Reviewer #2 (Public review):

Summary:

This manuscript investigates how neural cell development is affected in Lowe syndrome. Using neural cultures differentiated from human iPSCs carrying either an LS mutation or a genetically engineered mutation in OCRL, the authors show a depletion of mitochondrial DNA and a decrease in mitochondrial activities that correlate with an increased formation of astrocytes at the expense of neurons. Similar effects on mitochondria and on astrocyte development were observed in an LS mouse model. Moreover, these mutant brain cells are less likely to be ciliated and show a reduction in Sonic Hedgehog signalling.

Strengths/Weaknesses:

The study derives strength from the analyses of two different models of Lowe syndrome, both reaching similar conclusions. However, the observed changes in mitochondrial defects, neuronal/astrocytic development, and primary cilia are only correlated, with no attempt to investigate a causal relationship. Moreover, the mouse model is only analysed at the adult stage providing no insights into the development of the defects. Different brain regions are analysed with immunostainings and qRT-PCR making it challenging to draw clear correlations between these findings. The quality of the corresponding figures is often poor and the selection of markers is frequently inappropriate. Taken together, these limitations complicate the interpretations of the data and significantly limit the conclusions that can be drawn from the study.

eLife Assessment

This valuable study proposes a theoretical model of clathrin coat formation based on membrane elasticity that seeks to determine whether this process occurs by increasing the area of a protein-coated patch with constant curvature, or by increasing the curvature of a protein-coated patch that forms in an initially flat conformation (so called constant curvature or constant area models). Identifying energetically favorable pathways and comparing the obtained shapes with experiments provides solid support to the constant-area pathway. This work will be of interest for biologists and biophysicists interested in membrane remodelling and endocytosis. It provides an innovative approach to tackle the question of constant curvature vs. constant area coat protein formation, although some of the model's assumption are only partially supported by experimental evidence.

Reviewer #1 (Public review):

Summary:

The authors develop a set of biophysical models to investigate whether a constant area hypothesis or a constant curvature hypothesis explains the mechanics of membrane vesiculation during clathrin-mediated endocytosis.

Strengths:

The models that the authors choose are fairly well-described in the field and the manuscript is well-written.

Weaknesses:

One thing that is unclear is what is new with this work. If the main finding is that the differences are in the early stages of endocytosis, then one wonders if that should be tested experimentally. Also, the role of clathrin assembly and adhesion are treated as mechanical equilibrium but perhaps the process should not be described as equilibria but rather a time-dependent process. Ultimately, there are so many models that address this question that without direct experimental comparison, it's hard to place value on the model prediction.<br /> While an attempt is made to do so with prior published EM images, there is excessive uncertainty in both the data itself as is usually the case but also in the methods that are used to symmetrize the data. This reviewer wonders about any goodness of fit when such uncertainty is taken into account.

Reviewer #2 (Public review):

Summary:

In this manuscript, the authors employ theoretical analysis of an elastic membrane model to explore membrane vesiculation pathways in clathrin-mediated endocytosis. A complete understanding of clathrin-mediated endocytosis requires detailed insight into the process of membrane remodeling, as the underlying mechanisms of membrane shape transformation remain controversial, particularly regarding membrane curvature generation. The authors compare constant area and constant membrane curvature as key scenarios by which clathrins induce membrane wrapping around the cargo to accomplish endocytosis. First, they characterize the geometrical aspects of the two scenarios and highlight their differences by imposing coating area and membrane spontaneous curvature. They then examine the energetics of the process to understand the driving mechanisms behind membrane shape transformations in each model. In the latter part, they introduce two energy terms: clathrin assembly or binding energy, and curvature generation energy, with two distinct approaches for the latter. Finally, they identify the energetically favorable pathway in the combined scenario and compare their results with experiments, showing that the constant-area pathway better fits the experimental data.

Strengths:

The manuscript is well-written, well-organized, and presents the details of the theoretical analysis with sufficient clarity.<br /> The calculations are valid, and the elastic membrane model is an appropriate choice for addressing the differences between the constant curvature and constant area models.<br /> The authors' approach of distinguishing two distinct free energy terms-clathrin assembly and curvature generation-and then combining them to identify the favorable pathway is both innovative and effective in addressing the problem.<br /> Notably, their identification of the energetically favorable pathways, and how these pathways either lead to full endocytosis or fail to proceed due to insufficient energetic drives, is particularly insightful.

Weaknesses:

Membrane remodeling in cellular processes is typically studied in either a constant area or constant tension ensemble. While total membrane area is preserved in the constant area ensemble, membrane area varies in the constant tension ensemble. In this manuscript, the authors use the constant tension ensemble with a fixed membrane tension, σe. However, they also use a constant area scenario, where 'area' refers to the surface area of the clathrin-coated membrane segment. This distinction between the constant membrane area ensemble and the constant area of the coated membrane segment may cause confusion.

As mentioned earlier, the theoretical analysis is performed in the constant membrane tension ensemble at a fixed membrane tension. The total free energy E_tot of the system consists of membrane bending energy E_b and tensile energy E_t, which depends on membrane tension, σe. Although the authors mention the importance of both E_b and E_t, they do not present their individual contributions to the total energy changes. Comparing these contributions would enable readers to cross-check the results with existing literature, which primarily focuses on the role of membrane bending rigidity and membrane tension.

The authors introduce two different models, (1,1) and (1,2), for generating membrane curvature. Model 1 assumes a constant curvature growth, corresponding to linear curvature growth, while Model 2 relates curvature growth to its current value, resembling exponential curvature growth. Although both models make physical sense in general, I am concerned that Model 2 may lead to artificial membrane bending at high curvatures. Normally, for intermediate bending, ψ > 90, the bending process is energetically downhill and thus proceeds rapidly. the bending process is energetically downhill and thus proceeds rapidly. However, Model 2's assumption would accelerate curvature growth even further. This is reflected in the endocytic pathways represented by the green curves in the two rightmost panels of Fig. 4a, where the energy steeply increases at large ψ. I believe a more realistic version of Model 2 would require a saturation mechanism to limit curvature growth at high curvatures.

eLife Assessment

This study provides valuable quantitative data and analysis that reveals variations in 'Dorsal' nuclear dynamics along the dorso-ventral axis in the early Drosophila embryo. The evidence that supports that these variations are due to Dorsal/Cactus interactions in dorsal nuclei is convincing, albeit incomplete to understand the biological implications of these findings for developmental patterning.

Reviewer #1 (Public review):

Summary:

Al Asafen and colleagues apply a set of scanning fluorescence correlation spectroscopic approaches (Raster Image Correlation Spectroscopy (RICS), cross-correlation RICS, and pair-correlation function spectroscopy) to address the nuclear-cytoplasmic kinetics of the Dorsal (Dl) transcription factor in early Drosophila embryos. The Toll/Dl system has long been appreciated to establish dorsal-ventral polarity of the embryo through Toll-dependent control of Dl nuclear localization, and provides an example of a morphogen gradient produced with high enough precision to yield robust biophysical measurements of general transcription factor activity and function. By measuring GFP-tagged Dl protein, either in wild-type embryos or in mutant embryos with low/medium/high levels of Toll signaling, the authors report diffusivity of Dl in nuclear and cytoplasmic compartments of the embryo, as well as the fraction of mobile and immobile Dl, which can be correlated with DNA binding through cross-correlation RICS. A model is presented where Cactus/IkB is implicated in preventing Dl from binding to DNA.

Strengths:

The experiments on wild-type GFP-tagged Dorsal are performed well, are mostly reported well, and are interpreted fairly.

Weaknesses:

The discrepancy between experiment and theory as pertains to Michaelis-Menten kinetics is not fully motivated in the text, and could benefit from a more clear presentation. The experiments performed to distinguish between the contribution of Toll-dependent phosphorylation and Cactus interaction models for limiting Dorsal DNA binding are possibly confounded by the presence of wild-type, GFP-tagged Dorsal protein.

Reviewer #2 (Public review):

Summary:

In this manuscript, Al Asafen, Clark et al., use fluorescence correlation spectroscopy (FCS) to quantitatively analyze the mobility of Dl along the DV axis of the early Drosophila embryo. Dl is essential for dorsal-ventral (DV) patterning and its gradient initiates the activation of several genes and thereby orchestrates the formation of the Drosophila body plan. While the mechanisms underlying the formation of the Dl gradient have been extensively studied by this group and others, there are some observations for which there is not yet a mechanistic explanation. For example, the peak of the Dl gradient grows continuously during nuclear cycles 10-14. This is likely due to Cact-dependent Dl diffusion and Dl binding to DNA. However, the biophysical parameters governing Dl nuclear dynamics that would support these claims have not been previously measured. In this work, the authors provide evidence that GFP-tagged Dl may be separated into a mobile pool and an immobile pool. Interestingly, the fraction of immobile Dl is position-dependent along the DV axis, revealing more binding to DNA in the ventral than in the dorsal nuclei. This is either due to higher binding affinity in ventral locations (due to Toll-dependent Dl phosphorylation) or to higher Dl-Cact binding in dorsal nuclei that would prevent Dl from binding to DNA. Using dl-mutant alleles, the authors support the latter hypothesis.

Strengths:

The manuscript is well written and their conclusions are convincingly supported by their methodology and analysis. As a quantitative study, the biophysical analysis seems rigorous, in general.

Although this is not the first study that employs FSC to investigate the dynamics of a morphogen, it further exemplifies how these quantitative tools can be used to uncover mechanistic aspects of morphogen dynamics during development. In particular, the manuscript reports novel biophysical parameters of Dl dynamics that will be helpful in future hypotheses-driven modeling studies.

Weaknesses:

In my opinion, the main weakness of the manuscript is that the main biological implication of the study, namely that the asymmetry in the fraction of immobile Dl is a result of nuclear Dl-Cact binding which prevents Dl from binding DNA (Figure 5), occurs in a region of the embryo where there is very little Dl anyways (Figure 1A, 5A). While it is interesting that the fraction of immobile Dl increases (just a little, but significantly) in dorsal nuclei in mutants expressing a form of Dl with reduced Cact binding it is unclear what is the biological impact of this effect in a location where Dl is nearly absent. As can be seen in Figure 3F, the fraction of immobile is unaffected in Dl-mutant forms with reduced DNA binding, because it is already very low. It is unlikely that Dl binding to Cact in dorsal nuclei would affect shuttling as well since the fraction is very low anyway.

While the authors have a very clear understanding of the biology of the Dl gradient, I feel that the manuscript is more written as a 'tools' paper (i.e., to exemplify how FSC methods and analysis can be used for biological discovery). This is ok, but I think that the authors should discuss further what are the biological implications of these findings other than the contribution to uncovering the biophysical parameters. For example, I think that the implications of the rejected hypothesis (i.e., that Toll-dependent Dl phosphorylation does not seem to have an impact on Dl binding affinities to DNA) are important and should be further discussed (even if no additional experiments are performed). What is then the role of Dl phosphorylation? Perhaps it could have an impact on patterning robustness in lateral regions. The authors should report in Figure 5 also what happens to the fraction of Dl bound to DNA in lateral regions in the reduced Cact binding and reduced Toll phosphorylation mutants.

The way that position along the DV axis is reported using the nuclear-cytoplasmic-ratio (NCR) in Figures 1-3 is not incorrect, but I wonder if it is the best way of doing it. The reason is that it spreads out a relatively small region of the embryo (the ventral-most locations) and shrinks a relatively large region of the embryo (lateral and dorsal regions), see Figure 1A. Perhaps reporting the NCR in log_2 units would be more appropriate.

eLife Assessment

With compelling electrophysiological and behavioural evidence, this work establishes that the activity of insulin-producing cells (IPCs) depends on the nutritional state in Drosophila and that, like in mammals, there is also an incretin-like effect with IPCs responding to glucose feeding but not to glucose perfusion. Moreover, the authors demonstrate that DH44 neurons respond to glucose perfusion and, together with IPCs, modulate locomotor activity. This important study on the neuronal regulation of metabolic homeostasis will be of interest to both neuroscience and to medical research in diabetes.

Reviewer #1 (Public review):

Summary:

This study presents useful insights into the in vivo dynamics of insulin-producing cells (IPCs), key cells regulating energy homeostasis across the animal kingdom. The authors further provide compelling evidence using adult Drosophila melanogaster that IPCs, unlike neighboring DH44 cells, do not respond to glucose directly, but that glucose can indirectly regulate IPC activity after ingestion supporting an incretin-like mechanism in flies similar to mammals. The authors link decreased activity of IPCs to hyperactivity observed in starved flies, a locomotive behavior aimed to increase food search. Furthermore, the authors provide evidence that IPCs receive inhibitory inputs from Dh44 neurons, which are linked to increased locomotor activity.

This paper is of outstanding interest to scientists aiming to understand metabolic control of circuit dynamics, in particular for internal state-linked behaviors competing with the feeding state.

Strengths:

(1) By using whole cell patch clamp recording, the authors convincingly showed the activity pattern and regulation of IPCs and neighboring DH44 neurons under different feeding states and in various refeeding paradigms.<br /> (2) The paper provides compelling evidence that IPCs are not directly and acutely activated by glucose, but rather through a post-ingestive incretin-like mechanism. In addition, the authors show that Dh44 neurons located adjacent to the IPCs respond to bath application of nutritive sugars contrary to the IPCs.<br /> (3) The paper also provides useful data on the regulation of IPC activity by Dh44 neurons, which is useful to understand their regulation in vivo.

No major weaknesses remain in the revised version of this work.

Reviewer #2 (Public review):

Summary

In this study, Bisen et al. characterized the state-dependency of insulin-producing cells in the brain of Drosophila melanogaster. They successfully established that IPC activity is modulated by the nutritional state and age of the animal. Interestingly, they demonstrate that IPCs respond to the ingestion of glucose, rather than to perfusion with it, an observation reminiscent of the incretin effect in mammals. The study is well conducted and presented and the experimental data convincingly support the claims made.

Strengths

The study makes great use of the tools available in *Drosophila* research, demonstrating the effect that starvation and subsequent refeeding have on the physiological activity of IPCs as well as on the behavior of flies to then establish causal links by making use of optogenetic tools.<br /> It is particularly nice to see how the authors put their findings in context to published research and use for example TDC2 neuron activation or DH44 activity to establish baselines to relate their data to.

Reviewer #3 (Public review):

Although insulin release is essential in the control of metabolism, adjusted to nutritional state, and plays major roles in normal brain function as well as in aging and disease, our knowledge about the activity of insulin-producing (and releasing) cells (IPCs) in vivo in limited.

In this technically demanding study, IPC activity is studied in the Drosophila model system by fine in vivo patch clamp recordings with parallel behavioral analyses and various optogenetic as well as feeding manipulations.

The data provide compelling evidence that IPC activity is increased with a slow time course after feeding a high glucose diet. By contrast, IPC activity is not directly affected by rising blood glucose levels. This is reminiscent of the incretin effect known from vertebrates and points to a conserved mechanism in insulin production and release upon sugar feeding.

Moreover, the data confirm earlier studies that nutritional state strongly affects locomotion. Surprisingly, strong evidence shows that IPC activity makes only a negligible contribution to this. Instead, other modulatory neurons that are directly sensitive to blood glucose levels strongly affect locomotion. Together, these data reveal a network of multiple parallel and interacting neuronal layers to orchestrate the physiological, metabolic, and behavioral responses to the nutritional state. Together with the data from a previous study, this work sets the stage to dissect the architecture and function of this network.

Strengths:

State-of-the-art current clamp in situ patch clamp recordings in behaving animals are a demanding but powerful method to provide novel insight into the interplay of nutritional state, IPC activity, and locomotion. The patch clamp recordings and the parallel behavioral analyses are of high quality, as are the optogenetic manipulations. The data showing that starvation silences IPC activity in young flies (younger than 1 week) are excellent. The evidence for the claim that locomotor activity is not increased upon IPC activity but upon the activity of other blood glucose sensitive modulatory neurons (Dh44) is compelling, too. The study provides a great system to experimentally dissect the interplay of insulin production and release with metabolism, physiology, nutritional state, and behavior. Demonstrating the incretin effect in Drosophila provides novel experimental routes to further study it. During the revision process, compelling evidence has been added to underscore the incretin effect, the finding that IPCs themselves do not sense sugars, and that feeding a high sugar diet does not cause unspecific stress responses.

I found no more weaknesses: The authors have carefully addressed all of my previous critiques by adding compelling new data and carefully revising the text. This paper provides a prime example of how responsible authors can utilize this constructive (but relatively new) reviewing procedure to make a very good manuscript even better.

Author response:

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

Public Reviews: 

Reviewer #1 (Public Review): 

Summary: 

This study presents useful insights into the in vivo dynamics of insulin-producing cells (IPCs), key cells regulating energy homeostasis across the animal kingdom. The authors provide compelling evidence using adult Drosophila melanogaster that IPCs, unlike neighboring DH44 cells, do not respond to glucose directly, but that glucose can indirectly regulate IPC activity after ingestion supporting an incretin-like mechanism in flies, similar to mammals. The authors link the decreased activity of IPCs to hyperactivity observed in starved flies, a locomotive behavior aimed at increasing food search. 

Furthermore, there is supporting evidence in the paper that IPCs receive inhibitory inputs from Dh44 neurons, which are linked to increased locomotor activity. However, although the electrophysiological data underlying the dynamics of IPCs in vivo is compelling, the link between IPCs and other potential elements of the circuitry (e.g. octopaminergic neurons) regulating locomotive behaviors is not clear and would benefit from more rigorous approaches. 

This paper is of interest to cell biologists and electrophysiologists, and in particular to scientists aiming to understand circuit dynamics pertaining to internal state-linked behaviors competing with the feeding state, shown here to be primarily controlled by the IPCs. 

Strengths: 

(1) By using whole-cell patch clamp recording, the authors convincingly showed the activity pattern of IPCs and neighboring DH44 neurons under different feeding states. 

(2) The paper provides compelling evidence that IPCs are not directly and acutely activated by glucose, but rather through a post-ingestive incretin-like mechanism. In addition, the authors show that Dh44 neurons located adjacent to the IPCs respond to bath application of glucose contrary to the IPCs. 

(3) The paper provides useful data on the firing pattern of 2 key cell populations regulating foodrelated brain function and behavior, IPCs and Dh44 neurons, results which are useful to understand their in vivo function. 

Weaknesses: 

(1) The term nutritional state generally refers to the nutrients which are beneficial to the animal. In Figure 1, the authors showed that IPCs respond to glucose but not proteins. To validate the term nutritional state the authors could test the effect of a non-nutritive sugar (e.g. D-arabinose or L-Glucose) on the post-ingestive physiological responses of the IPCs.

We thank the referee for this insightful comment. Following their suggestion, we included two new experimental data sets, which we added to Figure 1: We show that IPCs do not respond to the non-nutritive sugar D-arabinose (Figure 1H). In order to further expand this data set and our conclusions, we additionally show that IPCs do respond to fructose – a second nutritive sugar in addition to glucose (Figure 1H). Together, these data sets permit the conclusion that IPCs are sensitive to the ingestion of nutritive sugars, and do not respond to ingestion of nonnutritive sugars or high protein diets. Thus, we validate the term nutritional state.

(2) It is difficult to grasp the main message from the figures in the result section as some figures have several results subsections referring to different points the authors want to make. The key results of a figure will be easier to understand if they are summarized in one section of the results. Alternatively, a figure can be split into 2 figures if there are several key messages in those figures, e.g. Figures 2 and 3.  

We appreciate this suggestion and have made several changes to our manuscript to add more clarity. Among other things, we have changed the order of data presentation in Figure 2, as suggested by the referee below, where we now start with the IPC activation data rather than the OAN activation. We also swapped the order of data presentation and split Figure S1 into Figures S1 & S2. Moreover, we re-arranged the panel order in supplementary figure S4. This significantly improved the flow of the results section. Since the figures the referee refers to contain comparative data, for example between diets (Figure 1) or neuron types (Figure 2), we prefer to keep these data sets together. However, we have carefully revised the results section to more clearly relate our statements to individual figure panels.

(3) The prime investigation of the paper is about the physiological response and locomotive behavioral readout linked to IPCs. The authors do not show a link between OANs and IPCs in terms of functional or behavioral readouts. In Figure 2 the authors first start with stating a link between OAN neurons and locomotion changes resulting from internal feeding states. The flow of the paper would be better if the authors focused on the effect of optogenetic activation of IPCs under different feeding states and their impact on fly locomotion. If the experiments done on optogenetic activation of OANs were to validate the experimental approach the data on OAN neurons is better suited for the supplement without the need of a subsection in the result section on the OANs.  

We agree with the reviewer’s suggestion and switched the order of the figure panels and text to aid the flow of the manuscript. We now show and discuss the IPC activation data first (Figure 2C-H) and OAN activation afterwards (Figure 2I-K). We did keep the OAN data in the main document, though, since that facilitates comparisons between the small effects of IPC activation and the large, well-established effects of OAN activation.

(4) Figure 2F shows that optogenetic activation of IPCs in fed flies does not influence their locomotor output. In the text, the conclusion linked to Figure 2F-H states that IPC activation reduces starvation-induced hyperactivity which is a statement more suited to Figure 2I-K. 

We edited the text accordingly.

(5) The authors show activation of Dh44 neurons leads to hyperpolarisation of the IPCs. What is the functional link between non-PI Dh44 neurons and the IPCs? Do IPCs express DH44R or is DH44 required for this effect on IPCs? Investigating a potential synaptic or peptidergic link between DH44 neurons and IPCs and its effect on behavior would benefit the paper, as it is so far not well connected. 

Although we have not performed any experiments dedicated to investigating the functional link between DH44Ns outside the PI and the IPCs in this study, there are two lines of evidence supporting that this connection is relatively direct. First, IPCs do express DH44R1 & R2, as we show in a parallel study in eLife (Held M, et al. ‘Aminergic and peptidergic modulation of Insulin-Producing Cells in Drosophila’. eLife. 2024;13. doi:10.7554/ELIFE.99548.1). Second, we performed functional connectivity experiments using a Leucokinin (LK) driver line in that paper. This driver line labels two pairs of non-PI DH44Ns in the VNC, which are DH44 and LK positive (Zandawala et al 2018). Activating that line leads to inhibition of IPCs, similar to the effect we observed here for DH44N activation. These two lines of evidence suggest that there could be a direct peptidergic connection between DH44+ neurons and IPCs. We have added a paragraph mentioning these experiments to our discussion:

‘Notably, the DH44^PINs express the DH44 peptide, as confirmed by anti-DH44 stainings(100). This also applies to a large fraction of neurons labelled in the broad DH44 driver line(100). However, a subset of neurons labelled in the broad line did not exhibit DH44 immunoreactivity(100), and might therefore not actually express the DH44 peptide. Hence, the inhibition of IPCs could be driven by neurons in the DH44 driver line that do not express DH44. A strong candidate for the inhibition are LK and DH44-positive neurons, which are labelled by the broad line(76). In a parallel study, we showed that LK-expressing neurons strongly inhibit IPCs(30), similar to the broad DH44 line used here. Furthermore, evidence from single-nucleus transcriptomic analysis shows that IPCs express DH44-R1 and DH44-R2 receptors(30). Therefore, it is possible that DH44Ns communicate with IPCs through a direct peptidergic connection. Notably, the inhibitory effect of non-PI DH44Ns on IPCs was very strong and fast, suggesting that a connection via classical synapses is more likely. Regardless, our results show that the glucose sensing DH44^PINs and IPCs act independently of each other.’

Reviewer #2 (Public Review): 

Summary: 

In this study, Bisen et al. characterized the state-dependency of insulin-producing cells in the brain of *Drosophila melanogaster*. They successfully established that IPC activity is modulated by the nutritional state and age of the animal. Interestingly, they demonstrate that IPCs respond to the ingestion of glucose, rather than to perfusion with it, an observation reminiscent of the incretin effect in mammals. The study is well conducted and presented and the experimental data convincingly support the claims made. 

Strengths: 

The study makes great use of the tools available in *Drosophila* research, demonstrating the effect that starvation and subsequent refeeding have on the physiological activity of IPCs as well as on the behavior of flies to then establish causal links by making use of optogenetic tools. 

It is particularly nice to see how the authors put their findings in context to published research and use for example TDC2 neuron activation or DH44 activity to establish baselines to relate their data to. 

Weaknesses: 

I find the inability of SD to rescue the IPC starvation effect in Figure 1G&H surprising, given that the fully fed flies were raised and kept on that exact diet. Did the authors try to refeed flies with SD for longer than 24 hours? I understand that at some point the age effect would also kick in and counteract potential IPC activity rescue. I think the manuscript would benefit if the authors could indicate the exact age of the SD refed flies and expand a bit on the discussion of that point.  

We have expanded the first paragraph of our discussion to tackle these questions, in particular the potential effect of aging, as suggested by the referee. We now also indicate the exact age of the flies. Moreover, we have conducted additional experiments in which we added either glucose or arabinose to our standard diet (Figure 1H). As we would have expected based on our hypothesis that the glucose concentration in our standard diet was too low to cause an increase in IPC activity after starvation, we find that feeding standard diet plus glucose increases IPC activity to the same level as glucose only, and that adding arabinose to the standard diet does not lead to increased IPC activity after starvation (Figure 1H).

The incretin-like effect is exciting and it will be interesting in the future to find out what might be the signal mediating this effect. It is interesting that IPCs in explants seem to be responsive to glucose. I think it would help if the authors could briefly discuss possible sources for the different findings between these in fact very different preparations. Could the the absence of the inhibitory DH44 feedback in the *ex-vivo* recordings for example play a role? 

We thank the referee for this interesting point and expanded our discussion accordingly. We included that, in particular in brain explants without a VNC, the inhibitory connection we describe might be absent, as the referee suggested: ‘Previous ex vivo studies suggested that IPCs, like pancreatic beta cells, sense glucose cell-autonomously(23,24). Consistent with this, we observed an increase in IPC activity after the ingestion of glucose (Figure 2B). However, IPC activity did not increase during the perfusion of glucose directly over the brain. Importantly, the fly preparations were kept alive for several hours allowing the glucose-rich saline to enter circulation and reach all body parts. Several factors may explain the difference between ex vivo and in vivo preparations. First, in ex vivo studies, certain regulatory feedback mechanisms present in vivo could be absent. For example, the strong inhibitory input IPCs receive from DH44Ns we found would likely be absent in brain explants without a VNC. A lack of inhibitory feedback might allow for more direct glucose sensing by IPCs ex vivo, whereas in vivo, the IPC response could be suppressed by more complex systemic feedback. Second, we attempted to use the intracellular saline formulation employed in a previous ex vivo study44. However, we observed that IPCs depolarized quickly using this saline, leading to unstable recordings that did not meet our quality standards for in vivo experiments. Another possible explanation for the lack of an effect of glucose might have been that the dominant circulating sugar in flies is trehalose(70,71) which is derived from glucose. When we extended our experiments, we found that trehalose perfusion did not affect IPC activity either, strengthening the idea that IPCs do not directly sense changes in hemolymph sugar levels. Therefore, our findings suggest that, similar to mammals, IPC activity and hence, insulin release, is not simply modulated by hemolymph sugar concentration in Drosophila.’ 

The incretin-like effect the authors observed seems to start only after 5h which seems longer than in mammals where, as far as I know, insulin peaks around 1h. Do the authors have ideas on how this timescale relates to ingestion and glucose dynamics in flies? 

We have now included the following section in the discussion to explicitly address the question of different activity dynamics in flies and mammals, but also the limitations of our electrophysiological approach in this regard: ‘We observed that IPC activity increased over a timescale of hours, which is longer compared to the fast insulin response in mammals, where insulin typically peaks within an hour of feeding(97). In flies, insulin levels rise within minutes of refeeding, followed by a drop after 30 min(20). Our experimental techniques limit our ability to capture these fast initial dynamics, since the preparation for intracellular recordings requires tens of minutes, so that we typically recorded IPC activity at least 20 min after the last food ingestion. Notably, studies in fasted mammals have shown that insulin peaks within minutes of refeeding, followed by a rapid decline, with levels stabilizing as feeding continues(98,99). We speculate a similar dynamic could be present in flies, but with our approach, we capture the steady-state reached tens of minutes after food ingestion rather than a potential initial peak.’ 

The authors mention "a decrease in the FV of IPC-activated starved flies even before the first optogenetic stimulation (Figure 2I),". Could this be addressed by running an experiment in darkness, only using the IR illumination of their behavioral assay? 

We thank the referee for pointing out this unexpected result. We discuss this in more detail in the new version of our manuscript and expand on the reasons for not performing these optogenetic activation experiments in the dark: First, the red LED required to activate CsChrimson triggers strong startle responses in dark-adapted flies, which mask other behavioral effects, in particular subtle ones such as those observed for IPCs. The startle response is much reduced when performing experiments under low background light conditions. Second, flies, at least in our hands, do not exhibit robust foraging behavior or starvation-induced hyperactivity in the dark, which is critical for our behavioral experiments. However, we also explain in our discussion that we believe the effect of background illumination is relatively small, since flies expressing CsChrimson in OANs or DH44Ns show comparable activity levels to controls. Hence, a part of this effect is likely attributable to leak currents induced by CsChrimson expression. We would like to point out though that we are careful in our description of the IPC effect on behavior, and focus on the fact that it is considerably smaller than the effects of other modulatory neurons (DH44Ns and OANs).

The authors show an inhibitory effect of DH44 neuron activation on IPC activity. They further demonstrate that DH44PI neurons are not the ones driving this and thus conclude that "...IPCs are inhibited by DH44Ns outside the PI.". As the authors mentioned the broad expression of the DH44-Gal4 line, can they be sure that the cells labeled outside the PI are actually DH44+? If so they should state this more clearly, if not they should adapt the discussion accordingly.   

We have substantially added to our discussion of this point, according to the referee’s great suggestion. In short, the broad line includes neurons that are DH44 positive and neurons that are not: ‘Notably, the DH44^PINs express the DH44 peptide, as confirmed by anti-DH44 stainings(100). This also applies to a large fraction of neurons labelled in the broad DH44 driver line(100). However, a subset of neurons labelled in the broad line did not exhibit DH44 immunoreactivity(100), and might therefore not actually express the DH44 peptide. Hence, the inhibition of IPCs could be driven by neurons in the DH44 driver line that do not express DH44.’

Reviewer #3 (Public Review): 

Although insulin release is essential in the control of metabolism, adjusted to nutritional state, and plays major roles in normal brain function as well as in aging and disease, our knowledge about the activity of insulin-producing (and releasing) cells (IPCs) in vivo is limited. 

In this technically demanding study, IPC activity is studied in the Drosophila model system by fine in vivo patch clamp recordings with parallel behavioral analyses and optogenetic manipulation. 

The data indicate that IPC activity is increased with a slow time course after feeding a high-glucose diet. By contrast, IPC activity is not directly affected by increasing blood glucose levels. This is reminiscent of the incretin effect known from vertebrates and points to a conserved mechanism in insulin production and release upon sugar feeding. 

Moreover, the data confirm earlier studies that nutritional state strongly affects locomotion. Surprisingly, IPC activity makes only a negligible contribution to this. Instead, other modulatory neurons that are directly sensitive to blood glucose levels strongly affect modulation. Together, these data indicate a network of multiple parallel and interacting neuronal layers to orchestrate the physiological, metabolic, and behavioral responses to nutritional state. Together with the data from a previous study, this work sets the stage to dissect the architecture and function of this network. 

Strengths: 

State-of-the-art current clamp in situ patch clamp recordings in behaving animals are a demanding but powerful method to provide novel insight into the interplay of nutritional state, IPC activity, and locomotion. The patch clamp recordings and the parallel behavioral analyses are of high quality, as are the optogenetic manipulations. The data showing that starvation silences IPC activity in young flies (younger than 1 week) are compelling. The evidence for the claim that locomotor activity is not increased upon IPC activity but upon the activity of other blood glucose-sensitive modulatory neurons (Dh44) is strong. The study provides a great system to experimentally dissect the interplay of insulin production and release with metabolism, physiology, and behavior. 

Weaknesses: 

Neither the mechanisms underlying the incretin effect, nor the network to orchestrate physiological, metabolic, and behavioral responses to nutritional state have been fully uncovered. Without additional controls, some of the conclusions would require significant downtoning. Controls are required to exclude the possibility that IPCs sense other blood sugars than glucose. The claim that IPC activity is controlled by the nutritional state would require that starvation-induced IPC silencing in young animals can be recovered by feeding a normal diet. At current firing in starvation, silenced IPCs can only be induced by feeding a high-glucose diet that lacks other important ingredients and reduces vitality. Therefore, feasible controls are needed to exclude that diet-induced increases in IPC firing rate are caused by stress rather than nutritional changes in normal ranges. The finding that refeeding starved flies with a standard diet had no effect on IPC activity but a strong effect on the locomotor activity of starved flies contradicts the statement that locomotor activity is affected by the same dietary manipulations that affect IPC activity. The compelling finding that starvation induces IPC firing would benefit from determining the time course of the effect. The finding that IPCs are not active in fed animals older than 1 week is surprising and should be further validated. 

We thank the referee for the thoughtful and constructive criticism of our experiments and conclusions. Below, we lay out how we tackled the individual points raised by the referee.

(1) ‘Controls are required to exclude the possibility that IPCs sense other blood sugars than glucose.’  

To address this point, we conducted experiments in which we perfused trehalose (Figure 3B), the main circulating hemolymph sugar in Drosophila and other insects. Our results clearly show that trehalose does not affect IPC activity upon perfusion, confirming our statements that IPCs do not sense key blood sugars directly.

(2) ‘Feasible controls are needed to exclude that diet-induced increases in IPC firing rate are caused by stress rather than nutritional changes in normal ranges’. 

We agree with the referee that this point was not completely fleshed out in our first submission. We have now performed additional experiments in which we added glucose (and fructose) to our standard diet (Figure 1H). Flies feeding on this diet received all necessary nutrients but still experienced high concentrations of sugars. The effects of high glucose in a standard diet background were indistinguishable from those of high glucose in agarose, confirming that the IPCs respond to sugar rather than stress. Another important observation in this context is that IPCs in flies kept on a high protein diet exhibited much lower spike rates than flies exhibiting the high glucose diet, even though they had a much shorter lifespan and therefore, presumably, experienced much higher stress levels (Figure 1H, Figure S1). These observations underline that stress is certainly not the primary factor here.

(3) ‘The finding that refeeding starved flies with a standard diet had no effect on IPC activity but a strong effect on the locomotor activity of starved flies contradicts the statement that locomotor activity is affected by the same dietary manipulations that affect IPC activity.’

We have revised the respective section of the results and discussion accordingly and are more careful and clearer in our interpretation of this behavioral dataset: ‘These results show that the locomotor activity was affected by the same dietary manipulations that had strong effects on IPC activity. However, IPC activity changes alone cannot explain the modulation of starvation-induced hyperactivity. On the one hand, high-glucose diets which drove the highest activity in IPCs were not sufficient to reduce locomotor activity back to baseline levels. On the other hand, refeeding flies with SD did not revert the effects of starvation on IPC activity (Figure 1H), but it was sufficient to reduce the locomotor activity below baseline levels (Figure 2B). This suggests that the modulation of starvation-induced hyperactivity is achieved by multiple modulatory systems acting in parallel.’

(4) ‘The compelling finding that starvation induces IPC firing would benefit from determining the time course of the effect.’

We followed the referee’s excellent suggestion and determined the time course of the starvation effect in three timesteps, similar to the experiments we did for refeeding (Figure 1G). In addition, we now also quantify the number of active IPCs (i.e., IPCs that fired at least one action potential during our five-minute analysis window), which further illustrates the dynamics of the starvation and refeeding effects. We find that the starvation effect is graded, and that IPC activity decreases with increasing starvation duration.

(5) ‘The finding that IPCs are not active in fed animals older than 1 week is surprising and should be further validated.’

To address the referee’s comment, we have added 14 new IPC recordings from flies in the 6–26-day range, such that we now have recordings from 9-14 IPCs for each age range (Figure S2B). They confirmed our previous analysis and strengthened the finding that IPC activity dramatically decreases after 8 days (on our standard diet). The total number of IPCs in this supplementary dataset was thus increased from 34 to 48.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors): 

(1) Do IPCs respond to glucose specifically after ingestion or generally to any other nutritive sugars? To tackle this question the IPC responses in starved flies can be recorded after refeeding flies with other nutritive sugars (fructose, sucrose). 

To address this important question, we have performed additional experiments in which we refed starved flies with fructose, as a nutritive sugar, and arabinose, as a non-nutritive sugar. As expected, IPCs responded to fructose but not arabinose and hence nutritive sugars in general. We describe and discuss these key results in the new version of our manuscript.

(2) In Figure 2, the x and y axes are not annotated on all subfigures, which might help improve clarity. 

We have annotated the subfigures as requested.

(3) In the discussion on page 9 ("...we observed an increase in IPC activity after the ingestion of glucose (Figure 2B)."), the authors refer to Figure 2B instead of 3C.

We have fixed this oversight.

Reviewer #2 (Recommendations For The Authors): 

Introduction 

I think it could be helpful for the reader if you would briefly state the number of IPCs and whether you are targeting all of them with Dilp2-Gal4. 

We included the numbers according to the suggestion. 14 IPCs are labeled in the driver line, and this is the number of IPCs commonly assumed to be present in the PI.

Figures 

In some Figures (for example 1D & E) the authors state the number of IPCs recorded (N) but not the number of animals used (n). This should be stated as the data from within an animal are dependent and might give insights about IPC heterogeneity. 

We have compiled tables for the supplementary material (Tables S5 & S6) in which we state the number of IPCs and DH44^PINs recorded and the number of different flies for each figure panel. We have recorded an average of 1.4 IPCs per fly (217 IPCs from 160 flies). We therefore expect the bias introduced by individual flies to be rather small. However, in our parallel study, we specifically investigate the heterogeneity of IPCs by maximizing the number of IPCs recorded per fly (Held M, et al. ‘Aminergic and peptidergic modulation of Insulin-Producing Cells in Drosophila’. eLife. 2024;13. doi:10.7554/ELIFE.99548.1). In the case of DH44PINs, we recorded 24 neurons in 21 flies – 1.1 neurons per fly.

- Figure 3D: There is some white visible among the cell bodies in the overlay. I assume this comes from projecting across layers rather than indicating DH44 - IPC overlap? It would help to explicitly state that. 

We have added a statement to the results section, in which we explain that most of the white is due to overlap in the z-projection rather than overlap in the driver lines. However, there are few cases (typically one to two cells per brain), in which neurons labeled by the DH44 line also stain positive for Dilp2, indicating they express both neuropeptides. We have added this information to the manuscript:  

Results: ‘DH44^PINs are anatomically similar to IPCs, and their cell bodies are located directly adjacent to those of IPCs in the PI, making them an ideal positive control for our experiments (Figure 3D). A small subset of DH44^PINs also expresses Dilp2(75), and our immunostainings confirmed colocalization of Dilp2 and DH44 in a single neuron (Figure 3D, white arrow).’

In figure caption: ‘UAS-myr-GFP was expressed under a DH44-GAL4 driver to label DH44 neurons. GFP was enhanced with anti-GFP (green), brain neuropils were stained with anti-nc82 (cyan), and IPCs were labelled using a Dilp2 antibody (magenta). White arrow indicates Dilp2 and DH44-GAL4 positive neuron. The other white regions in the image result from an overlap in z-projections between the two channels, rather than from antibody colocalization.’

- Figure 4I: One might get the impression that the fast onset peak of activity precedes the stimulation onset, using a thinner line width might help avoid that. 

This effect is due to a combination of using relatively heavy lines for clear visibility of the data and a gentle smoothing step (a 2s median filter, which corresponds to less than 1% of the 300s stimulation window) in our analysis of the behavioral data. However, inspection of the raw data clearly shows increases in velocity after the onset of the optogenetic activation. We clarified this in the figure caption: ‘Average FV across all DH44N activation trials based on two independent replications of the experiment in I. Note that the peak in average FV lies within the first frame of the stimulation window.’

- S3 panel letters do not match references in the text.

We fixed this oversight.

Formatting 

- Page 10: The paragraphs on the bottom of the page got switched around.

This has been fixed.

- Page 14: The first paragraph after the header "Free-walking assay" seems to be coming from elsewhere. 

We apologize for this slightly embarrassing mistake. We used our related bioRxiv preprint (Held et al.) as a template for formatting this paper, and accidentally left this part of the methods section in the manuscript. We have fixed this error in our resubmission.

Reviewer #3 (Recommendations For The Authors): 

Major suggestions: 

(1) The data show convincingly that IPC activity is decreased by starvation during the first week of adult life (Figures 1C and D). However, the conclusion that IPC activity is controlled by the nutritional state requires additional care. First, refeeding starved adult animals with a normal diet does not bring back normal IPC firing rates (Figure 1H). Therefore, IPC activity does not strictly follow changes in nutritional state, but IPCs are silenced by starvation. Second, from the second week of adult life on, IPCs are silent anyway, and thus unlikely responsive to changes in the nutritional state anymore (which might be different on a different standard diet?) The only effect of feeding on IPC activity is observed upon feeding starved, young animals with high glucose for 12-24 hrs (Figure 1G). However, it is not clear whether increased IPC firing is caused by the effects of high glucose on the nutritional state in a normal range, or because of diet-induced stress (the diet also severely shortens lifespan, Figure 1S). Does high glucose also increase IPC firing rate in young, fed animals? These would have strongly increased glucose concentrations but not suffer the stress of not getting any other nutrients. Such experiments would be required to make the statement that glucose feeding increases IPC firing rate. 

We have performed several experiments to address this criticism. First, we performed a time course analysis of the starvation effect. We show that the IPC activity reduction is graded, and that IPC activity declines already after two hours of starvation, a timepoint at which stress levels should still be relatively small (Figure 1G). Second, we refed flies with high glucose concentrations added to the standard diet (Figure 1H). This minimized any potential stress responses due to a lack in nutrients. Third, we now show that IPCs specifically respond to nutritive (glucose and fructose), but not to non-nutritive sugars (arabinose, Figure 1H). We believe that these data sets, in addition to the graded refeeding effect, make a strong case for the nutritional state dependent modulation of IPCs. 

(2) The testing of locomotor activity is well done, nicely recapitulates starvation-induced increases in locomotion, and adds interesting novel findings on refeeding with high glucose versus high protein diet. However, the statement that locomotor activity was affected by the same dietary manipulations that had strong effects on IPC activity does not reflect the data presented. Refeeding starved flies with a standard diet had no effect on IPC activity (Figure 1H) but a strong effect on locomotor activity of starved flies (a strong reduction, even stronger than high glucose diet, Figure 2B). 

We have revised the respective section of the results and discussion accordingly and are more careful and clearer in our interpretation of this behavioral dataset: ‘These results show that the locomotor activity was affected by the same dietary manipulations that had strong effects on IPC activity. However, IPC activity changes alone cannot explain the modulation of starvationinduced hyperactivity. On the one hand, high-glucose diets which drove the highest activity in IPCs were not sufficient to reduce locomotor activity back to baseline levels. On the other hand, refeeding flies with SD did not revert the effects of starvation on IPC activity (Figure 1H), but it was sufficient to reduce the locomotor activity below baseline levels (Figure 2B). This suggests that the modulation of starvation-induced hyperactivity is achieved by multiple modulatory systems acting in parallel.’

Related to points 1 and 2, a key statement that the results establish that IPC activity is controlled by the nutritional state requires care. What the data convincingly show is that IPC activity is near zero upon starvation. 

As described above, we have added several extensive data sets (fructose feeding, arabinose feeding, trehalose perfusion, starvation time course) to show that we indeed observe a nutritional state dependent modulation of IPCs and describe these new results in the results and discussion.

(3) The time course of nutritional state-dependent changes of IPC activity is claimed to be slow, several hours to days. Unless I have missed a figure, the underlying data are not presented (only for high glucose diet). It would be great if this could also be shown for a standard diet with higher glucose concentrations than the one used so that it rescues starvation-induced IPC silencing without shortening lifespan (if this is feasible?). The data showing starvation-induced IPC silencing are convincing, but, unless I have missed it, the time course has not been determined. It would be very nice to actually show this. Have different starvation times been tested in relation to IPC firing rate, and if yes, with what time resolution? Does IPC activity change already after 0.5 or 1 or a few hours of starvation? If starvation can silence IPCs faster than assumed, the nearzero IPC activity in animals older than a week could very well be caused by longer time intervals between meals. 

We have performed experiments to address both important points raised by the referee here. 1) We have added high glucose concentrations to our standard diet, and show that it has the same effect – a significant increase in IPC activity – as the high glucose diet (Figure 1H). 2) We have analyzed the time course of IPC activity reduction in response to starvation (Figure 1G). Indeed, we find that a few hours of starvation start reducing IPC activity. We discuss the possibility that reduced IPC activity in older flies could be due to reduced food intake: ‘One of our experiments demonstrated that IPC activity was heavily diminished in flies older than 10 days (Figure S2B). A possible explanation could be that flies feed less as they age. However, this only holds true for flies older than 14 days86. Therefore, reduced IPC activity in 10-11 day old flies is unlikely to result from reduced food intake and likely involves inhibition of insulin signaling.’

(4) The data on the proposed incretin effect are of high importance in potentially highlighting a highly conserved link between glucose ingestion and insulin release. An important control would be to test different sugars, such as trehalose, an important blood sugar of flies. If glucose is converted into trehalose and this is what IPCs sense, then perfusion of glucose has no effect. The fact fantastic experiments show that the DH44 neurons are sensitive to glucose perfusion does rule out that IPCs sense a different sugar. This would be very different from the incretin effect that requires additional hormones. In addition, as mentioned above, controls are required to show that high glucose affects IPCs as a nutrient and not as a stressor (see point 1), for example refeeding with a standard diet that contains a higher glucose concentration but does not reduce lifespan. Another great control to solidify the exciting claim on the incretin effect would be to knock out candidate Drosophila incretin hormones and test whether a high glucose diet stops increasing the IPC firing rate (although simpler controls might also do the job). 

We have performed the two key experiments suggested by the referee. 1) We perfused trehalose as the primary blood sugar of flies and showed that IPCs do not respond to trehalose perfusion (Figure 3B & C). This further strengthens the finding that IPC activity in flies shows an incretin-like effect. 2) We have added high concentrations of glucose to our standard diet to provide flies with a full diet that contains high glucose concentrations. IPC activity in these flies was indistinguishable from the activity in flies which consumed pure glucose diets. In contrast, IPC activity in flies kept on a high protein diet, which dramatically reduced lifespan, was very low. These results clearly show that higher IPC activity is not due to increased stress levels, but a function of nutritive sugar ingestion. We further validated this hypothesis by refeeding flies with fructose as a nutritive sugar, which increased IPC activity, and arabinose as a non-nutritive sugar, which did not affect IPC activity (Figure 1H).

Another point that might be relevant to this discussion is that IPC activity is almost entirely shut down during flight in Drosophila (which we showed in Liessem et al. 2023, Current Biology 33 (3), 449-463. e5). Several ‘stress hormones’ are released during flight, including octopamine. The fact that IPC activity is low in flying flies, starved flies, and flies kept on a pure protein diet (which all experience high stress levels), to us, very clearly suggests that stress is not the predominant factor here. We would also like to point out that, while the lifespan was reduced in flies kept on pure glucose diets, survival rates were at 100% until day 14, and we carried out our experiments on day 2 after starvation. Hence, these flies might not (yet) experience particularly high stress levels.

(5) The discussion relates the absence of IPC firing in animals older than 1 week to aging. However, given that the flies fed on a normal diet show the typical lifespan for Drosophila, a 10-dayold fly is still in its youth. Maybe flies at 10 days eat simply less and thus IPC spiking goes down as in starved flies, especially because the standard diet used contains low glucose. Do IPCs also become silent after a week if the animals are fed with a standard diet that contains a higher glucose concentration? Without additional controls, this part of the discussion is pretty speculative and should be revised. 

We agree with the reviewer, that it is not clear whether reduced IPC activity is a direct result of physiological changes that occur with aging, or an indirect effect of reduced food intake, which occur during aging. In both cases, in our view, it would be an age-related effect. Since this is a minor point of our manuscript, we decided not to perform additional experiments, other than significantly increasing the sample size for the aging data set already presented to shore up our findings (Figure S2B). We have, however, revisited the discussion of this point according to the referee’s suggestion: ‘One of our experiments demonstrated that IPC activity was heavily diminished in flies older than 10 days (Figure S2B). A possible explanation could be that flies feed less as they age. However, this only holds true for flies older than 14 days(85). Therefore, reduced IPC activity in 10-11 day old flies is unlikely to result from reduced food intake and likely involves inhibition of insulin signaling.’

Other suggestions: 

(6) For the mixed effects of octopamine and tyramine on larval locomotion that are referred to, it might be interesting to also look at Schützler et al 2019, PNAS because it shows that starvation activates TBH so that the octopamine to tyramine ratio is increased. 

We refer to Schützler et al. in the following paragraph of our discussion: ‘This intermittent locomotor arrest has been previously described in adult flies and is thought to be mediated by ventral unpaired median OANs, which have been suggested to suppress long-distance foraging behavior(69). Since these are not the only neurons we activate in the TDC2 line, we speculate that the stopping phenotype could also result from concerted effects of octopamine and tyramine modulating muscle contractions(65-67) and motor neuron excitability(68), as previously described in Drosophila larvae, or from OANs interfering with pattern generating networks in the ventral nerve cord (VNC) during longer activation(69).’  

(7) The reference list requires care. For example, reference 43 is identical to 67, reference 66 gives no information on incretin-like hormones in Drosophila as stated in the text 

We carefully double-checked our reference list and corrected the mistakes mentioned.

eLife Assessment

The manuscript presents valuable findings of bone remodeling under chronic unpredictable mild stress (CUMS). This is an interesting work on mental stress on bone health and osteoporosis, and the authors offer solid evidence of decreased bone mass mediated by miR-335-3p/Fos signaling in osteoclasts that are involved in the induction of bone loss caused by CUMS. This revised version provided new data that improved the quality of the manuscript and addressed the reviewers' concerns.

Reviewer #1 (Public review):

I have reviewed the manuscript "Psychological stress disturbs bone metabolism via miR-335-3p/Fos signaling in osteoclast" with interest. The described findings are relevant and useful for daily practice in periodontology. The paper is concise, professionally written, and easy to read. In this study, Jiayao et al. revealed the role of miR-335-3p in psychological stress-induced osteoporosis. CUMS mice were constructed to observe the femur phenotype, osteoclasts were identified as the main research object, and miRNA-seq was used to find the key miRNAs linking the brain and peripheral tissues. This study showed that miR-335-3p expression was simultaneously reduced in murine NAC, serum, and bone under psychological stress. The miR-335-3p/Fos/NFATC1 signaling pathway was validated in osteoclasts to reveal the potential mechanism of enhanced osteoclast activity under psychological stress. This study, from a new perspective of miRNAs, indicates a possible cause of disturbed bone metabolism due to psychological stress and may suggest a new approach to treating osteoporosis.

Reviewer #2 (Public review):

Zhang et al. established chronic unpredictable mild stress (CUMS) mouse model, which displayed osteoporosis phenotype, suggesting a potential correlation between psychological stress and bone metabolism. They found that miRNA candidate miR-335-3p is downregulated in the long bone of CUMS mice through microRNA sequencing experiments and qRT-PCR. They further demonstrated that miR-335-3p attenuates osteoclast activity via inhibiting Fos signaling, which can induce NFATC1 expression and regulate osteoclast activity.

My concerns have been addressed. And the quality of the manuscript is improved significantly.

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

I have reviewed, with interest, the manuscript "Psychological stress disturbs bone metabolism via miR-335-3p/Fos signaling in osteoclast". The described findings are relevant and useful for daily practice in periodontology. The paper is concise, professionally written, and easy to read. In this study, Jiayao et al. revealed the role of miR-335-3p in psychological stress-induced osteoporosis. CUMS mice were constructed to observe the femur phenotype, osteoclasts were identified as the primary research object, and miRNA-seq was used to find the key miRNAs linking the brain and peripheral tissues. This study showed that the expression of miR-335-3p was simultaneously reduced in mice's NAC, serum, and bone under psychological stress. The miR-335-3p/Fos/NFATC1 signaling pathway was validated in osteoclasts to reveal the potential mechanism of enhanced osteoclast activity under psychological stress. From a new perspective of miRNAs, this study indicates a possible cause of disturbed bone metabolism due to psychological stress and may suggest a new approach to treating osteoporosis.

We thank this reviewer for the instructive suggestions and encouragement.

Reviewer #2 (Public Review):

Zhang et al. established chronic unpredictable mild stress (CUMS) mouse model, which displayed osteoporosis phenotype, suggesting a potential correlation between psychological stress and bone metabolism. They found that miRNA candidate miR-335-3p is downregulated in the long bone of CUMS mice through microRNA sequencing and qRT-PCR experiments. They further demonstrated that miR-335-3p attenuates osteoclast activity via inhibiting Fos signaling, which can induce NFATC1 expression and regulate osteoclast activity.

Strengths:

The authors established CUMS mouse model and confirmed the osteoporosis phenotype through careful characterization of bone and analysis of osteoclast activity. They performed microRNA sequencing to identify the miRNA candidate regulating the bone loss in the CUMS mouse model. They also validated the expression of miR-335-3p and interfered with the function of miR-335-3p through an in vitro assay. Overall, the findings from this study provide important hints for the correlation between psychological stress and bone metabolism.

We thank this reviewer for the comprehensive summary and positive comment on our work.

Weakness:

The data provided by the authors are preliminary, especially the mechanistic insight, which needs to be enhanced. The authors have shown that miR-335-3p expression was altered in the CUMS mouse model and the change of its expression regulated osteoclast activity. The validation should be conducted in vivo, and the mechanism behind this should be investigated further.

We thank the reviewer’s important insight on the need for further in vivo validation of the role of miR-335-3p. Therefore, we designed and produced Antagomir-335-3p (antagonist) and Agomir-335-3p (agonist). Then, we injected them into the body through the tail vein for about 2 months and observed the bone phenotype in each group of mice. The results suggested that the decrease of miR-335-3p in vivo could lead to bone loss, which was consistent with our in vitro validation results (Figure 5H-I).

Reviewing Editor:

Method

(1) Bone histomorphometric analysis following ASBMR's guidelines Bone histomorphometric analysis of bone formation and bone resorption: The authors should follow ASBMR's guidelines for bone histomorphometry (PMCID: PMC3672237 and PMID: 3455637) to perform standard analyses of histomorphometry, rather than selected areas. They should also clearly describe a software used and define the areas analyzed.

We carefully re-analyzed bone histomorphometry according to ASBMR guidelines and combine this with our own understanding. At the same time, we improved the description of micro-CT and histological analysis in the method. If there is still any lack of standardization, we would be grateful for any constructive suggestions to improve this.

(2) Osteoclast cultures require nuclear staining to demonstrate multinucleated Trap positive cells.

We used the RAW264.7, a mouse macrophage-like cell line, for in vitro culture and induced its differentiation towards osteoclasts. Successfully induced osteoclasts showed enlarged cytoplasm and multinucleated fusion. Tartrate-resistant acid phosphatase (Trap) is the signature enzyme of osteoclasts. It can bind to the chromogen to exhibit a mauve color, based on the principle of azo-coupled immunohistochemistry. At the same time, small and rounded nuclei fused show a lighter color (author response image 1, yellow arrows). We attempted to stain the nuclei with hematoxylin based on this. However, it was unable to further distinguish the contours of the nuclei clearly due to the similar color to the Trap positive signals. Besides, many other scholars have assessed osteoclast activity in vitro experiments based solely on the results of Trap staining (area and number) (Cheng et al., 2022; Li et al., 2019; Ma et al., 2021; Zhong et al., 2023). Nevertheless, in the immunofluorescence staining of osteoclasts, the nuclei were labeled using a Hochest antibody to reflect the multinucleated fusion of osteoclasts (Figure 5G).  

(3) Osteoclast pit assays should be carried out to necessarily demonstrate the change of osteoclast resorption ability caused by miR-335-3p.

We added osteoclast pit assays to validate the role of miR-335-3p on osteoclast resorptive capacity (Figure 5D-E).

(4) Serum ELISA assay should be done to examine the global change of bone remodeling in the CUMS mice to assess bone formation and bone resorption that will support their claim.

We performed additional tests on serum concentrations of R-hydroxy glutamic acid protein (BGP), TRAP, Cathepsin K (CTSK), parathyroid hormone (PTH), calcium (CA), phosphate (P) in control and CUMS mice, which could better reflect the global change of bone remodeling in the CUMS mice (Figure 3— figure supplement 1).

(5) miR-RNA-seq: A labeled volcano plot should be used to replace the present one to show significant changes in differential gene expression.

We appreciate this great suggestion. We replaced the volcano plot that showed significant changes in differential gene expression (Figure 4B). We also uploaded the raw data to the GEO database (GSE253504), making the results clearer and more accessible.

Discussion

The authors should discuss previous works on the influences of hormones from the brain on chronic stress-induced bone loss and an association of these influences with their findings.

The discussion on the relationship between the bone metabolism regulation of both hormones and miR-335-3p in psychological stress was added in the second and fifth paragraphs of the discussion. To conclude, on the one hand, brain-derived and blood-transported miR-335-3p regulate bone metabolism synergistically. On the other hand, it exerted a more direct influence on bone under psychological stress.

Language

The language of the MS should be improved.

The manuscript has been carefully edited by a professional proofreader.

Reviewer #1 (Recommendations For The Authors):

(1) Figure 1F: The exact meaning of the Waveform Graph shown at left needs to be clarified for the not-so-experienced reader.

We added the more detailed meaning of the Waveform Graph in figure legends (Figure legend 1F).

(2) Is the concomitant increase in osteogenic and osteoblastic activity in this study consistent with that seen in similar disease studies? This could be added to the discussion.

In the fifth paragraph of the discussion section, we present the alterations of osteogenic and osteoblastic activity observed in other studies that are similar to ours. We also had a detailed discussion based on these observations.

(3) Figure 6A: Please highlight the key information to visualize the potential linkage among miR-335-3p, Fos, and osteoclast.

We highlighted the crucial linkage among miR-335-3p, Fos, and osteoclast with red arrows (Figure 6A)

4) Figure 6E: The specific area of the selected comparison needs to be clarified. Please add white dotted lines and lettering T (trabecular bone) and GP (growth plate) for the not-so-experienced reader. This will provide some orientation.

We used white dotted lines as well as letters to label the tissue in immunofluorescence staining images (Figure 6E).

(5) Line 350: "NAC derived and blood-trans, Ported miR-335-3p". There is a grammatical error. Please conduct general proofreading of the text and writing style.

Thank you for pointing this out. We have corrected this grammatical error, and we also checked the full text to correct similar errors.

Reviewer #2 (Recommendations For The Authors):

(1) miR-335-3p was downregulated in the femur in the CUMS mice. The possible mechanism for this outcome should be further discussed. In Figure 4B, the Volcano plot showed that only a few miRNA were differentially expressed between the control and CUMS mice. How do the authors explain this?

The chronic unpredictable mild stress (CUMS) model was constructed using normal mice. As the name of the model suggests, the stimulus is mild and does not cause developmental damage or teratogenic effects in mice. Conversely, CUMS has the potential to result in the chronic pathological conditions. Besides, in miRNA sequencing results from other tissues with similar models to ours, the number of differential miRNAs is also around a few dozen (Ma et al., 2019).

(2) The authors have demonstrated that miR-335-3p inhibits osteoclast differentiation based on an in vitro assay in Figure 5; however, an in vivo experiment is required to provide more solid evidence.

We strongly agree that in vivo experimental validation would bring more convincing results to this study. Therefore, we designed and produced Antagomir-335-3p (antagonist) and Agomir-335-3p (agonist), which were injected into mice via the tail vein every five days. Samples were collected at one and two months following the injection. We found that sustained two-month injections of antagomir could significantly lead to bone loss in mice (Figure 5H-I), which is consistent with our in vitro validation results.

However, the Agomir-miR-335-3p group did not exhibit a notable enhancement of bone mass. This may be attributed to the fact that the 11-week-old normal mice selected for this study were in their prime and did not have strong osteoclastic activity in vivo. Therefore, the osteoclastic inhibition of Agomir-335-3p could not be demonstrated.

In addition, no significant difference was seen one month after the injection. The main reason may be that the time is too short. On the one hand, the drug we injected was RNA preparation. They lacked stability resulting in poor delivery efficiency, which took some time to take effect. On the other hand, bone remodeling is also a time-consuming process.

(3) FOS and NFATC1 should be expressed in the nuclei of the cells, therefore, the quality of the images needs to be improved.

NFATC1 is a T-cell-activating nuclear factor that is activated in the nucleus to regulate the transcription of a variety of osteoclast-related genes, including ACP5, MMP9, etc. FOS could bind and interact with NFATC1, resulting in nuclear translocation and transcription activated. This could promote the differentiation and maturation of osteoclasts. They are both synthesized and processed in the cytoplasm and eventually enter the nucleus to perform their functions. Therefore, they are expressed in both the nucleus and the cytoplasm (Deng et al., 2022; Hounoki et al., 2008; Li et al., 2022).

In Figure 5G, we labeled cell nuclei with HOCHEST antibody with blue fluorescence, and more co-localized signals of nuclei (blue), FOS (red), and NFATC1 (green) were seen in the Inhibitor-miR-335-3p group, whereas the opposite result was observed in the Mimic-miR-335-3p group. These results indicated that inhibited miR-335-3p could promote osteoclast differentiation in vitro.

(4) The expression of FOS was elevated in CUMS group in Figure 6E; however, its mRNA level was unchanged, as shown in Figure 6 supplement; what is the explanation for this? How do the authors claim FOS is the downstream target if its mRNA expression is not impacted by CUMS?

The results demonstrated that miR-335-3p targeted binding to the mRNA of Fos did not result in mRNA degradation. Instead, this binding interferes with the protein translation process, which ultimately leads to the reduction of FOS protein.

(5) What would be the bone phenotype if a FOS inhibitor was injected into the control and CUMS mice? It is important to examine FOS function through an in vivo context.

The regulatory role of FOS for osteoclasts has been validated in numerous articles, both in vivo and in vitro(Aikawa et al., 2008; Cao et al., 2023; Cheng et al., 2022). For example, Aikawa et al. designed a small-molecule inhibitor of c-Fos/activator protein-1 (AP-1) using three-dimensional (3D) pharmacophore modeling, which helped verify the effect of FOS on osteoclasts in vivo(Aikawa et al., 2008).

We also strongly agree that in vivo injection of inhibitors of FOS, especially in CUMS mice, could further substantiate the role of miR-335-3p in osteoclasts under psychological stress. However, the study was constrained by the unavailability of commercially viable, efficacious small molecule inhibitors of FOS. In the future, we plan to design more precise therapeutic targets for psychological stress induced osteoporosis based on existing research ideas.

Reference

Aikawa, Y., Morimoto, K., Yamamoto, T., Chaki, H., Hashiramoto, A., Narita, H., Hirono, S., & Shiozawa, S. (2008). Treatment of arthritis with a selective inhibitor of c-Fos/activator protein-1. Nature Biotechnology, 26(7), 817-823. https://doi.org/10.1038/nbt1412

Cao, Z., Niu, X. B., Wang, M. H., Yu, S. W., Wang, M. K., Mu, S. L., Liu, C., & Wang, Y. X. (2023, Nov). Anemoside B4 attenuates RANKL-induced osteoclastogenesis by upregulating Nrf2 and dampens ovariectomy-induced bone loss [Article]. Biomedicine & Pharmacotherapy, 167, 12, Article 115454. https://doi.org/10.1016/j.biopha.2023.115454

Cheng, X., Yin, C., Deng, Y., & Li, Z. (2022). Exogenous adenosine activates A2A adenosine receptor to inhibit RANKL-induced osteoclastogenesis via AP-1 pathway to facilitate bone repair. Molecular Biology Reports, 49(3), 2003-2014. https://doi.org/10.1007/s11033-021-07017-1

Deng, W., Ding, Z., Wang, Y., Zou, B., Zheng, J., Tan, Y., Yang, Q., Ke, M., Chen, Y., Wang, S., & Li, X. (2022). Dendrobine attenuates osteoclast differentiation through modulating ROS/NFATc1/ MMP9 pathway and prevents inflammatory bone destruction. Phytomedicine : International Journal of Phytotherapy and Phytopharmacology, 96, 153838. https://doi.org/10.1016/j.phymed.2021.153838

Hounoki, H., Sugiyama, E., Mohamed, S. G.-K., Shinoda, K., Taki, H., Abdel-Aziz, H. O., Maruyama, M., Kobayashi, M., & Miyahara, T. (2008). Activation of peroxisome proliferator-activated receptor gamma inhibits TNF-alpha-mediated osteoclast differentiation in human peripheral monocytes in part via suppression of monocyte chemoattractant protein-1 expression. Bone, 42(4), 765-774. https://doi.org/10.1016/j.bone.2007.11.016

Li, Y., Yang, C., Jia, K., Wang, J., Wang, J., Ming, R., Xu, T., Su, X., Jing, Y., Miao, Y., Liu, C., & Lin, N. (2022). Fengshi Qutong capsule ameliorates bone destruction of experimental rheumatoid arthritis by inhibiting osteoclastogenesis. Journal of Ethnopharmacology, 282, 114602. https://doi.org/10.1016/j.jep.2021.114602

Li, Z., Huang, J., Wang, F., Li, W., Wu, X., Zhao, C., Zhao, J., Wei, H., Wu, Z., Qian, M., Sun, P., He, L., Jin, Y., Tang, J., Qiu, W., Siwko, S., Liu, M., Luo, J., & Xiao, J. (2019). Dual Targeting of Bile Acid Receptor-1 (TGR5) and Farnesoid X Receptor (FXR) Prevents Estrogen-Dependent Bone Loss in Mice. Journal of Bone and Mineral Research : the Official Journal of the American Society For Bone and Mineral Research, 34(4), 765-776. https://doi.org/10.1002/jbmr.3652

Ma, K., Zhang, H., Wei, G., Dong, Z., Zhao, H., Han, X., Song, X., Zhang, H., Zong, X., Baloch, Z., & Wang, S. (2019). Identification of key genes, pathways, and miRNA/mRNA regulatory networks of CUMS-induced depression in nucleus accumbens by integrated bioinformatics analysis. Neuropsychiatric Disease and Treatment, 15, 685-700. https://doi.org/10.2147/NDT.S200264

Ma, Q., Liang, M., Wu, Y., Luo, F., Ma, Z., Dong, S., Xu, J., & Dou, C. (2021). Osteoclast-derived apoptotic bodies couple bone resorption and formation in bone remodeling. Bone Research, 9(1), 5. https://doi.org/10.1038/s41413-020-00121-1

Zhong, L., Lu, J., Fang, J., Yao, L., Yu, W., Gui, T., Duffy, M., Holdreith, N., Bautista, C. A., Huang, X., Bandyopadhyay, S., Tan, K., Chen, C., Choi, Y., Jiang, J. X., Yang, S., Tong, W., Dyment, N., & Qin, L. (2023). Csf1 from marrow adipogenic precursors is required for osteoclast formation and hematopoiesis in bone. eLife, 12. https://doi.org/10.7554/eLife.82112

eLife Assessment

This study presents an advance in efforts to use histone post-translational modification (PTM) data to model gene expression and predict epigenetic editing activity. Such models are broadly useful to the research community, especially ones that can model epigenetic editing activity, which is novel; additionally, the authors have nicely integrated datasets across cell types into their model. The work is mostly solid, but it would be strengthened by performing rigorous comparisons to existing methods that predict gene expression from PTM data and from additional model validation beyond dCas9-p300 based perturbations.

Reviewer #1 (Public review):

Batra, Cabrera and Spence et al. present a model which integrates histone posttranslational modification (PTM) data across cell models to predict gene expression with the goal of using this model to better understand epigenetic editing. This gene expression prediction model approach is useful if a) it predicts gene expression in specific cell lines b) it predicts expression values rather than a rank or bin, c) if it helps us to better understand the biology of gene expression or d) it helps us to understand epigenome editing activity. Problematically for point a) and b) it is easier to directly measure gene expression than to measure multiple PTMs and so the real usefulness of this approach mostly relates to c) and d).

Other approaches have been published that use histone PTM to predict expression (e.g. PMID 27587684, 36588793). Is this model better in some way? No comparisons are made although a claim is made that direct comparisons are difficult. I appreciate that the authors have not used the histone PTM data to predict gene expression levels of an "average cell" but rather that they are predicting expression within specific cell types or for unseen cell types. Approaches that predict expression levels are much more useful whereas some previous approaches have only predicted expressed or not expressed or a rank order or bin-based ranking. The paper does not seem to have substantial novel insights into understanding the biology of gene expression.

The approach of using this model to predict epigenetic editor activity on transcription is interesting and to my knowledge novel although only examined in the context of a p300 editor. As the author point out the interpretation of the epigenetic editing data is convoluted by things like sgRNA activity scoring and to fully understand the results likely would require histone PTM profiling and maybe dCas9 ChIP-seq for each sgRNA which would be a substantial amount of work.

Furthermore from the model evaluation of H3K9me3 is seems the model is performing modestly for other forms of epigenetic or transcriptional editing- e.g. we know for the best studied transcriptional editor which is CRISPRi (dCas9-KRAB) that recruitment to a locus is associated with robust gene repression across the genome and is associated with H3K9me3 deposition by recruitment of KAP1/HP1/SETDB1 (PMID: 35688146, 31980609, 27980086, 26501517).

One concern overall with this approach is that dCas9-p300 has been observed to induce sgRNA independent off target H3K27Ac (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8349887/ see Figure S5D) which could convolute interpretation of this type of experiment for the model.

Reviewer #2 (Public review):

Summary:

The authors build a gene expression model based on histone post-translational modifications, and find that H3K27ac is correlated with gene expression. They proceed to perturb H3K27ac at 13 gene promoters in two cell types, and measure gene expression changes to test their model.

Strengths:

The combination of multiple methods to model expression, along with utilizing 6 histone datasets in 13 cell types allowed the authors to build a model that correlates between 0.7-0.79 with gene expression. They use dCas9-p300 fusions to perturb H3K27ac and monitor gene expression to test their model. Ranked correlations of the HEK293 data showed some support for the predictions after perturbation of H3K27ac.

Weaknesses:

The perturbation of 5 genes in K562 with perturb-seq data shows a modest correlation of ~0.5 and isn't included in the main figures. The authors are then left to speculate reasons why the outcome of epigenome editing doesn't fit their predictions, which highlights the limited value in the current version of this method.

As mentioned before, testing genes that were not expressed being most activated by dCas9-p300 weaken the correlations vs. looking at a broad range of different gene expression as the original model was trained on.<br /> If the authors want this method to be used to predict outcomes of epigenome editing, expanding to dCas9-KRAB and other CRISPRa methods (SAM and VPR) would be useful. Those datasets are published and could be analyzed for this manuscript.<br /> The authors don't compare their method to other prediction methods.

Author response:

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

Reviewer #1 (Public Review):

Batra, Cabrera, Spence et al. present a model which integrates histone posttranslational modification (PTM) data across cell models to predict gene expression with the goal of using this model to better understand epigenetic editing. This gene expression prediction model approach is useful if a) it predicts gene expression in specific cell lines b) it predicts expression values rather than a rank or bin, c) it helps us to better understand the biology of gene expression, or d) it helps us to understand epigenome editing activity. Problematically for points a) and b) it is easier to directly measure gene expression than to measure multiple PTMs and so the real usefulness of this approach mostly relates to c) and d).

We thank the reviewer for their comment and we agree that directly measuring gene expression (e.g., by performing RNA-seq) is easier than performing multiple PTMs in a new cell line. We designed our approach keeping in mind that the primary use case is to understand how epigenome editing would affect gene expression.

Other approaches have been published that use histone PTM to predict expression (e.g. 27587684, 36588793). Is this model better in some way? No comparisons are made. The paper does not seem to have substantial novel insights into understanding the biology of gene expression. The approach of using this model to predict epigenetic editor activity on transcription is interesting and to my knowledge novel but I doubt given the variability of the predictions (Figures 6 and S7&8) that many people will be interested in using this in a practical sense. As the authors point out, the interpretation of the epigenetic editing data is convoluted by things like sgRNA activity scoring and to fully understand the results likely would require histone PTM profiling and maybe dCas9 ChIP-seq for each sgRNA which would be a substantial amount of work.

We thank the reviewer for this insightful comment. We have included citations for a series of papers (PMIDs: 27587684, 30147283, 36588793) that performed gene expression prediction using histone PTM data. However, each of these methods perform classification of gene expression as opposed to predicting the actual gene expression value via regression. Additionally, the referenced studies all work with Roadmap Epigenomics read depth data as opposed to p-values obtained from the ENCODE pipelines, making it difficult to make direct comparisons.

We outline in the Discussion section that by creating a comprehensive dataset of epigenome editing outcomes, which include quantification of histone PTMs before and after in situ perturbations, will improve our understanding of the effects of dCas9-p300 on gene expression and assist in the design of gRNAs for achieving fine-tuned control over gene expression levels. 

Furthermore from the model evaluation of H3K9me3 it seems the model is not performing well for epigenetic or transcriptional editing- e.g. we know for the best studied transcriptional editor which is CRISPRi (dCas9-KRAB) that recruitment to a locus is associated with robust gene repression across the genome and is associated with H3K9me3 deposition by recruitment of KAP1/HP1/SETDB1 (PMID: 35688146, 31980609, 27980086, 26501517). However, it seems from Figures 2&4 that the model wouldn't be able to evaluate or predict this.

We thank the reviewer for their comment. We have included a supplementary figure, Figure 4 – figure supplement 1, that quantifies how sensitive the trained gene expression model is to perturbations in H3K9me3. Indeed our data suggests that the model predictions are sensitive to perturbations in H3K9me3. For instance, there is a clear decrease and a gradual increase as the position where the perturbation is performed moves from upstream to downstream of the TSS. Additionally, the magnitude of the predicted fold-change is a function of how much the H3K9me3 is perturbed and hence the magnitude of change would be even higher if the perturbation magnitude is increased. However, this precise magnitude is hard to estimate In the absence of experimental perturbation data for H3K9me3.

The model seems to predict gene expression for endogenous genes quite well although the authors sometimes use expression and sometimes use rank (e.g. Figure 6) - being clearer with how the model predicts expression rather than using rank or fold change would be very useful.

We thank the reviewer for this important suggestion. We have added text in the revised manuscript to clarify that the model predicts gene expression values, which can be interpreted as rank or fold change, depending on the use case.

One concern overall with this approach is that dCas9-p300 has been observed to induce sgRNA-independent off-target H3K27Ac (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8349887/ see Figure S5D) which could convolute interpretation of this type of experiment for the model.

This is an excellent point and indeed, we and others have observed that dCas9-p300 can result in off-target H3K27ac levels (both increased and suppressed) across the genome. However, p300 is one of the few known proteins that can catalyze H3K27ac in the human genome, and H3K27ac remains a proxy for active genomic regulatory elements. Nevertheless, dCas9-p300 off target activity could certainly convolute our approach. We have included language to address this caveat in our discussion. Interestingly, even though dCas9-p300 (and other epigenome editing enzymes) can lead to off-target chromatin modifications, these effects often occur without coincident disruptions to the transcriptome. This suggests that many chromatin modifications, while “supportive” or “instructive” of/for transcription, may be insufficient (either alone or in the context of dCas9-based fusions) for transcriptional effects.

Figure 2

It seems this figure presents known rather than novel findings from the authors' description. Please comment on whether there are any new findings in this figure. Please comment on differences in patterns of repressive and activating histone PTMs between cell lines (e.g. H1-Esc H3K27me3 green 25-50% is more enriched than red 0-25%).

Thank you for pointing out this issue. We have revised the text in both the Results and Discussion sections to better articulate that the goal of this figure is to validate the hypothesis that there are consistent patterns of histone PTMs with respect to gene expression across different human cell types.

In Figure 2, which illustrates the raw histone marks data, the non-monotonic behavior of H3K27me3 in H1-hESC cells is indicative of a real biological phenomenon. This interpretation is supported by the relatively low Pearson correlation for the H3K27me3 mark observed in these cells, as documented in Figure 1b of another study: https://www.biorxiv.org/content/10.1101/2024.03.29.587323v1.

Figure 3&4

There are a number of approaches including DeepChrome and TransferChrome that predict endogenous gene expression from histone PTMs. I appreciate that the authors have not used the histone PTM data to predict gene expression levels of an "average cell" but rather that they are predicting expression within specific cell types or for unseen cell types. But from what is presented it isn't clear that the author's model is better or enabling beyond other approaches. The authors should show their model is better than other approaches or make clear why this is a significant advance that will be enabling for the field. For example is it that in this approach they are actually predicting expression levels whereas previous approaches have only predicted expressed or not expressed or a rank order or bin-based ranking?

We thank the reviewer for this comment. We have added text to clarify the difference between our approach and existing approaches. There are two key differences between our model and other approaches. First, the gene expression model that we have trained here predicts gene expression values instead of gene expression levels as either high or low. Second, we have trained our models on ENCODE p-value data instead of read depths obtained from the Roadmap Epigenomics Consortium.

Figure 5

From the methods, it seems gene activation is measured by qpcr in hek293 transfected with individual sgRNAs and dCas9-p300. The cells aren't selected or sorted before qPCR so how are we sure that some of the variability isn't due to transfection efficiency associated with variable DNA quality or with variable transfection efficiency?

This is a good question. All DNA preps were generated using high-quality reagents and consistent protocols. In addition, the only variable that changed with respect to transfection efficiency was the gRNA-encoding vector used in qPCR assays. We have added new data which demonstrates that transfection efficiency is shared across experiments (Figure 5 – figure supplement 1). We have also added additional experimental data as well as computational analysis analyzing a new dCas9-p300 based Perturb-seq dataset to the manuscript (Figure 6 – figure supplement 1), which use lentiviral transduction and RNA-seq as readouts and thus, are buffered against the variances mentioned by the Reviewer.

Figure 6

The use of rank in 6D and 6E is confusing. In 6D a higher rank is associated with higher expression while in 6E a higher rank seems to mean a lower fold change e.g. CYP17A1 has a low predicted fold-change rank and qPCR fold-change rank but in Figure 5 a very high qPCR fold change. Labeling this more clearly or explaining it in the text further would be useful.

We thank the reviewer for their suggestion. We have made relevant changes to the caption of Figure 6 to clarify this.

Reviewer #2 (Public Review):

Summary:

The authors build a gene expression model based on histone post-translational modifications and find that H3K27ac is correlated with gene expression. They proceed to perturb H3K27ac at 8 gene promoters, and measure gene expression changes to test their model.

Strengths:

The combination of multiple methods to model expression, along with utilizing 6 histone datasets in 13 cell types allowed the authors to build a model that correlates between 0.7-0.79 with gene expression. This group also utilized a tool they are experts in, dCas9-p300 fusions to perturb H3K27ac and monitor gene expression to test their model. Ranked correlations showed some support for the predictions after the perturbation of H3K27ac.

Weaknesses:

The perturbation of only 8 genes, and the only readout being qPCR-based gene expression, as opposed to including H3K27ac, weakened their validation of the computational model. Likewise, the use of six genes that were not expressed being most activated by dCas9-p300 might weaken the correlations vs. looking at a broad range of different gene expressions as the original model was trained on.

We thank the reviewer for their comments. We have added additional experimental data as well as computational analysis analyzing a new dCas9-p300 based Perturb-seq dataset to the manuscript. We observe that the models we have developed are able to predict the fold-change rank across genes reasonably well (Figure 6 – figure supplement 1), similar to what we observe in Figure 6E.

Reviewer #1 (Recommendations For The Authors):

The authors should comment on how their model is different from or better than other models that use histone PTM data to predict gene expression.

We thank the reviewer for this insightful suggestion. We have included citations for a series of papers (PMIDs: 27587684, 30147283, 36588793) that performed gene expression prediction using histone PTM data. However, each of these methods perform classification of gene expression as opposed to predicting the actual gene expression value via regression. Additionally, the referenced studies all work with Roadmap Epigenomics read depth data as opposed to p-values obtained from the ENCODE pipelines, making it difficult to make direct comparisons.

The authors need to make clear whether their model will apply to other common epigenetic or transcriptional editors such as CRISPRi/H3K9me3 which is widely used.

In this study, we focus on the histone changes induced by p300. However, future studies may use the framework described in our manuscript and apply it to other transcriptional editors as well.

The authors need to be clearer about where they are predicting expression and where they are using rank. Ideally, show both.

We thank the reviewer for this important suggestion. We have added text in the revised manuscript to clarify that the model predicts gene expression values, which can be interpreted as rank or fold change, depending on the use case.

The authors should ideally show a case where they use the model to make a prediction of genes that can and can not be activated by dCas9-p300 or other epigenetic editors and then prove this with experiments.

Thank you for the excellent suggestion. While it is indeed relevant, exploring this would extend beyond the scope of our current study. We consider it a valuable topic for future research.

Reviewer #2 (Recommendations For The Authors):

The y-axis in 5C needs to be labeled. The authors state it is "relative mRNA" but these numbers correlated with fold changes shown in Table S2.

We have clarified the definition of the Y-axis in the caption for Figure 5C.

eLife Assessment

This valuable study presents a meta-analysis of the literature, confirming the relationship between the coupling of slow oscillations and fast spindles in memory formation, although the reported effects are weak and should be more clearly justified. Furthermore, while the evidence is convincing overall, the manuscript provides an incomplete description of the methods, which may challenge comprehension for readers unfamiliar with advanced statistical techniques. This study will be of interest to neuroscientists focusing on sleep and memory.

Reviewer #1 (Public review):

In this meta-analysis, Ng and colleagues review the association between slow-oscillation spindle coupling during sleep and overnight memory consolidation. The coupling of these oscillations (and also hippocampal sharp-wave ripples) have been central to theories and mechanistic models of active systems consolidation, that posit that the coupling between ripples, spindles, and slow oscillations (SOs) coordinate and drive the coordinated reactivation of memories in hippocampus and cortex, facilitating cross-regional information and ultimately memory strengthening and stabilisation.

Given the importance that these coupling mechanisms have been given in theory, this is a timely and important contribution to the literature in terms of determining whether these theoretical assumptions hold true in human data. The results show that the timing of sleep spindles relative to the SO phase, and the consistency of that timing, predicted overnight memory consolidation in meta-analytic models. The overall amount of coupling events did not show as strong a relationship. The coupling phase in particular was moderated by a number of variables including spindle type (fast, slow), channel location (frontal, central, posterior), age, and memory type. The main takeaway is that fast spindles that consistently couple close to the peak of the SO in frontal channel locations are optimal for memory consolidation, in line with theoretical predictions.

I did not follow the logic behind including spindle amplitude in the meta-analysis. This is not a measure of SO-spindle coupling (which is the focus of the review), unless the authors were restricting their analysis of the amplitude of coupled spindles only. It doesn't sound like this is the case though. The effect of spindle amplitude on memory consolidation has been reviewed in another recent meta-analysis (Kumral et al, 2023, Neuropsychologia). As this isn't a measure of coupling, it wasn't clear why this measure was included in the present meta-analysis. You could easily make the argument that other spindle measures (e.g., density, oscillatory frequency) could also have been included, but that seems to take away from the overall goal of the paper which was to assess coupling.

At the end of the first paragraph of section 3.1 (page 13), the authors suggest their results "... further emphasise the role of coupling compared to isolated oscillation events in memory consolidation". This had me wondering how many studies actually test this. For example, in a hierarchical regression model, would coupled spindles explain significantly more variance than uncoupled spindles? We already know that spindle activity, independent of whether they are coupled or not, predicts memory consolidation (e.g., Kumral meta-analysis). Is the variance in overnight memory consolidation fully explained by just the coupled events? If both overall spindle density and coupling measures show an equal association with consolidation, then we couldn't conclude that coupling compared to isolated events is more important.

It was very interesting to see that the relationship between the fast spindle coupling phase and overnight consolidation was strongest in the frontal electrodes. Given this, I wonder why memory promoting fast spindles shows a centro-parietal topography? Surely it would be more adaptive for fast spindles to be maximally expressed in frontal sites. Would a participant who shows a more frontal topography of fast spindles have better overnight consolidation than someone with a more canonical centro-parietal topography? Similarly, slow spindles would then be perfectly suited for memory consolidation given their frontal distribution, yet they seem less important for memory.

The authors rightly note the issues with multiple comparisons in sleep physiology and memory studies. Multiple comparison issues arise in two ways in this literature. First are comparisons across multiple electrodes (many studies now use high-density systems with 64+ channels). Second are multiple comparisons across different outcome variables (at least 3 ways to quantify coupling (phase, consistency, occurrence) x 2 spindle types (fast, slow). Can the authors make some recommendations here in terms of how to move the field forward, as this issue has been raised numerous times before (e.g., Mantua 2018, Sleep; Cox & Fell 2020, Sleep Medicine Reviews for just a couple of examples). Should researchers just be focusing on the coupling phase? Or should researchers always report all three metrics of coupling, and correct for multiple comparisons? I think the use of pre-registration would be beneficial here, and perhaps could be noted by the authors in the final paragraph of section 3.5, where they discuss open research practices.

Reviewer #2 (Public review):

Summary:

This article reviews the studies on the relationship between slow oscillation (SO)-spindle (SP) coupling and memory consolidation. It innovatively employs non-normal circular linear correlations through a Bayesian meta-analysis. A systematic analysis of the retrieved studies highlighted that co-coupling of SO and the fast SP's phase and amplitude at the frontal part better predicts memory consolidation performance. I only have a few comments that I recommend are addressed.

Major Comments:

Regarding the Moderator of Age: Although the authors discuss the limited studies on the analysis of children and elders regarding age as a moderator, the figure shows a significant gap between the ages of 40 and 60. Furthermore, there are only a few studies involving participants over the age of 60. Given the wide distribution of effect sizes from studies with participants younger than 40, did the authors test whether removing studies involving participants over 60 would still reveal a moderator effect?

Reviewer #3 (Public review):

This manuscript presents a meta-analysis of 23 studies, which report 297 effect sizes, on the effect of SO-spindle coupling on memory performance. The analysis has been done with great care, and the results are described in great detail. In particular, there are separate analyses for coupling phase, spindle amplitude, coupling strength (e.g., measured by vector length or modulation index), and coupling percentage (i.e., the percentage of SPs coupled with SOs). The authors conclude that the precision and strength of coupling showed significant correlations with memory retention.

There are two main points where I do not agree with the authors.

First, the authors conclude that "SO-SP coupling should be considered as a general physiological mechanism for memory consolidation". However, the reported effect sizes are smaller than what is typically considered a "small effect" (0.10<br /> Second, the study implements state-of-the-art Bayesian statistics. While some might see this as a strength, I would argue that it is the greatest weakness of the manuscript. A classical meta-analysis is relatively easy to understand, even for readers with only a limited background in statistics. A Bayesian analysis, on the other hand, introduces a number of subjective choices that render it much less transparent. This becomes obvious in the forest plots. It is not immediately apparent to the reader how the distributions for each study represent the reported effect sizes (gray dots). Presumably, they depend on the Bayesian priors used for the analysis. The use of these priors makes the analyses unnecessarily opaque, eventually leading the reader to question how much of the findings depend on subjective analysis choices (which might be answered by an additional analysis in the supplementary information). However, most of the methods are not described in sufficient detail for the reader to understand the proceedings. It might be evident for an expert in Bayesian statistics what a "prior sensitivity test" and a "posterior predictive check" are, but I suppose most readers would wish for a more detailed description. However, using a "Markov chain Monte Carlo (MCMC) method with the no-U-turn Hamiltonian Monte Carlo (HMC) sampler" and checking its convergence "through graphical posterior predictive checks, trace plots, and the Gelman and Rubin Diagnostic", which should then result in something resembling "a uniformly undulating wave with high overlap between chains" is surely something only rocket scientists understand. Whether this was done correctly in the present study cannot be ascertained because it is only mentioned in the methods and no corresponding results are provided. This kind of analysis seems not to be made to be intelligible to the average reader. It follows a recent trend of using more and more opaque methods. Where we had to trust published results a decade ago because the data were not openly available, today we must trust the results because the methods can no longer be understood with reasonable effort.

In one point the method might not be sufficiently justified. The method used to transform circular-linear r (actually, all references cited by the authors for circular statistics use r² because there can be no negative values) into "Z_r", seems partially plausible and might be correct under the H0. However, Figure 12.3 seems to show that under the alternative Hypothesis H1, the assumptions are not accurate (peak Z_r=~0.70 for r=0.65). I am therefore, based on the presented evidence, unsure whether this transformation is valid. Also, saying that Z_r=-1 represents the null hypothesis and Z_r=1 the alternative hypothesis can be misinterpreted, since Z_r=0 also represents the null hypothesis and is not half way between H0 and H1.

Author response:

Reviewer #1 (Public review):

I did not follow the logic behind including spindle amplitude in the meta-analysis. This is not a measure of SO-spindle coupling (which is the focus of the review), unless the authors were restricting their analysis of the amplitude of coupled spindles only. It doesn't sound like this is the case though. The effect of spindle amplitude on memory consolidation has been reviewed in another recent meta-analysis (Kumral et al, 2023, Neuropsychologia). As standardization this isn't a measure of coupling, it wasn't clear why this measure was included in the present meta-analysis. You could easily make the argument that other spindle measures (e.g., density, oscillatory frequency) could also have been included, but that seems to take away from the overall goal of the paper which was to assess coupling.

Indeed, spindle amplitude refers to all spindle events rather than only coupled spindles. This choice was made because we recognized the challenge of obtaining relevant data from each study—only 4 out of the 23 included studies performed their analyses after separating coupled and uncoupled spindles. This inconsistency strengthens the urgency and importance of this meta-analysis to standardize the methods and measures used for future analysis on SO-SP coupling and beyond. We agree that focusing on the amplitude of coupled spindles would better reveal their relations with coupling, and we will discuss this limitation in the manuscript.

Nevertheless, we believe including spindle amplitude in our study remains valuable, as it served several purposes. First, SO-SP coupling involves the modulation between spindle amplitude and slow oscillation phase. Different studies have reported conflicting conclusions regarding how spindle amplitude was related to coupling– some found significant correlations (e.g., Baena et al., 2023), while others did not (e.g., Roebber et al., 2022). This discrepancy highlights an indirect but potentially crucial insight into the role of spindle amplitude in coupling dynamics. Second, in studies related to SO-SP coupling, spindle amplitude is one of the most frequently reported measures along with other coupling measures that significantly correlated with oversleep memory improvements (e.g. Kurz et al., 2023; Ladenbauer et al., 2021; Niknazar et al., 2015), so we believe that including this measure can more comprehensively review of the existing literature on SO-SP coupling. Third, incorporating spindle amplitude allows for a direct comparison between the measurement of coupling and individual events alone in their contribution to memory consolidation– a question that has been extensively explored in recent research. (e.g., Hahn et al., 2020; Helfrich et al., 2019; Niethard et al., 2018; Weiner et al., 2023). Finally, spindle amplitude was identified as a key moderator for memory consolidation in Kumral et al.'s (2023) meta-analysis. By including it in our analysis, we sought to replicate their findings within a broader framework and introduce conceptual overlaps with existing reviews. Therefore, although we were not able to selectively include coupled spindles, there is still a unique relation between spindle amplitude and SO-SP coupling that other spindle measures do not have. 

Originally, we also intended to include coupling density or counts in the analysis, which seems more relevant to the coupling metrics. However, the lack of uniformity in methods used to measure coupling density posed a significant limitation. We hope that our study will encourage consistent reporting of all relevant parameters in future research, enabling future meta-analyses to incorporate these measures comprehensively. We will add this discussion to the manuscript in the revised version to further clarify these points.

References:

Roebber, J. K., Lewis, P. A., Crunelli, V., Navarrete, M. & Hamandi, K. Effects of anti-seizure medication on sleep spindles and slow waves in drug-resistant epilepsy. Brain Sci. 12, 1288 (2022). https://doi.org/10.3390/brainsci12101288

All other citations were referenced in the manuscript.

At the end of the first paragraph of section 3.1 (page 13), the authors suggest their results "... further emphasise the role of coupling compared to isolated oscillation events in memory consolidation". This had me wondering how many studies actually test this. For example, in a hierarchical regression model, would coupled spindles explain significantly more variance than uncoupled spindles? We already know that spindle activity, independent of whether they are coupled or not, predicts memory consolidation (e.g., Kumral meta-analysis). Is the variance in overnight memory consolidation fully explained by just the coupled events? If both overall spindle density and coupling measures show an equal association with consolidation, then we couldn't conclude that coupling compared to isolated events is more important.

While primary coupling measurements, including coupling phase and strength, showed strong evidence for their associations with memory consolidation, measures of spindles, including spindle amplitude, only exhibited limited evidence (or “non-significant” effect) for their association with consolidation. These results are consistent with multiple empirical studies using different techniques (e.g., Hahn et al., 2020; Helfrich et al., 2019; Niethard et al., 2018; Weiner et al., 2023), which reported that coupling metrics are more robust predictors of consolidation and synaptic plasticity than spindle or slow oscillation metrics alone. However, we agree with the reviewer that we did not directly separate the effect between coupled and uncoupled spindles, and a more precise comparison would involve contrasting the “coupling of oscillation events” with ”individual oscillation events” rather than coupling versus isolated events.

We recognized that Kumral and colleagues’ meta-analysis reported a moderate association between spindle measures and memory consolidation (e.g., for spindle amplitude-memory association they reported an effect size of approximately r = 0.30). However, one of the advantages of our study is that we actively cooperated with the authors to obtain a large number of unreported and insignificant data relevant to our analysis, as well as separated data that were originally reported under mixed conditions. This approach decreases the risk of false positives and selective reporting of results, making the effect size more likely to approach the true value. In contrast, we found only a weak effect size of r = 0.07 with minimal evidence for spindle amplitude-memory relation. However, we agree with the reviewer that using a more conservative term in this context would be a better choice since we did not measure all relevant spindle metrics including the density.

To improve clarity in our manuscript, we will revise the statement to: “Together with other studies included in the review, our results suggest a crucial role of coupling but did not support the role of spindle events alone in memory consolidation,” and provide relevant references. We believe this can more accurately reflect our findings and the existing literature to address the reviewer’s concern.

It was very interesting to see that the relationship between the fast spindle coupling phase and overnight consolidation was strongest in the frontal electrodes. Given this, I wonder why memory promoting fast spindles shows a centro-parietal topography? Surely it would be more adaptive for fast spindles to be maximally expressed in frontal sites. Would a participant who shows a more frontal topography of fast spindles have better overnight consolidation than someone with a more canonical centro-parietal topography? Similarly, slow spindles would then be perfectly suited for memory consolidation given their frontal distribution, yet they seem less important for memory.

Regarding the topography of fast spindles and their relationship to memory consolidation, we agree this is an intriguing issue, and we have already developed significant progress in this topic in our ongoing work. We share a few relevant observations: First, there are significant discrepancies in the definition of “slow spindle” in the field. Some studies defined slow spindle from 9-12 Hz (e.g. Mölle et al., 2011; Kurz et al., 2021), while others performed the event detection within a range of 11-13/14 Hz (e.g. Barakat et al., 2011; D'Atri et al., 2018). Compounding this issue, individual differences in spindle frequency are often overlooked, leading to challenges in reliably distinguishing between slow and fast spindles. Some studies have reported difficulty in clearly separating the two types of spindles altogether (e.g., Hahn et al., 2020). Moreover, a critical factor often ignored in past research is the traveling nature of both slow oscillations and spindles across the cortex, where spindles are coupled with significantly different phases of slow oscillations (see Figure 5). We believe a better understanding of coupling in the context of the movement of these waves will help us better understand the observed frontal relationship with consolidation. We will address this in our revised manuscript.

The authors rightly note the issues with multiple comparisons in sleep physiology and memory studies. Multiple comparison issues arise in two ways in this literature. First are comparisons across multiple electrodes (many studies now use high-density systems with 64+ channels). Second are multiple comparisons across different outcome variables (at least 3 ways to quantify coupling (phase, consistency, occurrence) x 2 spindle types (fast, slow). Can the authors make some recommendations here in terms of how to move the field forward, as this issue has been raised numerous times before (e.g., Mantua 2018, Sleep; Cox & Fell 2020, Sleep Medicine Reviews for just a couple of examples). Should researchers just be focusing on the coupling phase? Or should researchers always report all three metrics of coupling, and correct for multiple comparisons? I think the use of pre-registration would be beneficial here, and perhaps could be noted by the authors in the final paragraph of section 3.5, where they discuss open research practices.

There are indeed multiple methods that we can discuss, including cluster-based and non-parametric methods, etc., to correct for multiple comparisons in EEG data with spatiotemporal structures. In addition, encouraging the reporting of all tested but insignificant results, at least in supplementary materials, is an important practice that helps readers understand the findings with reduced bias. We agree with the reviewer’s suggestions and will add more information in section 3.5 to advocate for a standardized “template” used to analyze and report effect size in future research.

We advocate for the standardization of reporting all three coupling metrics– phase, consistency, and occurrence. Each coupling metric captures distinct properties of the coupling process and may interact with one another (Weiner et al., 2023). Therefore, we believe it is essential to report all three metrics to comprehensively explore their different roles in the “how, what, and where” of long-distance communication and consolidation of memory. As we advance toward a deeper understanding of the relationship between memory and sleep, we hope this work establishes a standard for the standardization, transparency, and replication of relevant studies.

Reviewer #2 (Public review):

Regarding the Moderator of Age: Although the authors discuss the limited studies on the analysis of children and elders regarding age as a moderator, the figure shows a significant gap between the ages of 40 and 60. Furthermore, there are only a few studies involving participants over the age of 60. Given the wide distribution of effect sizes from studies with participants younger than 40, did the authors test whether removing studies involving participants over 60 would still reveal a moderator effect?

We agree that there is an age gap between younger and older adults, as current studies often focus on contrasting newly matured and fully aged populations to amplify the effect, while neglecting the gradual changes in memory consolidation mechanisms across the aging spectrum. We suggest that a non-linear analysis of age effects would be highly valuable, particularly when additional child and older adult data become available.

In response to the reviewer’s suggestion, we re-tested the moderation effect of age after excluding effect sizes from older adults. The results revealed a decrease in the strength of evidence for phase-memory association due to increased variability, but were consistent for all other coupling parameters. The mean estimations also remained consistent (coupling phase-memory relation: -0.005 [-0.013, 0.004], BF10 = 5.51, the strength of evidence reduced from strong to moderate; coupling strength-memory relation: -0.005 [-0.015, 0.008], BF10 = 4.05, the strength of evidence remained moderate). These findings align with prior research, which typically observed a weak coupling-memory relationship in older adults during aging (Ladenbauer et al, 2021; Weiner et al., 2023) but not during development (Hahn et al., 2020; Kurz et al., 2021; Kurz et al., 2023). Therefore, this result is not surprising to us, and there are still observable moderate patterns in the data. We will report these additional results in the revised manuscript, and interpret “the moderator effect of age becomes less pronounced during development after excluding the older adult data”. We believe the original findings including the older adult group remain meaningful after cautious interpretation, given that the older adult data were derived from multiple studies and different groups.

Reviewer #3 (Public review):

First, the authors conclude that "SO-SP coupling should be considered as a general physiological mechanism for memory consolidation". However, the reported effect sizes are smaller than what is typically considered a "small effect" (0.10)

While we acknowledge the concern about the small effect sizes reported in our study, it is important to contextualize these findings within the field of neuroscience, particularly memory research. Even in individual studies, small effect sizes are not uncommon due to the inherent complexity of the mechanisms involved and the multitude of confounding variables. This is an important factor to be considered in meta-analyses where we synthesize data from diverse populations and experimental conditions. For example, the relationship between SO-slow SP coupling and memory consolidation in older adults is expected to be insignificant.

As Funder and Ozer (2019) concluded in their highly cited paper, an effect size of r = 0.3 in psychological and related fields should be considered large, with r = 0.4 or greater likely representing an overestimation and rarely found in a large sample or in a replication. Therefore, we believe r = 0.1 should not be considered as a lower bound of the small effect. Bakker et al. (2019) also advocate for a contextual interpretation of the effect size. This is particularly important in meta-analyses, where the results are less prone to overestimation compared to individual studies, and we cooperated with all authors to include a large number of unreported and insignificant results. In this context, small correlations may contain substantial meaningful information to interpret. Although we agree that effect sizes reported in our study are indeed small at the overall level, they reflect a rigorous analysis that incorporates robust evidence across different levels of moderators. Our moderator analyses underscore the dynamic nature of coupling-memory relationships, with certain subgroups demonstrating much stronger and more meaningful effects, especially after excluding slow spindles and older adults. For example, both the coupling phase and strength of frontal fast spindles with slow oscillations exhibited "moderate-to-large" correlations with the consolidation of different types of memory, especially in young adults, with r values ranging from 0.18 to 0.32. (see Table S9.1-9.4). We will add more discussion about the influence of moderators on the dynamics of coupling-memory associations. In addition, we will update the conclusion to be “SO-fast SP coupling should be considered as a general physiological mechanism for memory consolidation”.

Reference:

Funder, D. C. & Ozer, D. J. Evaluating effect size in psychological research: sense and nonsense. Adv. Methods Pract. Psychol. Sci. 2, 156–168 (2019). https://doi.org/10.1177/2515245919847202.

Bakker, A. et al. Beyond small, medium, or large: Points of consideration when interpreting effect sizes. Educ. Stud. Math. 102, 1–8 (2019). https://doi.org/10.1007/s10649-019-09908-4

Second, the study implements state-of-the-art Bayesian statistics. While some might see this as a strength, I would argue that it is the greatest weakness of the manuscript. A classical meta-analysis is relatively easy to understand, even for readers with only a limited background in statistics. A Bayesian analysis, on the other hand, introduces a number of subjective choices that render it much less transparent.

This kind of analysis seems not to be made to be intelligible to the average reader. It follows a recent trend of using more and more opaque methods. Where we had to trust published results a decade ago because the data were not openly available, today we must trust the results because the methods can no longer be understood with reasonable effort.

This becomes obvious in the forest plots. It is not immediately apparent to the reader how the distributions for each study represent the reported effect sizes (gray dots). Presumably, they depend on the Bayesian priors used for the analysis. The use of these priors makes the analyses unnecessarily opaque, eventually leading the reader to question how much of the findings depend on subjective analysis choices (which might be answered by an additional analysis in the supplementary information).

We appreciate the reviewer for sharing this viewpoint and we value the opportunity to clarify some key points. To address the concern about clarity, we will include a sub-section in the methods section explaining how to interpret Bayesian statistics including priors, posteriors, and Bayes factors, making our results more accessible to those less familiar with this approach.

On the use of Bayesian models, we believe there may have been a misunderstanding. Bayesian methods, far from being "opaque" or overly complex, are increasingly valued for their ability to provide nuanced, accurate, and transparent inferences (Sutton & Abrams, 2001; Hackenberger, 2020; van de Schoot et al., 2021; Smith et al., 1995; Kruschke & Liddell, 2018). It has been applied in more than 1,200 meta-analyses as of 2020 (Hackenberger, 2020). In our study, we used priors that assume no effect (mean set to 0, which aligns with the null) while allowing for a wide range of variation to account for large uncertainties. This approach reduces the risk of overestimation or false positives and demonstrates much-improved performance over traditional methods in handling variability (Williams et al., 2018; Kruschke & Liddell, 2018). Sensitivity analyses reported in the supplemental material (Table S9.1-9.4) confirmed the robustness of our choices of priors– our results did not vary by setting different priors.

As Kruschke and Liddell (2018) described, “shrinkage (pulling extreme estimates closer to group averages) helps prevent false alarms caused by random conspiracies of rogue outlying data,” a well-known advantage of Bayesian over traditional approaches. This explains the observed differences between the distributions and grey dots in the forest plots. Unlike p-values, which can be overestimated with a large sample size and underestimated with a small sample size, Bayesian methods make assumptions explicit, enabling others to challenge or refine them– an approach aligned with open science principles (van de Schoot et al., 2021). For example, a credible interval in Bayesian model can be interpreted as “there is a 95% probability that the parameter lies within the interval.”, while a confidence interval in frequentist model means “In repeated experiments, 95% of the confidence intervals will contain the true value.” We believe the former is much more straightforward and convincing for readers to interpret. We will ensure our justification for using Bayesian models is more clearly presented in the manuscript.

We acknowledge that even with these justifications, different researchers may still have discrepancies in their preferences for Bayesian and frequentist models. To increase the effort of transparent reporting, we have also reported the traditional frequentist meta-analysis results in Supplemental Material 10 to justify the robustness of our analysis, which suggested non-significant differences between Bayesian and frequentist models. We will include clearer references in the next version of the manuscript to direct readers to the figures that report the statistics provided by traditional models.

References:

Hackenberger, B.K. Bayesian meta-analysis now—let's do it. Croat. Med. J. 61, 564–568 (2020). https://doi.org/10.3325/cmj.2020.61.564

Sutton, A.J. & Abrams, K.R. Bayesian methods in meta-analysis and evidence synthesis. Stat. Methods Med. Res. 10, 277–303 (2001). https://doi.org/10.1177/096228020101000404

Williams, D.R., Rast, P. & Bürkner, P.C. Bayesian meta-analysis with weakly informative prior distributions. PsyArXiv (2018). https://doi.org/10.31234/osf.io/9n4zp

van de Schoot, R., Depaoli, S., King, R. et al. Bayesian statistics and modelling. Nat Rev Methods Primers 1, 1 (2021). https://doi.org/10.1038/s43586-020-00001-2

Smith, T.C., Spiegelhalter, D.J. & Thomas, A. Bayesian approaches to random-effects meta-analysis: a comparative study. Stat. Med. 14, 2685–2699 (1995). https://doi.org/10.1002/sim.4780142408

Kruschke, J.K. & Liddell, T.M. The Bayesian New Statistics: Hypothesis testing, estimation, meta-analysis, and power analysis from a Bayesian perspective. Psychon. Bull. Rev. 25, 178–206 (2018). https://doi.org/10.3758/s13423-016-1221-4

However, most of the methods are not described in sufficient detail for the reader to understand the proceedings. It might be evident for an expert in Bayesian statistics what a "prior sensitivity test" and a "posterior predictive check" are, but I suppose most readers would wish for a more detailed description. However, using a "Markov chain Monte Carlo (MCMC) method with the no-U-turn Hamiltonian Monte Carlo (HMC) sampler" and checking its convergence "through graphical posterior predictive checks, trace plots, and the Gelman and Rubin Diagnostic", which should then result in something resembling "a uniformly undulating wave with high overlap between chains" is surely something only rocket scientists understand. Whether this was done correctly in the present study cannot be ascertained because it is only mentioned in the methods and no corresponding results are provided. 

We appreciate the reviewer’s concerns about accessibility and potential complexity in our descriptions of Bayesian methods. Our decision to provide a detailed account serves to enhance transparency and guide readers interested in replicating our study. We acknowledge that some terms may initially seem overwhelming. These steps, such as checking the MCMC chain convergence and robustness checks, are standard practices in Bayesian research and are analogous to “linearity”, “normality” and “equal variance” checks in frequentist analysis. We have provided exemplary plots in the supplemental material and will add more details to explain the interpretation of these convergence checks. We hope this will help address any concerns about methodological rigor.

In one point the method might not be sufficiently justified. The method used to transform circular-linear r (actually, all references cited by the authors for circular statistics use r² because there can be no negative values) into "Z_r", seems partially plausible and might be correct under the H0. However, Figure 12.3 seems to show that under the alternative Hypothesis H1, the assumptions are not accurate (peak Z_r=~0.70 for r=0.65). I am therefore, based on the presented evidence, unsure whether this transformation is valid. Also, saying that Z_r=-1 represents the null hypothesis and Z_r=1 the alternative hypothesis can be misinterpreted, since Z_r=0 also represents the null hypothesis and is not half way between H0 and H1.

First, we realized that in the title of Figures 12.2 and 12.3. “true r = 0.35” and “true r = 0.65” should be corrected as “true Z_r”. The method we used here is to first generate an underlying population that has null (0), moderate (0.35), or large (0.65) Z_r correlations, then test whether the sampling distribution drawn from these populations followed a normal distribution across varying sample sizes. Nevertheless, the reviewer correctly noticed discrepancies between the reported true Z_r and its sampling distribution peak. This discrepancy arises because, when generating large population data, achieving exact values close to a strong correlation like Z_r = 0.65 is unlikely. We loop through simulations to generate population data and ensure their Z_r values fall within a threshold. For moderate effect sizes (e.g., Z_r = 0.35), this is straightforward using a narrow range (0.345 < Z_r < 0.355). However, for larger effect sizes like Z_r = 0.65, a wider range (0.6 < Z_r < 0.7) is required. therefore sometimes the population we used to draw the sample has a Z_r slightly deviated from 0.65. This remains reasonable since the main point of this analysis is to ensure that large Z_r still has a normal sampling distribution, but not focus specifically on achieving Z_r = 0.65.

We acknowledge that this variability of the range used was not clearly explained and it is not accurate to report “true Z_r = 0.65”. In the revised version, we will address this issue by adding vertical lines to each subplot to indicate the Z_r of the population we used to draw samples, making it easier to check if it aligns with the sampling peak. In addition, we will revise the title to “Sampling distributions of Z_r drawn from strong correlations (Z_r = 0.6-0.7)”. We confirmed that population Z_r and the peak of their sampling distribution remain consistent under both H0 and H1 in all sample sizes with n > 25, and we hope this explanation can fully resolve your concern.

We agree with the reviewer that claiming Z_r = -1 represents the null hypothesis is not accurate. The circlin Z_r = 0 is better analogous to Pearson’s r = 0 since both represent the mean drawn from the population with the null hypothesis. In contrast, the mean effect size under null will be positive in the raw circlin r, which is one of the important reasons for the transformation. To provide a more accurate interpretation, we will update Table 6 to describe the following strength levels of evidence: no effect (r < 0), null (r = 0), small (r = 0.1), moderate (r = 0.3), and large (r = 0.5).

eLife Assessment

This manuscript describes a fundamental investigation of the functioning of Cas9 and in particular on how variant xCas9 expands DNA targeting ability by an increase-flexibility mechanism. The authors provide compelling evidence to support their mechanistic models and the relevance of flexibility and entropy in recognition. This work can be of interest to a broad community of structural biophysicists, computational biologists, chemists, and biochemists.

Joint Public Review:

Summary:

Hossain and coworkers investigate the mechanisms of recognition of xCas9, a variant of Cas9 with expanded targeting capability for DNA. They do so by using molecular simulations and combining different flavors of simulation techniques, ranging from long classical MD simulations, to enhanced sampling, to free energy calculations of affinity differences. Through this, the authors are able to develop a consistent model of expanded recognition based on the enhanced flexibility of the protein receptor.

Strengths:

The paper is solidly based on the ability of the authors to master molecular simulations of highly complex systems. In my opinion, this paper shows no major weaknesses. The simulations are carried out in a technically sound way. Comparative analyses of different systems provide valuable insights, even within the well-known limitations of MD. Plus, the authors further investigate why xCas9 exhibits improved recognition of the TGG PAM sequence compared to SpCas9 via well-tempered metadynamics simulations focusing on the binding of R1335 to the G3 nucleobase and the DNA backbone in both SpCas9 and xCas9. In this context, the authors provide a free-energy profiling that helps support their final model.

The implementation of FEP calculations to mimic directed evolution improvement of DNA binding is also interesting, original and well-conducted.

Overall, my assessment of this paper is that it represents a strong manuscript, competently designed and conducted, and highly valuable from a technical point of view.

Weaknesses:

To make their impact even more general, the authors may consider expanding their discussion on entropic binding to other recent cases that have been presented in the literature recently (such as e.g. the identification of small molecules for Abeta peptides, or the identification of "fuzzy" mechanisms of binding to protein HMGB1). The point on flexibility helping adaptability and expansion of functional properties is important, and should probably be given more evidence and more direct links with a wider picture.

eLife Assessment

This study reports valuable findings that highlight the importance of data quality and data representation for ligand-based virtual screening experiments. The authors' claims are supported by solid evidence, although the conclusions have been inferred from only two datasets. The work would gain much impact if additional datasets were used. The main findings will be of interest to cheminformaticians and medicinal chemists working in QSAR modeling, and possibly in other areas related to machine learning.

Reviewer #1 (Public review):

Summary:

The work provides more evidence of the importance of data quality and representation for ligand-based virtual screening approaches. The authors have applied different machine learning (ML) algorithms and data representation using a new dataset of BRAF ligands. First, the authors evaluate the ML algorithms, and demonstrate that independently of the ML algorithm, predictive and robust models can be obtained in this BRAF dataset. Second, the authors investigate how the molecular representations can modify the prediction of the ML algorithm. They found that in this highly curated dataset the different molecule representations are adequate for the ML algorithms since almost all of them obtain high accuracy values, with Estate fingerprints obtaining the worst performing predictive models and ECFP6 fingerprints producing the best classificatory models. Third, the authors evaluate the performance of the models on subsets of different composition and size of the BRAF dataset. They found that given a finite number of active compounds, increasing the number of inactive compounds worsens the recall and accuracy. Finally, the authors analyze if the use of "less active" molecules affect the model's predictive performance using "less active" molecules taken from ChEMBl Database or using decoys from DUD-E. As results, they found that the accuracy of the model falls as the number of "less active" examples in the training dataset increases while the implementation of decoys in the training set generates results as good as the original models or even better in some cases. However, the use of decoys in the training set worsens the predictive power in the test sets that contain active and inactive molecules.

Strengths:

This is a highly relevant topic in medicinal chemistry and drug discovery. The manuscript is well-written, with a clear structure that facilitates easy reading, and it includes up-to-date references. The hypotheses are clearly presented and appropriately explored. The study provides valuable insights into the importance of deriving models from high-quality data, demonstrating that, when this condition is met, complex computational methods are not always necessary to achieve predictive models. Furthermore, the generated BRAF dataset offers a valuable resource for medicinal chemists working in ligand-based virtual screening.

Weaknesses:

While the work highlights the importance of using high-quality datasets to achieve better and more generalizable results, it does not present significant novelty, as the analysis of training data has been extensively studied in chemoinformatics and medicinal chemistry. Additionally, the inclusion of "AI" in the context of data-centric AI is somewhat unclear, given that the dataset curation is conducted manually, selecting active compounds based on IC50 values from ChEMBL and inactive compounds according to the authors' criteria.

Moreover, the conclusions are based on the analysis of only two high-quality datasets. To generalize these findings, it would be beneficial to extend the analysis to additional high-quality datasets (at least 10 datasets for a robust benchmarking exercise).

A key aspect that could be improved is the definition of an "inactive" compound, which remains unclear. In the manuscript, it is stated:

• "The inactives were carefully selected based on the fact that they have no known pharmacological activity against BRAF."<br /> Does the lack of BRAF activity data necessarily imply that these compounds are inactive?<br /> • "We define a compound as 'inactive' if there are no known pharmacological assays for the said compound on our target, BRAF."<br /> However, in the authors' response, they mention:<br /> • "We selected certain compounds that we felt could not possibly be active against BRAF, such as ligands for neurotransmitter receptors, as inactives."

Given that the definition of "inactive" is one of the most critical concepts in the study, I believe it should be clearly and consistently explained.

Lastly, while statistical comparison is not always common in machine learning, it would greatly enhance the value of this work, especially when comparing models with small differences in accuracy.

Reviewer #2 (Public review):

Summary:

The authors explored the importance of data quality and representation for ligand-based virtual screening approaches. I believe the results could be of potential benefit to the drug discovery community, especially to those scientists working in the field of machine learning applied to drug research. The in silico design is comprehensive and adequate for the proposed comparisons.

This manuscript by Chong A. et al describes that it is not necessary to resort to the use of sophisticated deep learning algorithms for virtual screening, since based on their results considering conventional ML may perform exceptionally well if feeded by the right data and molecular representations.

The article is interesting and well-written. The overview of the field and the warning about dataset composition are very well thought-out and should be of interest to a broad segment of the AI in drug discovery readership. This article further highlights some of the considerations that need to be taken into consideration for the implementation of data-centric AI for computer-aided drug design methods.

Strengths:

This study contributes significantly to the field of machine learning and data curation in drug discovery. The paper is, in general, well-written and structured. However, in my opinion, there are some suggestions regarding certain aspects of the data analyses.

Weaknesses:

The conclusions drawn in the study are based on the analysis of a two dataset. The authors chose BRAF as an example in this study, and expanded with BACE-1 dataset; however a benchmark with several targets would be suitable to evaluate reproducibility or transferability of the method. One concern could be the applicability of the method in other targets.

Reviewer #3 (Public review):

Summary:

The authors presented a data-centric ML approach for virtual ligand screening. They used BRAF as an example to demonstrate the predictive power of their approach.

Strengths:

The performance of predictive models in this study is superior (nearly perfect) with respect to exiting methods.

Comments on revisions:

In the revised manuscript, the presented approach has been robustly tested and can be very useful for ligand prediction.

Author response:

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

We thank the Editors and reviewers for their candid evaluation of our work. While it was suggested that we should demonstrate the validity of our approach with maybe 10 different datasets but we felt that this would place an undue burden on our resources. Generally, it takes about 4 to 6 months for us to build a dataset and this does not include the time taken to train and test our AI models. This would mean that it would take us another 3 to 5 years to complete this research project if we chose to provide 10 different datasets. Publishing a research on one dataset is definitely not unheard of: for example, Subramanian et al. (2016) published their widely-cited benchmark dataset for just BACE1 inhibitors. However, we hoped that the additional work where we showed that we were able to improve the benchmark dataset for BACE1 inhibitors and achieve the same high level of predictive performance for this dataset would convince the readers (and reviewers) of the reproducibility of our approach. Furthermore, we also showed that our approach is robust and does not rely on a large volume of data to achieve this near-perfect accuracy. As can be seen in the Supplemental section, even our AI models trained on ONLY 250 BRAF actives and 250 inactives could achieve 96.3% accuracy! Logically, if the model is robust then we would expect the model to be reproducible. As such, we do not feel it is necessary for us to test our approach on 10 different datasets. 

It was also suggested that we expand this study to other types of molecular representations to give a better idea of generalizability. We would like to point out that we tested, in total, 55 single fingerprints and paired combinations. Our goal was to create an approach that could give superior performance for virtual screening and we believe that we have achieved this. Based on the results of our study, we are of the opinion that molecular representations do not, in general, have an oversized effect on AI virtual screening. Although it is important to be aware that certain molecular representations may give SLIGHTLY better performance but we can see that with the exception of the 79-bit E-State fingerprint (which could still achieve an impressive 85% accuracy for the SVM model), nearly all molecular fingerprints and paired combinations that we used were able to achieve an accuracy of above 97%. Therefore, we do not share the reviewers' concern that our approach may not be useful when applied with other types of molecular representations.

It is true that our work involved manual curation of the datasets but the goal of this paper is to lay down some  ground rules for the future development of a data-centric AI approach. Although manual curation is a routine practice in AI/ML, but it should be recognised that there is good manual curation and bad manual curation, and rules need to be established to ensure we have good manual curation. Without these rules, we would also not be able to establish and train a data-centric AI. All manual curation involves a level of subjectiveness but that subjectiveness comes from one's experience and domain knowledge of the field in which the AI is being applied. For example, in the case of this study, we relied on our knowledge and understanding of pharmacology to determine whether a compound is pharmacologically inactive or active. This may seem somewhat arbitrary to the uninitiated but it is anything but arbitrary. It is through careful thought and assessment of the chemical compounds that we choose these compounds for training the AI. Unfortunately, this sort of subjective assessment cannot be easily or completely explained but we do show where current practices have failed when building a dataset for training an AI for virtual screening.

eLife Assessment

This important study used an automated system to collect eggs laid over the course of multiple days by individual female Drosophila to successfully reveal a robust yet noisy circadian rhythm of egg-laying. Their results show that the neural control of this rhythm is entirely different from the one that controls locomotor activity rhythmicity. Preliminary connectome-based analyses provide evidence for connections between the relevant clock neurons and neurons involved in oviposition. The evidence provided is solid, although using an independent tool for targeted knockdown of clock genes and including the time series of representative individuals for all genotypes tested would help interpret the results.

Joint Public Review:

Riva et al uncovered the neural substrate underlying the oviposition rhythm in Drosophila melanogaster using a novel device that automates egg collection from individual mated females over the course of multiple days. By systematically knocking down the clock gene period in specific clock neurons the authors show that three cryptochrome (cry) positive dorso-lateral neurons (LNds) present in each hemisphere of the fly brain are critical to generating a female, sex-specific rhythm in oviposition. Interestingly, these neurons are not essential for freerunning locomotor activity. By contrast, the LNvs (lateral ventral neurons), which are essential for freerunning locomotor activity rhythmicity, were not involved in controlling the circadian rhythmicity of oviposition. Thus, this work has identified the first truly sex-specific circadian circuit in Drosophila. Using available Drosophila hemibrain connectome data they identify bidirectional connections between cry-expressing LNd and oviposition-related neurons.

Strengths:

This paper established a new semi-automatic device to register egg-laying activity, in Drosophila and found a specific role for a subset of clock neurons in the control of a female-specific circadian behavior. They also lay the groundwork for understanding how these neurons are connected to the neurons that control egg laying.

Weaknesses:

(1) Controls for the genetic background are incomplete, leaving open the possibility that the observed oviposition timing defects may be due to targeted knockdown of the period (per) gene but from the GAL4, Gal80, and UAS transgenes themselves. To resolve this issue the authors should determine the egg-laying rhythms of the relevant controls (GAL4/+, UAS-RNAi/+, etc); this only needs to be done for those genotypes that produced an arrhythmic egg-laying rhythm.

(2) Reliance on a single genetic tool to generate targeted disruption of clock function leaves the study vulnerable to associated false positive and false negative effects: a) The per RNAi transgene used may only cause partial knockdown of gene function, as suggested by the persistent rhythmicity observed when per RNAi was targeted to all clock neurons. This could indicate that the results in Fig 2C-H underestimate the phenotypes of targeted disruption of clock function. b) Use of a single per RNAi transgene makes it difficult to rule out that off-target effects contributed significantly to the observed phenotypes. We suggest that the authors repeat the critical experiments using a separate UAS-RNAi line (for period or for a different clock gene), or, better yet, use the dominant negative UAS-cycle transgene produced by the Hardin lab (https://doi.org/10.1038/22566).

(3) The egg-laying profiles obtained show clear damping/decaying trends which necessitates careful trend removal from the data to make any sense of the rhythm. Further, the detrending approach used by the authors is not tested for artefacts introduced by the 24h moving average used.

(4) According to the authors the oviposition device cannot sample at a resolution finer than 4 hours, which will compel any experimenter to record egg laying for longer durations to have a suitably long time series which could be useful for circadian analyses.

(5) Despite reducing the interference caused by manually measuring egg-laying, the rhythm does not improve the signal quality such that enough individual rhythmic flies could be included in the analysis methods used. The authors devise a workaround by combining both strongly and weakly rhythmic (LSpower > 0.2 but less than LSpower at p < 0.05) data series into an averaged time series, which is then tested for the presence of a 16-32h "circadian" rhythm. This approach loses valuable information about the phase and period present in the individual mated females, and instead assumes that all flies have a similar period and phase in their "signal" component while the distribution of the "noise" component varies amongst them. This assumption has not yet been tested rigorously and the evidence suggests a lot more variability in the inter-fly period for the egg-laying rhythm.

(6) This variability could also depend on the genotype being tested, as the authors themselves observe between their Canton-S and YW wild-type controls for which their egg-laying profiles show clearly different dynamics. Interestingly, the averaged records for these genotypes are not distinguishable but are reflected in the different proportions of rhythmic flies observed. Unfortunately, the authors also do not provide further data on these averaged profiles, as they did for the wild-type controls in Figure 1, when they discuss their clock circuit manipulations using perRNAi. These profiles could have been included in Supplementary figures, where they would have helped the reader decide for themselves what might have been the reason for the loss of power in the LS periodogram for some of these experimental lines.

(7) By selecting 'the best egg layers' for inclusion in the oviposition analyses an inadvertent bias may be introduced and the results of the assays may not be representative of the whole population.

(8) An approach that measures rhythmicity for groups of individual records rather than separate individual records is vulnerable to outliers in the data, such as the inclusion of a single anomalous individual record. Additionally, the number of individual records that are included in a group may become a somewhat arbitrary determinant for the observed level of rhythmicity. Therefore, the experimental data used to map the clock neurons responsible for oviposition rhythms would be more convincing if presented alongside individual fly statistics, in the same format as used for Figure 1.

(9) The features in the experimental periodogram data in Figures 3B and D are consistent with weakened complex rhythmicity rather than arrhythmicity. The inclusion of more individual records in the groups might have provided the added statistical power to demonstrate this. Graphs similar to those in 1G and 1I, might have better illustrated qualitative and quantitative aspects of the oviposition rhythms upon per knockdown via MB122B and Mai179; Pdf-Gal80.

Wider context:

The study of the neural basis of oviposition rhythms in Drosophila melanogaster can serve as a model for the analogous mechanisms in other animals. In particular, research in this area can have wider implications for the management of insects with societal impact such as pests, disease vectors, and pollinators. One key aspect of D. melanogaster oviposition that is not addressed here is its strong social modulation (see Bailly et al.. Curr Biol 33:2865-2877.e4. doi:10.1016/j.cub.2023.05.074). It is plausible that most natural oviposition events do not involve isolated individuals, but rather groups of flies. As oviposition is encouraged by aggregation pheromones (e.g., Dumenil et al., J Chem Ecol 2016 https://link.springer.com/article/10.1007/s10886-016-0681-3) its propensity changes upon the pre-conditioning of the oviposition substrates, which is a complication in assays of oviposition rhythms that periodically move the flies to fresh substrate.

Author response:

(1) Controls for the genetic background are incomplete, leaving open the possibility that the observed oviposition timing defects may be due to targeted knockdown of the period (per) gene but from the GAL4, Gal80, and UAS transgenes themselves. To resolve this issue the authors should determine the egg-laying rhythms of the relevant controls (GAL4/+, UAS-RNAi/+, etc); this only needs to be done for those genotypes that produced an arrhythmic egg-laying rhythm.

We agree with this objection, and in the corrected version we plan to provide the assessment of the egg laying rhythms for the missing GAL4 controls as recommended only for Figure 3.

(2) Reliance on a single genetic tool to generate targeted disruption of clock function leaves the study vulnerable to associated false positive and false negative effects: a) The per RNAi transgene used may only cause partial knockdown of gene function, as suggested by the persistent rhythmicity observed when per RNAi was targeted to all clock neurons. This could indicate that the results in Fig 2C-H underestimate the phenotypes of targeted disruption of clock function. b) Use of a single per RNAi transgene makes it difficult to rule out that off-target effects contributed significantly to the observed phenotypes. We suggest that the authors repeat the critical experiments using a separate UAS-RNAi line (for period or for a different clock gene), or, better yet, use the dominant negative UAS-cycle transgene produced by the Hardin lab (https://doi.org/10.1038/22566).

We have recently acquired mutant flies with a dominant negative-cycle transgene (UAS-cycDN, Tanoue et al. 2004), and we plan to repeat our experiments with these mutants, in order to confirm our results.

(3) The egg-laying profiles obtained show clear damping/decaying trends which necessitates careful trend removal from the data to make any sense of the rhythm. Further, the detrending approach used by the authors is not tested for artefacts introduced by the 24h moving average used.

In the revised version we will show that the detrending approach used does not introduce any artefacts. The analysis of numerical simulations with an aperiodic stochastic signal superposed to a decaying signal shows that the detrending method used does not result in a spurious periodic signal. Furthermore, we can show that when the underlying signal is rhythmic, the correct period is obtained even when the moving average is a few hours larger or smaller than 24 h.

(4) According to the authors the oviposition device cannot sample at a resolution finer than 4 hours, which will compel any experimenter to record egg laying for longer durations to have a suitably long time series which could be useful for circadian analyses.

We apologize for not being clear enough. The device can in principle sample at any desired resolution. Notice, however, that the variable we are analyzing (number of eggs laid by a single female) has only a few possible values, which is one of the features that render the assessment of rhythmicity a particularly difficult task. If egg laying is sampled more often (say, at 2 h intervals) more time points will be available, but the values available for each time point will be much less. We will show an example where we compare both rates (2h and 4h). Even though the 2h sampling reveals the rhythmicity of the time series, the significance of the peaks obtained is less than when sampling at 4h intervals. We have found that a 4h sampling seems to provide the best compromise between frequency of the sampling and discreteness of the variable.

On the other hand, it is important to stress that sampling frequency and longer durations are not very correlated (see e.g. Cohen et al. Journal of Theoretical Biology 314, pp 182 [2012]). It has been shown that the best way to make accurate predictions of the period of a rhythmic signal is to have a series spanning many cycles, irrespective of the sampling frequency. In other words, it is not true that with a 2h sampling it would be possible to analyze shorter series than with 4h sampling. Unfortunately, egg laying records are usually less than 5 cycles long, which is one of the reasons for the difficulties in the assessment of their rhythmicity.

(5) Despite reducing the interference caused by manually measuring egg-laying, the rhythm does not improve the signal quality such that enough individual rhythmic flies could be included in the analysis methods used. The authors devise a workaround by combining both strongly and weakly rhythmic (LSpower > 0.2 but less than LSpower at p < 0.05) data series into an averaged time series, which is then tested for the presence of a 16-32h "circadian" rhythm. This approach loses valuable information about the phase and period present in the individual mated females, and instead assumes that all flies have a similar period and phase in their "signal" component while the distribution of the "noise" component varies amongst them. This assumption has not yet been tested rigorously and the evidence suggests a lot more variability in the inter-fly period for the egg-laying rhythm.

The assumption is difficult to test rigorously, since for individual flies the records seem to be so noisy that no information can be extracted. As shown in the paper, it is even very difficult to assess the presence of rhythmicity at the individual level. We consider that the appearance of a rhythm after averaging several records shows the presence of this rhythm at the individual level. But it could be argued that the presence of rhythmicity in the average record could be due to only a few (or even a single) rhythmic individuals. In order to show that this is probably not the case, in the revised version we will show that, when the individuals that are rhythmic are left out, the average of the remaining flies still shows a rhythm (albeit a weaker one, as was to be expected).

Regarding our assumption that all flies have the “same” period, the results on Fig. 1 F cannot really rule out this possibility, because with so few cycles, the determination of the period is not very accurate (see e.g. Cohen et al. Journal of Theoretical Biology 314, pp 182 [2012]). In our case, the error for the period is related to the width of the corresponding peak in the periodogram, which is typically 4 hs. In any case, in the revised version we will try to show, by using numerical simulations, that when the individual periods are not the same, but are distributed approximately as in Fig 1F, the average series is still rhythmic with the correct period.

(6) This variability could also depend on the genotype being tested, as the authors themselves observe between their Canton-S and YW wild-type controls for which their egg-laying profiles show clearly different dynamics. Interestingly, the averaged records for these genotypes are not distinguishable but are reflected in the different proportions of rhythmic flies observed. Unfortunately, the authors also do not provide further data on these averaged profiles, as they did for the wild-type controls in Figure 1, when they discuss their clock circuit manipulations using perRNAi. These profiles could have been included in Supplementary figures, where they would have helped the reader decide for themselves what might have been the reason for the loss of power in the LS periodogram for some of these experimental lines.

Even though we think that the individual records are in general too noisy to be really informative, we will provide all the individual egg profiles in the Supplementary Material of the revised version, in order to let the reader, check this for herself/himself.

(7) By selecting 'the best egg layers' for inclusion in the oviposition analyses an inadvertent bias may be introduced and the results of the assays may not be representative of the whole population.

We agree that this may introduce some bias in the results. But in our opinion this bias is very difficult to avoid, since for females that lay very few eggs, rhythmicity can even be difficult to define (some females can spend a whole day without laying a single egg). On the other hand, even when the results may not be representative of the whole population, they would be representative of the flies that lay most of the eggs in a population, which seems to be very relevant in ecological terms.

(8) An approach that measures rhythmicity for groups of individual records rather than separate individual records is vulnerable to outliers in the data, such as the inclusion of a single anomalous individual record. Additionally, the number of individual records that are included in a group may become a somewhat arbitrary determinant for the observed level of rhythmicity. Therefore, the experimental data used to map the clock neurons responsible for oviposition rhythms would be more convincing if presented alongside individual fly statistics, in the same format as used for Figure 1.

The question of possible rhythmic outliers has been addressed above, in question 5, where we discuss why we think that such outliers are not “determinant for the observed level of rhythmicity”. As also mentioned above, even though we think that they are too noisy to be informative, we plan to include all individual profiles in the Supplementary Material.

(9) The features in the experimental periodogram data in Figures 3B and D are consistent with weakened complex rhythmicity rather than arrhythmicity. The inclusion of more individual records in the groups might have provided the added statistical power to demonstrate this. Graphs similar to those in 1G and 1I, might have better illustrated qualitative and quantitative aspects of the oviposition rhythms upon per knockdown via MB122B and Mai179; Pdf-Gal80.

We assume that the features mentioned refer to the appearance in the periodograms of two small peaks under the significance lines. We are aware that in the studies of the rhythmicity of locomotor activity such features are usually interpreted as “complex rhythms”, i.e. as evidence of the existence of two different mechanisms producing two different rhythms in the same individual. In our case, however, at least two other possibilities should be taken into account. Since the periodograms we show assess the rhythmicity of the average time series of several individuals, the two small peaks could correspond to the periods of two different subpopulations. Another possibility could be that such peaks are simply an artifact of the method in the analysis of time series that consist of very few cycles (as explained above) and also few points per cycle. A cursory examination of the individual profiles, that will be provided in the new version, do not seem to support any of the first two possibilities mentioned. On the other hand, we will show evidence that the analysis of series that are perfectly random sometimes result in periodograms with some small peaks.

eLife Assessment

This important study demonstrates the ability for high-throughput recording and categorization of unconstrained and stimulus-based behaviors across a very large population of marmosets (n = 120 animals across 36 family units). The authors implement an analytical approach to identify "outlier" behavior that could be key in the development of next-generation precision psychiatry. While the strength of evidence appears solid overall, many key methodological details are incomplete.

Reviewer #1 (Public review):

Summary:

The authors demonstrate a fully unsupervised, high throughput (meaning very low human interaction required) approach to quantifying marmoset behavior in unconstrained environments.

Strengths:

The authors provide an approach that is scalable, easy to implement at face value, and highly robust. Currently, most behavioral quantification approaches do not work well on marmosets, or the published examples that do look promising do not scale towards high throughput as demonstrated by the authors.

While marmosets can certainly be a useful translational research model devoid of free behavior quantification, the authors make a compelling point about how this approach can be useful in the study of treatments of emerging marmoset disease models.

Overall this is a very exhaustive manuscript that overcomes significant shortcomings in previous work and speaks highly to the use of marmosets for unconstrained behavioral and neural assessment.

Weaknesses:

Recording marmoset behavior with a 60Hz frame rate is a significant limitation to the approach which is hopefully easily alleviated in the future through better cameras/reconstruction pipelines. Marmosets (in the reviewers' experience) have a lot of motion energy above the 30Hz nyquist limit imposed by this system and are agile to a degree requiring higher frame rates.

The manuscript neglects recent approaches to non-human primate behavioral quantification from other groups that should be included. Simians are simians after all.

As a minor weakness, this reviewer would have liked to see code shared for the reviewers to evaluate, especially pertaining to the high throughput and robustness of the approach.

Reviewer #2 (Public review):

In this manuscript, Menegas et al. classify the "control" behavior of captive marmosets. They combine behavioral screening from video recordings with audio and neural recordings (from the striatum) to better define what can be considered a typical behavioral repertoire for captive marmoset monkeys. A range of analyses is presented, investigating various aspects of behavior, such as social interactions and the detection of atypical individuals.

The manuscript is compelling in many respects, especially due to the richness of the dataset and the breadth of analyses presented. However, a significant issue with the manuscript lies in its writing: the results are conveyed in an overly succinct and superficial manner, and the "Methods" section is nearly absent. Key concepts are often undefined, and the mathematical details underlying the figures are not explained, leaving readers to guess the authors' approach.

Another issue is the vague use of the term "natural behavior." All data presented here appear to have been collected in small cages with limited climbing opportunities and enrichment. Thus, the authors should refrain from using "natural" to describe these conditions.

Below, we elaborate further on the lack of methodological detail. Based on these issues, we believe the manuscript, in its current form, does not meet the scientific standards necessary for proper review. We strongly encourage the authors to undertake an extensive revision.

Major Revision Points:

The methods and results require significantly more detail. A scientific publication should provide readers with enough information to reproduce the study. Here, the detail level is far too low to fully understand, or reproduce, the study, and in many instances, readers are left to guess how the figure panels were produced. Below is a non-exhaustive list of examples illustrating these issues:

(1) "we temporarily placed horizontal cage dividers to reduce the total cage size during data collection": What were the resulting (and initial) cage dimensions?

(2) "After training the network, we hierarchically clustered the latent space": What is the latent space? Based on Figure 2a, it appears related to the network's recurrent layer, but this is not clarified in the text.

(3) Alpha and perplexity parameters: Please define these terms. Since these concepts appear fundamental, readers should not have to consult external references.

(4) "We then traced cluster identities across hierarchical levels": What are hierarchical levels?

(5) "To understand how the input time series data was weighed in the bottleneck layer of the model": What is the bottleneck layer?

(6) "we measured the average attention allocation to previous time points": The authors should define "attention allocation."

(7) "we compared each neuron's firing rate distribution to shuffled data based on the overall frequency of each behavior during the session": This description is insufficient to understand the analysis.

(8) "we hierarchically clustered neurons according to their firing rate enrichment maps": No mathematical explanation is provided for neuron clustering, nor is the concept of a "firing rate enrichment map" clarified.

(9) "Cluster 4 showed higher activity when neurons were 'alone' or 'active'": This is vague and uses unclear jargon (e.g., "neurons alone"). Additionally, no mathematical explanation is provided for assigning neuronal activity to behavioral states.

(10) Figure 3f, right-side panels: The analysis seems to involve cage mate positioning, yet no description is provided.

(11) "we used motion watches to measure activity across all hours": Are these motion-sensitive watches physically attached to the animals? The methodology should be described, including data analysis details.

This list could continue, but we trust the authors understand the point. There is a wealth of analyses and information in this study, but the descriptions are too superficial. We understand that fully describing each analysis may require significant rewriting, including supplementary figures, and will likely make the manuscript longer. This is entirely acceptable, as the ideas presented here are worth the added rigor.

"Natural behavior": Typically, the term "natural" suggests that the dataset reflects the range of behaviors exhibited by animals in the wild. Here, however, recordings were made in a small cage with limited climbing opportunities and enrichment. Under these conditions, it's hard to justify describing the behavior as "natural". In a project aimed at classifying the behavioral repertoire of marmoset monkeys and making this dataset accessible to other laboratories, it would be helpful to include more detailed information about the animals' housing conditions. This might include cage sizes, temperature, humidity, and details on food quantities, quality, and feeding times.

Correlation versus causation: In the section titled "Large-scale data collection reveals variability across days and correlation between cagemates," the authors conclude: "Overall, these results indicate that measurements of animals' behavioral traits depend heavily on their social environment." This interpretation seems incorrect. We know that animal behavior varies throughout the day, with activity peaks typically occurring in the morning and afternoon. Such factors, or other external influences, could induce correlations between animals that are not caused by social interactions.

Figure 4g: What are we intended to conclude from this analysis?

Figure 5: Please specify the type of calls analyzed. For example, did you analyze only long-distance calls (aka 'loud phees' or 'shrills')? In "We split the audio data into 5-minute (non-continuous) segments and found that the average call rate in these segments varied from 0 calls per minute to 60 calls per minute (Fig. 5d-e)," does the call rate refer to individual animals or the entire cage?

"This implies that a high rate of calls in a room can interrupt animals during social resting states and cause them to preferentially exhibit more active/attentive states." Does it? This could simply indicate that more active animals produce more calls.

"We recorded neural activity in the striatum because it is known to contain diverse signals related to movement and social interactions." While I understand that the authors intend to publish neural data separately, a brief discussion of the striatum's role here would be helpful.

Author response:

We would like to thank the editors and reviewers for taking the time to help improve our manuscript. We appreciate the feedback and will definitely increase the level of methodological detail in a revised submission.

Here is a brief summary of our plan to address the points raised by the reviewers. We will respond to the comments in a point-by-point manner when we resubmit a revised manuscript.

Reviewer 1

This reviewer raised a question about the 60 Hz frame rate for recording. We agree that increasing the number of cameras and frame rate would improve the tracking quality, but this would come at the cost of scalability. In the current study (and other concurrent studies in the lab), we recorded from 10-20 families simultaneously to try to sample the distribution of behavioral responses to stimuli observed in animals in our colony. This was only possible logistically because of the lightweight equipment design allowing us to record data from animals without large disruptions to their home-cage environment.

One strategy for acquiring higher-resolution data is to build a small number of enclosures that are fully surrounded by cameras, and to cycle animals through these enclosures (1). However, this strategy limits throughput by reducing the number of animals per day that can be studied. If the size and cost of cameras and computers decreases in the future, then this recording strategy will be scalable to the whole-colony level. For our current study and analysis, we are limited by the resolution of our dataset. We do believe that our data (although not a perfect 3d reconstruction or an extremely high frame rate) is sufficient to label behavioral states with high accuracy. We will add a figure to more clearly show that behavioral state data can be accurately inferred from this imperfect data, which has also been recently highlighted by other groups (2).

Additionally, with recent progress in the application of deep learning to animal pose tracking, new models can infer 3d pose dynamics from 2d data (3) and leverage spatiotemporal structure to clean up noisy data (4). We believe that other groups will be able to use these types of approaches to extract much more value from this dataset. So, in summary, we do understand the concern related to reconstruction quality and will 1) more clearly define the usefulness of our current models, 2) release our data and code so that others can build upon it or repurpose it, and 3) plan future experiments with higher camera count and frame rate as permitted by logistical constraints. 

Reviewer 2

This reviewer asked for an increased level of methodological detail. We will try to address this in a few ways:

(1) Code and data sharing. We believe that many of the questions related to the methodology will be best answered by sharing the data and code directly. Because there is a large amount of code associated with this manuscript, it is impractical to list every step and every parameter in the paper. Along with our revised manuscript, we will make our data and code publicly available. That said, we will improve our description of key parameters in the paper as the reviewer suggested.

(2) More detailed Methods section. The reviewer asked us to provide more methodological detail. We understand that this is currently a weakness of our manuscript, and we will focus on addressing it. For instance, the reviewer rightly points out that we did not describe the motion watches used to generate the data in Figure S7. We will address this.

(3) Simplify the manuscript. The paper currently has 22 figures, and further analysis could be done based on the results shown in any of them. For instance, this reviewer asked us to add a comparison across females and males (similar to our comparison of juveniles and adults). While we plan to add that analysis, we recognize that there are several figures/panels that are not closely related to our intended goal of describing the patterns we found in our large dataset. We will simplify the manuscript by removing some excess figures/panels and focus on describing the parts of the analysis that are crucial to our conclusions in greater detail.

(4) More careful language. This reviewer pointed out that there were some inaccuracies with our descriptive language. For instance, we used the term "natural" behavior to describe the behavior of animals in captivity, which may more accurately be described as their home-cage behavior. We will be more careful to align our language to the standard for the field. For instance, several studies refer to unrestrained behavior in a laboratory setting as "spontaneous" behavior rather than "natural" behavior (5). In our case, the data consists of both spontaneously occurring behavior and responses to a set of stimuli. We will make sure that the descriptions are more precise in the revised manuscript.

(1) Bala, P. C. et al. Automated markerless pose estimation in freely moving macaques with OpenMonkeyStudio. Nat Commun 11, (2020).

(2) Weinreb, C. et al. Keypoint-MoSeq: parsing behavior by linking point tracking to pose dynamics. bioRxiv (2023) doi:10.1101/2023.03.16.532307.

(3) Gosztolai, A. et al. LiftPose3D, a deep learning-based approach for transforming two-dimensional to three-dimensional poses in laboratory animals. Nat Methods 18, 975–981 (2021).

(4) Wu, A. et al. Deep Graph Pose: a semi-supervised deep graphical model for improved animal pose tracking. Adv Neural Inf Process Syst 33, 6040–6052 (2020).

(5) Levy, D. R. et al. Mouse spontaneous behavior reflects individual variation rather than estrous state. Curr Biol 33, 1358-1364.e4 (2023).

eLife Assessment

This useful work identifies a key role for Tachykinin-1 parasubthalamic neurons in avoidance learning. At present, the evidence for the conclusions regarding fiber photometry, viral transfection, reporting of behavioral outcomes, and pathway-specificity is incomplete. This work will be of interest to neuroscientists studying neural mechanisms for avoidance and aversion.

Reviewer #1 (Public review):

This study is focused on a population of neurons in the mouse parasubthalamic nucleus (pSTN) that express Tackhykinin1 (Tac1). This gene has been used before to target pSTN for functional circuit studies because it is fairly selective for pSTN in this region, though it targets only a subset of pSTN neurons. Prior work has shown that activity in these neurons can impact motivated behaviors, including feeding and drinking behaviors, and that their activity is associated with aversion or avoidance behaviors. While not breaking much new ground, this study adds to that work by making use of a 2-way active avoidance assay, where a CS predicts a US (footshock), that the mice can escape. Using fiber photometry the authors show convincing evidence that Tac1 neurons in pSTN increase their activity in response to a US footshock, and that after some pairings the neurons will start responding to the CS too, though to a lesser extent than the US. Their most important data shows that either ablation or optogenetic inhibition of these cells can hugely block the active avoidance (escape) behavior, suggesting these neurons are key for the performance of this task, which they interpret as key for learning the task (but see more below). They show that optogenetic stimulation is aversive in a real-time place assay, and when paired with footshock can enhance active avoidance behavior. Finally, they show that Tac1 pSTN axons in PVT recapitulate these effects while showing that axons in CEA or PBN may only recapitulate some of these effects (more below). Overall I think the data is solid and shows that the activity of Tac1 pSTN neurons in the 2 way active avoidance task is causally related to avoidance behavior in the direction that would be predicted by recent literature. However, I think the authors overstate the conclusions in the title, abstract, and text. I do not think the data make a strong case for a role for these cells in learning, at least in any classical sense, as used in the title and abstract and elsewhere. Also the statement in the abstract that the pSTN mediates its effects 'differentially' through its downstream targets is not convincingly supported by data.

Major concerns:

(1) The authors infer that the activity in the Tac1 pSTN neurons is necessary for aversive or avoidance 'learning'. But this is not well defined, what exactly does that mean and what types of evidence would support or falsify such a hypothesis? Moreover, the authors show convincingly, and in line with prior reports, that these cells are activated by aversive stimuli (here footshock), and that activation of these cells is sufficient to induce avoidance behavior. Because manipulation of these cells can serve as a primary negative reinforcer, it becomes even more challenging and important to explain how experiments that manipulate these cells while measuring behavior/performance can discriminate between changes in: (1) primary aversion, (2) motivation to avoid, (3) associative learning, or (4) memory/retrieval. The authors seem to favor #3, but they don't make a clear case for this point of view or else what they mean by 'avoidance learning'. In my opinion, the data do not well discriminate between possibilities 1 through 3. The authors should clarify their logic and temper their conclusions throughout.

(2) Abstract line 37 is not well supported. The authors focus mostly on pSTN projections to PVT and show that the measurements or manipulation of these axons recapitulates the effects seen with pSTN cell bodies. The authors do fewer studies of axons in CeA and PBN, but do find that they can recapitulate the effects with opsin inhibition, but detect no effects with opsin stimulation. However, the lack of effect with opsin stimulation in Figure S7a-e proves very little on its own. It could be technical, due to inadequate expression or functional efficacy. It is not supported by histological and functional evidence that the manipulation was effective. Overall I can only conclude that the projections to these regions might be very similar (based on the inhibition data), or might be a little different. The data are thus inadequate to support the authors' claim that the pSTN mediates learning differentially through its downstream targets.

Other concerns:

(3) Line 93 is not adequately supported by data in Figure 1b. Additional data is needed that shows expression across cases, including any spread that may be visible when zooming out from pSTN. Additional methods are needed to indicate what exclusion criteria were applied and how many mice were excluded. These data could help support the statement on line 93 that expression was largely restricted within pSTN.

(4) From the results and methods it is not clear where the GFP signal would come from in the mice expressing Casp3 for the ablation studies. It is therefore not clear if the absence of GFP should be taken as evidence of cell loss. For example, it is not clear if multiple vectors were used, if volumes and titers were carefully matched between control groups, or if competition/occlusion between AAVs could be ruled out. It is also not clear how this was quantified, that is how many sections/subjects and how counting was done. It is not clear how long was waited between the AAV infusion, behavior, and euthanasia, perhaps especially important for the ablation done after avoidance learning occurred.

(5) The authors should consider showing individual measurements and not just mean/sem wherever feasible, for example, to support the statement on line 141 that 'all ablated mice showed...'.

(6) S3 is an important control for interpreting data in Figure 2d-i. Something similar is needed to support the inferences made in 2j-u. The very strong effect showing a lack of active avoidance in response to CS or the US when pSTN Tac1 neurons are inhibited during CS or during US suggests that something gross may be going on, such as a gross motor or sensory response that supersedes the effect of footshock. The authors do not comment on whether there are any gross behavioral responses to the inhibition, but an experiment as in S3 is needed, for example, to show that behavior is intact during pSTN inhibition if delivered after the mice already learned to associate CS with US.

(7) The authors use 100 shocks of 0.8 mA for 7 days. I think this is quite strong and in the pSTN inhibition experiments it seems to be functionally 'inescapable' and could thus produce behaviors similar to 'learned helplessness'. Can the authors consider whether this might contribute to the striking findings they observed in their opsin inhibition assays?

(8) The description of the experiment in S5 is inadequate. What are the adjacent areas? Where do the authors see spread? The use of the word 'case' in figure S5 implies an individual case, but the legend says 5 mice were used for 'case 1' and 3 mice were used for 'case 2'. The use of the word 'off-target in the figure implies that the expression was of the intended target. But the text of results and methods implies it was intentional targeting of unnamed and unshown adjacent regions. This should be clarified.

(9) The authors suggest the CPA study is divergent from Serra et al 2023. Though I think this could be due to how the conditioning was done, it would be helpful for the authors to include less processed data. This would aid in possible interpretations for any divergences across studies. Can the authors include raw data (in seconds of time spent) in each compartment for each group across baseline and test days?

Reviewer #2 (Public review):

Summary:

The manuscript by Hu et. al presents a clearly-designed examination of the role of tachykinin1-expressing neurons in the parasubthalamic nucleus of the lateral posterior hypothalamus (PTSN) in active avoidance learning. These glutamatergic neurons have previously been implicated in responding to negative stimuli. This manuscript expands the current understanding of PTSNTac1 neurons in learned responses to threats by showing their role in encoding and mediating the active avoidance response. The authors first use bulk fiber photometry imaging to show the encoding of the active avoidance procedure, followed by cell-type specific manipulations of PTSNTac1 neurons during active avoidance. Finally, they show that encoding and mediation of active avoidance in a downstream target of PTSNTac1 neurons, the PVT/intermediodorsal nuclei of the dorsal thalamus (IMD), has the same effect as what was discovered in the cell body. This contrasts other output regions of the PTSN, such as the PBN and CeA, which were not found to promote active avoidance learning. The experiments presented were well-designed to support the conclusions of the authors, however, the manuscript is missing several key control experiments and supplemental information to support their main findings.

Strengths:

The manuscript provides information on a brain region and downstream target that mediates active avoidance learning. The manuscript provides valuable information via necessity and sufficiency experiments to show the role of the population of interest (PTSNTac1 neurons) in active avoidance learning. The authors also performed most behavior experiments in male and female mice, with adequate power to address potential sex differences in the control of active avoidance by PTSNTac1 neurons. Finally, the manuscript provides valuable information about the specificity of the PTSNTac1 downstream target in regulating active avoidance learning, identifying the PVT/intermediodorsal nuclei of the dorsal thalamus as the key target and ruling out the PBN and CeA.

Weaknesses:

However, several main conclusions of the paper must be interpreted carefully due to missing or inadequate control experiments and histological verification.

(1) Inadequate presentation of viral localization. The authors state that expression was "largely restricted within PSTN" however there is no quantification of the amount of viral expression beyond the target region. Given that Tac1 is expressed in neighboring regions, it is critical to show the viral expression and fiber implant location data for all animals included in the figures. Furthermore, criteria for inclusion and exclusion based on mistargeting should be delineated. This should also be clearly outlined for the experiments in Figure S5, where "behavioral effects of activation of sparsely Tac1-expressing neurons in two adjacent areas of PSTN" was tested but the location of viral expression in those cases is unclear.

(2) Lack of motion artifact correction with isosbestic signal for GCamp recordings. It is appreciated that the authors included a separate EGFP-expressing group to compare to the GCamp-expressing group, however, additional explanation is required for the methods used to analyze the raw fluorescent signal. Namely, were fluorescent signals isosbestic-corrected prior to calculating ΔF/F? If no isosbestic signal was used to correct motion artifacts within a recording session, additional explanation is needed to explain how this was addressed. The lack of motion artifacts in the EGFP signal in a separate cohort is inadequate to answer this caveat as motion artifacts are within-animal.

(3) Missing control experiment demonstrating intact locomotor performance in caspase ablation experiments. The authors use caspase ablation of PTSNTac1 neurons prior to active avoidance learning to appraise the necessity of this cell population. However, a control experiment showing intact locomotor ability in ablated mice was not performed.

(4) Missing control experiment demonstrating [lack of] valence with PTSN silencing manipulations. The authors performed a real-time and conditioned place preference experiments for ChR2-expressing mice (Fig 3M) and found stimulation to be negatively-valenced and generate an aversive memory, respectively. Absent this control experiment with silencing, an alternative conclusion remains possible that optogenetic silencing via GtACR2 created nonspecific location preferences in the active avoidance apparatus, confounding the interpretation of those results.

(5) Incomplete analysis of sex differences. Data in female mice is conspicuously missing from inhibition experiments. The rationale for exclusion from this dataset would be useful for the interpretation of the other noted sex differences.

Reviewer #3 (Public review):

Summary:

This study by Hu et al. examined the role of tachykinin1 (Tac1)-expressing neurons in the para subthalamic nucleus (PSTH) in active avoidance of electric shocks. Bulk recording of PSTH Tac1 neurons or axons of these neurons in PVT showed activation of a shock-predicting tone and shock itself. Ablation of these neurons or optogenetic manipulation of these neurons or their projection to PVT suggests the causality of this pathway with the learning of active avoidance.

Strengths:

This work found an understudied pathway potentially important for active avoidance of electric shocks. Experiments were thoroughly done and the presentation is clear. The amount of discussion and references are appropriate.

Weaknesses:

Critical control experiments are missing for most experiments, and statistical tests are not clear or not appropriate in most parts. Details are shown below.

(1) There are some control experiments missing. Notably, optogenetic manipulation is not verified in any experiments. It is important to verify whether neural activation with optogenetic activation is at the physiological level or supra-physiological level, and whether optogenetic inhibition does not cause unwanted activity patterns such as rebound activation at the critical time window.

(2) Neural ablation with caspase was confirmed by GFP expression. However, from the present description, a different virus to express EITHER caspase or GFP was injected, and then the numbers of GFP-expressing neurons were compared. It is not clear how this can detect ablation.

(3) In many places, statistical approaches are not clear from the present figures, figure legends, and Methods. It seems that most statistics were performed by pooling trials, but it is not described, or multiple "n" are described. For example, it is explicitly mentioned in Figure 4H, "n = 3 mice, n = 213 avoidance trials and n = 87 failure trials". The authors should not pool trials, but should perform across-animal tests in this and other figures, and "n" for statistical tests should be clearly described in each plot.

(4) It is also unclear how the test types were selected. For example, in Figure 1K and O with similar datasets, one is examined by a paired test and the other is by an unpaired test. Since each animal has both early vs late trials, and avoidance vs failure trials, paired tests across animals should be performed for both.

(5) It is also strange to show violin plots for only 6 animals. They should instead show each dot for each animal, connected with a line to show consistent increases of activity in late vs early trials and avoidance vs failure trials.

(6) To tell specificity in avoidance learning, it is better to show escape in the current trials with optogenetic manipulation.

(7) For place aversion, % time decrease across days was tested. It is better to show the original number before normalization, as well.

(8) For anatomical results in Figure S6, it is important to show images with lower magnification, too.

(9) Inactivation of either pathway from PSTH to PBN or to CeA also inhibits active avoidance, but the authors conclude that these effects are "partial" compared to the inactivation of PSTH to PVT. It is not clear how the effects were compared since the effects of PSTH-CeA inactivation are quite strong, comparable to PSTH-PVT inactivation by eye. They should quantify the effects to conclude the difference.

(10) Supplementary table 1: as mentioned above, n for statistical tests should be clearer.

Author response:

Reviewer #1 (Public review):

This study is focused on a population of neurons in the mouse parasubthalamic nucleus (pSTN) that express Tackhykinin1 (Tac1). This gene has been used before to target pSTN for functional circuit studies because it is fairly selective for pSTN in this region, though it targets only a subset of pSTN neurons. Prior work has shown that activity in these neurons can impact motivated behaviors, including feeding and drinking behaviors, and that their activity is associated with aversion or avoidance behaviors. While not breaking much new ground, this study adds to that work by making use of a 2-way active avoidance assay, where a CS predicts a US (footshock), that the mice can escape. Using fiber photometry, the authors show convincing evidence that Tac1 neurons in pSTN increase their activity in response to a US footshock, and that after some pairings the neurons will start responding to the CS too, though to a lesser extent than the US. Their most important data shows that either ablation or optogenetic inhibition of these cells can hugely block the active avoidance (escape) behavior, suggesting these neurons are key for the performance of this task, which they interpret as key for learning the task (but see more below). They show that optogenetic stimulation is aversive in a real-time place assay, and when paired with footshock can enhance active avoidance behavior. Finally, they show that Tac1 pSTN axons in PVT recapitulate these effects while showing that axons in CEA or PBN may only recapitulate some of these effects (more below). Overall I think the data is solid and shows that the activity of Tac1 pSTN neurons in the 2 way active avoidance task is causally related to avoidance behavior in the direction that would be predicted by recent literature. However, I think the authors overstate the conclusions in the title, abstract, and text. I do not think the data make a strong case for a role for these cells in learning, at least in any classical sense, as used in the title and abstract and elsewhere. Also, the statement in the abstract that the pSTN mediates its effects 'differentially' through its downstream targets is not convincingly supported by data.

We are very pleased that Reviewer 1 thought our data is solid.

Major concerns:

(1) The authors infer that the activity in the Tac1 pSTN neurons is necessary for aversive or avoidance 'learning'. But this is not well defined, what exactly does that mean and what types of evidence would support or falsify such a hypothesis? Moreover, the authors show convincingly, and in line with prior reports, that these cells are activated by aversive stimuli (here footshock), and that activation of these cells is sufficient to induce avoidance behavior. Because manipulation of these cells can serve as a primary negative reinforcer, it becomes even more challenging and important to explain how experiments that manipulate these cells while measuring behavior/performance can discriminate between changes in: (1) primary aversion, (2) motivation to avoid, (3) associative learning, or (4) memory/retrieval. The authors seem to favor #3, but they don't make a clear case for this point of view or else what they mean by 'avoidance learning'. In my opinion, the data do not well discriminate between possibilities 1 through 3. The authors should clarify their logic and temper their conclusions throughout.

Thank you Reviewer 1 for providing us insightful suggestions. Based on our fiber photometry data that the activities of PSTN Tac1+ neurons show a significant increase in CS-evoked calcium fluorescent signals in late trials relative to those in early trials (Figure 1H-K) and our optogenetic inhibition experiments during CS (Figure 2N-Q), these results illustrate that the activities of PSTN Tac1+ neurons are modulated by learning and are required for active avoidance learning. Moreover, PSTN Tac1+ neurons are activated by footshock and activation of these cells is sufficient to induce avoidance behavior. These findings demonstrate that PSTN Tac1+ neurons encode aversive information. Together, our current data support that PSTN Tac1+ neurons encode both aversive event and its predicting cue. We will clarify our conclusions in the revised manuscript.

(2) Abstract line 37 is not well supported. The authors focus mostly on pSTN projections to PVT and show that the measurements or manipulation of these axons recapitulates the effects seen with pSTN cell bodies. The authors do fewer studies of axons in CeA and PBN, but do find that they can recapitulate the effects with opsin inhibition, but detect no effects with opsin stimulation. However, the lack of effect with opsin stimulation in Figure S7a-e proves very little on its own. It could be technical, due to inadequate expression or functional efficacy. It is not supported by histological and functional evidence that the manipulation was effective. Overall, I can only conclude that the projections to these regions might be very similar (based on the inhibition data), or might be a little different. The data are thus inadequate to support the authors' claim that the pSTN mediates learning differentially through its downstream targets.

In the revised version of manuscript, we will provide more histological and functional evidence for the PSTN-to-CeA and PSTN-to-PBN circuits to support our conclusion on the functional roles of these downstream targets. Similar with our anterograde experiment that the PSTN densely projects to CeA and PBN (Figure S6), optogenetic activation and inhibition experiments showed dense axonal terminals in the CeA and PBN from the PSTN and this line of data will be included in the revised manuscript. In addition, we will further examine these circuits by investigating the functional roles of CeA-projecting or PBN-Projecting PSTN neurons during 2-way active avoidance task.

Other concerns:

(3) Line 93 is not adequately supported by data in Figure 1b. Additional data is needed that shows expression across cases, including any spread that may be visible when zooming out from pSTN. Additional methods are needed to indicate what exclusion criteria were applied and how many mice were excluded. These data could help support the statement on line 93 that expression was largely restricted within pSTN.

In the revised version of manuscript, we will provide larger example images containing pSTN and its adjacent areas to demonstrate that the viral expression is well restricted into this brain area. Moreover, we will provide detailed information on the exclusion criteria and the number of mice excluded in the Method section.   

(4) From the results and methods it is not clear where the GFP signal would come from in the mice expressing Casp3 for the ablation studies. It is therefore not clear if the absence of GFP should be taken as evidence of cell loss. For example, it is not clear if multiple vectors were used, if volumes and titers were carefully matched between control groups, or if competition/occlusion between AAVs could be ruled out. It is also not clear how this was quantified, that is how many sections/subjects and how counting was done. It is not clear how long was waited between the AAV infusion, behavior, and euthanasia, perhaps especially important for the ablation done after avoidance learning occurred.

I totally agree with Reviewer 1’s concerns. We will perform immunohistochemistry or in situ hybridization for Tachykinin-1 itself and then measure colocalization of GFP with Tachykinin-1 inside and outside of the PTSN, and the degree of absence of Tachykinin-1 in Casp mice. In addition, we will provide more detailed experimental information in the revised manuscript.

(5) The authors should consider showing individual measurements and not just mean/sem wherever feasible, for example, to support the statement on line 141 that 'all ablated mice showed...'.

Thank you Reviewer 1 for this suggestion. We will re-plot the data as individual measurements in the revised manuscript.

(6) S3 is an important control for interpreting data in Figure 2d-i. Something similar is needed to support the inferences made in 2j-u. The very strong effect showing a lack of active avoidance in response to CS or the US when pSTN Tac1 neurons are inhibited during CS or during US suggests that something gross may be going on, such as a gross motor or sensory response that supersedes the effect of footshock. The authors do not comment on whether there are any gross behavioral responses to the inhibition, but an experiment as in S3 is needed, for example, to show that behavior is intact during pSTN inhibition if delivered after the mice already learned to associate CS with US.

Thank you Reviewer 1 for this insightful suggestion. During the review process, we have performed this line of experiment as in Figure S3. We measured the behavioral responses during pSTN optogenetic inhibition after the mice already learned to associate CS with US and found most GtACR-expressing mice showed unaffected avoidance learning. This data will be included in the revised manuscript.

(7) The authors use 100 shocks of 0.8 mA for 7 days. I think this is quite strong and in the pSTN inhibition experiments it seems to be functionally 'inescapable' and could thus produce behaviors similar to 'learned helplessness'. Can the authors consider whether this might contribute to the striking findings they observed in their opsin inhibition assays?

I agree with the Reviewer 1’s comment on the string findings in the optogenetic inhibition results. Indeed, based on the results on days 1 and 2, optogenetic inhibition of PSTN tac1+ neurons has significantly blocked GtACR-expressing animals’ behavioral performance during 2-way active avoidance task. To examine whether the effect by optogenetic inhibition of these neurons could possibly decline with prolonged training, we conducted additional 5-day training. We will discuss and add this comment in the revised manuscript.

(8) The description of the experiment in S5 is inadequate. What are the adjacent areas? Where do the authors see spread? The use of the word 'case' in figure S5 implies an individual case, but the legend says 5 mice were used for 'case 1' and 3 mice were used for 'case 2'. The use of the word 'off-target in the figure implies that the expression was of the intended target. But the text of results and methods implies it was intentional targeting of unnamed and unshown adjacent regions. This should be clarified.

We will add histological images and clarify these comments in the revised manuscript. The purpose of this experiment is to illustrate that even slightly spreading ChR2 viruses into Tac1+ neurons of the adjacent areas of the PSTN did not result in behavioral changes and this will indirectly support the main behavioral function caused by the PSTN tac1+ neurons rather than its neighboring areas. Because Tac1+ neurons outside the PSTN are sparsely expressed, it is quite difficult to completely restrict the viral expression in the PSTN from the anterior to the posterior. Thus, we will provide detailed information on the exclusion criteria and the number of mice excluded in the Method section.   

(9) The authors suggest the CPA study is divergent from Serra et al 2023. Though I think this could be due to how the conditioning was done, it would be helpful for the authors to include less processed data. This would aid in possible interpretations for any divergences across studies. Can the authors include raw data (in seconds of time spent) in each compartment for each group across baseline and test days?

We will follow Reviewer 1’s suggestion to include raw data (in seconds of time spent) in each compartment for each group across baseline and test days in the revised manuscript.

Reviewer #2 (Public review):

Summary:

The manuscript by Hu et. al presents a clearly-designed examination of the role of tachykinin1-expressing neurons in the parasubthalamic nucleus of the lateral posterior hypothalamus (PTSN) in active avoidance learning. These glutamatergic neurons have previously been implicated in responding to negative stimuli. This manuscript expands the current understanding of PTSNTac1 neurons in learned responses to threats by showing their role in encoding and mediating the active avoidance response. The authors first use bulk fiber photometry imaging to show the encoding of the active avoidance procedure, followed by cell-type specific manipulations of PTSNTac1 neurons during active avoidance. Finally, they show that encoding and mediation of active avoidance in a downstream target of PTSNTac1 neurons, the PVT/intermediodorsal nuclei of the dorsal thalamus (IMD), has the same effect as what was discovered in the cell body. This contrasts other output regions of the PTSN, such as the PBN and CeA, which were not found to promote active avoidance learning. The experiments presented were well-designed to support the conclusions of the authors, however, the manuscript is missing several key control experiments and supplemental information to support their main findings.

Strengths:

The manuscript provides information on a brain region and downstream target that mediates active avoidance learning. The manuscript provides valuable information via necessity and sufficiency experiments to show the role of the population of interest (PTSNTac1 neurons) in active avoidance learning. The authors also performed most behavior experiments in male and female mice, with adequate power to address potential sex differences in the control of active avoidance by PTSNTac1 neurons. Finally, the manuscript provides valuable information about the specificity of the PTSNTac1 downstream target in regulating active avoidance learning, identifying the PVT/intermediodorsal nuclei of the dorsal thalamus as the key target and ruling out the PBN and CeA.

We highly appreciate that Reviewer 2 thought that our experiments presented were well-designed to support the conclusions and provided valuable information in several aspects.

Weaknesses:

However, several main conclusions of the paper must be interpreted carefully due to missing or inadequate control experiments and histological verification.

(1) Inadequate presentation of viral localization. The authors state that expression was "largely restricted within PSTN" however there is no quantification of the amount of viral expression beyond the target region. Given that Tac1 is expressed in neighboring regions, it is critical to show the viral expression and fiber implant location data for all animals included in the figures. Furthermore, criteria for inclusion and exclusion based on mistargeting should be delineated. This should also be clearly outlined for the experiments in Figure S5, where "behavioral effects of activation of sparsely Tac1-expressing neurons in two adjacent areas of PSTN" was tested but the location of viral expression in those cases is unclear.

Similar with questions 3 and 8 of Reviewer 1. We will provide the viral expression and fiber implant location data for all animals included in the figures and histological images in Figure S5 in the revised manuscript. Moreover, we will provide detailed information on the exclusion criteria and the number of mice excluded in the Method section.  

2) Lack of motion artifact correction with isosbestic signal for GCamp recordings. It is appreciated that the authors included a separate EGFP-expressing group to compare to the GCamp-expressing group, however, additional explanation is required for the methods used to analyze the raw fluorescent signal. Namely, were fluorescent signals isosbestic-corrected prior to calculating ΔF/F? If no isosbestic signal was used to correct motion artifacts within a recording session, additional explanation is needed to explain how this was addressed. The lack of motion artifacts in the EGFP signal in a separate cohort is inadequate to answer this caveat as motion artifacts are within-animal.

We will follow Reviewer 2’s suggestion and perform isosbestic-correction for fluorescent signals prior to calculating ΔF/F. We will re-plot related figures and add this information in the revised manuscript.

(3) Missing control experiment demonstrating intact locomotor performance in caspase ablation experiments. The authors use caspase ablation of PTSNTac1 neurons prior to active avoidance learning to appraise the necessity of this cell population. However, a control experiment showing intact locomotor ability in ablated mice was not performed.

We will follow Reviewer 2’s suggestion to perform a control experiment showing intact locomotor ability in caspase 3-ablated mice and will include this data in the revised manuscript.

(4) Missing control experiment demonstrating [lack of] valence with PTSN silencing manipulations. The authors performed a real-time and conditioned place preference experiments for ChR2-expressing mice (Fig 3M) and found stimulation to be negatively-valenced and generate an aversive memory, respectively. Absent this control experiment with silencing, an alternative conclusion remains possible that optogenetic silencing via GtACR2 created nonspecific location preferences in the active avoidance apparatus, confounding the interpretation of those results.

Thank you Reviewer 2 for this useful suggestion. We will examine the valence with PTSN silencing manipulations by using a RTPP test and add this data in the revised manuscript.

(5) Incomplete analysis of sex differences. Data in female mice is conspicuously missing from inhibition experiments. The rationale for exclusion from this dataset would be useful for the interpretation of the other noted sex differences.

Thank you Reviewer 2 for this useful suggestion. During the review process, we have performed ablation and inhibition experiments in females, demonstrating similar behavioral effects as those in males. We will add these data in the revised manuscript.

Reviewer #3 (Public review):

Summary:

This study by Hu et al. examined the role of tachykinin1 (Tac1)-expressing neurons in the para subthalamic nucleus (PSTH) in active avoidance of electric shocks. Bulk recording of PSTH Tac1 neurons or axons of these neurons in PVT showed activation of a shock-predicting tone and shock itself. Ablation of these neurons or optogenetic manipulation of these neurons or their projection to PVT suggests the causality of this pathway with the learning of active avoidance.

Strengths:

This work found an understudied pathway potentially important for active avoidance of electric shocks. Experiments were thoroughly done and the presentation is clear. The amount of discussion and references are appropriate.

We are very pleased to have Reviewer 3’s positive comments on the manuscript.

Weaknesses:

Critical control experiments are missing for most experiments, and statistical tests are not clear or not appropriate in most parts. Details are shown below.

(1) There are some control experiments missing. Notably, optogenetic manipulation is not verified in any experiments. It is important to verify whether neural activation with optogenetic activation is at the physiological level or supra-physiological level, and whether optogenetic inhibition does not cause unwanted activity patterns such as rebound activation at the critical time window.

Thank you Reviewer 3 for this useful suggestion. We will perform in vitro slice recording experiments to verify optogenetic manipulations and add this line of evidence in the revised manuscript.

(2) Neural ablation with caspase was confirmed by GFP expression. However, from the present description, a different virus to express EITHER caspase or GFP was injected, and then the numbers of GFP-expressing neurons were compared. It is not clear how this can detect ablation.

Similar with question 4 of Reviewer 1. We will perform immunohistochemistry or in situ hybridization for Tachykinin-1 itself and then measure colocalization of GFP with Tachykinin-1 inside and outside of the PTSN, and the degree of absence of Tachykinin-1 in Casp-ablated mice. In addition, we will provide more detailed experimental information in the revised manuscript.

(3) In many places, statistical approaches are not clear from the present figures, figure legends, and Methods. It seems that most statistics were performed by pooling trials, but it is not described, or multiple "n" are described. For example, it is explicitly mentioned in Figure 4H, "n = 3 mice, n = 213 avoidance trials and n = 87 failure trials". The authors should not pool trials, but should perform across-animal tests in this and other figures, and "n" for should be clearly described in each plot.

We have provided all statistical information in the Supplementary Table 1. In the revised manuscript, we will perform across-animal tests, re-plot new figures and provide clear statistical information.

(4) It is also unclear how the test types were selected. For example, in Figure 1K and O with similar datasets, one is examined by a paired test and the other is by an unpaired test. Since each animal has both early vs late trials, and avoidance vs failure trials, paired tests across animals should be performed for both.

Following Reviewer 3’s suggestion, we will perform across-animal tests. In the first version of our manuscript, for fiber photometry experiments, we pooled trial data of each animal and performed statistics tests across trials. Because avoidance and failure trials were different, we thus selected an unpaired test for this kind of dataset.

(5) It is also strange to show violin plots for only 6 animals. They should instead show each dot for each animal, connected with a line to show consistent increases of activity in late vs early trials and avoidance vs failure trials.

Similar with question 4 of Reviewer 3, we pooled trial data of each animal and performed statistics tests across trials. We will perform across-animal tests and re-plot figures by connecting with a line to show consistent increases of activity in late vs early trials and avoidance vs failure trials for each animal.

(6) To tell specificity in avoidance learning, it is better to show escape in the current trials with optogenetic manipulation.

Thank you Reviewer 3 for this useful suggestion. We will follow this suggestion and add this analysis in the revised manuscript.

(7) For place aversion, % time decrease across days was tested. It is better to show the original number before normalization, as well.

Similar with question 9 of Reviewer 1, we will show the original number before normalization in the revised manuscript.

(8) For anatomical results in Figure S6, it is important to show images with lower magnification, too.

We will follow this suggestion and provide histological images with lower magnification in the revised manuscript.

(9) Inactivation of either pathway from PSTH to PBN or to CeA also inhibits active avoidance, but the authors conclude that these effects are "partial" compared to the inactivation of PSTH to PVT. It is not clear how the effects were compared since the effects of PSTH-CeA inactivation are quite strong, comparable to PSTH-PVT inactivation by eye. They should quantify the effects to conclude the difference.

We will quantify the effects of different downstream targets of the PSTN to make a precise conclusion.

(10) Supplementary table 1: as mentioned above, n for statistical tests should be clearer.

As mentioned above, we will perform across-animal tests and provide clear statistical information in the figure legends and supplementary table 1.

eLife Assessment

This study uses a large dataset from both recent isolates and genomes in databases to provide an important analysis of the population structure of the pathogen Salmonella gallinarum. The authors present convincing results regarding the regional adaptation and the evolutionary trajectory of the resistome and mobilome, even though some issues regarding the genomic analysis could be improved. This work will interest microbiologists and researchers working on genomics, evolution, and antimicrobial resistance.

Reviewer #1 (Public review):

Summary:

The investigators in this study analyzed the dataset assembly from 540 Salmonella isolates, and those from 45 recent isolates from Zhejiang University of China. The analysis and comparison of the resistome and mobilome of these isolates identified a significantly higher rate of cross-region dissemination compared to localized propagation. This study highlights the key role of the resistome in driving the transition and evolutionary history of S. Gallinarum.

Strengths:

The isolates included in this study were from 16 countries in the past century (1920 to 2023). While the study uses S. Gallinarun as the prototype, the conclusion from this work will likely apply to other Salmonella serotypes and other pathogens.

Weaknesses:

While the isolates came from 16 countries, most strains in this study were originally from China.

Comments on revisions:

This reviewer is happy with the detailed responses from the authors regarding revising this manuscript. I do not have further comments.

Reviewer #2 (Public review):

Summary:

The authors sequence 45 new samples of S. Gallinarum, a commensal Salmonella found in chickens, which can sometimes cause disease. They combine these sequences with around 500 from public databases, determine the population structure of the pathogen, and coarse relationships of lineages with geography. The authors further investigate known anti-microbial genes found in these genomes, how they associate with each other, whether they have been horizontally transferred, and date the emergence of clades.

Strengths:

- It doesn't seem that much is known about this serovar, so publicly available new sequences from a high burden region are a valuable addition to the literature.<br /> - Combining these sequences with publicly available sequences is a good way to better contextualise any findings.<br /> - The genomic analyses have been greatly improved since the first version of the manuscript, and appropriately analyse the population and date emergence of clades.<br /> - The SNP thresholds are contextualised in terms of evolutionary time.<br /> - The importance and context of the findings are fairly well described.

Weaknesses:

- There are still a few issues with the genomic analyses, although they no longer undermine the main conclusions:

(1) Although the SNP distance is now considered in terms of time, the 5 SNP distance presented still represents ~7yrs evolution, so it is unlikely to be a transmission event, as described. It would be better to use a much lower threshold or describe the interpretation of these clusters more clearly. Bringing in epidemiological evidence or external references on the likely time interval between transmissions would be helpful.

(2) The HGT definition has not fundamentally been changed and therefore still has some issues, mainly that vertical evolution is still not systematically controlled for. Using a 5kb window is not sufficient, as LD may extend across the entire genome. As the authors have now run gubbins correctly, they could use the results from this existing analysis to find recent HGT. To definite mobilisation, perhaps a standard pipeline such (e.g. https://github.com/EBI-Metagenomics/mobilome-annotation-pipeline) would be more convincing.

(3) The invasiveness index is better described, but the authors still did not provide convincing evidence that the small difference is actually biologically meaningful (there was no statistical difference between the two strains provided in response Figure 6). What do other Salmonella papers using this approach find, and can their links be brought in? If there is still no good evidence, a better description of this difference would help make the conclusions better supported.

In summary, the analysis is broadly well described and feels appropriate. Some of the conclusions are still not fully supported, although the main points and context of the paper now appear sound.

Author response:

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

Public Reviews: 

Reviewer #1 (Public review): 

Summary: 

The investigators in this study analyzed the dataset assembly from 540 Salmonella isolates, and those from 45 recent isolates from Zhejiang University of China. The analysis and comparison of the resistome and mobilome of these isolates identified a significantly higher rate of cross-region dissemination compared to localized propagation. This study highlights the key role of the resistome in driving the transition and evolutionary 

Thank you for summarizing our work. According to your comments, we carefully considered and responded to them and made corresponding revisions to the text. Additionally, to fully contextualize the background knowledge and clarify the major points in this study, we add some references.

Upon further review of our initial manuscript, we realized that the original submission did not strictly follow the lineage order proposed by Zhou et al. (Natl Sci Rev. 2023 Sep 2;10(10):nwad228). To avoid confusion and keep the uniform knowledge in the typing system, we have adjusted the lineage nomenclature along the revised manuscript to reflect the corrected order as follows:

Author response table 1.

To ensure consistency with previous studies, we have revised the nomenclature for the different lineages of bvSP.

Strengths: 

The isolates included in this study were from 16 countries in the past century (1920 to 2023). While the study uses S. Gallinarun as the prototype, the conclusion from this work will likely apply to other Salmonella serotypes and other pathogens. 

Thanks for the constructive comments and the positive reception of the manuscript.

Weaknesses: 

While the isolates came from 16 countries, most strains in this study were originally from China. 

We appreciate the reviewer's observation regarding the sampling distribution of isolates in this study. We acknowledge that while the isolates were collected from 15 different countries, with a significant proportion originated from China (Author response image 1). This focus is due to several reasons:

Author response image 1.

Geographic distribution of 580 S. Gallinarum. Different colors indicate the countries of origin for the 580 S. Gallinarum strains in the dataset. Darker shades represent higher numbers of strains.

(1) As once a globally prevalent pathogen across the 20th century, S. Gallinarum was listed by the World Organization for Animal Health (WOAH) due to its economic importance. After 30 years of implementation of the National Poultry Improvement Plan in the US, it was almost eradicated in high-income countries, and interestingly, it became an endemic pathogen with sporadic outbreaks in most low- or middle-income countries like China and Brazil. Given the vast expanse of China's land area and the country's economic factors, implementing the same measures remains challenging.  

(2) S. Gallinarum is an avian-specific pathogen, particularly affecting chickens, and its distribution is closely linked to chicken meat production in different countries. There are more frequent reports of fowl typhoid in some high chicken-producing developing countries. Data from the United States Department of Agriculture (USDA) on annual chicken meat production for 2023/2024 show that the global distribution of S. Gallinarum aligns closely with the overall chicken meat production of these countries (https://fas.usda.gov/data/production/commodity/0115000).

Author response image 2.

The United States Department of Agriculture (USDA) data on annual chicken meat production for 2023/2024 across different countries globally.

(3) Our primary objective was to investigate the localized resistome adaptation of S. Gallinarum in regions. Being a region with significant disease burden, China has reported numerous outbreaks (Sci Data. 2022 Aug 13;9(1):495; Sci Data. 2024 Feb 27;11(1):244) and a high AMR prevalence of this serovar (Natl Sci Rev. 2023 Sep 2;10(10):nwad228; mSystems. 2023 Dec 21;8(6):e0088323), making it an excellent example for understanding localized resistance mechanisms.

(4) As China is the primary country of origin for the strains in this study, it is necessary to ensure that the strains from China are consistent with the local geographic characteristics of the country. Therefore, we conducted a correlation analysis between the number of strains from different provinces in China and the total GDP/population size of those provinces (Author response image 3). The results show that most points fall within the 95% confidence interval of the regression line. Although some points exhibit relative unbalance in the number of S. Gallinarum strains, most data points for these regions have a small sample size (n < 15). Overall, we found that the prevalence of S. Gallinarum in different regions of China is consistent with the overall nationwide trend.

Author response image 3.

Correlation analysis between the number of S. Gallinarum collected from different provinces in China and the total GDP/population size. The figure depicts a series of points representing individual provinces. The x-axis indicates the number of S. Gallinarum included in the dataset, while the y-axis displays the values for total GDP and total population size, respectively.

Nevertheless, a search of nearly a decade of literature on PubMed and a summary of the S. Gallinarum genome available on public databases indicate that the dataset used is the most complete. Furthermore, focusing on a specific region within China allowed us to conduct a detailed and thorough analysis. However, we highly agree that expanding the study to include more isolates from other countries would enhance the generalizability of our findings, and we are actively collecting additional S. Gallinarum genome data. In the revised manuscript, we have further emphasized the limitations as follow:

Lines 427-429: “However, the current study has some limitations. Firstly, despite assembling the most comprehensive WGS database for S. Gallinarum from public and laboratory sources, there are still biases in the examined collection. The majority (438/580) of S. Gallinarum samples were collected from China, possibly since the WGS is a technology that only became widely available in the 21st century. This makes it impractical to sequence it on a large scale in the 20th century, when S. Gallinarum caused a global pandemic. So, we suspect that human intervention in the development of this epidemic is the main driving force behind the fact that most of the strains in the data set originated in China. In our future work, we aim to actively gather more data to minimize potential biases within our dataset, thereby improving the robustness and generalizability of our findings.”

Reviewer #2 (Public review): 

Summary: 

The authors sequence 45 new samples of S. Gallinarum, a commensal Salmonella found in chickens, which can sometimes cause disease. They combine these sequences with around 500 from public databases, determine the population structure of the pathogen, and coarse relationships of lineages with geography. The authors further investigate known anti-microbial genes found in these genomes, how they associate with each other, whether they have been horizontally transferred, and date the emergence of clades. 

Thank you for your constructive suggestions, which are valuable and highly beneficial for improving our paper. According to your comments, we carefully considered and responded to them and made corresponding revisions to the text. Furthermore, to fully contextualize the background knowledge and clarify the major points in this study, we add some references to support our findings and policy implications.

Upon further review of our initial manuscript, we realized that the original submission did not strictly follow the lineage order proposed by Zhou et al. (Natl Sci Rev. 2023 Sep 2;10(10):nwad228). To avoid confusion in the typing system, we have adjusted the lineage nomenclature in the revised manuscript to reflect the corrected order (see Author response table 1).

Strengths: 

(1) It doesn't seem that much is known about this serovar, so publicly available new sequences from a high-burden region are a valuable addition to the literature. 

(2) Combining these sequences with publicly available sequences is a good way to better contextualise any findings. 

Thank you so much for your thorough review and constructive comments on the manuscript.

Weaknesses: 

There are many issues with the genomic analysis that undermine the conclusions, the major ones I identified being: 

(1) Recombination removal using gubbins was not presented fully anywhere. In this diversity of species, it is usually impossible to remove recombination in this way. A phylogeny with genetic scale and the gubbins results is needed. Critically, results on timing the emergence (fig2) depend on this, and cannot be trusted given the data presented. 

We sincerely thank you for pointing out this issue. In the original manuscript, we aimed to present different lineages of S. Gallinarum within a single phylogenetic tree constructed using BEAST. However, in the revised manuscript, we have addressed this issue by applying the approach recommended by Gubbins to remove recombination events for each lineage defined by FastBAPs. Additionally, to better illustrate the removal of recombination regions in the genome, we have included a figure generated by Gubbins (New Supplementary Figure 12). 

Our results indicate that recombination events are relatively infrequent in Lineage 1, followed by Lineage 3, but occur more frequently in Lineage 2. In the revised manuscript, we have included additional descriptions in the Methods section to clarify this analysis. We hope these modifications adequately address the reviewer’s concerns and enhance the trustworthiness of our findings.

(2) The use of BEAST was also only briefly presented, but is the basis of a major conclusion of the paper. Plot S3 (root-to-tip regression) is unconvincing as a basis of this data fitting a molecular clock model. We would need more information on this analysis, including convergence and credible intervals. 

Thank you very much for raising this issue. We decided to reconduct separate BEAST analyses for each lineage, accurately presenting the evolutionary scale based on the abovementioned improvements. The implementation of individual lineage for BEAST analysis was conducted based on the following steps:

(1) Using R51 as the reference, a reference-mapped multiple core-genome SNP sequence alignment was created, and recombination regions were detected and removed as described above.

(2) TreeTime was used to assess the temporal structure by performing a regression analysis of the root-to-tip branch distances within the maximum likelihood tree, considering the sampling date as a variable (New Supplementary Figures 6). However, the root-to-tip regression analysis presented in New Supplementary Figures 6 was not intended as a basis for selecting the best molecular clock model; its purpose was to clean the dataset with appropriate measurements.

(3) To determine the optimal model for running BEAST, we tested a total of six combinations in the initial phase of our study. These combinations included the strict clock, relaxed lognormal clock, and three population models (Bayesian SkyGrid, Bayesian Skyline, and Constant Size). Before conducting the complete BEAST analysis, we evaluated each combination using a Markov Chain Monte Carlo (MCMC) analysis with a total chain length of 100 million and sampling every 10,000 iterations. We then summarized the results using NSLogAnalyser and determined the optimal model based on the marginal likelihood value for each combination. The results indicated that the model incorporating the Bayesian Skyline and the relaxed lognormal clock yielded the highest marginal likelihood value in our sample. Then, we proceeded to perform a timecalibrated Bayesian phylogenetic inference analysis for each lineage. The following settings were configured: the "GTR" substitution model, “4 gamma categories”, the "Relaxed Clock Log Normal" model, the "Coalescent Bayesian Skyline" tree prior, and an MCMC chain length of 100 million, with sampling every 10,000 iterations.

(4) Convergence was assessed using Tracer, with all parameter effective sampling sizes (ESS) exceeding 200. Maximum clade credibility trees were generated using TreeAnnotator. Finally, key divergence time points (with 95% credible intervals) were estimated, and the tree was visualized using FigTree. 

For the key lineages, L2b and L3b (carrying the resistome, posing antimicrobial resistance (AMR) risks, and exhibiting intercontinental transmission events), we have redrawn Figure 2 based on the updated BEAST analysis results (New Figure 2). For L1, L2a, and L3c, we have added supplementary figures to provide a more detailed visualization of their respective BEAST analysis outcomes (New Supplementary Figures 3-5). The revised BEAST analysis indicates that the origin of L3b in China can be traced back to as early as 1683 (95% CI: 1608 to 1839). In contrast, the earliest possible origin of L2b in China dates back to 1880 (95% CI: 1838 to 1902). This indicates that the previous manuscript's assumption that L2b is an older lineage compared to L3b may be inaccurate. 

Furthermore, In the revised manuscript, we specifically estimated the time points for the first intercontinental transmission events for the two major lineages, L2b and L3b. Our results indicate that L2b, likely underwent two major intercontinental transmission events. The first occurred around 1893 (95% CI: 1870 to 1918), with transmission from China to South America. The second major transmission event occurred in 1923 (95% CI: 1907 to 1940), involving the spread from South America to Europe. In contrast, the transmission pattern of L3b appears relatively more straightforward. Our findings show that L3b, an S. Gallinarum lineage originating in China, only underwent one intercontinental transmission event from China to Europe, likely occurring around 1790 (95% CI: 1661 to 1890) (New Supplementary Figure 7). Based on the more critical BEAST analysis for each lineage, we have revised the corresponding conclusions in the manuscript. We believe that the updated BEAST analysis, performed using a more accurate recombination removal approach, significantly enhances the rigor and credibility of our findings.

(3) Using a distance of 100 SNPs for a transmission is completely arbitrary. This would at least need to be justified in terms of the evolutionary rate and serial interval. 

Using single nucleotide polymorphism (SNP) distance to trace pathogen transmission is a common approach (J Infect Dis. 2015 Apr 1;211(7):1154-63) and in our previous studies (hLife 2024; 2(5):246-256. mLife 2024; 3(1):156-160.). When the SNP distance within a cluster falls below a set threshold, the strains in that cluster are considered to have a potential direct transmission link. It is generally accepted that the lower the threshold, the more stringent the screening process becomes. However, there is little agreement in the literature regarding what such a threshold should be, and the appropriate SNP cut-off for inferring transmission likely depends critically on the context (Mol Biol Evol. 2019 Mar 1;36(3):587-603).

In this study, we compared various thresholds (SNPs = 5, 10, 20, 25, 30, 35, 40, 50, 100) to ensure clustering in an appropriate manner. First, we summarized the tracing results under each threshold (Author response image 4), which demonstrated that, regardless of the threshold used, all strains associated with transmission events originated from the same location (New Figure 3a).

Author response image 4.

Clustering results of 45 newly isolated S. Gallinarum strains using different SNP thresholds of 5, 10, 15, 20, 25, 28, 30, 50, and 100 SNPs. The nine subplots represent the clustering results under each threshold. Each point corresponds to an individual strain, and lines connect strains with potential transmission relationships.

In response to your comments regarding the evolutionary rate, we estimated the overall evolutionary rate of the S. Gallinarum using BEAST. We applied the methodology described by Arthur W. Pightling et al. (Front Microbiol. 2022 Jun 16; 13:797997). The numbers of SNPs per year were determined by multiplying the evolutionary rates estimated with BEAST by the number of core SNP sites identified in the alignments. We hypothesize that a slower evolutionary rate in bacteria typically requires a lower SNP threshold when tracing transmission events using SNP distance analysis. Pightling et al.'s previous research found an average evolutionary rate of 1.97 SNPs per year (95% HPD, 0.48 to 4.61) across 22 different Salmonella serotypes. Our updated BEAST estimation for the evolutionary rate of S. Gallinarum suggests it is approximately 0.74 SNPs per year (95% HPD, 0.42 to 1.06). Based on these findings, and our previous experience with similar studies (mBio. 2023 Oct 31;14(5):e0133323.), we set a threshold of 5 SNPs in the revised manuscript.

Then, we adopted the newly established SNP distance threshold (n=5) to update Figure 3a and New Supplementary Figure 8. The heatmap on the far right of New Figure 3a illustrates the SNP distances among 45 newly isolated S. Gallinarum strains from two locations in Zhejiang Province (Taishun and Yueqing). New Supplementary Figure 8 simulates potential transmission events between the bvSP strains isolated from Zhejiang Province (n=95) and those from China with available provincial information (n=435). These analyses collectively demonstrate the localized transmission pattern of bvSP within China. Our analysis using the newly established SNP threshold indicates that the 45 strains isolated from Taishun and Yueqing exhibit a highly localized transmission pattern, with pairs of strains exhibiting potential transmission events below the set threshold occurring exclusively within a single location. Subsequently, we conducted the SNP distance-based tracing analysis for the 95 strains from Zhejiang Province and those from China with available provincial information (n=435) (New Supplementary Figure 8, New Supplementary Table S8). Under the SNP distance threshold (n=5), we identified a total of 91 potential transmission events, all of which occurred exclusively within Zhejiang Province. No inter-provincial transmission events were detected. Based on these findings, we revised the methods and conclusions in the manuscript accordingly. We believe that the updated version well addresses your concerns.

Nevertheless, the final revised and updated results do not change the conclusions presented in our original manuscript. Instead, applying a more stringent SNP distance threshold allows us to provide solid evidence supporting the localized transmission pattern of S. Gallinarum in China. 

(4) The HGT definition is non-standard, and phylogeny (vertical inheritance) is not controlled for.  

The cited method: 

'In this study, potentially recently transferred ARGs were defined as those with perfect identity (more than 99% nucleotide identity and 100% coverage) in distinct plasmids in distinct host bacteria using BLASTn (E-value {less than or equal to}10−5)' 

This clearly does not apply here, as the application of distinct hosts and plasmids cannot be used. Subsequent analysis using this method is likely invalid, and some of it (e.g. Figure 6c) is statistically very poor. 

Thank you for raising this important question. In our study, Horizontal Gene Transfer (HGT) is defined as the transfer of genetic information between different organisms, a process that facilitates the spread of antibiotic resistance genes (ARGs) among bacteria. This definition of HGT is consistent with that used in previous studies (Evol Med Public Health. 2015; 2015(1):193–194; ISME J. 2024 Jan 8;18(1):wrad032). In Salmonella, the transfer of antimicrobial resistance genes via HGT is not solely dependent on plasmids; other mobile genetic elements (MGEs), such as transposons, integrons, and prophages, also play significant roles. This has also  been documented in our previous work (mSystems. 2023 Dec 21;8(6):e0088323). Given the involvement of various MGEs in the horizontal transfer of ARGs, we propose that the criteria for evaluating horizontal transfer via plasmids can also be applied to ARGs mediated by other MGEs.

In this study, we adopted stricter criteria than those used by Xiaolong Wang et al. Specifically, we defined two ARGs as identical only if they exhibited 100% nucleotide identity and 100% coverage. To address concerns regarding the potential influence of vertical inheritance in our analysis, we have made the following improvements. In the revised manuscript, we provide a more detailed table that includes the co-localization analysis of each ARG with mobile genetic elements (New Supplementary Table 9). For prophages and plasmids, we required that ARGs be located directly within these elements. In contrast, for transposons and integrons, we considered ARGs to be associated if they were located within a 5 kb region upstream or downstream of these elements (Nucleic Acids Res. 2022 Jul 5;50(W1):W768-W773). 

In the revised manuscript, we first categorized a total of 621 ARGs carried by 436 bvSP isolates collected in China according to the aforementioned criteria and found that 415 ARGs were located on MGEs. After excluding the ARGs not associated with MGEs, we recalculated the overall HGT frequency of 10 types of ARGs in China, the horizontal ARGs transfer frequency in three key regions, and the horizontal ARGs transfer frequency within a single region (New Supplementary Table 7). Based on the results, we updated relevant sections of the manuscript and remade Figure 6. The updated manuscript describes the results of this section as follows:

“Horizontal transfer of resistome occurs widely in localized bvSP

Horizontal transfer of the resistome facilitates the acquisition of AMR among bacteria, which may record the distinct acquisition event in the bacterial genome. To compare these events in a geographic manner, we further investigated the HGT frequency of each ARG carried by bvSP isolated from China and explored the HGT frequency of resistome between three defined regions. Potentially horizontally transferred ARGs were defined as those with perfect identity (100% identity and 100% coverage) and were located on MGEs across different strains (Fig. 6a). We first categorized a total of 621 ARGs carried by 436 bvSP isolates collected in China and found that 415 ARGs were located on MGEs. After excluding the ARGs not associated with MGEs, our findings reveal that horizontal gene transfer of ARGs is widespread among Chinese bvSP isolates, with an overall transfer rate of 92%. Specifically, 50% of the ARGs exhibited an HGT frequency of 100%, indicating that these ARGs might underwent extensive frequent horizontal transfer events (Fig. 6b). It is noteworthy that certain resistance genes, such as tet(A), aph(3'')-Ib, and aph(6)-Id, appear to be less susceptible to horizontal transfer.

However, different regions generally exhibited a considerable difference in resistome HGT frequency. Overall, bvSP from the southern areas in China showed the highest HGT frequency (HGT frequency=95%). The HGT frequencies for bvSP within the eastern and northern regions of China are lower, at 92% and 91%, respectively (Fig. 6c). For specifical ARG type, we found tet(A) is more prone to horizontal transfer in the southern region, and this proportion was considerably lower in the eastern region. Interestingly, certain ARGs such as aph(6)-Id, undergo horizontal transfer only within the eastern and northern regions of China (Fig. 6d). Notably, as a localized transmission pathogen, resistome carried by bvSP exhibited a dynamic potential among inter-regional and local demographic transmission, especially from northern region to southern region (HGT frequency=93%) (Fig. 6e, Supplementary Table 7).”

We also modified the current version of the pipeline used to calculat the HGT frequency of resistance genes. In the revised pipeline, users are required to provide a file specifying the locations of mobilome on the genome before formally calculating the HGT frequency of the target ARGs. The specific code and data used in the calculation have been uploaded to https://github.com/tjiaa/Cal_HGT_Frequency.

However, we also acknowledge that the current in silico method has some limitations. This approach heavily relies heavily on prior information in existing resistome/mobilome databases. Additionally, the characteristics of second-generation sequencing data make it challenging to locate gene positions precisely. Using complete genome assemblies might be a crucial approach to address this issue effectively. In the revised manuscript, we have also provided a more detailed explanation of the implications of the current pipeline.

Regarding your second concern, "some of it (e.g., Figure 6c) is statistically very poor," the horizontal ARG transfer frequency calculation for each region was based on the proportion of horizontal transfer events of ARGs in that region to the total possible transfer events. As a result, we are unable to calculate the statistical significance between the two regions. Our aim with this approach is to provide a rough estimate of the extent of horizontal ARG transfer within the S. Gallinarum population in each region. In future studies, we will refine our conclusions by developing a broader range of evaluation methods to ensure more comprehensive assessment and validation.

(5) Associations between lineages, resistome, mobilome, etc do not control for the effect of genetic background/phylogeny. So e.g. the claim 'the resistome also demonstrated a lineage-preferential distribution' is not well-supported. 

Thank you for your comments. We acknowledge that the associations between lineages and the mobilome/resistome may be influenced by the genetic background or phylogeny of the strains. For instance, our conclusion regarding the lineage-preferential distribution of the resistome was primarily based on New Figure 4a, where L3 is clearly shown to carry the most ARGs. Furthermore, we observed that L3b tends to harbor bla_{_TEM-1B}, _sul2, and tet(A) more frequently than other lineages. However, we recognize that this evidence is insufficient to support a definitive conclusion of “demonstrated a lineage-preferential distribution”. Therefore, we have re-examined the current manuscript and described these findings as a potential association between the mobilome/resistome and lineages.

(6) The invasiveness index is not well described, and the difference in means is not biologically convincing as although it appears significant, it is very small. 

Thank you for pointing this out. For the invasiveness index mentioned in the manuscript, we used the method described in previous studies. (PLoS Genet. 2018 May 8;14(5), Nat Microbiol. 2021 Mar;6(3):327-338). Specifically, Salmonella’s ability to cause intestinal or extraintestinal infections in hosts is related to the degree of genome degradation. We evaluated the potential for extraintestinal infection by 45 newly isolated S. Gallinarum strains (L2b and L3b) using a model that quantitatively assesses genome degradation. We analyzed samples using the 196 top predictor genes, employing a machine-learning approach that utilizes a random forest classifier and delta-bitscore functional variant-calling. This method evaluated the invasiveness of S. Gallinarum towards the host, and the distribution of invasiveness index values for each region was statistically tested using unpaired t-test. The code used for calculating the invasiveness index is available at https://github.com/Gardner-BinfLab/invasive_salmonella. In the revised manuscript, we added a more detailed description of the invasiveness index calculation in the Methods section as follows:

Lines 592-603: “Specifically, Salmonella’s ability to cause intestinal or extraintestinal infections in hosts is related to the degree of genome degradation. We evaluated the potential for extraintestinal infection by 45 newly isolated S. Gallinarum strains (L2b and L3b) using a model that quantitatively assesses genome degradation. We analyzed each sample using the 196 top predictor genes for measuring the invasiveness of S. Gallinarum, employing a machine-learning approach that utilizes a random forest classifier and deltabitscore functional variant-calling. This method evaluated the invasiveness of S. Gallinarum towards the host, and the distribution of invasiveness index values for each region was statistically tested using unpaired t-test. The code used for calculating the invasiveness index is available at: https://github.com/Gardner-BinfLab/invasive_salmonella.”

Regarding the second question, 'the difference in means is not biologically convincing as although it appears significant, it is very small,' we believe that this difference is biologically meaningful. In our previous work, we infected chicken embryos with different lineages of S. Gallinarum (Natl Sci Rev. 2023 Sep 2;10(10):nwad228). The virulence of thirteen strains of Salmonella Gallinarum, comprising five from lineage L2b and eight from lineage L3b, was evaluated in 16-day-old SPF chicken embryos through inoculation into the allantoic cavity. Controls included embryos that inoculated with phosphate-buffered saline (PBS). The embryos were incubated in a thermostatic incubator maintained at 37.5°C with a relative humidity ranging from 50% to 60%. Prior to inoculation, the viability of the embryos was assessed by examining the integrity of their venous system and their movements; any dead embryos were excluded from the study. Overnight cultures resuspended in PBS at a concentration of 1000 CFU per 100 μL were administered to the embryos. Mortality was recorded daily for a period of five days, concluding upon the hatching of the chicks. 

It is generally accepted that strains with higher invasive capabilities are more likely to cause chicken embryo mortality. Our experimental results showed that the L2b, which exhibits higher invasiveness, with a slightly higher to cause chicken embryo death (Author response image 5). 

Author response image 5.

The survival curves of chicken embryos infected with bvSP isolates from S. Gallinarum L2b and S. Gallinarum L3b. Inoculation with Phosphate Buffer Saline (PBS) were considered controls. 

(7) 'In more detail, both the resistome and mobilome exhibited a steady decline until the 1980s, followed by a consistent increase from the 1980s to the 2010s. However, after the 2010s, a subsequent decrease was identified.' 

Where is the data/plot to support this? Is it a significant change? Is this due to sampling or phylogenetics? 

Thank you for highlighting these critical points. The description in this statement is based on New Supplementary Figure 11. On the right side of New Supplementary Figure 11, we presented the average number of Antimicrobial Resistance Genes (ARGs) and Mobile Genetic Elements (MGEs) carried by S. Gallinarum isolates from different years, and we described the overall trend across these years. However, we realized that this statement might overinterpret the data. Given that this sentence does not impact our emphasis on the overall increasing trends observed in the resistome and mobilome, as well as their potential association, we decided to remove it in the revised manuscript.

The revised paragraph would read as follows:

Lines 261-268: “Variations in regional antimicrobial use may result in uneven pressure for selecting AMR. The mobilome is considered the primary reservoir for spreading resistome, and a consistent trend between the resistome and the mobilome has been observed across different lineages, from L1-L3c. We observed an overall gradual rise in the resistome quantity carried by bvSP across various lineages, correlating with the total mobilome content (S11 Fig). Furthermore, we investigated the interplay between particular mobile elements and resistome types in bvSP.”

(8) It is not clear what the burden of disease this pathogen causes in the population, or how significant it is to agricultural policy. The article claims to 'provide valuable insights for targeted policy interventions.', but no such interventions are described. 

Thank you for your constructive suggestions. Salmonella Gallinarum is an avian-specific pathogen that induces fowl typhoid, a severe systemic disease characterized by high mortality rates in chickens, thereby posing a significant threat to the poultry industry, particularly in developing countries (Rev Sci Tech. 2000 Aug;19(2):40524). In our previous research, we conducted a comprehensive meta-analysis of 201 publications encompassing over 900 million samples to investigate the global impact of S. Gallinarum (Sci Data. 2022 Aug 13;9(1):495). Our findings estimated that the global prevalence of S. Gallinarum is 8.54% (with a 95% confidence interval of 8.43% to 8.65%), with notable regional variations in incidence rates.

Our previously analysis focused on the prevalence of S. Gallinarum (including biovars SP and SG) across six continents. The results revealed that all continents, except Oceania, exhibited positive prevalences of S. Gallinarum. Asia had the highest prevalence at 17.31%, closely followed by Europe at 16.03%. In Asia, the prevalence of biovar SP was higher than that of biovar SG, whereas in Europe, biovar SG was observed to be approximately two hundred times more prevalent than biovar SP. In South America, the prevalence of S. Gallinarum was higher than that of biovar SP, at 10.06% and 13.20% respectively. Conversely, the prevalence of S. Gallinarum was relatively lower in North America (4.45%) compared to Africa (1.10%) (Author response image 6).

Given the significant economic losses caused by S. Gallinarum to the poultry industry and the potential risk of escalating antimicrobial resistance, more targeted policy interventions are urgently needed. Further elaboration on this implication is provided in the revised “Discussion” section as follows:

Lines 401-416: “In summary, the findings of this study highlight that S. Gallinarum remains a significant concern in developing countries, particularly in China. Compared to other regions, S. Gallinarum in China poses a notably higher risk of AMR, necessitating the development of additional therapies, i.e. vaccine, probiotics, bacteriophage therapy in response to the government's policy aimed at reducing antimicrobial use ( J Infect Dev Ctries. 2014 Feb 13;8(2):129-36). Furthermore, given the dynamic nature of S. Gallinarum risks across different regions, it is crucial to prioritize continuous monitoring in key areas, particularly in China's southern regions where the extensive poultry farming is located. Lastly, from a One-Health perspective, controlling AMR in S. Gallinarum should not solely focus on local farming environments, with improved overall welfare on poultry and farming style. The breeding pyramid of industrialized poultry production should be targeted on the top, with enhanced and accurate detection techniques (mSphere. 2024 Jul 30;9(7):e0036224). More importantly, comprehensive efforts should be made to reduce antimicrobial usage overall and mitigate potential AMR transmission from environmental sources or other hosts (Vaccines (Basel). 2024 Sep 18;12(9):1067; Vaccines (Basel). 2023 Apr 18;11(4):865; Front Immunol. 2022 Aug 11:13:973224).”

Author response image 6.

A comparison of the global prevalence of S. gallinarum across continents.

(9) The abstract mentions stepwise evolution as a main aim, but no results refer to this. 

Thank you for raising this issue. In the revised manuscript, we have changed “stepwise evolution” to simply “evolution” to ensure a more accurate and precise description.

(10) The authors attribute changes in population dynamics to normalisation in China-EU relations and hen fever. However, even if the date is correct, this is not a strongly supported causal claim, as many other reasons are also possible (for example other industrial processes which may have changed during this period). 

Thank you for raising this critical issue. In the revised manuscript, we conducted a more stringent BEAST analysis for each lineage, as described earlier. This led to some changes in the inferred evolutionary timelines. Consequently, we have removed the corresponding statement from the “Results” section. Instead, we now only provide a discussion of historical events, supported by literature, that could have facilitated the intercontinental spread of L2b and L3b in the “Discussion” section. We believe these revisions have made the manuscript more rigorous and precise.

Lines 332-342: “_The biovar types of _S. Gallinarum have been well-defined as bvSP, bvSG, and bvSD historically ( J Vet Med B Infect Dis Vet Public Health. 2005 Jun;52(5):2148). Among these, bvSP can be further subdivided into five lineages (L1, L2a, L2b, L3b, and L3c) using hierarchical Bayesian analysis. Different sublineages exhibited preferential geographic distribution, with L2b and L3b of bvSP being predominant global lineage types with a high risk of AMR. The historical geographical transmission was verified using a spatiotemporal Bayesian framework. The result shows that L3b was initially spread from China to Europe in the 18^th-19^th century, which may be associated with the European hen fever event in the mid-19th century (Burnham GP. 1855. The history of the hen fever: a humorous record). L2b, on the other hand, appears to have spread to Europe via South America, potentially contributing to the prevalence of bvSP in the United States.”  

(11) No acknowledgment of potential undersampling outside of China is made, for example, 'Notably, all bvSP isolates from Asia were exclusively found in China, which can be manually divided into three distinct regions (southern, eastern, and northern).'.

Perhaps we just haven't looked in other places?

We appreciate the reviewer's observation regarding the sampling distribution of isolates in this study. We acknowledge that while the isolates were collected from 15 different countries with, a significant proportion originated from China (Author response image 1). This focus is due to several reasons:

(1) As once a globally prevalent pathogen across the 20th century, S. Gallinarum was listed by the World Organization for Animal Health (WOAH) due to its economic importance. After 30 years of implementation the National Poultry Improvement Plan in the US, it was almost eradicated in high-income countries, and interestingly, it became an endemic pathogen with sporadic outbreaks in most low- or middle-income countries like China and Brazil. Given the vast expanse of China's land area and the country's economic factors, implementing the same measures remains a challenging endeavour. 

(2) S. Gallinarum is an avian-specific pathogen, particularly affecting chickens, and its distribution is closely linked to chicken meat production in different countries. In some high chicken-producing developing countries, such as China and Brazil, there are more frequent reports of fowl typhoid. Data from the United States Department of Agriculture (USDA) on annual chicken meat production for 2023/2024 show that the global distribution of S. Gallinarum aligns closely with the overall chicken meat production of these countries (https://fas.usda.gov/data/production/commodity/0115000).  

(3) Our primary objective was to investigate the localized resistome adaptation of S. Gallinarum in regions. Being a region with significant disease burden, China has reported numerous outbreaks (Sci Data. 2022 Aug 13;9(1):495; Sci Data. 2024 Feb 27;11(1):244) and a high AMR prevalence of this serovar (Natl Sci Rev. 2023 Sep 2;10(10):nwad228; mSystems. 2023 Dec 21;8(6):e0088323), making it an excellent example for understanding localized resistance mechanisms. 

Nevertheless, a search of nearly a decade of literature on PubMed and a summary of the S. Gallinarum genome available on public databases indicate that the dataset used is the most complete. Furthermore, focusing on a specific region within China allowed us to conduct a detailed and thorough analysis. However, we highly agree that expanding the study to include more isolates from other countries would enhance the generalizability of our findings, and we are actively collecting additional S. Gallinarum genome data. In the revised manuscript, we modified this sentence to indicate that this phenomenon is only observed in the current dataset, thereby avoiding an overly absolute statement:

Lines 131-135: “For the bvSP strains from Asia included in our dataset, we found that all originated from China. To further investigate the distribution of bvSP across different regions in China, we categorized them into three distinct regions: southern, eastern, and northern (Supplementary Table 3)”.

(12) Many of the conclusions are highly speculative and not supported by the data. 

Thank you for your comment. We have carefully revised the manuscript to address your concerns. We hope that the changes made in the revised version meet your expectations and provide a clearer and more accurate interpretation of our findings.

(13) The figures are not always the best presentation of the data: 

a. Stacked bar plots in Figure 1 are hard to interpret, the total numbers need to be shown.

Panel C conveys little information. 

b. Figure 4B: stacked bars are hard to read and do not show totals. 

c. Figure 5 has no obvious interpretation or significance. 

Thank you for your comments. We have revised the figures to improve the clarity and presentation of the data.

In summary, the quality of analysis is poor and likely flawed (although there is not always enough information on methods present to confidently assess this or provide recommendations for how it might be improved). So, the stated conclusions are not supported. 

Thank you for your valuable feedback. We have carefully revised the manuscript to address your concerns. We hope that the updated figures and tables, and new data in the revised version meet your expectations and provide more appropriate interpretation of our findings.

Recommendations for the authors:  

Reviewer #1 (Recommendations for the authors): 

This reviewer enjoyed reading this well-written manuscript. The authors are encouraged to address the following comments and revise the manuscript accordingly. 

(1) Title: The authors use avian-restrict Salmonella to refer to Salmonella Gallinarum. Please consider using Salmonella Gallinarum in the title. Also, your analysis relates to resistome and mobilome. Would it make sense to add mobilome in the manuscript? 

Thank you for your guidance. In the revised manuscript, we have changed the title to “Avian-specific Salmonella enterica Serovar Gallinarum transition to endemicity is accompanied by localized resistome and mobilome interaction”. We believe that this revised title more accurately reflects the content of our study.

(2) Abstract: This study uses 45 isolates from your labs. However, you failed to include these 45 isolates in the Abstract. Also, please clarify the sources of these isolates (from dead chickens, or dead chicken embryos? You wrote in two different ways in this manuscript). Also, I am not entirely convinced how the results from these 45 isolates will support the overall conclusion of this work. 

Thank you for your thorough review and constructive comments on the manuscript. In the revised version, we have added a description of 45 newly isolated S. Gallinarum strains in the Abstract to provide readers with a clearer understanding of the dataset used in this study.

Lines 36-41: “Using the most comprehensive whole-genome sequencing dataset of Salmonella enterica serovar Gallinarum (S. Gallinarum) collected from 16 countries, including 45 newly recovered samples from two related local regions, we established the relationship among avian-specific pathogen genetic profiles and localization patterns.”

Furthermore, the newly isolated S. Gallinarum strains were obtained from dead chicken embryos. We think your second concern may arise from the following description in the manuscript: “All 734 samples of dead chicken embryos were collected from Taishun and Yueqing in Zhejiang Province, China. After the thorough autopsy, the liver, intestines, and spleen were extracted and added separately into 2 mL centrifuge tubes containing 1 mL PBS. The organs were then homogenized by grinding.” In fact, all the collected dead chicken embryos were aged 19 to 20 days. At this developmental stage, collecting the liver, intestines, and spleen for isolation and cultivation of S. Gallinarum is possible. To avoid any confusion, we have included a more detailed description of the dead chicken embryos in the revised manuscript as follows:

Lines 447-451: “All 734 samples of dead chicken embryos aged 19 to 20 days were collected from Taishun and Yueqing in Zhejiang Province, China. After a thorough autopsy, the liver, intestines, and spleen were extracted and added separately into 2 mL centrifuge tubes containing 1 mL PBS. The organs were then homogenized by grinding.”

Regarding your concern about the statement, “I am not entirely convinced how the results from these 45 isolates will support the overall conclusion of this work,” we would like to clarify the significance of these new isolates. Our research first identified distinct characteristics in the 45 newly isolated S. Gallinarum strains from Taishun and Yueqing, Zhejiang Province. Specifically, we found that most of the strains from Yueqing belonged to sequence type ST92, whereas the majority from Taishun were ST3717. Additionally, there were significant differences between these geographically close strains in terms of SNP distance and predicted invasion capabilities. These findings suggest that S. Gallinarum may exhibit localized transmission patterns, which forms the basis of the scientific question and hypothesis we originally aimed to address. Furthermore, in our previous work, we collected 325 S. Gallinarum strains. By incorporating the newly isolated 45 strains, we aim to provide a more comprehensive view of the population diversity, transmission pattern and potential risk of S. Gallinarum. We will continue to endeavour to understand the global genomic and population diversity in this field.

Finally, we revised the sentences that could potentially raise concerns for readers: 

Lines 175-177: “To investigate the dissemination pattern of bvSP in China, we obtained forty-five newly isolated bvSP from 734 samples (6.1% overall isolation rate) collected from diseased chickens at two farms in Yueqing and Taishun, Zhejiang Province.”  >  “To investigate the dissemination pattern of bvSP, we obtained forty-five newly isolated bvSP from 734 samples (6.1% overall isolation rate) collected from diseased chickens at two farms in Yueqing and Taishun, Zhejiang Province.”

(3) The manuscript uses nomenclature and classification into different sublineages. Did the authors establish the approaches for defining these sublineages in this group or did you follow the accepted standards? 

Thank you very much for raising this important issue. The biovar types of Salmonella Gallinarum have historically been well-defined as S. Gallinarum biovar

Pullorum (bvSP), S. Gallinarum biovar Gallinarum (bvSG), and S. Gallinarum biovar Duisburg (bvSD) (J Vet Med B Infect Dis Vet Public Health. 2005 Jun;52(5):214-8). However, there seems to be no widespread consensus on the population nomenclature for the key biovar bvSP. In a previous study, Zhou et al. classified bvSP into six lineages:

L1, L2a, L2b, L3a, L3b, and L3c (Natl Sci Rev. 2023 Sep 2;10(10):nwad228). However, our more comprehensive analysis of S. Gallinarum using a larger dataset and hierarchical Bayesian clustering revealed that L3a, previously considered a distinct lineage, is actually a sublineage of L3c. Upon further review of our initial manuscript, we realized that the original submission did not strictly follow the lineage order proposed by Zhou et al. To avoid confusion in the typing system, we have adjusted the lineage nomenclature in the revised manuscript to reflect the corrected order (see Author response table 1).

(4) This reviewer is convinced with the analysis approaches and conclusion of this work.

In the meantime, the authors are encouraged to discuss the application of the conclusion of this study: a) can the data be somehow used in the prediction model? b) would the conclusion from S. Gallinarum have generalized application values for other pathogens. 

Thank you for your constructive comments on the manuscript. 

a) can the data be somehow used in the prediction model?

We believe that genomic data can be effectively used for constructing prediction models; however, the success of such models largely depends on the specific traits being predicted. In this study, we utilized a random forest prediction model based on 196 top genes (PLoS Genet. 2018 May 8;14(5)) to predict the invasiveness of 45 newly isolated strains. In relation to the antimicrobial resistance (AMR) issue discussed in this paper, we also conducted relevant analyses. For instance, we explored the use of image-based models to predict whether a genome is resistant to specific antibiotics (Comput Struct Biotechnol J. 2023 Dec 29:23:559-565). We are confident that the incorporation of newly generated data will facilitate the development of future predictive models, and we plan to pursue further research in this area.

b) would the conclusion from S. Gallinarum have generalized application values for other pathogens.

This might be explained from two perspectives. First, the key role of the mobilome in facilitating the spread of the resistome, as emphasized in this study, has also been confirmed in research on other pathogens (mBio. 2024 Oct 16;15(10):e0242824). Thus, we believe that the pipeline we developed to assess the horizontal transfer frequency of different resistance genes across regions applies to various pathogens. On the other hand, due to distinct evolutionary histories, different pathogens exhibit varying levels of adaptation to their environments. In this study, we found that S. Gallinarum tends to spread highly localized; however, this conclusion may not necessarily hold for other pathogens.

Reviewer #2 (Recommendations for the authors): 

The authors would need to: 

(1) Address my concerns about genomic analyses listed in the public review. 

Thank you for your valuable feedback. We have carefully reviewed your concerns and made the necessary revisions to address the points raised about genomic analyses in the public review. We sincerely hope that these modifications meet your expectations and provide more robust analysis. We appreciate your thoughtful input and remain open to further suggestions to improve the manuscript.

(2) Add more detail on the genomic methods and their outputs, as suggested above. 

We have added further details to clarify the methodologies and outputs as mentioned above. Specifically, we expanded the description of the data processing, and the bioinformatic tools used for analysis. To ensure clarity, we also included an expanded discussion of the key outputs, highlighting their implications. We hope these revisions meet your expectations.

(3) Critically rewrite their introduction to make it clear what problem they are trying to address. 

Thank you for your guidance. In the revised manuscript, we have made the necessary modifications to the Introduction section to more clearly articulate the problem we aim to address.

(4) Critically rewrite their conclusions so they are supported by the data they present, and make it clear when claims are more speculative. 

Thank you for your guidance. In the revised manuscript, we have made the recommended modifications to the relevant sections of the conclusion as outlined above.

More minor issues I identified: 

(1) Typo in the title 'avian-restrict'. 

Done.

Line 1: “Avian-specific Salmonella enterica Serovar Gallinarum transition to endemicity is accompanied by localized resistome and mobilome interaction.”

(2) 'By utilizing the pipeline we developed' -- a pipeline has not been introduced at this point. 

In the revised manuscript, we have removed this section from the 'Abstract'.

Lines 46-48: “Notably, the mobilome-resistome combination among distinct lineages exhibits a geographical-specific manner, further supporting a localized endemic mobilome-driven process.”

(3) 'has more than 90% serovars' -- doesn't make sense. 

Revised.

Lines 82-83: “Salmonella, a pathogen with distinct geographical characteristics, has more than 90% of its serovars frequently categorized as geo-serotypes.”

(4) 'horrific mortality rates that remain a disproportionate burden'. 

Revised.

Lines 83-87: “Among the thousands of geo-serotypes, Salmonella enterica Serovar Gallinarum (S. Gallinarum) is an avian-specific pathogen that causes severe mortality, with particularly detrimental effects on the poultry industry in low- and middle-income countries.”

(5) What is the rate, what is a comparison, how is it disproportionate? 

Thank you for your valuable feedback. It is challenging to accurately estimate the specific prevalence of S. Gallinarum, particularly due to the lack of comprehensive data in many countries. Numerous cases likely go unreported. However, S. Gallinarum is more commonly detected in low- and middle-income countries. Here, we provide three evidence supporting this observation. First, in our previous research, we conducted a comprehensive meta-analysis of 201 studies, involving over 900 million samples, to evaluate the global impact of S. Gallinarum (Sci Data. 2022 Aug 13;9(1):495). The estimated prevalence in 17 countries showed that Bangladesh had the highest rate (25.75%) of S. Gallinarum infections. However, for biovar Pullorum (bvSP), Argentina (20.69%) and China (18.18%) reported the highest prevalence rates. Second, previous studies have also reported that S. Gallinarum predominantly occurs in low- and middleincome countries (Vet Microbiol. 2019 Jan:228:165-172; BMC Microbiol. 2024 Oct 18;24(1):414). Finally, S. Gallinarum was once a globally prevalent pathogen in the 20th century. Following the implementation of eradication programs in most high-income countries, it was listed by the World Organization for Animal Health and subsequently became an endemic pathogen with sporadic outbreaks. However, similar eradication efforts are challenging to implement in low- and middle-income countries, leading to a disproportionately higher incidence of S. Gallinarum in these regions.

In the revised manuscript, we have rephrased this sentence to enhance its accuracy:

Lines 83-87: “Among the thousands of geo-serotypes, Salmonella enterica serovar Gallinarum (S. Gallinarum) is an avian-specific pathogen that causes severe mortality, with particularly detrimental effects on the poultry industry in low- and middle-income countries.”

(6) 'we collected the most comprehensive set of 580 S. Gallinarum isolates', -> 'we collected the most comprehensive set S. Gallinarum isolates, consisting of 580 genomes'. 

Revised.

Lines 97-100: “To fill the gaps in understanding the evolution of S. Gallinarum under regional-associated AMR pressures and its adaptation to endemicity, we collected the most comprehensive set S. Gallinarum isolates, consisting of 580 genomes, spanning the period from 1920 to 2023.” 

(7) Sequence reads are not available, and use a non-standard database. The eLife policy states: 'Sequence reads and assembly must be included for reference genomes, while novel short sequences, including epitopes, functional domains, genetic markers and haplotypes should be deposited, together with surrounding sequences, into Genbank, DNA Data Bank of Japan (DDBJ), or EMBL Nucleotide Sequence Database (ENA). DNA and RNA sequencing data should be deposited in NCBI Trace Archive or NCBI Sequence Read Archive (SRA).' So the sequences assemblies and reads should ideally be mirrored appropriately. 

Thank you for your valuable suggestion regarding submitting the genome data for the newly isolated 45 S. Gallinarum strains. The genome data have been deposited in the NCBI Sequence Read Archive (SRA) under two BioProjects. The “SRA Accession number” for each strain have been added to New Supplementary Table 1. We believe this will ensure that the data are more readily accessible to a broader audience of researchers for download and analysis. We have revised the corresponding paragraph in the manuscript as follows:

Lines 606-608: “For the newly isolated 45 strains of Salmonella Gallinarum, genome data have been deposited in NCBI Sequence Read Archive (SRA) database. The “SRA Accession” for each strain are listed in Supplementary Table 1.”

(8) You should state at the start of the results which data is public, and how much is newly sequenced. 

Revised.

Lines 109-112: “To understand the global geographic distribution and genetic relationships of S. Gallinarum, we assembled the most comprehensive S. Gallinarum WGS dataset (n=580), comprising 535 publicly available genomes and 45 newly sequenced genomes.”

eLife Assessment

This valuable study tackles the well-established overflow metabolism issue by applying a coarse-grained metabolic flux model to predict how individual cells execute various energy strategies, such as respiration versus fermentation. The model's population average is convincing enough to align with experimental observations on overflow metabolism. However, the theoretical framework's reliance on single-cell growth rate variability must be questioned because of insufficient correlation with fluxes and the absence of regulatory mechanisms, highlighting the need for single-cell experimental validation to substantiate the proposed model.

Reviewer #1 (Public review):

Summary:

Cell metabolism exhibits a well-known behavior in fast-growing cells, which employ seemingly wasteful fermentation to generate energy even in the presence of sufficient environmental oxygen. This phenomenon is known as Overflow Metabolism or the Warburg effect in cancer. It is present in a wide range of organisms, from bacteria and fungi to mammalian cells.

In this work, starting with a metabolic network for Escherichia coli based on sets of carbon sources, and using a corresponding coarse-grained model, the author applies some well-based approximations from the literature and algebraic manipulations. These are used to successfully explain the origins of Overflow Metabolism, both qualitatively and quantitatively, by comparing the results with E. coli experimental data.

By modeling the proteome energy efficiencies for respiration and fermentation, the study shows that these parameters are dependent on the carbon source quality constants K_i (p.115 and 116). It is demonstrated that as the environment becomes richer, the optimal solution for proteome energy efficiency shifts from respiration to fermentation. This shift occurs at a critical parameter value K_A(C).<br /> This counter intuitive results qualitativelly explains Overflow Metabolism.

Quantitative agreement is achieved through the analysis of the heterogeneity of the metabolic status within a cell population. By introducing heterogeneity, the critical growth rate is assumed to follow a Gaussian distribution over the cell population, resulting in accordance with experimental data for E. coli. Overflow metabolism is explained by considering optimal protein allocation and cell heterogeneity.

The obtained model is extensively tested through perturbations: 1) Introduction of overexpression of useless proteins; 2) Studying energy dissipation; 3) Analysis of the impact of translation inhibition with different sub-lethal doses of chloramphenicol on Escherichia coli; 4) Alteration of nutrient categories of carbon sources using pyruvate. All model perturbations results are corroborated by E. coli experimental results.

Strengths:

In this work, the author effectively uses modeling techniques typical of Physics to address complex problems in Biology, demonstrating the potential of interdisciplinary approaches to yield novel insights. The use of Escherichia coli as a model organism ensures that the assumptions and approximations are well-supported in existing literature. The model is convincingly constructed and aligns well with experimental data, lending credibility to the findings. In this version, the extension of results from bacteria to yeast and cancer is substantiated by a literature base, suggesting that these findings may have broad implications for understanding diverse biological systems.

Weaknesses:

The author explores the generalization of their results from bacteria to cancer cells and yeast, adapting the metabolic network and coarse-grained model accordingly. In previous version this generalization was not completedly supported by references and data from the literature. This drawback, however, has been treated in this current version, where the authors discuss in much more detail and give references supporting this generalization.

Reviewer #2 (Public review):

In this version of manuscript, the author clarified many details and rewrote some sections. This substantially improved the readability of the paper. I also recognized that the author spent substantial efforts in the Appendix to answer the potential questions.

Unfortunately, I am not currently convinced by the theory proposed in this paper. In the next section, I will first recap the logic of the author and explain why I am not convinced. Although the theory fits many experimental results, other theories on overflow metabolism are also supported by experiments. Hence, I do not think based on experimental data we could rule in or rule out different theories.

Recap: To explain the origin of overflow metabolism, the author uses the following logic:

(1) There is a substantial variability of single-cell growth rate<br /> (2) The flux (J_r^E) and (J_f^E) are coupled with growth rate by Eq. 3<br /> (3) Since growth rate varies from cells to cells, flux (J_r^E) and (J_f^E) also varies<br /> (4) The variabilities of above fluxes in above create threshold-analog relation, and hence overflow metabolism.

My opinion:

The logic step (2) and (3) have caveats. The variability of growth rate has large components of cellular noise and external noise. Therefore, variability of growth rate is far from 100% correlated with variability of flux (J_r^E) and (J_f^E) at the single-cell level. Single-cell growth rate is a complex, multivariate functional, including (Jr^E) and (J_f^E) but also many other variables. My feeling is the correlation could be too low to support the logic here.

One example: ribosomal concentration is known to be an important factor of growth rate in bulk culture. However, the "growth law" from bulk culture cannot directly translate into the growth law at single-cell level [Ref1,2]. This is likely due to other factors (such as cell aging, other muti-stability of cellular states) are involved.

Therefore, I think using Eq.3 to invert the distribution of growth rate into the distribution of (Jr^E) and (J_f^E) is inapplicable, due to the potentially low correlation at single-cell level. It may show partial correlations, but may not be strong enough to support the claim and create fermentation at macroscopic scale.

Overall, if we track the logic flow, this theory implies overflow metabolism is originated from variability of k_cat of catalytic enzymes from cells to cells. That is, the author proposed that overflow metabolism happens macroscopically as if it is some "aberrant activation of fermentation pathway" at the single-cell level, due to some unknown partially correlation from growth rate variability.

Compared with other theories, this theory does not involve any regulatory mechanism and can be regarded as a "neutral theory". I am looking forward to seeing single cell experiments in the future to provide evidences about this theory.

[Ref1] https://www.biorxiv.org/content/10.1101/2024.04.19.590370v2<br /> [Ref2] https://www.biorxiv.org/content/10.1101/2024.10.08.617237v2

Author response:

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

Reviewer #1 (Public Review):

Summary:

Cell metabolism exhibits a well-known behavior in fast-growing cells, which employ seemingly wasteful fermentation to generate energy even in the presence of sufficient environmental oxygen. This phenomenon is known as Overflow Metabolism or the Warburg effect in cancer. It is present in a wide range of organisms, from bacteria and fungi to mammalian cells.

In this work, starting with a metabolic network for Escherichia coli based on sets of carbon sources, and using a corresponding coarse-grained model, the author applies some well-based approximations from the literature and algebraic manipulations. These are used to successfully explain the origins of Overflow Metabolism, both qualitatively and quantitatively, by comparing the results with E. coli experimental data.

By modeling the proteome energy efficiencies for respiration and fermentation, the study shows that these parameters are dependent on the carbon source quality constants K_i (p.115 and 116). It is demonstrated that as the environment becomes richer, the optimal solution for proteome energy efficiency shifts from respiration to fermentation. This shift occurs at a critical parameter value K_A(C).

This counterintuitive result qualitatively explains Overflow Metabolism.

Quantitative agreement is achieved through the analysis of the heterogeneity of the metabolic status within a cell population. By introducing heterogeneity, the critical growth rate is assumed to follow a Gaussian distribution over the cell population, resulting in accordance with experimental data for E. coli. Overflow metabolism is explained by considering optimal protein allocation and cell heterogeneity.

The obtained model is extensively tested through perturbations: 1) Introduction of overexpression of useless proteins; 2) Studying energy dissipation; 3) Analysis of the impact of translation inhibition with different sub-lethal doses of chloramphenicol on Escherichia coli; 4) Alteration of nutrient categories of carbon sources using pyruvate. All model perturbation results are corroborated by E. coli experimental results.

We appreciate the reviewer's highly positive comments and the accurate summary of our manuscript.

Strengths:

In this work, the author employs modeling methods typical of Physics to address a problem in Biology, standing at the interface between these two scientific fields. This interdisciplinary approach proves to be highly fruitful and should be further explored in the literature. The use of Escherichia coli as an example ensures that all hypotheses and approximations in this study are well-founded in the literature. Examples include the approximation for the Michaelis-Menten equation (line 82), Eq. S1, proteome partition in Appendix 1.1 (lines 68-69), and a stable nutrient environment in Appendix 1.1 (lines 83-84). The section "Testing the model through perturbation" heavily relies on bacterial data. The construction of the model and its agreement with experimental data are convincingly presented.

We appreciate the reviewer's highly positive comments. We have incorporated many of the reviewer's insightful suggestions and added citations in the appropriate contexts, which have significantly improved our manuscript.

Weaknesses:

In Section Appendix 6.4, the author explores the generalization of results from bacteria to cancer cells, adapting the metabolic network and coarse-grained model accordingly. It is argued that as a consequence, all subsequent steps become immediately valid. However, I remain unconvinced, considering the numerous approximations used to derive the equations, which the literature demonstrates to be valid primarily for bacteria. A more detailed discussion about this generalization is recommended. Additionally, it is crucial to note that the experimental validation of model perturbations heavily relies on E. coli data.

We appreciate the reviewer's insightful suggestions. We apologize for not clearly illustrating the generalization of results from bacteria to cancer cells in the previous version of our manuscript. Indeed, in our earlier version, there was no experimental validation of model results related to cancer cells.

Following the reviewer’s suggestions, we have now added Fig. 5 and Appendix-fig. 5, fully expanded the previous Appendix 6.4 into Appendix 9 in our current version, and added a new section entitled “Explanation of the Crabtree effect in yeast and the Warburg effect in cancer cells” in our main text to provide a detailed discussion of the generalization from bacteria to yeast and cancer cells. Through the derivations shown in Appendix 9 (Eqs. S180-S189), we arrived at Eq. 6 (or Eq. S190 in Appendix 9) to facilitate the comparison of our model results with experimental data in yeast and cancer cells. This comparison is presented in Fig. 5, where we demonstrate that our model can quantitatively explain the data for the Crabtree effect in yeast and the Warburg effect in cancer cells (related experimental data references: Shen et al., Nature Chemical Biology 20, 1123–1132 (2024); Bartman et al., Nature 614, 349-357 (2023)). These additions have significantly strengthened our manuscript.

Reviewer #2 (Public Review):

Summary

This paper has three parts. The first part applied a coarse-grained model with proteome partition to calculate cell growth under respiration and fermentation modes. The second part considered single-cell variability and performed population average to acquire an ensemble metabolic profile for acetate fermentation. The third part used model and simulation to compare experimental data in literature and obtained substantial consistency.

We thank the reviewer for the accurate summary and positive comments on our manuscript.

Strengths and major contributions

(i) The coarse-grained model considered specific metabolite groups and their interrelations and acquired an analytical solution for this scenario. The "resolution" of this model is in between the Flux Balanced Analysis/whole-cell simulation and proteome partition analysis.

(ii) The author considered single-cell level metabolic heterogeneity and calculated the ensemble average with explicit calculation. The results are consistent with known fermentation and growth phenomena qualitatively and can be quantitatively compared to experimental results.

We appreciate the reviewer’s highly positive comments.

Weaknesses

(i) If I am reading this paper correctly, the author's model predicts binary (or "digital") outcomes of single-cell metabolism, that is, after growth rate optimization, each cell will adopt either "respiration mode" or "fermentation mode" (as illustrated in Figure Appendix - Figure 1 C, D). Due to variability enzyme activity k_i^{cat} and critical growth rate λ_C, each cell under the same nutrient condition could have either respiration or fermentation, but the choice is binary.

The binary choice at the single-cell level is inconsistent with our current understanding of metabolism. If a cell only uses fermentation mode (as shown in Appendix - Figure 1C), it could generate enough energy but not be able to have enough metabolic fluxes to feed into the TCA cycle. That is, under pure fermentation mode, the cell cannot expand the pool of TCA cycle metabolites and hence cannot grow.

This caveat also appears in the model in Appendix (S25) that assumes J_E = r_E*J_{BM} where r_E is a constant. From my understanding, r_E can be different between respiration and fermentation modes (at least for real cells) and hence it is inappropriate to conclude that cells using fermentation, which generates enough energy, can also generate a balanced biomass.

We thank the reviewer for raising this question. Indeed, regarding energy biogenesis between respiration and fermentation, our model predicts binary outcomes at the single-cell level. However, this outcome does not hinder cell growth, as there are three independent possible fates for the carbon source (e.g., glucose) in metabolism: fermentation, respiration for energy biogenesis, and biomass generation. Each fate is associated with a distinct fraction of the proteome, with no overlap between them (see Appendix-figs. 1 and 5). Consequently, in a purely fermentative mode, a cell can still use the proteome dedicated to the biomass generation pathway to produce biomass precursors via the TCA cycle.

The classification of the carbon source’s fates into three independent pathways was initially introduced by Chen and Nielsen (Chen and Nielsen, PNAS 116, 17592-17597 (2019)). We apologize for the oversight in not citing their paper in this context in the previous version of our manuscript (although it was cited elsewhere). We have now included the citation in all appropriate places.

To illustrate this issue more clearly, we explicitly present the proteome allocation results for optimal growth in a fermentation mode below, where the proteome efficiency (i.e., the proteome energy efficiency in our previous version) in fermentation is higher than in respiration (i.e., ). We use the model shown in Fig. 1B as an example, with the relevant equations being Eqs. S26 and S28 in Appendix 2.1. By substituting Eq. S28 into Eq. S26, we arrive at Eq. 3 (or Eq. S29 in Appendix 2.1), which we restate here as Eq. R1:

For a given nutrient condition, i.e., for a specific value of κ_A at the single-cell level, the values of are determined (see Eqs. S20, S27, S31 and S32), while  ϕ and φ_max are constants (see Eq. S33 and Appendix 1.1). Therefore, if , then , since all coefficients are positive (i.e., ) and takes non-negative values. Hence, the solution for optimal growth is (see Eqs. S35-S36 in Appendix 2.2):

Here, the result signifies a pure fermentation mode with no respiration flux for energy biogenesis. Then, by combining Eq. R2 with Eqs. S28 and S30 from Appendix 2.1, we obtain the optimal proteome allocation results for this case:

where , while κ_A and take given values (see Eqs. S20 and S27). In Eq. R3, φ₃ corresponds to the fraction of the proteome devoted to carrying the carbon flux from Acetyl-CoA (the entry point of Pool b, see Fig. 1B and Appendix 1.2) to α-Ketoglutarate (the entry point of Pool c), with all of these being enzymes within the TCA cycle. The optimal growth solution is , which demonstrates that in a pure fermentation mode, the optimal growth condition includes the presence of enzymes within the TCA cycle capable of carrying the flux required for biomass generation.

Regarding Eq. S25, J_E represents the energy demand for cell proliferation, expressed as the stoichiometric energy flux in ATP. Although the influx of carbon sources (e.g., glucose) varies significantly between fermentation and respiration modes, J_BM and J_E  are the biomass and energy fluxes used to build cells, respectively. In bacteria, whether in fermentation or respiration mode, the proportion of maintenance energy used for protein degradation is roughly negligible (see Locasale and Cantley, BMC Biol 8, 88 (2010)). Consequently, the energy demand represented by J_E scales approximately linearly with the biomass production rate _J_BM (related experimental data reference: Ebenhöh et al., Life 14, 247 (2024)), regardless of the energy biogenesis mode. Therefore, _r_E can be regarded as roughly constant for bacteria. However, in eukaryotic cells such as yeast and mammalian cells, the proportion of maintenance energy is much more significant (see Locasale and Cantley, BMC Biol 8, 88 (2010)). Therefore, we have explicitly considered the contribution of maintenance energy in these cases and have extended the previous Appendix 6.4 into Appendix 9 in the current version.

(ii) The minor weakness of this model is that it assumes a priori that each cell chooses its metabolic strategy based on energy efficiency. This is an interesting assumption but there is no known biochemical pathway that directly executes this mechanism. In evolution, growth rate is more frequently considered for metabolic optimization. In Flux Balanced Analysis, one could have multiple objective functions including biomass synthesis, energy generation, entropy production, etc. Therefore, the author would need to justify this assumption and propose a reasonable biochemical mechanism for cells to sense and regulate their energy efficiency.

We thank the reviewer for raising this question and apologize for not explaining this point clearly enough in the previous version of our manuscript. Just as the reviewer mentioned, growth rate should be considered for metabolic optimization under the selection pressure of the evolutionary process. In fact, in our model, the sole optimization objective is exactly the cell growth rate. The determination of whether to use fermentation or respiration based on proteome efficiency (i.e., the proteome energy efficiency in our previous version) is not an a priori assumption in our model; rather, it is a natural consequence of growth rate optimization, as we detail below. 

For a given nutrient condition with a determined value of κ_A , as we have explained in the aforementioned responses, the constraint on the fluxes is summarized in Eq. 3 and is restated as Eq. R1. Mathematically, we can obtain the solution for the optimal growth strategy by combining Eq. R1 (i.e., Eq. 3) with the optimization on cell growth rate λ, and the solution can be obtained as follows: If the proteome efficiency in fermentation is larger than that in respiration, i.e., , then from Eq. R1, we obtain , since the values of ε_r , ε_f, Ψ, ϕ and φ_max are all fixed for a given κ_A_ , with ε_r , ε_f, Ψ, ϕ, φ_max > 0 . Hence, (since ), and note that . Therefore is the solution for optimal growth, where the growth rate can take the maximum value of . Similarly, for the case where the proteome efficiency in respiration is larger than that in fermentation (i.e ), is the solution for optimal growth. With this analysis, we have demonstrated that the choice between fermentation and respiration based on proteome efficiency is a natural consequence of growth rate optimization.

We have now revised the related content in our manuscript to clarify this point.

My feeling is that the mathematical structure of this model could be correct, but the single-cell interpretation for the ensemble averaging has issues. Each cell could potentially adopt partial respiration and partial fermentation at the same time and have temporal variability in its metabolic mode as well. With the modification of the optimization scheme, the author could have a revised model that avoids the caveat mentioned above.

We thank the reviewer for raising this question. In fact, in the above two responses, we have addressed the issues raised here, clarifying that the binary mode between respiration and fermentation does not hinder cell growth and that the sole optimization objective is the cell growth rate, as the reviewer suggested. Regarding temporal variability, due to factors such as cell cycle stages and the intrinsic noise arising from stochastic processes, temporal variability in the fermentation or respiration mode is indeed likely. However, at any given moment at the single-cell level, a binary choice between fermentation and respiration is what our model predicts for the optimal growth strategy. 

Discussion and impact for the field

Proteome partition models and Flux Balanced Analysis are both commonly used mathematical models that emphasize different parts of cellular physiology. This paper has ingredients for both, and I expect after revision it will bridge our understanding of the whole cell.

We appreciate the reviewer’s very positive comments. We have followed many of the good suggestions raised by the reviewer, and our revised manuscript is much improved as a result.

Reviewer #3 (Public Review):

Summary:

In the manuscript "Overflow metabolism originates from growth optimization and cell heterogeneity" the author Xin Wang investigates the hypothesis that the transition into overflow metabolism at large growth rates actually results from an inhomogeneous cell population, in which every individual cell either performs respiration or fermentation.

We thank the reviewer for carefully reading our manuscript and the accurate summary.

Weaknesses:

The paper has several major flaws. First, and most importantly, it repeatedly and wrongly claims that the origins of overflow metabolism are not known. The paper is written as if it is the first to study overflow metabolism and provide a sound explanation for the experimental observations. This is obviously not true and the author actually cites many papers in which explanations of overflow metabolism are suggested (see e.g. Basan et al. 2015, which even has the title "Overflow metabolism in E. coli results from efficient proteome allocation"). The paper should be rewritten in a more modest and scientific style, not attempting to make claims of novelty that are not supported. In fact, all hypotheses in this paper are old. Also the possiblility that cell heterogeneity explains the observed 'smooth' transition into overflow metabolism has been extensively investigated previously (see de Groot et al. 2023, PNAS, "Effective bet-hedging through growth rate dependent stability") and the random drawing of kcat-values is an established technique (Beg et al., 2007, PNAS, "Intracellular crowding defines the mode and sequence of substrate uptake by Escherichia coli and constrains its metabolic activity"). Thus, in terms of novelty, this paper is very limited. It reinvents the wheel and it is written as if decades of literature debating overflow metabolism did not exist.

We thank the reviewer for both the critical and constructive comments. Following the reviewer’s suggestion, we have revised our manuscript to adopt a more modest style. However, we respectfully disagree with the criticism regarding the novelty of our study, as detailed below.

First, while many explanations for overflow metabolism have been proposed, we have cited these in both the previous and current versions of our manuscript. We apologize for not emphasizing the distinctions between these previous explanations and our study in the main text of our earlier version, though we did provide details in Appendix 6.3. In fact, most of these explanations (e.g., Basan et al., Nature 528, 99-104 (2015); Chen and Nielsen, PNAS 116, 17592-17597 (2019); Majewski and Domach, Biotechnol. Bioeng. 35, 732-738 (1990); Niebel et al., Nat. Metab. 1, 125-132 (2019); Shlomi et al., PLoS Comput. Biol. 7, e1002018 (2011); Varma and Palsson, Appl. Environ. Microbiol. 60, 3724-3731 (1994); Vazquez et al., BMC Syst. Biol. 4, 58 (2010); Vazquez and Oltvai, Sci. Rep. 6, 31007 (2016); Zhuang et al., Mol. Syst. Biol. 7, 500 (2011)) heavily rely on the assumption that cells optimize their growth rate for a given rate of carbon influx under each nutrient condition (or certain equivalents) to explain the growth rate dependence of fermentation flux. However, this assumption—that cell growth rate is optimized for a given rate of carbon influx—is questionable, as the given factors in a nutrient condition are the identity and concentration of the carbon source, rather than the carbon influx itself.

Consequently, in our model, we purely optimize cell growth rate without imposing a special constraint on carbon influx. Our assumption that the given factors in a nutrient condition are the identity and concentration of the carbon source aligns with the studies by Molenaar et al. (Molenaar et al., Mol. Syst. Biol. 5, 323 (2009)), where they specified an identical assumption on page 5 of their Supplementary Information (SI); Scott et al. (Scott et al., Science 330, 1099-1102 (2010)), where the growth rate formula was derived for a culturing condition with a given nutrient quality; and Wang et al. (Wang et al., Nat. Comm. 10, 1279 (2019)), our previous study on microbial growth. Among these three studies, only Molenaar et al. addresses overflow metabolism. However, Molenaar et al. did not consider cell heterogeneity, resulting in their model predictions on the growth rate dependence of fermentation flux being a digital response, which is inconsistent with experimental data.

Furthermore, prevalent explanations such as those by Basan et al. (Basan et al., Nature 528, 99-104 (2015)) and Chen and Nielsen (Chen and Nielsen, PNAS 116, 17592-17597 (2019)) suggest that overflow metabolism originates from the proteome efficiency in fermentation always being higher than in respiration. However, Shen et al. (Shen et al., Nature Chemical Biology 20, 1123–1132 (2024)) recently discovered that the proteome efficiency measured at the cell population level in respiration is higher than in fermentation for many yeast and cancer cells, despite the presence of fermentation fluxes through aerobic glycolysis. This finding clearly contradicts the studies by Basan et al. (2015) and Chen and Nielsen (2019). 

Nevertheless, our model may resolve this puzzle by incorporating two important features. First, in our model, the proteome efficiency (i.e., the proteome energy efficiency in our previous version) in respiration is larger than that in fermentation when nutrient quality is low (Eqs. S174-S175 in Appendix 9). Second, and crucially, due to the incorporation of cell heterogeneity in our model, there could be a proportion of cells with higher proteome efficiency in fermentation than in respiration, even when the overall proteome efficiency at the cell population level is higher in respiration than in fermentation. As shown in the newly added Fig. 5A-B, our model results can quantitatively illustrate the experimental data from Shen et al., Nature Chemical Biology 20, 1123–1132 (2024).

Finally, regarding the criticism of the novelty of our hypothesis: As specified in our main text, cell heterogeneity has been widely reported experimentally in both microbes (e.g., Ackermann, Nat. Rev. Microbiol. 13, 497-508 (2015); Bagamery et al., Curr. Biol. 30, 4563-4578 (2020); Balaban et al., Science 305, 1622-1625 (2004); Nikolic et al., BMC Microbiol. 13, 1-13 (2013); Solopova et al., PNAS 111, 7427-7432 (2014); Wallden et al., Cell 166, 729-739 (2016)) and tumor cells (e.g., Duraj et al., Cells 10, 202 (2021); Hanahan and Weinberg, Cell 164, 681-694 (2011); Hensley et al., Cell 164, 681-694 (2016)). However, to the best of our knowledge, cell heterogeneity has not yet been incorporated into theoretical models for explaining overflow metabolism or the Warburg effect. The reviewer mentioned the study by de Groot et al. (de Groot et al., PNAS 120, e2211091120 (2023)) as studying overflow metabolism similarly to our work. We have carefully read this paper, including the main text and SI, and found that it is not directly relevant to either overflow metabolism or the Warburg effect. Instead, their model extends the work of Kussell and Leibler (Kussell and Leibler, Science 309, 2075-2078 (2005)), focusing on bet-hedging strategies of microbes in changing environments.

Regarding the criticism that random drawing of kcat-values is an established technique (Beg et al., PNAS 104, 12663-12668 (2007)), we need to stress that the distribution noise on kcat-values considered in our model is fundamentally different from that in Beg et al. In Beg et al., their model involved 876 reactions (see Dataset 1 in Beg et al.), of which only 109 had associated biochemical experimental data. Thus, their distribution of kcat-values pertains to different enzymes within the same cell. In contrast, we have the mean of the kcat-values from experimental data for each relevant enzymes, with the distribution of kcat-values representing the same enzyme in different cells.           

Moreover, the manuscript is not clearly written and is hard to understand. Variables are not properly introduced (the M-pools need to be discussed, fluxes (J_E), "energy coefficients" (eta_E), etc. need to be more explicitly explained. What is "flux balance at each intermediate node"? How is the "proteome efficiency" of a pathway defined? The paper continues to speak of energy production. This should be avoided. Energy is conserved (1st law of thermodynamics) and can never be produced. A scientific paper should strive for scientific correctness, including precise choice of words.

We thank the reviewer for the constructive comments. Following these, we have provided more explicit information and revised our manuscript to enhance readability. In our initially submitted version, the phrase "energy production" was borrowed from Nelson et al. (Nelson et al., Lehninger principles of biochemistry, 2008) and Basan et al. (Basan et al., Nature 528, 99-104 (2015)), and we chose to follow this terminology. We appreciate the reviewer’s suggestion and have now revised the wording to use more appropriate expressions.

The statement that the "energy production rate ... is proportional to the growth rate" is, apart from being incorrect - it should be 'ATP consumption rate' or similar (see above), a non-trivial claim. Why should this be the case? Such statements must be supported by references. The observation that the catabolic power indeed appears to increase linearly with growth rate was made, based on chemostat data for E.coli and yeast, in a recent preprint (Ebenhöh et al, 2023, bioRxiv, "Microbial pathway thermodynamics: structural models unveil anabolic and catabolic processes").

We thank the reviewer for the insightful suggestions. Following these, we have revised our manuscript and cited the suggested reference (i.e., Ebenhöh et al., Life 14, 247 (2024)).

All this criticism does not preclude the possibility that cell heterogeneity plays a role in overflow metabolism. However, according to Occam's razor, first the simpler explanations should be explored and refuted before coming up with a more complex solution. Here, it means that the authors first should argue why simpler explanations (e.g. the 'Membrane Real Estate Hypothesis', Szenk et al., 2017, Cell Systems; maximal Gibbs free energy dissipation, Niebel et al., 2019, Nature Metabolism; Saadat et al., 2020, Entropy) are not considered, resp. in what way they are in disagreement with observations, and then provide some evidence of the proposed cell heterogeneity (are there single-cell transcriptomic data supporting the claim?).

We thank the reviewer for raising these questions and providing valuable insights. Regarding the shortcomings of simpler explanations, as explained above, most proposed explanations (including the references mentioned by the reviewer: Szenk et al., Cell Syst. 5, 95-104 (2017); Niebel et al., Nat. Metab. 1, 125-132 (2019); Saadat et al., Entropy 22, 277 (2020)) rely heavily on the assumption that cells optimize their growth rate for a given rate of carbon influx under each nutrient condition (or its equivalents). However, this assumption is questionable, as the given factors in a nutrient condition are the identities and concentrations of the carbon sources, rather than the carbon influx itself.

Specifically, Szenk et al. is a perspective paper, and the original “membrane real estate hypothesis” was proposed by Zhuang et al. (Zhuang et al., Mol. Syst. Biol. 7, 500 (2011)). Zhuang et al. specified in Section 7 of their SI that their model’s explanation of the experimental results shown in Fig. 2C of their manuscript relies on the assumption of restrictions on carbon influx. In Niebel et al. (Niebel et al., Nat. Metab. 1, 125-132 (2019)), the Methods section specifies that the glucose uptake rate was considered a given factor for a growth condition. In Saadat et al. (Saadat et al., Entropy 22, 277 (2020)), Appendix A notes that their model results depend on minimizing carbon influx for a given growth rate, which is equivalent to the assumption mentioned above (see Appendix 6.3 in our manuscript for details). 

Regarding the experimental evidence for our proposed cell heterogeneity, Bagamery et al. (Bagamery et al., Curr. Biol. 30, 4563-4578 (2020)) reported non-genetic heterogeneity in two subpopulations of Saccharomyces cerevisiae cells upon the withdrawal of glucose from exponentially growing cells. This strongly indicates the coexistence of fermentative and respiratory modes of heterogeneity in S. cerevisiae cultured in a glucose medium (refer to Fig. 1E in Bagamery et al.). Nikolic et al. (Nikolic et al., BMC Microbiol. 13, 1-13 (2013)) reported a bimodal distribution in the expression of the acs gene (the transporter for acetate) in an E. coli cell population growing on glucose as the sole carbon source within the region of overflow metabolism (see Fig. 5 in Nikolic et al.), indicating the cell heterogeneity we propose. For cancer cells, Duraj et al. (Duraj et al., Cells 10, 202 (2021)) reported a high level of intra-tumor heterogeneity in glioblastoma using optical microscopy images, where 48%~75% of the cells use fermentation and the remainder use respiration (see Fig. 1C in Duraj et al.), which aligns with the cell heterogeneity picture of aerobic glycolysis predicted by our model.   

We have now added related content to the discussion section to strengthen our manuscript.

Reviewer #1 (Recommendations For The Authors): 

Some minor corrections:

(1) Adjusted the reference: (García-Contreras et al., 2012)

(2) Corrected line 255: Removed the duplicate "the genes"

We thank the reviewer for the suggestions and have implemented each of them to revise our manuscript. The reference in the form of García-Contreras et al., 2012, although somewhat unusual, is actually correct, so we have kept it unchanged.

General comment to the author:

Considering that this work exists at the interface between Physics and Biology, where a significant portion of the audience may not be familiar with the mathematical manipulations performed, it would enhance the paper's readability to provide more explicit indications in the text. For example, in line 91, explicitly define phi_A as phi_R; or in line 115, explain the K_i parameter in the text for better readability.

We thank the reviewer for the suggestion. Following this, we have now provided more explicit information for the definition of mathematical symbols to enhance readability.

Reviewer #2 (Recommendations For The Authors):

The current form of this manuscript is difficult to read for general readers. In addition, the model description in the Appendix can be improved for biophysics readers to keep track of the variables. Here are my suggestions:

a) In the main text, the author should give the definition of "proteome energy efficiency" explicitly both in English and mathematical formula - since this is the central concept of the paper. The biological interpretation of formula (4) should also be stated.

We thank the reviewer for the suggestion. Following this, we have now added definitions and biological interpretations to fix these issues.

b) I feel the basic model of the reaction network in the Appendix could be stated in a more concise way, by emphasizing whether a variable is extensive (exponential growing) or intensive (scale-invariant under exponential growth).

From my understanding, this work assumes balanced exponential growth and hence there is a balanced biomass vector Y* (a constant unit vector with all components sum to 1) for each cell. The steady-state fluxes {J} are extensive and all have growth rate λ. The proteome partition and relative metabolite fractions are ratios of different components of Y* and hence are intensive.

The normalized fluxes {J^(n)} (with respect to biomass) are a function of Y* and are all kept as constant ratios with each other. They are also intensive.

The biomass and energy production are linear combinations of {J} and hence are extensive and follow exponential growth. The biomass and energy efficiency are ratios between flux and proteome biomass, and hence are intensive.

We thank the reviewer for the insightful suggestion. Following this, we have now added the intensive and extensive information for all relevant variables in the newly added Appendix-table 3.

c) In the Appendix, the author should have a table or list of important variables, with their definition, units, and physiological values under respiration and fermentation.

We thank the reviewer for the very useful suggestion. Following this, we have now added Appendix-table 3 (pages 54-57 in the appendices) to illustrate the symbols used throughout our manuscript, as well as the model variables and parameter settings.   

d) Regarding the single-cell variability, the author ignored recent experimental measurements on single-cell metabolism. This includes variability on ATP, NAD(P)H in E. coli, which will be useful background for the readers, see below.

https://pubmed.ncbi.nlm.nih.gov/25283467/

https://pubmed.ncbi.nlm.nih.gov/29391569/

We thank the reviewer for the very useful suggestion. We have now cited these relevant studies in our manuscript.  

e) The choice between 100% respiration and 100% fermentation is based on the optimization of proteome energy efficiency, while the intermediate strategies are not favored in this model. This is similar to a concept in control theory called the bang-bang principle. This can be added to the Discussion.

We thank the reviewer for this suggestion. We have reviewed the concept and articles on the bang-bang principle. While the bang-bang principle is indeed relevant to binary choices, it is somewhat distant from the topic of metabolic strategies related to optimal growth. The elementary flux mode (see Müller et al., J. Theor. Biol. 347, 182190 (2014); Wortel et al., FEBS J. 281, 1547-1555 (2014)) is more pertinent to this topic, as it may lead to diauxic microbial growth (another binary metabolic strategy) in microbes grown on a mixture of two carbon sources from Group A (see Wang et al., Nat. Comm. 10, 1279 (2019)). Therefore, we have cited and mentioned only the elementary flux mode (Müller et al., J. Theor. Biol. 347, 182-190 (2014); Wortel et al., FEBS J. 281, 1547-1555 (2014)) in the introduction and discussion sections of our manuscript.

eLife Assessment

This important study explores the association between mother-child interactions and the development of children's social brain networks, specifically the theory of mind and social pain networks. The findings provide solid evidence for enhanced stimulus-evoked neural synchronization between child-caregiver dyads, while the evidence for the other variables is incomplete and could be strengthened with further analyses. The study effectively bridges brain development with children's behavior and parenting practices and would be of interest to broad research communities in social neuroscience and developmental psychology.

Reviewer #1 (Public review):

The authors sought to examine the associations between child age, reports of parent-child relationship quality, and neural activity patterns while children (and also their parents) watched a movie clip. Major methodological strengths include the sample of 3-8 year-old children in China (rare in fMRI research for both age range and non-Western samples), use of a movie clip previously demonstrated to capture theory of mind constructs at the neural level, measurement of caregiver-child neural synchrony, and assessment of neural maturity. Results provide important new information about parent-child neural synchronization during this movie and associations with reports of parent-child relationship quality. The work is a notable advance in understanding the link between the caregiving context and the neural construction of theory of mind networks in the developing brain.

There are several theoretical and methodological limitations of the manuscript in its current form:

(1) We appreciate that the authors wanted to show support for a mediational mechanism. However, we suggest that the authors drop the structural equation modeling because the data are cross-sectional so mediation is not appropriate. Other issues include the weak justification of including the parent-child neural synchronization as part of parenting.... it could just as easily be a mechanism of change or driven by the child rather than a component of parenting behavior. The paper would be strengthened by looking at associations between selected variables of interest that are MOST relevant to the imaging task in a regression type of model. Furthermore, the authors need to be more explicit about corrections for multiple comparisons throughout the manuscript; some of the associations are fairly weak so claims may need to be tempered if they don't survive correction.

(2) Reverse correlation analysis is sensible given what prior developmental fMRI studies have done. But reverse correlation analysis may be more prone to overfitting and noise, and lacks sensitivity to multivariate patterns. Might inter-subject correlation be useful for *within* the child group? This would minimize noise and allow for non-linear patterns to emerge.

(3) No learning effects or temporal lagged effects are tested in the current study, so the results do not support the authors' conclusions that the data speak to Bandura's social learning theory. The authors do mention theories of biobehavioral synchrony in the introduction but do not discuss this framework in the discussion (which is most directly relevant to the data). The data can also speak to other neurodevelopmental theories of development (e.g.,neuroconstructivist approaches), but the authors do not discuss them. The manuscript would benefit from significantly revising the framework to focus more on biobehavioral synchrony data and other neurodevelopmental approaches given the prior work done in this area rather than a social psychology framework that is not directly evaluated.

(4) The significance and impact of the findings would be clearer if the authors more clearly situated the findings in the context of (a) other movie and theory of mind fMRI task data during development; and (b) existing data on parent-child neural synchrony (often uses fNIRS or EEG). What principles of brain and social cognition development do these data speak to? What is new?

(5) There is little discussion about the study limitations, considerations about the generalizability of the findings, and important next steps and future directions. What can the data tell us, and what can it NOT tell us?

Reviewer #2 (Public review):

Summary:

This study investigates the impact of mother-child neural synchronization and the quality of parent-child relationships on the development of Theory of Mind (ToM) and social cognition. Utilizing a naturalistic fMRI movie-viewing paradigm, the authors analyzed inter-subject neural synchronization in mother-child dyads and explored the connections between neural maturity, parental caregiving, and social cognitive outcomes. The findings indicate age-related maturation in ToM and social pain networks, emphasizing the importance of dyadic interactions in shaping ToM performance and social skills, thereby enhancing our understanding of the environmental and intrinsic influences on social cognition.

Strengths:

This research addresses a significant question in developmental neuroscience, by linking social brain development with children's behaviors and parenting. It also uses a robust methodology by incorporating neural synchrony measures, naturalistic stimuli, and a substantial sample of mother-child dyads to enhance its ecological validity. Furthermore, the SEM approach provides a nuanced understanding of the developmental pathways associated with Theory of Mind (ToM).

Weaknesses:

(1) Upon reviewing the introduction, I feel that the first goal - developmental changes of the social brain and its relation to age - seems somewhat distinct from the other two goals and the main research question of the manuscript. The authors might consider revising this section to enhance the overall coherence of the manuscript. Additionally, the introduction lacks a clear background and rationale for the importance of examining age-related changes in the social brain.

(2) The manuscript uses both "mother-child" and "parent-child" terminology. Does this imply that only mothers participated in the fMRI scans while fathers completed the questionnaires? If so, have the authors considered the potential impact of parental roles (father vs. mother)?

(3) There is inconsistent usage of the terms ISC and ISS in the text and figures, both of which appear to refer to synchronization derived from correlation analysis. It would be beneficial to maintain consistency throughout the manuscript.

(4) Of the 50 dyads, 16 were excluded due to data quality issues, which constitutes a significant proportion. It would be helpful to know whether these excluded dyads exhibited any distinctive characteristics. Providing information on demographic or behavioral differences-such as Theory of Mind (ToM) performance and age range between the excluded and included dyads would enhance the assessment of the findings' generalizability.

(5) The article does not adhere to the standard practice of using a resting state as a baseline for subtracting from task synchronization. Is there a rationale for this approach? Not controlling for a baseline may lead to issues, such as whether resting state synchronization already differs between subjects with varying characteristics.

(6) The title of the manuscript suggests a direct influence of mother-child interactions on children's social brain and theory of mind. However, the use of structural equation modeling (SEM) may not fully establish causal relationships. It is possible that the development of children's social brain and ToM also enhances mother-child neural synchronization. The authors should address this alternative hypothesis of the potential bidirectional relationship in the discussion and exercise caution regarding terms that imply causality in the title and throughout the manuscript.

(7) I would appreciate more details about the 14 Theory of Mind (ToM) tasks, which could be included in supplemental materials. The authors score them on a scale from 0 to 14 (each task 1 point); however, the tasks likely vary in difficulty and should carry different weights in the total score (for example, the test and the control questions should have different weights). Many studies have utilized the seven tasks according to Wellman and Liu (2004), categorizing them into "basic ToM" and "advanced ToM." Different components of ToM could influence the findings of the current study, which should be further examined by a more in-depth analysis.

Reviewer #3 (Public review):

Summary:

The article explores the role of mother-child interactions in the development of children's social cognition, focusing on Theory of Mind (ToM) and Social Pain Matrix (SPM) networks. Using a naturalistic fMRI paradigm involving movie viewing, the study examines relationships among children's neural development, mother-child neural synchronization, and interaction quality. The authors identified a developmental pattern in these networks, showing that they become more functionally distinct with age. Additionally, they found stronger neural synchronization between child-mother pairs compared to child-stranger pairs, with this synchronization and neural maturation of the networks associated with the mother-child relationship and parenting quality.

Strengths:

This is a well-written paper, and using dyadic fMRI and naturalistic stimuli enhances its ecological validity, providing valuable insights into the dynamic interplay between brain development and social interactions. However, I have some concerns regarding the analysis and interpretation of the findings. I have outlined these concerns below in the order they appear in the manuscript, which I hope will be helpful for the revision.

Weaknesses:

(1) Given the importance of social cognition in this study, please cite a foundational empirical or review paper on social cognition to support its definition. The current first citation is primarily related to ASD research, which may not fully capture the broader context of social cognition development.

(2) It is standard practice to report the final sample size in the Abstract and Introduction, rather than the initial recruited sample, as high attrition rates are common in pediatric studies. For example, this study recruited 50 mother-child dyads, and only 34 remained after quality control. This information is crucial for interpreting the results and conclusions. I recommend reporting the final sample size in the abstract and introduction but specifying in the Methods that an additional 16 mother-child dyads were initially recruited or that 50 dyads were originally collected.

(3) In the "Neural maturity reflects the development of the social brain" section, the authors report the across-network correlation for adults, finding a negative correlation between ToM and SPM. However, the cross-network correlations for the three child groups are not reported. The statement that "the two networks were already functionally distinct in the youngest group of children we tested" is based solely on within-network positive correlations, which does not fully demonstrate functional distinctness. Including cross-network correlations for the child groups would strengthen this conclusion.

(4) The ROIs for the ToM and SPM networks are defined based on previous literature, applying the same ROIs across all age groups. While I understand this is a common approach, it's important to note that this assumption may not fully hold, as network architecture can evolve with age. The functional ROIs or components of a network might shift, with regions potentially joining or exiting a network or changing in size as children develop. For instance, Mark H. Johnson's interactive specialization theory suggests that network composition may adapt over developmental stages. Although the authors follow the approach of Richardson et al. (2018), it would be beneficial to discuss this limitation in the Discussion. An alternative approach would be to apply data-driven analysis to justify the selection of the ROIs for the two networks.

(5) The current sample size (N = 34 dyads) is a limitation, particularly given the use of SEM, which generally requires larger samples for stable results. Although the model fit appears adequate, this does not guarantee reliability with the current sample size. I suggest discussing this limitation in more detail in the Discussion.

(6) Based on the above comment, I believe that conclusions regarding the relationship between social network development, parenting, and support for Bandura's theory should be tempered. The current conclusions may be too strong given the study's limitations.

(7) The SPM (pain) network is associated with empathic abilities, also an important aspect of social skills. It would be relevant to explore whether (or explain why) SPM development and child-mother synchronization are (or are not) related to parenting and the parent-child relationship.

eLife Assessment

NeuroSCAN is an accessible and interactive tool for streamlined observation of neuronal morphology, membrane contact, and synaptic connectivity across developmental stages in the nematode C. elegans. This important tool relies on solid electron microscopy datasets. This resource will be of high interest to C. elegans researchers interested in nervous system wiring and circuit function.

Reviewer #1 (Public review):

Summary:

The authors present NeuroSCAN, an accessible and interactive tool for visualizing and summarizing data from multiple previously annotated C. elegans connectomes. NeuroSCAN provides a useful entry point for streamlined observation of neuronal morphology, and the membrane contacts and synaptic connectivity between neurons across developmental stages and individual connectomes readily extracted from existing data.

Strengths:

Koonce et al. have generated a web-based visualization tool for exploring C. elegans neuronal morphology, contact area between neurons, and synaptic connectivity data. Here, the authors integrate volumetric segmentation of neurons and visualization of contact area patterns of individual neurons generated from Diffusion Condensation and C-PHATE embedding based on previous work from adult volumetric electron microscopy (vEM) data, extended to available vEM data for earlier developmental stages, which effectively summarizes modularity within the collated C. elegans contactomes to date. Overall, NeuroSCAN's relative ease of use for generating visualizations, its ability to quickly toggle between developmental stages, and its integration of a concise visualization of individual neurons' contact patterns strengthen its utility.

Weaknesses:

NeuroSCAN provides an accessible and convenient platform. However, many of the characteristics of NeuroSCAN overlap with that of an existing tool for visualizing connectomics data, Neuroglancer, which is a widely-used and shared platform with data from other organisms. The authors do not make clear their motivation for generating this new tool rather than building on a system that has already collated previous connectomics data. Although the field will benefit from any tool that collates connectomics data and makes it more accessible and user-friendly, such a tool is only useful if it is kept up-to-date, and if data formatting for submitting electron microscopy data to be added to the tool is made clear. It is unclear from this manuscript whether NeuroSCAN will be updated with recently published and future C. elegans connectomes, or how additional datasets can be submitted to be added in the future.

The interface for visualizing contacts and synapses would be improved with better user access to the quantitative underlying data. When contact areas or synapses are added to the viewer, adding statistics on the magnitude of the contact area, the number of synapses, and the rank of these values among the neuron's top connections, would make the viewer more useful for hypothesis generation. Furthermore, synapses are currently listed individually, with names that are not very legible to the web user. Grouping them by pre- and postsynaptic neurons and linking these groups across developmental stages would also be an improvement.

While the DC/C-PHATE visualizations are a useful tool for the user, it is difficult to understand when grouping or splitting of cell contact patterns is biologically significant. DC is a deterministic algorithm applied to a contactome from a single organism, and the authors do not provide quantitative metrics of distances between individual neurons or a number of DC iterations on the C-PHATE plot, nor is the selection process for the threshold for DC described in this manuscript. In the application of DC/C-PHATE to larval stage nerve ring strata organization shown by the authors, qualitative observations of C-PHATE plots colored based on adult data seem to be the only evidence shown for persistent strata during development (Figure 3) or changing architectural motifs across stages (Figure 4). Quantitation of differences in neuron position within the DC hierarchy, or differences in modularity across stages, is needed to support these conclusions. Furthermore, illustrating the quantitative differences in C-PHATE plots used to make these conclusions will provide a more instructive guide for users of NeuroSCAN in generating future hypotheses.

While the case studies presented by the authors help to highlight the utility of the different visualizations offered by the NeuroSCAN platform, the authors need to be more careful with the claims they make from these correlative observations. For example, in Figure 4, the authors use C-PHATE clustering patterns to make conclusions about changes in clustering patterns of individual neurons across development based on single animal datasets. In this and many other cases presented in this study with the limited existing datasets, it is difficult to differentiate between developmental changes and individual variability between the neurite positions, contacts, and synapse differences within these data. This caveat needs to be clearly addressed.

Reviewer #2 (Public review):

Summary:

The past five years have seen the publication of both new (Witvliet et al., 2021) and newly analyzed (Cook et al., 2019; Moyle et al., 2021; Brittin et al., 2021) data for the C. elegans connectome. The increase in data availability for a single species allows researchers to examine variability due to both stochastic events and changes over development. The quantity of these data is huge. To help the community make these data more accessible, the authors present a new online tool that allows the examination of 3D models for C. elegans neurons in the central neuropil across development. In addition to visualizing the overall structure of the neuronal processes and locations of synapses, the NeuroSCAN tool also allows users to probe into the C-PHATE visualization results, which this group previously pioneered to describe similarities in neuron adjacency (Moyle et al., 2021).

Strengths:

The ability to visualize the data from both a connectomics and contactomics perspective across developmental time has significant power. The original C. elegans connectome (White et al., 1986) presented their circuits as line drawings with chemical and electrical synapses indicated through arrows and bars. While these line drawings remain incredibly useful, they were also necessary simplifications for a 2D publication and they lack details of the complex architecture seen within each EM image. Koonce et al take advantage of segmented image data of each neuronal process within the nerve ring to create a web interface where users can visualize 3D models for their neuron of choice. The C-PHATE visualization allows users to explore similarities among different neurons in terms of adjacency and then go directly to the 3D model for these neurons. The 3D models it generates are beautiful and will likely be showing up in many future presentations and publications. The tool doesn't require any additional downloading and is open source.

Weaknesses:

While it's impossible to create one tool that will satisfy all potential users, I found myself wanting to have numbers associated with the data. For example, knowing the number of connections or the total surface area of contacts between individual neurons wasn't possible through the viewer, which limits the utility of taking deep analytical dives. While connectivity data are readily accessible through other interfaces such as Nemanode and WormWiring, a more thorough integration may be helpful to some users.

There were several issues with the user interface that made it a bit clunky to use. For example, as I added additional neurons to the filter search box, the loading time got longer and longer. I ran an experiment uploading all of the amphid neurons, one pair at a time. Each additional neuron pair added an additional 5-10 seconds to the loading. By the time I got to the last pair, it took over a minute to load. Issues like these, some of which may be unavoidable given the size of the data, could be conveyed through better documentation. I did not find the tutorial very helpful and the supplementary movies lacked any voiceover, so it wasn't always clear what they were trying to show.

Reviewer #3 (Public review):

Summary:

This work provides graphical tools for reconstructing the detailed anatomy of a nervous system from a series of sections imaged by electron microscopy. Contact between neuronal processes can direct outgrowth and is necessary for connectivity and, thus function. A bioinformatic approach is used to group neurons according to shared features (e.g., contact, synapses) in a hierarchy of "relatedness" that can be interrogated at each step. In this work, Koonze et al analyze vEM data sets for the C. elegans nerve ring (NR), a dense fascicle of processes from181 neurons. In a bioinformatic approach, the clustering algorithm Diffusion Condensation (DC) groups neurons according to similar cell biological features in iterations that remove chunks of differences in feature data with each step ultimately merging all NR neurons in one cluster. DC results are displayed with C-Phate a 3D visualization tool to produce a trajectory that can be interrogated for cell identities and other features at each iterative step. In previous work by these authors, this approach was utilized to identify subgroups of neuronal processes or "strata" in the NR that can be grouped by physical contact and connectivity. Here they expand their analysis to include a series of available vEM data sets across C. elegans larval development. This approach suggests that strata initially established during embryonic development are largely preserved in the adult. Importantly, exceptions involving stage-specific reorganization of neuronal placement in specific strata were also detected. A case study featured in the paper demonstrates the utility of this approach for visualizing the integration of newly generated neurons into the existing NR anatomy. Visualization tools used in this work are publicly available at NeuroSCAN.

Strengths:

A web-based app, NeuroSCAN, that individual researchers can use to interrogate the structure and organization of the C. elegans nerve ring across development

Weaknesses:

In the opinion of this reviewer, only minor revisions are required.

eLife Assessment

This important study is of significant relevance to the fields of predictive processing, perception, and learning. The well-designed paradigm allows the authors to avoid several common confounds in investigating predictions, such as stimulus familiarity and adaptation. Using a state-of-the-art multivariate EEG approach, the authors test the opposing process theory and find evidence in support of it, but some elements - especially the interactions across block - have only incomplete support at present. This could be strengthened via further analyses and justification.

Reviewer #1 (Public review):

Summary:

In this lovely paper, McDermott and colleagues tackle an enduring puzzle in the cognitive neuroscience of perceptual prediction. Though many scientists agree that top-down predictions shape perception, previous studies have yielded incompatible results - with studies showing 'sharpened' representations of expected signals, and others showing a 'dampening' of predictable signals to relatively enhance surprising prediction errors. To deepen the paradox further, it seems like there are good reasons that we would want to see both influences on perception in different contexts.

Here, the authors aim to test one possible resolution to this 'paradox' - the opposing process theory (OPT). This theory makes distinct predictions about how the time course of 'sharpening' and 'dampening' effects should unfold. The researchers present a clever twist on a leading-trailing perceptual prediction paradigm, using AI to generate a large dataset of test and training stimuli so that it is possible to form expectations about certain categories without repeating any particular stimuli. This provides a powerful way of distinguishing expectation effects from repetition effects - a perennial problem in this line of work.

Using EEG decoding, the researchers find evidence to support the OPT. Namely, they find that neural encoding of expected events is superior in earlier time ranges (sharpening-like) followed by a relative advantage for unexpected events in later time ranges (dampening-like). On top of this, the authors also show that these two separate influences may emerge differently in different phases of learning - with superior decoding of surprising prediction errors being found more in early phases of the task, and enhanced decoding of predicted events being found in the later phases of the experiment.

Strengths:

As noted above, a major strength of this work lies in important experimental design choices. Alongside removing any possible influence of repetition suppression mechanisms in this task, the experiment also allows us to see how effects emerge in 'real-time' as agents learn to make predictions. This contrasts with many other studies in this area - where researchers 'over-train' expectations into observers to create the strongest possible effects or rely on prior knowledge that was likely to be crystallised outside the lab.

Weaknesses:

This study reveals a great deal about how certain neural representations are altered by expectation and learning on shorter and longer timescales, so I am loath to describe certain limitations as 'weaknesses'. But one limitation inherent in this experimental design is that, by focusing on implicit, task-irrelevant predictions, there is not much opportunity to connect the predictive influences seen at the neural level to the perceptual performance itself (e.g., how participants make perceptual decisions about expected or unexpected events, or how these events are detected or appear).

The behavioural data that is displayed (from a post-recording behavioural session) shows that these predictions do influence perceptual choice - leading to faster reaction times when expectations are valid. In broad strokes, we may think that such a result is broadly consistent with a 'sharpening' view of perceptual prediction, and the fact that sharpening effects are found in the study to be larger at the end of the task than at the beginning. But it strikes me that the strongest test of the relevance of these (very interesting) EEG findings would be some evidence that the neural effects relate to behavioural influences (e.g., are participants actually more behaviourally sensitive to invalid signals in earlier phases of the experiment, given that this is where the neural effects show the most 'dampening' a.k.a., prediction error advantage?)

Reviewer #2 (Public review):

Summary:

There are two accounts in the literature that propose that expectations suppress the activity of neurons that are (a) not tuned to the expected stimulus to increase the signal-to-noise ratio for expected stimuli (sharpening model) or (b) tuned to the expected stimulus to highlight novel information (dampening model). One recent account, the opposing process theory, brings the two models together and suggests that both processes occur, but at different time points: initial sharpening is followed by later dampening of the neural activity of the expected stimulus. In this study, the authors aim to test the opposing process theory in a statistical learning task by applying multivariate EEG analyses and finding evidence for the opposing process theory based on the within-trial dynamics.

Strengths:

This study addresses a very timely research question about the underlying mechanisms of expectation suppression. The applied EEG decoding approach offers an elegant way to investigate the temporal characteristics of expectation effects. A major strength of the study lies in the experimental design that aims to control for repetition effects, one of the common confounds in prediction suppression studies. The reported results are novel in the field and have the potential to substantially improve our understanding of expectation suppression in visual perception.

Weaknesses:

The strength in controlling for repetition effects by introducing a neutral (50% expectation) condition also adds a weakness to the current version of the manuscript, as this neutral condition is not integrated into the behavioral (reaction times) and EEG (ERP and decoding) analyses. This procedure remained unclear to me. The reported results would be strengthened by showing differences between the neutral and expected (valid) conditions on the behavioral and neural levels. This would also provide a more rigorous check that participants had implicitly learned the associations between the picture category pairings.

It is not entirely clear to me what is actually decoded in the prediction condition and why the authors did not perform decoding over trial bins in prediction decoding as potential differences across time could be hidden by averaging the data. The manuscript would generally benefit from a more detailed description of the analysis rationale and methods.

Finally, the scope of this study should be limited to expectation suppression in visual perception, as the generalization of these results to other sensory modalities or to the action domain remains open for future research.

Reviewer #3 (Public review):

Summary:

In their study, McDermott et al. investigate the neurocomputational mechanism underlying sensory prediction errors. They contrast two accounts: representational sharpening and dampening. Representational sharpening suggests that predictions increase the fidelity of the neural representations of expected inputs, while representational dampening suggests the opposite (decreased fidelity for expected stimuli). The authors performed decoding analyses on EEG data, showing that first expected stimuli could be better decoded (sharpening), followed by a reversal during later response windows where unexpected inputs could be better decoded (dampening). These results are interpreted in the context of opposing process theory (OPT), which suggests that such a reversal would support perception to be both veridical (i.e., initial sharpening to increase the accuracy of perception) and informative (i.e., later dampening to highlight surprising, but informative inputs).

Strengths:

The topic of the present study is of significant relevance to the field of predictive processing. The experimental paradigm used by McDermott et al. is well designed, allowing the authors to avoid several common confounds in investigating predictions, such as stimulus familiarity and adaptation. The introduction of the manuscript provides a well-written summary of the main arguments for the two accounts of interest (sharpening and dampening), as well as OPT. Overall, the manuscript serves as a good overview of the current state of the field.

Weaknesses:

In my opinion, several details of the methods, results, and manuscript raise doubts about the quality and reliability of the reported findings. Key concerns are:

(1) The results in Figure 2C seem to show that the leading image itself can only be decoded with ~33% accuracy (25% chance; i.e. ~8% above chance decoding). In contrast, Figure 2E suggests the prediction (surprisingly, valid or invalid) during the leading image presentation can be decoded with ~62% accuracy (50% chance; i.e. ~12% above chance decoding). Unless I am misinterpreting the analyses, it seems implausible to me that a prediction, but not actually shown image, can be better decoded using EEG than an image that is presented on-screen.

(2) The "prediction decoding" analysis is described by the authors as "decoding the predictable trailing images based on the leading images". How this was done is however unclear to me. For each leading image decoding the predictable trailing images should be equivalent to decoding validity (as there were only 2 possible trailing image categories: 1 valid, 1 invalid). How is it then possible that the analysis is performed separately for valid and invalid trials? If the authors simply decode which leading image category was shown, but combine L1+L2 and L4+L5 into one class respectively, the resulting decoder would in my opinion not decode prediction, but instead dissociate the representation of L1+L2 from L4+L5, which may also explain why the time-course of the prediction peaks during the leading image stimulus-response, which is rather different compared to previous studies decoding predictions (e.g. Kok et al. 2017). Instead for the prediction analysis to be informative about the prediction, the decoder ought to decode the representation of the trailing image during the leading image and inter-stimulus interval. Therefore I am at present not convinced that the utilized analysis approach is informative about predictions.

(3) I may be misunderstanding the reported statistics or analyses, but it seems unlikely that >10 of the reported contrasts have the exact same statistic of Tmax= 2.76. Similarly, it seems implausible, based on visual inspection of Figure 2, that the Tmax for the invalid condition decoding (reported as Tmax = 14.903) is substantially larger than for the valid condition decoding (reported as Tmax = 2.76), even though the valid condition appears to have superior peak decoding performance. Combined these details may raise concerns about the reliability of the reported statistics.

(4) The reported analyses and results do not seem to support the conclusion of early learning resulting in dampening and later stages in sharpening. Specifically, the authors appear to base this conclusion on the absence of a decoding effect in some time-bins, while in my opinion a contrast between time-bins, showing a difference in decoding accuracy, is required. Or better yet, a non-zero slope of decoding accuracy over time should be shown (not contingent on post-hoc and seemingly arbitrary binning).

(5) The present results both within and across trials are difficult to reconcile with previous studies using MEG (Kok et al., 2017; Han et al., 2019), single-unit and multi-unit recordings (Kumar et al., 2017; Meyer & Olson 2011), as well as fMRI (Richter et al., 2018), which investigated similar questions but yielded different results; i.e., no reversal within or across trials, as well as dampening effects with after more training. The authors do not provide a convincing explanation as to why their results should differ from previous studies, arguably further compounding doubts about the present results raised by the methods and results concerns noted above.

Impact:

At present, I find the potential impact of the study by McDermott et al. difficult to assess, given the concerns mentioned above. Should the authors convincingly answer these concerns, the study could provide meaningful insights into the mechanisms underlying perceptual prediction. However, at present, I am not entirely convinced by the quality and reliability of the results and manuscript. Moreover, the difficulty in reconciling some of the present results with previous studies highlights the need for more convincing explanations of these discrepancies and a stronger discussion of the present results in the context of the literature.

eLife Assessment

The findings are considered valuable and have theoretical implications for the interdisciplinary field of value-based social decision-making. Support for the main claims is incomplete and these should be supported by further analyses.

Reviewer #1 (Public review):

Summary:

The authors test the hypotheses, using an effort-exertion and an effort-based decision-making task, while recording brain dynamics with EEG, that the brain processes reward outcomes for effort differentially when they earned for themselves versus others.

Strengths:

The strengths of this experiment include what appears to be a novel finding of opposite signed effects of effort on the processing of reward outcomes when the recipient is self versus others. Also, the experiment is well-designed, the study seems sufficiently powered, and the data and code are publicly available.

Weaknesses:

Inferences rely heavily on the results of mixed effects models which may or may not be properly specified and are not supported by complementary analyses. Also, not all results hang together in a sensible way. For example, participants report feeling less subjective effort, but also more disliking of tasks when they were earning rewards for others versus self. Given that participants took longer to complete tasks when earning effort for others, it is conceivable that participants might have been working less hard for others versus themselves, and this may complicate the interpretation of results.

Reviewer #2 (Public review):

Summary:

Measurements of the reward positivity, an electrophysiological component elicited during reward evaluation, have previously been used to understand how self-benefitting effort expenditure influences the processing of rewards. The present study is the first to complement those measurements with electrophysiological reward after-effects of effort expenditure during prosocial acts. The results provide solid evidence that effort adds reward value when the recipient of the reward is the self but discounts reward value when the beneficiary is another individual.

Strengths:

An important strength of the study is that the amount of effort, the prospective reward, the recipient of the reward, and whether the reward was actually gained or not were parametrically and orthogonally varied. In addition, the researchers examined whether the pattern of results generalized to decisions about future efforts. The sample size (N=40) and mixed-effects regression models are also appropriate for addressing the key research questions. Those conclusions are plausible and adequately supported by statistical analyses.

Weaknesses:

Although the obtained results are highly plausible, I am concerned whether the reward positivity (RewP) and P3 were adequately measured. The RewP and P3 were defined as the average voltage values in the time intervals 300-400 ms and 300-440 ms after feedback onset, respectively. So they largely overlapped in time. Although the RewP measure was based on frontocentral electrodes (FC3, FCz, and FC4) and the P3 on posterior electrodes (P3, Pz, and P4), the scalp topographies in Figure 3 show that the RewP effects were larger at the posterior electrodes used for the P3 than at frontocentral electrodes. So there is a concern that the RewP and P3 were not independently measured. This type of problem can often be resolved using a spatiotemporal principal component analysis. My faith in the conclusions drawn would be further strengthened if the researchers extracted separate principal components for the RewP and P3 and performed their statistical analyses on the corresponding factor scores.

Reviewer #3 (Public review):

This study investigates how effort influences reward evaluation during prosocial behaviour using EEG and experimental tasks manipulating effort and rewards for self and others. Results reveal a dissociable effect: for self-benefitting effort, rewards are evaluated more positively as effort increases, while for other-benefitting effort, rewards are evaluated less positively with higher effort. This dissociation, driven by reward system activation and independent of performance, provides new insights into the neural mechanisms of effort and reward in prosocial contexts.

This work makes a valuable contribution to the prosocial behaviour literature by addressing areas that previous research has largely overlooked. It highlights the paradoxical effect of effort on reward evaluation and opens new avenues for investigating the mechanisms underlying this phenomenon. The study employs well-established tasks with robust replication in the literature and innovatively incorporates ERPs to examine effort-based prosocial decision-making - an area insufficiently explored in prior work. Moreover, the analyses are rigorous and grounded in established methodologies, further enhancing the study's credibility. These elements collectively underscore the study's significance in advancing our understanding of effort-based decision-making.

Despite these contributions, there are several gaps in the analysis that leave the conclusions incomplete and warrant further investigation. These issues can be summarized as follows:

(1) Incomplete EEG Reporting: The methods indicate that EEG activity was recorded for both tasks; however, the manuscript reports EEG results only for the first task, omitting the decision-making task. If the authors claim a paradoxical effect of effort on self versus other rewards, as revealed by the RewP component, this should also be confirmed with results from the decision-making task. Omitting these findings weakens the overall argument.

(2) Neural and Behavioural Integration: The neural results should be contrasted with behavioural data both within and between tasks. Specifically, the manuscript could examine whether neural responses predict performance within each task and whether neural and behavioural signals correlate across tasks. This integration would provide a more comprehensive understanding of the mechanisms at play.

(3) Success Rate and Model Structure: The manuscript does not clearly report the success rate in the prosocial effort task. If success rates are low, risk aversion could confound the results. Additionally, it is unclear whether the models accounted for successful versus unsuccessful trials or whether success was included as a covariate. If this information is present, it needs to be explicitly clarified. The exclusion criteria for unsuccessful trials in both tasks should also be detailed. Moreover, the decision to exclude electrodes as independent variables in the models warrants an explanation.

(4) Prosocial Decision Computational Modelling: The prosocial decision task largely replicates prior behavioural findings but misses the opportunity to directly test the hypotheses derived from neural data in the prosocial effort task. If the authors propose a paradoxical effect of effort on self-rewards and an inverse effect for prosocial effort, this could be formalised in a computational model. A model comparison could evaluate the proposed mechanism against alternative theories, incorporating the complex interplay of effort and reward for self and others. Furthermore, these parameters should be correlated with neural signals, adding a critical layer of evidence to the claims. As it is, the inclusion of the prosocial decision task seems irrelevant.

(5) Contradiction Between Effort Perception and Neural Results: Participants reported effort as less effortful in the prosocial condition compared to the self condition, which seems contradictory to the neural findings and the authors' interpretation. If effort has a discounting effect on rewards for others, one might expect it to feel more effortful. How do the authors reconcile these results? Additionally, the relationship between behavioural data and neural responses should be examined to clarify these inconsistencies.

Necessary Revisions to Manuscript: If the authors address the issues above, corresponding updates to the introduction and discussion sections could strengthen the narrative and align the manuscript with the additional analyses.

Author response:

(1) General Statements

We thank all three reviewers for their constructive comments and suggestions. We also thank reviewers #2 and #3 for considering our work to be timely and of interest to the field, not only for basic researchers, but also for translational scientists and industry. We are now providing additional results to further support our hypothesis and hope that all reviewers will find that our manuscript is now ready for publication. 

(2) Point-by-point description of the revisions

Reviewer #1 (Evidence, reproducibility and clarity (Required)): 

The manuscript by Coquel et al. investigates the effects of BKC and IBC, two compounds found in Psoralea corylifolia in DNA replication and the response to DNA damage, and explores their potential use in cancer treatment. These compounds have been previously shown to affect different cellular pathways and the authors use transformed cancer cells of different origins and a non-transformed cell line to question if their combination is toxic in cancer versus non-cancer cells. They propose that BKC inhibits DNA polymerases while IBC targets CHK2. Their results show that both compounds do affect DNA replication, inducing replication stress and affecting double strand break repair. They also show that their combined use increases their toxicity in a synergistic manner. 

However, there are some major conclusions that are still not very well supported by the data: first, the differential effect on cancer and non-transformed cells; second, the direct link of BKC to the inhibition of DNA polymerases; and third, it is unclear if CHK2 is the relevant target for IBC in this context. 

Regarding these points the authors should address the following issues: 

(1) Most of the experiments use BJ fibroblasts as a control cell line. In order to evaluate if these compounds are preferentially toxic for cancer cells, the use of more than one non-transformed cell line is necessary. In addition, BJ cells are fibroblasts while most of the cancer cell lines employed are of epithelial origin. The authors could use MCF10 and RPE cells (both of epithelial origin) as control cell lines to complement the results and better support this claim. 

We have now monitored the effect of IBC and BKC on the proliferation of MCF-7, MCF-10A and RPE-1 cells using the WST-1 assay and obtained similar results as for BJ and MCF-7 cells. These results are now included in the revised manuscript as Fig. S1A and S1B.

(2) In order to explore what are the targets of BKC and IBC Cellular Thermal Shift Assays (CETSA) could be used. Either by doing an unbiased mass spectrometry analysis of proteins stabilized by these compounds or by a direct analysis of candidate proteins by western blot (a similar approach has been used for IBC to show that it inhibits SIRT2 in Ren et al., 2024 Phytotherapy Res).

We thank this Reviewer for suggesting the use of the CETSA assay. We have now performed  CETSA on MCF-7 cells and found that IBC stabilizes CHK2 but not CHK1, to the same extent as the commercial CHK2 inhibitor BML-277 used here as a positive control. These results are now shown in new Fig. 4G and 4H.

(3) For BKC in vitro polymerase assays could be carried out to show the direct inhibition of the DNA polymerase delta, for instance. 

We have used high-speed Xenopus egg extracts to replicate ssDNA in vitro (Fig. S2C). This assay differs from the in vitro replication assay using low-speed Xenopus egg extracts (Fig. 2H) in that it only monitors elongation by replicative DNA polymerases (Pol δ and ε) and not earlier steps such as origin licensing and activation. The combined use of both low-speed and highspeed extracts strongly supports the view that BKC inhibits replicative DNA polymerases. 

To confirm this result, we have also used CETSA to monitor BKC binding to different subunits of DNA Polδ and Polε in MCF-7 cells and in Xenopus egg extracts (Fig. 3C-D Fig. S3). We found that BKC binds POLD1 and POLE, the catalytic subunits of Pol δ and ε respectively, but not the accessory subunit POLD3 nor PCNA. Together with our docking results and DNA fiber experiments, these data strongly support the view that BKC is a potent inhibitor of DNA Pol and Pol. 

(4) In addition, the authors could analyze the integrity of replication forks by PCNA immunofluorescence analysis. The colocalization of PCNA and POLD or POLE subunits could also support the role of DNA polymerases as targets of BKC. 

Our molecular docking results also show that BKC occupies the catalytic sites of DNA Pol δ and ε, which may not affect their subcellular localization and/or PCNA binding. Since our DNA replication assays, CETSA and DNA fiber analyses strongly support the view that BKC inhibits replicative DNA polymerases, we have not performed this additional experiment.

(5) In the case of IBC and the inhibition of CHK2, the authors should check the effect of IBC on the phosphorylation of BRCA1 on S988. The changes in CHK2 phosphorylation in Figure 3B are not convincing. The experiment should be repeated and the average of at least three experiments needs to be quantified. 

We now provide evidence that IBC inhibits BRCA1 phosphorylation on S988. Western blots and quantification for three biological replicates are shown in Fig. 4C and Fig. S4H. Densitometric quantification of CHK2 phosphorylation on S516 from 3 biological replicates, along with statistical analysis, is now shown in Fig. S4G.

(6) To prove that CHK2 is the relevant target for IBC the authors could test if ATM and CHK2 knockout cells are more resistant to this compound, since it would prevent the phosphorylation of CHK2. 

We have performed siRNA transfection targeting CHK2. The transfected cells died after 72 hours in culture, so we have been unable to determine whether CHK2-KD cells have increased resistance to IBC.  

In addition to these experiments, I would suggest some other major improvements in the manuscript: 

(1) The concentration of both compounds should be provided in molar units throughout the paper.

Thanks for pointing this out, we now use molar units throughout the paper.

(2) The authors do not clearly indicate the concentration that is employed in the different experiments, making it difficult to assess the results. For instance, Figure 2 does not include the concentration in the legend or in the text. Time and concentration need to be clearly shown for each experiment. 

The experimental conditions and inhibitor concentrations are now clearly indicated for each experiment.

(3) Some experiments are only repeated once (fiber assays) or twice (cell cycle analysis by flow cytometry). These experiments need to be repeated 3 times and the proper statistical analysis performed (comparison of the medians). 

Superplots with biological replicates for all DNA fiber assays are now displayed. The number of biological replicates is now indicated in the legends and appropriate statistical analyses are used.

Other minor points or suggestions: 

(1) Analyzing fork asymmetry would further support the direct effect of BKC on DNA polymerases. 

The effect of BKC on fork asymmetry is now shown in Fig. 2F. 

(2) A dose dependent analysis of BKC on the speed of DNA replication would also support this point. 

Superplots of DNA fiber assays showing the effect of different concentrations of BKC on fork speed from three biological replicates are now included in Fig. 2E.

(3) Page 7: BKC reduces fork speed ...two-fold. This sentence is not very clear, it would be better to say that speed is half of the control. 

This sentence was changed to “BKC reduced fork speed by a factor of two relative to untreated cells”.

(4) Figure 4G and S4D show contradictory results regarding the induction of Rad51 foci by IBC treatment. This needs to be clarified. 

Figure 4G and S4D (now Fig. 5G and S5D) do not show contradictory results. In both cases, IBC treatment impaired the induction of RAD51 foci by IR or bleomycin.  

(5) Page 12, Figure S5C is called for but it does not exist (probably meaning Figure S5B). 

We apologize for this error, which has now been corrected.  

Reviewer #1 (Significance): 

The work by Coquel et al. aims at elucidating the use of BKC and IBC as a combined therapy to induce cell death in cancer cells by targeting DNA replication and CHK2. Both BKC and IBC have been previously shown to affect the proliferation of cancer cells. BKC has been shown to induce S phase arrest in an ATR dependent manner in MCF7 cells (Li et al., 2016 Front Pharm), while IBC induces cell death in MDA-MB-231 cells (Wu et al., 2022 Molecules). In this regard, the more interesting contribution of the manuscript is the potential identification of the targets of these compounds in cancer cells. The inhibition of CHK2 by IBC is quite compelling although it needs to be further proven. In contrast, the hypothesis that BKC inhibits DNA polymerases remains highly speculative. The results offer a limited advance in the knowledge of the mechanism of action of these two compounds. Focusing on the action of IBC on CHK2 would increase the impact of the results. In this sense a very recent report has been published showing that IBC inhibits SIRT2 (Ren et al., 2024 Phyto Res), showing that IBC can affect multiple enzymes and processes. This should be taken into account for a further analysis of its mechanism of action. 

In addition to the identification of the targets of BKC and IBC, the authors also focus on their combination for cancer treatment. This is based on the idea that blocking the DSB repair and inducing replication stress at the same time is an efficient approach to induce cancer cell death. This is not a new concept, since the loss of ATM sensitizes cancer cells to the inhibition of the replication stress response and several combination therapies have been put forward with the idea of generating replication stress and preventing the subsequent repair of the double strand breaks induced in these cells. Thus, the novelty here is limited, especially considering that the effect of BKC on DNA replication has already been described. Further, since its mechanism of action is unclear, it is difficult to ascribe the observed synergy to the speculated hypothesis. A deeper analysis of IBC as a CHK2 inhibitor would be more interesting, and the potential combination with other chemotherapy agents such as replication stress inhibitors, HU or DNA damaging agents. Also, the lack of a good control of non-transformed cells also reduces the relevance of the work. 

In its current state, the interest of the manuscript is limited. The mechanistical advance is not strong enough and is not completely supported by the data, and the use of these compounds as a combination therapy does not provide new insights in cancer treatment. In my opinion, focusing on the inhibition of CHK2 by IBC and its potential use would broaden the impact of the results beyond the mere analysis of the action of these compounds. 

We thank this reviewer for his/her constructive and insightful comments. We have followed his/her advice and focused our analysis on the action of IBC on CHK2. Using CETSA, we confirmed that IBC binds CHK2 to the same extent as BML-277 inhibitor, but does not bind CHK1. We also show that IBC inhibits BRCA1 phosphorylation on S988 and CHK2 phosphorylation on S516. Together with the results presented in the initial version of the manuscript, these data support the view that CHK2 is a key IBC target. We have also applied CETSA to DNA polymerases and confirmed that BKC directly targets DNA Polδ and ε. Although it is unlikely that IBC and BKC will ever be used in combination therapies, the synergistic effect that we measured on cancer cells in vivo and in vitro indicates that IBC sensitizes cancer cells to endogenous replication stress and to exogenous sources of DNA damage, which could be used to replace BKC in combination therapies. For instance, our data indicate that IBC can be used in combination with drugs such as etoposide, doxorubicin or cyclophosphamide to potentiate their effect on drug-resistant lymphoma cell lines (DLBCL). As requested by this Reviewer, we have modified the discussion section to put more emphasis on IBC and CHK2 inhibitors and we hope that he/she will now find this revised version suitable for publication.

Reviewer #2 (Evidence, reproducibility and clarity): 

In the manuscript by Coquel et al., the authors report their findings on the effect of 2 natural compounds from Psoralea corylofolia plant extracts on cancer cells. They show that these compounds, bakuchiol (BKC) and isobavachalcone (IBC), inhibit proliferation of cancer cells and tumor development in xenografted mice, particularly when used in combination. They further show that BKC inhibited DNA polymerases and induced replication stress, and show evidence that IBC inhibits Chk2 kinase activity and downstream double-strand break repair. Based on their findings, the authors conclude that Chk2 inhibition and DNA replication inhibition represent a potential synergistic strategy to selecting target cancer cells. 

Major: 

(1) The data showing IBC is a Chk2 inhibitor is weak and more rigorous investigation is needed to establish this compound as a Chk2 inhibitor. 

As indicate in our response to Reviewer #1, we have now analyzed the binding of IBC to CHK2 using the Cellular Thermal Shift Assay (CETSA) in MCF-7 cells. Our data clearly show that IBC binds to CHK2 but not CHK1. These results are now shown in Fig. 4G and 4H.

For one, the authors mention they screened 43 cell cycle-related kinases in vitro, but only show data for 8 kinases in their kinase activity screens. Of these 8 kinases, Chk2 is the most strongly inhibited, but there are no data shown for the other 35 kinases. 

Data for all the protein kinases tested in the in vitro assay are now presented in Fig. S4D and S4E.  

Additionally, the purpose of the CHK2 mutants should be discussed in the text. 

The CHK2(I157T) mutation is linked to an increased risk of breast and colorectal cancers. CHK2(R145W) is associated with Li-Fraumeni Syndrome. Both mutations do not affect the basal kinase activity of CHK2. This information is now indicated in the legend of Fig. S4D. 

Secondly, the western blot in Fig 3B, appears to show a very modest effect of IBC on Chk2 autophosphorylation and not that different from the effect of IBC on Akt phosphorylation in Fig S3a. Yet, the authors claim that IBC inhibits Chk2 but not Akt. To strengthen these blots, a known Chk2 inhibitor, such as the one shown in Fig 4 (BML-277) should be included as a positive control for pChk2 similarly to what was shown for Akt with MK-2206. 

We have now replaced the western blot in Fig. 3B (now Fig. 4B) with another biological replicate. Quantifications and statistical analyses of biological replicates are shown in Fig. S4G. Overall, we observed a 50% reduction of CHK2 auto-phosphorylation in MCF7 cells treated with IBC, and a 20% reduction in AKT phosphorylation (Fig. S4A). There was no additional reduction in AKT phosphorylation when cells were treated with IBC in combination with MK-2206, compared to cells treated with MK-2206 alone. We now include the CHK2 inhibitor BML-277 as a positive control alongside with IBC to monitor CHK2 and CHK1 auto-phosphorylation in Fig. 4B, S4G, 4D and S4I, respectively.

Western blots showing a loss of phosphorylation of additional Chk2 targets is also needed. The manuscript mentions Brca1 S988 as a Chk2 substrate important for DSB repair. Showing the effect of IBC on this phosphorylation site would strengthen the conclusions. 

We now provide evidence that IBC inhibits BRCA1 phosphorylation at S988. Western blots and quantification for three biological replicates are shown in Fig. 4C and S4H. 

(2) The authors claim that the combination of IBC and BKC inhibit cell growth in a synergistic manner and that the "effect is more pronounce on cancer cells than on non-cancer cells." However, only 1 non-malignant cell line was used, and it was a fibroblast line. To make this claim, the authors need to show the effect in additional non-malignant cells, preferably with epithelial cell types. 

We have now monitored cell proliferation using the WST-1 assay in two additional non-malignant cell lines, namely MCF-10A and RPE-1 cells. Cells were treated with IBC/BKC and their growth was compared to that of MCF-7 cells. These experiments yielded similar results to those obtained with BJ fibroblasts. These new data are now included in the revised version as Fig. S1A and S1B. 

Minor: 

(1) Densitometry data for all western blots should be shown with mean+/- stdev of independent western blots. 

Densitometry data for all western blots with biological replicates are now shown in supplementary figures.

(2) In Figure 1B the statistical test used to analyze cell number was not stated. 

The statistical test is now indicated in Fig. 1B.

(3) In Figure 2A, the DAPI image for BKC is the merged image and should be replaced with just DAPI. 

This error has now been corrected.

(4) In Figure 2B, the y-axis label says "yH2AX foci (MFI)". MFI and foci are not the same thing, and for yH2AX, the signal is often not focal. MFI of yH2AX is an appropriate measurement for replication stress, it's just not appropriate to equate MFI to foci. 

We apologize for this labeling error, which has now been corrected.

(5) For the 53BP1 MFI and Rad51 MFI shown in Fig 4 and Fig S4, it is more appropriate to show the number of foci/cell as these are better indicators of breaks and repair sites. MFI is influenced by expression levels of the proteins and not necessarily the break/repair. 

The numbers of 53BP1 and RAD51 foci are now shown.

(6) The data in Figures 5B and 5C are very difficult to read. Perhaps color-coat the lines/symbols. 

We have now colored the graph to increase its readability. 

Reviewer #2 (Significance): 

The findings reported in this manuscript are timely, of interest to the field, and are mostly wellsupported by the experimental data. However, there are a few concerns that need to be addressed. 

We are grateful to Reviewer #2 for his positive assessment of our manuscript. We hope that we have adequately addressed all of his/her specific concerns and that he/she will agree with the need to put more emphasis on IBC and CHK2 inhibition as requested by Reviewer #1.

Reviewer #3 (Evidence, reproducibility and clarity): 

The manuscript: "Synergistic effect of inhibiting CHK2 and DNA replication on cancer cell growth" successfully demonstrates that the compounds BKC and IBC found in Psoralea corylifolia act synergistically to inhibit cancer cell proliferation, using a wide range of well-chosen methodologies. Moreover, the authors characterized the mechanisms of action of both drugs, which result in inhibition of cell proliferation. The use of multiple cell lines and the mice models makes the study robust and complete. The manuscript presents a well written study that offers new insights and contributions to the field. 

A few suggestions to improve the study: 

(1) Given that both compounds BKC and IBC have already been previously described in the literature, it would be helpful for the reader to have them described better at the beginning of the study. 

Thanks for pointing this out. We have now better described BKC and IBC at the beginning of the results section, as well as in the discussion. We agree that this could be helpful to readers.

(2) Addition of western blot quantifications over the number of experimental repeats is important specifically for Fig. 2C and Fig. 3C where partial effect of treatment on a signal level is reported. 

The densitometry analysis of data shown in Fig. 2C and biological replicates are now shown in Fig. S2B. Quantification for Fig. 3C (now Fig. 4D) is shown in Fig. S4I.

(3) The quantification of mean intensity for 53BP1 and RAD51 foci should be exchanged with the quantification of number of foci per cell. While the quantification of gH2AX signal intensity is a correct representation of induction of this signal upon damage, foci formed by protein recruitment to DNA damage sites should be quantified by counting the number of foci, rather than signal in the whole cell/nucleus. These proteins exist before damage and are re-located in response to the damage. 

Quantification of 53BP1 and RAD51 foci is now expressed as the number of foci per cell. 

(4) Materials & Methods section is missing the methods for the experiment described in Fig. 1B. In summary, after addressing our few concerns, we believe the manuscript should be accepted for publication. 

The WST-1 assay used for cell number quantification is included in “Reagents” in Material & Methods section.

Reviewer #3 (Significance):

The manuscript presents a well written study that offers new insights and contributions to the field. Although the inhibitors described have been known in science, the authors present convincingly their mode of action, which is either better characterized (for BKC) or inhibiting a different than previously suggested enzyme (for IBC). Authors also nicely pinpoint and explain the narrow window of concentrations when these two compounds act synergistically rather than additively. The analyses in multiple cell lines, mouse models and in combination with other cancer treatments, makes this study of interest not only for fundamental researchers but also for translational scientists and industry.

My field of expertise: DNA replication and replication stress across model systems. 

We are grateful to Reviewer #3 for his/her very positive assessment of our work and we hope that he/she will find this revised version suitable for publication.

eLife Assessment

This study presents important findings on the activity of two compounds, BKC and IBC, isolated from Psoralea corylifolia, which act synergistically to inhibit cancer cell proliferation. Using a spectrum of methods, the authors characterized the mechanisms of action of both drugs, providing convincing evidence that BKC targets DNA polymerases and IBC selectively inhibits CHK2. The study opens the possibility of improving the effectiveness of the combination of BKC and other damaging agents with IBC in cancer treatment.

[Editors' note: this paper was reviewed by Review Commons.]

Reviewer #1 (Public review):

The manuscript by Coquel et al. investigates the effects of BKC and IBC, two compounds found in Psoralea corylifolia in DNA replication and the response to DNA damage, and explores their potential use in cancer treatment. These compounds have been previously shown to affect different cellular pathways and the authors use transformed cancer cells of different origins and a non-transformed cell line to question if their combination is toxic in cancer versus non-cancer cells. They propose that BKC inhibits DNA polymerases while IBC targets CHK2. Their results show that both compounds do affect DNA replication, inducing replication stress and affecting double strand break repair. They also show that their combined use increases their toxicity in a synergistic manner.

Comments on current version:

The authors have addressed the main questions raised in the original manuscript. The new data provide stronger evidence supporting the inhibition of DNA polymerases by BKC and the effect of IBC on CHK2. In addition, the new data provides information about the potential mechanism of action of IBC in cells and xenograft models. Together, the revised manuscript has notably increased the relevance and impact of the results with stronger conclusions and better controlled experiments.

Reviewer #2 (Public review):

Summary:

The manuscript: "Synergistic effect of inhibiting CHK2 and DNA replication on cancer cell growth" successfully demonstrates that the compounds BKC and IBC found in Psoralea corylifolia act synergistically to inhibit cancer cell proliferation, using a wide range of well-chosen methodologies. Moreover, the authors characterized the mechanisms of action of both drugs, which result in inhibition of cell proliferation. The use of multiple cell lines and the mice models makes the study robust and complete.

Significance:

The manuscript presents a well written study that offers new insights and contributions to the field. Although the inhibitors described have been known in science, the authors present convincingly their mode of action, which is either better characterized (for BKC) or inhibiting a different than previously suggested enzyme (for IBC). Authors also nicely pinpoint and explain the narrow window of concentrations when these two compounds act synergistically rather than additively. The analyses in multiple cell lines, mouse models and in combination with other cancer treatments, make this study of interest not only for fundamental researchers but also for translational scientists and industry.

eLife Assessment

This valuable study uses AlphaFold2 to guide the structural modelling of different states of the human voltage-gated potassium channel KV11.1, a key pharmacological drug target. Follow-up molecular dynamics and drug-docking simulations, combined with experimental comparisons, offer convincing evidence supporting the models, showing that drugs bind more effectively to the inactivated state. The work shows potential for improving drug potency predictions in ion channel pharmacology, though its applicability to other systems remains uncertain.

Reviewer #1 (Public review):

Summary:

Ngo et. al use several computational methods to determine and characterize structures defining the three major states sampled by the human voltage-gated potassium channel hERG: the open, closed, and inactivated state. Specifically, they use AlphaFold and Rosetta to generate conformations that likely represent key features of the open, closed, and inactivated states of this channel. Molecular dynamics simulations confirm that ion conduction for structure models of the open but not the inactivated state. Moreover, drug docking in silico experiments show differential binding of drugs to the conformation of the three states; the inactivated one being preferentially bound by many of them. Docking results are then combined with a Markov model to get state-weighted binding free energies that are compared with experimentally measured ones.

Strengths:

The study uses state-of-the art modeling methods to provide detailed insights into the structure-function relationship of an important human potassium channel. AlphaFold modeling, MD simulations, and Markov modeling are nicely combined to investigate the impact of structural changes in the hERG channel on potassium conduction and drug binding.

Weaknesses:

(1) The selection of inactivated conformations based on AlphaFold modeling seems a bit biased. The authors base their selection of the "most likely" inactivated conformation on the expected flipping of V625 and the constriction at G626 carbonyls. This follows a bit of the "Streetlight effect". It would be better to have selection criteria that are independent of what they expect to find for the inactivated state conformations. Using cues that favour sampling/modeling of the inactivated conformation, such as the deactivated conformation of the VSD used in the modeling of the closed state, would be more convincing. There may be other conformations that are more accurately representing the inactivated state. I see no objective criteria that justify the non-consideration of conformations from cluster 3 of the inactivated state modeling. I am not sure whether pLDDT is a good selection criterion. It reports on structural confidence, but that may not relate to functional relevance.

(2) The comparison of predicted and experimentally measured binding affinities lacks an appropriate control. Using binding data from open-state conformations only is not the best control. A much better control is the use of alternative structures predicted by AlphaFold for each state (e.g. from the outlier clusters or not considered clusters) in the docking and energy calculations. Using these docking results in the calculations would reveal whether the initially selected conformations (e.g. from cluster 2 for the inactivated state) are truly doing a better job in predicting binding affinities. Such a control would strengthen the overall findings significantly.

(3) Figures where multiple datapoints are compared across states generally lack assessment of the statistical significance of observed trends (e,g. Figure 3d).

(4) Figure 3 and Figures S1-S4 compare structural differences between states. However, these differences are inferred from the initial models. The collection of conformations generated via the MD runs allow for much more robust comparisons of structural differences.

Reviewer #2 (Public review):

Summary:

Ngo et al. use AlphaFold2 and Rosetta to model closed, open, and inactive states of the human ion channel hERG. Subsequent MD simulations and comparisons with experiments support the plausibility of their models.

Strengths:

This is thorough work studied from many different angles. It provides a self-consistent picture of how conformational changes in hERG may affect its function and binding to different targets.

Weaknesses:

Though this work claims the methodologies can be generalized to other systems, it is not obvious how. Many modeling choices seem arbitrary and also seem to have required extensive expert knowledge of the system. This limits the applicability of the modeling strategy.

Reviewer #3 (Public review):

Summary:

The authors use Alphafold2, Rosetta, and Molecular Dynamics to model structures of the hERG K channel in open, inactive, and closed states. Experimental CryoEM data for open hERG (Wang and Mackinnon 2017), and closed EAG (Mandala and Mackinnon, 2002) were used as the main templates for channel models presented here. Given the importance of hERG as a safety pharmacology target, the identification of a robust simulation method to assess drug block is an important addition to the field.

Strengths

The key findings here are new inactivated and closed hERG channel conformations and hERG channel conformations with drugs docked in the inner vestibule below the selectivity filter. Amino acid pathways and interaction networks for different states are also presented.

The inactive state and drug block models are carefully correlated with experimental data for the inactivated state of hERG (Lau et al, 2024) and with experimental free energy data for drug binding and have overall good agreement.

It is remarkable that using cytoplasmic domain structures of hERG as a starting point revealed inactivation state structures in the hERG selectivity filter in Figures 2,3.

Weaknesses

Figure 6, if each data point is for a different drug, then perhaps identify each point.

The PAS domain was not included in the models as stated in Methods page 14 but the PAS does appear in some of the templates used as starting points for models in Figure 1 a,b,c. Perhaps mentioning that the PAS was not included in some (all?) of the final models should be moved into the main text and discussed.

The drug block of 1b channels (which do not contain PAS) has been reported to be slightly different than that for 1a channels (which contain PAS) and for 1a/1b channels (see London et al., 1997; https://doi.org/10.1161/01.RES.81.5.870 and Abi-Gerges et. al., 2011; DOI: 10.1111/j.1476-5381.2011.01378.x) and this should be discussed since the models presented here appear to be performed in the absence of the PAS.

It also appears that the N-linker region (between PAS and the S1) and distal C region of hERG (post CNBHD-COOH) are not included in models, please state this if correct, and discuss.

Author response:

Reviewer #1: 

Summary:

Ngo et. al use several computational methods to determine and characterize structures defining the three major states sampled by the human voltage-gated potassium channel hERG: the open, closed, and inactivated state. Specifically, they use AlphaFold and Rosetta to generate conformations that likely represent key features of the open, closed, and inactivated states of this channel. Molecular dynamics simulations confirm that ion conduction for structure models of the open but not the inactivated state. Moreover, drug docking in silico experiments show differential binding of drugs to the conformation of the three states; the inactivated one being preferentially bound by many of them. Docking results are then combined with a Markov model to get state-weighted binding free energies that are compared with experimentally measured ones.

Strengths:

The study uses state-of-the art modeling methods to provide detailed insights into the structure-function relationship of an important human potassium channel. AlphaFold modeling, MD simulations, and Markov modeling are nicely combined to investigate the impact of structural changes in the hERG channel on potassium conduction and drug binding.

We appreciate the reviewer’s recognition of our integration of state-of-the-art computational methods, including AlphaFold2, Rosetta, MD simulations, and Markov modeling. We are pleased that the reviewer found our approach to investigating the structure-function relationship of the hERG channel insightful.

Weaknesses:

(1) The selection of inactivated conformations based on AlphaFold modeling seems a bit biased. The authors base their selection of the "most likely" inactivated conformation on the expected flipping of V625 and the constriction at G626 carbonyls. This follows a bit of the "Streetlight effect". It would be better to have selection criteria that are independent of what they expect to find for the inactivated state conformations. Using cues that favour sampling/modeling of the inactivated conformation, such as the deactivated conformation of the VSD used in the modeling of the closed state, would be more convincing. There may be other conformations that are more accurately representing the inactivated state. I see no objective criteria that justify the non-consideration of conformations from cluster 3 of the inactivated state modeling. I am not sure whether pLDDT is a good selection criterion. It reports on structural confidence, but that may not relate to functional relevance.

We acknowledge the concern regarding the selection criteria for the inactivated state models. In the revised manuscript version, we plan to broaden our selection approach and explicitly include conformations from different clusters beyond those highlighted in the initial submission (e.g., from cluster 3). We will also incorporate structural metrics that do not solely depend on the known channel inactivation hallmarks or reply on the pLDDT scores to further justify our chosen representative inactivated state models.

(2) The comparison of predicted and experimentally measured binding affinities lacks an appropriate control. Using binding data from open-state conformations only is not the best control. A much better control is the use of alternative structures predicted by AlphaFold for each state (e.g. from the outlier clusters or not considered clusters) in the docking and energy calculations. Using these docking results in the calculations would reveal whether the initially selected conformations (e.g. from cluster 2 for the inactivated state) are truly doing a better job in predicting binding affinities. Such a control would strengthen the overall findings significantly.

We agree that a more rigorous control for our drug-binding predictions is desirable. To address this, we will include molecular docking simulations and associated drug binding affinity estimations for more hERG channel models, including alternate conformations from the initial clustering that were not chosen as the final models. This will allow us to test whether our inactivated state structure from cluster 2 indeed outperforms or differs significantly from other possible inactivated hERG channel conformations in reproducing experimental drug potencies.

(3) Figures where multiple datapoints are compared across states generally lack assessment of the statistical significance of observed trends (e,g. Figure 3d).

(4) Figure 3 and Figures S1-S4 compare structural differences between states. However, these differences are inferred from the initial models. The collection of conformations generated via the MD runs allow for much more robust comparisons of structural differences.

We will incorporate statistical analyses and measures of uncertainty for key comparisons. In Figures 3 and S1-S4 the consensus structural hERG channel models for open, inactivated and closed states are being compared, i.e. one representative model for each state. We believe this is a valid comparison, and the statistical analysis of the observed trends based on those models (e.g., in the bar plot of Figure 3d) alone might not be possible. However, we agree with the reviewer that instead of relying solely on those initial static models, we will also draw on the ensemble of states sampled during the MD simulations to quantify structural differences between different putative hERG channel states. Specifically, we will present ensemble-averaged measurements and highlight how these distributions differ significantly between states.

Reviewer #2:

Summary:

Ngo et al. use AlphaFold2 and Rosetta to model closed, open, and inactive states of the human ion channel hERG. Subsequent MD simulations and comparisons with experiments support the plausibility of their models.

Strengths:

This is thorough work studied from many different angles. It provides a self-consistent picture of how conformational changes in hERG may affect its function and binding to different targets.

We are grateful for the reviewer’s recognition of the thoroughness and multi-faceted nature of our study.

Weaknesses:

Though this work claims the methodologies can be generalized to other systems, it is not obvious how. Many modeling choices seem arbitrary and also seem to have required extensive expert knowledge of the system. This limits the applicability of the modeling strategy.

We appreciate the reviewer’s comment on the generalizability of our approach. In the revision, we will more explicitly discuss the rationale behind the modeling choices and the extent to which they reflect system-specific knowledge. We will clarify how the strategies we developed (e.g., iterative refinement with AlphaFold2 and Rosetta, followed by MD simulation validation) can be adapted to other ion channels or related proteins. We will also outline a more generalizable workflow, specifying which steps require system-specific information and which steps are broadly applicable.

Reviewer #3:

Summary:

The authors use Alphafold2, Rosetta, and Molecular Dynamics to model structures of the hERG K channel in open, inactive, and closed states. Experimental CryoEM data for open hERG (Wang and Mackinnon 2017), and closed EAG (Mandala and Mackinnon, 2002) were used as the main templates for channel models presented here. Given the importance of hERG as a safety pharmacology target, the identification of a robust simulation method to assess drug block is an important addition to the field.

Strengths

The key findings here are new inactivated and closed hERG channel conformations and hERG channel conformations with drugs docked in the inner vestibule below the selectivity filter. Amino acid pathways and interaction networks for different states are also presented.

The inactive state and drug block models are carefully correlated with experimental data for the inactivated state of hERG (Lau et al, 2024) and with experimental free energy data for drug binding and have overall good agreement.

It is remarkable that using cytoplasmic domain structures of hERG as a starting point revealed inactivation state structures in the hERG selectivity filter in Figures 2,3.

We thank the reviewer for highlighting the novelty and importance of our work, particularly regarding the identification of new inactivated and closed hERG channel conformations and the modeling of drug block. We are also pleased that the reviewer found the correlation with experimental data to be strong and the structural insights to be valuable.

Weaknesses

Figure 6, if each data point is for a different drug, then perhaps identify each point.

Thank you so much for this suggestion. Please note that Table 3 contains drug-specific data plotted in Figure 6 including drug names. We will provide a reference to Table 3 in the revised Figure 6 caption. We will also revise Figure 6 (and any similar figures) to clearly identify each data point with the corresponding drug and/or include a corresponding key in the Figure legend. This will make it easier to correlate each data point’s binding prediction with the experimental datasets.

The PAS domain was not included in the models as stated in Methods page 14 but the PAS does appear in some of the templates used as starting points for models in Figure 1 a,b,c. Perhaps mentioning that the PAS was not included in some (all?) of the final models should be moved into the main text and discussed.

The drug block of 1b channels (which do not contain PAS) has been reported to be slightly different than that for 1a channels (which contain PAS) and for 1a/1b channels (see London et al., 1997; https://doi.org/10.1161/01.RES.81.5.870 and Abi-Gerges et. al., 2011; DOI: 10.1111/j.1476-5381.2011.01378.x) and this should be discussed since the models presented here appear to be performed in the absence of the PAS.

It also appears that the N-linker region (between PAS and the S1) and distal C region of hERG (post CNBHD-COOH) are not included in models, please state this if correct, and discuss.

We appreciate the reviewer’s insightful comment regarding the PAS domain and the potential influence of other regions, such as the N-linker and distal C-region, on hERG channel drug binding and state transitions.

The PAS domain did appear in the starting templates used for initial structural modeling (as shown in Figure 1a, b, c), but it was not included in the final models used for subsequent analyses. Similarly, the N-linker and the distal C-region were also omitted from the final models. These omissions were primarily due to hardware constraints used for AlphaFold structural modeling, as including these additional protein regions would exceed the memory capacity of graphical processing unit (GPU) cards on our available intramural, external and cloud high-performance computing resources, leading to failures during the protein structure prediction step.

The PAS domain of hERG 1a isoform, even if not serving as a direct drug-binding site, can influence the gating kinetics of hERG channels as the reviewer pointed out. By altering the probability and duration with which those ion channels occupy specific conformational states, it can indirectly affect how well drugs bind. For example, if the presence of the PAS domain shifts channel gating so that more channels enter (and remain in) the inactivated state, drugs with a higher affinity for that state would appear to bind more potently, as observed in electrophysiological experiments. It is also plausible that the PAS domain could exert allosteric effects that alter the conformational landscape of the ion channel during gating transitions, potentially impacting drug accessibility or binding stability. This is an intriguing hypothesis and an important avenue for future research.

With access to more powerful computational resources, it would be valuable to explore the full-length hERG 1a channel, including the PAS domain and associated regions, to assess their potential contributions to drug binding and gating dynamics. We will incorporate a discussion of these points into the main text, acknowledging the limitations of our current models, citing the references provided by the reviewer, and highlighting the need for future studies to explore these protein regions in greater detail.

eLife Assessment

This is an important technical method paper that details the development and quality assessment of a 3D MERFISH method to enable spatial transcriptomics of thick tissues, representing a major step forward in the technical capacity of the MERFISH. The evidence presented is convincing.

Reviewer #1 (Public review):

Summary:

The study by Fang et al. reports a 3D MERFISH method that enable spatial transcriptomics for tissues up to 200um in thickness. MERFISH as well other spatial transcriptomics technologies have been mainly used for thin (e.g, 10um) tissue slices, which limits the dimension of spaital transcriptomics technique. Therefore, expanding the capacity of MERFISH to thick tissues represents a major technical advance to enable 3D spatial transcriptomics. Here the authors provide detailed technical descriptions of the new method, troubleshooting, optimization, and application examples to demonstrate its technical capacity, accuracy, sensitivity, and utility. The method will likely have major impact on future spatial transcriptomics studies to benefit diverse biomedical fields.

Strengths:

The study was well-designed, executed, and presented. Extensive protocol optimization and quality assessments were carried out and conclusions are well supported by the data. The methods were sufficiently detailed and the results are solid and compelling.

Weaknesses:

Thorough performance comparison with other existing technologies can be done in the future.

Comments on revisions:

The authors have sufficiently addressed the previous comments.

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Summary: 

The study by Fang et al. reports a 3D MERFISH method that enables spatial transcriptomics for tissues up to 200um in thickness. MERFISH, as well as other spatial transcriptomics technologies, have been mainly used for thin (e.g, 10um) tissue slices, which limits the dimension of spatial transcriptomics technique. Therefore, expanding the capacity of MERFISH to thick tissues represents a major technical advance to enable 3D spatial transcriptomics. Here the authors provide detailed technical descriptions of the new method, troubleshooting, optimization, and application examples to demonstrate its technical capacity, accuracy, sensitivity, and utility. The method will likely have a major impact on future spatial transcriptomics studies to benefit diverse biomedical fields. 

Strengths: 

The study was well-designed, executed, and presented. Extensive protocol optimization and quality assessments were carried out and conclusions are well supported by the data. The methods were sufficiently detailed, and the results are solid and compelling. 

Response: We thank the reviewer for the positive comments on our manuscript.  

Weaknesses: 

The biological application examples were limited to cell type/subtype classification in two brain regions. Additional examples of how the data could be used to address important biological questions will enhance the impact of the study. 

We appreciate the reviewer's suggestion that demonstrating the broader applications of our thick-tissue 3D MERFISH method to address important biological questions would enhance the impact of our study. In line with the reviewer's feedback, we have included discussions on how this method could be applied to address various biological questions in the summary (last) paragraph of our manuscript. These discussions highlight the versatility and utility of our approach in studying diverse biological processes beyond cell type classification. 

However, the goal of this work is to develop a method and establish its validity. While we are interested in applying it to addressing important biological questions in the future, we consider these applications beyond the scope of this work. 

Reviewer #2 (Public Review): 

Summary: 

In their preprint, Fang et al present data on extending a spatial transcriptomics method, MERFISH, to 3D using a spinning disc confocal. MERFISH is a well-established method, first published by Zhuang's lab in 2015 with multiple follow-up papers. In the last few years, MERFISH has been used by multiple groups working on spatial transcriptomics, including approximately 12 million cell maps measured in the mouse brain atlas project. Variants of MERFISH were used to map epigenetic information complementary to gene expression and RNA abundance. However, MERFISH was always limited to thin ~10um sections to this date.

The key contribution of this work by Fang et al. was to perform the optimization required to get MERFISH working in thick (100-200um) tissue sections. 

Major strengths and weaknesses: 

Overall the paper presents a technical milestone, the ability to perform highly multiplexed RNA measurements in 3D using MERFISH protocol. This is not the first spatial transcriptomics done in thick sections. Wang et al. 2018 - StarMAP used thick sections (150 um), and recently, Wang 2021 (EASI-FISH, not cited) performed serial HCR FISH on 300um sections. Data so far suggest that MERFISH has better sensitivity than in situ sequencing approaches (StarMAP) and has built-in multiplexing that EASI-FISH lacks. Therefore, while there is an innovation in the current work, i.e., it is a technically challenging task, the novelty, and overall contribution are modest compared to recently published work.  

The authors could improve the writing and the manuscript text that places their work in the right context of other spatial transcriptomics work. Out of the 25 citations, 12 are for previous MERFISH work by Zhuang's lab, and only one manuscript used a spatial transcriptomics approach that is not MERFISH. Furthermore, even this paper (Wang et al, 2018) is only discussed in the context of neuroanatomy findings. The fact that Wang et al. were the first to measure thick sections is not mentioned in the manuscript. The work by Wang et al. 2021 (EASI-FISH) is not cited at all, as well as the many other multiplexed FISH papers published in recent years that are very relevant. For example, a key difference between seqFISH+ and MERFISH was the fact that only seqFISH+ used a confocal microscope, and MERFISH has always been relying on epi. As this is the first MERFISH publication to use confocal, I expect citations to previous work in seqFISH and better discussions about differences. 

We thank the reviewer for recognizing our work as a technical milestone. Since the aim of this work is to build upon the strengths of MERFISH and address some of its limitations, we primarily cited previous MERFISH papers to clarify the specific improvements made in this work. Given the rapid growth of the spatial omics field, it has become impractical to comprehensively cite all method development papers. Instead, we cited a 2021 review article in the first sentence of the originally submitted manuscript and limited all discussions afterwards to MERFISH. In light of this reviewer’s suggestion to more broadly cite spatial transcriptomics work, we added two additional review articles on spatial omics. Spatial omics methods primarily include two categories: 1) imaging-based methods and 2) next-generation-sequencing based methods. The 2021 review article [Zhuang, Nat Methods 18,18–22 (2021)) included in the originally submitted manuscript is focused on imaging-based methods. The additional 2021 review article [Larsson et al., Nat Methods 18, 15–18 (2021)] that we now included in the revised manuscript is focused on next-generation-sequencing based methods. We also added a more recent review article published in 2023 [Bressan et al., Science 381:eabq4964 (2023)], which covers both categories of methods and include more recent technology developments. All three review articles are now cited in parallel in the first introductory paragraph of the manuscript.

Although we presented our work as an advance in MERFISH specifically, we do consider the reviewer’s suggestion of citing the 2018 STARmap paper [Wang et al., Science 361, eaat5961 (2018)] in the introduction part of our manuscript reasonable. This STARmap paper was already cited in the results part of our originally submitted manuscript, and we have now described this work in the introduction part of our revised manuscript (third paragraph), as this paper was the first to demonstrate 3D in situ sequencing in thick tissues. In addition, we thank the reviewer for bringing to our attention the EASI-FISH paper [Wang et al, Cell 184, 6361-6377 (2021)], which reported a method for thick-tissue FISH imaging and demonstrated imaging of 24 genes using multiple rounds of multi-color FISH imaging. We also recently became aware of a paper reporting 3D imaging of thick samples using PHYTOMap [Nobori et al, Nature Plants 9, 10261033 (2023)]. This paper, published a few days after we submitted our manuscript to eLife, demonstrated imaging of 28 genes in thick plant samples using multiple rounds of multicolor FISH and the probe targeting and amplification methods previously developed for in situ sequencing. We also included these two papers in the introduction section of our revised manuscript (third paragraph). In addition, we also expanded the discussion paragraph (last paragraph) of the manuscript to discuss these thick tissue imaging methods in more details, and in the same paragraph, we also included discussions on two recent bioRxiv preprints in thicktissue transcriptomic imaging [Gandin et al., bioRxiv, doi:10.1101/2024.05.17.594641 (2024); Sui et al., bioRxiv, doi:10.1101/2024.08.05.606553 (2024)]

However, we do not consider our use of confocal imaging in this work an advance in MERFISH because confocal microscopy, like epi-fluorescence imaging, is a commonly used approach that could be applied to MERFISH of thin tissues directly without any alteration of the protocol. Confocal imaging has been broadly used for both DNA and RNA FISH before any genomescale imaging was reported. Confocal and epi-imaging geometries have their distinct advantages, and which of these imaging geometries to use is the researcher’s choice depending on instrument availability and experimental needs. Thus, we do not find it necessary to cite specific papers just for using confocal imaging in spatial transcriptomic profiling. Our real advance related to confocal imaging is the use of machine-learning to increase the imaging speed. Without this improvement, 3D imaging of thick tissue using confocal would take a long time and likely degrade image quality due to photobleaching of out-of-focus fluorophores before they are imaged. We thus cited several papers that used deep learning to improve imaging quality and/or speed [(Laine et al., International Journal of Biochemistry & Cell Biology 140:106077 (2021); Ouyang et al., Nat Biotechnol 36:460–468 (2018); Weigert et al., Nat Methods 15:1090–1097 (2018)] in our original submission. Our unique contribution is the combination of machine learning with confocal imaging for 3D multiplexed FISH imaging of thick tissue samples, which had not been demonstrated previously.

To get MERFISH working in 3D, the authors solved a few technical problems. To address reduced signal-to-noise due to thick samples, Fang et al. used non-linear filtering (i.e., deep learning) to enhance the spots before detection. To improve registrations, the authors identified an issue specific to their Z-Piezo that could be improved and replaced with a better model. Finally, the author used water immersion objectives to mitigate optical aberrations. All these optimization steps are reasonable and make sense. In some cases, I can see the general appeal (another demonstration of deep learning to reduce exposure time). Still, in other cases, the issue is not necessarily general enough (i.e., a different model of Piezo Z stage) to be of interest to a broad readership. There were a few additional optimization steps, i.e., testing four concentrations of readout and encoder probes. So while the preprint describes a technical milestone, achieving this milestone was done with overall modest innovation. 

We appreciate the reviewer's recognition of the technical challenges we have overcome in developing this 3D thick-tissue MERFISH method. To achieve high-quality thick- tissue MERFISH imaging, we had to overcome multiple different challenges. We agree with the reviewer that the solutions to some of the above challenges are intellectually more impressive than the remaining ones that required relatively more mundane efforts. However, all of these are needed to achieve the overall goal, a goal that is considered a milestone by the reviewer.  We believe that the impact of a method should be evaluated based on its capabilities, potential applications, and its adaptability for broader adoption. In this regard, we anticipate that our reported method will be valuable and impactful contribution to the field of spatial biology.

Data and code sharing - the only link in the preprint related to data sharing sends readers to a deleted Dropbox folder. Similarly, the GitHub link is a 404 error. Both are unacceptable. The author should do a better job sharing their raw and processed data. Furthermore, the software shared should not be just the MERlin package used to analyze but the specific code used in that package.  

We shared the data through Dropbox as a temporary data-sharing approach for the review process, because of the potential needs to revise and/or add data during the paper revision process. We have now made all data publicly available at Dryad (https://doi.org/10.5061/dryad.w0vt4b922).

The GitHub link that we provided for the MERlin package was valid and when we clicked on it, it took us to the correct GitHub site. However, to make the code a permanent record, we also deposited the code to Zenodo (https://zenodo.org/records/13356944). Moreover, following the suggestion by the reviewer, in addition to the MERlin v2.2.7 package itself, we have also shared the specific code to utilize this package for analyzing the data taken in this work at Dryad (https://doi.org/10.5061/dryad.w0vt4b922). 

Recommendations For The Authors:

Reviewer #1 (Recommendations For The Authors): 

(1) It will be good to expand the application section to demonstrate the utility of 3D MERFISH to address diverse types of biological questions for the two brain regions examined. At present, it only examined the localization of various cell clusters in the tissues. Can it be used to examine both short and long-range interactions, for example? 

We appreciate the reviewer's feedback and agree that demonstrating the broader applications of our 3D thick-tissue MERFISH imaging method in addressing diverse biological questions would enhance the impact of our study.  

In line with the reviewer’s comments, one of the analyses we performed in the manuscript was examining short-range interactions based on soma contact between adjacent neurons in the two brain regions studied (see third-to-last and second-to-last paragraphs of the Main text). This analysis provided insights into the spatial organization of inhibitory neurons and potential interactions between the same type of interneurons in these brain regions. 

Although long-range interactions, for example synaptic interactions between neurons, would be of great interest, our current 3D MERFISH measurements does not allow such interactions to be determined. Future research to enable measurements of synaptic interactions between molecularly defined neuronal subtypes would be interesting, but we consider this to be out of the scope of the current study.

(2) For the nearest neighbor distance analysis in Figure 3, the method seems to be missing. Please add details about this analysis to allow better understanding. It is counterintuitive that the cell subtypes showed tight local distribution (Figure 3 - supplement 3), but the nearest neighbor distances with subtypes are not different from those between subtypes. Please explain. 

We apologize for the missing the nearest neighbor distance analysis in the Materials and Methods section.  We have added the detailed description of this analysis to the Materials and Methods section of the revised manuscript (last subsection of Materials and Methods).

Regarding the comment “It is counterintuitive that the cell subtypes showed tight local distribution (Figure 3 - supplement 3), but the nearest neighbor distances with subtypes are not different from those between subtypes”, this is not necessarily counter-intuitive given how we defined nearest-neighbor distances between the same subtype of neurons and nearestneighbor distances between different subtypes of neurons. Here is how we performed this analysis for interneurons. First, we determined the nearest-neighbor neurons for each interneuron and classified it as either having another interneuron of the same type as the nearest neighbor or having a different type of interneuron or an excitatory neuron as the nearest neighbor. We then determine the distributions for the distances between these two types of nearest neighbors and compared these distributions. When a neuronal subtype for a tight spatial cluster, such as the type-A cluster shown in the schematic below, the nearest-neighbor distances between nearest neighbor A-A pairs are indeed small. However, the distance between a type-A neuron and a different type of neurons (for example, type-B) is not necessarily bigger than those between two type-A neurons, if the nearest neighbor cell for this type-A neuron is a type-B neuron. These nearest-neighbor A-B pairs are likely formed between type-A neurons at the edge of the cluster with type-B neurons near the edge of the type-A cluster. If the distance of an A-B pair is not comparable to those of nearest-neighbor A-A pairs, it is unlikely a nearestneighbor pair by our definition as described above.

Author response image 1.

Reviewer #2 (Recommendations For The Authors): 

(1) The scholarship in this work is lacking. All of the non-MERFISH parts of the field of spatial transcriptomics are ignored. The work needs to be discussed in the context of the literature. 

We thank the reviewer for this suggestion and have included discussions of other spatial omics work, and other thick-tissue multiplexed imaging work in the Introduction and discussion section of the manuscript. Please see details in our response to the Public Review  portion of this reviewer’s comments.  

(2) The data/code sharing links are broken and need to be fixed. 

Response: We shared the data through Dropbox as a temporary data-sharing approach for the review process, because of the potential needs to revise and/or add data during the paper revision process We have now placed all data publicly available at Dryad (https://doi.org/10.5061/dryad.w0vt4b922). 

The GitHub link that we provided for the MERlin package was valid and when we clicked on it, it took us to the correct GitHub site. However, to make the code a permanent record, we also deposited the code to Zenodo (https://zenodo.org/records/13356944). Moreover, following the suggestion by the reviewer, in addition to the MERlin (MERFISH decoding package itself), we have also shared the specific code to utilize this package for analyzing the data taken in this work at Dryad (https://doi.org/10.5061/dryad.w0vt4b922) to ensure that the readers can fully reproduce the results presented in our manuscript.

eLife Assessment

This paper advances an important new concept in psoriasis pathogenesis and implicates Sema4a as a homeostatic regulator that is highly epithelial-specific. The findings are convincing and lend support for the biology described here as a mechanism with therapeutic implications.

Reviewer #1 (Public review):

Summary:

In this study, Kume et al examined the role of the protein Semaphorin 4a in steady state skin homeostasis and how this relates to skin changes seen in human psoriasis and imiquimod-induced psoriasis-like disease in mice. The authors found that human psoriatic skin has reduced expression of Sema4a in the epidermis. While Sema4a has been shown to drive inflammatory activation in different immune populations, this finding suggested Sema4a might be important for negatively regulating Th17 inflammation in the skin. The authors go on to show that Sema4a knockout mice have skin changes in key keratinocyte genes, increased gdT cells, and increased IL-17 similar to differences seen in non-lesional psoriatic skin, and that bone marrow chimera mice with WT immune cells and Sema4a KO stromal cells develop worse IMQ-induced psoriasis-like disease, further linking expression of Sema4a in the skin to maintaining skin homeostasis. The authors next studied downstream pathways that might mediate the homeostatic effects of Sema4a, focusing on mTOR given its known role in keratinocyte function. Like for the immune phenotypes, Sema4a KO mice had increased mTOR activation in the epidermis in a similar pattern to mTOR activation noted in non-lesional psoriatic skin. The authors next targeted the mTOR pathway and showed rapamycin could reverse some of the psoriasis-like skin changes in Sema4a KO mice, confirming the role of increased mTOR in contributing to the observed skin phenotype.

In the revised manuscript, the authors expand on the potential relevance to psoriasis by demonstrating similar findings in an IL-23-diriven model of skin inflammation, which is an orthogonal model of psoriasis to their original IMQ model. They also show that in addition to reversing steady state differences in skin thickness between Sema4a KO mice and WT mice, rapamycin improves metrics of disease in the IMQ model of psoriasis. These additional studies further bolster their conclusions that Sema4a may play a protective role in by preventing over-activation of mTOR in the skin in psoriasis.

Strengths:

The most interesting finding is the tissue-specific role for Sema4a, where it has previously been considered to play a mostly pro-inflammatory role in immune cells, this study shows that when expressed by keratinocytes, Sema4a plays a homeostatic role that when missing leads to development of psoriasis-like skin changes. This has important implications in terms of targeting Sema4a pharmacologically. It also may yield a novel mouse model to study mechanisms of psoriasis development in mice separate from the commonly used IMQ model. The included experiments are well-controlled and executed rigorously.

The new experiments provide additional data to support the conclusions through an orthogonal model of psoriasis and demonstrating rapamycin-induced reversal of changes in the IMQ disease model.

Weaknesses:

While the main weakness of these studies, lack of tissue-specific Sema4a knockout mice (e.g. in keratinocytes only), remains, generating these mice and performing the necessary experiments is beyond the scope of completing these particular studies. Similarly, it is understandable that additional bone marrow chimeras would be costly and labor intensive without adding much more in the absence of tissue-specific knockouts.

Reviewer #2 (Public review):

Summary:

Kume et al. found for the first time that Semaphorin 4A (Sema4A) was downregulated in both mRNA and protein levels in L and NL keratinocytes of psoriasis patients compared to control keratinocytes. In peripheral blood, they found that Sema4A is not only expressed in keratinocytes but is also upregulated in hematopoietic cells such as lymphocytes and monocytes in the blood of psoriasis patients. They investigated how the down-regulation of Sema4A expression in psoriatic epidermal cells affects the immunological inflammation of psoriasis by using a psoriasis mice model in which Sema4A KO mice were treated with IMQ. Kume et al. hypothesized that down-regulation of Sema4A expression in keratinocytes might be responsible for the augmentation of psoriasis inflammation. Using bone marrow chimeric mice, Kume et al. showed that KO of Sema4A in non-hematopoietic cells was responsible for the enhanced inflammation in psoriasis. The expression of CCL20, TNF, IL-17, and mTOR was upregulated in the Sema4AKO epidermis compared to the WT epidermis, and the infiltration of IL-17-producing T cells was also enhanced.

Strengths:

Decreased Sema4A expression may be involved in psoriasis exacerbation through epidermal proliferation and enhanced infiltration of Th17 cells, which helps understand psoriasis immunopathogenesis.

Weaknesses:

The mechanism of decreased Sema4A expression in psoriasis is not clear, although this does not affect the strength of this research.

Author response:

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

Public reviews:

Reviewer #1 (Public Review):

Summary:

In this study, Kume et al examined the role of the protein Semaphorin 4a in steady-state skin homeostasis and how this relates to skin changes seen in human psoriasis and imiquimod-induced psoriasis-like disease in mice. The authors found that human psoriatic skin has reduced expression of Sema4a in the epidermis. While Sema4a has been shown to drive inflammatory activation in different immune populations, this finding suggested Sema4a might be important for negatively regulating Th17 inflammation in the skin. The authors go on to show that Sema4a knockout mice have skin changes in key keratinocyte genes, increased gdT cells, and increased IL-17 similar to differences seen in non-lesional psoriatic skin, and that bone marrow chimera mice with WT immune cells and Sema4a KO stromal cells develop worse IMQ-induced psoriasis-like disease, further linking expression of Sema4a in the skin to maintaining skin homeostasis. The authors next studied downstream pathways that might mediate the homeostatic effects of Sema4a, focusing on mTOR given its known role in keratinocyte function. As with the immune phenotypes, Sema4a KO mice had increased mTOR activation in the epidermis in a similar pattern to mTOR activation noted in non-lesional psoriatic skin. The authors next targeted the mTOR pathway and showed rapamycin could reverse some of the psoriasis-like skin changes in Sema4a KO mice, confirming the role of increased mTOR in contributing to the observed skin phenotype.

Strengths:

The most interesting finding is the tissue-specific role for Sema4a, where it has previously been considered to play a mostly pro-inflammatory role in immune cells, this study shows that when expressed by keratinocytes, Sema4a plays a homeostatic role that when missing leads to the development of psoriasis-like skin changes. This has important implications in terms of targeting Sema4a pharmacologically. It also may yield a novel mouse model to study mechanisms of psoriasis development in mice separate from the commonly used IMQ model. The included experiments are well-controlled and executed rigorously.

Weaknesses:

A weakness of the study is the lack of tissue-specific Sema4a knockout mice (e.g. in keratinocytes only). The authors did use bone marrow chimeras, but only in one experiment. This work implies that psoriasis may represent a Sema4a-deficient state in the epidermal cells, while the same might not be true for immune cells. Indeed, in their analysis of non-lesional psoriasis skin, Sema4a was not significantly decreased compared to control skin, possibly due to compensatory increased Sema4a from other cell types. Unbiased RNA-seq of Sema4a KO mouse skin for comparison to non-lesional skin might identify other similarities besides mTOR signaling. Indeed, targeting mTOR with rapamycin reveres some of the skin changes in Sema4a KO mice, but not skin thickness, so other pathways impacted by Sema4a may be better targets if they could be identified. Utilizing WT→KO chimeras in addition to global KO mice in the experiments in Figures 6-8 would more strongly implicate the separate role of Sema4a in skin vs immune cell populations and might more closely mimic non-lesional psoriasis skin.

We sincerely appreciate your summary and for pointing out the strengths and weaknesses of our study. Although we were unfortunately unable to perform all these experiments due to limitations in our resources, we fully agree with the importance of studying tissue-specific Sema4A KO mice. As an alternative, we compared the IL-17A-producing potential of skin T cells between WT→KO mice and KO→KO mice following 4 consecutive days of IMQ treatment using flow cytometry. The results were comparable between the two groups. Additionally, we performed RNA-seq on the epidermis of WT and Sema4A KO mice. While we did not find similarities between Sema4A KO skin and non-lesional psoriasis except for S100a8 expression, we will further try to seek for the mechanisms how Sema4A KO skin mimics non-lesional psoriasis skin as a future project.

Although targeting mTOR with rapamycin did not reverse the epidermal thickness in Sema4A KO mice, rapamycin was effective in reducing epidermal thickness in a murine psoriasis model induced by IMQ in Sema4A KO mice. These results suggest potential clinical relevance for treating active, lesional psoriatic skin changes, which would be of interest to clinicians. Thank you once again for your valuable insights.

Reviewer #2 (Public Review):

Summary:

Kume et al. found for the first time that Semaphorin 4A (Sema4A) was downregulated in both mRNA and protein levels in L and NL keratinocytes of psoriasis patients compared to control keratinocytes. In peripheral blood, they found that Sema4A is not only expressed in keratinocytes but is also upregulated in hematopoietic cells such as lymphocytes and monocytes in the blood of psoriasis patients. They investigated how the down-regulation of Sema4A expression in psoriatic epidermal cells affects the immunological inflammation of psoriasis by using a psoriasis mice model in which Sema4A KO mice were treated with IMQ. Kume et al. hypothesized that down-regulation of Sema4A expression in keratinocytes might be responsible for the augmentation of psoriasis inflammation. Using bone marrow chimeric mice, Kume et al. showed that KO of Sema4A in non-hematopoietic cells was responsible for the enhanced inflammation in psoriasis. The expression of CCL20, TNF, IL-17, and mTOR was upregulated in the Sema4AKO epidermis compared to the WT epidermis, and the infiltration of IL-17-producing T cells was also enhanced.

Strengths:

Decreased Sema4A expression may be involved in psoriasis exacerbation through epidermal proliferation and enhanced infiltration of Th17 cells, which helps understand psoriasis immunopathogenesis.

Weaknesses:

The mechanism by which decreased Sema4A expression may exacerbate psoriasis is unclear as yet.

We greatly appreciate your summary and thoughtful feedback on the strengths and weaknesses of our study. In response, we have included the results of additional experiments on IL-23-mediated psoriasis-like dermatitis, which showed that epidermal thickness was significantly greater in KO mice compared to WT mice. When we analyzed the T cells infiltrating the ears using flow cytometry, the proportion of IL-17A producing Vγ2 and DNγδ T cells within the CD3 fraction of the epidermis was significantly higher in Sema4A KO mice, consistent with the results from IMQ-induced psoriasis-like dermatitis. Furthermore, we examined STAT3 expression in the epidermis of WT and Sema4A KO mice using Western blot analysis, and the results were comparable between the two groups. However, the mechanism by which decreased Sema4A expression may exacerbate psoriasis remains unclear. We have added some explanations and presumptions to the limitations section. Thank you once again for your valuable insights.

Recommendations For The Authors:

Reviewer #1 (Recommendations For The Authors):

Figure 1C

What statistics were used? The supplemental notes adjusted the P value, what correction for multiple comparisons was utilized? Could the authors instead show logFC for the DEGs between Ctl and L in each cluster? This might be best demonstrated with a volcano plot, highlighting SEMA4A, and other genes known to be DE in psoriasis.

We apologize for not including the detailed analysis methods in the original manuscript submission. We analyzed the scRNA-seq data using Cellxgene VIP with Welch’s t-test. Multiple comparisons were performed using the Benjamini-Hochberg procedure, setting the false discovery rate (FDR) at 0.05. These details are now explained in the MATERIALS AND METHODS section of the resubmitted manuscript. We also added a log2FC-log10 p-value graph for the DEGs in keratinocytes between Ctl and L to Figure 1-figure supplement 1D. The log2FC values in keratinocytes, dendritic cells, and macrophages were -0.07, 0.00, and -0.05, respectively. Although the log2FC is low in keratinocytes, the adjusted p-value (padj) for Sema4A is 2.83×10-39, indicating a statistically significant difference.

Page 8 Line 111 in the resubmitted manuscript:

“The adjusted p-value (padj) for SEMA4A in keratinocytes between Ctl and L was 2.83×10-39, indicating a statistically significant difference despite not being visually prominent in the volcano plot, which shows comprehensive differential gene expression in keratinocytes (Figure 1C; Figure 1-figure supplement 1D).”

Page 54: In the Figure legend of Figure 1-figure supplement 1D in the resubmitted manuscript:

“(D) The volcano plot displays changes in gene expression in psoriatic L compared to Ctl.”

Page 30 Line 481 in the resubmitted manuscript: In the “Data processing of single-cell RNA-sequencing and bulk RNA-sequencing” section.

“The data was integrated into an h5ad file, which can be visualized in Cellxgene VIP (K. Li et al., 2022). We then performed differential analysis between two groups of cells to identify differential expressed genes using Welch’s t-test. Multiple comparisons were controlled using the Benjamini-Hochberg procedure, with the false discovery rate set at 0.05 and significance defined as padj < 0.05.”

Figure 2B

The results narrative notes WT->WT is comparable to KO->WT. No statistics are given for this comparison. It appears the difference is less than the other comparisons, but still may be significant. Also, in the supplemental for Figure 2B, there appear to be missing columns for the 4 BM chimera groups (columns for WT and KO, but not 4 columns for each donor: recipient pair).

We sincerely apologize for any confusion. We presented the results of the chimeric mice in Figure 3, and Figure 3-source data 1 shows the 4 BM chimera groups. In Figure 3B, the p-value for the comparison between WT->WT mice and KO->WT mice was 0.7988, as indicated in Figure 3-source data 1.

Figure 3B

While ear skin is not easily obtainable at day 0 for comparison, why not also include back skin at Wk 8? If the back skin epidermis is thicker like the ear skin, it supports the ear skin conclusion and adds a more consistent comparison. If the back skin epidermis is not thicker, what would be the author's explanation as to the why only ear skin epidermis is thicker in KO mice at 8 weeks?

We appreciate and completely agree with the reviewer’s insightful comment. We have added images and dot plots of the back skin at Week 8 in Figure 4B. Since the back skin epidermis is thicker, similar to the ear skin, these results support the conclusion drawn from the ear skin data. Regarding Figure 4C, which shows the expression of Sema4a in the epidermis and dermis of 8-week-old WT mouse ear, we have modified the sentence in the manuscript to ‘the epidermis of WT ear at Week 8’ for clarification.

Page 12 Line 180 in the resubmitted manuscript:

“While epidermal thickness of back skin was comparable at birth (Figure 4B), on week 8, epidermis of Sema4AKO back and ear skin was notably thicker than that of WT mice (Figure 4B), suggesting that acanthosis in Sema4AKO mice is accentuated post-birth.”

Page 47: In the Figure legend of Figure 4B in the resubmitted manuscript:

“(B) Left: representative Hematoxylin and eosin staining of Day 0 back and Wk 8 back and ear. Scale bar = 50 μm. Right: Epi and Derm thickness in Day 0 back (n = 5) and Wk 8 back (n = 5) and ear (n = 8).”

Figures 3C&D, Figures 4 D-F

The figures might be easier to read if some of the data is moved to supplemental, especially in Figure 4, which has 36 panels just in D-F. Conversely, the dLN data is important in establishing the skin microenvironment as important in the accumulation of γδ cells and IL-17 production in the setting of Sema4a KO, so this might be more impactful if moved to the main figure.

We appreciate and agree with your comments. As recommended, we have moved data from Figure 3C and 4D-F to the supplemental section. The dLN data have been moved to the main figure as Figure 4E. This has improved the readability of the figures.

Figure 5 and Figure 6 might work better if combined. The differences in keratinocytes in psoriasis are well-known, so the novelty is how Sema4a KO skin appears to share similar differences. This would be easier to see if compared side-by-side in the same figure. Also, there is an opportunity to show this more rigorously by performing RNA-seq on WT vs Sema4a KO skin. Showing a larger set of DEGs that trend similarly between Ctl/NL psoriasis and WT/Sema4a KO skin in a heatmap would bolster the conclusion that Sema4a deficiency contributes to a psoriasis-like skin defect.

We appreciate your valuable suggestion. Following your recommendation, we have combined Figures 5 and 6 to facilitate a side-by-side comparison. This highlights the similarities between Sema4AKO skin and psoriasis, making it easier to observe differences in keratinocytes. Additionally, we performed RNA-seq on WT and Sema4a KO epidermis (n = 3 per group). We analyzed the raw count data using iDEP 2.0 (Ge S.X., BMC Bioinformatics, 2018), setting the minimal counts per million to 0.5 in at least one library. Differential gene expression analysis was conducted using DEseq2, with an FDR cutoff of 0.1 and a minimum fold change of 2. As a result, we identified 46 upregulated and 70 downregulated genes in Sema4AKO mice compared to WT mice (see the volcano plot and heat map). However, except for S100a8, we did not observe significant expression changes in non-lesional psoriasis-related genes between WT and Sema4AKO mice. In the future, we aim to identify subtle stimuli that could cause gene expression changes between these groups and we would like to perform additional RNA-seq experiments.

Author response image 1.

Author response image 2.

Page 48: The Figure title of Figure 5 in the resubmitted manuscript:

“Figure 5: Sema4AKO skin shares the features of human psoriatic NL.”

SEMA4A is not significantly DE between Ctl and NL in the psoriasis RNA-seq data. If a lower expression of SEMA4A in psoriasis skin is a driving part of the phenotype, why is this not observed in the RNA-seq data? Presumably, this could be explained by infiltration of immune cells with increased SEMA4A expression, like in the scRNA-seq data in Figure 1. If so, might it be useful to analyze WT->KO chimera mice similarly to global KO mice in Figures 6-8? This might more accurately reflect what is happening in psoriasis, if epidermal SEMA4A expression is low, but immune expression is not. The KO data on their own nicely show a skin phenotype, but these additional experiments might more closely mimic psoriatic disease and increase the rigor and impact of the study.

We really appreciate your insightful comments. Due to the limitations of the animal experimentation facility, we regret that we are unable to create additional chimeric mice. Although our analysis is limited, we compared IL-17A production from T cells of WT→KO mice and KO→KO mice following 4 consecutive days of IMQ treatment using flow cytometry (see Author response image 3 below; n = 6 for WT→KO, n = 4 for KO→KO). This comparison revealed that IL-17A production from T cells was comparable, regardless of whether they were derived from WT or Sema4AKO mice, when the skin constituent cells were derived from Sema4AKO. We appreciate the value of your advice, and agree that investigating keratinocyte differentiation and mTOR signaling in the epidermis, using either WT→KO chimeric mice or keratinocyte-specific Sema4A-deficient mice, is a crucial next step in our research.

Author response image 3.

Figure 8

Rapamycin was able to partially reverse the psoriasis-like skin phenotype in Sema4a KO mice. Would rapamycin also be effective in the more severe disease induced by IMQ in Sema4a KO mice? While partially reducing the effect of Sema4a KO on steady-state skin with rapamycin strengthens the link to mTOR dysregulation, it did not change skin thickness. It's unclear if this would be useful clinically for patients with well-controlled psoriasis (NL skin). Would it be useful to reverse active, lesional psoriatic skin changes? Testing this might yield results more relevant to clinicians and patients.

We are grateful for your valuable feedback. Rapamycin showed effectiveness in reducing epidermal thickness in a murine psoriasis model induced by IMQ in Sema4AKO mice. Rapamycin treatment downregulated the expression of Krt10, Krt14, and Krt16. We included these results to Figure 7-figure supplement 2. These results suggest potential clinical relevance for treating active, lesional psoriatic skin changes and may be of interest to clinicians and patients.

Page 17 Line 269 in the resubmitted manuscript:

“Next, we investigated whether intraperitoneal rapamycin treatment effectively downregulates inflammation in the IMQ-induced murine model of psoriasis in Sema4AKO mice (Figure 7-figure supplement 2A). Rapamycin significantly reduced epidermal thickness compared to vehicle treatment (Figure 7-figure supplement 2B). Additionally, rapamycin treatment downregulated the expression of Krt10, Krt14, and Krt16 (Figure 7-figure supplement 2C). While the upregulation of Il17a in the Sema4AKO epidermis in IMQ model was not clearly modified by rapamycin (Figure 7-figure supplement 2C), immunofluorescence revealed a decrease in the number of CD3 T cells in Sema4AKO epidermis by rapamycin (Figure 7-figure supplement 2D). In the naive states, mTORC1 primarily regulates keratinocyte proliferation, whereas mTORC2 mainly involved in the keratinocyte differentiation through Sema4A-related signaling pathways. Conversely, in the psoriatic dermatitis state, rapamycin downregulated both keratinocyte differentiation and proliferation markers. The observed similarities in Il17a expression following treatment with rapamycin and JR-AB2-011, regardless of additional IMQ treatment, suggest that Il17a production is not significantly dependent on Sema4A-related mTOR signaling.”

Page 29 Line 461 in the resubmitted manuscript: In the “Inhibition of mTOR” section.

“To analyze the preventive effectiveness of rapamycin in an IMQ-induced murine model of psoriatic dermatitis, Sema4AKO mice were administered either vehicle or rapamycin intraperitoneally from Day 0 to Day 17, and IMQ was topically applied to both ears for 4 days starting on Day 14. Then, on Day 18, ears were collected for further analysis.”

Page 71: Figure 7-figure supplement 2 in the resubmitted manuscript:

“Figure 7-figure supplement 2: Rapamycin treatment reduced the epidermal swelling observed in IMQ-treated Sema4AKO mice.

(A) Experimental scheme. (B) The Epi thickness on Day 18. (n = 10 for Ctl, n = 12 for Rapamycin). (C) Relative expression of keratinocyte differentiation markers and Il17a in Sema4AKO Epi (n = 10 for Ctl, n = 12 for Rapamycin). (D) The number of T cells in the Epi (left) and Derm (right), under Ctl or rapamycin and IMQ treatments (n = 10 for Ctl, n = 12 for Rapamycin). Each dot represents the sum of numbers from 10 unit areas across 3 specimens. A-C: *p < 0.05, **p < 0.01. NS, not significant.”

Reviewer #2 (Recommendations For The Authors):

(1) To know whether the decrease of Sema4A in the epidermis of psoriasis patients is a result or a cause of psoriasis, it is necessary to show how the expression of Sema4A in epidermal cells is regulated. Shouldn't the degree of change in the expression of essential molecules (which is the cause of psoriasis) be more pronounced in L than in NL?

We surveyed transcription factors of human Sema4A using GeneCards and found that NF-κB is the transcription factor most frequently associated with psoriasis. Wang et al. (Arthritis Res Ther. 2015) indicated NF-κB-dependent modulation of Sema4A expression in synovial fibroblasts of rheumatoid arthritis. However, since NF-κB expression is reportedly upregulated in psoriasis lesions, other transcription factors may function as key modulators of Sema4A expression in the epidermis.

Although the molecules causing psoriasis remain to be elucidated, we investigated the correlation between the expression of psoriasis-related essential molecules in keratinocytes—such as S100A7A, S100A7, S100A8, S100A9, and S100A12—and SEMA4A expression in L and NL samples using qRT-PCR. We could not identify a correlation between these molecules and SEMA4A expression. We added a note to the limitations section to acknowledge that we were not able to reveal how Sema4A expression is regulated and that we could not determine the relationships between Sema4A expression and the essential molecules upregulated in psoriatic keratinocytes.

Page 21 Line 328 in the resubmitted manuscript:

“We were not able to reveal how Sema4A expression is regulated. Although we showed that downregulation of Sema4A is related to the abnormal cytokeratin expression observed in psoriasis, we could not determine the relationships between Sema4A expression and the essential molecules upregulated in psoriatic keratinocytes.”

(2) Using bone marrow chimeric mice, it has already been reported that hematopoietic cells contain keratinocyte stem cells. Therefore, their interpretation is not supported by the results of their bone marrow chimeric mice experiment, and it is essential to generate keratinocyte-specific Sema4A knockout mice and perform similar experiments to support their interpretation.

We value the reviewer’s insightful comment. We have assessed the expression of Sema4a in the epidermis of WT→KO chimeric mice using qRT-PCR. Our findings indicate that Sema4a expression levels in the epidermis of these mice are minimal (cycle threshold values of Sema4a ranged from 31.9 to not detected in WT→KO chimeric mice, whereas they ranged from 24.5 to 26.2 in WT→ WT mice). Consequently, we believe that the impact of keratinocyte stem cells derived from WT-hematopoietic cells is limited in this model. We appreciate this opportunity to clarify our results and will consider the generation of keratinocyte-specific Sema4A knockout mice for future experiments to further substantiate our interpretation.

Page 11 Line 159 in the resubmitted manuscript:

“Since it has already been reported that bone marrow cells contain keratinocyte stem cells (Harris et al., 2004; Wu, Zhao, & Tredget, 2010), we confirmed that epidermis of mice deficient in non-hematopoietic Sema4A (WT→KO) showed no obvious detection of Sema4a, thereby ruling out the impact of donor-derived keratinocyte stem cells infiltrating the host epidermis (Figure 3-figure supplement 1A).”

Page 60: In the Figure legend of Figure 3-figure supplement 1A in the resubmitted manuscript:

“(A) Sema4a expression in the Epi of WT→ WT mice and WT→ KO mice (n = 8 for WT→ WT, n = 7 for WT→ KO).”

(3) Since Sema4A KO mice already have immunological and epidermal cell characteristics similar to psoriasis, albeit weak, it is possible that the nonspecific stimulus of simply topical IMQ may have appeared to exacerbate psoriasis. It is advisable to confirm whether a more psoriasis-specific stimulus, IL-23 administration, would produce similar results.

Thank you for your suggestion. Following your advice, we have analyzed IL-23-mediated psoriasis-like dermatitis. To induce the model, 20 μl of phosphate-buffered saline containing 500 ng of recombinant mouse IL-23 was injected intradermally into both ears for 4 consecutive days. Unlike with the application of IMQ, there was no significant difference in ear thickness. However, H&E staining revealed that the epidermal thickness was significantly greater in KO mice compared to WT mice. Although a longer period of IL-23 induction might result in more pronounced ear swelling, we conducted this experiment over the same duration as the IMQ application experiment to maintain consistency. When we analyzed the T cells infiltrating the ears using flow cytometry, the proportion of IL-17A producing Vγ2 and DNγδ T cells in CD3 fraction in the epidermis was significantly higher in Sema4A KO mice, consistent with the results from IMQ-induced psoriasis-like dermatitis.

The lack of significant difference in ear thickness changes with IL-23 administration might be due to IL-23 administration not reflecting upstream events of IL-23 production.

We consider that in psoriasis, the expression of Sema4A in keratinocytes is likely more important than in T cells. Therefore, it makes sense that the phenotype difference was more pronounced with IMQ, which likely has a greater effect on keratinocytes compared to IL-23.

Page 9 Line 137 in the resubmitted manuscript:

“Though the imiquimod model is well-established and valuable murine psoriatic model (van der Fits et al., 2009), the vehicle of imiquimod cream can activate skin inflammation that is independent of toll-like receptor 7, such as inflammasome activation, keratinocyte death and interleukin-1 production (Walter et al., 2013). This suggests that the imiquimod model involves complex pathway. Therefore, we subsequently induced IL-23-mediated psoriasis-like dermatitis (Figure2-figure supplement 2A), a much simpler murine psoriatic model, because IL-23 is thought to play a central role in psoriasis pathogenesis (Krueger et al., 2007; Lee et al., 2004). Although ear swelling on day 4 was comparable between WT mice and Sema4AKO mice (Figure2-figure supplement 2B), the epidermis, but not the dermis, was significantly thicker in Sema4AKO mice compared to WT mice (Figure2-figure supplement 2C). We found that the proportion of CD4 T cells among T cells was significantly higher in Sema4A KO mice compared to WT mice, while the proportion of Vγ2 and DNγδ T cells among T cells was comparable between them (Figure 2-figure supplement 2D). On the other hand, focusing on IL-17A-producing cells, the proportion of IL-17A-producing Vγ2 and DNγδ T cells in CD3 fraction in the epidermis was significantly higher in Sema4A KO mice, consistent with the results from imiquimod-induced psoriasis-like dermatitis. (Figure 2-figure supplement 2E).”

Page 24 Line 363 in the resubmitted manuscript: In the “Mice” section.

“To induce IL-23-mediated psoriasis-like dermatitis, 20 μl of phosphate-buffered saline containing 500 ng of recombinant mouse IL-23 (BioLegend, San Diego, CA) was injected intradermally into both ears of anesthetized mice using a 29-gauge needle for 4 consecutive days.”

Page 58: In the Figure legend of Figure 2-figure supplement 2 in the resubmitted manuscript:

“IL-23-mediated psoriasis-like dermatitis is augmented in Sema4AKO mice.

(A) An experimental scheme involved intradermally injecting 20 μl of phosphate-buffered saline containing 500 ng of recombinant mouse IL-23 into both ears of WT mice and KO mice for 4 consecutive days. Samples for following analysis were collected on Day 4. (B and C) Ear thickness (B) and Epi and Derm thickness (C) of WT mice and KO mice on Day 4 (n = 12 per group). (D and E) The percentages of Vγ3, Vγ2, DNγδ, CD4, and CD8 T cells (D) and those with IL-17A production (E) in CD3 fraction in the Epi (top) and Derm (bottom) of WT and KO ears (n = 5 per group). Each dot represents the average of 4 ear specimens. B-E: *p < 0.05, **p < 0.01. NS, not significant.”

(4) How is STAT3 expression in the epidermis crucial in the pathogenesis of psoriasis in Sem4AKO mice?

We appreciate your insightful comment. In our study, given the established role of activated STAT3 in psoriasis, we investigated both total STAT3 and phosphorylated STAT3 (p-STAT3) levels in the naive epidermis of WT and Sema4AKO mice (See the figure below). Our findings indicate that STAT3 activation does not occur in the epidermis of Sema4AKO mice. Therefore, we speculated that the hyperkeratosis observed in Sema4AKO mice is due to aberrant mTOR signaling rather than STAT3 activation. STAT3 may be relevant to other pathways independent of Sema4A signaling, or it may function as a complex with other molecules in the Sema4A signaling.

Author response image 4.

eLife Assessment

This paper presents the important discovery that lipid metabolic imbalance caused by Snail, an EMT-related transcription factor, contributes to the acquisition of chemoresistance in cancer cells. However, the incomplete support for the authors' claims is due to concerns about the causal relationship and lack of sufficient quantitative analysis. With strengthened evidence, this work would be of broad interest to researchers in the fields of cancer biology, lipid metabolism, and cell biology.

Reviewer #1 (Public review):

The authors focus on the molecular mechanisms by which EMT cells confer resistance to cancer cells. The authors use a wide range of methods to reveal that overexpression of Snail in EMT cells induces cholesterol/sphingomyelin imbalance via transcriptional repression of biosynthetic enzymes involved in sphingomyelin synthesis. The study also revealed that ABCA1 is important for cholesterol efflux and thus for counterbalancing the excess of intracellular free cholesterol in these snail-EMT cells. Inhibition of ACAT, an enzyme catalyzing cholesterol esterification, also seems essential to inhibit the growth of snail-expressing cancer cells.

However, It seems important to analyze the localization of ABCA1, as it is possible that in the event of cholesterol/sphingomyelin imbalance, for example, the intracellular trafficking of the pump may be altered.<br /> The authors should also analyze ACAT levels and/or activity in snail-EMT cells that should be increased. Overall, the provided data are important to better understand cancer biology.

Reviewer #2 (Public review):

Summary:

In this study, the authors discovered that the chemoresistance in RCC cell lines correlates with the expression levels of the drug transporter ABCA1 and the EMT-related transcription factor Snail. They demonstrate that Snail induces ABCA1 expression and chemoresistance, and that ABCA1 inhibitors can counteract this resistance. The study also suggests that Snail disrupts the cholesterol-sphingomyelin (Chol/SM) balance by repressing the expression of enzymes involved in very long-chain fatty acid-sphingomyelin synthesis, leading to excess free cholesterol. This imbalance activates the cholesterol-LXR pathway, inducing ABCA1 expression. Moreover, inhibiting cholesterol esterification suppresses Snail-positive cancer cell growth, providing potential lipid-targeting strategies for invasive cancer therapy.

Strengths:

This research presents a novel mechanism by which the EMT-related transcription factor Snail confers drug resistance by altering the Chol/SM balance, introducing a previously unrecognized role of lipid metabolism in the chemoresistance of cancer cells. The focus on lipid balance, rather than individual lipid levels, is a particularly insightful approach. The potential for targeting cholesterol detoxification pathways in Snail-positive cancer cells is also a significant therapeutic implication.

Weaknesses:

The study's claim that Snail-induced ABCA1 is crucial for chemoresistance relies only on pharmacological inhibition of ABCA1, lacking additional validation. The causal relationship between the disrupted Chol/SM balance and ABCA1 expression or chemoresistance is not directly supported by data. Some data lack quantitative analysis.

Author response:

Response to Reviewer 1

We will investigate the intracellular localization of ABCA1 in both EpH4 and EpH4-Snail cells. We will also examine the changes in ACAT expression levels within these cell lines.

Response to Reviewer 2

We will first investigate whether the chemoresistance exhibited by EpH4-Snail cells can be abolished not only through pharmacological inhibition of ABCA1 but also by knocking out the ABCA1 gene. Regarding causality, as demonstrated in Figure 2, we have already shown that reducing cholesterol levels in EpH4-Snail cells decreases ABCA1 expression. To further explore this relationship, we will assess whether increasing sphingomyelin levels by adding ceramide to the culture medium, thereby correcting the sphingomyelin-to-cholesterol ratio, would reduce ABCA1 expression. Furthermore, we will evaluate whether lowering cholesterol levels in EpH4-Snail cells via simvastatin treatment, along with normalization of the sphingomyelin-to-cholesterol ratio, attenuates their resistance to the anticancer drug nitidine chloride. Additionally, we will incorporate quantitative analyses for several experiments, as suggested in the reviewers’ comments, to enhance the robustness of our findings.

eLife Assessment

The authors use deep mutational scanning to assess the effect of all possible protein-coding variants in MC4R, a G protein-coupled receptor associated with obesity. They develop new, more precise approaches, enabling them to probe molecular phenotypes directly relevant to the development of drugs that target this receptor. In this important work, the authors provide convincing evidence that variants impact signaling through MC4R in different ways, that some defective variants are amenable to a corrector drug and that deep mutational scanning data could guide compound optimization.

Reviewer #1 (Public review):

Summary:

Howard et al. performed deep mutational scanning on the MC4R gene, using a reporter assay to investigate two distinct downstream pathways across multiple experimental conditions. They validated their findings with ClinVar data and previous studies. Additionally, they provided insights into the application of DMS results for personalized drug therapy and differential ligand responses across variant types.

Strengths:

They captured over 99% of variants with robust signals and investigated subtle functionalities, such as pathway-specific activities and interactions with different ligands, by refining both the experimental design and analytical methods.

Weaknesses:

While the study generated informative results, it lacks a detailed explanation regarding the input library, replicate correlation, and sequencing depth for a given number of cells. Additionally, there are several questions that it would be helpful for authors to clarify.

(1) It would be helpful to clarify the information regarding the quality of the input library and experimental replicates. Are variants evenly represented in the library? Additionally, have the authors considered using long-read sequencing to confirm the presence of a single intended variant per construct? Finally, could the authors provide details on the correlation between experimental replicates under each condition?

(2) Since the functional readout of variants is conducted through RNA sequencing, it seems crucial to sequence a sufficient number of cells with adequate sequencing saturation. Could the authors clarify the coverage depth used for each RNA-seq experiment and how this depth was determined? Additionally, how many cells were sequenced in each experiment?

(3) It appears that the frequencies of individual RNA-seq barcode variants were used as a proxy for MR4C activity. Would it be important to also normalize for heterogeneity in RNA-seq coverage across different cells in the experiment? Variability in cell representation (i.e., the distribution of variants across cells) could lead to misinterpretation of variant effects. For example, suppose barcode_a1 represents variant A and barcode_b1 represents variant B. If the RNA-seq results show 6 reads for barcode_a1 and 7 reads for barcode_b1, it might initially appear that both variants have similar effect sizes. However, if these reads correspond to 6 separate cells each containing 1 copy of barcode_a1, and only 1 cell containing 7 copies of barcode_b1, the interpretation changes significantly. Additionally, if certain variants occupy a larger proportion of the cell population, they are more likely to be overrepresented in RNA sequencing.

(4) Although the assay system appears to effectively represent MC4R functionality at the molecular level, we are curious about the potential disparity between the DMS score system and physiological relevance. How do variants reported in gnomAD distribute within the DMS scoring system?

(5) To measure Gq signaling, the authors used the GAL4-VPR relay system. Is there additional experimental data to support that this relay system accurately represents Gq signaling?

(6) Identifying the variants responsive to the corrector was impressive. However, we are curious about how the authors confirmed that the restoration of MC4R activity was due to the correction of the MC4R protein itself. Is there a possibility that the observed effect could be influenced by other factors affected by the corrector? When the corrector was applied to the cells, were any expected or unexpected differential gene expression changes observed?

(7) As mentioned in the introduction, gain-of-function (GoF) variants are known to be protective against obesity. It would be interesting to see further studies on the observed GoF variants. Do the authors have any plans for additional research on these variants?

Reviewer #2 (Public review):

Overview

In this manuscript, the authors use deep mutational scanning to assess the effect of ~6,600 protein-coding variants in MC4R, a G protein-coupled receptor associated with obesity. Reasoning that current deep mutational scanning approaches are insufficiently precise for some drug development applications, they focus on articulating new, more precise approaches. These approaches, which include a new statistical model and innovative reporter assay, enable them to probe molecular phenotypes directly relevant to the development of drugs that target this receptor with high precision and statistical rigor.

They use the resulting data for a variety of purposes, including probing the relationship between MC4R's sequence and structure, analyzing the effect of clinically important variants, identifying variants that disrupt downstream MC4R signaling via one but not both pathways, identifying loss of function variants are amenable to a corrector drug and exploring how deep mutational scanning data could guide small molecule drug optimization.

Strengths

The analysis and statistical framework developed by the authors represent a significant advance. In particular, the study makes use of barcode-level internally replicated measurements to more accurately estimate measurement noise.

The framework allows variant effects to be compared across experimental conditions, a task that is currently hard to do with rigor. Thus, this framework will be applicable to a large number of existing and future deep mutational scanning experiments.

The authors refine their existing barcode transcription-based assay for GPCR signaling, and develop a clever "relay" new reporter system to boost signaling in a particular pathway. They show that these reporters can be used to measure both gain of function and loss of function effects, which many deep mutational scanning approaches cannot do.

The use of systematic approaches to integrate and then interrogate high-dimensional deep mutational scanning data is a big strength. For example, the authors applied PCA to the variant effect results from reporters for two different MC4R signaling pathways and were able to discover variants that biased signaling through one or the other pathway. This approach paves the way for analyses of higher dimensional deep mutational scans.

The authors use the deep mutational scanning data they collect to map how different variants impact small molecule agonists activate MC4R signaling. This is an exciting idea, because developing small-molecule protein-targeting therapeutics is difficult, and this manuscript suggests a new way to map small-molecule-protein interactions.

Weaknesses

The authors derive insights into the relationship between MC4R signaling through different pathways and its structure. While these make sense based on what is already known, the manuscript would be stronger if some of these insights were validated using methods other than deep mutational scanning.

Likewise, the authors use their data to identify positions where variants disrupt MC4R activation by one small molecule agonist but not another. They hypothesize these effects point to positions that are more or less important for the binding of different small molecule agonists. The manuscript would be stronger if some of these insights were explored further.

Impact

In this manuscript, the authors present new methods, including a statistical framework for analyzing deep mutational scanning data that will have a broad impact. They also generate MC4R variant effect data that is of interest to the GPCR community.

Author response:

We thank the reviewers for their support of this work and insightful recommendations for how to improve it. We have provided specific responses to each reviewer comment below. To summarize how we intend to address the requested revisions:

Many of the reviewers’ comments requested additional technical or quality details about the DMS libraries or assays (e.g., number of cells tested, number of sequencing reads, assay replication, assay sensitivity, library balance), and we provide additional information and analyses that we can incorporate into the relevant portions of the text, supplementary tables, and supplementary figures to address these questions.

Some comments asked to clarify nomenclature/wording or provide additional labels to images, and we will make these changes as requested.

A few questions would require additional experimental data to address. Where experiments have already been performed, we will incorporate those results or cite relevant work previously reported in the literature.

Reviewer 1:

Summary

Howard et al. performed deep mutational scanning on the MC4R gene, using a reporter assay to investigate two distinct downstream pathways across multiple experimental conditions. They validated their findings with ClinVar data and previous studies. Additionally, they provided insights into the application of DMS results for personalized drug therapy and differential ligand responses across variant types.

Strengths

They captured over 99% of variants with robust signals and investigated subtle functionalities, such as pathway-specific activities and interactions with different ligands, by refining both the experimental design and analytical methods.

Weaknesses

While the study generated informative results, it lacks a detailed explanation regarding the input library, replicate correlation, and sequencing depth for a given number of cells.

Additionally, there are several questions that it would be helpful for authors to clarify.

(1) It would be helpful to clarify the information regarding the quality of the input library and experimental replicates. Are variants evenly represented in the library? Additionally, have the authors considered using long-read sequencing to confirm the presence of a single intended variant per construct? Finally, could the authors provide details on the correlation between experimental replicates under each condition?

Are variants evenly represented in the library?

We strive to achieve as evenly balanced library as possible at every stage of the DMS process (e.g., initial cloning in E. coli through integration into human cells). Below is a representative plot showing the number of barcodes per amino acid variant at each position in a given ~60 amino acid subregion of MC4R, which highlights how evenly variants are represented at the E. coli cloning stage.

Author response image 1.

We also make similar measurements after the library is integrated into HEK293T cell lines, and see similarly even coverage across all variants, as shown in the plot below.

Author response image 2.

Additionally, have the authors considered using long-read sequencing to confirm the presence of a single intended variant per construct?

We agree long-read sequencing would be an excellent way to confirm that our constructs contain a single intended variant. However, we elected for an alternate method (outlined in more detail in Jones et al. 2020) that leverages multiple layers of validation. First, the oligo chip-synthesized portions of the protein containing the variants are cloned into a sequence-verified plasmid backbone, which greatly decreases the chances of spuriously generating a mutation in a different portion of the protein. We then sequence both the oligo portion and random barcode using overlapping paired end reads during barcode mapping to avoid sequencing errors and to help detect DNA synthesis errors. At this stage, we computationally reject any constructs that have more than one variant. Given this, the vast majority of remaining unintended variants would come from somatic mutations introduced by the E. coli cloning or replication process, which should be low frequency. We have used our in-house full plasmid sequencing method, OCTOPUS, to sample and spot check this for several other DMS libraries we have generated using the same cloning methods. We have found variants in the plasmid backbone in only ~1% of plasmids in these libraries. Our statistical model also helps correct for this by accounting for barcode-specific variation. Finally we believe this provides further motivation for having multiple barcodes per variant, which dilutes the effect of any unintended additional variants.

Finally, could the authors provide details on the correlation between experimental replicates under each condition?

Certainly! In general, the Gs reporter had higher correlation between replicates than the Gq system (r ~ 0.5 vs r ~ 0.4). The plots below show two representative correlations at the RNA-seq stage of read counts for barcodes between the low a-MSH conditions. One important advantage of our statistical model is that it’s able to leverage information from barcodes regardless of the number of replicates they appear in.

Author response image 3.

Since the functional readout of variants is conducted through RNA sequencing, it seems crucial to sequence a sufficient number of cells with adequate sequencing saturation. Could the authors clarify the coverage depth used for each RNA-seq experiment and how this depth was determined? Additionally, how many cells were sequenced in each experiment?

This will be addressed by incorporating the following details into the manuscript:

We seeded 17 million cells per replicate at the start of each assay and, with a doubling of ~1.5x over the course of the assay, harvested ~25.5 million cells per replicate for RNA extraction and sequencing. We found this sufficient to get at least ~30-60x cellular coverage per amino acid variant.

Total mapped reads per replicate at RNA-seq stage

- Gs/CRE: 9.1-18.2 million mapped reads, median=12.3

- Gq/UAS: 8.6-24.1 million mapped reads, median=14.5

- Gs/CRE+Chaperone: 6.4-9.5 million mapped reads, median=7.5

Reads per barcode distribution

- Median read counts of 8, 10, and 6 reads per sample per barcode for Gs/CRE, Gq/UAS, and Gs/CRE+Chaperone assays, respectively.

Barcodes per variant distribution

- As reported, the median number of barcodes per variant across samples (the “median of medians”) is 56 for Gs/CRE and 28 for Gq/UAS

- Additionally, it is 44 for Gs/CRE+Chaperone

It appears that the frequencies of individual RNA-seq barcode variants were used as a proxy for MR4C activity. Would it be important to also normalize for heterogeneity in RNA-seq coverage across different cells in the experiment? Variability in cell representation (i.e., the distribution of variants across cells) could lead to misinterpretation of variant effects. For example, suppose barcode_a1 represents variant A and barcode_b1 represents variant B. If the RNA-seq results show 6 reads for barcode_a1 and 7 reads for barcode_b1, it might initially appear that both variants have similar effect sizes. However, if these reads correspond to 6 separate cells each containing 1 copy of barcode_a1, and only 1 cell containing 7 copies of barcode_b1, the interpretation changes significantly. Additionally, if certain variants occupy a larger proportion of the cell population, they are more likely to be overrepresented in RNA sequencing.

We account for this heterogeneity in several ways. First, as shown above (Response to Reviewer 1, Question 1), we aim to have even representation of variants within our libraries. Second, we utilize compositional control conditions like forskolin or unstimulated conditions to obtain treatment-independent measurements of barcode abundance and, consequently, of mutant-vs-WT effects that are due to compositional rather than biological variability. We expect that variability observed under these controls is due to subtle effects of molecular cloning, gene expression, and stochasticity. Using these controls, we observe that mutant-vs-WT effects are generally close to zero in these normalization conditions (e.g., in untreated Gq, see Supplementary Figure 3) as compared to drug-treated conditions. For example, pre-mature stops behave similar to WT in normalization conditions. This indicates that mutant abundance is relatively homogenous. Where there are barcode-dependent effects on abundance, we can use information from these conditions to normalize that effect. Finally, our mixed-effect model accounts for barcode-specific deviations from the expected mutant effect (e.g. a “high count” barcode consistently being high relative to the mean).

Although the assay system appears to effectively represent MC4R functionality at the molecular level, we are curious about the potential disparity between the DMS score system and physiological relevance. How do variants reported in gnomAD distribute within the DMS scoring system?

Figure 2D shows DMS scores (variant effect on Gs signaling) relative to human population frequency for all MC4R variants reported in gnomAD as of January 8, 2024.

To measure Gq signaling, the authors used the GAL4-VPR relay system. Is there additional experimental data to support that this relay system accurately represents Gq signaling?

The full Gq reporter uses an NFAT response element from the IL-2 promoter to regulate the expression of the GAL4-VPR relay. In this system, the activation of Gq signaling results in the activation of the NFAT response element, and this signal is then amplified by the GAL4-VPR relay. The NFAT response element has been previously well-validated to respond to the activation of Gq signaling (e.g., PMID: 8631834). We will add this reference to the text to further support the use of the Gq assay.

Identifying the variants responsive to the corrector was impressive. However, we are curious about how the authors confirmed that the restoration of MC4R activity was due to the correction of the MC4R protein itself. Is there a possibility that the observed effect could be influenced by other factors affected by the corrector? When the corrector was applied to the cells, were any expected or unexpected differential gene expression changes observed?

While we do not directly measure whether Ipsen-17 has effects on other signaling processes, previous work has shown that Ipsen-17 treatment does not indirectly alter signaling kinetics such as receptor internalization (Wang et al., 2014). Furthermore, our analysis methods inherently account for this by normalizing variant effects to WT signaling levels. Any observed rescue of a given variant inherently means that the variant is specifically more responsive to Ipsen-17 than WT, and the fact that different variants exhibit different levels of rescue is reassuring that the mechanism is on target to MC4R. Lastly, Ipsen-17 is known to be an antagonist of alpha-MSH activity and is thought to bind directly to the same site on MC4R (Wang et al., 2014).

As mentioned in the introduction, gain-of-function (GoF) variants are known to be protective against obesity. It would be interesting to see further studies on the observed GoF variants. Do the authors have any plans for additional research on these variants?

We agree this would be an excellent line of inquiry, but due to changes in company priorities we unfortunately do not have any plans for additional research on these variants.

Reviewer 2:

Overview

In this manuscript, the authors use deep mutational scanning to assess the effect of ~6,600 protein-coding variants in MC4R, a G protein-coupled receptor associated with obesity. Reasoning that current deep mutational scanning approaches are insufficiently precise for some drug development applications, they focus on articulating new, more precise approaches. These approaches, which include a new statistical model and innovative reporter assay, enable them to probe molecular phenotypes directly relevant to the development of drugs that target this receptor with high precision and statistical rigor.

They use the resulting data for a variety of purposes, including probing the relationship between MC4R's sequence and structure, analyzing the effect of clinically important variants, identifying variants that disrupt downstream MC4R signaling via one but not both pathways, identifying loss of function variants are amenable to a corrector drug and exploring how deep mutational scanning data could guide small molecule drug optimization.

Strengths

The analysis and statistical framework developed by the authors represent a significant advance. In particular, the study makes use of barcode-level internally replicated measurements to more accurately estimate measurement noise.

The framework allows variant effects to be compared across experimental conditions, a task that is currently hard to do with rigor. Thus, this framework will be applicable to a large number of existing and future deep mutational scanning experiments.

The authors refine their existing barcode transcription-based assay for GPCR signaling, and develop a clever "relay" new reporter system to boost signaling in a particular pathway. They show that these reporters can be used to measure both gain of function and loss of function effects, which many deep mutational scanning approaches cannot do.

The use of systematic approaches to integrate and then interrogate high-dimensional deep mutational scanning data is a big strength. For example, the authors applied PCA to the variant effect results from reporters for two different MC4R signaling pathways and were able to discover variants that biased signaling through one or the other pathway. This approach paves the way for analyses of higher dimensional deep mutational scans.

The authors use the deep mutational scanning data they collect to map how different variants impact small molecule agonists activate MC4R signaling. This is an exciting idea, because developing small-molecule protein-targeting therapeutics is difficult, and this manuscript suggests a new way to map small-molecule-protein interactions.

Weaknesses

The authors derive insights into the relationship between MC4R signaling through different pathways and its structure. While these make sense based on what is already known, the manuscript would be stronger if some of these insights were validated using methods other than deep mutational scanning.

Likewise, the authors use their data to identify positions where variants disrupt MC4R activation by one small molecule agonist but not another. They hypothesize these effects point to positions that are more or less important for the binding of different small molecule agonists. The manuscript would be stronger if some of these insights were explored further.

Impact

In this manuscript, the authors present new methods, including a statistical framework for analyzing deep mutational scanning data that will have a broad impact. They also generate MC4R variant effect data that is of interest to the GPCR community.

eLife Assessment

This important study demonstrates the therapeutic potential of recombinant human PDGF-AB/BB proteins in reducing the senescent signatures of primary human intervertebral disc cells. The study represents a significant advance in the treatment of intervertebral disc degeneration and can be applied broadly to other degenerative musculoskeletal tissues through the suppression of senescence. Solid evidence, primarily based on transcriptomic analysis and direct protein measurements from relatively homogeneous cell populations, supports the therapeutic potential of PDGF, but more experimental details would be needed to make the evidence stronger.

Reviewer #1 (Public review):

The authors, Zhang et al., demonstrate the beneficial effects of treating degenerate human primary intervertebral disc (IVD) cells with recombinant human PDGF-AB/BB on the senescence transcriptomic signatures. Utilizing a combination of degenerate cells from elderly humans and experimentally induced senescence in young, healthy IVD cells, the authors show the therapeutic effects on mRNA transcription as well as cellular processes through informatics approaches.

One notable strength of this study is the use of human primary cells and recombinant forms of human PDGF-AB/BB proteins, which increases the translational potential of these in vitro studies. The manuscript is well-written, and the informatics analyses are thorough and clearly presented.

However, in its current form, the study does not provide sufficient experimental details, and clarifications are needed. These are as follows:

(1) The source of PDGF-AB/BB proteins is not detailed.<br /> (2) The irradiation parameters are not adequately reported - the authors should consider (PMCID: PMC5495460) for the parameters that should be reported.<br /> (3) The criteria for young and old patient donors are not explicitly described - though from the table, one presumes the cut-off for young is 27 years old.<br /> (4) What is the rationale for using different concentrations of PDGF-AB/BB in the degenerate cell and irradiation experiments?

There are also a number of other issues the authors could consider. First, in the title and throughout the manuscript, the effects of PDGF-AB/BB are described as protective, yet in all the experiments, PDGF-AB/BB appears to be administered following either in vivo degeneration or in vitro irradiation, where protective effects (e.g., administration prior to insult) were not tested. Therefore, the effects of PDGF-AB/BB may be more accurately described as mitigating or therapeutic rather than protective.

The authors state that the focus on NP (nucleus pulposus) cell studies is due to NP being the first site impacted during degeneration. However, this reviewer believes that this is because changes in the NP are more clinically evident (by imaging methods), despite degeneration often initiating from the AF (annulus fibrosus), e,g. through tears/microtears.

A prior study has examined the effects of X-ray irradiation on NF-kB signaling in young and aged IVDs (PMCID: PMC5495460), and the authors may wish to consider this work.

Reviewer #2 (Public review):

Summary:

This work highlights a novel role for platelet-derived growth factor (PDGF) in mitigating cellular senescence associated with age-related and painful intervertebral disc degeneration. Prior literature has demonstrated the importance of the accumulation of senescent cells in mediating many of the pathological effects associated with the degenerate disc joint such as inflammation and tissue breakdown. In this study, the authors treat clinically relevant human nucleus pulposus and annulus fibrosus cells from patients undergoing discectomy with recombinant PDGF-AB/BB for 5 days and then deep phenotyped the outcomes using bulk RNA sequencing. In addition, they irradiated healthy human disc cells which they subsequently treated with PDGF-AB/BB examining the expression of SASP-related markers and also PDGFRA receptor gene expression. Overall PDGF was able to down-regulate many senescent-associated pathways and the degenerate phenotype in IVD cells. Altered pathways were associated with neurogenesis, mechanical stimuli, metabolism, cell cycle, reactive oxygen species, and mitochondrial dysfunction. Overall the authors achieved their aims and the results by and large support their conclusions although improvements could be made to enhance the rigor of the study and findings.

Strengths:

A major strength of this study is the use of human cells from patients undergoing discectomy for disc herniation as well as access to healthy human cells. Investigating the role of PDGF regarding cellular senescence in the degenerate disc joint is a novel and underexplored area of research which is a significant contribution to the field of spine. This study highlights a potential target for addressing cellular senescence where most of the prior focus has been on senolytic drugs. Such studies have broad implications for other age-related diseases where senescence plays a major role. The use of transcriptomics and therefore an unbiased approach to investigating the role of PDGF is also considered a strength as are the follow-up studies involving irradiating healthy human disc cells and treating these cells with PDGF. The combined assessment of both nucleus pulposus and annulus fibrosus cells in the context of these studies adds to the impact.

Weaknesses:

A weakness of these studies lies in the lack of experimental details provided in the methodology including the rationale for such methods/conditions. Many details such as the specific culture models utilized, substrates, cell density, and media components are missing which impacts rigor. Such details would strengthen the manuscript and the ability to replicate and build on such work/findings. An additional weakness relates to qualitative data presented such as the B-galactosidase assay and immunofluorescence of senescence-associated proteins such as P21 and P16. Quantification of such data sets would greatly strengthen the studies and lend further support to the hypotheses. The study in its current form could be strengthened by the inclusion of mechanistic studies probing the downstream PDGF receptor-associated pathways for example specifically targeting or modulating the activity of the PDGF receptor PDGFRA including validation of the gene data for PDGFRA with protein level data to determine if the transcripts are being translated to protein. The claim that in annulus fibrosus cells, PDGF do not mediate their effects via the PDGFRA does not appear to be supported by the current data as only gene expression for the receptor was assessed with no functional or mechanistic studies being performed. Further discussion, interpretation, and direct comparison of the nucleus pulposus and annulus fibrosus data sets would be helpful for the readers. The magnitude of changes related to the effects of PDGF-BB on the S-phase in irradiated NP and AF cells between control and treated groups seem small making interpretation of these findings challenging.

eLife Assessment

In this manuscript, Yao et al. present solid evidence to show that MnMYL3 may serve as a receptor for NNV via macropinocytosis. This manuscript is valuable for understanding the molecular mechanisms of NNV cell entry. However, the manuscript will benefit from broader implications of these findings for cell entry of other viruses.

Reviewer #1 (Public review):

Summary:

In this manuscript, the authors discovered MYL3 of marine medaka (Oryzias melastigma) as a novel NNV entry receptor, elucidating its facilitation of RGNNV entry into host cells through macropinocytosis, mediated by the IGF1R-Rac1/Cdc42 pathway.

Strengths:

In this manuscript, the authors have performed in vitro and in vivo experiments to prove that MnMYL3 may serve as a receptor for NNV via macropinocytosis pathway. These experiments with different methods include Co-IP, RNAi, pulldown, SPR, flow cytometry, immunofluorescence assays, and so on. In general, the results are clearly presented in the manuscript.

Weaknesses:

For the writing in the introduction and discussion sections, the author Yao et al mainly focus on the viral pathogens and fish in Aquaculture, the meaning and novelty of results provided in this manuscript are limited, and not broad in biology. The authors should improve the likely impact of their work on the viral infection field, maybe also in the evolutionary field with the fish model.

(1) Myosin is a big family, why did authors choose MYL3 as a candidate receptor for NNV?

(2) What is the relationship between MmMYL3 and MmHSP90ab1 and other known NNV receptors? Why does NNV have so many receptors? Which one is supposed to serve as the key entry receptor?

(3) In vivo knockout of MYL3 using CRISPR-Cas9 should be conducted to verify whether the absence of MYL3 really inhibits NNV infection. Although it might be difficult to do it in marine medaka as stated by the authors, the introduction of zebrafish is highly recommended, since it has already been reported that zebrafish could serve as a vertebrate model to study NNV (doi: 10.3389/fimmu.2022.863096).

(4) The results shown in Figure 6 are not enough to support the conclusion that "RGNNV triggers macropinocytosis mediated by MmMYL3". Additional electron microscopy of macropinosomes (sizes, morphological characteristics, etc.) will be more direct evidence.

(5) MYL3 is "predominantly found in muscle tissues, particularly the heart and skeletal muscles". However, NNV is a virus that mainly causes necrosis of nervous tissues (brain and retina). If MYL3 really acts as a receptor for NNV, how does it balance this difference so that nervous tissues, rather than muscle tissues, have the highest viral titers?

Reviewer #2 (Public review):

Summary:

The manuscript offers an important contribution to the field of virology, especially concerning NNV entry mechanisms. The major strength of the study lies in the identification of MmMYL3 as a functional receptor for RGNNV and its role in macropinocytosis, mediated by the IGF1R-Rac1/Cdc42 signaling axis. This represents a significant advance in understanding NNV entry mechanisms beyond previously known receptors such as HSP90ab1 and HSC70. The data, supported by comprehensive in vitro and in vivo experiments, strongly justify the authors' claims about MYL3's role in NNV infection in marine medaka.

Strengths:

(1) The identification of MmMYL3 as a functional receptor for RGNNV is a significant contribution to the field. The study fills a crucial gap in understanding the molecular mechanisms governing NNV entry into host cells.

(2) The work highlights the involvement of IGF1R in macropinocytosis-mediated NNV entry and downstream Rac1/Cdc42 activation, thus providing a thorough mechanistic understanding of NNV internalization process. This could pave the way for further exploration of antiviral targets.

Reviewer #3 (Public review):

Summary:

The manuscript presents a detailed study on the role of MmMYL3 in the viral entry of NNV, focusing on its function as a receptor that mediates viral internalization through the macropinocytosis pathway. The use of both in vitro assays (e.g., Co-IP, SPR, and GST pull-down) and in vivo experiments (such as infection assays in marine medaka) adds robustness to the evidence for MmMYL3 as a novel receptor for RGNNV. The findings have important implications for understanding NNV infection mechanisms, which could pave the way for new antiviral strategies in aquaculture.

Strengths:

The authors show that MmMYL3 directly binds the viral capsid protein, facilitates NNV entry via the IGF1R-Rac1/Cdc42 pathway, and can render otherwise resistant cells susceptible to infection. This multifaceted approach effectively demonstrates the central role of MmMYL3 in NNV entry.

eLife Assessment

This important work develops a new protocol to experimentally perturb target genes across a quantitative range of expression levels in cell lines. The evidence supporting their new perturbation approach is compelling, and the computational analyses to better understand dosage response relationships between genes are convincing. The study will be of broad interest to scientists in the fields of functional genomics and biotechnology. However, the evidence supporting the conclusions can be further improved.

Reviewer #1 (Public review):

In this manuscript, Domingo et al. present a novel perturbation-based approach to experimentally modulate the dosage of genes in cell lines. Their approach is capable of gradually increasing and decreasing gene expression. The authors then use their approach to perturb three key transcription factors and measure the downstream effects on gene expression. Their analysis of the dosage response curve of downstream genes reveals marked non-linearity.

One of the strengths of this study is that many of the perturbations fall within the physiological range for each cis gene. This range is presumably between a single-copy state of heterozygous loss-of-function (log fold change of -1) and a three-copy state (log fold change of ~0.6). This is in contrast with CRISPRi or CRISPRa studies that attempt to maximize the effect of the perturbation, which may result in downstream effects that are not representative of physiological responses.

Another strength of the study is that various points along the dosage-response curve were assayed for each perturbed gene. This allowed the authors to effectively characterize the degree of linearity and monotonicity of each dosage-response relationship. Ultimately, the study revealed that many of these relationships are non-linear, and that the response to activation can be dramatically different than the response to inhibition.

To test their ability to gradually modulate dosage, the authors chose to measure three transcription factors and around 80 known downstream targets. As the authors themselves point out in their discussion about MYB, this biased sample of genes makes it unclear how this approach would generalize genome-wide. In addition, the data generated from this small sample of genes may not represent genome-wide patterns of dosage response. Nevertheless, this unique data set and approach represents a first step in understanding dosage-response relationships between genes.

Another point of general concern in such screens is the use of the immortalized K562 cell line. It is unclear how the biology of these cell lines translates to the in vivo biology of primary cells. However, the authors do follow up with cell-type-specific analyses (Figures 4B, 4C, and 5A) to draw a correspondence between their perturbation results and the relevant biology in primary cells and complex diseases.

The conclusions of the study are generally well supported with statistical analysis throughout the manuscript. As an example, the authors utilize well-known model selection methods to identify when there was evidence for non-linear dosage response relationships.

Gradual modulation of gene dosage is a useful approach to model physiological variation in dosage. Experimental perturbation screens that use CRISPR inhibition or activation often use guide RNAs targeting the transcription start site to maximize their effect on gene expression. Generating a physiological range of variation will allow others to better model physiological conditions.

There is broad interest in the field to identify gene regulatory networks using experimental perturbation approaches. The data from this study provides a good resource for such analytical approaches, especially since both inhibition and activation were tested. In addition, these data provide a nuanced, continuous representation of the relationship between effectors and downstream targets, which may play a role in the development of more rigorous regulatory networks.

Human geneticists often focus on loss-of-function variants, which represent natural knock-down experiments, to determine the role of a gene in the biology of a trait. This study demonstrates that dosage response relationships are often non-linear, meaning that the effect of a loss-of-function variant may not necessarily carry information about increases in gene dosage. For the field, this implies that others should continue to focus on both inhibition and activation to fully characterize the relationship between gene and trait.

Reviewer #2 (Public review):

Summary:

This work investigates transcriptional responses to varying levels of transcription factors (TFs). The authors aim for gradual up- and down-regulation of three transcription factors GFI1B, NFE2, and MYB in K562 cells, by using a CRISPRa- and a CRISPRi line, together with sgRNAs of varying potency. Targeted single-cell RNA sequencing is then used to measure gene expression of a set of 90 genes, which were previously shown to be downstream of GFI1B and NFE2 regulation. This is followed by an extensive computational analysis of the scRNA-seq dataset. By grouping cells with the same perturbations, the authors can obtain groups of cells with varying average TF expression levels. The achieved perturbations are generally subtle, not reaching half or double doses for most samples, and up-regulation is generally weak below 1.5-fold in most cases. Even in this small range, many target genes exhibit a non-linear response. Since this is rather unexpected, it is crucial to rule out technical reasons for these observations.

Strengths:

The work showcases how a single dataset of CRISPRi/a perturbations with scRNA-seq readout and an extended computational analysis can be used to estimate transcriptome dose responses, a general approach that likely can be built upon in the future.

Weaknesses:

(1) The experiment was only performed in a single replicate. In the absence of an independent validation of the main findings, the robustness of the observations remains unclear.

(2) The analysis is based on the calculation of log-fold changes between groups of single cells with non-targeting controls and those carrying a guide RNA driving a specific knockdown. How the fold changes were calculated exactly remains unclear, since it is only stated that the FindMarkers function from the Seurat package was used, which is likely not optimal for quantitative estimates. Furthermore, differential gene expression analysis of scRNA-seq data can suffer from data distortion and mis-estimations (Heumos et al. 2023 (https://doi.org/10.1038/s41576-023-00586-w), Nguyen et al. 2023 (https://doi.org/10.1038/s41467-023-37126-3)). In general, the pseudo-bulk approach used is suitable, but the correct treatment of drop-outs in the scRNA-seq analysis is essential.

(3) Two different cell lines are used to construct dose-response curves, where a CRISPRi line allows gene down-regulation and the CRISPRa line allows gene upregulation. Although both lines are derived from the same parental line (K562) the expression analysis of Tet2, which is absent in the CRISPRi line, but expressed in the CRISPRa line (Figure S3A) suggests substantial clonal differences between the two lines. Similarly, the PCA in S4A suggests strong batch effects between the two lines. These might confound this analysis.

(4) The study uses pseudo-bulk analysis to estimate the relationship between TF dose and target gene expression. This requires a system that allows quantitative changes in TF expression. The data provided does not convincingly show that this condition is met, which however is an essential prerequisite for the presented conclusions. Specifically, the data shown in Figure S3A shows that upon stronger knock-down, a subpopulation of cells appears, where the targeted TF is not detected anymore (drop-outs). Also Figure 3B (top) suggests that the knock-down is either subtle (similar to NTCs) or strong, but intermediate knock-down (log2-FC of 0.5-1) does not occur. Although the authors argue that this is a technical effect of the scRNA-seq protocol, it is also possible that this represents a binary behavior of the CRISPRi system. Previous work has shown that CRISPRi systems with the KRAB domain largely result in binary repression and not in gradual down-regulation as suggested in this study (Bintu et al. 2016 (https://doi.org/10.1126/science.aab2956), Noviello et al. 2023 (https://doi.org/10.1038/s41467-023-38909-4)).

(5) One of the major conclusions of the study is that non-linear behavior is common. This is not surprising for gene up-regulation, since gene expression will reach a plateau at some point, but it is surprising to be observed for many genes upon TF down-regulation. Specifically, here the target gene responds to a small reduction of TF dose but shows the same response to a stronger knock-down. It would be essential to show that his observation does not arise from the technical concerns described in the previous point and it would require independent experimental validations.

(6) One of the conclusions of the study is that guide tiling is superior to other methods such as sgRNA mismatches. However, the comparison is unfair, since different numbers of guides are used in the different approaches. Relatedly, the authors point out that tiling sometimes surpassed the effects of TSS-targeting sgRNAs, however, this was the least fair comparison (2 TSS vs 10 tiling guides) and additionally depends on the accurate annotation of TSS in the relevant cell line.

(7) Did the authors achieve their aims? Do the results support the conclusions?: Some of the most important conclusions are not well supported because they rely on accurately determining the quantitative responses of trans genes, which suffers from the previously mentioned concerns.

(8) Discussion of the likely impact of the work on the field, and the utility of the methods and data to the community:<br /> Together with other recent publications, this work emphasizes the need to study transcription factor function with quantitative perturbations. Missing documentation of the computational code repository reduces the utility of the methods and data significantly.

eLife Assessment

This study presents valuable data on the identification and function of a protein complex present at the Maurer's cleft organelles of Plasmodium falciparum-infected red blood cells. The evidence supporting the findings is solid, but would benefit from greater rigor in presentation and analysis.

Reviewer #1 (Public review):

Summary:

In this paper, Blancke Soares and Stäcker et al serendipitously identify a domain of the Plasmodium falciparum protein MSRP6 that mediates both export from the parasite into the infected red blood cell and association with the Maurer's cleft organelles found in the infected cell. The authors use this domain to identify a putative complex of proteins at the Maurer's cleft via proximity biotinylation. Six members of the complex are confirmed to interact with MSRP6 by co-immunoprecipitation.

The functions of select proteins of this complex are further investigated with regard to the formation of Maurer's clefts. Disruption of PeMP2, PIESP2, and Pf332 resulted in morphological changes to the Maurer's clefts and prevented the anchoring of the Maurer's clefts to the infected red blood cell plasma membrane that normally occurs in the trophozoite stage. Curiously, disruption of MSRP6, the central member of the complex, did not affect Maurer's cleft anchoring. Mechanistically, how this complex affects Maurer's cleft structure and anchoring remains unclear.

Finally, the authors show that the loss of Maurer's cleft anchoring observed upon disruption of PIESP2 or Pf332 does not affect cytoadherence of infected red blood cells via PfEMP1, arguing against a prior assumption that cleft tethering is required for the presentation of parasite-exported proteins on the infected red blood cell surface.

Strengths:

Maurer's clefts are enigmatic organelles found in red blood cells infected by Plasmodium falciparum that are presumed to play a role in trafficking exported parasite proteins to the surface of the red blood cells, though little is known about their biogenesis and function. The authors here convincingly identify a protein complex present at the Maurer's clefts using multiple orthogonal tools, and carry out assays that indicate this protein complex has a role in shaping and anchoring the Maurer's clefts at their final location at the red blood cell membrane. The data indicating that Maurer's cleft anchoring is dispensable for trafficking of P. falciparum exported proteins to the infected red blood cell membrane has implications for understanding the function of this organelle.

Weaknesses:

In many instances, the data lack appropriate controls that would be desirable for the highest level of rigor. Many, if not most, fluorescence microscopy assays lack untagged/parental controls (prepared in parallel and captured with the same settings) that are necessary to determine the validity of the data - that the observed signal is specific to the protein of interest and not due to autofluorescence or bleed-through from other channels. In other cases, wild-type controls are missing where data from disruption mutants are presented. Additionally, while some phenotypes are quantified, others are only qualitatively described where a more thorough quantitative investigation would be valuable. Finally, where phenotypes have been quantified, in many instances it is not clear that the analyses have included biological replicates as would be expected.

Reviewer #2 (Public review):

Summary:

Soares et al characterize several P. falciparum exported proteins that localize to Maurer's Clefts (MCs), membrane structures formed in the host erythrocyte cytosol. MCs are thought to act as sorting stations that mediate the trafficking of effector proteins to the erythrocyte membrane, such as the surface adhesin and major virulence factor PfEMP1. While initially mobile within the host cytosol, MCs become anchored at the erythrocyte periphery around the time PfEMP1 appears on the RBC surface. While MC immobilization is thought to be important for the delivery of PfEMP1 onto the erythrocyte surface, this hypothesis has remained untested due to the lack of mutants that prevent anchoring. The study begins by determining the sequence features able to mediate the export of PF3D7_0830300 and MSRP6, both PEXEL-Negative Exported Proteins (PNEPs) with signal peptides. The authors show that in both proteins, a region downstream of the signal peptide is sufficient to mediate export, indicating the mature N-terminus is also important for the translocation of this type of PNEP, similar to other classes of exported proteins. Surprisingly, an additional C-terminal region of MSRP6 is also sufficient to mediate export when placed downstream of the signal peptide in the absence of other MSRP6 features. This region also mediates recruitment to MCs and was used as BioID bait to identify proximal MC proteins, several of which form a complex with MSRP6. Strikingly, disruption of certain MSRP6 interacting proteins (PeMP2, PIESP2, and Pf332) abolishes MC anchoring and in some cases also results in major changes in MC morphology. Surprisingly, neither PfEMP1 surface display nor cytoadhesion of infected RBCs is impacted in these mutants. This study features an impressive array of genetically modified parasites and will be of broad interest in providing the first functional analysis of MC anchoring, challenging the prevailing model for PfEMP1 trafficking within the infected RBC.

Strengths:

(1) The first section of the paper presents an in-depth dissection of the features that enable the export of signal peptide-containing PNEPs, confirming the mature N-terminus is sufficient for export across all known types of exported proteins. While it remains unknown how these features enable export, the results reinforce the universal importance of the mature N-terminus, whether generated by signal peptidase or Plasmepsin 5.

(2) The discovery that a C-terminal region of MSRP6 (MAD) is also sufficient for export is novel. The authors suggest this may be the result of piggybacking on another exported protein, although the discussion acknowledges there are challenges with this model since unfolding by PTEX would be expected to disrupt these interactions. An alternative might be considered: the related protein MSRP7 is also exported but consists essentially of a signal peptide and MSP7-like domain without the large N-terminal region found in MSRP6. Presumably, the mature N-terminus of MSRP7 mediates export. If MSRP6 is derived from an exported predecessor composed only of the MSP7-like domain (like MSRP7), the MAD domain might retain the ancestral export information near the beginning of the MSP7-like domain. If this were the case, then the MAD domain (3cd region) should only be sufficient to mediate export when positioned immediately after the signal peptide as in the experiment in Fig 3C (SP-3cd-GFP). It would be interesting to determine if an SP-GFP-3cd construct is exported.

(3) Disruption of PeMP2, PIESP2 or Pf332 is found to prevent MC anchoring. This is the most exciting part of the study as it provides the first set of mutants that interfere with anchoring, enabling the surprising observation that MC immobilization is not important for PfEMP1 surface display or cytoadhesion. The MC movement assay is a nice way to visualize anchoring and would be strengthened by a quantitative measure of colocalization between the time-lapse images (ie, Pearson correlation coefficient) to enable a statistical test. The use of SLI to specifically activate a var gene of choice is an exciting new approach that will be of great use to the PfEMP1 field together with the semi-automated binding assay that helps to increase throughput and reduce bias.

Weaknesses:

(1) At least two of the MSRP6 complex members were found to depend on other complex members for MC trafficking: PeMP3 depends on MSRP6 and Pf332 depends on PIESP2 (previously shown by Zhang et al 2018 and confirmed in the present study). While the authors disrupted all seven MSRP6 complex members, the impact on the trafficking of the other complex members was not systematically investigated. It would be particularly interesting to know which (if any) complex members are required for MC recruitment of PeMP2 since this protein is also needed for MC anchoring.

(2) Some images of exported puncta are interpreted as localization to the MCs without a co-marker. Since other compartments have been identified in the RBC cytosol in addition to MCs (ie, J dots), an MC co-marker would help to verify these actually correspond to MCs. For example, in Figure 5B, GEXP18 gives an exported punctate appearance but lack of co-localization with SBP1 in Fig S2B shows that this does not correspond to MCs.

(3) The authors show MAHRP2 localization is not impacted in their PIESP2 and Pf332 mutants and this is interpreted to indicate the tether structures are not disrupted. However, this conclusion requires actual analysis of the tether structures by electron microscopy since MAHRP2 association to MCs may not require tether integrity and could persist even if the tethers are altered or disrupted. Otherwise, this statement should be adjusted. Additionally, since T2A skipping efficiency can vary between constructs, it would be a good idea to perform a western blot to ensure that the SBP1-GFP and MAHRP2-mScarlet signals in Figure 8D,F reflect separated proteins.

(4) The trypsin assays to monitor PfEMP1 surface display would benefit from a more detailed explanation of how the results were interpreted. For instance, though perhaps less intense than in the PIESP2, Pf332, and MSRP6 mutants, a Var01-protected fragment is also seen in the SBP1 mutant. Additionally, a protected fragment is indicated for most of the SBP1N controls (asterisk). As per the author's experimental design (lines 956-957), does this indicate that the RBC membrane was partially compromised during the experiment? In line 505, the trypsin assay data in the mutants is interpreted relative to the parent IT4var01-HA line but no data is shown for the parent.

Reviewer #3 (Public review):

Summary:

Malaria is caused by Plasmodium falciparum parasites that infect, grow, and reproduce inside red blood cells. The parasites extensively modify the blood cells they infect, by exporting hundreds of proteins into the red blood cell compartment. One of the most important modifications made by the parasite is to display adhesive proteins on the blood cell surface which attach the infected cells to walls of small blood vessels. This can lead to organ damage resulting in serious disease complications and there is great interest in blocking the adhesive process to reduce disease. This study investigates the function of an atypical, exported protein that along with other proteins maintains the integrity of membranous sacs formed by the parasite in the blood cell compartment. These sacs are widely believed to help organise the display of the adhesive proteins on the infected blood cell surface. This study challenges this dogma by showing that disruption of the sacs does not prevent the display of the adhesive proteins suggesting alternative pathways are likely involved in adhesive protein display.

Strengths:

The conclusions are supported by a beautiful series of live parasite images.

Weaknesses:

No major weaknesses were identified by this reviewer.

eLife Assessment

Morphological characteristics and phenotypes of mutations in key developmental genes suggest that head, trunk, and tail development are regulated by discernible modules. Gdf11 signalling plays a crucial role in orchestrating the transition from trunk to tail tissues in vertebrate embryos. This important study presents convincing evidence that Tgfbr1 acts upstream of Isl1 (a pivotal effector of Gdf11 signalling) and regulates blood vessels, the lateral plate mesoderm, and the endoderm associated with the trunk-to-tail transition. Together with the previous studies, this work identifies a key signal that acts as the pivot of the trunk-to-tail transition.

Joint Public Review:

Previously, this group showed that Tgfbr1 regulates the reorganization of the epiblast and primitive streak into the chordo-neural hinge and tailbud during the trunk-to-tail transition. Gdf11 signaling plays a crucial role in orchestrating the transition from trunk to tail tissues in vertebrate embryos, including the reallocation of axial progenitors into the tailbud and Tgfbr1 plays a key role in mediating its signaling activity. Progenitors that contribute to the extension of the neural tube and paraxial mesoderm into the tail are located in this region. In this work, the authors show that Tgfbr1 also regulates the reorganization of the posterior primitive streak/base of allantois and the endoderm as well.

By analyzing the morphological phenotypes and marker gene expression in Tgfbr1 mutant mouse embryos, they show that it regulates the merger of somatic and splanchnic layers of the lateral plate mesoderm, the posterior streak derivative. They also present evidence suggesting that Tgfbr1 acts upstream of Isl1 (key effector of Gdf11 signaling for controlling differentiation of lateral mesoderm progenitors) and regulates the remodelling of the major blood vessels, the lateral plate mesoderm and endoderm associated with the trunk-to-tail transition. Through a detailed phenotypic analysis, the authors observed that, similarly to Isl1 mutants, the lack of Tgfbr1 in mouse embryos hinders the activation of hindlimb and external genitalia maker genes and results in a failure of lateral plate mesoderm layers to converge during tail development. As a result, they interpret that ventral lateral mesoderm, which generates the peri cloacal mesenchyme and genital tuberculum, fails to specify.

They also show defects in the morphogenesis of the dorsal aorta at the trunk/tail juncture, resulting in an aberrant embryonic/extraembryonic vascular connection. Endoderm reorganization defects following abnormal morphogenesis of the gut tube in the Tgfbr1 mutants cause failure of tailgut formation and cloacal enlargement. Thus, Tgfbr1 activity regulates the morphogenesis of the trunk/tail junction and the morphogenetic switch in all germ layers required for continuing post-anal tail development. Taken together with the previous studies, this work places Gdf11/8 - Tgfbr1 signaling at the pivot of trunk-to-tail transition and the authors speculate that critical signaling through Tgfbr1 occurs in the posterior-most part of the caudal epiblast, close to the allantois.

The data shown is solid with excellent embryology/developmental biology. This work demonstrates meticulous execution and is presented in a comprehensive and coherent manner. Although not completely novel, the results/conclusions add to the known function of Gdf11 signaling during the trunk-to-tail transition.

Author response:

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

Joint Public Review:

Previously, this group showed that Tgfbr1 regulates the reorganization of the epiblast and primitive streak into the chordo-neural hinge and tailbud during the trunk-to-tail transition. Gdf11 signaling plays a crucial role in orchestrating the transition from trunk to tail tissues in vertebrate embryos, including the reallocation of axial progenitors into the tailbud and Tgfbr1 plays a key role in mediating its signaling activity. Progenitors that contribute to the extension of the neural tube and paraxial mesoderm into the tail are located in this region. In this work, the authors show that Tgfbr1 also regulates the reorganization of the posterior primitive streak/base of allantois and the endoderm as well. 

By analyzing the morphological phenotypes and marker gene expression in Tgfbr1 mutant mouse embryos, they show that it regulates the merger of somatic and splanchnic layers of the lateral plate mesoderm, the posterior streak derivative. They also present evidence suggesting that Tgfbr1 acts upstream of Isl1 (key effector of Gdf11 signaling for controlling differentiation of lateral mesoderm progenitors) and regulates the remodelling of the major blood vessels, the lateral plate mesoderm and endoderm associated with the trunk-to-tail transition. Through a detailed phenotypic analysis, the authors observed that, similarly to Isl1 mutants, the lack of Tgfbr1 in mouse embryos hinders the activation of hindlimb and external genitalia maker genes and results in a failure of lateral plate mesoderm layers to converge during tail development. As a result, they interpret that ventral lateral mesoderm, which generates the peri cloacal mesenchyme and genital tuberculum, fails to specify. 

They also show defects in the morphogenesis of the dorsal aorta at the trunk/tail juncture, resulting in an aberrant embryonic/extraembryonic vascular connection. Endoderm reorganization defects following abnormal morphogenesis of the gut tube in the Tgfbr1 mutants cause failure of tailgut formation and cloacal enlargement. Thus, Tgfbr1 activity regulates the morphogenesis of the trunk/tail junction and the morphogenetic switch in all germ layers required for continuing post-anal tail development. Taken together with the previous studies, this work places Gdf11/8 - Tgfbr1 signaling at the pivot of trunk-to-tail transition and the authors speculate that critical signaling through Tgfbr1 occurs in the posterior-most part of the caudal epiblast, close to the allantois. 

Strengths: 

The data shown is solid with excellent embryology/developmental biology. This work demonstrates meticulous execution and is presented in a comprehensive and coherent manner. Although not completely novel, the results/conclusions add to the known function of Gdf11 signaling during the trunk-to-tail transition. 

Weaknesses: 

The authors rely on the expression of a small number of key regulatory genes to interpret the developmental defects. The alternative possibilities remain to be ruled out thoroughly. The manuscript is also quite descriptive and would benefit from more focused highlighting of the novelty regarding the absence of Tgfbr1 in the mouse embryo. They should also strengthen some of their conclusions with more details in the results.

Although we used a limited number of key regulatory genes to interpret the phenotype, these genes were carefully chosen to focus on specific processes involving the lateral mesoderm, its derivatives, and the endoderm. In addition to these markers, we included references to other relevant markers that were previously analyzed and initially led us to examine the lateral plate mesoderm and tail gut in Tgfbr1 mutants. To strengthen our analysis, we have now incorporated additional data to clarify specific phenotypes. For instance, in situ hybridization (ISH) for Shh further confirms abnormalities at the caudal end of the endoderm in mutant embryos, while no endodermal defects are observed in the trunk region. We also included an analysis of the intermediate mesoderm, which shows abnormalities at the same level as those found in the lateral plate mesoderm and endoderm of Tgfbr1 mutants.

It’s important to note that using additional markers to assess the epiblast/primitive streak of Tgfbr1 mutants at E7.5–E8.5, as suggested by a reviewer, is unlikely to yield new insights. At these early stages, Tgfbr1 mutant embryos do not display observable phenotypes in the main body axis. Data in this manuscript already demonstrate the absence of abnormalities at this stage, as shown in Figure 3 and Supplementary Figure 6. Additionally, the expression of certain genes showing abnormalities when the embryo would enter tail development, in the trunk their expression remains unaffected, indicating that trunk extension is not significantly impacted by Tgfbr1 deficiency. While transcriptomic analysis of these Tgfbr1 mutants could provide interesting insights, it would be more appropriate to focus on later developmental stages, which would be beyond the scope of the current study.

The second major critique was that the manuscript is primarily descriptive. We disagree with this assessment. Several hypotheses were rigorously tested using genetic approaches, including Isl1 knockout experiments, cell tracing from the primitive streak with a newly generated Cre driver to activate a reporter from the ROSA26 locus, and assessment of extraembryonic endoderm fate in Tgfbr1 mutants by introducing the Afp-GFP transgene into the Tgfbr1 mutant background. Additionally, we conducted tracing analyses of tail bud cell contributions to the tail gut via DiI injection and embryo incubation. To address potential concerns regarding this experiment, we have included data showing the DiI position immediately after injection to confirm that it does not contact the tail gut. We also considered and accounted for potential DiI leakage into neuromesodermal progenitors to clarify the endodermal results.

Our genetic and DiI experiments were specifically designed to differentiate between alternative hypotheses and to confirm hypotheses generated from other analyses. Additionally, improvements in some of the imaging data have helped address remaining concerns.

Reviewer #1 (Recommendations For The Authors): 

I have listed my suggestions as queries. The authors may perform experiments or clarify by editing the text to address them. 

The authors state on Page 11 and elsewhere that the ventral lateral mesoderm is absent in the Tgfbr1 mutant. What is the basis for this conclusion? Are there specific markers for PCM or GT primordium? 

The specific marker of PCM and GT primordium is Isl1. The absence of this marker in the Tgfbr1 mutants is shown in (Dias et al, 2020). The reference is introduced in the manuscript.

A schematic illustrating the VLM and the expression patterns of Tgfbr1, Gdf11, etc., would be helpful. 

Characterization of Gdf11 expression has been previously reported (e.g. McPherron et al 1999, cited in our manuscript). It is expressed in the region containing of axial progenitors before the trunk to tail transition and not expressed in the VLM. As for Tgfbr1 expression is hard to detect, likely because it is ubiquitously expressed at low level. We include in this document some pictures of an ISH, including a control using the Tgfbr1 mutants to illustrate that the staining resembling background actually represents Tgfbr1 expression. If the reviewers find it important, we can also incorporate these data into the manuscript. Under these circumstances, we feel that a schematic might not be very informative.

Author response image 1.

Image showing an example of an ISH procedure with a probe against Tgfbr1, showing widespread and low expression. The lower picture shows a ventral view of a stained wild type E10.5 embryo.

Foxf1+ cells in the 'extended LPM' of Tgfbr1 mutants suggest fate transformation, or does it indicate the misexpression of marker gene otherwise suppressed by Tgfbr1 activity? The authors suggest that Foxf1+ cells are VLM progenitors from posterior PS trapped in the extended LPM. Do they continue to express PS markers? 

The observation that both in wild type and Tgfbr1 mutant embryos Foxf1 expression in the trunk is restricted to the splanchnic LPM indicates that the absence of this marker in the somatic LPM is not the result of a suppression of its expression by Tgfbr1. In wild type embryos Foxf1 is also expressed in the posterior PS, regulated independently of its expression in the LPM (i.e. Shh-independent) and later in the pericloacal mesoderm (our supplementary figure 2). As Foxf1 expression in the posterior PS was not suppressed in the Tgfbr1 mutants, together with the absence of pericloacal mesoderm, we interpret that the Foxf1-positive cells in the two layers around the extended celomic cavity in the posterior end of the mutant embryos derived from the posterior PS, resulting from the absence of its normal progression through the embryonic tissues.

We did not find expression of PS markers giving rise to paraxial mesoderm, like Tbxt, further suggesting that those cells could derive from the restricted set of cells within the posterior PS that contribute to the pericloacal mesoderm

For example, the misexpression of Apela is interpreted as mis-localized endoderm cells. They show scattered Keratin 8 misexpression to support the interpretation. It would be more convincing if the authors tested the expression of other endoderm markers. 

As indicated in the manuscript, we suggest that these cells are endoderm progenitors (p. 13), like those present at the posterior end of the gut tube at E9.5 and E10.5, that are unable to incorporate into the gut tube. Apela is not a general endodermal marker: it is expressed in the foregut pocket and the nascent cells of the hindgut/tail gut, becoming down regulated as cells take typical endodermal signatures. The presence of ectopic Apela expression in the extended LPM of the mutant embryos might indeed indicate the presence of progenitors that failed to downregulate Apela resulting from the lack differentiation-associated downregulation. This would also implicate the absence of definitive endodermal markers.

The Nodal signaling pathway in the anterior PS drives endoderm development. It acts through Alk7. Does Tgfbr1 (Alk5) mutation impact endoderm development, in general? It isn't easy to assess this from the Foxa2 in situ RNA hybridization shown in Figures 6A and B. It would be helpful for the readers if the authors clarified this point. 

In the pictures shown in Figure 7D-D’ it is already shown that the endoderm is mostly preserved until the region of the trunk to tail transition. The presence of a rather normal endoderm in the embryonic trunk can also be seen with Shh, a figure added as Supplementary Fig.5.

Reviewer #2 (Recommendations For The Authors): 

The authors mention two interesting novel points which they should develop in the discussion, and probably also in the results. 

(1) The authors speculate about the possible involvement of the posterior PS as a mediator of Gdf11/Tgfbr1 signaling activity. However, as mentioned in the manuscript, their experiments do not allow regional sublocalization within the PS... Here it would be important to assess/discuss in more detail which progenitors respond to this signaling activity and when they do it. At the very least, the authors should provide high-resolution spatiotemporal data of the expression of Tgfbr1 in the PS. 

Tgfbr1 expression at this embryonic stage does not give clear differential patterns. The data reported for this expression in Andersson et al 2006 is very low quality and we have not been able to reproduce the reported pattern. On the contrary, all our efforts over the years provided a very general staining that could even be interpreted as background. When we now included Tgfbr1 mutants as controls, it became clear that the ubiquitous and low level signal observed in wild type embryos indeed represent Tgfbr1 expression pattern: low level and ubiquitous. We are attaching a figure to this document illustrating these observations. If required, this can also be included in the manuscript as a supplementary figure. 

Also, the work of Wymeersch et al., 2019 regarding the lateral plate mesoderm progenitors (LPMPs) should be referred to and discussed here. 

This was now added in the results (page 11) and in discussion (page 16). 

For instance, are the LPMP transcriptomic differences detected between E7.5 and E8.5 caused by Tgfbr1 signaling activity? This question could be easily answered through a comparative bulk RNAseq analysis of the posterior-most region of the PS of mutant and WT embryos. The possible colocalization of Tgfb1 (Wymeersch et al., 2019) and Tgfbr1 in the LPMPs should also be addressed. 

We agree with the suggestion that RNA-seq in the posterior PS of WT and mutant embryos might be informative. However, it is very likely that within the proposed timeframe (E7.5 to E8.5) that there are no significant differences between the wild type and the Tgfbr1 mutant embryos because there is no apparent axial phenotype in Tgfbr1 mutant embryos before the trunk to tail transition. Therefore, at this stage, we think that this experiment is out of the scope of the present manuscript. 

(2) The activity of Tgfbr1 during the trunk-to-tail transition is critical for the development of tail endodermal tissues. Here the authors suggest again the involvement of the posterior PS/allantois region, but a similar phenotype can also be observed for instance in the absence of Snai1 in the caudal epiblast (Dias et al., 2020)... It would be important to assess/discuss the origin of those morphogenetic problems in the gut. Is it due to the reallocation of NMC cells into the CNH? The tailbud-EMT process? LPMPs specification?... Regional mutations or gain of functions of Snai1 or Tgfbr1 in the caudal epiblast would help answer the question.  

The endodermal phenotype in the Snai1 mutants is different to that observed in the Tgfbr1 mutants. As can be observed in Figures 3, 4 and 5 of Dias et al. the absence of tailbud is replaced by a structure that extends the epiblast. As a consequence, the endoderm finishes at the base of that structure, even expanding to make a structure resembling the cloaca, which is different to what is seen in the Tgfbr1 mutants. In this case, the lack of tail gut is likely to result either from the lack of formation of the progenitors of the gut endoderm or from the dissociation of what would be the tail bud from the LPM. Actually, hindlimb/pericloacal mesoderm markers, like Tbx4, are preserved in the Snai1 mutant. As for the gain of function of Snai1 experiment, already reported also in Dias et al 2020, the destiny of these cells is not clear. The ISH for Foxa2 showed extra signals but as it is not an exclusive marker for endoderm it is not possible to know whether any of these signals correspond to endodermal tissues.

Regarding the development of tail endodermal tissues, the authors suggest that it occurs from a structure derived from the PS that is located posteriorly, in the tailbud, after the tip of the growing gut. This is an important and novel point as it suggests that the primordia of the endoderm is not wholly specified during gastrulation. So the observation should be well supported. How can Anastasiia et al. distinguish such "structure" from the actual developing gut? Does it have a distinct molecular signature or any morphological landmark that enables its separation from the actual gut? The data suggests that the region highlighted in Supplementary Figure 4Ab contains part of the actual gut tube (the same is suggested in Figure 5B). If the authors think otherwise, they must characterize that region of the tailbud by doing a thorough morphological and gene/protein expression analysis and assess its potency, via transplantation experiments. Also, the authors' claim mostly relies on the DiI experiments and those have three problems: #1 Anastasiia et al. assess "tail" endodermal growth at E9.5 when the correct stage to do it is after E10.5 (after tailbud formation). 2# Incongruencies, low number (only three embryos), and diversity in the results shown in Figure 8 and Supplementary Figure 4. For instance, despite similar staining at 0h, the extension and amount of DiI present in the gut tube after 20h varies significantly amongst the differently labeled embryos. A possible explanation lies in the abnormal leakiness of the DiI labelings and that is confirmed by the observations shown in Supplementary Figure 4M-O; the same for Supplementary Figure 4G, which shows a substantial amount of DiI in the neural tube. 3# The authors must provide high-quality data showing which tissues/regions were labelled at time 0h, including transversal and sagittal sections as they did for the 20h time-point. Additionally, it is important to re-orient the sagittal optical sections to a position that also shows the neural tube (like a mid-sagittal section) and include information concerning the AP/DV axis, as well as the location of the transversal optical sections in the sagittal image. 

As described in the reply to reviewer 1, Apela is expressed in the nascent tail gut endoderm but not in more anterior areas except for a foregut pocket, and becomes downregulated as the tube acquires endodermal signatures. Therefore, the structure to which the reviewer refers to might indeed represent a group of progenitors that extend the tail gut. And the observation that this property is observed only in the tail gut as it grows, already separates this region of the gut, which in the end do not contribute to mature organs, from more anterior areas of the endoderm (essentially anterior to the cloaca) that will become a relevant tissue of the intestinal organs. Our DiI labelling experiment was aimed to test whether this pool of cells contributes to the gut but does not allow to determine the nature of those cells, a question that will require further research (discussed on p. 17) and we think is beyond the scope of the present manuscript.

Regarding the labelling at E10.5, we agree that the tail bud in terms of NMCs is not completely formed, for example, at E9.5 the neuropore is not yet closed. However, we are more interested in regression of the epiblast, which is complete by E9.5. Injecting at E9.5 also has technical advantages for us, first, because in our hands earlier embryos grow better in culture, and second, because it is easier to inject in the tailbud at E9.5 because it is a little bit bigger than at E10.5. Therefore, injecting at E9.5 is less prone to technical artifacts due to injection inaccuracy and compromised growth in culture.

We agree that the injected DiI could also leak into NMPs, which might be located in the same area. However, while this could result in labeling of the neural tube, it would not affect the interpretation of the finding of labeled cells in the tail gut. Indeed, the presence of this label in the gut epithelium indicates the presence of progenitors in the injected region of the tail gut. We added some considerations of this the possible leakage into the results section of the manuscript (p. 15). We thank the reviewer for drawing our attention to this issue. 

We also now provide high quality data showing labelled tissue at 0h in Supplementary figure 8A-c’, higher magnification images in Fig. 8, and reoriented optical sections in Fig.6 and in Supplementary Fig. 7, including axis and location of the sections as suggested by the reviewer.

Minor concerns/comments: 

(1) The abstract is quite long, though this might be fine for this journal. 

(2) In relation to the comment on the abstract, the manuscript needs an initial Figure descrbing the events that are described in the introduction. Otherwise, the manuscript will only be accessible to mouse embryologists.

We have a figure summarizing the results at the end of the manuscript, we think that including similar figure in the beginning might be redundant. What we could do, if required, is to include this type of schematic as a graphical abstract.

(3) The authors need to clarify what they mean when they use the following expressions "PS fate" and "fate of the posterior PS".

I do not think that we have used such expressions. Indeed, they did not come out when we run a “find” in the word document. However, they would mean the tissue that would come out from them at later developmental stages.

(4) The assessment of Isl1 expression in Tgfbr1 mutant and transgenic mouse embryos would be better indicative of their molecular relationship than a comparative phenotypic analysis. 

These data have been reported in Dias et al 2020 and Jurberg et al 2013, both cited in the manuscript.  

(5) The authors should explain or discuss what the upregulation of Foxa2 in the posterior end of Tgfbr1 mutants means.

While an upregulation is apparent in the figure, looking at other pictures we cannot be sure of this being a significantly quantifiable up-regulation. We therefore removed the statement from the text.

(6) What happens to the intermediate mesoderm during the trunk-to-tail transition? Is Tgfbr1 involved in the regulation of its development?

We have tested this using Pax2 and added the relevant data in Supplementary Fig. 1 and described in the results.

(7) The term "potential" should not be used during the description of DiI labeling experiments as this technique only assesses cell fate.

Corrected

(8) Some figures lack AP/DV axis information (e.g. Figures 6, C, and D).

Corrected

eLife Assessment

This study provides a methodological report on a modified adaptive sampling approach, multiple walker supervised molecular dynamics (mwSuMD), and its application to G protein-coupled receptors (GPCRs), which are the most abundant membrane proteins and key targets for drugs. The mwSuMD approach assists in sampling complex binding processes, leading to useful findings for GPCR activity, although results may be considered incomplete, given the high RMSD values reported and lack of validation using experimental data. The manuscript also needs corrected descriptions of high-resolution PDB structures and better relation to existing computational literature.

Reviewer #1 (Public review):

Summary:

The authors investigate ligand and protein-binding processes in GPCRs (including dimerization) by the multiple walker supervised molecular dynamics method. The paper is interesting and it is very well written.

Strengths:

The authors' method is a powerful tool to gain insight on the structural basis for the pharmacology of G protein-coupled receptors.

Reviewer #2 (Public review):

The study by Deganutti and co-workers is a methodological report on an adaptive sampling approach, multiple walker supervised molecular dynamics (mwSuMD), which represents an improved version of the previous SuMD.<br /> Case-studies concern complex conformational transitions in a number of G protein Coupled Receptors (GPCRs) involving long time-scale motions such as binding-unbinding and collective motions of domains or portions. GPCRs are specialized GEFs (guanine nucleotide exchange factors) of heterotrimeric Gα proteins of the Ras GTPase superfamily. They constitute the largest superfamily of membrane proteins and are of central biomedical relevance as privileged targets of currently marketed drugs.<br /> MwSuMD was exploited to address:

a) binding and unbinding of the arginine-vasopressin (AVP) cyclic peptide agonist to the V2 vasopressin receptor (V2R);<br /> b) molecular recognition of the β2-adrenergic receptor (β2-AR) and heterotrimeric GDP-bound Gs protein;<br /> c) molecular recognition of the A1-adenosine receptor (A1R) and palmotoylated and geranylgeranylated membrane-anchored heterotrimeric GDP-bound Gi protein;<br /> d) the whole process of GDP release from membrane-anchored heterotrimeric Gs following interaction with the glucagon-like peptide 1 receptor (GLP1R), converted to the active state following interaction with the orthosteric non-peptide agonist danuglipron.

The revised version has improved clarity and rigor compared to the original also thanks to the reduction in the number of complex case studies treated superficially.<br /> The mwSuMD method is solid and valuable, has wide applicability and is compatible with the most world-widely used MD engines. It may be of interest to the computational structural biology community.<br /> The huge amount of high-resolution data on GPCRs makes those systems suitable, although challenging, for method validation and development.<br /> While the approach is less energy-biased than other enhanced sampling methods, knowledge, at the atomic detail, of binding sites/interfaces and conformational states is needed to define the supervised metrics, the higher the resolution of such metrics is the more accurate the outcome is expected to be. Definition of the metrics is a user- and system-dependent process.

Reviewer #3 (Public review):

Summary:

In the present work Deganutti et al. report a structural study on GPCR functional dynamics using a computational approach called supervised molecular dynamics.

Strengths:

The study has potential to provide novel insight into GPCR functionality. Example is the interaction between D344 and R385 identified during the Gs coupling by GLP-1R. However, validation of the findings, even computationally through for instance in silico mutagenesis study, is advisable.

Weaknesses:

No significant advance of the existing structural data on GPCR and GPCR/G protein coupling is provided. Most of the results are reproductions of the previously reported structures.

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

The authors investigate ligand and protein-binding processes in GPCRs (including dimerization) by the multiple walker supervised molecular dynamics method. The paper is interesting and it is very well written.

Strengths:

The authors' method is a powerful tool to gain insight into the structural basis for the pharmacology of G protein-coupled receptors.

Weaknesses:

Cholesterol may play a fundamental role in GPCR dimerization (as cited by the authors, Prasanna et al, "Cholesterol-Dependent Conformational Plasticity in GPCR Dimers"). Yet they do not use cholesterol in their simulations of the dimerization.

We thank Reviewer #1 for the positive comment on mwSuMD.

In the revised version of the manuscript, the section about the A_2A/D2 receptors dimerization has been removed because largely speculative. We agree that the lack of cholesterol in those simulations added uncertainty to the presented results.

Reviewer #2 (Public Review):

The study by Deganutti and co-workers is a methodological report on an adaptive sampling approach, multiple walker supervised molecular dynamics (mwSuMD), which represents an improved version of the previous SuMD.

Case-studies concern complex conformational transitions in a number of G protein Coupled Receptors (GPCRs) involving long time-scale motions such as binding-unbinding and collective motions of domains or portions. GPCRs are specialized GEFs (guanine nucleotide exchange factors) of heterotrimeric Gα proteins of the Ras GTPase superfamily. They constitute the largest superfamily of membrane proteins and are of central biomedical relevance as privileged targets of currently marketed drugs.

MwSuMD was exploited to address:

(1) Binding and unbinding of the arginine-vasopressin (AVP) cyclic peptide agonist to the V2 vasopressin receptor (V2R);

(2) Molecular recognition of the β2-adrenergic receptor (β2-AR) and heterotrimeric GDPbound Gs protein;

(3) Molecular recognition of the A1-adenosine receptor (A1R) and palmitoylated and geranylgeranylated membrane-anchored heterotrimeric GDP-bound Gi protein;

(4) The whole process of GDP release from membrane-anchored heterotrimeric Gs following interaction with the glucagon-like peptide 1 receptor (GLP1R), converted to the active state following interaction with the orthosteric non-peptide agonist danuglipron;

(5) The heterodimerization of D2 dopamine and A2A adenosine receptors (D2R and A2AR, respectively) and binding to a bi-valent ligand.

The mwSuMD method is solid and valuable, has wide applicability, and is compatible with the most world-widely used MD engines. It may be of interest to the computational structural biology community.

The huge amount of high-resolution data on GPCRs makes those systems suitable, although challenging, for method validation and development.

While the approach is less energy-biased than other enhanced sampling methods, knowledge, at the atomic detail, of binding sites/interfaces and conformational states is needed to define the supervised metrics, the higher the resolution of such metrics is the more accurate the outcome is expected to be. The definition of the metrics is a user- and system-dependent process.

The too many and ambitious case-studies undermine the accuracy of the output and reduce the important details needed for a methodological report. In some cases, the available CryoEM structures could have been exploited better.

The most consistent example concerns AVP binding/unbinding to V2R. The consistency with CryoEM data decreases with an increase in the complexity of the simulated process and involved molecular systems (e.g. receptor recognition by membrane-anchored G protein and the process of nucleotide exchange starting from agonist recognition by an inactive-state receptor). The last example, GPCR hetero-dimerization, and binding to a bi-valent ligand, is the most speculative one as it does not rely on high-resolution structural data for metrics supervision.

We praise Reviewer #2 for the detailed comment on the manuscript. In this revised version, the hetero-dimerization between A_2AR and D₂R has been removed. Also, results about GPCR case studies other than GLP-1R have been reduced and downgraded in importance to focus on the fundamental key points of the adaptive sampling method.  We agree that the consistency with cryoEM data tends to decrease with an increase in the complexity of the simulated process and involved molecular systems. While it is possible to approximate cryoEM results  our unbiased adaptive sampling technique finds its most interesting application in mechanistically unknown out-of-equilibrium processes rather than reproducing known experimental data perfectly. The simulated case studies we present showcase the versatility, speed and consistency of our adaptive method to explore energetically unbiased transitions.

Reviewer #3 (Public Review):

Summary:

In the present work, Deganutti et al. report a structural study on GPCR functional dynamics using a computational approach called supervised molecular dynamics.

Strengths:

The study has the potential to provide novel insight into GPCR functionality. An example is the interaction between loops of GPCR and G proteins, which are not resolved experimentally, or the interaction between D344 and R385 identified during the Gs coupling by GLP-1R. However, validation of the findings, even computationally through for instance in silico mutagenesis study, is advisable.

Weaknesses:

In its current form, the manuscript seems immature and in particular, the described results grasp only the surface of the complex molecular mechanisms underlying GPCR activation. No significant advance of the existing structural data on GPCR and GPCR/G protein coupling is provided. Most of the results are a reproduction of the previously reported structures.

We thank Reviewer #3 for the positive comment on the work. The revised manuscript focuses more on the GLP-1R and Gs case studies. We believe it addresses the weaknesses raised by showing the behaviour of key structural motifs and providing new hypotheses about GDP release.  

Reviewer #2 (Recommendations For The Authors):

In this methodological report, Deganutti and co-workers propose an improved version of supervised molecular dynamics (SuMD), named multiple walker SuMD (mwSuMD). Such an adaptive sampling method was challenged in simulations of complex transitions involving GPCRs, which are out of reach by classical MD.

Although less energy-biased than other enhanced sampling methods, mwSuMD requires knowledge of the atomic detail of the ligand-protein or protein-protein binding site/interfaces and the structural hallmarks of the states whose conversion the method is going to address. Such knowledge is, indeed, necessary to define the supervised metrics (e.g. distances, RMSD, etc), which is a user- and system-dependent process.

We classify mwSuMD as an adaptive, rather than enhanced, sampling method as it does not use any energy bias. We agree with the Reviewer that some knowledge of the system is required to productively set up the simulations, but this is the case for almost any MD advanced methods.  

The text requires improvement in the essential methodological details and cleaning of those parts is not properly instrumental in method validation.

While attempting to prove the widest possible applicability of the method, the authors exaggerated the number of examples, which, in spite of the increasing complexity were only summarily described. Please, limit the case studies to AVP binding/unbinding to V2R and the whole process of GDP release from membrane-anchored Gs following activation of GLP1R by danuglipron. The latter case, indeed, involves small ligand binding (danuglipron), small ligand dissociation (GDP), receptor activation, and activated receptor binding to membraneanchored G protein and G protein conformational transition instrumental to nucleotide depletion, which is already too much. In this framework, the cases of Gs-β2AR and Gi-A2R recognition are redundant. Most importantly, the case of D2R-A2AR heterodimerization and binding to a bi-valent ligand must be eliminated. The reason is that the case is not entirely based on the mwSuMD and the biased protein-protein interface does not rely on highresolution data (i.e. no structural model of D2R-A2AR dimer has been determined so far). Last but not least, the high intrinsic flexibility of the bi-valent ligand adds further indetermination to the computational experiment. Being too speculative, the case-study does not serve to model validation.

We thank the Reviewer for the suggestion. In the current revised form, the manuscript focuses on AVP binding/unbinding to V2R and the GLP-1R activation, Gs recognition and GDP release.

While eliminating the three case studies mentioned above, the remaining ones should be described more extensively and clearly, highlighting the most productive setup for each system. Incidentally, listing the performance parameters (e.g. distribution mode and minimum RMSD) of each simulation setting in Table S1 is worth doing.

More accuracy in the methodological description is needed.

As for the supervised metrics, the rationale behind the choice of a particular index and whether it is the outcome of a number of trials must be declared and the selected indices must be better defined. Here there are a few examples.

AVP-V2R case. It is not clear why the AVP centroids were computed on residues C1-Q4 (I suppose the Cα-atoms) and not on the Cα-atoms of the whole cyclic part (C1-C6). Along the same line, the choice of the Cα-atoms of four amino acid residues to compute the receptor binding-site centroids requires justification.

We have amended the text to clarify that all the heavy atoms of AVP residues C1-Q4, which are anticipated to bind deep into V₂R, were considered alongside V₂R residues part of the peptide binding site (Cα atoms only). From our experience, the choice of including side chains or not for the definition of centroids usually does not affect the supervision output. It should only affect the output of mwSuMD simulations based on the RMSD which considers the specific relative distance from the reference. However, a benchmark of the differences produced by divergent selections is beyond the scope of the present work.

GLP1R case. The statement: "Since the opening of TM1-ECL1 was observed in two replicas out of four, we placed the ligand in a favorable position for crossing that region of GLP-1R" is rather weak as a strategy to manually (?) define the input position of the ligand.

As stated in the manuscript, placing the agonist in that position was driven by preliminary 8 μs of classic MD simulations that pointed out the possible path for binding.  We agree with the Reviewer that there is still some degree of arbitrarity in it and for this reason, we have not presented structural details of the F06882961 binding path.

As for the supervised metrics, what does it mean "the distance between the ligand and GLP-1R TM7 residues L3797.34-F3817.36"? Was the distance computed between ligand and L379-F381 centroids? Also: "In the supervised stages, the distance between residues M386-L394 Gas of helix 5 (α5) and the GLP-1R intracellular residues R1762.46, R3486.37, S3526.41, and N4057.60 was monitored" was it an inter-centroid distance? Furthermore, "supervising the distance between AHD residues G70-R199 Gas and K300-L394Gas" was it the distance between the centroid of the AHD and the centroid of the C-terminal half of the Ras-like domain? In general, when more than two atoms are involved in distance calculation, please, specify if the distance is inter-centroid.

Also: "During the third phase, the RMSD of PF06882961, as well as the RMSD of ECL3 (residues A3686.57-T3787.33, Ca atoms), were supervised" was the RMSD computed without superimposing the ligand to estimate its roto-translations?

We have added details about the selections used for computing centroids throughout the methods section. For example, all the heavy atoms of F06882961 and the Ca atoms of L379-F381 were considered. RMSD values during GLP-1R activation were computed after superimposition on TM2, ECL1, and TM3 residues 170-240 (Ca atoms). This now has been specified in the text.

The authors considered the 7LCJ GLP1R-danuglipron complex as a fully active reference state instead of considering the receptor from a ternary complex with Gs. The ternary complex (7LCI) was indeed considered as a reference only in simulations of receptor-G protein recognition. 

7LCJ and 7LCI are both fully active states. The main difference is that in 7LCJ, Gs coordinates were not deposited. Indeed, their RMSD computed on the TMD Ca atoms and F06882961 is 0.63 Å and 0.54 Å, respectively.

Most importantly, the ternary complex chosen by the authors is not adequate as a reference for simulating the "opening" of the AHD because it bears a miniGs, hence, missing the AHD. In that framework, such an opening is rather vague and was not properly supervised by mwSuMD. The authors must repeat metrics supervisions by using, as a reference, the 6X1A ternary complex, which bears a displaced AHD. This would likely lead to a different path of GDP release.

To the best of our knowledge, there is no evidence that a specific open conformation of the AHD is linked to GDP release. In support, we note that in GPCR ternary complexes, the AHD is usually not modelled because of its high flexibility. The only body of evidence we are aware of is that AHD must open up to allow GDP release. For this reason,  we decided to supervise the distance between AHD and the Ras domain without using a reference.

In the statement: "The AHD opening was simulated starting from the best GLP-1R:Gs binding mwSuMD replica" the definition "best binding" requires clarification.

This has been amended, specifying that Replica 2 was considered the “best replica” due to the closed deviation to the cryoEM structure.

As for the case study on β2-AR-Gs recognition, I strongly suggest to eliminate it. However, I'd like to make some comments. The sentence: "the adrenergic β2 receptor (b2 AR) in an intermediate active state was downloaded from GPCRdb (https://gpcrdb.org/)" is vague as it does not indicate what intermediate active state structure was used. Since the goal of the case study was to probe the method in simulating receptor-G protein binding, it would have been better to start with a fully active state of the receptor like the 4LDO structure, employed by the authors only to extract epinephrine.

mwSuMD is designed to provide insights into structural transitions. We started from an intermediate active state of β2-AR in complex with adrenaline because resembling the most populated state stabilised by a full agonist according to NMR studies (DOI:10.1016/j.cell.2015.08.045); the fully-active β2-AR conformation is stabilized only after Gs binding. However, following the Reviewer’s suggestion, we have reduced the presented results for the β2-AR-Gs recognition.

Also in this case, it is not clear if the supervised receptor-G protein distance is between the centroid of the whole 7-helix bundle and the centroid of Gs α5. It is not clear why the TM6 RMSD concerned only the cytosolic end of the helix and did not include the kink region. With that selection, to estimate the outward displacement, RMSD should have been computed without superimposing the considered portion (once all remaining Cα-atoms of the receptors are superimposed).

As the Reviewer pointed out above, some knowledge of the system is required to set up mwSuMD. Using more generic metrics as we did in this case, like the distance between the whole TMD and Gs α5 represents a general approach applicable to other GPCRs, that should allow orthogonal metrics to evolve independently from the supervision.

As now specified in the text, the superimposition for RMSD calculation was performed on residues 40 to 140 Ca atoms, hence not considering TM6.

As for the A1R-Gi recognition, as already stated, I strongly suggest eliminating it. However, I'd like to add some comments. I would discourage the employment of an AlphaFold model for simulations deputed to model validation in general and, in particular, when highresolution structures are available. In this case, the authors would have used the 1GP2 structure of heterotrimeric Gi no matter if from the rat species.

Following the Reviewer’s suggestion, we have dramatically reduced the results presented for the A1R-Gi recognition. We considered 1GP2 for the simulations but H5 lacks the Cterminal six residues and therefore some extent of modelling was still necessary. However, we take the Reviewer’s comment on board and consider it for future work.

Also, the palmitoylation and geranylgeranylation process is quite tortuous and it is not clear why the NVT ensemble was employed in the second stage of equilibration. This is reflected also on the GLP1R case study.

We have amended the text to clarify this passage. The second NVT stage is required for stabilizing the G protein and its orientation in the simulation box. The figure below shows that a plateau of the Ca RMSD during the NVT step was reached after 700 ns for both Gi (black) and Gs (orange).

Author response image 1.

Here, it is not clear if the RMSD of α5 of Gi was computed with or without superposition.

The RMSD of α5  was computed after superimposing on A₁R residues 40-140 Ca atoms (the less flexible region of the receptor). We have now amended the text to report this information. 

Reviewer #3 (Recommendations For The Authors):  

Points to address:

(1) Root Mean Square Deviation (RMSD) data are often reported as minimum values. It would be useful to provide the average value along the stable part of the trajectories. From the plots in Figure 2ab, it seems that the minimum values reported in the paper are very far from the average ones and thus represent special cases that are seldom reached during simulation. The authors should clarify this point;

For the revised manuscript, we moved Figure 2 to the supplementary material and added average RMSD values for the most notable replicas in Figures 4e and S8a,b. As a reference, in the text, we now report RMSDs from our previous classic MD simulations (https://doi.org/10.1038/s41467-021-27760-0) of Gs:GLP-1R cryoEM structure (G_α = 6.18 ± 2.40 Å; G_β \= 7.22 ± 3.12 Å; G_γ = 9.30 ± 3.65 Å) which show how flexible G proteins bound to GPCRs are and give better context to the RMSD values we measured during mwSuMD simulations.

(2) The RMSD values reported in the paper always refer to single molecules or proteins. It would be useful to also report the RMSD computed over the whole complexes (ligand/GPCR or GPCR/G protein). It would provide a better metric for understanding the general distance between the results and the reference experimental structures;

We have now removed the results sections for A₁R and β₂ AR to focus on GLP-1R, whose RMSD is analyzed in detail in Figures 2, 3 and 4.

(3) A number of computational works investigated the GPCR/G protein interaction and these studies should be cited and discussed. Examples are the works from Mafi et al. 2023 (doi: 10.1038/s41557-023-01238-6), Fleetwood et al. 2020 (doi: 10.1021/acs.biochem.9b00842), Calderon et al. 2023 and 2024 (doi: 10.1021/acs.jcim.3c00805 and doi: 10.1021/acs.jcim.3c01574), Maria-Solano and Choi 2023 (doi: 10.7554/eLife.90773.1), Mitrovic et al. 2023 (doi: 10.1021/acs.jpcb.3c04897), and D'Amore et al. 2023 (doi: 10.1101/2023.09.14.557711). Many of these works focused on the activation of B2AR and the interaction with its G protein. In addition, Maria-Solano and Choi 2023 and D'Amore et al. 2023 also characterized the rotation of TM6 during the A1R and A2AR activation. Therefore, the claim "To the best of our knowledge, this is the first time an MD simulation captures the TM6 rotation upon receptor activation as results reported so far are largely limited to the TM6 opening and kinking55." is untimely;

We thank the Reviewer for the suggested references. We have added them to the introduction as examples of energy-biased (Calderon et al. 2023 and 2024, Maria-Solano and Choi, Mitrovic et al., D'Amore et al) or adaptive sampling (Fleetwood et al) approaches to GPCR. Since the above articles focus on β₂  AR and A₁R, we do not discuss them in detail because the results sections for A₁R and b₂ AR have been drastically reduced in the manuscript.

We note that among the suggested references, only Mafi et al report about a simulated G protein (in a pre-formed complex) and none of the work sampled TM6 rotation without input of energy. However, we have removed the claim from the text.   

(4) In the discussion section, the authors claim that a distance-based approach can be employed when the structural data of the endpoints is limited. However, the results obtained from the distance-based protocol during the validation of the approach, which was done using V2R as a reference, are unsatisfying, as acknowledged by the authors themselves. For instance, the RMSD mode value reported for the AVP C alpha atoms with respect to 7DW9 is high, 0.7 nm, whereas the minimum value is 0.38 nm. In addition, some side chains are not oriented in the experimental conformation and might have a different interaction pattern with the receptor if compared with the experimental structure. Considering that in this case the endpoint is known, it is plausible that the performance of the method would degrade even further when data about the target structure is limited. In a real case scenario, the ligand binding mode is unknown and in such a case no RMSD matrix can be used. This represents the major concern of this study that is no prediction is provided, but only - rather inaccurate - reproduction of the known structural data;

The goal of the first part of the work was to compare mwSuMD to SuMD to justify its application on ligand binding using a challenging case study like vasopressin. The general validation of the parent method SuMD as a predictive tool for ligand binding mode has been extensively reported over the years (a few examples: https://doi.org/10.1021/ci400766b ; https://doi.org/10.1021/acs.jcim.5b00702 ; https://doi.org/10.1038/s41598-020-77700-z) and fell beyond the scope of this work. 

(5) In the discussion, the authors write "A complete characterization of the possible interfaces between GPCR monomers, which falls beyond the goal of the present work, should be achieved by preparing different initial unbound states characterized by divergent relative orientations between monomers to dynamically dock." It would be useful for the reader to refer to and cite here advanced computational approaches that allow a comprehensive sampling of GPCR dimerization independently from the starting conformation of the receptors. One example is coarse-grained metadynamics as shown in doi: 10.1038/s41467-023-42082-z;

The A<sub>2A</sub/D<sub>2</sub receptors dimerization has been removed from the manuscript. 

(6) In many cases, it is not reported how residues missing from the experimental structures used to model the proteins were reconstructed. This information is important, considering that the authors comment on the results of their calculations on addressing these regions, such as in the case of B2AR. Furthermore, the authors did not report how their initial models were validated. The authors should also explain why they did not model the IC loops of A2AR and D2R;

In the current version of the manuscript, for V2R ECL2 and GLP-1R, we specify that we produced 10 solutions with Modeller and considered the best one in terms of the DOPE score. 

The only receptor model used,  β₂ AR, is now presented as preliminary data focusing on Gs and avoiding any structural detail of the Gs recognition. 

As reported above the A2A-D2 dimerization has been removed from the manuscript.

(7) In several cases, the authors state that residues never investigated before play an important role in the interaction between different proteins. An example is provided on page 6 for the B2AR/G protein association. Since this claim is quite significant, it would benefit from validation, at least for further calculations such as in silico mutagenesis studies. Another example is at the end of page 10 where the authors report a hidden interaction between D344 and R385 that is pivotal for Gs coupling by GLP-1R. Is there other evidence supporting this result (previously reported literature data, conservation rate of these residues, etc.)?;

We have removed the supplementary table reporting B2AR/G protein interactions to reduce speculations and added a reference that reports GLP-1 EC50 reduction upon mutation of position 344 to Ala (https://doi.org/10.1021/acscentsci.3c00063).

(8) The authors should provide a deeper discussion about the conformational rearrangement of GPCR and G protein during the coupling. In detail, the conformational changes of microswitich amino acids of GPCR (e.g., PIF, NPxxY, inactivating ionic lock) and alpha helix 5 of G proteins should be discussed in relation to the literature data and experimental structures;

We have removed the A1R and b2 AR results to focus on GLP-1R. Key structural motifs in the polar central network and TM6 kink are analyzed more in detail in Figure 3.

(9) The chronology of the conformational changes of GLP-1R is arbitrarily chosen. During the simulation, the RMSD values reported in Fig. 3 are high and do not demonstrate the full accomplishment of the simulation of the activation process of the receptor;

We agree with the Reviewer that the GLP-1R inactive to active transition was not fully accomplished, compared to other work on class A GPCRs.  Unlike class A, class B GPCRs represent a challenging system to work with in silico because inactive starting conformations (e.g 6LN2) are extremely distant from the active one (e.g 7LCJ, 7LCI or 6X18), as demonstrated in Figure S6 for GLP-1R. Here we report the first attempt to model a class B GPCR activation mechanism starting from the inactive state, and even if not fully achieved, we believe it represents state-of-the-art simulations for this class of receptors.

(10) It would be helpful for the reader not familiar with the employed technique that the authors explain in one sentence in the main text the pros and cons of using multiple walkers instead of single walker SuMD;

We thank the Reviewer for the excellent suggestion. In the Discussion, we have now commented that: “more extensive sampling obtainable by seeding multiple parallel short simulations instead of a single simulation for batch”, while in the Methods we explain that “mwSuMD is designed to increase the sampling from a specific configuration by seeding user-decided parallel replicas (walkers) rather than one short simulation as per SuMD. Since one replica for each batch of walkers is always considered productive, mwSuMD gives more control than SuMD on the total wall-clock time used for a simulation. On the flip side, mwSuMD requires multiple GPUs to be the most effective, although any multi-threaded GPU can run more walkers on the same hardware keeping the sampling variety.”.

Minor points to address:

(11) Page 3: the following sentence is duplicated (also found on page 2) "GPCRs preferentially couple to very few G proteins out of 23 possible counterparts";

(12) Page 20: Figure S13 refers to the QM validation of PF06882961 torsional angle, not to the image of the receptor conformational changes, which is instead Figure S14 (please correct figure caption).

We thank the Reviewer for the accurate reading of the manuscript. These typos have been corrected.

eLife Assessment

This study describes a novel mechanism for how collagen fibrils are formed. The authors present compelling evidence that collagen-I fibrillogenesis relies on a functional endocytic system for recycling collagen-I, with circadian-regulated VPS33b and integrin-α11 being critical for fibril assembly. This is an important study for the understanding of the pathophysiology of collagen fibrillogenesis.

Reviewer #1 (Public review):

Summary:

The authors describe that the endocytic pathway is crucial for ColI fibrillogenesis. ColI is endocytosed by fibroblasts, prior to exocytosis and formation of fibrils, which can include a mixture of endogenous/nascent ColI chains and exogenous ColI. ColI uptake and fibrillogenesis are regulated by circadian rhythm as described by the authors in 2020, thanks to the dependence of this pathway on circadian-clock-regulated protein VPS33B. Cells are capable of forming fibrils with recently endocytosed ColI along when nascent chains are not available. Previously identified VPS33B is demonstrated not to have a role in endocytosis of ColI, but to play a role in fibril formation, which the authors demonstrate by showing the loss of fibril formation in VPS33B KO, and an excess of insoluble fibrils - along-side a decrease in soluble ColI secretion - in VPS33B overexpression conditions. A VPS33B binding protein VIPAS39 is also shown to be required for fibrillogenesis and to colocalise with ColI. The authors thus conclude that ColI is internalised into endosomal structures within the cell, and that ColI, VPS33B and VIPA39 are co-trafficked to the site of fibrillogenesis, where along with ITGA11, which by mass spectrometric analysis is shown to be regulated by VPS33B levels, ColI fibrils are formed. Interestingly, in involved human skin sections from idiopathic pulmonary fibrosis (IPF) patients, ITGA11 and VPS33B expression is increased compared to healthy tissue, while in patient-derived fibroblasts, uptake of fluorescently-labelled ColI is also increased. This suggests that there may be a significant contribution of endocytosis-dependent fibrillogenesis in the formation of fibrotic and chronic wound-healing diseases in humans.

Strengths:

This is an interesting paper that contributes an exciting novel understanding of the formation of fibrotic disease, which despite its high occurrence, still has no robust therapeutic options. The precise mechanisms of fibrillogenesis are also not well understood, so a study devoted to this complex and key mechanism is well appreciated. The dependence of fibrillogenesis on VPS33B and VIPA39 is convincing and robust, while the distinction between soluble ColI secretion and insoluble fibrillar ColI is interesting and informative.

Weaknesses:

There are a number of limitations to this study in its current state. Inhibition of ColI uptake is performed using Dyngo4a, which although proposed as an inhibitor of Clathrin-dependent endocytosis is known to be quite un-specific. This may not be a problem however, as the endocytic mechanism for ColI also does not seem to be well defined in the literature, in fact, the principle mechanism described in the papers referred to by the authors is that of phagocytosis. It would be interesting to explore this important part of the mechanism further, especially in relation to the intracellular destination of ColI. The circadian regulation does not appear as robust as the authors last paper, however, there could be a larger lag between endocytosis of ColI and realisation of fibrils. The authors state that the endocytic pathway is the mechanism of trafficking and that they show ColI, VPS33B and VIPA39 are co-trafficked. However, the only link that is put forward to the endosomes is rather tenuously through VPS33B/VIPA39. There is no direct demonstration of ColI localisation to endosomes (ie. immunofluorescence), and this is overstated throughout the text. Demonstrating the intracellular trafficking and localisation of ColI, and its actual relationship to VPS33B and VIPA39, followed by ITGA11, would broaden the relevance of this paper significantly to incorporate the field of protein trafficking. Finally, the "self-formation" of ColI fibrils is discussed in relation to the literature and the concentration of fluorescently-tagged ColI, however as the key message of the paper is the fibrillogenesis from exocytosed colI, I do not feel like it is demonstrated to leave no doubt. Specific inhibition of intracellular trafficking steps, or following the progressive formation of ColI fibrils over time by immunofluorescence would demonstrate without any further doubt that ColI must be endocytosed first, to form fibrils as a secondary step, rather than externally-added ColI being incorporated directly to fibrils, independent of cellular uptake.

Reviewer #3 (Public review):

Summary:

Chang et al. investigated the mechanisms governing collagen fibrillogenesis, firstly demonstrating that cells within tail tendons are able to uptake exogenous collagen and use this to synthesize new collagen-1 fibrils. Using an endocytic inhibitor, the authors next showed that endocytosis was required for collagen fibrillogenesis and that this process occurs in a circadian rhythmic manner. Using knockdown and overexpression assays, it was then demonstrated that collagen fibril formation is controlled by vacuolar protein sorting 33b (VPS33b), and this VPS33b-dependent fibrillogenesis is mediated via Integrin alpha-11 (ITGA11). The authors also demonstrated increased expression of VPS33b and ITGA11 at the gene level in fibroblasts from patients with idiopathic pulmonary fibrosis (IPF), and greater expression of these proteins in both lung samples from IPF patients and in chronic skin wounds, indicating that endocytic recycling is disrupted in fibrotic diseases. Finally, the authors performed knockdown assays in patient derived IPF fibroblasts to confirm that silencing of VPS33b and ITGA11 results in a decrease in recycling of exogenous collagen-1

Strengths:

The authors have performed a comprehensive functional analysis of the regulators of endocytic recycling of collagen, providing compelling evidence that VPS33b and ITGA11 are crucial regulators of this process, and that this endocytic recycling becomes disrupted in fibrotic diseases.

Author response:

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

Overall authors’ response

We would like to thank the 3 reviewers for a thorough critique of our manuscript, and acknowledging the novelty and importance of our studies, in particular the relevance to collagenrelated pathologies such as idiopathic pulmonary fibrosis and chronic skin wound. We appreciate that there are shortcomings in these studies, as highlighted by reviewers; we have rewritten parts of our manuscript to clarify any misunderstandings, and conducted additional experiments to address concerns raised by reviewers (please see below red text within each response), which have been incorporated into our revised manuscript (modified text highlighted in yellow in revised manuscript). We believe that the revision had made our manuscript stronger in support of our original conclusions. 

Public Reviews: 

Reviewer #1 (Public Review): 

Summary: 

The authors describe that the endocytic pathway is crucial for ColI fibrillogenesis. ColI is endocytosed by fibroblasts, prior to exocytosis and formation of fibrils, which can include a mixture of endogenous/nascent ColI chains and exogenous ColI. ColI uptake and fibrillogenesis are regulated by circadian rhythm as described by the authors in 2020, thanks to the dependence of this pathway on circadian-clock-regulated protein VPS33B. Cells are capable of forming fibrils with recently endocytosed ColI when nascent chains are not available. Previously identified VPS33B is demonstrated not to have a role in endocytosis of ColI, but to play a role in fibril formation, which the authors demonstrate by showing the loss of fibril formation in VPS33B KO, and an excess of insoluble fibrils - along-side a decrease in soluble ColI secretion - in VPS33B overexpression conditions. A VPS33B binding protein VIPAS39 is also shown to be required for fibrillogenesis and to colocalise with ColI. The authors thus conclude that ColI is internalised into endosomal structures within the cell, and that ColI, VPS33B, and VIPA39 are co-trafficked to the site of fibrillogenesis, where along with ITGA11, which by mass spectrometric analysis is shown to be regulated by VPS33B levels, ColI fibrils are formed. Interestingly, in involved human skin sections from idiopathic pulmonary fibrosis (IPF) patients, ITGA11 and VPS33B expression is increased compared to healthy tissue, while in patient-derived fibroblasts, uptake of fluorescently-labelled ColI is also increased. This suggests that there may be a significant contribution of endocytosis-dependent fibrillogenesis in the formation of fibrotic and chronic wound-healing diseases in humans. 

Strengths: 

This is an interesting paper that contributes an exciting novel understanding of the formation of fibrotic disease, which despite its high occurrence, still has no robust therapeutic options. The precise mechanisms of fibrillogenesis are also not well understood, so a study devoted to this complex and key mechanism is well appreciated. The dependence of fibrillogenesis on VPS33B and VIPA39 is convincing and robust, while the distinction between soluble ColI secretion and insoluble fibrillar ColI is interesting and informative. 

Weaknesses: 

There are a number of limitations to this study in its current state. Inhibition of ColI uptake is performed using Dyngo4a, which although proposed as an inhibitor of Clathrin-dependent endocytosis is known to be quite un-specific. This may not be a problem however, as the endocytic mechanism for ColI also does not seem to be well defined in the literature, in fact, the principle mechanism described in the papers referred to by the authors is that of phagocytosis.

We thank the reviewer for pointing this out. Macropinocytosis or phagocytosis could be modelled using high molecular weight dextran, and we have used fluorescently-labelled dextran to investigate potential co-localisation with exogenous collagen to investigate the involvement of these mechanisms in addition to endocytosis, and showed very little co-localisation (revised Figure S2B, lines 123-126). Further, we have performed a competition experiment where unlabelled collagen was added in excess at the same time as labelled collagen and showed that excess unlabelled collagen led to a retention of labelled collagen at the cell periphery (revised Figure S2C, lines 126-129). This is suggestive of collagen-I uptake utilises a different pathway to dextran (i.e. fluid-phase endocytosis) and is a receptor-mediated process.  

It would be interesting to explore this important part of the mechanism further, especially in relation to the intracellular destination of ColI.

We agree with the reviewer that the intracellular destination of ColI is very interesting, which is what the current Chang lab is investigating, although we believe the research findings fall out of scope for the revised manuscript here. However, we have included additional immunofluorescence data to support that collagen is indeed taken up into endosomal compartments using GFP-tagged Rab5 constructs (revised Figure 1D, Figure S6A).

The circadian regulation does not appear as robust as the authors' last paper, however, there could be a larger lag between endocytosis of ColI and realisation of fibrils.

The authors state that the endocytic pathway is the mechanism of trafficking and that they show ColI, VPS33B, and VIPA39 are co-trafficked. However, the only link that is put forward to the endosomes is rather tenuously through VPS33B/VIPA39.

We would like to clarify that we meant the post-Golgi compartment. We did not mean VPS33b/VIPAS39 as an endosome marker; however as we see collagen entering the cell in intracellular compartments, which is then recycled, we take that as convention, the endosome would be involved. This is further supported that we see some colocalisation with the classic Rab5 endosome marker.

There is no direct demonstration of ColI localisation to endosomes (ie. immunofluorescence), and this is overstated throughout the text.

We appreciate the comment and have modified overstatements in the revised manuscript as appropriate. As stated above, we have included additional immunofluorescence data to support that collagen is indeed taken up into endosomal compartments.

Demonstrating the intracellular trafficking and localisation of ColI, and its actual relationship to VPS33B and VIPA39, followed by ITGA11, would broaden the relevance of this paper significantly to incorporate the field of protein trafficking. Finally, the "self-formation" of ColI fibrils is discussed in relation to the literature and the concentration of fluorescently-tagged ColI, however as the key message of the paper is the fibrillogenesis from exocytosed colI, I do not feel like it is demonstrated to leave no doubt. Specific inhibition of intracellular trafficking steps, or following the progressive formation of ColI fibrils over time by immunofluorescence would demonstrate without any further doubt that ColI must be endocytosed first, to form fibrils as a secondary step, rather than externally-added ColI being incorporated directly to fibrils, independent of cellular uptake.

We appreciate the concern raised here. This is precisely why we trypsinised and replated cells as part of the workflow, so we can make sure that there is no residual exogenous collagen which is not endocytosed being incorporated onto pre-existing fibrils. We have new data using flow imaging, which showed that cells that don’t endocytose exogenous collagen has accumulation of said collagen at the periphery of the cells, which is greatly reduced after trypsinisation. This new data is in a more detailed methodology-based study which is under preparation, which will allow future studies to further dissect the collagen intracellular trafficking process, and thus is not included in the revised manuscript. 

Reviewer #2 (Public Review): 

Summary: 

In this manuscript, the authors describe a mechanism, by which fluorescently-labelled Collagen type

I is taken up by cells via endocytosis and then incorporated into newly synthesized fibers via an ITGA11 and VPS33B-dependent mechanism. The authors claim the existence of this collagen recycling mechanism and link it to fibrotic diseases such as IPF and chronic wounds. 

Strengths: 

he manuscript is well-written, and experimentally contains a broad variation of assays to support their conclusions. Also, the authors added data of IPF patient-derived fibroblasts, patient-derived lung samples, and patient-derived samples of chronic wounds that highlight a potential in vivo disease correlation of their findings. 

The authors were also analyzing the membrane topology of VPS33B and could unravel a likely 'hairpin' like conformation in the ER membrane. 

Weaknesses: 

Experimental evidence is missing that supports the non-degradative endocytosis of the labeled collagen.

We thank the reviewer for raising this. We would like to clarify that we do not think that all endocytosed collagen-I is recycled, but rather sorted in the endosome which determines the fate of endocytosed collagen. Interestingly, results from Kadler’s group has shown that blocking lysosome function (through chloroqine and bafilomycin) significantly reduced endogenous collagen fibril formation (https://www.biorxiv.org/content/10.1101/2024.05.09.593302v1), suggesting a nondegradative role for lysosome in fibrillogenesis.   

The authors show and mention in the text that the endocytosis inhibitor Dyngo®4a shows an effect on collagen secretion. It is not clear to me how specific this readout is if the inhibitor affects more than endocytosis. This issue was unfortunately not further discussed.

We thank the reviewer for this comment and have included in discussion the specificity of Dyngo4a (revised manuscript lines 383392). The ponceau stain suggests that Dyngo4a treatment did not affect global secretion and thus the effects are specific to collagen-I (Fig 2B).

The authors use commercial rat tail collagen, it is unclear to me which state the collagen is in when it's endocytosed. Is it fully assembled as collagen fiber or are those single heterotrimers or homotrimers?

We apologise for the confusion and will clarify in our revision. These would be single helical trimers from acid-extracted rat tail collagen. We have performed additional light scattering and CD spectra to confirm the molecular weight and helicity, and confirm that adding fluorescent tags did not alter the readout. We have included this in the revised manuscript (revised Figure S1A-C, manuscript lines 82-86).    

The Cy-labeled collagen is clearly incorporated into new fibers, but I'm not sure whether the collagen is needed to be endocytosed to be incorporated into the fibers or if that is happening in the extracellular space mediated by the cells.

We appreciate the concern raised here, which is also raised by reviewer 1. As answered above, this is why we trypsinised and replated cells as part of the workflow, so we can make sure that there is no residual exogenous collagen being incorporated onto pre-existing fibrils. We also have new data using flow imaging, which shows that cells that don’t endocytose exogenous collagen has accumulation of said collagen at the periphery of the cells, which is greatly reduced after trypsinisation. This new data is in a methodology-based manuscript which is under preparation, thus will not be included in the revised manuscript.  

In general for the collagen blots, due to the lack of molecular weight markers, what chain/form of collagen type I are you showing here?

Apologies for the lack of molecular weight markers, it was an oversight by the authors and have been included in the revised figures.  

Besides the VPS33B siRNA transfected cells the authors also use CRISPR/Cas9-generated KO. The KO cells do not seem to be a clean system, as there is still a lot of mRNA produced. Were the clones sequenced to verify the KO on a genomic level?

Yes, the clones were verified and used in our previous paper on circadian control of collagen homeostasis. There are instances where despite knockout at the protein level, mRNA is still persistent; however these transcripts are likely then directed to degradation through nonsense-mediated mRNA decay. To fully understand this mechanism is beyond the scope of this paper. 

For the siRNA transfection, a control blot for efficiency would be great to estimate the effect size. To me it is not clear where the endocytosed collagen and VPS33B eventually meet in the cells and whether they interact. Or is ITGA11 required to mediate this process, in case VPS33B is not reaching the lumen?

This is an interesting question. We have conducted experiments with Col1-GFP11 containing conditioned media incubated with VPS33b-barrell in the revised paper, which showed that they interact within the cell and not at the cell periphery (revised Figure 6G, lines 293-296), again highlighting that VPS33b is not involved in the endocytosis step but interacts with endocytosed collagen-I intracellularly. We have attempted colocliasation studies using the split GFP approach with VPS33B and ITGA11 to investigate where they interact, but as the ITGA11 construct we used did not localise to the cell surface as expected, we are not confident that this system is appropriate for investigating how/if VPS33B interacts with ITGA11, and there are simply no good antibody for VPS33B for staining. 

The authors show an upregulation of ITGA11 and VPS33B in IPF patients-derived fibroblasts, which can be correlated to an increased level of ColI uptake, however, it is not clear whether this increased uptake in those cells is due to the elevated levels of VPS33B and/or ITGA11.

We would like to clarify here that we do not think collagen-I uptake is due to VPS33B and/or ITGA11, as siITGA11 and VPS33B in fibroblasts showed no consistent changes in uptake as determined by flow cytometry, which was included in the original manuscript (now revised Figure 6H, 7I). VPS33B and ITGA11 are involved in the ‘outward’ arm of recycled collagen-I, i.e. directing to fibrillogenesis route. We agree that the inclusion of additional functional studies using IPF patient-derived patient fibroblasts would add to the manuscript, and have performed siRNA against VPS33B and ITGA11 on IPF fibroblasts, and demonstrated a late of endocytic recycling events (revised Figure 8D, S6B, lines 351-353).  

Reviewer #3 (Public Review): 

Summary: 

Chang et al. investigated the mechanisms governing collagen fibrillogenesis, firstly demonstrating that cells within tail tendons are able to uptake exogenous collagen and use this to synthesize new collagen-1 fibrils. Using an endocytic inhibitor, the authors next showed that endocytosis was required for collagen fibrillogenesis and that this process occurs in a circadian rhythmic manner. Using knockdown and overexpression assays, it was then demonstrated that collagen fibril formation is controlled by vacuolar protein sorting 33b (VPS33b), and this VPS33b-dependent fibrillogenesis is mediated via Integrin alpha-11 (ITGA11). Finally, the authors demonstrated increased expression of VPS33b and ITGA11 at the gene level in fibroblasts from patients with idiopathic pulmonary fibrosis (IPF), and greater expression of these proteins in both lung samples from IPF patients and in chronic skin wounds, indicating that endocytic recycling is disrupted in fibrotic diseases. 

Strengths: 

The authors have performed a comprehensive functional analysis of the regulators of endocytic recycling of collagen, providing compelling evidence that VPS33b and ITGA11 are crucial regulators of this process. 

Weaknesses: 

Throughout the study, several different cell types have been used (immortalised tail tendon fibroblasts, NIHT3T cells, and HEK293T cells). In general, it is not clear which cells have been used for a particular experiment, and the rationale for using these different cell types is not explained. In addition, some experimental details are missing from the methods.

We thank the reviewer for pointing out the lack of clarity, and have filled in missing information in the methods. HEK293T cells were used for virus production for the VPSoe system, and we have clarified the cell types used in figure legends (predominantly iTTF). We have also provided justification when NIH3T3 cells were used (revised lines 290-291).    

There is also a lack of functional studies in patient-derived IPF fibroblasts which means the link between endocytic recycling of collagen and the role of VPS33b and ITGA11 cannot be fully established.

We thank the reviewer for this comment, which was also raised by reviewer 2 above. We agree that the inclusion of additional functional studies using IPF patient-derived patient fibroblasts would add to the manuscript and have performed siRNA against VPS33B and ITGA11 on IPF fibroblasts, and demonstrated a late of endocytic recycling events (revised Figure 8D, S6B, lines 351-353).  

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors): 

The authors inhibit Clathrin-dependent endocytosis with dyngo4a. It is well known that this inhibitor is not highly specific for this pathway. It is also not explained why the authors only inhibit the Clathrin uptake pathway, and not pinocytosis or Clathrin-independent endocytosis too. The authors refer to papers that describe pinocytosis for collagen endocytosis.

We thank the reviewer for raising this question. Based on the fact that inhibition of clathrin-dependent pathway does not completely abrogate endocytosis of collagen-I, we anticipate that other pathways are involved in mediating collagen-I uptake, although additional data suggested this is unlikely through fluid-phase endocytosis, and is receptor mediated (revised Figure S2B, C).  

Where does the ColI go in the cell? Depending on the uptake pathway, it is likely to pass through endocytic carriers to endosomes, where it may be recycled to the PM or degraded. From the start, the authors describe the ColI as being in vesicular structures, however, the imaging data that this is based on is not co-labelled with anything to determine the potential structure/localisation. This is not done at any point in the paper, until IF is shown of ColI with VIPA39, however without the relevant controls, this IF is unconvincing, as the general pattern of ColI and VIPA39 as an endosomal marker are not classically recognisable. Additionally, VPS33B is described as a late endosome/lysosome marker, which would have different connotations on ColI trafficking or destination than other types of endosomes.

We thank the reviewer for pointing out the weaknesses in our original IF. We have included new confocal images showing labelled collagen co-localisation with GFP-tagged Rab5 through transient transfection, which is a more traditional endosome marker (revised Figure 1D, Figure S6A).  

We are currently characterising the compartments to where ColI is trafficked to, which is being prepared as part of a methodology-based manuscript. We believe that this characterisation would be too detailed to be included in a revised version of this manuscript. The Kadler lab also have data suggesting that the lysosome is involved in collagen fibrillogenesis instead of its canonical degradation function, which is in another submitted manuscript (https://www.researchsquare.com/article/rs-1336021/v1). It was not included in this manuscript due to our focus (i.e. endocytic-recycling).   

In Figure 5H, the pattern of Cy5-ColI staining looks like it could even be ER/Golgi in the VPSKO zoom panel, but in the absence of co-labelling, we cannot conclude anything. In order for the authors to conclude that ColI is within the endosomes, co-labelled If should be performed to demonstrate ColIendosomal colocalization. Likewise for the role of VPS33B in ColI fibrillogenesis: dependence of the process is demonstrated, but the relationship is not defined. This could be clarified using IF. This would also support the authors' statements of co-trafficking between ColI, VPS33B, and VIPA39, which as the paper stands, is not demonstrated.

We would like to clarify that our hypothesis is that the endosome controls how collagen is being deposited outside the cell, i.e. whether it’s protomeric secretion or fibrillogenesis, and that the decision of whether an endocytosed collagen is recycled or degraded lies in this compartment. The reviewer is correct that it may not be just the endosome that endocytosed collagen-I ends up in, as we have new data suggesting involvement of other intracellular compartment, although the detailed mechanism is beyond the scope of this manuscript. Nonetheless, we have included new data showing co-localisation of endocytosed collagen with Rab5 in this revised manuscript (revised Figure 1D, Figure S6A).  

The basis of this paper is that endocytosis of ColI must occur before re-exocytosis as fibrillar ColI. The authors show this through pulse-chase experiments, with a trypsinisation step to remove any externally bound ColI. The authors also show nice time progression by flow cytometry, but it would truly demonstrate this point if they showed 0 timepoint, or low timepoint of IF to show progressive lengthening of ColI fibrils. This is used early on in Figure 1D, although the presentation here is not very clear. This is especially important as the authors address the self-seeding capabilities of Collagen in cell-free conditions in Figure 1F.

We would like to thank the reviewer for this suggestion.  From previous endogenously tagged collagen data, we know that the appearance of collagen fibrils is rather rapid, thus it may not be a gradual lengthening as expected, but rather a depletion of endocytosed collagen in the initial seeding/growth step (please see https://www.researchsquare.com/article/rs-1336021/v1). We have included an image of replated fibroblasts after 18 hours showing no appearance of extracellular collagen, endogenous or otherwise (revised Figures S2A, line 110).  

Finally, although the involvement of ITGA11 is interesting, it is not well described, and its role is not well demonstrated. This could likely be clarified by an additional introduction to ITGA11 and its role in collagen exocytosis/fibrillogenesis.

We would like to thank the reviewer for pointing this out and have included additional sentences to specifically introduce ITGA11 and its role in fibrillogenesis (see lines 320, 321; 446-450).  

Specific points: 

Line 73: You haven't compared reuse vs production, so you can't say that reuse is central rather than production. They may be both as important or production still may be the most crucial, maybe it depends on cell/collagen type. Using the ColI KD or CHX to block nascent synthesis, you could directly compare the impact of both.

We would like to clarify that we are not referring to reuse/recycling here. We meant that production of collagen (i.e. single hetero/homotrimer molecules within the cell) is not as crucial as the utilisation (i.e. are these being secreted as protomers, or assembled into fibrils) of these building blocks by the cells, which was supported by our finding that production (as suggested by mRNA levels) of IPF fibroblasts are similar to that in control fibroblasts (now revised Figure 8A). We have conducted ColI siRNA to block nascent synthesis in the original manuscript and showed that fibroblasts can efficiently make new fibrils by recycling exogenous collagen (Figure 3B, C), although we appreciate that siRNA may not completely inhibit endogenous production. Thus, we have also included new data using collagen-I knockout cells to support our hypothesis that without endogenous production, fibroblasts can still effectively make collagen fibrils if they can reuse what is available in the extracellular space (revised Figure 4, Figure S3C, D; lines 178-199).  

Lines 83-87: The rationale for this experiment is not clear. Cy3-ColI is added, taken up into cells, and incorporated into fibrils coming from cells. 5FAM-ColI is added at a later stage, then at 2 days (when incorporation is demonstrated in Fig 1B), it is also incorporated into cells as expected. Why does this comment on ColI not being degraded any more than Cy3-ColI alone?

We believe that the pulse chase experiment using the differently tagged collagen demonstrated a dimension of dynamics that is not demonstrated with Cy3-ColI alone. In this case, Cy3-ColI was initially added, and removed after 3 days; 5FAM-ColI is then added and incubated for 2 more days. Thus after 5 days since the initial pulse, the Cy3-ColI persisted and was not degraded. We would like to apologise for causing this confusion, and have clarified in the revised manuscript (lines 542-549; Figure S1D figure legend).  

Figure 1A: I would like to see a negative control: either dark colI or no Cy3-Col, or timescale. Is B quantified from these images?

We thank the reviewer for this comment. We have added the nocollagen control image in our revision (revised Figure S1D). 1B is not quantified from the ex vivo tendon experiments, but rather the in vitro cell culture experiments (i.e. those from 1D-1F, although they are all from independent experiments).  

Figure 1B: in iTTF cells (immortalised tendon cells) Corrected to max: What does that mean?

As there are variations between individual experiments (e.g. changes in the amount of collagen added due to pipetting) we have normalised to the maximum value obtained in each individual experiments so that we can display all biological repeats within the same graph.  

Figure 1C: You can't say ColI is in vesicular structures from this, they are spots, yes, but that could also be in Golgi/ER (unlikely to be cytosolic but not impossible).

We appreciate this comment and have change the wording accordingly and call them intracellular/punctate structures.

Figure 1D: Not the best presentation: The cell mask has structures: what are these? It's not clear if this is a single cell, would be better with a defined marker (endocytic marker, lysosome etc). Instead of a low-resolution 3D view, it would be clearer with normal confocal XY and zooms of "vesicular structures" using appropriate markers as 3D reconstructions I think it could be removed.

This is a single cell and the cell mask is staining plasma membrane. We didn’t use defined marker as we wanted to visualise the whole intracellular cell compartment. We appreciate that further proof is needed to verify the location of the endocytosed collagen, and have included additional confocal imaging data to support the localisation of collagen into Rab5 positive intracellular compartments (revised Figure 1D, Figure S6B).  

Figure 1 E/F: Cy3 is only visible in extracellular structure, not also intracellular. Why? Would be useful to see the time points of incorporation at the end of the pulse, then at an early point into the chase, to demonstrate 1) Cy3-ColI uptake into cells and progressive incorporation rather than potential direct binding of ColI-Cy3 to ECM, or other non-specific factors. Showing the image at 0t would demonstrate an absence of external labelled colI and therefore its appearance later could be presumed that it had been internalised before.

As the cells were trypsinized and replated after one hour labelled collagen feeding to ensure we are only tracking endocytosed collagen, t=0 in this case would be cells that are unattached. We have included t=18hr images post replate instead to show baseline level of collagen (revised Figures S2A, line 110).

Figure S1A: yellow box: doesn't show only Cy3-ColI, there is red and yellow in the central cell, and large yellow blobs in the cell above. These images do not support this claim, including the Fiber Zoom box. They should also be shown in single channels to demonstrate the authors' points better.

Apologies for the confusion – this is to show that newly added FAM5 Collagen is also co-localising with previously endocytosed Cy3-ColI, i.e. the Cy3-ColI is persisting rather than being degraded.  

Line 92: endocytosed into distinct structures: These images are very vague, but I don't think you can call them distinct structures, all you can say from this is that they are spots.

We have changed the wording to ‘distinct puncta’.  

It is not clear why the authors use Cy3, Cy5, and 5FAM labelled colI. A brief explanation would be useful.

Apologies for the confusion, we initially included our justification (to show that the fluorescence labels do not change the way collagen is internalised) but removed it in the final manuscript due to length. We have added the justification (revised line 101-102).   

Figure 1F: It would be useful to see a quantification of the Cy3 channel here: I agree with the conclusions, and find the 0.5 ug/ml condition more convincing than 0.1 actually, although there is some feint Cy3 in cell-free samples there seems to be quite a big increase in the presence of cells, and this would look more convincing if quantified.

We thank the reviewer for this suggestion and have included quantification in the revised manuscript (revised Figure 1G-I).  

Figure 2B: Dyng is not an abbreviation of Dyng. Standardise Dyng/Dyngo/Dyngo4a. WB is soluble colI and represents little (if any) insoluble col. IF is more or less the other way round. How do they compare this?

Thank you for pointing out the inconsistencies, we have corrected this in the revised manuscript. We took the conditioned media from the same experiment where cells are fixed for IF and carried out Western blot analyses. The IF showed some collagen still present, albeit significantly reduced. This is in agreement with the western blot results (i.e. Dyng4a inhibits both soluble and insoluble forms of collagen deposition).  

Figure 2C: not an image series. Quant: no cells/independent exps and STATS?

Apologies for the missing experimental details in figure legends, it should say ‘representative of N=3 experiments’. We are not sure what the reviewer meant by Figure 2C not being an image series, as we meant it to be an image series of the individual fluorescence channels. We have changed this terminology to avoid confusion, and have included statistical analyses in the methods section. The statistical analyses of the fibril quantification is next to the fluorescence images.  

Figures 2D/E: The authors show that internalised ColI peaks at 20h and decreases to 60h, Fibers peak at 40h. How is this measured? ECM removed? Why would there be less in the cells, degradation? Whats the synchronisation?

We apologise for omitting the synchronisation method in methods section, and have included in our revised manuscript (revised lines 542-544). This is through dexamethasone addition (and removal after 1hr incubation) as standard. The internalised Col-I is measured using Cy3ColI so the cells would have both nascent and external collagen. Total intracellular collagen at the different time points would likely be higher than represented as a result, but here we are demonstrating that internalisation is a rhythmic event using the external labelled collagen. Fibers are measured using standard IF and then fibril counting.  

Please note that we are only overlaying the two graphs to form our hypothesis that endocytosis may be used for accumulation of collagen protomers that then allows for efficient fibrillogenesis. They are not directly comparable as the quantification are of different things (internalised Cy3-ColI, total collagen fibrils). We have clarified this in our discussion (revised lines 399-401).  

Discussion: Where does the ColI go? Solubilised? Degraded? Taken up by other cells? 

The inverse correlation is not very tight. In fact, at 38h where fiber count peaks, Cy3-ColI also peaks (esp in normalised data, Figure S2D).

We thank the reviewer for this comment and have reworded our main text to reflect this, and included additional discussion in our revised manuscript (revised lines 401-404).  

Line 123: What is the turnover rate of Fibrils? Don't know for how long the transcription has been done, or when this would affect the fibril number. You have the quant for Fn1, where is the quant for ColI?

We have included the quantification of collagen-I in original Figure 2A. We appreciate that it might cause confusion in Figure 2C (as we co-stained ColI and Fn1 in the same experiment) we have removed the collagen-I panel from the revised Figure 2C. We know from previous results that the number of fibrils fluctuate over 24hour period, although the turnover of one specific fibril is unlikely going to be 24 hours (https://www.biorxiv.org/content/10.1101/331496v2)

Line 124: no accumulation of col in extracellular space, but you don't know how much endogenous colI (or other endogenous ECM proteins) they're taking up as it isn't measured here. If the author wants to comment on this, should use either exogenous col to monitor take up and resection or block transcription/translation to show fibril formation endo/exocytosis independent of endogenous synthesis.

This experiment has been done in the original manuscript – siCol1a1 experiment was done with two rounds of siRNA, first round is normal transfection followed by reverse transfection onto fresh coverslips (this will ensure no prior ECM is being deposited, see Figure 3). However we appreciate that there may still be low levels of endogenous collagen-I, and thus have included new data using collagen-I knock-out fibroblasts to strengthen our findings (revised Figure 4).  

Line 142: Why is fibronectin synthesis also decreased in Col KD? This is clear in the image but no explanation/reference is given.

Due to the dynamic and complex nature of ECM, it is unsurprising if there is a knockon effect when knocking down one matrix protein. However, we have quantified the amount of fibronectin fibril deposited by scr and siCol1a1 fibroblasts, and showed that there was in fact no significant change between the two treatments (revised Figure 3A).

Figure 3A: Need labels for which colour/protein is shown. Needs quantifying, especially as the Fn1 decrease is not so obvious here, it is consistent between Figure 3A and 2C?

We have provided quantification in the revision (revised Figure 3A). Figure 3A and 2C are two separate experiments (one is Dyngo treatment and one is siCol1a1), and neither showed significant changes in fibronectin fibril areas.   

Figure 3B: Line 151: the text states that "The observation of fibrillar Cy3 signals in siCol1a1 cells showed that the cells can repurpose collagen into fibrils without the requirement for intrinsic collagen-I production (red arrow Figure 3B), however, there is clearly endogenous colI here too (along the fiber and also strongly at each end). Does the ColI antibody recognise the exogenous ColI?

In our hands the ColI antibody does not recognise exogenous ColI, as the cell-free Cy3-ColI images were also stained with ColI antibody to ensure the two experimental conditions were treated exactly the same.

This conclusion could only be made in the true absence of collagen: either in knock-out cells, or where collagen production/trafficking has been blocked (ie knockout of ColI chaperone or ERES block), or in a cell type that produces collagens but not ColI. Alternatively, if there are any fibrils seen that are completely negative, they should be shown in the figure and quantified (number of Cy3-ColI+-ColI+ vs Cy3-ColI+-ColI-).

We thank the reviewer for this suggestion. We have included new data from collagen knock-out fibroblasts in this revision (revised Figure 4).  

Figure S4A: the quality of this blot isn't very high, the result is not very clear and the high intensity (unspecific?) band below confounds the interpretation. In the author's previous paper (NCB 2020) the blots for VPS33B were much clearer, as is Fig S4D. It would be nice to include a clearer blot, maybe from the other repeats.

This is the only blot that we used to select which knockout clones to use for our previous paper, which is why the quality is not as high. Knockout clones were all verified with additional western blots, and we do not think that endogenous VPS33b is expressed at high levels (also verified by MS analyses).  Fig S4D is overexpression of VPS33b, which is much easier to detect.  

Figure S4D: This blot is much clearer, it would be useful to include a high gain to show the VPS33B band in CT to be able to understand the true increase.

From the qPCR data one can see that the increase at mRNA is 20+ fold increase; we’ve always had problems trying to detect endogenous VPS33b using western blot or mass spectrometry analysis.  

Figure 4A: The fibrils here in the CT are not obvious, and the difference between CT and KOs is not appreciable. Would this be clearer shown at a lower magnification, with zooms where needed? Or immunogold labelling/CLEM to label the ColI?

It is not trivial to carry out immunogold labelling/CLEM. These are cell-derived matrices in culture and thus lower magnification may not show as many collagen fibrils as one would expect. We are not confident that lower magnification will provide more information as the characteristic D-banded collagen pattern will be lost.  

Line 167/Figure 4B: It looks like there is more internal ColI in KO, but the images are not good enough to tell. This could be better shown by flow cytometry.

We have previously seen that VPSKO leads to accumulation of collagen-I in intracellular punctas (NCB2020) which is also seen here. Flow cytometry data for internalisation of external collagen is already included in original Figure 5G (revised Figure 6H).  

Again you mention intercellular vesicles, but based on these images, it is not possible to conclude this. These large spots could be aggregation elsewhere in the cell. Specific localisation should be shown by co-labelled IF/confocal, or it could be nicely shown by EM + fluorescent element (CLEM / Immunogold), or these statements removed from the text.

We appreciate that the term ‘vesicles’ is very defined in the trafficking field, and have changed it to ‘intracellular compartments’.  

Line 173-174 / Figure 4E: Why do you think the matrix mass is not increased in VPSoe by the approach shown in E when there is seemingly a huge increase by IF? E must also measure other ECM matrix proteins, which do you expect to be secreted by these cells? Could this confound the data if they too are affected by VPSoe?

IF is showing specifically collagen-I. Hydroxyproline detects multiple collagens, and shows a trend of increase (although not significant due to one outlier). Matrix mass is a very generic measurement of total ECM deposited based on decellularized ECM weight. The reviewer is correct that VPSoe may also affect other ECM deposition, however here we are focussing specifically with its effect on collagen-I. How VPSoe changes other types of ECM deposition would be something that could be addressed in future studies and is not within scope of this manuscript.   

Are the results in E paired?

Individual values between control and VPSoe in each separate experiments are paired.  

Figure 4F: Is quantification from IF shown in D? Specify which kind of microscopy it is based on.

Quantification is based on fibril counting using standard fluorescence microscopy, as used in our previous paper. D is independent of F, as F is specifically looking at synchronised circadian effects, and D (and elsewhere) we are looking at global collagen deposition effects, irrespective of what time of day the cells are in.  

Figure S5F: What do the yellow/red spots in the blots represent?

We apologise for the initial unclear description of what the yellow/magenta circles depict in relation to the phosphoimages of the radiolabelled cell free translation products displayed in Supplementary Figure 5, panels F, G and I. These circles indicate non-glycosylated (yellow) and N-glycosylated (magenta) species respectively, as is now clearly descried in the revised manuscript.

Figure 5 title: You can't conclude this from these images, need confocal and PM or cytosolic marker.

We have changed the title to ‘VPS33B co-trafficks with collagen-I”. There is no good commercial VPS33b antibody for immunofluorescence staining, which is why we used the split GFP approach in this paper, and the images were acquired using confocal imaging (Olympus SpinSR system).  

Figure 5E: The authors describe that ColI is in endosomes throughout most of the paper, and this is based on the involvement of VPS33B in the colI pathway. VPS33B is thought to be at the late endosome/lysosome. However, these images do not look like classic endosomes or lysosomes, or other normal organelle IF phenotypes. The fluorescent intensity looks saturated, and it is difficult to conclude anything from these images. It is unclear where in the cell the largest blob in the zoom would be localised and in which cell. I would suggest that this image is replaced and proper controls included (IgG controls and single channels) as well as using different markers for other potential intracellular structures.

We appreciate the reviewers comment with regards to the classification of VPS33b localisation in the endosome compartment. We did not mean to use VPS33b as an endosome marker, as the focus of our studies are the function of VPS33b in directing endogenous or exogenous collagen to fibrillogenesis. With live imaging we could see endocytosed collagen moving in intracellular compartments, and have conducted additional staining to show co-localisation with Rab5 (revised Figure 1), which we take to indicate, through convention, that it is occupying an endosome compartment. We have included single channel images in the revised manuscript (revised Figure 6E).

Line 255/ Figure 5G: no consistent change in uptake. Why are the results so varied in the KO and oe, here and in Fig 4C/E? N=4, what does that mean? 4 cells? 4 independent exps?

In all cases, “N” represents independent biological experiments in this manuscript. Thus “N=4” in this case is 4 independent biological experiments, with at least 10,000 cells analysed per experiment. 

We don’t know why there is a variation in response, however that is also why we concluded that it is unlikely that VPS33B is directly involved with collagen uptake. We have changed 5G (now revised Figure 5H) to a paired line graph for better representation.  

Figure 5H shows the uptake of Cy5ColI. At this resolution, VP2ko looks like the col is ER, in one of the cells in the zoom, it looks like it is at Golgi. I think that the uptake route of ColI needs to be better defined, as there is no way to tell here where the colI goes. ColI being recycled/degraded would be most likely. But this figure looks like that might not be the case. It is also not clear where the zooms come from, they should be indicated with dashed boxes in the lower mag image

We thank the reviewer for this comment, and agree that we need to define the uptake route of ColI. This is currently being assembled as a methodology manuscript, and how ColI is being recycled/degraded is one major research area of the Chang lab. 

We have added dashed boxes in the lower mag images to indicate where the zooms derived from, and we would also like to thank the reviewer for pointing this out as we realised we have accidentally cropped the image to a slightly different area for the VPSko image, and have now corrected this.  

Line 257: Based on this data, it could be trafficking through the cell as well as into the extracellular space.

We think that VPS33B is involved in trafficking collagen through the cell to plasma membrane but not secreted, as based on our split-GFP experiment we never observed extracellular GFP signal, which suggests VPS33b is not deposited extracellularly.

Line 259: "highlighting the role in recycling col to fibril formation sites" is an overstatement based on the data shown here, there is no data on colI trafficking or its regulation

We respectfully disagree that we have not shown data on col-I trafficking or regulation by VPS33b – split GFP highlighted cotrafficking to the plasma membrane, and we have shown a clear relationship between VPS33b and collagen-I fibril formation, with minimal changes to collagen-I mRNA levels. We acknowledge that we have not shown specifically the location of VPS33b at fibrillogenic sites and have modified this statement in revised manuscript (revised line 302).  

Line 262: "Having identified VPS33B as specifically driving collagen-I fibril formation" is also an overstatement.

We refer here the data that VPS33b is not controlling collagen-I secretion (as demonstrated by the CM westerns) and specifically fibrillogenesis. We have clarified this in the revised text (revised line 304).  

Line 286: It would be useful to have a brief intro to PLOD3.

We have included a brief intro to PLOD3 in the introduction, as well as the results highlighted by the reviewer, in our revised manuscript (revised line 54-58).  

Line 289/290: There could be other explanations for disruption to exo-endocytosis when disrupting col trafficking. Is VPS33B controlling exocytosis in general? Why should it be specific to col? Likewise with siITGA11 KD? Hypothesis for ITGA11 and fibrillogenesis?

The relationship between ITGA11 and collagen fibrillogenesis is currently in a manuscript by Donald Gullberg and Cedric Zeltz, under revision at Matrix Biology (see reference 63 in revised manuscript). We do not think that VPS33b is controlling exocytosis in general, which is supported by the minimal change in ponceau stain of the western blots in the manuscript. Previously it has been shown that VPS33B co-trafficks with PLOD3, a collagen-I modifier.  

Figure 6I: Why only quant Scr + siITGA11, not in VPSoe? It looks like there is still an increase in intracellular or fibril formation in VPSoe + siITGA11, which would be a key result to discuss.

We would like to clarify that 6I (now revised Figure 7I) is on the endocytosis of exogenous collagen-I, not quantification of Figure 6H.  

Line 307: Discuss fibrillogenic sites, what are they?

As we have not shown direct evidence of VPS33B delivering endocytosed collagen at the site of fibrillogenesis, we have decided to alter the text to avoid overstatement, as suggested from previous reviewers’ comments.  

Figure 8: What does pentachrome label?

Pentachrome staining allows for simultaneous staining of multiple species: collagen in red, sulphated mucopolysaccharides in violet, red blood cells in yellow, muscle in orange, nuclei in green.

Line 326: "In this study we have identified the endosome as a major protagonist in..." This is an overstatement and cant be drawn from this data.

We have modified this statement to “In this study we have identified an endocytic recycling mechanism for type I collagen fibrillogenesis that is under circadian regulation”

Line 330/331: "Collagen-I co-traffics with VPS33B in a VIPAS-containing endosomal compartment that directs collagen-I to sites of fibril assembly," This is also an overstatement that cannot be drawn from this data.

We have modified this statement to “Collagen-I co-traffics with VPS33B to the plasma membrane for fibrillogenesis”.  

Line 340: again, the demonstration of the involvement of the endocytic pathway is very limited.

We have provided new evidence in the revised manuscript that support the involvement of classical endosomal compartments.  

Line 366: You cant conclude this, you have not manipulated these proteins to show a functional effect or modulation of fibrillogenesis, it could still be a secondary effect.

We have provided new evidence in the revised manuscript that supports this conclusion. 

Line 569: "Unless otherwise stated, incubation and washes were done at room temperature." Which incubations? Specify if this is just post-fixation during the EM prep or during cell culture.

This is specific to the EM preparation and we have clarified in the revised manuscript (revised line 663).  

Small text alterations:

Overall we would like to thank the reviewer for highlighting these errors and mistakes in our manuscript, and have corrected them in our revised manuscript.  

Figure 1E: Fluoro image series? This is only one image.

We wrote this to mean single channel images, we have corrected the terminology.  

Line 111: Ref for Dyngo4a?

We have included this in the revised manuscript  

Line 121: introduction/abbreviation definition for Fn1? Instead it is on Line 140.

Thank you for highlighting this, we have corrected this in revised manuscript.  

Figure S2C: Alignment of labels cleaves x-axis.

We thank the reviewer for catching this and have corrected this with our revised manuscript.  

Figure S4F and G should be inverted to mention sequentially in the text.

We thank the reviewer for catching this and have corrected this in our revised manuscript.  

Line 182: Figure 4J should be G.

We thank the reviewer for catching this and have corrected this in our revised manuscript.

Line 209: typo: N-glycosylated.

We have corrected this typo in our revised manuscript.

Fig 6E: Very big as a figure element compared to others.

We have made this smaller in the revised manuscript to fit better with rest of the figure.  

Line 313: Figure 7E not F.

Thank you for spotting this, we have corrected it.  

Line 555: Typo: Scraped.

We have corrected this typo in our revised manuscript.

Line 562: missing )

We have corrected this typo in our revised manuscript.

Standardise

We thank the reviewer for spotting the mistakes below and have corrected in our revised manuscript.  

Legends: Include numbers of repeats and STATs throughout. 

Terminology: Dyng etc. 

Scale bars: some included as editable lines, some with size on top, small/large etc.

In certain cases we have positioned the scale bars in different regions of the figures to ensure no obscuring of the images.

VPS33b v B. 

Reviewer #2 (Recommendations For The Authors):  

The authors can improve the experimental part of the manuscript the following: 

-  For all the western blots please include molecular weight markers.

We thank the reviewer for noticing this omission and have included molecular weight markers in the revised manuscript.  

- Performing immunofluorescence and western blot analysis of endocytosed collagen -/+ inhibitors for lysosomal degradation (BafA1 or E64d+PepstatinA) in order to exclude endocytosis for degradation.

We thank the reviewer for this comment, another paper from the lab has identified lysosome to be involved in collagen fibrillogenesis (https://www.biorxiv.org/content/10.1101/2024.05.09.593302v1), thus  

- Figure out how Dyngo4a is affecting Col1 secretion in the first place? Does it interfere with the secretory pathway. Alternatively, use a different model to block endocytosis (e.g. siRNA Dynamin).

We thank the reviewer for raising this. The Dyngo CM blot for total ponceau stain (revised Figure 2B) showed minimal changes, which suggest that global secretion is not affected.  

- Further characterization of the VPS33B / collagen vesicles by immunofluorescence containing markers for early, late, and recycling endosomes. Block endocytic recycling by depletion of either Rabs or e.g. EHD1.

There are no good VPS33b antibody for staining. We have included images of GFP-tagged Rab5 co-localisation with labelled collagen-I (revised Figure 1D, Figure S6B).  

- Further clarify the status of the VPS33B knockouts e.g. by sequencing. also provide a readout of the siRNA KD, besides the mRNA levels, since there the difference is not striking.

The knockout cell lines were characterised previously in our 2020 paper, which is referred to in our revised manuscript. We have always had issues detecting endogenous VPS33b due to reagents limitations, which is why we resorted to mRNA as the key readout.  

- Doing siRNA knockdowns and endocytosis inhibition in the IPF fibroblasts to further strengthen the link between elevated expression of VPS33B/ ITGA11 and increased collagen uptake.

We thank the reviewer for suggesting these experiments. Due to limitations of the patient-derived fibroblasts (cell numbers and passage numbers) we had to prioritise experiments, and thus have performed siRNA against VPS33B and ITGA11 in the IPF fibroblasts. We showed that in both cases the amount of recycled labelled-collagen in collagen fibrils is significantly reduced (revised Figure 8D).  

Reviewer #3 (Recommendations For The Authors): 

Major points 

(1) Choice of cells: Please provide a rationale for why each cell line was used, and make sure that it is clear throughout the manuscript which cell line was used for each particular experiment. The HEK293T cell line is also missing from the reagent table.

We thank the reviewer for pointing out this omission, and have clarified in our revised manuscript which cell lines were used in each experiment. We used HEK293T to generate lentiviruses as described in the methods section.  

(2) Missing information from methods. Experimental details are missing from the methods in several places, making it difficult for someone to replicate an experiment. For example, no details are given in the methods describing the explant culture of murine tail tendons (described in results lines 78100), and there are no details on how the skin samples were obtained or stained. Further, no ethical approval details are provided for the use of human skin tissue.

We apologise for leaving the ethical approval details and skin sample collection out, this was an oversight and will be included in the revised manuscript. We have also included the method to how murine tail tendons were cultured ex vivo (revised lines 527-531, 546-553).  

(3) Functional studies in patient-derived cells. To fully establish the role of VPS33b and ITGA11 in fibrotic diseases, functional studies including the knockdown/overexpression of these genes could be performed to establish if the same response is seen as in non-diseased cells.

We agree that this will add much to the paper, and have performed siRNA against VPS33B and ITGA11 in the IPF fibroblasts. We showed that in both cases the amount of recycled labelled-collagen in collagen fibrils is significantly reduced (revised Figure 8D).

Minor Points

We thank the reviewer for pointing out these mistakes, and have corrected and included additional details in the revised manuscript.  

(1) Lines 51-52. Wording of this sentence is unclear, please rephrase. 

(2) Line 182. Should this be Fig 4G rather than J? 

(3) Line 209. Correct spelling of glycosylated. 

(4) Line 463. Incomplete brackets and details missing? 

(5) Line 590. Correct tense - was rather than are. 

(6) Line 593. Specify centrifugation speed. 

(7) Line 619. Nuclei rather than nucleus. 

(8) Ln 650. Statistical analysis - was normality tested? 

(9) Figure 1e - Difficult to read labels for coll/DAPI.

eLife Assessment

This valuable study discusses a hot topic in post-endoscopic retrograde cholangiopancreatography pancreatitis. The new score for predicting post-ERCP pancreatitis offers an idea about the risk of pancreatitis before the procedure. Although most scores depend on intraprocedural manoeuvres, such as the number of attempts to cannulate the papilla, this is a solid retrospective single-center study in one country. To be validated in the future, this score will need to be done in many countries and on large numbers of patients.

Joint Public Review:

Summary:

This work provides a new general tool for predicting post-ERCP pancreatitis before the procedure depending on pancreatic calcification, female sex, intraductal papillary mucinous neoplasm, a native papilla of Vater, or the use of pancreatic duct procedures. Even though it is difficult for the endoscopist to predict before the procedure which case might have post-ERCP pancreatitis, this new model score can help with the maneuver and when the patient is at high risk of pancreatitis, sometimes can be deadly), so experienced endoscopists can do the procedure from the start. This paper provides a model for stratifying patients before the ERCP procedure into low, moderate, and high risk for pancreatitis. To be validated, this score should be done in many countries and on large numbers of patients. Risk factors can also be identified and added to the score to increase rank.

Strengths:

(1) One of the severe complications of endoscopic retrograde cholangiopancreatography procedure is pancreatitis, so investigators try all the time to find a score that can predict which patients will probably have pancreatitis after the procedure. Most scores depend on the intraprocedural maneuver. Some studies discuss the preprocedural score that can predict pancreatitis before the procure. This study discusses a new preprocedural score for post-ERCP pancreatitis.

(2) Depending on this score that identifies low, moderate, and high-risk patients for post-pancreatitis, so from the start, experienced and well-trained endoscopists can do the procedure or can refer patients to tertiary hospitals or use interventional radiology or endoscopic retrograde cholangiopancreatography.

(3) The number of patients in this study is sufficient to analyze data correctly.

Weaknesses:

(1) It is a single-country, retrospective study.

(2) Many cases were excluded, so the score cannot be applied to those patients.

Comments on revised version:

Depending on old references cannot help us know the current situation. What if there are better more recent predictive tools? It would be better to test the validity of that score against, if present, a proven score to check its validity.

Author response:

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

Joint Public review:

Summary:

This work provides a new general tool for predicting post-ERCP pancreatitis before the procedure depending on pancreatic calcification, female sex, intraductal papillary mucinous neoplasm, a native papilla of Vater, or the use of pancreatic duct procedures. Even though it is difficult for the endoscopist to predict before the procedure which case might have post-ERCP pancreatitis, this new model score can help with the maneuver and when the patient is at high risk of pancreatitis, sometimes can be deadly), so experienced endoscopists can do the procedure from the start. This paper provides a model for stratifying patients before the ERCP procedure into low, moderate, and high risk for pancreatitis. To be validated, this score should be done in many countries and on large numbers of patients. Risk factors can also be identified and added to the score to increase rank.

Thank you for reviewing our manuscript. We hope that this score will be validated in other countries from now on.

Strengths

(1) One of the severe complications of endoscopic retrograde cholangiopancreatography procedure is pancreatitis, so investigators try all the time to find a score that can predict which patients will probably have pancreatitis after the procedure. Most scores depend on the intraprocedural maneuver. Some studies discuss the preprocedural score that can predict pancreatitis before the procure. This study discusses a new preprocedural score for post-ERCP pancreatitis.

Thank you for evaluating our manuscript and raising a strength of this manuscript.

(2) Depending on this score that identifies low, moderate, and high-risk patients for post-pancreatitis, so from the start, experienced and well-trained endoscopists can do the procedure or can refer patients to tertiary hospitals or use interventional radiology or endoscopic retrograde cholangiopancreatography.

Thank you for evaluating our manuscript and raising a strength of this manuscript.

(3) The number of patients in this study is sufficient to analyze data correctly.

Thank you for evaluating our manuscript and raising a strength of this manuscript.

Weaknesses:

(1) It is a single-country, retrospective study.

Thank you for this comment. It’s exactly as you said. This is a limitation (Lines 326-327).

(2) Many cases were excluded, so the score cannot be applied to those patients.

Thank you for this valuable comment. The predictive PEP score is not necessary for the excluded patients. The reasons were as follows. Biliary duct cannulation was not attempted in patients for whom it was difficult to identify the Vater papilla. The biliary tract was separated from the pancreas in patients with a past history of choledochojejunostomy, pancreatojejunostomy, or pancreatogastrostomy. PEP risk was thought to be low in these patients and patients who underwent bile duct cannulation via the choledochoduodenal fistula. PEP diagnosis is difficult in patients with acute pancreatitis, whose diagnosis is currently in progress. We added these explanations (Lines 98-106).

(3) Many other studies, e.g., https://link.springer.com/article/10.1007/s00464-021-08491-1, https://pubmed.ncbi.nlm.nih.gov/36344369/, that have been published before discussing the same issue, so what is the new with this score?

Thank you for raising the new reference written by Archibugi et al. in 2023. The novelty of our score is that it is calculated using the factors that are investigated before ERCP procedures. The study written by Archibugi et al. involved procedure time and cannulation attempts for PEP prediction. These two factors are unknown before ERCP procedures. Therefore, a preprocedural predictive risk model for PEP was not created before our study was performed. We added the content of the past study written by Archibugi and included the report as a reference (Lines 65-67, 73-74).

(4) The discussion section needs reformulation to express the study's aim and results.

Thank you for this valuable comment. I have rewritten the first paragraph of the discussion. In the paragraph, we showed that the study achieved the aim on the basis of the results (Lines 245-255).

(5) Why did the authors select these items in their scoring system and did not add more variables?

Thank you for this valuable comment. We selected the items listed in the Japanese guidelines for acute pancreatitis and post-ERCP pancreatitis. We added this description (Lines 123-126). The original references of the guidelines were cited in the first draft version.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Comment1. Please revise these documents: copyright, disclaimer, ethics approval, consent to participate, consent for publication, data and material availability, competing interests, funding, authors' contributions, and acknowledgments.

First, thank you for reviewing our manuscript. We have already described the required information in the “author information” section. The sentences containing this information were proofread in English.

Reviewer #2 (Recommendations for the authors):

Comment 1. It would be best if you did this study in a Prospective way for more validation.

First, thank you for reviewing our manuscript. We have revised our manuscript according to your comments. It’s exactly as you said. These points are limitations (Lines 312-318, lines 326-327). We hope that future validation studies over wider geographic regions will prove our opinions.

Comment 2. The model name should be Acronyum (the first letter of the five items in the risk model).

Thank you for this valuable comment. Sorry, we could not create a memorable model name using the first letter of the five items.

Comment 3. You say that you include the pre-procedure criteria that predict PEP. You state one of the items, pancreatic duct procedure. Do you mean it is a history?

Thank you for this valuable comment. This means that the main purpose is the pancreatic duct. Therefore, the pancreatic duct procedure is listed as “planned pancreatic duct procedures” in Figure 2 (Lines 40-41, 231-234). When an unintended pancreatic duct procedure is performed, we can calculate the risk score by adding two points for “planned pancreatic duct procedures” (Lines 48-49, 247-250).

Comment 4. Regarding calcification, do you mean chronic pancreatitis? It needs more clarification regarding its degree.

Thank you for this valuable comment. We regard pancreatic calcification as a finding of chronic pancreatitis. Pancreatic calcification was defined as the degree that was confirmed by imaging, such as CT, MRI, and EUS. These definitions have been written in the first draft version (Lines 134-137).

Comment 5. Why don't you include young age in the model? Your result found that age less than 50 is significantly associated with PEP.

Thank you for this valuable comment. We selected the PEP risk factors listed in the Japanese guidelines for acute pancreatitis and post-ERCP pancreatitis. Age less than 50 years was listed as a PEP risk factor in the Japanese guidelines for acute pancreatitis. We added this description (Lines 123-126).

Comment 6. There is an ancient reference, some of them in 1994,1996.

Sorry for the old references. These references were written by Cotton et al. 1991, Freeman et al. 1996, and Loperfido et al. 1998. These are still important today. The diagnostic criteria for PEP were determined in the report written by Cotton et al., which is Cotton’s criteria. The other two references are representative reports that described risk factors for PEP, and these two reports were cited in the Japanese guidelines for pancreatitis written by Takada et al. 2022 (Lines 123-126).

Comment 7. In the introduction, you say that the first score includes one of the items for PEP pain during the procedure. It is a little bit strange.

Thank you for this comment. The first PEP risk score did not involve PEP pain but involved pain during the procedure (Line 68).

Comment 8. We know that once ERCP is indicated, you justify the importance of the risk model, stating that if one or more risks are found, we can do EUS or PTD. It is not reasonable to abort the procedure in case of frequent pancreatic duct cannulation or cancel ERCP if pt has one or more risk factors.

Thank you for this valuable comment. If ERCP is performed for high-risk patients, prophylaxes for PEP, such as procedures by experts, pancreatic stent placement, and NSAID suppository insertion, should be performed as much as possible (Lines 281-287, 308-311).

Comment 9. Regarding ERCP pancreatitis criteria, does it include amylase 3t or lipase?

Thank you for this comment. We used Cotton’s criteria for diagnosing PEP. Cotton’s criteria include hyperamylasemia (more than three times the normal upper limit) at least 24 hours after ERCP (114-116).

Comment 10. It is well known that pr with functional biliary disorder has a high incidence of PEP; it doesn't need a manometer for diagnosis. It needs to be included.

Thank you for this comment. Moreover, functional biliary disorders are difficult to diagnose before ERCP procedures (Lines 259-262). The factor that is not apparent before ERCP could not be included in the predictive PEP scoring system.

Comment 11: What is gabexare and nafamost.

Thank you for this comment, and sorry for our insufficient explanation. These compounds include gabexate masilate and nafamostat masilate, which are protease inhibitors. In some institutions, protease inhibitors are used as prophylaxis for PEP. We added “protease inhibitors” (Lines 138-139, Tables 1 and 2).

Reviewer #3 (Recommendations for the authors):

Comment 1. The sample size needs clarification.

First, thank you for reviewing our manuscript. The sample size has been included in the “Methods” section (Lines 157-165).

Comment 2. They need to be mentioned cause they depend on old references in discussion and background.

Thank you for this comment. The previous references were written by Cotton et al. 1991, Freeman et al. 1996, and Loperfido et al. 1998. These are still important today. The diagnostic criteria for PEP were determined in the report written by Cotton et al., which is Cotton’s criteria. The other two references are representative reports that described risk factors for PEP, and these two reports were cited in the Japanese guidelines for pancreatitis written by Takada et al. 2022 (Lines 122-126). In the background and discussion, we added new recent references and information related to the references (Lines 65-67, 285-287, 291-295, 308-311).

Comment 3. Case definition should be added to the methodology.

Thank you for this comment. We added patient information. Please refer to the response against the eLife assessment, weakness, (2).

Comment 4. Do you include all who met the inclusion criteria, or was there any random sampling technique?

No, we did not use random sampling techniques.

Comment 5. What is the value of comparing the development and validation groups? I do not think it adds anything new as if you want to exclude confounders. Has the comparison revealed that a confounder does exist? What was your point of view concerning that?

Thank you for this valuable comment, and sorry for the insufficient explanation. The differences between the development cohort and the validation cohort are important because the goodness of fit for the score could be confirmed in significantly different groups. We added this explanation (Lines 197-199, 251-253).

eLife Assessment

This important study provides one mechanism that can explain the rapid diversification of poison-antidote pairs in fission yeast: recombination between existing pairs. The evidence is largely solid, but the study can benefit from demonstrating that the novel poison-antidote constructed by the authors can serve as a meiotic driver. The work is of interest to colleagues studying genetic incompatibilities.

Reviewer #1 (Public review):

Summary

The authors determine the phylogenetic relation of the roughly two dozen wtf elements of 21 S. pombe isolates and show that none of them in the original S. pombe are essential for robust mitotic growth. It would be interesting to test their meiotic function by simply crossing each deletion mutant with the parent and analyzing spores for non-Mendelian inheritance. If this has been reported already, that information should be added to the manuscript. If not, I suggest the authors do these simple experiments and add this information.

Strengths:

The most interesting data (Figure 4) show that one recombinant (wtfC4) between wtf18 and wtf23 produces in mitotic growth a poison counteracted by its own antidote but not by the parental antidotes. Again, it would be interesting to test this recombinant in a more natural setting - meiosis between it and each of the parents.

Weaknesses:

In the opinion of this reviewer, some minor rewriting is needed.

Reviewer #2 (Public review):

Summary:

This important study provides a mechanism that can explain the rapid diversification of poison-antidote pairs (wtf genes) in fission yeast: recombination between existing genes.

Strengths:

The authors analyzed the diversity of wtf in S. pombe strains, and found pervasive copy number variations. They further detected signals of recurrent recombination in wtf genes. To address whether recombination can generate novel wtf genes, the authors performed artificial recombination between existing wft genes, and showed that indeed a new wtf can be generated: the poison cannot be detoxified by the antidotes encoded by parental wtf genes but can be detoxified by own antidote.

Weaknesses:

The study can benefit from demonstrating that the novel poison-antidote constructed by the authors can serve as a meiotic driver.

Reviewer #3 (Public review):

Summary:

In this manuscript, Wang and colleagues explore factors contributing to the diversification of wtf meiotic drivers. wtf genes are autonomous, single-gene poison-antidote meiotic drivers that encode both a spore-killing poison (short isoform) and an antidote to the poison (long isoform) through alternative transcriptional initiation. There are dozens of wtf drivers present in the genomes of various yeast species, yet the evolutionary forces driving their diversification remain largely unknown. This manuscript is written in a straightforward and effective manner, and the analyses and experiments are easy to follow and interpret. While I find the research question interesting and the experiments persuasive, they do not provide any deeper mechanistic understanding of this gene family.

Strengths:

(1) The authors present a comprehensive compendium and analysis of the evolutionary relationships among wtf genes across 21 strains of S. pombe.

(2) The authors found that a synthetic chimeric wtf gene, combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves like a meiotic driver that could only be rescued by the chimeric antidote but neither of the parental antidotes. This is a very interesting observation that could account for their inception and diversification.

Weaknesses:

(1) Deletion strains

The authors separately deleted all 25 Wtf genes in the S. pombe ference strain. Next, the authors performed a spot assay to evaluate the effect of wtf gene knockout on the yeast growth. They report no difference to the WT and conclude that the wtf genes might be largely neutral to the fitness of their carriers in the asexual life cycle at least in normal growth conditions.

The authors could have conducted additional quantitative growth assays in yeast, such as growth curves or competition assays, which would have allowed them to detect subtle fitness effects that cannot be quantified with a spot assay. Furthermore, the authors do not rule out simpler explanations, such as genetic redundancy. This could have been addressed by crossing mutants of closely related paralogs or editing multiple wtf genes in the same genetic background.

Another concern is the lack of detailed information about the 25 knockout strains used in the study. There is no information provided on how these strains were generated or, more importantly, validated. Many of these wtf genes have close paralogs and are flanked by repetitive regions, which could complicate the generation of such deletion strains. As currently presented, these results would be difficult to replicate in other labs due to insufficient methodological details

(2) Lack of controls

The authors found that a synthetic chimeric wtf gene, constructed by combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves as a meiotic driver that can be rescued only by its corresponding chimeric antidote, but not by either of the parental antidotes (Figure 4F). In contrast, three other chimeric wtf genes did not display this property (Figure 4C-E). No additional experiments were conducted to explain these differences, and basic control experiments, such as verifying the expression of the chimeric constructs, were not performed to rule out trivial explanations. This should be at the very least discussed. Also, it would have been better to test additional chimeras.

3. Statistical analyses

In line 130 the authors state that: "Given complex phylogenetic mixing observed among wtf genes (Figure 1E), we tested whether recombination occurred. We detected signals of recombination in the 25 wtf genes of the S. pombe reference genome (p = 0) and in the wtf genes of the 21 S. pombe strains (p = 0) using pairwise homoplasy index (HPI) test. ". Reporting a p-value of 0 is not appropriate. Exact P-values should be reported.

Author response:

Public Reviews:

Reviewer #1 (Public review):

Summary

The authors determine the phylogenetic relation of the roughly two dozen wtf elements of 21 S. pombe isolates and show that none of them in the original S. pombe are essential for robust mitotic growth. It would be interesting to test their meiotic function by simply crossing each deletion mutant with the parent and analyzing spores for non-Mendelian inheritance. If this has been reported already, that information should be added to the manuscript. If not, I suggest the authors do these simple experiments and add this information.

Thanks for the great summary! Most of the wtf genes have been tested for meiotic drive phenotypes previously by Bravo Nunez et al. (2020; http://doi.org/10.1371/journal.pgen.1008350). The reference was cited in our original manuscript, and we added the details in the revised manuscript.

Strengths:

The most interesting data (Figure 4) show that one recombinant (wtfC4) between wtf18 and wtf23 produces in mitotic growth a poison counteracted by its own antidote but not by the parental antidotes. Again, it would be interesting to test this recombinant in a more natural setting - meiosis between it and each of the parents.

We will test the meiotic driver phenotype of the wtfC4 we constructed in S. pombe as suggested.

Weaknesses:

In the opinion of this reviewer, some minor rewriting is needed.

We did the rewriting as this reviewer suggested in the comments to authors.

Reviewer #2 (Public review):

Summary:

This important study provides a mechanism that can explain the rapid diversification of poison-antidote pairs (wtf genes) in fission yeast: recombination between existing genes.

Thanks!

Strengths:

The authors analyzed the diversity of wtf in S. pombe strains, and found pervasive copy number variations. They further detected signals of recurrent recombination in wtf genes. To address whether recombination can generate novel wtf genes, the authors performed artificial recombination between existing wft genes, and showed that indeed a new wtf can be generated: the poison cannot be detoxified by the antidotes encoded by parental wtf genes but can be detoxified by own antidote.

Thanks for the great summary!

Weaknesses:

The study can benefit from demonstrating that the novel poison-antidote constructed by the authors can serve as a meiotic driver.

We will test the meiotic driver phenotype of the wtfC4 we constructed in S. pombe as suggested.

Reviewer #3 (Public review):

Summary:

In this manuscript, Wang and colleagues explore factors contributing to the diversification of wtf meiotic drivers. wtf genes are autonomous, single-gene poison-antidote meiotic drivers that encode both a spore-killing poison (short isoform) and an antidote to the poison (long isoform) through alternative transcriptional initiation. There are dozens of wtf drivers present in the genomes of various yeast species, yet the evolutionary forces driving their diversification remain largely unknown. This manuscript is written in a straightforward and effective manner, and the analyses and experiments are easy to follow and interpret. While I find the research question interesting and the experiments persuasive, they do not provide any deeper mechanistic understanding of this gene family.

Thanks! Please see the following for our point-to-point response.

Strengths:

(1) The authors present a comprehensive compendium and analysis of the evolutionary relationships among wtf genes across 21 strains of S. pombe.

(2) The authors found that a synthetic chimeric wtf gene, combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves like a meiotic driver that could only be rescued by the chimeric antidote but neither of the parental antidotes. This is a very interesting observation that could account for their inception and diversification.

Thanks for the great summary!

Weaknesses:

(1) Deletion strains

The authors separately deleted all 25 Wtf genes in the S. pombe ference strain. Next, the authors performed a spot assay to evaluate the effect of wtf gene knockout on the yeast growth. They report no difference to the WT and conclude that the wtf genes might be largely neutral to the fitness of their carriers in the asexual life cycle at least in normal growth conditions.

The authors could have conducted additional quantitative growth assays in yeast, such as growth curves or competition assays, which would have allowed them to detect subtle fitness effects that cannot be quantified with a spot assay. Furthermore, the authors do not rule out simpler explanations, such as genetic redundancy. This could have been addressed by crossing mutants of closely related paralogs or editing multiple wtf genes in the same genetic background.

Another concern is the lack of detailed information about the 25 knockout strains used in the study. There is no information provided on how these strains were generated or, more importantly, validated. Many of these wtf genes have close paralogs and are flanked by repetitive regions, which could complicate the generation of such deletion strains. As currently presented, these results would be difficult to replicate in other labs due to insufficient methodological details

We will generate growth curves for all the 25 wtf deletion strains. We will also provide detailed for wtf gene knockout. However, for 25 wtf genes, there are too many combinations for editing two genes, and it is technically challenging to knock out multiple wtf together. Nevertheless, our results suggest single wtf gene has little effect on the host fitness under normal condition.  

(2) Lack of controls

The authors found that a synthetic chimeric wtf gene, constructed by combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves as a meiotic driver that can be rescued only by its corresponding chimeric antidote, but not by either of the parental antidotes (Figure 4F). In contrast, three other chimeric wtf genes did not display this property (Figure 4C-E). No additional experiments were conducted to explain these differences, and basic control experiments, such as verifying the expression of the chimeric constructs, were not performed to rule out trivial explanations. This should be at the very least discussed. Also, it would have been better to test additional chimeras.

We will verify the expression of the chimeric genes, and test the phenotype of meiotic diver for wtfC4 in S. pombe.

(3) Statistical analyses

In line 130 the authors state that: "Given complex phylogenetic mixing observed among wtf genes (Figure 1E), we tested whether recombination occurred. We detected signals of recombination in the 25 wtf genes of the S. pombe reference genome (p = 0) and in the wtf genes of the 21 S. pombe strains (p = 0) using pairwise homoplasy index (HPI) test. ". Reporting a p-value of 0 is not appropriate. Exact P-values should be reported.

We will report the exact p values in the revised manuscript.

eLife Assessment

This important chronobiological study in mice suggests that light modulated activity of Cdk5 activity on the PKA-CaMK-CREB signaling pathway provides missing molecular mechanistic details to understand light-induced circadian clock phase delays during the early night, but not for phase advances in the morning. The authors provide convincing evidence bridging from behavioral to molecular/cellular experiments to neural activity imaging.

Reviewer #1 (Public review):

In the manuscript Cyclin-dependent kinase 5 (Cdk5) activity is modulated by light and gates rapid phase shifts of the circadian clock Brenna et al., study the role of Cdk5 on circadian rhythms, the authors aim to elucidate the role of Cyclin-Dependent Kinase 5 (Cdk5) in modulating circadian rhythms, particularly in response to light cues. They hypothesized that Cdk5 acts as a gatekeeper, regulating the sensitivity of the circadian clock to light-induced phase shifts.

Strengths:

• Novelty: The study presents a novel mechanism by which Cdk5 influences circadian rhythms, particularly its role in modulating the light-induced phase-shifting response.<br /> • Experiments: The authors have employed a combination of molecular, cellular, and behavioural techniques, including genetic manipulations, biochemical assays, and electrophysiology, to investigate the role of Cdk5. The set of experiments performed in this work is non-trivial, done to a high standard and the additional experiments, data and textual alterations presented following the 1st round of review needs to be lauded.<br /> • Data: The data is well-presented in clear figures and appropriately described in the text.

Weaknesses:

• Although I found the data on Cdk5 gating light responses highly convincing there could be additional mechanisms which the authors have duly acknowledged and discussed in their text.<br /> In my assessment, the authors have convincingly demonstrated that Cdk5 plays a critical role in gating the light-induced phase-shifting response of the circadian clock. Their results strongly support their conclusions, as evidenced by their findings:<br /> This study provides valuable insights into the molecular mechanisms underlying circadian rhythm regulation and the impact of light on the circadian clock. The findings have the potential to influence future research in the field of chronobiology and may have implications for understanding and treating circadian rhythm disorders.<br /> The methods and data presented in this study are valuable to the field and can be used to further investigate the role of Cdk5 and other signalling pathways in circadian rhythm regulation.<br /> Broader context<br /> The circadian clock is a fundamental biological process that regulates various physiological functions, including sleep-wake cycles, hormone secretion, and metabolism. Disruptions to the circadian clock have been linked to a variety of health problems, such as sleep disorders, metabolic disorders, and cancer. Understanding the molecular mechanisms that underlie circadian rhythm regulation is essential for developing effective treatments for these disorders.

All in all, I have no reservations regarding the manuscript titled "Cyclin-dependent kinase 5 (Cdk5) activity is modulated by light and gates rapid phase shifts of the circadian clock by Brenna et al. After consideration of the authors' revisions, I believe the manuscript has been significantly improved. I commend the authors for their diligence in addressing the reviewers' comments and for the quality of their research.

Reviewer #2 (Public review):

Summary:

Definition of the role of CdK5 in circadian locator activity and light induced neural activity in the mouse SCN in-vivo revealing its mode of action through PKA-CaMK-CREB signaling pathway.

Strengths:

The experimental approaches are carried from in-vivo, to cellular and molecular level and provide first evidence for the specific involvement of CdK5 in light-induced phase advance of the free-running rhythm.

Weaknesses:

The behavioral analyses are limited to some selected parameters.

Downstream effects on circadian oscillation of gene expression and physiological functions in other brain regions, organs is missing.

Comments on revisions:

I am happy with the manuscript in its present form.

Author response:

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

Reply to reviewer comments:

(1) Given the interpretations of this study hinge on the specificity of the antibodies used in immune fluorescence, the authors should provide full western-blot images of all their antibodies in supplementary information. 

The commercial antibodies have been validated by the provider. 

Additionally, we did our own tests. Of note is that proper validation of any antibody is only possible by using a knockout mouse for each protein analyzed (i.e. for pPKA wt vs. pka ko mice). This is not possible, because we do not have all these knock-out strains. However, specific proteins like pPKA, pCAMKII, and pCAMKIV are known to be increased by a light pulse. We show by western blot that pPKA (Fig. 2a, b) and pCamKII (Fig. S2a, b) are increased in wt animals mirroring what we observed in the immunofluorescence. These results suggest that the signal is specific to these antibodies. We provide a full panel of western blots, including the other proteins studied by immunofluorescence such as pCamKIV, pCREB, CaV 3.1, and pDARP32 and show that they detect a protein of the expected size. Full Western-blots mentioned in the manuscript are shown in Supplementary Figure 7. Below are additional validations of antibodies used in the immunofluorescence experiments.

Author response image 1.

Author response image 2.

(2) The explanation in the results section surrounding Fig. 4 seems to be specific for the representative trace rather than the group. Specifically, does the following statement apply to all the replicates?  " A Ca2+ transient was observed right before the light was given at ZT14 (Fig. 4b), which showed the same magnitude as those observed during and after the light stimulus". 

If not this should be corrected.  

We have replaced now Fig. 4b with an average trace of all experiments. The individual traces can be seen in supplementary figure 4d.

(3) Are lines 236 -244 and figure 5A/B demonstrating shCDK5 being similar to no-calcium or EGTA conditions at the level of CREB not contradicting Figure 3 which argues that the reason behind the increase in CAMK-phosphorylation and pCREB following shCDK5 is increased basal calcium? If this is the case then why does removing the external calcium phenocopy shCDK5 in these cells? The authors need to clarify this and give an explanation. 

(4) The authors should explain why they see an equivalent level (or more) of CREB activation, 5 minutes following forskolin activation in Ca2+-free condition (apparent in the case of shCDK5 and EGTA) in the FRET assay. Does this not imply PKA is the most likely candidate mediating this reaction at this stage? Given this interaction has been demonstrated in multiple (other) experiments including in vitro isolated enzyme experiments involving CREB and PKA (E.G. fig 6A in PMID: 2900470) an absence of p-PKA pulldown is not sufficient to justify the non-involvement of PKA (PMID: 22583753). This statement needs support in the form of positive data or acknowledging the limitations in the text (conditions, single technique, etc). 

(5) The authors should better explain the fret pairs used in the experiments involving ICAP for the reader's benefit - a reduction in fluorescence as a function of CREB activation is non-intuitive.

We answer all three questions (3-5) together since they belong to the same concept.

(1) How FRET works.

The Forster resonance energy transfer (FRET) technique is widely used to investigate molecular interactions between proteins such as CREB: CBP in living cells. We used a sensor called ICAP (an Indicator of CREB Activation due to Phosphorylation) published by Friedrich and colleagues in 2010

(https://doi.org/10.1074/jbc.M110.124545). The sensor is composed of three different elements: 1) the KID domain of CREB containing the Ser-133, which is phosphorylated upon forskolin induction in our experimental setup, 2) the KIX domain of CBP, which is responsible for the dimerization with phospho-CREB and 3) a short linker that separates the KID with the KIX domain. KID is flanked by a cyan fluorescent protein (CFP), while KIX is flanked by a yellow fluorescent protein (YFP). When KID is not phosphorylated, the ICAP conformation allows CFP - stimulated by blue UV light - to transfer energy to YFP, producing FRET resulting in yellow light emission. Therefore, the ratiometric analysis FRET/CFP shows FRET > CFP. After a stimulus (forskolin), the serine-133 in KID is phosphorylated and KID can bind to KIX. The dimerization separates CFP from YFP, resulting in decreased FRET and increased CFP-dependent blue light emission (see Author response image 3 below). Therefore, the ratiometric analysis FRET/CFP shows FRET<CFP over time (usually within 20’ after the forskolin stimulus).

Author response image 3.

FRET model. On the left is a schematic representation of how ICAP works. On the right, an example of the quantified FRET decrease associated with increased KID: KIX interaction.

(2) The ‘apparent’ contradiction between Figure 5A and Fig 3.

As mentioned before, the chosen FRET method is ratiometric, meaning that a relative FRET signal in fluorescence is measured compared to the baseline (absence of forskolin, assay buffer). The FRET experiment can only tell whether there is a change in the phosphorylation state of KID during the live imaging comparing the baseline to the period after the forskolin treatment. The result produces a delta [ (time after forskolin)(baseline)]. The higher the delta, the more KID is phosphorylated after forskolin treatment. If KID phosphorylation is not increased compared to the baseline, the FRET signal tends to return to the baseline with a reduced delta [ (time after forskolin)-(baseline)]. Therefore, the experiment does not tell at the quantitative level the amount of KID (CREB domain) phosphorylation before the stimulus. It only tells whether after the stimulus the phosphorylation is increased producing or not a delta. This means that the lack of delta can be caused by: A) high KID phosphorylation in the baseline which does not further increase after the forskolin stimulus; B) very low KID phosphorylation in the baseline which does not increase after the forskolin stimulus. In Fig. 5A, wt cells (orange trace, lines, and double arrow) show a higher delta compared to the ko cells (blue trace, lines, and double arrow). The result indicated that the phosphorylation of CREB (KID domain) is increased after the forskolin stimulus only in the wt. To that extent, the results are in line with the experiment that we show in Figure 3. Indeed, the increased delta in CREB phosphorylation is observed only in the scramble animals, where it is lost in the ko (the blue double arrow indicates the delta in the scramble). 

Author response image 4.

(3) The FRET signal within 3 minutes after forskolin stimulation

The signal mentioned by the reviewers at 5’ is an artifact given by the light diffraction promoted by the addition of Forskolin in DMSO which propagates through the plate. The same effect is observed in the only DMSO treatment (Fig.S5). Therefore, it needs not to be taken into account. The amplitude of this signal in this window of time is due to many independent variables (buffer composition, cell shape, room temperature, pipetting), therefore it is not possible to speculate any consideration about it. We never consider this time window for describing our results.

Author response image 5.

(4) Role of PKA and considerations about experiments performed in Fig. 5a and b

To answer the question about the role of PKA, we believe it is a pivotal player. Our results indicate that PKA might promote CaV3.1, the entrance of calcium, and therefore, CAM Kinase pathway activation leading to CREB phosphorylation (Fig. 5). However, if the calcium is depleted, even a channel activation mediated by PKA cannot propagate the signal. For that reason, when we deplete calcium in wt cells as we do in the experiment performed in Figure 5B the activation of PKA alone cannot promote the CREB phosphorylation associated with a reduction of the FRET signal. As mentioned before, the FRET method gives a binary answer. It means either a higher or lower delta comparing time after forskolin to baseline. It cannot give stoichiometric info about the level of calcium and/or phosphorylation in the baseline. To that extent, the FRET experiment in Figure 5A cannot be connected to the experiment in Figure 5B. The method is the same, but the scientific questions are different. In Figure 5A we demonstrate that CDK5 plays a role in the PKA activation pathway. In Figure 5B we demonstrate that the general pathway needs calcium.

We modified the text accordingly.

(6) The presentation of the data in Figure 6 seems to be divergent from the rest of the data presentations. Please make it more consistent and also provide more explanations. Specifically, the authors suggest increased P-CREB nuclear localization (and an increase in phosphorylated PKA/CAMK) following shCDK5. Won't this lead to an increase in Per1, Dec1, cFos, and Sik1 basally (pre-light pulse)?

We followed the reviewer's suggestion and present data in Figure 6 as done before in the manuscript. The reviewers should also consider our papers published before (Brenna et al., 2019; Brenna et al., 2021). In these papers, we demonstrate two important concepts that are in line with this manuscript. First, the lack of CDK5 promotes PER2 degradation and lack of nuclear translocation (Brenna et al., 2019). Second, PER2 plays a scaffold role in promoting the formation of the CREB transcriptional complex involved in the regulation of the expression of light-dependent genes (Brenna et al., 2021). Therefore, the take-home message here is that even if a lack of Cdk5 promotes a higher basal level of CREB phosphorylation, it also promotes PER2 degradation. Therefore, without PER2, the CREB-dependent gene expression is reduced. For this reason, we say that CDK5 gates phase shift (via PKA-CAM Kinases-CREB axis) of the circadian clock (via PER2).

(7) The authors should discuss why calcium-sensitive phosphatases such as PP2A (PMID: 23752926) or calcineurin (PMID: 10217279) are not considered candidates for dephosphorylation of DARPP32 as these are described previously (CDK5) and conditions of increased calcium as seen here would favour these enzymes. The phospho-T75 data are supportive, but such additional discussion could be important given the past demonstrations.

We thank the reviewers for the great insight. The pathway that promotes the T75 phosphorylation/dephosphorylation indeed includes many players as calcineurin and PPA2A. We mention this in the discussion now as follows:

However, phosphatases such as PP2A and calcineurin, which de-phosphorylate DARPP32 including the Cdk5 phosphorylation site, may be involved in this process as well (Girault and Nairn, 2021). Upon light treatment and increase of Ca2+ these phosphatases would dephosphorylate DARPP32 and thereby inactivate it, leading to PKA activation. This process may occur in parallel to the Cdk5 regulation of DARPP32 contributing to a sustained activation of the light signaling pathway via PKA activation.

(8) additional details on the knock-downs would be helpful: 

- the relative amount of reduction in gene expression upon shRNA treatment should be provided  - How was the exact viral delivery and reduction in shRNA-induced knock-down confirmed for the individual animals?  

The validation of Cdk5 knockdown was widely performed in the previous paper (Brenna et al., 2019, Fig2-Fig supp1, and Fig3-Fig suppl2). We used the same mice. We confirmed the goodness of the silencing also in the supp figure 1A of the current paper.

(9) The authors only focus on male mice. This is rather incomplete, as it leaves away an important half of biological reality. Testing relevant aspects of the work in female mice would close this significant gap and also increase the number of biological replicates, which can still be considered relatively low. 

We thank the reviewers for the suggestion. We injected female mice and performed the Ashoff type-II light pulse experiment at ZT14 and observe the same phenotype as for male mice. This is stated now in the paper and the data are shown in supplemental figure 1 e-f.

(10) Given the roles of CdK5 in circadian clock period length regulation, but also light-induced phase delays, it would be interesting for a broader audience to discuss possible expectations of CdK5's roles, e.g. 

(a) How will other circadian parameters, eg. activity bouts (numbers, length, activity onset/ offset) be affected? 

(b) How does that relate to sleep, sleep phases? 

(c) What is the expected impact on other physiological rhythms, eg food intake, cortisol levels? 

(d) What are the expected effects on circadian oscillation of gene expression in other brain regions, organs? 

We thank the reviewers for the observations. 

a) The activity was discussed in the previous paper (Brenna et al. 2019). ShCdk5 mice show a reduced activity in both DD and LD 12:12 compared to wt, mirroring the Per2 brdm phenotype (Figure- Suppl3, with the difference mostly observed at night time (Figure 2-suppl4).

We also demonstrate in Suppl Fig1 b, c of the current paper that light pulse does not affect the period length either in scramble mice or in sh Cdk5.

b) We performed preliminary experiments with SCN shCdk5 knock-down animals and compared them to scr control mice using the Piezo sleep system. Total sleep was not different, however during the dark phase shCdk5 animals tended to sleep a bit more, similar to the neuronal Per2 KO animals (Wendrich et al., 2023 https://doi.org/10.3390/clockssleep5020017 ). After sleep-deprivation no differences were observed between shCdk5 and scr animals. This was comparable to the neuronal Per2 KO animals that also showed no phenotype after sleep deprivation.

c) and d) We did not investigate food intake, cortisol, or other parameters involving peripheral clocks. We did not investigate the gene expression in other brain regions because the SCN is the main brain region involved in the regulation of the circadian clock phase shift. However future studies will address these questions.

eLife Assessment

This study examined the role of the dorsal medial prefrontal cortex of mice in anticipating reward-value switch points in a two-armed bandit task. They demonstrate the dorsal medial prefrontal cortex is involved in task performance and use model-based methods to uncover the algorithmic processes affected by prefrontal cortex perturbations. If the claims were supported, this would be a valuable finding. Unfortunately, the reviewers recognised significant issues with the task design and analyses, making the evidence to support the role of the prefrontal cortex in anticipation of switches inadequate.

Reviewer #1 (Public review):

Summary:

In this manuscript, the authors train mice on a two-armed bandit task, in which the reward value associated with the arms suddenly switches in a pseudorandom fashion. Their first finding is that the mice are able to anticipate the reward value switch points after long blocks, evident both prior to the switch point with higher rates of switching to the less-rewarded arm, and after the switch point with faster transition to the more-rewarded arm. They next find that unilateral ACAd/MO lesion / optogenetic silencing (surprisingly) causes greater anticipation of reward switch points, both prior to and after the switch point. They use behavioral modeling to argue that the unilateral ACAd/MO lesion effects are due to an increase in the contralateral hazard rate. Finally, they found that bilateral lesions did not have any effect on the hazard rate, suggesting that the unilateral lesion effect is due to balancing between hemispheres. This manuscript employed a clever behavioral design and analysis approach, though the effects were somewhat difficult to interpret and the author's interpretation relies heavily on the accuracy of their underlying behavioral model.

Strengths:

This paper employs a well-designed task that allows the researchers to detect whether mice have noticed a change in reward value both before and after the change takes place. The use of unilateral and bilateral inactivation experiments allowed the authors to test the role of the ACAd/MO region in the change point estimation. They found that unilateral inactivation, but not bilateral inactivation, had a significant effect on behavior. They performed sophisticated behavioral analysis to determine how ACAd/MO perturbations affect decision-making variables. This topic is of interest to the field, and the results are presented clearly and generally convincing.

Weaknesses:

The observed effects of the lesions are somewhat counterintuitive, with lesions appearing to affect persistence within a block more than change point detection itself-the mice actually adjusted more quickly to changes in reward values. Moreover, they had no issue detecting change points after bilateral inactivation. As a result, I'm not sure if the main framing of the article (including the title) is supported by their findings. Finally, I was unsure how the differences between unilateral and bilateral inactivation could be explained by their behavioral model.

Reviewer #2 (Public review):

Summary:

The manuscript by Murphy et al. titled "Change point estimation by the mouse medial prefrontal cortex during probabilistic reward learning" investigated the role of the mPFC in the exploitation of task structure. Previous work had shown that monkeys and humans exploit predictable task structures (e.g., switching rapidly when heavily trained a reversal learning task), but whether this was also the case for mice was not known. To test this, Murphy et al. trained head-fixed mice on a two-armed bandit task in which the contingencies reversed when mice met a performance criterion (10 trials choosing the better option) plus an additional random number of trials (referred to as Lrandom). They found that as the length of Lrandom increased, mice began to exhibit pre-emptive switching in their choices as if they were expecting and/or anticipating the reversal to occur. They report that unilateral lesions of the mPFC (ACC + MO) led to earlier pre-emptive switching (although I found this part of the manuscript the most challenging to understand) and faster post-reversal switching that they argue reflects an impairment in the proper estimation of the reversal. They also report that this requires inter-hemispheric coordination because bilateral lesions did not further impair this estimation. Optogenetic inhibition just prior to the mouse making a choice recapitulated some of the behavioral metrics observed in the mPFC lesioned animals. Finally, the authors developed a novel hybrid belief-choice kernel model to provide a computational approach to quantifying these behavioral differences.

Strengths:

The paper is extremely well written and was an absolute pleasure to read. The results are novel and provide exciting (although not surprising) evidence that mice exploit task structures to earn rewards. Moreover, the experiments were well-designed and included appropriate controls and/or control conditions that support their findings.

Weaknesses:

Some of the results need to be clarified and/or language changed to ensure that readers will understand. Restricting analyses to expert mice that show the predicted effect is problematic.

Reviewer #3 (Public review):

Summary:

The authors examine the role of the medial frontal cortex of mice in exploiting statistical structure in tasks. They claim that mice are "proactive": they predict upcoming changes, rather than responding in a "model-free" way to environmental changes. Further, they speculate that the estimation of future change (i.e., prediction of upcoming events, based on learning temporal regularities) might be "a main ... function of dorsal medial frontal cortex (dmFC)." Unfortunately, the current manuscript contains flaws such that the evidence supporting these claims is inadequate.

Strengths:

Understanding the neural mechanisms by which we learn about statistical structure in the world is an important goal. The authors developed an interesting task and used model-based techniques to try to understand the mechanisms by which perturbation of dmFC influenced behavior. They demonstrate that lesions and optogenetic silencing of dmFC influence behavior, showing that this region has a causal influence on the task.

Weaknesses:

I was concerned that the main behavioral effects shown in Figure 1F were a statistical artifact. By requiring the Geometric block length to be preceded by a performance-based block, the authors introduce a dependence that can generate the phenomena they describe as anticipation.

To demonstrate this, I simulated their task with an agent that does not have any anticipation of the change point (Reviewer image 1). The agent repeats the previous action with probability `p(repeat)` (similar to the choice kernel in the author's models). If the agent doesn't repeat then the next choice depends on the previous outcome. If the previous choice was rewarded, it stays with `P(WS)` and chooses randomly with `1-P(WS)`. If the previous choice was unrewarded, it switches with `P(LS)` and chooses randomly with `1-P(LS)`.

Review image 1.

An agent with `P(WS)=P(LS)=P(repeat)=0.85` shows the same phenomena as the mice: a difference in performance before the block switch and "earlier" crossing of the midpoint after the switch. https://imgdrop.io/image/aHn6y. The phenomena go away in the simulations when a fixed block length of 20 trials is followed by a Geometric block length.

The authors did not completely rely on the phenomena of Figure 1F for their conclusions. They did a model comparison to provide evidence that animals are anticipating the switch. Unfortunately, the authors did not use state-of-the-art methods in this section of the paper. In particular, they failed to show that under a range of generative parameters for each model class, the model selection process chooses the correct model class (i.e. a confusion matrix). A more minor point, they used BIC instead of a more robust cross-validated metric for model selection. Finally, instead of comparing their "best" anticipating model to their 2nd best model (without anticipation), they compared their best to their 4th best (Supp Fig 3.5). This seems misleading.

Given all of the the above issues, it is hard to critically evaluate the model-based analysis of the effects of lesions/optogenetics.

Author response:

We appreciate the reviewers' thoughtful and constructive comments. In this provisional response, we aim to address what we see as the key critiques, with a detailed, point-by-point reply to be provided alongside the revised manuscript. Below, we outline how we intend to address these critiques in the revised manuscript.

(1) We will revise sections of the manuscript to ensure that all results, particularly those concerning the effects of lesions, are described more clearly and with sufficient context. This includes providing additional visualizations and rewording any ambiguous statements.

(2) In this study, we examined a subset of 7,396 blocks where animals quickly adapted after block switches (achieving LCriterion in 20 or fewer trials), thereby focusing on expert-level performance and avoiding periods that might be affected by low motivation. It is valid to question whether the same observations would hold if the full dataset were analyzed. To address this, we expanded our analysis to include a supplementary figure Supplementary Figure 1.1 that illustrate the same relationships based on block length (BL) instead of LRandom, both with and without the restriction on LCriterion (n = 9,156 blocks in which the block length is under 100 trials, without any LCriterion restrictions), and based on LRandom without any LCriterion restrictions and with a less stringent LCriterion restriction (with ≤ 50 Trials for the criterion). This method allowed us to include all trials in our dataset. We observed similar effects of block length on choice behavior around switches (Figure 3), confirming the consistency of our findings across different analytical conditions.

(3) We agree that robust validation of model selection is crucial. To address this, we will generate a confusion matrix to assess whether our model selection process accurately identifies the correct model class across a range of generative parameters. Include additional model selection metrics, such as cross-validation, to complement the BIC analysis and provide a more robust comparison of models.

(4) We acknowledge the concern regarding our comparison of the "best" and the "4th best" models. The "4th best" model was chosen because it is the most widely recognized in the literature. Our intention was to demonstrate the performance of the most commonly used model, but we understand how this may have been misleading. To address this, we will revise our comparison to focus on the "best" and the "2nd best" models, ensuring greater clarity in the manuscript. Additionally, we will include supplementary simulation results and figures to provide a more comprehensive analysis on models.

eLife Assessment

This valuable study presents a thorough analysis of protein abundance changes caused by amino acid substitutions, using structural context to improve predictive accuracy. By deriving substitution response matrices based on solvent accessibility, the authors demonstrate that simple structural features can predict abundance effects with accuracy comparable to complex methods such as free energy calculations. The strength of the evidence is solid, supported by robust experimental design and comprehensive analyses. This work is expected to be of interest to broad audiences as it offers practical tools for analyzing mutational effects and insights into the structural basis of proteostasis.

Reviewer #1 (Public review):

Of course, there is always another layer of the onion, VAMP-seq measures contributions from isolated thermodynamic stability, stability conferred by binding partners (small molecule and protein), synthesis/degradation balance (especially important in "degron" motifs), etc. Here the authors' goal is to create simple models that can act as a baseline for two main reasons:<br /> (1) how to tell when adding more information would be helpful for a global model;<br /> (2) how to detect when a residue/mutation has an unusual profile indicative of an unbalanced contribution from one of the factors listed above.

As such, the authors state that this manuscript is not intended to be a state-of-the-art method in variant effect prediction, but rather a direction towards considering static structural information for the VAMP-seq effects. At its core, the method is a fairly traditional asymmetric substitution matrix (I was surprised not to see a comparison to BLOSUM in the manuscript) - and shows that a subdivision by burial makes the model much more predictive. Despite only having 6 datasets, they show predictive power even when the matrices are based on a smaller number. Another success is rationalizing the VAMPseq results on relevant oligomeric states.

Specific Feedback:

Major points:

The authors spend a good amount of space discussing how the six datasets have different distributions in abundance scores. After the development of their model is there more to say about why? Is there something that can be leveraged here to design maximally informative experiments?

They compare to one more "sophisticated model" - RosettaddG - which should be more correlated with thermodynamic stability than other factors measured by VAMP-seq. However, the direct head-to-head comparison between their matrices and ddG is underdeveloped. How can this be used to dissect cases where thermodynamics are not contributing to specific substitution patterns OR in specific residues/regions that are predicted by one method better than the other? This would naturally dovetail into whether there is orthogonal information between these two that could be leveraged to create better predictions.

Perhaps beyond the scope of this baseline method, there is also ThermoMPNN and the work from Gabe Rocklin to consider as other approaches that should be more correlated only with thermodynamics.

I find myself drawn to the hints of a larger idea that outliers to this model can be helpful in identifying specific aspects of proteostasis. The discussion of S109 is great in this respect, but I can't help but feel there is more to be mined from Figure S9 or other analyses of outlier higher than predicted abundance along linear or tertiary motifs.

Reviewer #2 (Public review):

Summary:

This study analyzes protein abundance data from six VAMP-seq experiments, comprising over 31,000 single amino acid substitutions, to understand how different amino acids contribute to maintaining cellular protein levels. The authors develop substitution matrices that capture the average effect of amino acid changes on protein abundance in different structural contexts (buried vs. exposed residues). Their key finding is that these simple structure-based matrices can predict mutational effects on abundance with accuracy comparable to more complex physics-based stability calculations (ΔΔG).

Major strengths:

(1) The analysis focuses on a single molecular phenotype (abundance) measured using the same experimental approach (VAMP-seq), avoiding confounding factors present when combining data from different phenotypes (e.g., mixing stability, activity, and fitness data) or different experimental methods.

(2) The demonstration that simple structural features (particularly solvent accessibility) can capture a significant portion of mutational effects on abundance.

(3) The practical utility of the matrices for analyzing protein interfaces and identifying functionally important surface residues.

Major weaknesses:

(1) The statistical rigor of the analysis could be improved. For example, when comparing exposed vs. buried classification of interface residues, or when assessing whether differences between prediction methods are significant.

(2) The mechanistic connection between stability and abundance is assumed rather than explained or investigated. For instance, destabilizing mutations might decrease abundance through protein quality control, but other mechanisms like degron exposure could also be at play.

(3) The similar performance of simple matrix-based and complex physics-based predictions calls for deeper analysis. A systematic comparison of where these approaches agree or differ could illuminate the relationship between stability and abundance. For instance, buried sites showing exposed-like behavior might indicate regions of structural plasticity, while the link between destabilization and degradation might involve partial unfolding exposing typically buried residues. The authors have all the necessary data for such analysis but don't fully exploit this opportunity.

(4) The pooling of data across proteins to construct the matrices needs better justification, given the observed differences in score distributions between proteins (for example, PTEN's distribution is shifted towards high abundance scores while ASPA and PRKN show more binary distributions).

(5) Some key methodological choices require better justification. For example, combining "to" and "from" mutation profiles for PCA despite their different behaviors, or using arbitrary thresholds (like 0.05) for residue classification.

The authors largely achieve their primary aim of showing that simple structural features can predict abundance changes. However, their secondary goal of using the matrices to identify functionally important residues would benefit from more rigorous statistical validation. While the matrices provide a useful baseline for abundance prediction, the paper could offer deeper biological insights by investigating cases where simple structure-based predictions differ from physics-based stability calculations.

This work provides a valuable resource for the protein science community in the form of easily applicable substitution matrices. The finding that such simple features can match more complex calculations is significant for the field. However, the work's impact would be enhanced by a deeper investigation of the mechanistic implications of the observed patterns, particularly in cases where abundance changes appear decoupled from stability effects.

Reviewer #3 (Public review):

"Effects of residue substitutions on the cellular abundance of proteins" by Schulze and Lindorff-Larsen revisits the classical concept of structure-aware protein substitution matrices through the scope of modern protein structure modelling approaches and comprehensive phenotypic readouts from multiplex assays of variant effects (MAVEs). The authors explore 6 unique protein MAVE datasets based on protein abundance (and thus stability) by utilizing structural information, specifically residue solvent accessibility and secondary structure type, to derive combinations of context-specific substitution matrices predicting variant abundance. They are clear to outline that the aim of the study is not to produce a new best abundance predictor but to showcase the degree of prediction afforded simply by utilizing information on residue accessibility. The performance of their matrices is robustly evaluated using a leave-one-out approach, where the abundance effects for a single protein are predicted using the remaining datasets. Using a simple classification of buried and solvent-exposed residues, and substitution matrices derived respectively for each residue group, the authors convincingly demonstrate that taking structural solvent accessibility contexts into account leads to more accurate performance than either a structure-unaware matrix, secondary structure-based matrix, or matrices combining both solvent accessibility or secondary structure. Interestingly, it is shown that the performance of the simple buried and exposed residue substitution matrices for predicting protein abundance is on par with Rosetta, an established and specialized protein variant stability predictor. More importantly, the authors finish off the paper by demonstrating the utility of the two matrices to identify surface residues that have buried-like substitution profiles, that are shown to correspond to protein interface residues, post-translational modification sites, functional residues, or putative degrons.

Strengths:

The paper makes a strong and well-supported main point, demonstrating the utility of the authors' approach through performance comparisons with alternative substitution matrices and specialized methods alike. The matrices are rigorously evaluated without introducing bias, exploring various combinations of protein datasets. Supplemental analyses are extremely comprehensive and detailed. The applicability of the substitution matrices is explored beyond abundance prediction and could have important implications in the future for identifying functionally relevant sites.

Comments:

(1) A wider discussion of the possible reasons why matrices for certain proteins seem to correlate better than others would be extremely interesting, touching upon possible points like differences or similarities in local environments, degradation pathways, post-translation modifications, and regulation. While the initial data structure differences provide a possible explanation, Figure S17A, B correlations show a more complicated picture.

(2) The performance analysis in Figure 2D seems to show that for particular proteins "less is more" when it comes to which datasets are best to derive the matrix from (CYP2C9, ASPA, PRKN). Are there any features (direct or proxy), that would allow to group proteins to maximize accuracy? Do the authors think on top of the buried vs exposed paradigm, another grouping dimension at the protein/domain level could improve performance?

(3) While the matrices and Rosetta seem to show similar degrees of correlation, do the methods both fail and succeed on the same variants? Or do they show a degree of orthogonality and could potentially be synergistic?

Overall, this work presents a valuable contribution by creatively utilizing a simple concept through cutting-edge datasets, which could be useful in various.

eLife Assessment

This manuscript describes work that is fundamental to our understanding of the mechanism of how stress regulates the noradrenergic system and the CRH system. Using a combination of ex vivo physiology and in vitro methods, the study provides compelling evidence as to the signaling mechanisms of how glucocorticoids impact adrenergic GPCRs in CRH cells via receptor trafficking. While the ex vivo work is specific to the hypothalamus, the mechanisms here could be extrapolated and investigated in other brain regions that may have similar signaling regulation.

Reviewer #1 (Public review):

Summary:

In this manuscript, the authors investigate the cellular mechanism underlying suppression of adrenergic effects on excitatory transmission onto hypothalamic CRH neurons by stress. Experiments in ex-vivo slices show that this is a long-lasting effect that occurs through endocytosis of receptors. The authors then move into an immortalized hypothalamic cell line to enable investigation of the mechanism of changes in receptor trafficking. They use a series of immunohistochemistry, FRET, and biochemical experiments to show that application of corticosterone increases targeting of alpha1 adrenergic receptors to the late endosome and lysosome rather than the recycling endosome. Perhaps most interesting, they find that alpha1 receptors and glucocortioid receptors form a complex that is ultimately transferred to the nucleus.

Strengths:

Overall, the studies in this manuscript are rigorous and well-conducted. The data supports their conclusions, and they've shown convincingly that glucocorticoid signaling affects trafficking of alpha1 receptors in the culture system they are using. These findings are important for the field of stress research, both in understanding how two components of the stress system (norepinephrine and HPA axis) interact with each other and in neuromodulatory modulation of hypothalamic CRH neurons. Their finding that alpha1 receptors and glucocorticoid receptors form a complex is particularly interesting and maybe impactful outside of the immediate application in the hypothalamus.

Weaknesses:

The study has two primary weaknesses. First, the majority of the experiments were conducted in an immortalized hypothalamic cell line. This was necessary to conduct the type of experiments needed to test the author's hypothesis, but it remains unclear how closely these cells resemble CRH neurons, or how the same mechanism may be preserved or altered in an intact circuit. Further discussion of these points would strengthen the manuscript.<br /> Second, while experiments are generally well-designed, the authors do not show that the effects of corticosterone can be blocked with a glucocorticoid receptor antagonist. This is fairly standard pharmacology and would strengthen confidence in the findings presented in the study.

Reviewer #2 (Public review):

Summary:

In this manuscript, the authors report novel and exciting findings delineating a non-transcriptional mechanism whereby glucocorticoids desensitize CRH neurons to NE in response to somatic stress. The authors find that this desensitization induced by CORT 1. persists more than 18h, 2. reduces surface expression of AR1bR (NE receptors) by redirecting trafficking from rapid recycling to late endosomal pools and lysosomes, 3. is dependent on NE binding to the AR1bR, 4. involves cellular nitrosylation, 5. involves ubiquitination of beta-arrestin, and 5. involves interactions between glucocorticoid receptors and AR1bs, glucocorticoid receptors and ubiquitinated beta-arrestin, and AR1b and ubiquitinated beta-arrestin. While the authors do not directly provide evidence for a trimeric complex composed of these three proteins, their data that CORT causes translocation of these dimeric complexes to the cell nucleus suggests it is likely. Overall, these results are highly informative for understanding novel mechanisms mediating glucocorticoid regulation of GPCRs.

Strengths:<br /> - Good rationale for each experiment, which describes many parts of the CORT-NE desensitization mechanism<br /> - Great discussion of limitations of the approaches and the parts of the mechanism we do not fully understand yet<br /> - Appropriate approaches for questions being answered<br /> - Describes a highly novel CORT mechanism that non-transcriptionally switches GPCR trafficking dynamics, something that could have far reaching implications for other GPCRs involved in stress responses

Weaknesses:<br /> - Unclear how this mechanism would generalize to other stressor modalities. Restraint stress is a somatic stressor, but can also be considered a psychological stressor (model of depression-like behavior). A purely somatic stressor might increase the robustness of this phenomenon.<br /> - Remains unknown how nitrosylation plays into the mechanism in terms of specific proteins affected by CORT (GRK2, endophilin, clathrin possibilities)

Reviewer #3 (Public review):

Summary:

In this manuscript, Weiss et al describe a mechanism through which glucocorticoids desensitize CRH neurons in the PVN to norepinephrine. This follows on from previous work from this lab showing rapid glucocorticoid suppression of adrenergic signaling in CRH neurons specific to somatic stress activation, and modality-selective glucocorticoid negative feedback.

Specifically, their previous work shows that:<br /> (1) NE increases glutamate drive to CRH neurons<br /> (2) CORT blunts the effects of NE through a dynamin-dependent mechanism<br /> (3) This contributes to loss of NE signalling after stress (specifically when the second stressor is a physiological one)

Here they extend this line of interrogation by showing that CORT redistributes Ara1b receptors from rapid recycling endosomes to late endosomes and lysosomes. They show a time window of CORT actions and provide additional mechanistic details implicating nitric oxide-dependent nitrosylation in receptor trafficking.

Strengths:

Builds on existing work to provide additional mechanistic details.<br /> The experiments are well done and data are compelling.<br /> The link to nitrosylation is novel (but see below)

Weaknesses:

(1) The link to nitrosylation is interesting, but a bit confusing. If I understand correctly, inhibiting the production of NO or using NEM increases receptor internalization, suggesting that NO-dependent nitrosylation prevents ligand-dependent internalization. What is unclear to me is how CORT is linked to this step. I note the authors show a decrease in nitrosylation with CORT. So, does CORT decrease the activity of NOS and, thus, the production of NO? If so, then exogenously activating this system in the presence of CORT should result in a recovery of NE-dependent increase in glutamate release. Or is the GCR directly decreasing nitrosylation? Linking these elements is critical in terms of furthering our mechanistic understanding of this process.

(2) It's not clear why/how blockade of Ara1 after CORT-induced cytosolic accumulation results in a reversal of effect. Unless I misunderstood something, this requires further explanation.

Author response:

We appreciate the expression of enthusiasm for our paper by the editors and the three reviewers and the suggestions on how to improve the study. Here we outline how we will address the reviewers’ concerns and suggestions in a planned revision of our manuscript.

Reviewer #1 listed two primary weaknesses:

(1) the need for discussion of the extent to which the cell line we used resembles CRH neurons and

(2) that we did not test for the effect of blockade of the glucocorticoid receptor.

(1) As the reviewer acknowledges, our experiments called for the use of a cell line to dissect intracellular trafficking of the α1 adrenoreceptor. We selected the N42 cell line for this purpose because it is an immortalized hypothalamic cell line (developed by Belsham and colleagues, Belsham et al., 2004) that expresses CRH. We have used this cell line successfully in the past to study transcriptional and rapid non-genomic actions of glucocorticoids, which indicated that, in addition to expressing CRH, these cells also express both the nuclear glucocorticoid receptor and a membrane-associated receptor that binds glucocorticoids (Rainville et al., 2019; Weiss et al., 2019). We believe that this hypothalamic cell line is the most closely related to native PVN CRH neurons of any cell line available. As requested, we will add to the Discussion of the manuscript to further justify our choice of cells.

(2) We agree that this experiment should be performed. We will test the classical GR (and progesterone) antagonist RU486 (mifepristone) for its effect on the cort regulation of α1 adrenoreceptor trafficking. Our ex vivo electrophysiology studies have indicated that the rapid glucocorticoid effect in native hypothalamic CRH neurons is not blocked by RU486 and is not, therefore, dependent on activation of the classical nuclear GR (Di et al., 2003; Di et al., 2016).

Reviewer #2 also listed two main weaknesses of the study:

(1) that we did not test whether the adrenoreceptor desensitization by restraint stress generalizes to other stress modalities and might be more robust with a pure somatic stressor, and

(2) the lack of identification of a target protein as a mechanism for the role of nitrosylation.

(1) We used restraint stress as a means to elicit corticosterone release, which desensitized the HPA response to a NE-dependent somatic stressor (lipopolysaccharide injection) but not to a NEindependent psychological stressor (predator odor) (Jiang et al., 2021). We got a near-complete loss of the sensitivity of CRH neurons to NE with restraint (i.e., near ceiling effect), such that a different stressor, including a more purely somatic stressor, should not increase the Cort-induced desensitization further. For that reason, we would argue that testing other stressors would not add value to the current study. That said, we plan and have received new funding to test in the future whether the Cort desensitization of the HPA response to LPS stress generalizes to other somatic stressors. We also have future plans to test for the Cort desensitization of other Gq-coupled receptors.

(2) We agree that finding the molecular target of nitrosylation as the mechanism for Cort desensitization of α1 adrenoreceptors would significant improve the study, but this is a potentially enormous undertaking as it will require the screening and validation of multiple proteins involved in protein trafficking to find the one(s) targeted for nitrosylation by Cort. We tested β-arrestin as a possible target in the paper, but did not find Cort to regulate β-arrestin nitrosylation. We plan to undertake a general nitrosylation screen of proteins to identify multiple possible targets, but prefer to defer this and the validation of possible targets to a future, more thorough analysis.

Reviewer #3 also pointed out two main weaknesses of our study:

(1) that the glucocorticoidnitrosylation link was confusing, and

(2) that it was unclear how blocking α1 adrenoreceptors reversed the Cort-induced cytosolic accumulation of the receptor.

We appreciate the reviewer pointing out these deficiencies in our interpretation and explanation of our findings. We plan to address them directly in the revised version of the paper. 

References

Belsham DD, Cai F, Cui H, Smukler SR, Salapatek AMF, Shkreta L (2004) Generation of a phenotypic array of hypothalamic neuronal cell models to study complex neuroendocrine disorders. Endocrinology 145:393–400.

Weiss GL, Rainville JR, Zhao Q, Tasker JG (2019) Purity and stability of the membrane-limited glucocorticoid receptor agonist dexamethasone-BSA. Steroids 142:2-5. 

Rainville JR, Weiss GL, Evanson N, Herman JP, Vasudevan N, Tasker JG (2019) Membrane-initiated nuclear trafficking of the glucocorticoid receptor in hypothalamic neurons. Steroids 142:55-64.

Di S, Malcher-Lopes R, Halmos KCs, Tasker JG (2003) Non-genomic glucocorticoid inhibition via endocannabinoid release in the hypothalamus: a fast feedback mechanism. Journal of Neuroscience 23:4850-4857.

Di S, Itoga CA, Fisher MO, Solomonow J, Roltsch EA, Gilpin NW, Tasker JG (2016) Acute stress suppresses inhibition and increases anxiety via endocannabinoid release in the basolateral amygdala. Journal of Neuroscience 36:8461-8470.

Jiang Z, Chen C, Weiss GL, Fu X, Stelly CE, Sweeten BLW, Tirrell PS, Pursell I, Stevens CR, Fisher MO, Begley JC, Harrison LM, Tasker JG (2022) Stress-induced glucocorticoid desensitizes adrenoreceptors to gate the neuroendocrine response to somatic stress in male mice. Cell Reports 41(3):111509.

eLife assessment

This study reports analysis of the formation of electrosensory ampullary organs in non-model organisms, the sterlet sturgeon. By using a combination of targeted gene knock-out and inhibition, the study provides overall convincing evidence for differential roles of BMP signaling in lateral-line development, with few aspects that could be improved. The study is particularly valuable for understanding the development of a still-mysterious sensory system, and for its evolutionary implications.

Reviewer #1 (Public Review):

The authors were curious about the formation of the electrosensory lateral line, which is found in non-traditional model organisms. This issue has traditionally hampered studies because those organisms are not amenable to controlled experimental work.

The authors skillfully use CRIPR-based technologies to overcome this limitation. Together with exceptionally good whole-mount in situ hybridisation, they produced a well-supported conclusion that Bmp signalling has different roles in the development of electrosensory ampullary organs.

I would not entirely agree that Bmp signalling has "opposing" roles because the authors do not show evidence of opposition via gain-of-function experiments at different developmental times. Instead, they are simply different at different periods of organogenesis.

The study is important for understanding the development of a still-mysterious sensory system, and for its implications in evolutionary biology more generally.