french is present on 5 continents (+ english)

I think this sentence is useful to contextualize just how important and useful the French language can be.

I think this is a strong way to end the article since it emphasizes just how wide spread the French language is.

I find it interesting how Quebec is now listed here but Canada is not on the list at all. This shows how densely spoken French is in Quebec.

I find it interesting how French is not in the top 5 most spoken languages but is more wide spread than almost all of them.

I think its interesting that french is all around the globe and its made an impression in cultures everywhere

What the author is trying to say by saying "Galaxie francophone" is that french is all around the world and isnt going to slow down any time

This graph is saying that french is growing very rapidly

French is spoken in all sorts of different social occasions

French is spoken on 5 different continents

cognate

Reviewer #2 (Public review):

Summary:

This manuscript proposes that the use of a latent cause model for assessment of memory-based tasks may provide improved early detection in Alzheimer's Disease as well as more differentiated mapping of behavior to underlying causes. To test the validity of this model, the authors use a previously described knock-in mouse model of AD and subject the mice to several behaviors to determine whether the latent cause model may provide informative predictions regarding changes in the observed behaviors. They include a well-established fear learning paradigm in which distinct memories are believed to compete for control of behavior. More specifically, it's been observed that animals undergoing fear learning and subsequent fear extinction develop two separate memories for the acquisition phase and the extinction phase, such that the extinction does not simply 'erase' the previously acquired memory. Many models of learning require the addition of a separate context or state to be added during the extinction phase and are typically modeled by assuming the existence of a new state at the time of extinction. The Niv research group, Gershman et al. 2017, have shown that the use of a latent cause model applied to this behavior can elegantly predict the formation of latent states based on a Bayesian approach, and that these latent states can facilitate the persistence of the acquisition and extinction memory independently. The authors of this manuscript leverage this approach to test whether deficits in production of the internal states, or the inference and learning of those states, may be disrupted in knock-in mice that show both a build-up of amyloid-beta plaques and a deterioration in memory as the mice age.

Strengths:

I think the authors' proposal to leverage the latent cause model and test whether it can lead to improved assessments in an animal model of AD is a promising approach for bridging the gap between clinical and basic research. The authors use a promising mouse model and apply this to a paradigm in which the behavior and neurobiology are relatively well understood - an ideal situation for assessing how a disease state may impact both the neurobiology and behavior. The latent cause model has the potential to better connect observed behavior to underlying causes and may pave a road for improved mapping of changes in behavior to neurobiological mechanisms in diseases such as AD.<br /> The authors also compare the latent cause model to the Rescorla-Wagner model and a latent state model allowing for better assessment of the latent cause model as a strong model for assessing reinstatement.

Weaknesses:

I have several substantial concerns which I've detailed below. These include important details on how the behavior was analyzed, how the model was used to assess the behavior, and the interpretations that have been made based on the model.<br /> (1) There is substantial data to suggest that during fear learning in mice separate memories develop for the acquisition and extinction phases, with the acquisition memory becoming more strongly retrieved during spontaneous recovery and reinstatement. The Gershman paper, cited by the authors, shows how the latent causal model can predict this shift in latent causes by allowing for the priors to decay over time, thereby increasing the posterior of the acquisition memory at the time of spontaneous recovery. In this manuscript, the authors suggest a similar mechanism of action for reinstatement, yet the model does not appear to return to the acquisition memory after reinstatement, at least based on the simulation and examples shown in figures 1 and 3. More specifically, in figure 1, the authors indicate that the posterior probability of the latent cause, z_A (the putative acquisition memory), increases, partially leading to reinstatement. This does not appear to be the case as test 3 (day 36) appears to have similar posterior probabilities for z_A as well as similar weights for the CS as compared to the last days of extinction. Rather, the model appears to mainly modify the weights in the most recent latent cause, z_B - the putative the 'extinction state', during reinstatement. The authors suggest that previous experimental data have indicated that spontaneous recovery or reinstatement effects are due to an interaction of the acquisition and extinction memory. These studies have shown that conditioned responding at a later time point after extinction is likely due to a balance between the acquisition memory and the extinction memory, and that this balance can shift towards the acquisition memory naturally during spontaneous recovery, or through artificial activation of the acquisition memory or inhibition of the extinction memory (see Lacagnina et al. for example). Here the authors show that the same latent cause learned during extinction, z_B, appears to dominate during the learning phase of reinstatement, with rapid learning to the context - the weight for the context goes up substantially on day 35 - in z_B. This latent cause, z_B, dominates at the reinstatement test, and due to the increased associative strength between the context and shock, there is a strong CR. For the simulation shown in figure 1, it's not clear why a latent cause model is necessary for this behavior. This leads to the next point.

(2) The authors compared the latent cause model to the Rescorla-Wagner model. This is very commendable, particularly since the latent cause model builds upon the RW model, so it can serve as an ideal test for whether a more simplified model can adequately predict the behavior. The authors show that the RW model cannot successfully predict the increased CR during reinstatement (Appendix figure 1). Yet there are some issues with the way the authors have implemented this comparison:<br /> (2A) The RW model is a simplified version of the latent cause model and so should be treated as a nested model when testing, or at a minimum, the number of parameters should be taken into account when comparing the models using a method such as the Bayesian Information Criterion, BIC.<br /> (2B) The RW model provides the associative strength between stimuli and does not necessarily require a linear relationship between V and the CR. This is the case in the original RW model as well as in the LCM. To allow for better comparison between the models, the authors should be modeling the CR in the same manner (using the same probit function) in both models. In fact, there are many instances in which a sigmoid has been applied to RW associative strengths to predict CRs. I would recommend modeling CRs in the RW as if there is just one latent cause. Or perhaps run the analysis for the LCM with just one latent cause - this would effectively reduce the LCM to RW and keep any other assumptions identical across the models.<br /> (2C) In the paper, the model fits for the alphas in the RW model are the same across the groups. Were the alphas for the two models kept as free variables? This is an important question as it gets back to the first point raised. Because the modeling of the reinstatement behavior with the LCM appears to be mainly driven by latent cause z_B, the extinction memory, it may be possible to replicate the pattern of results without requiring a latent cause model. For example, the 12-month-old App NL-G-F mice behavior may have a deficit in learning about the context. Within the RW model, if the alpha for context is set to zero for those mice, but kept higher for the other groups, say alpha_context = 0.8, the authors could potentially observe the same pattern of discrimination indices in figure 2G and 2H at test. Because the authors don't explicitly state which parameters might be driving the change in the DI, the authors should show in some way that their results cannot simply be due to poor contextual learning in the 12 month old App NL-G-F mice, as this can presumably be predicted by the RW model. The authors' model fits using RW don't show this, but this is because they don't consider this possibility that the alpha for context might be disrupted in the 12-month-old App NL-G-F mice. Of course, using the RW model with these alphas won't lead to as nice of fits of the behavior across acquisition, extinction, and reinstatement as the authors' LCM, the number of parameters are substantially reduced in the RW model. Yet the important pattern of the DI would be replicated with the RW model (if I'm not mistaken), which is the important test for assessment of reinstatement.

(3) As stated by the authors in the introduction, the advantage of the fear learning approach is that the memory is modified across the acquisition-extinction-reinstatement phases. Although perhaps not explicitly stated by the authors, the post-reinstatement test (test 3) is the crucial test for whether there is reactivation of a previously stored memory, with the general argument being that the reinvigorated response to the CS can't simply be explained by relearning the CS-US pairing, because re-exposure the US alone leads to increase response to the CS at test. Of course there are several explanations for why this may occur, particularly when also considering the context as a stimulus. This is what I understood to be the justification for the use of a model, such as the latent cause model, that may better capture and compare these possibilities within a single framework. As such, it is critical to look at the level of responding to both the context alone and to the CS. It appears that the authors only look at the percent freezing during the CS, and it is not clear whether this is due to the contextual-US learning during the US re-exposure or to increased responding to the CS - presumably caused by reactivation of the acquisition memory. The authors do perform a comparison between the preCS and CS period, but it is not clear whether this is taken into account in the LCM. For example, the instance of the model shown in figure 1 indicates that the 'extinction cause', or cause z6, develops a strong weight for the context during the reinstatement phase of presenting the shock alone. This state then leads to increased freezing during the final CS probe test as shown in the figure. If they haven't already, I think the authors must somehow incorporate these different phases (CS vs ITI) into their model, particularly since this type of memory retrieval that depends on assessing latent states is specifically why the authors justified using the latent causal model. In more precise terms, it's not clear whether the authors incorporate a preCS/ITI period each day the cue is presented as a vector of just the context in addition to the CS period in which the vector contains both the context and the CS. Based on the description, it seemed to me that they only model the CRs during the CS period on days when the CS is presented, and thereby the context is only ever modeled on its own (as just the context by itself in the vector) on extinction days when the CS is not presented. If they are modeling both timepoints each day that the CS I presented, then I would recommend explicitly stating this in the methods section.

(4) The authors fit the model using all data points across acquisition and learning. As one of the other reviewers has highlighted, it appears that there is a high chance for overfitting the data with the LCM. Of course, this would result in much better fits than models with substantially fewer free parameters, such as the RW model. As mentioned above, the authors should use a method that takes into account the number of parameters, such as the BIC.

(5) The authors have stated that they do not think the Barnes maze task can be modeled with the LCM. Whether or not this is the case, if the authors do not model this data with the LCM, the Barnes maze data doesn't appear valuable to the main hypothesis. The authors suggest that more sophisticated models such as the LCM may be beneficial for early detection of diseases such as Alzheimer's, so the Barnes maze data is not valuable for providing evidence of this hypothesis. Rather, the authors make an argument that the memory deficits in the Barnes maze mimic the reinstatement effects providing support that memory is disrupted similarly in these mice. Although, the authors state that the deficits in memory retrieval are similar across the two tasks, the authors are not explicit as to the precise deficits in memory retrieval in the reinstatement task - it's a combination of overgeneralizing latent causes during acquisition, poor learning rate, over differentiation of the stimuli.

ccording to all known laws of aviation, there is no way a bee should be able to fly. Its wings are too small to get its fat little body off the ground. The bee, of course, flies anyway because bees don't care what humans think is impossible. Yellow, black. Yellow, black. Yellow, black. Yellow, black. Ooh, black and yellow! Let's shake it up a little. Barry! Breakfast is ready! Coming! Hang on a second. Hello? - Barry? - Adam? - Can you believe this is happening? - I can't. I'll pick you up. Looking sharp. Use the stairs. Your father paid good money for those. Sorry. I'm excited. Here's the graduate. We're very proud of you, son. A perfect report card, all B's. Very proud. Ma! I got a thing going here. - You got lint on your fuzz. - Ow! That's me! - Wave to us! We'll be in row 118,000. - Bye! Barry, I told you, stop flying in the house! - Hey, Adam. - Hey, Barry. - Is that fuzz gel? - A little. Special day, graduation. Never thought I'd make it. Three days grade school, three days high school. Those were awkward. Three days college. I'm glad I took a day and hitchhiked around the hive. You did come back different. - Hi, Barry. - Artie, growing a mustache? Looks good. - Hear about Frankie? - Yeah. - You going to the funeral? - No, I'm not going. Everybody knows, sting someone, you die. Don't waste it on a squirrel. Such a hothead. I guess he could have just gotten out of the way. I love this incorporating an amusement park into our day. That's why we don't need vacations. Boy, quite a bit of pomp... under the circumstances. - Well, Adam, today we are men. - We are! - Bee-men. - Amen! Hallelujah! Students, faculty, distinguished bees, please welcome Dean Buzzwell. Welcome, New Hive City graduating class of... ...9:15. That concludes our ceremonies. And begins your career at Honex Industries! Will we pick our job today? I heard it's just orientation. Heads up! Here we go. Keep your hands and antennas inside the tram at all times. - Wonder what it'll be like? - A little scary. Welcome to Honex, a division of Honesco and a part of the Hexagon Group. This is it! Wow. Wow. We know that you, as a bee, have worked your whole life to get to the point where you can work for your whole life. Honey begins when our valiant Pollen Jocks bring the nectar to the hive. Our top-secret formula is automatically color-corrected, scent-adjusted and bubble-contoured into this soothing sweet syrup with its distinctive golden glow you know as... Honey! - That girl was hot. - She's my cousin! - She is? - Yes, we're all cousins. - Right. You're right. - At Honex, we constantly strive to improve every aspect of bee existence. These bees are stress-testing a new helmet technology. - What do you think he makes? - Not enough. Here we have our latest advancement, the Krelman. - What does that do? - Catches that little strand of honey that hangs after you pour it. Saves us millions. Can anyone work on the Krelman? Of course. Most bee jobs are small ones. But bees know that every small job, if it's done well, means a lot. But choose carefully because you'll stay in the job you pick for the rest of your life. The same job the rest of your life? I didn't know that. What's the difference? You'll be happy to know that bees, as a species, haven't had one day off in 27 million years. So you'll just work us to death? We'll sure try. Wow! That blew my mind! "What's the difference?" How can you say that? One job forever? That's an insane choice to have to make. I'm relieved. Now we only have to make one decision in life. But, Adam, how could they never have told us that? Why would you question anything? We're bees. We're the most perfectly functioning society on Earth. You ever think maybe things work a little too well here? Like what? Give me one example. I don't know. But you know what I'm talking about. Please clear the gate. Royal Nectar Force on approach. Wait a second. Check it out. - Hey, those are Pollen Jocks! - Wow. I've never seen them this close. They know what it's like outside the hive. Yeah, but some don't come back. - Hey, Jocks! - Hi, Jocks! You guys did great! You're monsters! You're sky freaks! I love it! I love it! - I wonder where they were. - I don't know. Their day's not planned. Outside the hive, flying who knows where, doing who knows what. You can't just decide to be a Pollen Jock. You have to be bred for that. Right. Look. That's more pollen than you and I will see in a lifetime. It's just a status symbol. Bees make too much of it. Perhaps. Unless you're wearing it and the ladies see you wearing it.

Author response:

The following is the authors’ response to the original reviews

Reviewer #1 (Public Review):

Summary:

In this study, López-Jiménez and colleagues demonstrated the utility of using high-content microscopy in dissecting host and bacterial determinants that play a role in the establishment of infection using Shigella flexneri as a model. The manuscript nicely identifies that infection with Shigella results in a block to DNA replication and protein synthesis. At the same time, the host responds, in part, via the entrapment of Shigella in septin cages.

Strengths:

The main strength of this manuscript is its technical aspects. They nicely demonstrate how an automated microscopy pipeline coupled with artificial intelligence can be used to gain new insights regarding elements of bacterial pathogenesis, using Shigella flexneri as a model system. Using this pipeline enabled the investigators to enhance the field's general understanding regarding the role of septin cages in responding to invading Shigella. This platform should be of interest to those who study a variety of intracellular microbial pathogens.

Another strength of the manuscript is the demonstration - using cell biology-based approaches- that infection with Shigella blocks DNA replication and protein synthesis. These observations nicely dovetail with the prior findings of other groups. Nevertheless, their clever click-chemistry-based approaches provide visual evidence of these phenomena and should interest many.

We thank the Reviewer for their enthusiasm on technical aspects of this paper, regarding both the automated microscopy pipeline coupled with artificial intelligence and the click-chemistry based approaches to dissect DNA replication and protein synthesis by microscopy.

Weaknesses:

There are two main weaknesses of this work. First, the studies are limited to findings obtained using a single immortalized cell line. It is appreciated that HeLa cells serve as an excellent model for studying aspects of Shigella pathogenesis and host responses. However, it would be nice to see that similar observations are observed with an epithelial cell line of intestinal, preferably colonic origin, and eventually, with a non-immortalized cell line, although it is appreciated that the latter studies are beyond the scope of this work.

The immortalized cell line HeLa is widely regarded as a paradigm to study infection by Shigella and other intracellular pathogens. However, we agree that future studies beyond the scope of this work should include other cell lines (eg. epithelial cells of colonic origin, macrophages, primary cells). 

The other weakness is that the studies are minimally mechanistic. For example, the investigators have data to suggest that infection with Shigella leads to an arrest in DNA replication and protein synthesis; however, no follow-up studies have been conducted to determine how these host cell processes are disabled. Interestingly, Zhang and colleagues recently identified that the Shigella OspC effectors target eukaryotic translation initiation factor 3 to block host cell translation (PMID: 38368608). This paper should be discussed and cited in the discussion.

We appreciate the Reviewer’s concern about the lack of follow up work on observations of host DNA and protein synthesis arrest upon Shigella infection, which will be the focus of future studies. We acknowledge the recent work of Zhang et al. (Cell Reports, 2024) considering their similar results on protein translation arrest, and this reference has been more fully discussed in the revised version of the manuscript.

Reviewer #2 (Public Review):

Summary:

Septin caging has emerged as one of the innate immune responses of eukaryotic cells to infections by intracellular bacteria. This fascinating assembly of eukaryotic proteins into complex structures restricts bacteria motility within the cytoplasm of host cells, thereby facilitating recognition by cytosolic sensors and components of the autophagy machinery. Given the different types of septin caging that have been described thus far, a single-cell, unbiased approach to quantify and characterise septin recruitment at bacteria is important to fully grasp the role and function of caging. Thus, the authors have developed an automated image analysis pipeline allowing bacterial segmentation and classification of septin cages that will be very useful in the future, applied to study the role of host and bacterial factors, compare different bacterial strains, or even compare infections by clinical isolates.

Strengths:

The authors developed a solid pipeline that has been thoroughly validated. When tested on infected cells, automated analysis corroborated previous observations and allowed the unbiased quantification of the different types of septin cages as well as the correlation between caging and bacterial metabolic activity. This approach will prove an essential asset in the further characterisation of septin cages for future studies.

We thank the Reviewer for their positive comments, and for highlighting the strength of our imaging and analysis pipeline to analyse Shigella-septin interactions.

Weaknesses:

As the main aim of the manuscript is to describe the newly developed analysis pipeline, the results illustrated in the manuscript are essentially descriptive. The developed pipeline seems exceptionally efficient in recognising septin cages in infected cells but its application for a broader purpose or field of study remains limited.

The main objective of this manuscript is the development of imaging and analysis tools to study Shigella infection, and in particular, Shigella interactions with the septin cytoskeleton. In future work we will provide more mechanistic insight with novel experiments and broader applicability, using different cell lines (in agreement with Reviewer 1), mutants or clinical isolates of Shigella and different bacteria species (eg. Listeria, Salmonella, mycobacteria).

Reviewer #3 (Public Review):

Summary:

The manuscript uses high-content imaging and advanced image-analysis tools to monitor the infection of epithelial cells by Shigella. They perform some analysis on the state of the cells (through measurements of DNA and protein synthesis), and then they focus on differential recruitment of Sept7 to the bacteria. They link this recruitment with the activity of the bacterial T3SS, which is a very interesting discovery. Overall, I found numerous exciting elements in this manuscript, and I have a couple of reservations. Please see below for more details on my reservations. Nevertheless, I think that these issues can be addressed by the authors, and doing so will help to make it a convincing and interesting piece for the community working on intracellular pathogens. The authors should also carefully re-edit their manuscript to avoid overselling their data (see below for issues I see there). I would consider taking out the first figure and starting with Figure 3 (Figure 2 could be re-organized in the later parts)- that could help to make the flow of the manuscript better.

Strengths:

The high-content analysis including the innovative analytical workflows are very promising and could be used by a large number of scientists working on intracellular bacteria. The finding that Septins (through SEPT7) are differentially regulated through actively secreting bacteria is very exciting and can steer novel research directions.

We thank the Reviewer for their constructive feedback and excitement for our results, including our findings on T3SS activity and Shigella-septin interactions. In accordance with the Reviewer’s comments, we avoid overselling our data in the revised version of the manuscript.

Weaknesses:

The manuscript makes a connection between two research lines (1: Shigella infection and DNA/protein synthesis, 2: regulation of septins around invading Shigella) that are not fully developed - this makes it sometimes difficult to understand the take-home messages of the authors.

We agree that the manuscript is mostly technical and therefore some of our experimental observations would benefit from follow up mechanistic studies in the future. We highlight our vision for broader applicability in response to weaknesses raised by Reviewer 2.

It is not clear whether the analysis that was done on projected images actually reflects the phenotypes of the original 3D data. This issue needs to be carefully addressed.

We agree with the Reviewer that characterizing 3D data using 2D projected images has limitations.

We observe an increase in cell and nuclear surface that does not strictly imply a change in volume. This is why we measure Hoechst intensity in the nucleus using SUM-projection (as it can be used as a proxy of DNA content of the cell). However, we agree that future use of other markers (such as fluorescently labelled histones) would make our conclusions more robust.

Regarding the different orientation of intracellular bacteria, we agree that investigation of septin recruitment is more challenging when bacteria are placed perpendicular to the acquisition plane. In a first step, we trained a Convolutional Neural Network (CNN) using 2D data, as it is easier/faster to train and requires fewer annotated images. In doing so, we already managed to correctly identify 80% of Shigella interacting with septins, which enabled us to observe higher T3SS activity in this population. In future studies, we will maximize the 3D potential of our data and retrain a CNN that will allow more precise identification of Shigella-septin interactions and in depth characterization of volumetric parameters.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

(1) To conclude that cell volume is indeed increased, the investigators should consider staining the cells with markers that demarcate cell boundaries and/or are confined to the cytosol, i.e., a cell tracker dye.

Staining using our SEPT7 antibody enables us to define cell boundaries for cellular area measurements (Novel Figure 1 - figure supplement 1A). However, we agree with the Reviewer that staining cells with additional markers (such as a cell tracker dye) would be required to conclude that cell volume is increased. We therefore adjust our claims in the main text (lines 107-115 and 235-246).

(2) Line 27: I understand what is meant by "recruited to actively pathogenic bacteria with increased T3SS activation." However, one could argue that there are many different roles of the intracytosolic bacteria in pathogenesis in terms of pathogenesis, not just actively secreting effectors.

T3SS secretion by cytosolic bacteria is tightly regulated and both T3SS states (active, inactive) likely contribute to the pathogenic lifestyle of S. flexneri. In agreement with this, we removed this statement from the manuscript (lines 27, 225 and 274).

(3) Line 88: Please clarify in the text that HeLa cells are being studied.

We explicitly mention that the epithelial cell line we study is HeLa in the main text (line 93), in addition to the Materials and methods (line 328).

(4) Line 97: is it possible to quantify the average distance of the nuclei from the cell perimeter? This would help provide some context as to what it means to be a certain distance from the nucleus, i.e., is there another way to point out that distance from nuclei correlates with movement inward post-invasion at the periphery?

To provide more context to the inward movement of bacteria to the cell centre, we provide calculations based on measurements in Figure 1G, I. If we approximate geometric shape of both cells and nucleus to a circle, the median radius of a HeLa cell is 31.1 µm² (uninfected cell) and 36.3 µm² (infected cell). Similarly, the median radius of the nucleus is 22.2 µm² (uninfected cell) and 24.57 µm² (infected cell).

However, we note that Figure 1F shows distance of bacteria to the centroid of the cell, which is the geometric centre of the cell, and which does not necessarily coincide with the geometric centre of the nucleus. We also note that nuclear area increases with infection (in a bacterial dose dependent manner). Finally, we note that these measurements are performed on max projections of 3D Z-stacks. In this case we cannot fully appreciate distance to the nucleus for bacteria located above it.

(5) Lines 212-213 - there is no Figure 9A, B - I think this should be Figure 7A, B.

Text has been updated (lines 216-217).

Reviewer #2 (Recommendations For The Authors):

Testing the analysis pipeline as a proof-of-concept question such as the comparison of caging around the laboratory strain as compared to one or a few clinical isolates or mutants of interest would help stress the relevance of this new, remarkable tool.

We thank the Reviewer for their enthusiasm.

Future research in the Mostowy lab will capitalise on the high-content tools generated here to explore the frequency and heterogeneity of septin cage entrapment for a wide variety of S. flexneri mutants and Shigella clinical isolates.

The sentence in line 215 ends with "in agreement with" followed by a reference.

Text has been updated (line 219).

The sentence in line 217 on the correlation between caging and T3SS is not very clear.

Text has been clarified (lines 221-223).

There is a typo in line 219 : "protrusSions"

Text has been updated (line 223).

Reviewer #3 (Recommendations For The Authors):

Major points

The quantitative analysis approach in Figure 1 has multiple issues. Some examples:<br /> (1) How was the cell area estimated? Normally, a marker for the whole cell (CellMask or similar) or cells expressing GFP would be good indicators. Here it is not clear to me what was done.

The cell area was estimated using SEPT7 antibody staining which is enriched under the cell cortex. CellProfiler was used to segment cells based on SEPT7 staining, using a propagation method from the identified nucleus based on Otsu thresholding. To provide more clarity on how this was performed, we now include a new figure (Figure 1- figure supplement 1A) showing a representative image of HeLa cells stained with SEPT7 and the corresponding cell segmentation performed with CellProfiler software, together with an updated figure legend explaining the procedure (lines 784–787).

(2) The authors use Hoechst and integrated z-projections (Figure 1 S1) as a proxy to estimate nuclear volume. Hoechst staining depends on the organization of the DNA within the nucleus and I find that the authors need to do better controls to estimate nuclear size - this would be possible with cells expressing fluorescently labeled histones, or even better with a fluorescently tagged nuclear pore/envelope marker. The current quantification approach is misleading.

We understand Reviewer #3’s concerns about using Hoechst staining as a proxy of nuclear volume, due to potential differences in DNA organisation within the nucleus.

Following the recommendation of Reviewer #3 in the following point 3, text has been updated (lines 107–115 and 235-246).

(3) Was cell density assessed for the measurements? If cells are confluent, bacteria could spread between cells within 3 hrs, if cells are less dense, this does not occur. When epithelial cells are infected for some hours, they have the tendency to round up a bit (and to appear thicker in z), but a bit smaller in xy. My suggestion to the authors (as they use these findings to follow up with experiments on the underlying processes) would be to tone down their statements - eg, Hoechst staining could be simply indicated as altered, but not put in a context of size (this would require substantial control experiments).

Local cell density was not directly measured, but the experiment was set up to infect at roughly 80% confluency (cells were seeded at 10⁴ cells/well 2 days prior to infection in a 96-well microplate, as described in the Materials and methods section) and to ensure bacterial spread between cells.

In agreement with Reviewer #3 we tone down statements in the main text (see response to point 2 above).

In addition, I found Figure 1 (and parts of Figure 2) disconnected from the rest of the manuscript, and it may even be an idea to take it out of the manuscript (that could also help to deal with my feedback relating to Figure 1). I would suggest starting the manuscript with the current Figure 3 and building the biological story with a stronger focus on SEPT7 (and its links with T3 secretion and actively pathogenic bacteria) from there on. As it stands, the two parts of the manuscript are not well connected.

We carefully considered this comment but following revisions we have not reorganised the manuscript. We believe that high-content characterisation of S. flexneri infection in Figure 1 and 2 provides insightful information about changes in host cells in response to infection. Following this, we move onto characterising intracellular bacteria (and in particular those entrapped in septin cages) in the second part of the manuscript (Figure 3-7). Similar methods were used to analyse both host and bacterial cells and results obtained offer complementary views on host-pathogen interactions.

My major reservation with the experimental work of the current version of the manuscript relates to Figure 5: The analysis of the septin phenotypes in Figure 5 seems to be problematic - to me, it appears that analysis and training were done on projected image stacks. As bacteria are rod-shaped their orientation in space has an enormous impact on how the septin signal appears in a projection - this can lead to wrong interpretation of the phenotypes. The authors need to do some quantitative controls analyzing their data in 3D. To be more clear: the example "tight" (second row) shows a bacterium that appears short. It may be that it's actually longer if one looks in 3D, and the septin signal could possibly fall in the category "rings" or even "two poles".

The deep learning training and subsequent analysis of septin-cage entrapment is done on projected Z-stacks, which presents limitations. Future work in the Mostowy lab will exploit this first study and dive deeper into 3D aspects of the data.

To address Reviewer #3’s concern, we include a sentence explaining that this analysis was performed using 2D max projections (lines 708 and 724), as well as acknowledging its limitations in the main text (lines 259-262).

Minor points

The scale bar in Fig 1 is very thin.

We corrected the scale bar in Fig. 1 to make it more visible.

Could it be that Figure 1F is swapped with Figure1E in the description?

Descriptions for Figure 1E and F are correct.

Line 27: what does "actively pathogenic bacteria" mean? I propose to change the term.

We agree with Reviewer #3 that “actively pathogenic bacteria” should be removed from the text. This update is also in agreement with Reviewer #1 (see Reviewer #1 point 2).

Line 28: "dynamics" can be confusing as it relates to dynamic events imaged by time-lapse.

Although we are making a snapshot of the infection process at 3 hpi, we capture asynchronous processes in both host and bacterial cells (eg. host cells infected with different bacterial loads, bacterial cells undergoing actin polymerisation or septin cage entrapment). We agree that we are not following dynamics of full events over time. However, our high content approach enables us to capture different stages of dynamic processes. To avoid confusion, we replace “dynamics” by “diverse interactions” (line 28), and we discuss the importance of follow-up studies studying microscopy timelapses (line 274).

Paragraph 59 following: the concept of heterogeneity was investigated in some detail for viral infection by the Pelkmans group (PMID: 19710653) using advanced image analysis tools. Advanced machine-learning-based analysis was then performed on Salmonella invasion by Voznica and colleagues (PMID: 29084895). It would be great to include these somewhat "old" works here as they really paved the way for high-content imaging, and the way analyses were performed then should be also discussed in light of how analyses can be performed now with the approaches developed by the authors.

We agree. These landmark studies have now been included in the main text (lines 71-74).

Line 181: I do not know what "morphological conformations" means, perhaps the authors can change the wording or clarify.

We substituted the phrase “morphological conformations” by “morphological patterns” to improve clarity in the main text (lines 185).

The authors claim (eg in the abstract) that they are measuring the dynamic infection process. To me, it appears that they look at one time-point, so no dynamic information can be extracted. I suggest that the authors tone down their claims.

Please note our response above (Minor points, Line 28) which also refers to this question.

Reviewer #1 (Public review):

This is my first review of this manuscript. The authors included previous reviews for a different journal with a length of 90 and 39 pages; I did not review this reply in my assessment of the paper itself. Influenza prediction is not my area of expertise.

A major concern is that the model is trained in the midst of the COVID-19 pandemic and its associated restrictions and validated on 2023 data. The situation before, during, and after COVID is fluid, and one may not be representative of the other. The situation in 2023 may also not have been normal and reflective of 2024 onward, both in terms of the amount of testing (and positives) and measures taken to prevent the spread of these types of infections. A further worry is that the retrospective prospective split occurred in October 2020, right in the first year of COVID, so it will be impossible to compare both cohorts to assess whether grouping them is sensible.

The outcome of interest is the number of confirmed influenza cases. This is not only a function of weather, but also of the amount of testing. The amount of testing is also a function of historical patterns. This poses the real risk that the model confirms historical opinions through increased testing in those higher-risk periods. Of course, the models could also be run to see how meteorological factors affect testing and the percentage of positive tests. The results only deal with the number of positive (only the overall number of tests is noted briefly), which means there is no way to assess how reasonable and/or variable these other measures are. This is especially concerning as there was massive testing for respiratory viruses during COVID in many places, possibly including China.

(1) Although the authors note a correlation between influenza and the weather factors. The authors do not discuss some of the high correlations between weather factors (e.g., solar radiation and UV index). Because of the many weather factors, those plots are hard to parse.

(2) The authors do not actually compare the results of both methods and what the LSTM adds.

Minor comments:

(3) The methods are long and meandering. They could be cleaned up and shortened. E.g., there is no need for 30 lines on PCR testing; the study area should come before the study design. The authors discuss similar elements in multiple places; this whole section can be shortened considerably without affecting the content.

(4) How reliable is the "Our Word in Data" website for subnational coverage of restrictions? Some of the authors are from Putian and should be able to confirm the accuracy for both studied areas.

(5) Figure 2A is hard to parse; it would make more sense to plot these as line plots (y=count, x=month).

Author response:

Reviewer # 1 (Public review):

A major concern is that the model is trained in the midst of the COVID-19 pandemic and its associated restrictions and validated on 2023 data. The situation before, during, and after COVID is fluid, and one may not be representative of the other. The situation in 2023 may also not have been normal and reflective of 2024 onward, both in terms of the amount of testing (and positives) and measures taken to prevent the spread of these types of infections. A further worry is that the retrospective prospective split occurred in October 2020, right in the first year of COVID, so it will be impossible to compare both cohorts to assess whether grouping them is sensible.

We fully concur with the reviewer that the COVID-19 pandemic represents a profound confounding factor that fundamentally impacts the interpretation and generalizability of our model. This is a critical point that deserves a more thorough treatment. In the revised manuscript, we will add a dedicated subsection in the Discussion to explicitly analyze the pandemic’s impact. We will reframe our model’s contribution not as a universally generalizable tool for a hypothetical “normal” future, but as a robust framework demonstrated to capture complex epidemiological dynamics under the extreme, non-stationary conditions of a real-world public health crisis. We will argue that its strong performance on the 2023 validation data, a unique post-NPI “rebound” year, specifically showcases its utility in modeling volatile periods.

The outcome of interest is the number of confirmed influenza cases. This is not only a function of weather, but also of the amount of testing. The amount of testing is also a function of historical patterns. This poses the real risk that the model confirms historical opinions through increased testing in those higher-risk periods. Of course, the models could also be run to see how meteorological factors affect testing and the percentage of positive tests. The results only deal with the number of positive (only the overall number of tests is noted briefly), which means there is no way to assess how reasonable and/or variable these other measures are. This is especially concerning as there was massive testing for respiratory viruses during COVID in many places, possibly including China.

The reviewer raises a crucial point regarding surveillance bias, which is inherent in studies using reported case data. We acknowledge this limitation and will address it more transparently.

(1) Clarification of Available Data: Our manuscript states that over the six-year period, a total of 20,488 ILI samples were tested, yielding 3,155 positive cases (line 471; Figure 1). We will make this denominator more prominent in the Methods section. However, the reviewer is correct that our models for Putian and the external validation for Sanming utilize the daily positive case counts as the outcome. The reality of our surveillance data source is that while we have the aggregate total of tests over six years, obtaining a reliable daily denominator of all respiratory virus tests conducted (not just for ILI patients as per the surveillance protocol) is not feasible. This is a common constraint in real-world public health surveillance systems.

(2) Justification and Discussion: We will add a detailed paragraph to the Limitations section to address this. We will justify our use of case counts as it is the most direct metric for assessing public health burden and planning resource allocation (e.g., hospital beds, antivirals). We will also explain that modeling the positivity rate presents its own challenges, as the ILI denominator is also subject to biases (e.g., shifts in healthcare-seeking behavior, co-circulation of other pathogens causing similar symptoms). We will thus frame our work as forecasting the direct surveillance signal that public health officials monitor daily.

Although the authors note a correlation between influenza and the weather factors. The authors do not discuss some of the high correlations between weather factors (e.g., solar radiation and UV index). Because of the many weather factors, those plots are hard to parse.

This is an excellent point. Our preliminary analysis (Supplementary Figure S2) indeed confirms a strong positive correlation between solar radiation and the UV index. Perhaps the reviewer overlooked the contents of the supplementary information document. We have included the figure for their review. Our original discussion did explicitly address this multicollinearity, summarized as follows: We acknowledge the high correlation between certain meteorological variables. We then explain that our two-stage modeling approach is designed to mitigate this issue. In the first stage, the DLNM models assess the impact of each variable individually, thus isolating their non-linear and lagged effects without being confounded by interactions. In the second stage, the LSTM network, by its nature, is a powerful non-linear function approximator that is robust to multicollinearity and can learn the complex, interactive relationships between all input features, including correlated ones.

Figure S2. Scatterplot matrix illustrating correlations between Influenza cases and meteorological factors. This comprehensive scatterplot matrix visualizes the relationships between influenza-like illness (ILI) cases, influenza A and B cases, and multiple meteorological variables, including average temperature, humidity, precipitation, wind speed, wind direction, solar radiation, and ultraviolet (UV) index. The figure is composed of three distinct sections that collectively provide an in-depth analysis of these relationships:

(1) Upper-right triangle: This section presents a Pearson correlation coefficient matrix, with color intensity reflecting the strength of correlations between the variables. Red cells represent positive correlations, while green cells represent negative correlations. The closer the coefficient is to 1 or -1, the darker the cell and the stronger the correlation, with statistically significant correlations marked by asterisks. This matrix allows for a rapid identification of notable relationships between influenza cases and meteorological factors.

(2) Lower-left triangle: This section contains scatterplots of pairwise comparisons between variables. These scatterplots facilitate the visual identification of potential linear or non-linear relationships, as well as any outliers or anomalies. This visualization is essential for evaluating the nature of interactions between meteorological factors and influenza cases.

(3) Diagonal: The diagonal displays the density distribution curves for each individual variable. These curves provide an overview of the distribution characteristics of each variable, revealing central tendencies, variance, and any skewness present in the data.

The authors do not actually compare the results of both methods and what the LSTM adds.

We thank the reviewer for this comment and realize we may not have signposted the comparison clearly enough. Our manuscript does present a direct comparison between the LSTM and ARIMA models in the Results section (lines 737-745) and Table 2, where performance metrics (MAE, RMSE, MAPE, SMAPE) for both models on the 2023 validation set are detailed, showing LSTM’s superior performance, particularly for Influenza A. Furthermore, Figure 6 (panels A and B) visualizes the LSTM’s predictions against observed values, and Supplementary Figure S3 does the same for the ARIMA model, allowing for a visual comparison of their fit.

To address the reviewer’s concern, in the revised manuscript, we will:

(1) Add a more explicit comparative statement in the Results section, directly contrasting the key metrics and highlighting the LSTM’s advantages in capturing peak activities.

(2) Consider combining the visualizations from Figure 6 and Supplementary Figure S3 into a single, more powerful comparative figure that shows the observed data, the LSTM predictions, and the ARIMA predictions on the same plot.

Meandering methods; reliability of “Our Word in Data”; Figure 2A is hard to parse.

We will address these points comprehensively.

(3) Methods: We will significantly streamline and restructure the Methods section. We also wish to provide context that the manuscript’s current structure reflects an effort to incorporate feedback from multiple rounds of peer review across different journals, which may have led to some repetition. We will perform a thorough edit to improve its conciseness and logical flow.

(4) Data Reliability: The reviewer raises a crucial and highly insightful question regarding the validity of using a national-level index to represent local public health interventions. This is a critical aspect of our model’s construction, and we are grateful for the opportunity to provide a more thorough justification.

We acknowledge that the ideal variable would be a daily, quantitative, city-level index of non-pharmaceutical interventions (NPIs). However, the practical reality of the data landscape in China is that such granular, publicly accessible databases for subnational regions do not exist. Given this constraint, our choice of the Our World in Data (OWID) national stringency index was the result of a careful consideration process, and we believe it serves as the best available proxy for our study context.

In the revised manuscript, we will significantly expand the Methods section to articulate our rationale, which is threefold:

National Policy Coherence: During the COVID-19 pandemic in mainland China, core NPIs, particularly mandatory face-covering policies in shared public spaces, were implemented with a high degree of national uniformity. While local governments had some autonomy, they operated within a centrally defined framework, ensuring a baseline level of policy consistency across the country.

Local Context Alignment: A key factor supporting the use of this national proxy is the specific epidemiological context of Putian during the study period. For the vast majority of the pandemic, Putian was classified as a low-risk area with only sporadic COVID-19 cases. Consequently, the city’s public health measures consistently aligned with the standard national guidelines. It did not experience prolonged or exceptionally strict local lockdowns that would cause a significant deviation from the national-level policy trends captured by the OWID index.

Validation by Local Public Health Experts: Most critically, and to directly address your suggestion, our co-authors from the Putian Center for Disease Control and Prevention have meticulously reviewed the OWID stringency index against their on-the-ground, institutional knowledge of the mandates that were in effect. They have confirmed that the categorical levels (0-4) and the temporal trends of the OWID index provide a faithful representation of the public health restrictions concerning face coverings as experienced by the population of Putian.

Therefore, we will revise our manuscript to make it clear that the use of the OWID index was not a choice of convenience, but a necessary and well-vetted decision. Given the unavailability of official local data, the OWID index, cross-validated by our local experts, represents the most rigorous and appropriate variable available to account for the profound impact of NPIs on influenza transmission in our model.

(5) Figure 2A: We agree completely and will replace the heatmap with a multi-line plot or a stacked area chart to better visualize the temporal dynamics of influenza subtypes.

We have preliminarily completed the redrawing of Figure 3A. The new and old versions are presented for your review to determine which figure is more suitable for this manuscript in terms of scientific accuracy and visual impact.

Reviewer #2 (Public review):

Weakness (1):

The rationale of the study is not clearly stated.

We appreciate the reviewer’s critique and acknowledge that the unique contribution of our study needs to be articulated more forcefully. Our introduction (lines 105-140) attempted to outline the limitations of existing studies, but we will revise it to be much sharper. The revised introduction will state unequivocally that our study’s rationale is to address a confluence of specific, unresolved gaps in the literature: 1) The persistent challenge of forecasting influenza in subtropical regions with their erratic seasonality; 2) The lack of studies that build subtype-specific models for Influenza A and B, which we show have distinct meteorological drivers; 3) The methodological gap in integrating the explanatory power of DLNM with the predictive power of a rigorously, Bayesian-optimized LSTM network; and 4) The unique opportunity to develop and test a model on data that encompasses the unprecedented disruption of the COVID-19 pandemic, a critical test of model robustness.

Weakness (2):

Several issues with methodological and data integration should be clarified.

We interpret this as a general statement, with the specific issues detailed in the reviewer’s subsequent points and the “Recommendations for the authors” section. We will meticulously address each of these specific points in our revision. For instance, as a demonstration of our commitment to clarification, we will provide a much more detailed justification for our choice of benchmark model (ARIMA), as detailed in our response to Recommendation #11.

Reviewer #2 (Recommendation  for the authors):

The authors should justify why the baseline model selection was made by comparing the LSTM model only with ARIMA? How the outcomes could be sensitive to other commonly used machine learning methods, such as Random Forest or XGBoost, etc, as a benchmark for their performance.

The reviewer raises a highly pertinent question regarding the selection of our benchmark model. A robust comparison is indeed essential for contextualizing the performance of our proposed LSTM network. Our choice to benchmark against the ARIMA model was a deliberate and principled decision, grounded in the specific literature of influenza forecasting at the intersection of climatology and epidemiology.

In the revised manuscript, we will expand our justification within the Methods section and reinforce it in the Discussion. Our rationale is as follows:

(1) ARIMA as the Established Standard: As we briefly noted in our original introduction (lines 110-113), the ARIMA model is arguably the most widely established and frequently cited statistical method for time-series forecasting of influenza incidence, including studies investigating meteorological drivers. It serves as the conventional benchmark against which novel methods in this specific domain are often evaluated. Therefore, demonstrating superiority over ARIMA is the most direct and scientifically relevant way to validate the incremental value of our deep learning approach.

(2) A Focused Scientific Hypothesis: Our primary hypothesis was that the LSTM network, with its inherent ability to capture complex non-linearities and long-term dependencies, could overcome the documented limitations of linear autoregressive models like ARIMA in the context of climate-influenza dynamics. Our study was designed specifically to test this hypothesis.

(3) Avoiding a “Bake-off” without a Clear Rationale: While other machine learning models like Random Forest or XGBoost are powerful, they are not established as the standard baseline in this particular niche of literature. Including them would shift the focus from a targeted comparison against the conventional standard to a broader, less focused “bake-off” of various algorithms. Such an exercise, while potentially interesting, would risk diluting the core message of our paper and would be undertaken without a clear, literature-driven hypothesis for why one of these specific tree-based models should be the next logical benchmark.

Therefore, we will argue in the revised manuscript that our focused comparison with ARIMA provides the clearest and most meaningful assessment of our model’s contribution to the existing body of work on climate-informed influenza forecasting. We will, however, explicitly acknowledge in the Discussion that future work could indeed benefit from a broader comparative analysis as the field continues to evolve and adopt a wider array of machine learning techniques.

Similarly, for some of the reviewer’s recommendations that do not require significant time and effort to implement, such as recommendation 7, we have also redrawn Figure 3 based on your feedback. It is provided for your review.

Figure 3 presents the time series of the cases. I wonder whether the data for these factors and outcomes are daily or aggregated by week/month? I suggest representing it in 9x1 format with a single x-axis to compare, instead of 3x3 format. Authors can refer similar plot in https://doi.org/ 10.1371/journal.pcbi.1012311 in Figure 1.

We are deeply grateful for the reviewer’s valuable suggestion and thoughtful provision of reference illustrations. Based on their input, we have redrawn Figure 3 and have included it for their review.

Weakness (3):

Validation of the models is not presented clearly.

We were concerned by this comment and conducted a thorough self-assessment of our manuscript. We believe we have performed a multi-faceted validation, but we have evidently failed to present it with sufficient clarity and structure. Our validation strategy, detailed across the Methods and Results sections, includes:

• Internal Out-of-Time Validation: Using 2023 data as a hold-out set to test the model trained on 2018-2022 data (lines 695-696, 705-710; Figure 6A, B).

• External Validation: Testing the trained model on an independent dataset from a different city, Sanming (lines 730-736; Figure 6I, J).

• Benchmark Model Comparison: Quantitatively comparing the LSTM’s performance against the standard ARIMA model using multiple error metrics (lines 737-745; Table 2).

• Interpretability Validation (Sanity Check): Using SHAP analysis to ensure the model’s predictions are driven by epidemiologically plausible factors (lines 746-755; Figure 6E-H).

To address the reviewer’s valid critique of our presentation, we will significantly restructure the relevant parts of the Results section. We will create explicit subheadings such as “Internal Validation,” “External Validation,” and “Comparative Performance against ARIMA Benchmark” to make our comprehensive validation process unambiguous and easy to follow.

Weakness (4):

The claim for providing tools for 'early warning' was not validated by analysis and results.

We agree with this assessment entirely. This aligns with the eLife Assessment and comments from Reviewer #1. Our primary revision will be to systematically recalibrate the manuscript's language. We will replace all instances of “early warning tool” with more accurate and modest phrasing, such as “high-performance forecasting framework” or “a foundational model for future warning systems.” We will ensure that our revised title, abstract, and conclusions precisely reflect what our study has delivered: a robust predictive model, not a field-ready public health intervention tool.

Súmula nº 416/TST - MANDADO DE SEGURANÇA. EXECUÇÃO. LEI Nº 8.432/1992. ART. 897, § 1º, DA CLT. CABIMENTO - Devendo o agravo de petição delimitar justificadamente a matéria e os valores objeto de discordância, não fere direito líquido e certo o prosseguimento da execução quanto aos tópicos e valores não especificados no agravo.

Obs.: Considerando a obrigação legal do agravante em delimitar as matérias e valores impugnados, sob pena de imediata execução da parte não impugnada, não há que se falar em violação de direito líquido e certo se o agravante não especifica adequadamente suas impugnações.

Não há suspensão do processo de execução pela interposição de Agravo de Instrumento contra negação ao Agravo de Petição.

A não fruição das férias no período correto gera o dever de pagar o dobro da remuneração.

Obs.: Estava vigente a Súmula nº 450/TST, que previa o pagamento em dobro no caso de inobservância do prazo para pagamento das verbas relativas às férias, para além da hipótese legal do pagamento dobrado para fruição de férias fora do período adequado. Porém, foi julgada inconstitucional pela ADPF 501.

Ementa: ARGUIÇÃO DE DESCUMPRIMENTO DE PRECEITO FUNDAMENTAL. CONSTITUCIONAL E TRABALHISTA. SÚMULA 450 DO TRIBUNAL SUPERIOR DO TRABALHO. PAGAMENTO DA REMUNERAÇÃO DE FÉRIAS EM DOBRO QUANDO ULTRAPASSADO O PRAZO DO ART. 145 DA CLT. IMPOSSIBILIDADE DE O PODER JUDICIÁRIO ATUAR COMO LEGISLADOR POSITIVO. AUSÊNCIA DE LACUNA. INTERPRETAÇÃO RESTRITIVA DE NORMA SANCIONADORA. OFENSA À SEPARAÇÃO DE PODERES E AO PRINCÍPIO DA LEGALIDADE. PROCEDÊNCIA. - 1. Os poderes de Estado devem atuar de maneira harmônica, privilegiando a cooperação e a lealdade institucional e afastando as práticas de guerrilhas institucionais, que acabam minando a coesão governamental e a confiança popular na condução dos negócios públicos pelos agentes públicos. Precedentes. - 2. Impossibilidade de atuação do Poder Judiciário como legislador positivo, de modo a ampliar o âmbito de incidência de sanção prevista no art. 137 da CLT para alcançar situação diversa, já sancionada por outra norma. - 3. Ausência de lacuna justificadora da construção jurisprudencial analógica. Necessidade de interpretação restritiva de normas sancionadoras. Proibição da criação de obrigações não previstas em lei por súmulas e outros enunciados jurisprudenciais editados pelo Tribunal Superior do Trabalho e pelos Tribunais Regionais do Trabalho (CLT, art. 8º, § 2º). - 4. Arguição julgada procedente.

(ADPF 501, Relator(a): ALEXANDRE DE MORAES, Tribunal Pleno, julgado em 08-08-2022, PROCESSO ELETRÔNICO DJe-163 DIVULG 17-08-2022 PUBLIC 18-08-2022)

A empresa devera participar, avisar, comunicar, o empregado a respeito da concessão de férias com antecedência <u>mínima</u> de 30 dias.

Deverá haver recibo desse aviso de concessão de férias.

Súmula nº128/TST - DEPÓSITO RECURSAL (incorporadas as Orientações Jurisprudenciais nºs 139, 189 e 190 da SBDI- - I - É ônus da parte recorrente efetuar o depósito legal, integralmente, <u>em relação a cada novo recurso interposto</u>, sob pena de deserção. Atingido o valor da condenação, nenhum depósito mais é exigido para qualquer recurso. (ex-Súmula nº 128 - alterada pela Res. 121/2003, DJ 21.11.03, que incorporou a OJ nº 139 da SBDI-I - inserida em 27.11.1998)

• II - Garantido o juízo, na fase executória, a exigência de depósito para recorrer de qualquer decisão viola os incisos II e LV do art. 5º da CF/1988. Havendo, porém, elevação do valor do débito, exige-se a complementação da garantia do juízo. (ex-OJ nº 189 da SBDI-I - inserida em 08.11.2000)
• III - Havendo condenação <u>solidária</u> de duas ou mais empresas, o depósito recursal efetuado por uma delas aproveita as demais, quando a empresa que efetuou o depósito não pleiteia sua exclusão da lide. (ex-OJ nº 190 da SBDI-I - inserida em 08.11.2000)

Obs.: Essa súmula define que, na hipótese de condenação solidária, o depósito recursal efetuado por apenas uma devedora aproveita às demais acaso a empresa que efetuou o depósito não demande sua exclusão da lide. Isso porque o depósito recursal tem natureza de garantia do juízo, razão pela qual a empresa que demanda sua exclusão da lide, acaso se reconheça a sua ilegitimidade, o valor garantido seria perdido acaso as demais devedoras não tivesse realizado o depósito recursal.

No mais, a súmula determina o depósito recursal para cada novo recurso, sob pena de deserção, até o limite do valor da condenação.

Afirma-se que, na fase executória, é descabida a exigência de depósito recursal se o juízo já tiver sido garantido, não obstando, lado outro, a complementação da garantia na hipótese de alteração do valor da condenação.

Note que a súmula, sobre o depósito recursal em condenação de duas ou mais empresas, fala em aproveitamento de depósito recursal na hipótese de condenação solidária. Esse mesmo entendimento deve ser estendido à hipótese de condenação subsidiária, desde que a depositante não pretenda à sua exclusão da lide. Vide:

Tema 146/TST: - O depósito recursal efetuado pelo devedor principal, desde que não tenha requerido sua exclusão da lide, aproveita ao responsável <u>subsidiário</u>.

Alguns outros temas importantes aprovados pelo TST em relação ao depósito recursal:

Tema 158: - O comprovante de agendamento bancário não é suficiente para demonstrar o recolhimento das custas processuais e do depósito recursal e não cabe a concessão de prazo para regularização. Obs.: É considerado defeito insanável, insuscetível de concessão de prazo para regularização, a juntada de comprovante de agendamento de depósito para fins de pagamento de custas e depósito recursal, sendo insuficiente para se provar o efeito pagamento.

Via de regra, a jurisprudência consolidada do TST somente admite a concessão de prazo de regularização acaso haja algum valor pago, concedendo-se prazo para a complementação. No entanto, é tido por insanável a ausência de qualquer pagamento de custa ou depósito recursal, razão pela qual, nessas hipóteses, o recurso será tido por deserto. (Tema 271)

Tema 173: - A substituição do depósito recursal por seguro-garantia, nos termos do art. 899, § 11, da CLT, sem a inclusão do acréscimo de 30% exigido pelo art. 3º, II, do Ato Conjunto TST/CSJT/CGJT nº 1/2019, impõe a intimação do recorrente para complementação da garantia, sob pena de deserção, conforme dispõe o art. 1.007, § 2º, do CPC/2015.

Obs.: No processo do trabalho, é admissível a substituição do depósito recursal por seguro-garantia, desde que o valor do seguro abranja o principal + 30%, na mesma sistemática do CPC. O seguro que assegure apenas o valor nominal será considerado insuficiente, razão pela qual deve haver prazo para complementação, sob pena de deserção.

Tema 271: - É incabível a concessão de prazo para regularização do preparo nos casos de <u>total ausência</u> de comprovação do recolhimento das custas ou do depósito recursal no prazo do recurso, não se aplicando o disposto no art. 1.007, §§ 2º, 4º e 7º, do CPC.

RESOLUÇÃO Nº 203, DE 15 DE MARÇO DE 2016. - Edita a Instrução Normativa n° 39, que dispõe sobre as normas do Código de Processo Civil de 2015 aplicáveis e inaplicáveis ao Processo do Trabalho, de forma não exaustiva

Art. 9º O cabimento dos embargos de declaração no Processo do Trabalho, para impugnar qualquer decisão judicial, rege-se pelo art. 897-A da CLT e, supletivamente, pelo Código de Processo Civil (arts. 1022 a 1025; §§ 2º, 3º e 4º do art. 1026), excetuada a garantia de prazo em dobro para litisconsortes (§ 1º do art. 1023). - Parágrafo único. A omissão para fins do prequestionamento ficto a que alude o art. 1025 do CPC dá-se no caso de o Tribunal Regional do Trabalho, mesmo instado mediante embargos de declaração, recusar-se a emitir tese sobre questão jurídica pertinente, na forma da Súmula nº 297, item III, do Tribunal Superior do Trabalho.

Súmula nº 297/TST PREQUESTIONAMENTO. OPORTUNIDADE. CONFIGURAÇÃO

• I. Diz-se prequestionada a matéria ou questão quando na decisão impugnada haja sido adotada, explicitamente, tese a respeito.

• II. Incumbe à parte interessada, desde que a matéria haja sido invocada no recurso principal, opor embargos declaratórios objetivando o pronunciamento sobre o tema, sob pena de preclusão.

• III. Considera-se prequestionada a questão jurídica invocada no recurso principal sobre a qual se omite o Tribunal de pronunciar tese, não obstante opostos embargos de declaração.

Súmula nº184/TST EMBARGOS DECLARATÓRIOS. OMISSÃO EM RECURSO DE REVISTA. PRECLUSÃO - Ocorre preclusão se não forem opostos embargos declaratórios para suprir omissão apontada em recurso de revista ou de embargos.

Súmula nº 278/TST EMBARGOS DE DECLARAÇÃO. OMISSÃO NO JULGADO - A natureza da omissão suprida pelo julgamento de embargos declaratórios pode ocasionar efeito modificativo no julgado.

Súmula nº421/TST EMBARGOS DE DECLARAÇÃO. CABIMENTO. DECISÃO MONOCRÁTICA DO RELATOR CALCADA NO ART. 932 do CPC DE 2015. ART. 557 DO CPC de 1973. (atualizada em decorrência do CPC de 2015) - I – Cabem embargos de declaração da decisão monocrática do relator prevista no art. 932 do CPC de 2015 (art. 557 do CPC de 1973), se a parte pretende tão somente juízo integrativo retificador da decisão e, não, modificação do julgado. - II – Se a parte postular a revisão no mérito da decisão monocrática, cumpre ao relator converter os embargos de declaração em agravo, em face dos princípios da fungibilidade e celeridade processual, submetendo-o ao pronunciamento do Colegiado, após a intimação do recorrente para, no prazo de 5 (cinco) dias, complementar as razões recursais, de modo a ajustá-las às exigências do art. 1.021, § 1º, do CPC

OJ-SDI1-142 EMBARGOS DE DECLARAÇÃO. EFEITO MODI-FICATIVO. VISTA PRÉVIA À PARTE CONTRÁRIA (cancelado o item II em decorrência do CPC de 2015) - - É passível de nulidade decisão que acolhe embargos de declaração com efeito modificativo sem que seja concedida oportunidade de manifestação prévia à parte contrária.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

Nahas et al. investigated the roles of herpes simplex virus 1 (HSV-1) structural proteins using correlative cryo-light microscopy and soft X-ray tomography. The authors generated nine viral variants with deletions or mutations in genes encoding structural proteins. They employed a chemical fixation-free approach to study native-like events during viral assembly, enabling observation of a wider field of view compared to cryo-ET. The study effectively combined virology, cell biology, and structural biology to investigate the roles of viral proteins in virus assembly and budding.

Strengths:

(1) The study presented a novel approach to studying viral assembly in cellulo.

(2) The authors generated nine mutant viruses to investigate the roles of essential proteins in nuclear egress and cytoplasmic envelopment.

(3) The use of correlative imaging with cryoSIM and cryoSXT allowed for the study of viral assembly in a near-native state and in 3D.

(4) The study identified the roles of VP16, pUL16, pUL21, pUL34, and pUS3 in nuclear egress.

(5) The authors demonstrated that deletion of VP16, pUL11, gE, pUL51, or gK inhibits cytoplasmic envelopment.

(6) The manuscript is well-written, clearly describing findings, methods, and experimental design.

(7) The figures and data presentation are of good quality.

(8) The study effectively correlated light microscopy and X-ray tomography to follow virus assembly, providing a valuable approach for studying other viruses and cellular events.

(9) The research is a valuable starting point for investigating viral assembly using more sophisticated methods like cryo-ET with FIB-milling.

(10) The study proposes a detailed assembly mechanism and tracks the contributions of studied proteins to the assembly process.

(11) The study includes all necessary controls and tests for the influence of fluorescent proteins.

Weaknesses:

Overall, the manuscript does not have any major weaknesses, just a few minor comments:

(1) The gel quality in Figure 1 is inconsistent for different samples, with some bands not well resolved (e.g., for pUL11, GAPDH, or pUL20).

We thank the reviewer for their suggestion. We tried to resolve the bands several times, but unfortunately this was the best outcome we could achieve.

(2) The manuscript would benefit from a summary figure or table to concisely present the findings for each protein. It is a large body of manuscript, and a summary figure showing the discovered function would be great.

We thank the reviewer for their suggestion. We have created a summary table (Table 2).

(3) Figure 2 lacks clarity on the type of error bars used (range, standard error, or standard deviation). It says, however, range, and just checking if this is what the authors meant.

We thank the reviewer for double-checking, but it is meant to be range, as reported in the legend. We used range because there are only two data points for each time point, which are insufficient to calculate standard deviation or standard error.

(4) The manuscript could be improved by including details on how the plasma membrane boundary was estimated from the saturated gM-mCherry signal. An additional supplementary figure with the data showing the saturation used for the boundary definition would be helpful.

We appreciate the suggestion and have included an example of how saturated gM-mCherry signal was used to delineate the cytoplasm in Supp. Fig. 4A.

(5) Additional information or supplementary figures on the mask used to filter the YFP signal for Figure 4 would be helpful.

Thanks, we have adapted the text in the results section to clarify: “eYFP-VP26 signal was manually inspected to determine threshold values that filtered out background and included pixels containing individual or clustered puncta that represent capsids.”

(6) The figure legends could include information about which samples are used for comparison for significance calculations. As the colour of the brackets is different from the compared values (dUL34), it would be great to have this information in the figure legend.

Thanks, we have adapted Fig. 4B to make the colour of the brackets match the colour used for the ΔUL34 mutant, and we have included labels next to the brackets for clarity. We have applied similar adjustments to Fig. 5D & E and Supp. Fig. 4C.

(7) In Figure 5B, the association between YFP and mCherry signals is difficult to assess due to the abundance of mCherry signal; single-channel and combined images might improve visualization.

Thanks, we have provided split and combined channel views in Supp. Fig. 4B to improve visualization.

(8) In Figure 6D, staining for tubulin could help identify the cytoskeleton structures involved in the observed virus arrays.

We thank the reviewer for their suggestion, which we think would be interesting future work to build on the current study. Given the competitive nature of access to the cryoSIM and cryoSXT, CLXT, including staining for tubulin was outside the scope of additional experiments we were able to conduct at this time.

(9) It is unclear in Figure 6D if the microtubule-associated capsids are with the gM envelope or not, as the signal from mCherry is quite weak. It could be made clearer with the split signals to assess the presence of both viral components.

We have provided split channels to the figure to aid with visualization.

(10) The representation of voxel intensity in Figure 8 is somewhat confusing. Reversion of the voxel intensity representation to align brighter values with higher absorption, which would simplify interpretation.

We thank the reviewer for this suggestion. In contrast to fluorescence microscopy where high intensities reflect signal, low intensities represent signal (absorbance of X-rays) in cryoSXT. We respectfully decided not to reverse the values, as we believe that could cause more confusion. We have instead added a black-to-white gradient bar to illustrate that low voxel intensities correspond to dark signal in Fig 8.

(11) The visualization in panel I of Figure 8 might benefit from a more divergent colormap to better show the variation in X-ray absorbance.

We thank the reviewer for their suggestion. We experimented with a few different colour schemes but concluded that the current one produced the clearest results and was most accessible for color-blind viewers.

(12) Figure 9 would be enhanced by images showing the different virus sizes measured for the comparative study, which would help assess the size differences between different assembly stages.

We thank the reviewer for their suggestion and have included images to accompany the graph.

Overall, this is an excellent manuscript and an enjoyable read. It would be interesting to see this approach applied to the study of other viruses, providing valuable insights before progressing to high-resolution methods.

Reviewer #2 (Public review):

Summary:

For centuries, humans have been developing methods to see ever smaller objects, such as cells and their contents. This has included studies of viruses and their interactions with host cells during processes extending from virion structure to the complex interactions between viruses and their host cells: virion entry, virus replication and virion assembly, and release of newly constructed virions. Recent developments have enabled simultaneous application of fluorescence-based detection and intracellular localization of molecules of interest in the context of sub-micron resolution imaging of cellular structures by electron microscopy.

The submission by Nahas et al., extends the state-of-the-art for visualization of important aspects of herpesvirus (HSV-1 in this instance) virion morphogenesis, a complex process that involves virus genome replication, and capsid assembly and filling in the nucleus, transport of the nascent nucleocapsid and some associated tegument proteins through the inner and outer nuclear membranes to the cytoplasm, orderly association of several thousand mostly viral proteins with the capsid to form the virion's tegument, envelopment of the tegumented capsid at a virus-tweaked secretory vesicle or at the plasma membrane, and release of mature virions at the plasma membrane.

In this groundbreaking study, cells infected with HSV-1 mutants that express fluorescently tagged versions of capsid (eYFP-VP26) and tegument (gM-mCherry) proteins were visualized with 3D correlative structured illumination microscopy and X-ray tomography. The maturation and egress pathways thus illuminated were studied further in infections with fluorescently tagged viruses lacking one of nine viral proteins.

Strengths:

This outstanding paper meets the journal's definitions of Landmark, Fundamental, Important, Valuable, and Useful. The work is also Exceptional, Compelling, Convincing, and Solid. The work is a tour de force of classical and state-of-the-art molecular and cellular virology. Beautiful images accompanied by appropriate statistical analyses and excellent figures. The numerous complex issues addressed are explained in a clear and coordinated manner; the sum of what was learned is greater than the sum of the parts. Impacts go well beyond cytomegalovirus and the rest of the herpesviruses, to other viruses and cell biology in general.

Reviewer #3 (Public review):

Summary:

Kamal L. Nahas et al. demonstrated that pUL16, pUL21, pUL34, VP16, and pUS3 are involved in the egress of the capsids from the nucleous, since mutant viruses ΔpUL16, ΔpUL21, ΔUL34, ΔVP16, and ΔUS3 HSV-1 show nuclear egress attenuation determined by measuring the nuclear:cytoplasmic ratio of the capsids, the dfParental, or the mutants. Then, they showed that gM-mCherry+ endomembrane association and capsid clustering were different in pUL11, pUL51, gE, gK, and VP16 mutants. Furthermore, the 3D view of cytoplasmic budding events suggests an envelopment mechanism where capsid budding into spherical/ellipsoidal vesicles drives the envelopment.

Strengths:

The authors employed both structured illumination microscopy and cellular ultrastructure analysis to examine the same infected cells, using cryo-soft-X-ray tomography to capture images. This combination, set here for the first time, enabled the authors to obtain holistic data regarding a biological process, as a viral assembly. Using this approach, the researchers studied various stages of HSV-1 assembly. For this, they constructed a dual-fluorescently labelled recombinant virus, consisting of eYFP-tagged capsids and mCherry-tagged envelopes, allowing for the independent identification of both unenveloped and enveloped particles. They then constructed nine mutants, each targeting a single viral protein known to be involved in nuclear egress and envelopment in the cytoplasm, using this dual-fluorescent as the parental one. The experimental setting, both the microscopic and the virological, is robust and well-controlled. The manuscript is well-written, and the data generated is robust and consistent with previous observations made in the field.

Weaknesses:

It would be helpful to find out what role the targeted proteins play in nuclear egress or envelopment acquisition in a different orthoherpesvirus, like HSV-2. This would confirm the suitability of the technical approach set and would also act as a way to validate their mechanism at least in one additional herpesvirus beyond HSV-1. So, using the current manuscript as a starting point and for future studies, it would be advisable to focus on the protein functions of other viruses and compare them.

We appreciate the suggestion and agree that this would be a great starting point for future studies. At present, we do not have a panel of mutant viruses in HSV-2 or another orthoherpesvirus, and it would be significant work to generate them, so we consider this outside the scope of the current study.

Recommendations for the authors:

Reviewer #2 (Recommendations for the authors):

(1) There are enough uncommon abbreviations in the text to justify the inclusion of an abbreviation list.

We thank the reviewer for the suggestion, but we define all uncommon abbreviations at first mention and an abbreviations list is not part of eLife’s house style.

(2) The complex paragraph on p. 7 would be much easier to digest if broken into smaller chunks. Consider similar treatment for other lengthy landmark-free blocks of text, e.g., the one that begins on p. 14. Subheadings would help.

We thank the reviewer for this suggestion. We have divided large paragraphs into more easily digestible chunks throughout the manuscript, for example in the discussion where the previous monolithic 3rd paragraph has been divided into five shorter, focussed paragraphs.

(3) Table 1 needs units.

We thank the reviewer for noticing our omission and apologise for the oversight - the table has been updated accordingly.

Reviewer #3 (Recommendations for the authors):

(1) Toward the end of the manuscript, I missed some lines attempting to speculate on the origin/nature of the spherical/ellipsoidal vesicles providing the envelopment. Would it be possible to incorporate this in the Discussion section?

Thank you for noticing that omission. We have now included a few lines speculating that they may represent recycling endosomes, trans-Golgi network vesicles, or a hybrid compartment.

(2) I congratulate the authors. The work is robust, and I personally highlight the way they managed to include others' results merged with their own, providing a complete view of the story.

We thank the reviewer for their kind words.

Note to editors

In addition to these responses to the reviewer’s comments, we have also now included in the methods section details of the Tracking of Indels by Decomposition (TIDE) analysis we performed (data in Supplementary Figure 3) that was omitted by mistake from the original submission.

Cortes is claiming him and Montezuma are friends.

The captain only speaks Spanish.

There wasn't enough signatures to delay LA County's minimum wage increases: 22.50 in 25 to 25 in 26, 27.50 in 27 and $30 and 28. Also a mandatory healthcare benefits payment willl being in 26

You love her these three days, You love her these three days, My heart to you, My heart to you, You love her these three days

Author response:

The following is the authors’ response to the previous reviews

Reviewer #1 (Public review):

Summary:

This computational modeling study builds on multiple previous lines of experimental and theoretical research to investigate how a single neuron can solve a nonlinear pattern classification task. The authors construct a detailed biophysical and morphological model of a single striatal medium spiny neuron, and endow excitatory and inhibitory synapses with dynamic synaptic plasticity mechanisms that are sensitive to (1) the presence or absence of a dopamine reward signal, and (2) spatiotemporal coincidence of synaptic activity in single dendritic branches. The latter coincidence is detected by voltage-dependent NMDA-type glutamate receptors, which can generate a type of dendritic spike referred to as a "plateau potential." In the absence of inhibitory plasticity, the proposed mechanisms result in good performance on a nonlinear classification task when specific input features are segregated and clustered onto individual branches, but reduced performance when input features are randomly distributed across branches. Interestingly, adding inhibitory plasticity improves classification performance even when input features are randomly distributed.

Strengths:

The integrative aspect of this study is its major strength. It is challenging to relate low-level details such as electrical spine compartmentalization, extrasynaptic neurotransmitter concentrations, dendritic nonlinearities, spatial clustering of correlated inputs, and plasticity of excitatory and inhibitory synapses to high-level computations such as nonlinear feature classification. Due to high simulation costs, it is rare to see highly biophysical and morphological models used for learning studies that require repeated stimulus presentations over the course of a training procedure. The study aspires to prove the principle that experimentally-supported biological mechanisms can explain complex learning.

Weaknesses:

The high level of complexity of each component of the model makes it difficult to gain an intuition for which aspects of the model are essential for its performance, or responsible for its poor performance under certain conditions. Stripping down some of the biophysical detail and comparing it to a simpler model may help better understand each component in isolation.

We greatly appreciate your recognition of the study’s integrative scope and the challenges of linking detailed biophysics to high-level computation. We acknowledge that the model’s complexity can obscure the contribution of individual components. However, as stated in the introduction the principles already have been shown in simplified theoretical models for instance  in Tran-Van-Minh et al. 2015. Our aim here was to extend those ideas into a more biologically detailed setting to test whether the same principles still hold under realistic constraints. While simplification can aid intuition, we believe that demonstrating these effects in a biophysically grounded model strengthens the overall conclusion. We agree that further comparisons with reduced models would be valuable for isolating the contribution of specific components and plan to explore that in future work.  

Reviewer #2 (Public review):

Summary:

The study explores how single striatal projection neurons (SPNs) utilize dendritic nonlinearities to solve complex integration tasks. It introduces a calcium-based synaptic learning rule that incorporates local calcium dynamics and dopaminergic signals, along with metaplasticity to ensure stability for synaptic weights. Results show SPNs can solve the nonlinear feature binding problem and enhance computational efficiency through inhibitory plasticity in dendrites, emphasizing the significant computational potential of individual neurons. In summary, the study provides a more biologically plausible solution to single-neuron learning and gives further mechanical insights into complex computations at the single-neuron level.

Strengths:

The paper introduces a novel learning rule for training a single multicompartmental neuron model to perform nonlinear feature binding tasks (NFBP), highlighting two main strengths: the learning rule is local, calcium-based, and requires only sparse reward signals, making it highly biologically plausible, and it applies to detailed neuron models that effectively preserve dendritic nonlinearities, contrasting with many previous studies that use simplified models.

Thank you for highlighting the biological plausibility of our calcium- and dopamine-dependent learning rule and its ability to exploit dendritic nonlinearities. Your positive assessment reinforces our commitment to refining the rule and exploring its implications in larger, more diverse settings.

Reviewer #1 (Recommendations for the authors):

Major recommendations:

P9: When introducing the excitatory learning rule, the reader is referred to the Methods. I suggest moving Figure 7A-D, "Excitatory plasticity" to be more prominently presented in the main body of the paper where the reader needs to understand it. There are errors in the current Figure 7, and wrong/confusing acronyms. The abbreviations "LTP-K" and "MP-K" are not intuitive. In A, I would spell out "LTP kernel" and "Theta_LTP adaptation".  In B, I would spell out "LTD kernel" and "Theta_LTD adaptation".

We have clarified the terminology in Figure 7 by replacing “LTP-K” with “LTP kernel” and “MP-K” with “metaplasticity kernel”.  While we kept Figure 7 in the Methods section to maintain the flow of the main text, we agree that an earlier introduction of the learning rule improves clarity. To that end, we added a simplified schematic to Figure 3 in the Results section, which provides readers with an accessible overview of the excitatory plasticity mechanism at the point where it is first introduced.

In C, for simplicity and clarity, I would only show the initial and updated LTP kernel and Calcium and remove the Theta_LTP adaptation curve, it's too busy and not necessary. Similarly in D, I would show only the initial and updated LTD kernel and Calcium and remove the Theta_LTD adaptation curve. In the current version of the Figure, panel B, right incorrectly labels "Theta_LTD" as "Theta_LTP". Panel D incorrectly labels "LTD kernel" as "LTP/MP-K" in the subheading and "MP/LTP-K" in the graph.

To avoid confusion and better illustrate the interactions between calcium signals, kernels, and thresholds, we have added a movie showing how these components evolve during learning. The figure panels remain as originally designed, since the LTP kernel governs both potentiation and depression through metaplastic threshold adaptation, while the LTD kernel remains fixed.

P17: Again, instead of pointing the reader to the Methods, I would move Figure 7E, "Inhibitory plasticity" to the main body of the paper where the reader needs to understand it. For clarity, I would label "C_TL" and "Theta_Inh,low" and "C_TH" as "Theta_Inh,high". The right panel could be better labeled "Inhibitory plasticity kernel". The left panel could be better labeled "Theta_Inh adaptation", with again replacing the acronyms "C_TL" and "C_TH". The same applies to Fig. 5D on P19.

We have updated the labeling in Figures 5D and 7E for clarity, including replacing "C_TL" and "C_TH" with "Theta_Inh,low" and "Theta_Inh,high". In addition, we added a simplified schematic of the inhibitory plasticity rule to Figure 5 to assist the reader’s understanding when presenting the results. Figure 7E remains in the Methods section to preserve the flow of the main text.

P12: I would suggest simplifying Fig. 3 panels and acronyms as well. Remove "MP-K" from C and D. Relabel "LTP-K" as "LTP kernel". The same applies to Fig. 5E on P19 and Fig. 3 - supplement 1 on P46 and Fig 6 - supplement 1 on P49.

We have simplified the labeling across all relevant figures by replacing “MP-K” with “metaplasticity kernel” and “LTP-K” with “LTP kernel.” To maintain clarity, we retained these terms in only one panel as a reference.

Minor recommendations:

P4: "Although not discussed much in more theoretical work, our study demonstrates the necessity of metaplasticity for achieving stable and physiologically realistic synaptic weights." This sentence is jarring. BCM and metaplasticity has been discussed in hundreds of theory papers! Cite some. This sentence would more accurately read, "Our study corroborates prior theory work (citations) demonstrating that metaplasticity helps to achieve stable and physiologically realistic synaptic weights."

We have followed the reviewers suggestion and updated the sentence to: Previous theoretical studies (Bienenstock et al., 1982; Fusi et al., 2005; Clopath et al., 2010; Benna & Fusi, 2016; Zenke & Gerstner, 2017) demonstrate the essential role of metaplasticity in maintaining stability in synaptic weight distributions. (page 2 line 49-51, page 3 line 1)

P9: Grammar. "The neuron model was during training activated..." should read "During training, the neuron model was activated..."

Corrected

P17: Lovett-Barron et al., 2012 is appropriately cited here. Milstein et al., Neuron, 2015 also showed dendritic inhibition regulates plateau potentials in CA1 pyramidal cells in vitro, and Grienberger et al., Nat. Neurosci., 2017 showed it in vivo.

P19 vs P16 vs P21. Fig. 4B, Fig. 5B, and Fig. 6B choose different strategies to show variance across seeds. Please choose one strategy and apply to all comparable plots.

We thank the reviewer for these helpful points.

We have added the suggested citations (Milstein et al., 2015; Grienberger et al., 2017) alongside Lovett-Barron et al., 2012. 

Variance across seeds is now displayed uniformly (mean is solid line STD is shaded area) in Figures 4B, 5B, and 6B.

Reviewer #2 (Recommendations for the authors):

Major Points:

(1)  Quality of Scientific Writing:

i. Mathematical and Implementation Details:

I appreciate the authors' efforts in clarifying the mathematical details and providing pseudocode for the learning rule, significantly improving readability and reproducibility. The reference to existing models via GitHub and ModelDB repositories is acceptable. However, I suggest enhancing the presentation quality of equations within the Methods section-currently, they are low-resolution images. Please consider rewriting these equations using LaTeX or replacing them with high-resolution images to further improve clarity.

We appreciate the reviewer’s comment regarding clarity and reproducibility. In response, we have rewritten all equations in LaTeX to improve their readability and presentation quality in the Methods section.

ii. Figure quality.

I acknowledge the authors' effort to improve figure clarity and consistency throughout the manuscript. However, I notice that the x-axis label "[Ca]_v (μm)" in Fig. 7E still appears compressed and unclear. Additionally, given the complexity and abundance of hyperparameters or artificial settings involved in your experimental design and learning rule (such as kernel parameters, metaplasticity kernels, and unspecific features), the current arrangement of subfigures (particularly Fig. 3C, D and Fig. 5D, E) still poses readability challenges. I recommend reordering subfigures to present primary results (e.g., performance outcomes) prominently upfront, while relegating visualizations of detailed hyperparameter manipulations or feature weight variations to later sections or the discussion, thus enhancing clarity for readers.

We thank the reviewer for pointing out the readability issue. We have corrected the x-axis label in Figure 7D. We hope this new layout with a simplified rule in Fig 3 and Fig 5   presents the key findings while retaining full mechanistic detail to make it easier to understand the model behavior.  

iii. Writing clarity.

The authors have streamlined the "Metaplasticity" section and reduced references to dopamine, which is a positive step. However, the broader issue remains: the manuscript still appears overly detailed and more like a technical report of a novel learning rule, rather than a clearly structured scientific paper. I strongly recommend that the authors further distill the manuscript by clearly focusing on one or two central scientific questions or hypotheses-for instance, emphasizing core insights such as "inhibitory inputs facilitate nonlinear dendritic computations" or "distal dendritic inputs significantly contribute to nonlinear integration." Clarifying and highlighting these primary scientific questions early and consistently throughout the manuscript would substantially enhance readability and impact.

We appreciate the reviewer’s guidance on improving the manuscript’s clarity and focus.In response, we now highlight two central questions at the end of the Introduction and have retitled the main Results subsections to follow this thread, thereby sharpening the manuscript’s focus while retaining necessary technical detail (page3 line 20-28).We have also removed redundant passages and simplified technical details to improve overall readability .

Minor:

(1) The [Ca]NMDA in Figure 2A and 2C can have large values even when very few synapses are activated. Why is that? Is this setting biologically realistic?

The authors acknowledge that their simulated [Ca²⁺] levels exceed typical biological measurements but claim that the learning rule remains robust across variations in calcium concentrations. However, robustness to calcium variations was not explicitly demonstrated in the main figures. To convincingly address this concern, I recommend the authors explicitly test and present whether adopting biologically realistic calcium concentrations (~1 μM) impacts the learning outcomes or synaptic weight dynamics. Clarifying this point with a supplemental analysis or an additional figure panel would significantly strengthen their argument regarding the model's biological plausibility and robustness.

We thank the reviewer for the comment. The elevated [Ca^²⁺]_NMDA values reflect localized transients in spine heads with narrow necks and high NMDA conductance. These values are not problematic for our model, as the plasticity rule depends on relative calcium differences rather than absolute levels as the metaplasticity kernel will adjust. In future versions of our detailed neuron model, we will likely decrease the spine axial resistance of the spine neck.

In fact there is no difference to other plot designs, you can do exactly the same also in fixed area or nested plots... If you calculate the expansion factor per tree and expand e.g. tree basal area to one hectar, then you can sum this over all of your trees (from multiple plots) and divide by n. Same result! Sum(y_i) (plot aggregate) is here equal to Sum(y_ij) (sum over trees). Only that you need no expansion factor here, since you are already counting on a per ha basis

I am confused: this is only for counting factor k=1? For measurement in meters and cm the ratios for k=1,2,4 are 1:50, 1:34.5 and 1:25 respectively (both in same units). Means, using k=1 (every counted tree is 1m²/ha) a tree with dbh of 35cm (0.35m) has a maximum distance of 0.35*50=17.5 meters SORRY, just realized that k is not counting factor (as used here in Germany). Our "k" is your BAF and what you notate as k is what we call c...

Ship of supplies was burned down and there are some serious consequences.

eLife Assessment

This valuable work explores how synaptic activity encodes information during memory tasks. All reviewers agree that the work is of very high quality and that the methodological approach is praiseworthy. Although the experimental data support the possibility that phospholipase diacylglycerol signaling and synaptotagmin 7 (Syt7) dynamically regulate the vesicle pool required for presynaptic release, a concern remains that the central finding of paired-pulse depression at very short intervals could be due to a mechanism that does not depend on exocytosis, such as Ca²⁺ channel inactivation, rather than vesicle pool depletion. Overall, this is a solid study although the results still warrant consideration of alternative interpretations.

Reviewer #3 (Public review):

To summarize: The authors' overfilling hypothesis depends crucially on the premise that the very-quickly reverting paired-pulse depression seen after unusually short rest intervals of << 50 ms is caused by depletion of release sites whereas Dobrunz and Stevens (1997) concluded that the cause was some other mechanism that does not involve depletion. The authors now include experiments where switching extracellular Ca2+ from 1.2 to 2.5 mM increases synaptic strength on average, but not by as much as at other synapse types. They contend that the result supports the depletion hypothesis. I didn't agree because the model used to generate the hypothesis had no room for any increase at all, and because a more granular analysis revealed a mixed population with a subset where: (a) synaptic strength increased by as much as at standard synapses; and yet (b) the quickly reverting depression for the subset was the same as the overall population.

The authors raise the possibility of additional experiments, and I do think this could clarify things if they pre-treat with EGTA as I recommended initially. They've already shown they can do this routinely, and it would allow them to elegantly distinguish between pv and pocc explanations for both the increases in synaptic strength and the decreases in the paired pulse ratio upon switching Ca2+ to 2.5 mM. Plus/minus EGTA pre-treatment trials could be interleaved and done blind with minimal additional effort.

Showing reversibility would be a great addition too, because, in our experience, this does not always happen in whole-cell recordings in ex-vivo tissue even when electrical properties do not change. If the goal is to show that L2/3 synapses are less sensitive to changes in Ca2+ compared to other synapse types - which is interesting but a bit off point - then I would additionally include a positive control, done by the same person with the same equipment, at one of those other synapse types using the same kind of presynaptic stimulation (i.e. ChRs).

Specific points (quotations are from the Authors' rebuttal)

(1) Regarding the Author response image 1, I was instead suggesting a plot of PPR in 1.2 mM Ca2+ versus the relative increase in synaptic strength in 2.5 versus in 1.2 mM. This continues to seem relevant.

(2) "Could you explain in detail why two-fold increase implies pv < 0.2?"

a. start with power((2.5/(1 + (2.5/K1) + 1/2.97)),4) = 2*power((1.3/(1 + (1.3/K1) + 1/2.97)),4);

b. solve for K1 (this turns out to be 0.48);

c. then implement the premise that pv -> 1.0 when Ca2+ is high by calculating Max = power((C/(1 + (C/K1) + 1/2.97)),4) where C is [Ca] -> infinity.

d. pv when [Ca] = 1.3. mM must then be power((1.3/(1 + (1.3/K1) + 1/2.97)),4)/Max, which is <0.2.

Note that modern updates of Dodge and Rahamimoff typically include a parameter that prevents pv from approaching 1.0; this is the gamma parameter in the versions from Neher group.

(3) "If so, we can not understand why depletion-dependent PPD should lead to PPF."

When PPD is caused by depletion and pv < 0.2, the number of occupied release sites should not be decreased by more than one-fifth at the second stimulus so, without facilitation, PPR should be > 0.8. The EGTA results then indicate there should be strong facilitation, driving PPR to something like 1.2 with conservative assumptions. And yet, a value of < 0.4 is measured, which is a large miss.

(4) Despite the authors' suggestion to the contrary, I continue to think there is a substantial chance that Ca2+-channel inactivation is the mechanism underlying the very quickly reverting paired-pulse depression. However, this is only one example of a non-depletion mechanism among many, with the main point being that any non-depletion mechanism would undercut the reasoning for overfilling. And, this is what Dobrunz and Stevens claimed to show; that the mechanism - whatever it is - does not involve depletion. The most effective way to address this would be affirmative experiments showing that the quickly reverting depression is caused by depletion after all. Attempting to prove that Ca2+-channel inactivation does not occur does not seem like a worthwhile strategy because it would not address the many other possibilities.

(5) True that Kusick et al. observed morphological re-docking, but then vesicles would have to re-prime and Mahfooz et al. (2016) showed that re-priming would have to be slower than 110 ms (at least during heavy use at calyx of Held).

should look like. We believe in diversity.

7ww w.si deways.earth

Main benefits

• Exchange and commerce
• Cross-organization dispute resolution
• International recognition
• Member hosting and benefits across organizations
• Values alignment / adherence to standards

DAOs vouched

you can have the same basal area with a single big tree or hundreds of small trees. There is not necessarily a relation between BA and density

herramienta de diseño de cocinas 3D Faire la même modification dans le titre en dessus

¿Dudas? ¡Te contestamos!

La herramienta 3D de ixina permite diseñar una cocina en unos pocos clics, incluso aunque no tengas experiencia previa con este tipo de herramientas online. Su interfaz intuitiva te va guiando paso a paso. Te ayuda a tener en cuenta aspectos fundamentales como la ergonomía: la altura ideal de las encimeras según tu estatura, la accesibilidad de los muebles altos o la ubicación óptima de los electrodomésticos respecto a las zonas de trabajo de la cocina. Cada paso está diseñado para que el resultado sea funcional, estético y perfectamente adecuado a tus necesidades.

DOI: 10.1038/s41418-019-0378-6

Resource: Bloomington Drosophila Stock Center (RRID:SCR_006457)

Curator: @bdscstockkeepers

SciCrunch record: RRID:SCR_006457

What is this?

DOI: 10.1038/s41418-019-0378-6

Resource: Bloomington Drosophila Stock Center (RRID:SCR_006457)

Curator: @bdscstockkeepers

SciCrunch record: RRID:SCR_006457

What is this?

DOI: 10.1161/ATVBAHA.124.322893

Resource: RRID:IMSR_JAX:026284

Curator: @scibot

SciCrunch record: RRID:IMSR_JAX:026284

What is this?

DOI: 10.1101/2025.08.29.673187

Resource: La Jolla Institute for Immunology Next Generation Sequencing Core Facility (RRID:SCR_023107)

Curator: @scibot

SciCrunch record: RRID:SCR_023107

What is this?

DOI: 10.1016/j.xgen.2025.100980

Resource: Sony MA900 Multi-Application Cell Sorter (RRID:SCR_026300)

Curator: @scibot

SciCrunch record: RRID:SCR_026300

What is this?

Consultar el catélogo online y dejarme guiar para crear mi proyecto

Reviewer #1 (Public review):

Summary:

The authors investigate how methicillin-resistant (MRSA) and sensitive (MSSA) Staphylococcus aureus adapt to a new host (C. elegans) in the presence or absence of a low dose of the antibiotic oxacillin. Using an "Evolve and Resequence" design with 48 independently evolving populations, they track changes in virulence, antibiotic resistance, and other fitness-related traits over 12 passages. Their key finding is that selection from both the host and the antibiotic together, rather than either pressure alone, results in the evolution of the most virulent pathogens. Genomically, they find that this adaptation repeatedly involves mutations in a small number of key regulatory genes, most notably codY, agr, and saeRS.

Strengths:

The main advantage of the research lies in its strong and thoroughly replicated experimental framework, enabling significant conclusions to be drawn based on the concept of parallel evolution. The study successfully integrates various phenotypic assays (virulence, growth, hemolysis, biofilm formation) with whole-genome sequencing, offering an extensive perspective on the adaptive landscape. The identification of certain regulatory genes as common targets of selection across distinct lineages is an important result that indicates a level of predictability in how pathogens adapt.

Weaknesses:

(1) The main limitation of the paper is that its findings on the function of specific genes are based on correlation, not cause-and-effect evidence. While the parallel evolution evidence is strong, the authors have not yet performed the definitive tests (i.e., reconstruction of ancestral genes) to ensure that the mutations identified in isolation are enough to account for the virulence or resistance changes observed. This makes the conclusions more like firm hypotheses, not confirmed facts.

(2) In some instances, the claims in the text are not fully supported by the visual data from the figures or are reported with vagueness. For example, the display of phenotypic clusters in the PCA (Figure 6A) and the sweeping generalization about the effect of antibiotics on the mutation rates (Figure S5) can be more precise and nuanced. Such small deviations dilute the overall argument somewhat and must be corrected.

Reviewer #2 (Public review):

Summary:

This study attempts to dissect the contributions of type I and type III IFNs to the antiviral response in chickens. The first part of the study characterises the generation of IFNAR and IFNLR KO chicken strains and describes basic differences. Four different viruses are then tested in chicken embryos, while the subsequent analysis of the antiviral response in vivo is performed with one influenza H3N1 strain.

Strengths:

Having these two KO chicken strains as a tool is a great achievement. The initial analysis is solid. Clear effect of IFNAR deficiency in in vivo infection, less so for IFNLR deficiency.

Weaknesses:

(1) The antibody induction by KLH immunisation: No data indicated whether or not this vaccination induces IFN responses in wt mice, so the effects observed may be due to steady-state differences or to differential effects of IFN induced during the vaccination phase. No pre-immune results are shown. The differences are relatively small and often found at only one plasma dilution - the whole of Figure 4 could be condensed into one or two panels by proper calculation of Ab titers - would these titres be significantly different? This, as all of the other in vivo experiments, has not been repeated, if I understand the methods section correctly.

(2) The basic conundrum here and in later figures is never addressed by the authors: Situations where IFN type 1 and 3 signalling deficiency each have an independent effect (i.e., Figure 4d) suggest that they act by separate, unrelated mechanisms. However, all the literature about these IFN families suggests that they show almost identical signalling and gene induction downstream of their respective receptors. How can the same signalling, clearly active here downstream of the receptors for IFN type 1 or type 3, be non-redundant, i.e., why does the unaffected IFN family not stand in? This is a major difference from the mouse studies, which showed a rather subtle phenotype when only one of the two IFN systems was missing, but a massive reduction in virus control in double KO mice (the correct primary paper should be quoted here, not only the review by McNab). Reasons could be a direct effect of IFNab on B cells and an indirect effect of IFNL through non-B cells, timing issues, and many other scenarios can be envisaged. The authors do not address this question, which limits the depth of analysis.

(3) In the one in vivo experiment performed with chickens, only one virus was tested; more influenza strains should be included, as well as non-influenza viruses.

(4) The basic conundrum of point 2 applies equally to Figure 6a; both KOs have a phenotype. Again in 6d, both IFNs appear to be separately required for Mx induction. An explanation is needed.

(5) Line 308, where are the viral titers you refer to in the text? The statement that the results demonstrate that excessive IFNab has a negative impact is overstretched, as no IFN measurements of the infected embryos are shown here.

(6) The in vivo infection is the most interesting experiment, and the key outcome here is that IFN type 1 is crucial for anti-H3N1 protection in chickens, while type 3 is less impactful. However, this experiment suffers from the different time points when chickens were culled, so many parameters are impossible to compare (e.g., weight loss, histopathology, IFN measurements, and more). Many of these phenomena are highly dynamic in acute virus infections, so disparate time points do not allow a meaningful comparison between different genotypes. What are the stats in 7b? Is the median rather than the mean indicated by the line? Otherwise, the lines appear in surprising places. SD must be shown, and I find it difficult to believe that there is a significant difference in weight, for e.g., IFNAR KO, unless maybe with a paired t test. What is the statistical test?

(7) Figures 7e,f: these comparisons are very difficult to interpret as the virus loads at these time points already differ significantly, so any difference could be secondary to virus load differences.

la cocina de tus sueños

Je pense que les titres ici ne sont pas bien placés. Muebles 10 años garantía - > la garantía au lieu de la calidad Aprés le √ -> la calidad au lieu de la garantía

Un diseñador realiza tu proyecto a medida y el plano en 3D Disfruta de un asesoramiento personal para crear la cocina de tus sueños. Un asesor te acompañará en casa paso, te ayudará a elegir los muebles, la configuración ideal, los accesorios y los electrodomésticos según tus usos y necesidades… todo ello adaptado a TU presupuesto. El asesor te ayudará a tener una visión más clara, especialmente gracias a un diseño en 3D de tu cocina que te permitirá proyectarla mejor. Podrás tomarte el tiempo que necesites: tu asesor se adapta a TI. Además, también podrá ayudarte a encontrar opciones de financiación a corto, medio o largo plazo.

la tuya

Author response:

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

Reviewer #1 (Public review):

Summary:

Chao et al. produced an updated version of the SpliceAI package using modern deep learning frameworks. This includes data preprocessing, model training, direct prediction, and variant effect prediction scripts. They also added functionality for model fine-tuning and model calibration. They convincingly evaluate their newly trained models against those from the original SpliceAI package and investigate how to extend SpliceAI to make predictions in new species. While their comparisons to the original SpliceAI models are convincing on the grounds of model performance, their evaluation of how well the new models match the original's understanding of non-local mutation effects is incomplete. Further, their evaluation of the new calibration functionality would benefit from a more nuanced discussion of what set of splice sites their calibration is expected to hold for, and tests in a context for which calibration is needed.

Strengths:

(1) They provide convincing evidence that their new implementation of SpliceAI matches the performance of the original model on a similar dataset while benefiting from improved computational efficiencies. This will enable faster prediction and retraining of splicing models for new species as well as easier integration with other modern deep learning tools.

(2) They produce models with strong performance on non-human model species and a simple, well-documented pipeline for producing models tuned for any species of interest. This will be a boon for researchers working on splicing in these species and make it easy for researchers working on new species to generate their own models.

(3) Their documentation is clear and abundant. This will greatly aid the ability of others to work with their code base.

We thank the reviewer for these positive comments.  

Weaknesses:

(1) The authors' assessment of how much their model retains SpliceAI's understanding of "nonlocal effects of genomic mutations on splice site location and strength" (Figure 6) is not sufficiently supported. Demonstrating this would require showing that for a large number of (non-local) mutations, their model shows the same change in predictions as SpliceAI or that attribution maps for their model and SpliceAI are concordant even at distances from the splice site. Figure 6A comes close to demonstrating this, but only provides anecdotal evidence as it is limited to 2 loci. This could be overcome by summarizing the concordance between ISM maps for the two models and then comparing across many loci. Figure 6B also comes close, but falls short because instead of comparing splicing prediction differences between the models as a function of variants, it compares the average prediction difference as a function of the distance from the splice site. This limits it to only detecting differences in the model's understanding of the local splice site motif sequences. This could be overcome by looking at comparisons between differences in predictions with mutants directly and considering non-local mutants that cause differences in splicing predictions.

We agree that two loci are insufficient to demonstrate preservation of non-local effects. To address this, we have extended our analysis to a larger set of sites: we randomly sampled 100 donor and 100 acceptor sites, applied our ISM procedure over a 5,001 nt window centered at each site for both models, and computed the ISM map as before. We then calculated the Pearson correlation between the collection of OSAI_MANE and SpliceAI ISM importance scores. We also created 10 additional ISM maps similar to those in Figure 6A, which are now provided in Figure S23.

Follow is the revised paragraph in the manuscript’s Results section:

First, we recreated the experiment from Jaganathan et al. in which they mutated every base in a window around exon 9 of the U2SURP gene and calculated its impact on the predicted probability of the acceptor site. We repeated this experiment on exon 2 of the DST gene, again using both SpliceAI and OSAI_MANE . In both cases, we found a strong similarity between the resultant patterns between SpliceAI and OSAI_MANE, as shown in Figure 6A. To evaluate concordance more broadly, we randomly selected 100 donor and 100 acceptor sites and performed the same ISM experiment on each site. The Pearson correlation between SpliceAI and OSAI_MANE yielded an overall median correlation of 0.857 (see Methods; additional DNA logos in Figure S23). 

To characterize the local sequence features that both models focus on, we computed the average decrease in predicted splice-site probability resulting from each of the three possible singlenucleotide substitutions at every position within 80bp for 100 donor and 100 acceptor sites randomly sampled from the test set (Chromosomes 1, 3, 5, 7, and 9). Figure 6B shows the average decrease in splice site strength for each mutation in the format of a DNA logo, for both tools.

We added the following text to the Methods section:

Concordance evaluation of ISM importance scores between OSAI_MANE and SpliceAI

To assess agreement between OSAI_MANE  and SpliceAI across a broad set of splice sites, we applied our ISM procedure to 100 randomly chosen donor sites and 100 randomly chosen acceptor sites. For each site, we extracted a 5,001 nt window centered on the annotated splice junction and, at every coordinate within that window, substituted the reference base with each of the three alternative nucleotides. We recorded the change in predicted splice-site probability for each mutation and then averaged these Δ-scores at each position to produce a 5,001-score ISM importance profile per site.

Next, for each splice site we computed the Pearson correlation coefficient between the paired importance profiles from ensembled OSAI_MANE and ensembled SpliceAI. The median correlation was 0.857 for all splice sites. Ten additional zoom-in representative splice site DNA logo comparisons are provided in Supplementary Figure S23.

(2) The utility of the calibration method described is unclear. When thinking about a calibrated model for splicing, the expectation would be that the models' predicted splicing probabilities would match the true probabilities that positions with that level of prediction confidence are splice sites. However, the actual calibration that they perform only considers positions as splice sites if they are splice sites in the longest isoform of the gene included in the MANE annotation. In other words, they calibrate the model such that the model's predicted splicing probabilities match the probability that a position with that level of confidence is a splice site in one particular isoform for each gene, not the probability that it is a splice site more broadly. Their level of calibration on this set of splice sites may very well not hold to broader sets of splice sites, such as sites from all annotated isoforms, sites that are commonly used in cryptic splicing, or poised sites that can be activated by a variant. This is a particularly important point as much of the utility of SpliceAI comes from its ability to issue variant effect predictions, and they have not demonstrated that this calibration holds in the context of variants. This section could be improved by expanding and clarifying the discussion of what set of splice sites they have demonstrated calibration on, what it means to calibrate against this set of splice sites, and how this calibration is expected to hold or not for other interesting sets of splice sites. Alternatively, or in addition, they could demonstrate how well their calibration holds on different sets of splice sites or show the effect of calibrating their models against different potentially interesting sets of splice sites and discuss how the results do or do not differ.

We thank the reviewer for highlighting the need to clarify our calibration procedure. Both SpliceAI and OpenSpliceAI are trained on a single “canonical” transcript per gene: SpliceAI on the hg 19 Ensembl/Gencode canonical set and OpenSpliceAI on the MANE transcript set. To calibrate each model, we applied post-hoc temperature scaling, i.e. a single learnable parameter that rescales the logits before the softmax. This adjustment does not alter the model’s ranking or discrimination (AUC/precision–recall) but simply aligns the predicted probabilities for donor, acceptor, and non-splice classes with their observed frequencies. As shown in our reliability diagrams (Fig. S16-S22), temperature scaling yields negligible changes in performance, confirming that both SpliceAI and OpenSpliceAI were already well-calibrated. However, we acknowledge that we didn’t measure how calibration might affect predictions on non-canonical splice sites or on cryptic splicing. It is possible that calibration might have a detrimental effect on those, but because this is not a key claim of our paper, we decided not to do further experiments. We have updated the manuscript to acknowledge this potential shortcoming; please see the revised paragraph in our next response.

(3) It is difficult to assess how well their calibration method works in general because their original models are already well calibrated, so their calibration method finds temperatures very close to 1 and only produces very small and hard to assess changes in calibration metrics. This makes it very hard to distinguish if the calibration method works, as it doesn't really produce any changes. It would be helpful to demonstrate the calibration method on a model that requires calibration or on a dataset for which the current model is not well calibrated, so that the impact of the calibration method could be observed.

It’s true that the models we calibrated didn’t need many changes. It is possible that the calibration methods we used (which were not ours, but which were described in earlier publications) can’t improve the models much. We toned down our comments about this procedure, as follows.

Original:

“Collectively, these results demonstrate that OSAIs were already well-calibrated, and this consistency across species underscores the robustness of OpenSpliceAI’s training approach in diverse genomic contexts.”

Revised:

“We observed very small changes after calibration across phylogenetically diverse species, suggesting that OpenSpliceAI’s training regimen yielded well‐calibrated models, although it is possible that a different calibration algorithm might produce further improvements in performance.”

Reviewer #2 (Public review):

Summary:

The paper by Chao et al offers a reimplementation of the SpliceAI algorithm in PyTorch so that the model can more easily/efficiently be retrained. They apply their new implementation of the SpliceAI algorithm, which they call OpenSpliceAI, to several species and compare it against the original model, showing that the results are very similar and that in some small species, pretraining on other species helps improve performance.

Strengths:

On the upside, the code runs fine, and it is well documented.

Weaknesses:

The paper itself does not offer much beyond reimplementing SpliceAI. There is no new algorithm, new analysis, new data, or new insights into RNA splicing. There is no comparison to many of the alternative methods that have since been published to surpass SpliceAI. Given that some of the authors are well-known with a long history of important contributions, our expectations were admittedly different. Still, we hope some readers will find the new implementation useful.

We thank the reviewer for the feedback. We have clarified that OpenSpliceAI is an open-source PyTorch reimplementation optimized for efficient retraining and transfer learning, designed to analyze cross-species performance gains, and supported by a thorough benchmark and the release of several pretrained models to clearly position our contribution.

Reviewer #3 (Public review):

Summary:

The authors present OpenSpliceAI, a PyTorch-based reimplementation of the well-known SpliceAI deep learning model for splicing prediction. The core architecture remains unchanged, but the reimplementation demonstrates convincing improvements in usability, runtime performance, and potential for cross-species application.

Strengths:

The improvements are well-supported by comparative benchmarks, and the work is valuable given its strong potential to broaden the adoption of splicing prediction tools across computational and experimental biology communities.

Major comments:

Can fine-tuning also be used to improve prediction for human splicing? Specifically, are models trained on other species and then fine-tuned with human data able to perform better on human splicing prediction? This would enhance the model's utility for more users, and ideally, such fine-tuned models should be made available.

We evaluated transfer learning by fine-tuning models pretrained on mouse (OSAI_Mouse), honeybee (OSAI_Honeybee), Arabidopsis (OSAI_Arabidopsis), and zebrafish (OSAI_Zebrafish) on human data. While transfer learning accelerated convergence compared to training from scratch, the final human splicing prediction accuracy was comparable between fine-tuned and scratch-trained models, suggesting that performance on our current human dataset is nearing saturation under this architecture.

We added the following paragraph to the Discussion section:

We also evaluated pretraining on mouse (OSAI_Mouse), honeybee (OSAI_Honeybee), zebrafish (OSAI_Zebrafish), and Arabidopsis (OSAI_Arabidopsis) followed by fine-tuning on the human MANE dataset. While cross-species pretraining substantially accelerated convergence during fine-tuning, the final human splicing-prediction accuracy was comparable to that of a model trained from scratch on human data. This result indicates that our architecture seems to capture all relevant splicing features from human training data alone, and thus gains little or no benefit from crossspecies transfer learning in this context (see Figure S24).

Reviewer #1 (Recommendations for the authors):

We thank the editor for summarizing the points raised by each reviewer. Below is our point-bypoint response to each comment:

(1) In Figure 3 (and generally in the other figures) OpenSpliceAI should be replaced with OSAI_{Training dataset} because otherwise it is hard to tell which precise model is being compared. And in Figure 3 it is especially important to emphasize that you are comparing a SpliceAI model trained on Human data to an OSAI model trained and evaluated on a different species.

We have updated the labels in Figures 3, replacing “OpenSpliceAI” with “OSAI_{training dataset}” to more clearly specify which model is being compared.

(2) Are genes paralogous to training set genes removed from the validation set as well as the test set? If you are worried about data leakage in the test set, it makes sense to also consider validation set leakage.

Thank you for this helpful suggestion. We fully agree, and to avoid any data leakage we implemented the identical filtering pipeline for both validation and test sets: we excluded all sequences paralogous or homologous to sequences in the training set, and further removed any sequence sharing > 80 % length overlap and > 80 % sequence identity with training sequences. The effect of this filtering on the validation set is summarized in Supplementary Figure S7C.

Reviewer #3 (Recommendations for the authors):

(1) The legend in Figure 3 is somewhat confusing. The labels like "SpliceAI-Keras (species name)" may imply that the model was retrained using data from that species, but that's not the case, correct?

Yes, “SpliceAI-Keras (species name)” was not retrained; it refers to the released SpliceAI model evaluated on the specified species dataset. We have revised the Figure 3 legends, changing “SpliceAI-Keras (species name)” to “SpliceAI-Keras” to clarify this.

(2) Please address the minor issues with the code, including ensuring the conda install works across various systems.

We have addressed the issues you mentioned. OpenSpliceAI is now available on Conda and can be installed with:  conda install openspliceai. 

The conda package homepage is at: https://anaconda.org/khchao/openspliceai We’ve also corrected all broken links in the documentation.

(3) Utility:

I followed all the steps in the Quick Start Guide, and aside from the issues mentioned below, everything worked as expected.

I attempted installation using conda as described in the instructions, but it was unsuccessful. I assume this method is not yet supported.

In Quick Start Guide: predict, the link labeled "GitHub (models/spliceai-mane/10000nt/)" appears to be incorrect. The correct path is likely "GitHub (models/openspliceaimane/10000nt/)".

In Quick Start Guide: variant (https://ccb.jhu.edu/openspliceai/content/quick_start_guide/quickstart_variant.html#quick-startvariant), some of the download links for input files were broken. While I was able to find some files in the GitHub repository, I think the -A option should point to data/grch37.txt, not examples/data/input.vcf, and the -I option should be examples/data/input.vcf, not data/vcf/input.vcf.

Thank you for catching these issues. We’ve now addressed all issues concerning Conda installation and file links. We thank the editor for thoroughly testing our code and reviewing the documentation.

tu asesor, tienes la posibilidad de crear tu libro de proyecto para compartirlo con tu asesor.

En ella encontrarás todas las instrucciones, paso a paso, para montar la cocina. Una herramienta fácil y completa, con toda la información que necesitas saber y tener en cuenta para el montaje. Ir a la guía

El artículo explica que no se avanza de forma lineal, sino que se retrocede constantemente para replantear preguntas, ajustar objetivos o redefinir enfoques. Esto refleja que la investigación no es estática, sino dinámica, y permite ver los retrocesos como parte del aprendizaje y no como errores definitivos.

El texto subraya que la elección del tema de investigación es el punto de partida esencial en cualquier proyecto académico. No se trata únicamente de seleccionar algo que interese al estudiante, sino de considerar factores como la disponibilidad de bibliografía, la comprensión de los conceptos implicados y la factibilidad metodológica. De esta manera, la investigación no solo será significativa en lo personal, sino también realizable en la práctica y con valor científico.

Otra parte que me parece sumamente interesante es la manera en la que el texto nos indica como se debe plantear una pregunta, el cual es uno de los pasos principales al comenzar una investigación. Se menciona que las principales preguntas son demasiado abstractas y generales, ya que al comenzar solo se tiene una pequeña noción sobre el tema que se quiere investigar, sin embargo es de suma importancia que estas preguntas iniciales se enfoquen a la búsqueda de una respuesta más puntual, es decir, debemos aprender a realizar preguntas más especificas y especializadas, debido a que está pregunta funcionará como uno de los puntos principales sobre los cuáles vamos a comenzar a investigar.

La representación gráfica de como debe ser un correcto proceso de investigación me parece sumamente útil, pues muchas veces pensamos y concebimos a la investigación como un proceso lineal, en el cual se busca solucionar una problemática, para posteriormente realizar una investigación y obtener una respuesta, sin embargo la figura nos muestra que muchas veces el proceso de investigación puede ser repetitivo, pues a medida que vamos investigando y obteniendo más información, podemos ver el tema de investigación con un enfoque más crítico, generando nuevas interrogantes, hipótesis, planteamientos y conclusiones.

En base a mi propia experiencia coincido completamente con lo que expresa el texto, pues muchas veces el problema al realizar el desarrollo de una investigación no es la investigación en si (ya que al plantear una metodología clara se puede tener una mejor noción sobre como se va a trabajar y los objetivos que se busca alcanzar) pues el principal problema es plantear la problemática inicial de la investigación, así como encontrar soluciones creativas, originales e innovadoras que puedan dar una base solida y sustento a nuestra investigación.

Dice que plantear el problema de investigación no es solo tener una idea, sino estructurarla de manera clara, concreta y organizada. Me parece muy útil la recomendación de formular preguntas relacionadas con el tema, jerarquizarlas y construir una red conceptual, porque así se clarifica el objeto de estudio y se delimita la investigación.

Ayuda a entender que investigar es un proceso ordenado y reflexivo, donde identificar la situación problemática y formular el problema correctamente es fundamental. Me parece interesante que resalte que todo conocimiento científico surge de preguntas claras y que, sin una buena formulación del problema, es difícil avanzar en la investigación de manera sólida.

Por eso es importante conocer la historia para poder entender los problemas, buscar soluciones y no repetir los errores cometidos en el pasado.

Es importante tener claro el tema, el problema y la metodología en una investigación. Nos podemos llegar a confundir porque hay muchas formas de hacer investigación, pero al final todo depende de plantear bien el problema y seguir un proceso ordenado. También me parece interesante el proceso en espiral de la investigación donde se aprende y se mejora constantemente.

En la parte del texto donde se dice: "Antes, tiene que for mular el problema específico en términos concretos y explícitos", es decir que debes de conocer primeramente que es lo que estas buscando con exactitud y en qué grado. Cómo resultado encontraras una respuesta más satisfaria y aterrizada a lo que encuentras.

Es muy importante tomas en cuentas estas tres dimensiones cuando se quiere investigar ya que los seres humanos somos seres sociales y estas dimensiones nos pueden ayudar mucho a entender como nos relacionamos y a su vez nos permite entender el porque de muchos problemas que se originan.

Una de las principales ventajas que adquieres al investigar es conocer la realidad y los antecedentes que preceden los sucesos. Conocer todo esto es una ventaja, principalmente para no cometer los mismos errores que las personas que nos prescindieron, si no que tambien es una forma de impulsarnos a tomar una perspectiva diferente a las problemáticas que se pudieran encontrar en función a lo previamente establecido.

El propósito es la respuesta a la pregunta ¿Para qué?

No sabia el significado de epistemología, así que la busque, significa ciencia que indaga lo científico, reflexionando la profundidad del conocimiento, su origen, su forma y como debería ser.

Creo que una de las fundamentales cosas que se tiene que hacer a la hora de investigar un tema es conocer del mismo, como resultado de eso se tiene que hacer una investigación previa en la cual puedas conocer todo acerca del tema, no es necesario volverse un experto, pero tienes que entender de lo que estás hablando para saber qué es lo que estás haciendo y el porque de tu investigación.

* one who = quien. * Se usa one por que volviste a mencionar la misma persona y who no puede estar después del verbo to be.

* Dumbass = Idiota, de falta de inteligencia. * proof = prueba. * ya (contracción) de you * talking all high and mighty like that (modismo) = hablar con tanto aire de superioridad.

In this sentence Sandra cisneros tell us her father love her ,he said to her daughter Que bueno ,mi ha that meant a lot to me , his thought about her study and her life patener also .she was lucky because ,she was bless with caring father.

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

Learn more at Review Commons

Reply to the reviewers

Response to reviewers

We thank the reviewers for their constructive feedback, which has greatly improved the clarity and rigor of our manuscript. We have carefully addressed each comment below, indicating changes made to the text, figures, or supplementary material where appropriate. References to line numbers correspond to the revised version of the manuscript.

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

* In this paper, the authors focus on the role of Reticulon-1C in concert with Spastin in response to axonal injury. In data mining, they find axonal mRNAs encoding for ER-associated proteins including Rtn-1. They establish a knockdown targeting both Rtn-1 isoforms Rtn-1A and Rtn-1C. They observe decreased beta-3-Tubulin levels in the soma while axonal protein levels are unchanged. In microfluidic devices, they characterise the effect of a compartment-specific Rtn-1 KD on axonal outgrowth in the axonal compartment. The authors quantify axonal outgrowth, seeing increased outgrowth in an axonal compartment-specific Rtn-1 KD, while the effect seems to be reversed when applying the KD construct in the somatic compartment. When focussing on the axonal growth cone, they find the Rtn-1 KD shows differences in several morphological features of the growth cone. They find an increase in Tubulin levels in an axonal compartment-specific, but a decrease in a somatic compartment-specific Rtn-1 KD. Colocalisation of Rtn-1C and Spastin is shown to be monolaterally increased following axotomy. Combining axotomy with the Rtn-1 KD shows increases in dynamic microtubule growth rates and track lengths. In another model system, neuron balls, they show Rtn1-C, but not Rtn1-A to be present in the axon. In a puro-PLA assay they also show it can be synthesised in the axonal compartment. To investigate the mechanism enabling the cooperation between Spastin and Rtn-1C, they move to a cell line model in which they see a correlating distribution between Spastin and Rtn-1C but not Rtn-1A. Finally, they use in silico modelling to speculate on binding between Spastin domains and Rtn-1 isoforms.*

Major comment:

The rationale behind the work is convincing, however some interpretations are presented as more robust than some data allow. Most notably, while the interaction between Rtn-1 and Spastin has been shown prior to this study, it is only presented here through in silico analysis. In figure 5, an increase in the growth rate of dynamic microtubules is observed in either a Rtn-1C KD or by using a Spastin-inhibitor. Due to a described increase in colocalisation between Rtn-1C and Spastin (5A), the increase in growth rate is displayed as caused by Rtn-1 promoting Spastin's severing ability. This result might however be correlative. Further in the injured samples, Spastin-levels seemingly increase (in the representative images) and it is thus not surprising that the level of Rtn-1C colocalising with Spastin increases as well. This might not be indicative of a cooperation and further experimental evidence are required.

R: We thank the reviewer for this thoughtful comment. We agree that our interpretation should be more cautious, and we have revised the Title, Results and Discussion sections accordingly. In particular:

1. Following yours and other reviewer comments, we have analyzed a new set of experiments regarding the STED images of non-injured and injured axons. To eliminate the risk of artifactual descriptions, we have avoided deconvolution and worked directly with raw STED images (Figure 5A). Under these conditions, the distribution of Spastin and its intensity in distal axons are not modified by injury, nor those of Rtn-1C and Spastin (Supplementary figure 4). We emphasize in the revised text that the in silico modeling we present is supportive, but not definitive, of a direct interaction. To address this concern, we clarify that our study builds on prior evidence of biochemical interaction between Rtn-1C and Spastin (Mannan et al., 2006), and that our own data demonstrate: i) compatible subcellular distribution in axons by super-resolution (STED microscopy, Figure 5A);ii) a potential functional interplay in axons (rescue of β3-tubulin levels by Spastin inhibition, Figure 5B), and iii) isoform-specific co-distribution with Spastin in heterologous cells that is associated with changes on microtubule integrity (see improved Figure 7). Together, these results go beyond correlative localization, but we acknowledge that they do not directly demonstrate a molecular complex in axons. Thus, we now indicate that "Although we did not directly test their molecular association, these results are consistent with Rtn-1C and Spastin sharing a similar subcellular localization, potentially enabling their functional interaction in distal axons" (lines 285-287)

We would like to clarify a possible misunderstanding: in our experiments, the increase in microtubule growth rate was observed after axonal Rtn-1 KD. Spastazoline (SPTZ) only prevented the reduction in β3-tubulin levels induced by Rtn-1 KD, while leaving the KD-driven increase in growth rate and track length unaffected (Figures 5B-E). Thus, our interpretation is that axonal Rtn-1 KD correlates with increased Spastin function. (lines 307-309)

Other comments:

• Generally, graphs would benefit from individual values plotted as well as the summary. Font sizes and types (but rarely) are sometimes inconsistent. Proteins should be consistently written (capitalised or not).

__R: __ We agree with the reviewer and thank for taking the time for noticing these inconsistencies as it significantly affects the quality of the work. We have improved several figures and added graphs plotting individual values (Figures: 2 C, 2E; 4 (A-E); 5E; 6D). We have reviewed the Font size and types more carefully and capitalized the proteins accordingly.

• *Table 1 and figure 1 present data collected from a vast amount of resources. It should be highlighted that datasets from which data was obtained includes many different models, different DIVs and neuronal cell types. Figure 1B may benefit from a different colour scheme. "Ex-vivo" should be "Ex vivo". For "ER mRNAs are a relevant category" it is not described what "relevant" would mean in this context. The title might remove this small part or describe it in the text. It should be described how it is decided that mRNAs are "common". *

• *

__R: __We have now highlighted in the result section the diverse origins of the analyzed samples; We removed the indicated part from the text and explained that common mRNAs were chosen based on the Benjamini-Hochberg (Ben) analysis. (Page 33, lines 1299-1304).

* - Figure 2: add description to y-axis to describe what fold change is displayed, applies to multiple figures. Will improve readability of the figures. In 2C, the ROI showing neuronal somata should be increased to show part of the axon and not cut off the soma.*

• *

__R: __We thank the reviewer for taking the time to highlight this. We have included this modification in figure 2 and throughout the article. We have also enlarged the indicated ROIs in figure 2C as requested. (Page 34)

• *Figure 3: Three out of four axonal compartments seem to be comprised of dying or damaged axons. Especially the axonal KD scrambled image. It should be ensured that neuronal cultures are healthy. *

• *

__R: __We completely agree with the reviewer that the selected images were not describing the general good health of axons which has been accredited by the lack of fragmentation and functional responsiveness shown in (Figure 4 and 5 B, C, E). Thus, we have now replaced the previous axonal fields by more representative ones (Figure 3). (page 36)

• *

Typo in "intersections". The schematic of 3B is a great addition to explain the graphs above. Perhaps it could be a bit refined as it is currently hard to see whether this is a neuron or a growth cone without context. Maybe show where the axon connects to the depicted growth cones and change the third icon which looks like it was crossed out. Small formatting issues: remove additional space bar before "Figure 3." And add after "Bar"

__R: __Many thanks for these great suggestions. We have now improved the figures as suggested and changed the indicated formatting issues. (page 36)

- Figure 4: If not misunderstanding what is depicted, in 4A and B, different lookup tables are used to depict the same signal. Only one of each images is necessary. Do the axons have more tiny branches in the Rtn-1 KD condition in 4A? Unclear why Rtn-1 levels are increased in the Rtn-1 KD (4C), please clarify.

• *

__R: __We thank the reviewer for these observations. The reviewer is correct that different lookup tables were initially applied to the same image. Our intention was to highlight the fine distribution of axonal Rtn-1, but since this aspect is already clearly shown in previous figures, we now retain only a single lookup table. The appearance of tiny branches in the Rtn-1 KD condition represents an isolated observation and does not reflect a consistent or robust phenotype associated with Rtn-1 KD.

As the reviewer points out, the increase of Rtn-1 in the cell bodies of injured neurons following axonal KD was initially surprising to us. However, this was a consistent phenomenon, as shown in the improved Figure 4. Of note, previous studies have reported that total Rtn-1C (but not Rtn-1A) levels increase in response to injury in cortical neurons(Fan et al., 2018). In our case, we interpret this as a compensatory somatic response triggered by the local reduction of Rtn-1 in injured axons. This interpretation is also consistent with the apparent lack of effect of siRNA on distal axonal Rtn-1 levels when applied locally after injury (while somatic application of the same siRNA does reduce axonal Rtn-1). Thus, after 24 hours of KD, the somatic upregulation of Rtn-1 may partially compensate for its expected local synthesis decrease. We have clarified this assumption in the revised text. (lines 247-251)

- Figure 5: It may be easier to understand what "axotomy" samples are if just referred to as "injured" as later in the same figure. The procedure could also very briefly be explained in the results. 5C should depict AUC in µm2 not µm. 5D Spastin is barely visible, brightness and contrast should be adjusted to enhance visibility.

• *

__R: __We thank the reviewer for these helpful suggestions and have implemented the requested changes in Figure 5. Specifically:

We now consistently refer to "axotomy" samples as "injured" throughout the figure and article. In addition, a brief explanation of the axotomy procedure has been added before Figure 2 and before figure 5, also the description has been clarified in Materials and methods. (lines 191-192) and (lines 289-290) and (lines 779-787)

To improve the reproducibility of our outgrowth measurements, we revised this analysis approach. Based on previous work from a co-autor (McCurdy et al., 2019), instead of reporting the "relative number of intersections," we now present the total counts obtained from Sholl analysis of binarized axons (see Methods). To this end, we took advantage of the NeuroAnatomy plugin of FIJI, which more precisely tracks axon trajectories and makes the measurement more independent of axon width. Also, this new approach avoids the conflict we had with what we considered the "first line" after the groove ends, which was a bit of arbitrary. Accordingly, the correct term is now "summation of intersections (sum.)" at different distance bins, as reflected in Figure 5D. (page 40)

For the former Figure 5D (now Figure 5B), we have improved the acquisition of representative images and applied a different set of lookup tables to enhance visibility. (page 40)

- Figure 6: It should be made clear why it is necessary to switch to another model system just for 6A, please indicate this in the text. PCR bands seem very pixelated, check the quality. It is unclear why soma genes/proteins were only tested with either PCR or WB others with both. Rtn-1C and Rtn1-A should be presented in the same order in the PCR and WB panel. Correct "Rtn1-1A" typo. In 6D, 1.5 dots per soma seems like a low number. When normalized to the area the soma vs the axon occupies, the compartmentalization does not work? Maybe it makes sense to refine analysis or apply puromycin in the somatic compartment and analyze the axonal compartment as comparison?

__R: __Many thanks for these observations. We have now included the following clarification in the text: "We sought to characterize the isoform expression of Rtn-1 mRNA and protein in both axons and cell bodies. Because microfluidic chambers yield only limited cellular material, we adopted an alternative culture approach using 'neuronballs.' This method enables the segregation of an axon-enriched fraction by mechanically separating axons from somato-dendritic structures" (lines 375-376).

The resolution of PCR bands has been improved in the revised figure. Note that because the amount of cellular material is relatively scarce, we did not obtain too strong bands.

The difference in the genes/proteins used for characterizing RNA and protein samples reflects our intention to treat both approaches as complementary. The PCR markers were primarily included to confirm sample purity, which also applies to the WB samples since they derive from the same preparation. In both assays, we used MAP2 as a dendritic marker to demonstrate axonal purity. While we acknowledge that the same genes could have been tested by both methods, we believe the results as presented adequately demonstrate the effective isolation of axons.

We have switched the order of Rtn-1C/1A for consistency across PCR and WB panels and corrected the indicated typo in Figure 6A.

We agree with the reviewer that an average of 1.5 puncta per soma initially appeared low. We have identified at least three reasons for this:

First, the signal derives from only a 15-minute puromycin pulse, which is a very short labeling window. Second, our puro-PLA assay is particularly stringent, as ligation relies directly on puromycin- and Rtn-1C-labeled primary antibodies, without the additional spacing normally introduced by secondary antibodies. In standard PLA, the critical distance for amplification is ~30-40 nm, whereas in our assay this distance is even more restrictive. Third, in our initial analysis we applied an overly cautious threshold to define "true" amplification. We have now refined this threshold using a baseline defined by the absence of puromycin stimulation. With this improved criterion, we now quantify an average of ~5 puncta per soma and ~10 puncta per 1000 µm² of axonal area (Figure 6D and Supplementary Figure 3D). Assuming a neuronal soma diameter of 15 µm (area ≈ 176.71 µm²), this yields ~0.028 puncta per µm² in soma. In comparison, axons display ~0.01 puncta per µm², approximately one-third of the soma value, which is compatible with the idea thar cell bodies dominate neuronal protein synthesis.

Following the reviewer's valuable suggestion, we performed additional quantifications in which puromycin was applied exclusively to the somatic compartment. Under these conditions, we still observed amplification in axons (~5 puncta per 1000 µm²), although this value was significantly lower than when puromycin was applied directly to axons. This analysis provided a novel appreciation of the puro-PLA technique in neurons: at least half of the signal originates in the axonal compartment, while a portion may reflect proteins synthesized in soma and transported anterogradely to the axon through yet-unknown mechanisms (potentially involving rapid anterograde transport) (Figure 6D). (page 42)

• Figure 7: 7A shows two images depicting the same information that may not be needed. Can probably be removed. In 7B there is no negative (or any) correlation between Spastin levels and Tubulin, however later it is mentioned that Rtn-1C transports Spastin thus causing a decrease in Tubulin at certain locations? It is nclear if Spastin levels vary intensely between different samples. Mean intensity of the somatic area may be beneficial to rule this out. 7B Tubulin on the right top panel seems to have a decrease in Tubulin levels which is not visible due to the Y axis of Tubulin being set to a different range than the middle and lower panel. The average of line scans from multiple cells may be helpful to determine whether there is indeed no colocalization between Rtn-1A and Spastin. The provided representative images seem to show similar degrees of colocalization between Spastin and Rtn-1A/C.

• *

__R: __We thank the reviewer for these valuable observations and acknowledge that Figure 7 may have caused confusion. We have eliminated the fluorescence line-scan traces, as they can be biased depending on the region of the cell analyzed. Although this may not have been sufficiently emphasized in the text, we had already performed a quantitative colocalization analysis across multiple cells and independent experiments, using Mander's coefficients (Figure 7B). These analyses showed higher colocalization between Rtn-1C and Spastin compared to Rtn-1A. Regarding the concerns about variability in Spastin levels or possible bias from Y-axis scaling, we have eliminated those traces by the risk of bias. Also, we had already quantified the total tubulin fluorescence intensity across all the z-stacks and from multiple cells from independent experiments as shown in Figure 7C. To further rule out artifacts caused by variable transfection efficiency, we quantified total fluorescence intensity in both RFP and GFP channels across conditions. As shown in Supplementary Figure 6, no significant differences were observed, suggesting that the changes in tubulin reflect specific effects of Spastin/Rtn-1C co-expression rather than variability in expression levels.

Results: - It would be helpful to reiterate the hypothesis at the start to ease the reading flow.

__ R: __Many thanks, we have introduced a line reiterating the hypothesis as suggested (lines 117-118)

- There seems to be minor redundancy in lines 132-138.

• *

__R: __Indeed, we have now removed the indicated phrase.

• There are several spellings, proof-reading is recommended. For example, in line 136 should be "promotes". 160 "localla", 192 should be "the actin cytoskeleton".,194 should be "we first examined", 195 should be "Different", 223 "using", 259 "axons".

__R: __We apologize for the spellings; we have now performed a careful proof-reading and introduced these corrections.

- 154-155: Unclear, why the lower MW Rtn-1C was seen as more important.

__R: __We apologize for not being clear enough. It is not necessarily more important, but we just took the Rtn-1C molecular weight as reference for the analysis considering that this isoform is the predominant in axons. In any case we have found a significant effect for both isoforms at least on siRNA 2 (data not shown), which is now expressed in the text (line 165-169) : "We also examined the 180 kDa band and found that siRNA 1 reduced expression to a mean of 0.41 relative to Scr, showing a strong trend that did not reach statistical significance (p = 0.05; N = 3; Wilcoxon test compared to 1, data not shown). In contrast, siRNA 2 further reduced expression to a mean of 0.29, which was statistically significant (p = 0.04; N = 3; Wilcoxon test compared to 1, data not shown)."

- 167 results of 2E not stated before interpreting them.

• *

__R: __We have corrected this mistake.

- 181 would suggest "outline" instead of "perimeter".

• *

__R: __We have considered this suggestion and included "outline", nevertheless the morphometric parameter is defined as perimeter, so we retained the term, but with the suggested clarification.

• *

- 183-184 "longest shortest path" is a confusing term.

__R: __We agree that it is a confusing term, thus have now introduced multiple clarifications for the term in the leyend of figure 3 (page 36), and with more detail in a new section of Materials and methods (lines 697-699).

• figure 4B should be referenced earlier in the sentence.

__R: __We have corrected the sentence in the text.

- 243-244 may be correlation. Rtn-1 and Spastin do not necessarily interact so that this result is achieved.

• *

__R: __Thanks for the clarification, we are aware that so far in the manuscript the conclusion is not correct, thus now we have stated at the end of the paragraph: "Together, these observations suggest that axonal Rtn-1 KD correlates with higher Spastin microtubule severing" (lines 307-309)

- 246: In figure 1 the KD seemed to influence both Rtn-1 isoforms, why not here anymore? 259 "axons". 284 "counteract" instead of "suppress"?

• *

__R: __We acknowledge the confusion at this point of the article because of measuring a specific isoform. We now indicate that we will focus on Rtn-1C because of previous evidence of the literature pointing to an interaction of Rtn-1C with Spastin (line 264-267). Later we show that Rtn-1C is the predominant isoform in axons (Figure 6). We have corrected all the suggestions in the manuscripts.

- 485: rephrase as the interaction between Rtn-1C with Spastin has not been shown directly in these experiments.

__R: __Many thanks for the relevant clarification. Now, we have corrected:" Here, we have described an emerging mechanism relating Rtn-1C with the activity of Spastin, which is the most frequently mutated isoform in HSP (Hazan et al., 1999; Mannan et al., 2006)." (line 632-634). * Methods: 535 "in PBS". 543 citation error. 689-699 is it necessary to add a gaussian blur?*

• *

__R: __We have corrected the words and removed the wrong reference. Regarding the use of Gaussian blur, it is a very important point. We used this approach because, in our experimental conditions, it was critical to highlight moving particles that otherwise would go unnoticed by the noise. This was particularly manifest for the seemingly more "unorganized" movements of axonal microtubules after injury.

References: Mannan, A U et al. appears twice in the citation list (36 and 44).

* *R: Many thanks for the observation. Now we have corrected it.

Reviewer #1 (Significance (Required)):

Overall, this manuscript describes novel fundings which will be interesting to the neuronal cell biology community and scientists working on the field of neuronal injury and regeneration. It is well structured, and the data are mostly well presented but sometimes conclusions are over-interpreted. However, several points need to be addressed in a more convincing way.

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

Axonal mRNA localization and localized translation support many neuronal functions and is an important determinant of the regenerative potential of axons after injury. How this works mechanistically remains unclear. The authors present a well performed and technically challenging study in which they identify RTN-1 as a regulator of axonal outgrowth after injury. They provide evidence using experiments in microfluidic chambers that RTN1 is locally synthesized in axons. Interestingly, they identify a (local) interplay between RTN1 and Spastin which affects microtubules and thereby regulates the outgrowth of cortical axons after injury. This study provides an interesting new link between a locally synthesized protein (RTN1) and a microtubule-regulating protein Spastin that is changed upon axon injury. This provides an advance in our understanding in axon regeneration after injury and provides the basis for new studies that can further investigate this interplay. Although interesting, I have several concerns that should be clarified and are needed to substantiate the findings and model presented in this study.

Major concerns:

2. In figure 1, the authors provide an analysis of overlapping axonal mRNAs. There are more axonal transcriptome studies and a recent study by von Kugelgen and Chekulaeva (2020; doi: 10.1002/wrna.1590) already performed such an analysis, which included more studies. It would be good to mention this. It can be perceived that studies were now chosen to get the outcome that Rtn-1 is present in all studies. For example, von Kugelgen finds mRNA coding for RTN3, another ER structural protein, as present in 16 out of 20 studies analyzed. That said, the authors present more reasons to look at Rtn-1, so the selection to continue with this protein remains valid but can be written up differently so not to present it as the 'sole' ER-shaping protein consistently present in axonal transcriptomes. __R: __We appreciate this important observation to enrich the article; we are aware that the transcriptome data can be even further expanded to more recent studies. Thus, we have now included this reference in the main text and highlighted the relevant finding of RTN3. However, Kugelgen and Chekulaeva used data from dendrites/axons (neurites). Thus, we indicate that "...On a similar approach, but combining data from dendrites and axons, it was found that Reticulon-3 *mRNA is present in 16 out of 20 studies, further suggesting a wider presence of other mRNAs coding for ER structural proteins in axons " (line 128-131)

3. The description of methods is currently insufficient and incomplete and does not allow for reproducibility of this study. For example, different Rtn-1 antibodies seem to be used in this study. Is the same antibody used for staining and WB? There is no listing of any of the antibodies used in the study and which one is used for which technique/experiment. This should be clarified and should be easy to do so in the methods section (antibody name, origin/company, dilution used) to enhance reproducibility of this study. This is not limited to primary antibodies and any information on secondary antibodies, including what was used for STED is completely missing.*

4. *

__R: __Thanks for these critical comments. First, we apologize for the former method version which was mistakenly not as accurate as it should. We have now revisited it and improved several points throughout this section. Regarding the use of primary and secondary antibodies, plasmids, siRNAs, and general reagents, they are all indicated in the Supplementary material, including company and dilution ("Reagent tables").

• The timeline of KD experiments in Figures 2 and 3 are unclear. For the Western blot KD is performed at DIV7 and collected 48 hours later. However, this is not specified for the stainings done in Figure 2C-E. Is this also at DIV7 and then for 48 hours? In figure 3 the siRNA is added at DIV8 (together with axotomy) and outgrowth is measured 24 hours later. Is 24 hours sufficient to achieve knockdown? Is this also what was done for stainings? Later on in Figure 5B, 48 hours of KD is again used. It is unclear what the rationale of these differing timepoints is. Why was this chosen? Is the timeline also the reason for the difference in segment lengths chosen? In Figure 3, there is a significant effect on outgrowth in the KD in the 'mid-range' which is not present in Figure 5.*

__R: __We regret the confusion, now all this information is explicitly clarified in the main text (lines 297-299) and the corresponding figure legends. We have strong reasons to have used these different time points. Figure 2 A-B is aimed at validating the siRNA against Rtn-1 thus we treated 7 DIV cultures for 48 hours to be sure of revealing a global effect by WB. In figure 2 C-D, we used the same 7 DIV cultures, but only for 24 hours. The reason for this is that, once the RNAi was validated, we explored its control on local synthesis in a shorter period based in previous literature supporting that axonal KD for 24 hours is sufficient for regulating axonal transcripts (Batista et al., 2017; Gracias et al., 2014; Lucci et al., 2020). We are also confident of using this time point based in the new supplementary figure 3D that shows a significant decrease on puro-PLA signal (indicative of Rtn-1C synthesis) 24 hours after axonal KD.

In figure 3, we performed axotomy thus we had to wait a longer period for axons to grow (8 DIV) before fully cut them out, in this case we performed axonal KD from 8 to 9 DIVs. This is the same period used for the staining and quantifications shown in figure 4. All this is properly clarified in the main text and figures.

In Figure 5 we performed a more challenging experiment that required to transfect cells with an EB3-GFP plasmid, then perform axotomy along with axonal KD as well as pharmacological treatment selectively in axonal compartment. First, we tried to measure microtubule dynamics under the same temporal frame of figure 3. Nevertheless, expression levels of EB3-GFP were not adequate for axonal measurements by live-cell imaging. Therefore, compared to figure 3, we increased the time frame after axotomy 24 hours (from 9 to 10 DIV) by this technical reason, but also to explore whether the changes on tubulin intensity might be revealed more clearly (which was the case, figure 5B). These considerations are now included in the main text

Regarding the significant effect on outgrowth in the KD in the 'mid-range' which is not present in Figure 5. Given that in figure 5D axons are left growing for two days instead of one, the number of intersections and the differences between conditions is modified compared to figure 3, while retaining the overall trends. Note that to improve the reproducibility of our outgrowth measurements, we revised this analysis approach. Based on previous work of a co-autor (McCurdy et al., 2019), instead of reporting the "relative number of intersections," we now present the total counts obtained from the Sholl analysis of binarized axons (see Materials and methods). To this end, we took advantage of the NeuroAnatomy plugin of FIJI, which precisely tracks axon trajectories and makes the measurements more independent of axon width segmentation. Also, this new approach avoids the conflict we had with what we considered the "first line" after the groove ends, which was a bit of arbitrary. Accordingly, the correct term is now "summation of intersections (sum.)" at different distance bins, as reflected in the new Figure 5D.

Could the authors provide a rescue condition for their siRNA (using a siRNA-resistant construct) to show that their siRNA is specific for RTN1. They nicely show the efficiency of the siRNA but not its specificity. This is crucial because if not specific, this will affect a large part of their study. They already have RTN1A and RTN1C constructs available. Such a rescue experiment should ideally also be performed for one or more of their phenotypic experiments, such as the one presented in Figure 3A or 5 to show that the phenotype is really RTN1 dependent. If done by re-expressing either RTN1A or RTN1C, this could provide insightful information on the relevant isoforms.

__R: __We agree with the reviewer that this is a critical point. A major challenge in demonstrating the functional role of axonally synthesized proteins using a KD approach is that the rescue may also need to occur locally. Since axonal Rtn-1 appears to play a distinct role compared to its somato-dendritic counterpart (Figure 3), a siRNA-resistant construct would ideally require an axon-targeting sequence to restore local synthesis. As this is technically demanding, we have not yet been able to perform such an experiment, but we are actively working on identifying the optimal sequence to direct Rtn-1C to axons. Importantly, studies performing axonal KD typically rely on at least two independent siRNA sequences, thereby minimizing the likelihood that a phenotype arises from off-target effects. Thus, we have now validated a third siRNA (siRNA 3), which selectively downregulates Rtn-1C. Then, following the same experimental frame of figure 3, we performed axonal Rtn-1 KD after injury and observed that siRNA 3 also significantly increases the outgrowth of injured axons (Supplementary figure 2). This suggests that, at least this phenotype, is not product of an off-target effect. Complementarily, pharmacological rescue with the Spastin inhibitor SPTZ mitigated both the reduction in distal axonal β3-tubulin and the increase on axon outgrowth, supporting that the observed phenotypes are unlikely to arise from off-target effects. If these effects were due to random interference with unrelated mRNA targets, inhibition of an ostensibly independent target such as Spastin would not be expected to yield such a consistent rescue. Accordingly, SPTZ treatment alone did not increase β3-tubulin, indicating that its action is specifically contingent upon Rtn-1 KD. Taken together, the pharmacological rescue in axons (Figure 5B) and the Rtn-1C/Spastin co-distribution in heterologous cells, which correlates with preserved microtubules (improved Figure 7), provide converging evidence to suggest that Rtn-1C-Spastin interplay may underly the observed phenotypes in axons.

• I find the data presented in Figure 4A/B confusing. Axonal RTN-1 KD does not reduce axonal RTN1 levels, but somatic KD does. I understand that this implies most protein comes from the soma, and the authors indeed present an explanation that increased somatic RTN1 occurs after axonal KD as a compensation mechanism. However, this can also be interpreted that there is no axonal synthesis of RTN1 after injury and axonal KD has indirect or even aspecific effects. Their model depends on this difference. Their data in Figure 6 could provide supporting evidence if it shows RTN1 puro-PLA after injury. Along these same lines, in Figure 6, they nicely include a compartment control for puro-PLA. It therefore seems doable to include a somatic puromycin control for their axonal puro-PLA, to exclude and diffusion/transport of the newly synthesized peptides. This is especially considering two recent papers reporting on this possible phenomenon, although these studies were not performed in neurons.*

__R: __We consider the possibility that after injury there is no axonal Rtn-1 synthesis as a plausible and relevant appreciation. Unfortunately, we could not perform a puro-PLA experiment after injury, which would have provided a more definite answer. However, now we are more confident of regulating Rtn-1 synthesis before injury as supported by a new supplementary figure 3D that shows a significant decrease on puro-PLA signal (indicative of Rtn-1C synthesis) 24 hours after axonal KD. Thus, based on the similar phenotypes observed before and after injury, we consider our results are still compatible with Rtn-1 axonal synthesis being downregulated, but not absent after injury. First, axonal Rtn-1 KD decreased β3-tubulin levels before and after injury according to figure 5B and the improved statistical analysis performed on figure 2E. Similarly, axonal Rtn-1KD significantly increases microtubule growth rate before and after injury according to the current statistical comparisons (Figure 5E). Second, if β3-tubulin decrease was a merely unspecific siRNA targeting, it is unlikely that SPTZ treatment should increase and restore β3-tubulin levels only in the context of axonal Rtn-1 KD (Figure 5B). We have now included these considerations in the discussion (lines 537-543). Although on a different track, the mechanistic relationship between Rtn-1C and Spastin suggested in Figure 7 could make more plausible that a similar phenomenon regarding the control of tubulin levels may occur locally in axons.

Following the reviewer's valuable suggestion, we performed additional quantifications in which puromycin was applied exclusively to the somatic compartment. Under these conditions, we still observed amplification in axons (~4 puncta per 1000 µm²), although this value was significantly lower than when puromycin was applied directly to axons (~10 puncta per 1000 µm²). This analysis provided a novel appreciation of the puro-PLA technique in neurons: at least half of the signal originates in the axonal compartment, while a portion may reflect proteins synthesized in soma and transported anterogradely to the axon through yet-unknown mechanisms (potentially involving rapid anterograde transport). Note that we revised the criteria for detecting true amplification spots based in staining without puromycin, which increased true amplification numbers. Still, these seemingly low values are compatible with reflecting a limited amount of time (only 15´ of puromycin pulse) and the stringent conditions of this experiment in which secondary antibodies were avoided by directly labeling primary ones. This approach makes the classical 30-40nm distance for PLA even narrower, thus reducing signal. In any case, assuming a neuronal soma diameter of 15 µm (area ≈ 176.71 µm²), this yields ~0.028 puncta per µm² in somata. In comparison, axons display ~0.01 puncta per µm², approximately one-third of the soma value, which makes sense for the expected difference in ribosome density.

• In Figure 5A the authors find an increased co-localization (RTN1/Spastin) after axotomy. From their images, it seems that the amount of Spastin is hugely increased, which would by default increase the chance of (random) colocalization of RTN1 on Spastin. Could the authors comment on this?*

__R: __Thanks for this relevant and constructive critique. We formerly based our colocalization analysis on deconvolved images. However, after performing several quantifications through different deconvolution parameters, we were not convinced about the robustness of this finding and the performed staining. Thus, we performed a new set of experiments and found that non-deconvolved images from the STED microscope were more informative about the expected tubular morphology of the axonal ER. Thus, we improved figure 5A, and now the main conclusion is just that both proteins are closely distributed in distal axons before and after injury.

• In figure 5E and 5F, the condition of scr + SPTZ is omitted. What is the reason for this? The explanation of results in these figures is confusing. The authors report a 'clear trend' in increase in comet track length and lifetime upon addition of SPTZ to axonal RTN-1 KD. This is however not significant. The comparisons that are made afterwards are confusing (e.g. increase in comet lifetime of SPTZ in non-injured axons with RTN1 KD compared to Scr+DMSO and KD + DMSO in injured axons). Their conclusion is axonal RTN-1 synthesis in injured axons (see my concern in the points above on this) governs microtubules growth rate beyond Spastin activity yet blocking Spastin activity still completely blocks the effect of KD on outgrowth.*

* *__R: __We thank this observation and fully agree that the general description provided in figure 5 E wasn't satisfactory. We have re-organized the descriptions of these results and performed more relevant statistical comparisons (lines 338-359). Based on the reviewer observation, we now conclude: "Together, these results suggest that axonal Rtn-1 synthesis controls microtubule dynamics in both non-injured and injured axons, mostly independently of Spastin-mediated microtubule severing." (lines 357-359).

Other/minor concerns:

- The gene ontology analysis in Figure 1A contains the category 'Endoplasmic reticulum'. In this category are mainly ribosomal proteins. Although in a gene ontology analysis these proteins will be included in this category, it is misleading in this respect since they are just as likely to be coming from cytoplasmic ribosomes. Although it cannot be excluded that these are ER-bound ribosomes, not in the last place because a recent study (Koppers et al., 2024, doi: 10.1016/j.devcel.2024.05.005) found ribosomes attached to the ER in axons, I believe the category should be adapted or at the least clarified in the text.

• *

__R: __Many thanks for the suggestion, which is now included in the text. "Note that several of the identified transcripts in the category 'endoplasmic reticulum' code for cytoplasmic ribosomal components, which indeed can be attached to the axonal ER (Koppers et al., 2024) and be locally synthesized in axons (Shigeoka et al., 2019)." (lines 125-128)

- Is RTN-1C isoform still an ER-shaping protein or rather an ER protein with alternative functions? The final sentence in the abstract makes a statement that a locally synthesized ER-shaping protein lessens microtubule dynamics. Could the authors provide a clearer description and discussion of the evidence in literature for this? RTN1C has been suggested to perform alternative functions in which case the statement that the local synthesis of an ER-shaping protein is important for axonal outgrowth should be adapted.

R: We agree with the reviewer and are aware that some non-canonical roles of Rtn-1C may partially explain the observed phenotypes. Thus, we have rephrased the last statement of the abstract: "These findings uncover a mechanism by which axonal protein synthesis provides fine control over the microtubule cytoskeleton in response to injury.". Also, we have modified the discussion section introducing new references accordingly..." Some studies have pointed to a non-canonical role for Rtn-1C in the nucleus, including DNA binding and histone deacetylase inhibition (Nepravishta et al., 2010, 2012). It is tempting to speculate that these still emerging roles may also contribute to the observed phenotypes. Of note, different axonally synthesized proteins exert transcriptional control in response to injury or local cues (Twiss et al., 2016)." (lines 576-580).

• Is there a difference in RTN1 distribution or levels pre- and post-axotomy?

R: Thanks for the suggestion, with the new analysis we have only found slight reorganization of Rtn-1C and Spastin in distal axons (Figure 5A). We have also included now quantification of their levels and found no significant differences for both proteins (Supplementary figure 4)

- Line 100/101 states 'the interactome of the axonal ER provides...'. To my knowledge there has been no study looking at the interactome of the axonal ER specifically. Surely axonal ER proteins are known but there is a difference.

• *

__R: __We agree with the reviewer that the phrase was misleading, so we rephrased it in the introduction "...Different lines of evidence support that the protein components of the axonal ER may interact with proteins that regulate microtubule dynamics"

* - Typo line 160 'localla'*

• *

__R: __Thanks for taking the time, we have now corrected it.

- In Figure S1 B, please add the DIVs to make it clearer what each graph corresponds to. The legend of S1B states different distances from the cell body but the graph shows distances from the tip.

• *

__R: __We have now corrected the legend accordingly.

- Figure 2C, why does B3 tubulin decrease in soma, aspecific effect of siRNA?

• *

__R: __This was indeed an unexpected finding. However, we do not observe unspecific or global changes in β3-tubulin levels (see Figure 2A and Supplementary Figure 2). Considering our other results linking Rtn-1 to the regulation of the microtubule cytoskeleton, we interpret this decrease as an indirect effect of Rtn-1 depletion rather than an off-target action of the siRNA. Moreover, if the effect were unspecific, both proteins would likely be reduced in the cell body, given that the siRNA was specifically designed to target Rtn-1 as its primary sequence-specific target.

- What is the rationale on the opposite effect found in outgrowth in Figure 3?

• *

__R: __The apparent opposite outcomes observed in Figure 3 - where axonal versus somatic Rtn-1 knockdown leads to divergent effects on axonal outgrowth - can be explained by compartment-specific environments and isoform distribution. The siRNA targets the conserved RHD region, reducing both Rtn-1A and Rtn-1C. Axons are enriched in Rtn-1C. Thus, axonal KD preferentially reduces Rtn-1C. In contrast, somatic KD reduces both isoforms. Rtn-1A, predominant in cell bodies, may probably engage other signaling pathways (Kaya et al., 2013). Interestingly, it was reported by Nozumi et al. (2009b) that global Rtn-1 depletion reduces axonal outgrowth in developing cortical neurons. This aligns with the notion that somatic KD mimics a more global loss of function, whereas axonal KD reveals a compartmentalized, pro-regenerative effect due to local Rtn-1C regulation. (All the references indicated here are in the main manuscript). These considerations are now included in the discussion ( lines 581-593).

* - Missing word 'we' on line 194*

• *

__R: __ We have corrected it.

- Typo line 629 'witmn h', please proofread the entire manuscript carefully.

• *

__R: __ We apologize for the spellings, now we have carefully revised the manuscript.

- Could the authors comment on why, in Figure 7B/C, GFP only is colocalizing with Spastin-RFP? In general, GFP should be diffusive and not display punctate colocalization with Spastin.

• *

We appreciate the reviewer's comment. Under normal conditions, GFP displays a diffuse cytoplasmic distribution. However, in our experimental setup, we observed punctate GFP signals only in the context of co-expression with Spastin-RFP. This is consistent with prior reports showing that soluble GFP can occasionally be sequestered into late endosomal structures (Sahu et al., 2011), which are also known to harbor the M87 Spastin isoform (Allison et al., 2013; Allison et al., 2019). To rigorously exclude the possibility of unspecific fluorescence crosstalk, we independently acquired each fluorophore channel and confirmed that GFP puncta were genuine and not due to bleed-through (Supplementary Figure 5). Further, cells expressing only GFP or only Spastin-RFP did not show overlapping puncta, and co-expression of GFP with Rtn-1A-RFP did not produce any apparent overlap, indicating that the punctate GFP pattern is specifically associated with Spastin co-expression. Thus, the observed GFP colocalization with Spastin reflects a biological phenomenon potentially linked to the endosomal localization of M87 Spastin, and not an artifact of imaging or fluorophore bleed-through.

Reviewer #2 (Significance (Required)):

* Axonal mRNA localization and localized translation support many neuronal functions and is an important determinant of the regenerative potential of axons after injury. How this works mechanistically remains unclear. The authors present a well performed and technically challenging study in which they identify RTN-1 as a regulator of axonal outgrowth after injury. They provide evidence using experiments in microfluidic chambers that RTN1 is locally synthesized in axons. Interestingly, they identify a (local) interplay between RTN1 and Spastin which affects microtubules and thereby regulates the outgrowth of cortical axons after injury. This study provides an interesting new link between a locally synthesized protein (RTN1) and a microtubule-regulating protein Spastin that is changed upon axon injury. This provides an advance in our understanding in axon regeneration after injury and provides the basis for new studies that can further investigate this interplay. Although interesting, I have several concerns that should be clarified and are needed to substantiate the findings and model presented in this study.*

*

The audience for this study will be mainly basic research in the fields of both axonal protein synthesis and axon regeneration. My expertise is in the field of mRNA localization and local protein synthesis.*

Batista, A. F. R., Martínez, J. C., & Hengst, U. (2017). Intra-axonal synthesis of SNAP25 is required for the formation of presynaptic terminals. Cell Reports, 20(13), 3085. https://doi.org/10.1016/J.CELREP.2017.08.097

Fan, X. xuan, Hao, Y. ying, Guo, S. wen, Zhao, X. ping, Xiang, Y., Feng, F. xue, Liang, G. ting, & Dong, Y. wei. (2018). Knockdown of RTN1-C attenuates traumatic neuronal injury through regulating intracellular Ca2+ homeostasis. Neurochemistry International, 121, 19-25. https://doi.org/10.1016/J.NEUINT.2018.10.018

Gracias, N. G., Shirkey-Son, N. J., & Hengst, U. (2014). Local translation of TC10 is required for membrane expansion during axon outgrowth. Nature Communications 2014 5:1, 5(1), 1-13. https://doi.org/10.1038/ncomms4506

Lucci, C., Mesquita-Ribeiro, R., Rathbone, A., & Dajas-Bailador, F. (2020). Spatiotemporal regulation of GSK3β levels by miRNA-26a controls axon development in cortical neurons. Development (Cambridge), 147(3). https://doi.org/10.1242/DEV.180232,

Reviewer #3 (Evidence, reproducibility and clarity (Required)):

This manuscript investigates the relationship between the endoplasmic reticulum morphogen reticulon-1 (Rtn-1) and the microtubule severing protein spastin in axons after injury. The main message and conclusion of the paper is that local axonal synthesis of Rtn-1 plays a role in regulating the microtubule severing activity of spastin by interacting with spastin and inhibiting its activity. This mechanism would be important after injury by regulating axonal growth.

* The conclusions of the paper are based on the following claims:*

* 1) Rtn-1 is synthesized locally in axons.*

* 2) Specific downregulation in Rtn-1 in axons using microfluidic chambers affects microtubules abundance (measured by beta-3 tubulin) and promotes axon growth after injury.*

* 3) Inhibition of spastin MT-severing activity with a specific drug rescues the growth effect induced by axonal downregulation of Rtn-1.*

* 4) Rtn-1c interacts with spastin-M87 to limit its MT-severing activity in a cellular system upon overexpression.*

*

*

Major comments:

1) Evidence that Rtn-1 is synthesized in axons comes from two experiments. Initially, the authors show that Rtn-1 siRNA transfection in the axonal compartment of microfluidic chambers reduces Rtn-1 levels in axons, suggesting that there is some local synthesis. Although this method is very attractive, I am concerned about the statistical analysis. The graphs show bars rather than individual data points from the average of many neurons (about 300). The plots also show the SEM instead of the SD, thus covering all the variability that is inherent in this type of experiment. The statistics are probably not performed on the 3 biological replicates, but consider the individual neurons as N. This is obviously not correct, since neurons in an experiment may all be affected by the same technical problem and are not independent replicates. For this reason, I am a bit skeptical about this quantification. Another problem is that the quantification of the fluorescence intensity of the sample does not take the nuclei into account. Are the nuclei removed for analysis? Are the images single planes? Addressing the quantification issues is crucial also for data in Figure 4, where the authors show a different effect of Rtn-1 axonal KD after injury.

* The second experiment is the Puro-PLA in Figure 6D. This experiment shows an average of 1.5 dots of signal per soma, which is a very low level of translation for this compartment where most of the synthesis should be taking place. In the axons, it is not clear how they calculate the axonal area. Again, the number of dots detected is very low and the physiological significance is questionable. A control with a known mRNA translated in axons would be important.*

* Finally, as an important control, the authors should show the presence of Rtn-1 mRNA by FISH in their experimental system.*

__R: __We appreciate the critical points addressed here as they moved us to improve the quality of the findings. We analyzed cells/axons as statistical units to increase statistical power given the subtle nature of these local changes. We agree with the reviewer that this approach may increase the risk of finding false positives. To address this point, i) we plotted the individual data points and colored them according with the different experimental dates (all the dates showed a similar trend) ii) We indicated SD instead of SEM iii) We analyzed our data using linear mixed-effects models, with experimental date included as a random effect. This approach allows to preserve the granularity and statistical power, while avoiding pseudoreplication. To exclude artifactual changes, we now analyzed the intensity fold change of total fluorescence normalized to Scr. Our former quantifications were based on the corrected fluorescence intensity used to construct the plot profiles, which could be adding some distortion to the measurements. These changes were applied throughout figures 2 and 4 (pages 34 and 38, respectively). After these new analyses the formerly presented results remain valid.

We thank the reviewer for raising concerns about the quantification of fluorescence intensity in cell bodies. We now specify in Materials and methods that fluorescence intensity analysis of distal axons (always isolated by the microfluidic chambers) and of cell bodies was performed using the wide-field configuration of the microscope. In all the cases, a single (epifluorescent) plane was analyzed to reflect the total fluorescence of a cell or axon. We did not exclude the nuclear region from the quantifications, as this would also remove cytoplasmic signal located above or below the nucleus.

We also understand the concerns about puro-PLA experiments. We agree with the reviewer that an average of 1.5 puncta per soma initially appeared low. We have identified at least three reasons for this. First, the signal derives from only a 15-minute puromycin pulse, which is a short labeling window. Second, our puro-PLA assay is particularly stringent, as ligation relied directly on puromycin- and Rtn-1C-labeled primary antibodies, without the additional spacing normally introduced by secondary antibodies. In standard PLA, the critical distance for amplification is ~30-40 nm, whereas in our assay this distance is even more restrictive. Third, in our initial analysis we applied an overly cautious threshold to define "true" amplification. We have now refined this threshold using a baseline defined by the absence of puromycin stimulation. With this improved criterion, we now quantify an average of ~5 puncta per soma and ~10 puncta per 1000 µm² of axonal area (Supplementary Figure 3D). As it is now included in methods, we calculated the axonal area by binarizing β3-tubulin staining and only counted the true amplification spots inside this region. Assuming a neuronal soma diameter of 15 µm (area ≈ 176.71 µm²), this yields ~0.028 puncta per µm² in somata. In comparison, axons display ~0.01 puncta per µm², approximately one-third of the soma value which seems more reasonable. This is also compatible with most of Rtn-1C synthesis comes from the cell body.

Unfortunately, we could not be able to perform puro-PLA of other axonally synthesized proteins. Nevertheless, to further validate our puro-PLA signal, we tested the specificity of the Rtn-1C antibody we used for this assay by WB, IF, and Rtn-1 KD (Supplementary figure 3 A-C). In addition, we performed axonal Rtn-1 KD in microfluidic chambers for twenty-four hours, which elicited a significant decrease in puro PLA signal compared to Scr (Supplementary figure 3D). Together, these results strongly indicate that the quantified signal reflects Rtn-1C synthesis. To prove that Rtn-1 mRNA is present in these conditions, we now included a RT-PCR performed on RNA isolated from the somato-dendritic and pure axonal fractions of 8 DIV microfluidic chambers (Supplementary figure 3D). Note that the presence of this mRNA in axons has been supported by several studies, one of them using cortical neurons of similar DIV and cultured in microfluidic chambers (Table I and figure 1).

2) The effects on tubulin following Rtn-1 downregulation in axons is potentially very interesting, but the authors should be careful because it could also mean that the axons are suffering. Can they also stain for other cytoskeletal markers?

R: Regarding this concern, we are aware that in the former Figure 3 we mistakenly selected axonal fields that did not display healthy axons, which was not the dominant trend. This is accredited by the lack of fragmentation and by the functional responsiveness (microtubule dynamics) shown in Figures 4 and 5B, C, E. We have now replaced the previous axonal fields in Figure 3 with more representative axons (healthy), devoid of varicosities and fragmentation (page 37)

3) The results using SPTZ are very interesting and implicate spastin microtubule severing activity in the observed phenotype. In my opinion these experiments however do not prove that "axonal Rtn-1 is indeed promoting the severing of microtubules by spastin", but simply that the blocking spastin activity prevents the appearance of the microtubular phenotype (which appears still with a mysterious mechanism). What happens if they try to stabilize the cytoskeleton by another mean (with taxol for example?). The authors should rephrase this conclusion.

__R: __We completely agree with the reviewer's appreciation. We now explicitly indicate in the main text that this is (so far in the manuscript) a still correlative phenomenon that suggests an interplay with Spastin activity "..Together, these results suggest that locally synthesized Rtn-1 normally acts to suppress the outgrowth of injured axons, a process that could involve the microtubule-severing activity of Spastin." (lines 321-323). Later in the article, with the improved Figure 7, we further propose that these findings may reflect a causal relationship, although this mechanism has not yet been directly demonstrated in axons.

4) The last experiment (Figure 7) that aims to connect Rtn-1 and spastin function is very artificial, since it is based on overexpression. Why should spastin M87 interact with an ER morphogen? Endogenously it is conceivable that spastin M1 which localizes to the ER would interact with Rtn-1. Moreover, this experiment needs further controls and quantifications. First, it is quite obvious from panel 7C that there is crossover of signal in the two fluorescence channels (see GFP and spastin). Controls need to be shown, where only one of the two fluorescent proteins is expressed, and the specificity of the laser is tested. This experiment is based on only 1 cell shown where co-localisation is detected based on a line that is placed in a specific area of the cell. The effects on the microtubular network needs quantification.

__R: __We have now improved Figure 7 and added the requested controls to rule out crosstalk as indicated in Supplementary Figure 5 and in the main text. We agree that under normal conditions GFP should display a diffuse cytoplasmic distribution. However, in our experimental setup, we observed punctate GFP signals only in the context of co-expression with Spastin-RFP. This is consistent with prior reports showing that soluble GFP can occasionally be sequestered into late endosomal structures (Sahu et al., 2011), which are also known to harbor the M87 Spastin isoform (Allison et al., 2013; Allison et al., 2019). To exclude the possibility of unspecific fluorescence crosstalk, we independently acquired each fluorophore channel and confirmed that GFP puncta were genuine and not due to bleed-through (Supplementary Figure 5). Further, cells expressing only GFP or only Spastin-RFP did not show overlapping puncta (arrowheads), and the co-expression of GFP with Rtn-1A-RFP did not produce any apparent overlap, indicating that the punctate pattern of GFP is specifically associated with Spastin co-expression. Thus, we consider that the observed GFP colocalization with Spastin potentially reflects a true phenomenon and not an artifact of imaging or fluorophore bleed-through.

We thank for these observations and apologize for the confusion in the outline of the former figure 7 and the lack of a better description. As the reviewer indicates, one interesting aspect of the M87 isoform is that lacks the ER morphogen domain (so is soluble or cytoplasmic in principle). However, it also harbors endosome and microtubule binding domains which according to previous literature (now included in the main text) may render it a punctate rather than a homogeneous pattern. Also, M87 is the most abundant isoform in the nervous system, particularly at early development. This is the reason why we selected this isoform to test our model. To clarify this point, we based our colocalization analysis in different cells and experimental dates and analyzed all the z-stacks for each cell (see new figure 7B and methods), the intensity plots (now removed) were only for graphical purposes. Similarly, we had already quantified the total tubulin intensity in COS cells based on many cells from different dates and included the sum projections of all the z-stacks from these cells (see new figure 7C). Thus, we removed the intensity profiles as they were clearly misleading (see new figure 7).

We agree that over-expressing constructs may force interactions or co-distribution of proteins. However, in this case, if the observed results were mainly due to over-expression, we should see a similar trend with isoform A as both constructs are under the control of the same strong promoter (CMV) and harbor the same ER morphogen domain (RHD). Nevertheless, the distribution of M87 tightly mirrors Rtn-1C, which is not the case for Rtn-1A. Only as a theoretical prediction, our molecular modeling suggests that Rtn-1C may be associated with Spastin through its microtubule binding domain (Figure 7E). This would suppose that Spastin "decorates" ER-tubules rather than being in the same ER membranous structure. This discrete pattern of Spastin is more coherent with the distribution of both proteins that is now more clearly observed in distal axons by STED super-resolution (new figure 5A). So, despite a bit unexpected, these results suggest a novel interaction mechanism between these two proteins that deserves further validation.

5) What is exactly the model proposed? The title implies that axonal synthesis of Rtn-1 is important during injury, but the data in the paper rather suggest that upon injury the majority of Rtn-1 is not locally synthesized. If the levels of Rtn-1 do not change, why the effect on the microtubules should be specific? Why would a siRNA against Rtn-1 in axons not affect the levels of Rtn-1, but those of tubulin? The authors should be careful, and test other control siRNAs, and Rtn-1 siRNAs, since it is well known even in more simple cellular systems that the toxicity of individual siRNAs can vary greatly.

We consider the possibility that after injury there is no axonal Rtn-1 synthesis as a plausible and relevant appreciation. Unfortunately, we could not perform a puro-PLA experiment after injury, which would have provided a more definite answer. However, now we are more confident of regulating Rtn-1 synthesis before injury as supported by a Supplementary figure 3D that shows a significant decrease on puro-PLA signal (indicative of Rtn-1C synthesis) 24 hours after axonal KD. Thus, based on some similar phenotypes before and after injury, we consider our results are still compatible with Rtn-1 axonal synthesis being downregulated, but not fully absent (the mRNA is still detected, as described by Taylor 2009). As such, axonal Rtn-1 KD decreased β3-tubulin levels before and after injury according to figure 5B and the improved statistical analysis performed on figure 2E. Similarly, axonal Rtn-1KD significantly increases microtubule growth rate before and after injury according to the current statistical comparisons (Figure 5E). in complement, if β3-tubulin decrease was merely due to unspecific siRNA targeting, it is unlikely that SPTZ treatment should restore β3-tubulin only in the context of axonal Rtn-1 KD (Figure 5B). Although on a different track, the mechanistic relationship between Rtn-1C and Spastin suggested in Figure 7 could make more plausible that a similar phenomenon regarding the control of tubulin levels could be occurring locally in axons. We have now included these considerations in the discussion (lines 535-543).

To discard off-targets effects, we have now validated a third siRNA sequence (siRNA 3) specifically designed against Rtn-1 and showed that it selectively downregulates Rtn-1C but not β3-tubulin in cultured cortical neurons. Then, following the same experimental frame of figure 3, we performed axonal Rtn-1 KD after injury and observed that siRNA 3 also significantly increases the outgrowth of injured axons (Supplementary figure 2). This suggests that, at least this phenotype, is not product of an off-target effect. Thus, the pharmacological rescue of β3-tubulin levels by SPTZ (Figure 5B) and the Rtn-1C/Spastin co-distribution in heterologous cells, which correlates with preserved microtubules (improved Figure 7), provide converging evidence to suggest that Rtn-1C-Spastin interplay may underly the observed phenotypes in axons.

Minor comments:

In Figure 5A, it would be helpful to indicate the border of the axon. The figure is not really convincing.

Following yours and other reviewer comments, we have analyzed a new set of experiments regarding the STED images of non-injured and injured axons. To eliminate the risk of artifactual descriptions, we have avoided deconvolution and worked directly with raw STED images (Figure 5A). Under these conditions, distribution of Spastin and its intensity in distal axons are not modified by injury, nor those of Rtn-1C and Spastin (Supplementary figure 4). Despite these results, data still supports that both proteins are restricted to similar domains subcellular domains before and after injury.

Reviewer #3 (Significance (Required)):

The manuscript uses complex methods to address an interesting cell biological question of relevance to understand axonal growth regulation upon injury. A limitation of the study is the statistical analysis, which triggers some doubts about the reproducibility of the data. Further experiments and the addition of controls would be important to support the claims of the authors.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

This manuscript investigates the relationship between the endoplasmic reticulum morphogen reticulon-1 (Rtn-1) and the microtubule severing protein spastin in axons after injury. The main message and conclusion of the paper is that local axonal synthesis of Rtn-1 plays a role in regulating the microtubule severing activity of spastin by interacting with spastin and inhibiting its activity. This mechanism would be important after injury by regulating axonal growth.

The conclusions of the paper are based on the following claims:

1. Rtn-1 is synthesized locally in axons.
2. Specific downregulation in Rtn-1 in axons using microfluidic chambers affects microtubules abundance (measured by beta-3 tubulin) and promotes axon growth after injury.
3. Inhibition of spastin MT-severing activity with a specific drug rescues the growth effect induced by axonal downregulation of Rtn-1.
4. Rtn-1c interacts with spastin-M87 to limit its MT-severing activity in a cellular system upon overexpression.

Major comments:

1. Evidence that Rtn-1 is synthesized in axons comes from two experiments. Initially, the authors show that Rtn-1 siRNA transfection in the axonal compartment of microfluidic chambers reduces Rtn-1 levels in axons, suggesting that there is some local synthesis. Although this method is very attractive, I am concerned about the statistical analysis. The graphs show bars rather than individual data points from the average of a large number of neurons (about 300). The plots also show the SEM instead of the SD, thus covering all the variability that is inherent in this type of experiment. The statistics are probably not performed on the 3 biological replicates, but consider the individual neurons as N. This is obviously not correct, since neurons in an experiment may all be affected by the same technical problem and are not independent replicates. For this reason, I am a bit skeptical about this quantification. Another problem is that the quantification of the fluorescence intensity of the sample does not take the nuclei into account. Are the nuclei removed for analysis? Are the images single planes? Addressing the quantification issues is crucial also for data in Figure 4, where the authors show a different effect of Rtn-1 axonal KD after injury. The second experiment is the Puro-PLA in Figure 6D. This experiment shows an average of 1.5 dots of signal per soma, which is a very low level of translation for this compartment where most of the synthesis should be taking place. In the axons, it is not clear how they calculate the axonal area. Again, the number of dots detected is very low and the physiological significance is questionable. A control with a known mRNA translated in axons would be important. Finally, as an important control, the authors should show the presence of Rtn-1 mRNA by FISH in their experimental system.
2. The effects on tubulin following Rtn-1 downregulation in axons is potentially very interesting, but the authors should be careful because it could also mean that the axons are suffering. Can they also stain for other cytoskeletal markers?
3. The results using SPTZ are very interesting and implicate spastin microtubule severing activity in the observed phenotype. In my opinion these experiments however do not prove that "axonal Rtn-1 is indeed promoting the severing of microtubules by spastin", but simply that the blocking spastin activity prevents the appearance of the microtubular phenotype (which appears still with a mysterious mechanism). What happens if they try to stabilize the cytoskeleton by another mean (with taxol for example?). The authors should rephrase this conclusion.
4. The last experiment (Figure 7) that aims to connect Rtn-1 and spastin function is very artificial, since it is based on overexpression. Why should spastin M87 interact with an ER morphogen? Endogenously it is conceivable that spastin M1 which localizes to the ER would interact with Rtn-1. Moreover, this experiment needs further controls and quantifications. First, it is quite obvious from panel 7C that there is crossover of signal in the two fluorescence channels (see GFP and spastin). Controls need to be shown, where only one of the two fluorescent proteins is expressed and the specificity of the laser is tested. This experiment is based on only 1 cell shown where co-localisation is detected based on a line that is placed in a specific area of the cell. The effects on the microtubular network needs quantification.
5. What is exactly the model proposed? The title implies that axonal synthesis of Rtn-1 is important during injury, but the data in the paper rather suggest that upon injury the majority of Rtn-1 is not locally synthesized. If the levels of Rtn-1 do not change, why the effect on the microtubules should be specific? Why would a siRNA against Rtn-1 in axons not affect the levels of Rtn-1, but those of tubulin? The authors should be careful, and test other control siRNAs, and Rtn-1 siRNAs, since it is well known even in more simple cellular systems that the toxicity of individual siRNAs can vary greatly.

Minor comments:

In Figure 5A, it would be helpful to indicate the border of the axon. The figure is not really convincing.

Significance

The manuscript uses complex methods to address an interesting cell biological question of relevance to understand axonal growth regulation upon injury. A limitation of the study is the statistical analysis, which triggers some doubts about the reproducibility of the data. Further experiments and the addition of controls would be important to support the claims of the authors.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Axonal mRNA localization and localized translation supports many neuronal functions and is an important determinant of the regenerative potential of axons after injury. How this works mechanistically remains unclear. The authors present a well performed and technically challenging study in which they identify RTN-1 as a regulator of axonal outgrowth after injury. They provide evidence using experiments in microfluidic chambers that RTN1 is locally synthesized in axons. Interestingly, they identify a (local) interplay between RTN1 and Spastin which affects microtubules and thereby regulates the outgrowth of cortical axons after injury. This study provides an interesting new link between a locally synthesized protein (RTN1) and a microtubule-regulating protein Spastin that is changed upon axon injury. This provides an advance in our understanding in axon regeneration after injury and provides the basis for new studies that can further investigate this interplay. Although interesting, I have several concerns that should be clarified and are needed to substantiate the findings and model presented in this study.

Major concerns:

1. In figure 1, the authors provide an analysis of overlapping axonal mRNAs. There are more axonal transcriptome studies and a recent study by von Kugelgen and Chekulaeva (2020; doi: 10.1002/wrna.1590) already performed such an analysis, which included more studies. It would be good to mention this. It can be perceived that studies were now chosen to get the outcome that Rtn-1 is present in all studies. For example, von Kugelgen finds mRNA coding for RTN3, another ER structural protein, as present in 16 out of 20 studies analyzed. That said, the authors present more reasons to look at Rtn-1, so the selection to continue with this protein remains valid but can be written up differently so not to present it as the 'sole' ER-shaping protein consistently present in axonal transcriptomes.
2. The description of methods is currently insufficient and incomplete and does not allow for reproducibility of this study. For example, different Rtn-1 antibodies seem to be used in this study. Is the same antibody used for staining and WB? There is no listing of any of the antibodies used in the study and which one is used for which technique/experiment. This should be clarified and should be easy to do so in the methods section (antibody name, origin/company, dilution used) to enhance reproducibility of this study. This is not limited to primary antibodies and any information on secondary antibodies, including what was used for STED is completely missing.
3. The timeline of KD experiments in Figure 2 and 3 are unclear. For the Western blot KD is performed at DIV7 and collected 48 hours later. However, this is not specified for the stainings done in Figure 2C-E. Is this also at DIV7 and then for 48 hours? In figure 3 the siRNA is added at DIV8 (together with axotomy) and outgrowth is measured 24 hours later. Is 24 hours sufficient to achieve knockdown? Is this also what was done for stainings? Later on in Figure 5B, 48 hours of KD is again used. It is unclear what the rationale of these differing timepoints is. Why was this chosen? Is the timeline also the reason for the difference in segment lengths chosen? In Figure 3, there is a significant effect on outgrowth in the KD in the 'mid-range' which is not present in Figure 5.
4. Could the authors provide a rescue condition for their siRNA (using a siRNA-resistant construct) to show that their siRNA is specific for RTN1. They nicely show the efficiency of the siRNA but not its specificity. This is crucial because if not specific, this will affect a large part of their study. They already have RTN1A and RTN1C constructs available. Such a rescue experiment should ideally also be performed for one or more of their phenotypic experiments, such as the one presented in Figure 3A or 5 to show that the phenotype is really RTN1 dependent. If done by re-expressing either RTN1A or RTN1C, this could provide insightful information on the relevant isoforms.
5. I find the data presented in Figure 4A/B confusing. Axonal RTN-1 KD does not reduce axonal RTN1 levels but somatic KD does. I understand that this implies most protein comes from the soma and the authors indeed present an explanation that increased somatic RTN1 occurs after axonal KD as a compensation mechanism. However, this can also be interpreted that there is no axonal synthesis of RTN1 after injury and axonal KD has indirect or even aspecific effects. Their model depends on this difference. Their data in Figure 6 could provide supporting evidence if it shows RTN1 puro-PLA after injury. Along these same lines, in Figure 6, they nicely include a compartment control for puro-PLA. It therefore seems doable to include a somatic puromycin control for their axonal puro-PLA, to exclude and diffusion/transport of the newly synthesized peptides. This is especially in light of two recent papers reporting on this possible phenomenon, although these studies were not performed in neurons.
6. In Figure 5A the authors find an increased co-localization (RTN1/Spastin) after axotomy. From their images, it seems that the amount of Spastin is hugely increased, which would by default increase the chance of (random) colocalization of RTN1 on Spastin. Could the authors comment on this?
7. In figure 5E and 5F, the condition of scr + SPTZ is omitted. What is the reason for this? The explanation of results in these figures is confusing. The authors report a 'clear trend' in increase in comet track length and lifetime upon addition of SPTZ to axonal RTN-1 KD. This is however not significant. The comparisons that are made afterwards are confusing (e.g. increase in comet lifetime of SPTZ in non-injured axons with RTN1 KD compared to Scr+DMSO and KD + DMSO in injured axons). Their conclusion is axonal RTN-1 synthesis in injured axons (see my concern in the points above on this) governs microtubules growth rate beyond Spastin activity yet blocking Spastin activity still completely blocks the effect of KD on outgrowth.

Other/minor concerns:

• The gene ontology analysis in Figure 1A contains the category 'Endoplasmic reticulum'. In this category are mainly ribosomal proteins. Although in a gene ontology analysis these proteins will be included in this category, it is misleading in this respect since they are just as likely to be coming from cytoplasmic ribosomes. Although it cannot be excluded that these are ER-bound ribosomes, not in the last place because a recent study (Koppers et al., 2024, doi: 10.1016/j.devcel.2024.05.005) found ribosomes attached to the ER in axons, I believe the category should be adapted or at the least clarified in the text.
• Is RTN-1C isoform still an ER-shaping protein or rather an ER protein with alternative functions? The final sentence in the abstract makes a statement that a locally synthesized ER-shaping protein lessens microtubule dynamics. Could the authors provide a clearer description and discussion of the evidence in literature for this? RTN1C has been suggested to perform alternative functions in which case the statement that the local synthesis of an ER-shaping protein is important for axonal outgrowth should be adapted.
• Is there a difference in RTN1 distribution or levels pre- and post-axotomy?
• Line 100/101 states 'the interactome of the axonal ER provides...'. To my knowledge there has been no study looking at the interactome of the axonal ER specifically. Surely axonal ER proteins are known but there is a difference.
• Typo line 160 'localla'
• In Figure S1 B, please add the DIVs to make it more clear what each graph corresponds to. The legend of S1B states different distances from the cell body but the graph shows distances from the tip.
• Figure 2C, why does B3 tubulin decrease in soma, aspecific effect of siRNA?
• What is the rationale on the opposite effect found in outgrowth in Figure 3?
• Missing word 'we' on line 194
• Typo line 629 'witmn h', please proofread the entire manuscript carefully.
• Could the authors comment on why, in Figure 7B/C, GFP only is colocalizing with Spastin-RFP? In general, GFP should be diffusive and not display punctate colocalization with Spastin.

Significance

Axonal mRNA localization and localized translation supports many neuronal functions and is an important determinant of the regenerative potential of axons after injury. How this works mechanistically remains unclear. The authors present a well performed and technically challenging study in which they identify RTN-1 as a regulator of axonal outgrowth after injury. They provide evidence using experiments in microfluidic chambers that RTN1 is locally synthesized in axons. Interestingly, they identify a (local) interplay between RTN1 and Spastin which affects microtubules and thereby regulates the outgrowth of cortical axons after injury. This study provides an interesting new link between a locally synthesized protein (RTN1) and a microtubule-regulating protein Spastin that is changed upon axon injury. This provides an advance in our understanding in axon regeneration after injury and provides the basis for new studies that can further investigate this interplay. Although interesting, I have several concerns that should be clarified and are needed to substantiate the findings and model presented in this study.

The audience for this study will be mainly basic research in the fields of both axonal protein synthesis and axon regeneration. My expertise is in the field of mRNA localization and local protein synthesis.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

In this paper, the authors focus on the role of Reticulon-1C in concert with Spastin in response to axonal injury. In data mining, they find axonal mRNAs encoding for ER-associated proteins including Rtn-1. They establish a knockdown targeting both Rtn-1 isoforms Rtn-1A and Rtn-1C. They observe decreased beta-3-Tubulin levels in the soma while axonal protein levels are unchanged. In microfluidic devices, they characterise the effect of a compartment-specific Rtn-1 KD on axonal outgrowth in the axonal compartment. The authors quantify axonal outgrowth, seeing increased outgrowth in an axonal compartment-specific Rtn-1 KD, while the effect seems to be reversed when applying the KD construct in the somatic compartment. When focussing on the axonal growth cone, they find the Rtn-1 KD shows differences in several morphological features of the growth cone. They find an increase in Tubulin levels in an axonal compartment-specific, but a decrease in a somatic compartment-specific Rtn-1 KD. Colocalisation of Rtn-1C and Spastin is shown to be monolaterally increased following axotomy. Combining axotomy with the Rtn-1 KD shows increases in dynamic microtubule growth rates and track lengths. In another model system, neuron balls, they show Rtn1-C, but not Rtn1-A to be present in the axon. In a puro-PL assay they also show it can be synthesised in the axonal compartment. To investigate the mechanism enabling the cooperation between Spastin and Rtn-1C, they move to a cell line model in which they see a correlating distribution between Spastin and Rtn-1C but not Rtn-1A. Finally, they use in silico modelling to speculate on binding between Spastin domains and Rtn-1 isoforms.

Major comment:

The rationale behind the work is convincing, however some interpretations are presented as more robust than some data allow. Most notably, while the interaction between Rtn-1 and Spastin has been shown prior to this study, it is only presented here through in silico analysis. In figure 5, an increase in the growth rate of dynamic microtubules is observed in either a Rtn-1C KD or by using a Spastin-inhibitor. Due to a described increase in colocalisation between Rtn-1C and Spastin (5A), the increase in growth rate is displayed as caused by Rtn-1 promoting Spastin's severing ability. This result might however be correlative. Further in the injured samples, Spastin-levels seemingly increase (in the representative images) and it is thus not surprising that the level of Rtn-1C colocalising with Spastin increases as well. This might not be indicative of a cooperation and further experimental evidence are required.

Other comments:

• Generally, graphs would benefit from individual values plotted as well as the summary. Font sizes and types (but rarely) are sometimes inconsistent. Proteins should be consistently written (capitalised or not).
• Table 1 and figure 1 present data collected from a vast amount of resources. It should be highlighted that datasets from which data was obtained includes many different models, different DIVs and neuronal cell types. Figure 1B may benefit from a different colour scheme. "Ex-vivo" should be "Ex vivo". For "ER mRNAs are a relevant category" it is not described what "relevant" would mean in this context. The title might remove this small part or describe it in the text. It should be described how it is decided that mRNAs are "common".
• Figure 2: add description to y-axis to describe what fold change is displayed, applies to multiple figures. Will improve readability of the figures. In 2C, the ROI showing neuronal somata should be increased to show part of the axon and not cut off the soma.
• Figure 3: Three out of four axonal compartments seem to be comprised of dying or damaged axons. Especially the axonal KD scrambled image. It should be ensured that neuronal cultures are healthy. Typo in "intersections". The schematic of 3B is a great addition to explain the graphs above. Perhaps it could be a bit refined as it is currently hard to see whether this is a neuron or a growth cone without context. Maybe show where the axon connects to the depicted growth cones and change the third icon which looks like it was crossed out. Small formatting issues: remove additional space bar before "Figure 3." And add after "Bar"
• Figure 4: If not misunderstanding what is depicted, in 4A and B, different lookup tables are used to depict the same signal. Only one of each images is necessary. Do the axons have more tiny branches in the Rtn-1 KD condition in 4A? Unclear why Rtn-1 levels are increased in the Rtn-1 KD (4C), please clarify.
• Figure 5: It may be easier to understand what "axotomy" samples are if just referred to as "injured" as later in the same figure. The procedure could also very briefly be explained in the results. 5C should depict AUC in µm2 not µm. 5D Spastin is barely visible, brightness and contrast should be adjusted to enhance visibility.
• Figure 6: It should be made clear why it is necessary to switch to another model system just for 6A, please indicate this in the text. PCR bands seem very pixelated, check the quality. It is unclear why soma genes/proteins were only tested with either PCR or WB others with both. Rtn-1C and Rtn1-A should be presented in the same order in the PCR and WB panel. Correct "Rtn1-1A" typo. In 6D, 1.5 dots per soma seems like a low number. When normalised to the area the soma vs the axon occupies, the compartmentalisation does not work? May be it make sense to refine analysis or apply puromycin in the somatic compartment and analyse the axonal compartment as comparison?
• Figure 7: 7A shows two images depicting the same information that may not be needed. Can probably be removed. In 7B there is no negative (or any) correlation between Spastin levels and Tubulin, however later it is mentioned that Rtn-1C transports Spastin thus causing a decrease in Tubulin at certain locations? It is nclear if Spastin levels vary intensely between different samples. Mean intensity of the somatic area may be beneficial to rule this out. 7B Tubulin on the right top panel seems to have a decrease in Tubulin levels which is not visible due to the Y axis of Tubulin being set to a different range than the middle and lower panel. The average of line scans from multiple cells may be helpful to determine whether there is indeed no colocalization between Rtn-1A and Spastin. The provided representative images seem to show similar degrees of colocalization between Spastin and Rtn-1A/C.

Results:

• It would be helpful to reiterate the hypothesis at the start to ease the reading flow.
• There seems to be minor redundancy in lines 132-138.
• There are several spellings, proof-reading is recommended. For example, in line 136 should be "promotes". 160 "localla", 192 should be "the actin cytoskeleton".,194 should be "we first examined", 195 should be "Different", 223 "using", 259 "axons". ...
• 154-155: Unclear, why the lower MW Rtn-1C was seen as more important.
• 167 results of 2E not stated before interpreting them.
• 181 would suggest "outline" instead of "perimeter".
• 183-184 "longest shortest path" is a confusing term.
• figure 4B should be referenced earlier in the sentence.
• 243-244 may be correlation. Rtn-1 and Spastin do not necessarily interact so that this result is achieved.
• 246: In figure 1 the KD seemed to have an effect on both Rtn-1 isoforms, why not here anymore? 259 "axons". 284 "counteract" instead of "suppress"?
• 485: rephrase as the interaction between Rtn-1C with Spastin has not been shown directly in these experiments.

Methods: 535 "in PBS". 543 citation error. 689-699 is it necessary to add a gaussian blur?

References: Mannan, A U et al. appears twice in the citation list (36 and 44).

Significance

Overall, this manuscript describes novel fundings which will be interesting to the neuronal cell biology community and scientists working on the field of neuronal injury and regeneration. It is well structured, and the data are mostly well presented but sometimes conclusions are over-interpreted. However, several points need to be addressed in a more convincing way.

I agree, till this day that's what it means, they all think that were being disrespectful, especially in a black Household.

Social patterns have actually been used in both ways stated it is unavoidable social response and imposed by a governing body for its own purpose. Both get so mixed that they can't be distinguished in some cases. Its one thing when it happens naturally it another thing when it is rigged. To identify them is the first step to making changes if necessary.

Lots of human error then, or personal bias

Again, very Ad Hoc at a time where there was enough demand for a systematic procedure

I read with interest the article by Bascur, Costas, and Verberne, which examines the use of diverse data sources to influence topic emergence in science maps.

At the outset, I note that I am affiliated with Digital Science, the owner and operator of both Altmetric and Dimensions—two of the data sources analysed in the study.

The article focuses on the mapping of articles onto science maps characterised by clusters of topical areas. These are typically visualised in two dimensions, where the relative positions of topics are determined by a selected distance metric. This area of study has seen considerable development in recent years, and science maps continue to serve as a compelling tool in various analytical and strategic contexts.

While the paper’s focus on mapping research outputs onto science maps is timely and relevant, I was disappointed to see that key foundational works in this field were not cited. In particular, I believe the following references are highly relevant, especially the last, which explores the use of Wikipedia in a manner closely related to the current paper:

• Börner, K., Chen, C., & Boyack, K. W. (2005). Visualizing knowledge domains. Annual Review of Information Science and Technology, 37(1), 179. https://doi.org/10.1002/aris.1440370106

• Boyack, K., Klavans, R., & Börner, K. (2005). Mapping the backbone of science. Scientometrics, 64, 351–374. https://doi.org/10.1007/s11192-005-0255-6

• Holloway, T., Bozicevic, M., & Börner, K. (2007). Analyzing and visualizing the semantic coverage of Wikipedia and its authors. Complexity, 12, 3–85. https://doi.org/10.1002/cplx.20164

The paper is framed as an exploration of how networks emerge, which is an important and intriguing subject. Visual representations play a crucial role in knowledge communication and decision-making, and I consider this work both significant and valuable.

That said, I initially interpreted the paper as an exploration of science map construction via bibliometric coupling informed by different data sources. However, the paper does not appear to explore alternative embedding metrics, topological variations, or the graphs that might result from different coupling strategies. Instead, it primarily assesses how faithfully papers can be mapped onto an existing topic structure using alternative data sources. I use the term "faithful" here in its group-theoretic sense—capturing both purity and effectiveness—as it conveys the intended meaning more precisely, in my opinion.

This focus differs somewhat from the broader ambitions implied in the title and abstract. I recommend the authors reassess whether the current framing accurately reflects the content, or alternatively, provide a more explicit explanation of the embedding method used and how it relates to the structural similarity being evaluated. Even if the scope is narrower than anticipated, the findings remain rigorous, well-articulated, and represent a valuable contribution.

If the paper is best understood as a study of the faithfulness of mapping unclassified papers to an existing clustering structure using different linking mechanisms (e.g., social data vs. textual or citation-based), then its key result appears to be that BERT-based methods offer a more faithful reproduction of the existing map than bibliometric approaches using the alternative data sources.

Given the clustering methodology behind the target map, this result is sensible. The use of social media data, as discussed in the paper, is more likely to yield alternative representations rather than a faithful reproduction. By contrast, BERT embeddings naturally align more closely with textual structures already reflected in the map. This outcome is consistent with the analytic approach adopted.

As an aside, the paper does not appear to reference the work of Evans and Lambiotte (https://doi.org/10.1103/PhysRevE.80.016105), which investigates the use of bipartite graphs and their duals for community detection. Their work is directly relevant to the paper’s discussion of faithfulness, particularly in terms of minimising overlap in cluster assignments.

Finally, I believe the paper would benefit from a stronger articulation of the contexts in which science mapping is applied. I have previously explored this in (https://doi.org/10.1162/qss_a_00244), and I believe the current work holds particular promise in evaluative settings. Preserving classification consistency over time is often vital for longitudinal comparisons, and the paper’s approach could be valuable in assessing alternative coupling strategies against a stable reference framework.

In summary, this paper presents a thoughtful and well-executed study. I find the technical development in the methodology around the refinement of purity to be helpful and something that those in the field will want to explore further. I recommend the authors consider refining the framing and expanding the contextualisation to strengthen the contribution and clarify its position within the existing literature.

https://www.linkedin.com/feed/update/urn:li:activity:7283852658128429056?commentUrn=urn%3Ali%3Acomment%3A%28activity%3A7283852658128429056%2C7288857497979748352%29&dashCommentUrn=urn%3Ali%3Afsd_comment%3A%287288857497979748352%2Curn%3Ali%3Aactivity%3A7283852658128429056%29

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

Learn more at Review Commons

Reply to the reviewers

Reviewer #1 (Evidence, reproducibility and clarity):

This manuscript described the translational responses to single and combined BCAA shortages in mouse cell lines. Using Ribo-seq and RNA-seq analysis, the authors found selective ribosome pausing at codons that encode the depleted amino acids, where the pausing at valine codons was prominent at both a single and triple starvations whereas isoleucine codons showed pausing only under a single depletion. They analyzed the mechanisms of the unexpected selective pausing and proposed that the positional codon usage bias could shape the ribosome stalling and tRNA charging patterns across different amino acids. They also examined the stress responses and the changes in the protein expression levels under BCAA starvation.

The manuscript was well-written, and the findings are interesting, especially their model that positional codon usage bias could be a regulator of ribosome pausing and tRNA charging levels. Although different translational responses to distinct amino acid starvation have been widely documented, the positional codon usage bias is an interesting aspect. The manuscript's central message could have been made clearer. The authors may consider emphasizing this point more explicitly in the abstract. The rich multi-omics dataset in this work provides valuable resources for the translation field.

We thank the reviewer for the thoughtful and positive evaluation of our work.

Major comments

1. The abstract may need to be revised since it is hard to immediately catch the authors' main point. If the authors regard this work as a resource paper, the current version is fine. But it could be better to point out the positional codon usages the authors found, which is a strong point of the current manuscript.

Response: We thank the reviewer for highlighting the importance of positional codon usage, which indeed represents a key finding of our study. We revised the abstract, and we now emphasize this aspect more clearly. However, in response to review #2, we have framed the observed positional effects and the idea of an elongation bottleneck as one possible contributing mechanism among others and relate it specifically to the attenuation of isoleucine-specific stalling under triple starvation.

1. Page 18 "Beyond these tRNA dynamics, our data also highlight the importance of the codon positional context within mRNAs, indicating that where a codon is located within the CDS can influence both the extent of ribosomal stalling and overall translation efficiency during nutrient stress." This idea is interesting. To what extent the authors think this could be generalized? The authors may discuss whether they think their proposed model is specific to the different ribosome stalling patterns between valine and isoleucine codons or generalized to other codon combinations. For example, the positional codon usage bias will be different among different organisms, and are there any previous reports on ribosome behaviors that align with their model?

Response: We thank the reviewer for raising these important points. While our study primarily focuses on the differential stalling patterns of valine and isoleucine codons, we believe the underlying principle, that the position of codons within the CDS can modulate the extent of ribosome stalling, may under very specific circumstances extend beyond this amino acid pair. We expect this positional effect to be potentially relevant for combinations in which one amino acid has considerable enrichment near the 5′ end of coding sequences, coupled with starvation-sensitive tRNA isoacceptors, while the other does not. In our case, valine meets these criteria (see Fig. S11A and Fig. 6). In contrast, isoleucine and leucine codons, although also relatively frequent, show more variable positional distributions and are both decoded by isoacceptors that appear more resistant to starvation, as illustrated in Fig. 6 and reported for mammals and bacteria in Saikia et al. 2016; Darnell, Subramaniam, and O’Shea 2018; Elf et al. 2003; Dittmar et al. 2005. To explore the generalizability of this model, we have now included a transcriptome-wide analysis of codon position biases in mouse for all codons in the revised manuscript (Supplementary Figures 10 and 11). This analysis may serve as a basis to identify additional candidate codons for future studies. Furthermore, we now mention in the Discussion that amino acids with similar properties to valine regarding their positional distribution and tRNA isoacceptors, such as phenylalanine, and glutamine, whose tRNA isoacceptors are predicted to be fully deacylated under their respective starvation in bacteria (Elf et al. 2003), could be promising candidates for testing this model, in combination with amino acids, whose tRNAs are expected to remain partially charged under starvation or to be depleted at the start of the CDS such as i.e. His (Supplementary Fig.11C).

Even if the authors think this model can be applied to BCAA starvation, would it be possible to explain the different isoleucine codon responses between single and double starvation? The authors may discuss why the ribosome stalling at isoleucine AUU and AUC codons was slightly attenuated under double starvation. And how about the different leucine codon responses among single, double, and triple starvations, although the pausing is not as strong as isoleucine and valine codons?

Response: Regarding the attenuated isoleucine stalling under double starvation, we believe this is primarily due to stronger inhibition of the mTORC1 pathway when leucine is co-depleted (i.e., in the double starvation condition; Fig. 2D–F). This results in a more substantial suppression of global translation, reducing overall tRNA demand and thereby mitigating stalling (Darnell, 2018). A similar effect may explain the only mild leucine codon stalling observed under single leucine starvation, which also triggers strong mTORC1 inhibition and reduced initiation. In contrast, triple starvation does not suppress mTORC1 to the same extent, and thus reduced initiation alone cannot explain the absence of leucine codon stalling. Instead, we propose that additional features, such as the relative sensitivity of tRNA isoacceptors to starvation and their aminoacylation dynamics, must be considered. Valine tRNAs, for example, are known to be highly sensitive and become strongly deacylated under starvation in bacteria (Elf et al. 2003), a pattern that we also find in our own data (Fig. 6). Leucine tRNAs, by contrast, appear more resistant, possibly due to better amino acid recycling or isoacceptor-specific differences in charging kinetics, though further validation would be needed. However, combined with the strong stalling at 5′-enriched valine codons, this could reduce downstream ribosome traffic and limit exposure of leucine codons, thus preventing stalling. However, our new analysis of the positional relationship between valine and leucine codons within individual transcripts (now shown in Supplementary Figure 11B) did not reveal as strong a pattern as we observed for valine and isoleucine codons. We now discuss these points and their implications in the revised Discussion.

Experimental validation using artificial reporters carrying biased sequences may also be considered.

Response: We appreciate the reviewer’s suggestion. In fact, we explored this experimentally using a dual-fluorescent reporter system (GFP–RFP) (Juszkiewicz and Hegde 2017) containing consecutive Val or Ile codons. However, the constructs yielded variable and non-reproducible results under starvation conditions. In addition, testing the role of codon position would require placing the same codons at multiple defined positions within a single transcript and performing ribosome profiling directly on the reporter. This type of targeted experimental validation is technically challenging and falls beyond the scope of the current study. We now mention this explicitly in the revised Discussion as an interesting direction for future work.

1. Page 13 "Moreover, we noticed that DT changes extend beyond the ribosomal A-site, including the P-site, E-site, and even further positions (Supplementary Fig. 2A), consistent with other studies on single amino acid starvation 39 (Supplementary Fig. 2B-C)." Could the widespread DT changes be due to Ribo-DT pipeline they used or difficulties in offset determination? Indeed the authors showed that this feature was found in other datasets, but it seems that the datasets were processed and analyzed in the same way as their data. The original Ribo-DT paper (Gobet and Naef, 2022, Methods) also showed some widespread DT changes even from RNA-seq. Another analysis method like the codon subsequence abundant shift as a part of diricore analysis (Loayza-Puch et al., 2016, Nature) did not show that broad changed regions. The authors are encouraged to re-analyze the data sets using different methods.

Response: We agree with the reviewer that the fact that DT changes beyond the ribosomal A-site is puzzling, but this has already been seen in other papers using other approaches (Darnell, Subramaniam, and O’Shea 2018). To validate that this shift is not due to our A-site assignment, enrichment analysis, or DT method, we applied the Diricore pipeline to our Ribo-Seq data. The output of the pipeline provides either 5’-end ribosome density or “subsequence” analysis using an A-site offset for each read size based on the metagene profile at the start codon. Both analyses show the same enriched codons across the different conditions as in our analyses, and the broad shift is similar, with the maximum signal at E, -1 position (Fig. R1).

1. Page 13 "Intriguingly, only two of the three isoleucine codons (AUU and AUC) showed increased DTs upon Ile starvation (p < 0.01), while just one leucine codon (CUU) exhibited a modest but significant DT increase (p < 0.01) under Leu starvation (Figure 1A-B, Supplementary Figure 2A)." How can the authors explain the different strengths of ribosome pausing at Ile codons under Ile and double starvation? The AUA codon did not show any pausing under either of the starvation conditions. Throughout the manuscript, the authors mainly describe the difference between amino acids but it is desirable to discuss the codon-level difference as well.

Response: Thank you for raising this point. The observed differences in stalling between the isoleucine codons can likely be explained by differences in tRNA isoacceptor charging and positional bias within transcripts. The AUA codon is decoded by a distinct tRNAIle isoacceptor (tRNAIleUAU), which, according to our tRNA charging data (Fig. 6), remains largely charged during Ile starvation. This observation aligns with previous reports suggesting that this isoacceptor is more resistant to starvation-induced deacylation in mammalian cells and bacteria (Saikia et al. 2016; Elf et al. 2003). In contrast, the AUU and AUC codons are primarily decoded by the tRNAIleAAU isoacceptor, which we find to be strongly deacylated under Ile starvation, likely contributing to the observed codon-specific ribosome pausing. Additionally, we found that the AUA codons are relatively rare in general and particularly underrepresented near the 5′ ends of coding sequences. Our new spatial analysis (now included in Supplementary Figure 11B) confirms that AUA codons tend to occur downstream of AUU and AUC codons within transcripts. This potentially further reduces stalling on these codons and further diminishes their apparent DT increase under starvation. In order to better explain these important points, we have now expanded the codon-level discussion of these differences in the revised manuscript.

1. Page 13 "We examined the effects of single amino acid starvations (-Leu, -Ile and -Val), as well as combinations, including a double starvation of leucine and isoleucine (hereafter referred to as "double") and a starvation of leucine, isoleucine, and valine ("triple"), allowing us to identify potential non-additive effects." The different double starvations, isoleucine and valine, and leucine and valine, will further support their hypothesis on the effects of the positional codon usage bias on ribosome pausing and tRNA charging patterns. Although this could be beyond the scope of the current manuscript, the authors are encouraged to provide a rationale for the chosen combination.

Response: Our experimental design evolved stepwise: we initially focused on leucine and isoleucine depletion as we found that despite their structure similarity these had respectively short and long dwell times in our previous work in the mouse liver (Gobet et al. 2020). Valine was included at a later stage to cover all the BCAAs. At the time, we did not anticipate valine to yield particularly striking effects in cells, and therefore we did not include systematic pairwise depletions involving valine. However, the strong and unexpected stalling observed at valine codons, especially under triple starvation, became a central aspect of the study. Thus, we agree that additional combinations, such as Leu/Val or Val/Ile, could be informative and now mention this in the Discussion as a potential direction for future studies.

Minor comments

Page 16 "these results imply that BCAA deprivation lowers protein output through multiple pathways: a combination of reduced initiation, direct elongation blocks (stalling), and possibly an increased proteolysis" This conclusion is totally right but may be too general. Could the authors summarize BCAA-specific features of the events including reduced initiation, stalling, and proteolysis that all contribute to protein outputs? This is not well discussed in the latter sections including Discussion.

Response: We thank the reviewer for this helpful suggestion. We agree that the original statement was too general and have revised the relevant section to more clearly delineate the distinct responses observed under each BCAA starvation condition. Specifically, we now summarize that valine starvation is characterized by strong, positionally biased ribosome stalling; leucine starvation primarily impacts translation initiation, likely via mTORC1 repression; and isoleucine starvation shows a mixed phenotype, with features of both impaired initiation and codon-specific elongation delays. We also clarify that while protein stability or degradation may contribute to the observed changes in protein output, our current data do not allow for quantitative assessment of proteolytic effects (e.g., changes in protein half-life). Therefore, we refrain from making direct quantitative conclusions about the differential modulations of proteolysis and instead focus our discussion on the translational mechanisms supported by our data.

Reviewer #1 (Significance):

The manuscript was well-written, and the findings are interesting, especially their model that positional codon usage bias could be a regulator of ribosome pausing and tRNA charging levels. Although different translational responses to distinct amino acid starvation have been widely documented, the positional codon usage bias is an interesting aspect. The manuscript's central message could have been made clearer. The authors may consider emphasizing this point more explicitly in the abstract. The rich multi-omics dataset in this work provides valuable resources for the translation field.

We thank the reviewer for the encouraging comments and share the view that positional codon-usage bias is an important result; accordingly, we now underscore this point explicitly in the revised Abstract. We also emphasise that our other observations are, to our knowledge, novel: only a handful of multi-omics studies have combined ribosome-pausing profiles with direct tRNA-aminoacylation measurements, and none has systematically examined multiple amino-acid-deprivation conditions as presented here.

Reviewer #2 (Evidence, reproducibility and clarity):

This study examines the consequences of starvation for the BRCAAs, either singly, for Leu & Ile, or for all three simultaneously in HeLa cells on overall translation rates, decoding rates at each codon, and on ribosome density, protein expression, and distribution of ribosome stalling events across the CDS for each expressed gene. The single amino acid starvation regimes specifically reduce the cognate intracellular amino acid pool and lead to deacylation of at least a subset of the cognate tRNAs in a manner dependent on continuing protein synthesis. They also induce the ISR equally and decrease bulk protein synthesis equally in a manner that appears to occur largely at the initiation level for -Leu and -Val, judging by the decreased polysome:monsome ratio, but at both the initiation and elongation levels for -Ile-a distinction that remains unexplained. Only -Leu appears to down-regulate mTORC1 and TOP mRNA translation.There is a significant down-regulation of protein levels for 50-200 genes, which tend to be unstable in nutrient-replete cells, only a fraction of which are associated with reduced ribosome occupancies (RPFs measured by Ribo-Seq) on the corresponding mRNAs in the manner expected for reduced initiation, suggesting that delayed elongation is responsible for reduced protein levels for the remaining fraction of genes. All three single starvations lead to increased decoding times for a subset of the cognate "hungry" codons: CUU for -Leu, AUU and AUC for -Ile, and all of the Val codons, in a manner that is said to correspond largely to the particular tRNA isoacceptors that become deacylated, although this correspondence was not explained explicitly and might not be as simple as claimed. All three single starvations also evoke skewing of RPFs towards the 5' ends of many CDSs in a manner correlated with an enrichment within the early regions of the CDSs for one or more of the cognate codons that showed increased decoding times for -Ile (AUC codon) and -Val (GUU, GUC, and GUG), but not for -Leu-of which the latter was not accounted for. These last findings suggest that, at least for -Val and -Ile, delays in decoding N-terminal cognate codons cause elongating ribosomes to build-up early in the CDS. They go on to employ a peak calling algorithm to identify stalling sites in an unbiased way within the CDS, which are greatest in number for -Val, and find that Val codons are enriched in the A-sites (slightly) and adjacent 5' nucleotides (to a greater extent) for -Val starvation; and similarly for Ile codons in -Ile conditions, but not for -Leu starvation-again for unknown reasons. It's unclear why their called stalling sites have various other non-hungry codons present in the A sites with the cognate hungry codons being enriched further upstream, given that stalling should occur with the "hungry" cognate codon in the A site. The proteins showing down-regulation are enriched for stalling sites only in the case of the -Val starvation in the manner expected if stalling is contributing to reduced translation of the corresponding mRNA. It's unclear why this enrichment apparently does not extend to -Ile starvation which shows comparable skewing of RPFs towards the 5'ends, and this fact diminishes the claim that pausing generally contributes to reduced translation for genes with abundant hungry codons. All of the same analyses were carried out for the Double -Ile/-Leu and Triple starvations and yield unexpected results, particularly for the triple starvation wherein decoding times are increased only at Val codons, skewing of RPFs towards the 5' ends of CDSs is correlated only with an enrichment for Val codons within the early regions of the CDSs, and stall sites are enriched only for Val codons at nearly upstream sites, all consistent with the finding that only Val tRNAs become deacylated in the Triple regime. To explain why only Val tRNA charging is reduced despite the observed effective starvation for all three amino acids, they note first that stalling at Val codons is skewed towards the 5'ends of CDS for both -Val and triple starvations more so than observed for Ile or -Leu starvation, which they attribute to a greater frequency of Val codons vs Ile codons in the 5' ends of CDSs. As such, charged Val tRNAs are said to be consumed in translating the 5'ends of CDSs and the resulting stalling prevents ribosomes from reaching downstream Ile and Leu codons at the same frequencies and thus prevents deacylation of the cognate Ile and Leu tRNAs. It's unclear whether this explanation is adequate to explain the complete lack of Ile or Leu tRNA deacylation observed even when amino acid recycling by the proteasome is inhibited-a treatment shown to exacerbate deacylation of cognate tRNAs in the single amino acid starvations and of Val tRNA in the triple starvation. As such, the statement in the Abstract "Notably, we could show that isoleucine starvation-specific stalling largely diminished under triple starvation, likely due to early elongation bottlenecks at valine codons" might be too strong and the word "possibly" would be preferred over "likely". It's also unclear why the proteins that are down-regulated in the triple starvation are not significantly enriched for stalling sites (Fig. 5B) given that the degree of skewing is comparable or greater than for -Val. This last point seems to undermine their conclusion in the Abstract that "that many proteins downregulated under BCAA deprivation harbor stalling sites, suggesting that compromised elongation contributes to decreased protein output." In the case of the double -Ile/-Leu starvation, a related phenomenon occurs wherein decoding rates are decreased for only the AUU Ile codon and only the AAU Ile tRNA becomes deacylated; although in this case increased RPFs in the 5' ends are not correlated with enrichment for Ile or Leu codons and, although not presented, apparently stall sites are not associated with the Ile codon in the double starvation. In addition, stalling sites are not enriched in the proteins down-regulated by the double starvation. Moreover, because Ile codons are not enriched in the 5'ends of CDS, it doesn't seem possible to explain the selective deacylation of the single Ile tRNA observed in the double starvation by the same "bottleneck" mechanism proposed to explain selective deacylation of only Val tRNAs during the triple starvation. This is another reason for questioning their "bottleneck" mechanism.

We thank the reviewer for their deep assessment, exhaustive reading, and constructive feedback, which have greatly contributed to improving the clarity and contextualization of our manuscript. We would first like to clarify that all experiments in this study were conducted in NIH3T3 mouse fibroblasts, not HeLa cells; we assume this was a misunderstanding and have verified that the correct cell line is consistently indicated throughout the manuscript. We also clarify that our data show that -Leu, double starvation, and to a lesser extent -Ile, downregulate mTORC1 signaling and TOP mRNA translation, whereas valine -Val and triple starvation had minimal effects on these pathways. We agree that some of our conclusions and observed phenomena were not explained in sufficient detail in the original version. To address this, we have significantly reworked the discussion, added complementary figures and clarified key points throughout the text, to better convey the underlying rationale and biological interpretation of our findings. We address each of the reviewer’s points in detail in the point-by-point responses below.

Specific comments (some of which were mentioned above):

-The authors have treated cells with CHX in the Ribo-Seq experiments, which has been shown to cause artifacts in determining the locations of ribosome stalling in vivo owing to continued elongation in the presence of CHX (https://doi.org/10.1371/journal.pgen.1005732 ). The authors should comment on whether this artifact could be influencing some of their findings, particular the results in Fig. 5C where the hungry codons are often present upstream of the A sites of called stalling sites in the manner expected if elongation continued slowly following stalling in the presence of CHX.

Response: We thank the reviewer for raising this important concern. We would like to clarify that our ribosome profiling protocol did not include CHX pretreatment of live cells. CHX was added only during the brief PBS washes immediately before lysis and in the lysis buffer itself. This approach aligns with best practices aimed at minimizing post-lysis ribosome run-off, and is intended to prevent the downstream ribosome displacement artifacts described by Hussmann et al. 2015, which result from pre-incubation of live cells with CHX for several minutes before harvesting. Furthermore, recent studies have demonstrated that CHX-induced biases are species-specific. For instance, Sharma et al. 2021 found that human (and mice) ribosomes are not susceptible to conformational restrictions by CHX, nor does CHX distort gene-level measurements of ribosome occupancy. This suggests that the use of CHX in the lysis buffer, as performed in our protocol, is unlikely to introduce significant artifacts in our ribosome profiling data. To further support this, we reanalyzed data from Darnell, Subramaniam, and O’Shea 2018, where the ribosome profiling samples were prepared without any CHX pretreatment or CHX in the wash buffer, and still observed similar upstream enrichments in their stalling profiles (see Supplementary Figure 2B-C in our manuscript). Additionally, in our previous work (Gobet et al. 2020), we compared ribosome dwell times with and without CHX in the lysis buffer and found no significant differences, reinforcing the notion that CHX use during lysis does not substantially affect the measurement of ribosome stalling. Given these considerations, we believe that CHX-related artifacts, such as downstream ribosome movement, are unlikely to explain the enrichment of hungry codons upstream of identified stalling sites in our data. We have now adjusted the Methods section to clarify this point.

-p. 12: "These starvation-specific DT and ribosome density modulations were also evident at the individual transcript level, as exemplified by Col1a1, Col1a2, Aars, and Mki67 which showed persistent Val-codon-specific ribosome density increases but lost Ile-codon-specific increases under triple starvation (Supplementary Figure 3A-D). " This conclusion is hard to visualize for any but Val codons. It would help to annotate the relevant peaks of interest for -Ile starvation with arrows.

Response: We agree and thank the reviewer for this observation. We have now annotated exemplary peaks in Supplementary Figure 3A–D to highlight ribosome pileups over Ile codons. However, we agree that it is still hard to visualize in the given Figure. Therefore, we added scatter plots for each of the transcripts that show the RPM of each position in the Ctrl vs starvation to allow for a better illustration of the milder effects upon Ile starvation (Supplementary Figure 4).

-To better make the point that codon-specific stalling under BCAA starvation appears to be not driven by codon usage, rather than the analysis in Fig. 1H, wouldn't it be better to examine the correlation between increases in DT under the single amino acid starvation conditions and the codon frequencies across all codons?

Response: We appreciate the suggestion. We have now added an additional analysis correlating the change in DT with codon usage frequency for each starvation condition. This is included in Supplementary Figure 5A-D and supports our interpretation that codon frequency alone does not explain the observed stalling behavior.

-p. 13, entire paragraph beginning with "Our RNA-seq and Ribo-seq revealed a general activation of stress response pathways across all starvations..." It is difficult to glean any important conclusions from this lengthy analysis, and the results do not appear to be connected to the overall topic of the study. If there are important conclusions here that relate to the major findings then these connections should be made or noted later in the Discussion. If not, perhaps the analysis should be largely relegated to the Supplemental material.

Response: We thank the reviewer for this comment. The paragraph in question is intended to provide a global overview of transcriptional and translational responses across the starvation conditions. It serves both as a quality control (e.g., PCA clustering and global shifts in RPF/RNA-seq profiles), and to confirm that expected starvation-induced responses are among the strongest detectable signals separating the starved samples from the control. Indeed, these observations establish that the perturbations are effective and that hallmark nutrient stress responses are globally engaged across conditions. Importantly, very few studies to date have examined transcriptional and translational responses under single or combined branched-chain amino acid (BCAA) starvation conditions. It therefore remains unclear to what extent BCAA depletion broadly remodels gene expression and translation. Our analysis contributes to addressing this gap, revealing that while certain stress pathways are commonly induced, others show condition-specific patterns such as we observed for -Ile starvation. To maintain focus, we have kept the detailed pathway analyses and transcript-level enrichments in the Supplement and rewritten the corresponding text in a more compact manner, reducing it by more than one third.

-p. 15: "Together, these findings highlight that BCAA starvation triggers a combination of effects on initiation and elongation, with varying dynamics by amino acid starvation." I take issue with this statement as it appears that translation is reduced primarily at the initiation step for all conditions except -Ile. As noted above, these data are never menitioned in the DISCUSSION as to why only -Ile would show a marked elongation component to the inhibition whereas -Val gives the greatest amount of ribosome stalling.

Response: We acknowledge the reviewer’s point. While the polysome profiles (Figure 3F-H) directly indicate that most conditions repress initiation, codon- and condition-specific elongation defects can still contribute to reduced protein output, even if they are not always detectable as global polysome shifts. Polysome profiles reflect the combined outcome of reduced initiation (which decreases polysome numbers) and ribosome stalling (which can, but does not always have to, increase ribosome density on individual transcripts, potentially counteracting the effects of reduced initiation). For valine starvation strong stalling occurs very early in the CDS (Figure 5F). This bottleneck restricts overall ribosome movement to downstream regions. Thus, while elongation is profoundly impaired, the total number of ribosomes per transcript (which polysome signals largely reflect) may appear low due to reduced overall ribosome traffic. In contrast, isoleucine codon stalling tends to occur also further downstream on the transcript (Figure 5F), allowing ribosomes to accumulate in larger numbers on the mRNA, leading to a clearer "elongation signature" in polysome profiles (Figure 3F, H). Additionally, we observed slightly higher inter-replicate variance for isoleucine starvation (Supplementary Figure 6B), which may have reduced the number of statistically significant stalling sites extracted compared to valine. We have revised the main text and discussion to clarify these points.

-I cannot decipher Fig. 4D and more detail is required to indicate the identity of each column of data.

Response: We thank the reviewer for pointing this out. Figure 4D (now Figure 4E) presents an UpSet plot, which is a scalable alternative to Venn diagrams commonly used to visualize intersections across multiple sets. Briefly, each bar in the upper plot represents the number of transcripts with increased 5′ ribosome coverage (Δpi < -0.15; p < 0.05) shared across the conditions indicated in the dot matrix below. Each column in the dot matrix highlights the specific combination of conditions contributing to a given intersection (e.g., dots under “Val” and “Triple” show the overlap between these two). To improve clarity, we have expanded the figure legend accordingly and now refer to the UpSetR methodology in the main text.

-In Fig. 4E, one cannot determine what the P values actually are, which should be provided in the legend to confirm statistical significance.

Response: Thank you for pointing that out. The legend in Figure 4E (now Figure 4F) for the p-values was accidentally removed during figure editing. We have added the legend back, so that the statistical significance is clear.

-It's difficult to understand how the -Leu condition and the Double starvation can produce polarized RPFs (Fig. 4A) without evidence of stalling at the cognate hungry codons (Fig. 4E), despite showing later in Fig. 5A that the numbers of stall sites are comparable in those cases to that found for -Ile.

Response: We appreciate this comment, which points to an important property of RPF profiles under nutrient stress. As shown in Figure 4A, all starvation conditions induce a degree of 5′ ribosome footprint polarization, a pattern that can be observed under various stress conditions and perturbations (Allen et al. 2021; Hwang and Buskirk 2017; Li et al. 2023). This general 5′ bias likely reflects a combination of slowed elongation and altered ribosome dynamics and is not necessarily linked to codon-specific stalling. However, Val and Triple starvation show a much stronger and more asymmetric polarization, characterized by pronounced 5′ accumulation and 3′ depletion of ribosome density. To better illustrate this, we have updated the visualization of polarity scores and added a new bar chart summarizing the number of transcripts showing strong 5′ polarization under each condition. This quantification highlights that the effect is markedly more prevalent under Val and Triple conditions than under Leu or Double starvation. In addition, Figure 4F demonstrates that this polarity is codon-specific under Val and Triple starvation. We clarify that this analysis tests for enrichment of specific codons near the start codon among the polarized transcripts and does not directly assess stalling. The observed enrichment of Val codons in the 5′ regions of polarized transcripts supports the interpretation that early elongation delays contribute to the RPF shift. In contrast, no such enrichment is observed for Leu starvation, reinforcing that Leu-induced polarity is not driven by stalling at Leu codons. While Figure 5 shows a similar number of peak-called stalling sites in -Leu, -Ile, and Double starvation, we note that Ribo-seq signal variability under Ile starvation was higher, which may have limited statistical power for detecting stalling sites, even though clear dwell time increases were observed at specific codons. Additionally, we have improved the metagene plots depicting total ribosome footprint density in Figure 4A. The previous version incorrectly showed sharp drops at CDS boundaries due to binning artifacts. The updated version more accurately reflects the density distribution and further highlights the stronger polarization in Val and Triple conditions. Together, these clarifications and improvements within the main text now more clearly distinguish between general polarity effects and codon-specific stalling.

-Fig. 5B: the P values should be given for all five columns, and it should be explained here or in the Discussion why the authors conclude that stalling is an important determinant for reduced translation when a significant correlation seems to exist only for the -Val condition and not even for the Triple condition.

Response: We thank the reviewer for this important observation. In response, we have revised both the text and the figures to provide a clearer and biologically more meaningful representation of the relationship between ribosome stalling and reduced protein output. Specifically, we have replaced the previous Figure 5B with a new analysis that stratifies transcripts based on the number of identified stalling sites. This updated analysis, now shown in Figure 5B, reveals that under Val and Triple starvation conditions, proteins that are downregulated tend to originate from transcripts with multiple stalling sites. Importantly, the corresponding p-values for all five conditions are now explicitly shown in the figure (as red lines). As the reviewer correctly notes, only the Val condition shows a statistically significant enrichment when considering overall overlap. Triple starvation shows a similarly high proportion of overlap (72.3%) but does not reach statistical significance, likely due to the more complex background composition under combined starvation, which increases the expected overlap and reduces statistical power. By stratifying transcripts by the number of stalling sites, we uncover that transcripts with ≥2 stalling sites are enriched among downregulated proteins specifically under Val and Triple conditions, providing a more robust indication of the link between stalling and translation repression under Valine deprivations. We believe this refined approach, prompted by the reviewer’s comment, offers a clearer and biologically more relevant perspective on the role of ribosome stalling. The original analysis previously shown in Figure 5B is now provided as Supplemental Figure 10C for transparency and comparison. We have clarified this in the revised text and now interpret the relationship more cautiously.

-p. 17: "Of note, in cases where valine or isoleucine codons were present just upstream (rather than at) the stalling position, we noted a strong bias for GAG (E), GAA (E), GAU (D), GAC (D), AAG (K), CAG (Q), GUG (V) and GGA (G) (Val starvation) and AAC (N), GAC (D), CUG (L), GAG (E), GCC (A), CAG (Q), GAA (E) and AAG (K) (Ile starvation) at the stalling site (Supplementary Figure 7B)." The authors fail to explain why these codons would be present in the A sites at stalling sites rather than the hungry codons themselves, especially since it is the decoding times of the hungry codons that are increased according to Fig. 1A-E. As suggested above, is this a CHX artifact?

Response: We agree that the observation that the listed codons are enriched at identified stalling positions (now Supplementary Figure 10C), while the depleted amino acid codon is located upstream, is a finding that needs more detailed explanation. Importantly, this phenomenon is not attributable to CHX artifacts, as our Ribo-seq protocol employs CHX solely during brief washes and lysis to prevent post-lysis ribosome run-off, rather than live-cell pre-treatment. Instead, we propose two hypotheses to explain this pattern: Firstly, many of these enriched codons are already inherently slow-decoded with longer DTs even under control conditions (Supplementary Figure 5H, newly added). Together with the upstream hungry codons they might form a challenging consecutive decoding environment, which results in an attenuated ribosome slowdown downstream after the hungry codon. Second, ribosome queuing may further explain this pattern. When a ribosome encounters a critically hungry codon and stalls, subsequent ribosomes can form a queue. The codon within the A-site of the queued ribosome would be (more or less) independent of the identity of the hungry codon itself that caused the initial stall. Since the listed codons have a high frequency within the transcriptome (Supp. Fig 5B), they therefore have an increased likelihood of appearing at this “stalling site”. Importantly, both of these phenomena are not necessarily represented by a general increase of DT on all of the listed codons and would therefore only be captured by the direct extraction of stalling sites but might be averaged out in the global dwell time analysis. We mention this phenomenon now in the Discussion.

-Fig. 5D: P values for the significance, or lack thereof, of the different overlaps should be provided.

Response: Thanks for pointing out this omission. We have now computed hypergeometric p-values for comparisons shown in Figure 5D and Figure 5E, and report them directly in the main text. As described, the overlap in stalling sites between Val and triple starvation is highly significant (2522 positions, p < 2.2×10⁻¹⁶), while overlaps involving Ile-specific stalling positions are smaller but still statistically robust (e.g., 149 positions for Ile – Triple, p = 1.77×10⁻⁵²). Notably, we also calculated p-values at the transcript level and found that a large fraction of transcripts with Ile-specific stalling under single starvation also stall under triple starvation, though often at different positions (1806 transcripts, p = 1.78×10⁻⁵⁸). These values are now included in the revised results section to support the interpretation of these overlaps.

-p. 17: "Nonetheless, when we examined entire transcripts rather than single positions, many transcripts that exhibited isoleucine-related stalling under Ile starvation also stalled under triple starvation, but at different sites along the CDS (Figure 5E). This finding is particularly intriguing, as it suggests that while Ile-starvation-specific stalling sites may shift under triple starvation, the overall tendency of these transcripts to stall remains." The authors never come back to account for this unexpected result.

Response: Thank you for highlighting this point. We've incorporated this finding as part of the proposed "bottleneck" scenario. While the isoleucine-specific stalling sites identified under Ile starvation do shift or disappear under triple starvation, we've observed that the same transcripts still tend to exhibit stalling. However, this now primarily occurs at upstream valine codons. We interpret this as a consequence of early elongation stalling caused by strong pausing at Val codons. This restriction on ribosome progression effectively prevents ribosomes from reaching the original Ile stalling sites. Therefore, the stalling sites identified under triple starvation are largely explained by the Val codons, reflecting a redistribution of stalling rather than its loss. To further clarify this crucial point, we've now explicitly mentioned Figure 5D-E again in the subsequent paragraph, which introduces the bottleneck theory.

-It seems very difficult to reconcile the results in Fig. 5F with those in Fig. 4A, where similar polarities in RPFs are observed for -Ile and -Val in Fig, 4A but dramatically different distributions of stalling sites in Fig. 5F. More discussion of these discrepancies is required.

Response: Thank you for pointing this out. The apparent discrepancy between the RPF profiles shown in Figure 4A and the stalling site distributions in Figure 5F likely reflects the fact that RPF polarization includes both general (unspecific) and codon-specific components. Figure 4A displays total ribosome footprint density, capturing both broad stress-induced effects and codon-specific contributions, whereas Figure 5F focuses specifically on peak-called stalling sites, representing localized and statistically significant pauses. Importantly, we would like to emphasise that Fig 4 shows that -Val and -Ile starvation exhibit different responses and not the same patterns. To make these differences even clearer, we have now updated the visualizations in Figure 4, including improved polarity plots and a new bar chart summarizing the number of transcripts with strong 5′ polarization. These additions highlight that the RPF profiles under -Val starvation are more pronounced and asymmetric, particularly due to 3′ depletion, while the polarity under -Ile is milder and a distinct, much smaller subset of transcripts appears to show polarity score shifts. We believe the updated figures and accompanying explanations now make these distinctions clearer.

• p. 18: " These isoacceptor-specific patterns correlate largely with the particular subsets of leucine and isoleucine codons that stalled (Figure 1A)." This correlation needs to be addressed for each codon-anticodon pair for all of the codons showing stalling in Fig. 1A.

Response: We thank the reviewer for this important comment. In the revised manuscript, we have expanded the relevant sections to address codon–anticodon relationships more thoroughly. We now explicitly match codons that exhibited increased dwell times under starvation to the corresponding tRNA isoacceptors whose charging was affected, and we provide a clearer discussion of the caveats involved. As noted by the reviewer, this correlation is not straightforward, as it is complicated by wobble base pairing, anticodon modifications, and the fact that multiple codons can be decoded by more than one isoacceptor, and vice versa. Moreover, in our qPCR-based tRNA charging assay, certain isoacceptors cannot be distinguished due to highly similar sequences (e.g., LeuAAG and LeuUAG, and LeuCAA and LeuCAG), which limits resolution for exact pairing. In addition, we did not assess absolute tRNA abundance, which may further influence decoding capacity. Nevertheless, where resolution is possible, the patterns align well: All tRNAVal isoacceptors became uncharged under Val and triple starvation, matching the consistent dwell time increases across all Val codons. Only tRNAIleAAU (decoding AUU and AUC) was deacylated, matching to these codons showing increased dwell times, while AUA (decoded by still-charged tRNAIleUAU) did not. Only CUU (decoded by uncharged tRNALeuGAA) showed increased dwell time. A mild deacylation of the other Leu isoacceptors was observed, but isoacceptor-level resolution is limited by assay constraints. However, these rather minimal tRNA and DT changes were consistent with more dominant initiation repression rather than elongation stalls. To support this analysis, we included an illustrative figure (now in Supplementary Figure 12F) summarizing the codon–anticodon matches.

-p. 19: "For instance, in our double starvation condition, unchanged tRNA charging levels (Figure 6E) may result from a pronounced downregulation of global translation initiation, likely driven by the activation of stress responses (Figure 2), subsequently lowering the demand for charged tRNAs as it has been observed previously for Leu starvation 39.” This seems at odds with the comparable down-regulation of protein synthesis for the Double starvation and -Leu and -Ile single starvations shown in Fig. 3C. Also, in the current study, Leu starvation does lower charging of certain Leu tRNAs.

Response: We thank the reviewer for raising this important point. In the revised manuscript, we have clarified this section and now offer a more refined interpretation of the tRNA charging patterns observed under double starvation. While Figure 3C shows a comparable reduction in global protein synthesis across the -Leu, -Ile, and double starvation conditions, it needs to be considered that the OPP assay has limited sensitivity. It operates in a relatively low fluorescence intensity range and is subject to background signal, which may obscure subtle differences between conditions. Moreover, other factors such as changes in protein stability or turnover could also contribute to the observed differences. Therefore, inter-condition differences in translation repression should be interpreted with caution. However, based on our stress response analysis (Figure 2), mTORC1 inactivation appears strongest under double starvation, likely leading to more profound suppression of translation initiation. This would reduce the overall demand for charged tRNAs and could explain why no detectable tRNA deacylation was observed under double starvation, even though mild uncharging of Leu isoacceptors occurred under -Leu, which exhibited a milder stress response. This distinction is consistent with the observed mild dwell time increases for one Leu codon under -Leu, but not in the double condition. Similarly, the absence of Ile codon stalling and tRNA deacylation under double starvation may be attributed to stress-driven reductions in elongation demand, preventing the tRNA depletion and codon-specific delays observed under single Ile starvation. A more direct clarification is now included in the revised manuscript.

Reviewer #2 (Significance):

The results here are significant in showing that starvation for a single amino acid does not lead to deacylation of all isoacceptors for that amino acid and in revealing that starvation for one amino acid can prevent deacylation of tRNAs for other amino acids, as shown most dramatically for the selective deacylation of only Val tRNAs in the triple BRCAA starvation condition. For the various reasons indicated above, however, I'm not convinced that their "bottleneck" mechanism is adequate to explain this phenomenon, especially in the case of the selective deacylation of Ile vs Leu tRNA in the Double starvation regime. It's also significant that deacylation leads to ribosome build-up near the 5'ends of CDS, which seems to be associated with an enrichment for the hungry codons in the case of Val and Ile starvation, but inexplicably, not for Leu or the Double starvations. This last discrepancy makes it hard to understand how the -Leu and Double starvations produce RPF buildups near the 5 ends of CDSs. In addition, the claim in the Discussion that "our data also highlight the importance of the codon positional context within mRNAs, indicating that where a codon is located within the CDS can influence both the extent of ribosomal stalling and overall translation efficiency during nutrient stress" overstates the strength of evidence that the stalling events lead to substantial decreases in translational efficiencies for the affected mRNAs, as the stalling frequency and decreased protein output are significantly correlated only for the -Val starvation, and the data in Fig. 3 D-H suggest that the reductions in protein synthesis generally occur at the level of initiation, even for -Val starvation, with a contribution from slow elongation only for -Ile-which is in itself difficult to understand considering that stalling frequencies are highest in -Val. Thus, while many of the results are very intriguing and will be of considerable interest to the translation field, it is my opinion that a number of results have been overinterpreted and that important inconsistencies and complexities have been overlooked in concluding that a significant component of the translational inhibition arises from the increased decoding times at hungry codons during elongation and that the selective deacylation of Val tRNAs in the Triple starvation can be explained by the "bottleneck" mechanism. The complexities and limitations of the data and their intepretations should be discussed much more thoroughly in the Discussion, which currently is devoted mostly to other phenomena often of tangential importance to the current findings. A suitably revised manuscript would clearly state the limitations and caveats of the proposed mechanisms and consider other possible explanations as well.

Again, we thank the reviewer for the valuable insights and constructive critiques. We believe that the concerns regarding potential overinterpretation and inconsistencies have now been addressed through clearer explanations and more cautious interpretation throughout the revised manuscript. We also agree that the original Discussion included aspects that, while interesting, were of secondary importance. In light of the reviewer’s suggestions, we have restructured and rebalanced the Discussion to focus more directly on the key findings and their implications. Importantly, we wish to clarify that we do not propose the elongation bottleneck model as a general mechanism across all conditions. In particular, for double (Leu/Ile) starvation, we attribute the observed effects primarily to stress response–mediated translational repression, and not to codon-specific stalling or tRNA depletion. We believe that this distinction is now more clearly conveyed in the revised manuscript.

Reviewer #3 (Evidence, reproducibility and clarity):

Summary

Worpenberg and colleagues investigated the translational consequences of branched-chain amino acid (BCAA) starvation in mouse cells. Limitation of individual BCAAs has been reported to cause codon-specific and global translational repression. In this paper, the authors use RNA-seq, ribosome profiling (Ribo-seq), proteomics, and tRNA charging assays to characterize the impacts of individual and combined depletion of leucine, isoleucine, and valine on translation. They find that BCAA starvation increases codon-specific ribosome dwell times, activates global translational stress responses and reduces global protein synthesis. They infer that this effect is due to decreased translation initiation and codon-specific translational stalling. They find that the effects of simultaneous depletion are non-additive. In valine and triple (valine, leucine, and isoleucine) depletion, they show that affected transcripts have a high density of valine codons early in their coding sequences, creating an "elongation bottleneck" that obscures the impact of starvation of other amino acids. Finally, they identify isoacceptor-specific differences in tRNA charging that help explain the codon-specific effects that they observe.

We find the major findings convincing and clear. We find that some results are incompletely explained. We suggest an additional experiment and also have some minor comments that we hope will improve clarity and rigor.

We thank the reviewer for the thorough and constructive feedback. We appreciate the recognition of our main findings and the helpful suggestions for improving the manuscript. Below we address each point in detail.

Major comments

Figure 3O: In this figure and the associated text, the authors try to determine whether differences in protein degradation can explain why some proteins have higher ribosome density but lower proteomic expression. However, since this analysis relies on published protein half-lives from non-starvation conditions and on the assumption that protein synthesis has entirely stopped, we are not convinced it is informative for this experimental context. It does not distinguish between a model in which protein synthesis has been reduced by stalling and a model in which both protein synthesis and degradation rate have increased, which are both consistent with their Ribo-seq and proteomic data. To address this issue, the authors should either perform protein half-life measurements under their starvation conditions, or more clearly explain these two models in the text and acknowledge that they cannot distinguish between them.

Response: We agree with the reviewer that our current analysis, which is based on protein half-lives obtained under non-starvation conditions, can not definitively separate the effects of reduced translation from those of increased protein degradation. We have revised the relevant section in the manuscript to more clearly state that this analysis is correlative in nature and serves only to explore one possible explanation for the observed disconnect between ribosome density and protein levels. We now also explicitly acknowledge that our dataset does not allow us to distinguish between a model in which protein output is reduced due to stalling and one in which both translation and degradation rates are altered. However, the observed log2FC in the proteomics data are often milder than expected based on complete-medium condition half-life alone, which would be difficult to reconcile with a dominant contribution from global protein destabilization. That said, we also acknowledge that protein degradation is highly context- and protein-specific, and that proteolytic regulation might still play a role. Performing a direct protein half-life measurement under our starvation conditions would indeed be required to rigorously test this, but such an experiment is outside the scope of this study. We now highlight this as a limitation and a valuable direction for future work, and we have softened any interpretations in the main text to reflect the uncertainty regarding the contribution of protein stability changes.

Minor comments

Figure 1G: Why does intracellular valine seem to be less depleted under starvation conditions than intracellular leucine or isoleucine? Are the limits of detections different for different amino acids? The authors should acknowledge this discrepancy and comment on whether it has any implications for interpretation of their results.

Response: We thank the reviewer for this important point. While valine appears slightly less depleted than leucine or isoleucine in Figure 1G, the fold changes and absolute reductions are strong for all three BCAAs, including valine. To further illustrate this, we have added a supplementary bar chart showing the measured intracellular concentrations in µmol/L, including mean and variance across five biological replicates (Supplementary Figure 5A). We believe that the variation may reflect technical factors, such as differences in detection sensitivity or ionization efficiency between amino acids in the targeted metabolomics assay and, therefore, that the observed difference does not have a meaningful impact on the interpretation of our results. We now directly acknowledge these differences in the main text.

Figure 1H: These data do not appear to meet the assumptions for linear regression. We suggest either reporting a Spearman R correlation (as the data appears linear in rank but not absolute value), or remove it entirely - we think the plot without statistics is sufficient.

Response: We thank the reviewer for the suggestion. In the revised manuscript, we removed the statistical annotation and retained only the trend line to illustrate the general pattern. We agree that this visualization alone is sufficient to support the qualitative point we aimed to convey.

Figure 2B: The in-text description of this figure states that "most" ISR genes show a "robust induction," but only three genes are shown in the figure, two of which are upregulated. The authors should instead specify that 2 out of the 3 genes profiled were robustly induced.

Response: We have rephrased the sentence to say “two of the three genes profiled…” for precision and consistency with the data shown.

Figure 2D: Please include the full, uncropped blots in the supplementary materials.

Response: We have now added the full, uncropped western blots to the supplementary material (Supplementary Figure 8).

Figure 2E: Swap the positions of the RPS6 and 4E-BP1 plots so they line up with their respective blots to make these figures easier to interpret. Authors should consider doing a one-way ANOVA and post-hoc analysis, if we correctly understand that they are making a conclusion about the difference between multiple groups in aggregate.

Response: We thank the reviewer for the suggestion. The alignment of the RPS6 and 4E-BP1 plots with their respective blots has been corrected. As this panel focuses on comparisons to the control condition only, we have retained the original presentation.

Figure 4B: Panel A in this figure is very convincing, and these plots don't add additional information. The authors could consider removing them. If this panel stays in, we suggest removing the "mid index" plot, since it is never referenced in the text and doesn't seem relevant to the message of the figure.

Response: We appreciate the feedback. While we considered removing panel B as suggested, we decided to retain it because it provides a useful summary of panel A. To improve clarity and visual interpretation, we replaced the original boxplot with a bar plot displaying mean values and SEM error bars. We believe the bar plot now nicely illustrates that Val and Triple starvation lead to stronger effects, especially in the reduction of the 3′ index. The “mid index” plot, which was not referenced in the text and did not contribute to the central message, has been removed as suggested.

Figure 4E: Why is there a reduction in frequency of a Leu and a Val codon under Ile starvation?

Response: Thank you for highlighting this observation. The reduction in the frequency of a specific Leu and Val codon under Ile starvation in Figure 4F (former Figure 4E) is indeed intriguing. This figure reflects codon usage in the first 20% of the CDSs among the subset of transcripts that exhibit a footprint polarization under each starvation condition. As such, the observed depletion likely arises from the specific transcript composition of the polarized subset under -Ile, which differs from that under -Val or other conditions. Importantly, this pattern is not consistently observed when analyzing the full transcripts (another Leu codon is affected), indicating that it is not a systematic depletion of these codons. One possibility is that an increased frequency of Ile codons (AUC) within the constrained region may lead to a relative underrepresentation of other codons, such as Leu and Val. Alternatively, this may reflect non-random codon co-occurrence patterns within specific transcripts. While our current data do not allow us to investigate this further, we acknowledge these as speculative explanations and now mention this point in the Discussion as a potential avenue for future study.

Figure 5G: There appears to be one Val codon early in the Hint1 transcript without much stalling under triple or valine starvation conditions. The authors should acknowledge this and comment on why this may be.

Response: We thank the reviewer for pointing this out. While the Hint1 transcript indeed contains a valine codon early in its CDS, no clear stalling peak was observed at that position under valine or triple starvation. Several factors may contribute to this: local sequence context can influence ribosome pausing, and not all cognate codons necessarily lead to detectable stalling even under amino acid starvation. Additionally, coverage at the 5′ end of Hint1 is relatively sparse in our dataset, and potential mappability limitations, such as regions with low complexity or repetitive elements, may further reduce resolution at specific sites. We now briefly mention this in the manuscript to clarify the possible causes.

Figure 5B: In the text referencing this figure, the authors state that "a high number of downregulated proteins with associated ribosome stalling sites did not show an overall decreased mean RPF count...as it would be expected from translation initiation defects, linking these stalling sites directly to proteomic changes." However, RPF is affected both by stalling (increases RPF) and initiation defects (decreases RPF). A gene with both stalling and decreased initiation may appear to have no RPF change. The data does suggest a contribution from stalling, but the authors should also acknowledge that reduced initiation may also be playing a role.

Response: We agree with the reviewer comment. Our cited statement should indeed be more nuanced. The reviewer correctly points out that RPFs are influenced by both increased ribosome density due to stalling and decreased ribosome density due to reduced initiation. Therefore, a gene experiencing both stalling and reduced initiation might appear to have no net change in RPF, or even a slight increase if stalling is dominant. Thus, while the presence of stalling sites strongly suggests a contribution from compromised elongation to reduced protein output, we cannot definitively rule out a concurrent role for reduced initiation, even in cases where RPF counts are not globally decreased. We revised this section in the manuscript to acknowledge this interplay.

Figure 5E: the black text on dark brown in the center of the Venn diagram is difficult to read. The diagram should either have a different color scheme, or the text in the center should be white instead of black for higher contrast.

Response: We have adjusted the text color for better contrast and improved readability.

Supplementary Figure 1C: The ribosome dwell time data in this study is described as "highly correlated" with another published dwell time dataset, but the P and E site data do not seem strongly correlated. The authors should remove the word "highly."

Response: We have removed the word “highly” to have a more cautious interpretation in the text.

Supplementary Figure 3E: Not all of the highlighted codons in this figure are ones with prolonged dwell times. To clarify the point that dwell time change is not related to codon frequency, this figure should only highlight codons that have a significantly prolonged dwell time in at least one starvation condition.

Response: We thank the reviewer for pointing this out. To improve clarity, we have revised the figure and now specifically highlight codons with significantly prolonged dwell times with stars.

Supplementary Figure 5C: The gene Chop is mentioned in the main text when referencing this figure, but is absent from the heatmap.

Response: We thank the reviewer for noting this. The gene Chop is annotated under its alternative name Ddit3 in the current version of the heatmap and is indeed present. To avoid confusion, we have now updated the label in the figure to display Chop (Ddit3) directly.

Supplementary Figure 7A: The authors could clarify this figure by adding additional language to either the figure panel or the figure legend specifying that the RPM metric being used comes from Ribo-seq.

Response: We have updated the legend to explicitly state that the RPM values shown are derived from Ribo-seq data.

Supplementary Figure 7D: The metric used to describe the spatial relationship between the first valine and isoleucine codons in transcripts in this figure seems to be describing something conceptually similar to the stalling sites in Figure 5G, but uses a different metric. These figures would be easier to interpret if these spatial relationships were presented in a consistent way throughout the manuscript.

Response: We thank the reviewer for this helpful observation. Supplementary Figure 7D (now Supplementary Figure 11B) originally used a gene-length-normalized metric to describe codon spacing, whereas Figure 5G depicted absolute nucleotide distances to stalling sites. To ensure consistency across the manuscript, we have now updated Supplementary Figure 11B to also use absolute distances. We believe this adjustment improves clarity and allows for a more direct comparison between spatial codon patterns and stalling events.

Discussion:

Reader understanding would be improved if the relevance of paragraphs were established in the first sentence. For instance, in the paragraphs about adaptive misacylation and posttranscriptional modifications, it is unclear until the end of the paragraph how these topics are relevant. Introducing the relevant aspects of the study (the fact that some starvation conditions have less severe effects and the observation about m6A-related mRNAs) at the beginning of these paragraphs would improve clarity.

Response: We thank the reviewer for this helpful comment. We agree that the flow and clarity of the Discussion can be improved by making the relevance of each paragraph clearer from the outset. In the revised manuscript, we have restructured these sections to better highlight the connection between each topic and our main findings. These changes also align with suggestions from Reviewer 2, and we believe they help to focus the Discussion more tightly around the core insights of our study.

The authors should provide more information and speculation about possible physiological relevance of their findings, particularly about the way that the effects of triple starvation are highly valine-dependent. Are there physiological conditions under which starvation of all three BCAAs is more likely than starvation of one or two of them? If so, are there any reasons why a valine-based bottleneck might be advantageous?

Response: We appreciate the reviewer's insightful question regarding the physiological relevance of our findings, particularly the valine-dependent bottleneck observed under triple BCAA starvation. This prompts a crucial discussion on the broader biological context of our work.

While complete starvation of all three BCAAs might be less frequent than individual deficiencies, such conditions are physiologically relevant in several contexts. In prolonged fasting, starvation, or severe cachectic states associated with chronic diseases (e.g., advanced cancer, critical illness), systemic amino acid pools, including BCAAs, can become significantly depleted due to increased catabolism and insufficient intake (Yu et al. 2021). Moreover, certain specialized diets or therapeutic strategies aim to modulate BCAA levels. For instance, in some Maple Syrup Urine Disease (MSUD) management protocols, BCAA intake is severely restricted to prevent the accumulation of toxic BCAA metabolites (Mann et al. 2021). Similarly, emerging cancer therapies sometimes explore nutrient deprivation strategies to selectively target tumor cells, which could involve broad BCAA reduction (e.g. Sheen et al. 2011; Xiao et al. 2016).

In these contexts, a valine-based bottleneck, as we describe, could indeed represent an adaptive strategy. If valine-tRNAs are particularly susceptible to deacylation and valine codons are strategically enriched at the 5' end of transcripts, stalling at these early positions could serve as a rapid "gatekeeper" for global translation. This early-stage inhibition would conserve cellular energy and available amino acids by quickly reducing the overall demand for charged tRNAs. Such a mechanism could potentially prioritize the translation of a subset of proteins that might have different codon usage biases or are translated via alternative, less valine-dependent mechanisms. This aligns with the concept of a multi-layered translational control where global initiation repression (as reflected in mTORC1 inhibition and polysome profiles) is complemented by specific elongation checkpoints, allowing for a more nuanced and adaptive response to severe nutrient stress.

Reviewer #3 (Significance):

Nature and significance of the advance

The main contribution of this work is to demonstrate that depletion of multiple amino acids simultaneously impacts translation elongation in ways that are not necessarily additive. These impacts can depend on the distribution of codons in a transcript. It adds to a growing body of work showing that essential amino acid starvation can cause codon-specific ribosome stalling. The authors suggest that the position-dependent stalling they observe could be a novel regulatory mechanism to alleviate the effects of multi-amino acid starvation. However, it is not fully clear from the paper what the significance of a valine-based regulatory adaptation to BCAA starvation is, or whether simultaneous starvation of all three BCAAs is of particular physiological relevance. The paper's primary contribution is mainly focused on the similarity between valine and triple BCAA starvation, and it provides limited insight into the effects of combined depletion of two BCAAs.

Context of existing literature

Although ribosome profiling does not distinguish between actively-elongating and stalled ribosomes, sites with higher read coverage, and thereby higher inferred dwell time, can be used to infer ribosome stalling (Ingolia 2011). Various downstream effects of essential amino acid depletion have been documented, such as leucine deficiency being sensed by mTORC1 via leucyl-tRNA synthetase (Dittmar 2005, Han 2012), and shared transcriptional responses among many amino acid depletion conditions (Tang 2015). These authors have previously measured the translational effects of nutrient stress using ribosome profiling (e.g., Gobet 2020), as have others (Darnell 2018, Kochavi et al. 2024). The present work represents the first study (to our knowledge) combining BCAA depletions, representing an incremental and useful contribution to our understanding of translational responses to stress conditions.

Audience

This work is of interest to investigators studying the response of human cells in stress conditions, such as in human disease, as well as investigators studying the basic biology of eukaryotic translational control.

Reviewer expertise: mRNA decay and translation regulation in bacteria.

We hope the authors have found our comments thoughtful and useful. We welcome further discussion or clarification via email: Juliana Stanley (julianst@mit.edu) and Hannah LeBlanc (leblanch@mit.edu).

We sincerely thank the reviewers for their thoughtful and constructive feedback, as well as for their careful and thorough reading of our manuscript. We also gratefully acknowledge the invitation for further discussion and would be happy to engage in future correspondence.

References

Allen, George E., Olesya O. Panasenko, Zoltan Villanyi, Marina Zagatti, Benjamin Weiss, Lucile Pagliazzo, Susanne Huch, et al. 2021. “Not4 and Not5 Modulate Translation Elongation by Rps7A Ubiquitination, Rli1 Moonlighting, and Condensates That Exclude eIF5A.” Cell Reports 36 (9): 109633. https://doi.org/10.1016/j.celrep.2021.109633.

Darnell, Alicia M., Arvind R. Subramaniam, and Erin K. O’Shea. 2018. “Translational Control through Differential Ribosome Pausing during Amino Acid Limitation in Mammalian Cells.” Molecular Cell 71 (2): 229-243.e11. https://doi.org/10.1016/j.molcel.2018.06.041.

Dittmar, Kimberly A., Michael A. Sørensen, Johan Elf, Måns Ehrenberg, and Tao Pan. 2005. “Selective Charging of tRNA Isoacceptors Induced by Amino-Acid Starvation.” EMBO Reports 6 (2): 151–57. https://doi.org/10.1038/sj.embor.7400341.

Elf, Johan, Daniel Nilsson, Tanel Tenson, and Mans Ehrenberg. 2003. “Selective Charging of tRNA Isoacceptors Explains Patterns of Codon Usage.” Science (New York, N.Y.) 300 (5626): 1718–22. https://doi.org/10.1126/science.1083811.

Gobet, Cédric, Benjamin Dieter Weger, Julien Marquis, Eva Martin, Nagammal Neelagandan, Frédéric Gachon, and Felix Naef. 2020. “Robust Landscapes of Ribosome Dwell Times and Aminoacyl-tRNAs in Response to Nutrient Stress in Liver.” Proceedings of the National Academy of Sciences of the United States of America 117 (17): 9630–41. https://doi.org/10.1073/pnas.1918145117.

Hussmann, Jeffrey A., Stephanie Patchett, Arlen Johnson, Sara Sawyer, and William H. Press. 2015. “Understanding Biases in Ribosome Profiling Experiments Reveals Signatures of Translation Dynamics in Yeast.” Edited by Michael Snyder. PLOS Genetics 11 (12): e1005732. https://doi.org/10.1371/journal.pgen.1005732.

Hwang, Jae-Yeon, and Allen R. Buskirk. 2017. “A Ribosome Profiling Study of mRNA Cleavage by the Endonuclease RelE.” Nucleic Acids Research 45 (1): 327–36. https://doi.org/10.1093/nar/gkw944.

Juszkiewicz, Szymon, and Ramanujan S. Hegde. 2017. “Initiation of Quality Control during Poly(A) Translation Requires Site-Specific Ribosome Ubiquitination.” Molecular Cell 65 (4): 743-750.e4. https://doi.org/10.1016/j.molcel.2016.11.039.

Li, Fajin, Jianhuo Fang, Yifan Yu, Sijia Hao, Qin Zou, Qinglin Zeng, and Xuerui Yang. 2023. “Reanalysis of Ribosome Profiling Datasets Reveals a Function of Rocaglamide A in Perturbing the Dynamics of Translation Elongation via eIF4A.” Nature Communications 14 (1): 553. https://doi.org/10.1038/s41467-023-36290-w.

Mann, Gagandeep, Stephen Mora, Glory Madu, and Olasunkanmi A. J. Adegoke. 2021. “Branched-Chain Amino Acids: Catabolism in Skeletal Muscle and Implications for Muscle and Whole-Body Metabolism.” Frontiers in Physiology 12 (July):702826. https://doi.org/10.3389/fphys.2021.702826.

Saikia, Mridusmita, Xiaoyun Wang, Yuanhui Mao, Ji Wan, Tao Pan, and Shu-Bing Qian. 2016. “Codon Optimality Controls Differential mRNA Translation during Amino Acid Starvation.” RNA (New York, N.Y.) 22 (11): 1719–27. https://doi.org/10.1261/rna.058180.116.

Sharma, Puneet, Jie Wu, Benedikt S. Nilges, and Sebastian A. Leidel. 2021. “Humans and Other Commonly Used Model Organisms Are Resistant to Cycloheximide-Mediated Biases in Ribosome Profiling Experiments.” Nature Communications 12 (1): 5094. https://doi.org/10.1038/s41467-021-25411-y.

Sheen, Joon-Ho, Roberto Zoncu, Dohoon Kim, and David M. Sabatini. 2011. “Defective Regulation of Autophagy upon Leucine Deprivation Reveals a Targetable Liability of Human Melanoma Cells In Vitro and In Vivo.” Cancer Cell 19 (5): 613–28. https://doi.org/10.1016/j.ccr.2011.03.012.

Xiao, Fei, Chunxia Wang, Hongkun Yin, Junjie Yu, Shanghai Chen, Jing Fang, and Feifan Guo. 2016. “Leucine Deprivation Inhibits Proliferation and Induces Apoptosis of Human Breast Cancer Cells via Fatty Acid Synthase.” Oncotarget 7 (39): 63679–89. https://doi.org/10.18632/oncotarget.11626.

Yu, Deyang, Nicole E. Richardson, Cara L. Green, Alexandra B. Spicer, Michaela E. Murphy, Victoria Flores, Cholsoon Jang, et al. 2021. “The Adverse Metabolic Effects of Branched-Chain Amino Acids Are Mediated by Isoleucine and Valine.” Cell Metabolism 33 (5): 905-922.e6. https://doi.org/10.1016/j.cmet.2021.03.025.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary

Worpenberg and colleagues investigated the translational consequences of branched-chain amino acid (BCAA) starvation in mouse cells. Limitation of individual BCAAs has been reported to cause codon-specific and global translational repression. In this paper, the authors use RNA-seq, ribosome profiling (Ribo-seq), proteomics, and tRNA charging assays to characterize the impacts of individual and combined depletion of leucine, isoleucine, and valine on translation. They find that BCAA starvation increases codon-specific ribosome dwell times, activates global translational stress responses and reduces global protein synthesis. They infer that this effect is due to decreased translation initiation and codon-specific translational stalling. They find that the effects of simultaneous depletion are non-additive. In valine and triple (valine, leucine, and isoleucine) depletion, they show that affected transcripts have a high density of valine codons early in their coding sequences, creating an "elongation bottleneck" that obscures the impact of starvation of other amino acids. Finally, they identify isoacceptor-specific differences in tRNA charging that help explain the codon-specific effects that they observe.

We find the major findings convincing and clear. We find that some results are incompletely explained. We suggest an additional experiment and also have some minor comments that we hope will improve clarity and rigor.

Major comments

Figure 3O: In this figure and the associated text, the authors try to determine whether differences in protein degradation can explain why some proteins have higher ribosome density but lower proteomic expression. However, since this analysis relies on published protein half-lives from non-starvation conditions and on the assumption that protein synthesis has entirely stopped, we are not convinced it is informative for this experimental context. It does not distinguish between a model in which protein synthesis has been reduced by stalling and a model in which both protein synthesis and degradation rate have increased, which are both consistent with their Ribo-seq and proteomic data. To address this issue, the authors should either perform protein half-life measurements under their starvation conditions, or more clearly explain these two models in the text and acknowledge that they cannot distinguish between them.

Minor comments

Figure 1G: Why does intracellular valine seem to be less depleted under starvation conditions than intracellular leucine or isoleucine? Are the limits of detections different for different amino acids? The authors should acknowledge this discrepancy and comment on whether it has any implications for interpretation of their results.

Figure 1H: These data do not appear to meet the assumptions for linear regression. We suggest either reporting a Spearman R correlation (as the data appears linear in rank but not absolute value), or remove it entirely - we think the plot without statistics is sufficient.

Figure 2B: The in-text description of this figure states that "most" ISR genes show a "robust induction," but only three genes are shown in the figure, two of which are upregulated. The authors should instead specify that 2 out of the 3 genes profiled were robustly induced.

Figure 2D: Please include the full, uncropped blots in the supplementary materials.

Figure 2E: Swap the positions of the RPS6 and 4E-BP1 plots so they line up with their respective blots to make these figures easier to interpret. Authors should consider doing a one-way ANOVA and post-hoc analysis, if we correctly understand that they are making a conclusion about the difference between multiple groups in aggregate.

Figure 4B: Panel A in this figure is very convincing, and these plots don't add additional information. The authors could consider removing them. If this panel stays in, we suggest removing the "mid index" plot, since it is never referenced in the text and doesn't seem relevant to the message of the figure.

Figure 4E: Why is there a reduction in frequency of a Leu and a Val codon under Ile starvation?

Figure 5G: There appears to be one Val codon early in the Hint1 transcript without much stalling under triple or valine starvation conditions. The authors should acknowledge this and comment on why this may be.

Figure 5B: In the text referencing this figure, the authors state that "a high number of downregulated proteins with associated ribosome stalling sites did not show an overall decreased mean RPF count...as it would be expected from translation initiation defects, linking these stalling sites directly to proteomic changes." However, RPF is affected both by stalling (increases RPF) and initiation defects (decreases RPF). A gene with both stalling and decreased initiation may appear to have no RPF change. The data does suggest a contribution from stalling, but the authors should also acknowledge that reduced initiation may also be playing a role.

Figure 5E: the black text on dark brown in the center of the Venn diagram is difficult to read. The diagram should either have a different color scheme, or the text in the center should be white instead of black for higher contrast.

Supplementary Figure 1C: The ribosome dwell time data in this study is described as "highly correlated" with another published dwell time dataset, but the P and E site data do not seem strongly correlated. The authors should remove the word "highly."

Supplementary Figure 3E: Not all of the highlighted codons in this figure are ones with prolonged dwell times. To clarify the point that dwell time change is not related to codon frequency, this figure should only highlight codons that have a significantly prolonged dwell time in at least one starvation condition.

Supplementary Figure 5C: The gene Chop is mentioned in the main text when referencing this figure, but is absent from the heatmap.

Supplementary Figure 7A: The authors could clarify this figure by adding additional language to either the figure panel or the figure legend specifying that the RPM metric being used comes from Ribo-seq.

Supplementary Figure 7D: The metric used to describe the spatial relationship between the first valine and isoleucine codons in transcripts in this figure seems to be describing something conceptually similar to the stalling sites in Figure 5G, but uses a different metric. These figures would be easier to interpret if these spatial relationships were presented in a consistent way throughout the manuscript.

Discussion:

Reader understanding would be improved if the relevance of paragraphs were established in the first sentence. For instance, in the paragraphs about adaptive misacylation and posttranscriptional modifications, it is unclear until the end of the paragraph how these topics are relevant. Introducing the relevant aspects of the study (the fact that some starvation conditions have less severe effects and the observation about m6A-related mRNAs) at the beginning of these paragraphs would improve clarity.<br /> The authors should provide more information and speculation about possible physiological relevance of their findings, particularly about the way that the effects of triple starvation are highly valine-dependent. Are there physiological conditions under which starvation of all three BCAAs is more likely than starvation of one or two of them? If so, are there any reasons why a valine-based bottleneck might be advantageous?

We hope the authors have found our comments thoughtful and useful. We welcome further discussion or clarification via email: Juliana Stanley (julianst@mit.edu) and Hannah LeBlanc (leblanch@mit.edu).

Significance

Nature and significance of the advance

The main contribution of this work is to demonstrate that depletion of multiple amino acids simultaneously impacts translation elongation in ways that are not necessarily additive. These impacts can depend on the distribution of codons in a transcript. It adds to a growing body of work showing that essential amino acid starvation can cause codon-specific ribosome stalling. The authors suggest that the position-dependent stalling they observe could be a novel regulatory mechanism to alleviate the effects of multi-amino acid starvation. However, it is not fully clear from the paper what the significance of a valine-based regulatory adaptation to BCAA starvation is, or whether simultaneous starvation of all three BCAAs is of particular physiological relevance. The paper's primary contribution is mainly focused on the similarity between valine and triple BCAA starvation, and it provides limited insight into the effects of combined depletion of two BCAAs.

Context of existing literature

Although ribosome profiling does not distinguish between actively-elongating and stalled ribosomes, sites with higher read coverage, and thereby higher inferred dwell time, can be used to infer ribosome stalling (Ingolia 2011). Various downstream effects of essential amino acid depletion have been documented, such as leucine deficiency being sensed by mTORC1 via leucyl-tRNA synthetase (Dittmar 2005, Han 2012), and shared transcriptional responses among many amino acid depletion conditions (Tang 2015). These authors have previously measured the translational effects of nutrient stress using ribosome profiling (e.g., Gobet 2020), as have others (Darnell 2018, Kochavi et al. 2024). The present work represents the first study (to our knowledge) combining BCAA depletions, representing an incremental and useful contribution to our understanding of translational responses to stress conditions.

Audience

This work is of interest to investigators studying the response of human cells in stress conditions, such as in human disease, as well as investigators studying the basic biology of eukaryotic translational control.

Reviewer expertise: mRNA decay and translation regulation in bacteria.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary and General Critique:

This study examines the consequences of starvation for the BRCAAs, either singly, for Leu & Ile, or for all three simultaneously in HeLa cells on overall translation rates, decoding rates at each codon, and on ribosome density, protein expression, and distribution of ribosome stalling events across the CDS for each expressed gene. The single amino acid starvation regimes specifically reduce the cognate intracellular amino acid pool and lead to deacylation of at least a subset of the cognate tRNAs in a manner dependent on continuing protein synthesis. They also induce the ISR equally and decrease bulk protein synthesis equally in a manner that appears to occur largely at the initiation level for -Leu and -Val, judging by the decreased polysome:monsome ratio, but at both the initiation and elongation levels for -Ile-a distinction that remains unexplained. Only -Leu appears to down-regulate mTORC1 and TOP mRNA translation. There is a significant down-regulation of protein levels for 50-200 genes, which tend to be unstable in nutrient-replete cells, only a fraction of which are associated with reduced ribosome occupancies (RPFs measured by Ribo-Seq) on the corresponding mRNAs in the manner expected for reduced initiation, suggesting that delayed elongation is responsible for reduced protein levels for the remaining fraction of genes. All three single starvations lead to increased decoding times for a subset of the cognate "hungry" codons: CUU for -Leu, AUU and AUC for -Ile, and all of the Val codons, in a manner that is said to correspond largely to the particular tRNA isoacceptors that become deacylated, although this correspondence was not explained explicitly and might not be as simple as claimed. All three single starvations also evoke skewing of RPFs towards the 5' ends of many CDSs in a manner correlated with an enrichment within the early regions of the CDSs for one or more of the cognate codons that showed increased decoding times for -Ile (AUC codon) and -Val (GUU, GUC, and GUG), but not for -Leu-of which the latter was not accounted for. These last findings suggest that, at least for -Val and -Ile, delays in decoding N-terminal cognate codons cause elongating ribosomes to build-up early in the CDS. They go on to employ a peak calling algorithm to identify stalling sites in an unbiased way within the CDS, which are greatest in number for -Val, and find that Val codons are enriched in the A-sites (slightly) and adjacent 5' nucleotides (to a greater extent) for -Val starvation; and similarly for Ile codons in -Ile conditions, but not for -Leu starvation-again for unknown reasons. It's unclear why their called stalling sites have various other non-hungry codons present in the A sites with the cognate hungry codons being enriched further upstream, given that stalling should occur with the "hungry" cognate codon in the A site. The proteins showing down-regulation are enriched for stalling sites only in the case of the -Val starvation in the manner expected if stalling is contributing to reduced translation of the corresponding mRNA. It's unclear why this enrichment apparently does not extend to -Ile starvation which shows comparable skewing of RPFs towards the 5'ends, and this fact diminishes the claim that pausing generally contributes to reduced translation for genes with abundant hungry codons.<br /> All of the same analyses were carried out for the Double -Ile/-Leu and Triple starvations and yield unexpected results, particularly for the triple starvation wherein decoding times are increased only at Val codons, skewing of RPFs towards the 5' ends of CDSs is correlated only with an enrichment for Val codons within the early regions of the CDSs, and stall sites are enriched only for Val codons at nearly upstream sites, all consistent with the finding that only Val tRNAs become deacylated in the Triple regime. To explain why only Val tRNA charging is reduced despite the observed effective starvation for all three amino acids, they note first that stalling at Val codons is skewed towards the 5'ends of CDS for both -Val and triple starvations more so than observed for Ile or -Leu starvation, which they attribute to a greater frequency of Val codons vs Ile codons in the 5' ends of CDSs. As such, charged Val tRNAs are said to be consumed in translating the 5'ends of CDSs and the resulting stalling prevents ribosomes from reaching downstream Ile and Leu codons at the same frequencies and thus prevents deacylation of the cognate Ile and Leu tRNAs. It's unclear whether this explanation is adequate to explain the complete lack of Ile or Leu tRNA deacylation observed even when amino acid recycling by the proteasome is inhibited-a treatment shown to exacerbate deacylation of cognate tRNAs in the single amino acid starvations and of Val tRNA in the triple starvation. As such, the statement in the Abstract "Notably, we could show that isoleucine starvation-specific stalling largely diminished under triple starvation, likely due to early elongation bottlenecks at valine codons" might be too strong and the word "possibly" would be preferred over "likely". It's also unclear why the proteins that are down-regulated in the triple starvation are not significantly enriched for stalling sites (Fig. 5B) given that the degree of skewing is comparable or greater than for -Val. This last point seems to undermine their conclusion in the Abstract that "that many proteins downregulated under BCAA deprivation harbor stalling sites, suggesting that compromised elongation contributes to decreased protein output."<br /> In the case of the double -Ile/-Leu starvation, a related phenomenon occurs wherein decoding rates are decreased for only the AUU Ile codon and only the AAU Ile tRNA becomes deacylated; although in this case increased RPFs in the 5' ends are not correlated with enrichment for Ile or Leu codons and, although not presented, apparently stall sites are not associated with the Ile codon in the double starvation. In addition, stalling sites are not enriched in the proteins down-regulated by the double starvation. Moreover, because Ile codons are not enriched in the 5'ends of CDS, it doesn't seem possible to explain the selective deacylation of the single Ile tRNA observed in the double starvation by the same "bottleneck" mechanism proposed to explain selective deacylation of only Val tRNAs during the triple starvation. This is another reason for questioning their "bottleneck" mechanism.

Specific comments (some of which were mentioned above):

• The authors have treated cells with CHX in the Ribo-Seq experiments, which has been shown to cause artifacts in determining the locations of ribosome stalling in vivo owing to continued elongation in the presence of CHX (https://doi.org/10.1371/journal.pgen.1005732 ). The authors should comment on whether this artifact could be influencing some of their findings, particular the results in Fig. 5C where the hungry codons are often present upstream of the A sites of called stalling sites in the manner expected if elongation continued slowly following stalling in the presence of CHX.
• p. 12: "These starvation-specific DT and ribosome density modulations were also evident at the individual transcript level, as exemplified by Col1a1, Col1a2, Aars, and Mki67 which showed persistent Val-codon-specific ribosome density increases but lost Ile-codon-specific increases under triple starvation (Supplementary Figure 3A-D). " This conclusion is hard to visualize for any but Val codons. It would help to annotate the relevant peaks of interest for -Ile starvation with arrows.
• To better make the point that codon-specific stalling under BCAA starvation appears to be not driven by codon usage, rather than the analysis in Fig. 1H, wouldn't it be better to examine the correlation between increases in DT under the single amino acid starvation conditions and the codon frequencies across all codons?
• p. 13, entire paragraph beginning with "Our RNA-seq and Ribo-seq revealed a general activation of stress response pathways across all starvations..." It is difficult to glean any important conclusions from this lengthy analysis, and the results do not appear to be connected to the overall topic of the study. If there are important conclusions here that relate to the major findings then these connections should be made or noted later in the Discussion. If not, perhaps the analysis should be largely relegated to the Supplemental material.
• p. 15: "Together, these findings highlight that BCAA starvation triggers a combination of effects on initiation and elongation, with varying dynamics by amino acid starvation." I take issue with this statement as it appears that translation is reduced primarily at the initiation step for all conditions except -Ile. As noted above, these data are never menitioned in the DISCUSSION as to why only -Ile would show a marked elongation component to the inhibition whereas -Val gives the greatest amount of ribosome stalling.
• I cannot decipher Fig. 4D and more detail is required to indicate the identify of each column of data.
• In Fig. 4E, one cannot determine what the P values actually are, which should be provided in the legend to confirm statistical significance.
• It's difficult to understand how the -Leu condition and the Double starvation can produce polarized RPFs (Fig. 4A) without evidence of stalling at the cognate hungry codons (Fig. 4E), despite showing later in Fig. 5A that the numbers of stall sites are comparable in those cases to that found for -Ile.
• Fig. 5B: the P values should be given for all five columns, and it should be explained here or in the Discussion why the authors conclude that stalling is an important determinant for reduced translation when a significant correlation seems to exist only for the -Val condition and not even for the Triple condition.
• p. 17: "Of note, in cases where valine or isoleucine codons were present just upstream (rather than at) the stalling position, we noted a strong bias for GAG (E), GAA (E), GAU (D), GAC (D), AAG (K), CAG (Q), GUG (V) and GGA (G) (Val starvation) and AAC (N), GAC (D), CUG (L), GAG (E), GCC (A), CAG (Q), GAA (E) and AAG (K) (Ile starvation) at the stalling site (Supplementary Figure 7B)." The authors fail to explain why these codons would be present in the A sites at stalling sites rather than the hungry codons themselves, especially since it is the decoding times of the hungry codons that are increased according to Fig. 1A-E. As suggested above, is this a CHX artifact?
• Fig. 5D: P values for the significance, or lack thereof, of the different overlaps should be provided.
• p. 17: "Nonetheless, when we examined entire transcripts rather than single positions, many transcripts that exhibited isoleucine-related stalling under Ile starvation also stalled under triple starvation, but at different sites along the CDS (Figure 5E). This finding is particularly intriguing, as it suggests that while Ile-starvation-specific stalling sites may shift under triple starvation, the overall tendency of these transcripts to stall remains." The authors never come back to account for this unexpected result.
• It seems very difficult to reconcile the results in Fig. 5F with those in Fig. 4A, where similar polarities in RPFs are observed for -Ile and -Val in Fig, 4A but dramatically different distributions of stalling sites in Fig. 5F. More discussion of these discrepancies is required.
• p. 18: " These isoacceptor-specific patterns correlate largely with the particular subsets of leucine and isoleucine codons that stalled (Figure 1A)." This correlation needs to be addressed for each codon-anticodon pair for all of the codons showing stalling in Fig. 1A.
• p. 19: "For instance, in our double starvation condition, unchanged tRNA charging levels (Figure 6E) may result from a pronounced downregulation of global translation initiation, likely driven by the activation of stress responses (Figure 2), subsequently lowering the demand for charged tRNAs as it has been observed previously for Leu starvation 39. This seems at odds with the comparable down-regulation of protein synthesis for the Double starvation and -Leu and -Ile single starvations shown in Fig. 3C. Also, in the current study, Leu starvation does lower charging of certain Leu tRNAs.

Significance

The results here are significant in showing that starvation for a single amino acid does not lead to deacylation of all isoacceptors for that amino acid and in revealing that starvation for one amino acid can prevent deacylation of tRNAs for other amino acids, as shown most dramatically for the selective deacylation of only Val tRNAs in the triple BRCAA starvation condition. For the various reasons indicated above, however, I'm not convinced that their "bottleneck" mechanism is adequate to explain this phenomenon, especially in the case of the selective deacylation of Ile vs Leu tRNA in the Double starvation regime. It's also significant that deacylation leads to ribosome build-up near the 5'ends of CDS, which seems to be associated with an enrichment for the hungry codons in the case of Val and Ile starvation, but inexplicably, not for Leu or the Double starvations. This last discrepancy makes it hard to understand how the -Leu and Double starvations produce RPF buildups near the 5 ends of CDSs. In addition, the claim in the Discussion that "our data also highlight the importance of the codon positional context within mRNAs, indicating that where a codon is located within the CDS can influence both the extent of ribosomal stalling and overall translation efficiency during nutrient stress" overstates the strength of evidence that the stalling events lead to substantial decreases in translational efficiencies for the affected mRNAs, as the stalling frequency and decreased protein output are significantly correlated only for the -Val starvation, and the data in Fig. 3 D-H suggest that the reductions in protein synthesis generally occur at the level of initiation, even for -Val starvation, with a contribution from slow elongation only for -Ile-which is in itself difficult to understand considering that stalling frequencies are highest in -Val. Thus, while many of the results are very intriguing and will be of considerable interest to the translation field, it is my opinion that a number of results have been overinterpreted and that important inconsistencies and complexities have been overlooked in concluding that a significant component of the translational inhibition arises from the increased decoding times at hungry codons during elongation and that the selective deacylation of Val tRNAs in the Triple starvation can be explained by the "bottleneck" mechanism. The complexities and limitations of the data and their intepretations should be discussed much more thoroughly in the Discussion, which currently is devoted mostly to other phenomena often of tangential importance to the current findings. A suitably revised manuscript would clearly state the limitations and caveats of the proposed mechanisms and consider other possible explanations as well.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

This manuscript described the translational responses to single and combined BCAA shortages in mouse cell lines. Using Ribo-seq and RNA-seq analysis, the authors found selective ribosome pausing at codons that encode the depleted amino acids, where the pausing at valine codons was prominent at both a single and triple starvations whereas isoleucine codons showed pausing only under a single depletion. They analyzed the mechanisms of the unexpected selective pausing and proposed that the positional codon usage bias could shape the ribosome stalling and tRNA charging patterns across different amino acids. They also examined the stress responses and the changes in the protein expression levels under BCAA starvation.

The manuscript was well-written, and the findings are interesting, especially their model that positional codon usage bias could be a regulator of ribosome pausing and tRNA charging levels. Although different translational responses to distinct amino acid starvation have been widely documented, the positional codon usage bias is an interesting aspect. The manuscript's central message could have been made clearer. The authors may consider emphasizing this point more explicitly in the abstract. The rich multi-omics dataset in this work provides valuable resources for the translation field.

Major comments

1. The abstract may need to be revised since it is hard to immediately catch the authors' main point. If the authors regard this work as a resource paper, the current version is fine. But it could be better to point out the positional codon usages the authors found, which is a strong point of the current manuscript.
2. Page 18 "Beyond these tRNA dynamics, our data also highlight the importance of the codon positional context within mRNAs, indicating that where a codon is located within the CDS can influence both the extent of ribosomal stalling and overall translation efficiency during nutrient stress."<br /> This idea is interesting. To what extent the authors think this could be generalized? The authors may discuss whether they think their proposed model is specific to the different ribosome stalling patterns between valine and isoleucine codons or generalized to other codon combinations. For example, the positional codon usage bias will be different among different organisms, and are there any previous reports on ribosome behaviors that align with their model? Even if the authors think this model can be applied to BCAA starvation, would it be possible to explain the different isoleucine codon responses between single and double starvation? The authors may discuss why the ribosome stalling at isoleucine AUU and AUC codons was slightly attenuated under double starvation. And how about the different leucine codon responses among single, double, and triple starvations, although the pausing is not as strong as isoleucine and valine codons? Experimental validation using artificial reporters carrying biased sequences may also be considered.
3. Page 13 "Moreover, we noticed that DT changes extend beyond the ribosomal A-site, including the P-site, E-site, and even further positions (Supplementary Fig. 2A), consistent with other studies on single amino acid starvation 39 (Supplementary Fig. 2B-C)." Could the widespread DT changes be due to Ribo-DT pipeline they used or difficulties in offset determination? Indeed the authors showed that this feature was found in other datasets, but it seems that the datasets were processed and analyzed in the same way as their data. The original Ribo-DT paper (Gobet and Naef, 2022, Methods) also showed some widespread DT changes even from RNA-seq. Another analysis method like the codon subsequence abundant shift as a part of diricore analysis (Loayza-Puch et al., 2016, Nature) did not show that broad changed regions. The authors are encouraged to re-analyze the data sets using different methods.
4. Page 13 "Intriguingly, only two of the three isoleucine codons (AUU and AUC) showed increased DTs upon Ile starvation (p < 0.01), while just one leucine codon (CUU) exhibited a modest but significant DT increase (p < 0.01) under Leu starvation (Figure 1A-B, Supplementary Figure 2A)." How can the authors explain the different strengths of ribosome pausing at Ile codons under Ile and double starvation? The AUA codon did not show any pausing under either of the starvation conditions. Throughout the manuscript, the authors mainly describe the difference between amino acids but it is desirable to discuss the codon-level difference as well.
5. Page 13 "We examined the effects of single amino acid starvations (-Leu, -Ile and -Val), as well as combinations, including a double starvation of leucine and isoleucine (hereafter referred to as "double") and a starvation of leucine, isoleucine, and valine ("triple"), allowing us to identify potential non-additive effects." The different double starvations, isoleucine and valine, and leucine and valiene, will further support their hypothesis on the effects of the positional codon usage bias on ribosome pausing and tRNA charging patterns. Although this could be beyond the scope of the current manuscript, the authors are encouraged to provide a rationale for the chosen combination.

Minor comments

Page 16 "these results imply that BCAA deprivation lowers protein output through multiple pathways: a combination of reduced initiation, direct elongation blocks (stalling), and possibly an increased proteolysis" This conclusion is totally right but may be too general. Could the authors summarize BCAA-specific features of the events including reduced initiation, stalling, and proteolysis that all contribute to protein outputs? This is not well discussed in the latter sections including Discussion.

Significance

The manuscript was well-written, and the findings are interesting, especially their model that positional codon usage bias could be a regulator of ribosome pausing and tRNA charging levels. Although different translational responses to distinct amino acid starvation have been widely documented, the positional codon usage bias is an interesting aspect. The manuscript's central message could have been made clearer. The authors may consider emphasizing this point more explicitly in the abstract. The rich multi-omics dataset in this work provides valuable resources for the translation field.

Typo : c'est "la" propriété

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

Learn more at Review Commons

Reply to the reviewers

Prior to the point-by-point response to the reviewer, we would like to sincerely thank all the peer reviewers for their overwhelmingly positive comments and helpful suggestions. The recommendations have undoubtedly improved our initial submission, and we have done our best to incorporate as many of the suggestions as possible.

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

*Jones et al. have submitted a manuscript detailing the role of Coenzyme A in the regulation of macrophage polarization. Overall, the manuscript is well designed, and the conclusions are well supported by the data. I find no major or minor deficiencies that need to be corrected. *

* Reviewer #1 (Significance (Required)): *

For decades the immunology community has boldly stated that mitochondrial metabolism not only provides the bioenergetics for cell expansion but also dictates cell fate. This has been especially true for fatty acid beta oxidation. Macrophage, T-cell and B-cell polarization have all been shown to require FAO for their polarization, but all based on one inhibitor. NONE of these observations hold up with more rigorous experimentation. The Divakaruni group has previously suggested that intracellular CoA homeostasis was the driver of macrophage differentiation as they could reverse the inhibitory effects by providing heroic levels of CoA extracellularly. Here, they have clarified the role of CoA. Intracellular CoA does not affect macrophage polarization/differentiation. This was done with cleaver manipulation of the CoA pools. Rather, extracellular CoA can act as a weak TLR4 ligand. This work nicely clarifies their previous work and further demonstrates a role for this metabolite as an endogenous activator of type 1 macrophages.

We are thrilled by the positive comments about our work, and we are grateful the reviewer found our submission to be clarifying for the field and significant in the larger context of immunometabolism research.

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

*This is a fairly straightforward manuscript that indicated CoA acts as a "weak" TLR4 agonist and primes macrophages for alternative activation. Overall, the experiments are well done and clear enough. There are two major issues that need to be addressed: *

We thank the reviewer for their positive comments regarding the quality and clarity of our work.

1. *Previous work has shown the following pathway: LPS>IL10>STAT3>IL4Ra>>>increased responsiveness to IL4/IL13 and increased expression of M2 associated markers (please note, this pathway does not apply to Arg1, often erroneously associated with M2 macrophages - LPS induces Arg1 far more than IL4 and this is independent of the STAT6 pathway - Dichtl et al., Science Advances and El Kasmi et al. Nature Immunology, and others). This pathway was first described in Lang et al. 2002 J. Immunol. Subsequently, other groups showed IL6 (Jens Brüning) and OSM (Carl Richards) do the same thing, which is not surprising given that they are STAT3 activators. Thus, Il4ra is a STAT3 target gene; this also makes sense in the kinetic evolution of macrophages from inflammatory to tissue reparative (if they survive). In my view, the authors have most likely found the same pathway. In Jones, expression of the IL4Ra was not quantified. Thus, the pathway described above needs to be accounted for. It may not apply here but seems the easiest explanation of the data. *

This is an excellent and important experiment suggested by the reviewer, and we address this in our revised Supplemental Figure 5. To determine whether the effect of CoA can be explained simply by a STAT3-mediated effect on the IL-4 receptor, we treated cells with the well-characterized STAT3 inhibitor Napabucasin and measured whether CoA could enhance the macrophage IL-4 response. Two results are clear from the data:

• Treatment with Napabucasin reduced the expression of IL-4-linked cell surface markers and the IL-4 target gene Ccl8. This serves as an important control consistent with the Il4ra gene being a STAT3 target that increases IL-4 responsiveness.
• Despite STAT3 inhibition and a reduced IL-4 response, CoA provision still augmented the IL-4-induced expression of Ccl8 and the percentage of CD206+/CD301+ cells, indicating a STAT3-independent mechanism. The result aligns with our ATAC-Seq data in Figure 6 that shows broad changes in chromatin accessibility that cannot be completely explained by expression-level changes in the IL-4 receptor.

*Can the authors come up with a meaningful in vivo experiment to corroborate their data. Pantothenate-deficient mice have many phenotypes (not fully explored at all - PMID 31918006, for example) and pantothenate metabolism can be manipulated in different ways. Obviously, a complex in vivo experiment is not feasible here. But this should be discussed. What happens in human macrophages, where "polarization" is a completely different beast? *

We thank the reviewer for these thoughtful comments, and address the questions regarding in vivo proof-of-concept and polarization of human macrophages separately:

• Regarding the question of whether CoA can enhance the phenotype of IL-4-activated human macrophages, this is an excellent suggestion and we have added the data as Figure 1h. Indeed, Coenzyme A dramatically amplifies expression of the human IL-4 responsive genes CCL17, TGM2, and PDCD1LG2 (similarly to mouse macrophages). The result substantially expands the significance of our work by showing the phenotype is reproducible in both mouse and human macrophages – unlike many immunometabolic phenotypes – and we thank the reviewer again for suggesting this experiment.
• With respect to an in vivo experiment to corroborate our data, we entirely agree with the reviewer regarding both the importance, but also the difficulty in interpretation, of an experiment genetically manipulating CoA synthesis in vivo. As they have suggested, we raise these issues in the discussion on Lines 370-377 of the revised manuscript. Here, we note the following points:
• Wherever possible/appropriate (e.g. Figures 1g, 3f&g, 5g&h), we have sought to corroborate our in vitro findings with in vivo/ex vivo proofs-of-concept.
• Studying immune phenotypes in pantothenate-deficient mice would be an exciting experiment in principle, but difficult to interpret if conducted. As noted by the reviewer in the work from Drs. Rock and Jackowski, knockout of one of four isoforms of pantothenate kinase (PANK) shows mild phenotypes consistent with compensation across isoforms for CoA provision. Global double knockout of PANK1 and PANK2, however, is postnatally lethal. Regardless, a tissue-specific double knockout in myeloid cells is unlikely to show a phenotype given our results showing that manipulating intracellular CoA levels in BMDMs does not alter the IL-4 response (Figs. 2h-j).
• Given the established role of CoA in postnatal development, it would be difficult to attribute any immunologic phenotypes in genetically modified mice to direct effects of CoA as a metabolic DAMP as opposed to indirect effects from a chronically altered immune system.

Reviewer #2 (Significance (Required)): *This is a fairly straightforward manuscript that indicated CoA acts as a "weak" TLR4 agonist and primes macrophages for alternative activation. Overall, the experiments are well done and clear enough.

*

We reiterate our gratitude for the comments on the quality and clarity of our work.

Reviewer #3 (Evidence, reproducibility and clarity (Required)):

Summary: In this manuscript on enhancement of mIL-4 polarization by exogenous CoA, the authors follow up on their previous studies that had shown a correlation between Etomoxir-driven block in mIL-4 and a reduction of intracellular CoA levels. The results obtained (lack of enhancement of IL-4-induced changes in oxidative phosphorylation and glycolysis; lack of impact of pharmacological decrease/increase of intracellular CoA levels) led them to discard their initial hypothesis. Instead, the presence of a proinflammatory gene signature in macrophages treated with IL-4+CoA triggered experiments testing the involvement of TLR-Myd88 signaling and the identification of CoA as a weak agonist for TLR4 (which is consistent with a preprint manuscript posted in 2022 by others and showing induction of proinflammatory gene express in a TLR2/4-dependent manner).

• Significance: Overall, these results are novel and interesting, although the use of yeast-derived CoA preparations raises a question about the contribution of contaminants that is only partially controlled by data obtained with a synthetic CoA. Regarding a biological role for CoA in macrophage biology in vivo, the authors propose that CoA may act as a DAMP upon release from dying/dead cells and thereby modify transcriptional polarization of m(IL-4). I have several comments related to specific experimental conditions and interpretation that should be addressed. Most importantly, the key findings of the manuscript should be demonstrated using synthetic CoA as described in comment #5. *

We are heartened that the reviewer found our initial submission to be novel and interesting, and are grateful for their suggestions to reinforce our existing data with more studies comparing yeast-derived and synthetically-derived Coenzyme A. We have done our best to address each of the individual questions below:

Major comments:

1. *Increasing/decreasing intracellular CoA levels does not alter IL-4-induced CD206 expression (Fig. 2i/j. However, the impact of CoA addition to mIL-4 is stronger for Ccl8 and Mgl2 mRNA (Fig. 1a) than for the CD206+ cell fraction (Fig. 1d). Therefore, it would be better (higher sensitivity) to include expression of these genes as readout after CPCA/PZ-2891 treatment. *

This is a helpful suggestion, and we have now conducted gene expression studies to complement our flow cytometry and mass spectrometry studies while manipulating the intracellular CoA pool. In line with our previous work, neither CPCA (which decreases intracellular free CoA) or PZ-2891 (which increases intracellular free CoA) meaningfully alter expression of IL-4-linked genes including Ccl8 or Mgl2. In fact, the only (statistically insignificant) trend refutes the hypothesis, as gene expression with CPCA leads to marginally increased gene expression. These results are now included in Supplemental Figure S2f. We thank the reviewer for this helpful suggestion, as it has strengthened our conclusion that intracellular CoA levels do not adjust the macrophage IL-4 response.

• The CoA-induced proinflammatory gene expression in Fig. 3c is relatively weak (e.g. compared to LPS). The authors use CoA throughout the manuscript at a concentration of 1 mM, and we do not know how much of it is required to cause an effect at all. Therefore, dose-response curves for the stimulation of macrophages with titrated amounts of CoA should be provided. In addition, *

We thank the reviewer for bringing up this point so we could clarify and add to our existing data. We should note that Supplemental Figures 1b&c of our previous submission (and resubmitted manuscript) detail a concentration-response curve showing that at little as 62.5 mM CoA – the lowest concentration tested – was sufficient to enhance IL-4 cell surface marker expression.

However, it is an excellent suggestion as the reviewer notes, to conduct a similar concentration-response to determine if this lines up with CoA inducing a pro-inflammatory response. The full data set is presented in the answer to reviewer question 4 (comparing CoA purchased from Sigma vs. Avanti Polar Lipids), though we now show in Supplemental Figure S3 that 62.5 mM CoA is sufficient to elicit a pro-inflammatory response. Though it is indeed a weak effect as noted by the reviewer, our data suggest that the relatively mild stimulus is crucial for the effect. Given the results with the TLR3 agonist Poly I:C (Figure 5), which engages a Type 1 interferon response, strong TLR4 agonists that engage the TRIF/Type I interferon arm of the TLR4 response are likely to blunt or block the IL-4 response.

• Related question: we are informed that the concentration of CoA in the mitochondrial matrix is 5mM, whereas cytosol contains 100µM. For CoA to act as DAMP, I would like to know the concentration of it in supernatants of cell cultures (live vs. dying/dead cells) and from tissues. *

This is an important point brought up by the reviewer, and we agree that the implicit issue raised (i.e. “do the concentrations of CoA required to see an effect reconcile with a physiological role as a DAMP?”) should be more thoroughly addressed in the manuscript. Tissue concentrations of free CoA (in ng/mg tissue) are well established for mice and range from >100 nmol/g tissue (liver, heart, brown adipose tissue) to Nonetheless, the reviewer’s larger point is very well reasoned, and we address it in the following ways in the discussion on __Lines 378-391. __

• In light of the reviewer’s comment, we now mention specific instances in the discussion where CoA acting as a DAMP may reasonably play a physiological role (e.g. acetaminophen-induced acute liver injury or other forms of sterile liver injury given that DAMPs are known to be important factors and liver tissue contains relatively high concentrations of CoA).
• Although cytoplasmic concentrations of CoA may only be 50-100 mM, our work establishes a framework for how ubiquitous metabolic co-factors can activate pattern recognition receptors. Put another way, although CoA itself may not be a physiologically relevant DAMP, discovering this pathway could inform how other nucleotide or nucleoside analogs (e.g. adenine- or adenosine-containing molecules present at millimolar concentrations) exert their effects on innate immunity.
• Our newly obtained data with HMDMs (Figure 1h) shows that the CoA response in human macrophages – boosting IL-4-linked gene expression by 10-100X – may be much stronger than the 1.5-5X effect observed in mouse BMDMs. As such, it is exciting to speculate that CoA may have a more potent effect on the IL-4 response in humans relative to mice. We trust the reviewer understands the limitations of obtaining human macrophages that preclude conducting a thorough concentration-response analysis given the restrictions of a manuscript revision.

• It is very good that the authors validate the findings obtained using the yeast-derived CoA with the synthetic molecule. It is very conceivable that the 15% contaminating substances in the yeast CoA could be causing the observed changes in m(IL-4). The fact that synthetic CoA has higher activity in proinflammatory gene expression by BMM (Suppl. Fig. S3) is reassuring, however, it raises the question why this is the case. One possibility is that the concentrations of the different CoA preps cannot directly be compared. Therefore, dose response curves should also be provided for synthetic CoA. *

This is an astute observation by the reviewer and we thank them for reading our manuscript with such detailed attention to pick this up. We are reassured that the reviewer shares our interpretation that the effect of CoA is not due to a contaminating TLR4 agonist in the yeast-derived preparation (from Sigma-Aldrich; ~85% pure) given a negative Limulus Test (Supplemental Figure S4b). Moreover, the synthetically-derived preparation (from Avanti Polar Lipids; ~99% pure) yields a stronger TLR4 response.

An exploration of the follow-on question regarding why the effect is greater than 15% is presented below. These experiments have been added to Supplemental Figure S4c&d. The summary of our data suggests the individual concentrations indeed cannot be compared – matched concentrations of synthetic Avanti CoA have greater than a 15% effect than yeast-derived Sigma CoA. There are likely multiple factors that could explain this, some of which are listed below.

• The physiological effect of a TLR agonist need not be linear with its concentration, as demonstrated by the sigmoidal calibration curves for the TLR-expressing HEK-blue cells (Figures 4b, S4a). This likely does not explain the dramatic difference between the two CoA preparations but is worth noting.
• While we have determined that the 15% contaminating substances in the yeast-derived CoA are not causing the observed changes in the IL-4 response, it is formally possible that there are contaminating substances blunting the pro-inflammatory response and therefore limiting the effect of CoA purchased from Sigma-Aldrich relative to that from Avanti Polar Lipids. Importantly, however, our data in response to Reviewer Question #5 show there is no difference in amplifying the IL-4 response between the yeast- and synthetically-derived CoA.

• The difference in activity of yeast and synth. CoA could also be caused by the additional biologically active molecules in the yeast CoA. Therefore, it is important to show that the key findings in the paper (enhancement of m(IL-4) associated gene expression and CD206+ upregulation in vitro and in vivo) are also induced by synth. CoA. This is even more important in the context of the Myd88-independence of CD206+ upregulation in BMM treated with CoA (Suppl. Fig. S4). The experiment should be repeated with synth. CoA. If the enhancement of CD206+ cells induced by CoA is indeed unchanged in Myd88 KO BMM, then the title of the manuscript "CoA enhances alternative macrophage activation via Myd88" would not be supported by the data and needed to be changed. Activation of the TLR4 reporter cell line should also be tested using the synth. CoA molecule.*

We are grateful for this suggestion by the reviewer to further cement the idea that our observation of CoA enhancing the macrophage IL-4 response was not due to a contaminant in the Sigma-Aldrich CoA preparation. The reviewer makes a few points in this question which we address individually here.

• The suggestion to confirm that the CoA-induced enhancement of M(IL-4) is not due to a contaminating substance in the Sigma-Aldrich CoA is excellent and necessary. Here we show that synthetically derived CoA (99% pure, purchased from Avanti Polar Lipids) quantitatively reproduces the effect from yeast-derived CoA from Sigma-Aldrich in Supplemental Figure S4e. The response is noteworthy because synthetic CoA has profoundly stronger pro-inflammatory response than yeast-derived CoA, yet both have a similar effect on augmenting M(IL-4). This suggests that any appropriate pro-inflammatory response – irrespective of the relative strength or weakness – is sufficient to maximize the effect. This can also be observed with the range of MyD88-linked TLR agonists used in Figures 5 and S6a.
• Similarly, we also conducted experiments to show that the effect of synthetic CoA on M(IL-4) is independent of MyD88 similarly to yeast-derived CoA. These data are present in Supplemental Figure S6b&c. Here again, we should note that the effect of synthetic CoA is quantitatively similar to the effect of yeast CoA and Imiquimod (Supplemental Figure S6a).
• Activation of the TLR4 reporter cell line is available in Supplemental Figure S4c.
• Regarding the title of the manuscript, we acknowledge that we struggled a bit with how to frame our findings. Importantly, our findings support a model where (i) CoA provision enhances the IL-4 response not via metabolic changes but rather by acting as a mild pro-inflammatory stimulus, and (ii) MyD88 signaling augments the IL-4 response. We should also note that our findings simply show that CoA does not exclusively enhance the IL-4 response via MyD88 signaling, and there may be other redundant pathways (similarly to MyD88 agonist imiquimod but unlike the MyD88 agonists Pam3-CSK4 and low concentrations of LPS). We are open to working journal editors to strike the right balance of scientific accuracy and representation of the work when deciding on a final title.

• The results from the tumor model in Fig. 5 are presented to show a stronger tumor-promoting effect of m(IL-4) stimulated with Pam3. However, the variability of the data is high and 2 out of 6 mice in the +Pam3 group appear to actually have a lower tumor weight than the control mice. Therefore, these data are quite superficial and preliminary, and would benefit from a replicate experiment. Furthermore, for the evaluation of CoA as a biologically relevant DAMP, it would be important to know whether CoA-treated m(IL-4) show the same tumor-promoting effect in vivo as Pam3. *

We thank the reviewer for their comment, and agree that our in vivo work is indeed preliminary. Our goal with this report was to focus on the initial discovery of this molecular pathway and its first, broad characterization using a range of techniques (e.g. in vivo outcomes, ATAC-Seq, etc.), many of which can spur more detailed follow-up studies for future papers. As detailed in the manuscript discussion (Lines 415-419), future work beyond our initial discovery is warranted to thoroughly explore the physiological outcomes of CoA as a metabolic DAMP in relevant model systems such as acute liver injury. As an initial proof-of-concept to show that MyD88 signaling can enhance alternative activation, however, we believe our two discrete experiments (sterile inflammation and tumor formation) are sufficient to indicate the phenotype is likely relevant in animal models. In vivo syngeneic tumor models display natural variability in tumor size due to differences in implantation efficiency, host immune responses, and tumor-intrinsic growth kinetics. Nonetheless, our statistical analysis demonstrates that, with high confidence, that the observed differences are reproducible and not attributable to random variation.

Minor comments:

1. • Fig. 1b: where the gates for CD206/CD301 set based on isotype control stainings? *

We thank the reviewer for pointing out this oversight in our methods. The gates were indeed set on isotype control stains, and this is now mentioned in Lines 519-521 of the revised manuscript.

The formatting not cohesive m(IL-4) vs. M(IL-4)

Again, this is an embarrassing oversight on our part and we have done our very best to copy edit the piece and remove any inconsistencies and errors.

*Methods: primer sequences are not shown. They should be provided. *

We thank the reviewer for pointing this out, and now include all primer sequences used in Supplemental Table 1 of the revised manuscript.

Description of flowcytometry (L/D staining after surface? No washing steps after addition of L/D staining)

We thank the reviewer for pointing out another oversight in our methods, and have provided a more detailed description of the flow cytometric analysis in Lines 509-521 of the revised manuscript.

Statistics: the methods section states that variability is indicated by SD, but the Figure legends always mention SEM. Please correct.

We are grateful for the reviewer’s helpful attention to detail, and have corrected the methods to line up with the figure legends.

*A multitude of typos and editorial inconsistencies (e.g. spelling of m(IL-4), punctation and capitalization) should be corrected/streamlined. *

We are grateful for the reviewer’s helpful attention to detail, and have done our best to copy edit the manuscript prior to resubmission.

Reviewer #3 (Significance (Required)):

strengths: I like that the authors follow up their previous work on Etomoxir and CoA, now finding again an unexpected twist in how the effect on m(IL-4) is brought about. This makes the story more complicated, but is important to get to a more precise and realistic understanding of metabolic and transcriptomic regulation and how they are interconnected (or not). In addition, the use of a relatively broad set of methods including ATACseq and mass spectrometry is a strength.

weakness: the use of the not very pure yeast derived CoA prep, which is controlled for induction of proinflammatory cytokines by one experiment with synth. CoA. This validation needs to be expanded (see comments above) to substantiate the main message of the manuscript.

The scope of the manuscript is quite focussed on the mechanism of CoA enhanced m(IL-4). The finding that CoA appears not to act by changing intracellular macrophage metabolism but instead after its release by activation TLR4 widens the scope and suggests a new function for CoA as DAMP. This aspect would need to be further substantiated to be convincing.

Audience: scientists working at the intersection between metabolism and innate immunity will be interested in the results.

We thank the reviewer for their kind comments regarding the precision, credibility, and breadth of our manuscript. We hope they find our revised manuscript an improvement over our previous submission regarding both the new experiments and modified text. The comments have undoubtedly improved our manuscript and we are grateful to the reviewer for the considerable effort they put into reading our submission.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary:

In this manuscript on enhancement of mIL-4 polarization by exogenous CoA, the authors follow up on their previous studies that had shown a correlation between Etomoxir-driven block in mIL-4 and a reduction of intracellular CoA levels. The results obtained (lack of enhancement of IL-4-induced changes in oxidative phosphorylation and glycolysis; lack of impact of pharmacological decrease/increase of intracellular CoA levels) led them to discard their initial hypothesis. Instead, the presence of a proinflammatory gene signature in macrophages treated with IL-4+CoA triggered experiments testing the involvement of TLR-Myd88 signaling and the identification of CoA as a weak agonist for TLR4 (which is consistent with a preprint manuscript posted in 2022 by others and showing induction of proinflammatory gene express in a TLR2/4-dependent manner).

Significance:

Overall, these results are novel and interesting, although the use of yeast-derived CoA preparations raises a question about the contribution of contaminants that is only partially controlled by data obtained with a synthetic CoA. Regarding a biological role for CoA in macrophage biology in vivo, the authors propose that CoA may act as a DAMP upon release from dying/dead cells and thereby modify transcriptional polarization of m(IL-4). I have several comments related to specific experimental conditions and interpretation that should be addressed. Most importantly, the key findings of the manuscript should be demonstrated using synthetic CoA as described in comment #5.

Major comments:

1. Increasing/decreasing intracellular CoA levels does not alter IL-4-induced CD206 expression (Fig. 2i/j. However, the impact of CoA addition to mIL-4 is stronger for Ccl8 and Mgl2 mRNA (Fig. 1a) than for the CD206+ cell fraction (Fig. 1d). Therefore, it would be better (higher sensitivity) to include expression of these genes as readout after CPCA/PZ-2891 treatment.
2. The CoA-induced proinflammatory gene expression in Fig. 3c is relatively weak (e.g. compared to LPS). The authors use CoA throughout the manuscript at a concentration of 1 mM, and we do not know how much of it is required to cause an effect at all. Therefore, dose-response curves for the stimulation of macrophages with titrated amounts of CoA should be provided. In addition,
3. Related question: we are informed that the concentration of CoA in the mitochondrial matrix is 5mM, whereas cytosol contains 100µM. For CoA to act as DAMP, I would like to know the concentration of it in supernatants of cell cultures (live vs. dying/dead cells) and from tissues.
4. It is very good that the authors validate the findings obtained using the yeast-derived CoA with the synthetic molecule. It is very conceivable that the 15% contaminating substances in the yeast CoA could be causing the observed changes in m(IL-4). The fact that synthetic CoA has higher activity in proinflammatory gene expression by BMM (Suppl. Fig. S3) is reassuring, however, it raises the question why this is the case. One possibility is that the concentrations of the different CoA preps cannot directly be compared. Therefore, dose response curves should also be provided for synthetic CoA.
5. The difference in activity of yeast and synth. CoA could also be caused by the additional biologically active molecules in the yeast CoA. Therefore, it is important to show that the key findings in the paper (enhancement of m(IL-4) associated gene expression and CD206+ upregulation in vitro and in vivo) are also induced by synth. CoA. This is even more important in the context of the Myd88-independence of CD206+ upregulation in BMM treated with CoA (Suppl. Fig. S4). The experiment should be repeated with synth. CoA. If the enhancement of CD206+ cells induced by CoA is indeed unchanged in Myd88 KO BMM, then the title of the manuscript "CoA enhances alternative macrophage activation via Myd88" would not be supported by the data and needed to be changed. Activation of the TLR4 reporter cell line should also be tested using the synth. CoA molecule.
6. The results from the tumor model in Fig. 5 are presented to show a stronger tumor-promoting effect of m(IL-4) stimulated with Pam3. However, the variability of the data is high and 2 out of 6 mice in the +Pam3 group appear to actually have a lower tumor weight than the control mice. Therefore, these data are quite superficial and preliminary, and would benefit from a replicate experiment. Furthermore, for the evaluation of CoA as a biologically relevant DAMP, it would be important to know whether CoA-treated m(IL-4) show the same tumor-promoting effect in vivo as Pam3.

Minor comments:

1. Fig. 1b: where the gates for CD206/CD301 set based on isotype control stainings?
2. The formatting not cohesive m(IL-4) vs. M(IL-4)
3. Methods: primer sequences are not shown. They should be provided.
4. Description of flowcytometry (L/D staining after surface? No washing steps after addition of L/D staining)
5. Statistics: the methods section states that variability is indicated by SD, but the Figure legends always mention SEM. Please correct.
6. A multitude of typos and editorial inconsistencies (e.g. spelling of m(IL-4), punctation and capitalization) should be corrected/streamlined.

Significance

Strengths: I like that the authors follow up their previous work on Etomoxir and CoA, now finding again an unexpected twist in how the effect on m(IL-4) is brought about. This makes the story more complicated, but is important to get to a more precise and realistic understanding of metabolic and transcriptomic regulation and how they are interconnected (or not). In addition, the use of a relatively broad set of methods including ATACseq and mass spectrometry is a strength.

Weakness: the use of the not very pure yeast derived CoA prep, which is controlled for induction of proinflammatory cytokines by one experiment with synth. CoA. This validation needs to be expanded (see comments above) to substantiate the main message of the manuscript.

The scope of the manuscript is quite focussed on the mechanism of CoA enhanced m(IL-4). The finding that CoA appears not to act by changing intracellular macrophage metabolism but instead after its release by activation TLR4 widens the scope and suggests a new function for CoA as DAMP. This aspect would need to be further substantiated to be convincing.

Audience: scientists working at the intersection between metabolism and innate immunity will be interested in the results.

“¿Educamos para pensar o para obedecer?” En muchas aulas, el silencio es confundido con disciplina, y la repetición con aprendizaje. Se espera que los estudiantes memoricen, repitan y aprueben, pero no siempre que cuestionen. ¿Qué ocurre cuando un estudiante levanta la mano para disentir? ¿Lo celebramos como pensamiento crítico o lo corregimos como falta de respeto? Educar no debería ser domesticar. El pensamiento crítico nace cuando se permite la duda, cuando se enseña a leer entre líneas, a identificar intenciones, a reconocer que todo texto tiene un contexto y todo autor una postura. Si queremos formar ciudadanos libres, debemos enseñarles a pensar, no solo a responder.

Reviewer #1 (Public review):

This work provides a valuable toolkit for endogenous isolation of projection neuron subtypes. With further validation, it could present a solid method for low-input ribosome affinity purification using a ribosomal RNA (rRNA) antibody. The experimental evidence for the distinct ribosomal complexes is limited to this method and indirect support from complementary analyses of pre-existing data. However, with additional experimental data to support the specificity of ribosomal complex pulldown and confirmation of the putative ribosomal complex proteins of interest, the study would provide compelling evidence for translation regulation of neuronal development through compositional ribosome heterogeneity. This work would be of interest to neuroscientists, developmental biologists, and those studying translational networks underlying gene regulation.

Strengths

(1) This in vivo labeling of specific projection neurons and ribosomal rRNA affinity purification method accommodates a low input of <100K somata per replicate, which is useful for the study of neuronal subtypes with limited input. In principle, this set of techniques could work across different cell types with limited input, depending on the molecule used for cell type labeling.

(2) The authors are also able to isolate endogenous neurons with minimal perturbation up to the point of collection, preserving the native state for the neuron in vivo as long as possible prior to processing.

(3) This study identified over a dozen potential non-ribosomal proteins associated with SCPN ribosomal complexes, as well as a ribosomal protein enriched in CPN.

Limitations

(1) In this study, the authors address the advantages of their ribosomal complex isolation method in SCPN and CPN against RPL22-HA affinity purification. While this does show more pull-down of the ribosomal RNA by the Y10B rRNA antibody, the authors claim this method identifies cell-type-specific ribosomal complex proteins without demonstrating a positive control for the method's specificity. There are very limited experiments to truly delineate how "specific" this method is working and whether there could be contamination from other complexes bound by the antibody. I see this as the major limitation that should be addressed. To boost their claims of capturing cell-type-specific ribosomal complexes, the authors could consider applying their rRNA affinity purification pipeline to compare cell types with well-characterized ribosome-associated proteins, like mouse embryonic stem cells and HELA cells. The reviewer can completely appreciate the elegance in the neural characterization here, but it seems there needs to be a solid foothold on the specificity of the method, perhaps facilitated by cell types that can be more readily scaled up and tested.

(2) The authors followed up on their differentially enriched ribosomal complex proteins by analyzing the ribosome association of these proteins in external datasets. While this analysis supports the ribosome-association of these proteins, there is limited experimental validation of physical association with the ribosome, much less any functional characterization. The reciprocal pulldown of PRKCE is promising; however, I would recommend orthogonal validation of several putative ribosomal complex proteins to increase confidence. Specifically, the authors could use sucrose gradient fractionation of SCPN and CPN, followed by a western blot to identify the putative interaction with the 80S monosome or polysomes. This would also provide evidence towards the pulldown capturing association with mature ribosome species, which is currently unclear. This experiment would provide substantial evidence for the direct association of these non-ribosomal proteins with subtype-specific ribosomal complexes.

(3) The authors state interest in learning more about the differences underlying translational regulation of projection neuron development. This method only captures neuronal somata, which will only capture ribosomes in the main cell body. There are also ribosomes regulating local translation in the axons, which may also play a critical role in axonal circuit establishment and activity. These ribosomal complex interactions may also be rather transient and difficult to capture at only one developmental stage. Therefore, this method is currently limited to a single developmental snapshot of ribosomal complexes at P3 within the main cell body. It would be exciting to see the extended utility of this method to sample neurites and additional developmental stages to gain further resolution on the developmental translation regulation of these projection neurons.

Likely impact of the work on the field, and the utility of the methods and data to the community:

The authors introduce a unique pipeline of techniques to identify cell-type-specific ribosomal complex compositions. With more validation, there is certainly potential for those studying neuronal translation to leverage this method in limited primary cells as an alternative to existing methods that do not rely on ribosomal protein tagging, such as ARC-MS (Bartsch et al., 2023), RAPIDASH (Susanto and Hung et al., 2024), and RAPPL (Nature Communications, 2025).

Siento que los sentimientos expresados ​​en estas tarjetas muestran lo poco que cambian las cosas. En aquel entonces se quejaban de los inmigrantes del medio oeste y ahora se quejan de los inmigrantes de otros países. En ambos casos hay una reacción violenta, pero aun así la gente tiene que vivir su vida.

Esta imagen da la impresión de la mirada de mil millas. La mujer parece agotada pero todavía está intentando asegurarse de que su bebé esté abrigado. Esta imagen transmite una fuerte sensación de realismo.

Esta imagen transmite una sensación diferente a muchas otras. Creo que se debe a que está ambientada en una ciudad, no en una granja, y la gente aquí está bien vestida y parece más esperanzada que en otras imágenes.

Otra triste realidad. Una cosa que noto es la inocencia del niño, ya que sonríen incluso en la situación en la que se encuentran. En cuanto a la madre, ella conoce la triste realidad.

Lo más interesante de todas estas imágenes para mí son los ojos de las personas. Hay muchas personas diferentes con todo tipo de experiencias de vida diferentes, pero hay una forma específica en que la cámara captura sus ojos que hace que todas estas imágenes sean familiares.

Esta foto muestra la dura realidad de lo que algunas personas enfrentaron en la gran depresión. Me entristece ver a niños y familias sin un hogar.

Cuando vi esta fotografía, lo me comunicó el estrés de la pobreza puede poner en una familia. La distancia entre los miembros y las apariencias cansadas y estresadas de cada miembro me comunicó esto sentimiento. Además, un otro detalle importante que esta fotografía fue tocado en 1941 - un año que la mayoridad de estadounidenses ahora lo considera el fin del Gran Depresión. Pero para muchas familias estadounidenses negras como la, ellos no recibieron los beneficios del New Deal en el EEUU y continuaban a sufrir en pobreza. Hay un episodio bueno de Crash Course: Black American History por Clint Smith que explican la experiencia del Gran Depresión para los estadounidenses negros.

Me recuerda de los otros intentos a erradicar la existencia de los pobres por el gobierno y otros partidos. Para mí, el ejemplo más sutil pero lo más mayor es que se llama "la arquitectura hostil"; cuando los oficiales ciudadanos establecen reglas para crear bancos, paredes y aceras que prohiben la oportunidad a dormir para los sin techo. Por lo tanto, la arquitectura hostil fuerza a los sin techo a emigrar en otros partes de la ciudad donde ellos no serán visibles para el público genera. La arquitectura hostil puede ser oscuro pero la puede ser permanente también y puede tener una relación costo-eficacia más mayor de las redadas.

Yo creé muchas interpretaciones de esta fotografía. Los paredes son cubiertos con periódicos sobre los eventos más mayores contemporáneos en el mundo, pero los no son el sujeto de la fotografía. La familia pobre es el sujeto, y la ironía de que el mundo continua sin importar la pobreza y la pena de la familia no lo olvidé. También puedes interpretarlo como un comentario sobre la cultura política del EEUU; enfocamos en los eventos espectaculares pero olvidamos los millones de familias que sufren como resultado de la pobreza y quien luchan para un porvenir más mejor porque sus luchas no tan emocionantes como el drama político.

Los niño no parecen feliz, ni bien cuidada, no por su salud, ellos parecen bien alimentados pero la suciedad de. los cuerpos y caras. Tal vez los niños están así porque tuvieron trabajar todo el día el chiquitico también. ¡Guau!

Aquí está mamá es bien vestida, ella y el bebé. Ella es más joven y también está afuera de la casa. Tal vez ella tiene un poquito más éxito en este tiempo. Su cara no es tan mal como las otras mamás. Entonces tal vez su vida es un poco mejor.

De nuevo todos parecen aburridos, pero la casa parece muy tranquilla , limpia y las chicas relajadas. Pero todavía faltan felicidad.

En esta foto, también se siente diferente. Toda la family están juntos. Puedes ver que es un tiempo de dificultad pero hay sonrisas chiquitas en las caras de las mujeres. Pero hay gravedad en las caras de los demás. En está foto se siente más completa y unificada. Como la familia está soportándolo juntas.

Creo que esta foto es diferente porque esta muestra amistad. Algo que no está en las fotos de las mujeres. Ellas están solas con sus hijos, sin amigas o hombres. En pequeño espacios no en la comunidad.

La mujer en esta foto tiene una vibra diferente, como ella tiene un poco de propósito. Las niñas no tienen felicidad para nada y parecen aburridas.

Las mamás parecen triste o enojada y las caras de los niños son son felices realmente tampoco. Esta chica tiene un poco de sonrisa en la cara, pero tal vez alguien esta hablando con ella porque todavía hay un poco de tristeza a tras de los ojos.

Siento que en cada foto las mujeres parecen más mayor. Las caras tienen arrugas profundas. ¿Pensé es que la vida es duro o estas mamás tuvieron hijo mas tarde en sus vidas?

Esta foto me resulta muy interesante. Todas las demás fotos de esta lista me parecen muy deprimentes, tristes y sucias. Sin embargo, en esta, todos están bien vestidos, el hombre está feliz y todo parece estar limpio. ¿Por qué esta es diferente y significativa para la lista? Solo sé que esta foto fue tomada durante la Segunda Guerra Mundial y están haciendo fila para ver una película.

Me gusta que esta foto muestre un poco de alegría en esa época. Ver a tanta gente haciendo fila para ver una película sin duda despierta la felicidad en muchos. Y esta foto lo demuestra a la perfección.

La inclusión de una familia afroamericana en estas fotografías dice mucho sobre el contexto histórico de la época.

Una mirada a la realidad de la vida rural durante la Gran Depresión. Esta cabaña de Ozark, que alberga a seis personas, habla por sí sola sobre la resiliencia y las dificultades que experimentaron.

Este foto puede instilar empatía entre nosotros porque es difícil para mover todas las cosas entre la casa, pero puedo realizo que hay muchas niños que era afectaban también durante el tiempo difícil.

En este foto, creo que las personas son contentos y queríamos celebrar los tiempos mejores durante la gran depresión. Que interesante las personas pueden usar la cine para escapar.

Creo que en este foto podemos ver la dificultad que la gran depresión supuso para las familias, especialmente para las madres cuando miramos a los ojos de ella.

* It's = esto es (falso sujeto). * It's just killing = es solo matar. <br /> El falso sujeto* sucede con el pronombre it y sirve para introducir situaciones donde el gerundio que le prosigue toma la función de sujeto.

* well, who care = bien, a quien le importa. * nice to meet ya = encantado de conocerte.

* grateful (adjetivo) = agradecido * to you = se usa con algunos adjetivos para expresar el por quien esta así. * just like = justo como * treats = golosina * happily (adverbio) = felizmente * thing is = la cosa es

• Starting today = a partir de hoy.
• Subject feel object up (modismo) significa tocar o manosear de manera sexual, en la mayoría de los casos sin su consentimiento .

* buddy = compañero / amigo (en contexto informal). * ya know (contracción) de do you know.

* either = cualquier(cuando no va al final de la oración).

* Give up on (modismo) = rendirme con. * El with en el ingles es mas limitado que en el español. Solo se usa para indicar compañía como un adicional a la oración que complementa sin cambiar el significado de esta. Ejemplo: I wrote a thesis with my best friend.

• Cuando quieras usar los otros usos de con de las otras formas ,como en el español, hara referencia a una phrasal verb o preposicional-phrasal verb que lleve on.

Auschwitz was the main concentration camp of all the camps. More people in Auschwitz died compared to other prisoners in other camps. It is sad how these prisoners were worked to die in the end

HISTORIA QUE SUSTENTA LA CREACION DE MOLECULAS ORGÁNICAS EN UN PLANETA COMO EL NUESTRO

Differentiated professional learning: mentorship adapts to each teacher’s needs. High-quality urban teacher preparation: focus on equity, diversity, and culturally responsive teaching. Research and innovation: mentors and residents reflect and test new practices to improve learning. Simultaneous renewal: universities and schools transform together through partnership

Le genre de commentaire que tes collègues ont pu juger trop « politique » en période de électorale ? Cela me fait penser à la discussion avec mon directeur de mémoire sur la possibilité de faire un travail de recherche engagé sans être normatif c’était intéressant :)

ojo con la numeración de títulos

Creo que a esta sección la llamaría "Cohesión Horizontal". El objetivo principal es la operacionalización, para lo cual se parte de un marco conceptual. En la estructura actual es raro que se llame marco conceptual, cuando de eso hay solo un párrafo y una cita.

Le falta mucho para revisar aún. Falta contexto y orientación a audiencia (=otro_s), no se entiende a menos que uno sea parte de este equipo, y la idea es que sea entendido por público general. Esto se le está contando a alguien que tiene que ser capaz de entenderlo. Partiendo por "en esta sección...". Y en todas las secciones, en particluar la de propuestas donde no hay nada escrito.

redundante, mejor "En el caso de la simpatía hacia migrantes, ...

for - citation - book - Patrick Harpur - Daimonic Reality: A field guide to the otherworld - to - book Daimonic Reality: A field guide to the otherworld - Patrick Harpur - adjacency - realm between fact and fiction - Donald Hoffman interview - Deep Humanity - self / other gestalt - the Indyweb - physiosphere - symbolosphere - this is exactly the intetwingledness of - the subject and the object - consciousness and phenomenal reality - Deep Humanity - the individual / collective gestalt - the self / other gestalt - symbolosphere / physiosphere - to - Youtube - The Diary of a CEO - Donald Hoffman interview - https://hyp.is/go?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DW0vTZrZny6A&group=world - internet Archive - https://hyp.is/egkk-IvhEfCpxyM0mIOqLA/archive.org/details/daimonicrealityf0000harp - Patrick Harpur - book webpage - https://hyp.is/1iPUDovhEfC4PStyYJoYnQ/www.harpur.org/x1Daimonic.htm

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

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2025-02922

Corresponding author(s): Christian Specht

[Please use this template only if the submitted manuscript should be considered by the affiliate journal as a full revision in response to the points raised by the reviewers.

• *

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

1. General Statements [optional]

This section is optional. Insert here any general statements you wish to make about the goal of the study or about the reviews.

• *

We thank the reviewers for their thorough and constructive evaluation of our work. We have revised the manuscript carefully and addressed all the criticisms raised, in particular the issues mentioned by several of the reviewers (see point-by-point response below). We have also added a number of explanations in the text for the sake of clarity, while trying to keep the manuscript as concise as possible.

• *

In our view, the novelty of our research is two-fold. From a neurobiological point of view, we provide conclusive evidence for the existence of glycine receptors (GlyRs) at inhibitory synapses in various brain regions including the hippocampus, dentate gyrus and sub-regions of the striatum. This solves several open questions and has fundamental implications for our understanding of the organisation and function of inhibitory synapses in the telencephalon. Secondly, our study makes use of the unique sensitivity of single molecule localisation microscopy (SMLM) to identify low protein copy numbers. This is a new way to think about SMLM as it goes beyond a mere structural characterisation and towards a quantitative assessment of synaptic protein assemblies.

2. Point-by-point description of the revisions

This section is mandatory. *Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. *

• *

__Reviewer #1 (Evidence, reproducibility and clarity (Required)): __

In this manuscript, the authors investigate the nanoscopic distribution of glycine receptor subunits in the hippocampus, dorsal striatum, and ventral striatum of the mouse brain using single-molecule localization microscopy (SMLM). They demonstrate that only a small number of glycine receptors are localized at hippocampal inhibitory synapses. Using dual-color SMLM, they further show that clusters of glycine receptors are predominantly localized within gephyrin-positive synapses. A comparison between the dorsal and ventral striatum reveals that the ventral striatum contains approximately eight times more glycine receptors and this finding is consistent with electrophysiological data on postsynaptic inhibitory currents. Finally, using cultured hippocampal neurons, they examine the differential synaptic localization of glycine receptor subunits (α1, α2, and β). This study is significant as it provides insights into the nanoscopic localization patterns of glycine receptors in brain regions where this protein is expressed at low levels. Additionally, the study demonstrates the different localization patterns of GlyR in distinct striatal regions and its physiological relevance using SMLM and electrophysiological experiments. However, several concerns should be addressed.

The following are specific comments:

1. Colocalization analysis in Figure 1A. The colocalization between Sylite and mEos-GlyRβ appears to be quite low. It is essential to assess whether the observed colocalization is not due to random overlap. The authors should consider quantifying colocalization using statistical methods, such as a pixel shift analysis, to determine whether colocalization frequencies remain similar after artificially displacing one of the channels. *Following the suggestion of reviewer 1, we re-analysed CA3 images of Glrbeos/eos hippocampal slices by applying a pixel-shift type of control, in which the Sylite channel (in far red) was horizontally flipped relative to the mEos4b-GlyRβ channel (in green, see Methods). As expected, the number of mEos4b-GlyRβ detections per gephyrin cluster was markedly reduced compared to the original analysis (revised__ Fig. 1B__), confirming that the synaptic mEos4b detections exceed chance levels (see page 5). *

Inconsistency between Figure 3A and 3B. While Figure 3B indicates an ~8-fold difference in the number of mEos4b-GlyRβ detections per synapse between the dorsal and ventral striatum, Figure 3A does not appear to show a pronounced difference in the localization of mEos4b-GlyRβ on Sylite puncta between these two regions. If the images presented in Figure 3A are not representative, the authors should consider replacing them with more representative examples or providing an expanded images with multiple representative examples. Alternatively, if this inconsistency can be explained by differences in spot density within clusters, the authors should explain that.

*The pointillist images in Fig. 3A are essentially binary (red-black). Therefore, the density of detections at synapses cannot be easily judged by eye. For clarity, the original images in Fig. 3A have been replaced with two other examples that better reflect the different detection numbers in the dorsal and ventral striatum. *

• *

Quantification in Figure 5. It is recommended that the authors provide quantitative data on cluster formation and colocalization with Sylite puncta in Figure 5 to support their qualitative observations.

*This is an important point that was also raised by the other reviewers. We have performed additional experiments to increase the data volume for analysis. For quantification, we used two approaches. First, we counted the percentage of infected cells in which synaptic localisation of the recombinant receptor subunit was observed (Fig. 5C). We found that mEos4b-GlyRa1 consistently localises at synapses, indicating that all cells express endogenous GlyRb. When neurons were infected with mEos4b-GlyRb, fewer cells had synaptic clusters, meaning that indeed, GlyR alpha subunits are the limiting factor for synaptic targeting. In cultures infected with mEos4b-GlyRa2, only very few neurons displayed synaptic localisation (as judged by epifluorescence imaging). We think this shows that GlyRa2 is less capable of forming heteromeric complexes than GlyRa1, in line with our previous interpretation (see pp. 9-10, 13). *

• *

Secondly, we quantified the total intensity of each subunit at gephyrin-positive domains, both in infected neurons as well as non-infected control cultures (Fig. 5D). We observed that mEos4b-GlyRa1 intensity at gephyrin puncta was higher than that of the other subunits, again pointing to efficient synaptic targeting of GlyRa1. Gephyrin cluster intensities (Sylite labelling) were not significantly different in GlyRb and GlyRa2 expressing neurons compared to the uninfected control, indicating that the lentiviral expression of recombinant subunits does not fundamentally alter the size of mixed inhibitory synapses in hippocampal neurons. Interestingly, gephyrin levels were slightly higher in hippocampal neurons expressing mEos4b-GlyRa1. In our view, this comes from an enhanced expression and synaptic targeting of mEos4b-GlyRa1 heteromers with endogenous GlyRb, pointing to a structural role of GlyRa1/b in hippocampal synapses (pp. 10, 13).

• *

The new data and analyses have been described and illustrated in the relevant sections of the manuscript.

Potential for pseudo replication. It's not clear whether they're performing stats tests across biological replica, images, or even synapses. They often quote mean +/- SEM with n = 1000s, and so does that mean they're doing tests on those 1000s? Need to clarify.

All experiments were repeated at least twice to ensure reproducibility (N independent experiments). Statistical tests were performed on pooled data across the biological replicates; n denotes the number of data points used for testing (e.g., number of synaptic clusters, detections, cells, as specified in each case). We have systematically given these numbers in the revised manuscript (n, N, and other experimental parameters such as the number of animals used, coverslips, images or cells). Data are generally given as mean +/- SEM or as mean +/- SD as indicated.

• *

Does mEoS effect expression levels or function of the protein? Can't see any experiments done to confirm this. Could suggest WB on homogenate, or mass spec?

The Glrbeos/eos knock-in mouse line has been characterised previously and does not to display any ultrastructural or functional deficits at inhibitory synapses (Maynard et al. 2021 eLife). GlyRβ expression and glycine-evoked responses were not significantly different to those of the wild-type. The synaptic localisation of mEos4b-GlyRb in KI animals demonstrates correct assembly of heteromeric GlyRs and synaptic targeting. Accordingly, the animals do not display any obvious phenotype. We have clarified this in the manuscript (p. 4). In the case of cultured neurons, long-term expression of fluorescent receptor subunits with lentivirus has proven ideal to achieve efficient synaptic targeting. The low and continuous supply of recombinant receptors ensures assembly with endogenous subunits to form heteropentameric receptor complexes (e.g. [Patrizio et al. 2017 Sci Rep]). In the present study, lentivirus infection did not induce any obvious differences in the number or size of inhibitory synapses compared to control neurons, as judged by Sylite labelling of synaptic gephyrin puncta (new__ Fig. 5D__).

Quantification of protein numbers is challenging with SMLM. Issues include i) some of FP not correctly folded/mature, and ii) dependence of localisation rate on instrument, excitation/illumination intensities, and also the thresholds used in analysis. Can the authors compare with another protein that has known expression levels- e.g. PSD95? This is quite an ask, but if they could show copy number of something known to compare with, it would be useful.

We agree that absolute quantification with SMLM is challenging, since the number of detections depends on fluorophore maturation, photophysics, imaging conditions, and analysis thresholds (discussed in Patrizio & Specht 2016, Neurophotonics). For this reason, only very few datasets provide reliable copy numbers, even for well-studied proteins such as PSD-95. One notable exception is the study by Maynard et al. (eLife 2021) that quantified endogenous GlyRb-containing receptors in spinal cord synapses using SMLM combined with correlative electron microscopy. The strength of this work was the use of a KI mouse strain, which ensures that mEos4b-GlyRb expression follows intrinsic regional and temporal profiles. The authors reported a stereotypic density of ~2,000 GlyRs/µm² at synapses, corresponding to ~120 receptors per synapse in the dorsal horn and ~240 in the ventral horn, taking into account various parameters including receptor stoichiometry and the functionality of the fluorophore. These values are very close to our own calculations of GlyR numbers at spinal cord synapses that were obtained slightly differently in terms of sample preparation, microscope setup, imaging conditions, and data analysis, lending support to our experimental approach. Nevertheless, the obtained GlyR copy numbers at hippocampal synapses clearly have to be taken as estimates rather than precise figures, because the number of detections from a single mEos4b fluorophore can vary substantially, meaning that the fluorophores are not represented equally in pointillist images. This can affect the copy number calculation for a specific synapse, in particular when the numbers are low (e.g. in hippocampus), however, it should not alter the average number of detections (Fig. 1B) or the (median) molecule numbers of the entire population of synapses (Fig. 1C). We have discussed the limitations of our approach (p. 11).

Rationale for doing nanobody dSTORM not clear at all. They don't explain the reason for doing the dSTORM experiments. Why not just rely on PALM for coincidence measurements, rather than tagging mEoS with a nanobody, and then doing dSTORM with that? Can they explain? Is it to get extra localisations- i.e. multiple per nanobody? If so, localising same FP multiple times wouldn't improve resolution. Also, no controls for nanobody dSTORM experiments- what about non-spec nb, or use on WT sections?

*As discussed above (point 6), the detection of fluorophores with SMLM is influenced by many parameters, not least the noise produced by emitting molecules other than the fluorophore used for labelling. Our study is exceptional in that it attempts to identify extremely low molecule numbers (down to 1). To verify that the detections obtained with PALM correspond to mEos4b, we conducted robust control experiments (including pixel-shift as suggested by the reviewer, see point 1, revised__ Fig. 1B__). The rationale for the nanobody-based dSTORM experiments was twofold: (1) to have an independent readout of the presence of low-copy GlyRs at inhibitory synapses and (2) to analyse the nanoscale organisation of GlyRs relative to the synaptic gephyrin scaffold using dual-colour dSTORM with spectral demixing (see p. 6). The organic fluorophores used in dSTORM (AF647, CF680) ensure high photon counts, essential for reliable co-localisation and distance analysis. PALM and dSTORM cannot be combined in dual-colour mode, as they require different buffers and imaging conditions. *

The specificity of the anti-Eos nanobody was demonstrated by immunohistochemistry in spinal cord cultures expressing mEos4b-GlyRb and wildtype control tissue (Fig. S3). In response to the reviewer's remarks, we also performed a negative control experiment in Glrbeos/eos slices (dSTORM), in which the nanobody was omitted (new__ Fig. S4F,G__). Under these conditions, spectral demixing produced a single peak corresponding to CF680 (gephyrin) without any AF647 contribution (Fig. S4F). The background detection of "false" AF647 detections at synapses was significantly lower than in the slices labelled with the nanobody. We conclude that the fluorescence signal observed in our dual-colour dSTORM experiments arises from the specific detection of mEos4b-GlyRb by the nanobody, rather than from background, cross-reactivity or wrong attribution of colour during spectral demixing. We have added these data and explanations in the results (p. 7) and in the figure legend of Fig. S4F,G.

What resolutions/precisions were obtained in SMLM experiments? Should perform Fourier Ring Correlation (FRC) on SR images to state resolutions obtained (particularly useful for when they're presenting distance histograms, as this will be dependent on resolution). Likewise for precision, what was mean precision? Can they show histograms of localisation precision.

This is an interesting question in the context of our experiments with low-copy GlyRs, since the spatial resolution of SMLM is limited also by the density of molecules, i.e. the sampling of the structure in question (Nyquist-Shannon criterion). Accordingly, the priority of the PALM experiments was to improve the sensibility of SMLM for the identification of mEos4b-GlyRb subunits, rather than to maximize the spatial resolution. The mean localisation precision in PALM was 33 +/- 12 nm, as calculated from the fitting parameters of each detection (Zeiss, ZEN software), which ultimately result from their signal-to-noise ratio. This is a relatively low precision for SMLM, which can be explained by the low brightness of mEos4b compared to organic fluorophores together with the elevated fluorescence background in tissue slices.

• *

In the case of dSTORM, the aim was to study the relative distribution of GlyRs within the synaptic scaffold, for which a higher localisation precision was required (p. 6). Therefore, detections with a precision ≥ 25 nm were filtered during analysis with NEO software (Abbelight). The retained detections had a mean localisation precision of 12 +/- 5 for CF680 (Sylite) and 11 +/- 4 for AF647 (nanobody). These values are given in the revised manuscript (pp. 18, 22).

Why were DBSCAN parameters selected? How can they rule out multiple localisations per fluor? If low copy numbers (

Multiple detections of the same fluorophore are intrinsic to dSTORM imaging and have not been eliminated from the analysis. Small clusters of detections likely represent individual molecules (e.g. single receptors in the extrasynaptic regions, Fig. 2A). DBSCAN is a robust clustering method that is quite insensitive to minor changes in the choice of parameters. For dSTORM of synaptic gephyrin clusters (CF680), a relatively low length (80 nm radius) together with a high number of detections (≥ 50 neighbours) were chosen to reconstruct the postsynaptic domain with high spatial resolution (see point 8). In the case of the GlyR (nanobody-AF647), the clustering was done mostly for practical reasons, as it provided the coordinates of the centre of mass of the detections. The low stringency of this clustering (200 nm radius, ≥ 5 neighbours) effectively filters single detections that can result from background noise or incorrect demixing. An additional reference explaining the use of DBSCAN including the choice of parameters is given on p. 22 (see also R2 point 4).

For microscopy experiment methods, state power densities, not % or "nominal power".

*Done. We now report the irradiance (laser power density) instead of nominal power (pp. 18, 21). *

In general, not much data presented. Any SI file with extra images etc.?

*The original submission included four supplementary figures with additional data and representative images that should have been available to the reviewer (Figs. S1-S4). The SI file has been updated during revision (new Fig. S4E-G). *

Clarification of the discussion on GlyR expression and synaptic localization: The discussion on GlyR expression, complex formation, and synaptic localization is sometimes unclear, and needs terminological distinctions between "expression level", "complex formation" and "synaptic localization". For example, the authors state:"What then is the reason for the low protein expression of GlyRβ? One possibility is that the assembly of mature heteropentameric GlyR complexes depends critically on the expression of endogenous GlyR α subunits." Does this mean that GlyRβ proteins that fail to form complexes with GlyRα subunits are unstable and subject to rapid degradation? If so, the authors should clarify this point. The statement "This raises the interesting possibility that synaptic GlyRs may depend specifically on the concomitant expression of both α1 and β transcripts." suggests a dependency on α1 and β transcripts. However, is the authors' focus on synaptic localization or overall protein expression levels? If this means synaptic localization, it would be beneficial to state this explicitly to avoid confusion. To improve clarity, the authors should carefully distinguish between these different aspects of GlyR biology throughout the discussion. Additionally, a schematic diagram illustrating these processes would be highly beneficial for readers.

We thank the reviewer to point this out. We are dealing with several processes; protein expression that determines subunit availability and the assembly of pentameric GlyRs complexes, surface expression, membrane diffusion and accumulation of GlyRb-containing receptor complexes at inhibitory synapses. We have edited the manuscript, particularly the discussion and tried to be as clear as possible in our wording.

• *

We chose not to add a schematic illustration for the time being, because any graphical representation is necessarily a simplification. Instead, we preferred to summarise the main numbers in tabular form (Table 1). We are of course open to any other suggestions.

Interpretation of GlyR localization in the context of nanodomains. The distribution of GlyR molecules on inhibitory synapses appears to be non-homogeneous, instead forming nanoclusters or nanodomains, similar to many other synaptic proteins. It is important to interpret GlyR localization in the context of nanodomain organization.

The dSTORM images in Fig. 2 are pointillist representations that show individual detections rather than molecules. Small clusters of detections are likely to originate from a single AF647 fluorophore (in the case of nanobody labelling) and therefore represent single GlyRb subunits. Since GlyR copy numbers are so low at hippocampal synapses (≤ 5), the notion of nanodomain is not directly applicable. Our analysis therefore focused on the integration of GlyRs within the postsynaptic scaffold, rather than attempting to define nanodomain structures (see also response to point 8 of R1). A clarification has been added in the revised manuscript (p. 6).

__Reviewer #1 (Significance (Required)): __

The paper presents biological and technical advances. The biological insights revolve mostly on the documentation of Glycine receptors in particular synapses in forebrain, where they are typically expressed at very low levels. The authors provide compelling data indicating that the expression is of physiological significance. The authors have done a nice job of combining genetically-tagged mice with advanced microscopy methods to tackle the question of distributions of synaptic proteins. Overall these advances are more incremental than groundbreaking.

We thank the reviewer for acknowledging both the technical and biological advances of our study. While we recognize that our work builds upon established models, we consider that it also addresses important unresolved questions, namely that GlyRs are present and specifically anchored at inhibitory synapses in telencephalic regions, such as the hippocampus and striatum. From a methodological point of view, our study demonstrates that SMLM can be applied not only for structural analysis of highly abundant proteins, but also to reliably detect proteins present at very low copy numbers. This ability to identify and quantify sparse molecule populations adds a new dimension to SMLM applications, which we believe increases the overall impact of our study beyond the field of synaptic neuroscience.

__Reviewer #2 (Evidence, reproducibility and clarity (Required)): __

In their manuscript "Single molecule counting detects low-copy glycine receptors in hippocampal and striatal synapses" Camuso and colleagues apply single molecule localization microscopy (SMLM) methods to visualize low copy numbers of GlyRs at inhibitory synapses in the hippocampal formation and the striatum. SMLM analysis revealed higher copy numbers in striatum compared to hippocampal inhibitory synapses. They further provide evidence that these low copy numbers are tightly linked to post-synaptic scaffolding protein gephyrin at inhibitory synapses. Their approach profits from the high sensitivity and resolution of SMLM and challenges the controversial view on the presence of GlyRs in these formations although there are reports (electrophysiology) on the presence of GlyRs in these particular brain regions. These new datasets in the current manuscript may certainly assist in understanding the complexity of fundamental building blocks of inhibitory synapses.

However I have some minor points that the authors may address for clarification:

1) In Figure 1 the authors apply PALM imaging of mEos4b-GlyRß (knockin) and here the corresponding Sylite label seems to be recorded in widefield, it is not clearly stated in the figure legend if it is widefield or super-resolved. In Fig 1 A - is the scale bar 5 µm? Some Sylite spots appear to be sized around 1 µm, especially the brighter spots, but maybe this is due to the lower resolution of widefield imaging? Regarding the statistical comparison: what method was chosen to test for normality distribution, I think this point is missing in the methods section.

*This is correct; the apparent size of the Sylite spots does not reflect the real size of the synaptic gephyrin domain due to the limited resolution of widefield imaging including the detection of out-of-focus light. We have clarified in the legend of Fig. 1A that Sylite labelling was with classic epifluorescence microscopy. The scale bar in Fig. 1A corresponds to 5 µm. Since the data were not normally distributed, nonparametric tests (Kruskal- Wallis one-way ANOVA with Dunn’s multiple comparison test or Mann-Whitney U-test for pairwise comparisons) were used (p. 23). *

Moreover I would appreciate a clarification and/or citation that the knockin model results in no structural and physiological changes at inhibitory synapses, I believe this model has been applied in previous studies and corresponding clarification can be provided.

The Glrbeos/eos mouse model has been described previously and does not exhibit any structural or physiological phenotypes (Maynard et al. 2021 eLife). The issue was also raised by reviewer R1 (point 5) and has been clarified in the revised manuscript (p. 4).

2) In the next set of experiments the authors switch to demixing dSTORM experiments - an explanation why this is performed is missing in the text - I guess better resolution to perform more detailed distance measurements? For these experiments: which region of the hippocampus did the authors select, I cannot find this information in legend or main text.

Yes, the dSTORM experiments enable dual-colour structural analysis at high spatial resolution (see response to R1 point 7). An explanation has been added (p. 6).

3) Regarding parameters of demixing experiments: the number of frames (10.000) seems quite low and the exposure time higher than expected for Alexa 647. Can the authors explain the reason for chosing these particular parameters (low expression profile of the target - so better separation?, less fluorophores on label and shorter collection time?) or is there a reference that can be cited? The laser power is given in the methods in percentage of maximal output power, but for better comparison and reproducibility I recommend to provide the values of a power meter (kW/cm2) as lasers may change their maximum output power during their lifetime.

Acquisition parameters (laser power, exposure time) for dSTORM were chosen to obtain a good localisation precision (~12 nm; see R1 point 8). The number of frames is adequate to obtain well sampled gephyrin scaffolds in the CF680 channel. In the case of the GlyR (nanobody-AF647), the concept of spatial resolution does not really apply due to the low number of targets (see R1, point 13). Power density (irradiance) values have now been given (pp. 18, 21).

4) For analysis of subsynaptic distribution: how did the authors decide to choose the parameters in the NEO software for DBSCAN clustering - was a series of parameters tested to find optimal conditions and did the analysis start with an initial test if data is indeed clustered (K-ripley) or is there a reference in literature that can be provided?

DBSCAN parameters were optimised manually, by testing different values. Identification of dense and well-delimited gephyrin clusters (CF680) was achieved with a small radius and a high number of detections (80 nm, ≥ 50 neighbours), whereas filtering of low-density background in the AF647 channel (GlyRs) required less stringent parameters (200 nm, ≥ 5) due to the low number of target molecules. Similar parameters were used in a previous publication (Khayenko et al. 2022, Angewandte Chemie). The reference has been provided on p. 22 (see also R1 point 9).

5) A conclusion/discussion of the results presented in Figure 5 is missing in the text/discussion.

*This part of the manuscript has been completely overhauled. It includes new experimental data, quantification of the data (new Fig.5), as well as the discussion and interpretation of our findings (see also R1, point 3). In agreement with our earlier interpretation, the data confirm that low availability of GlyRa1 subunits limits the expression and synaptic targeting of GlyRa1/b heteropentamers. The observation that GlyRa1 overexpression with lentivirus increases the size of the postsynaptic gephyrin domain further points to a structural role, whereby GlyRs can enhance the stability (and size) of inhibitory synapses in hippocampal neurons, even at low copy numbers (pp. 13-14). *

6) in line 552 "suspension" is misleading, better use "solution"

Done.

__Reviewer #2 (Significance (Required)): __

Significance: The manuscript provides new insights to presence of low-copy numbers by visualizing them via SMLM. This is the first report that visualizes GlyR optically in the brain applying the knock-in model of mEOS4b tagged GlyRß and quantifies their copy number comparing distribution and amount of GlyRs from hippocampus and striatum. Imaging data correspond well to electrophysiological measurements in the manuscript.

Field of expertise: Super-Resolution Imaging and corresponding analysis

__Reviewer #4 (Evidence, reproducibility and clarity (Required)): __

In this study, Camuso et al., make use of a knock-in mouse model expressing endogenously mEos4b-tagged GlyRβ to detect endogenous glycine receptors using single-molecule localization microscopy. The main conclusion from this study is that in the hippocampus GlyRβ molecules are barely detected, while inhibitory synapses in the ventral striatum seem to express functionally relevant GlyR numbers.

I have a few points that I hope help to improve the strength of this study.

• In the hippocampus, this study finds that the numbers of detections are very low. The authors perform adequate controls to indicate that these localizations are above noise level. Nevertheless, it remains questionable that these reflect proper GlyRs. The suggestion that in hippocampal synapses the low numbers of GlyRβ molecules "are important in assembly or maintenance of inhibitory synaptic structures in the brain" is on itself interesting, but is not at all supported. It is also difficult to envision how such low numbers could support the structure of a synapse. A functional experiment showing that knockdown of GlyRs affects inhibitory synapse structure in hippocampal neurons would be a minimal test of this.

*It is not clear what the reviewer means by “it remains questionable that these reflect proper GlyRs”. The PALM experiments include a series of stringent controls (see R1, point 1) demonstrating the existence of low-copy GlyRs at inhibitory synapses in the hippocampus (Fig. 1) and in the striatum (Fig. 3), and are backed up by dSTORM experiments (Fig. 2). We have no reason to doubt that these receptors are fully functional (as demonstrated for the ventral striatum (Fig. 4). However, due to their low number, a role in inhibitory synaptic transmission is clearly limited, at least in the hippocampus and dorsal striatum. *

• *

We therefore propose a structural role, where the GlyRs could be required to stabilise the postsynaptic gephyrin domain in hippocampal neurons. This is based on the idea that the GlyR-gephyrin affinity is much higher than that of the GABAAR-gephyrin interaction (reviewed in Kasaragod & Schindelin 2018 Front Mol Neurosci). Accordingly, there is a close relationship between GlyRs and gephyrin numbers, sub-synaptic distribution, and dynamics in spinal cord synapses that are mostly glycinergic (Specht et al. 2013 Neuron; Maynard et al. 2021 eLife; Chapdelaine et al. 2021 Biophys J). It is reasonable to assume that low-copy GlyRs could play a similar structural role at hippocampal synapses. A knockdown experiment targeting these few receptors is technically very challenging and beyond the scope of this study. However, in response to the reviewer's question we have conducted new experiments in cultured hippocampal neurons (new__ Fig. 5__). They demonstrate that overexpression of GlyRa1/b heteropentamers increases the size of the postsynaptic domain in these neurons, supporting our interpretation of a structural role of low-copy GlyRs (p. 14).

• The endogenous tagging strategy is a very strong aspect of this study and provides confidence in the labeling of GlyRβ molecules. One caveat however, is that this labeling strategy does not discriminate whether GlyRβ molecules are on the cell membrane or in internal compartments. Can the authors provide an estimate of the ratio of surface to internal GlyRβ molecules?

Gephyrin is known to form a two-dimensional scaffold below the synaptic membrane to which inhibitory GlyRs and GABAARs attach (reviewed in Alvarez 2017 Brain Res). The majority of the synaptic receptors are therefore thought to be located in the synaptic membrane, which is supported by the close relationship between the sub-synaptic distribution of GlyRs and gephyrin in spinal cord neurons (e.g. Maynard et al. 2021 eLife). To demonstrate the surface expression of GlyRs at hippocampal synapses we labelled cultured hippocampal neurons expressing mEos4b-GlyRa1 with anti-Eos nanobody in non-permeabilised neurons (see Figure below for the reviewer only). The close correspondence between the nanobody (AF647) and the mEos4b signal confirms that the majority of the GlyRs are indeed located in the synaptic membrane.

• *

Figure (for the reviewer only).* Left: Lentivirus expression of mEos4b-GlyRa1 in fixed and non-permeabilised hippocampal neurons (mEos4b signal). Right: Surface labelling of the recombinant subunit with anti-Eos nanoboby (AF647). *

• 'We also estimated the absolute number of GlyRs per synapse in the hippocampus. The number of mEos4b detections was converted into copy numbers by dividing the detections at synapses by the average number of detections of individual mEos4b-GlyRβ containing receptor complexes'. In essence this is a correct method to estimate copy numbers, and the authors discuss some of the pitfalls associated with this approach (i.e., maturation of fluorophore and detection limit). Nevertheless, the authors did not subtract the number of background localizations determined in the two negative control groups. This is critical, particularly at these low-number estimations.

We fully agree that background subtraction can be useful with low detection numbers. In the revised manuscript, copy numbers are now reported as background-corrected values. Specifically, the mean number of detections measured in wildtype slices was used to calculate an equivalent receptor number, which was then subtracted from the copy number estimates across hippocampus, spinal cord and striatum. This procedure is described in the methods (p. 20) and results (p. 5, 8), and mentioned in the figure legends of Fig. 1C, 3C. The background corrected values are given in the text and Table 1.

Furthermore, the authors state that "The advantage of this estimation is that it is independent of the stoichiometry of heteropentameric GlyRs". However, if the stoichometry is unknown, the number of counted GlyRβ subunits cannot simply be reported as the number of GlyRs. This should be discussed in more detail, and more carefully reported throughout the manuscript.

*The reviewer is right to point this out. There is still some debate about the stoichiometry of heteropentameric GlyRs. Configurations with 2a:3b, 3a:2b and 4a:1b subunits have been advanced (e.g. Grudzinska et al. 2005 Neuron; Durisic et al. 2012 J Neurosci; Patrizio et al. 2017 Sci Rep; Zhu & Gouaux 2021 Nature). We have therefore chosen a quantification that is independent of the underlying stoichiometry. Since our quantification is based on very sparse clusters of mEos4b detections that likely originate from a single receptor complex (irrespective of its stoichiometry), the reported values actually reflect the number of GlyRs (and not GlyRb subunits). We have clarified this in the results (p. 5) and throughout the manuscript (Table 1). *

• The dual-color imaging provides insights in the subsynaptic distribution of GlyRβ molecules in hippocampal synapses. Why are similar studies not performed on synapses in the ventral striatum where functionally relevant numbers of GlyRβ molecules are found? Here insights in the subsynaptic receptor distribution would be of much more interest as it can be tight to the function.

This is an interesting suggestion. However, the primary aim of our study was to identify the existence of GlyRs in hippocampal regions. At low copy numbers, the concept of sub-synaptic domains (SSDs, e.g. Yang et al. 2021 EMBO Rep) becomes irrelevant (see R1 point 13). It should be pointed out that the dSTORM pointillist images (Fig. 2A) represent individual GlyR detections rather than clusters of molecules. In the striatum, our specific purpose was to solve an open question about the presence of GlyRs in different subregions (putamen, nucleus accumbens).

• It is unclear how the experiments in Figure 5 add to this study. These results are valid, but do not seem to directly test the hypothesis that "the expression of α subunits may be limiting factor controlling the number of synaptic GlyRs". These experiments simply test if overexpressed α subunits can be detected. If the α subunits are limiting, measuring the effect of α subunit overexpression on GlyRβ surface expression would be a more direct test.

Both R1 and R2 have also commented on the data in Fig. 5 and their interpretation. We have substantially revised this section as described before (see R1 point 3) including additional experiments and quantification of the data (new Fig. 5). The findings lend support to our earlier hypothesis that GlyR alpha subunits (in particular GlyRa1) are the limiting factor for the expression of heteropentameric GlyRa/b in hippocampal neurons (pp. 13-14). Since the GlyRa1 subunit itself does not bind to gephyrin (Patrizio et al. 2017 Sci Rep), the synaptic localisation of the recombinant mEos4b-GlyRa1 subunits is proof that they have formed heteropentamers with endogenous GlyRb subunits and driven their membrane trafficking, which the GlyRb subunits are incapable of doing on their own.

__Reviewer #4 (Significance (Required)): __

These results are based on carefully performed single-molecule localization experiments, and are well-presented and described. The knockin mouse with endogenously tagged GlyRβ molecules is a very strong aspect of this study and provides confidence in the labeling, the combination with single-molecule localization microscopy is very strong as it provides high sensitivity and spatial resolution.

The conceptual innovation however seems relatively modest, these results confirm previous studies but do not seem to add novel insights. This study is entirely descriptive and does not bring new mechanistic insights.

This study could be of interest to a specialized audience interested in glycine receptor biology, inhibitory synapse biology and super-resolution microscopy.

my expertise is in super-resolution microscopy, synaptic transmission and plasticity

As we have stated before, the novelty of our study lies in the use of SMLM for the identification of very small numbers of molecules, which requires careful control experiments. This is something that has not been done before and that can be of interest to a wider readership, as it opens up SMLM for ultrasensitive detection of rare molecular events. Using this approach, we solve two open scientific questions: (1) the demonstration that low-copy GlyRs are present at inhibitory synapses in the hippocampus, (2) the sub-region specific expression and functional role of GlyRs in the ventral versus dorsal striatum.

• *

• *

The following review was provided later under the name “Reviewer #4”. To avoid confusion with the last reviewer from above we will refer to this review as R4-2.

__Reviewer #4-2 (Evidence, reproducibility and clarity (Required)): __

Summary:

Provide a short summary of the findings and key conclusions (including methodology and model system(s) where appropriate).

The authors investigate the presence of synaptic glycine receptors in the telencephalon, whose presence and function is poorly understood.

Using a transgenically labeled glycine receptor beta subunit (Glrb-mEos4b) mouse model together with super-resolution microscopy (SLMM, dSTORM), they demonstrate the presence of a low but detectable amount of synaptically localized GLRB in the hippocampus. While they do not perform a functional analysis of these receptors, they do demonstrate that these subunits are integrated into the inhibitory postsynaptic density (iPSD) as labeled by the scaffold protein gephyrin. These findings demonstrate that a low level of synaptically localized glycerine receptor subunits exist in the hippocampal formation, although whether or not they have a functional relevance remains unknown.

They then proceed to quantify synaptic glycine receptors in the striatum, demonstrating that the ventral striatum has a significantly higher amount of GLRB co-localized with gephyrin than the dorsal striatum or the hippocampus. They then recorded pharmacologically isolated glycinergic miniature inhibitory postsynaptic currents (mIPSCs) from striatal neurons. In line with their structural observations, these recordings confirmed the presence of synaptic glycinergic signaling in the ventral striatum, and an almost complete absence in the dorsal striatum. Together, these findings demonstrate that synaptic glycine receptors in the ventral striatum are present and functional, while an important contribution to dorsal striatal activity is less likely.

Lastly, the authors use existing mRNA and protein datasets to show that the expression level of GLRA1 across the brain positively correlates with the presence of synaptic GLRB.

The authors use lentiviral expression of mEos4b-tagged glycine receptor alpha1, alpha2, and beta subunits (GLRA1, GLRA1, GLRB) in cultured hippocampal neurons to investigate the ability of these subunits to cause the synaptic localization of glycine receptors. They suggest that the alpha1 subunit has a higher propensity to localize at the inhibitory postsynapse (labeled via gephyrin) than the alpha2 or beta subunits, and may therefore contribute to the distribution of functional synaptic glycine receptors across the brain.

Major comments:

• Are the key conclusions convincing?

The authors are generally precise in the formulation of their conclusions.

• They demonstrate a very low, but detectable, amount of a synaptically localized glycine receptor subunit in a transgenic (GlrB-mEos4b) mouse model. They demonstrate that the GLRB-mEos4b fusion protein is integrated into the iPSD as determined by gephyrin labelling. The authors do not perform functional tests of these receptors and do not state any such conclusions.
• The authors show that GLRB-mEos4b is clearly detectable in the striatum and integrated into gephyrin clusters at a significantly higher rate in the ventral striatum compared to the dorsal striatum, which is in line with previous studies.
• Adding to their quantification of GLRB-mEos4b in the striatum, the authors demonstrate the presence of glycinergic miniature IPSCs in the ventral striatum, and an almost complete absence of mIPSCs in the dorsal striatum. These currents support the observation that GLRB-mEos4b is more synaptically integrated in the ventral striatum compared to the dorsal striatum.

• The authors show that lentiviral expression of GLRA1-mEos4b leads to a visually higher number of GLR clusters in cultured hippocampal neurons, and a co-localization of some clusters with gephyrin. The authors claim that this supports the idea that GLRA1 may be an important driver of synaptic glycine receptor localization. However, no quantification or statistical analysis of the number of puncta or their colocalization with gephyrin is provided for any of the expressed subunits. Such a claim should be supported by quantification and statistics A thorough analysis and quantification of the data in Fig.5 has been carried out as requested by all the other reviewers (e.g. R1, point 3). The new data and results have been described in the revised manuscript (pp. 9-10, 13-14).

• Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?

One unaddressed caveat is the fact that a GLRB-mEos4b fusion protein may behave differently in terms of localization and synaptic integration than wild-type GLRB. While unlikely, it is possible that mEos4b interacts either with itself or synaptic proteins in a way that changes the fused GLRB subunit’s localization. Such an effect would be unlikely to affect synaptic function in a measurable way, but might be detected at a structural level by highly sensitive methods such as SMLM and STORM in regions with very low molecule numbers (such as the hippocampus). Since reliable antibodies against GLRB in brain tissue sections are not available, this would be difficult to test. Considering that no functional measures of the hippocampal detections exist, we would suggest that this possible caveat be mentioned for this particular experiment.

*This question has also been raised before (R1, point 5). According to an earlier study the mEos4b-GlyRb knock-in does not cause any obvious phenotypes, with the possible exception of minor loss of glycine potency (Maynard et al. 2021 eLife). The fact that the synaptic levels in the spinal cord in heterozygous animals are precisely half of those of homozygous animals argues against differences in receptor expression, heteropentameric assembly, forward trafficking to the plasma membrane and integration into the synaptic membrane as confirmed using quantitative super-resolution CLEM (Maynard et al. 2021 eLife). Accordingly, we did not observe any behavioural deficits in these animals, making it a powerful experimental model. We have added this information in the revised manuscript (p. 4). *

In addition, without any quantification or statistical analysis, the author’s claims regarding the necessity of GLRA1 expression for the synaptic localization of glycine receptors in cultured hippocampal neurons should probably be described as preliminary (Fig. 5).

As mentioned before, we have substantially revised this part (R1, point 3). The quantification and analysis in the new Fig. 5 support our earlier interpretation.

• Would additional experiments be essential to support the claims of the paper? Request additional experiments only where necessary for the paper as it is, and do not ask authors to open new lines of experimentation.

The authors show that there is colocalization of gephyrin with the mEos4b-GlyRβ subunit using the Dual-colour SMLM. This is a powerful approach that allows for a claim to be made on the synaptic location of the glycine receptors. The images presented in Figure 1, together with the distance analysis in Figure 2, display the co-localization of the fluorophores. The co-localization images in all the selected regions, hippocampus and striatum, also show detections outside of the gephyrin clusters, which the authors refer to as extrasynaptic. These punctated small clusters seem to have the same size as the ones detected and assigned as part of the synapse. It would be informative if the authors analysed the distribution, density and size of these non-synaptic clusters and presented the data in the manuscript and also compared it against the synaptic ones. Validating this extrasynaptic signal by staining for a dendritic marker, such as MAP-2 or maybe a somatic marker and assessing the co-localization with the non-synaptic clusters would also add even more credibility to them being extrasynaptic.

The existence of extrasynaptic GlyRs is well attested in spinal cord neurons (e.g. Specht et al. 2013 Neuron; this study see Fig. S2). The fact that these appear as small clusters of detections in SMLM recordings results from the fact that a single fluorophore can be detected several times in consecutive image frames and because of blinking. Therefore, small clusters of detections likely represent single GlyRs (that can be counted), and not assemblies of several receptor complexes. Due to their diffusion in the neuronal membrane, they are seen as diffuse signals throughout the somatodendritic compartment in epifluorescence images (e.g. Fig. 5A). SMLM recordings of the same cells resolves this diffuse signal into discrete nanoclusters representing individual receptors (Fig. 5B). It is not clear what information co-localisation experiments with specific markers could provide, especially in hippocampal neurons, in which the copy numbers (and density) of GlyRs is next to zero.

In addition we would encourage the authors to quantify the clustering and co-localization of virally expressed GLRA1, GLRA2, and GLRB with gephyrin in order to support the associated claims (Fig. 5). Preferably, the density of GLR and gephyrin clusters (at least on the somatic surface, the proximal dendrites, or both) as well as their co-localization probability should be quantified if a causal claim about subunit-specific requirements for synaptic localization is to be made.

Quantification of the data have been carried out (new Fig.5C,D). The results have been described before (R1, point 3) and support our earlier interpretation of the data (pp. 13-14).

Lastly, even though it may be outside of the scope of such a study analysing other parts of the hippocampal area could provide additional important information. If one looks at the Allen Institute’s ISH of the beta subunit the strongest signal comes from the stratum oriens in the CA1 for example, suggesting that interneurons residing there would more likely have a higher expression of the glycine receptors. This could also be assessed by looking more carefully at the single cell transcriptomics, to see which cell types in the hippocampus show the highest mRNA levels. If the authors think that this is too much additional work, then perhaps a mention of this in the discussion would be good.

We have added the requested information from the ISH database of the Allen Institute in the discussion as suggested by the reviewer (p. 12). However, in combination with the transcriptomic data (Fig. S1) our finding strongly suggest that the expression of synaptic GlyRs depends on the availability of alpha subunits rather than on the presence of the GlyRb transcript. This is obvious when one compares the mRNA levels in the hippocampus with those in the basal ganglia (striatum) and medulla. While the transcript concentrations of GlyRb are elevated in all three regions and essentially the same, our data show that the GlyRb copy numbers *at synapses differ over more than 2 orders of magnitude (Fig. 1B, Table 1). *

• Are the suggested experiments realistic in terms of time and resources? It would help if you could add an estimated cost and time investment for substantial experiments.

Since the labeling and some imaging has been performed already, the requested experiment would be a matter of deploying a method of quantification. In principle, it should not require any additional wet-lab experiments, although it may require additional imaging of existing samples.

• Are the data and the methods presented in such a way that they can be reproduced?

Yes, for the most part.

• Are the experiments adequately replicated and statistical analysis adequate?

Yes

Minor comments:

• Specific experimental issues that are easily addressable.

N/A

• Are prior studies referenced appropriately?

Yes

• Are the text and figures clear and accurate?

Yes, although quantification in figure 5 is currently not present.

A quantification has been added (see R1, point 3).

• Do you have suggestions that would help the authors improve the presentation of their data and conclusions?

This paper presents a method that could be used to localize receptors and perhaps other proteins that are in low abundance or for which a detailed quantification is necessary. I would therefore suggest that Figure S4 is included into Figure 2 as the first panel, showcasing the demixing, followed by the results.

We agree in principle with this suggestion. However, the revised Fig. S4 is more complex and we think that it would distract from the data shown in Fig. 2. Given that Fig. S4 is mostly methodological and not essential to understand the text, we have kept it in the supplement for the time being. We leave the final decision on this point to the editor.

__Reviewer #4-2 (Significance (Required)): __

[This review was supplied later]

• Describe the nature and significance of the advance (e.g. conceptual, technical, clinical) for the field.

Using a novel and high resolution method, the authors have provided strong evidence for the presence of glycine receptors in the murine hippocampus and in the dorsal striatum. The number of receptors calculated is small compared to the numbers found in the ventral striatum. This is the first study to quantify receptor numbers in these region. In addition it also lays a roadmap for future studies addressing similar questions.

• Place the work in the context of the existing literature (provide references, where appropriate).

This is done well by the authors in the curation of the literature. As stated above, the authors have filled a gap in the presence of glycine receptors in different brain regions, a subject of importance in understanding the role they play in brain activity and function.

• State what audience might be interested in and influenced by the reported findings.

Neuroscientists working at the synaptic level, on inhibitory neurotransmission and on fundamental mechanisms of expression of genes at low levels and their relationship to the presence of the protein would be interested. Furthermore, researchers in neuroscience and cell biology may benefit from and be inspired by the approach used in this manuscript, to potentially apply it to address their own aims.

*We thank the reviewer for the positive assessment of the technical and biological implications of our work, as well as the interest of our findings to a wide readership of neuroscientists and cell biologists. *

• Define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate.

Synaptic transmission, inhibitory cells and GABAergic synapses functionally and structurally, cortex and cortical circuits. No strong expertise in super-resolution imaging methods.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

In this manuscript, the authors investigate the nanoscopic distribution of glycine receptor subunits in the hippocampus, dorsal striatum, and ventral striatum of the mouse brain using single-molecule localization microscopy (SMLM). They demonstrate that only a small number of glycine receptors are localized at hippocampal inhibitory synapses. Using dual-color SMLM, they further show that clusters of glycine receptors are predominantly localized within gephyrin-positive synapses. A comparison between the dorsal and ventral striatum reveals that the ventral striatum contains approximately eight times more glycine receptors and this finding is consistent with electrophysiological data on postsynaptic inhibitory currents. Finally, using cultured hippocampal neurons, they examine the differential synaptic localization of glycine receptor subunits (α1, α2, and β). This study is significant as it provides insights into the nanoscopic localization patterns of glycine receptors in brain regions where this protein is expressed at low levels. Additionally, the study demonstrates the different localization patterns of GlyR in distinct striatal regions and its physiological relevance using SMLM and electrophysiological experiments. However, several concerns should be addressed.

The following are specific comments:

1. Colocalization analysis in Figure 1A. The colocalization between Sylite and mEos-GlyRβ appears to be quite low. It is essential to assess whether the observed colocalization is not due to random overlap. The authors should consider quantifying colocalization using statistical methods, such as a pixel shift analysis, to determine whether colocalization frequencies remain similar after artificially displacing one of the channels.
2. Inconsistency between Figure 3A and 3B. While Figure 3B indicates an ~8-fold difference in the number of mEos4b-GlyRβ detections per synapse between the dorsal and ventral striatum, Figure 3A does not appear to show a pronounced difference in the localization of mEos4b-GlyRβ on Sylite puncta between these two regions. If the images presented in Figure 3A are not representative, the authors should consider replacing them with more representative examples or providing an expanded images with multiple representative examples. Alternatively, if this inconsistency can be explained by differences in spot density within clusters, the authors should explain that.
3. Quantification in Figure 5. It is recommended that the authors provide quantitative data on cluster formation and colocalization with Sylite puncta in Figure 5 to support their qualitative observations.
4. Potential for pseudo replication. It's not clear whether they're performing stats tests across biological replica, images, or even synapses. They often quote mean +/- SEM with n = 1000s, and so does that mean they're doing tests on those 1000s? Need to clarify.
5. Does mEoS effect expression levels or function of the protein? Can't see any experiments done to confirm this. Could suggest WB on homogenate, or mass spec?
6. Quantification of protein numbers is challenging with SMLM. Issues include i) some of FP not correctly folded/mature, and ii) dependence of localisation rate on instrument, excitation/illumination intensities, and also the thresholds used in analysis. Can the authors compare with another protein that has known expression levels- e.g. PSD95? This is quite an ask, but if they could show copy number of something known to compare with, it would be useful.
7. Rationale for doing nanobody dSTORM not clear at all. They don't explain the reason for doing the dSTORM experiments. Why not just rely on PALM for coincidence measurements, rather than tagging mEoS with a nanobody, and then doing dSTORM with that? Can they explain? Is it to get extra localisations- i.e. multiple per nanobody? If so, localising same FP multiple times wouldn't improve resolution. Also, no controls for nanobody dSTORM experiments- what about non-spec nb, or use on WT sections?
8. What resolutions/precisions were obtained in SMLM experiments? Should perform Fourier Ring Correlation (FRC) on SR images to state resolutions obtained (particularly useful for when they're presenting distance histograms, as this will be dependent on resolution). Likewise for precision, what was mean precision? Can they show histograms of localisation precision.
9. Why were DBSCAN parameters selected? How can they rule out multiple localisations per fluor? If low copy numbers (<10), then why bother with DBSCAN? Could just measure distance to each one.
10. For microscopy experiment methods, state power densities, not % or "nominal power".
11. In general, not much data presented. Any SI file with extra images etc.?
12. Clarification of the discussion on GlyR expression and synaptic localization: The discussion on GlyR expression, complex formation, and synaptic localization is sometimes unclear, and needs terminological distinctions between "expression level", "complex formation" and "synaptic localization". For example, the authors state:"What then is the reason for the low protein expression of GlyRβ? One possibility is that the assembly of mature heteropentameric GlyR complexes depends critically on the expression of endogenous GlyR α subunits." Does this mean that GlyRβ proteins that fail to form complexes with GlyRα subunits are unstable and subject to rapid degradation? If so, the authors should clarify this point. The statement "This raises the interesting possibility that synaptic GlyRs may depend specifically on the concomitant expression of both α1 and β transcripts." suggests a dependency on α1 and β transcripts. However, is the authors' focus on synaptic localization or overall protein expression levels? If this means synaptic localization, it would be beneficial to state this explicitly to avoid confusion. To improve clarity, the authors should carefully distinguish between these different aspects of GlyR biology throughout the discussion. Additionally, a schematic diagram illustrating these processes would be highly beneficial for readers.
13. Interpretation of GlyR localization in the context of nanodomains. The distribution of GlyR molecules on inhibitory synapses appears to be non-homogeneous, instead forming nanoclusters or nanodomains, similar to many other synaptic proteins. It is important to interpret GlyR localization in the context of nanodomain organization.

Significance

The paper presents biological and technical advances. The biological insights revolve mostly on the documentation of Glycine receptors in particular synapses in forebrain, where they are typically expressed at very low levels. The authors provide compelling data indicating that the expression is of physiological significance. The authors have done a nice job of combining genetically-tagged mice with advanced microscopy methods to tackle the question of distributions of synaptic proteins. Overall these advances are more incremental than groundbreaking.

for - book - More Everything Forever - Adam Becker - from - Youtube - Essentia Foundation - interview - Alex Gomez-Marin - Neuroscientist speaks out on the hidden war on consciousness - https://hyp.is/ile8TIvJEfCl35MW3f5B8Q/www.youtube.com/watch?v=K7NIicE_h9w

Summary - Interesting adjacency with another video I've been watching, that focused on a Western monk's practice of Tibetan Buddhism, who after 12 years, entered a 4 year retreat and panicked - His demons emerged in the first 2 years of the retreat and he left but returned - This monk emphasized accepting the relationship with his demons instead of averting them and how craving and desire emphasized by Western civilllization is the cause of modernity's meaning crisis - to - Youtube - Diary of a CEO - Your brain is lying to you - Interview - Gerong Tupton - https://hyp.is/go?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DvIbLQQ1i56Y&group=world

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

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2025-03111

Corresponding author(s): Qingyin Qian and Ryusuke Niwa

1. General Statements [optional]

We would like to thank reviewers for their feedback on our initial submission. Changes in figures were noted in the point-to-point reply. For submission of our current revised manuscript, we provide two Word files, which are the “clean” and “Track-and-Change” files. Page and line numbers described below correspond to those of the “clean” file. The “Track-and-Change” file might be helpful for Reviewers to find what we have changed for the current revision.

In the revised manuscript, major changes in the text were tracked, while minor edits in figure numbers and legends were not tracked. In the Discussion, the section “Xrp1-mediated EE plasticity…” was moved before “Xrp1, a transcription factor …”, to follow the order of the Results, and was split into two: “EE plasticity …” and “Xrp1-mediated EE plasticity …”.

2. Description of the planned revisions

- The authors should investigate the regenerative growth of the adult midgut after irradiation. Is there an impact on ISCs proliferation or cell turn over. Is Xrp1 in EEs required in this adaptive response. It would be elegant to use the recently generated tracing method by Tobias Reiff lab to observe overall impact on tissue renewal (rapport-tracing esglexReDDM esg-lexA, 13xLexAop2-CD8::GFP, 13xLexAop2-H2B::mCherry::HA, tub-Gal80ts on the second chromosome. It can be combined with any EEs Gal4-driver (see Nat Commun 2025, https://doi.org/10.1038/s41467-024-55664-2, the stock is already existing, see table1). This reviewer thinks that it is a key experiment to support the proposed model.

2.1. Author response:

We will conduct the following experiments to answer these criticisms.

(1) We will investigate the ISC behavior, proliferation and differentiation, after 100 Gy of radiation by examining changes in the number of progenitor cells and their progenies, using esgtsF/O (esg-Gal4, UAS-GFP, tub-Gal80ts; Act>Cd2>Gal4, UAS-Flp) generated in the study (Jiang et al. Cell 2009 DOI: 10.1016/j.cell.2009.05.014) or esgReDDM (esg-Gal4, UAS-CD8::GFP; UAS-H2B::RFP, tubGal80ts) generated in the study (Antonello et al. EMBO J. 2015 DOI: 10.15252/embj.201591517). Flies will have progenitor cell lineages traced for 7 days, irradiated on day 6, and examined at different time points after radiation, following the design shown in Fig. 2A. Based on the previous findings (Sharma et al. Sci. Rep. 2020 DOI: 10.1038/s41598-020-75867-z; Pyo et al. Radiat. Res. 2014 DOI: 10.1667/RR13545.1), we anticipate that radiation compromises ISCs’ proliferation and differentiation. Should this be the case, our results can be interpreted in relation to those earlier studies.

(2) In parallel, we will examine whether Xrp1 expression in EEs affects radiation-induced ISC behaviors. As suggested, we will use “EE Rapport” (esg-lexA, 13xLexAop2-CD8::GFP, 13xLexAop2-H2B::mCherry::HA, tub-Gal80ts; Rab3-Gal4) generated in the study (Zipper et al. Nat. Commun. 2025 DOI: 10.1038/s41467-024-55664-2) and compare control flies to flies with Xrp1 knocked down in EEs to assess the impact on ISC behaviors.

- Is p53 required for Xrp1 induction in the gut after irradiation?

2.2. Author response:

To answer this point, we will perform immunostaining of anti-Xrp1 antibody to examine whether p53 is required for Xrp1 induction in irradiated flies with p53 knocked down in EEs.

- Xrp1 over expression has been shown to induce upd3 ligand and nutrient-driven dedifferentiation of enteroendocrine cells is occuring by activation of the JAK-STAT pathway (DOI: 10.1016/j.devcel.2023.08.022). Could the authors test the function of this signaling pathway during irradiation (upd3-lacZ and Stat-GFP can be used in parallel of upd3 RNAi and UAS Dome-DN.

2.3. Author response:

We will conduct the following experiments to answer these points.

(1) We will examine the cell type in which upd3 ligand induction occurs after radiation by using the upd3.1-LacZ reporter generated in the study (Jiang et al. Cell Stem Cell 2011 DOI: doi.org/10.1016/j.stem.2010.11.026).

(2) One possibility is that upd3.1-LacZ is detected in EEs. In this case, we will examine the requirement of upd3 in EEs for radiation-induced EE plasticity by knocking down upd3. Another possibility is that upd3.1-LacZ is detected in non-EE cells. If so, we will examine the requirement of the JAK-STAT pathway in EEs by overexpressing dome[△cyt] generated in the study (Brown et al. Curr. Biol. 2001 DOI: 10.1016/s0960-9822(01)00524-3) or knocking down Stat92E in EEs. Because these conditions are not mutually exclusive, both approaches may be pursued, with the latter relating our results to nutrient-driven EE dedifferentiation.

- Xrp1 is known for its role in cell competition and elimination of looser cells by induction of apoptosis. It would be interesting to check for induction of cell death and/or caspase activation in the fly gut after irradiation and verify a non apoptotic role of DRONC activation in this context using a Dronc RNAi (as proposed by Bergmann lab (https://doi.org/10.1038/s41598-021-81261-0) or Baena-Lopez lab (DOI: 10.15252/embr.201948892)). Overexpression of Xrp1 could be combined with UAS-p35.

2.4. Author response:

To address these points, we will investigate apoptosis induction following radiation with anti-cleaved Dcp-1 immunostaining. Based on the previous finding (Sharma et al. Sci. Rep. 2020 DOI: 10.1038/s41598-020-75867-z), we anticipate seeing increased cleaved Dcp-1 signals in all cell types after radiation. We intend to clarify whether radiation increases the ratio of apoptotic EEs among EEs; however, we cannot yet be certain whether it will be feasible.

Regarding Dronc activation, we previously requested the antibody used in the study (Wilson et al. Nat. Cell Biol. 2002 DOI: 10.1038/ncb799; Lindblad et al. Sci. Rep. 2021 DOI: 10.1038/s41598-021-81261-0) and tested it in our context, after radiation and by Xrp1-S O/E in EEs. We present our data below. In the anterior midgut, anti-Dronc signals were not observed under both control conditions. After radiation and by Xrp1-S O/E in EEs, anti-Dronc signals were seen in part of past EEs (#2 past) and progenitor cells (#3 prgn), implying their EB identity. However, anti-Dronc signals were never observed in current EEs (#1 current), suggesting Dronc does not act directly downstream to Xrp1.

We will address UAS-p35 in 3.3. Author response and Dronc-RNAi in 4.2. Author response.

- The authors do not justify or explain why they used 100 Gy of radiation. This is higher than doses used in comparable regeneration studies in adult Drosophila (e.g., PMID25959206, PMID: 28925355). The authors should clarify why this dose was chosen.

2.5. Author response:

Our initial rationale was based on the paper (Sharma et al. Sci. Rep. 2020 DOI: 10.1038/s41598-020-75867-z), where the authors claimed that ISC proliferation was inhibited and the ISC number was decreased by 100 Gy of radiation.

Nevertheless, we understand the reviewer’s concern and will examine 50 Gy of radiation as used in the papers the reviewer listed. We will examine radiation-induced changes in EE lineages and ISC behaviors. Depending on the results, we will evaluate whether and how they should be incorporated into the manuscript.

- Fig. 2C, the number of past EE’s increased transiently so that baseline number is restored at 18 hr after IR. The authors conclude that fate plasticity is a transient event. Can they rule out loss due to cell death?

2.6. Author response:

In our system, past EEs were detected transiently but did not persist. We agree that we cannot distinguish whether the transient appearance of past EEs reflects transient adoption of another identity that ends in cell death or reversible plasticity.

To partially address this criticism, as noted in 2.4. Author response, we will examine the apoptosis marker cleaved Dcp-1, which also tests whether cleaved Dcp-1-positive cells can be past EEs. However, regardless of detecting apoptosis markers in past EEs, we have changed “transient” into “temporary” to describe a short-lived cell state (see Page 8, Line 178; Page 15, Line 338).

- They authors interpret fate-conversion as beneficial for tissue repair but never test whether blocking this process impairs recovery or organismal survival or whether promoting it improves outcomes.

2.7. Author response:

We have removed this potentially misleading interpretation (see Page 4, removed the last part of the previous introduction, “and propose the possibility that such plasticity contributes to tissue repair”). We present below the data showing a severe reduction of the ISC number in 7-day post-radiation guts, suggesting the inability of tissue repair. We will add this to the manuscript together with results from the following experiments.

(1) We will examine if the blockage of radiation-induced EE plasticity, via knocking down Xrp1 in EEs, alters the epithelial cell number and cell junction protein localization.

(2) To complement the result of plasticity inhibition, we attempt to promote plasticity by overexpressing Xrp1 in EEs, to test whether this rescues ISC loss or restores junctions.

Should knockdown worsen ISC loss and junction integrity, or overexpression rescue them, we will describe EE plasticity as beneficial; otherwise, we will present it as a radiation-induced response without inferring benefits, while noting our limitations.

We will address organismal survival in 4.3. Author response.

- Related to the above, it would be helpful to know if fate-converted cells function as true ISCs or ECs (e.g., through proliferation or absorption assays).

2.8. Author response:

To partially answer this criticism, we will examine whether EE-derived ISCs are proliferative by examining whether they can be positive for the mitotic marker phospho-histone 3.

We will address absorption assays in 4.4. Author response.

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

- It is surprising to observe EEs dedifferentiation at a steady state during homeostasis, a condition in which Xrp1 is not detected in the gut. Can the authors comment this point in the discussion?

3.1. Author response:

We have added our thoughts in terms of Xrp1 being not detectable in homeostatic EE lineages (see Page 15, Line 350 - 356). We have also added our thoughts regarding observation of EE plasticity in homeostatic guts (see Page 14, Line 322 - 332).

- Xrp1 is existing as a short of long isoforms. The short form has been recently proposed to be required for cell competition (https://doi.org/10.1101/2025.06.15.659587) whereas Xrp1 long isoform may be responsible for reduced cell growth. Could the authors test which isoform is induced in the gut after irradiation? Is the overexpression of Xrp1 long isoform having the same effect that the short isoform used by the authors.

3.2. Author response:

We have added data on the effect of Xrp1 long isoform overexpression on EE plasticity (see Fig. 5A - 5B, Page 12, Line 276 - 278), showing that overexpression of the Xrp1 long isoform caused a similar increase in past EEs. In addition, we have changed Xrp1 O/E to Xrp1-S O/E in the contents related to Figs 4, 5, S4, and S5.

We will address radiation-induced Xrp1 isoforms in 4.1. Author response.

- Xrp1 is known for its role in cell competition and elimination of looser cells by induction of apoptosis. It would be interesting to check for induction of cell death and/or caspase activation in the fly gut after irradiation and verify a non apoptotic role of DRONC activation in this context using a Dronc RNAi (as proposed by Bergmann lab (https://doi.org/10.1038/s41598-021-81261-0) or Baena-Lopez lab (DOI: 10.15252/embr.201948892)). Overexpression of Xrp1 could be combined with UAS-p35.

3.3. Author response:

We have added data regarding p35 O/E combined with Xrp1 O/E, showing that p35 O/E did not further increase the number of past EEs, thereby suggesting that Xrp1-driven EE plasticity has a non-apoptotic nature (see Fig. 5C - 5D, Page 13, Line 293 - 297).

- Line 221: fig S3E should be S3F

- Line 230: fig S3F-G should be S3G-H

• • Line 230, Fig S3F-G should be Fig S3G-H.*

3.4. Author response:

We have fixed this error.

- The posterior gut region R4 is more proliferative than the anterior part and is usually used for testing regenerative growth. What is happening there after irradiation?

3.5. Author response:

We present below radiation-induced changes in EE lineages and ISC number in the R4bc gut region. Radiation did not alter the proportion of past EEs among EE lineages but reduced the ISC number. We acknowledge differences between anterior and posterior gut regions, but we do not plan to further analyze regional differences or underlying mechanisms.

- The authors’ explanation for cells with weak GFP in Figure 1 is not convincing. Induction of GFP is an all or nothing event as it results from Pros-driven FLPase and a recombination that removes the transcription stop signals to express GFP from a Ubi promotor. Once that happens, it should not matter how strong or weak Pros is, GFP should be the same. So, another explanation is needed. Nuclear staining of cell #2 in Fig 1B resembles a metaphase chromosome arrangement. Nuclear GFP may appear ‘weak’ in mitosis as the nuclear envelope breaks down. It is positive for the purple Pros/Dl stain, which makes it hard to tell if it is Pros+ or Pros- even though the authors state that cells with weak GFP are Pros- in line 104 (see the point above regarding confusing same-color stain for ISC and EE markers). Could cell #2 be a pre-EE that is undergoing mitosis since the lineage tracer marks both EE and pre-EE cells (line 119)? Or do the authors mean recombination on one or both homologs? This should not be possible since the cells are heterozygotes for the Ubi-GFP locus.

3.6. Author response:

For cell #5, RFP- GFPweak may result from the leakiness of the G-TRACE system. We have added our observations of the G-TRACE strains and changed our previous explanation (see Fig. S1B - S1C, Page 5, Line 94 - 97, 103 - 106).

For cell #2, we agree that RFP+ GFPweak cells may either be a cell turning on pros expression just before sample preparation or a pre-EE undergoing mitosis. Nevertheless, it is not a past EE that has lost the EE marker Pros, so it is considered a current EE. We have removed our previous interpretation of cell #2 (see Page 5, removed “which likely had not yet fully activated recombination”), and changed the image to avoid confusion (see Fig. 1C).

- Fig. 2C, if past-EE’s increased in number while current EE’s stayed the same, where are new past-EE’s coming from? There cannot be compensatory proliferations since EE’s are post-mitotic. For fate conversion, one would expect the generation of each past-EE to accompany loss of one current EE.

3.7. Author response:

We agree that the generation of one past EE should be accompanied by the loss of one current EE. We do not have a clear answer to this question. Our data showed cell numbers per ROI rather than the total cell number across the whole gut. To address this, we have changed the number to the proportion, calculated from [past EE] / ([past EE] + [current EE]), in experiments examining damage-induced EE plasticity, which provides a more informative measure for EE fate conversion (see Fig. 2C, also Fig. S2B and 3E).

- Fig. 2E. Dl+ past-EE cell number declined at 14 and 18 h after IR and because cell sized increased, the authors conclude that EE cells that de-differentiated into ISCs subsequently re-differentiated into EC’s. To reach this conclusion, the authors should count past-EEs that are positive for EC markers. Cell size alone is insufficient evidence.

3.8. Author response:

We have added data quantifying the proportion of past EEs that are positive for the EC marker Pdm1, showing that past EEs were more likely to be ECs in guts examined 14 h after radiation (see Fig. 2F - 2G, Page 9, Line 189).

- Fig. 6. Where are the % numbers for ISC, EB and EE’s coming from? And wouldn’t these change with time after IR, etc?

3.9. Author response:

The numbers came from the calculation of the percentage of the absolute values of control and 14 h post-IR conditions from Fig. 2E. These numbers changed with time after radiation. We realized that the precise numbers were misleading. We therefore have removed such illustration and instead added phrases “more current EEs → past EEs, more past EEs being ISCs → past EEs being ECs” to describe the increase in past EE cell number and the shift in the composition of past EEs (see Fig. 6).

- Improve Figure 1B: Pros and Dl are shown in the same color, creating confusion. If both are stained together, different colors or clearer labeling should be used. Clarify how cells are identified as Pros+ vs Dl+.

3.10. Author response:

Anti-Pros and anti-Dl antibodies were produced from the same host species and were detected with the same secondary antibody, so they were in the same color. We have stated that solid nuclear staining indicates Pros, whereas punctate cytoplasmic staining indicates Dl (see Page 5, Line 100, 102, and 103). Such staining has been reported in previous studies (for example, Fig. 2A - 2B, Veneti et al. Nat. Commun. 2024 DOI: 10.1038/s41467-024-46119-9).

- Why is Dl (supposed to be cytoplasmic) overlapping with nuclear GFP in cells #3 and 4 in Fig. 1B?

3.11. Author response:

Because Dl signals were located apically to DAPI/GFP signals, the overlap was likely due to Z-projection from stacked slices. We present below orthogonal slices along the z-axis, from top to bottom by row, and composite and individual color channels, from left to right by columns, for cell #3 (left) and cell #4 (right).

For cell #3, Dl signals were present in slices 1/8 and 2/8 and disappeared in slice 3/8, whereas DAPI signals appeared from slice 2/8. For cell #4, Dl signals surrounded DAPI signals when viewed separately. In addition, we realized that nuclear GFP signals slightly outgrew DAPI signals, despite our confirmation that the GFP channel was not saturated.

We have included separate color channels for DAPI signals and Pros, Dl and DAPI merged channels, showing that Dl signals were absent from the nucleus. For cell #3, in which the nuclear DAPI and cytoplasmic Dl cannot be distinguished in the stacked view, we show the images from a single orthogonal slice in the main panel, and the image from stacked slices as insets (see Fig. 1C).

- Fig. S1E and F. Very hard to see what the authors describe about Arm and Cora. One problem is that cell boundaries are not visible, just the nuclei, so it is hard to know whether cell-cell interactions the authors describe as normal are really normal. Another problem is the overlap of Arm (supposed to be cytoplasmic) with the nuclear GFP signal. What is that?

3.12. Author response:

Regarding the invisibility of cell boundaries, we have improved the image of anti-Cora staining and added anti-Mesh staining and a separate color channel for DAPI signals to reinforce junction integrity (see Fig. S1H - S1I).

Regarding the overlap of Arm signals with nuclear GFP signals, we realized similar problems as those noted in 3.11. Author response. We present below orthogonal slices along the z-axis and combined and individual color channels, for cell #2 (left) and cell #3 (right). For both cells, Arm signals did not overlap with DAPI signals. We have adjusted the maximum intensity projection to include slices 1-4 instead of 1-8 and added a separate color channel for DAPI signals to avoid the signals appearing to overlap (see Fig. S1G).

- Include a simple schematic of ISC to EE/EC lineages for readers unfamiliar with Drosophila gut biology.

3.13. Author response:

We have included a schematic (see Fig. 1A). Although not requested, we have also improved Fig. 1B to enhance clarity.

- Discuss the regional difference in Xrp1 efficacy (R2a vs R2b). Is there something known about gene expression differences in different gut regions that can explain the results?

3.14. Author response:

At present, we do not have an explanation for these results. We have refined our discussion regarding such regional differences (see Page 16 - 17, Line 381 - 390).

- Consider moving scRNAseq (Fig. S1G) into main paper: this is a central part of the conclusion.

3.15. Author response:

We have moved Fig. S1G, as well as Fig. S1H and S1I, into the main figure (see Fig. 1G - 1I).

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

- Xrp1 is existing as a short of long isoforms. The short form has been recently proposed to be required for cell competition (https://doi.org/10.1101/2025.06.15.659587) whereas Xrp1 long isoform may be responsible for reduced cell growth. Could the authors test which isoform is induced in the gut after irradiation? Is the overexpression of Xrp1 long isoform having the same effect that the short isoform used by the authors.

4.1. Author response:

We prefer not to distinguish whether the long or short Xrp1 isoform is induced in the gut after radiation. This presents technical challenges and falls outside the scope of the present study. As noted in 3.2. Author response, we instead report in the revised manuscript that both isoforms similarly promote EE plasticity.

- Xrp1 is known for its role in cell competition and elimination of looser cells by induction of apoptosis. It would be interesting to check for induction of cell death and/or caspase activation in the fly gut after irradiation and verify a non apoptotic role of DRONC activation in this context using a Dronc RNAi (as proposed by Bergmann lab (https://doi.org/10.1038/s41598-021-81261-0) or Baena-Lopez lab (DOI: 10.15252/embr.201948892)). Overexpression of Xrp1 could be combined with UAS-p35.

4.2. Author response:

We prefer not to perform Dronc-RNAi, because we did not observe Dronc activation downstream to Xrp1, as shown in 2.4. Author response.

- They authors interpret fate-conversion as beneficial for tissue repair but never test whether blocking this process impairs recovery or organismal survival or whether promoting it improves outcomes.

4.3. Author response:

We prefer not to examine organismal survival. We agree that organismal survival would be informative, but our study focuses on epithelial cell number, which will be tested as noted in 2.7. Author response. We will not mention broad claims at the organismal level.

- Related to the above, it would be helpful to know if fate-converted cells function as true ISCs or ECs (e.g., through proliferation or absorption assays).

4.4. Author response:

We prefer not to perform absorptive assays due to technical challenges. We will instead test proliferation, as noted in 2.8. Author response, and note our limitations.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Adult tissue homeostasis refers to the process by which tissues maintain a stable and functional state over time. This usually depends on stem cell activity and the balance between cell proliferation and differentiation to ensure that tissues can repair damage, replace old or dead cells, and maintain their structure and function.

Damage-induced plasticity plays an important role in restoring tissue homeostasis. Cellular plasticity is the ability of differentiated cells to acquire alternative phenotypic identities. It is typically constrained under homeostatic conditions but can be activated in response to tissue damage to support regeneration. In this study entitled "Xrp1 drives damage-induced cellular plasticity of enteroendocrine cells in the adult Drosophila midgut", Qian Q. et al., describe damage-induced plasticity of secretory enteroendocrine cells (EEs) in the adult Drosophila midgut. They found that ionizing radiation enhances EE plasticity, enabling EEs to dedifferentiate into intestinal stem cells (ISCs), which subsequently re-differentiate into absorptive enterocytes (ECs). Mechanistically, radiation triggers the expression of Xrp1, a stress-responsive transcription factor, within EE lineages. Xrp1 upregulation is necessary for initiating EE plasticity by expressing progenitor specific genes (like escargot for example), as verified by single-cell RNA sequencing of midguts with EE-specific Xrp1 overexpression. This is suggesting that Xrp1 reprograms EEs by promoting progenitor-like transcriptional states.

The authors nicely describe the dedifferentiation of EEs using the G-TRACE system in response to irradiation and the role of Xrp1 in this process. Yet, the authors need to show the requirement of the EEs dedifferenciation during regenerative growth.

Major comments:

• The authors should investigate the regenerative growth of the adult midgut after irradiation. Is there an impact on ISCs proliferation or cell turn over. Is Xrp1 in EEs required in this adaptive response. It would be elegant to use the recently generated tracing method by Tobias Reiff lab to observe overall impact on tissue renewal (rapport-tracing esglexReDDM esg-lexA, 13xLexAop2-CD8::GFP, 13xLexAop2-H2B::mCherry::HA, tub-Gal80ts on the second chromosome. It can be combined with any EEs Gal4-driver (see Nat Commun 2025, https://doi.org/10.1038/s41467-024-55664-2, the stock is already existing, see table1). This reviewer thinks that it is a key experiment to support the proposed model.
• It is surprising to observe EEs dedifferentiation at a steady state during homeostasis, a condition in which Xrp1 is not detected in the gut. Can the authors comment this point in the discussion?

Minor comments:

• Is p53 required for Xrp1 induction in the gut after irradiation?
• Xrp1 is existing as a short of long isoforms. The short form has been recently proposed to be required for cell competition (https://doi.org/10.1101/2025.06.15.659587) whereas Xrp1 long isoform may be responsible for reduced cell growth. Could the authors test which isoform is induced in the gut after irradiation? Is the overexpression of Xrp1 long isoform having the same effect that the short isoform used by the authors.
• Xrp1 over expression has been shown to induce upd3 ligand and nutrient-driven dedifferentiation of enteroendocrine cells is occuring by activation of the JAK-STAT pathway (DOI: 10.1016/j.devcel.2023.08.022). Could the authors test the function of this signaling pathway during irradiation (upd3-lacZ and Stat-GFP can be used in parallel of upd3 RNAi and UAS Dome-DN.
• Xrp1 is known for its role in cell competition and elimination of looser cells by induction of apoptosis. It would be interesting to check for induction of cell death and/or caspase activation in the fly gut after irradiation and verify a non apoptotic role of DRONC activation in this context using a Dronc RNAi (as proposed by Bergmann lab (https://doi.org/10.1038/s41598-021-81261-0) or Baena-Lopez lab (DOI: 10.15252/embr.201948892)). Overexpression of Xrp1 could be combined with UAS-p35.
• Line 221: fig S3E should be S3F
• Line 230: fig S3F-G should be S3G-H
• The posterior gut region R4 is more proliferative than the anterior part and is usually used for testing regenerative growth. What is happening there after irradiation?

Significance

Altogether, the paper present compiling lines of evidence supporting the proposed model. The experiments are well designed and are convincing. The papers is interesting and relevant for a broad audience.

DOI: 10.7554/eLife.46315

Resource: RRID:BDSC_35155

Curator: @bdscstockkeepers

SciCrunch record: RRID:BDSC_35155

What is this?

DOI: 10.7150/thno.113382

Resource: Addgene (RRID:SCR_002037)

Curator: @olekpark

SciCrunch record: RRID:SCR_002037

What is this?

DOI: 10.1101/gr.246884.118

Resource: RRID:BDSC_28218

Curator: @bdscstockkeepers

SciCrunch record: RRID:BDSC_28218

What is this?

DOI: 10.1002/advs.202415181

Resource: RRID:Addgene_48138

Curator: @olekpark

SciCrunch record: RRID:Addgene_48138

What is this?

DOI: 10.7554/eLife.45631

Resource: (Sevrioukov EA; Mol Cell. 1999 Cat# car, RRID:AB_2569524)

Curator: @scibot

SciCrunch record: RRID:AB_2569524

What is this?

DOI: 10.4103/NRR.NRR-D-23-01953

Resource: (ABclonal Cat# A11250, RRID:AB_2758477)

Curator: @scibot

SciCrunch record: RRID:AB_2758477

What is this?

DOI: 10.3389/fimmu.2025.1638500

Resource: (BioLegend Cat# 101804, RRID:AB_312837)

Curator: @scibot

SciCrunch record: RRID:AB_312837

What is this?