulation.

• It is not possible to evaluate the validity of the conclusions that are based on Figure 6. We need to know how long the cells were treated with HU to disrupt the interaction between E6AP and MASTL. Is the timing of this in the range of the timing of MASTL increase after damage? A time course experiment is required here.

• The data obtained on E6AP-S218 phosphorylation and with the S218A mutant during damage and recovery look very promising. But again, the release from HU is confusing me as to what to conclude from them. Also, the authors should show how S218A expression affects MASTL levels (before and after damage). Also, a time course of ATM/ATR activation is required to decide if initial or late ATM/ATR signaling is affected.

Detailed time point information is now added in the figure legends for E6AP phosphorylation and E6AP-MASTL dissociation data. We were able to observe these events during early stages (e.g., 1 hr, or 2-4 hr) of the DNA damage response, prior to significant MASTL protein accumulation.

• The conclusion that "and was not likely to be caused by the completion of DNA repair, as judged by the phosphorylation of replication protein A" (page 5) is based on western blots that represent the average across the entire population. It is possible that MASTL expression is still low in the cells that have not completed repair, while it's increase on blots comes from a subset of cells where repair is complete. The authors should perform immunofluorescence so that expression levels of MASTL can be directly compared to levels of phospho-RPA in individual cells. In fact, the manuscript could benefit a lot from a more in-depth single-cell (microscopy)-based analysis of the relations over time between ATM/ATR activation, E6AP phosphorylation, MASTL stabilization versus the checkpoint arrest and subsequent recovery.

Time point analyses were provided for DNA damage-induced RPA phosphorylation and ATM/ATR substrate phosphorylation (Fig. 1). These data showed MASTL accumulation in the presence of active DNA damage checkpoint signaling. To further strengthen this point, as suggested by the reviewer, we have included quantification of fluorescent measurements (Fig. 2, supplemental 1 A-C). IF data showed MASTL upregulation in correlation with ATM/ATR activation.

Minor points:

It's not "ionized radiation", but "ionizing radiation" (page 5)

We have made the correction as pointed out by the reviewer.

Expression levels of MASTL should be quantified over time after DNA damage. In some of the experiments the increase seems to plateau relatively quick (HU treatment, fig 1B, 1-2 hours), while in others the levels continue to increase over longer periods (HU treatment, fig 1D, 6 hours). This is relevant to the timer function of MASTL that is proposed here.

The kinetics of MASTL upregulation is generally consistent among all cell lines. As suggested, quantification of immunoblots is provided (Fig. 1, supplemental 1E); additional quantification of IF signals is also included (Fig. 2, supplemental 1 A-C).

The experiment executed with caffeine (page 5) should be repeated with more selective/potent ATM/ATR inhibitors that are commercially available.

Specific ATM inhibitor was used to confirm the caffeine result in Fig. 7 supplemental 1B&C.

"a potential binding pattern" (page 6) should be "a potential binding partner"

We have made the correction as pointed out by the reviewer.

Reviewer #2 (Recommendations For The Authors):

1) All western blots require size markers. The FACS blots shown do not have any axis labels.

We have included size markers for blots, at the first appearance of each antibody. Labels are added for FACS blots.

2) The quantification of mitotic cells does not indicate how many cells were counted and if this was done by eye or using software.

The missing experimental information is included in the figure legends, as suggested.

3) The western blots demonstrating ubiquitylation of Greatwall (Figure 4D) are of very poor quality and impossible to interpret.

The ubiquitination of MASTL did not show clear ladders, possibly due to its relative protein size.

Reviewer #3 (Recommendations For The Authors):

Specific suggestions to improve the manuscript:

1) Include literature regarding known mediators of DNA damage checkpoint recovery, including MASTL/Greatwall and PP2A, in the manuscript and discuss the observations from this manuscript in relationship with the literature.

Related literatures are included in the discussion.

2) The increase in MASTL protein levels upon DNA damage are not always clear, for example Fig. 1A. The same for MASTL stability after DNA damage, such as in Fig. 2C. Quantification of the westerns would help demonstrating a significant effect.

As suggested by the reviewer, we have included quantification of fluorescent measurements (Fig. 2, supplemental 1 A-C). Quantification of immunoblots for MASTL upregulation was also added in Fig. 1, supplemental 1E.

3) The E6AP-MASTL in vitro interaction studies shown in Fig. 3 raise doubts. First, beads only are used as negative control, whereas MBP only-beads are a better control. The westerns in top panels of 3B (MASTL), 3C (GST-MASTL) and 3D (MASTL) should be improved. In addition, in Fig. 3C, different GSTMASTL fragments are used in an MBP-E6AP pull down, but the GST-MASTL input does not show any specific band to demonstrate that these fragments are correct. The same for the GFP-E6AP fragments in Fig. 3 Suppl. 1C The input does not show any proteins, there is no N fragment present in the IP and the size of the fragment N3 in the IP GFP does not seem correct.

Altogether, it makes me doubt that the interaction between E6AP and MASTL is direct. Better data with appropriate controls should show whether the interaction is direct or mediated via another protein.

Purified proteins used for the in vitro interaction had significant degradation, causing many bands in the input. We included a lighter exposure of the input here as Author response image 1. MBP alone did not bind MASTL, as both M and C segments of MASTL were MBP-tagged, and did not pull down MASTL. We agree with the reviewer that our direct interaction data showed rather weak MASTL/E6AP interaction, suggesting the interaction is dynamic or possibly mediated by additional binding proteins. We have included this statement in the revised manuscript “Taken together, our data characterized MASTL-E6AP association which was likely mediated via direct protein interaction, although the potential involvement of additional binding partners was not excluded”.

Author response image 1.

4) Fig. 4B. Overexpression of HA-E6AP results in a decrease in MASTL protein levels. Can this effect be rescued by treatment with proteasome inhibitor MG132?

As expected, MG132 stabilized MASTL, with or without E6AP overexpression. We have added this new data in Fig. 4, supplemental 1B.

5) Fig. 4G. MASTL interacts with HA-ubiquitin in WT, but not E6AP KO cells. These cells are treated with MG132, so if E6AP really ubiquitinates MASTL, I would expect MASTL to be polyubiquitinated. However, the "interaction signal" does not show polyubiquitination. In fact, this band actually runs lower than MASTL in input samples, which even could be an artifact. Please explain.

The ubiquitination of MASTL did not show clear ladders, possibly due to its relative protein size. As the reviewer noted, the band position in the HA-Ub IP lanes seemed slightly shifted, compared to the input. We have noticed in many experiments that bands in the IP lanes did not perfectly align with the input lanes.

6) The DNA damage recovery experiments measuring mitotic index after washing off etoposide (Fig. 5A and Fig. 8A): What are the time points taken? And importantly, why are there no error bars on these intermediate time points, but only on the 4 hour time point?

As suggested, time point information and additional error bars are included.

7) Fig. 5E. According to the authors, depletion of MASTL rescues the effect of KO of E6AP. However, no increase in pATM/ATR substrate signal is seen upon etoposide treatment in these samples so I am not convinced this experiment demonstrates a rescue.

The rescue was evident, especially for many high molecular weight bands which were more effectively detected by this phospho-specific antibody.

8) Fig. 5C and 8D strongly suggest that E6AP is involved in checkpoint activation. How do these data relate to DNA damage recovery? Is the recovery in E6AP KO cells faster as a consequence of reduced checkpoint signaling or is the recovery effect really specific by stabilization of MASTL? These data should be explained, also taken the data from Wong et al. (Sci. Rep. 2016) into account, that demonstrate that G2 checkpoint activation is independent of MASTL.

The expression level of MASTL is not upregulated during the activation stage of the DNA damage checkpoint (unless E6AP is depleted). DNA damage signaling, via ATM-dependent E6AP phosphorylation, caused MASTL accumulation over time. This ultimately shifts the balance toward checkpoint recovery and cell cycle re-entry. As such, the role of MASTL (and E6AP-depletion) in suppressing DNA damage checkpoint is in harmony with the proposed role of MASTL upregulation in promoting checkpoint recovery. We have made additional clarifications about this point in the revised manuscript.

9) The model presented in Fig. 9 is puzzling because there does not seem to be a difference between phosphorylation of E6AP and the interaction with MASTL on early versus late times after DNA damage. And this exactly is what is missing in the manuscript: A more detailed evaluation of the timing of E6APSer218 phosphorylation and the E6AP-MASTL interaction in response to DNA damage.

More clarification is given to explain this model in the figure legend of Fig. 9.<br /> Time point analyses were provided for DNA damage-induced RPA phosphorylation and ATM/ATR substrate phosphorylation (Fig. 1). These data showed MASTL accumulation in the presence of active DNA damage checkpoint signaling. To further strengthen this point, we have included quantification of fluorescent measurements (Fig. 2, supplemental 1 A-C). IF data showed MASTL upregulation in correlation with ATM/ATR activation. Time point information was also added for Ser-218 phosphorylation and MASTL-ENSA dissociation which were observed in early stages of the DNA damage response (1 hr, or 2-4 hr).

Author Response

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

Reviewer #1 (Public Review):

Membrane receptor guanylyl cyclases are important for many physiological processes but their structures in full-length and their mechanism are poorly understood. Caveney et al. determined the cryo-EM structure of a highly engineered GC-C in a complex with endogenous HSP90 and CDC37. The structural work is solid and the structural information will be useful for the membrane receptor guanylyl cyclases field and the HSP90 field. However, a detailed characterization of the protein sample is lacking. Moreover, the physiological significance of this structure is not fully exploited by supporting experiments and the mechanistic insight is currently limited.

We thank Reviewer #1 for constructive reviews and agree that this work forms the basis for future exploration by the guanylyl cyclase and HSP90 fields.

1) The characterization of the protein sample is lacking. SDS-PAGE would be useful to identify potential proteolysis, leading to the dissociation of GC dimer. Further size-exclusion chromatography would be helpful to estimate the molecular weight of the complex and to determine if only GC-C monomer is purified.

We have included a representative SDS-PAGE gel in our revised version of the manuscript (Figure 1—figure supplement 1). While we agree that SEC could be beneficial to further explore the stoichiometry of the imaged sample, we see no significant degradation of the guanylyl cyclase via SDS-PAGE, and therefore believe that the zippered construct would remain dimeric. Relatively poor yields of these samples precluded further exploration in this regard.

2) The orientation distribution of the particles is not homogenous in Fig. S1D. It would be helpful to present the 3DFSC curve to evaluate the effect of preferred orientation on the reconstruction.

While the orientational distribution is not perfectly uniform, the provided angles allowed for sufficient reconstruction of maps with no notable anisotropy. We have included 3DFSC curves in our revised version of Figure 1—figure supplement 1.

3) Description of protein expression details is lacking. Did the author use transient transfection, stable cell line or virus-mediated transduction?

We have clarified that these cells were expressed using transiently transfected ExpiCHO cells.

4) HSP90 binds ATP and is often co-purified with endogenous ATP/ADP. Is there ATP or ADP present in the sample/cryo-EM maps? Is the conformation of NBD similar to ATP-bound HSP90? The author needs to include the description/figures about the nucleotide state of HSP90.

There is clear density for present nucleotide in our reconstruction. Given the mechanistic role for ATP turnover in the release of HSP90 client (Young, Hartl, 2000 – PMID 11060043) and the resolved density, we believe the identity for this nucleotide is ATP. We have added comment to this regard in the revised manuscript: “…the C2 pseudosymmetric, ATP bound, closed state Hsp90 dimer.”

5) The catalytic domains of GC have to be dimerized to perform cyclase function. The presence of only one GC-PK monomer in the cryo-EM structure indicates the structure does not represent an active state of GC. These results suggest the GC expressed in this way is not functional. The authors need to explain why most of the GC protein is trapped in this inactive form.

Indeed, we do believe that this regulatory state is non-functional, as observed for active kinases. We have clarified this in the revised manuscript: “In addition, this disruption of the native state of GC-C, as observed in our structure, would likely leave GC domains out of each other’s proximity, precluding their catalytic activity while Hsp90 is bound.”

6) The GC-C construct used here is a highly engineered "artificial" construct, which has not been fully characterized in this work. Does this construct have similar activity as the activated wt GC-C? Does the protein (this engineered construct) expressed in CHO cells show activity?

While our original goal in developing this construct was to create an imageable construct that was locked in the active state, our current interpretation of the data is that the leucine-zipper induced, putative active geometry leads to the majority of this construct falling into the regulatory state with HSP90 binding. We make no claim to have resolved an active conformation in this work, yet believe that this state is of note due to the previously unresolved nature of these regulatory complexes for guanylyl cyclase receptors.

7) Are the residues on the interface between GC and HSP conserved in other members of membrane receptor guanylyl cyclases? Would mutations on this interface affect the activity of GC?

Given the role this structure plays in our understanding that HSP90 client recruitment is largely not driven by specific residue interactions and the ~30% identity of GC-C to NPR-A and NPR-B, we do not believe that mutations that do not significantly change the stability or fold of the PK domain would significantly modify recruitment to HSP.

8) The authors propose that targeting HSP90 would tune the activity of GC. Is there any experimental data supporting this idea?

Based on the work of Kumar et al., 2001 (PMID 11152473), we do believe that there is a functional link between HSP90 recruitment and GC activity. We hope that this work will spark further exploration of these concepts.

9) The model in Fig. S3 is largely speculative due to the lack of supporting functional data. In addition, it would be better to change the title to "structure of the protein kinase domain of guanylyl cyclase receptor in complex with HSP90 and cdc37" because the mechanistic insight is limited.

We agree that our supplemental figure is more speculative. We have referenced this in the discussion section of the manuscript and put this figure in the supplement to ensure that this is understood to be more speculative in nature.

Reviewer #2 (Public Review):

Caveney et al have overexpressed an engineered construct of the human membrane receptor guanyl cyclase GC-C in hamster cells and co-purified it with the endogenous HSP90 and CDC37. They have then determined the structure of the resultant complex by single particle cryoEM reconstruction at sufficient resolution to dock existing structures of HSP90 and CDC37, plus an AlphaFold model of the pseudo-kinase domain of the guanylyl cyclase. The novelty of the work stems from the observation that the pseudo-kinase domain of GC-C associates with CDC37 and HSP90 similarly to how the bona fide protein kinases CDK4, CRAF and BRAF have been previously shown to interact.

The experimentation is limited to the cryoEM analysis, and is lacking additional studies that would give deeper insight into the oligomeric nature - if any - of the GC-C when bound to HSP90-CDC37 as compared to the free protein. This is relevant, as the dimerization domain downstream of the pseudokinase, is evident in the maps - albeit not well resolved - and it is not clear whether it is still able to mediate dimerization with a second free or HSP90-CDC37bound GC-C. It would also be good to see some experimentation that asks whether association with HSP90-CDC37 inhibits the guanyl cyclase activity. It is clear from previous work that HSP90-CDC37 silence the kinase activity of their bound client kinases, but in this case the catalytic guanyl cyclase is not directly associated with the chaperone complex and may still be able to function.

Given the geometry of the interaction, the dimerization domain of the GC would likely be monomerized, albeit with global dimerization remaining – contributed by the ECD, or in our case the liganded-ECD mimicking leucine zipper. Experimentally, it has been shown in live cells (Kumar et al., 2001, PMID 11152473) that the HSP90 association is required for maximal GC-A function. This is likely due to some sort of resetting nature to the associating to allow further activity, as opposed to activity during the association – given the latter is unlikely based on our structure, where the two GC domains would not be able to form the active dimerized state. Further dissection of this, while outside the scope of the current work, is of great interest.

Although the sequence alignment presented in SuppFig 2 shows that GC-C conserves the classic DFG motif that plays a critical role in the regulation of most kinases, the numbering of the sequence is absent, making it very difficult to relate this to the structural detail shown in Fig 2B. This needs to be clarified, as the interaction of CDC37-Trp31 with the DFG motifs and downstream activation loops in CRAF and BRAF have been proposed as important features of the selectivity of these kinases for the HSP90-CDC37 system, and it would be good to be able to see clearly how much of this is also conserved in the GC-C pseudokinase domain interaction. For example, is the much shorter activation segment (DFG -> APE) ordered in the complex or disordered?

We have clarified Figure 2—figure supplement 1 with additional numbering. While we agree that the DFG motif may play a role in recognition, only the first residue of this motif is interacting with CDC37 in our structure, so it may be likely that the role of this motif is more structural in maintaining a CDC37 complementary fold, as opposed to direct residue interactions. Additionally, many kinases which are not regulated by CDC37/HSP90 contain this motif. The shorter DFGAPE of GC-C is traceable with the exception of N613, S614, I615, though the density in this region reflects this loop not being well stabilized.

It was not easy to follow what was in the sample used for cryoEM. The cloning of the guanylyl cyclase (GC) component is described in the methods and they have shown some illustrations in fig 1 but a proper numbered figure of the domain organisation clearly showing domain boundaries and linker segments is really needed for a reader not familiar with the structure of GCs, especially since they have replaced the ECD with a leucine zipper in their construct. It is important to show a domain figure of what this construct looks like as well, as from the illustrations in fig 1 for examples its hard to see what's PK, DD, GC domains. It would also be helpful to see in the supplementary a gel of complex they put on the grids, to make it clearer what exactly the sample is and to reassure that the GC-C domains that are not resolved in the cryoEM are nonetheless present in the sample.

We have added in a gel figure to the supplement and clarified the content of the imaged construct in the methods section: “This construct contains all domains of the native GC-C, with the exception of the ECD.”

Overall there is only minimal proposal of mechanism or biological function based on the structure. The speculation in the Discussion of two fates - PP5 dephosphorylation or E3 ligase recruitment, is not supported by any experimentation, which is reasonable for speculation, but is also not underpinned by reference to any previously published work suggesting that these additional processes may be important. In the absence of any work by the authors can they put these speculations more in context with previously published work that supports the importance of these processes specifically for GC regulation?

We have ensured that these potential pathways only appear in the discussion section. It has been observed, for instance by Oberoi et al., 2022 that phosphatases can act on all components of a HSP90–CDC37–client system. Given there are well characterized phosphorylation sites for membrane GC receptors, we believe this is worth discussing in this manuscript, to stimulate further exploration of these mechanisms in the field. In addition, it has been reported that many E3 ligases are recruited to HSP90 complexes and can degrade rather non-specifically. It has been shown that one can generate PROTAC-like molecules to target non-specific clients to HSP90–E3 ligase machinery for degradation (Li et al., 2023). Given this proximity induced nature to E3 degradation of HSP90 clients, it would be highly likely that, at least in some cases, mGCs would be degraded by this mechanism as well.

Reviewer #3 (Public Review):

A detailed understanding of how membrane receptor guanylyl cyclases (mGC) are regulated has been hampered by the absence of structural information on the cytoplasmic regions of these signaling proteins. The study by Caveney et al. reports the 3.9Å cryo-EM structure of the human mGC cyclase, GC-C, bound to the Hsp90-Cdc37 chaperone complex. This structure represents a first view of the intracellular functional domains of any mGC and answers without doubt that Hsp90-Cdc37 recognizes mGCs via their pseudokinase (PK) domain. This is the primary breakthrough of this study. Additionally, the new structural data reveals that the manner in which Hsp90-Cdc37 recognizes the GC-C PK domain C-lobe is akin to how kinase domains of soluble kinases docks to the chaperone complex. This is the second major finding of this study, which provides a concrete framework to understand, more broadly, how Hsp90-Cdc37 recruits a large number of other diverse client proteins containing kinase or pseudokinase domains. Finally, the Hsp90-Cdc37-GC-C structure offer clues as to how GC-C may be regulated by phosphorylation and/or ubiquitinylation by serving as a platform for recruitment of PP5 and/or E3 ligases.

Comments:

1) The authors used an interesting approach to obtain the GC-C-Hsp90-Cdc37 complex. Flagtagged human GC-C was overexpressed in CHO cells with the expectation of co-purifying endogenous hamster homologs of Hsp90 and Cdc37. There are several points worth noting:

a) It is not clear from the data presented (Figure 1C, Suppl Fig 1A) or the Methods the percentage of particles in the cryo-EM specimen that represent the GC-C-Hsp90-Cdc37 complex. Presumably, some fraction of GC-C isolated will not be associated with Hsp90Cdc37. If a very large portion of GC-C is associated with Hsp90-Cdc37, it would be good to explain why this is to be expected. Are 2D/3D classes corresponding to the activated GC-C dimer found? If not, why?

While we see some traces of GC-C not bound by Hsp90, there is, in the least, a significant alignment bias for the Hsp90 bound complex. We believe that the engineered construct, which we designed to be locked in a putative active conformation, is going through catalytic cycles to some point where the regulatory mechanism is kicking in. It may be that for proper resetting of the receptor, the receptor needs to cycle back through an unliganded, inactive conformation, which our leucine zipper construct is unable to allow, thus locking our GC in the regulatory complex, though this is speculation.

b) Figure 1A suggests that GC-C is phosphorylated before recruitment of Hsp90-Cdc37. What is the phosphorylation status of the GC-C specimen that was imaged by cryo-EM?

We had placed the P in grey in this figure to represent the potential for the active state to be phosphorylated. For GC-C in particular, the phosphorylation state does not affect activity as much as GC-A and GC-B for example. We have removed this P from the figure for clarity.

c) The resolution of the cryo-EM map (3.9 Å) is too low for unambiguous identification of proteins. Please provide more precise justification for the claim that the densities observed do in fact correspond to hamster Hsp90 and Cdc37.

While we agree that the resolution is limiting for protein identification, the fact that we are using a very stringent FLAG purification allows confidence in the ID for our target, GC-C. For Hsp90 and Cdc37, we are confident that they are endogenous hamster Hsp90 and Cdc37, given the large structural similarity observed in comparison to prior Hsp90/Cdc37/client complex structures, and the ID/register well confirmed by the placement of bulky residues.

d) The authors state that human GC-C pulls down hamster Hsp90-cdc37 but soluble kinases cannot, despite the high sequence identity between human and hamster Hsp90-cdc37. Is this because GC-C recognition is more promiscuous? Can this difference be understood in light of the new structural information presented?

“This native pulldown strategy contrasts with the structures of Hsp90–Cdc37 in complex with soluble kinases (García-Alonso et al., 2022; Oberoi et al., 2022; Verba et al., 2016), for which Hsp90 and Cdc37 had to be overexpressed to obtain complex suitable for imaging.”

It is our understanding, from reading the papers cited above, that Hsp90/Cdc37 needed to be overexpressed to obtain these samples for imaging. We use a different strategy because our sample does not require overexpression of Hsp90 and Cdc37. This may be because of something specific to hamster cells, which were (presumably) not tested in the above studies, or it could be something specific to do with GC-C.

2) A large portion of the enforced GC-C dimer was not visible in the cryo-EM maps. It is not easy to learn from Figure 1 exactly which parts of the GC-C construct was sufficiently ordered and observed structurally. Please improve Figure 1.

We have adjusted Figure 1 to better depict what is observed in the cryoEM density.

3) On page 4, the authors claim that they are able to orient the GC-C-Hsp90-Cdc37 complex "as it would sit on a membrane" and referred to Figure 1B. It is not clear what is implied here. Does Hsp90-Cdc37 binding constrain the complex to face the inner leaflet of the membrane in a specific orientation as shown in Figure 1B? If true, this could potentially have important functional implications. Please illustrate how this was deduced based on the information available.

Given the observed density for the PK domain, which is membrane proximal, we can safely assume that the TM would be located immediately above this region. Given the size of Hsp90 and assuming the soluble Hsp90 must sit below the membrane, we can determine, with some accuracy the relative orientation of this complex next to the membrane. This orientation is depicted in Figure 1B.

4) Also on page 4, it is stated that it is sterically unlikely an additional Hsp90-Cdc37 complex is associated with the other copy of GC-C in the leucine zippered dimer. It is not obvious to the reader how this may be the case. An additional figure could help make this more clear. Additional biochemical evidence will also help. The absence of GC-C-Hsp90-Cdc37 dimers in cryo-EM micrographs can also support the argument.

We have clarified this: “is sterically unlikely that an additional regulatory complex is forming on the second GC-C in a concurrent fashion, given the large size of the first Hsp90–Cdc37 and the requisite proximity of the second GC-C.”

5) Some comments on Figure 2:

a) NTD and CTD are mislabeled in Figure 2A.

Thank you for catching this, we have fixed this.

b) The authors should show cryo-EM density to support their modeling of GC-C in Figures 2B and C.

We have provided maps and models to the reviewer and will release these maps and models upon publication so that all relevant densities can be interpreted to their fullest extent by readers. In addition, we have added representative density panels to Figure 1-figure supplement 2.

6) The authors claim that Hsp90-Cdc37 clients are more similar structurally near the cdc37 interface. Please illustrate this with additional figures. Suppl. Figure 2 is inadequate for this purpose.

We have added a structural overlay to Figure 2—figure supplement 1A to illustrate this.

The authors can also consider adding a more detailed discussion comparing the interactions between the pseudokinase/kinase C-lobe and Cdc37 in known structures. Is shape/charge complementarity a universal feature of cdc37-dependent kinase/pseudokinase recruitment? It would be interesting to also consider if it would be possible to predict which of the ~60 human pseudokinases are possible Hsp90-Cdc37 clients. New structural findings from this study and publicly available AI-predicted protein structures could help.

While the use of AI to predict pseudokinase interactions would indeed be interesting, we believe this is outside the scope of this work. Given methodology is in place for determination of kinase clients for Hsp90 (Taipale et al., 2012), this could be an additional route to obtain this information in future work.

Reviewer #2 (Recommendations For The Authors):

In Figure 1B the authors show a large unaccounted-for region of density which they speculate may be due to the dimerization domain. That this is lost in the sharpened maps suggests that it is more mobile than the core which probably dominates the automatic mask generation used by cryoSPARC. It would be very interesting to try and resolve this region further by using focussed classification and refinement - probably in RELION. This would add further novelty, as so far in the three HSP90-CDC37 kinase complexes previously described, little is seen outside the C-terminal lobe of the kinase (or in this case pseudokinase) lobe.

Given the structurally uncharacterized nature of the DD and GC domains for mGCs, using computational means to further our understanding of these regions was attempted. Across several software packages, these attempts were unsuccessful. We will be uploading these micrographs to EMPIAR shortly after publication, which will allow for other groups to re-process this data as they see fit and as new software techniques emerge in this rapidly developing field. We believe that the partially unfolded nature of the PK domain is providing too much of a hinge point prior to the DD for the software to be able to resolve this currently.

Author Response

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

Reviewer # 1

Specific comments

1) Figure 1: it is unclear how many mice were used for the described phenotypic analyses (panels D and E). Please clarify.

We acknowledge that we made a mistake in failing to clearly describe the phenotypic analyses. In Figure 1D and E, we performed statistical analysis on the number of TEBs in whole mammary mounts. One mouse stained a mammary whole mount with Carmine-alum staining. Thus, “n” represents the 10 mice we analyzed. We have modified the legend of Figure 1 to " D, E. Quantification of the average number of TEBs and bifurcated TEBs in littermate Crb3fl/fl (n=10) and Crb3fl/fl;MMTV-Cre (n=10) mice at 8 weeks old" in lines 909-911.

2) Figure 2: in panels B and C it is unclear how the data was quantified; the legend states "n=10", does this mean the experiment in B was done 10 times? And that 10 acini per condition were measured in panel C? In panel D a difference in 0.3% between NC and shCRB3 seems miniscule; do the authors mean 30% instead? And how many acini were counted per condition per (how many) experiments? Same applies to panels G and H, it is unclear how many cells were analyzed per (how many) experiments.

Thanks for your suggestions. We failed to describe the details of the statistical analysis well in the experimental method. To provide a brief overview of our statistical analysis method, we took 3-4 random bright-field micrographs of each well in the chamber slide system and repeated the experiment three times. We then counted the number of acini in all micrographs (Figure 2B) and examined the diameter of all acini in each photograph, averaging the values as data (Figure 2C). We also determined the percentage of aberrant acini in each photograph, which was used as an analysis value (Figure 2D). We carefully confirmed that the vertical axis of Figure 3D was indeed mislabeled and should mean 30%, and revised the original figure. For IF analysis of the mitotic spindle orientation during lumen formation, we examined the division angle of one cell in one acinus that was mitotically dividing, 3-4 acini were randomly examined in each well in the chamber slide system, and this experiment was repeated three times (Figure 2G and H). Therefore, we have provided a detailed description of these issues in the Figure 2 legend. The revised parts are found in lines 922-924, lines 926-927, lines 929-930, and line 932.

3) Figure 2: it would be desirable if authors were able to quantify the data in panels E and I.

Thank you for your comments. According to your suggestions, we performed the quantitative analysis of Figure 2E and I, which is now presented in the new Figure 2D and H.

4) For all cell-based assays using shRNA to knock down CRB3 (Fig. 2A-H; Fig. 3A-F; Fig. 4C-E; Fig. 5G-J; Fig. 6C; Fig. 7C, D; Fig. 8E-G), it would be desirable to perform rescue experiments to ensure that the observed phenotype of CRB3 depleted cells is specific and not due to off-target effects of the shRNA.

Yes, rescue experiments involving overexpression of CRB3 in CRB3 depleted cells can accurately account for the specific phenotype as well as eliminate the off-target effects of shRNA. However, our group has long focused on the role of the cell polarity protein CRB3 in contact inhibition and tumorigenesis. Our previous studies have ruled out the off-target effects of shRNA and reported that CRB3 regulates contact inhibition and tumorigenesis through Hippo or Wnt signaling pathways (Cell Death Dis 2017;8(1):e2546, Oncogenesis 2017;6(4):e322, J Cell Mol Med 2018;22(7):3423-33). Therefore, we will pay close attention to rescue experiments to ensure experimental integrity and phenotypic specificity in our subsequent studies.

5) Figure 3: how many cells were counted/measured per condition (in how many experiments) in panels B, D, H, F, G and H? In panels C and D, what is the CRB3 protein level in these cells? This is of relevance as protein overexpression per se could impinge on ciliation frequency. This question could be addressed by performing a western blot analysis with CRB3 antibody.

We did not clearly describe the measurement and statistical analysis methods in the previous manuscript. Similarly, we took 3-4 random IF and SEM micrographs of each sample in one experiment, and this experiment was repeated three times. Subsequently, the number of ciliated cells and total cells were counted, and the proportion of ciliated cells was calculated (Figure 3B, D and F). In these figures, the cilium length of representative ciliated cells was measured in each photograph. In the knockout mouse model, we needed to find the intact mammary ductal lumen and renal tubule in IF staining of mouse mammary and renal tissue sections, with 5-6 random fields micrographs taken per slice, and the proportion of ciliated cell was measured by counting and taking the average. A total of ten mice were repeated in these experiments (Figure 3G and H). Therefore, the legend of Figure 3G and H has been partially modified and a detailed description has been added to the Figure 3 legend. The revised parts are in lines 945-946, lines 950-951, line 953.

Thank you for your suggestions that we perform a western blot analysis with CRB3 antibody in Figure 3C and D. And we have added the western blotting with CRB3 analysis in the new Supplementary Figure 3A.

6) Figure 3G: it is very difficult to see that the red stained structures are primary cilia.

Yes, the staining structure of primary cilia in mammary ductal lumen are less clear than that of individual cells and in renal tubule in Figure 3G. We used recognized acetylated tubulin and γ-tubulin to stain the primary cilia, which were clearly labeled in individual cells. However, the labeled primary cilia in renal tubule were longer length and demonstrated a more pronounced structure than those in the mammary ductal lumen. In the mammary ductal lumen of the 10 mice we analyzed, the primary cilia showed shorter length and staining structure than the others shown in Figure 3G. This difference may be due to the distinct characteristics of primary cilia in different tissues.

7) Figure 4B: how many cells were analyzed in how many experiments?

Our statistical methods for analyzing cellular experiments using IF were essentially the same. We randomly selected 3-4 IF micrographs of each sample in one experiment, and this experiment was repeated three times. Subsequently, the number of colocalization cells and total cells were counted, and the proportion of cells with pericentrin and CRB3 colocalization was calculated (Figure 4B). The detailed description has been added to the Figure 4 legend. The revised part is in lines 962-963.

8) Lines 217-219: since the cells were not stained with a cilia marker, only a centrosome marker, the claim that CRB3 localizes to the base of cilia is unsubstantiated.

Thank you for your comments. The base of cilia is the basal body, which develops from the mother centriole of the centrosome (Cancer Res. 2006;66(13): 6463-7). Firstly, we found colocalization of CRB3 and pericentrin, a centrosome marker, in MCF10A cells (Figure 4A and B). Secondly, we verified the colocalization of CRB3 with γ-tubulin, a marker of basal body in primary cilia, in confluent quiescence cells (Figure 4C and D). In addition, we found that CRB3 was localized at the base of primary cilia labeled with acetylated tubulin (Figure 4E and F). Due to the species of commercialized CRB3 antibody, we were able to indirectly claim that CRB3 localizes to the base of cilia through these experiments.

9) Figure 3 and Figure 4: is it problematic to use gamma tubulin as centrosome marker if CRB3 depletion causes reduced centrosomal recruitment of gamma tubulin ring complex components? Also, in Figure S3A no gamma tubulin staining can be seen in the lower panel, why?

Thank you for your positive comments. As is well known, γ-tubulin is a marker of the centrosome, and we found that CRB3 depletion causes reduced centrosomal recruitment of gamma tubulin ring complex components. However, Our Figure 3 was illustrated the effect of CRB3 on ciliary assembly, and Figure 4 was analyzed the localization of CRB3 in primary cilia. In some reports on ciliary assembly, the fluorescent double staining of acetylated tubulin and γ-tubulin have been used to label primary cilia, and the effect of target genes on ciliary number and assembly were analyzed by these markers (Nature. 2013;502(7470): 254-7, Cell. 2007;130(4): 678-90 and so on). Although CRB3 affects the recruitment of gamma tubulin ring complex components, it does not affect the analysis of ciliary number and localization in Figures 3 and 4.

In Figure S3A, green staining labeled with γ-tubulin could be clearly found in the lower left panel. The representative area from the left amplification may have been poorly selected, resulting in no γ-tubulin staining on the right side. We have updated the lower right panel in the new Supplementary Figure 3B.

10) Figure S4A: the grouping of indicated proteins is factually wrong. For example, FBF1, SCLT1 and ODF2 are not IFT-B components, and several of the proteins indicated as localizing to the basal body also localize to (unciliated) centrioles. In contrast, CP110 is usually only found on unciliated centrioles and not mature basal bodies. Authors should consult the relevant literature and correct the figure accordingly. Alternatively, this misleading text/grouping could be removed from the figure. Furthermore, in the legend to Figure S4 there is no information provided about this quantitative analysis (how many independent experiments, which cells were analyzed etc.).

Thank you for your helpful suggestions. We have taken your advice and removed this misleading information from the manuscript, Supplementary Figure 4A and its corresponding legend. In the legend to Supplementary Figure 4A, we have added the detailed information for this quantitative analysis in the legend. The revised legend is shown in lines 1098-1100.

11) Figure S4B: how do authors know which of the bands correspond to CRB3 fusion protein?

Based on the construction strategy of the CRB3-GFP fusion protein (Figure 6D) and its base sequence, we were able to calculate its molecular weight. Then the molecular weight of CRB3-GFP fusion protein was verified by western blotting (Figure 6F and 7A). Meanwhile, exogenous overexpression allowed for the production of the CRB3-GFP fusion protein in large quantities. Due to these features, we could know that the band indicated by the black arrow is most likely CRB3-GFP fusion proteins. In order to check the molecular weight, we have labeled the key molecular weight markers in the new Supplementary Figure 4B.

12) Lines 251-253: this seems like data overinterpretation.

Thank you for your comments. We have revised this sentence in lines 252-254.

13) Lines 260-261: the data showing perturbed gamma tubulin localization is not convincing as data was not quantified.

According to your suggestions, we performed the quantitative analysis of Figure 4C, which is now presented in the new Figure 4E.

14) Figure 5H and Figure 6C: to show that the GCP6 IP actually worked, these blots should be probed also for GCP6.

Thank you for your good suggestions. We have added these blots probed for GCP6 in new Figure 5H and 6C.

15) Figure 5I: how many cells were analyzed in how many experiments?

Our statistical methods for analyzing cellular experiments using IF were essentially the same. We took 3-4 random IF micrographs of each sample in one experiment, and this experiment was repeated three times. The detailed description has been added to the Figure 5 legend. The revised part is in lines 992-994.

16) Figure S5: it looks like GPC6 and Rab11 are localizing all over the cell, are the antibodies used for the IFMs specific for these proteins?

After checking the specificity of these antibodies used for the IFMs, we have decided to delete the corresponding results in the Supplementary Figure 5 and their description in the original manuscript.

17) Lines 43, 89, and 314-315: the claim that CRB3 directly binds Rab11 is not supported by the data. The data provided only shows that these proteins interact indirectly. To show direct interaction, yeast-2-hybrid analysis or pull-down assays with purified proteins would be required.

Thank you for your positive comments. Since we were unable to complete the relevant experiments to demonstrate direct interaction of two proteins, we have revised our conclusions. Replace " CRB3 directly binds Rab11" with " CRB3 binds Rab11" in the manuscript.

18) Figure 6G and lines 314-315: this result is surprising as it indicates GTP- and GDP-locked versions of Rab11 have the same inhibitory effect on CRB3 binding? Please comment, and also indicate how data in Figure 6G was quantified (and how many independent experiments were used for the quantification).

We were also puzzled by the results shown in Figure 6G. Based on the western blotting bands, we suspected that there may have been some issues with the experiment. Specifically, we believed that the inefficient transfection of Flag-Rab11aWT, Flag-Rab11a[Q70L], Flag-Rab11a[S20V], and Flag-Rab11a[S25N] plasmids, as well as the insufficient amount of GFP antibody used in the co-IP experiment, led to the corresponding bands being too weak and masking the true differences.

To address this, we optimized the experimental conditions, strictly increased the experimental control, and repeated the experiment in triplicate. The new results are shown in the revised Figure 6G. The statistics from the three independent experiments revealed that CRB3b had a stronger interaction with Rab11a[Q70L] and Rab11a[S20V], while showing a weaker interaction with Rab11a[S25N], compared to Rab11aWT. As this result, we revised the original manuscript in lines 308-310 and added a detailed description to the Figure 6 legend in lines 1012-1013.

19) Figure 8G: data needs to be quantified.

Thank you for your comments. We replaced the unattractive bands in the western blotting of Figure 8G with better quality ones. The statistical analysis of the Figure 8G data is shown in Supplementary Figure 6.

Further minor comments

1) Abstract should indicate that this study describes conditional knockout of Crb3 in mouse mammary gland epithelial cells.

This is good writing advice. We have added the relevant description in lines 40-42.

2) Line 87: specify which gland (mammary?).

We have modified to " mammary gland" in line 87.

3) Line 140: sentence states that knockout of Crb3 is essential for branching morphogenesis in mammary gland development, I do not think this is correct.

We have removed the inappropriate finding.

4) Line 152: "formed more number" should be "formed more" or "formed higher number of".

We modified "formed more number" to "formed more" in line 154.

5) Lines 157-163: text and logic are difficult to follow for a non-expert.

We have modified the logic of this paragraph, as detailed in lines 158-165.

6) Figure 4A, C: figure resolution could be improved. It is difficult to see what the authors claim these figures are showing.

The clarity of the original images in Figure 4A and C is acceptable, while the images on the right are electronically enlarged. Although there is a decrease in pixels, it can still display our findings.

7) Figure 7D, E: images look pixelated.

The clarity of the original images in Figure 7D and E is acceptable using a laser confocal microscope, while the images on the right are electronically enlarged.

8) Line 222: unclear what authors mean by "detected a series".

We modified "detected a series" to "some important" in line 226.

9) Lines 221-225: which cells were used for the analysis in Fig. S4?

We used MCF10A cells for the analysis in Supplementary Figure 4, and modified its legend in line 1098.

10) Line 245: what is "cytomembrane"?

We modified "cytomembrane" to "cell membrane" in lines 246-247.

11) Lines 246-250: wording is unclear/difficult to understand.

We have modified this paragraph, as detailed in lines 248-251.

12) Line 273: should "regimented" be "sedimented"?

We modified "regimented" to "sedimented" in line 274.

13) Line 287-288: sentence does not make sense.

We have removed this sentence.

14) Figure 5A: it would be desirable to show the original dataset (Excel file) used for generating this figure.

To maintain data integrity, we should provide the original dataset (Excel file). However, there are some unpublished data in this file that we must withhold for the time being. If needed, the corresponding author can be requested to provide the file.

15) Lines 298-299: wording is unclear.

We have modified this sentence, as detailed in lines 296-298.

16) Lines 285-287: replace "instead of" with "but not".

We modified "instead of" to "but not" in line 286.

17) For all IFMs showing merged images of the green and red channel, please also show the red and green channel separately.

Most of our fluorescence images are presented separately for each channel in this manuscript, with only a few merged images due to space limitations. This type of presentation is commonly used in published papers.

18) Lines 326 and 327: replace "bonded" with "bound".

We have modified in lines 322-323.

19) Lines 327-328 and 361-364: wording is unclear/grammatically incorrect.

We have modified these paragraphs, as detailed in line 323 and lines 357-360.

20) Line 342: what is meant by "the combination of"?

We modified "the combination of" to "the binding of" in line 338.

21) Line 365: localization of what?

This means "subcellular localization" in lines 360-361.  

Reviewer # 2

Major points

1) CRB3 is present in mammals as 2 isoforms, A and B, originating from alternative splicing. In this study, the authors never mention this fact and when using approaches to KO or KD CRB3A/B they are likely to deplete both isoforms which have been shown to have different C-terminal domains and functions (Fan et al., 2007). This is also important for the CRB3 antibodies used in the study since according to the material and methods section they are either against the extracellular domain common to both isoforms or the intracellular domain which is only similar in the domain close to transmembrane between the 2 isoforms. Since the antibodies used in each figure are not detailed it is impossible to know if the authors are detecting CRB3A or B or both. Please provide the information and correct for the actual isoform detected in the data and conclusions.

Thanks for your positive comments. In mammals, CRB3 has two isoforms, CRB3a and CRB3b, distinguished by alternative splicing within the fourth exon of the CRB3 gene, which in turn produces a protein with 23 amino acid differences at the C terminus. Both CRB3a and CRB3b have mostly identical amino acid sequences, and have indistinguishable molecular weight sizes. As a result, the knockout mouse construction strategy and the design principles of RNAi sequences target both CRB3a and CRB3b. This is described in lines 100-104 and lines 149-150. Additionally, commercially available antibodies detect both CRB3a and CRB3b, as mentioned in line 123 and lines 636-637 in revised manuscript.

However, it should be noted that our CRB3 overexpression, as shown in the CRB3 structural domain in Figure 6D, refers specifically to the sequence of CRB3b. As a result, we have updated the original manuscript as well as the legends of Figures 3C, 3E, 4A, 5A, 5B, 6D-G, 7A, 7B and Supplementary Figure 2F-H, 3A, 4B, 6B to reflect this change. All instances of overexpressed CRB3 have been changed to CRB3b.

2) CRB3A and B have been localized in the cilium itself (Fan et al., 2004; 2007) but in the study CRB3A/B does not enter the cilium but is localized in the basal body (figure 4). How the authors reconcile these different localizations?

Indeed, we found that CRB3 is mainly localized at the basal body of the primary cilium, which differs from previous reports in the literature (Curr Biol. 2004;14(16):1451-61 and J Cell Biol. 2007;178(3):387-98). However, upon closer examination of one of these reports (Curr Biol. 2004;14(16):1451-61), it appears that CRB3 was actually scattered on the primary cilia, with a strong focus at the basal body. Additionally, in rat kidney collecting ducts, the localization of CRB3 on primary cilia was significantly reduced, with obvious localization at the basal body. Another study (J Cell Biol. 2007;178(3):387-98) also reported the co-localization of CRB3b and γ-tubulin in MDCK cells, which is consistent with our conclusion. We further verified the co-localization of CRB3 with the centrosome by overexpressing CRB3b in mammary epithelial cells, indicating that CRB3 mainly localizes to the basal body of the primary cilium. This information is discussed in the Discussion section of the manuscript (lines 400-410).

3) The authors use GFP-CRB3A/B, it is not stated which isoform, over-expression to localize CRB3A/B in MCF10A cells (figure 4A). The levels of expression appear to be very high in the GFP panel and it is likely that the secretory pathway of the cells is clogged with GFP-CRB3A/B in transit from the ER to the plasma membrane. Thus, the colocalization with pericentrin might be due to the accumulation of ER and Golgi around the centrosome. This colocalization should be done with the endogenous CRB3A/B and with a better resolution.

Thank you for your comments. We were also interested in the co-localization of endogenous CRB3 and centrosome proteins. However, the only commercial CRB3 antibody available is the rabbit species, and the pericentrin antibody (Abcam, ab4448) that is very useful is also the rabbit species. We had difficulty finding commercial centrosome-associated antibodies for other species. Therefore, we examined the co-localization of endogenous CRB3 with γ-tubulin in Figure 4C and combined the results with those of exogenous CRB3 to illustrate the co-localization of CRB3 with centrosomes.

4) The staining for CRB3A/B in figure 4C (red) is striking with a very strong accumulation in an undefined intracellular structure and the authors do not provide any explanation for such a difference with the GFP-CRB3A/B just above.

Thank you for your good suggestions. The immunofluorescence images of GFP-CRB3 in Figure 4a were obtained using a fluorescence microscope, while the images of endogenous CRB3 were obtained using a laser confocal microscope. The fluorescence microscope excites a fluorescent dye to emit a signal, which is amplified into a visible light signal and presents a full fluorescent signal. In Figure 4a, we can clearly see the full distribution of exogenous CRB3 in MCF10A cells, including its tight junctional localization consistent with previous reports in the literature and its co-localization with centrosomal proteins. On the other hand, laser confocal microscopy uses a laser as the light source to excite the fluorescence within the sample point by point. It employs a precision pinhole filtering technique with strong laminar imaging capabilities. In the specific analysis of endogenous CRB3 co-localization studies with centrosomes and primary cilium, signals at tight junctions must be excluded. Therefore, Figure 4c represents the fluorescence signal at the level of intracellular CRB3 co-localization with γ-tubulin. The two methods use different detection means and techniques, and are not directly comparable.

5) The staining in figure 4E is also different from those shown in figure 4F in which the CRB3A/B staining is right at the base of the axoneme while it is not the case in figure 4E where we can see a red dot close to but not right at the base of the axoneme.

Thank you for your comments. The new Figure 4F displays the localization relationship between CRB3 and primary cilium, analyzed using laser confocal microscopy. With the unique single-level detection function of this microscope, the problem of level selection may cause the red dots to appear close to, rather than right at the basal body of the primary cilium. However, the new Figure 4G, based on the use of 3D reconstruction scanning technique, clearly demonstrates the localization of CRB3 at the basal body of the primary cilium under the same cells and conditions.

6) The authors claim that CRB3A/B interacts directly with Rab11 but they only show co-immunoprecipitation experiments from cell lysates which do not support direct interactions. The only way to show a direct interaction is to produce both proteins in vitro. Thus, the term direct interaction should be removed.

Thank you for your positive comments. Since we were unable to complete the relevant experiments to demonstrate direct interaction of two proteins, we have revised our conclusions. Replace " CRB3 directly binds Rab11" with " CRB3 binds Rab11" in the manuscript.

7) In addition, the authors claim (Line 251/252) that Rab11 is necessary for the transport of CRB3A/B but they should KD Rab11 to show this.

Thank you for your good suggestions. It is essential to observe CRB3 trafficking after knockdown Rab11. However, in Figure 5C, we used the endocytosis inhibitor dynasore, which also inhibits Rab11-positive endosomes. This result shows that dynasore can significantly inhibit CRB3 trafficking in MCF10A cells. We believe that this experiment partially demonstrates that inhibiting Rab11 function can affect CRB3 trafficking.

8) The domain of CRB3A/B that is necessary for the interaction with Rab11 is the N-terminal part of the extracellular domain. This domain is thus inside the transport vesicles and not accessible from the cytoplasm. Given that Rab11 is a cytoplasmic protein, how the 2 proteins could interact across the membrane? The authors do not even discuss this essential point for their hypothesis.

Thank you for your positive comments. As shown in the schematic model in Figure 9, we believe that when cells form tight junctions, CRB3 is primarily located on the cell membrane. Subsequently, endosomes are involved in the intracellular degradation process of CRB3 on the cell membrane. Intracellular CRB3 can bind to Rab11 through the extracellular domain, which in turn participates in primary cilia assembly. We have made detailed modifications to lines 418-421.

9) Figures are not numbered.

Thank you for your comments. We have updated the numbers in the original manuscript as well as the legends of Figures 1D, 1E, 2B, 2D, 2F, 2G, 3B, 3D, 3F-H, 4B, 4E, 5I, 6, 8G and Supplementary Figure 1E, 2, 3C, 4A, 5B, 6.

Minor points

1) The authors cite several studies showing that a down regulation of CRB3A/B in human cells promotes cancer but other studies show the contrary: Lin et al., 2015 for example. Please discuss these discrepancies.

Thanks for your good suggestion. We have included additional studies with contrasting results in the discussion section, specifically in lines 378-380.

2) Line 98: "exhibit smaller" smaller than what?

We modified "exhibit smaller" to "exhibit smaller size" in line 97.

3) Line 152: "form more number, ..." ???

We modified "formed more number" to "formed more" in line 154.

4) Line 180: "Compared with the control, the number of cells with primary cilium was significantly increased ». To me it is the contrary! This part is not clear at all. Please rewrite.

We have revised the sentence in lines 183-185.

5) Authors should check and review extensively for improvements to the use of English.

Thanks for your good writing advice. We have carefully reviewed and revised the entire manuscript to improve its readability.

Author Response

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

We thank the reviewers and appreciate their recommendations to improve this work.

Reviewer 1:

Reviewer 1 recognizes that ‘This is an important finding that is relevant to the actions of VDR on colorectal cancer. The data presented to support the presented conclusion is convincing’.

Reviewer 1 identifies as a major weakness ‘that the site of SIRT1 regulatory lysine acetylation is defined by mutational analysis rather than by direct biochemical analysis.

However, as the reviewer mentions “previous reports of K610 acetylation using mass spec https://www.phosphosite.org/proteinAction.action?id=5946&showAllSites=true), and the absence of SIRT1 mutant K610R in the immunoprecipitates using anti-acetylated lysine antibodies presented in Fig. 4E clearly overcome this weakness”.

In addition, overall SIRT1 acetylation is reduced by vitamin D and by the specific SIRT1 activator SRT1720 as shown by decreased SIRT1 in the anti-acetyl-lysine immunoprecipates, (Fig. 4A and B). The second weakness identified by Reviewer 1 concerns “the use of only one shRNA to deplete VDR in CRC cells.”

We have made efforts to demonstrate that the results are specific, though we do not have results with alternative shRNAs for a variety of reasons. To mitigate this issue, we have compared two colon cancer cells originating from the same patient which differ in the presence/absence of VDR. SW480, derive from the primary tumor and express VDR, whereas SW620 cells were derived from a lung metastasis and lack VDR. Similar, to the comparison of HCT116 with shVDR HCT116 cells presented in this study, VD induced SIRT1 levels in SW480 in contrast to a lack of induction in SW620, as shown in Author response image 1. This result provides support for the specificity of the shVDR.

Author response image 1.

Vitamin D requires the presence of VDR to increase SIRT1 protein levels. SW480 and SW620 cell lines derive from the same patient, from primary tumor and lung metastasis respectively and differ in their VDR content. 1α,25-dihydroxyvitamin D3 (1,25(OH)2D3) was added at 100 nM for 24 h. Representative western-blot, where TBP was used as a loading control, of four biological replicates. Statistical analysis by ANOVA and values represent mean ± SEM; *p<0.05; *** p<0.001.

The referee noticed the inclusion of an siRNA for SIRT1 in Table 1. We apologize for that, since this is an error, and no results are presented in this study with SIRT1 depletion. Table 1 has been modified accordingly.

Concerning the third and fourth weaknesses that Reviewer 1 identifies, we agree that mapping the interacting domains in both VDR and SIRT1 and in vitro reconstitution would improve the present study. However, we believe that these would constitute long-term studies that themselves are not strictly necessary at this stage. Consequently, we favor the publication of the present body of work. In vitro reconstitution of the present work and the putative relevance of the proposed mechanism of vitamin D action via SIRT1 on types of cancer other than colon (eg breast etc), are certainly very interesting and warrant further investigation.

Reviewer 2:

This reviewer acknowledges that “…this study provides very interesting and solid information on the link between vitamin D and colorectal cancer. It is likely that this study will provide insight into the importance of vitamin D in other types of cancer. It may also lead to new therapeutic strategies for specific cases. This article is convincing, although the authors can improve their study as outlined…”

We acknowledge the proposed changes and recommendations, and have changed the text and Figures as suggested the by Reviewer as follows:

Figure 1

Figure 1E and F: the cell lines used were described in the figure legend, but we agree that including the name in the figure brings more clarity and these are now added.

Figure 1G: the statistical analysis was for all panels of Figure 1 as described in the Figure legend (lines 731-32), We have amended the original omission of panels 1G and 1H. In panel G, * represents statistical analysis by ANOVA (comparing the four groups) whereas # was the analysis by Students t test (comparing the two indicated groups), where * or #p<0.05. We hope to have clarified this point now.

Figure 2

Figure 2C: We showed originally the SIRT1/VDR interaction by immunoprecipitation of VDR and detection of SIRT1 in immunoprecipitates. We also showed immunoprecipitation of exogenously expressed Myc-SIRT1 (WT or mutants) and detection of VDR in immunoprecipitates (Figure 4F). The reviewer requests that we perform the inverse IP for endogenous SIRT1, that is immunoprecipitate SIRT1 and detect VDR in the immunoprecipates, which we now supply for the reviewer in Author response image 2.

Author response image 2

Immunoprecipitation of endogenous SIRT1 to show interaction with VDR. 1α,25-dihydroxyvitamin D3 (1,25(OH)2D3) was added at 100 nM for 24 h. Representative western-blots, where TBP was used as a loading control.

Figure 3

• Figure 3D: ‘The authors should indicate the color of the different stainings’. Immunostainings have been revealed with DAB (diaminebenzidine); thus, positiveness is highlighted by light or dark brown according to their low or high protein expression. Counterstaining has been performed with hematoxilin, which stains nuclei in dark blue and cytoplasm in light blue.

Do the authors mean that the secondary antibody marks in brown/red? If so, these results are inconsistent with the text considering that hematoxylin was used for non-tumor tissue. This part needs to be clarified.

We thank the Reviewer for asking us to clarify this issue. Neither the primary nor anti-Ig horseradish peroxidase-conjugated secondary antibodies presented positiveness resulting from these antibodies individually. Therefore, secondary antibody does not mark in any color. Hematoxylin has been used as counterstaining for both non-tumor as well as for tumor tissues.

What about the level of FOXO3A in these tissues/tumors?

We did not prove the tumor sections for specific SIRT1 substrates such as FoxO3A since their levels may not entirely depend on SIRT1 specific deacetylation.

What is the level of 1,25(OH)2D3 in these patients?

We agree with this referee that this information would be very useful, but unfortunately, we do not have data on vitamin D levels for these patients since they were not specifically recruited for this study and vitamin D levels are not routinely measured.

Figure 3D, the following information is missing: "A detailed amplification is shown in the lower left of each micrograph."

We decided not to include the amplification in micrographs because the aim of the manuscript is focused on protein levels, not localization and including the amplification was more confusing than enlightening. This has been amended now in the text.

Figure 3E, it says p=0.325, in the legend p<0.01, and in the text there is a trend. Which is the correct version?

We really apologize for this misunderstanding. As stated in the Figure, p=0.325 and therefore it does not reach statistical significance. We have amended the main text and figure legend to report that differences between SIRT1 expression levels of healthy and cancer human colon samples are not statistically significant.

Figure 4

Figure 4F. The quality of the presented blots is not optimal. It needs to be improved. In addition, the number of independent biological experiments is not indicated.

We have substituted the representative western-blot and included statistical analysis of four independent biological replicates. Since 4F is now a bigger panel, it has required a slight reorganization of the whole Figure, but the rest of panels remain with the originals. Now we indicate in the figure legend that at least three independent biological replicas were analyzed. In addition, we supply below the four experiments for the reviewer in Author response image 3.

Author response image 3

Immunoprecipitation of exogenous myc-tagged SIRT1 to show interaction with VDR of wild type (WT) or mutants. 1α,25-dihydroxyvitamin D3 (1,25(OH)2D3) was added at 100 nM for 24 h. FT: Flow Through. TBP as a loading control.

Regarding the last general comment concerning the number of independent experiments performed, this is indicated in the Figure legends (lines 732-36, 757-58, 82324, 840-41). All the in vitro experiments were performed at least as three independent experiments and not by repeating a western blot. A representative western blot is shown, and the statistical analysis corresponds to the analysis of the three biological replicates. For experiments with patient samples, the number of patients appears clearly indicated in the corresponding panel.

Author Response

Reviewer #1 (Public Review):

The objective of this investigation was to determine whether experimental pain could induce alterations in cortical inhibitory/facilitatory activity observed in TMS-evoked potentials (TEPs). Previous TMS investigations of pain perception had focused on motor evoked potentials (MEPs), which reflect a combination of cortical, spinal, and peripheral activity, as well as restricting the focus to M1. The main strength of this investigation is the combined use of TMS and EEG in the context of experimental pain. More specifically, Experiment 1 investigated whether acute pain altered cortical excitability, reflected in the modulation of TEPs. The main outcome of this study is that relative to non-painful warm stimuli, painful thermal stimuli led to an increase on the amplitude of the TEP N45, with a larger increase associated with higher pain ratings. Because it has been argued that a significant portion of TEPs could reflect auditory potentials elicited by the sound (click) of the TMS, Experiment 2 constituted a control study that aimed to disentangle the cortical response related to TMS and auditory activity. Finally, Experiment 3 aimed to disentangle the cortical response to TMS and reafferent feedback from muscular activity elicited by suprathreshold TMS applied over M1. The fact that the authors accompanied their main experiment with two control experiments strengthens the conclusion that the N45 TEP peak could be implicated in the perception of painful stimuli.

Perhaps, the addition of a highly salient but non-painful stimulus (i.e. from another modality) would have further ruled out that the effects on the N45 are not predominantly related to intensity/saliency of the stimulus rather than to pain per se.

We thank the reviewer for their comment on the possibility of whether stimulus salience influences the N45 as opposed to pain per se. However, we note that in Experiment 1, despite the same level of stimulus salience/intensity for all participants (46 degrees), individual differences in pain ratings were associated with the change in the N45 amplitude, suggesting that the results cannot be explained by stimulus intensity/salience.

Reviewer #2 (Public Review):

The authors have used transcranial magnetic stimulation (TMS) and motor evoked potentials (MEPs) and TMS-electroencephalography (EEG) evoked potentials (TEPs) to determine how experimental heat pain could induce alterations in these metrics.In Experiment 1 (n = 29), multiple sustained thermal stimuli were administered over the forearm, with the first, second, and third block of stimuli consisting of warm but non-painful (pre-pain block), painful heat (pain block) and warm but non-painful (post-pain block) temperatures respectively. Painful stimuli led to an increase in the amplitude of the fronto-central N45, with a larger increase associated with higher pain ratings. Experiments 2 and 3 studied the correlation between the increase in the N45 in pain and the effects of a sham stimulation protocol/higher stimulation intensity. They found that the centro-frontal N45 TEP was decreased in acute pain.

The study comes from a very strong group in the pain fields with long experience in psychophysics, experimental pain, neuromodulation, and EEG in pain. They are among the first to report on changes in cortical excitability as measured by TMS-EEG over M1.

While their results are in line with reductions seen in motor-evoked responses during pain and effort was made to address possible confounding factors (study 2 and 3), there are some points that need attention. In my view the most important are:

1) The method used to calculate the rest motor threshold, which is likely to have overestimated its true value : calculating highly abnormal RMT may lead to suprathreshold stimulations in all instances (Experiment 3) and may lead to somatosensory "contamination" due to re-afferent loops in both "supra" and "infra" (aka. less supra) conditions.

The method used to assess motor threshold was the TMS motor threshold Assessment Tool (Awiszus et al., 2003). This was developed as a quicker alternative for calculating motor threshold compared to the traditional Rossini-Rothwell method which involves determining the lowest intensity that evokes 5/10 MEPs of at least 50 microvolts. The method has been shown to achieve the same accuracy of determining motor threshold as the traditional Rossini-Rothwell method, but with fewer pulses (Qi et al., 2011; Silbert et al., 2013). Therefore, the high RMTs in our study cannot be explained by the threshold assessment method. Instead, they are likely explained by aspects of the experimental setup that increased the distance between the TMS coil and the scalp, including the layer of foam placed over the coil, the EEG cap and the fact that the electrodes we used had a relatively thick profile.

Awiszus, F. (2003). TMS and threshold hunting. In Supplements to Clinical neurophysiology (Vol. 56, pp. 13-23). Elsevier.

Qi, F., Wu, A. D., & Schweighofer, N. (2011). Fast estimation of transcranial magnetic stimulation motor threshold. Brain stimulation, 4(1), 50-57.

Silbert, B. I., Patterson, H. I., Pevcic, D. D., Windnagel, K. A., & Thickbroom, G. W. (2013). A comparison of relative-frequency and threshold-hunting methods to determine stimulus intensity in transcranial magnetic stimulation. Clinical Neurophysiology, 124(4), 708-712.

2) The low number of pulses used for TEPs (close to ⅓ of the usual and recommended)

We agree that increasing the number of pulses can increase the signal to noise ratio. During piloting, participants were unable to tolerate the painful stimulus for long periods of time and we were required to minimize the number of pulses per condition.

We note that there is no set advised number of trials in TMS-EEG research. According to the recommendations paper, the number of trials should be based on the outcome measure e.g., TEP peaks vs. frequency domain measures vs. other measures and based on previous studies investigating test-retest reliability (Hernandez-Pavon et al., 2023). The choice of 66 pulses per condition was based on the study by Kerwin et al., (2018) showing that optimal concordance between TEP peaks can be found with 60-100 TMS pulses delivered in the same run (as in the present study). The concordance was particularly higher for the N40 peak at prefrontal electrodes, which was the key peak and electrode cluster in our study.

Further supporting the reliability of the TEP data in our experiment, we note that the scalp topographies of the TEPs for active TMS at various timepoints (Figures 5, 7 and 9) were similar across all three experiments, especially at 45 ms post-TMS (frontal negative activity, parietal-occipital positive activity).

In addition to this, the interclass correlation coefficient (Two-way fixed, single measure) for the N45 to active suprathreshold TMS across timepoints for each experiment was 0.90 for Experiment 1 (across pre-pain, pain, post-pain time points), 0.74 for Experiment 2 (across pre-pain and pain conditions), and 0.95 for Experiment 3 (across pre-pain conditions). This suggests that even with the fluctuations in the N45 induced by pain, the N45 for each participant was stable across time, further supporting the reliability of our data. These ICCs will be reported in the next revision.

Hernandez-Pavon, J. C., Veniero, D., Bergmann, T. O., Belardinelli, P., Bortoletto, M., Casarotto, S., ... & Ilmoniemi, R. J. (2023). TMS combined with EEG: Recommendations and open issues for data collection and analysis. Brain Stimulatio, 16(3), 567-593

Kerwin, L. J., Keller, C. J., Wu, W., Narayan, M., & Etkin, A. (2018). Test-retest reliability of transcranial magnetic stimulation EEG evoked potentials. Brain stimulation, 11(3), 536-544.

Lack of measures to mask auditory noise.

In TMS-EEG research, various masking methods have been proposed to suppress the somatosensory and auditory artefacts resulting from TMS pulses, such as white noise played through headphones to mask the click sound (Ilmoniemi and Kičić, 2010), and a thin layer of foam placed between the TMS coil and EEG cap to minimize the scalp sensation (Massimini et al., 2005). However, recent studies have shown that even when these methods are used, sensory contamination of TEPs is still present, as shown by studies that show commonalities in the signal between active and sensory sham conditions that mimic the auditory/somatosensory aspects of real TMS (Biabani et al., 2019; Conde et al., 2019; Rocchi et al., 2021). This has led many authors (Biabani et al., 2019; Conde et al., 2019) to recommend the use of sham conditions to control for sensory contamination. To separate the direct cortical response to TMS from sensory evoked activity, Experiment 2 (n = 10) included a sham TMS condition that mimicked the auditory/somatosensory aspects of active TMS to determine whether any alterations in the TEP peaks in response to pain were due to changes in sensory evoked activity associated with TMS, as opposed to changes in cortical excitability. Therefore, the lack of auditory masking does not impact the main conclusions of the paper.

Ilmoniemi, R. J., & Kičić, D. (2010). Methodology for combined TMS and EEG. Brain topography, 22, 233-248.

Massimini, M., Ferrarelli, F., Huber, R., Esser, S. K., Singh, H., & Tononi, G. (2005). Breakdown of cortical effective connectivity during sleep. Science, 309(5744), 2228-2232.

Biabani, M., Fornito, A., Mutanen, T. P., Morrow, J., & Rogasch, N. C. (2019). Characterizing and minimizing the contribution of sensory inputs to TMS-evoked potentials. Brain stimulation, 12(6), 1537-1552.

Conde, V., Tomasevic, L., Akopian, I., Stanek, K., Saturnino, G. B., Thielscher, A., ... & Siebner, H. R. (2019). The non-transcranial TMS-evoked potential is an inherent source of ambiguity in TMS-EEG studies. Neuroimage, 185, 300-312.

Rocchi, L., Di Santo, A., Brown, K., Ibáñez, J., Casula, E., Rawji, V., ... & Rothwell, J. (2021). Disentangling EEG responses to TMS due to cortical and peripheral activations. Brain stimulation, 14(1), 4-18.

3) A supra-stimulus heat stimulus not based on individual HPT, that oscillates during the experiment and that lead to large variations in pain intensity across participants is unfortunate.

The choice of whether to calibrate or fix stimulus intensity is a contentious question in experimental pain research. A recent discussion by Adamczyk et al., (2022) explores the pros and cons of each approach and recommends situations where one method may be preferred over the other. That paper suggests that the choice of the methodology is related to the research question – when the main outcome of the research is objective (neurophysiological measures) and researchers are interested in the variability in pain ratings, the fixed approach is preferrable. Given we explored the relationship between MEP/N45 modulation by pain and pain intensity, this question is better explored by using the same stimulus intensity for all participants, as opposed to calibrating the intensity to achieve a similar of pain across participants.

Adamczyk, W. M., Szikszay, T. M., Nahman-Averbuch, H., Skalski, J., Nastaj, J., Gouverneur, P., & Luedtke, K. (2022). To calibrate or not to calibrate? A methodological dilemma in experimental pain research. The Journal of Pain, 23(11), 1823-1832.

So is the lack of report on measures taken to correct for a fortuitous significance (multiple comparison correction) in such a huge number of serial paired tests.

Note that we used a Bayesian approach for all analyses as opposed to traditional frequentist approach. In contrast to the frequentist approach, the Bayesian approach does not require corrections for multiple comparisons (Gelman et al., 2000) given that they provide a ratio representing the strength of evidence for the null vs. alternative hypotheses as opposed to accepting or rejecting the null hypothesis based on p-values. As such, throughout the paper, we frame our interpretations and conclusions based on the strength of evidence (e.g. anecdotal/weak, moderate, strong, very strong) as opposed to referring to the significance of the effects.

Gelman A, Tuerlinckx F. (2000). Type S error rates for classical and Bayesian single and multiple comparison procedures. Computational statistics, 15(3):373-90.

Reviewer #3 (Public Review):

The present study aims to investigate whether pain influences cortical excitability. To this end, heat pain stimuli are applied to healthy human participants. Simultaneously, TMS pulses are applied to M1 and TMS-evoked potentials (TEPs) and pain ratings are assessed after each TMS pulse. TEPs are used as measures of cortical excitability. The results show that TEP amplitudes at 45 msec (N45) after TMS pulses are higher during painful stimulation than during non-painful warm stimulation. Control experiments indicate that auditory, somatosensory, or proprioceptive effects cannot explain this effect. Considering that the N45 might reflect GABAergic activity, the results suggest that pain changes GABAergic activity. The authors conclude that TEP indices of GABAergic transmission might be useful as biomarkers of pain sensitivity.

Pain-induced cortical excitability changes is an interesting, timely, and potentially clinically relevant topic. The paradigm and the analysis are sound, the results are mostly convincing, and the interpretation is adequate. The following clarifications and revisions might help to improve the manuscript further.

1) Non-painful control condition. In this condition, stimuli are applied at warmth detection threshold. At this intensity, by definition, some stimuli are not perceived as different from the baseline. Thus, this condition might not be perfectly suited to control for the effects of painful vs. non-painful stimulation. This potential confound should be critically discussed.

In Experiment 3, we also collected warmth ratings to confirm whether the pre-pain stimuli were perceived as different from baseline. We did not include this data initially in the first submission, but will do so in the supplemental material in our next revision. This data showed warmth ratings were close to 2/10 on average. This confirms that the non-painful control condition produced some level of non-painful sensation.

2) MEP differences between conditions. The results do not show differences in MEP amplitudes between conditions (BF 1.015). The analysis nevertheless relates MEP differences between conditions to pain ratings. It would be more appropriate to state that in this study, pain did not affect MEP and to remove the correlation analysis and its interpretation from the manuscript.

The interindividual relationship between changes in MEP amplitude and individual pain rating is statistically independent from the overall group level effect of pain on MEP amplitude. Therefore, conclusions for the individual and group level effects can be made independently.

It is also important to note that in the pain literature, there is now increasing emphasis placed on investigating the individual level relationship between changes in cortical excitability and pain as opposed to the group level effect (Seminowicz et al., 2019; Summers et al., 2019). As such, it is important to make these results readily available for the scientific community.

Summers, S. J., Chipchase, L. S., Hirata, R., Graven-Nielsen, T., Cavaleri, R., & Schabrun, S. M. (2019). Motor adaptation varies between individuals in the transition to sustained pain. Pain, 160(9), 2115-2125.

Seminowicz, D. A., Thapa, T., & Schabrun, S. M. (2019). Corticomotor depression is associated with higher pain severity in the transition to sustained pain: a longitudinal exploratory study of individual differences. The Journal of Pain, 20(12), 1498-1506.

3) Confounds by pain ratings. The ISI between TMS pulses is 4 sec and includes verbal pain ratings. Considering this relatively short ISI, would it be possible that verbal pain ratings confound the TEP? Moreover, could the pain ratings confound TEP differences between conditions, e.g., by providing earlier ratings when the stimulus is painful? This should be carefully considered, and the authors might perform control analyses.

It is unlikely that the verbal ratings contaminated the TEP response as the subsequent TMS pulse was not delivered until the verbal rating was complete and given that each participant was cued by the experimenter to provide the pain rating after each pulse (rather than the participant giving the rating at any time). As such, it would not be possible for participants to provide earlier ratings to more painful stimuli. We will make this part of the protocol clearer in the next revision of the manuscript.

4) Confounds by time effects. Non-painful and painful conditions were performed in a fixed order. Potential confounds by time effects should be carefully considered.

Previous research suggests that pain alters neural excitability even after pain has subsided. In a recent meta-analysis (Chowdhury et al., 2022) we found effect sizes of 0.55-0.9 for MEP reductions 0-30 minutes after pain had resolved. As such, we avoided intermixing pain and warm blocks given subsequent warm blocks would not serve as a valid baseline, as each subsequent warm block would have residual effects from the previous pain blocks.

At the same time, given there was no conclusive evidence for a difference in N45 amplitude between pre-pain and post-pain conditions of Experiment 1 (Supplementary Figure 1), it is unlikely that the effect of pain was an artefact of time i.e., the explanation that successive thermal stimuli applied to the skin results an increase in the N45, regardless of whether they are painful or not. We will make this point in our next revision.

Chowdhury, N. S., Chang, W. J., Millard, S. K., Skippen, P., Bilska, K., Seminowicz, D. A., & Schabrun, S. M. (2022). The Effect of Acute and Sustained Pain on Corticomotor Excitability: A Systematic Review and Meta-Analysis of Group and Individual Level Data. The Journal of Pain, 23(10), 1680-1696.

5) Data availability. The authors should state how they make the data openly available.

We will upload the MEP, TEP and pain data on the Open science framework at the time of the next revision.

Author Response

Reviewer #1 (Public Review):

Sun and colleagues investigated the cross-reactive antibodies between E. coli and the host in severe alcoholic hepatitis (SAH). The study found that IgA and IgG were deposited in the liver of SAH patients. Complements C3d and C4d were also deposited in the SAH patient's liver. Moreover, they found that the Ig accumulated in the SAH liver, but not in the SAH serum, induced hepatocyte killing, suggesting that liver Ig is important. Then, they found that these Ig can recognize both human and E. coli antigens. Very interestingly, SAH-derived Ig shows cross-reactivity to both human and E. coli antigens, suggesting E. coli-primed Ig in SAH may damage hepatocytes through host antigen recognition. These Ig are not observed in alcoholic cirrhosis patients. The liver RNA-seq data suggested that Ig was also produced in the liver, not only gut-derived Ig. This is a very interesting study showing the novel mechanism of SAH mediated by the Ig with the cross-reactivity with bacteria and host antigens, which is not observed in AC patients. Overall, the study design is reasonable and the data are consistent to support their central hypothesis. There are a few comments.

We thank the Reviewer for his/her positive comments on our manuscript!

Specific comments:

1) Figures 1 and 2 show Ig deposition in the liver (it seems on hepatocytes). Not only Ig reaction to the specific antigen but also non-specific Fc receptor-mediated binding to hepatocytes could have contributed.

2) Similarly, in Figure 2G Ig-mediated hepatocyte killing, Fc receptor-mediated hepatocyte killing may be involved.

Anti-IgG antibody (ab200699) recognizes a protein of 75 kDa, identified as gamma heavy chain of human immunoglobulins. It is possible that non-specific Fc receptor-mediated binding to hepatocytes in the SAH liver can also be recognized by this anti-IgG antibody staining.

However, no IgG or IgA deposition in the healthy donor livers was identified by anti-IgG or IgA staining. These results suggest that there was no antigen specific or Fc receptor-mediated binding to healthy hepatocytes.

In the ADCC assay, hepatocytes isolated from healthy donor livers were used as the target cells. Immune cell (NK) mediated ADCC is mainly triggered by IgG (binding to antigens of hepatocytes) through the interaction between IgG Fc fragment and Fc-receptors (FcγRs) of NK cells. If IgG deposition in the SAH liver were mainly due to non-specific Fc receptor-mediated binding to hepatocytes, we would expect IgG binding to FcγRs of hepatocytes and no activation of NK cells. Ig-mediated hepatocyte killing (Figure 2G) indicates the Ig (from SAH liver) reaction to the specific antigens.

3) The study examined the possibility of liver resident B cell and plasma cell-mediated Ig. As the authors mentioned in the discussion, B cells may be translocated from the intestine to the liver. Or the resident B cells (not from the gut) are also involved.

We agree with the Reviewer at this point. The resident B cells may be also involved in the Ig production.

Reviewer #2 (Public Review):

In this paper, Ahmadi et al demonstrated that antibodies produced locally in the liver by infiltrating B cells can enhance liver damage caused by fat accumulation. The main finding is that human samples extracted from severe alcoholic hepatitis showed antibody accumulation that may be related to an enhanced immune response to self-antigens, which could ultimately fuel liver damage - which was already present due to alcohol consumption. Their data are corroborated by arrays and gene ontology assays, and I strongly believe that these data could add to the future options we have to treat patients.

We thank the Reviewer for his/her positive comments on our manuscript!

Author Response:

We thank the reviewers and eLife editorial team for their valuable assessment. While additional experiments could further strengthen the theoretical framework proposed in this study, we believe that we have successfully established the delayed nuclear export of hemagglutinin and neuraminidase mRNAs by quantifying the FISH observation with the mathematical model. We agree that this finding raises a further important question to be addressed regarding the molecular mechanism underlying the prolonged nuclear retention of these segments. Our ongoing investigation is focusing on identifying potential cis-elements that contribute to the delay of these segments.