eLife Assessment

This fundamental study measures the functional specialization of distinct subregions within the mouse posterior parietal cortex (PPC) using mesoscopic two-photon calcium imaging during visual discrimination and choice history-dependent tasks. It presents compelling evidence supporting the existence of functional specialized subregions within the PPC. The work will be of interest to system and computational neuroscientists interested in decision-making, working memory, and multisensory integration.

Reviewer #1 (Public review):

Summary:

This study examined the functional organization of the mouse posterior parietal cortex (PPC) using meso-scale two-photon calcium imaging during visually-guided and history-guided tasks. The researchers found distinct functional modules within the medial PPC: area A, which integrates somatosensory and choice information, and area AM, which integrates visual and choice information. Area A also showed a robust representation of choice history and posture. The study further revealed distinct patterns of inter-area correlations for A and AM, suggesting different roles in cortical communication. These findings shed light on the functional architecture of the mouse PPC and its involvement in various sensorimotor and cognitive functions.

Strengths:

Overall, I find this manuscript excellent. It is very clearly written and built up logically. The subject is important, and the data supports the conclusions without overstating implications. Where the manuscript shines the most is the exceptionally thorough analysis of the data. The authors set a high bar for identifying the boundaries of the PPC subareas, where they combine both somatosensory and visual intrinsic imaging. There are many things to compliment the authors on, but one thing that should be applauded in particular is the analysis of the body movements of the mice in the tube. Anyone working with head-fixed mice knows that mice don't sit still but that almost invariable remains unanalyzed. Here the authors show that this indeed explained some of the variance in the data.

Comments on revisions:

I only had minor comments on the first version of the manuscript and these concerns were fully addressed after revision.

Reviewer #2 (Public review):

Summary:

The posterior parietal cortex (PPC) has been identified as an integrator of multiple sensory streams and guides decision making. Hira et al observe that dissection of the functional specialization of PPC subregions requires simultaneous measurement of neuronal activity throughout these areas. To this end, they use widefield calcium imaging to capture the activity of thousands of neurons across the PPC and surrounding areas. They begin by delineating the boundaries between the primary sensory and higher visual areas using intrinsic imaging and validate their mapping using calcium imaging. They then conduct imaging during a visually guided task to identify neurons that respond selectively to visual stimuli or choice. They find that vision and choice neurons intermingle primarily in the anterior medial (AM) area, and that AM uniquely encodes information regarding both the visual stimulus and the previous choice, positioning AM as the main site of integration of behavioral and visual information for this task.

Strengths:

There is an enormous amount of data and results reveal very interesting relationships between stimulus and choice coding across areas and how network dynamics relate to task coding.

Weaknesses:

The enormity of the data and the complexity of the analysis makes the manuscript hard to follow. Sometimes it reads like a laundry list of results as opposed a cohesive story.

Comments on revisions:

The authors have addressed our concerns.

Reviewer #3 (Public review):

Summary:

This work from Hira et al leverages mesoscopic 2-photon imaging to study large neural populations in different higher visual areas, in particular areas A and AM of the parietal cortex. The focus of the study is to obtain a better understanding of the representation of different task-related parameters, such as choice formation and short-term history, as well as visual responses in large neural populations across different cortical regions to obtain a better understanding of the functional specialization of neural populations in each region as well as the interaction of neural populations across regions. The authors image a large number of neurons in animals that either perform a visual discrimination or a history-dependent task to test how task demands affect neural responses and population dynamics. Furthermore, by including a behavioral perturbation of animal posture they aim to dissociate the neural representation of history signals from body posture. Lastly, they relate their functional findings to anatomical data from the Allen connectivity atlas and show a strong relation of functional correlations on anatomical connectivity patterns.

Strengths:

Overall, the study is very well done and tackles a problem that should be of high interest to the field by aiming to obtain a better understanding of the function and spatial structure of different regions in the parietal cortex. The experimental approach and analyses are sound and of high quality and the main conclusions are well supported by the results. Aside from the detailed analyses, a particular strength is the additional experimental perturbation of posture to isolate history-related activity which supports the conclusion that both posture and history signals are represented in different neurons within the same region.

Weaknesses:

The work does not focus on functional overlap at the single-cell level but on the spatial distribution of functional classes across areas. A minor weakness is therefore that it does not explicitly address how the finding of functional clusters relate to established notions of mixed selectivity within PPC.

Author response:

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

We sincerely appreciate your constructive feedback. Based on the comments from the three reviewers, we were able to substantially improve the manuscript. Below, we provide our point-by-point responses.

Public Reviews:

Reviewer #1 (Public review):

Summary:

This study examined the functional organization of the mouse posterior parietal cortex (PPC) using meso-scale two-photon calcium imaging during visually-guided and history-guided tasks. The researchers found distinct functional modules within the medial PPC: area A, which integrates somatosensory and choice information, and area AM, which integrates visual and choice information. Area A also showed a robust representation of choice history and posture. The study further revealed distinct patterns of inter-area correlations for A and AM, suggesting different roles in cortical communication. These findings shed light on the functional architecture of the mouse PPC and its involvement in various sensorimotor and cognitive functions.

Strengths:

Overall, I find this manuscript excellent. It is very clearly written and built up logically. The subject is important, and the data supports the conclusions without overstating implications. Where the manuscript shines the most is the exceptionally thorough analysis of the data. The authors set a high bar for identifying the boundaries of the PPC subareas, where they combine both somatosensory and visual intrinsic imaging. There are many things to compliment the authors on, but one thing that should be applauded in particular is the analysis of the body movements of the mice in the tube. Anyone working with head-fixed mice knows that mice don't sit still but that almost invariable remains unanalyzed. Here the authors show that this indeed explained some of the variance in the data.

Weaknesses:

I see no major weaknesses and I only have minor comments.

Reviewer #2 (Public review):

Summary:

The posterior parietal cortex (PPC) has been identified as an integrator of multiple sensory streams and guides decision-making. Hira et al observe that dissection of the functional specialization of PPC subregions requires simultaneous measurement of neuronal activity throughout these areas. To this end, they use wide-field calcium imaging to capture the activity of thousands of neurons across the PPC and surrounding areas. They begin by delineating the boundaries between the primary sensory and higher visual areas using intrinsic imaging and validate their mapping using calcium imaging. They then conduct imaging during a visually guided task to identify neurons that respond selectively to visual stimuli or choices. They find that vision and choice neurons intermingle primarily in the anterior medial (AM) area, and that AM uniquely encodes information regarding both the visual stimulus and the previous choice, positioning AM as the main site of integration of behavioral and visual information for this task.

Strengths:

There is an enormous amount of data and results reveal very interesting relationships between stimulus and choice coding across areas and how network dynamics relate to task coding.

Weaknesses:

The enormity of the data and the complexity of the analysis make the manuscript hard to follow. Sometimes it reads like a laundry list of results as opposed to a cohesive story.

Reviewer #3 (Public review):

Summary: This work from Hira et al leverages mesoscopic 2-photon imaging to study large neural populations in different higher visual areas, in particular areas A and AM of the parietal cortex. The focus of the study is to obtain a better understanding of the representation of different task-related parameters, such as choice formation and short-term history, as well as visual responses in large neural populations across different cortical regions to obtain a better understanding of the functional specialization of neural populations in each region as well as the interaction of neural populations across regions. The authors image a large number of neurons in animals that either perform visual discrimination or a history-dependent task to test how task demands affect neural responses and population dynamics. Furthermore, by including a behavioral perturbation of animal posture they aim to dissociate the neural representation of history signals from body posture. Lastly, they relate their functional findings to anatomical data from the Allen connectivity atlas and show a strong relation between functional correlations on anatomical connectivity patterns.

Strengths:

Overall, the study is very well done and tackles a problem that should be of high interest to the field by aiming to obtain a better understanding of the function and spatial structure of different regions in the parietal cortex. The experimental approach and analyses are sound and of high quality and the main conclusions are well supported by the results. Aside from the detailed analyses, a particular strength is the additional experimental perturbation of posture to isolate history-related activity which supports the conclusion that both posture and history signals are represented in different neurons within the same region. Weaknesses: The main point that I found hard to understand was the fairly strong language on functional clusters of neurons while also stating that neurons encoded combinations of different types of information and leveraging the encoding model to dissociate these contributions. Do the authors find mixed selectivity or rather functional segregation of neural tuning in their data? More details on this and some other points are below.

We thank the three reviewers for their accurate and expert evaluations.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) It wasn't clear to me why the authors focused on areas A and AM, but not RL. After all, at the beginning of the results, the authors ask: "PPC has been reported to have functions including visually guided decision-making and working memory. Do these functions differ among RL, A, and AM?".

Thank you for the comment. The manuscript first characterizes AM as a region involved in visually guided decision-making and A as a region related to history and/or working memory. Subsequently, when discussing correlation structure, we stated the following:

“In particular, based on the critical functional differences between A and AM that we found, A and AM may belong to distinct cortical networks that consist of different sets of densely interacting cortical areas.”

Thus, the logical flow of our analysis is to first reveal the functional contrast between A and AM through comparative functional analyses across RL, A, and AM, and then to focus on this contrast. We speculate that RL may exhibit more distinctive functional properties in tasks that rely on whisker-based processing or related modalities. We have therefore revised the text as described below to avoid the impression that the manuscript places disproportionate emphasis on RL.

Line 137: “PPC has been reported to have functions including visually guided decisionmaking and working memory. Do these functions differ among A, AM, and RL?”

(2) Figures 2 E, F, and Figure 3A, could the authors indicate the trial structure better on these plots?

Thank you for the comment. We have added explanations of the bar meanings to the figure legends.

Figure 2:

“(E) Representative vision neurons (ROI 1-4 in I). The red bars indicate sampling periods during video presentation, and the brown bars indicate sampling periods without video stimulation. Vertical black lines mark the onset of the sampling period. F. Representative choice neuron (ROI 5-8 in I) and a non-selective neuron (ROI 9). Light blue lines indicate the response periods in trials with left choices, and purple lines indicate the response periods in trials with right choices. Vertical black lines mark the onset of the response period.”

Figure 3:

“(A) The representative history neurons. Numbers correspond to that of panel B and C. Light blue lines indicate rewards delivered from the left lick port, and purple lines indicate rewards delivered from the right lick port. Vertical white lines mark the onset of the sampling period.”

(3) There are several typos that need correcting. Also, small and big capital letters to demark the panel names in the legends have been mixed.

Thank you for the comment. We have corrected the panel labels as described below.

Figure 2 legend:

“Representative choice neuron (ROI 5-8 in I) and a non-selective neuron (ROI 9)”

Figure 3 legend:

“..than the next choice. I. The decoding accuracy of the next choice …”

Figure 3 legend:

“Error bars, mean ± s.e.m. in I, 95% confidence interval in G. M, and O.”

Supplementary Figure 6：

“…neurons with rt ≥ 0.3 (blue) were shown. B. Trial-to-trial activity fluctuation … (rt ≥ 0.3, panel B) was color coded…”

We thoroughly checked the manuscript for typographical errors and corrected the issues.

(4) Many in the field still use the Paxinos nomenclature for PPC subfields, could the authors write something short about how these two nomenclatures correspond?

We have described the relationship between our area definitions and those of Paxinos in the main text as follows.

Line 702: “In addition to our definition, previous studies have also defined posterior parietal cortex (PPC) to include the higher visual areas A, AM, and RL (Glickfeld and Olsen, 2017; Wang et al., 2011). These areas partially overlap with the parietal association regions defined in the Paxinos atlas, including MPtA, LPtA, PtPD, and PtPR. For a detailed discussion of the correspondence and variability among these regional definitions, see Lyamzin and Benucci (2019).”

(5) Analyzing choice history may be affected by the long fluorescence Ca transients and will depend on excellent event deconvolution. Could the authors show some more zoomed-in examples of how well their deconvolution works?

We provide enlarged, trial-by-trial activity traces of the four example neurons shown in Figure 3A in Supplementary Figure 3G. In all neurons, multiple small calcium transients occur repeatedly throughout the delay period, which lasts longer than 10 s. If the sustained activity during the delay were simply due to a long decay time constant, one would expect a large calcium transient in the preceding trial that slowly decays over the delay period. However, such a pattern is not observed in the actual data. Also, since the decay time constant of GCaMP6s is on the order of ~1 s, signals persisting for ~10 s cannot be explained by slow decay alone.

(6) The authors write: "the history neurons exhibited properties of working memory." However, note that this is not a working memory task since the mice don't need to keep evidence in memory, the direction to lick can be made at the very beginning of a trial.

Behaviorally, demonstrating that an animal maintains working memory requires showing that its behavior changes based on retained information when new information is introduced, as in delayed match-to-sample tasks. In the present task, however, the correct action for the next trial is determined at the moment the action in the previous trial is completed, such that animals can simply switch to motor preparation at that point. Thus, from a strictly behavioral perspective, working memory is not required.

On the other hand, during the inter-trial interval (ITI), information from the previous trial dominates over information from the upcoming trial (Fig. 3H), which is more consistent with retention of past information than with motor preparation. Moreover, trials in which neural activity maintained information about the previous trial’s action were associated with a higher probability of correct performance in the subsequent trial. In other words, retaining past information contributes to guiding correct behavior in the next trial.

Based on these neural analyses, we interpret that mice retain information about their previous trial’s action history in working memory and use it to determine behavior in the subsequent trial. Accordingly, we consider ITI activity in PPC to reflect working memory rather than motor preparation. Nevertheless, we acknowledge that your concern is valid, and we have therefore revised the text as follows:

Line 234: “These results suggest that the history neurons exhibited properties of working memory.”

(7) In the section about the Choice History Task, the authors write: "Since the visual stimuli were randomly presented during the sampling period, the mice had to ignore the visual stimuli." Why continue to present the visual stimuli?

Thank you for the suggestion. By designing the vision task and the history task to have identical structures, we can apply the same encoding and decoding models to both tasks, which facilitates direct comparison between them. This design makes it easier to examine how neuronal activity patterns change depending on task demands.

Reviewer #2 (Recommendations for the authors):

(1) I don't understand the logic of Figure S7 and the neuropil analysis in general. Neuropil activity is purported to represent input, so it seems unsurprising that nearby neurons would exhibit similar dynamics.

Thank you for your comment. Your argument is correct, and it is not at all surprising that neuropil signals correlate with the activity of surrounding neurons. Here, we quantitatively examined the relationship between neuropil activity and the average activity of nearby neurons. In addition, in a separate analysis, we clarified the relationship between connectome information and neuropil activity. Taken together, these analyses reveal the relationship between connectome information and the local average of neuronal activity. We describe this point as follows:

“Indeed, the trial-to-trial variation of a neuropil activity could be approximated by the average of 1,000–10,000 neurons within several hundred micrometers from the center (Figure S7).”

Although we analyzed this phenomenon in the cases of areas A and AM, this finding should not be considered specific to A and AM but instead has broader, general significance. Accordingly, we added a new Results subsection and revised the manuscript as follows.

Line 448: “Constraints and limits of anatomical connectivity on neuronal population activity Although we have so far focused on the differences between A and AM, our data provide broader insights into the relationship between anatomical connectivity and neuronal population activity. First, based on Figure S7 and the considerations above, anatomical input correlations strongly constrain the correlations between local averages of activity across thousands of neurons. We then asked whether this anatomical constraint extends beyond mean activity, and how anatomical input correlations relate to relationships between neuronal population activities (population vectors).

The correlation between CC_t and r_anatomy was moderate (r = 0.60, Figure 6L). This moderate correlation did not change when the coupling neurons were eliminated (r = 0.61). Interestingly, the largest canonical component was the most unpredictable from the anatomical data (Figure 6M). Thus, while inter-area correlations based on the mean activity of neuronal populations are largely determined by anatomical input correlations, correlations between population vectors contain additional structure that cannot be captured by anatomical input correlations alone.

One possible source of this additional structure is globally shared activity, which may reflect behavior, brain state, or levels of neuromodulators. To evaluate the contribution of global activity on the canonical correlation between areas, we first compared the canonical coefficient vectors (CCV). We found that the first CCV had a similar orientation, regardless of the paired areas (Figure6N). This indicates that the largest components of correlated activity in the CCA analysis are globally shared fluctuations. We also directly evaluated the correlated activity components across all 8 areas with generalized canonical correlation analysis. The first CCV also had a similar orientation to the first generalized canonical coefficient vector (GCCV) (Figure 6O). These results indicate that the largest canonical component reflects a global correlation across all cortical areas imaged. Such global correlations may be driven by factors beyond cortico-cortical or thalamo-cortical inputs, such as the animal’s behavioral state as we recently characterized (H. Imamura et al., 2025; F. Imamura et al., 2025). We also confirmed the robustness of these results by repeating analyses using only the 40% highly active neurons after denoising with non-negative deconvolution (36828 out of 91397 neurons; Figure S9).”

(2) Furthermore, the neuropil signal likely contains signals from out-of-focus neurons that are presumably functioning similarly to the in-focus cells. Wouldn't the interesting question be to what extent the local neuropil signal in, for example, area A resembled that of neuronal activity in S1t?

Thank you very much for your comment. We agree with your point. Based on the evaluation in Figure S7, the neuropil signal likely contains the average activity of several thousand local neurons, including out-of-focus contributions. The neuropil signal in area A may also partially reflect neuronal activity from the neighboring S1t area. In particular, neurons that show little correlation with the local population average (i.e., the neuropil signal) within the same area are sometimes referred to as “soloists” (M. Okun et al., 2015). If such soloist neurons were found to exhibit strong correlations with the neuropil signal of an adjacent area, this would be a highly interesting result. However, such an analysis would go beyond the scope of the present manuscript and would require a new line of discussion; therefore, we plan to address this issue in future work.

(3) I generally found the final Results section (Relationship between mesoscale functional correlation and anatomical connections) to be hard to follow. The motivation for this analysis should be better explained.

We fully incorporated your suggestion and rewrote the final section of the Results accordingly. Please refer to our responses to the two comments above.

(4) The question of brain state/neuromodulation as a driver of the globally shared activity may be addressable by considering its correlation with pupillometry data.

We fully agree with your suggestion. In our experiments, visual stimuli change continuously, and thus pupil diameter changes are most likely driven primarily by changes in visual input. Although state-dependent fluctuations of brain activity may also be present, they are likely masked by the larger effects induced by visual stimulation. Therefore, analyzing pupil-linked signals as a factor of globally shared activity would be more appropriately addressed in experiments without visual stimulation. We plan to investigate this issue in future studies. Here, we have added the following description regarding pupil dynamics and their associated relationships.

Line 292: “We found that the neurons related to the tail and forepaws were similarly distributed around the parietal cortex including S1 and A, while the pupil-size related neurons were mapped around visual areas (Figure 4C). Changes in pupil diameter may influence neuronal activity through multiple mechanisms, including behavioral state or noradrenergic level [REF], nonlinear interactions with visual stimulation, and changes in the amount of light reaching the retina.”

Minor issues

(1) The authors deploy sophisticated mathematical techniques with essentially no explanation outside the Methods section. A brief introduction of jPCA and CCA in the main text would help the reader understand the value of these analyses.

Thank you for the comment. We added the following explanation.

Line 238: “In this task, left and right selection are alternated, so the activity of the history neuron is a sequence that repeats in two consecutive trials. We used jPCA⁴⁹ to visualize and quantify this activity pattern (Figure 3K). jPCA identifies low-dimensional projections of population activity that maximize rotational dynamics across time.”

Line 374: “Next, to investigate r_t of the population activity (r_{t_population}), we first reduced the dimension of population activity in each area into 10 by using PCA (principal component analysis) (Figure S6B,C). Then, “fluctuation activity” was recalculated for each dimension and trial type, analogous to the single-neuron analysis described above, but here representing noise in population-level activation patterns. We applied CCA (canonical correlation analysis) to each pair of areas and obtained an average of 10 canonical correlations (CC_t) as r_{t_population}. CCA identifies pairs of linear combinations of population activity from two areas that maximize their correlation across trials, thereby capturing shared population-level fluctuations. The CC_t structure between areas was similar across task types (Figure 5H) indicating that this structure reflects the underlying functional connectivity independent of the task. The CC_t between A and S1t was the largest among all the pairs (Figure 5H), whereas when the CC_t was averaged across all connections for each area, A and AM had the largest and second largest C_t, respectively (Figure 5I). The dominance in CC_t in A and AM disappeared when the neurons with r_{t_single} >0.3 were removed. Notably, the CC_t of AM and the other areas was uniform regardless of the paired areas across all 10 canonical components (Figure 5J). Thus, area AM is an integration hub of interareal communication, whereas A simply coupled with S1t, and such correlation structure at the population level critically depends on this subset of neurons.”

(2) The manuscript contains numerous typos ("hoice"), spelling errors ("parameters", "costom"), abbreviations that are not defined (ex: RL/rostrolateral), and minor grammatical issues that should be addressed by a round of copy editing.

We thank the reviewer for pointing this out. We have thoroughly corrected these typographical and grammatical errors, and have described the revisions in detail in our response to Reviewer 1, comment (3). In addition, we have clarified the abbreviations in the manuscript as follows.

Line 94: “rostrolateral area (RL)”

Figure 1 legend: “Abbreviations: RL, rostrolateral HVA; PM, posteromedial HVA; RSC, retrosplenial cortex.“

(3) Figure 3K unlabeled axes.

Thank you for the comment. We have added the axis labels.

(4) Figure 3K caption, first "(right)" should be "(left)".

Thank you very much for your careful attention to detail. We have made the requested correction.

(5) Figure 6 is hard to read. Panel A is too small, and the interpretation of G is difficult.

- For panel A, we added an enlarged view with images from a larger number of trials in Figure S7A.

- G represents the connectivity matrix. The sources correspond to the injection sites, and the targets correspond to voxels in the cerebral cortex. Because the latter may not be immediately clear, we explicitly indicated in the figure that the targets are cortical voxels.

(6) Figure S4C has a double compass.

Thank you for the comment. We have revised the manuscript accordingly.

Reviewer #3 (Recommendations for the authors):

While I have some questions and additional suggestions to further improve the clarity of the manuscript, I already found it to be highly interesting and well done in its current form.

Major points:

(1) The t-SNE comes up rather abruptly and is not well-explained in the main text or the figure caption. It would be good to provide some more information on the rationale of this analysis and how to interpret it. In particular, I don't see clear clusters in Figure 2H although the description of the authors seems to indicate that they observe clear functional classes such as choice, stimulus, and history neurons. Similarly, in Figure 3B, I don't see a clear separation between history and choice neurons in the t-SNE map. The example cells in Figure 3A appear to be delayed or long-tailed choice neurons rather than a dedicated group of 'history neurons'. It would be helpful for the interpretation of the t-SNE plots to show different PSTHs for different regions of the t-SNE map to better illustrate what different regions within the t-SNE projection represent and what distinguishes these cells.

Thank you for the comment. The absence of clearly defined clusters in the t-SNE map suggests that neuronal activity forms a continuum rather than discrete classes. Importantly, the purpose of the t-SNE map here is not to identify sharp clusters, but to demonstrate that the functional categorization provided by our encoding model broadly and comprehensively spans the major structures present in the unsupervised t-SNE map. We have revised the relevant text in the manuscript accordingly as follows.

Line 158: “To examine whether the neuron groups labeled by this model broadly capture the diversity of neuronal activity, we performed unsupervised clustering of neuronal activity using t-SNE. The functional labels revealed by this encoding model were consistent with the t-SNE clusters, indicating the validity of the encoding model (Figure 2H; Figure S4B; materials and methods).”

The issue regarding History neurons was also raised in Reviewer #1’s comment (5). We provide an enlarged view of Figure 3A in Figure S3A. Each History neuron exhibits multiple calcium transients repeatedly and asynchronously following the previous reward acquisition. Therefore, rather than being “choice neurons with a long tail,” these neurons are better interpreted as neurons whose activity is sustained during this delay period.

(2) Although the authors mention that neurons represent a mixture of features, they then use the encoding model to isolate clusters, such as vision or choice neurons. In general, the language throughout the manuscript suggests that there are various clusters of functionally segregated neurons (vision, choice, history, or coupling neurons). However, it is not clear to me to what extent this is supported by the data. Couldn't a choice neuron also be a vision neuron if both variables make significant contributions to the model? Similarly, are 'history' and 'choice' separate labels from the encoding model, or could a cell be given multiple labels? If a cell could be given multiple labels how did the authors create the colored plots on the right-hand side of Figures 2H and 3B? The example history cells in Figure 3J also appear to be highly selective for the contralateral choice, so again this seems to argue against a clear separation of choice and history neurons.

Each label is assigned based on whether the corresponding coefficient is significant in the encoding model, and therefore neurons that are both vision- and choice-selective do exist. The presence of mixed selectivity neurons in PPC is well established (e.g., MJ Goard et al., 2016 elife). In this manuscript, however, we focus not on functional overlap at the single neuron level, but on the spatial distribution of functional classes, and thus do not explicitly address mixed selectivity. Although the colors in Figure 2H and Figure 3B overlap, the underlying data for each are presented separately in Figure S4B and S4D, respectively. As shown there, each color generally occupies distinct regions in the t-SNE map.

(3) The decoding analysis in Figure 3F also suggests that a potential reason why there are more choice history signals in areas S1 and A is that neural activity is simply larger rather than due to the activity of a dedicated group of history neurons. Are the authors interpreting this differently? Could the duration of stored choice information also be affected by the dynamics of the calcium indicator?

Thank you for the comment. Simply having larger neural activity in S1t or A would not result in calcium transients with a ~1-s time constant persisting throughout a delay period lasting up to 10 seconds. As also noted in comment (1), History neurons exhibit sustained and repeated calcium transients, and therefore their activity cannot be explained merely by elevated neural activity levels. One could argue that all cortical areas carry history-related information but that the signal-to-noise ratio is higher in S1t or A, which might make such signals more detectable there. If this were the case, however, differences across areas in all forms of selectivity should similarly depend on signal-to-noise ratio. This is not what we observe in our data.

(4) I'm confused as to why the decoding accuracy is so high for areas A and S1t at time -3 relative to the choice in Figure 3F. Shouldn't this be the same as predicting the next choice in Figure 3H? Why is the decoding accuracy lower in this case?

Thank you for the comment. The analysis shown in Figure 3F includes only trials in which the choice was correct. This is the reason why the decoding performance in Figure 3H is lower. We have added this clarification to the main text.

Figure 3F: “Decoding accuracy of choice, outcome, and visual stimuli by the activity of 20 neurons from each area using only correct trials, before and after the choice onset, reward delivery, and the end of the visual stimuli, respectively. Line colors corresponded to the areas shown in panel G.”

(5) In general, the text is not very detailed about the statistics. While test scores and p-values are mentioned, it would be good to also state what is actually compared and what the n is (e.g. how many neurons, neuron pairs, areas, sessions, or animals) for each case. How do the authors account for the nested experiment design where many neurons are coming from a low number of animals?

Thank you for the comment. In our decoding analyses, we generally treat the number of animals as the independent variable. In contrast, for the encoding model analyses, we treat the number of neurons as the independent variable. As you correctly pointed out, because we recorded activity from a large number of neurons, statistical tests that treat individual neurons as independent samples can readily yield significant p-values even with a small number of animals. We have therefore confirmed that our conclusions are not driven by a large effect from a single animal. When making qualitative claims, we rely not only on statistical significance (p-values) but also require clear differences in effect size. We have added the following clarification to the Statistics section accordingly.

Line 1049: ”For the decoding analyses, the number of animals was treated as the independent variable, whereas for the encoding model analyses, the number of neurons was treated as the independent variable. To ensure that the results were not driven by a single animal, we repeated the statistical tests while systematically excluding data from one animal at a time and confirmed that statistical significance was preserved in all cases. Furthermore, qualitative interpretations were made only when differences in effect size were clearly observed.”

(6) How was the grouping in Figure 2O done? Specifically, how were the thresholds for the dashed lines selected to separate PM and V1 from AM and RL as association areas? It seems to me like this grouping was done rather arbitrarily as the difference in choice decoding accuracy is not particularly large between these areas.

This line does not have a specific quantitative basis, but we consider it useful as an illustrative aid. We have added this clarification to the figure legend.

Figure 2O: “Decoding accuracies of time in video presentation and choice direction indicate that AM would be the best position for associating these two signals. The background color and dashed lines are provided as visual aids for illustrative purposes.”

(7) The fact that neurons with high rt_single tend to share the same function might also indicate the approach is insufficient to remove all effects of tuning to trial types from the neural data. Since the authors subtract the average of each trial type, the average trial-type related information is removed but type-specific variations that are not equally presented in the average might remain. For choice neurons for example, attentive vs in-attentive choices could be represented differently and thus remain in the data since the average would be a mixture of both. The same goes for other factors that would drive a particular modulation in the choice - or stimulus - related part of the trial which could still tie these neurons together. One way to circumvent this concern could be to first compute the mean activity for all time points in each trial and then compute the trial-to-trial variability across all trials of the same type. Alternatively, I would be curious how the results play out when using data when the animal is not actively performing the task to compute rt_single.

Thank you for the comment. The concern raised by the reviewer applies to all noise-correlation analyses and highlights an important limitation of this approach, namely that factors other than the observed variables are treated as noise. By subtracting the trial-averaged activity, information related to sensory input and the direction of the first lick at choice can be removed. However, other factors cannot be eliminated if they are not observed. For example, if right hindlimb movements tend to occur only in trials with visual stimulation combined with left choice, such effects cannot be removed because they are not measured. The same issue remains even when restricting the analysis to a single trial type. Based on these considerations, we have added the following text to the manuscript.

Line 932: “Correlation of trial-to-trial variance of activity between a pair of single neurons was defined as r_{t_single}. To calculate r_{t_single}, we averaged the activity of individual neurons over the sampling period, and the average across each trial type was subtracted from this value. The trial types consisted of four sets of pairs of stimuli and responses, that is, the video stimulation and left choice, the video stimulation and right choice, the black screen and left choice, and the black screen and right choice. By this operation, we extracted the fluctuating components of single-neuron activity that are independent of the trial types. Although the finding that neurons with high r_{t_single} tend to share the functional properties we propose is not a trivial consequence of the analysis. At the same time, it remains possible that high r_{t_single} reflects the degree to which neurons share unobserved features, and that such features are correlated with our functional classification. Thus, while this analysis suggests that correlated fluctuations across cortical areas may contribute to the determination of functional types, establishing an exclusive conclusion will require more fine-grained behavioral measurements, tighter control of internal states, and causal identification through targeted interventions.”

Minor points:

(1) Why did the authors use the activity of 50 neurons for the decoder analysis in Figure 2K? Didn't they have many more neurons available? How were these selected?

We found that the conclusions were identical when using datasets consisting of either 50 neurons or 20 neurons across all analyses. Because the total number of recorded PM neurons did not reach 100 in at least one mouse, we standardized the analyses to 50 neurons in order to match the number of neurons across all cortical areas and animals.

(2) The authors mention that some PPC neurons showed complex dynamics rather than encoding a specific feature such as visual or choice information but do not mention actual numbers on this point. It would be good to quantify to what extent neurons in different regions represent such mixed selectivity and whether there are clear differences in selectivity. This would also be interesting to discuss in context to earlier work on mixed selectivity in the parietal cortex, such as Raposo et al 2015.

Thank you for the comment. Your point is entirely valid. However, as explained in our response to your major comment, our analyses focus not on how individual neurons are classified, but rather on the spatial distribution of these functional categories.

(3) I have a hard time understanding what the length of the bars in the right panel of Figure 2k indicates. Does this plot show more than the decoder accuracy before and after the choice? Is the bar length related to the standard deviation? The same question for the visualization in panel 2n. It looks nice but I'm confused about what it shows exactly.

These bars represent confidence intervals. Although this is stated at the end of the Figure 2 legend, we agree that it may not be sufficiently clear, and we have therefore added this information to the Statistics section.

Line 1046: “In Figure 2K and N, and Figure 3G, L, M, and O, the bars indicate the 95% confidence intervals. All other bars denote s.e.m., unless otherwise noted.”

(4) Is Figure 3D showing the same association index as in Figure 2j, thus showing the same result as in the vision task or is this meant to show something new? It was not clear to me from the wording, so it would be good to clarify.

You are correct that the magenta trace in Fig. 3D is the same as in Fig. 2J. This panel was included to explicitly illustrate that, in areas A and AM, the separation between History and Association approximately overlaps. We have added the following clarification to the figure legend accordingly.

Figure 3D: “The percentage of history neurons and the association index (as defined in Fig. 2J) were overlaid for comparison.”

(5) When computing the Pseudo R2 for regressor contribution, how was the null model computed? From shuffling all regressors in the model? I think this is fine but it's not fully clear what the intended effect of this procedure is. For the description of Figure 4C it would be good to add a sentence explaining how to interpret the pseudo R^2.

The null model predicts a fixed value that is independent of the explanatory variables, i.e., it predicts only the intercept. This provides a useful correction term when performing cross-validation, particularly in cases where baseline values differ across folds. In Figure 4C, the analysis shows the contribution of adding body part positions and pupil diameter to the model for predicting neural activity. We have added the following text to the Methods section.

Line 881: “To estimate the contribution of parameters for the left forelimb, the right forelimb, the tail, and the pupil, we repeated the same analysis with a reduced model where each set of predictors was eliminated from the full model (Figure 4B). Then, the pseudo-R² was obtained for each set of predictors by (MSE_reduced ‒ MSE_full) /MSE_null, where MSE is the mean squared error, MSE_reduced is MSE for the reduced model, MSE_full is the MSE of the full model, and MSE_null is the null model. The null model predicts a fixed value that is independent of the explanatory variables; specifically, it simply outputs the mean of the training data. For example, we constructed a regression model without the parameters regarding the left forelimb (green shade of Figure 4B), obtained MSE_reduced for the left forelimb, and the pseudo-R² was calculated as above by comparing the MSE of the full model and the null model. This value reflects the extent to which the position of the left forelimb contributes to the prediction of neuronal activity.”

(6) It seems surprising that the pupil-size-related neurons were mapped around visual areas although the pupil should carry clear luminance information. Is this because the luminancerelated information in the pupil can also be explained by the stimulus variable in the model?

Pupil size changed markedly before and after visual stimulus presentation (Figure S5C), dilating during the black stimulus and constricting during the video stimulus. This likely reflects changes relative to the luminance of the gray screen presented in the absence of visual stimuli. In our encoding model, visual stimuli are included as independent regressors for each corresponding time window. Therefore, pupil fluctuations that are temporally locked to visual stimulation are explained by these visual regressors. Neuronal activity that is better explained by pupil size changes not accounted for by the visual regressors is classified as pupil-related. At least three mechanisms may underlie the influence of pupil size on neuronal activity. First, fluctuations in pupil diameter have been linked to behavioral state or noradrenergic level [REF], which can act as variables independent of visual stimulation. Second, pupil fluctuations may be amplified in a stimulus-dependent manner, reflecting nonlinear interactions between visual input and brain state. Third, changes in pupil diameter alter the amount of light reaching the retina, which can modulate activity in visual cortical areas. The latter two mechanisms are therefore expected to predominantly affect visual areas and may explain why pupil-related neurons are more frequently observed there. The first mechanism is likely related to global brain state, and its association with behavior may account for the presence of pupil-related neurons in S1. However, these interpretations require confirmation through more refined causal manipulations. Accordingly, we limited the addition to the manuscript to the following statement.

Line 292: “We found that the neurons related to the tail and forepaws were similarly distributed around the parietal cortex including S1 and A, while the pupil-size related neurons were mapped around visual areas (Figure 4C). Changes in pupil diameter may influence neuronal activity through multiple mechanisms, including behavioral state or noradrenergic level [REF], nonlinear interactions with visual stimulation, and changes in the amount of light reaching the retina.”

(7) What is meant by 'external control parameters such as a video frame' when explaining the encoding model?

Thank you for the comment. We added the following explanation.

Line 151: “In the encoding model, the activity of each neuron was fitted by a weighted sum of external control parameters, such as video frames, and behavioral parameters, such as choice and reward direction. Because the visual stimulus changes continuously over time, sliding time windows were placed during the visual stimulus period.”

(8) What does the trace in Figure 2G show? Is this a single-cell example? What are the axes here?

We added an explanation to the figure legend.

Figure 2G: “Schematic of our encoding model. The bottom right panel shows an example of single-neuron activity with an overlay of the fitting obtained by the encoding model.”

(9) There seems to be a word missing in the sentence that describes the results for Figure 3O in the main text.

Thank you for the comment. We added the following description related to Fig. 3O.

Line 247: “resulting in the decoding accuracy of time after a specific choice being lower than in A (Figure 3O).”

(10) The abbreviation RP is used when describing Figure S5A. It should be mentioned that this refers to the response period.

Thank you for the comment. We added the following description related to Figure S5A.

Line 283: “We found that the angle of the tail was significantly different from the baseline values several seconds after the response period (RP) (Figure S5A)”

(11) I can't see the color difference between the traces in Figure 2E. There are probably red and green but this is hard to see for readers with red-green color blindness. Does the black indicate the time of visual stimulation? Is the line in Figure 2F the time when the spouts move in?

Thank you for the comment. In Fig. 2E, we improved visibility by changing the line opacity. In addition, the vertical line in Fig. 2E indicates the onset of the visual stimulus, and the vertical line in Fig. 2F indicates the onset of the response period. We have added the following explanations to the figure legend.

Figure 2: E. “Representative vision neurons (ROI 1-4 in I). The red bars indicate sampling periods during video presentation, and the brown bars indicate sampling periods without video stimulation. Vertical black lines mark the onset of the sampling period. F. Representative choice neuron (ROI 5-8 in I) and a non-selective neuron (ROI 9). Light blue lines indicate the response periods in trials with left choices, and purple lines indicate the response periods in trials with right choices. Vertical black lines mark the onset of the response period.”

(12) It might be useful to provide a short explanation in the results or methods of why the harmonic mean was used for the computation of the association index. I think it makes sense but since it is not commonly used this could be helpful for the reader to understand the approach.

Thank you for the comment. We added the following explanation to the main text.

Line 869: “The association index was determined by the harmonic mean of the rates of vision neurons and choice neurons. The harmonic mean approaches the arithmetic mean when the two values are similar, but becomes closer to the smaller value when the two values differ substantially. Therefore, the association index takes a large value when both vision neurons and choice neurons are abundant.”

(13) I don't fully understand how coupling diversity is computed. If there are six preference vectors, what is meant by taking the average of angles between all pairs of the two vectors?

Which two are meant here?

Thank you for the comment. We revised the explanation as follows.

Line 950: “To quantify the diversity of coupling patterns across clusters, we computed the angle between every pair of preference vectors. We then averaged these pairwise angles and defined this quantity as the “coupling diversity.”

(14) The results text states that the high correlation between r_anatomy and r_neuropil (Figure 6I) is evidence for the functional correlations being driven by cortico-cortical connectivity. However, Figure 6J shows that correlations for either cortico-cortical or thalamo-cortical connectivity are below 0.94 and generally higher for thalamo-cortical connectivity. This doesn't negate the general point of the authors but it would be good to clarify this section so it is easier to understand if r_anatomy includes both cortico-cortical and thalamo-cortical data and how the results in Figure I and J go together with the description in the results section.

You are correct. We have revised the text to clarify that the analysis reflects the combined effects of both cortico-cortical and thalamo-cortical inputs.

Line 436: “This correspondence suggests that the mesoscale interarea correlation is determined by the cortico-cortical and thalamo-cortical common input at mesoscale. Figure S8: A. Using Allen connectivity atlas, the axonal density of cortico-cortical and thalamo-cortical projection was analyzed.”

(15) I'm not very familiar with canonical correlation analysis and found this part hard to follow. Some additional explainer sentences would be helpful here. For example, what does it mean to take the average of the top 10 canonical correlations as rt_population? What exactly are the canonical correlation vectors? It was also not clear to me what exactly the results in Figure 5J signify.

Thank you for the comment. We have clarified the description in the main text related to CCA and the associated analyses as follows.

Line 374: “Next, to investigate r_t of the population activity (r_{t_population}), we first reduced the dimension of population activity in each area into 10 by using PCA (principal component analysis) (Figure S6B,C). Then, “fluctuation activity” was recalculated for each dimension and trial type, analogous to the single-neuron analysis described above, but here representing noise in population-level activation patterns. We applied CCA (canonical correlation analysis) to each pair of areas and obtained an average of 10 canonical correlations (CC_t) as r_{t_population}. CCA identifies pairs of linear combinations of population activity from two areas that maximize their correlation across trials, thereby capturing shared population-level fluctuations. The CC_t structure between areas was similar across task types (Figure 5H) indicating that this structure reflects the underlying functional connectivity independent of the task. The CC_t between A and S1t was the largest among all the pairs (Figure 5H), whereas when the CC_t was averaged across all connections for each area, A and AM had the largest and second largest CC_t, respectively (Figure 5I). The dominance in CC_t in A and AM disappeared when the neurons with r_t,single >0.3 were removed. Notably, the CC_t of AM and the other areas was uniform regardless of the paired areas across all 10 canonical components (Figure 5J). Thus, area AM is an integration hub of interareal communication, whereas A simply coupled with S1t, and such a correlation structure at the population level critically depends on this subset of neurons.”

eLife Assessment

This valuable study proposes a novel rapid-entry mechanism for S. aureus, involving the rapid release of calcium from lysosomes. The paper's strength lies in its very interesting hypothesis. The methods used are solid and adequately support the conclusions.

Reviewer #2 (Public review):

In the manuscript Ruhling et al propose a rapid uptake pathway that is dependent on lysosomal exocytosis, lysosomal Ca2+ and acid sphingomyelinase, and further suggest that the intracellular trafficking and fate of the pathogen is dictated by the mode of entry. Overall, this is manuscript argues for an important mechanism of a 'rapid' cellular entry pathway of S.aureus that is dependent on lysosomal exocytosis and acid sphingomyelinase and links the intracellular fate of bacterium including phagosomal dynamics, cytosolic replication and host cell death to different modes of uptake.

Key strength is the nature of the idea proposed, while continued reliance on inhibitor treatment combined with lack of phenotype / conditional phenotype for genetic knock out is a major weakness.

In the revised version, the authors perform experiments with ASM KO cells to provide genetic evidence of the role for ASM in S. aureus entry through lysosomal modulation. The key additional experiment is the phenotype of reduced bacterial uptake in low serum, but not in high serum conditions. The authors suggest this could be due to the SM from serum itself affecting the entry. While this explanation is plausible, prolonged exposure of cells to low serum is well documented to alter several cellular functions, particularly in the context of this manuscript, lysosomal positioning, exocytosis and Ca2+ signaling. A better control here could be WT cells grown in low serum. If SM in serum can interfere, why do they see such pronounced phenotype on bacterial entry in WT cells upon chemical inhibition?

While the authors argue a role for undetectable nano-scale Cer platforms on the cell surface caused by ASM activity, results do not rule out a SM independent role in the cellular uptake phenotype of ASM inhibitors.

The authors have attempted to address many of the points raised in the previous revision. While the new data presented provide partial evidence, the reliance on chemical inhibitors and lack of clear results directly documenting release of lysosomal Ca2+, or single bacterial tracking, or clear distinction between ASM dependent and independent processes dampen the enthusiasm.

I acknowledge the author's argument of different ASM inhibitors showing similar phenotypes across different assays as pointing to a role for ASM, but the lack of phenotype in ASM KO cells is concerning. The author's argument that altered lipid composition in ASM KO cells could be overcoming the ASM-mediated infection effects by other ASM-independent mechanisms is speculative, as they acknowledge, and moderates the importance of ASM-dependent pathway. The SM accumulation in ASM KO cells does not distinguish between localized alterations within the cells. If this pathway can be compensated, how central is it likely to be ?

The authors allude to lower phagosomal escape rate in ASM KO cells compared to inhibitor treatment, which appears to contradict the notion of uptake and intracellular trafficking phenotype being tightly linked. As they point out, these results might be hard to interpret. Could an inducible KD system recapitulate (some of) the phenotype of inhibitor treatment? If S. aureus does not escape phagosome in macrophages, could it provide a system to potentially decouple the uptake and intracellular trafficking effects by ASM (or its inhibitor treatment) ?

The role of ASM on cell surface remains unclear. The hypothesis proposed by the authors that the localized generation of Cer on the surface by released ASM leads to generation of Cer-enriched platforms could be plausible, but is not backed by data, technical challenges to visualize these platforms notwithstanding. These results do not rule out possible SM independent effects of ASM on the cell surface, if indeed the role of ASM is confirmed by controlled genetic depletion studies.

The reviewer acknowledges technical challenges in directly visualizing lysosomal Ca2+ using the methods outlined. Genetically encoded lysosomal Ca2+ sensor such as Gcamp3-ML1 might provide better ways to directly visualize this during inhibitor treatment, or S. aureus infection.

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #2 (Public review):

In the manuscript, Ruhling et al propose a rapid uptake pathway that is dependent on lysosomal exocytosis, lysosomal Ca2+ and acid sphingomyelinase, and further suggest that the intracellular trafficking and fate of the pathogen is dictated by the mode of entry. Overall, this is manuscript argues for an important mechanism of a 'rapid' cellular entry pathway of S.aureus that is dependent on lysosomal exocytosis and acid sphingomyelinase and links the intracellular fate of bacterium including phagosomal dynamics, cytosolic replication and host cell death to different modes of uptake.

Key strength is the nature of the idea proposed, while continued reliance on inhibitor treatment combined with lack of phenotype for genetic knock out is a major weakness.

We agree with the reviewer that a S. aureus invasion phenotype in ASM K.O. cells would unequivocally demonstrate the importance of ASM for the process. In the revised manuscript, we report an invasion phenotype in ASM K.O. cells. The absence of an invasion phenotype in ASM K.O. cells in our original experiments was likely caused by SM accumulation in ASM-depleted cells originating from FBS (see Figure 2I, in the revised manuscript).

We thus cultured cells for up to three days in 2% FBS and then reduced the concentration to 1% FBS one day prior to experimentation. Under these conditions reduced S. aureus invasion in ASM K.O.s was observed when compared to wildtype cells.

This was not detected when we cultured the cells in medium containing the common concentration of 10% FBS. Our new data supports the results we acquired with three different ASM inhibitors.

The invasion defect in ASM K.O.s cultured in low FBS was more pronounced at 10 min p.i. when compared to the 30 minute time point (Figure 2K), further corroborating that the ASM-dependent invasion pathway is relevant early in infection. This is consistent with the invasion dynamics we observed upon interference with lysosomal Ca²⁺ signaling [TPC1 K.O. (Figure 1C), BAPTA-AM (Figure 3D)], lysosomal exocytosis [Syt7 K.O. (Figure 2F), Ionomycin (Figure 3D)] and ASM activity by inhibitor treatment (Figure 3D).

Originally, we had hypothesized that changes in the sphingolipidome induced by absence of ASM may have caused the lack of an S. aureus invasion phenotype. We thus compared the sphingolipidome of ASM K.O.s cultured in 1% and 10% FBS. Indeed, SM accumulation was less severe when we cultured the cells in 1% FBS (Figure 2M and Supp. Figure 3). Hence, we think that strong SM accumulations in ASM K.O. cells cultured in 10% FBS may facilitate ASM-independent invasion mechanisms and thus, the absence of ASM-dependent invasion could not be detected by analyzing the number of invaded bacteria. This is supported by experiments, where we treated ASM K.O.s with the ASM inhibitor ARC39, which only slightly affected S. aureus invasion, whereas we detected a strong reduction of internalized bacteria by ARC39 treatment of WT cells (Figure 2 J). We think that this experiment and the reduced invasion in ASM K.O.s rule out an ASM/SM-independent effect of the inhibitors.

- While the authors argue a role for undetectable nano-scale Cer platforms on the cell surface caused by ASM activity, results do not rule out a SM independent role in the cellular uptake phenotype of ASM inhibitors.

We agree with reviewer that we do not show formation of ceramide-enriched platforms, and we thus changed the manuscript accordingly (see below).

- The authors have attempted to address many of the points raised in the previous revision. While the new data presented provide partial evidence, the reliance on chemical inhibitors and lack of clear results directly documenting release of lysosomal Ca2+, or single bacterial tracking, or clear distinction between ASM dependent and independent processes dampen the enthusiasm.

We shared the reviewer’s desire to discriminate between ASM-dependent and ASM-independent processes, but we are limited by cell biology and the simultaneous occurrence of processes - here the uptake of bacteria by multiple pathways.

However, we were able to address ASM-dependency of our rapid uptake mechanism by observing a genetic phenotype in SMPD1 knockout-cells.

We here do not make any assumptions on the centrality of the pathway and its importance in vivo. As scientists we were interested in the fact that such an ASM dependent pathway existed. In different as of yet still unidentified cell lines such a pathway may pose the main entry point for bacteria. Or maybe it represent an ASM-dependent mode of receptor uptake which we have identified with the bacteria piggy-backing into the cells.

- I acknowledge the author's argument of different ASM inhibitors showing similar phenotypes across different assays as pointing to a role for ASM, but the lack of phenotype in ASM KO cells is concerning. The author's argument that altered lipid composition in ASM KO cells could be overcoming the ASM-mediated infection effects by other ASM-independent mechanisms is speculative, as they acknowledge, and moderates the importance of ASM-dependent pathway. The SM accumulation in ASM KO cells does not distinguish between localized alterations within the cells. If this pathway can be compensated, how central is it likely to be?

We are convinced that our new genetic evidence of an S. aureus invasion phenotype in ASM K.O.s will eliminate the reviewer’s concerns about the role of ASM during the bacterial invasion.

The new lipidomics data of ASM K.O.s cultured in 1% and 10% FBS (Figure 2, M, Supp. Figure 3) and inhibitor-treated WT cells (Figure 2L, Supp. Figure 3) show a correlation between SM accumulation and the invasion phenotype.

We agree with the reviewer, however, that the reason why changes in sphingolipidome increase ASM-independent S. aureus internalization by host cells remains elusive. One possible explanation is a dysfunction of the lipid raft-associated protein caveolin-1 upon strong SM accumulation, which was previously shown to appear in ASM-deficient cells (1, 2). A lack of caveolin-1 results in strongly increased host cell entry of S. aureus (3, 4). Characterization of the mechanism behind these observations requires further experimentation and is beyond the scope of the current manuscript.

Host cells possess mechanisms to prevent infections, while pathogens developed strategies to circumvent these defense processes. In the present scenario, a physiological membrane composition of the host cell represents such a pathogen defense mechanism (as shown e.g. for caveolin-1 that restricts invasion of S. aureus in healthy cells). If a defense mechanism is disabled (as we speculate it is the case upon strong SM accumulation in ASM K.O.s cultured in 10%FBS), infection is facilitated. In healthy WT cells, these mechanisms (e.g. caveolin-1) are functional and, hence, we would not expect a “compensation” of ASM-dependent invasion. We here analyze invasion events that cannot be prevented by host defense mechanisms as they occur in untreated WT cells and are absent upon interfering with the ASM-dependent invasion pathway (by inhibitors and genetic K.O.). Thus, we think the ASM-dependent pathway, which mediates 50-70% of bacteria internalized by healthy WT cells 10 min p.i., is central for the infection.

- The authors allude to lower phagosomal escape rate in ASM KO cells compared to inhibitor treatment, which appears to contradict the notion of uptake and intracellular trafficking phenotype being tightly linked. As they point out, these results might be hard to interpret.

We measured phagosomal escape of S. aureus JE2 in ASM K.O. cells cultured in 1% FBS. Again, we infected cells for 10 or 30 min and determined the escape rates 3h p.i. However, the results are similar to escape rates determined with 10% FBS (Author response image 1).

Escape rates of S. aureus were significantly decreased in absence of ASM regardless of the FBS concentration in the medium. We therefore think that prolonged absence of ASM has other side effects. For instance, certain endocytic pathways could be up- or down-regulated to adapt for the absence of ASM or could be affected by other changes in the lipidome (that can be minimized but not completely prevented by culturing cells in 1% FBS). This could, for instance, affect maturation of S. aureus-containing phagosomes and hence phagosomal escape.

Author response image 1.

As it is unclear how prolonged absence of ASM can affect cellular processes, we think other experiments investigating the role of ASM-dependent invasion for phagosomal escape are more reliable. Most importantly, bacteria that enter host cell early during infection (and thus, predominantly via the “rapid” ASM-dependent pathway) possess lower phagosomal escape rates than bacteria that entered host cells later during infection (Figure 5, D and E). This is confirmed by higher escapes rates upon blocking ASM-dependent invasion with Vacuolin-1 (Figure 4E) and three different ASM inhibitors (Figure 4C and D). We further demonstrate that sphingomyelin on the plasma membrane during invasion influences phagosomal escape, while sphingomyelin levels in the phagosomal membrane did not change phagosomal escape (Figure5 a and b). This is summarized in Figure 5F.

- Could an inducible KD system recapitulate (some of) the phenotype of inhibitor treatment ? If S. aureus does not escape phagosome in macrophages, could it provide a system to potentially decouple the uptake and intracellular trafficking effects by ASM (or its inhibitor treatment)?

Inducible knock-downs in our laboratory are based on the vector pLVTHM in cells co-expressing the repressor TetR fused to a KRAB domain. It needs to be stated that for optimal knock-downs the induction has to be performed by doxycycline supplementation in the medium for 7 days thus leading to several days of growth of the cells, which will allow the cells to adapt their lipid metabolism thus reflecting a situation that we encounter for the K.O.s.

ASM-dependent uptake of S. aureus in macrophages has been demonstrated before (5). However, the course of infection in macrophages differs from non-professional phagocytes (6). E.g. in macrophages, S. aureus replicates within phagosomes, whereas in non-professional phagocytes replicates in the host cytosol. Absence of ASM therefore may influence the intracellular infection of macrophages with S. aureus in a distinct manner.

- The role of ASM on cell surface remains unclear. The hypothesis proposed by the authors that the localized generation of Cer on the surface by released ASM leads to generation of Cer-enriched platforms could be plausible, but is not backed by data, technical challenges to visualize these platforms notwithstanding. These results do not rule out possible SM independent effects of ASM on the cell surface, if indeed the role of ASM is confirmed by controlled genetic depletion studies.

We agree with the reviewer that we do not show generation of ceramide-enriched platforms. We thus changed Figure 6F in the revised manuscript to make clear that it remains elusive whether ceramide-enriched platforms are formed. We also added a sentence to the discussion (line 615) to emphasize that the existence of these microdomains is still debated in lipid research.

We think that the following observations support SM-dependent effects of ASM during S. aureus invasion:

(i) reduced invasion upon removing SM from the plasma membrane (Figure 2N, Supp. Figure 2M)

(ii) increased invasion in TPC1 and Syt7 K.O. (Figure 2, P) in presence of exogenously added SMase.

However, we agree with the reviewer that we do not directly demonstrate ASM-mediated SM cleavage during S. aureus invasion. Hence, we added a sentence to the discussion that mentions a possible SM-independent role of ASM for invasion (line 556) that reads:

“Since it remains elusive to which extent ASM processes SM on the plasma membrane during S. aureus invasion, one may speculate that ASM could also have functions other than SM metabolization during host cell entry of the pathogen. However, we did not detect a direct interaction between S. aureus and ASM in an S. aureus-host interactome screen (7).”

- The reviewer acknowledges technical challenges in directly visualizing lysosomal Ca2+ using the methods outlined. Genetically encoded lysosomal Ca2+ sensor such as Gcamp3-ML1 might provide better ways to directly visualize this during inhibitor treatment, or S. aureus infection.

We thank the reviewer for this suggestion. We included the following section in our discussion (line 593):

“Since fluorescent calcium reporters allow to monitor this process microscopically (8, 9) ,future experiments may visualize this process in more detail and contribute to our understanding of the underlying signaling. mechanisms.”

References

(1) J. Rappaport, C. Garnacho, S. Muro, Clathrin-mediated endocytosis is impaired in type A-B Niemann-Pick disease model cells and can be restored by ICAM-1-mediated enzyme replacement. Mol Pharm 11, 2887-2895 (2014).

(2) J. Rappaport, R. L. Manthe, C. Garnacho, S. Muro, Altered Clathrin-Independent Endocytosis in Type A Niemann-Pick Disease Cells and Rescue by ICAM-1-Targeted Enzyme Delivery. Mol Pharm 12, 1366-1376 (2015).

(3) C. Hoffmann et al., Caveolin limits membrane microdomain mobility and integrin-mediated uptake of fibronectin-binding pathogens. J Cell Sci 123, 4280-4291 (2010).

(4) L.-P. Tricou et al., Staphylococcus aureus can use an alternative pathway to be internalized by osteoblasts in absence of β1 integrins. Scientific Reports 14, 28643 (2024).

(5) C. Li et al., Regulation of Staphylococcus aureus Infection of Macrophages by CD44, Reactive Oxygen Species, and Acid Sphingomyelinase. Antioxid Redox Signal 28, 916-934 (2018).

(6) A. Moldovan, M. J. Fraunholz, In or out: Phagosomal escape of Staphylococcus aureus. Cell Microbiol 21, e12997 (2019).

(7) M. Rühling, F. Schmelz, A. Kempf, K. Paprotka, J. Fraunholz Martin, Identification of the Staphylococcus aureus endothelial cell surface interactome by proximity labeling. mBio 0, e03654-03624 (2025).

(8) D. Shen et al., Lipid storage disorders block lysosomal trafficking by inhibiting a TRP channel and lysosomal calcium release. Nat Commun 3, 731 (2012).

(9) L. C. Davis, A. J. Morgan, A. Galione, NAADP-regulated two-pore channels drive phagocytosis through endo-lysosomal Ca(2+) nanodomains, calcineurin and dynamin. EMBO J 39, e104058 (2020).

eLife Assessment

This paper reports new data on the structure of the human CTF18-RFC clamp loader complex bound to the PCNA clamp. The new and convincing data complement previous reports of CTF-RFC-PCNA structures and as such, represents an important contribution.

Reviewer #1 (Public review):

Summary:

The authors report the structure of the human CTF18-RFC complex bound to PCNA. Similar structures (and more) have been reported by the O'Donnell and Li labs. This study should add to our understanding of CTF18-RFC in DNA replication and clamp loaders in general. However, there are numerous major issues that I recommend the authors fix.

Strengths:

The structures reported are strong and useful for comparison with other clamp loader structures that have been reported lately.

Reviewer #2 (Public review):

Summary

Briola and co-authors have performed a structural analysis of the human CTF18 clamp loader bound to PCNA. The authors purified the complexes and formed a complex in solution. They used cryo-EM to determine the structure to high resolution. The complex assumed an auto-inhibited conformation, where DNA binding is blocked, which is of regulatory importance and suggests that additional factors could be required to support PCNA loading on DNA. The authors carefully analysed the structure and compared it to RFC and related structures.

Strength & Weakness

Their overall analysis is of high quality, and they identified, among other things, a human-specific beta-hairpin in Ctf18 that flexible tethers Ctf18 to Rfc2-5. Indeed, deletion of the beta-hairpin resulted in reduced complex stability and a reduction in the rate of primer extension assay with Pol ε. Moreover, the authors identify that the Ctf18 ATP-binding domain assumes a more flexible organisation.

The data are discussed accurately and relevantly, which provides an important framework for rationalising the results.

All in all, this is a high-quality manuscript that identifies a key intermediate in CTF18-dependent clamp loading.

Reviewer #3 (Public review):

Summary:

CTF18-RFC is an alternative eukaryotic PCNA sliding clamp loader which is thought to specialize in loading PCNA on the leading strand. Eukaryotic clamp loaders (RFC complexes) have an interchangeable large subunit which is responsible for their specialized functions. The authors show that the CTF18 large subunit has several features responsible for its weaker PCNA loading activity, and that the resulting weakened stability of the complex is compensated by a novel beta hairpin backside hook. The authors show this hook is required for the optimal stability and activity of the complex.

Relevance:

The structural findings are important for understanding RFC enzymology and novel ways that the widespread class of AAA ATPases can be adapted to specialized functions. A better understanding of CTF18-RFC function will also provide clarity into aspects of DNA replication, cohesion establishment and the DNA damage response.

Strengths:

The cryo-EM structures are of high quality enabling accurate modelling of the complex and providing a strong basis for analyzing differences and similarities with other RFC complexes. They use complementary pre-steady state FRET and polymerase primer extension assays to investigate the role of a unique structural element in CTF18.

Weaknesses:

The manuscript would have benefited from a more detailed biochemical analysis using mutagenesis and assays to tease apart the functional relevance of the many differences with the canonical RFC complex.

Overall appraisal:

Overall, the work presented here is solid and important. The data is sufficient to support the stated conclusions.

Author response:

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

Public Reviews: 

Reviewer #1 (Public review): 

Summary: 

The authors report the structure of the human CTF18-RFC complex bound to PCNA. Similar structures (and more) have been reported by the O'Donnell and Li labs. This study should add to our understanding of CTF18-RFC in DNA replication and clamp loaders in general. However, there are numerous major issues that I recommend the authors fix. 

Strengths: 

The structures reported are strong and useful for comparison with other clamp loader structures that have been reported lately. 

Comments on revisions: 

The revised manuscript is greatly improved. The comparison with hRFC and the addition of direct PCNA loading data from the Hedglin group are particular highlights. I think this is a strong addition to the literature.

We thank the reviewer for their positive comments.  

I only have minor comments on the revised manuscript. 

(1) The clamp loading kinetic data in Figure 6 would be more easily interpreted if the three graphs all had the same x axes, and if addition of RFC was t=0 rather than t=60 sec.

We now analyze and plot EFRET as a function of time after complex addition, effectively setting the loader addition to t = 0 for each trace (Figure 6 and Figs S10-14 in the new manuscript). Baseline (Ymin) and plateau (Ymax) EFRET values were obtained by averaging the stable signal regions immediately before and after clamp-loader addition, respectively. Traces are normalized to their own dynamic range before fitting.

(2) The author's statement that "CTF18-RFC displayed a slightly faster rate than RFC" seems to me a bit misleading, even though this is technically correct. The two loaders have indistinguishable rate constants for the fast phase, and RFC is a bit slower than CTF18-RFC in the slow phase. However, the data also show that RFC is overall more efficient than CTF18-RFC at loading PCNA because much more flux through the fast phase (rel amplitudes 0.73 vs 0.36). Because the slow phase represents such a reduced fraction of loading events, the slight reduction in rate constant for the slow phase doesn't impact RFC's overall loading. And because the majority of loading events are in the fast phase, RFC has a faster halftime than CTF18-RFC. (Is it known what the different phases correspond to? If it is known, it might be interesting to discuss.)

We removed the quoted statement. We avoid comparing amplitude partitions (A₁/A_T) for CTF18-RFC because (i) a substantial fraction of the reaction occurs within the <7 s dead time, and (ii) single- vs double-exponential identifiability differs across complexes. Instead, we report model-minimal progress times: RFC t_0.5 ≤ 7 s (faster onset), CTF18-RFC ~ 8 s, CTF18^Δ165–194-RFC ~ 12 s; completion (t_0.95): RFC ≈ 77 s, CTF18-RFC ≈ 77 s, mutant ≈ 145 s. This shows RFC has the steeper onset, while CTF18-RFC catches up in completion, and the mutant is slower overall. We briefly note that RFC’s phases have been assigned in prior stopped-flow work and are consistent with a rapid entry step and a slower repositioning/complex release phase; we do not assign phases for CTF18-RFC here and instead rely on model-minimal timing comparisons to avoid over-interpretation. 

(3) AAA+ is an acronym for "ATPases Associated with diverse cellular Activities" rather than "Adenosine Triphosphatase Associated". 

Corrected to ATPases Associated with diverse cellular Activities (AAA+).

Reviewer #2 (Public review): 

Summary 

Briola and co-authors have performed a structural analysis of the human CTF18 clamp loader bound to PCNA. The authors purified the complexes and formed a complex in solution. They used cryo-EM to determine the structure to high resolution. The complex assumed an auto-inhibited conformation, where DNA binding is blocked, which is of regulatory importance and suggests that additional factors could be required to support PCNA loading on DNA. The authors carefully analysed the structure and compared it to RFC and related structures. 

Strength & Weakness 

Their overall analysis is of high quality, and they identified, among other things, a humanspecific beta-hairpin in Ctf18 that flexible tethers Ctf18 to Rfc2-5. Indeed, deletion of the beta-hairpin resulted in reduced complex stability and a reduction in a primer extension assay with Pol ε. Moreover, the authors identify that the Ctf18 ATP-binding domain assumes a more flexible organisation. 

The data are discussed accurately and relevantly, which provides an important framework for rationalising the results. 

All in all, this is a high-quality manuscript that identifies a key intermediate in CTF18-dependent clamp loading. 

Comments on revisions: 

The authors have done a nice job with the revision. 

We thank the reviewer for their very positive comments.

Reviewer #3 (Public review): 

Summary: 

CTF18-RFC is an alternative eukaryotic PCNA sliding clamp loader which is thought to specialize in loading PCNA on the leading strand. Eukaryotic clamp loaders (RFC complexes) have an interchangeable large subunit which is responsible for their specialized functions. The authors show that the CTF18 large subunit has several features responsible for its weaker PCNA loading activity, and that the resulting weakened stability of the complex is compensated by a novel beta hairpin backside hook. The authors show this hook is required for the optimal stability and activity of the complex. 

Relevance: 

The structural findings are important for understanding RFC enzymology and novel ways that the widespread class of AAA ATPases can be adapted to specialized functions. A better understanding of CTF18-RFC function will also provide clarity into aspects of DNA replication, cohesion establishment and the DNA damage response. 

Strengths: 

The cryo-EM structures are of high quality enabling accurate modelling of the complex and providing a strong basis for analyzing differences and similarities with other RFC complexes. 

Weaknesses: 

The manuscript would have benefited from a more detailed biochemical analysis using mutagenesis and assays to tease apart the differences with the canonical RFC complex. Analysis of the FRET assay could be improved. 

Overall appraisal: 

Overall, the work presented here is solid and important. The data is mostly sufficient to support the stated conclusions.

We thank the reviewer for their mainly positive assessment. Following this reviewer suggestion, we have re-analysed the FRET assay data and amended the manuscript accordingly.

Comments on revisions: 

While the authors addressed my previous specific concerns, they have now added a new experiment which raises new concerns. 

The FRET clamp loading experiments (Fig. 6) appear to be overfitted so that the fitted values are unlikely to be robust and it is difficult to know what they mean, and this is not explained in this manuscript. Specifically, the contribution of two exponentials is floated in each experiment. By eye, CTF18-RFC looks much slower than RFC1-RFC (as also shown previously in the literature) but the kinetic constants and text suggest it is faster. This is because the contribution of the fast exponential is substantially decreased, and the rate constants then compensate for this. There is a similar change in contribution of the slow and fast rates between WT CTF18 and the variant (where the data curves look the same) and this has been balanced out by a change in the rate constants, which is then interpreted as a defect. I doubt the data are strong enough to confidently fit all these co-dependent parameters, especially for CTF18, where a fast initial phase is not visible. I would recommend either removing this figure or doing a more careful and thorough analysis. 

We appreciate the reviewer’s concern regarding potential overfitting of the kinetic data in Figure 6. To address this, we performed a model-minimal re-analysis designed specifically to avoid parameter covariance and over-interpretation (Figure 6 and Figs S11-14 in the new manuscript). Only data recorded after the instrument’s <7 s dead time were included in the fits, thereby excluding the partially obscured early region of the reaction. For each clamp loader complex, we selected the minimal kinetic model that produced residuals randomly distributed about zero. This approach yielded a single-exponential fit for CTF18-RFC, whereas RFC and CTF18^Δ165–194-RFC required double-exponential fits; single-exponential models for the latter two complexes left structured residuals, clearly indicating the presence of an additional kinetic phase.

Rather than relying on co-dependent amplitude and rate parameters, we quantified the reactions by reporting progress times (t_0.5, t_0.90, t_0.95), which provide a model-independent measure of reaction speed. This directly addresses the reviewer’s concern and allows a fair comparison of the relative kinetics among the complexes.

From this analysis, RFC exhibited the fastest onset (t_0.5 ≤ 7 s; lower bound), while CTF18RFC and CTF18^Δ165–194-RFC showed progressively slower half-times of approximately 8 s and 12 s, respectively. Completion times further emphasized these differences: both RFC and CTF18-RFC reached 95 % completion at ~77 s, whereas the mutant required ~145 s. Despite these kinetic distinctions, CTF18-RFC and its β-hairpin deletion mutant achieved similar EFRET plateaus, indicating that the mutation slows reaction progression but does not reduce the overall extent of PCNA loading.

Finally, we emphasize that our interpretation is deliberately conservative. We do not assign distinct kinetic phases to CTF18-RFC, as their molecular basis remains unresolved. RFC’s phases have been characterized in prior stopped-flow studies, but CTF18-RFC likely follows a distinct or simplified pathway. Our conclusions are thus limited to what the data unambiguously support: deletion of the Ctf18 β-hairpin decreases the rate—but not the extent—of PCNA loading, consistent with the reduced stimulation of Pol ε primer extension observed under single-turnover conditions.

eLife Assessment

This important study reports that EEG recordings of the earliest stage of information processing in human visual cortex can be used to predict subsequent choice responses. The findings provide novel, convincing evidence for integrative processing in low-level sensory cortices at the level of scalp-recorded potentials, with the exact nature of the neural signals at the single cell level to be determined. The paper is likely to be of interest to neuroscientists interested in the contribution of early sensory signals to decision making.

Reviewer #1 (Public review):

General assessment of the work

In this manuscript, Mohr and Kelly show that the C1 component of the human VEP is correlated with binary choices in a contrast discrimination task, even when the stimulus is kept constant and confounding variables are considered in the analysis. They interpret this as evidence for the role V1 plays during perceptual decision formation. Choice-related signals in single sensory cells are enlightening because they speak to the spatial (and temporal) scale of the brain computations underlying perceptual decision making. However, similar signals in aggregate measures of neural activity offer a less direct window and thus less insight into these computations. The authors do a good job justifying their focus on the C1 component and illustrating how it may behave under different simulated scenarios. The results are interesting, although it is difficult to specify which reasonable hypothesis is exactly ruled out by these results. One interpretation is that V1 activity directly guides perceptual decisions in this task. Alternatively, higher-level areas may do this, provided that their activity largely reflects their V1-inputs. This certainly seems possible in a simple task like this.

Summary of substantive concerns

I have no substantive concerns about the revised version of the paper.

Reviewer #2 (Public review):

Summary:

Mohr and Kelly report a high-density EEG study in healthy human volunteers in which they test whether correlations between neural activity in primary visual cortex and choice behavior can be measured non-invasively. Participants performed a contrast discrimination task on large arrays of Gabor gratings presented in the upper left and lower right quadrants of the visual field. The results indicate that single-trial amplitudes of C1, the earliest cortical component of the visual evoked potential in humans, predict forced-choice behavior over and beyond other behavioral and electrophysiological choice-related signals. These results constitute an important advance for our understanding of the nature and flexibility of early visual processing.

Strengths:

The findings suggest a previously unsuspected role for aggregate early visual cortex activity in shaping behavioral choices.

The authors extend well-established methods for assessing covariation between neural signals and behavioral output to non-invasive EEG recordings.

The effects of initial afferent information in primary visual cortex on choice behavior is carefully assessed by accounting for a wide range of potential behavioral and electrophysiological confounds.

Caveats and limitations are transparently addressed and discussed.

Weaknesses:

Due to the inherent limitations of scalp-recorded visual evoked potentials, the results cannot be directly compared to invasive recordings in animal models.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

General assessment of the work:

In this manuscript, Mohr and Kelly show that the C1 component of the human VEP is correlated with binary choices in a contrast discrimination task, even when the stimulus is kept constant and confounding variables are considered in the analysis. They interpret this as evidence for the role V1 plays during perceptual decision formation. Choice-related signals in single sensory cells are enlightening because they speak to the spatial (and temporal) scale of the brain computations underlying perceptual decision-making. However, similar signals in aggregate measures of neural activity offer a less direct window and thus less insight into these computations. For example, although I am not a VEP specialist, it seems doubtful that the measurements are exclusively picking up (an unbiased selection of) V1 spikes. Moreover, although this is not widely known, there is in fact a long history to this line of work. In 1972, Campbell and Kulikowski ("The Visual Evoked Potential as a function of contrast of a grating pattern" - Journal of Physiology) already showed a similar effect in a contrast detection task (this finding inspired the original Choice Probability analyses in the monkey physiology studies conducted in the early 1990's). Finally, it is not clear to me that there is an interesting alternative hypothesis that is somehow ruled out by these results. Should we really consider that simple visual signals such as spatial contrast are *not* mediated by V1? This seems to fly in the face of well-established anatomy and function of visual circuits. Or should we be open to the idea that VEP measurements are almost completely divorced from task-relevant neural signals? Why would this be an interesting technique then? In sum, while this work reports results in line with several single-cell and VEP studies and perhaps is technically superior in its domain, I find it hard to see how these findings would meaningfully impact our thinking about the neural and computational basis of spatial contrast discrimination.

We agree that single cell measurements allow for a spatially more detailed analysis, but they are not feasible in humans. Assuming we value insights into the relationship between neural activity and decision making in the human as well as non-human brain, we are restricted to non-invasive measurements such as EEG, which inevitably showcase the neural underpinnings of decision making at a coarser level of analysis. This was the challenge we met with our paradigm design. For example, we chose contrast as the task-relevant stimulus feature in this study because monotonic contrast response functions exist for sensory neurons throughout the visual system, and the aggregated measures that we could attain with EEG would reflect that contrast-sensitivity and hence provide a window onto the encoding of the main decision-relevant quantity. We were specifically interested in initial afferent, contrast-dependent V1 activity reflected in the C1 component (80-90 ms). As we point out in the Introduction, the C1 is unusual among EEG signals in the extent to which it is dominated by a single visual area, V1 (Jeffreys & Axford, 1972; Clark et al., 1994; Di Russo et al., 2002; Ales et al., 2010; Mohr et al., 2024), and even if other downstream areas also make a minor contribution in the C1 time period, it still represents a very low-level sensory response early in the sensory analysis pipeline, appropriate for addressing our primary question of whether such a low-level signal is used in the formation of perceptual decisions. The alternative hypothesis, that early responses are passed over in decision readout, relates to a fundamental debate about whether early sensory responses are separated from cognition. The possibility that late, but not early, representations are correlated with choices does not imply that the later sensory representations are divorced from the earlier ones, only that there is a noise component that is not shared between the two, such as that produced by the ensuing computations that generate the later representations. Instead, a lack of choice probability in early representations would imply that decision readout is selective in where it sources sensory evidence from, with some possible reasons being to maintain high quality standards for sensory evidence or to impose a layer of separation between cognition and sensation.

As the reviewer points out, the animal literature is highly mixed on the topic of choice probability in V1. Even for orientation discrimination tasks where V1 is ostensibly highly suited given the existence of orientation columns in V1, and even when measurements are taken from V1 neurons with good neurometric performance and/or aggregated across a V1 population (Jasper et al 2019), some studies have reported little to no V1 choice probability. If our alternative hypothesis of no EEG-indexed V1 choice probability flies in the face of well-established anatomy and function of visual circuits, then so also do these empirical findings in the animal neurophysiology literature. 

Although there are important aspects of choice probability that are accessible in single cell studies but not in EEG (e.g. noise correlations, details of circuit physiology), our EEG measurements tap into the same phenomenon, just at a different level of analysis, i.e. the neural population level. At this level, we have been able to address whether the full body of sensory responses at a particular stage of visual analysis is systematically related to perceptual decision outcomes. Very similar questions are in fact sometimes addressed in the animal neurophysiology literature; for example, Kang and Maunsell (2020) aggregated single-cell choice probability measurements within visual areas to investigate whether choice probability strength at the level of an entire visual area was sensitive to task demands. The global vantage point of EEG comes with the additional benefit of picking up signatures of other potentially mediating processes such as attention and being able to control for them in our analysis. Our human study thus provides a valuable complementary viewpoint alongside animal neurophysiology work in this area.

Summary of substantive concerns:

(1) The study of choice probability in V1 cells is more extensive than portrayed in the paper's introduction. In recent years, choice-related activity in V1 has also been studied by Nienborg & Cumming (2014), Goris et al (2017), Jasper et al (2019), Lange et al (2023), and Boundy-Singer et al (2025). These studies paint a complex picture (a mixture of positive, absent, and negative results), but should be mentioned in the paper's introduction.

We thank the reviewer for highlighting these papers bearing on choice-related activity in V1, only two of which we had cited. The three additional studies do indeed lend further support to our description of the complex picture around V1-CP effects in the literature and we have now included them.

(2) The very first study to conduct an analysis of stimulus-conditioned neural activity during a perceptual decision-making task was, in fact, a VEP study: Campbell and Kulikowski (1972). This study never gained the fame it perhaps deserves. But it would be appropriate to weave it into the introduction and motivation of this paper.

We are aware of this paper, and indeed we ourselves have shown steady-state VEP (SSVEP) correlations with timing and selection of decision reports (O'Connell et al 2012; Grogan et al 2023), but SSVEPs do not provide an index of initial afferent V1 activity in the way that the C1 of the transient VEP does. SSVEPs are evoked by a rapid sequence of stimulus onsets, so that activity cannot be attributed to a particular stimulus onset nor its bottom-up latency resolved, and, being a response to an ongoing stimulus, it combines top-down and bottom-up influences from striate and extra striate areas (Di Russo et al 2007). Indeed, in Campbell and Kulikowski (1972) the SSVEP was almost entirely eliminated when the stimulus was undetected. This is in keeping with robust modulations of the SSVEP by spatial attention (Muller and Hillyard 2000). Cognitive influences of this magnitude are never observed in the C1, and in fact are often not observed at all even when later VEP components show robust modulations (Luck et al 2000), which motivated a recent meta-analysis to address the issue (Qin et al 2022). This highlights the important distinction between the earliest transient VEP activity reflecting mainly the initial afferent response in V1, and steady-state sensory activity reflecting a mix of bottom-up and top-down influences across visual cortex. Because of the importance of this distinction, we have added a reference to the above SSVEP papers to the 3rd paragraph of the introduction along with a statement about the distinction.

(3) What are interesting alternative hypotheses to be considered here? I don't understand the (somewhat implicit) suggestion here that contrast representations late in the system can somehow be divorced from early representations. If they were, they would not be correlated with stimulus contrast.

This same conundrum applies to single-cell studies of choice probability. Do studies showing choice probability in V4 but not V1 for example demonstrate that V4 is divorced from V1? In such studies, measurements are typically taken from large representative samples of neurons from both areas with good neurometric performance in both cases and the task often (though not always) involves a target stimulus feature that is encoded in V1 such as orientation. Why then should V4 but not V1 show choice probability when we know the vast majority of input to the visual cortex passes through V1? It must be that feature representation and choice formation are different things with one not inferring the other. This is true for an EEG study as much as it is for a single-cell study.

The alternative hypothesis in our study is that the early sensory responses indexed by the C1 are not directly used in the formation of the perceptual decision at hand. As outlined in our comments above, this does not imply that those early responses are divorced from later responses. Of course, both are correlated with stimulus contrast and so would correlate with each other across changing contrast but this does not necessitate that their noise is correlated when contrast is held constant because new instantiations of noise can be generated by the computations performed at each stage of visual processing. Thus, the interesting alternative hypothesis is that information contained in the sensory representation generated during initial afferent V1 activity is not used directly to form decisions, and instead, decisions are read out from the outputs of computations performed further downstream. Such an outcome, if it had arisen in our data, would have been consistent with a separation between cognition and early visual processing. Instead, our results suggest a certain level of cognitive interfacing at the lowest and earliest cortical levels of visual processing. We have now added text to the Introduction to highlight the distinction between sensory representation and decision readout in order to make the alternative hypothesis clearer.

(4) I find the arguments about the timing of the VEP signals somewhat complex and not very compelling, to be honest. It might help if you added a simulation of a process model that illustrated the temporal flow of the neural computations involved in the task. When are sensory signals manifested in V1 activity informing the decision-making process, in your view? And how is your measure of neural activity related to this latent variable? Can you show in a simulation that the combination of this process and linking hypothesis gives rise to inverted U-shaped relationships, as is the case for your data?

We thank the reviewer for this suggestion of a simulation, which we carried out using the Matlab code. We have also included new Figure 1-Figure Supplement 1 in the revised manuscript.

In our view, sensory signals in V1 are informing the decision-making process in this task from at least as early as the initial afferent response. The main point about C1 latency in relation to the response-time contingency of the choice probability effect is that the more time that elapses without a decision made (and therefore the more additional sensory processing that contributes to the decision), the more diluted is the contribution of the C1 to the decision by contributions from later representations, and thus choice probability reduces. Likewise, when response times are too quick for C1 evidence to contribute, choice probability is also absent, hence the inverted-U-shaped curve. Moreover, if the C1-choice correlation is mediated by a top-down factor such as attention rather than readout, the inverted-U-shaped curve is not expected because in such a case the relative timing of the C1 and choice commitment would not be relevant.

Reviewer #2 (Public review):

Summary:

Mohr and Kelly report a high-density EEG study in healthy human volunteers in which they test whether correlations between neural activity in the primary visual cortex and choice behavior can be measured non-invasively. Participants performed a contrast discrimination task on large arrays of Gabor gratings presented in the upper left and lower right quadrants of the visual field. The results indicate that single-trial amplitudes of C1, the earliest cortical component of the visual evoked potential in humans, predict forced-choice behavior over and beyond other behavioral and electrophysiological choice-related signals. These results constitute an important advance for our understanding of the nature and flexibility of early visual processing.

Strengths:

(1) The findings suggest a previously unsuspected role for aggregate early visual cortex activity in shaping behavioral choices.

(2) The authors extend well-established methods for assessing covariation between neural signals and behavioral output to non-invasive EEG recordings.

(3) The effects of initial afferent information in the primary visual cortex on choice behavior are carefully assessed by accounting for a wide range of potential behavioral and electrophysiological confounds.

(4) Caveats and limitations are transparently addressed and discussed.

We would like to thank the reviewer for these positive remarks.

Weaknesses:

(1) It is not clear whether integration of contrast information across relatively large arrays is a good test case for decision-related information in C1. The authors raise this issue in the Discussion, and I agree that it is all the more striking that they do find C1 choice probability. Nevertheless, I think the choice of task and stimuli should be explained in more detail.

We thank the reviewer for raising this point about the large stimulus arrays. As we said in our Discussion, it would seem that aggregation across a large stimulus region would be better suited to a downstream visual area with larger receptive fields, yet our setting of a strict deadline would put the emphasis back on earlier sensory representations. We now elaborate on this matter in the discussion, to say that although the small receptive fields and short, slow horizontal connections in V1 mean that the aggregation necessary for performing the task is unlikely to happen within V1 during the C1 timeframe, the aggregation would be readily achieved simply by convergence of the outputs of all relevant V1 neurons for a given stimulus array on the same decision process. In this sense, the design of our paradigm was such that the globally-measured C1 component on the scalp reflected the same aggregated evidence input as the summed V1 readout that we suppose would be entering the decision process.  

We have also added further rationale in the Methods section on the practical benefits of the stimulus design, as the reviewer anticipates in their subsequent point, of yielding robust C1 signals. This concern was paramount in the design of this study because we expected the C1 difference metric that was of interest to be very small. We also needed a robust C1 to be measured in both the upper and lower visual field in as many individuals as possible and, in our experience, this is true less often when using smaller stimuli, even with a pre-mapping procedure.

It also helped to homogenize C1 topography across individuals and ensure that topographies from the upper and lower visual field had sufficient overlap that there were electrodes with strong loading from both topographies where the C1 difference as a function of which array was brighter would be maximal.

We have updated the methods section to provide these rationales while we describe the stimulus design.

(2) In a similar vein, while C1 has canonical topographical properties at the grand-average level, these may differ substantially depending on individual anatomy (which the authors did not assess). This means that task-relevant information will be represented to different degrees in individuals' single-trial data. My guess is that this confound was mitigated precisely by choosing relatively extended stimulus arrays. But given the authors' impressive track record on C1 mapping and modeling, I was surprised that the underlying rationale is only roughly outlined. For example, given the topographies shown and the electrode selection procedure employed, I assume that the differences between upper and lower targets are mainly driven by stimulus arms on the main diagonal. Did the authors run pilot experiments with more restricted stimulus arrays? I do not mean to imply that such additional information needs to be detailed in the main article, but it would be worth mentioning.

We thank the reviewer for their thoughtful consideration of this issue about individual variability in C1 retinotopy. Indeed, as the reviewer anticipated we expected the large stimulus coverage to mitigate this issue and we think that our response to the point above and the changes we made to the manuscript in response address this point also. Although we did not show this in the manuscript, we did in fact find that C1 topography was much more similar across individuals than it has been in previous C1 experiments we have carried out with smaller stimuli.

However, we acknowledge the reviewer’s point that the signal measured at a specific electrode likely has a variable loading strength from the various gratings in the stimulus array and that the gratings of maximal loading may indeed vary from subject to subject. Such inter-subject variability cannot confound the choice probability effects because the latter are measured within-subject. Nevertheless, it could be a source of noise. We believe the impact of this is unlikely to be substantial for the following reasons:

i) We designed the spatial spread of contrasts in such a way as to encourage participants to aggregate across the full array. In essence, to match the property of the C1 as an aggregate measure of V1 activity, we designed a task that involved aggregating across stimulus elements. Therefore, the decision weighting applied to any particular grating should be representative of the weighting applied to all gratings and, as such, the specific gratings that contribute most to the C1 signal for a particular participant should be relatively inconsequential.

ii) By avoiding the horizontal and vertical meridians we avoided the regions of space where the shifts in C1 topography are largest.

(3) Also, the stimulus arrangement disregards known differences in conduction velocity between the upper and lower visual fields. While no such differences are evident from the maximal-electrode averages shown in Figure 1B, it is difficult to assess this issue without single-stimulus VEPs and/or a dedicated latency analysis. The authors touch upon this issue when discussing potential pre-C1 signals emanating from the magnocellular pathway.

Indeed, there are important differences in V1 properties between the upper and lower visual fields, visual acuity being another example in addition to conduction velocity as the reviewer points out. However, these differences appeared to be quite minimal in this case (Figure 1B does in fact include a single-stimulus VEP – the “1-stim” entry in the legend). Perhaps this is also due to the large stimulus array which may include a range of conduction velocities within it and thereby blur overall differences between the upper and lower visual field. The variability of contrast within each array was also quite high (+/-20% from the midpoint), which would have further increased within-array conduction velocity variability and blurred differences between arrays.

Our staircasing procedure may have also helped in this regard to some extent as it included a bias parameter between the arrays to account for any behavioural response biases. Although the small contrast changes it usually incurred are likely much too small to change conduction velocities, it corrected for any effect on behaviour they may have.

(4) I suspect that most of these issues are at least partly related to a lack of clarity regarding levels of description: the authors often refer to 'information' contained in C1 or, apparently interchangeably, to 'visual representations' before, during, or following C1. However, if I understand correctly, the signal predicting (or predicted by) behavioral choice is much cruder than what an RSA-primed readership may expect, and also cruder than the other choice-predictive signals entered as control variables: namely, a univariate difference score on single-trial data integrated over a 10 ms window determined on the basis of grand-averaged data. I think it is worth clarifying and emphasizing the nature of this signal as the difference of aggregate contrast responses that *can* only be read out at higher levels of the visual system due to the limited extent of horizontal connectivity in V1. I do not think that this diminishes the importance of the findings - if anything, it makes them more remarkable.

This is true that a univariate measure may stick out in a field increasingly favouring multivariate analyses with the spread of machine learning, and so we have added a short qualifier in the methods section where we describe the C1 measurement to explicitly state that it is a scalar variable. What we have done in using this univariate measure is leverage the rich prior knowledge about V1 anatomy and neurophysiology, rather than trust in data-driven classifiers; interestingly, we found that such a classifier trained on all electrodes discriminates choices less well than our informed univariate measure during the C1 time-frame. 

We also thank the reviewer for raising an interesting point about the nature of aggregation and readout in the context of our stimulus. We agree that it is not feasible that V1 activity would be aggregated locally in V1 across such large regions of space prior to being readout within the C1 time period. As we say above, the aggregation may instead be carried out through convergent transmission of the parallel, spatially-local V1 information to the decision process.

(5) Arguably even more remarkable is the finding that C1 amplitudes themselves appear to be influenced by choice history. The authors address this issue in the Discussion; however, I'm afraid I could not follow their argument regarding preparatory (and differential?) weighting of read-outs across the visual hierarchy. I believe this point is worth developing further, as it bears on the issue of whether C1 modulations are present and ecologically relevant when looking (before and) beyond stimulus-locked averages.

We thank the reviewer for their positive appraisal of this additional finding, which we also found remarkable. We agree that our description of our interpretation was too brief and lacked clarity. We have reworded it and expressed it in terms of the speed accuracy trade-off, with the new explanation given below. However, it is important to remember that this account is speculative and serves only to explain the response-time contingency of the bias. That the bias was present and constitutes a modulation of the C1 does not rest on this argument:

[…] “to explain the RT contingency for the C1 bias, we speculate that the speed-accuracy trade-off could fluctuate from trial to trial and that the corresponding decision bound fluctuations (Heitz and Schall 2012) could be implemented by pre-determining decision weights across visual areas. For example, to achieve faster decisions, the sensory evidence requirement could be reduced by placing greater emphasis on initial afferent V1 evidence. In such a case, the RT contingency of the above choice history bias could be explained if the C1 bias is exerted in proportion with the planned emphasis of C1 evidence for the upcoming decision.”

Recommendations to the Authors:

Reviewer #2 (Recommendations for the authors):

(1) As someone whose first language is not English, I am somewhat hesitant to bring this up, but I found the use of 'readout' as both noun and verb somewhat confusing. I thought read-out was defined as 'that which is read out'.

We agree that this dual use of the word readout may cause confusion. To avoid this, we have edited the manuscript to replace verbal forms of the word “readout” with “read out”.

(2) I found it difficult to follow the reasoning for why intermediate RTs should be the ones most affected by C1-related information. Perhaps this could be described in more detail for the uninitiated reader.

We appreciate that our reasoning for why intermediate RTs should be the ones most affected by C1-related information was difficult to follow. We have now added a simulation to showcase this rationale more clearly - see response to reviewer 1, and new figure supplement to figure 1. 

(3) It would be interesting to compare the effect sizes observed here to those seen in single-cell studies and to discuss this comparison with regard to differences in the nature of EEG signals and single-cell firing rates.

While we agree that such a comparison would be interesting if feasible, it would have to be for the same task settings, which have not been used in a single-cell study, and  the very different nature and extent of noise between the two recording modalities would make such a comparison difficult to interpret, e.g. background noise in EEG from ongoing processes unrelated to the task. 

(4) Figure 1: It may be worth mentioning in the legend that only parts of the peripheral stimulus grid are shown for better visibility, as the Methods speak of 9 x 9 grids. Also, in panel B, it should be mentioned that waveshapes are calculated using individually selected maximal-difference electrodes.

We thank the reviewer for spotting these. We have updated the caption for this figure to reflect these two observations.

(5) Figure 4: The different shades of green may be difficult to distinguish when printed.

Although this may be true, we chose shades of green that differ in luminance so they should still be distinguishable. Different colours may in fact be less distinguishable if they had the same luminance and the print was black-and-white. We chose different shades of the same colour to reflect the fact that we were plotting the same signals at different difficulty levels. In our opinion, this takes precedence since eLife is an online journal so the majority of readers will likely read it digitally.

(6) Methods/Task: While the ITI of 780 ms is substantial, I was wondering why the authors decided against jittering this interval? It would be helpful to briefly discuss whether contrast adaptation for slow periodic stimulation may have affected the findings.

We opted against jittering the ITI to avoid an additional source of inter-trial variability. While this may allow for adaptation effects of this source, this would be approximately constant across trials and therefore less of a concern for our design. We have added text to the methods section to state this rationale.

(7) Methods/Stimuli: The authors convincingly argue that focusing on single arms of the stimuli is an unlikely strategy, but did they ask for participants' strategies during debriefing?

We are glad that the reviewer found our argument about whether or not participants may have focused on a single arm of the stimuli convincing. We did not ask participants about their strategies but even with such a debriefing, there would still remain a possibility that a participant may have used that strategy but were unaware that they were doing so. In any case, if participants were doing this it would have dampened the strength of our choice probability result. 

(8) Methods/Procedure, Difficulty Titration: Why did the authors opt for manually adapting the difficulty level in a separate session rather than constantly and automatically titrating difficulty?

We did this because calculating choice probability requires a comparison of trials with different choice outcomes but the same stimulus so continuously staircasing difficulty level during the experiment would have created a confound. Although this could have been corrected for in our regression, this would have entailed greater noise that we could avoid by staircasing in advance.

eLife Assessment

This important study links allelic expression imbalance with replication timing, suggesting a stochastic model for haploinsufficiency in dosage-sensitive disease. The integration of allele-specific RNA-seq and replication timing in clonal systems provides solid evidence for an association between asynchronous replication and allelic imbalance, although the scope and generality of some conclusions require more cautious interpretation. This study will interest epigeneticists and genome regulation researchers studying replication timing and monoallelic expression, as well as developmental biologists and human geneticists concerned with clonal heterogeneity, haploinsufficiency, and variable disease penetrance.

[Editors' note: this paper was reviewed by Review Commons.]

Reviewer #1 (Public review):

Summary:

The existence of VERT regions is well supported, but the number of regions called as ISCs may be inflated by permissive thresholds (e.g., AEI {greater than or equal to} 0.8 or {less than or equal to} 0.2 in a single clone). This risks conflating transient stochastic differences with stable ISCs. Similarly, the claim of cell-type specificity is not convincingly demonstrated given the small sample size (n=4) and strong batch confounding between lymphoblastoid and cartilage progenitors. While syntenic VERT regions across mouse and human are intriguing, they complicate interpretation of strong clustering by cell type. Sampling depth may also have exaggerated allelic imbalance calls.

The proposed role of ISCs in haploinsufficiency is conceptually interesting but remains speculative; developmental stochasticity and founder population size may play larger roles than replication timing. The claim that autosomal inactivation is mechanistically distinct from XCI, however, is reasonable and supported.

Some conclusions should be more explicitly qualified as preliminary. Cell-type specificity and mitotic stability both require stronger evidence; the latter is inferred indirectly from clonal expansion rather than shown directly, and orthogonal experiments (e.g., allele-specific ChIP-seq, DNA methylation) would be required. Estimated genomic coverage of ISCs should also be re-evaluated, as single-clone observations may inflate counts.

Replication is limited. Hierarchical clustering is confounded by batch and based on presence/absence calls that lack quantitative resolution. More robust approaches would include using magnitude of imbalance, annotating VERTs by genomic location, applying stricter thresholds for replication timing, and benchmarking AEI distributions against the X chromosome. These are realistic re-analyses requiring no new data and could be completed in ~1 month.

Methods are generally well described and reproducible. Figures and text would benefit from improved clarity: axis labels are missing in places (e.g., Fig. 1c, Fig. 2g), legends should explain chromosome arm colors, and cluttered figures such as Fig. 1j could be re-visualized for interpretability. Gene set enrichment analysis should be restricted to avoid inflated significance from overly broad categories. A useful citation for XCI timing (pmid=39420003) could be added to strengthen background.

Significance:

Conceptually, this work introduces ISC-like phenomena in human and mouse progenitor lines, coupling allelic expression imbalance with replication timing. Technically, it combines allele-specific RNA-seq with Repli-seq in genotyped, clonal, single-cell-derived lines. Clinically, it suggests an alternative model for haploinsufficiency, relevant to dosage-sensitive diseases where stochastic transcriptional delays could shape penetrance.

The study builds on prior work in allelic exclusion (e.g., HLA, olfactory receptors) and random monoallelic expression, generalizing these phenomena into ISC/vert frameworks and proposing mitotic stability of allele choice. By extending beyond expression to replication timing, the authors suggest a broader paradigm for epigenetic regulation at autosomal loci.

The paper will be of interest to epigeneticists studying XCI, allelic exclusion, and monoallelic expression; to developmental biologists examining replication timing and differentiation; and to clinicians concerned with dosage-sensitive and haploinsufficient disorders.

Reviewer #2 (Public review):

Summary:

- This is a complicated research topic that touches on a few sub-fields of biology, and thus to make the paper more approachable I would recommend a careful edit of the text for clarity and precision of language.<br /> - Authors point out that this is a decades-old field; it would make sense to use terminology established within the field rather than inventing their own. Allelic imbalance has been referred to as AI, MAE (monoallelic expression), RMAE (random monoallelic expression) etc. The paper whose mouse data the authors make use of uses Asynchronous Stochastic Replication Timing (ASRT) instead of VERT to refer to the same phenomenon. Creating unnecessary jargon makes the paper more difficult to read and adds needless complexity to an already complex field.<br /> - Methods do not provide sufficient detail to fully evaluate or reproduce these experiments.<br /> - It is helpful to show representative loci as the authors do in Fig 1F and G and Fig 2, but these panels are very densely rendered and thus difficult to process visually - even the cartoon version (1D) is thick with overlapping lines. The point that allelic imbalance is enriched in VERTs would be enhanced if the authors could present the allelic ratio for all genes found in all VERTs, demonstrating how replication timing on either chromosome affects the allelic ratio.<br /> - The authors make the important point that VERTs are unlikely to be shared among different cell types and tissues (Fig 1i) but then find an enrichment for neuronal and immune genes in VERT regions identified in ACPs. It follows that these same genes are unlikely to be in such regions in the tissues where they are relevant. Some of the GO terms presented are too broad to suggest any biological significance to the result, even if there is statistical significance (for example, the top term for LCL clones 'Cytoplasm' is associated with 12,000 genes, and the second term for mouse clones 'Membrane' is associated with 10,000). It would be helpful to focus on GO terms lower in the GO hierarchy.<br /> - Figure 3 highlights the association of related gene clusters with VERTs but the VERTs are assigned based on variable replication timing in just 1 or 2 clones. This is an interesting observation, but to make the point that "VERT regions frequently coincide with gene clusters in the human genome" there needs to be a systematic assessment of replication timing at all gene clusters across all clones, and a statistical test for significance.<br /> - It is an interesting hypothesis that VERTs are conserved between species at synentic loci. If such regions are really conserved, one would expect that replication timing at these sites would be consistently asynchronous. However, the data presented shows that in human clones these VERTs can be specific to an individual donor (as in 5A) or an individual clone (as in 5H).<br /> - Again, the finding that VERTs coincide with neurodevelopmental disease genes in immune and cartilage cells is at odds with the previous statements and data about the tissue specificity of VERTs. In order to support the claim that neurodevelopmental disease associated genes reside in asynchronously replicating regions, and are thus more prone to allelic imbalance, the authors would need to demonstrate this phenomenon in neuronal cells.

Significance:

The authors pair analysis of replication timing and allele-specific expression in clonal populations of primary human cells. They combine these data with previously published data on clones from transformed human cell lines. They identify a number of genomic regions that display asynchronous replication timing in at least one clone and correlate these regions with allele-specific expression of genes within them. They also observe that several interesting gene sets, including genes that are associated with human diseases, map to asynchronously replicating regions. This is a good experimental approach that builds on already published data demonstrating the connection between allelic imbalance and replication timing. However, the authors consistently lean on thin evidence (i.e. a single clone) within a modestly sized dataset (4 clones from 2 donors each) to propose a new model for haploinsufficiency in human disease. The consistent focus on limited elements in the data and perhaps an overreach in the interpretation makes it difficult to appreciate what is in fact a very good experiment.

Author response:

General Statements

We thank the reviewers for their thoughtful and constructive comments, which will substantially improve our manuscript. In response, we will revise the text and figures throughout to address the points raised. Specifically, we will:

i. Refine our definition of Inactivation/Stability Centers (I/SCs): We will limit this designation to loci where both Allelic Expression Imbalance (AEI) and Variable Epigenetic Replication Timing (VERT) are detected, either in the present study or in previously published work.

ii. Expand methodological clarity: We will provide detailed descriptions of how VERT regions were identified, annotated, and quantified, including thresholds for allelic imbalance, replication timing variability, and sampling depth. We also justify the ≥80% AEI cutoff, which is based on recent studies showing that modest allelic biases can have biological and clinical significance.

iii. Enhanced benchmarking and validation: In addition to the analysis of X inactivation in female ACP cells, we will include comparisons between imprinted and non-imprinted regions to benchmark the magnitude of allelic replication timing imbalance, demonstrating that the magnitude of imbalance observed at imprinted loci is comparable to that at the non-imprinted VERT regions.

iv. Address tissue specificity and sampling limitations: We will discuss the limited number of clones, tissues, and individuals analyzed, emphasizing that while our data identify robust AEI and VERT patterns, additional tissues and individuals will be required to capture the full diversity of I/SC regulation.

v. Clarify biological relevance: We will expand our discussion to highlight the consistency of AEI findings across cell types, including examples of genes implicated in neurodevelopmental and neurodegenerative disorders, and we will clarify our model of how I/SC regulation may contribute to haploinsufficiency, variable expressivity, and incomplete penetrance in human disease.

vi. Improved figures and supplemental data: We will update figure legends for clarity, add a new supplementary figure comparing imprinted and non-imprinted regions, and cross-reference all supplemental tables.

We believe these revisions strengthen the manuscript conceptually and experimentally, and we thank the reviewers and editors for their valuable feedback.

Description of the planned revisions

Reviewer #1:

The existence of VERT regions is well supported, but the number of regions called as ISCs may be inflated by permissive thresholds (e.g., AEI {greater than or equal to} 0.8 or {less than or equal to} 0.2 in a single clone). This risks conflating transient stochastic differences with stable ISCs.

We selected the >80% (or <20%) allelic imbalance threshold, along with the requirement of at least one biallelic clone, as our criterion for significant AEI. This choice was guided by a recent study demonstrating that allelic imbalance as low as a 65%/35% is enough to effect disease penetrance in humans (Nature 2025; 637:1186–1197). For completeness, results obtained using more stringent thresholds (>90% and >95% imbalance) are presented in Supplementary Table 2.

Furthermore, it is unlikely that transient stochastic differences in allelic expression, such as those detected by single-cell RNA sequencing assays (Nat. Rev. Genet. 2015; 16:653–664), would be captured by our approach. Each clone in our study was expanded from a single cell to over one million cells before both RNA-seq and Repli-seq analysis, effectively averaging out transient transcriptional and/or replication fluctuations, and thus reflecting stable, mitotically heritable epigenetic states.

More robust approaches would include using magnitude of imbalance, annotating VERTs by genomic location, applying stricter thresholds for replication timing, and benchmarking AEI distributions against the X chromosome.

All VERT regions identified in this study were annotated according to both the magnitude of allelic imbalance and their genomic coordinates, using 250 kb windows for the human samples and 50 kb windows for the mouse samples (see Supplementary Tables 1 and 6). Figure 1c directly compares the magnitude of imbalance, defined as outliers in the standard deviation, for both allelic replication timing and allelic expression across autosomal and X-linked loci in female ACP cells.

In addition, we will benchmark the magnitude of replication timing imbalance using autosomal imprinted regions as a second internal control. We detected allelic replication imbalance at 13 known imprinted loci, and the standard deviation of replication timing at these loci, measured in 250 kb windows, is comparable to that observed across the >350 VERT regions detected at non-imprinted sites. To illustrate this comparison, we will include a supplementary figure directly comparing imprinted and non-imprinted regions.

Figures and text would benefit from improved clarity: axis labels are missing in places (e.g., Fig. 1c, Fig. 2g), legends should explain chromosome arm colors, and cluttered figures such as Fig. 1j could be re-visualized for interpretability.

Figure labels will be added to Figs. 1c and 2g, and legends will be modified for clarity.

“the claim of cell-type specificity is not convincingly demonstrated given the small sample size (n=4) and strong batch confounding between lymphoblastoid and cartilage progenitors.” And “Hierarchical clustering is confounded by batch and based on presence/absence calls that lack quantitative resolution.”

We agree that the limited number of individuals and clones, as well as the comparison between only two distinct tissue types (LCLs and ACPs), have quantitative limitations. Our primary intent was to evaluate whether any I/SCs were shared between independently derived clonal datasets and to determine whether there is evidence of tissue-specific I/SC usage, rather than to make quantitative claims about global cell-type specificity.

To address this concern, we will replace the hierarchical clustering analysis currently shown in Figure 1i with a Venn diagram that more directly illustrates the overlap and tissue-specific distribution of VERT regions detected in the different clonal sets. This revised representation avoids assumptions about clustering relationships and removes batch-driven bias, while still conveying the key observation that many VERT regions are shared across tissues and others appear tissue-restricted.

While syntenic VERT regions across mouse and human are intriguing, they complicate interpretation of strong clustering by cell type. Sampling depth may also have exaggerated allelic imbalance calls.

We note that the human LCLs used in our study are B cells, and immunoglobulin gene rearrangements were used to confirm the clonal uniqueness of each line. Similarly, the mouse replication timing data analyzed here was generated from pre-B cells, which also undergo immunoglobulin gene rearrangement. Thus, both the human LCL and mouse pre-B cell datasets were derived from B-cell lineages, providing a consistent cellular context for comparative analysis.

Sequencing depth is an important consideration for all variant base calls. Without fully haplotype-resolved genomes, previous studies relied on calculating per-SNP calls of allelic imbalance based on reads covering a single nucleotide locus. To improve sequencing depth supporting the identification of VERT and AEI regions, we utilized fully haplotype-resolved genomes that allowed all informative allele-specific reads to be pooled across all heterozygous SNPs within genomic windows or expressed genes. For AEI, we set a minimum threshold of 20 informative allele-specific reads per gene, a minimum FDR-corrected p-value of <=0.05, and a minimum of 80% vs 20% allelic imbalance. Importantly, a recent study has shown that allelic imbalance as low as a 65%/35% is enough to effect disease penetrance in humans (Nature 2025; 637:1186–1197). We reiterate that more stringent thresholds (>90% and >95% imbalance) are presented in Supplementary Table 2.

Gene set enrichment analysis should be restricted to avoid inflated significance from overly broad categories.

Reviewer #2:

Some of the GO terms presented are too broad to suggest any biological significance to the result, even if there is statistical significance (for example, the top term for LCL clones 'Cytoplasm' is associated with 12,000 genes, and the second term for mouse clones 'Membrane' is associated with 10,000). It would be helpful to focus on GO terms lower in the GO hierarchy.

We will include our complete Gene Ontology analysis, with more specific biological categories, in Supplemental Table 5.

Allelic imbalance has been referred to as AI, MAE (monoallelic expression), RMAE (random monoallelic expression) etc. The paper whose mouse data the authors make use of uses Asynchronous Stochastic Replication Timing (ASRT) instead of VERT to refer to the same phenomenon. Creating unnecessary jargon makes the paper more difficult to read and adds needless complexity to an already complex field.

While we agree that allelic expression imbalance has been described by different investigators using many different phrases, we believe that MAE, RMAE and AI do not represent an accurate description of the phenomenon. In our study [and our previous study; Nat Commun. 2022; 13(1):6301] we used clonal analysis of allele-specific expression and found that while some clones display equivalent levels of expression between alleles of a given gene (i.e. bi-allelic expression) other clones express only one allele (i.e. mono-allelic expression), and yet other clones have undetectable expression (i.e. silent on both alleles). This pattern of allele-restricted expression indicates that each allele independently adopts either an expressed or silent state. Importantly, because these expression states are mitotically stable, allele-autonomous, and independent of parental origin, we refer to the choice of the expressed allele as stochastic. Given this variability, we believe that the phrase “Allelic Expression Imbalance” (AEI) represents a more accurate descriptor for this phenomenon. We also point out that “Allelic Expression Imbalance” has been used >120 times in the Pubmed database.

In addition, the replication asynchrony that exists at these loci is not consistent with purely ASynchronous Replication Timing (ASRT) between alleles. We found that each allele can independently adopt either earlier or later replication timing in different clones. This variability results in some clones exhibiting pronounced asynchrony between alleles, while in others, the two alleles replicate synchronously, with both adopting either the earlier or later timing state. As reported in our previous study (Nat. Commun. 2022; 13:6301), this behavior reflects a stochastic and allele-autonomous process, leading us to describe these loci as exhibiting Variable Epigenetic Replication Timing (VERT), which we believe is a more accurate descriptor of this phenomenon.

The point that allelic imbalance is enriched in VERTs would be enhanced if the authors could present the allelic ratio for all genes found in all VERTs, demonstrating how replication timing on either chromosome affects the allelic ratio.

The stochastic nature of allelic expression and replication timing observed at VERT loci indicates that each allele independently acquires its epigenetic state. Specifically, the expressed or silent status of one allele does not predict the replication timing or expression status of the opposite allele. Accordingly, the Early/Late pattern of replication timing that we detect, both in this study and in our previous work (Nat. Commun. 2022; 13:6301), is not correlated with which allele is transcriptionally active. This supports our conclusion that asynchronous replication timing is not a downstream consequence of monoallelic transcription, but rather an independent epigenetic feature of I/SCs. Regardless, we will provide the combined expression ratios for all transcripts that are located within the VERT regions in a Supplemental Table.

In addition, our analysis of imprinted loci reveals that even at genomic regions with parent-of-origin–specific expression, replication timing does not align with allelic activity: both early- and late-replicating alleles can be transcriptionally active, depending on the gene. This observation is consistent with the complex organization of many imprinted domains, where genes on opposite alleles exhibit reciprocal expression patterns. To illustrate this point, we will include a new supplemental figure demonstrating that imprinted loci harbor genes expressed from both the earlier- and later-replicating alleles.

Figure 3 highlights the association of related gene clusters with VERTs but the VERTs are assigned based on variable replication timing in just 1 or 2 clones. This is an interesting observation, but to make the point that "VERT regions frequently coincide with gene clusters in the human genome" there needs to be a systematic assessment of replication timing at all gene clusters across all clones, and a statistical test for significance.

Our intent in Figure 3 was not to suggest that all gene clusters are subject to VERT and AEI, but rather to highlight that several well-characterized multigene families that are known to exhibit random AEI, such as olfactory receptor and HLA gene clusters, coincide with VERT regions at their genomic locations. These examples serve as representative illustrations demonstrating that I/SC-associated regulation occurs at established AEI loci organized in gene clusters.

To clarify this point, we will revise the text to explicitly state that Figure 3 presents illustrative examples of known AEI-associated gene clusters overlapping with VERT regions, rather than a comprehensive or statistically exhaustive analysis of all gene clusters across the genome.

It is an interesting hypothesis that VERTs are conserved between species at synentic loci. If such regions are really conserved, one would expect that replication timing at these sites would be consistently asynchronous. However the data presented shows that in human clones these VERTs can be specific to an individual donor (as in 5A) or an individual clone (as in 5H).

As discussed in our Limitations section, our analysis was restricted to a limited number of cell types, clones, and individuals, which may not capture the full diversity of I/SC usage across tissues and populations. While our dataset was sufficient to identify robust patterns of AEI and VERT, it likely represents only a subset of the broader landscape of I/SC regulation in both humans and mice. We anticipate that future studies incorporating a wider range of tissues, individuals, and clonal analyses will uncover an even greater degree of conservation and diversity in I/SC usage across genomes.

In order to support the claim that neurodevelopmental disease associated genes reside in asynchronously replicating regions, and are thus more prone to allelic imbalance, the authors would need to demonstrate this phenomenon in neuronal cells.

We make two points that address this critique: First, many of the neurodevelopmental disease genes located within or adjacent to VERT regions are not exclusively expressed in neuronal cells and have already been shown to exhibit AEI in non-neuronal contexts. For example, Gimelbrant and Chess (Science, 2007; 318:1136–1140) demonstrated AEI of the Parkinson disease genes SNCA and LRRK2 in lymphoblastoid cell lines (LCLs), and in our previous study, we detected AEI of DNAJC6, another Parkinson disease gene, in LCL cells (Nat. Commun. 2022; 13:6301). In the present study that used ACP cells, we identified VERT and AEI of several epilepsy-associated genes, including SCN1A, SCN2A (Fig. 6b), GABRA1(Fig. 6e), and SAMD12 (Fig. 6j), as well as a gene implicated in autism and neurodevelopmental disorders, SEMA5A (Fig. 5c).

Second, independent studies from the E. Heard laboratory have provided further evidence that AEI occurs in neuronal lineages. Using mouse neural progenitor cells (NPCs), they identified genes subject to AEI (Dev. Cell, 2014; 28:366–380) and they later evaluated AEI of syntenic human neurodevelopmental disease genes, including Snca, App, Eya4, and Grik2 (Nat. Commun. 2021; 12:5330). In addition, they used the phrase “Allelic Expression Imbalance” to describe the epigenetic expression biases at these genes.

Together, these findings reinforce that AEI, and by extension I/SC regulation, is not restricted to specific cell types, but rather represents a generalizable mechanism of stochastic epigenetic regulation that includes genes relevant to neurodevelopment and disease.

However, the authors consistently lean on thin evidence (i.e. a single clone) within a modestly sized dataset (4 clones from 2 donors each) to propose a new model for haploinsufficiency in human disease. The consistent focus on limited elements in the data and perhaps an overreach in the interpretation makes it difficult to appreciate what is in fact a very good experiment.

We agree that our analysis was conducted on a modest number of clones and individuals, which we explicitly acknowledge as a limitation of the present study. However, several key points support the robustness and broader relevance of our conclusions:

i. Clonal Design and Replication: The strength of our approach lies in its clonal resolution. Each clone represents a single-cell–derived population expanded to over a million cells, enabling direct detection of stable, mitotically heritable allele-specific epigenetic states that would not be apparent in population-averaged data. Importantly, many of the VERT regions we identified are shared between independent clones from different donors and across distinct cell types (ACP and LCL), demonstrating reproducibility and biological consistency.

ii. Cross-Species Validation: We further identified syntenic VERT regions in mouse pre-B cell clones, including at loci known to exhibit AEI in prior studies, providing independent validation and evolutionary conservation of the phenomenon.

iii. Integration with Published Evidence: Our findings extend prior observations of AEI and variable replication timing (e.g. Gimelbrant et al. Science 2007; Heskett et al. Nat. Commun. 2022) and are fully consistent with known stochastic allelic expression imbalance of autosomal genes. We also draw parallels with the absence of cellular selection mechanisms that dictate dominant inheritance patterns for loss of function alleles for X linked disease genes (reviewed in: J Clin Invest, 2008, 20-23; and Nat Rev Genet. 2025, 26, 571–580). Our proposed model linking I/SC regulation to haploinsufficiency is therefore a synthesis of our results with an extensive body of published data, not an inference drawn from isolated observations.

iv. Scope and Framing: We will revise the manuscript to clarify that our proposed model represents a mechanistic framework, not a definitive or exclusive explanation, for how stochastic allelic regulation could contribute to dosage-sensitive disease phenotypes. We will also explicitly discuss the need for larger datasets and additional tissues to refine and test this model.

In summary, while we recognize the limited sampling inherent to clonal analyses, the consistency of our observations across donors, cell types, and species, together with prior corroborating studies, supports the validity of the conclusions and justifies the broader conceptual implications.

Description of analyses that authors prefer not to carry out

Reviewer #1:

Cell-type specificity and mitotic stability both require stronger evidence; the latter is inferred indirectly from clonal expansion rather than shown directly, and orthogonal experiments (e.g., allele-specific ChIP-seq, DNA methylation) would be required.

We disagree with this reviewer that the mitotic stability of the epigenetic states are “inferred indirectly from clonal expansion rather than shown directly”. Our experimental design inherently captures mitotically stable, allele-specific states because each clonal line is derived from a single progenitor cell and expanded to millions of cells before analysis. The allele-specific replication timing and expression profiles observed in these clones therefore reflect epigenetic states that are stably inherited across many cell divisions, rather than transient or stochastic fluctuations. This approach was also validated in our previous study (Nat. Commun. 2022; 13:6301), where the same clonal strategy demonstrated stable allele-restricted replication and expression patterns over extended passages.

We agree that orthogonal assays such as allele-specific ChIP-seq or DNA methylation analyses would provide additional mechanistic detail on the nature of I/SC-associated regulation. However, these experiments fall outside the scope of the present study, which was designed specifically to identify and map autosomal loci that exhibit coordinated AEI and VERT, the defining epigenetic features of I/SCs. While we fully acknowledge that defining the precise molecular marks (e.g., histone modifications, DNA methylation, chromatin accessibility) that underlie I/SC regulation will be an important future direction, our current data provide a genome-wide, allele-resolved foundation upon which such mechanistic studies can build.

In summary, the current dataset achieves the central goal of defining the genomic distribution and conservation of I/SCs based on functional readouts of replication timing and expression. Future work will extend these findings using allele-specific epigenomic profiling to characterize the epigenetic modifications associated with I/SC stability and cell-type specificity.

eLife Assessment

This manuscript reports the development and characterization of iGABASnFR2, a genetically encoded GABA sensor that demonstrates substantially improved performance compared to its predecessor, iGABASnFR1. The work is comprehensive and methodologically rigorous, combining high-throughput mutagenesis, functional screening, structural analysis, biophysical characterization, and in vivo validation. The significance of the findings is fundamental, and the supporting evidence is compelling. iGABASnFR2 represents a notable advance in GABA sensor engineering, enabling enhanced imaging of GABA transmission both in brain slices and in vivo, and constitutes a timely, technically robust addition to the molecular toolkit for neuroscience research.

Reviewer #1 (Public review):

Summary:

This manuscript by Kolb and Hasseman et al. introduces a significantly improved GABA sensor, building on the pioneering work of the Janelia team. Given GABA's role as the main inhibitory neurotransmitter and the historical lack of effective optical tools for real-time in vivo GABA dynamics, this development is particularly impactful. The new sensor boasts an enhanced signal-to-noise ratio (SNR) and appropriate kinetics for detecting GABA dynamics in both in vitro and in vivo settings. The study is well-presented, with convincing and high-quality data, making this tool a valuable asset for future research into GABAergic signaling.

Strengths:

The core strength of this work lies in its significant advancement of GABA sensing technology. The authors have successfully developed a sensor with higher SNR and suitable kinetics, enabling the detection of GABA dynamics both in vitro and in vivo. This addresses a critical gap in neuroscience research, offering a much-needed optical tool for understanding the most important inhibitory neurotransmitter. The clear representation of the work and the convincing, high-quality data further bolster the manuscript's strengths, indicating the sensor's reliability and potential utility. We anticipate this tool will be invaluable for further investigation of GABAergic signaling.

Weaknesses:

Despite the notable progress, a key limitation is that the current generation of GABA sensors, including the one presented here, still exhibits inferior performance compared to state-of-the-art glutamate sensors. While this work is a substantial leap forward, it highlights that further improvements in GABA sensor would still be highly beneficial for the field to match the capabilities seen with glutamate sensors.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript by Kolb and Hasseman et al. introduces a significantly improved GABA sensor, building on the pioneering work of the Janelia team. Given GABA's role as the main inhibitory neurotransmitter and the historical lack of effective optical tools for real-time in vivo GABA dynamics, this development is particularly impactful. The new sensor boasts an enhanced signal-to-noise ratio (SNR) and appropriate kinetics for detecting GABA dynamics in both in vitro and in vivo settings. The study is well-presented, with convincing and high-quality data, making this tool a valuable asset for future research into GABAergic signaling.

Strengths:

The core strength of this work lies in its significant advancement of GABA sensing technology. The authors have successfully developed a sensor with higher SNR and suitable kinetics, enabling the detection of GABA dynamics both in vitro and in vivo.

This addresses a critical gap in neuroscience research, offering a much-needed optical tool for understanding the most important inhibitory neurotransmitter. The clear representation of the work and the convincing, high-quality data further bolster the manuscript's strengths, indicating the sensor's reliability and potential utility. We anticipate this tool will be invaluable for further investigation of GABAergic signaling.

Weaknesses:

Despite the notable progress, a key limitation is that the current generation of GABA sensors, including the one presented here, still exhibits inferior performance compared to state-of-the-art glutamate sensors. While this work is a substantial leap forward, it highlights that further improvements in GABA sensors would still be highly beneficial for the field to match the capabilities seen with glutamate sensors.

We thank Reviewer 1 for the positive assessment. We agree that further improvements in GABA sensor performance remain desirable. We acknowledge this limitation and outline directions for future development in the Discussion paragraph beginning "There are several promising avenues that could be taken to further optimize iGABASnFR."

Reviewer #2 (Public review):

Summary:

This manuscript presents the development and characterization of iGABASnFR2, a genetically encoded GABA sensor with markedly improved performance over its predecessor, iGABASnFR1. The study is comprehensive and methodologically rigorous, integrating high-throughput mutagenesis, functional screening, structural analysis, biophysical characterization, and in vivo validation. iGABASnFR2 represents a significant advancement in GABA sensor engineering and application in imaging GABA transmission in slice and in vivo. This is a timely and technically strong contribution to the molecular toolkit for neuroscience.

Strengths:

The authors apply a well-established sensor optimization pipeline and iterative engineering strategy from single-site to combinatorial mutants to engineer iGABASnFR2. The development of both positive and negative going variants (iGABASnFR2 and iGABASnFR2n) offers experimental flexibility. The structure and interpretation of the key mutations provide insights into the working mechanism of the sensor, which also suggest optimization strategies. Although individual improvements in intrinsic properties are incremental, their combined effect yields clear functional gains, enabling detection of direction-selective GABA release in the retina and volume-transmitted GABA signaling in somatosensory cortex, which were challenging or missed using iGABASnFR1.

Weaknesses:

With minor revisions and clarifications, especially regarding membrane trafficking, this manuscript will be a valuable resource for probing inhibitory transmission.

We thank Reviewer 2 for the positive assessment. Regarding membrane trafficking, we appreciate the suggestion to test different trafficking motifs. While such optimization represents a valuable direction for future development, it was beyond the scope of the present study and not feasible with the available time and resources. A different imaging modality would be needed to assess membrane trafficking efficiency or membrane-restricted expression, as the images presented in the manuscript (Figure 2a) are wide-field epifluorescence images, which lack the axial resolution required to distinguish membrane-localized signal from cytosolic fluorescence.

We expect that the current characterization of iGABASnFR2 will nevertheless provide a strong foundation for future efforts to optimize membrane targeting and expression using alternative trafficking strategies.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) We noted an interesting inconsistency in the response of iGABASnFR1 and iGABASnFR2 when expressed as purified protein versus in mammalian cells. Such discrepancies are not uncommon for proteins exhibiting different behaviors in E. coli versus mammalian expression systems. We appreciate the authors' diligent effort in performing screening within a neuronal context. Similarly, the stark difference between the absolute affinity in purified form (∼0.778 μM) and on-cell measurements (6.4 μM) warrants further discussion. The authors may consider commenting on these observations in the discussion section.

We have revised the Discussion (lines 401-410 in the ‘Tracked Changes’ document) to address the discrepancy between measurements obtained with purified protein and those from expression on the neuronal surface. As noted by the reviewer, such discrepancies are common, and our revision is intended to convey our empirical experience with this phenomenon rather than to offer a definitive mechanistic explanation.

One factor to appreciate is that, when on the surface of neurons, the sensor is tethered to the membrane by an additional 60 amino acids. In addition to altering the local chemical environment, membrane tethering could impose entropic or mechanical constraints on the sensor. These constraints may damp conformational motions that underlie ligand binding and fluorescence changes. Beyond this, the local environment experienced by a membrane-anchored sensor differs substantially from that of soluble protein. There are potential electrostatic and steric effects arising from the plasma membrane and extracellular matrix, as well as post-translational modifications associated with mammalian expression. These effects on sensor performance are not readily predictable in either magnitude or direction, as illustrated by iGluSnFR, which exhibits a higher apparent affinity when membrane-tethered than in soluble form (Aggarwal et al 2023). For these reasons, we place greater emphasis on neuronal measurements as the most informative indicator of in vivo sensor performance.

(2) Although iGABASnFR2 fluorescence exhibits pH dependence, its response appears less pH-dependent compared to the first-generation sensor. To enhance clarity, we suggest plotting the normalized response of both sensors across different pH values. This visual representation would be highly informative for readers.

Thank you - we have implemented this, now showing the (F_sat - F_apo)/F_apo response as a function of pH for all three sensors in Fig 4 fig. supp 3b. This visualization nicely illustrates that the apo-to-sat response of iGABASnFR1 is much more influenced by pH than either iGABASnFR2 or iGABASnFR2n, which we note on lines 252-253 of the ‘Tracked Changes’ document.

(3) To provide a more comprehensive characterization of the sensors, we recommend including a quantification of the decay times for all three versions of the sensors in Figure 2, specifically after panel 2c.

Thank you - we now provide this in Fig 2d.

(4) For improved readability of Figure 3a, we suggest adding distinct labels for iGABASnFR1 and iGABASnFR2 with corresponding colors.

Good suggestion - we matched the color of the backbones to the rest of the manuscript (orange and green). We also added labels on the figure to ensure clarity.

(5) The GABA released by SAC cells in Figure 5 looks amazing! We propose a minor modification to the cartoon in Figure 5b: mirroring the image horizontally (left to right). Given that the subsequent panels (e, h, and k) set the preferred direction of SAC movement as rightward, the current cartoon in Figure 5b inadvertently suggests stronger inhibition by SAC-released GABA when the spot moves left. Mirroring the image would align the cartoon more accurately with the subsequent data representations.

Thanks - this is a nice streamlining. We have implemented the change.

Reviewer #2 (Recommendations for the authors):

(1) As sensor performance differs substantially between purified protein and neurons, a summary table comparing key properties (e.g., EC50, ∆F/F _ax, response amplitude to # of AP) across purified protein and neurons would be highly informative.

We discuss differences in sensor performance between purified protein and neurons in the Discussion (lines 401-410 in ‘Tracked Changes document) and, for the reasons outlined there, consider neuronal measurements to be far more predictive of in vivo performance. We therefore chose not to include a summary table directly comparing purified protein and neuronal data, as this would risk over-emphasizing in vitro measurements that we view primarily as qualitative signposts rather than more directly informative indicators of functional performance.

(2) The authors should comment on the observed differences in performance between purified protein and neuronal expression. Would HEK293 cell measurements serve as a better predictor of in vivo performance than in vitro titrations? Insights here would benefit future sensor development pipelines.

We have revised the Discussion to address this point (lines 401-410 in the ‘Tracked Changes’ document). We often observe differences in sensor performance between purified protein measurements and cellular or in vivo contexts. In our experience, titrations in primary neurons provide a better predictor of in vivo performance than in vitro protein titrations, as they more closely reflect relevant cellular factors. We do not have direct evidence that expression in heterologous systems such as HEK293 cells is generally more predictive, although this seems plausible; however, predictions inevitably become less reliable as sensors are translated to fully in vivo conditions.

(3) Improved membrane localization likely contributes to the enhanced sensitivity of iGABASnFR2 in neurons beyond changes in EC50. In Figure 2a, membrane trafficking appears suboptimal. The authors should explore alternative trafficking motifs (e.g., ER2, Kv2.1, or motifs from other sensors) to further improve the membrane expression and consider adding a second fluorescent protein for quantifying membrane-localized brightness.

Figure 2a presents wide-field epifluorescence images, which lack the axial resolution required to distinguish membrane-localized signal from cytosolic fluorescence. We therefore do not consider this imaging modality suitable for assessing membrane trafficking efficiency or membrane-restricted expression.

We appreciate the suggestion to test different trafficking motifs to attempt to better capture biological signals. While such optimization represents a valuable direction for future development, it was beyond the scope of the present study and not feasible with the available time and resources. We expect that the current characterization of iGABASnFR2 will nevertheless provide a strong foundation for future efforts to further optimize membrane targeting and expression using alternative trafficking strategies.

(4) Figure 4 - Supplement 2: The apparent EC50 of iGABASnFR2 seems affected by buffer composition and the presence of high concentrations of unrelated compounds. The authors should comment on this.

We thank the reviewer for raising this point. Upon closer inspection, the EC50 of iGABASnFR2 in Fig 4 Supp 2 is measured at 1.4 μM, while in Fig 4a it is 1.1 μM - these mean values are quite close to one another, and within the range of experimental variability we expect for experiments done weeks or months apart. What differs most noticeably in this dataset is the shape of the dose–response curve rather than the EC50 itself; the origin of this difference is currently unclear. We have revised the Results text (lines 226-231 in ‘Tracked Changes document) to clarify this point and to emphasize that the key observation of Fig. 4–figure supplement 2 is that none of the additional compounds tested substantially impair GABA binding, indicating that they do not act as strong non-competitive allosteric antagonists or inhibitors.

(5) The negative-going variant, iGABASnFR2n, is introduced but only briefly characterized. Including additional data or even a conceptual use case would clarify its potential utility.

We have modified the discussion to provide more examples of conceptual use cases, clarifying how such a sensor could indeed be highly impactful. The full passage is lines 372-387 in the ‘Tracked Changes’ document; to summarize: a key application of the negative-going sensor is detecting decreases in ‘GABA tone’, which plays a key role in setting the excitation-inhibition balance across brain circuits. Reductions in extrasynaptic GABA are a well-documented feature of several biologically important brain-state transitions, including arousal, experience-dependent plasticity, and stress-related modulation of inhibition, and iGABASnFR2n could be an important tool for investigating these processes.

eLife Assessment

In this important contribution, Yan and colleagues describe a powerful and compelling strategy to generate concatamers of the BK channel and their fusion constructs with the auxiliary gamma subunits, which allows exploring contributions of individual subunits of the tetrameric channel to its gating and the study of heteromeric channel complexes of defined composition. Distinct examples are presented, which illustrate great diversity in the stoichiometric control of BK channel gating, depending on the site and nature of molecular perturbations. The molecular approaches could be extended to other membrane proteins whose N and C termini face opposite sides of the membrane.

Reviewer #1 (Public review):

Summary:

BK channels are widely distributed and involved in many physiological functions. They have also proven a highly useful tool for studying general allosteric mechanisms for gating and modulation by auxiliary subunits. Tetrameric BK channels are assembled from four separate alpha subunits which would be identical for homozygous alleles and of potentially five different combinations for heterozygous alleles Geng et al . (2023), (https://doi.org/10.1085/jgp.202213302). Construction of BK channels with concatenated subunits in order to strictly control heteromeric subunit composition had not yet been used because the N-terminus in BK channels is extracellular whereas the C-terminus is intracellular. In this new work, Chen, Li, and Yan devise clever methods to construct and assemble BK channels of known subunit composition, as well as to fix the number of γ1 axillary subunits per channel. With their novel molecular approaches, Chen, Li and Yan report that a single γ1 axillary subunit is sufficient to fully modulate a BK channel, that the deep conducting pore mutation L312A exhibited a graded effect on gating with each addition mutated subunit replacing a WT subunit in the channel adding an additional incremental left shift in activation, and that the V288A mutation at the selectivity filter must be present on all four alpha subunits in order to induce channel inactivation. Chen, Li, and Yan have been successful in introducing new molecular tools to generate BK channels of known stoichiometry and subunit composition. They validate their methods and provide three different examples of stoichiometric modulation by LRRC26, the selectivity filter, and the pore.

Strengths:

Powerful new molecular tools for study of channel gating are developed and validated in the study.

Weaknesses:

One example each of auxiliary, deep pore, and selectivity filter allosteric actions are presented, but this is sufficient for the purposes of the paper to establish their methods and present specific examples of applicability.

Reviewer #2 (Public review):

Summary:

This manuscript describes novel BK channel concatemers as a tool to study the stoichiometry of gamma subunit and mutations in modulation of the channel. Taking the advantage of modular design of BK channel alpha subunit the authors connected S1-S6/1st RCK as two- and four-subunit concatemers and coexpressed with S0-RCK2 to form normal function channels. These concatemers avoided the difficulty that the extracellular N-terminus of S0 was unable to connect with the cytosolic C-terminus of the alpha or gamma subunit, allowing a single gamma subunit to be connected to the concatemers. The concatemers also helped reveal the required stoichiometry of mutant BK subunits in modulating channel function. These include L312A in the deep pore region that altered channel function additively with each additional subunit harboring the mutation, and V288A at the selectivity filter that altered channel function cooperatively only when all four subunits being mutated. These results demonstrate that the concatemers are robust and effective in studying BK channel function and molecular mechanisms related to stoichiometry. The different requirement of the gamma subunit and the mutations stoichiometry for altering channel function is interesting, revealing fundamental mechanisms of how different motifs of the channel protein control function.

Strengths:

The manuscript presents well designed experiments with high quality data, which convincingly demonstrate the BK channel concatemers and their utility. The results are clearly written.

Weaknesses:

This reviewer did not identify any major concerns with the manuscript.

Editors' note: We thank you for addressing some of the concerns, adding clarifications and more complete discussions, including further details about experimental protocols. The revised version is significantly improved. Some concerns linger that the biophysical/structural mechanisms underlying the observed phenotypes remain unclear and in some ways are phenomenological. However, the current study is more about the methodology and the mechanisms underlying the stoichiometry dependent effects are perhaps left for a separate study, with more detailed exploration. Congratulations for the excellent work.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary:

BK channels are widely distributed and involved in many physiological functions. They have also proven a highly useful tool for studying general allosteric mechanisms for gating and modulation by auxiliary subunits. Tetrameric BK channels are assembled from four separate alpha subunits, which would be identical for homozygous alleles and potentially of five different combinations for heterozygous alleles (Geng et al., 2023, https://doi.org/10.1085/jgp.202213302). Construction of BK channels with concatenated subunits in order to strictly control heteromeric subunit composition had not yet been used because the N-terminus in BK channels is extracellular, whereas the C-terminus is intracellular. In this new work, Chen, Li, and Yan devise clever methods to construct and assemble BK channels of known subunit composition, as well as to fix the number of γ1 axillary subunits per channel. With their novel molecular approaches, Chen, Li and Yan report that a single γ1 axillary subunit is sufficient to fully modulate a BK channel, that the deep conducting pore mutation L312A exhibited a graded effect on gating with each addition mutated subunit replacing a WT subunit in the channel adding an additional incremental left shift in activation, and that the V288A mutation at the selectivity filter must be present on all four alpha subunits in order to induce channel inactivation. Chen, Li, and Yan have been successful in introducing new molecular tools to generate BK channels of known stoichiometry and subunit composition. They validate their methods and provide three examples of their use with useful observations.

Strengths:

Powerful new molecular tools for the study of channel gating have been developed and validated in the study.

Weaknesses:

(1) One example each of auxiliary, deep pore, and selectivity filter allosteric actions is presented, but this is sufficient for the purposes of the paper to establish their methods and present specific examples of applicability.

We sincerely thank Reviewer #1 for the thoughtful and supportive evaluation of our work. We greatly appreciate the reviewer’s clear summary of the study and the recognition of the novelty and utility of our molecular concatemer strategy for controlling BK channel subunit composition and stoichiometry.

We also appreciate the reviewer’s positive assessment that the three examples (auxiliary subunit modulation, deep pore mutation, and selectivity filter mutation) are sufficient to establish the method and demonstrate its applicability. We are encouraged that the reviewer found the new molecular tools to be powerful and well validated.

We have no further changes to make in response to this review, but we are grateful for the reviewer’s constructive and encouraging comments.

Reviewer #2 (Public review):

Summary:

This manuscript describes novel BK channel concatemers as a tool to study the stoichiometry of the gamma subunit and mutations in the modulation of the channel. Taking advantage of the modular design of the BK channel alpha subunit, the authors connected S1-S6/1st RCK as two- and four-subunit concatemers and coexpressed with S0-RCK2 to form normal function channels. These concatemers avoided the difficulty that the extracellular N-terminus of S0 was unable to connect with the cytosolic C-terminus of the gamma subunit, allowing a single gamma subunit to be connected to the concatemers. The concatemers also helped reveal the required stoichiometry of mutant BK subunits in modulating channel function. These include L312A in the deep pore region that altered channel function additively with each additional subunit harboring the mutation, and V288A at the selectivity filter that altered channel function cooperatively only when all four subunits were mutated. These results demonstrate that the concatemers are robust and effective in studying BK channel function and molecular mechanisms related to stoichiometry. The different requirement of the gamma subunit and the mutations stoichiometry for altering channel function is interesting, which may relate to the fundamental mechanism of how different motifs of the channel protein control function.

Strengths:

The manuscript presents well-designed experiments with high-quality data, which convincingly demonstrate the BK channel concatemers and their utility. The results are clearly presented.

Weaknesses:

This reviewer did not identify any major concerns with the manuscript.

We sincerely thank Reviewer #2 for the careful reading of our manuscript and for the highly positive and supportive comments. We appreciate the reviewer’s detailed summary of our concatemer design strategy and its use in studying gamma subunit stoichiometry and mutation-dependent modulation of BK channel function.

We are especially grateful for the reviewer’s recognition that the experiments are well designed, the data are of high quality, and the results demonstrate the robustness and utility of the concatemer approach. We also appreciate the reviewer’s thoughtful note on the mechanistic implications of the distinct stoichiometric requirements observed for the gamma subunit, L312A, and V288A.

We are pleased that the reviewer identified no major concerns. We have no further changes to make in response to this review, and we thank the reviewer again for the positive evaluation.

Recommendations for the authors:

Reviewing Editor Comments:

While the study presents a great methodological advancement, the phenomenological examples described could perhaps benefit from a little more mechanistic description/discussion. In particular, the functional effect of the V288A mutant is very novel. It could be useful to discuss whether this mutant impacts channel selectivity/conductance. It could be beneficial to also contrast the subunit dependence of V288A with that of the W434F mutant of the Shaker channel. In the latter, C-type inactivation gating is accelerated even when the mutant is present in a single subunit, which contrasts with the effect in V288A.

We greatly appreciate the editor’s and reviewers’ thorough and constructive evaluation, and we have revised the manuscript accordingly.

We added discussion with citation about the potential effect of V288A on selectivity (lines 348349). We also added the reported stoichiometric effects of mutations in Shaker and hERG1 channels on C-inactivation in discussion (lines 336-351). From these studies and our findings with V288A in BK channels, it is interesting to note that the stoichiometric effects of these mutations varies and those located near or within selectivity filter signature exhibited an all-or-none effect in both hERG1 and BK channels.

The authors might also want to consider performing and showing immunoblots with the alpha_deltaM fragment co-expressed with the other channel fragments. Together with the GFP tag, this alpha_deltaM would perhaps be a ~90 kDa protein. It should be captured by anti-V5 IP and resolved on an SDS-PAGE gel (at least with the quad construct).

We added supplemental data (Fig.1 – figure supplement 1) to show co-expression and co-IP of the α^ΔM-GFP construct and a FLAG-tagged α_M construct. The α^ΔM-GFP displayed right size on SDS-PAGE. It is of note that the single unit α_M construct tended to oligomerize even under denatured condition on SDS-PAGE.

For Figure 4, providing details about the inter-pulse intervals and interpulse holding voltage would be helpful. I was not able to find this information in the methods or text.

The inter-pulse intervals and holder voltage are now added in Fig. 4 legend (line 638).

Reviewer #1 (Recommendations for the authors):

(1) Submitted papers should have page numbers to facilitate reviewing.

Both page and line numbers are added.

(2) The designation of the various channel types, such as BKα and BKαM should be identical in the text and figures, so either drop BK in the text or add BK in the figures. Maybe drop BK in the text, as it is known that BK channels are the topic of this study.

We appreciate the suggestion to be consistent in text and figures. We have dropped “BK” for “BKα_M” throughout the text.

(3) "Single Boltzmann fits of G-V curves" would be consistent with a homogenous channel population but do not necessarily suggest a single homogenous channel population of BK channels, as was shown by Geng et al. (2023) (https://doi.org/10.1085/jgp.202213302) where the G-V curve for simultaneous expression of five BK channel types with different V1/2s for each channel type was well approximated by a single Boltzmann function. The dogma that a single Boltzmann fit suggests one channel type needs to be reset. So wave a red flag here: whereas a single Boltzmann fit is consistent with a single channel type, it does not establish a single channel type nor even suggest a single channel type.

We fully agree that a good Single Boltzmann fit doesn’t mean homogenous channel population. We have changed “suggesting” to “consistent with” (line 203) and “reflecting” to “agreeing with” (line 205).

(4) Geng et al. (2023) demonstrated that the pore mutation G375R in BK channels gave a left shift in activation linearly related to the number of WT subunits replaced with mutant subunits. This should incremental shift in activation for G375R should be mentioned, as it is consistent with the incremental effects of the L312A deep pore mutation on activation as reported by the authors in their Figure 3D.

We appreciate the pointing-out of this highly relevant publication. We have now included this reference and discussed together with L312A mutation (lines 309-313).

(5) I went back and looked at the Lingle laboratory papers on the gamma subunit. An additional sentence or two on what the Lingle lab found and didn't find would be useful here for readers.

In the Introduction, we have listed the Lingle lab’s findings and the limitations of their experimental methods that warrants the development of a concatenated construct method as proposed in this study (lines 84-88). We prefer to not discuss further in the Discussion as it will be redundant.

(6) For the two examined mutations L312A and V288A, include in the Methods a 21 amino acid sequence for each mutation with the amino acid to be mutated (L or V) in the center, with beginning and end numbering at the beginning and end of each list. This will allow the reader/experimenter to readily locate the mutated residue on their BK amino acid sequences, which may have different numbering than U11058. Interestingly, for the so-called canonical sequence Q12791 · KCMA1_HUMAN that I found in UniProt starting with U11058, there is an L312, but I found no V288, but an F288. Am I doing this correctly? Do I have the correct sequence/isoform? The only sure way to identify an AA is with an extensive pre and post-sequence so that the chance of misidentification approaches zero.

We verified that the listed Gene Bank IDs of U11058 for cDNA and AAB65837 for protein should point to the right sequences. In the section of Results, we have now included the peptide sequences of the selectivity filter signature motif and part of the S6 TM where V288 and L312A are located, respectively (lines 179 and 220).

Reviewer #2 (Recommendations for the authors):

The different stoichiometry of the gamma subunit and the mutations in regulating channel function raise important questions. For instance, what are the structural and energetic bases for their different stoichiometric requirements? Does the structure motif, such as the selectivity filter or deep pore, act as a unit? Or does a specific residue, such as V288 or L312, act individually to determine the different stoichiometric requirements? What molecular interactions are involved for these residues and subunit to influence the cooperativity among the four alpha subunits in channel function? Some of these questions are discussed in the manuscript, but it may help the readers to clarify what aspects of the mechanistic bases for the findings in this manuscript are known and what aspects remain to be studied.

We agree that these are all important questions. We have now cited more previous studies on C-inactivation in other K⁺ channels and on deep pore mutations in BK channels in terms of subunit stoichiometry (lines 336-351). The results appear to be consistent, suggesting shared properties among residues within the selectivity filter motif or among residues in deep pore region.

Some minor comments are as follows.

(1) Page 7, 2nd paragraph: "Page 2B" change to "Page 3B"? Also, "delay in deactivation" is not precise. The term "Delay" in channel kinetics has a specific meaning, and the use of this word here causes some confusion. The authors may want to delete "substantial delay in deactivation evident as a”.

Corrected by changing Fig. 2B to Fig. 3B and deleting “a substantial delay in deactivation evident as” (line 191).

(2) Page 9, 1st paragraph: "used in the voltage protocol used". Drop one of the instances of used".

Corrected by deleting the first “used” (line 246).

(3) Page 12, 1st paragraph: "Nonetheless, the tight inter-subunit cooperativity observed at the selectivity filter makes it a plausible candidate for serving as the activation gate, a property not yet demonstrated for the lower S6 segment." This seems to be an interesting idea. However, it is not clearly explained. The authors may want to clarify how the cooperativity is related to the activation gate.

We have now added a sentence with citations to discuss the requirement of intersubunit cooperativity for an activation gate to function (lines 354-357).

Other major changes: We updated immunoblot figures Fig1C and Fig2C for better presentation.

eLife Assessment

This fundamental study presents experimental evidence on how geomagnetic and visual cues are integrated in a nocturnally migrating insect. The evidence supporting the conclusions is compelling. The work will be of broad interest to researchers studying animal migration and navigation.

Reviewer #1 (Public review):

Summary

The manuscript by Ma et al. provides robust and novel evidence that the noctuid moth Spodoptera frugiperda (Fall Armyworm) possesses a complex compass mechanism for seasonal migration that integrates visual horizon cues with Earth's magnetic field (likely its horizontal component). This is an important and timely study: apart from the Bogong moth, no other nocturnal Lepidoptera has yet been shown to rely on such a dual-compass system. The research therefore expands our understanding of magnetic orientation in insects with both theoretical (evolution and sensory biology) and applied (agricultural pest management, a new model of magnetoreception) significance.

The study uses state-of-the-art methods and presents convincing behavioural evidence for a multimodal compass. It also establishes the Fall Armyworm as a tractable new insect model for exploring the sensory mechanisms of magnetoreception, given the experimental challenges of working with migratory birds. Overall, the experiments are well designed, the analyses are appropriate, and the conclusions are generally well supported by the data.

Strengths

• Novelty and significance: First strong demonstration of a magnetic-visual compass in a globally relevant migratory moth species, extending previous findings from the Bogong moth and opening new research avenues in comparative magnetoreception.

• Methodological robustness: Use of validated and sophisticated behavioural paradigms and magnetic manipulations consistent with best practices in the field. The use of 5 min bins to study a dynamic nature of magnetic compass which is anchored to a visual cue but updated with latency of several minutes is an important finding and a new methodological aspect in insect orientation studies.

• Clarity of experimental logic: The cue-conflict and visual cue manipulations are conceptually sound and capable of addressing clear mechanistic questions.

• Ecological and applied relevance: Results have implications for understanding migration in an invasive agricultural pest with expanding global range.

• Potential model system: Provides a new, experimentally accessible species for dissecting the sensory and neural bases of magnetic orientation.

Weaknesses

Overall, this is a strong study, and the authors have completed an excellent major revision that has undoubtedly addressed most major and minor issues. The remaining points below are minor recommendations, and I acknowledge that differences in opinion are always possible:

(1) Structure and Presentation of Results

• I recommend reordering the visual-cue experiments to progress from simpler conditions (no cues) to more complex ones (cue-conflict). This would improve narrative logic and accessibility for non-specialist readers. The authors have chosen not to implement this suggestion, which I respect, but my recommendation stands.

(2) Ecological Interpretation

• The authors should expand their discussion on how the highly simplified, static cue setup translates to natural migratory conditions, where landmarks are dynamic, transient, or absent. Specifically, further consideration is needed on how the compass might function when landmarks shift position, become obscured, or are replaced by celestial cues. Additionally, the discussion would benefit from a more consolidated section with concrete suggestions for future experiments involving transient, multiple, or more naturalistic visual cues.

This point was addressed partially in one paragraph of the Discussion, which reads as follows:

"In nature, they are likely to encounter a range of luminance-gradient visual cues, including relatively stable celestial cues as well as transient or shifting local features encountered en route. Although such natural cues differ from our simplified laboratory stimulus, they may represent intermittently sampled visual inputs that can be optimally integrated with magnetic information, with the congruency between visual and magnetic cues likely playing a key role in maintaining a stable compass response. Whether the cues are static or changing, brief periods without them may still allow the subsequent recovery of a stable long-distance orientation strategy. Determining which types of natural visual cues support the magnetic-visual compass, and how they interact with magnetic information, including how their momentary alignment or angular relationship is integrated and how such visual cue-magnetic field interactions may require time to influence orientation, together with elucidating the genetic and ecological bases of multimodal orientation, will be important objectives for future research."

While this paragraph is informative, the wording remains lengthy, somewhat unclear, and vague. Shorter, clearer statements would improve readability and impact. For example:

• How could moths maintain direction during periods when only the magnetic field is present and visual landmarks are absent?

• Could celestial cues (e.g., stars) compensate, and what happens if these are also obscured?

• What role does saliency play when multiple visual landmarks are present simultaneously?

• How might a complex skyline without salient landmarks affect orientation?

Including simple, concise sentences that pose concrete open questions and suggest experimental designs would strengthen the discussion without creating space issues. In my view, a comprehensive discussion of how the simplified, static cue setup relates to natural migratory conditions-where landmarks are dynamic, transient, or absent-would add significant value to the paper.

(3) Methodological Details and Reproducibility

• The lack of luminance level measurements should be explicitly highlighted.

• The authors chose not to adjust figure legends by replacing "magnetic South" with "magnetic North." While I believe this would be more conventional and preferable, this is ultimately a minor stylistic issue.

(4) Conceptual Framing and Discussion

• Although the authors made a good attempt to explain the limitations of using an artificial visual cue, I believe there is room for a more explicit argument. For example, it could be stated clearly that this species is unlikely to encounter a situation in nature where a single, highly salient landmark coincides with its migratory direction. Therefore, how these findings translate to real migratory contexts remains an open question. A sentence or two making this point directly would strengthen the discussion.

(5) Technical and Open-Science Points

• Sharing the R code openly (e.g., via GitHub) should be seriously considered. The code does not need to be perfectly formatted, but making it available would be highly beneficial from an open-science perspective.

Reviewer #2 (Public review):

Summary:

The work titled "Geomagnetic and visual cues guide seasonal migratory orientation in the nocturnal fall armyworm, the world's most invasive insect" provided experimental evidence on how geomagnetic and visual cues are integrated, and visual cues are indispensable for magnetic orientation in the nocturnal fall armyworm.

Strengths:

It has been demonstrated that the Australian Bogon moth could integrate global stellar cues with the geomagnetic field for long distance navigation. However, data are lacking for other insects. This study suggested that the integration of geomagnetic and visual cues may represent a conserved navigational mechanism broadly employed across migratory insects.

Weaknesses:

The visual cues used in the indoor experimental system designed by the authors may have some limitations in ecological relevance. The author may need more explanations on this experimental system.

In the revised manuscript, the authors have added explanations in the discussion section. I am fine with the revision.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary

The manuscript by Ma et al. provides robust and novel evidence that the noctuid moth Spodoptera frugiperda (Fall Armyworm) possesses a complex compass mechanism for seasonal migration that integrates visual horizon cues with Earth's magnetic field (likely its horizontal component). This is an important and timely study: apart from the Bogong moth, no other nocturnal Lepidoptera has yet been shown to rely on such a dual-compass system. The research therefore expands our understanding of magnetic orientation in insects with both theoretical (evolution and sensory biology) and applied (agricultural pest management, a new model of magnetoreception) significance.

The study uses state-of-the-art methods and presents convincing behavioural evidence for a multimodal compass. It also establishes the Fall Armyworm as a tractable new insect model for exploring the sensory mechanisms of magnetoreception, given the experimental challenges of working with migratory birds. Overall, the experiments are well-designed, the analyses are appropriate, and the conclusions are generally well supported by the data.

Strengths

(1) Novelty and significance: First strong demonstration of a magnetic-visual compass in a globally relevant migratory moth species, extending previous findings from the Bogong moth and opening new research avenues in comparative magnetoreception.

(2) Methodological robustness: Use of validated and sophisticated behavioural paradigms and magnetic manipulations consistent with best practices in the field. The use of 5-minute bins to study the dynamic nature of the magnetic compass which is anchored to a visual cue but updated with a latency of several minutes, is an important finding and a new methodological aspect in insect orientation studies.

(3) Clarity of experimental logic: The cue-conflict and visual cue manipulations are conceptually sound and capable of addressing clear mechanistic questions.

(4) Ecological and applied relevance: Results have implications for understanding migration in an invasive agricultural pest with an expanding global range.

(5) Potential model system: Provides a new, experimentally accessible species for dissecting the sensory and neural bases of magnetic orientation.

Weaknesses

While the study is strong overall, several recommendations should be addressed to improve clarity, contextualisation, and reproducibility:

We thank Reviewer #1 for the positive and encouraging evaluation of our study. We appreciate the recognition of our work’s strengths and are grateful for the constructive feedback on the remaining weaknesses, which will guide and strengthen our revisions.

Structure and presentation of results

Requires reordering the visual-cue experiments to move from simpler (no cues) to more complex (cue-conflict) conditions, improving narrative logic and accessibility for non-specialists.

Thank you for this thoughtful suggestion. While we appreciate the rationale for presenting results from simpler to more complex conditions, we kept the original sequence because it aligns with the logic of our study. Our initial aim was to determine whether fall armyworms use a magnetic compass integrated with visual cues, as shown in the Bogong moth. After establishing this phenotype, we then examined whether visual cues are required for maintaining magnetic orientation. We have also clarified in the Introduction that magnetic orientation in the Bogong moth relies on integration with visual cues, which provides readers with clearer context and improves the overall narrative flow.

Ecological interpretation

(a) The authors should discuss how their highly simplified, static cue setup translates to natural migratory conditions where landmarks are dynamic, transient or absent.

Thank you for raising this important point. We agree that natural migratory environments provide visual information that is often dynamic, transient, or intermittently absent, in contrast to the simplified and static cue used in our indoor experiments. Our intention in using a minimal, static cue was to isolate and test the fundamental presence of magnetic–visual integration in fall armyworms under fully controlled conditions.To address the reviewer’s concern, we have added a brief note in the Discussion indicating that fall armyworms may encounter both static and dynamic luminance-based visual cues in nature, such as light–dark gradients created by terrain features or more stable celestial patterns. Although these natural cues differ from our simplified laboratory stimulus, they may similarly provide asymmetric visual structure that can be integrated with magnetic information. We also note that determining which natural visual cues support the magnetic–visual compass will be an important direction for future work.

(b) Further consideration is required regarding how the compass might function when landmarks shift position, are obscured, or are replaced by celestial cues. Also, more consolidated (one section) and concrete suggestions for future experiments are needed, with transient, multiple, or more naturalistic visual cues to address this.

Thank you for this constructive suggestion. We appreciate the reviewer’s point that additional consideration of how the compass might function under shifting, obscured, or celestial visual cues would strengthen the manuscript. Given the limited evidence currently available for this species, we have incorporated a concise and appropriately cautious discussion addressing these possibilities.

Methodological details and reproducibility

(a) It would be better to move critical information (e.g., electromagnetic noise measurements) from the supplementary material into the main Methods.

Thank you for this helpful suggestion. In the revised manuscript, we have added the key electromagnetic noise measurements information to the main Methods section.

(b) Specifying luminance levels and spectral composition at the moth's eye is required for all visual treatments.

Thank you for this helpful comment. We have clarified in the Methods as well as the legend of Fig. S3 that both luminance levels and spectral composition were measured at the position corresponding to the moth’s head.

(c) Details are needed on the sex ratio/reproductive status of tested moths, and a map of the experimental site and migratory routes (spring vs. fall) should be included.

Thanks. We have added the reproductive status of the tested moths in the Methods, specifying that all individuals used were unmated 2-day-old adults.

(d) Expanding on activity-level analyses is required, replacing "fatigue" with "reduced flight activity," and clarifying if such analyses were performed.

Thank you for this comment. In this context, the term “fatigue” referred to the possibility that moths might gradually lose motivation or attention to orient when flying for an extended period in a simplified, artificial environment with limited sensory cues. Such a decrease in orientation motivation over time could, in theory, lead to a loss of individual orientation and consequently to the observed loss of group orientation. To test this possibility, we analyzed the orientation performance of each individual moth across different phases using the Rayleigh test. The r-value was used as a measure of individual directedness (higher r-values indicate stronger orientation). Our results showed that mean r-values did not differ significantly among the experimental phases (multiple comparisons, Table S2). This indicates that 25min measurement itself was not responsible for the loss of orientation. We did not perform a quantitative activity-level analysis in this study. However, as mentioned in Methods, flight activity was continuously monitored during the experiments by observing fluctuations in the pointer values on the experimental software, which corresponded to the moth’s rotational movements. If the pointer values remained unchanged for more than 10 seconds, the experimenter checked for wing vibrations by sound; if the moth had stopped flying, gentle tapping on the arena wall was used to stimulate renewed flight. Only individuals that maintained active flight throughout the experiment, with fewer than four instances of wingbeat cessation, were included in the analysis. We also mentioned that activity level analysis was not performed due to technical difficulties in the revised manuscript.

Figures and data presentation

(a) The font sizes on circular plots should be increased; compass labels (magnetic North), sample sizes, and p-values should be included.

Thank you for this helpful suggestion. Regarding the compass labels and statistical reporting, our analysis provides significance levels as ranges rather than exact p-values; therefore, we clarified in the figure legends that the two dashed circles correspond to thresholds for statistical significance p = 0.05 and p = 0.01, respectively. Sample sizes are already indicated within each panel. To avoid visual clutter caused by displaying both magnetic North and South, we show only the magnetic South direction (mS) consistently across panels, which can improve readability.

(b) More clarity is required on what "no visual cue" conditions entail, and schematics or photos should be provided.

Thank you for this comment. In our study, the “no visual cue” condition refers to the absence of the black triangular landmark inside the flight simulator. To improve clarity, we have updated the legend of Fig. 4 to explicitly state this and have referred readers to the schematic in Fig. 1, which illustrates the structure of the flight simulator. These additions clarify what the “no visual cue” condition entails without requiring additional schematics.

(c) The figure legends should be adjusted for readability and consistency (e.g., replace "magnetic South" with magnetic North, and for box plots better to use asterisks for significance, report confidence intervals).

Thank you. Regarding the choice of compass labeling, we intentionally used magnetic South (mS) rather than magnetic North (mN) because the main population tested in our experiments represents the autumn migratory generation. During autumn, fall armyworms orient southward when visual and magnetic cues are aligned. Using magnetic South in the plots therefore provides a clearer representation of cue alignment in this season and avoids potential confusion when interpreting the combined visual–magnetic information.

Conceptual framing and discussion

(a) Generalisations across species should be toned down, given the small number of systems tested by overlapping author groups.

Thank you for this valuable comment. In the revised manuscript, we have softened such statements in both abstract and maintext.

(b) It requires highlighting that, unlike some vertebrates, moths require both magnetic and visual cues for orientation.

Thank you for this helpful suggestion. We have added a sentence to the Discussion explicitly highlighting that, unlike some vertebrates capable of using magnetic information in the absence of visual cues, moths require the integration of both magnetic and visual cues for accurate orientation. This clarification emphasizes the distinct multimodal nature of compass use in migratory moths.

(c) It should be emphasised that this study addresses direction finding rather than full navigation.

Thank you for this important clarification. We have now made it explicit in the manuscript that our experiments address direction finding (i.e., orientation) rather than full navigation. This distinction is stated in both the Introduction and Discussion to clearly define the scope of the study.

(d) Future Directions should be integrated and consolidated into one coherent subsection proposing realistic next steps (e.g., more complex visual environments, temporal adaptation to cue-field relationships).

Thank you for this constructive suggestion. We agree that outlining realistic next steps is valuable. However, given the limited scope of the current data, we have only slightly expanded the existing forward-looking statements in the Discussion.

(e) The limitations should be better discussed, due to the artificiality of the visual cue earlier in the Discussion.

Thank you for this comment. We agree that the artificiality of the visual cue is an important limitation of the present study. Rather than extending speculative discussion, we have clarified this limitation in the revised Discussion and highlighted the key questions that future work must address.

Technical and open-science points

Appropriate circular statistics should be used instead of t-tests for angular data shown in the supplementary material.

Thank you for this comment. We have addressed this point (Fig. S1) in the revised supplementary material.

Details should be provided on light intensities, power supplies, and improvements to the apparatus.

Thank you. Light intensities are reported as spectral irradiance measurements in Supplementary Materials, which provide full wavelength-resolved information for the illumination used, although a separate measurement of total illuminance (lux) was not performed. We have also added the requested information on the power supplies.

The derivation of individual r-values should be clarified.

Thanks. We have clarified in the revised manuscript.

Share R code openly (e.g., GitHub).

Thanks. We are in the process of organizing the relevant R code, but have not been able to upload it to GitHub before the current revision deadline. The code is available from the corresponding author upon request.\

Some highly relevant - yet missing - recent and relevant citations should be added, and some less relevant ones removed..

Thanks. We added one recent relevant reference to the revised manuscript.

Reviewer #2 (Public review):

Summary:

This work provided experimental evidence on how geomagnetic and visual cues are integrated, and visual cues are indispensable for magnetic orientation in the nocturnal fall armyworm.

Strengths:

Although it has been demonstrated previously that the Australian Bogon moth could integrate global stellar cues with the geomagnetic field for long-distance navigation, the study presented in this manuscript is still fundamentally important to the field of magnetoreception and sensory biology. It clearly shows that the integration of geomagnetic and visual cues may represent a conserved navigational mechanism broadly employed across migratory insects. I find the research very important, and the results are presented very well.

We thank Reviewer #2 for the positive and encouraging evaluation of our study. We appreciate the recognition of our work’s strengths.

Weaknesses:

The authors developed an indoor experimental system to study the influence of magnetic fields and visual cues on insect orientation, which is certainly a valuable approach for this field. However, the ecological relevance of the visual cue may be limited or unclear based on the current version. The visual cues were provided "by a black isosceles triangle (10 cm high, 10 cm 513 base) made from black wallpaper and fixed to the horizon at the bottom of the arena". It is difficult to conceive how such a stimulus (intended to represent a landmark like a mountain) could provide directional information for LONG-DISTANCE navigation in nocturnal fall armyworms, particularly given that these insects would have no prior memory of this specific landmark. It might be a good idea to make a more detailed explanation of this question.

We appreciate the constructive feedback on the weaknesses, which will guide and strengthen our revisions. To address the reviewer’s concern, we have added a brief note in the Discussion indicating that fall armyworms may encounter both static and dynamic luminance-based visual cues in nature, such as light–dark gradients created by terrain features or more stable celestial patterns. Although such natural cues differ from our simplified laboratory stimulus, they may represent intermittently sampled visual inputs that can be optimally integrated with magnetic information, whether the cues are static or changing, and brief periods without them may still allow the subsequent recovery of a stable long-distance orientation strategy.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Major to Medium Suggestions

(a) Reordering of Visual Cue Tests

The manuscript currently presents cue-conflict experiments before the simpler "no visual cue" tests. For non-specialist readers, it would be more logical to start with the basic condition (no visual cues) and then move to progressively more complex ones. This provides a clearer and more logically sound narrative.

For example, the results could first demonstrate that without visual cues, the moths fail to orient (both in darkness and uniform light), and then show that introducing a single salient cue (a triangle on the horizon) restores directed behaviour. This would help readers understand the logic of the progression and should be better integrated throughout the Results and Discussion.

Thanks. We have responded this comment in Public Reviews.

(b) Translating Key Findings to Realistic Scenarios (LL 333-344 or where suitable in Discussion, and mentioning that we utilised a reductionist principle first in Intro, but clearly articulated that it is very simplified)

The main text (eg Discussion) should address how these findings translate to real-world conditions. The experimental design used a single, highly salient, and static cue, always aligned with the migratory direction. In nature, such a consistent landmark is unlikely-mountains or other features would shift position relative to the moth's trajectory as it flies.

Key questions arise which need to be addressed:

- How would the compass system adapt to changing landmark positions as the moth moves?

- What happens when no landmarks are visible (e.g. over flat plains or cloudy nights)?

- Would stellar or other cues take over in such cases? Your hypotheses, please.

Addressing these points - and proposing specific future experiments (e.g. with transient or multiple visual cues)-would strengthen the ecological relevance of the findings and show a clear way forward.

Thanks for your kind comments. We now explicitly state in the Introduction that our study employs a reductionist approach using a simplified visual environment to isolate magnetic-visual interactions. As the ecological questions raised by the reviewer cannot be addressed with the current dataset, we avoid extended speculation but have added brief clarification in the Discussion and addressed these points in the Public Reviews response. We also indicate that future work will need to examine the types of visual cues that can support magnetic orientation and how such cues couple with geomagnetic information.

Technical and Methodological Points

(a) Incomplete Methods Section

Critical technical information (e.g. electromagnetic noise measurements) currently appears only in supplementary figure legends. All such details should be included in the main Methods section if the word count allows (or include a short section in the main text with reference to more details in the supplementary material).

Thanks for your kind comments. We have addressed this as suggested in the Public Reviews.

(b) Lighting Conditions

Specify luminance levels (the amount of light emitted and passing through in quanta per unit of surface, eg m2) at the moth's eye and indicate whether spectral composition was consistent between treatments (with and without the visual cue).

Thanks for your comments. We have responded to this point in the Public Reviews.

(c) Figures

- Increase font sizes on circular histograms.

- Add compass labels (ideally magnetic North, mN, not south, etc, as it is usual in pertinent literature), sample sizes, and p-values on each panel.

- Replace "magnetic South" (mS) indicators with magnetic North (mN) to align with convention.

Thanks for your comments. We have responded to this point in the Public Reviews.

(d) Migratory Expectations

Include expected compass bearings for spring and autumn migrations (with citations) to relevant figures (Figure 2, 4, S2).

Thanks for your comments. We have added the information that “We recently found that fall armyworms from the year-round range in Southwest China (Yunnan) exhibit seasonally appropriate migratory headings when flown outdoors in virtual flight simulators, heading northward in the spring and southward in the fall, and this seasonal reversal is controlled by photoperiod (Chen et al., 2023).” in Introduction. Thus, we didn’t offer expected seasonal compass bearings in Results section.

(e) Add a map showing the experimental site and known migratory routes, clearly labelling spring vs fall routes. It would help justify expected headings.

Thank you for this suggestion. At present, there are no experimentally validated migratory routes (e.g., through mark-release-recapture or tracking approaches) for the specific fall armyworm population used in our study. Because these routes have not been biologically confirmed, we didn’t offer a presumed migratory map that may imply unwarranted certainty.

(f) Composition of Test Groups

Indicate sex ratios and reproductive status (mated/unmated) of tested moths, if known or comment if unknown, as both can affect migratory motivation and behaviour.

Thank you for this suggestion. We have responded to this point in the Public Reviews.

(g) Role and Nature of Visual Cues

While the results clearly show that orientation disappears without visual cues, the triangle cue is highly artificial. Well-studied Bogong moths are known to rely on views of Australian mountain ranges during their nocturnal migrations, but there is no evidence that armyworms use a similar strategy. Even for bogongs, it is not just one salient mountain always in front of them on migration. Discuss whether Fall Armyworm would encounter comparable natural cues in the field along their migratory route, or whether the triangle might simply provide a frame of reference rather than a true landmark.

Thank you for this comments. We have responded to this point in the Public Reviews.

(h) Future work could test:

- More naturalistic sky cues (moonlight, star fields).

- Varying the landmark's position relative to the magnetic field - slowly moving along - transient landmarks. Also, less salient landmarks and a more complex skyline, as it is usually more complex than just a single salient peak.

Thank you for this comments. We have responded to this point in the Public Reviews. Brief discussion as suggested has been added to the revised manuscript.

Minor Comments and Line-by-Line Suggestions

L70 - Check citation (possibly Mouritsen 2018). Missing in the list of references.

Thanks. This point has been addressed.

L75 - Consider citing the new and highly relevant preprint:

Pakhomov, A., Shapoval, A., Shapoval, N., & Kishkinev, D. (2025). Not All Butterflies Are Monarchs: Compass Systems in the Red Admiral (Vanessa atalanta). bioRxiv.

Thanks. We have cited this reference.

LL81-82 - Clarify vague phrasing; specify criteria for "good" vs "poor" orientation ability. Or reword/leave out.

Thanks for your comments.

L85 - "but one," not "bar one." 

Thanks. Corrected.

L124 - The 2 genetic citations are weakly linked to magnetoreception. We do not have a clear understanding of the insect magnetoreceptor and its underlying mechanism, so we simply cannot interpret genetic associations very well to underpin them to magnetoreception. For example, does noctuid's magnetic sense require a magnetised-based receptor and genes involved in biomineralization? Consider removing or softening claims. 

Thanks. Adressed.

LL123-126 - Define what for YOU constitutes "strong evidence" for magnetoreception (e.g. adaptive directional behaviour consistent with migratory orientation?). Is there such a thing as strong evidence at all?

Thanks for your comments. We agree that terms such as “confirmed” or “strong evidence” can overstate the certainty of magnetoreception findings, given the ongoing debates in the field. In the revised manuscript, we have toned down.

L153 - Indicate whether coils in NMF condition were powered or inactive.

Thanks for your comments. Addressed.

L163 - Justify use of multiple 5-min phases (e.g. temporal resolution of behaviour). It is confusing at the start, where first mentioned, and becomes clearer only towards the end, but it should be clearer at the start.

Thanks for your comments. The assay was divided into these 5-min segments to provide the temporal resolution needed to detect changes in flight orientation as the relative alignment of magnetic and visual cues was systematically altered. We now clarify this earlier in the Results.

LL167-171 - This is a good place where you can provide a map (main or supplementary with referencing) showing the study site and migration routes.

Thanks for your suggestion. We have responded to this point in the Public Reviews.

L174 - Avoid repetition of "expected."

Thanks. Addressed.

LL176-177 - Report 95% confidence intervals or equivalent and clarify which test (e.g. Moore's paired test) each p-value refers to.

Thanks for your suggestion.

LL189-191 - explain what fatigue means. I would remove fatigue and substitute it with "lowered flight activity". Also, the same statement comes later, so avoid repetitiveness and remove it in one place. The analysis of directedness is good throughout, but what about the analysis of activity level? Could you explain whether you did it or not, and if not, why, or if angular changes can serve as an activity proxy? Replace "fatigue" with "reduced flight activity." Avoid repetition. Clarify if activity level analysis was performed or if it was not, e.g. due to technical difficulties.

Thanks for your comments. We have responded to this point in the Public Reviews.

L196 - Note whether 95% CI overlaps with the expected direction. This is a crucial outcome.

Thanks for your comments.

LL203-205 - unclear, better to stick to "congruency", especially "initial congruency for the relationship between mN and visual cue" throughout.

Thanks for your suggestions.

L206 - Better to introduce a new subheading: "Laboratory-Reared Animals.".

Thanks for your suggestion. A new subheading has been added in the revised manuscript.

LL207-208 - Clarify which cues were available in Chen et al. (2023) and how they differ here.

Thanks for your comments. In Chen et al. (2023), the moths oriented under an artificial starry sky together with optic flow cues. In contrast, our experiments intentionally removed both the starry-sky pattern and optic flow to avoid introducing additional visual information when testing magnetic-visual integration for orientation. We have added further clarification regarding the conditions used in Chen et al. (2023) in the revised manuscript.

L228 - Use "lab-reared" consistently throughout the entire MS. Do not mix with lab-raised.

Thanks. Addressed by consistently using “lab-raised”.

Figure 2 - Confusing in parts, especially for people coming from birds and other vertebrates orientation background. At 12 o'clock, you usually expect either mN / gN (magnetic or geographic North) or the animal's own initial directional response used as control to compare the same animal's direction post-treatment. Here, your 6 o'clock is magnetic South in the first place - non-conventional. At 12 o'clock, better use mN or gN. Avoid using non-conventional references such as magnetic south. Remind readers of seasonally appropriate headings and refer to the map.

Thanks. We have responded to this point in the Public Reviews.

LL232-234 - Emphasize that cue-magnetic congruency is key. Highlight the most important point that the congruency between the seasonal migratory direction and visual cues is key, not that in spring/fall, visual cues must be towards or opposite to the migratory goal. But the visual cue could be in the migratory direction or opposite, or at an angle - this is for future direction.

Thanks. We have responded to this point in the Public Reviews.

Figure 2 and associated main text - highlight that you only tested the designs when in all seasons the salient and single visual cue was in the migratory direction (in spring it coincided with mN but in fall it was towards the magnetic south). Other directions of visual cues have not been tested, but for simplicity and consistency, you chose to do these ones as the first step, perhaps.

Thank you for this insightful comment. Yes, our experiments tested only the conditions in which the salient and single visual cue was aligned with the migratory direction. Other angular relationships between visual cues and the magnetic field were not examined in this study. For simplicity and consistency, we focused on this alignment as a first step toward understanding magnetic-visual cue integration in migratory orientation. We now highlight this in the Fig. 2 legend.

Figures captures/legends - hard to tell from the main text now, better to italicize figure caption text and visually space them from the main text.

Thanks for your suggestions.

LL 250-251 - mention to people more familiar with r - lowercase - what is the expected range for R uppercase. It is not bound 0-1 as r. Could it be negative? How large can it be?

Thanks. Thanks for the comment. After revisiting Moore (1980) we think that R* cannot take negative values. However, since R* = R*/N^ (3/2), it is not bounded between 0 and 1. We didn’t find any concept of an upper bound in the paper (https://doi.org/10.2307/2335330).

Figure 3 - Consider adding a horizontal line indicating the 5% significance threshold.

Thanks for your suggestions.

L 261 - need to have some narrative after the subheading before you insert Figure 3.

Thanks. Addreseed.

LL274-275 - highlight that the timeline of this congruency between mN and a landmark and the effect of this on directedness is not explored here, but worth doing in future. How long does a new congruency or a relationship between mN and a visual cue need to be exposed to the animal to regain its directional response? Clearly, it is just a question of time of exposure so that a new association is established. Suggest future work on time-dependent adaptation to new cue-field relationships.

Thanks for your suggestion. We have now included this point as a future direction in the revised Discussion.

Figure 4 & S4 - Replace letters with asterisks/brackets for significance. The use of the letter is confusing and unconventional.

Thanks for your suggestion.

Figure 4 caption - Clarify the main takeaway.

Thanks for your suggestion.

Figure 4 - bare minimum is confusing. I understand that you wanted to avoid "no visual cues" because, as long as the animal sees things, there are things to be used as visual cues, even if this is not the intention of the experimenter. However, it needs clarification and rewording. Better to be more specific, like "no black triangle and horizon were used, just the uniformly white cylinder", or something like that.

Thanks for your comments. In our setup it accurately describes the intentional removal of both the black triangle and the horizon, leaving only the uniformly white cylinder as the visual environment. This wording was chosen to reflect the practical limitations of producing a perfectly symmetrical flight simulator under laboratory conditions, and we therefore prefer to retain the original phrasing.

L328 - Remove Xu et al. (2021) citation (not relevant). This is an in vitro study with a protein which may not work exactly as it is claimed in the paper in vivo.

Thanks. Citation removed.

L349-350 - Clarify what "no visual cue" means (e.g., uniformly white cylinder, no horizon line). Include a photo or a schematic of the inner surface of the cylinder for this condition in the Supplementary Materials.

Thanks. We have responded to this point in the Public Reviews.

L380 & throughout - Replace "barely minimum visual cues" (BMVC) with "no visual cues", clarifying limitations in Methods, meaning that you can explain that absolutely no visual cues is practically impossible because, as long as there is light, animals can use some asymmetries as cues even if this is not the intention of the experimenter.

Thank you for this comment. We have decided to retain the term “barely minimum visual cues (BMVC)” because it accurately describes our experimental condition, which is distinct from a true “no visual cues” environment. In the revised Figure legend, we now clarify that BMVC refers to conditions in which obvious visual cues (i.e., features such as the black triangle in Fig. 1) were removed, while acknowledging that complete elimination of all visual information is not possible under illuminated conditions.

L396 - Be cautious when generalizing from two species tested by a research group that is not absolutely independent (some authors in bogong and armyworm works overlap). We saw examples in diurnal migratory butterflies (Monarchs), a more studied species than the armyworm, that the findings do not entirely translate to Red Admirals (Pakhomov et al. 2025 preprint mentioned). Suggestion to tone down any claims of broad generalisation throughout the manuscript.

Thank you for this comment. We have responded to this point in the Public Reviews.

LL402-407 - Note that, unlike birds (e.g. European robins), moths appear to require both magnetic and visual cues for orientation, whereas birds, mole rats and some other animals can use magnetic cues alone.

Thank you for this comment. We have responded to this point in the Public Reviews.

L410 - Specify that this is correct only in the Northern Hemisphere.

Thank you for this comment. Addressed.

LL415-416 - Acknowledge artificiality of single-cue setup (see the major comments above); integrate earlier in the Discussion.

Thank you for this comment. We have responded to this point in the Public Reviews.

LL420-425 - Consolidate Future Directions into a single subsection; include more concrete experimental ideas, for example, using more naturalistic, numerous transient landmarks (could be done in a virtual maze with LEDs on the wall of the cylinder with cues moving with time). Multiple visual cues. Manipulating with salience of cues - less simplistic, less salient.

Thank you for this comment. We have responded to this point in the Public Reviews.

L431 - Does this paper support this statement? I think it just tested the use of stellar cues in a zero magnetic field. It also dealt with direction finding, not navigation, which is a position-finding ability - a much more complex feat and might not be the ability of moths (requires further studies like with geographic and magnetic displacements, etc). Reword and check this. Show the distinction between direction finding and navigation.

Thank you for this comment. We have reworded the relevant sentence to use “orientation” instead of “navigation”.

L436-437 - Specify "global visual cues" (stellar, lunar, etc.) and merge all future directions into one coherent section.

Thank you for this comment. Addressed.

LL443-446 - A bit early to plan such studies because migratory direction could well be a complex multigenetic trait, so that you cannot approach it simply with the knock out of a single gene. The genetic basis of magnetic direction needs to be first demonstrated, which leads you to the Future Directions section.

Thank you for this helpful comment. We fully agree that migratory direction is likely a complex multigenic trait, and our intention was not to imply that knocking out a single gene would be sufficient to explain magnetic or migratory orientation. Our statement aimed only to highlight that identifying candidate genes is an important first step toward understanding the genetic basis of magnetic orientation.

Line 496 - Clarify whether optic flow was used (unlike previous studies).

Thank you for pointing this out. Clarified.

LL499-511 - Clarify the improvements done in Chen's system and their relevance.

Thank you for pointing this out. We reworded this sentence “The Flash flight simulator system was developed based on the early design of the Mouritsen-Frost flight simulator and adapted for our experiments in Yuanjiang”.

Line 531 - Report and compare light intensities between indoor and outdoor experiments.

Thanks for this comment. Unfortunately, due to the sensitivity limits of our current equipment, we were unable to reliably measure outdoor light intensities at night. However, we did not perform any open-top outdoor flight-simulator experiments; instead, we used field-captured moths but conducted all behavioral tests indoors.

L549 - Add make/model of power supplies.

Thanks. Addressed.

LL582-585 - Specify whether R code will be shared; recommend open access (e.g., GitHub, other open repositories). Reiterate the importance of open science and sharing all scripts. Also here, add citations to some studies where MMRT has been used recently.

Thank you for this comment. We have responded to this point in the Public Reviews.

Line 592 - Explain how individual r-values were derived from optical encoder data.

Thank you for this comment. Addressed.

L842-843 - t-tests are inappropriate for angular data; use circular tests (Watson-Williams, Mardia-Watson-Wheeler, etc.).

Thank you for this comment. Addressed.

L865 - Reword to avoid repetition of "fall." Example: "In field captured armyworms during fall migration".

Thank you for this comment. Addressed.

LL882-885 - Improve phrasing and language here. Confirming that - no colon after. "Both the acrylic plate and diffusion paper." Confirm relevance of spectra to moth visual sensitivity - add relevant citation to original studies showing that.

Thank you for this comment. Addressed.

L886 - Reword "uniform" - does not look uniform to me.

Thank you for this comment. Addressed.

Reviewer #2 (Recommendations for the authors):

The first two sentences of the abstract ("The navigational mechanisms employed by nocturnal insect migrants remain to be elucidated in most species. Nocturnal insect migrants are often considered to use the Earth's geomagnetic field for navigation, yet the underlying mechanisms of magnetoreception in insects remain elusive") are somewhat redundant. The authors may consider rewriting them.

Thank you for pointing this out. We have rewritten this opening to provide a more concise and non-repetitive introduction.

eLife Assessment

Laaker et al. make an important finding that the cribriform plate acts as a unique neuroimmune interface that shapes local myeloid cell states during EAE-induced neuroinflammation. Using immuohistochemistry, flow cytometry, and single-cell RNAseq of doublets to interrogate cell-cell interactions, the authors provide solid evidence that macrophages, migratory dendritic cells (DCs), and fibroblasts interact at the site of CSF outflow, with DCs showing characteristics of immune tolerance. While the functional consequences of these cell states remain to be established, the work shows that the cribiform plate can play a key role in influencing immune cell composition and interactions with stromal cells.

Reviewer #1 (Public review):

Summary:

Laaker et al. investigate the immunological role of the cribriform plate during neuroinflammation using the EAE model. The authors combine immunohistochemistry, flow cytometry, and single-cell RNA sequencing to characterize CD11b+CD11c+ myeloid cells that accumulate at podoplanin (PDPN)-rich meningeal-lymphatic niches surrounding olfactory nerve bundles. They identified distinct populations of migratory dendritic cells (DCs) and macrophages retained at the cribriform plate that exhibit transcriptional signatures consistent with immune tolerance, reduced interferon signaling, and programmed cell death, including Pdcd1 (PD-1) expression. In parallel, CCR2+ monocytes and alternatively activated (M2-like) Arg1+/CHI3L3+ macrophages integrate into this niche, suggesting the establishment of a locally immunosuppressive myeloid network.

Strengths:

(1) Overall, the study postulates a novel model in which the cribriform plate functions as a specialized perineural immune interface that reshapes myeloid phenotypes during neuroinflammation.

(2) Suggests broader relevance for shaping peripheral immunity and therapeutic targeting. If DCs are being "tuned" at this exit site, it could influence what reaches cervical lymph nodes and how peripheral responses are set during CNS autoimmunity; the authors explicitly position this as relevant to CNS autoimmunity and possibly other CNS diseases (while acknowledging the need for human validation).

(3) Technical sound and highly original work. Convergent multi-method support: the central narrative is backed by immunohistochemistry + flow cytometry + scRNA-seq, rather than a single assay. The headline conclusion (tolerogenic/suppressive skew at the cribriform plate during EAE) is explicitly built from these combined modalities.

Weaknesses:

(1) In Figure 1, the manuscript would be strengthened by quantification of CSF1R-GFP+ and CD11c-eYFP+ cells in PDPN+LYVE1- versus PDPN+LYVE1+ regions in both control and EAE mice. This would demonstrate selective accumulation or retention of myeloid cells at the cribiform plate niche.

(2) While the PostContact-seq strategy is innovative (Figure 3), additional justification is needed to demonstrate that tissue dissociation did not artificially disrupt PDPN-myeloid contacts. The relatively small proportion of live PDPN-rich doublets (~2.5% total aggregates and ~18% PDPN+ within total aggregates) raises questions about representativeness compared with in situ observations. The authors should also more explicitly elaborate on why PostContact-seq was favored over alternative approaches such as PIC-seq.

(3) The authors stated that results regarding cell-cell interactions were integrated across four intercellular communication methodologies (Figure 4B), but this integration is not clearly described in either the Results or Method sections. This needs clarification. Moreover, the interaction analysis in Figure 4B seems to rely on TALKIEN, which does not incorporate prior ligand-receptor knowledge. Given the availability of widely used tools, such as CellChat and NicheNet, the authors may consider cross-referencing their findings.

(4) Given the increase in CCR2+ cells in PDPN+ regions (Figure S4), a pseudotime trajectory analysis may be valuable to test whether CCR2+ monocytes preferentially differentiate into CHI3L3+ immunosuppressive macrophages, PD-1+ DCs, or other myeloid subsets in post-contact versus no contact.

(5) Validation of immunosuppressive signatures in macrophages (Fig. 4G-H) using the same FACS-based post-contact versus no-contact sorting strategy (as in Figure 3A) would strengthen the conclusions.

(6) The identity of CD45IV+ cells in contact with PDPN+ cells is unclear (Figure 6B-C). The authors should provide a gating strategy demonstrating that these cells are CD11b+CD11c+ DCs within the PDPN+ doublet population, and ideally show whether these dying cells are PD-1+. Furthermore, co-labeling in tissue sections for PD-1, cleaved caspase-3, and CD11c-eYP would provide important spatial validation of flow cytometry findings (Figure 6E).

(7) In Figures 1F-H, the authors should comment on the morphological differences of CD11c+ cells in the olfactory bulb versus those infiltrating the cribriform plate.

Reviewer #2 (Public review):

Summary:

In this article, Laaker et al described diverse populations of macrophages and dendritic cells found in and around the cribriform plate in the context of a neuroinflammation caused by an autoimmune disease (EAE). The authors utilize elegant histochemical staining and a nifty approach to sort doublets to interrogate cells that are in contact with one another, presumably in vivo. Notably, they uncover a population of CD11c+CD11b+ cells interacting with M2 macrophages and PDPN+ fibroblasts and lymphatics. These cells are heterogenous but some of these DCs express PD-1, and transcriptional profiling suggests they may have immunosuppressive behavior. Altogether, this article explains well the complexity of cell populations found around the cribriform plate during inflammation, and is suggestive of different interactions that trigger these different phenotypes from immune cells.

Strengths:

Beautiful images of a unique CNS: peripheral interface that support a novel scRNA approach to understanding how different cell populations engage in functional interactions in vivo.

Weaknesses:

It's currently unclear how the sorted populations reflect in vivo interactions or a propensity to form aggregates during ex vivo processing. The authors address both podoplanin-expressing cells as stromal cells and as lymphatic endothelial cells, but at times it's unclear which of these two populations is being analyzed and which is the most relevant. While novel observations, most of these findings are descriptive and lack functional correlates, and in places, the potential implications could use further discussion.

Author response:

We would like to thank the reviewers for their supportive comments which largely agree with our main finding that a heterogeneous population of dendritic cells and Th2-skewed macrophages interact with the PDPN+ niche at the cribriform plate during EAE neuroinflammation. Additionally, they have provided several meaningful critiques to our study which we are now working on addressing in a newly revised manuscript.

eLife Assessment

In this valuable study, DNA and RNA are co-imaged in single cells to show that the proximity of topologically associated domain (TAD) boundaries is uncoupled from the transcriptional activity of nearby genes. The evidence supporting these conclusions is convincing for the regions examined, with high-throughput imaging providing robust statistics. This work will be of interest to researchers studying genome architecture and its relationship to gene regulation.

Reviewer #1 (Public review):

Summary:

This is an important study that employs high-throughput single-cell imaging to directly investigate the relationship between topologically associating domain (TAD) boundaries and gene regulation. The authors rigorously test the prevailing model that TAD boundaries functionally regulate gene activity by modulating chromatin interactions. Their core finding is that, under their specific experimental conditions, the physical distance between TAD boundaries shows no consistent correlation with the transcriptional bursting activity of a gene within the TAD. However, the authors' leap from this specific observation to the broad conclusion that "TAD boundary architecture and gene activity are uncoupled" risks conceptual overgeneralization and may lead to misinterpretation, as it seemingly contradicts substantial prior evidence supporting the regulatory role of TAD structures.

Strengths:

The major strength of this work lies in its innovative high-throughput, multi-colour imaging platform, which enables the simultaneous detection of spatial distances between specific DNA elements (TAD boundaries) and transcriptional activity at the same genomic locus in single cells and single alleles. The high-throughput nature makes the results convincing. A second key strength is the incorporation of perturbations, including global transcriptional inhibition, cell-type comparison, and degradation of key architectural proteins (CTCF, cohesin). This provides a comprehensive methodological framework to examine the relationship between boundary proximity and gene activity from multiple angles under defined conditions.

Weaknesses:

(1) Conceptual framing and interpretation:

The central conclusion may require more precise framing to avoid potential overreach. The authors' interpretation equating "physical distance between TAD boundaries" with overall "TAD boundary architecture," and "transcriptional bursting events" with broader "gene activity," could benefit from clarification. This framing may not fully capture the temporal dynamics of transcription or the regulatory complexity within TADs. Furthermore, the broad conclusion of an uncoupled relationship appears to challenge extensive prior evidence from perturbation studies showing that disrupting TAD boundaries can alter gene expression. The authors' own observation of reduced gene activity upon RAD21 degradation suggests that global TAD disruption can affect transcription. A more precise and limited conclusion, acknowledging that their data demonstrate a lack of detectable correlation between boundary distance and bursting activity in their system, would be more accurate and help reconcile these findings with the existing literature.

(2) Technical methods and data presentation:

(2.1) Accuracy and dimensionality of distance measurements: The manuscript does not clearly state whether distances are measured in 2D or 3D, nor does it sufficiently address precision limits. The stated Z-step size (1 µm) may be inadequate for accurately measuring sub-micron chromatin distances in 3D.

(2.2) Probe design and systematic error: The genomic coverage size of the BAC probes used for DNA FISH is not explicitly stated. Large probe coverage could inherently blur the precise spatial location of adjacent DNA loci. The reported average distance (~300 nm) may be influenced by the physical size of the probes, as well as systematic expansion or distortion introduced by sample fixation and FISH processing. Although such technical limitations are currently unavoidable, the authors should clarify how these factors might affect their ability to detect subtle distance changes.

(2.3) Data Visualization: The manuscript would benefit from including representative, zoomed-in regions of interest from the raw imaging data. This would allow readers to visually assess measured distance differences against background noise.

(2.4) Potential impact of resolution limits: In Figure 5, the micro-C data reveal a clear difference in interaction patterns inside versus outside the VARS2 locus TAD, yet the imaging data show no corresponding distance difference. This strongly suggests that the current imaging system, limited by optical resolution, probe size, and localisation accuracy, may be unable to resolve finer-scale spatial reorganizations associated with specific chromatin conformations (e.g., enhancer-promoter loops). The authors should explicitly discuss that their conclusion of "no coupling observed" may be constrained by the resolution and sensitivity of their method and does not preclude the possibility of detecting such associations with higher-precision measurements or in live-cell dynamics.

In summary, this study provides a valuable single-cell perspective. However, the authors should more cautiously define the scope of their findings in the manuscript and provide a more balanced discussion situating their work within the broader field.

Reviewer #2 (Public review):

Summary:

Almansour et al. investigate whether the proximity of TAD boundaries is directly linked to gene activity. The authors use high-throughput imaging to simultaneously measure the gene activity and physical distances between boundary regions in an allele-specific manner. Using transcriptional inhibitors, expression induction, and acute depletion of CTCF and cohesin, they test whether proximity of boundaries affects, or is affected by, gene activity.

Strengths:

The combined use of DNA and RNA imaging enabled simultaneous measurement of boundary proximity and transcriptional status at individual alleles. This allows single-allele correlation between boundary proximity and gene activity at multiple loci across thousands of alleles.

The use of both transcription inhibitors and transcription stimulation provides compelling and consistent evidence that boundary proximity can be disconnected from a gene's activity. The data convincingly support the conclusion that stable proximity between boundary regions is not required for ongoing transcription at the loci and timescales examined.

This work strengthens the emerging view that genome organization at the level of domain boundaries does not impose a deterministic control over transcription.

Weaknesses:

In untreated cells, the distribution of distance measurements between boundary probes is exceptionally narrow. While depletion of RAD21 clearly demonstrates an ability to detect changes in this distribution, this tight baseline distribution may limit sensitivity to more subtle changes (like those one might expect from transcriptional influences). In addition, the correlation analysis is asymmetric, primarily stratifying by transcriptional status and then comparing boundary distances. Given the central claim that boundary architecture does not influence gene activity, the analysis should be done from the opposite perspective (stratifying by boundary distance).

Strong disruption of boundary distances is only observed upon depletion of cohesin. Notably, this corresponds with the largest changes in gene activity. In contrast, depletion of CTCF actually had minimal impact on boundary distances and also had minimal impact on gene activity. This makes sense in light of previous work, where live cell imaging demonstrated that cohesin is more important for domain-structure, whereas CTCF is only important for blocking cohesin from continuing on, such that the fully formed loop occurs in a very small percentage of cells. Therefore, the fact that disruption of cohesin (more important for internal domain structure) affects gene activity while disruption of CTCF does not is exceptionally interesting but is lacking from the discussion.

On a related note, this approach primarily tests the role of boundary interactions rather than domain organization as a whole, and it should be acknowledged that internal domain structures are not directly assessed.

The comparison to work in other organisms (particularly the comparisons made to Drosophila) should be handled with care. The mechanisms underlying domain formation differ substantially across these systems, particularly regarding the differences in CTCF's role.

eLife Assessment

This study presents a high-quality cryo-EM structure of the human kinase PINK1 in complex with the HSP90-CDC37 chaperone complex, capturing a partially folded intermediate in which the C-lobe and C-terminal extension are structured while the N-lobe remains unfolded and engaged by the HSP90 clamp. The structural data are broadly consistent with a recently published structure of the same complex, providing useful insight into early steps of PINK1 maturation and highlighting residues linked to familial Parkinson's disease. However, the mechanistic conclusions remain incomplete because the manuscript does not experimentally validate key hypotheses raised by the structure, including the functional roles of the C-lobe interface, the HPNI motif, the C-terminal extension, or the proposed competition between HSP90 and TOM20.

Reviewer #1 (Public review):

Summary:

The ubiquitin kinase PINK1 accumulates on damaged mitochondria to signal the initiation of mitophagy. While we know what PINK1 looks like when it is stabilised on damaged mitochondria, not much is known about how it gets there. In this study, Okatsu et al. solve a cryoEM structure of partially folded PINK1 in complex with its chaperones HSP90 and CDC37 to a resolution of 3.08 Å. This structure captures PINK1 in a state whereby the C-lobe of its kinase domain is folded, while the N-lobe remains unfolded and stabilised by an HSP90 dimer. According to the authors' model, their structure represents cytosolic PINK1 on its way to the mitochondria. This structure also demonstrates how PINK1 is folded in a step-wise mechanism and proposes a role for residues that are mutated in Parkinson's disease.

Strengths:

PINK1 is known to be a client of the HSP90 chaperone system. Here, Okatsu et al. present a solid structural dataset showing how PINK1 interacts with HSP90 and CDC37, and they describe key residues and motifs predicted to facilitate the interactions between PINK1 and the chaperones. Notably, two key residues within interacting regions on PINK1 are also mutated in Parkinson's disease. The structure by Okatsu et al. is in line with another recently published structure of the same complex (Tian et al. Nat Comms, 2025), which appears very similar, further supporting the findings. Together, these two studies represent the first observations of cytosolic PINK1 in a semi-folded state, which provides a novel insight into PINK1 at an earlier stage within the signalling cascade.

Weaknesses:

This paper is not the first to describe the structure of the PINK1-HPS90-CDC37 complex. A study by Tian et al. was published in early December 2025 in Nature Communications, reporting a 2.84 Å structure of PINK1-HSP90-CDC37, as well as a structure of PINK1 with HSP90 and another HSP90 co-chaperone, FKBP51. It would be important to acknowledge this comparable study and to discuss how the structure in this study compares with the Tian et al. structures and whether it reveals any additional information.

Although they make claims about the functional relevance of PINK1-interacting residues, the study by Okatsu et al. does not include any biochemical or functional validation of the structure. To support their claims, the authors should test the PINK1-HSP90-CDC37 interaction using their recombinant proteins for mutants of the conserved hydrophobic PINK1 residues in the PINK1 c-lobe, H352, L353, H360, I382, D384, as well as the PINK1 HPNI motif, especially the PD mutation H271Q. The PINK1 PD mutation L347P, which interacts with the CDC37 HPNI moti,f is also worth testing.

A major question that arises from this work is whether the PINK1-HSP90-CDC37 complex is newly translated PINK1 on its way to mitochondria (as suggested by the authors) or PINK1 that has already entered mitochondria, been cleaved and then retrotranslocated. This latter scenario is the favoured model proposed by Tian et al. (Nat Comms) based on their biochemical experiments. The discrepancies between the two models should at least be discussed, and the authors should also attempt to demonstrate experimentally whether their model is correct. This question is important to address because it would allow this structural information to be placed in the greater context of PINK1 signalling.

It is also unclear what the consequences are of disrupted PINK1-HSP90-CDC37 interactions on the PINK1 signalling process more broadly - does PINK1 accumulate in the cytosol? Is there less of it? Can it still be degraded via the N-end rule? What happens during mitophagy? Perhaps some of these questions can be answered with cell-based studies using a selection of the PINK1 mutants mentioned above that disrupt the PINK1-HSP90-CDC37 complex formation.

Reviewer #2 (Public review):

Summary:

Okatsu et al report the cryoEM structure of the PINK1-HSP90-CDC37 complex at 3.08A. To do so, they mutated the PARL cleavage site (F104M) and removed the N-terminal 103 a.a. The construct was co-expressed with HSP90beta and CDC37 in insect cells, as performed previously for other kinase-HSP90-CDC37 complexes (e.g. Raf1). Molybdate was added to prevent cycling between open and closed HSP90 conformations. The initial characterization by single particle cryoEM reveals two HSP90 conformations: closed with CDC37 dissociated, and open with the CTD of HSP90 separated. Thus, the authors crosslinked the complex, which yielded a more homogenous closed structure with clearly visible density for HSP90, CDC37, and PINK1. The structure shows an immature or partially folded kinase domain conformation for PINK1, with the C-lobe bound to HSP90 and the N-lobe unfolded. The C-lobe binds to HSP90 via the HPNI motif in CDC37, which mimics the HPNI motif found in the N-lobe of kinases, and which is conserved across kinases. The main novelty here is the interaction between the C-terminal extension (CTE) of PINK1, which must adopt another conformation than in the folded state, which would otherwise clash with HSP90. The interaction with the CTE is notably mediated by the flexible charged linker (FCL) of HSP90, which is partially disordered. In this conformation, HSP90 would clash with TOM20 binding.

Strengths:

Overall, this is well-executed structural biology work, which brings insight into the elements required to fold PINK1. The protein engineering used in this study is of great value and will help others in the field explore the function of PINK1 folding. Understanding the mode of activation of PINK1 is important, and this work brings forward hypotheses that are worthy of testing.

Weaknesses:

In the absence of functional assays, the study does not bring much novelty or biological insights. Furthermore, there are already several structures of HSP90-CDC37 bound to partially folded kinases, and a simple superposition of these structures on the model of HsPINK1 allows similar conclusions to be drawn, i.e. that it would bind a folded C-lobe and unfolded N-lobe. Furthermore, a very similar structure of PINK1 bound to HSP90-CDC37 (and FKBP5) was published in Nature Communications in December 2025 by another group. The main novelty from this work (and the paper published in December) is that the CTE adopts a different conformation compared to the mature form, but the implications of this are not explored. Furthermore, the authors propose that HSP90 would compete with TOM20, but what dictates the outcome of this competition? More importantly, how do these results help understand how PINK1 become active? Again, this is not explored.

eLife Assessment

This valuable study demonstrates molecular changes associated with age related impairment in oligodendrocyte differentiation and ability to myelinate. The identification of particular genes that are associated with this decline will provide potential future targets for therapeutic interventions. The reviewers felt that the quality of the evidence was solid while identifying some minor weaknesses that if addressed would enhance the rigor of the study.

Reviewer #1 (Public review):

Summary:

This manuscript by Ghosh and colleagues investigates the transcriptional changes within the oligodendrocyte lineage that contribute to age-related declines in oligodendrocyte differentiation and myelination. Combining bulk RNA-Seq on acutely purified oligodendrocyte lineage cells with bioinformatic approaches, the authors identify groups of genes that show different patterns of dynamic regulation during differentiation (which they term "switch" genes, or "switches"). A subset of these switch genes is differentially regulated with age. The authors identify two transcription factors, Bcl11a and Foxm1, that are downregulated during differentiation, have predicted binding site enrichment at other switch genes, and are downregulated in aged OPCs. Functionally testing Bcl11a, the authors show that Bcl11a knockdown inhibits the differentiation of young OPCs in culture, whereas overexpression promotes the differentiation of aged OPCs. Viral expression of Bcl11a in Sox10-expressing cells accelerates the formation of Plp1+ oligodendrocytes in aged rodents following lysolecithin induced demyelination.

Strengths:

The work is clearly presented and addresses an important biological problem. The bioinformatic approaches used in the manuscript are powerful, and the identification of Bcl11a as a modulator of oligodendrocyte differentiation is a novel finding. The combined in vitro and in vivo approaches to assess the function of Bcl11a in oligodendrocyte differentiation are a substantial strength of the work.

Weaknesses:

Although the PCA plots show distinct and reproducible global gene expression differences between the different isolated cell populations, the authors do not present a figure showing expression levels of typical stage-specific markers (e.g., Pdgfra, Pcdh15, C1ql1 for OPCs, Bcas1, Enpp6, Gpr17 for preOLs, Mobp, Mog, etc. for OLs) or confirm the absence of markers of other lineages (astrocytes, neurons, microglia, etc.). This makes it difficult to evaluate the success of their cell isolation strategy at different ages without reanalyzing the raw data. In addition, other publicly available datasets (e.g., the Barres lab bulk RNA-Seq datasets from PMID 25186741 or the Castelo-Branco lab single cell datasets from PMID 27284195) do not show downregulation of Bcl11a during OL differentiation as is described here - this apparent discrepancy is not discussed.

Reviewer #2 (Public review):

Aging poses a significant challenge to the regenerative capacity of oligodendrocyte precursor cells (OPCs) to differentiate and myelinate neuronal axons. Myelin abnormalities accumulate with age, and it is likely that the ability of OPCs to differentiate into myelinating oligodendrocytes becomes progressively impaired during aging, leading to inefficient turnover of damaged myelin and oligodendrocytes, as well as reduced adaptive myelination. Understanding the molecular mechanisms underlying the compromised capacity of aged OPCs is therefore critical for addressing age-related white matter decline.

This study aims to decipher the intrinsic molecular changes that occur in aged OPCs. By profiling differentially expressed transcription factors (TFs) between young and aged OPCs, and by employing a novel bioinformatic tool to identify key TFs that undergo dynamic changes across distinct stages of OPC differentiation, the authors identify Bcl11a as a potential regulator. Bcl11a is highly expressed in young OPCs but markedly reduced in aged cells. Functional experiments further demonstrate that while Bcl11a does not affect OPC proliferation, it significantly promotes the differentiation of aged OPCs. Importantly, this effect is also observed in vivo following demyelinating injury in aged mice.

While the study provides compelling evidence that BCL11A represents a limiting factor for OPC differentiation during ageing, the downstream targets and molecular mechanisms through which BCL11A exerts its effects are not directly addressed. As such, the work should be interpreted primarily as identifying a key regulatory node rather than a fully defined molecular pathway.

Overall, this study offers valuable insights into the age-related loss of regenerative capacity in the central nervous system and introduces a computational framework that may be broadly useful for investigating dynamic gene regulation in other biological contexts.

Major Points:

(1) MACS mouse anti-A2B5 microbeads are not OPC-specific and may also label astrocyte precursor cells or immature astrocytes. How do the authors justify this caveat? Could some of the claimed "OPC-specific" switch genes in fact be enriched in astrocyte lineage cells?

(2) Overall, Figures 1 and 2 are not very informative in terms of biological insight. The authors should provide more detail in the main figures regarding the enriched gene sets associated with each of the Type 1-4 switch categories. For example, summarizing the top Gene Ontology terms for each switch type would greatly enhance interpretability.

(3) A similar issue applies to Figure 3. The authors should explicitly specify the transcription factors in the main figure, particularly the 27 TFs identified through the ENCODE/ReMap2 analysis.

(4) Have the authors validated Bcl11a expression across different CNS cell types and between young and aged conditions using independent methods such as qPCR, immunofluorescence, or western blotting?

(5) Regarding OPC aging, an open question is whether the reduced differentiation capacity of aged OPCs is an intrinsic property of the cells themselves or whether it results from prolonged exposure to an aging environment that induces non-cell-autonomous epigenetic or genetic changes, thereby rendering OPCs less efficient at differentiating. It would be helpful if the authors could expand on this point in the Discussion, with reference to relevant previous studies and experimental evidence.

(6) Do the authors observe a change in the number or density of OPCs between young and aged mice?

(7) The in vivo characterization of Bcl11a overexpression using the AAV-based approach appears incomplete. Do aged mice overexpressing Bcl11a in Sox10⁺ cells exhibit reduced age-related myelin degeneration under baseline conditions? In the LPC model, do the authors observe differences in lesion size and/or remyelination efficiency?

(8) Are the authors presenting gSWITCH for the first time in this manuscript? Given that the gSWITCH framework is novel and central to the study, its conceptual contribution could be emphasized more strongly. A brief comparison with existing trajectory- or pattern-based methods-ideally in the main text around Figure 1-would help readers better appreciate its novelty.

(9) The evolutionary analysis also appears somewhat disconnected from the rest of the study. Could the authors leverage available public datasets to test whether a similar Bcl11a expression trajectory is observed in human oligodendrocyte lineage cells?

eLife Assessment

This useful study develops an individual-based model to investigate the evolution of division of labor in vertebrates, comparing the contributions of group augmentation and kin selection. The findings are solid in showing that, within the specific structure of the model and the parameter space explored, group augmentation can robustly favor the evolution of differentiated helper roles, particularly when age-dependent task switching and dominance dynamics are allowed to evolve. However, the evidence only partially supports the authors' broader claim that group augmentation is the primary driver of vertebrate division of labor. Several modelling assumptions, including the limited scope for synergistic task benefits, the restriction of helper effects to group-size-mediated benefits, and the relatively narrow exploration of cost and benefit parameters, constrain the potential for kin selection to generate division of labor and limit the generality of the conclusions.

Reviewer #2 (Public review):

Summary:

This paper formulates an individual-based model to understand the evolution of division of labor in vertebrates. The model considers a population subdivided in groups, each group has a single asexually-reproducing breeder, other group members (subordinates) can perform two types of tasks called "work" or "defense", individuals have different ages, individuals can disperse between groups, each individual has a dominance rank that increases with age, and upon death of the breeder a new breeder is chosen among group members depending on their dominance. "Workers" pay a reproduction cost by having their dominance decreased, and "defenders" pay a survival cost. Every group member receives a survival benefit with increasing group size. There are 6 genetic traits, each controlled by a single locus, that control propensities to help and disperse, and how task choice and dispersal relate to dominance. To study the effect of group augmentation without kin selection, the authors cross-foster individuals to eliminate relatedness. The paper allows for the evolution of the 6 genetic traits under some different parameter values to study the conditions under which division of labour evolves, defined as the occurrence of different subordinates performing "work" and "defense" tasks. The authors envision the model as one of vertebrate division of labor.

The main conclusion of the paper is that group augmentation is the primary factor causing the evolution of vertebrate division of labor, rather than kin selection. This conclusion is drawn because, for the parameter values considered, when the benefit of group augmentation is set to zero, no division of labor evolves and all subordinates perform "work" tasks but no "defense" tasks.

Strengths:

The model incorporates various biologically realistic details, including the possibility to evolve age polytheism where individuals switch from "work" to "defence" tasks as they age or vice versa, as well as the possibility of comparing the action of group augmentation alone with that of kin selection alone.

Weaknesses from the previous round of review::

The model and its analysis are limited, which in my view makes the results insufficient to reach the main conclusion that group augmentation and not kin selection is the primary cause of the evolution of vertebrate division of labour. There are several reasons.

First, although the main claim that group augmentation drives the evolution of division of labour in vertebrates, the model is rather conceptual in that it doesn't use quantitative empirical data that applies to all/most vertebrates and vertebrates only. So, I think the approach has a conceptual reach rather than being able to achieve such conclusion about a real taxon.

Second, I think that the model strongly restricts the possibility that kin selection is relevant. The two tasks considered essentially differ only by whether they are costly for reproduction or survival. "Work" tasks are those costly for reproduction and "defense" tasks are those costly for survival. The two tasks provide the same benefits for reproduction (eqs. 4, 5) and survival (through group augmentation, eq. 3.1). So, whether one, the other, or both helper types evolve presumably only depends on which task is less costly, not really on which benefits it provides. As the two tasks give the same benefits, there is no possibility that the two tasks act synergistically, where performing one task increases a benefit (e.g., increasing someone's survival) that is going to be compounded by someone else performing the other task (e.g., increasing that someone's reproduction). So, there is very little scope for kin selection to cause the evolution of labour in this model. Note synergy between tasks is not something unusual in division of labour models, but is in fact a basic element in them, so excluding it from the start in the model and then making general claims about division of labour is unwarranted. In their reply, the authors point out that they only consider fertility benefits as this, according to them, is what happens in cooperative breeders with alloparental care; however, alloparental care entails that workers can increase other's survival *without group augmentation*, such as via workers feeding young or defenders reducing predator-caused mortality, as a mentioned in my previous review but these potentially kin-selected benefits are not allowed here.

Third, the parameter space is understandably little explored. This is necessarily an issue when trying to make general claims from an individual-based model where only a very narrow parameter region of a necessarily particular model can be feasibly explored. As in this model the two tasks ultimately only differ by their costs, the parameter values specifying their costs should be varied to determine their effects. In the main results, the model sets a very low survival cost for work (yh=0.1) and a very high survival cost for defense (xh=3), the latter of which can be compensated by the benefit of group augmentation (xn=3). Some limited variation of xh and xn is explored, always for very high values, effectively making defense unevolvable except if there is group augmentation. In this revision, additional runs have been included varying yh and keeping xh and xn constant (Fig. S6), so without addressing my comment as xn remains very high. Consequently, the main conclusion that "division of labor" needs group augmentation seems essentially enforced by the limited parameter exploration, in addition to the second reason above.

Fourth, my view is that what is called "division of labor" here is an overinterpretation. When the two helper types evolve, what exists in the model is some individuals that do reproduction-costly tasks (so-called "work") and survival-costly tasks (so-called "defense"). However, there are really no two tasks that are being completed, in the sense that completing both tasks (e.g., work and defense) is not necessary to achieve a goal (e.g., reproduction). In this model there is only one task (reproduction, equation 4,5) to which both helper types contribute equally and so one task doesn't need to be completed if completing the other task compensates for it; instead, it seems more fitting to say that there are two types of helpers, one that pays a fertility cost and another one a survival cost, for doing the same task. So, this model does not actually consider division of labor but the evolution of different helper types where both helper types are just as good at doing the single task but perhaps do it differently and so pay different types of costs. In this revision, the authors introduced a modified model where "work" and "defense" must be performed to a similar extent. Although I appreciate their effort, this model modification is rather unnatural and forces the evolution of different helper types if any help is to evolve.

I should end by saying that these comments don't aim to discourage the authors, who have worked hard to put together a worthwhile model and have patiently attended to my reviews. My hope is that these comments can be helpful to build upon what has been done to address the question posed.

[Editors' note: the authors have provided responses to the each of these points.]

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #2 (Public review):

Summary:

This paper formulates an individual-based model to understand the evolution of division of labor in vertebrates. The model considers a population subdivided in groups, each group has a single asexually-reproducing breeder, other group members (subordinates) can perform two types of tasks called "work" or "defense", individuals have different ages, individuals can disperse between groups, each individual has a dominance rank that increases with age, and upon death of the breeder a new breeder is chosen among group members depending on their dominance. "Workers" pay a reproduction cost by having their dominance decreased, and "defenders" pay a survival cost. Every group member receives a survival benefit with increasing group size. There are 6 genetic traits, each controlled by a single locus, that control propensities to help and disperse, and how task choice and dispersal relate to dominance. To study the effect of group augmentation without kin selection, the authors cross-foster individuals to eliminate relatedness. The paper allows for the evolution of the 6 genetic traits under some different parameter values to study the conditions under which division of labor evolves, defined as the occurrence of different subordinates performing "work" and "defense" tasks. The authors envision the model as one of vertebrate division of labor.

The main conclusion of the paper is that group augmentation is the primary factor causing the evolution of vertebrate division of labor, rather than kin selection. This conclusion is drawn because, for the parameter values considered, when the benefit of group augmentation is set to zero, no division of labor evolves and all subordinates perform "work" tasks but no "defense" tasks.

Strengths:

The model incorporates various biologically realistic details, including the possibility to evolve age polytheism where individuals switch from "work" to "defense" tasks as they age or vice versa, as well as the possibility of comparing the action of group augmentation alone with that of kin selection alone.

Weaknesses:

The model and its analysis are limited, which in my view makes the results insufficient to reach the main conclusion that group augmentation and not kin selection is the primary cause of the evolution of vertebrate division of labor. There are several reasons.

(1) First, although the main claim that group augmentation drives the evolution of division of labor in vertebrates, the model is rather conceptual in that it doesn't use quantitative empirical data that applies to all/most vertebrates and vertebrates only. So, I think the approach has a conceptual reach rather than being able to achieve such a conclusion about a real taxon.

We appreciate the reviewer’s point that our model does not incorporate quantitative empirical data across vertebrate taxa. This is indeed a limitation and reflects the current lack of fine-scale datasets on task division, the influence of life-history traits, and the fitness consequences of different cooperative activities in vertebrates. One of our aims, however, is precisely to stimulate such empirical work by highlighting the value of examining division of labor in species inhabiting harsh environments, considering age/size/dominance structure when evaluating variation in cooperative activities, and incorporating defense behaviors more consistently into analyses of helping, especially since defenders are often overlooked relative to the classic helpers-at-the-nest that provision offspring. The model therefore remains directly relevant to vertebrate systems because it departs from insect-inspired approaches that focus on fitness outcomes based solely in maximizing colony productivity. Instead, it incorporates direct fitness benefits to group members, an essential feature of vertebrate cooperative breeding and of other systems with fertile “workers,” as we clarified in the discussion.

(2) Second, I think that the model strongly restricts the possibility that kin selection is relevant. The two tasks considered essentially differ only by whether they are costly for reproduction or survival. "Work" tasks are those costly for reproduction and "defense" tasks are those costly for survival. The two tasks provide the same benefits for reproduction (eqs. 4, 5) and survival (through group augmentation, eq. 3.1). So, whether one, the other, or both helper types evolve presumably only depends on which task is less costly, not really on which benefits it provides. As the two tasks give the same benefits, there is no possibility that the two tasks act synergistically, where performing one task increases a benefit (e.g., increasing someone's survival) that is going to be compounded by someone else performing the other task (e.g., increasing that someone's reproduction). So, there is very little scope for kin selection to cause the evolution of labor in this model. Note synergy between tasks is not something unusual in division of labor models, but is in fact a basic element in them, so excluding it from the start in the model and then making general claims about division of labor is unwarranted. In their reply, the authors point out that they only consider fertility benefits as this, according to them, is what happens in cooperative breeders with alloparental care; however, alloparental care entails that workers can increase other's survival *without group augmentation*, such as via workers feeding young or defenders reducing predator-caused mortality, as a mentioned in my previous review but these potentially kin-selected benefits are not allowed here.

We understand the reviewer’s concern that our model restricts the scope for kin-selected benefits by not including task-specific synergy effects—specifically, help that directly increases the survival of group members (e.g., load-lightening via feeding young, or predator defense that reduces mortality of breeders or offspring independently of group augmentation). We agree that such effects can occur in some cooperative breeders, and that they can, in principle, generate indirect fitness benefits. However, even when helpers increase the survival of breeders or reduce parental investment per offspring, these effects generally translate into higher breeder productivity—either via increased fecundity, increased survival to the next breeding attempt, or increased investment in subsequent broods. Thus, although we treat benefits in terms of enhanced breeder productivity, this formulation implicitly captures a range of help-related effects that ultimately improve the reproductive output of the breeders, including those mediated through increased survival. For this reason, we believe that the model remains relevant for vertebrate systems despite not representing each pathway separately.

(3) Third, the parameter space is understandably little explored. This is necessarily an issue when trying to make general claims from an individual-based model where only a very narrow parameter region of a necessarily particular model can be feasibly explored. As in this model the two tasks ultimately only differ by their costs, the parameter values specifying their costs should be varied to determine their effects. In the main results, the model sets a very low survival cost for work (yh=0.1) and a very high survival cost for defense (xh=3), the latter of which can be compensated by the benefit of group augmentation (xn=3). Some limited variation of xh and xn is explored, always for very high values, effectively making defense unevolvable except if there is group augmentation. In this revision, additional runs have been included varying yh and keeping xh and xn constant (Fig. S6), so without addressing my comment as xn remains very high. Consequently, the main conclusion that "division of labor" needs group augmentation seems essentially enforced by the limited parameter exploration, in addition to the second reason above.

As we have explained in previous revisions, the costs associated with work and defense are not directly comparable because they affect different fitness components: work costs reduce dominance, whereas defense costs reduce survival. Whether a particular cost is “high” or “low” can only be evaluated by examining the evolved reaction norms and identifying the ranges over which these norms change. For this reason, we focused on parameter ranges that actually generate shifts in reaction norms rather than presenting large regions of parameter space where nothing changes.

We also reiterate that we did in fact explore broader parameter ranges than those shown in the main text. Additional analyses, including those specifically designed to identify conditions under which division of labor evolves under kin selection alone, are provided in the Supplementary Material. Specifically, Figure S1 addresses the point raised by the “need” of group augmentation benefits for defense to evolve, by increasing the baseline survival x₀.

We now include one additional figure in the Supplementary Material with a lower value for the benefit of group size (x_n = 1 instead of x_n = 3), and we extended the range of x_h to include lower values (x_h = 1). As we can see in Figure S7 and Table S8, group augmentation benefits are still the primary reason for individuals to group (see dispersal values). For low benefits of group augmentation, defense evolves in harsh environments in the absence of kin selection, and in benign environments when both direct and indirect fitness benefits take place. We have also now expanded the results section to include these last results. Note that we also checked even lower values for x_h under the only kin selection implementation, with results being qualitatively similar, but chose not to include them in the manuscript since it is already a very long Supplementary Material. Here are the averages for two examples with x_h = 0.1 and when we promote division of labor:

Author response table 1.

In short, the conclusion that division of labor requires group augmentation is not an artifact of limited parameter exploration. It arises because kin selection alone favors division of labor only under highly restrictive parameter combinations, whereas including direct fitness benefits substantially expands the conditions under which division of labor evolves. This pattern is consistent across the full set of parameter combinations we examined.

(4) Fourth, my view is that what is called "division of labor" here is an overinterpretation. When the two helper types evolve, what exists in the model is some individuals that do reproduction-costly tasks (so-called "work") and survival-costly tasks (so-called "defense"). However, there are really no two tasks that are being completed, in the sense that completing both tasks (e.g., work and defense) is not necessary to achieve a goal (e.g., reproduction). In this model there is only one task (reproduction, equation 4,5) to which both helper types contribute equally and so one task doesn't need to be completed if completing the other task compensates for it; instead, it seems more fitting to say that there are two types of helpers, one that pays a fertility cost and another one a survival cost, for doing the same task. So, this model does not actually consider division of labor but the evolution of different helper types where both helper types are just as good at doing the single task but perhaps do it differently and so pay different types of costs. In this revision, the authors introduced a modified model where "work" and "defense" must be performed to a similar extent. Although I appreciate their effort, this model modification is rather unnatural and forces the evolution of different helper types if any help is to evolve.

In previous models of division of labor in eusocial insects, the implicit benefit is also colony-level productivity (see Beshers & Fewell, 2001, for a review of division of labor in insects). Even in humans, division of labor functions as a means to increase efficiency toward achieving a shared goal. Our model adopts this same interpretation, as outlined in the Introduction, but extends it by considering that different tasks may impose different fitness costs, an aspect that has been largely overlooked in the existing literature. It is precisely because fitness outcomes are not fully shared among group members in vertebrates that distinguishing these cost structures matters. Unlike eusocial insects with sterile workers, vertebrate helpers can obtain direct fitness benefits, and the model explicitly accounts for these direct benefits—something absent from most insect-inspired approaches even when direct fitness benefits can also arise in some of those systems. Thus, our framework is not simply evolving “two types of helpers doing the same task,” but instead evolving specialization in different cooperative roles that carry different fitness consequences. It is therefore suitable for our model to treat contributions to breeder productivity as a common currency, while allowing individuals to specialize in different cost-distinct forms of help.

Finally, regarding synergy: with the extension introduced in the previous revision, we now incorporate the requirement that multiple forms of help must be performed for the group to achieve maximal reproductive output. This directly addressed the reviewer’s concern about synergistic dependencies between tasks and aligns our framework with the kinds of complementarity highlighted in other models of division of labor.

In summary, the structure of the model is consistent with both the theoretical literature on division of labor and the biological realities of vertebrate cooperative systems. We believe it is important for future models to explicitly consider the different fitness benefits and costs associated with distinct cooperative behaviors, and hope that our framework encourages more targeted empirical research on division of labor in vertebrates (e.g. inclusion of data on defense, life-history traits and environmental challenges) to better inform future modelling efforts.

I should end by saying that these comments don't aim to discourage the authors, who have worked hard to put together a worthwhile model and have patiently attended to my reviews. My hope is that these comments can be helpful to build upon what has been done to address the question posed.

We appreciate the reviewer’s thoughtful and constructive comments, as well as the time invested in evaluating our work. These insights have greatly helped us improve the clarity and overall quality of the manuscript. We hope that the revisions and additional clarifications we have provided adequately address all remaining concerns.

eLife Assessment

This manuscript provides valuable novel insights into the role of interpersonal guilt in social decision-making by showing that responsibility for a partner's bad lottery outcomes influences happiness. Through the integration of neuroimaging and computational modelling methods, and by combining findings from two studies, the authors provide solid support for their claims. The findings will be of interest to researchers in the field of social neuroscience and decision making.

Reviewer #1 (Public review):

Summary:

The authors aimed to characterize neurocomputational signals underlying interpersonal guilt and responsibility. Across two studies, one behavioral and one fMRI, participants made risky economic decisions for themselves or for themselves and a partner; they also experienced a condition in which the partners made decisions for themselves and the participant. The authors also assessed momentary happiness intermittently between choices in the task. Briefly, the results demonstrated that participants' self-reported happiness decreased after disadvantageous outcomes for themselves and when both they and their partner were affected; and this effect was exacerbated when participants were responsible for their partner's low outcome, rather than the opposite, reflecting experienced guilt. Consistent with previous work, BOLD signals in the insula correlated with experienced guilt and insula-right IFG connectivity was enhanced when participants made risky choices for themselves and safe choices for themselves and a partner.

Strengths:

This study implements an interesting approach to investigating guilt and responsibility; the paradigm in particular is well-suited to approach this question, offering participants the chance to make risky vs. safe choices that affect both themselves and others. I appreciate the assessment of happiness as a metric for assessing guilt across the different task/outcome conditions, as well as the implementation of both computational models and fMRI.

Weaknesses:

In spite of the overall strengths of the study, I think there are a few areas in which the paper fell a bit short and could be improved.

Comment on the revised submission:

I appreciate the authors' attention to all of my comments and questions regarding the initial version of the paper. However, I still do not believe that the point about the small volume correction in the insula has been adequately addressed. The authors claim that because the SVC was done using an anatomically defined ROI, that it is valid and not double dipping. I understand where the authors are coming from. However, there are a few issues here. First, any use of ROIs is best done via pre-registration (Gentili et al., 2021, European Journal of Neuroscience). Second, the whole set of analyses in this section leading up to the SVC seems somewhat circular. The first step was a whole brain contrast of lottery vs. safe outcomes, which revealed activation in many areas including the insula. Then, it appears that the parameter estimates from the insula were extracted and submitted offline to linear mixed models probing for effects of outcome magnitude, social condition and time, which revealed that the insula activation demonstrated the 'sought after' effect. Next, the manuscript states that the authors attempted to confirm these results with a univariate analysis for the so-called guilt effect within regions showing a stronger response to outcomes of risky relative to safe outcomes, which again showed activation in the insula (not surprisingly), and then a small volume correction was applied to these insula voxels. While an anatomical ROI from a different study was used for the correction, the issue is that multiple analyses already revealed that the insula was involved in the effect of interest. It is unclear why this is even necessary given that the LMM analysis demonstrated the expected result.

Reviewer #2 (Public review):

Summary

This manuscript focuses on the role of social responsibility and guilt in social decision making by integrating neuroimaging and computational modeling methods. Across two studies, participants completed a lottery task in which they made decisions for themselves or for a social partner. By measuring momentary happiness throughout the task, the authors show that being responsible for a partner's bad lottery outcome leads to decreased happiness compared to trials in which the participant was not responsible for their partner's bad outcome. At the neural level, this guilt effect was reflected in increased neural activity in the anterior insula, and altered functional connectivity between the insula and the inferior frontal gyrus. Using computational modeling, the authors show that trial by trial fluctuations in happiness were successfully captured by a model including participant and partner rewards and prediction errors (a 'responsibility' model), and model-based neuroimaging analyses suggested that prediction errors for the partner were tracked by the superior temporal sulcus. Taken together, these findings suggest that responsibility and interpersonal guilt influence social decision making.

Strengths

This manuscript investigates the concept of guilt in social decision making through both statistical and computational modeling. It integrates behavioral and neural data, providing a more comprehensive understanding of the psychological mechanisms. For the behavioral results, data from two different studies is included, and although minor differences are found between the two studies, the main findings remain consistent. The authors share all their code and materials, leading to transparency and reproducibility of their methods.

The manuscript is well-grounded in prior work. The task design is inspired by a large body of previous work on social decision making, and includes the necessary conditions to support their claims (i.e., Solo, Social, and Partner conditions). The computational models used in this study are inspired by previous work, and build on well-established economic theories of decision making. The research question and hypotheses clearly extend previous findings, and the more traditional univariate results align with prior work.

The authors conducted extensive analyses, as supported by the inclusion of different linear models and computational models described in the supplemental materials. Psychological concepts like risk preferences are defined and tested in different ways, and different types of analyses (e.g., univariate and multivariate neuroimaging analyses) are used to try to answer the research questions. The inclusion and comparison of different computational models provides compelling support for the claim that partner prediction errors indeed influence task behavior, as illustrated by the multiple model comparison metrics and the good model recovery.

The authors did a good job acknowledging other factors that could differ between the conditions, including the role of other emotions (like empathy) or agency in the decision making process. These additional analyses and nuances strengthen the manuscript and the interpretability of the findings.

Weaknesses

As the authors already note, they did not directly ask participants to report their feelings of guilt. The authors clearly describe this limitation, and also note that in addition to guilt, other emotions like empathy could also be at play in interpersonal decisions. Despite this limitation, this study provides insights into the neural and behavioral mechanisms of responsibility and guilt in social decision making, and how they influence behavior.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary

The authors aimed to characterize neurocomputational signals underlying interpersonal guilt and responsibility. Across two studies, one behavioral and one fMRI, participants made risky economic decisions for themselves or for themselves and a partner; they also experienced a condition in which the partners made decisions for themselves and the participant. The authors also assessed momentary happiness intermittently between choices in the task. Briefly, results demonstrated that participants' self-reported happiness decreased after disadvantageous outcomes for themselves and when both they and their partner were affected; this effect was exacerbated when participants were responsible for their partner's low outcome, rather than the opposite, reflecting experienced guilt. Consistent with previous work, BOLD signals in the insula correlated with experienced guilt, and insula-right IFG connectivity was enhanced when participants made risky choices for themselves and safe choices for themselves and a partner.

Strengths:

This study implements an interesting approach to investigating guilt and responsibility; the paradigm in particular is well-suited to approach this question, offering participants the chance to make risky v. safe choices that affect both themselves and others. I appreciate the assessment of happiness as a metric for assessing guilt across the different task/outcome conditions, as well as the implementation of both computational models and fMRI.

We thank Reviewer 1 for their positive assessment of our manuscript.

Weaknesses:

In spite of the overall strengths of the study, I think there are a few areas in which the paper fell a bit short and could be improved.

We thank Reviewer 1 for their comments, which we have used to improve our manuscript. We hope that these changes address the issues raised by the Reviewer.

(1) While the framing and goal of this study was to investigate guilt and felt responsibility, the task implemented - a risky choice task with social conditions - has been conducted in similar ways in past research that were not addressed here. The novelty of this study would appear to be the additional happiness assessments, but it would be helpful to consider the changes noted in risk-taking behavior in the context of additional studies that have investigated changes in risky economic choice in social contexts (e.g., Arioli et al., 2023 Cerebral Cortex; Fareri et al., 2022 Scientific Reports).

We certainly agree that several previously published studies have relied on risky choice tasks with social conditions. In this revised version, we now mention these two studies in the substantially revised Introduction.

(2) The authors note they assessed changes in risk preferences between social and solo conditions in two ways - by calculating a 'risk premium' and then by estimating rho from an expected utility model. I am curious why the authors took both approaches (this did not seem clearly justified, though I apologize if I missed it). Relatedly, in the expected utility approach, the authors report that since 'the number of these types of trials varied across participants', they 'only obtained reliable estimates for [gain and loss] trials in some participants' - in study 1, 22 participants had unreliable estimates and in study 2, 28 participants had unreliable estimates. Because of this, and because the task itself only had 20 gains, 20 losses, and 20 mixed gambles per condition, I wonder if the authors can comment on how interpretable these findings are in the Discussion. Other work investigating loss aversion has implemented larger numbers of trials to mitigate the potential for unreliable estimates (e.g., Sokol-Hessner et al., 2009).

We agree that we have not clearly justified why we have taken two approaches to assess risk preferences. In short, while the expected utility approach is a more comprehensive method to model a participant’s choices, we had not sufficiently considered the need for the large number of trials required to fit such models when designing our experiment. Calculating the risk premium was the less comprehensive, simpler alternative that we could calculate for all participants. We have now mentioned this fact in the Results section. As the only difference in risk aversion across conditions was found in Study 1 using the expected utility method, which could only be successfully applied in a minority of participants, we believe that this difference should not be taken as a strong finding. We have now mentioned this fact in the revised Discussion.

(3) One thing seemingly not addressed in the Discussion is the fact that the behavioral effect did not replicate significantly in study 2.

We agree that we had not sufficiently discussed the fact that there were (slight but significant) differences in risk preferences between the Solo and Social conditions in Study 1 but not in Study 2. We now do so in the revised Discussion, and write the following:

“Participants made slightly more risk-seeking choices when deciding for themselves than for both themselves and the partner in Study 1, but this difference disappeared in Study 2. The ρ parameter on which this finding in Study 1 is based could only be estimated in a minority of participants due to a relatively low number of trials, which suggests that this finding may not be very reliable. The simpler and more robust method (evaluation of a risk premium) showed no difference in risk aversion across conditions in either study. Overall, we believe that we do not have strong evidence of differences in risk preferences across conditions.”

(4) Regarding the computational models, the authors suggest that the Reponsibility and Responsibility Redux models provided the best fit, but they are claiming this based on separate metrics (e.g., in study 1, the redux model had the lowest AIC, but the responsibility only model had the highest R^2; additionally, the basic model had the lowest BIC). I am wondering if the authors considered conducting a direct model comparison to statistically compare model fits.

We agree that we should run formal, direct model comparison tests. We now ran likelihood-ratio tests which showed that the Responsibility model was the best. We now report this in the Results section, just below Table 1:

“A likelihood ratio test (Equation 9) revealed that the Responsibility model fitted better than all the other models, including the Responsibility Redux model (Study 1: all LR ≥ 47.36, p < 0.0001; Study 2: all LR ≥ 77.83, p < 0.0001).”

(5) In the reporting of imaging results, the authors report in a univariate analysis that a small cluster in the left anterior insula showed a stronger response to low outcomes for the partner as a result of participant choice rather than from partner choice. It then seems as though the authors performed small volume correction on this cluster to see whether it survived. If that is accurate, then I would suggest that this result be removed because it is not recommended to perform SVC where the volume is defined based on a result from the same whole-brain analysis (i.e., it should be done a priori).

As indicated in the manuscript, the small insula cluster centered at [-28 24 -4] and shown in Figure 4F survived corrections for multiple tests within the anatomically-defined anterior insula (based on the anatomical maximum probability map described in Faillenot et al., 2017), which is independent of the result of our analysis. Functionally defining the small volume based on the same data would indeed be circular and misleading “double-dipping”. We have most certainly NOT done this. The reason why we selected the anterior insula is because it is one of the regions most frequently associated with guilt (see the explanations in our Introduction, which refers for example to Bastin et al., 2016; Lamm & Singer, 2010; Piretti et al., 2023). Thus we feel that performing small-volume correction within the anatomically-defined anterior insula is a valid analysis. We fully acknowledge that, independently of any correction, the effect and the cluster are small. We now write:

“We found a weak response in a small cluster within the left anterior insula (peak T = 3.95, d = 0.59, 22 voxels, peak intensity at [-28 24 -4]; Figure 4F). Given the documented association between anterior insula and guilt (see Introduction), we proceeded to test whether this result survived correction for family-wise errors due to multiple comparisons restricted to the left anterior insula gray matter [defined anatomically and thus independently from our findings, as the anterior short gyrus, middle short gyrus, and anterior inferior cortex in an anatomical maximum probability map (Faillenot et al., 2017)]. This correction resulted in a p value of 0.024. This result, although it is only a small effect in a small cluster, is consistent with the mixed model analysis reported earlier.”

Reviewer #2 (Public review):

Summary

This manuscript focuses on the role of social responsibility and guilt in social decision-making by integrating neuroimaging and computational modeling methods. Across two studies, participants completed a lottery task in which they made decisions for themselves or for a social partner. By measuring momentary happiness throughout the task, the authors show that being responsible for a partner's bad lottery outcome leads to decreased happiness compared to trials in which the participant was not responsible for their partner's bad outcome. At the neural level, this guilt effect was reflected in increased neural activity in the anterior insula, and altered functional connectivity between the insula and the inferior frontal gyrus. Using computational modeling, the authors show that trial-by-trial fluctuations in happiness were successfully captured by a model including participant and partner rewards and prediction errors (a 'responsibility' model), and model-based neuroimaging analyses suggested that prediction errors for the partner were tracked by the superior temporal sulcus. Taken together, these findings suggest that responsibility and interpersonal guilt influence social decision-making.

Strengths

This manuscript investigates the concept of guilt in social decision-making through both statistical and computational modeling. It integrates behavioral and neural data, providing a more comprehensive understanding of the psychological mechanisms. For the behavioral results, data from two different studies is included, and although minor differences are found between the two studies, the main findings remain consistent. The authors share all their code and materials, leading to transparency and reproducibility of their methods.

The manuscript is well-grounded in prior work. The task design is inspired by a large body of previous work on social decision-making and includes the necessary conditions to support their claims (i.e., Solo, Social, and Partner conditions). The computational models used in this study are inspired by previous work and build on well-established economic theories of decision-making. The research question and hypotheses clearly extend previous findings, and the more traditional univariate results align with prior work.

The authors conducted extensive analyses, as supported by the inclusion of different linear models and computational models described in the supplemental materials. Psychological concepts like risk preferences are defined and tested in different ways, and different types of analyses (e.g., univariate and multivariate neuroimaging analyses) are used to try to answer the research questions. The inclusion and comparison of different computational models provide compelling support for the claim that partner prediction errors indeed influence task behavior, as illustrated by the multiple model comparison metrics and the good model recovery.

We thank the reviewer very much for their comprehensive description of our study and the positive assessment of our study and approach.

Weaknesses

As the authors already note, they did not directly ask participants to report their feelings of guilt. The decrease in happiness reported after a bad choice for a partner might thus be something else than guilt, for example, empathy or feelings of failure (not necessarily related to guilt towards the other person). Although the patterns of neural activity evoked during the task match with previously found patterns of guilt, there is no direct measure of guilt included in the task. This warrants caution in the interpretation of these findings as guilt per see.

We fully agree that not directly asking participants about feelings of guilt is a clear limitation of our study. While we already mention this in our Discussion, we have expanded our discussion of the consequences on the interpretation of our results along the lines described by the reviewer in the revised manuscript. We would like to thank the reviewer for proposing these lines of thought, and have now made the following changes to the text:

In the first paragraph of the discussion, we now write: “Being responsible for choosing a lottery that yielded a low outcome for a partner made our participants feel worse than witnessing the same outcome resulting from their partner’s choice, which we interpret as interpersonal guilt; although we note that we have not asked participants specifically about which emotion they felt in these situations.

Later on, in the third paragraph focusing on the anterior insula, we now write: “This replicates a large body of evidence associating aIns with feelings of guilt evoked during social decisions (see Introduction). Because we have neither asked our participants specifically what they felt in these situations, nor specifically whether they experienced guilt, we cannot exclude the possibility that they have instead or in addition felt empathy for their partner, a feeling of failure or bad luck, or some other emotion.”

As most comparisons contrast the social condition (making the decision for your partner) against either the partner condition (watching your partner make their decision) or the solo condition (making your own decision), an open question remains of how agency influences momentary happiness, independent of potential guilt. Other open questions relate to individual differences in interpersonal guilt, and how those might influence behavior.

How agency influences momentary happiness or variations thereof during the course of an experiment such as ours is an interesting question in itself. We now ran linear mixed models assessing agency (i.e. we compared happiness in conditions Solo & Social conditions vs. Partner condition), which revealed lower happiness in Solo and Social conditions (i.e. when it was the participant’s turn to decide) in both studies. This is interesting in itself and may reflect the drive behind responsibility aversion reported by Edelson et al.’s 2018 study: being assigned the role of the decider in a social setting may make people slightly unhappy, perhaps due to “weight of the responsibility”. We now report these findings in the Results section, including this proposed explanation; because we were not specifically interested in responsibility aversion, we do not discuss this further in the Discussion. The edited text is under the new subsection entitled ‘Momentary happiness: effects of agency, responsibility and guilt’, on page 12:

“Next, we assessed whether happiness varied depending on the participant’s agency (Social + Solo vs. Partner), and found happiness to be lower when the participant chose, independent of the outcome (Study 1: t(3600) = -3.92, p = 0.00009, β = -0.14, 95% CI = [-0.20 -0.07]; Study 2: t(2870) = -6.07, p = 0.000000001, β = -0.24, 95% CI = [-0.31 -0.16]). . This is interesting in itself and may reflect the drive behind responsibility aversion reported by Edelson et al.’s 2018 study: being assigned the role of the decider in a social setting may make people slightly unhappy, perhaps due to “weight of the responsibility”. To specifically search for a sign of interpersonal guilt, [...]”

Regarding individual differences: this is a very interesting topic that we have not addressed here due to the (relatively) small number of participants in our studies, but we might consider this for future follow-up studies, which we mention in the Discussion paragraph regarding open questions.

This manuscript is an impressive combination of multiple approaches, but how these different approaches relate to each other and how they can aid in answering slightly different questions is not very clearly described. The authors could improve this by more clearly describing the different methods and their added value in the introduction, and/or by including a paragraph on implications, open questions, and future work in the discussion.

We thank the reviewer for their appreciation of our complementary approach, and agree that we had not sufficiently explained the reasons why we used several methods. We have now added a paragraph explaining this at the end of the Introduction (page 5):

“We analysed our behavioural data using several complementary methods: choices were modelled with mixed-effects regressions serving as manipulation checks; risk preferences expressed in choices were assessed using a comprehensive expected utility model as well as with a simpler, more robust “risk premium” approach; and happiness data were fitted, in addition to the computational models, with several linear mixed models to assess the impact of both the participant’s and their partner’s rewards, the impact of agency and their interactions. Inspired by findings reported in previous neuroimaging of social emotions, we also used several methods to analyse our fMRI data, including conventional methods (both region-of-interest and mass univariate); mixed-effects regression models; computational model-based analyses (inspired by e.g. Konovalov et al., 2021; Rutledge et al., 2014); and functional connectivity (e.g. Edelson et al., 2018; Konovalov et al., 2021). The behavioural modelling is thus complemented by neuroimaging analyses that offer insight about both the activity in regions associated with guilt as well as their place in a wider network, providing an in-depth comprehensive analysis of the mechanisms behind guilt evoked by social responsibility.”

In addition, as suggested we added the following paragraph on open questions and future work in the Discussion:

“Several open questions remain at the end of this study. As discussed above, asking participants directly about which emotions they have felt during the different stages of this task would allow us to link subjective experience with our analytical measures. Testing more participants would allow us to assess the impact of inter-individual variations in personality traits on the experience as well as the behavioural and neural correlates of guilt and responsibility. Using more trials in the experiment would allow separate modelling of risk preferences in gain and loss trials in each experimental condition using expected utility models, and could allow testing whether changes in momentary happiness affect subsequent choices. Varying partner identities (friends, strangers, artificial agent) could reveal the impact of social discounting on guilt and responsibility. In sum, we believe that this experimental approach lends itself very well to the study of several aspects of social emotions.”

However, taken together, this study provides useful insights into the neural and behavioral mechanisms of responsibility and guilt in social decision-making and how they influence behavior. 

We thank the reviewer again for their appreciation of our work and hope that our revisions improved the manuscript.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

The majority of my suggestions are in the public review, so I will not repeat them here. But in general, I like the paper, and in addition to my other comments, I think that there should be more discussion of the potential limitations of the study and conclusions that can be drawn. I also thought parts of the results were a little hard to follow, particularly in the 'momentary happiness' section. Perhaps an additional subsection here might help with flow.

We agree that we could have discussed further the limitations of our study and the conclusions that can be drawn from it, which we have now done in the last paragraphs of the Discussion in this revised version.

To improve the structure of the section on ‘momentary happiness’, we separated this section into two, entitled: ‘Momentary happiness: links to reward‘ and ‘Momentary happiness: effects of agency, responsibility and guilt’, which should facilitate the reading of this long section. We proceeded in a similar manner for the Choices section, which is now subdivided into ‘Choices: manipulation check’ and ‘Choices: risk preferences’. We believe that these changes have indeed improved the readability of our manuscript.

Reviewer #2 (Recommendations for the authors):

Overall, I believe this manuscript was well-designed, consists of extensive analyses, and provides interesting new insights into the mechanisms underlying social decision-making. I mostly have some clarifying questions and minor comments, which are described below. 

(1) Integration of prior findings in the first paragraphs of the Introduction. Although all the previous work described in the 2nd-5th paragraph introduction is interesting, it felt a bit like an enumeration of findings rather than an integrated introduction leading to the current research question. At the end of paragraph 5, it becomes clear how these findings relate to the current research question, but I believe it will improve the flow and readability of the introduction if this becomes clear earlier on.

We agree that we could have integrated the cited previous work into the Introduction so that the text builds up to the research question. We have now extensively reworked several paragraphs in the Introduction (pages 3-5) and hope that these changes have made it easier to follow.

(2) For the risk attitudes (Choices), you describe pooling the gains and losses and then comparing the social and solo conditions. I was wondering whether you also looked at potential differences between gains and losses (delta measure) for social versus the solo condition (so a comparison of the delta). Based on prior work, I can imagine that the difference in risk attitudes for gains and losses might differ when making decisions for yourself versus when you're doing it for a partner. In general, I was wondering how you explain these findings, as there is also a lot of work showing differences in risk-taking patterns for gains and losses.

We agree that we could have compared delta measures between solo and social conditions. However, as we describe in the Results section and comment on in the Discussion, the relatively low number of trials made separate fitting of gain and loss trials across conditions difficult. While this question could thus be addressed in subsequent versions of our experiment with more trials, such a fine-grained analysis of the decisions was not the focus of our current study.

(3) On page 11, you state: "in particular the partner's reward prediction errors resulting from the participants' decisions, i.e. those pRPE for which participants were responsible." From the results described in the paragraph above, this doesn't become clear (e.g., there's no distinction made between social_pRPE and partner_pRPE in the text), as it only discusses differences in weights between pRPE and sRPE. I would recommend including some more information in the main text on these main modeling findings, so one doesn't have to go to the Supplemental Materials to understand them.

We did indeed fail to report these findings in the text! We thank the reviewer for pointing this out. We have now edited this passage as follows:

“Crucially, we find here that the partner’s reward prediction errors (social_pRPE and partner_pRPE) contributed to explaining changes in participants’ momentary happiness: the Responsibility and ResponsibilityRedux models explained the data better than the models without these parameters (see Table 1). In particular, the partner’s reward prediction errors resulting from the participants’ decisions (social_pRPE), i.e. those pRPE for which participants were responsible, contributed to explaining our data (weights for social_pRPE were greater than 0: Responsibility model: Study 1: Z = 2.85, p = 0.004, Study 2: Z = 3.26, p = 0.001; Responsibility Redux model: Study 1: Z = 2.93, p = 0.003, Study 2: Z = 3.30, p = 0.001; weights for social_pRPE tended to be higher than weights for partner_pRPE: Responsibility model: Study 1: Z = 2.14, p = 0.033; Study 2: Z = 1.41, p = 0.16).”

(4) The functional connectivity findings seem to come out of nowhere and are not introduced or described anywhere prior in the manuscript. It is therefore not completely clear why you conducted these analyses, or what they add above and beyond previous analyses. Already introducing this method earlier on would fix that.

We agree that we could have introduced functional connectivity analyses earlier in the text, particularly given the many previous studies in our field using this technique. We have now done this at the end of a new last paragraph of the Introduction:

“Inspired by findings reported in previous neuroimaging of social emotions, we also used several methods to analyse our fMRI data, including conventional methods (both region-of-interest and mass univariate); mixed-effects regression models; computational model-based analyses (inspired by e.g. Konovalov et al., 2021; Rutledge et al., 2014); and functional connectivity (e.g. Edelson et al., 2018; Konovalov et al., 2021). The behavioural modelling is thus complemented by neuroimaging analyses that offer insight about both the activity in regions associated with guilt as well as their place in a wider network, providing an in-depth comprehensive analysis of the mechanisms behind guilt evoked by social responsibility.”

(5) For the functional connectivity findings: I was wondering why you only looked at the choice phase, and not at the feedback phase. I understand that previous work focused on the choice phase, but for the purpose of this study (focus on guilt), I can imagine it is also interesting to see what happens with feedback. In the discussion, you also state "How we feel when we witness our decisions' consequences on others is an important signal to consider when attempting to make good social decisions." (p. 19), which is more focused on the feedback rather than choice, and also supports the idea that looking at the feedback moment might be relevant.

We agree that we could also have looked at the functional connectivity during the feedback phase. The main reason why we had originally not done so was time constraints. At the current time we would in addition point out that the manuscript is already very long and contains many analyses of behavioural and fMRI data. Adding this analysis would cost additional time and would further delay the publication of our manuscript, which we would prefer to avoid. However, one could of course look at these effects in subsequent analyses of the same data or in subsequent versions of this experiment. We have now mentioned this in the Discussion, in the paragraphs on open questions.

Minor comments:

(1) For some of the Figures, it would be helpful if the subtitles were more informative. For Figure 2 and Figure 3 for example, it would be nice if Study 1 and Study 2 were not only mentioned in the figure description but also in the actual figure. For Figures 3 and 4, it would be helpful to have significance stars for the bar plots as well.

We agree that these changes make the figures more easily understandable and have implemented them all, except for adding stars on Figure 4, because all bar plots in panels C and E would have been labeled with two or more stars, which would have made the figure difficult to read. We have now mentioned the fact that all these coefficients were significant in the figure legend.

(2) For some of the Supplementary Results, it would be very helpful if there was a legend or description. This is already the case for most of the SR, but not for all.

We have now added a legend to all elements of the Supplementary Results.

Some questions that came to mind while going through them:

- Supplementary Table 1: which p-values correspond to the significance stars? This information is included for Supplementary Table 2, but not for ST1. 

We have now added the missing information in ST1.

- Supplementary Figure 1: do the colors correspond to different participants? 

We have now specified that the colors do indeed correspond to different participants.

- Supplementary Table 5 (final table): what do the - represent? As in, why is there no value for "run" for the MPFC? At first, I thought you only included the significant values, but then I noticed a few non-significant values as well, so it wasn't completely clear to me why some of the values were missing. This also applies to Supplementary Table 6.

We have indeed forgotten to explain this. The ‘-’ in Supplementary Tables 4 and 6 indicate that the linear mixed model without the factor ‘run’ was the better-fitting one. We have now added the following explanation in the text accompanying Supplementary Table 4:

“We tested these models both with and without the factor Run and associated interaction, and we report the best-fitting model in the table below: a dash (‘-’) in the row displaying parameters for the run and socialVsSolo:run regressors indicates that the model without factor run was better-fitting for this ROI.”

(3) I came across a few minor typos or sentences that were not completely clear to me.

- On page 3: "Patients with damage to ventromedial prefrontal cortex (vmPFC) seem insensitive to guilt when playing social economic games (Krajbich et al., 2009)." This sentence felt a bit out of nowhere and doesn't logically follow from the previous sentences. 

We have now revised the descriptions of this previous study as well as several others and how they fit into the research question.

- On page 3: "In another study, participant errors in a difficult perception task lead to a partner feeling pain and evoked activations in left aIns and dlPFC (Koban et al., 2013)." This sentence doesn't really flow, and from the wording, it is not completely clear whether it's the errors or the partner pain that led to the aIns and dlPFC activation.

We have now revised the description of this study as well, as follows:

“In another study, partners received painful stimuli when participants made errors during a difficult perception task. These errors evoked activations in the left aIns and dlPFC in the participants (Koban et al., 2013).”

- Supplementary Figure 1: there is a missing period after the sentence "We then compared these new estimated parameters to the actual parameters from which the synthetic data were generated"

We have now added a missing comma after “generated”.

- On page 5: "We ran two experiments, Study 1 outside fMRI and Study 2 during fMRI, with separate groups of participants." I would change "outside fMRI" to outside the MRI scanner or something like that, as it's not completely correct to say "outside fMRI".

We have changed the sentence to “outside the MRI scanner”.

- On page 6: for the first result, there are currently two p-values reported (p < 2.5e-20 and p < 2e-16). I believe this is an error?

This was indeed an error! We have re-run this analysis, noticed that also the degrees of freedom were miscalculated, and have updated this result and the effect of condition (solo vs social). Results are almost identical as previously and all conclusions hold. We have also checked the other analyses reported in this paragraph – all results replicate exactly.

- On page 6: "Supplemental Table 1" should be "Supplementary Table 1" (for consistency).

Done.

On page 8: "participants in both conditions of both studies", I would change "of both studies" to "for both studies".

Done.

On page 8: for the "Momentary Happiness" paragraph, it would be helpful if you could briefly describe the Rutledge method here, for people who are unfamiliar with the approach.

We now write the following at the beginning of this paragraph:

“Following Rutledge and colleagues’ methodology, which considers that changes in momentary happiness in response to outcomes of a probabilistic reward task are explained by the combined influence of recent reward expectations and prediction errors arising from those expectations, we fitted computational models to each participant’s happiness data.”

On page 10: "Wilkoxon sign-rank tests", should be "Wilcoxon".

Done.

We thank the reviewer for their careful reading of our manuscript. We believe that these changes have indeed improved our manuscript.

eLife Assessment

This study investigates how RNA molecules modulate phase separation, aggregation, and cytotoxicity of the staphylococcal virulent peptide PSMα3 and the human host‑defence peptide LL‑37 using an array of biophysical and cell‑based assays. If validated, these findings would be important, as they suggest that nucleic acids can tune the material state and bioactivity of amyloids, with implications for host-pathogen interactions and for the design of therapeutics that target phase behaviour. However, the evidence is incomplete: many key claims rest on qualitative imaging and contested assumptions about "functional" amyloids, and the absence of quantitative binding data, phase diagrams, and appropriate controls limits confidence in the conclusions.

Reviewer #1 (Public review):

Summary:

The manuscript by Rayan et al. aims to elucidate the role of RNA as a context-dependent modulator of liquid-liquid phase separation (LLPS), aggregation, and bioactivity of the amyloidogenic peptides PSMα3 and LL-37, motivated by their structural and functional similarities.

Strengths:

The authors combine extensive biophysical characterization with cell-based assays to investigate how RNA differentially regulates peptide aggregation states and associated cytotoxic and antimicrobial functions.

Weaknesses:

While the study addresses an interesting and timely question with potentially broad implications for host-pathogen interactions and amyloid biology, several aspects of the experimental design and data analysis require further clarification and strengthening.

Major Comments:

(1) In Figure 1A, the author showed "stronger binding affinity" based on shifts at lower peptide concentrations, but no quantitative binding parameters (e.g., apparent Kd, fraction bound, or densitometric analysis) are presented. This claim would be better supported by including: (i) A binding curve with quantification of free vs bound RNA band intensities (ii) Replicates and error estimates (mean {plus minus} SD).

(2) The authors report droplet formation at low RNA (50 ng/µL) but protein aggregation at high RNA (400 ng/µL) through fluorescence microscopy. However, no intermediate RNA concentrations (e.g., 100-300 ng/µL) are tested or discussed, leaving a critical gap in understanding the full phase diagram and transition mechanisms. Additionally, the behaviour of PSMα3 in the absence of RNA under LLPS conditions is not shown. Without protein-only data, it is difficult to assess if droplets are RNA-induced or if protein has a weak baseline LLPS that RNA tunes. The saturation concentration (csat) for PSMα3 phase separation, either in the absence or presence of RNA, should be reported.

(3) For a convincing LLPS claim, it is important to show: Quantitative FRAP curves (mobile fraction and half-time of recovery) rather than only microscopy images and qualitative statements.

(4) The manuscript highly relies on fluorescence microscopy to show colocalization. However, the colocalization is presented in a qualitative manner only. The manuscript would benefit from the inclusion of quantitative metrics (e.g., Pearson's correlation coefficient, Manders' overlap coefficients, or intensity correlation analysis).

(5) In Figures 3 B and 3C, the contrast between "no AT630 at 30 min, strong at 2 h" (50 ng/μL) and "strong at 30 min" (400 ng/μL) is compelling, but a simple quantification (e.g., mean fluorescence intensity per area) would greatly increase rigor.

(6) In Figure S3 ssCD data, if possible, indicate whether the α-helical signal increases with RNA concentration or shows a non-linear dependence, which might link to the LLPS vs solid aggregate regimes.

(7) In Figure 5B, FRAP recovery in dying cells may reflect artifactual mobility rather than biological relevance. Additionally, the absence of quantification data limits interpretation; providing recovery curves would clarify relevance.

(8) The narrative conflates cytotoxicity endpoints (membrane damage, PI staining, aggregates) with localization data (nucleolar foci), creating ambiguity about whether nucleolar targeting drives toxicity or is a consequence of cell death. Separating toxicity assessment from localization analysis, or clearly demonstrating that nucleolar accumulation precedes cytotoxicity, would resolve this ambiguity.

(9) In Figure 8, to strengthen the LLPS assignment for LL-37, additional evidence, such as FRAP analysis or observation of droplet fusion events, would be valuable. This is particularly relevant given that the heat shock conditions (65{degree sign}C for 15 minutes) could potentially induce partial denaturation or nonspecific coacervation.

Reviewer #2 (Public review):

In this paper, Rayan et al. report that RNA influences cytotoxic activity of the staphylococcal secreted peptide cytolysin PSMalpha3 versus human cells and E. coli by impacting its aggregation. The authors used sophisticated methods of structural analysis and described the associated liquid-liquid phase separation. They also compare the influence of RNA on the aggregation and activity of LL-37, which shows differences from that on PSMalpha3.

Strengths:

That RNA impacts PSM cytotoxicity when co-incubated in vitro becomes clear.

Weaknesses:

I have two major and fundamental problems with this study:

(1) The premise, as stated in the introduction and elsewhere, that PSMalpha3 amyloids are biologically functional, is highly debatable and has never been conclusively substantiated. The property that matters most for the present study, cytotoxicity, is generally attributed to PSM monomers, not amyloids. The likely erroneous notion that PSM amyloids are the predominant cytotoxic form is derived from an earlier study by the authors that has described a specific amyloid structure of aggregated PSMalpha3. Other authors have later produced evidence that, quite unsurprisingly, indicated that aggregation into amyloids decreases, rather than increases, PSM cytotoxicity. Unfortunately, yet other groups have, in the meantime, published in-vitro studies on "functional amyloids" by PSMs without critically challenging the concept of PSM amyloid "functionality". Of note, the authors' own data in the present study, which show strongly decreased cytotoxicity of PSMalpha3 after prolonged incubation, are in agreement with monomer-associated cytotoxicity as they can be easily explained by the removal of biologically active monomers from the solution.

(2) That RNA may interfere with PSM aggregation and influence activity is not very surprising, given that PSM attachment to nucleic acids - while not studied in as much detail as here - has been described. Importantly, it does not become clear whether this effect has biologically significant consequences beyond influencing, again not surprisingly, cytotoxicity in vitro. The authors do show in nice microscopic analyses that labeled PSMalpha3 attaches to nuclei when incubated with HeLa cells. However, given that the cells are killed rapidly by membrane perturbation by the applied PSM concentrations, it remains unclear and untested whether the attachment to nucleic acids in dying cells makes any contribution to PSM-induced cell death or has any other biological significance.

Overall, the findings can be explained in a much more straightforward way with the common concept of cytotoxicity being due to monomeric PSMs, and the impact of nucleic acids on cytotoxicity being due to lowering of the concentration of that active form by RNA attachment. Further limiting the significance of the findings, whether this interaction has any biological significance on the physiology or infectivity of the PSM producer remains largely unexplored.

Reviewer #3 (Public review):

Summary:

The manuscript by Rayan et al. aims to investigate the role of RNA in modulating both virulent amyloid and host-defense peptides, with the objective of understanding their self-assembly mechanisms, morphological features, and aggregation pathways.

Strengths:

The overall content is well-structured with a logical flow of ideas that effectively conveys the research objectives.

Weaknesses:

(1) Figure 2 displays representative FRAP images demonstrating fluorescence recovery within seconds. To gain a more comprehensive understanding of how recovery after photobleaching varies under different conditions, it is recommended to supplement these images with corresponding quantitative fluorescence recovery curves for analysis.

(2) Ostwald ripening typically leads to the shrinkage or even disappearance of smaller droplets, accompanied by the further growth of large droplets. However, the droplet size in Figure 2D decreases significantly after 2 h of incubation. This observation prompts the question, what is the driving force underlying RNA-regulated phase separation and phase transition?

(3) The manuscript aims to study the role of RNA in modulating PSMα3 aggregation by using solution-state NMR to obtain residue-specific structural information. The current NMR data, as described in the method and figure captions, were recorded in the absence of RNA. Whether RNA binding induces conformational changes of PSMα3, and how these changes alter the NMR spectra? Also, the sequential NOE walk between neighboring residues can be annotated on the spectrum for clarity.

(4) The authors claim that LL-37 shares functional, sequence, and structural similarities with PSMα3. However, no droplet formation was observed of LL-37 in the presence of RNA only. The authors then applied thermal stress to induce phase separation of LL-37. What are the main factors contributing to the different phase behaviors exhibited by LL-37 and PSMα3? What are the differences in the conformation of amyloid aggregates and the kinetics of aggregation between the condensation-induced aggregation in the presence of RNA and the conventional nucleation-elongation process in the absence of RNA for these two proteins?

Author response:

We thank the reviewers for their thoughtful and constructive comments, which greatly helped us to clarify, quantify, and strengthen both our findings and interpretations. Below, we provide a point-by-point response to each comment and describe the corresponding changes made.

Public Reviews:

Reviewer #1 (Public review):

Summary:

The manuscript by Rayan et al. aims to elucidate the role of RNA as a context-dependent modulator of liquid-liquid phase separation (LLPS), aggregation, and bioactivity of the amyloidogenic peptides PSMα3 and LL-37, motivated by their structural and functional similarities.

Strengths:

The authors combine extensive biophysical characterization with cell-based assays to investigate how RNA differentially regulates peptide aggregation states and associated cytotoxic and antimicrobial functions.

Weaknesses:

While the study addresses an interesting and timely question with potentially broad implications for host-pathogen interactions and amyloid biology, several aspects of the experimental design and data analysis require further clarification and strengthening.

Major Comments:

(1) In Figure 1A, the author showed "stronger binding affinity" based on shifts at lower peptide concentrations, but no quantitative binding parameters (e.g., apparent Kd, fraction bound, or densitometric analysis) are presented. This claim would be better supported by including: (i) A binding curve with quantification of free vs bound RNA band intensities ,(ii) Replicates and error estimates (mean {plus minus} SD).

We thank the reviewer for this suggestion. To quantitatively support the binding differences observed in Figure 1A, we have now performed densitometric analysis of the EMSA data and included the results in Figure S1. The analysis showed that the Kd for PSMα3 binding to polyAU and polyA RNA is in the same order of magnitude but lower for the polyAU, indicating a stronger binding. A description was added to the results in lines 137-145 of the revised version.

(2) The authors report droplet formation at low RNA (50 ng/µL) but protein aggregation at high RNA (400 ng/µL) through fluorescence microscopy. However, no intermediate RNA concentrations (e.g., 100-300 ng/µL) are tested or discussed, leaving a critical gap in understanding the full phase diagram and transition mechanisms.

Our initial choice of 50 ng/µL (low RNA) and 400 ng/µL (high RNA) was guided by a broader RNA titration performed by turbidity measurements across 0, 10, 20, 50, 100, 200, and 400 ng/µL (Figure S2 in the revised version). In this screen, turbidity increased up to 50 ng/µL and then decreased dose-dependently from 100–400 ng/µL. We interpret this non-monotonic behavior as consistent with a transition from a dropletrich regime (maximal light scattering at intermediate dense-phase volume) toward conditions where assemblies become larger and/or more compact and sediment out of the optical path. This is described in lines 158-161 of the revised version.

Of note, additional intermediate RNA conditions (100 and 200 ng/µL) are included in Figure S14 (of the revised version). While these experiments were performed under the heat-shock perturbation, they nevertheless support the central point that RNA tunes assembly state across intermediate concentrations rather than producing a binary low/high outcome.

Importantly, we agree with the reviewer that a full phase diagram would be the most rigorous way to define the transition mechanism. However, establishing csat and constructing a complete phase diagram would require systematic measurements of dilute-phase concentrations (e.g., centrifugation/quantification or fluorescence calibration), controlled ionic strength titrations, and time-resolved mapping, which is beyond the scope of the present study. We have therefore revised the text to avoid implying that we provide a complete phase diagram. Instead, we frame our results as a qualitative with multi-assay characterization showing that RNA concentration drives a shift from liquid-like condensates (at low RNA) toward solid-like assemblies (at high RNA), with an intermediate regime suggested by the turbidity transition and supported by additional imaging under stress. Finally, to address the “critical gap” concern directly, we add a sentence (lines 239-241) stating that: “Future work will be required to quantitatively define the phase boundaries and delineate the dominant mechanisms, such as sedimentation, dissolution, or coarsening/aging, across intermediate RNA concentrations.

(3) Additionally, the behaviour of PSMα3 in the absence of RNA under LLPS conditions is not shown. Without protein-only data, it is difficult to assess if droplets are RNA-induced or if protein has a weak baseline LLPS that RNA tunes. The saturation concentration (csat) for PSMα3 phase separation, either in the absence or presence of RNA, should be reported.

In response to the reviewer’s request, we have added Figure 2F, which shows PSMα3 alone in the absence of RNA under the same conditions. PSMα3 does not form droplets in this condition, indicating that condensate formation is RNA-dependent in the tested conditions. This is referred to in the text in lines 190-193 of the revised version. Please see our response about determining the csat in the response to the previous comment.

(4) For a convincing LLPS claim, it is important to show: Quantitative FRAP curves (mobile fraction and half-time of recovery) rather than only microscopy images and qualitative statements.

We have included quantitative FRAP analysis in Figure S4 of the revised version, showing normalized recovery curves along with extracted mobile fractions and half-times of recovery (t₁/₂). These quantitative measurements support the dynamic nature of the PSMα3–RNA. This is referred to in the text in lines 179-184 of the revised version.

(5) The manuscript highly relies on fluorescence microscopy to show colocalization. However, the colocalization is presented in a qualitative manner only. The manuscript would benefit from the inclusion of quantitative metrics (e.g., Pearson's correlation coefficient, Manders' overlap coefficients, or intensity correlation analysis).

In response, we have added quantitative colocalization analysis to the revised manuscript. Specifically, we now report Pearson’s correlation coefficients and Manders’ overlap coefficients for the dual-channel fluorescence microscopy datasets in Figure S5 of the revised version. These metrics provide an objective measure of codistribution and complement the qualitative imaging.

The analysis supports that at low RNA concentrations (droplet/condensate conditions), PSMα3 and RNA show strong colocalization, consistent with RNA being incorporated within, or closely associated with, the peptide-rich phase. In contrast, at high RNA concentrations, where the assemblies are more solid-like/amyloid-positive, the quantitative coefficients decrease, consistent with reduced overlap and an apparent spatial demixing in which RNA becomes partially excluded from the peptide-rich structures. This is referred to in the text in lines 194-203 of the revised version.

(6) In Figures 3 B and 3C, the contrast between "no AT630 at 30 min, strong at 2 h" (50 ng/μL) and "strong at 30 min" (400 ng/μL) is compelling, but a simple quantification (e.g., mean fluorescence intensity per area) would greatly increase rigor.

We have included quantitative analysis of AmyTracker630 fluorescence intensity in Figure S6 of the revised version, reporting the mean fluorescence intensity per area for the indicated conditions and time points. This quantification supports the qualitative differences observed in Figures 3B and 3C. This is now referred to in the text in lines 233-236 of the revised version.

(7) In Figure S3 ssCD data, if possible, indicate whether the α-helical signal increases with RNA concentration or shows a non-linear dependence, which might link to the LLPS vs solid aggregate regimes.

The ssCD spectra displayed in Figure S7 in the revised version (corresponding to Figure S3 in the original submission) show that the α-helical signature of PSMα3 is markedly enhanced in the presence of RNA compared to peptide alone, as evidenced by increased signal intensity, deeper minima, and more pronounced spectral features characteristic of α-helical structure. Importantly, this enhancement is more pronounced at 400 ng/µL Poly(AU) RNA than at 50 ng/µL, particularly after 2 hours of coincubation, indicating that RNA concentration influences the stabilization of α-helical assemblies. This is now more specifically detailed in the text in lines 258-263 of the revised version.

We note that solid-state CD does not allow direct quantitative deconvolution of secondary structure content (e.g., % helix) in the same manner as solution CD, due to sample anisotropy, scattering, and orientation effects inherent to dried or aggregated films. Consequently, our interpretation is qualitative rather than strictly quantitative. The ssCD data therefore suggest a non-linear dependence on RNA concentration, rather than a simple linear dose–response. This is also expected considering that phase transition, suggested by the other findings, is intrinsically non-linear.

(8) In Figure 5B, FRAP recovery in dying cells may reflect artifactual mobility rather than biological relevance. Additionally, the absence of quantification data limits interpretation; providing recovery curves would clarify relevance.

We added quantitative FRAP analysis of the effect on PSMα3 within HeLa cells, shown in Figure S8 of the revised version. Compared to PSMα3 assemblies in vitro, nucleolar PSMα3 exhibits slower fluorescence recovery and a reduced mobile fraction. The nucleolus represents a highly crowded, RNA-rich cellular environment, which is expected to impose additional constraints on molecular mobility and likely contributes to the slower recovery kinetics observed in cells. This is now more specifically detailed in the text in lines 324-333 and discussed in lines 597-607 of the revised version.

(9) The narrative conflates cytotoxicity endpoints (membrane damage, PI staining, aggregates) with localization data (nucleolar foci), creating ambiguity about whether nucleolar targeting drives toxicity or is a consequence of cell death. Separating toxicity assessment from localization analysis, or clearly demonstrating that nucleolar accumulation precedes cytotoxicity, would resolve this ambiguity.

We thank the reviewer for raising this important point. We agree that, in the current dataset, cytotoxicity readouts (membrane damage, PI staining, aggregate formation) and subcellular localization (nucleolar accumulation) are observed in close temporal proximity, which limits our ability to unambiguously assign causality. In the experiments presented here, PSMα3 was applied at concentrations known to induce rapid membrane disruption and cytotoxicity in HeLa cells. Under these conditions, PSMα3 accumulates on cellular membranes and penetrates into the cell and nucleus on very short timescales (seconds to minutes), likely preceding the temporal resolution accessible by standard live-cell fluorescence microscopy. As a result, nucleolar accumulation and cytotoxic endpoints are detected essentially concurrently, precluding a definitive determination of whether nucleolar association actively drives toxicity or occurs as a downstream consequence of membrane permeabilization and cell damage.

We therefore emphasize that, in this study, nucleolar localization is presented as a phenomenological observation consistent with RNA-rich compartment association, rather than as a demonstrated causal mechanism of cytotoxicity. We have revised the Discussion (lines 597-607) to clarify this distinction and to avoid implying that nucleolar targeting is the primary driver of cell death.

We agree that resolving this ambiguity would require systematic time-resolved and concentration-dependent experiments, including analysis at sub-toxic PSMα3 concentrations below the membrane-disruptive threshold, combined with orthogonal imaging approaches. Such experiments are planned for future work but are beyond the scope of the present study.

(10) In Figure 8, to strengthen the LLPS assignment for LL-37, additional evidence, such as FRAP analysis or observation of droplet fusion events, would be valuable. This is particularly relevant given that the heat shock conditions (65 °C for 15 minutes) could potentially induce partial denaturation or nonspecific coacervation.

In response to this comment, we have added FRAP analysis of LL-37 assemblies in the revised manuscript (Figure S12), including representative images and corresponding fluorescence recovery curves. The FRAP measurements show minimal fluorescence recovery over the acquisition window, indicating that the LL-37–RNA assemblies formed under these conditions are largely immobile and solid-like, rather than liquid-like droplets. This is now referred to in the text in lines 458-462 of the revised version.

Reviewer #2 (Public review):

In this paper, Rayan et al. report that RNA influences cytotoxic activity of the staphylococcal secreted peptide cytolysin PSMalpha3 versus human cells and E. coli by impacting its aggregation. The authors used sophisticated methods of structural analysis and described the associated liquid-liquid phase separation. They also compare the influence of RNA on the aggregation and activity of LL-37, which shows differences from that on PSMalpha3.

Strengths:

That RNA impacts PSM cytotoxicity when co-incubated in vitro becomes clear.

Weaknesses:

I have two major and fundamental problems with this study:

(1) The premise, as stated in the introduction and elsewhere, that PSMalpha3 amyloids are biologically functional, is highly debatable and has never been conclusively substantiated. The property that matters most for the present study, cytotoxicity, is generally attributed to PSM monomers, not amyloids. The likely erroneous notion that PSM amyloids are the predominant cytotoxic form is derived from an earlier study by the authors that has described a specific amyloid structure of aggregated PSMalpha3. Other authors have later produced evidence that, quite unsurprisingly, indicated that aggregation into amyloids decreases, rather than increases, PSM cytotoxicity. Unfortunately, yet other groups have, in the meantime, published in-vitro studies on "functional amyloids" by PSMs without critically challenging the concept of PSM amyloid "functionality". Of note, the authors' own data in the present study, which show strongly decreased cytotoxicity of PSMalpha3 after prolonged incubation, are in agreement with monomer-associated cytotoxicity as they can be easily explained by the removal of biologically active monomers from the solution.

We thank the reviewer for this important critique and agree that direct cytotoxicity is most plausibly mediated by soluble PSM species, while extensive fibrillation generally reduces toxicity by depleting these forms, a conclusion supported by our data and by other studies (e.g., Zheng et al 2018 and Yao et al 2019). We do not propose mature amyloid fibrils as the primary toxic entities. Rather, we use the term functional amyloid in a regulatory sense, consistent with other biological amyloids whose fibrillar states modulate activity (e.g., hormone storage amyloids or RNA-binding proteins).

In line with emerging findings, we interpret PSMα3 toxicity as arising from a dynamic assembly process rather than from a single static molecular species. We previously showed that PSMα3 forms cross-α fibrils that are thermodynamically and mechanically less stable than cross-β amyloids and readily disassemble upon heat stress, fully restoring cytotoxic activity (Rayan et al., 2023). This behavior contrasts with PSMα1, which forms highly stable cross-β fibrils that do not recover activity after heat shock, suggesting that the limited thermostability of PSMα3 is an evolved feature enabling reversible switching between inactive (stored) and active states.

Consistent with this view, both PSMα1 and PSMα3 are cytotoxic in their soluble states, yet mutants unable to fibrillate lose activity, indicating that fibrillation is required but not itself the toxic end state (Tayeb-Fligelman et al., 2017, 2020; Malishev et al., 2018). Our other studies further show that cytotoxicity toward human cells correlates with inherent or lipid-induced α-helical assemblies, rather than with inert β-sheet amyloids (RagonisBachar et al., 2022, 2026; Salinas 2020, Bücker 2022). Together, these findings support a model in which membrane-associated, dynamic α-helical assembly, which requires continuous exchange between soluble species and growing fibrils, drives membrane disruption, potentially through lipid recruitment or extraction, analogous to mechanisms proposed for human amyloids such as islet amyloid polypeptide (Sparr et al., 2004).

In the present study, we further show that RNA reshapes this dynamic landscape: while PSMα3 alone progressively loses activity upon incubation, co-incubation with RNA preserves cytotoxicity by stabilizing bioactive polymorphs and condensate-like states, whereas high RNA concentrations promote solid aggregation but nevertheless preserve activity. Thus, aggregation is neither inherently functional nor toxic, but context-dependent and environmentally regulated. Taken together, our data support a model in which PSMα3 amyloids act as a dynamic reservoir, enabling S. aureus to tune virulence by reversibly shifting between dormant and active states in response to environmental cues such as heat or RNA.

This is now discussed in lines 56-76 and 523-553 of the revised version.

(2) That RNA may interfere with PSM aggregation and influence activity is not very surprising, given that PSM attachment to nucleic acids - while not studied in as much detail as here - has been described. Importantly, it does not become clear whether this effect has biologically significant consequences beyond influencing, again not surprisingly, cytotoxicity in vitro. The authors do show in nice microscopic analyses that labeled PSMalpha3 attaches to nuclei when incubated with HeLa cells. However, given that the cells are killed rapidly by membrane perturbation by the applied PSM concentrations, it remains unclear and untested whether the attachment to nucleic acids in dying cells makes any contribution to PSM-induced cell death or has any other biological significance.

We thank the reviewer for this important point and agree that PSM–nucleic acid interactions are not unexpected and that our data do not support a direct intracellular role for RNA binding in mediating cytotoxicity. Accordingly, we do not propose nucleolar or nuclear association of PSMα3 as a causal mechanism of cell death. At the concentrations used, PSMα3 induces rapid membrane disruption, and nucleic acid association is observed along with membrane attachment, precluding conclusions about intracellular function. This limitation is now explicitly clarified in the revised manuscript. The biological significance of our findings lies instead in extracellular and environmental contexts, where PSMα3 encounters abundant nucleic acids, such as RNA or DNA released from damaged host cells or present in biofilms as now addressed in lines 622631. Our data show that RNA modulates PSMα3 aggregation trajectories, shifting the balance between liquid-like condensates and solid aggregates, and thereby regulates the persistence and timing of cytotoxic activity. In this framework, RNA acts as a context-dependent regulator of virulence, rather than as an intracellular cytotoxic cofactor, an aspect which would be studied in depth in future work. This is now addressed in the text in lines 597-607 of the revised version.

Reviewer #3 (Public review):

Summary:

The manuscript by Rayan et al. aims to investigate the role of RNA in modulating both virulent amyloid and host-defense peptides, with the objective of understanding their self-assembly mechanisms, morphological features, and aggregation pathways.

Strengths:

The overall content is well-structured with a logical flow of ideas that effectively conveys the research objectives.

Weaknesses:

(1) Figure 2 displays representative FRAP images demonstrating fluorescence recovery within seconds. To gain a more comprehensive understanding of how recovery after photobleaching varies under different conditions, it is recommended to supplement these images with corresponding quantitative fluorescence recovery curves for analysis.

In response to this comment, we have supplemented the representative FRAP images with quantitative fluorescence recovery curves, reporting normalized recovery kinetics for the indicated conditions. These data are now provided in Figure S4 of the revised manuscript, allowing direct comparison of recovery behavior across conditions (shown by microscopy in Figure 2). In addition, we have included quantitative FRAP analyses for the cellular imaging shown in Figure 5 (presented in Figure S8) and for LL-37 assemblies formed under heat-shock conditions (Figure S12). Together, these additions provide a quantitative framework for interpreting the FRAP results and strengthen the distinction between liquid-like and solid-like assembly states.

(2) Ostwald ripening typically leads to the shrinkage or even disappearance of smaller droplets, accompanied by the further growth of large droplets. However, the droplet size in Figure 2D decreases significantly after 2 h of incubation. This observation prompts the question, what is the driving force underlying RNA-regulated phase separation and phase transition?

We thank the reviewer for this observation. Across multiple samples, we consistently observe a coexistence of small droplets and larger aggregates, rather than systematic growth of larger droplets at the expense of smaller ones or a uniform decrease in droplet size. In addition, the timescales examined do not allow us to reliably assess whether diffusion-driven droplet coalescence is fast enough to draw firm conclusions about droplet size evolution. This is now addressed in the text in lines 181-184 of the revised version.

A decrease in droplet size over time is nevertheless observed in some instances and is more consistent with a time-dependent conversion of initially liquid-like condensates into more solid-like assemblies, which would reduce molecular mobility and suppress droplet coalescence. In parallel, progressive fibril formation may act as a sink for soluble peptide, leading to partial dissolution or shrinkage of less mature condensates. Together, these observations are consistent with a non-equilibrium aging process, in which RNAregulated assemblies evolve from dynamic condensates toward more solid structures rather than following equilibrium Ostwald ripening.

(3) The manuscript aims to study the role of RNA in modulating PSMα3 aggregation by using solution-state NMR to obtain residue-specific structural information. The current NMR data, as described in the method and figure captions, were recorded in the absence of RNA. Whether RNA binding induces conformational changes of PSMα3, and how these changes alter the NMR spectra? Also, the sequential NOE walk between neighboring residues can be annotated on the spectrum for clarity.

The solution-state NMR experiments were performed specifically to characterize the potential binding of EGCG to PSMα3. Due to the strong tendency of PSMα3 to undergo rapid aggregation and line broadening upon RNA addition, solutionstate NMR spectra in the presence of RNA could not be obtained at sufficient quality for residue-specific analysis. As suggested, we have updated and annotated the sequential NOE walk between neighboring residues on the relevant NOESY spectra to improve clarity.

(4) The authors claim that LL-37 shares functional, sequence, and structural similarities with PSMα3. However, no droplet formation was observed of LL-37 in the presence of RNA only. The authors then applied thermal stress to induce phase separation of LL-37. What are the main factors contributing to the different phase behaviors exhibited by LL37 and PSMα3? What are the differences in the conformation of amyloid aggregates and the kinetics of aggregation between the condensation-induced aggregation in the presence of RNA and the conventional nucleation-elongation process in the absence of RNA for these two proteins?”

We appreciate this important question and have clarified both the basis of the comparison and the origin of the divergent phase behaviors of LL-37 and PSMα3. While PSMα3 and LL-37 share key properties as short, cationic, amphipathic α-helical peptides that self-assemble and interact with nucleic acids, they differ fundamentally in their assembly architectures. PSMα3 is an amyloidogenic peptide that forms cross-α amyloid fibrils, in which α-helices stack perpendicular to the fibril axis. In contrast, LL-37 can form fibrillar or sheet-like assemblies (observed in cryo grids), but these lack canonical amyloid features without clear cross-α or cross-β amyloid order, as so far observed by crystal structures. This is now clarified in different parts of the text of the revised version. Thus, the comparison between the two peptides is functional and physicochemical rather than implying identical amyloid mechanisms. These structural differences likely underlie their distinct phase behaviors.

Because LL-37 does not follow a classical amyloid nucleation–elongation pathway, and high-resolution structural information (e.g., cryo-EM) is currently lacking, partly due to its sheet-like, non-twisted morphology (unpublished results), it is not possible to directly compare aggregation kinetics or nucleation mechanisms between LL-37 and PSMα3. It is possible that amyloidogenic systems such as PSMα3 exhibit greater flexibility in prefibrillar and fibrillar polymorphism, enabling RNA-regulated phase behavior, whereas nonamyloid assemblies such as LL-37 are more prone to stress-induced solid aggregation. We note that this interpretation is necessarily tentative and does not imply a general rule, but rather reflects differences evident in the present system.

eLife Assessment

In this valuable study, the authors present traces of bone modification on ~1.8 million-year-old proboscidean remains from Tanzania, which they infer to be the earliest evidence for stone-tool-assisted megafaunal consumption by hominins. Challenging published claims, the authors argue that persistent megafaunal exploitation roughly coincided with the earliest Acheulean tools. Notwithstanding the rich descriptive and spatial data, the behavioral inferences about hominin agency rely on traces (such as bone fracture patterns and spatial overlap) that are not unequivocal; the evidence presented to support the inferences thus remains incomplete. Given the implications of the timing and extent of hominin consumption of nutritious and energy-dense food resources, as well as of bone toolmaking, the findings of this study will be of interest to paleoanthropologists and other evolutionary biologists.

Reviewer #1 (Public review):

Domínguez-Rodrigo and colleagues make a moderately convincing case for habitual elephant butchery by Early Pleistocene hominins at Olduvai Gorge (Tanzania), ca. 1.8-1.7 million years ago. They present this at the site scale (the EAK locality, which they excavated), as well as across the penecontemporaneous landscape, analyzing a series of findspots that contain stone tools and large-mammal bones. The latter are primarily elephants, but giraffids and bovids were also butchered in a few localities. The authors claim that this is the earliest well-documented evidence for elephant butchery; doing so requires debunking other purported cases of elephant butchery in the literature, or in one case, reinterpreting elephant bone manipulation as being nutritional (fracturing to obtain marrow) rather than technological (to make bone tools). The authors' critical discussion of these cases may not be consensual, but it surely advances the scientific discourse. The authors conclude by suggesting that an evolutionary threshold was achieved at ca. 1.8 ma, whereby regular elephant consumption rich in fats and perhaps food surplus, more advanced extractive technology (the Acheulian toolkit), and larger human group size had coincided.

The fieldwork and spatial statistics methods are presented in detail and are solid and helpful, especially the excellent description (all too rare in zooarchaeology papers) of bone conservation and preservation procedures. However, the methods of the zooarchaeological and taphonomic analysis - the core of the study - are peculiarly missing. Some of these are explained along the manuscript, but not in a standard Methods paragraph with suitable references and an explicit account of how the authors recorded bone-surface modifications and the mode of bone fragmentation. This seems more of a technical omission that can be easily fixed than a true shortcoming of the study. The results are detailed and clearly presented.

By and large, the authors achieved their aims, showcasing recurring elephant butchery in 1.8-1.7 million-year-old archaeological contexts. Nevertheless, some ambiguity surrounds the evolutionary significance part. The authors emphasize the temporal and spatial correlation of (1) elephant butchery, (2) Acheulian toolkits, and (3) larger sites, but do not actually discuss how these elements may be causally related. Is it not possible that larger group size or the adoption of Acheulian technology have nothing to do with megafaunal exploitation? Alternative hypotheses exist, and at least, the authors should try to defend the causation, not just put forward the correlation. The only exception is briefly mentioning food surplus as a "significant advantage", but how exactly, in the absence of food-preservation technologies? Moreover, in a landscape full of aggressive scavengers, such excess carcass parts may become a death trap for hominins, not an advantage. I do think that demonstrating habitual butchery bears very significant implications for human evolution, but more effort should be invested in explaining how this might have worked.

Overall, this is an interesting manuscript of broad interest that presents original data and interpretations from the Early Pleistocene archaeology of Olduvai Gorge. These observations and the authors' critical review of previously published evidence are an important contribution that will form the basis for building models of Early Pleistocene hominin adaptation.

Reviewer #2 (Public review):

The authors argue that the Emiliano Aguirre Korongo (EAK) assemblage from the base of Bed II at Olduvai Gorge shows systematic exploitation of elephants by hominins about 1.78 million years ago. They describe it as the earliest clear case of proboscidean butchery at Olduvai and link it to a larger behavioral shift from the Oldowan to the Acheulean.

The manuscript makes a valuable contribution to the Olduvai Gorge record, offering a detailed description of the EAK faunal assemblage. In particular, the paper provides a high-resolution record of a juvenile Elephas recki carcass, associated lithic artifacts, and several green-broken bone specimens. These data are inherently valuable and will be of significant interest to researchers studying Early Pleistocene taphonomy.

Comments on previous round of revisions:

The revised manuscript does a good job of using less definitive language, particularly by adding "possible" qualifiers to several interpretations. This addresses the concern about overstatement.

The main issue raised in the original review, however, remains unresolved. Only two elephant bone specimens at EAK show green-bone breakage interpreted as anthropogenic, and the diagnostic basis for that interpretation is not demonstrated clearly on the EAK material itself. The manuscript discusses a suite of fracture attributes described as diagnostic of dynamic percussive breakage, but these attributes are not explicitly documented on the EAK specimens. Instead, the diagnostic traits are illustrated using material from other Olduvai contexts, and that behavior is then extrapolated to make similar claims at EAK. For a paper making a potentially important behavioral argument, the key diagnostic evidence is not clearly demonstrated at the focal assemblage.

This problem is evident in the presentation of the EAK specimens. In their response, the authors state that one EAK specimen shows "overlapping scars" and constitutes a "long bone flake"; however, these features are not clearly identifiable in the figures or captions as currently presented. The authors state that Figures S21-S23 clearly indicate human agency, including a long bone flake with overlapping scars and a view of the medullary surface, but it is unclear which specimens or surfaces these descriptions refer to. Figure S21 does appear to show green fracture and is described only as an "elephant-sized flat bone fragment with green-bone curvilinear break." Figure S22 shows the same bone and cortical surface in a different orientation, providing no additional information. In Figure S23, I cannot clearly identify a medullary surface or evidence of green-bone fracture from this image. None of these images clearly demonstrates overlapping scars, and the figures would be substantially improved by explicitly identifying the features described in the text. Even if both EAK specimens are accepted as green-broken, they do not demonstrate the co-occurrence of multiple diagnostic fracture traits such as multiple green breaks, large step fractures, hackle marks, and overlapping scars that the authors state is required to attribute dynamic percussive activity to hominins and address equifinality.

I appreciate that the authors are careful to state that spatial association between stone tools and fossils alone does not demonstrate hominin behavior, and that they treat the spatial analyses as supportive rather than decisive. While the association is intriguing, the problem is downstream: spatial association is used to strengthen an interpretation of butchery at EAK that still depends on fracture evidence that is not clearly documented at the assemblage level.

The critique concerning Nyayanga is not addressed in the revision. The manuscript proposes alternative explanations for the Nyayanga material but does not demonstrate why these are more plausible than the interpretation advanced by Plummer et al. (2023). I am not arguing that the Nyayanga material should be accepted as butchery; rather, showing that trampling is possible does not establish it as more probable than cut marks. In contrast, the EAK material is treated as evidence of butchery on the basis of evidence that, in my opinion, is more limited and less clearly demonstrated. Even if this is not the authors' intention, the uneven treatment removes an earlier megafaunal case from the comparison and strengthens the case for interpreting EAK as marking a behavioral shift toward megafaunal butchery by excluding other early cases.

While I remain concerned about how the EAK evidence is documented and interpreted, I think the manuscript is appropriate for publication and will generate useful discussion. Readers can then assess for themselves whether the available evidence supports the strength of the behavioral claims.

[Editors' note: the authors are encouraged to make this version the Version of Record.]

Author response:

The following is the authors’ response to the previous reviews

Reviewer #2 (Public review):

This problem is evident in the presentation of the EAK specimens. In their response, the authors state that one EAK specimen shows "overlapping scars" and constitutes a "long bone flake"; however, these features are not clearly identifiable in the figures or captions as currently presented. The authors state that Figures S21-S23 clearly indicate human agency, including a long bone flake with overlapping scars and a view of the medullary surface, but it is unclear which specimens or surfaces these descriptions refer to. Figure S21 does appear to show green fracture and is described only as an "elephant-sized flat bone fragment with green-bone curvilinear break." Figure S22 shows the same bone and cortical surface in a different orientation, providing no additional information. In Figure S23, I cannot clearly identify a medullary surface or evidence of green-bone fracture from this image. None of these images clearly demonstrates overlapping scars, and the figures would be substantially improved by explicitly identifying the features described in the text. Even if both EAK specimens are accepted as green-broken, they do not demonstrate the co-occurrence of multiple diagnostic fracture traits such as multiple green breaks, large step fractures, hackle marks, and overlapping scars that the authors state is required to attribute dynamic percussive activity to hominins and address equifinality.

We appreciate the reviewer’s careful evaluation of the EAK specimens. We acknowledge that the overlapping scars and medullary surface of the specimen originally shown in Figure S23 were not sufficiently clear. To address this, we have extensively revised Figure S23. In the updated Supplementary File, we have provided new annotations and line drawings that explicitly trace the outlines of the overlapping scars and clearly shows the green-bone fracture features. These enhancements ensure that the diagnostic traits discussed in the text are now directly identifiable in the visual record. This demonstrates the co-occurrence of traits: green-broken outlines and overlapping scars, which meet the criteria for identifying dynamic percussive activity. This is so following Reviewer´s 2 partial handling of our arguments; since we argued in our previous response that clear simple green-broken elephant long limb bones were an anthropogenic signature per se, given that currently no durophagous predator/scavenger (including spotted hyenas) are able to produce them. Additional secondary features like hackle marks are supportive but not necessary to attribute human agency.

I appreciate that the authors are careful to state that spatial association between stone tools and fossils alone does not demonstrate hominin behavior, and that they treat the spatial analyses as supportive rather than decisive. While the association is intriguing, the problem is downstream: spatial association is used to strengthen an interpretation of butchery at EAK that still depends on fracture evidence that is not clearly documented at the assemblage level.

The association is inferred (not demonstrated) by the strong statistical spatial association between lithics and bones. Additional taphonomic evidence (like cut marks or green-broken bones) do further support the inference but they do not demonstrate it, given the highly subjective nature of cut mark identification and the plethora of alternative scenarios: one green-broken bone would not demonstrate complete elephant butchery (it could result from a marginal exploitation of just that bone); one cutmarked bone could equally reflect several alternative access types to the remains. The reviewer recognized above the presence of green-broken elements at EAK; again, this supports anthropogenic agency better than any other alternative scenario, because one of the green-broken bones is a long bone and modern hyenas are not able to produce this kind of specimens.

The critique concerning Nyayanga is not addressed in the revision. The manuscript proposes alternative explanations for the Nyayanga material but does not demonstrate why these are more plausible than the interpretation advanced by Plummer et al. (2023). I am not arguing that the Nyayanga material should be accepted as butchery; rather, showing that trampling is possible does not establish it as more probable than cut marks. In contrast, the EAK material is treated as evidence of butchery on the basis of evidence that, in my opinion, is more limited and less clearly demonstrated. Even if this is not the authors' intention, the uneven treatment removes an earlier megafaunal case from the comparison and strengthens the case for interpreting EAK as marking a behavioral shift toward megafaunal butchery by excluding other early cases.

Again, it was never our intention to “demonstrate” anything. The reviewer is misusing this term. These types of arguments are epistemologically impossible to demonstrate. One can just discuss the heuristics of alternative scenarios. The point that we tried to make was that the Nyayanga purported cut marks on megafaunal remains are (as identified and published) impossible to differentiate from natural sedimentary abrasive marks (like trampling). Therefore, they cannot be argued to represent anthropogenic butchery on a secure basis. Especially, when they do not occur in conjunction with green-broken elements of clear dynamic loading nature.

eLife Assessment

Mitochondrial DNA (mtDNA) exhibits a degree of resistance to mutagenesis under genotoxic stress, and this study on the mitochondrial Transcription Factor A (TFAM) presents important data concerning the possible mechanisms involved. The presented data are solid, technically rigorous, and consistent with established literature findings. The experiments are well-executed, providing convincing evidence on the change of TFAM-DNA interactions following UVC irradiation.

Reviewer #1 (Public review):

Summary:

The authors investigate how UVC induced DNA damage alters the interaction between the mitochondrial transcription factor TFAM and mtDNA. Using live-cell imaging, qPCR, atomic force microscopy (AFM), fluorescence anisotropy, and high-throughput DNA-chip assays, they show that UVC irradiation reduces TFAM sequence specificity and increases mtDNA compaction without protecting mtDNA from lesion formation. From these findings the authors suggest that TFAM acts as a "sensor" of damage rather than a protective or repair-promoting factor.

Strengths:

(1) The focus on UVC damage offers a clean system to study mtDNA damage sensing independently of more commonly studied repair pathways, such as oxidative DNA damage. The impact of UVC damage is not well understood in the mitochondria and this study fills that gap in knowledge.

(2) In particular, the custom mitochondrial genome DNA chip provides high resolution mapping of TFAM binding and reveals a global loss of sequence specificity following UVC exposure.

(3) The combination of in vitro TFAM DNA biophysical approaches combined with cellular responses (gene expression, mtDNA turnover) provides a coherent multi-scale view.

(4) The authors demonstrate that TFAM induced compaction does not protect mtDNA from UVC lesions, an important contribution given assumptions about TFAM providing protection.

Weaknesses:

(1) The authors show a decrease in mtDNA levels and increased lysosomal colocalization but do not define the pathway responsible for degradation. Distinguishing between replication dilution, mitophagy, or targeted degradation would strengthen the interpretation and justifies future experiments.

(2) The manuscript briefly notes enrichment of TFAM at certain regions of the mitochondrial genome but provides little interpretation of why these regions are favored. Discussion of whether high-occupancy sites correspond to regulatory or structural elements would add valuable context.

(3) The authors provide a discrepancy between the anisotropy and binding array results. The reason for this is not clear and one wonders if an orthogonal approach for the binding experiments would elucidate this difference (minor point).

Assessment of conclusions:

The manuscript successfully meets its primary goal of testing whether TFAM protects mtDNA from UVC damage and the impact this has on the mtDNA. While their data points to an intriguing model that TFAM acts as a sensor of damaged mtDNA, the validation of this model requires further investigation to make the model more convincing. This is likely warranted for a followup study. Also the biological impact of this compaction, such as altering transcription levels is not clear in this study.

Impact and utility of the methods:

This work advances our understanding of how mitochondria manage UVC genome damage and proposes a structural mechanism for damage "sensing" independent of canonical repair. The methodology, including the custom TFAM DNA chip, will be broadly useful to the scientific community.

Context: The study supports a model in which mitochondrial genome integrity is maintained not only by repair factors, but also by selective sequestration or removal of damaged genomes. The demonstration that TFAM compaction correlates with damage rather than protection reframes an interesting role in mtDNA quality control.

Comments on revised version:

The authors addressed all concerns during the revision.

Reviewer #2 (Public review):

Summary:

King et al. present several sets of experiments aimed to address potential impact of UV irradiation on human mitochondrial DNA as well as possible role of mitochondrial TFAM protein in handling UV irradiated mitochondrial genomes. The carefully worded conclusion derived from the results of experiments performed with human HeLa cells, in vitro small plasmid DNA, with PCR-generated human mitochondrial DNA and with UV-irradiated small oligonucleotides is presented in the title of the manuscript: "UV irradiation alters TFAM binding to mitochondrial DNA". Authors also interpret results of somewhat unconnected experimental approaches to speculate that "TFAM as a potential DNA damage sensing protein in that it promotes UVC-dependent conformational changes in the [mitochondrial] nucleoids, making them more compact. They further propose that such a proposed compaction might trigger removal of UV-damaged mitochondrial genomes as well as facilitates replication of undamaged mitochondrial genomes.

Strengths:

(1) Authors presented convincing evidence that a very high dose (1500 J/m2) of UVC applied to oligonucleotides covering the entire mitochondrial DNA genome alleviates sequence specificity of TFAM binding (Figure 3). This high dose was sufficient to cause UV-lesions in a large fraction of individual oligonucleotides. The method has been developed in the lab of one of the corresponding authors (ref. 74) and is technically well refined. This result can be published as is or in combination with other data.

(2) Manuscript also presents AFM evidence (Figure 4) that TFAM, which was long known to facilitate compaction of mitochondrial genome (Alam et al., 2003; PMID 12626705 and follow up citations), causes in vitro compaction of a small pUC19 plasmid and that approximately 3 UVC lesions per plasmid molecule results in slight albeit detectable increase in TFAM compaction of the plasmid.

Both results are discussed in line of a possible extrapolation to in vivo phenomena. The revised version of the discussion includes a clear statement that no in vivo support was provided within the set of experiments presented in the manuscript.

Weaknesses:

The experiments presented on Figures 3 and 4 may support the speculation that TFAM can carry protective role of eliminating mitochondrial genomes with bulky lesions by way of excessive compaction and removal damaged genomes from the in vivo pool, however extensive additional studies that would go well beyond the experiments described in this paper are needed to fill the gap between this set of results and the proposed explanations.

Reviewer #3 (Public review):

Summary

The study is grounded in the observation that mitochondrial DNA (mtDNA) shows some resistance to mutagenesis under genotoxic stress. The manuscript focuses on the effects of UVC-induced DNA damage on TFAM-DNA binding in vitro and in cells. The authors demonstrate increased TFAM-DNA compaction following UVC irradiation in vitro, as assessed by high-throughput protein-DNA binding assays and atomic force microscopy (AFM). The authors did not observe a similar trend in fluorescence polarization assays and attributed the difference in the extent of TFAM oligomerization as a potential reason. In cells, the authors found that UVC exposure increased mRNA levels of TFAM, POLG, and POLRMT without altering mitochondrial membrane potential. Overexpressing TFAM in cells or varying TFAM concentration in reconstituted nucleoids did not alter the accumulation or disappearance of mtDNA damage. Based on their data, the authors proposed a plausible model: following UVC-induced DNA damage, TFAM facilitates nucleoid compaction, which may signal damage in the mitochondrial genome. The proposed model may inspire future follow-up studies to further study the role of TFAM in sensing UVC-induced damage.

Comments on revised version:

The authors have addressed the reviewer's concerns.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary:

The authors investigate how UVC-induced DNA damage alters the interaction between the mitochondrial transcription factor TFAM and mtDNA. Using live-cell imaging, qPCR, atomic force microscopy (AFM), fluorescence anisotropy, and high-throughput DNA-chip assays, they show that UVC irradiation reduces TFAM sequence specificity and increases mtDNA compaction without protecting mtDNA from lesion formation. From these findings, the authors suggest that TFAM acts as a "sensor" of damage rather than a protective or repair-promoting factor.

Strengths:

(1) The focus on UVC damage offers a clean system to study mtDNA damage sensing independently of more commonly studied repair pathways, such as oxidative DNA damage. The impact of UVC damage is not well understood in the mitochondria, and this study fills that gap in knowledge.

(2) In particular, the custom mitochondrial genome DNA chip provides high-resolution mapping of TFAM binding and reveals a global loss of sequence specificity following UVC exposure.

(3) The combination of in vitro TFAM DNA biophysical approaches, combined with cellular responses (gene expression, mtDNA turnover), provides a coherent multi-scale view.

(4) The authors demonstrate that TFAM-induced compaction does not protect mtDNA from UVC lesions, an important contribution given assumptions about TFAM providing protection.

Weaknesses:

(1) The authors show a decrease in mtDNA levels and increased lysosomal colocalization but do not define the pathway responsible for degradation. Distinguishing between replication dilution, mitophagy, or targeted degradation would strengthen the interpretation

We thank the reviewer for their careful reading of our manuscript and thoughtful suggestions. We agree that distinguishing between replication dilution, mitophagy, and/or targeted degradation would strengthen our understanding of how UV-induced DNA damage is handled in the mitochondria. Currently we are undertaking experiments to tease this apart, but consider the scope of those experiments to be beyond this manuscript and expect to publish them in a subsequent paper rather than this one. We added text explicitly stating that these possibilities are not distinguished by our results in pages 8-9 in the Discussion under the subsection ‘Mitochondria respond to UVC-induced mtDNA damage in the absence of apparent mitochondrial dysfunction’.

(2) The sudden induction of mtDNA replication genes and transcription at 24 h suggests that intermediate timepoints (e.g., 12 hours) could clarify the kinetics of the response and avoid the impression that the sampling coincidentally captured the peak.

We agree and have added additional timepoints of 12 hours and 18 hours post exposure. We have updated Figure 2 to include the new data and have added text on page 4 to include these results.

(3) The authors report no loss of mitochondrial membrane potential, but this single measure is limited. Complementary assays such as Seahorse analysis, ATP quantification, or reactive oxygen species measurement could more fully assess functional integrity.

We focused on membrane potential because loss of membrane potential is such a well-understood of mechanism for triggering mitophagy, but agree that these additional measurements are useful. We have added experiments to assess ATP levels, but did not see changes; we have added this data to Figure 2. We have also added text highlighting that we previously assessed mtROS following the same levels of UV exposure and observed no changes (in the results section on page 5 and in the discussion section on page 9). Given that we observe no changes in membrane potential or ATP, we have opted to not move forward with Seahorse analysis for the purposes of this paper.

(4) The manuscript briefly notes enrichment of TFAM at certain regions of the mitochondrial genome but provides little interpretation of why these regions are favored. Discussion of whether high-occupancy sites correspond to regulatory or structural elements would add valuable context.

We agree a discussion of these findings provides context and insight into where the field is currently in understanding TFAM sequence specificity. We have updated text in the discussion (pages 9-10) to include our thoughts on the drivers of TFAM sequence specificity with regard to the discrepancy with the anisotropy data and the lack of overlap with regulatory/structural elements.

(5) It remains unclear whether the altered DNA topology promotes TFAM compaction or vice versa. Addressing this directionality, perhaps by including UVC-only controls for plasmid conformation, would help disentangle these effects if UVC is causing compaction alone.

We have added an additional control making this comparison and updated the text on page 7 in the results section. UVC by itself (without TFAM being present) does not alter the plasmid compaction; see new supplemental Figure S16.

(6) The authors provide a discrepancy between the anisotropy and binding array results. The reason for this is not clear, and one wonders if an orthogonal approach for the binding experiments would elucidate this difference (minor point).

The discrepancy between anisotropy and the binding array results is certainly unusual and contrary to previous studies that have used these arrays. In addition to the anisotropy experiments, we selected a ‘high occupancy’ and ‘low occupancy’ sequence from the binding array and performed oligomerization experiments using atomic force microscopy, which allowed us to detect small changes in cooperativity (see supplemental Figure S15). We previously only discussed this briefly in the results section on page 6, but we have now updated the discussion section (pages 9-10) to highlight this finding and put forth ideas for the field as to why we think this might be the case. While we do see that the binding array data aligns with oligomerization and cooperativity of TFAM, we still do not know what it is about these sequences that would drive such differences in TFAM binding, but we speculate that it could have something to do with flexibility of the DNA sequences.

Assessment of conclusions:

The manuscript successfully meets its primary goal of testing whether TFAM protects mtDNA from UVC damage and the impact this has on the mtDNA. While their data points to an intriguing model that TFAM acts as a sensor of damaged mtDNA, the validation of this model requires further investigation to make the model more convincing. This is likely warranted for a follow-up study. Also, the biological impact of this compaction, such as altering transcription levels, is not clear in this study.

We have updated wording in the Abstract, Introduction, and elsewhere in the text (as detailed in other portions of our response) to make as explicit and clear as possible which results are supported by the in vitro versus in vivo data, and which parts are conclusions supported by the data versus hypothesized models to be tested in future work.

Impact and utility of the methods:

This work advances our understanding of how mitochondria manage UVC genome damage and proposes a structural mechanism for damage "sensing" independent of canonical repair. The methodology, including the custom TFAM DNA chip, will be broadly useful to the scientific community.

Context:

The study supports a model in which mitochondrial genome integrity is maintained not only by repair factors, but also by selective sequestration or removal of damaged genomes. The demonstration that TFAM compaction correlates with damage rather than protection reframes an interesting role in mtDNA quality control.

Reviewer #2 (Public review):

Summary:

King et al. present several sets of experiments aimed to address the potential impact of UV irradiation on human mitochondrial DNA as well as the possible role of mitochondrial TFAM protein in handling UV-irradiated mitochondrial genomes. The carefully worded conclusion derived from the results of experiments performed with human HeLa cells, in vitro small plasmid DNA, with PCR-generated human mitochondrial DNA, and with UV-irradiated small oligonucleotides is presented in the title of the manuscript: "UV irradiation alters TFAM binding to mitochondrial DNA". The authors also interpret results of somewhat unconnected experimental approaches to speculate that "TFAM is a potential DNA damage sensing protein in that it promotes UVC-dependent conformational changes in the [mitochondrial] nucleoids, making them more compact." They further propose that such a proposed compaction triggers the removal of UV-damaged mitochondrial genomes as well as facilitates replication of undamaged mitochondrial genomes.

Strengths:

(1) The authors presented convincing evidence that a very high dose (1500 J/m2) of UVC applied to oligonucleotides covering the entire mitochondrial DNA genome alleviates sequence specificity of TFAM binding (Figure 3). This high dose was sufficient to cause UV lesions in a large fraction of individual oligonucleotides. The method was developed in the lab of one of the corresponding authors (reference 74) and is technically well-refined. This result can be published as is or in combination with other data.

(2) The manuscript also presents AFM evidence (Figure 4) that TFAM, which was long known to facilitate compaction of the mitochondrial genome (Alam et al., 2003; PMID 12626705 and follow-up citations), causes in vitro compaction of a small pUC19 plasmid and that approximately 3 UVC lesions per plasmid molecule result in a slight, albeit detectable, increase in TFAM compaction of the plasmid. Both results can be discussed in line with a possible extrapolation to in vivo phenomena, but such a discussion should include a clear statement that no in vivo support was provided within the set of experiments presented in the manuscript.

We thank this reviewer for their careful reading and interpretation of the manuscript. We agree that discussion of in vivo implications and extrapolations need clear statements indicating where there is not currently in vivo support. We have updated the text throughout the paper to include this.

Weaknesses:

Besides the experiments presented in Figures 3 and 4, other results do not either support or contradict the speculation that TFAM can play a protective role, eliminating mitochondrial genomes with bulky lesions by way of excessive compaction and removing damaged genomes from the in vivo pool.

To specify these weaknesses:

(1) Figure 1 - presents evidence that UVC causes a reduction in the number of mitochondrial spots in cells. The role of TFAM is not assessed.

We are working to understand the role of TFAM in vivo following UV irradiation, but believe that work should be included in follow up studies rather than this publication.

(2) Figure 2 - presents evidence that UVC causes lesions in mitochondrial genomes in vivo, detectable by qPCR. No direct assessment of TFAM roles in damage repair or mitochondrial DNA turnover is assessed despite the statements in the title of Figure 2 or in associated text. Approximately 2-fold change in gene expression of TFAM and of the three other genes does not provide any reasonable support to suggestion about increased mitochondrial DNA turnover over multiple explanations on related to mitochondrial DNA maintenance.

We agree and have updated the title of Figure 2 to better reflect the findings outlined in the figure as well as the text.

The new title is, “UVC causes mtDNA damage that decreases over time and is associated with upregulation of mtDNA replication genes, in the absence of apparent mitochondrial dysfunction.”

We agree that there are numerous mechanistic hypotheses that could explain the decrease in mtDNA damage over time. In Figure 1, we show that there is an overall decrease in mtDNA spots, and an increase in mtDNA-lysosome colocalization, suggestive of mtDNA degradation, which could serve to remove damaged genomes. One possibility is that TFAM is playing a role in the damage removal (but not repair per cell as these lesions are not repaired). Another is changes in mtDNA turnover via increasing the replication machinery in order the synthesize non-damaged mtDNA molecules to dilute out damage. These and other possibilities are not mutually exclusive. We have added text (pages 8-9) to make explicit that additional work will be required to distinguish these possibilities. We note that we have also added an additional experiment showing that TFAM knockdown affects mtDNA damage at baseline, as well as after UVC exposure (Figure 5J).

(3) Figure 5. Shows that TFAM does not protect either mitochondrial nucleoids formed in vitro or mitochondrial DNA in vivo from UVC lesions as well as has no effect on in vivo repair of UV lesions.

We agree that Figure 5 shows that TFAM does not protect DNA from UVC-induced lesions, and that a roughly 2-fold increase in TFAM protein does not alter damage reduction over time. We have added new data showing that in vivo, knockdown of TFAM results in an increase in baseline (control conditions) mtDNA damage, and also alters the rate of decrease of mtDNA damage over time after UVC (Figure 5J).

(4) Figure 6: Based on the above analysis, the model of the role of TFAM in sensing mtDNA damage and elimination of damaged genomes in vivo appears unsupported.

We have updated the legend for Figure 6 in which we outline our hypothesized role of TFAM in sensing mtDNA damage to ensure that readers know this has yet to be fully tested in vivo. We have also updated the Figure legend title from “proposed model” to “hypothesized model,” and changed the wording in the conclusion section (page 11) to highlight more clearly that this is a working model.

(5) Additional concern about Figure 3 and relevant discussion: It is not clear if more uniform TFAM binding to UV irradiated oligonucleotides with varying sequence as compared to non-irradiated oligonucleotides can be explained by just overall reduced binding eliminating sequence specific peaks.

We do not believe this is the case given the similar K_D values for the sequences tested. In our hands and in other publications (reviewed in PMID: 34440420), it has been well established that TFAM binds damaged DNA very well—essentially just as well as nondamaged DNA or better.

Additionally, a reduction in overall binding on these DNA arrays tends to make sequence specific peaks more apparent. We ran our experiments at both 30 nM and 300 nM TFAM specifically to be able to assess this question. The 300 nM data can be found in supplemental Figure S7. In this figure, we notice that the peaks appear more uniform at the high concentration (comparing Figure 3A to Figure S7A). That is presumably because there is so much more binding happening across the array that the peaks associated with the strongest binders become less pronounced. For the sake of brevity, we have not added this reasoning to the text, but are willing to do so if the Reviewers and Editor feel that it is important to include.

Reviewer #3 (Public review):

Summary:

The study is grounded in the observations that mitochondrial DNA (mtDNA) exhibits a degree of resistance to mutagenesis under genotoxic stress. The manuscript focuses on the effects of UVC-induced DNA damage on TFAM-DNA binding in vitro and in cells. The authors demonstrate increased TFAM-DNA compaction following UVC irradiation in vitro based on high-throughput protein-DNA binding and atomic force microscopy (AFM) experiments. They did not observe a similar trend in fluorescence polarization assays. In cells, the authors found that UVC exposure upregulated TFAM, POLG, and POLRMT mRNA levels without affecting the mitochondrial membrane potential. Overexpressing TFAM in cells or varying TFAM concentration in reconstituted nucleoids did not alter the accumulation or disappearance of mtDNA damage. Based on their data, the authors proposed a plausible model that, following UVC-induced DNA damage, TFAM facilitates nucleoid compaction, which may serve to signal damage in the mitochondrial genome.

Strengths:

The presented data are solid, technically rigorous, and consistent with established literature findings. The experiments are well-executed, providing reliable evidence on the change of TFAM-DNA interactions following UVC irradiation. The proposed model may inspire future follow-up studies to further study the role of TFAM in sensing UVC-induced damage.

Weaknesses:

The manuscript could be further improved by refining specific interpretations and ensuring terminology aligns precisely with the data presented.

(1) In line 322, the claim of increased "nucleoid compaction" in cells should be removed, as there is a lack of direct cellular evidence. Given that non-DNA-bound TFAM is subject to protease digestion, it is uncertain to what extent the overexpressed TFAM actually integrates into and compacts mitochondrial nucleoids in the absence of supporting immunofluorescence data.

We would like to thank this reviewer for their comments and suggestions. We feel these specific language changes have strengthened the interpretability of the text. The TFAM overexpression cells used in this experiment were given to us by Isaac et al., who demonstrated that when TFAM was overexpressed in this specific cell line, the nucleoids were indeed more compact, measured by Fiber-seq (Isaac et al., 2024; PMID: 38347148). We have removed the claim “increased compaction” from the section title, Figure 5 legend title, and from line 322 (now on page 8), and have also added an additional sentence to ensure the reader knows these cells have been shown to have presumed increased compaction by other groups.

(2) In lines 405 and 406, the authors should avoid equating TFAM overexpression with compaction in the cellular context unless the compaction is directly visualized or measured.

We have updated the text to ensure that it is clear that this was tested by other groups. We also changed the wording to “inaccessible (presumably compacted) nucleoids.” While we did not demonstrate altered compaction in our study, we think that based on the results from Isaac et al., it is likely that there was increased compaction. In addition, some readers might not have the context to make the connection between compaction and accessibility, so eliminating all reference to compaction could obscure the point.

(3) In lines 304 and 305 (and several other places throughout the manuscript), the authors use the term "removal rates". A "removal rate" requires a direct comparison of accumulated lesion levels over a time course under different conditions. Given the complexity of UV-induced DNA damage-which involves both damage formation and potential removal via multiple pathways-a more accurate term that reflects the net result of these opposing processes is "accumulated DNA damage levels." This terminology better reflects the final state measured and avoids implying a single, active 'removal' pathway without sufficient kinetic data.

We agree and have updated the language throughout the text as well as the results heading for this section.

(4) In line 357, the authors refer to the decrease in the total DNA damage level as "The removal of damaged mtDNA". The decrease may be simply due to the turnover and resynthesis of non-damaged mtDNA molecules. The term "removal" may mislead the casual reader into interpreting the effect as an active repair/removal process.

We agree and have restructured this sentence for clarity. We do believe there is some removal happening, given the increase in mtDNA colocalization in lysosomes alongside decrease of mtDNA spots in our live cell imaging. We have written it to reflect the inclusion of removal and resynthesis of nondamaged mtDNA molecules (see pages 8-9).

Recommendations for the authors:

Reviewing Editor Comments:

The reviewers appreciate the quality of the presented data but concur that they do not support the primary claims in the title and abstract. The reviewers also realize that in vivo evidence for the model would require extensive new experimentation that goes beyond a reasonable revision. The recommendation is to change the title and significantly revise text, figure titles and legends for transparency, and conclusions within results and discussion sections.

We thank the editor and all the reviewers for their feedback. We have added additional experiments, updated text throughout the entire paper to ensure our claims are supported, and revised our title. We feel that the changes we have made have indeed made the paper stronger, more transparent, and that the evidence put forth in this paper provides support for all claims made.

Reviewer #1 (Recommendations for the authors):

(1) Clarify mitochondrial response kinetics by adding an intermediate (e.g., 12 hrs) recovery timepoint for transcriptional analysis to resolve when TFAM and replication genes are induced.

We have added additional timepoints of 12 and 18 hours following exposure in Figure 2. These results strengthen our finding that the nuclear transcriptional program supporting mtDNA replication appears to be activated prior to the nuclear transcriptional program supporting mitochondrial transcription, in that POLG and TFAM come up before POLRMT and ND1.

(2) Strengthen functional readouts by assessing additional parameters of mitochondrial function to substantiate the claim that UVC does not impair mitochondrial performance.

We have referenced our previously-published data on mtROS and added a measurement of ATP following UVC exposure in Figure 2.

(3) Consider exploring whether mtDNA degradation occurs via mitophagy, nucleoid-phagy, or another pathway-potentially by using inhibitors or markers of these processes.

While we agree that this is an important follow up question and are currently working on experiments to address this, those experiments are outside the scope of this manuscript.

(4) Provide additional details for the high occupancy TFAM sites. Provide brief annotation or discussion of genomic regions showing strong TFAM binding under non-irradiated conditions that are lost during UVC treatment. This would be helpful to the field as a whole.

We have updated our discussion section to include this.

(5) Include or discuss a control using UVC irradiated pUC19 without TFAM to confirm that observed compaction categories are TFAM dependent rather than an UVC induced DNA distortion.

We have added in a supplemental figure (Figure S16) containing comparison of area analysis of control pUC19 and UV-irradiated pUC19 and we have added associated text in the results section of the paper.

(6) It would be interesting to explore the link between compaction to transcriptional output. In the TFAM overexpression model, the authors could measure expression of mtDNA encoded transcripts (e.g., ND1, COX1) to connect increased compaction with altered mitochondrial transcription.

While we agree that understanding how the compactional status alters mitochondrial transcription is worthwhile, we believe this is beyond the scope of this paper. Furthermore, this connection has previously been shown by Bruser et al., 2021 (PMID: 34818548) who showed that more compact nucleoids are not undergoing active transcription. It will be interesting to see in future work if mtDNA damage drives changes in both compaction as well as transcriptional activity.

(7) Clarify quantitative presentation in figure 2F to explicitly note whether the observed increase in fluorescence intensity was statistically insignificant and confirm that the assay sensitivity is sufficient to detect small potential changes. As presented it is not clear if there is a change.

We have changed the presentation of Figure 2F. There is a slight increase in membrane potential at the 24-hour time point and we have made that clear in the text as well. We included FCCP as a (standard) positive control, for which we can detect the associated decrease in membrane potential for. While it is always possible that a very small decrease occurred that we were unable to detect, we note that none of the six UVC-exposed groups that we tested even trended towards a decrease in MMP, making it less likely that there was an effect that we simply lacked the power or sensitivity to detect.

(8) It would be interesting if the authors can comment on whether TFAM induced compaction after UVC might shield mtDNA from other, repairable lesions (e.g., oxidative or alkylation damage), offering a broader context for this mechanism beyond just UVC.

In theory, we believe this is possible. It will also be interesting to see if the increased compaction following UVC also protects or shields the mtDNA from other enzymatic processes, such as repair proteins that may be searching for repairable lesions such as oxidative or alkylation damage. In this case, it seems as though the increased compaction would prevent the repair from happening at genomes harboring damage.

In this study we show with our in vitro nucleoids that the increased compaction does not protect against UVC, but this is likely because UVC does not need physical access to the DNA in order to damage it, as the wavelengths of UVC (centered in this case at 254nm) are readily absorbed by proteins and thus can go right through the proteins. Currently, we know that increased compaction by TFAM makes the DNA inaccessible to the enzymes required to methylate DNA used in Fiber-seq (PMID: 38347148), but we do not know if the compaction is tight enough to prevent ROS or alkylating agents from damaging the DNA. We have updated text in the discussion on page 10 to highlight some of these ideas.

Reviewer #2 (Recommendations for the authors):

Please, go over all display items and text and clarify details that can help readers to understand important specifics of the experiments. Examples are provided below:

(1) Abstract and Introduction - indicate species and cell line

We have updated the text to include this information.

(2) Table 1 "TFAM KD measurements"- title and footnotes are entirely cryptic. Please, clarify the experimental design, question(s) addressed and conclusions drawn from data.

We have updated the title of Table 1 to "Binding of TFAM to array sequences, measured using fluorescence anisotropy,” and clarified the footnotes to make sure it is clear which sequences were selected for AFM oligomerization experiments.

(3) Figure 3 and Material and Methods - specify UVC dose.

We have added this information to both the figure legend and the methods section.

(4) Figure 4 - specify UVC dose.

We have added this information to the figure legend.

(5) Figure 5. Panel B indicate which band is TFAM and which is HA-tag; Indicate clearly which panel is showing in vivo or in vitro results.

We have updated the figure to label the untagged TFAM and HA-tagged TFAM and changed the panel titles to specify if they are in vivo results.

eLife Assessment

This valuable study provides convincing evidence that MgdE, a conserved mycobacterial nucleomodulin, downregulates inflammatory gene transcription by interacting with the histone methyltransferase COMPASS complex and altering histone H3 lysine methylation. This work will interest microbiologists as well as cell and cancer biologists.

Reviewer #1 (Public review):

Summary:

This fundamental study identifies a new mechanism that involves a mycobacterial nucleomodulin manipulation of the host histone methyltransferase COMPASS complex to promote infection. Although other intracellular pathogens are known to manipulate histone methylation, this is the first report demonstrating specific targeting the COMPASS complex by a pathogen. The rigorous experimental design using of state-of-the art bioinformatic analysis, protein modeling, molecular and cellular interaction and functional approaches, culminating with in vivo infection modeling provide convincing, unequivocal evidence that supports the authors claims. This work will be of particular interest to cellular microbiologist working on microbial virulence mechanisms and effectors, specifically nucleomodulins, and cell/cancer biologists that examine COMPASS dysfunction in cancer biology.

Strengths:

(1) The strengths of this study include the rigorous and comprehensive experimental design that involved numerous state-of-the-art approaches to identify potential nucleomodulins, define molecular nucleomodulin-host interactions, cellular nucleomodulin localization, intracellular survival, and inflammatory gene transcriptional responses, and confirmation of the inflammatory and infection phenotype in a small animal model.

(2) The use of bioinformatic, cellular and in vivo modeling that are consistent and support the overall conclusions is a strengthen of the study. In addition, the rigorous experimental design and data analysis including the supplemental data provided, further strengthens the evidence supporting the conclusions.

Comments on revisions:

The authors have previously addressed the weaknesses that were identified by this reviewer by providing rational explanation and specific references that support the findings and conclusions.

Reviewer #2 (Public review):

Summary:

The manuscript by Chen et al addresses an important aspect of pathogenesis for mycobacterial pathogens, seeking to understand how bacterial effector proteins disrupt the host immune response. To address this question the authors sought to identify bacterial effectors from M. tuberculosis (Mtb) that localize to the host nucleus and disrupt host gene expression as a means of impairing host immune function. Their revised manuscript has strengthened their observations by performing additional experiments with BCG strains expressing tagged MgdE.

Strengths:

The researchers conducted a rigorous bioinformatic analysis to identify secreted effectors containing mammalian nuclear localization signal (NLS) sequences, which formed the basis of quantitative microscopy analysis to identify bacterial proteins that had nuclear targeting within human cells. The study used two complementary methods to detect protein-protein interaction: yeast two-hybrid assays and reciprocal immunoprecipitation (IP). The combined use of these techniques provides strong evidence of interactions between MgdE and SET1 components and suggests the interactions are in fact direct. The authors also carried out rigorous analysis of changes in gene expression in macrophages infected with MgdE mutant BCG. They found strong and consistent effects on key cytokines such as IL6 and CSF1/2, suggesting that nuclear-localized MgdE does in fact alter gene expression during infection of macrophages. The revised manuscript contains additional biochemical analyses of BCG strains expressing tagged MgdE that further supports their microscopy findings.

Reviewer #3 (Public review):

In this study, Chen L et al. systematically analyzed the mycobacterial nucleomodulins and identified MgdE as a key nucleomodulin in pathogenesis. They found that MgdE enters into host cell nucleus through two nuclear localization signals, KRIR108-111 and RLRRPR300-305, and then interacts with COMPASS complex subunits ASH2L and WDR5 to suppress H3K4 methylation-mediated transcription of pro-inflammatory cytokines, thereby promoting mycobacterial survival.

Comments on revisions:

The authors have previously adequately addressed previous concerns through additional experimentation. The revised data robustly support the main conclusions, demonstrating that MgdE engages the host COMPASS complex to suppress H3K4 methylation, thereby repressing pro-inflammatory gene expression and promoting mycobacterial survival. This work represents a significant conceptual advance.

Author response:

The following is the authors’ response to the previous reviews

Reviewer #2 (Recommendations for the authors):

Major:

Over-interpretation of data. There are a few instances of this:

The authors claim "Our work shows that MgdE interacts with both WDR5 and ASH2L and inhibits the methyltransferase activity of the COMPASS complex" (Line 318). However, they provide no biochemical analysis of methyltransferase activity to support this claim. While they cite Figure 4A-C and Figure 5, these data simply show (slightly) decreased cellular levels of H3K4Me. There are multiple ways H3K4Me could decrease including blocking recruitment of COMPASS to promoters or the enzymatic activity of MgdE itself.

The data itself related to H3K4Me changes (Figure 5D) is difficult to interpret in light of the controls they now provide. Examining the blot itself there seems to be a massive increase in H3K4Me in control cells expressing GFP that is not reflected in the quantification that shows only a ~2x increase in GFP-expressing cells. In addition， there is very little decrease in H3K4Me in the MgdE-expressing cells relative to controls or site-mutant (no change apparent visually and ~10% change per their quantification). However, the authors interpret this as," revealed that cells expressing WT MgdE exhibited lower levels of H3K4me3". In both these cases I would recommend the authors consider modifying their interpretation of the data.

We thank the reviewer for the comment.

(1) We have now revised this interpretation in the manuscript as follows:

Lines 311-312: “Our work shows that MgdE interacts with both WDR5 and ASH2L, leading to a decrease in H3K4me3 levels.”

(2) Figure 5D presents the results of three independent biological replicates. The bar graph shows the average signal intensity of H3K4me3 normalized to the corresponding loading controls. Accordingly, we have revised the analysis and description of the experimental results.

Lines 214-217: “Immunoblot analysis of nuclear extracts showed that cells expressing WT MgdE had ~25% lower H3K4me3 levels than EGFP-expressing cells and ~40% lower levels than those expressing the D244A/H47A mutant (Figure 5D).”

Minor

What is "CK"? Please clarify (Figure 2F).

We thank the reviewer for the comment. In this context, "CK" refers to the uninfected control group, which serves as the negative control in the experiment. We have revised the label in Figure 2F.

How many times was the BCG mouse experiment performed? This should be indicated in the figure legend? (Figure 7A).

We thank the reviewer for the comment. The BCG mouse experiment was performed once, and we have added this information to the figure legend of Figure 7A.

It is unclear why the secreted protein (after signal peptide removal) migrates at the same size as the full-length protein (Figure S2).

We thank the reviewer for the comment. The precursors of secreted proteins after translation in the cytoplasm will be translated into the periplasm immediately. Therefore, MgdE or Ag85B obtained from the whole-cell lysate (Figure S2A) mostly have had the signal peptides removed. This is also validated in the case of Rv0455c secretion by Mtb (Zhang et al., Nature Communications, 2022). This explains why MgdE (or Ag85B) proteins from whole-cell lysates or from supernatants show same size in SDS-PAGE gels.

It is still unclear why the transcripts with very little fold-change in expression (in grey) have the most significant p-values for being different (Figure 6).

We thank the reviewer for the comment. The p-value calculation takes into account not only the magnitude of expression change but also the consistency of expression levels within each group and the number of biological replicates. When the variation among replicates is minimal, even a small difference in group means can result in a statistically significant p-value. In our RNA-seq analysis, we used DESeq2 with three biological replicates per group. DESeq2 employs a model based on the negative binomial distribution and accounts for multiple factors, including the mean expression level, within-group variance (dispersion), sample size, and normalization accuracy. As a result, it is common to observe that genes with small variability and strong consistency between replicates may show significant p-values even with modest fold changes. Conversely, genes with larger fold changes but greater variability might not reach statistical significance.

Reference

Zhang L, Kent JE, Whitaker M, Young DC, Herrmann D, Aleshin AE, Ko YH, Cingolani G, Saad JS, Moody DB, Marassi FM, Ehrt S, Niederweis M (2022) A periplasmic cinched protein is required for siderophore secretion and virulence of Mycobacterium tuberculosis Nat Commun 13(1):2255.

eLife Assessment

In their valuable study, Beaudet, Berger and Hendricks provide a mechanistic link between disease-associated tau hyperphosphorylation, loss of cooperative tau envelope formation on microtubules, and dysregulation of axonal transport prior to aggregation. Using complementary in vitro reconstitution and human iPSC-derived neuronal assays with phosphodeficient and phosphomimetic tau constructs targeting 14 disease-relevant sites, the authors convincingly show that phosphorylation state alters tau organization on microtubules and differentially impacts kinesin- and lysosome-based transport. The evidence is solid and well aligned with the conclusions, yet the work could be further strengthened by incorporating additional controls and motor-specific assays to refine the mechanistic depth.

Reviewer #1 (Public review):

Summary:

This work by Beaudet and colleagues aims at exploring the effect of phosphorylation on the formation of tau envelopes and consequently on axonal transport, both in vitro on reconstituted microtubules and in human excitatory neurons derived from IPSCs.

The authors found that a relatively widely used construct in which 14 serine or threonine residues, often hyperphosphorylated in Alzheimer's disease, are mutated to alanines (phosphodeficient), increases the density of tau envelopes compared to wildtype tau, whereas a phosphomimetic (same residues mutated to glutamic acid) reduces envelope density both in vitro and in human excitatory neurons derived from IPSCs.

By analysing the trafficking of different kinesins (KIF1a and KIF5C), they observed different effects of tau phosphorylation status on the movement of these two motors.

They then analyse transport of lysosomes by employing live imaging of lysotracker in human excitatory neurons derived from IPSCs transfected with wildtype, phosphodeficient or phosphomimetic tau, observing that phosphodeficient tau seems to reduce transport of lysosomes while phosphomimetic increases transport compared to wildtype tau.

Strengths:

(1) The work aims to study a novel and underexplored topic in the tau field, tau envelopes, and investigate their relevance to Alzheimer's disease pathology.

(2) Experiments are well conducted and of high quality.

Weaknesses:

Relying only on in vitro reconstituted microtubules and human neurons derived from IPSCs leaves some doubts about the relevance of these results for Alzheimer's disease, considering the embryonic state of IPSCs-derived neurons.

Reviewer #2 (Public review):

This manuscript examines how disease-associated hyperphosphorylation disrupts tau's role as a cooperative microtubule-binding regulator of intracellular transport. Using in vitro reconstitution assays and live-cell imaging in iPSC-derived neurons, the authors employ phosphomutant tau constructs (E14 to mimic hyperphosphorylation, AP to prevent phosphorylation) at 14 disease-associated residues to isolate phosphorylation effects independent of expression system-dependent PTM heterogeneity. The results show that hyperphosphorylated tau fails to form cooperative envelope-like structures on microtubules, instead binding diffusely and dissociating rapidly. In contrast, wild-type and phospho-resistant tau form cohesive envelopes that regulate motor protein access. At the single-molecule level, hyperphosphorylation reduces KIF5C inhibition while maintaining or enhancing KIF1A inhibition through altered processivity and detachment rates. In live neurons, hyperphosphorylated tau phenocopies tau knockout conditions, weakening tau-mediated inhibition of lysosome transport and increasing processive motility. The authors quantify tau binding using Gaussian mixture model-based image analysis and measure tau kinetics via FRAP, demonstrating that hyperphosphorylation-induced loss of cooperative binding correlates with dysregulated organelle transport. These findings establish a mechanism by which phosphorylation-driven disruption of tau's gatekeeper function on microtubules compromises axonal transport prior to aggregation in tauopathies. The paper provides interesting new knowledge for the field, but there are outstanding concerns that could be further addressed by the authors to strengthen and clarify the current manuscript:

(1) Lack of Phosphatase-Treated Control and Explicit WT Phosphorylation Quantification

Wild-type tau expressed in insect and mammalian cells is known to be phosphorylated by endogenous kinases (eg, GSK3, CDK5, MARK). The manuscript acknowledges this in the Discussion but provides no phosphatase-treated lysate control or quantification of endogenous phosphorylation on WT tau via phospho-specific Western blots. This leaves ambiguity about whether observed differences between WT and E14 reflect purely the introduced mutations or confounding baseline differences in phosphostate content.

(2) Limited Normalization of Motor Effects to Measured Tau Lattice Occupancy

Although kinesin trajectories are classified inside vs. outside tau envelopes (inherently normalizing to local tau density), motor parameters are not systematically reported as functions of tau fluorescence intensity across all constructs. Co-purifying MAPs or microtubule-modifying enzymes in cell lysates is not quantified or excluded, leaving residual uncertainty about tau-specificity of observed motor inhibition. This should be at least acknowledged in the results section.

(3) Insufficient Citation of Prior Neuronal Tau Envelope Evidence

In the Introduction, the authors state, "it was an open question if tau forms envelopes in neurons," but this understates existing evidence. Tan et al. (2019) report tau neuronal staining consistent with envelope formation, while Siahaan et al. (2021) provide more direct evidence in non-neuronal cells. The framing should acknowledge and integrate these prior findings.

(4) Unclear Wording on Expression System-Dependent Phosphorylation

The sentence "The phosphostate of tau is strongly dependent on the expression system" requires rewording. It is ambiguous whether this refers to the final phosphostate achieved after expression or the inherent phosphorylating capacity of each system. Clearer language would strengthen the methodological justification.

(5) Insufficient Quantification of Motor and Lysosome Transport Effect Magnitudes in Results Section

The data on molecular motor motility and lysosome transport are densely described. The magnitude of effects (fold-changes, percentage differences) should be explicitly stated in the Results section when first presenting findings to orient readers to biological significance. For example, effect magnitudes for lysosome run lengths, velocities, and directional bias should be quantified in text, not left to figure inspection.

(6) Incomplete Discussion of Projection Domain Necessity for Envelope Formation

The Discussion states the projection domain is "a critical regulator of both tau-tau and tau-microtubule interactions," but does not engage with prior domain dissection work. Tan et al. (2019) found that the entire projection domain is not necessary for envelope formation in vitro. The authors should discuss which projection domain regions are specifically regulated by phosphorylation vs. required for cooperativity, providing a more nuanced interpretation than implied by their current framing.

Author response:

We thank the reviewers for their thoughtful and constructive feedback. Addressing these points will strengthen the manuscript and improve its clarity.

A primary concern involved the justification for using COS7 cell lysates in reconstitution approaches and iPSC-derived neuronal model systems as models for AD. We will clarify the language throughout the manuscript to more explicitly state the study’s goals, emphasize that these systems were selected as robust, well-controlled platforms to test the mechanisms through which tau hyperphosphorylation affects microtubule interactions and tau’s role in regulating intracellular transport, and the limitations of in vitro and iPSC models.

Reviewers also raised the possibility that background phosphorylation could contribute to the effects observed in the pseudo-phosphorylation model. We cite two recent preprints that provide insight into this question through quantitatively assessing tau phosphorylation across expression systems. In the revised manuscript, we will elaborate on how their assessment of tau phosphorylation fits within the scope of our approach and clarify how our experimental controls effectively minimize uncertainty related to background phosphorylation.

Another point concerned the potential influence of other microtubule-associated proteins in lysates and the impact of tau lattice occupancy on motility outcomes. To further strengthen this aspect, we will include additional analyses correlating tau intensity along microtubules with kinesin intensity and motility behavior, and we will more clearly explain how the AP and WT controls provide confidence in the robustness of the system.

Detailed responses to each reviewer comment are provided below point by point. The planned revisions, which include clearer language, stronger justification of the experimental approaches, and additional supporting analyses, will substantially improve the clarity, rationale, and overall impact of the study.

Public Reviews:

Reviewer #1 (Public review):

Summary:

This work by Beaudet and colleagues aims at exploring the effect of phosphorylation on the formation of tau envelopes and consequently on axonal transport, both in vitro on reconstituted microtubules and in human excitatory neurons derived from IPSCs.

The authors found that a relatively widely used construct in which 14 serine or threonine residues, often hyperphosphorylated in Alzheimer's disease, are mutated to alanines (phosphodeficient), increases the density of tau envelopes compared to wildtype tau, whereas a phosphomimetic (same residues mutated to glutamic acid) reduces envelope density both in vitro and in human excitatory neurons derived from IPSCs.

By analysing the trafficking of different kinesins (KIF1a and KIF5C), they observed different effects of tau phosphorylation status on the movement of these two motors.

They then analyse transport of lysosomes by employing live imaging of lysotracker in human excitatory neurons derived from IPSCs transfected with wildtype, phosphodeficient or phosphomimetic tau, observing that phosphodeficient tau seems to reduce transport of lysosomes while phosphomimetic increases transport compared to wildtype tau.

Strengths:

(1) The work aims to study a novel and underexplored topic in the tau field, tau envelopes, and investigate their relevance to Alzheimer's disease pathology.

(2) Experiments are well conducted and of high quality.

Weaknesses:

Relying only on in vitro reconstituted microtubules and human neurons derived from IPSCs leaves some doubts about the relevance of these results for Alzheimer's disease, considering the embryonic state of IPSCs-derived neurons.

We agree with the reviewer that iPSC-derived neurons represent an immature state compared with the neurons affected in Alzheimer’s disease. However, iPSC-derived neurons, together with in vitro reconstitution, provide insight into (1) whether tau hyperphosphorylation influences its association with microtubules and its ability to form envelope-like structures thought to regulate transport, (2) how tau hyperphosphorylation affects the motility of kinesin motors that are strongly inhibited by tau, and (3) how transport of endogenous degradative organelles such as lysosomes are impacted by tau hyperphosphorylation. We hope that our studies will help to inform future studies examining how tau-related dysfunction evolves in more mature neurons and contributes to the more severe pathological effects observed at later disease stages.

We will include a paragraph in the Discussion section addressing the limitations of this study to better contextualize our findings within the broader effort to understand tauopathies and Alzheimer’s disease.

Reviewer #2 (Public review):

This manuscript examines how disease-associated hyperphosphorylation disrupts tau's role as a cooperative microtubule-binding regulator of intracellular transport. Using in vitro reconstitution assays and live-cell imaging in iPSC-derived neurons, the authors employ phosphomutant tau constructs (E14 to mimic hyperphosphorylation, AP to prevent phosphorylation) at 14 disease-associated residues to isolate phosphorylation effects independent of expression system-dependent PTM heterogeneity. The results show that hyperphosphorylated tau fails to form cooperative envelope-like structures on microtubules, instead binding diffusely and dissociating rapidly. In contrast, wild-type and phospho-resistant tau form cohesive envelopes that regulate motor protein access. At the single-molecule level, hyperphosphorylation reduces KIF5C inhibition while maintaining or enhancing KIF1A inhibition through altered processivity and detachment rates. In live neurons, hyperphosphorylated tau phenocopies tau knockout conditions, weakening tau-mediated inhibition of lysosome transport and increasing processive motility. The authors quantify tau binding using Gaussian mixture model-based image analysis and measure tau kinetics via FRAP, demonstrating that hyperphosphorylation-induced loss of cooperative binding correlates with dysregulated organelle transport. These findings establish a mechanism by which phosphorylation-driven disruption of tau's gatekeeper function on microtubules compromises axonal transport prior to aggregation in tauopathies. The paper provides interesting new knowledge for the field, but there are outstanding concerns that could be further addressed by the authors to strengthen and clarify the current manuscript:

(1) Lack of Phosphatase-Treated Control and Explicit WT Phosphorylation Quantification

Wild-type tau expressed in insect and mammalian cells is known to be phosphorylated by endogenous kinases (eg, GSK3, CDK5, MARK). The manuscript acknowledges this in the Discussion but provides no phosphatase-treated lysate control or quantification of endogenous phosphorylation on WT tau via phospho-specific Western blots. This leaves ambiguity about whether observed differences between WT and E14 reflect purely the introduced mutations or confounding baseline differences in phosphostate content.

Tau contains ~85 putative phosphorylation sites and is modified by several kinases in cells. Studies by Siahaan et al. (2024) and Fan et al. (2025) provide detailed insight into tau phosphorylation, its role in protecting the microtubule lattice from severing enzymes, and the implications of phosphorylation patterns for aggregate formation. Specifically, Fan et al. (2025) show that HEK-expressed tau is phosphorylated by endogenous kinases at 58 residues, with most phospho-occupancy levels below 15%, indicating substantial heterogeneity among individual tau molecules. In the revised manuscript, we will (1) provide justification for the use of the pseudo-phosphorylation model system as an approach to limit heterogeneity among tau molecules, (2) clarify the importance of the WT and AP controls, (3) discuss that E14, WT, and AP tau likely exhibit similar degrees of background phospho-heterogeneity, with WT tau likely exhibiting some overlap between background phosphorylation and the 14 AD-associated sites examined, and (4) expand the discussion to emphasize that although background phosphorylation is present, our results do not suggest that it contributes significantly to the observations reported in this study.

(2) Limited Normalization of Motor Effects to Measured Tau Lattice Occupancy

Although kinesin trajectories are classified inside vs. outside tau envelopes (inherently normalizing to local tau density), motor parameters are not systematically reported as functions of tau fluorescence intensity across all constructs. Co-purifying MAPs or microtubule-modifying enzymes in cell lysates is not quantified or excluded, leaving residual uncertainty about tau-specificity of observed motor inhibition. This should be at least acknowledged in the results section.

The reviewer raises a valid point. It is challenging to compare conditions where the occupancy of tau on microtubules is similar across conditions, as phosphorylation strongly effects the interaction between tau and microtubules. We will quantify and report tau intensity in single-molecule motility assays. On the second point, while effects from other MAPs or motor proteins could potentially affect kinesin motility, we would expect that these effects would be similar for all tau phosphomutant constructs, such that the effect of tau phospho-states on kinesin motility can be assessed.

(3) Insufficient Citation of Prior Neuronal Tau Envelope Evidence

In the Introduction, the authors state, "it was an open question if tau forms envelopes in neurons," but this understates existing evidence. Tan et al. (2019) report tau neuronal staining consistent with envelope formation, while Siahaan et al. (2021) provide more direct evidence in non-neuronal cells. The framing should acknowledge and integrate these prior findings.

We agree with the reviewer that evidence from several studies using reconstitution systems, fixed neurons, and live cultured cells provides evidence of tau envelope formation in neurons. Specifically, tau envelopes have been observed along taxol-stabilized or GMPCPP-capped GDP microtubules in vitro (e.g., Dixit et al., 2008; Monroy et al., 2018; Tan et al., 2019; Siahaan et al., 2019), in 4% PFA-fixed and Triton X-100–extracted DIV7 mouse hippocampal neurons (Tan et al., 2019), and in live, non-neuronal U-2 OS cells following taxol treatment (Siahaan et al., 2022) or elevated pH (Siahaan et al., 2024). However, to our knowledge, our study is the first to demonstrate tau envelope formation in live neuronal cells under normal cell culture conditions. We will revise this sentence in the manuscript to more precisely position our findings within the context of prior studies.

(4) Unclear Wording on Expression System-Dependent Phosphorylation

The sentence "The phosphostate of tau is strongly dependent on the expression system" requires rewording. It is ambiguous whether this refers to the final phosphostate achieved after expression or the inherent phosphorylating capacity of each system. Clearer language would strengthen the methodological justification.

We agree that the wording here is ambiguous and requires clarification. In the revised manuscript, we will clarify that tau phosphorylation depends on the expression system used; bacterial systems lack the capacity for many post-translational modifications compared with insect and mammalian systems. We will also emphasize that in insect and mammalian expression systems, tau phosphorylation occurs heterogeneously, as demonstrated in previous studies by Siahaan et al. (2024) and Fan et al. (2025).

(5) Insufficient Quantification of Motor and Lysosome Transport Effect Magnitudes in Results Section

The data on molecular motor motility and lysosome transport are densely described. The magnitude of effects (fold-changes, percentage differences) should be explicitly stated in the Results section when first presenting findings to orient readers to biological significance. For example, effect magnitudes for lysosome run lengths, velocities, and directional bias should be quantified in text, not left to figure inspection.

Our initial justification for omitting quantitative data from the results text was to improve readability; however, in doing so, we may have reduced the accessibility and clarity regarding the significance of the findings. In the revised manuscript, we will incorporate the relevant quantifications and statistical significance for the motility data in the text.

(6) Incomplete Discussion of Projection Domain Necessity for Envelope Formation

The Discussion states the projection domain is "a critical regulator of both tau-tau and tau-microtubule interactions," but does not engage with prior domain dissection work. Tan et al. (2019) found that the entire projection domain is not necessary for envelope formation in vitro. The authors should discuss which projection domain regions are specifically regulated by phosphorylation vs. required for cooperativity, providing a more nuanced interpretation than implied by their current framing.

We agree with the reviewer. Tan et al. (2019) demonstrated that the proline-rich region (residues 198–244) within the projection domain of full-length 2N4R tau is the minimal region required to maintain tau’s ability to form envelopes along microtubules. We will incorporate this work on the dissection of the projection domain and discuss how the phosphorylation sites examined in our study are primarily located within this region. Together, these data highlight the proline-rich region as a potential major regulator of tau–tau cooperativity.

eLife Assessment

This study is a valuable contribution that comprehensively identifies and characterizes LC3B-binding peptides through a bacterial cell-surface display screen covering approximately 500,000 human peptides. The data presented are solid, although this approach has limitations (e.g., it cannot assess the effects of post-translational modifications, which are often relevant to LIR-mediated interactions). Validation of the newly identified binding peptides by demonstrating their interactions with full-length proteins in cells would further strengthen this manuscript.

Reviewer #1 (Public review):

Summary:

This study uses high-throughput bacterial cell-surface display to identify LC3B-interacting peptides in the human proteome. The screen is unbiased, and this type of assay has not previously been used for selecting LC3B-interacting peptides. The screen was done with a library of 500,000 peptides, and they ended up with 427 peptides that they scored as high-confidence LC3B binders. The experiments performed are solid, and data are analyzed using well-documented methods and statistics.

The aim of the authors was to isolate LC3B-interacting peptides from the human proteome, and the screen succeeded in doing so. The selected set of peptides included several previously reported LIR motifs, but also many novel LC3B binding peptides that either contained or did not contain the canonical core LIR motif [WFY]xx[LVI].

Another aim was to identify binding determinants important for the LC3B interaction, and they made an interesting sequence logo based on selected LIR-containing peptides. However, this study does not really extend our knowledge related to binding determinants essential for LIR motifs in LC3B binding. They basically verify known characteristics, including the importance of varied types of electrostatic interactions supporting the docking of the core LIR into the LDS of LC3B.

Strengths:

The approach used here (high-throughput bacterial-surface-display) is new. The screen is unbiased, and the fact that peptides are directly tested for LC3B binding may facilitate the discovery of non-canonical LIR motifs. The screen appears to be highly selective and manages to distinguish between peptides that interact with LC3B and peptides that do not interact.

Weaknesses:

It is a limitation that no proteins are analyzed in this study. Further work is therefore needed to verify that identified LIR motifs are functional in full-length proteins and in cells.

Reviewer #2 (Public review):

Summary:

To discover peptides that interact with autophagy-related protein LC3B and profile the key binding determinants, the authors screened a library of ~500,000 36-residue peptides derived from the human proteome using bacterial cell-surface display. Analysis of the screening data revealed exceptions to the reported LIR motif and a strong preference for negatively charged residues adjacent to the LIR.<br /> These results support a refinement of the LIR motif definition and expand the network of candidate LC3B interaction partners.

Strengths:

High-throughput approach.

Weaknesses:

Lack of in vitro data and molecular dynamics simulations.

Reviewer #3 (Public review):

Summary:

The LC3 family of proteins, which includes LC3B, are ubiquitin-like proteins that are covalently linked to phosphatidylethanolamine in the expanding autophagosomal membrane during autophagy. LC3 family members bind to short sequences of amino acids that reside within dynamic regions in a wide variety of proteins. These sequences, termed LC3 Interacting Regions (LIRs), were initially thought to function primarily to link LIR-containing autophagy cargo receptors to LC3 family members to help facilitate their capture during autophagy. However, the functional importance of LIRs in autophagy has broadened to include more general functions in autophagy as well. While a general consensus for LIR sequences has been described as [FWY]0-X1-X2-[LVI]3, recent work has suggested that additional sequences outside of the canonical LIR sequence can bind LC3 family members and play important roles in autophagy. In this manuscript by Kosmatka et al, the authors perform a high-throughput screen using bacterial surface display coupled with fluorescence-associated cell sorting to identify which human sequences can bind to LC3B. They identify a variety of peptides capable of binding LC3B, including peptides from proteins that have not previously been described as LC3B-binding proteins. The results from the bacterial surface display were then used to guide sequence analysis, mutational analysis, and structural studies to further characterize the range of LIR sequences that are capable of binding LC3B. Taken together, this work adds to the growing knowledge of how LIR sequences interact with LC3 family members and demonstrates which amino acids both inside and outside of the LIR sequence aid in binding. This work also identifies new potential LC3 binding proteins, which may play unknown roles in autophagy regulation. Lastly, this work reinforces the importance of alternative LIR sequences such as the [WFY]0-X1-X2-[WFY]3 sequence, which the authors have dubbed the LIR+ sequence.

Strengths:

The manuscript uses a robust approach to identify and characterize different peptide sequences that can interact with LC3B. They validate a large number of sequences using biolayer interferometry (BLI) and attempt to correlate different amino acids with their binding affinity for LC3B. The large number of LC3B binding sequences and their dissociation constants adds significant new information to the field that will help others understand what sequences can bind to LC3B. The authors are also very careful to accurately report on their data and not overly interpret their findings.

Weaknesses:

After the authors identify proteins from their bacterial display assay, the remainder of the manuscript is focused on characterizing the different types of sequences that are identified in addition to validating the LC3B-LIR interactions using biochemical approaches, including BLI and X-ray crystallography. However, it's not entirely clear if the screen identified novel LC3B binders that interact with LC3B in cells. While I acknowledge that the focus of the manuscript is on the characterization of LIR sequences that can bind LC3B, it seems like a missed opportunity not to validate a few of the novel LC3B binders in vivo. This may result in the demonstration of novel binders of LC3B in cells and may further demonstrate the strength of this approach for identifying LC3 family member binding partners. Therefore, it would be helpful to look at a few proteins identified in the HC set that have not previously been identified as LC3B binders in cells to determine if they CO-IP with LC3B or interact with LC3B using a different approach.

eLife Assessment

The work convincingly demonstrates the role of the mycobacterial secreted effector protein MmpE, which translocates to the host nucleus and exhibits phosphatase activity. The study is particularly valuable in showing that both the nuclear localization signal sequences and residues critical for phosphatase function are essential for host gene regulation, lysosomal biogenesis, and intracellular survival. Future studies will be needed to explore additional host pathways modulated by MmpE, particularly in the context of infection with a fully virulent Mycobacterium tuberculosis strain.

Reviewer #1 (Public review):

Summary:

The study provides insightful characterization of the mycobacterial secreted effector protein MmpE which translocates to the host nucleus and exhibits phosphatase activity. The study characterizes the nuclear localization signal sequences and residues critical for the phosphatase activity, both of which are required for intracellular survival

Strengths:

(1) The study addresses the role of nucleomodulins, an understudied aspect in mycobacterial infections.

(2) The authors employ a combination of biochemical and computational analyses along with in vitro and in vivo validations to characterize the role of MmpE.

Weaknesses:

(1) While the study establishes that the phosphatase activity of MmpE operates independently of its NLS, there is a clear gap in understanding how this phosphatase activity supports mycobacterial infection. The investigation lacks experimental data on specific substrates of MmpE or pathways influenced by this virulence factor.

(2) The study does not explore whether the phosphatase activity of MmpE is dependent on the NLS within macrophages, which would provide critical insights into its biological relevance in host cells. Conducting experiments with double knockout/mutant strains and comparing their intracellular survival with single mutants could elucidate these dependencies and further validate the significance of MmpE's dual functions.

(3) The study does not provide direct experimental validation of the MmpE deletion on lysosomal trafficking of the bacteria.

(4) The role of MmpE as a mycobacterial effector would be more relevant using virulent mycobacterial strains such as H37Rv.

Comments on revisions:

I appreciate the work the authors have done to address reviewers comments. The revised manuscript looks significantly improved. My major concern in the revised version is the microscopy data where the BCG staining using the DiD fluorescent stain does not bring out the rod-shaped bacilli structure. I suggest the authors either use a GFP reporter or some other fluorescent stain to address this issue.

Reviewer #2 (Public review):

Summary:

In this paper, the authors have characterized Rv2577 as a Fe3+/Zn2+ -dependent metallophosphatase and a nucleomodulin protein. The authors have also identified His348 and Asn359 as critical residues for Fe3+ coordination. The authors show that the proteins encode for two nuclease localization signals. Using C-terminal Flag expression constructs, the authors have shown that MmpE protein is secretory. The authors have prepared genetic deletion strains and show that MmpE is essential for intracellular survival of M. bovis BCG in THP-1 macrophages, RAW264.7 macrophages and mice model of infection. The authors have also performed RNA-seq analysis to compare the transcriptional profiles of macrophages infected with wild type and mmpE mutant strain. The relative levels of ~ 175 transcripts were altered in mmpE mutant infected macrophages and majority of these were associated with various immune and inflammatory signalling pathways. Using these deletion strains, the authors proposed that MmpE inhibits inflammatory gene expression by binding to the promoter region of vitamin D receptor. The authors also showed that MmpE arrests phagosome maturation by regulating the expression of several lysosome associated genes such as TFEB, LAMP1, LAMP2 etc. These findings reveal a sophisticated mechanism by which a bacterial effector protein manipulates gene transcription and promotes intracellular survival.

Strength:

The authors have used a combination of cell biology, microbiology and transcriptomics to elucidate the mechanisms by which Rv2577 contributes to intracellular survival.

Weakness:

The authors should thoroughly check the mice data and show individual replicate values in bar graphs.

Comments on revisions:

Thanks to the authors for addressing the concerns raised during the review of the original manuscript. The data is now presented with clarity, and discrepancies in mouse experiments have also been addressed with additional experiments.

Reviewer #3 (Public review):

Summary:

In this manuscript titled "Mycobacterial Metallophosphatase MmpE Acts as a Nucleomodulin to Regulate Host Gene Expression and Promote Intracellular Survival", Chen et al describe biochemical characterisation, localisation and potential functions of the gene using a genetic approach in M. bovis BCG and perform macrophage and mice infections to understand the roles of this potentially secreted protein in the host cell nucleus. The findings demonstrate the role of a secreted phosphatase of M. bovis BCG in shaping the transcriptional profile of infected macrophages, potentially through nuclear localisation and direct binding to transcriptional start sites, thereby regulating the inflammatory response to infection.

Strengths:

The authors demonstrate using a transient transfection method that MmpE when expressed as a GFP-tagged protein in HEK293T cells, exhibits nuclear localisation. The authors identify two NLS motifs that together are required for nuclear localisation of the protein. A deletion of the gene in M. bovis BCG results in poorer survival compared to the wild type parent strain, which is also killed by macrophages. Relative to the WT strain infected macrophages, macrophages infected with the mmpE strain exhibited differential gene expression. Overexpression of the gene in HEK293T led to occupancy of the transcription start site of several genes, including the Vitamin D Receptor. Expression of VDR in THP1 macrophages was lower in case of mmpE infection compared to WT infection. This data supports the utility of the overexpression system in identifying potential target loci of MmpE using the HEK293T transfection model. The authors also demonstrate that the protein is a phosphatase and the phosphatase activity of the protein is partially required for bacterial survival but not for regulation of the VDR gene expression.

Weaknesses:

There are significant differences in lysosomal retention between M. tuberculosis and M. bovis BCG. This study uses BCG and MMPE overexpression to draw conclusions about the impact of the MMPE gene on host gene expression and the bacteria's lysosomal localisation. While the authors have convincingly supported their claims with this model system, the relevance of this mechanism in M. tuberculosis infection remains unaddressed.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary:

Review of the manuscript titled " Mycobacterial Metallophosphatase MmpE acts as a nucleomodulin to regulate host gene expression and promotes intracellular survival".

The study provides an insightful characterization of the mycobacterial secreted effector protein MmpE, which translocates to the host nucleus and exhibits phosphatase activity. The study characterizes the nuclear localization signal sequences and residues critical for the phosphatase activity, both of which are required for intracellular survival.

Strengths:

(1) The study addresses the role of nucleomodulins, an understudied aspect in mycobacterial infections.

(2) The authors employ a combination of biochemical and computational analyses along with in vitro and in vivo validations to characterize the role of MmpE.

Weaknesses:

(1) While the study establishes that the phosphatase activity of MmpE operates independently of its NLS, there is a clear gap in understanding how this phosphatase activity supports mycobacterial infection. The investigation lacks experimental data on specific substrates of MmpE or pathways influenced by this virulence factor.

We thank the reviewer for this insightful comment and agree that identification of the substrates of MmpE is important to fully understand its role in mycobacterial infection. MmpE is a putative purple acid phosphatase (PAP) and a member of the metallophosphoesterase (MPE) superfamily. Enzymes in this family are known for their catalytic promiscuity and broad substrate specificity, acting on phosphomonoesters, phosphodiesters, and phosphotriesters (Matange et al., Biochem J, 2015). In bacteria, several characterized MPEs have been shown to hydrolyze substrates such as cyclic nucleotides (e.g., cAMP) (Keppetipola et al., J Biol Chem, 2008; Shenoy et al., J Mol Biol, 2007), nucleotide derivatives (e.g., AMP, UDP-glucose) (Innokentev et al., mBio, 2025), and pyrophosphate-containing compounds (e.g., Ap4A, UDP-DAGn) (Matange et al., Biochem J., 2015). Although the binding motif of MmpE has been identified, determining its physiological substrates remains challenging due to the low abundance and instability of potential metabolites, as well as the limited sensitivity and coverage of current metabolomic technologies in mycobacteria.

(2) The study does not explore whether the phosphatase activity of MmpE is dependent on the NLS within macrophages, which would provide critical insights into its biological relevance in host cells. Conducting experiments with double knockout/mutant strains and comparing their intracellular survival with single mutants could elucidate these dependencies and further validate the significance of MmpE's dual functions.

We thank the reviewer for the comment. Deletion of the NLS motifs did not impair MmpE’s phosphatase activity in vitro (Figure 2F), indicating that MmpE's enzymatic function operates independently of its nuclear localization. Indeed, we confirmed that Fe³⁺-binding ability via the residues H348 and N359 is required for enzymatic activity of MmpE. We have expanded on this point in the Discussion section “MmpE is a bifunctional virulence factor in Mtb”.

(3) The study does not provide direct experimental validation of the MmpE deletion on lysosomal trafficking of the bacteria.

We thank the reviewer for the comment. To validate the role of MmpE in lysosome maturation during infection, we conducted fluorescence colocalization assays in THP-1 macrophages infected with BCG strains, including WT, ∆MmpE, Comp-MmpE, Comp-MmpE^ΔNLS1, Comp-MmpE^ΔNLS2, Comp-MmpE^ΔNLS1-2. These strains were stained with the lipophilic membrane dye DiD, while macrophages were treated with the acidotropic probe LysoTracker^TM Green (Martins et al., Autophagy, 2019). The result indicated that ΔMmpE and MmpE^NLS1-2 mutants exhibited significantly higher co-localization with LysoTracker compared to WT and Comp-MmpE strains (New Figure 5G), suggesting that MmpE deletion leads to enhanced lysosomal maturation during infection.

(4) The role of MmpE as a mycobacterial effector would be more relevant using virulent mycobacterial strains such as H37Rv.

We thank the reviewer for the comment. Previously, the role of Rv2577/MmpE as a virulence factor has been demonstrated in M. tuberculosis CDC 1551, where its deletion significantly reduced bacterial replication in mouse lungs at 30 days post-infection (Forrellad et al., Front Microbiol, 2020). However, that study did not explore the underlying mechanism of MmpE function. In our study, we found that MmpE enhances M. bovis BCG survival in macrophages (THP-1 and RAW264.7 both) and in mice (Figure 3, Figure 7A), consistent with its proposed role in virulence. To investigate the molecular mechanism by which MmpE promotes intracellular survival, we used M. bovis BCG as a biosafe surrogate and this model is widely accepted for studying mycobacterial pathogenesis (Wang et al., Nat Immunol, 2015; Wang et al., Nat Commun, 2017; Péan et al., Nat Commun, 2017).

Reviewer #2 (Public review):

Summary:

In this paper, the authors have characterized Rv2577 as a Fe3+/Zn2+ -dependent metallophosphatase and a nucleomodulin protein. The authors have also identified His348 and Asn359 as critical residues for Fe3+ coordination. The authors show that the proteins encode for two nuclease localization signals. Using C-terminal Flag expression constructs, the authors have shown that the MmpE protein is secretory. The authors have prepared genetic deletion strains and show that MmpE is essential for intracellular survival of M. bovis BCG in THP-1 macrophages, RAW264.7 macrophages, and a mouse model of infection. The authors have also performed RNA-seq analysis to compare the transcriptional profiles of macrophages infected with wild-type and MmpE mutant strains. The relative levels of ~ 175 transcripts were altered in MmpE mutant-infected macrophages and the majority of these were associated with various immune and inflammatory signalling pathways. Using these deletion strains, the authors proposed that MmpE inhibits inflammatory gene expression by binding to the promoter region of a vitamin D receptor. The authors also showed that MmpE arrests phagosome maturation by regulating the expression of several lysosome-associated genes such as TFEB, LAMP1, LAMP2, etc. These findings reveal a sophisticated mechanism by which a bacterial effector protein manipulates gene transcription and promotes intracellular survival.

Strength:

The authors have used a combination of cell biology, microbiology, and transcriptomics to elucidate the mechanisms by which Rv2577 contributes to intracellular survival.

Weakness:

The authors should thoroughly check the mice data and show individual replicate values in bar graphs.

We kindly appreciate the reviewer for the advice. We have now updated the relevant mice data in the revised manuscript.

Reviewer #3 (Public review):

Summary:

In this manuscript titled "Mycobacterial Metallophosphatase MmpE Acts as a Nucleomodulin to Regulate Host Gene Expression and Promote Intracellular Survival", Chen et al describe biochemical characterisation, localisation and potential functions of the gene using a genetic approach in M. bovis BCG and perform macrophage and mice infections to understand the roles of this potentially secreted protein in the host cell nucleus. The findings demonstrate the role of a secreted phosphatase of M. bovis BCG in shaping the transcriptional profile of infected macrophages, potentially through nuclear localisation and direct binding to transcriptional start sites, thereby regulating the inflammatory response to infection.

Strengths:

The authors demonstrate using a transient transfection method that MmpE when expressed as a GFP-tagged protein in HEK293T cells, exhibits nuclear localisation. The authors identify two NLS motifs that together are required for nuclear localisation of the protein. A deletion of the gene in M. bovis BCG results in poorer survival compared to the wild-type parent strain, which is also killed by macrophages. Relative to the WT strain-infected macrophages, macrophages infected with the ∆mmpE strain exhibited differential gene expression. Overexpression of the gene in HEK293T led to occupancy of the transcription start site of several genes, including the Vitamin D Receptor. Expression of VDR in THP1 macrophages was lower in the case of ∆mmpE infection compared to WT infection. This data supports the utility of the overexpression system in identifying potential target loci of MmpE using the HEK293T transfection model. The authors also demonstrate that the protein is a phosphatase, and the phosphatase activity of the protein is partially required for bacterial survival but not for the regulation of the VDR gene expression.

Weaknesses:

(1) While the motifs can most certainly behave as NLSs, the overexpression of a mycobacterial protein in HEK293T cells can also result in artefacts of nuclear localisation. This is not unprecedented. Therefore, to prove that the protein is indeed secreted from BCG, and is able to elicit transcriptional changes during infection, I recommend that the authors (i) establish that the protein is indeed secreted into the host cell nucleus, and (ii) the NLS mutation prevents its localisation to the nucleus without disrupting its secretion.

We kindly appreciate the reviewer for this insightful comment. To confirm the translocation of MmpE into the host nucleus during BCG infection, we first detected the secretion of MmpE by M. bovis BCG, using Ag85B as a positive control and GlpX as a negative control (Zhang et al., Nat commun, 2022). Our results showed that MmpE- Flag was present in the culture supernatant, indicating that MmpE is secreted by BCG indeed (new Figure S1C).

Next, we performed immunoblot analysis of the nuclear fractions from infected THP-1 macrophages expressing FLAG-tagged wild-type MmpE and NLS mutants. The results revealed that only wild-type MmpE was detected in the nucleus, while MmpE^ΔNLS1, MmpE^ΔNLS2 and MmpE^ΔNLS1-2 were not detectable in the nucleus (New Figure S1D). Taken together, these findings demonstrated that MmpE is a secreted protein and that its nuclear translocation during infection requires both NLS motifs.

Demonstration that the protein is secreted: Supplementary Figure 3 - Immunoblotting should be performed for a cytosolic protein, also to rule out detection of proteins from lysis of dead cells. Also, for detecting proteins in the secreted fraction, it would be better to use Sauton's media without detergent, and grow the cultures without agitation or with gentle agitation. The method used by the authors is not a recommended protocol for obtaining the secreted fraction of mycobacteria.

We kindly appreciate the reviewer for the advice. To avoid the effects of bacterial lysis, we cultured the BCG strains expressing MmpE-Flag in Middlebrook 7H9 broth with 0.5% glycerol, 0.02% Tyloxapol, and 50 µg/mL kanamycin at 37 °C with gentle agitation (80 rpm) until an OD₆₀₀ of approximately 0.6 (Zhang et al., Nat Commun, 2022). Subsequently, we assessed the secretion of MmpE-Flag in the culture supernatant, using Ag85B as a positive control and GlpX as a negative control (New Figure S1C). The results showed that GlpX was not detected in the supernatant, while MmpE and Ag85B were detected, indicating that MmpE is indeed a secreted protein in BCG.

Demonstration that the protein localises to the host cell nucleus upon infection: Perform an infection followed by immunofluorescence to demonstrate that the endogenous protein of BCG can translocate to the host cell nucleus. This should be done for an NLS1-2 mutant expressing cell also.

We thank the reviewer for the suggestion. We agree that this experiment would be helpful to further verify the ability of MmpE for nuclear import. However, MmpE specific antibody is not available for us for immunofluorescence experiment. Alternatively, we performed nuclear-cytoplasmic fractionation for the THP-1 cells infected with the M. bovis BCG strains expressing FLAG-tagged wild-type MmpE, as well as NLS deletion mutants (MmpE^ΔNLS1, MmpE^ΔNLS2, and MmpE^ΔNLS1-2). The WT MmpE is detectable in both cytoplasmic and nuclear compartments, while MmpE^ΔNLS1, MmpE^ΔNLS2 or MmpE^ΔNLS1-2 were almost undetectable in nuclear fractions (New Figure S1D), suggesting that both NLS motifs are necessary for nuclear import.

(2) In the RNA-seq analysis, the directionality of change of each of the reported pathways is not apparent in the way the data have been presented. For example, are genes in the cytokine-cytokine receptor interaction or TNF signalling pathway expressed more, or less in the ∆mmpE strain?

We thank the reviewer for the comment. The KEGG pathway enrichment diagrams in our RNA-seq analysis primarily reflect the statistical significance of pathway enrichment based on differentially expressed genes, but do not indicate the directionality of genes expression changes. To address this concern, we conducted qRT-PCR on genes associated with the cytokine-cytokine receptor interaction pathway, specifically IL23A, CSF2, and IL12B. The results showed that, compared to the WT strain, infection with the ΔMmpE strain resulted in significantly increased expression levels of these genes in THP-1 cells (Figure 4F, Figure S4B), consistent with the RNA-seq data. Furthermore, we have submitted the complete RNA-seq dataset to the NCBI GEO repository [GSE312039], which includes normalized expression values and differential expression results for all detected genes.

(3) Several of these pathways are affected as a result of infection, while others are not induced by BCG infection. For example, BCG infection does not, on its own, produce changes in IL1β levels. As the author s did not compare the uninfected macrophages as a control, it is difficult to interpret whether ∆mmpE induced higher expression than the WT strain, or simply did not induce a gene while the WT strain suppressed expression of a gene. This is particularly important because the strain is attenuated. Does the attenuation have anything to do with the ability of the protein to induce lysosomal pathway genes? Does induction of this pathway lead to attenuation of the strain? Similarly, for pathways that seem to be downregulated in the ∆mmpE strain compared to the WT strain, these might have been induced upon infection with the WT strain but not sufficiently by the ∆mmpE strain due to its attenuation/ lower bacterial burden.

We thank the reviewer for the comment. Previous studies have shown that wild-type BCG induces relatively low levels of IL-1β, while retaining partial capacity to activate the inflammasome (Qu et al., Sci Adv, 2020). Our data (Figures 3G) show that infection with the ΔMmpE strain results in enhanced IL-1β expression, consistent with findings by Master et al. (Cell Host Microbe, 2008), in which deletion of zmp1 in BCG or M. tuberculosis led to increased IL-1β levels due to reduced inhibition of inflammasome activation.

In the revised manuscript, we have provided additional qRT-PCR data using uninfected macrophages as a baseline control. These results demonstrate that the WT strain suppresses lysosome-associated gene expression, whereas the ΔMmpE strain upregulates these genes, indicating that MmpE inhibits lysosome-related genes expression (Figure 4G). Furthermore, bacterial burden analysis revealed that ∆mmpE exhibited ~3-fold lower intracellular survival than the WT strain in THP-1 cells. However, when lysosomal maturation was inhibited, the difference in bacterial load between the two strains was reduced to ~1-fold (New Figures S6B and C). These findings indicate that MmpE promotes intracellular survival primarily by inhibiting lysosomal maturation, which is consistent with a previous study (Chandra et al., Sci Rep, 2015).

(4) CHIP-seq should be performed in THP1 macrophages, and not in HEK293T. Overexpression of a nuclear-localised protein in a non-relevant line is likely to lead to several transcriptional changes that do not inform us of the role of the gene as a transcriptional regulator during infection.

We thank the reviewer for the comment. We performed ChIP-seq in HEK293T cells based on their high transfection efficiency, robust nuclear protein expression, and well-annotated genome (Lampe et al., Nat Biotechnol, 2024; Marasco et al., Cell, 2022). These characteristics make HEK293T an ideal system for the initial identification of genome-wide chromatin binding profiles by MmpE.

Further, we performed comprehensive validation of the ChIP-seq findings in THP-1 macrophages. First, CUT&Tag and RNA-seq analyses in THP-1 cells revealed that MmpE modulates genes involved in the PI3K–AKT signaling and lysosomal maturation pathways (Figure 4C; Figure S5A-B). Correspondingly, we found that infection with the ΔMmpE strain led to reduced phosphorylation of AKT (S473), mTOR (S2448), and p70S6K (T389) (New Figure 5E-F), and upregulation of lysosomal genes such as TFEB, LAMP1, and LAMP2 (Figure 4G), compared to infection with the WT strain, and lysosomal maturation in cells infected with the ΔMmpE strain more obviously (New Figure 5G). Additionally, CUT&Tag profiling identified MmpE binding at the promoter region of the VDR gene, which was further validated by EMSA and ChIP-qPCR. Also, qRT-PCR demonstrated that MmpE suppresses VDR transcription, supporting its role as a transcriptional regulator (Figure 6). Collectively, these data confirm the biological relevance and functional significance of the ChIP-seq findings obtained in HEK293T cells.

(5) I would not expect to see such large inflammatory reactions persisting 56 days post-infection with M. bovis BCG. Is this something peculiar for an intratracheal infection with 1x107 bacilli? For images of animal tissue, the authors should provide images of the entire lung lobe with the zoomed-in image indicated as an inset.

We thank the reviewer for the comment. The lung inflammation peaked at days 21–28 and had clearly subsided by day 56 across all groups (New Figure 7B), consistent with the expected resolution of immune responses to an attenuated strain like M. bovis BCG. This temporal pattern is in line with previous studies using intravenous or intratracheal BCG vaccination in mice and macaques, which also demonstrated robust early immune activation followed by resolution over time (Smith et al., Nat Microbiol, 2025; Darrah et al., Nature, 2020).

In this study, the infectious dose (1×10⁷ CFU intratracheal) was selected based on previous studies in which intratracheal delivery of 1×10⁷ CFU produced consistent and measurable lung immune responses and pathology without causing overt illness or mortality (Xu et al., Sci Rep, 2017; Niroula et al., Sci Rep, 2025). We have provided whole-lung lobe images with zoomed-in insets in the source dataset.

(6) For the qRT-PCR based validation, infections should be performed with the MmpE-complemented strain in the same experiments as those for the WT and ∆mmpE strain so that they can be on the same graph, in the main manuscript file. Supplementary Figure 4 has three complementary strains. Again, the absence of the uninfected, WT, and ∆mmpE infected condition makes interpretation of these data very difficult.

We thank the reviewer for the comment. As suggested, we have conducted the qRT-PCR experiment including the uninfected, WT, ∆mmpE, Comp-MmpE, and the three complementary strains infecting THP-1 cells (Figure 4F and G; New Figure S4B–D).

(7) The abstract mentions that MmpE represses the PI3K-Akt-mTOR pathway, which arrests phagosome maturation. There is not enough data in this manuscript in support of this claim. Supplementary Figure 5 does provide qRT-PCR validation of genes of this pathway, but the data do not indicate that higher expression of these pathways, whether by VDR repression or otherwise, is driving the growth restriction of the ∆mmpE strain.

We thank the reviewer for the comment. In the updated manuscript, we have provided more evidence. First, the RNA-seq analysis indicated that MmpE affects the PI3K-AKT signaling pathway (Figure 4C). Second, CUT&Tag analysis suggested that MmpE binds to the promoter regions of key pathway components, including PRKCB, PLCG2, and PIK3CB (Figure S5A). Third, confocal microscopy showed that ΔMmpE strain promotes significantly increased lysosomal maturation compared to the WT, a process downstream of the PI3K-AKT-mTOR axis (New Figure 5G).

Further, we measured protein phosphorylation for validating activation of the pathway (Zhang et al., Stem Cell Reports, 2017). Our results showed that cells infected with WT strains exhibited significantly higher phosphorylation of Akt, mTOR, and p70S6K compared to those infected with ΔMmpE strains (New Figures 5E and F). Moreover, the dual PI3K/mTOR inhibitor BEZ235 abolished the survival advantage of WT strains over ΔMmpE mutants in THP-1 macrophages (New Figure S6B and C). Collectively, these results support that MmpE activates the PI3K–Akt–mTOR signaling pathway to enhance bacterial survival within the host.

(8) The relevance of the NLS and the phosphatase activity is not completely clear in the CFU assays and in the gene expression data. Firstly, there needs to be immunoblot data provided for the expression and secretion of the NLS-deficient and phosphatase mutants. Secondly, CFU data in Figure 3A, C, and E must consistently include both the WT and ∆mmpE strain.

We thank the reviewer for the comment. We have now added immunoblot analysis for expression and secretion of MmpE mutants. The result show that NLS-deficient and phosphatase mutants can detected in supernatant (New Figure S1C). Additionally, we have revised Figures 3A, 3C, and 3E to consistently include both the WT and ΔMmpE strains in the CFU assays (Figures 3A, 3C, and 3E).

Recommendations for the authors:

Reviewer #2 (Recommendations for the authors):

The authors should attempt to address the following comments:

(1) Please perform densitometric analysis for the western blot shown in Figure 1E.

We sincerely thank the reviewer for the suggestion. In the updated manuscript, we have performed densitometric analysis of the western blot shown in New Figure 1F and G.

(2) Is it possible to measure the protein levels for MmpE in lysates prepared from infected macrophages.

We thank the reviewer for the comment. In the revised manuscript, we performed immunoblot analysis to measure MmpE levels in lysates from infected macrophages. The results demonstrated that wild-type MmpE was present in both the cytoplasmic and nuclear fractions during infection in THP-1 cells (New Figure S1D).

(3) The authors should perform circular dichroism studies to compare the secondary structure of wild type and mutant proteins (in particular MmpEHis348 and MmpEAsn359.

We thank the reviewer for this valuable suggestion. We agree that circular dichroism spectroscopy could provide useful information in comparison of the differences on the secondary structures. However, due to the technical limitations, we instead compared the structures of wild-type MmpE and the His348 and Asn359 mutant proteins predicted by AlphaFold. These structural models showed almost no differences in secondary structures between the wild-type and mutants (Figure S1B).

(4) The authors should perform more experiments to determine the binding motif for MmpE in the promoter region of VDR.

We thank the reviewer for this suggestion. In the current study, we have identified the MmpE-binding motif within the promoter region of VDR using CUT&Tag sequencing. This prediction was further validated by ChIP-qPCR and EMSA (Figure 6). These complementary approaches collectively support the identification of a specific MmpE-binding motif and demonstrate its functional relevance. Such approach was acceptable in many publications (Wen et al., Commun Biol, 2020; Li et al., Nat Commun, 2022).

(5) Were the transcript levels of VDR also measured in the lung tissues of infected animals?

We thank the reviewer for this suggestion. In the revised manuscript, we have performed qRT-PCR to assess VDR transcript levels in the lung tissues of infected mice (New Figure S8B).

(6) How does MmpE regulate the expression of lysosome-associated genes?

We thank the reviewer for this question. Our experiments suggested that MmpE suppresses lysosomal maturation probably by activating the host PI3K–AKT–mTOR signaling pathway (New Figure 5E–I). This pathway is well established as a negative regulator of lysosome biogenesis and function (Yang et al., Signal Transduct Target Ther, 2020; Cui et al., Nature, 2023; Cui et al., Nature, 2025). During infection, THP-1 cells infected with the WT showed increased phosphorylation of Akt, mTOR, and p70S6K compared to those infected with ΔMmpE (New Figure S5C, New Figure 5E and F), and concurrently downregulated key lysosomal maturation markers, including TFEB, LAMP1, LAMP2, and multiple V-ATPase subunits (Figure 4G). Given that PI3K–AKT–mTOR signaling suppresses TFEB activity and lysosomal gene transcription (Palmieri et al., Nat Commun, 2017), we propose that MmpE modulates lysosome-associated gene expression and lysosomal function probably by PI3K–AKT–mTOR signaling pathway.

(7) Mice experiment:

(a) The methods section states that mice were infected intranasally, but the legend for Figure 6 states intratracheally. Kindly check?

(b) Supplementary Figure 7 - this is not clear. The legend says bacterial loads in spleens (CFU/g) instead of DNA expression, as shown in the figure.

(c) The data in Figure 6 and Figure S7 seem to be derived from the same experiment, but the number of animals is different. In Figure 6, it is n = 6, and in Figure S7, it is n=3.

We thank the reviewer for the comments.

(a) The infection was performed intranasally, and the figure legend for New Figure 7 has now been corrected.

(b) We adopted quantitative PCR method to measure bacterial DNA levels in the spleens of infected mice. We have now revised the legend.

(c) We have conducted new experiments where each experiment now includes six mice. The results are showed in Figure 7B and C, as well as in the new Figure S8.

(8) The authors should show individual values for various replicates in bar graphs (for all figures).

We thank the reviewer for this helpful suggestion. We have now updated all relevant bar graphs to include individual data points for each biological replicate.

(9) The authors should validate the relative levels of a few DEGs shown in Figure 3F, Figure 3G, and Figure S4C, in the lung tissues of mice infected with wild-type, mutant, and complemented strains.

We thank the reviewer for this suggestion. In the revised manuscript, we have performed qRT-PCR to validate the expression levels of selected DEGs, including inflammation-related and lysosome-associated genes, in lung tissues from mice infected with wild-type, mutant, and complemented strains (New Figure S8C-H).

(10) Did the authors perform an animal experiment using a mutant strain complemented with the phosphatase-deficient MmpE (Comp-MmpE-H348AN359H)?

We appreciate the reviewer's comment. We agree that an additional animal experiment would be useful to assess the effects of the phosphatase. However, our study mainly focused on interpreting the function of the nuclear localization of MmpE during BCG infection. Additionally, we have assessed the role of the phosphatase of MmpE during infection with cell model (Figure 3E).

Minor comment:

The mutant strain should be verified by either Southern blot or whole genome sequencing.

We thank the reviewer for this comment. We verified deletion of mmpE gene by PCR method (Figure S3A-D) which was acceptable in many publications (Zhang et al., PLoS Pathog, 2020; Zhang et al., Nat Commun, 2022).

Reviewer #3 (Recommendations for the authors):

(1) Line 195: cytokine.

We thank the reviewer for the comments. We have now corrected it.

(2) Line 225: rewording required.

Corrected.

(3) Figure 4A. "No difference" instead of "No different".

Corrected.

(4) "KommpE" should be replaced with "∆mmpE strain" (∆=delta symbol).

Corrected.

(5) Supplementary Figure 7. The figure legend states CFU assays, but the y-axis and the graph seem to depict IS1081 quantification.

We thank the reviewer for the comment. The figure is based on IS1081 quantification using qRT-PCR, not CFU assays. We have now revised the legend for New Figure S8A.

References

Chandra P, Ghanwat S, Matta SK, Yadav SS, Mehta M, Siddiqui Z, Singh A, Kumar D (2015) Mycobacterium tuberculosis Inhibits RAB7 Recruitment to Selectively Modulate Autophagy Flux in Macrophages Sci Rep 5:16320.

Darrah PA, Zeppa JJ, Maiello P, Hackney JA, Wadsworth MH 2nd, Hughes TK, Pokkali S, Swanson PA 2nd, Grant NL, Rodgers MA, Kamath M, Causgrove CM, Laddy DJ, Bonavia A, Casimiro D, Lin PL, Klein E, White AG, Scanga CA, Shalek AK, Roederer M, Flynn JL, Seder RA (2020) Prevention of tuberculosis in macaques after intravenous BCG immunization Nature 577:95-102. 

Forrellad MA, Blanco FC, Marrero Diaz de Villegas R, Vázquez CL, Yaneff A, García EA, Gutierrez MG, Durán R, Villarino A, Bigi F (2020) Rv2577 of Mycobacterium tuberculosis Is a virulence factor with dual phosphatase and phosphodiesterase functions Front Microbiol 11:570794.

Innokentev A, Sanchez AM, Monetti M, Schwer B, Shuman S (2025) Efn1 and Efn2 are extracellular 5'-nucleotidases induced during the fission yeast response to phosphate starvation mBio 16: e0299224.

Keppetipola N, Shuman S (2008) A phosphate-binding histidine of binuclear metallophosphodiesterase enzymes is a determinant of 2',3'-cyclic nucleotide phosphodiesterase activity J Biol Chem 283:30942-9.

Lampe GD, King RT, Halpin-Healy TS, Klompe SE, Hogan MI, Vo PLH, Tang S, Chavez A, Sternberg SH (2024) Targeted DNA integration in human cells without double-strand breaks using CRISPR-associated transposases Nat Biotechnol 42:87-98.

Li Z, Sheerin DJ, von Roepenack-Lahaye E, Stahl M, Hiltbrunner A (2022) The phytochrome interacting proteins ERF55 and ERF58 repress light-induced seed germination in Arabidopsis thaliana Nat Commun 13:1656.

Marasco LE, Dujardin G, Sousa-Luís R, Liu YH, Stigliano JN, Nomakuchi T, Proudfoot NJ, Krainer AR, Kornblihtt AR (2022) Counteracting chromatin effects of a splicing-correcting antisense oligonucleotide improves its therapeutic efficacy in spinal muscular atrophy Cell 185:2057-2070.e15.

Martins WK, Santos NF, Rocha CS, Bacellar IOL, Tsubone TM, Viotto AC, Matsukuma AY, Abrantes ABP, Siani P, Dias LG, Baptista MS (2019) Parallel damage in mitochondria and lysosomes is an efficient way to photoinduce cell death Autophagy 15:259-279.

Master SS, Rampini SK, Davis AS, Keller C, Ehlers S, Springer B, Timmins GS, Sander P, Deretic V (2008) Mycobacterium tuberculosis prevents inflammasome activation Cell Host Microbe 3:224-32.

Matange N, Podobnik M, Visweswariah SS (2015) Metallophosphoesterases: structural fidelity with functional promiscuity Biochem J 467:201-16.

Niroula N, Ghodasara P, Marreros N, Fuller B, Sanderson H, Zriba S, Walker S, Shury TK, Chen JM (2025) Orally administered live BCG and heat-inactivated Mycobacterium bovis protect bison against experimental bovine tuberculosis Sci Rep 15:3764.

Palmieri M, Pal R, Nelvagal HR, Lotfi P, Stinnett GR, Seymour ML, Chaudhury A, Bajaj L, Bondar VV, Bremner L, Saleem U, Tse DY, Sanagasetti D, Wu SM, Neilson JR, Pereira FA, Pautler RG, Rodney GG, Cooper JD, Sardiello M (2017) mTORC1-independent TFEB activation via Akt inhibition promotes cellular clearance in neurodegenerative storage diseases Nat Commun 8:14338.

Péan CB, Schiebler M, Tan SW, Sharrock JA, Kierdorf K, Brown KP, Maserumule MC, Menezes S, Pilátová M, Bronda K, Guermonprez P, Stramer BM, Andres Floto R, Dionne MS (2017) Regulation of phagocyte triglyceride by a STAT-ATG2 pathway controls mycobacterial infection Nat Commun 8:14642.

Qu Z, Zhou J, Zhou Y, Xie Y, Jiang Y, Wu J, Luo Z, Liu G, Yin L, Zhang XL (2020) Mycobacterial EST12 activates a RACK1-NLRP3-gasdermin D pyroptosis-IL-1β immune pathway Sci Adv 6: eaba4733.

Shenoy AR, Capuder M, Draskovic P, Lamba D, Visweswariah SS, Podobnik M (2007) Structural and biochemical analysis of the Rv0805 cyclic nucleotide phosphodiesterase from Mycobacterium tuberculosis J Mol Biol 365:211-25.

Smith AA, Su H, Wallach J, Liu Y, Maiello P, Borish HJ, Winchell C, Simonson AW, Lin PL, Rodgers M, Fillmore D, Sakal J, Lin K, Vinette V, Schnappinger D, Ehrt S, Flynn JL (2025) A BCG kill switch strain protects against Mycobacterium tuberculosis in mice and non-human primates with improved safety and immunogenicity Nat Microbiol 10:468-481.

Wang J, Ge P, Qiang L, Tian F, Zhao D, Chai Q, Zhu M, Zhou R, Meng G, Iwakura Y, Gao GF, Liu CH (2017) The mycobacterial phosphatase PtpA regulates the expression of host genes and promotes cell proliferation Nat Commun 8:244.

Wang J, Li BX, Ge PP, Li J, Wang Q, Gao GF, Qiu XB, Liu CH (2015) Mycobacterium tuberculosis suppresses innate immunity by coopting the host ubiquitin system Nat Immunol 16:237–245

Wen X, Wang J, Zhang D, Ding Y, Ji X, Tan Z, Wang Y (2020) Reverse Chromatin Immunoprecipitation (R-ChIP) enables investigation of the upstream regulators of plant genes Commun Biol 3:770.

Xu X, Lu X, Dong X, Luo Y, Wang Q, Liu X, Fu J, Zhang Y, Zhu B, Ma X (2017) Effects of hMASP-2 on the formation of BCG infection-induced granuloma in the lungs of BALB/c mice Sci Rep 7:2300.

Zhang L, Hendrickson RC, Meikle V, Lefkowitz EJ, Ioerger TR, Niederweis M. (2020) Comprehensive analysis of iron utilization by Mycobacterium tuberculosis PLoS Pathog 16: e1008337.

Zhang L, Kent JE, Whitaker M, Young DC, Herrmann D, Aleshin AE, Ko YH, Cingolani G, Saad JS, Moody DB, Marassi FM, Ehrt S, Niederweis M (2022) A periplasmic cinched protein is required for siderophore secretion and virulence of Mycobacterium tuberculosis Nat Commun 13:2255.

Zhang X, He X, Li Q, Kong X, Ou Z, Zhang L, Gong Z, Long D, Li J, Zhang M, Ji W, Zhang W, Xu L, Xuan A (2017) PI3K/AKT/mTOR Signaling Mediates Valproic Acid-Induced Neuronal Differentiation of Neural Stem Cells through Epigenetic Modifications Stem Cell Reports 8:1256-1269.

eLife Assessment

The authors provide a useful resource and approach to identify early-stage biomarkers of MASLD progression, notably when no other apparent symptoms have arisen. The strength of evidence to support new MASLD signatures is solid as the work combines metabolomic and transcriptomic measures in blood and liver biopsies.

[Editors' note: this paper was reviewed by Review Commons.]

Reviewer #1 (Public review):

Summary:

Metabolic dysfunction-associated steatotic liver disease (MASLD) ranges from simple steatosis, steatohepatitis, fibrosis/cirrhosis, and hepatocellular carcinoma. In the current study, the authors aimed to determine the early molecular signatures differentiating patients with MASLD associated fibrosis from those patients with early MASLD but no symptoms. The authors recruited 109 obese individuals before bariatric surgery. They separated the cohorts as no MASLD (without histological abnormalities) and MASLD. The liver samples were then subjected to transcriptomic and metabolomic analysis. The serum samples were subjected to metabolomic analysis. The authors identified dysregulated lipid metabolism, including glyceride lipids, in the liver samples of MASLD patients compared to the no MASLD ones. Circulating metabolomic changes in lipid profiles slightly correlated with MASLD, possibly due to the no MASLD samples derived from obese patients. Several genes involved in lipid droplet formation were also found elevated in MASLD patients. Besides, elevated levels of amino acids, which are possibly related to collagen synthesis, were observed in MASLD patients. Several antioxidant metabolites were increased in MASLD patients. Furthermore, dysregulated genes involved in mitochondrial function and autophagy were identified in MASLD patients, likely linking oxidative stress to MASLD progression. The authors then determined the representative gene signatures in the development of fibrosis by comparing this cohort with the other two published cohorts. Top enriched pathways in fibrotic patients included GTPas signaling and innate immune responses, suggesting the involvement of GTPas in MASLD progression to fibrosis. The authors then challenged human patient derived 3D spheroid system with a dual PPARa/d agonist and found that this treatment restored the expression levels of GTPase-related genes in MASLD 3D spheroids. In conclusion, the authors suggested the involvement of upregulated GTPase-related genes during fibrosis initiation.

Significance:

Overall, the current study might provide some new resources regarding transcriptomic and metabolomic data derived from obese patients with and without MASLD. The MASLD research community will be interested in the resource data.

Comments on revised version:

I have no further comments. Thank you.

Reviewer #3 (Public review):

Summary:

Metabolic dysfunction associated liver disease (MASLD) describes a spectrum of progressive liver pathologies linked to life style-associated metabolic alterations (such as increased body weight and elevated blood sugar levels), reaching from steatosis over steatohepatitis to fibrosis and finally end stage complications, such as liver failure and hepatocellular carcinoma. Treatment options for MASLD include diet adjustments, weight loss, and the receptor-β (THR-β) agonist resmetirom, but remain limited at this stage, motivating further studies to elucidate molecular disease mechanisms to identify novel therapeutic targets.

In their present study, the authors aim to identify early molecular changes in MASLD linked to obesity. To this end, they study a cohort of 109 obese individuals with no or early-stage MASLD combining measurements from two anatomic sides: 1. bulk RNA-sequencing and metabolomics of liver biopsies, and 2. metabolomics from patient blood. Their major finding is that GTPase-related genes are transcriptionally altered in livers of individuals with steatosis with fibrosis compared to steatosis without fibrosis.

Major comments:

(1) Confounders (such as (pre-)diabetes)

The patient table shows significant differences in non-MASLD vs. MASLD individuals, with the latter suffering more often from diabetes or hypertriglyceridemia. Rather than just stating corrections, subgroup analyses should be performed (accompanied with designated statistical power analyses) to infer the degree to which these conditions contribute to the observations. I.e., major findings stating MASLD-associated changes should hold true in the subgroup of MASLD patients without diabetes/of female sex and so forth (testing for each of the significant differences between groups).

Post-rebuttal update: The authors have performed the requested sub-group analysis and find the gene signatures hold for the non-diabetic sub-cohort, but not the diabetic subgroup. They denote a likely interaction between fibrosis and diabetes, that was not corrected for in the original analysis.

Post-post-rebuttal update: I thank the authors for having added Figure 5-figure supplement 2 to show this analysis.

(2) External validation

Additionally, to back up the major GTPase signature findings, it would be desirable to analyze an external dataset of (pre)diabetes patients (other biased groups) for alternations in these genes. It would be important to know if this signature also shows in non-MASLD diabetic patients vs. healthy patients or is a feature specific to MASLD. Also, could the matched metabolic data be used to validate metabolite alterations that would be expected under GTPase-associated protein dysregulation?

Post-rebuttal update: The authors confirm that with the present data, insulin resistance cannot be fully ruled out as a confounder to the GTP-ase related gene signature. They however plan future mouse model experiments to study whether the GTPase-fibrosis signature differs in diabetic vs. non-diabetic conditions.

(3) 3D liver spheroid MASH model, Fig. 6D/E

This 3D experiment is technically not an external validation of GTPase-related genes being involved in MASLD, since patient-derived cells may only retain changes that have happened in vivo. To demonstrate that the GTPase expression signature is specifically invoked by fibrosis the LX-2 set up is more convincing, however, the up-regulation of the GTPase-related genes upon fibrosis induction with TGF-beta, in concordance with the patient data, needs to be shown first (qPCR or RNA-seq). Additionally, the description of the 3D model is too uncritical. The maintenance of functional PHHs is a major challenge (PMID: 38750036, PMID: 21953633, PMID: 40240606, PMID: 31023926). It cannot be ruled out that their findings are largely attributable to either 1) the (other present) mesenchymal cells (i.e., mesenchyme-derived cells, such as for example hepatic stellate cells, not to be confused with mesenchymal stem cells, MSCs), or 2) related to potential changes in PHHs in culture, and these limitations need to be stated.

Post-rebuttal update: To address the concern of other cells than hepatocytes contributing to the observed effects in culture, the authors performed TGF-beta treatment in independent mono-cultures (Figure R4): LX-2 and hepatocytes, and the spheroid system. Surprisingly, important genes highlighted in Figure 6E for the spheroid system (RAB6A, ARL4A, RAB27B, DIRAS2) are all absent from this qPCR(?) validation experiment. The authors evaluate instead RAC1, RHOU, VAV1, DOCK2, RAB32. ­In spheroids, RHOU and RAB32 are down-regulated with TGF-B. In hepatocytes DOCK2 and RAC seemed up-regulated. They find no difference in these genes in LX-2 cells. Surprisingly, ACTA2 expression values are missing for LX-2 cells. Together, it is hard to judge which individual cell type recapitulates the changes observed in patients in this validation experiment, as the major genes called out in Figure 6E are not analyzed.

Post-post-rebuttal update: I thank the authors for having added Figure 6-figure supplement 5 to show qPCR results for this question.

Unfortunately, the 3D liver spheroid model used (as presente­d in PMID39605182) lacks important functional validation tests of maintained hepatocyte identity in culture (at the very least Albumin expression and secretion plus CYP3A4 assay). This functional data (acquired at the time point in culture when the RNA expression analysis in 6E was performed) is indispensable prior to stating that mature hepatocytes cause the observed effects.

Post-post-rebuttal update: I thank the authors for having added more references, I still think a quick functional validation of the system (at the time point in culture when the RNA expression analysis in 6E was performed) would be beneficial.

(4) Novelty / references

Similar studies that also combined liver and blood lipidomics/metabolomics in obese individuals with and without MASLD (e.g. PMID 39731853, 39653777) should be cited. Additionally, it would benefit the quality of the discussion to state how findings in this study add new insights over previous studies, if their findings/insights differ, and if so, why.

Post-rebuttal update: The authors have included the studies into their discussion.

Overall post-post-rebuttal update: I thank the authors for having added more data, important discussion points, and references, and have no further requests.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Thank you for the authors' responses to my concerns. I do not have any further comments.

We thank this reviewer for the positive and constructive evaluation of our manuscript.

Reviewer #2 (Public Review):

I have no further comment about this amended version, aside from suggesting to add (if known) the time at which biopsies were collected. Time-of-day is an important yet often overlooked parameter of gene expression variation, and along the same line, the imposed fasting to bariatric surgery patients is also a matter of variation of gene expression and of metabolite abundance. It is hoped that future investigations will more precisely characterize the role of the newly identified targets in MASLD.

We agree with this and are fully aware that metabolism in the liver is controlled by circadian rhythm and therefore the time-of-day is an important parameter when liver samples are collected. All liver samples were collected between 8am and 1pm, and this information has been added to the Methods section. We are already working on the characterization of the newly identified targets. Thank you for the positive and constructive evaluation of our manuscript.

Reviewer #3 (Public Review):

(1) Confounders (such as (pre-)diabetes)

The patient table shows significant differences in non-MASLD vs. MASLD individuals, with the latter suffering more often from diabetes or hypertriglyceridemia. Rather than just stating corrections, subgroup analyses should be performed (accompanied with designated statistical power analyses) to infer the degree to which these conditions contribute to the observations. I.e., major findings stating MASLD-associated changes should hold true in the subgroup of MASLD patients without diabetes/of female sex and so forth (testing for each of the significant differences between groups).

Post-rebuttal update: The authors have performed the requested sub-group analysis and find the gene signatures hold for the non-diabetic sub-cohort, but not the diabetic subgroup. They denote a likely interaction between fibrosis and diabetes, that was not corrected for in the original analysis.

(2) External validation

Additionally, to back up the major GTPase signature findings, it would be desirable to analyze an external dataset of (pre)diabetes patients (other biased groups) for alternations in these genes. It would be important to know if this signature also shows in non-MASLD diabetic patients vs. healthy patients or is a feature specific to MASLD. Also, could the matched metabolic data be used to validate metabolite alterations that would be expected under GTPase-associated protein dysregulation?

Post-rebuttal update: The authors confirm that with the present data, insulin resistance cannot be fully ruled out as a confounder to the GTPase related gene signature. They however plan future mouse model experiments to study whether the GTPase-fibrosis signature differs in diabetic vs. non-diabetic conditions.

(3) 3D liver spheroid MASH model, Fig. 6D/E

This 3D experiment is technically not an external validation of GTPase-related genes being involved in MASLD, since patient-derived cells may only retain changes that have happened in vivo. To demonstrate that the GTPase expression signature is specifically invoked by fibrosis the LX-2 set up is more convincing, however, the up-regulation of the GTPase-related genes upon fibrosis induction with TGF-beta, in concordance with the patient data, needs to be shown first (qPCR or RNA-seq). Additionally, the description of the 3D model is too uncritical. The maintenance of functional PHHs is a major challenge (PMID: 38750036, PMID: 21953633, PMID: 40240606, PMID: 31023926). It cannot be ruled out that their findings are largely attributable to either 1) the (other present) mesenchymal cells (i.e., mesenchyme-derived cells, such as for example hepatic stellate cells, not to be confused with mesenchymal stem cells, MSCs), or 2) related to potential changes in PHHs in culture, and these limitations need to be stated.

Post-rebuttal update: To address the concern of other cells than hepatocytes contributing to the observed effects in culture, the authors performed TGF-beta treatment in independent mono-cultures (Figure R4): LX-2 and hepatocytes, and the spheroid system. Surprisingly, important genes highlighted in Figure 6E for the spheroid system (RAB6A, ARL4A, RAB27B, DIRAS2) are all absent from this qPCR(?) validation experiment. The authors evaluate instead RAC1, RHOU, VAV1, DOCK2, RAB32. -In spheroids, RHOU and RAB32 are down-regulated with TGF-B. In hepatocytes DOCK2 and RAC seemed up-regulated. They find no difference in these genes in LX-2 cells. Surprisingly, ACTA2 expression values are missing for LX-2 cells. Together, it is hard to judge which individual cell type recapitulates the changes observed in patients in this validation experiment, as the major genes called out in Figure 6E are not analyzed.

All biological experiments show variations and especially when analyzing various cell types (lines), we are not completely surprised that not all results are completely aligned. In other words, some of the GTPases will be upregulated in hepatocytes, while other may be upregulated in hepatic stellate cells due to the complex signaling arrangement in each cell. To address this reviewer’s concerns, we have done qPCR for RAB6A, ARL4A, RAB27B, DIRAS2 in LX-2 cells and the results are shown in the revised now Figure 6– figure supplement 5. To align all three graphs displaying the same genes analyzed, we have now depicted the gene expression for the co-culture (hepatocytes, hepatic stellate cells, and Kupffer cells) and mono-culture (hepatocytes only) from RNAseq analysis.

Unfortunately, the 3D liver spheroid model used (as presente-d in PMID39605182) lacks important functional validation tests of maintained hepatocyte identity in culture (at the very least Albumin expression and secretion plus CYP3A4 assay). This functional data (acquired at the time point in culture when the RNA expression analysis in 6E was performed) is indispensable prior to stating that mature hepatocytes cause the observed effects.

We agree that the characterization of the liver spheroid model derived from human patient samples is important. The functional characterization has already been published in these papers:

(1) Bell, C. C. et al. Transcriptional, Functional, and Mechanistic Comparisons of Stem Cell–Derived Hepatocytes, HepaRG Cells, and Three-Dimensional Human Hepatocyte Spheroids as Predictive In Vitro Systems for Drug-Induced Liver Injury. Drug Metab. Dispos. 45, 419–429 (2017).

(2) Bell, C. C. et al. Characterization of primary human hepatocyte spheroids as a model system for drug-induced liver injury, liver function and disease. Sci. Rep. 6, 25187 (2016). 3.Vorrink, S. U. et al. Endogenous and xenobiotic metabolic stability of primary human hepatocytes in long‐term 3D spheroid cultures revealed by a combination of targeted and untargeted metabolomics. FASEB J. 31, 2696–2708 (2017).

(4) Messner, S. et al. Transcriptomic, Proteomic, and Functional Long-Term Characterization of Multicellular Three-Dimensional Human Liver Microtissues. Appl. In Vitro Toxicol. 4, 1–12 (2018).

(5) Bell, C. C. et al. Comparison of Hepatic 2D Sandwich Cultures and 3D Spheroids for Long-term Toxicity Applications: A Multicenter Study. Toxicol. Sci. 162, 655–666 (2018). We have mentioned this now in the manuscript on page 18 to make this point clear.

(4) Novelty / references

Similar studies that also combined liver and blood lipidomics/metabolomics in obese individuals with and without MASLD (e.g. PMID 39731853, 39653777) should be cited. Additionally, it would benefit the quality of the discussion to state how findings in this study add new insights over previous studies, if their findings/insights differ, and if so, why.

Post-rebuttal update: The authors have included the studies into their discussion.

Recommendations for the authors:

Reviewer #3 (Recommendations for the authors):

(1) Add the plots showing diabetes/non-diabetes sub-group analysis and power estimates to the Supplementary Figures (rather than just as a Supplementary table)

We have added this as Figure 5-figure supplement 2 in the revised manuscript (R2).

(2) Add a short note on the validity of the results limiting to the non-diabetes subgroup to the limitations section

We have done this in the revised manuscript (R2).

(3) Add a short note on the missing adjustment for fibrosis/diabetes interactions in the study to the limitations paragraph

We appreciate the reviewer’s suggestion to address the lack of adjustment for potential fibrosis–diabetes interaction. We added a note to the limitations paragraph in the Limitations section. Although diabetes considerably modulates the risk for steatohepatitis, only a small number of participants had diabetes (29 of 109) in our study, undermining statistical power to detect meaningful interaction effects.

Author response table 1.

(4) Fig S10/6E: In vitro TGF-b stimulation on spheroids, LX-2 cells, hepatocytes: evaluate expression of RAB6A, ARL4A, RAB27B, DIRAS2 genes from 6E to create consistency between the findings. Confirm ACTA2 up-regulation in LX-2 cells treated with TGF-β as a positive control. Also specify methods for gene expression analysis in spheroids and the cell types in the figure legends (RNA-Seq? qPCR?)

To address this reviewer’s concerns, we have done qPCR for RAB6A, ARL4A, RAB27B, DIRAS2 in LX-2 cells stimulated with TGF-β and the results are shown in the revised now Figure 6–figure supplement 5. To align all three graphs displaying the same genes analyzed, we have now depicted the gene expression for the co-culture (hepatocytes, hepatic stellate cells, and Kupffer cells) and mono-culture (hepatocytes only) from RNAseq analysis. We have also updated the methods that we used in the figure legend.

(5) Validate the functionality of hepatocytes in the 3D liver spheroid model used (PMID: 39605182) at the time points of which the experiments have been performed (e.g. Albumin secretion, CYP-assays).

We agree that the characterization of the liver spheroids from human patients using fully differentiated cells, is important but this has already been done and is published in these papers:

(1) Bell, C. C. et al. Transcriptional, Functional, and Mechanistic Comparisons of Stem Cell–Derived Hepatocytes, HepaRG Cells, and Three-Dimensional Human Hepatocyte Spheroids as Predictive In Vitro Systems for Drug-Induced Liver Injury. Drug Metab. Dispos. 45, 419–429 (2017).

(2) Bell, C. C. et al. Characterization of primary human hepatocyte spheroids as a model system for drug-induced liver injury, liver function and disease. Sci. Rep. 6, 25187 (2016). 3.Vorrink, S. U. et al. Endogenous and xenobiotic metabolic stability of primary human hepatocytes in long‐term 3D spheroid cultures revealed by a combination of targeted and untargeted metabolomics. FASEB J. 31, 2696–2708 (2017).

(4) Messner, S. et al. Transcriptomic, Proteomic, and Functional Long-Term Characterization of Multicellular Three-Dimensional Human Liver Microtissues. Appl. In Vitro Toxicol. 4, 1–12 (2018).

(5) Bell, C. C. et al. Comparison of Hepatic 2D Sandwich Cultures and 3D Spheroids for Long-term Toxicity Applications: A Multicenter Study. Toxicol. Sci. 162, 655–666 (2018).

We have mentioned this now in the manuscript on page 18 and also the Limitation section to make this point clear.

(6) Add a note on limitations of the PHH-spheroid and cell line in vitro models to the limitations section and discuss the need for future experiments to examine the cellular crosstalk and cell types potentially responsible for the proposed GTPase-gene dysregulation.

We have added this to the limitation section on page 13 this in the revised manuscript (R2).

eLife Assessment

Kambali et al use optogenetic manipulations to examine whether the ventral hippocampal Schaffer collateral (vCA3-to-vCA1) and temporoammonic (EC-to-vCA1) pathways regulate anxiety- and fear-related behaviors in mice. They find that both pathways regulate the expression of fear (freezing) responses to a context and auditory conditioned stimulus paired with foot shock (trace conditioning protocol), but only the Schaffer collateral pathway regulates the expression of anxiety-related behaviors in the elevated plus maze, open field test, and Vogel conflict test. Overall, the study is valuable: it detects bidirectional effects of optogenetic excitation and inhibition in both pathways. However, the strength of the evidence in support of its main claims is incomplete.

Reviewer #1 (Public review):

Summary:

The hippocampus, especially the ventral subregion, has been related to emotional processing. However, the specific circuitry involved deserves further investigation. By using a bidirectional optogenetic modulation, Kambali et al. have investigated the role of different inputs to vCA1 (i.e., from vCA3 and entorhinal cortex) in anxiety- and fear-related responses. The major findings of this work suggested that both inputs to vCA1 control fear-related responses, whereas only the projection between vCA3 and vCA1 controls anxiety-related behavior. Overall, the authors used an advanced methodological approach, which allows them to modulate specific brain circuits, to study specific hippocampal projections, providing some new information regarding the hippocampal function in anxiety and fear.

Strengths:

(1) The manuscript is well written, clear and has a detailed and specific discussion.

(2) Results from each optogenetic manipulation are clear in different anxiety- and fear-related tasks, demonstrating the robustness of the findings.

(3) The overall conclusions are very interesting and might be relevant for the field of mental health disorders accompanied by anxiety- and fear-related alterations.

Weaknesses:

(1) The major differences in basal behavioral performance in the different paradigms between the two optogenetic modulations prevent the achievement of strong conclusive results.

(2) Data presentation and representative figures need a major revision.

(3) No analysis has been performed to analyze potential sex differences in behavioral domains where sex is important.

Reviewer #2 (Public review):

Summary:

This paper uses an optogenetic approach to either activate or inhibit separate neural pathways projecting to the ventral CA1 hippocampal subregion, from either CA3 or the entorhinal cortex. The authors report that manipulation of the vCA3→vCA1 pathway affected behavioural performance on a number of tasks: elevated plus maze, open field, Vogel conflict test and freezing behaviour to both context and a trace CS cue. In contrast, optogenetic manipulation of neural activity in the EC→vCA1 pathway only affected behaviour on the trace CS/context fear memory test but had no effect on the elevated plus maze, open field or Vogel conflict test. The authors suggest different roles for these two ventral hippocampal pathways in fear versus anxiety.

Strengths:

This is an interesting study addressing an important question in a highly topical subject area. The experiments are well conducted and have generated interesting and important data.

Weaknesses:

While I am broadly sympathetic to the overall narrative of the paper, I have some questions/comments around the specific interpretation of the results presented. In my view, the authors' claims may not be completely supported by their data, but the data are interesting nonetheless.

In terms of the framework presented by the authors for interpreting their data, many would argue that freezing (or at least reduced activity/behavioural inhibition) to the context provides a readout of conditioned anxiety rather than fear. In this sense, the context is a signal of potential threat (i.e. the context becomes associated with both shock and with the absence of shock) and thus generates anxiety rather than fear. Likewise, the trace CS cue could be considered as an ambiguous predictor of shock in that the shock doesn't occur straight away. In contrast, a punctate CS cue which co-terminates with shock would be a reliable signal of imminent threat and thus generates a fear response. Thus, it might be argued that all of the assays adopted by the authors are readouts of anxiety (albeit comprising tests of both conditioned and unconditioned anxiety). For example, from the authors' perspective, it is not clear a priori why the Vogel conflict test is considered anxiety, but contextual freezing is considered fear? Indeed, in the Discussion, the authors mention another study in which the data from the Vogel conflict test align with fear assays rather than anxiety tests. Can the authors elaborate on their distinction? I appreciate that, in practice, it might be difficult to distinguish between fear and anxiety at the behavioural level in rodents (although opposing effects of fear and anxiety on pain responses might be one option). At the very least, this issue merits further discussion.

Another question is whether rather than representing a qualitative difference between the contributions of the vCA3→vCA1 and EC→vCA1 pathways to different aspects of fear/anxiety behaviours, the different results reflect a quantitative difference between the magnitude of effects in vCA1 that are generated from optogenetic manipulation of the two pathways, coupled with the possibility that behaviour on the trace CS/context fear memory task is more sensitive to manipulation than the "anxiety tests". The possibility that vCA3→vCA1 stimulation is more effective is potentially supported by the c-fos measurements in vCA1. vCA3→vCA1 stimulation produced a much bigger vCA1 c-fos response (approx. 350% c-fos cell activation; see Figure 1E) compared to activation of the EC→vCA1 pathway (approx. 170% c-fos cell activation; see Figure 4E).

Furthermore, in some studies, there seem to be quite large differences between the laser OFF conditions for the different groups (which presumably one would not expect to be different). For example, compare laser OFF for the Inhibition group for time in open arms of EPM in Figure 5C (> 40%) versus laser OFF for the Inhibition group for time in open arms of EPM in Fig. 2C (< 20%). This could potentially result in ceiling effects, such that it is very hard to see a further increase in time in the open arms from a level already above 40% when the laser is then switched on. This could complicate the interpretation of the laser ON condition.

Likewise, there is a big difference between the behavioral performance of the two SHAM groups in Figure 3 (compare SHAM in 3 B, C and SHAM in 3 D, E). How is this explained? Could this generate a ceiling effect? This may also merit some discussion. More details on the SHAM procedure(s) in the main manuscript may also be helpful.

According to Figure 3A, the test of freezing response to the trace Tone CS is conducted in a different context from the conditioning context. The data presented in Figure 3 for tone fear are the levels of freezing during the presentation of this cue in the different contexts. It would be important to present both pre-CS and CS freezing levels here to determine how much of the freezing is actually driven by the punctate tone CS. The pre-CS freezing levels in this different context would also provide a nice control for the contextual fear conditioning.

Reviewer #3 (Public review):

Summary:

In their paper entitled "Ventral hippocampal temporoammonic and Schaffer collateral pathways differential control fear- and anxiety-related behaviors" the authors use a bidirectional optogenetic approach to elucidate the role of temporammonic (TA) and Schaffer collateral (SC) inputs to the ventral hippocampus (CA1) in modulating both fear and anxiety-related behaviors. While fear and anxiety behaviors are often considered on a continuous spectrum, identifying neural pathways that are differentially activated represents an important open question in the field. The authors find that optogenetic stimulation or inhibition of the Schaffer Collateral pathway in the ventral hippocampus (CA3-CA1) bidirectionally modulates both fear-related and anxiety-related behavioral paradigms. More specifically, optogenetic excitation of the CA3-CA1 pathway using ChR2-expressing viral constructs increases anxiety-like behaviors in numerous behavioral paradigms (elevated plus maze, open field, Vogel conflict test). Conversely, optogenetic inhibition using halorhodopsin reduced anxiety-like behaviours. To examine fear behaviors, the authors examined contextual and trace fear conditioning. Similar to their results with anxiety-like behaviors, the authors observed bidirectional fear modulation following optogenetic stimulation of the vCA3-vCA1 pathway. The authors next examined the temporammonic pathway originating from the lateral entorhinal cortex to vCA1. Unlike with SC stimulation, stimulation of the TA pathway had no effect on anxiety-like behaviors but did bidirectionally modulate contextual fear conditioning. Together, these results differentiate the SC and TA pathways in the ventral hippocampus as distinct regulators of affective behavior.

Strengths:

The paper has numerous technical strengths, including dissecting the role of both excitation and inhibition of both pathways and the use of behavioral measures of anxiety and fear. This balanced and internally controlled design allows readers to evaluate the effects of both pathways in a single study, thereby reducing technical complications from experiments being completed across laboratories and experimental conditions.

Weaknesses:

There are a few limitations of the study, however, which bear discussion.

(1) The authors use halorhodopsin to achieve optogenetic inhibition. Halorhodopsin is generally considered a first-generation optogenetic actuator, as it is a Cl- pump rather than an ion channel. This limits the degree of inhibition (i.e. by preventing shunting inhibition) and can result in altered chloride gradients in the period immediately following optogenetic stimulation. This is of particular concern in this paper as the stimulation parameters and behavioral analysis are not temporally correlated, therefore confounds of disrupted chloride cannot be experimentally accounted for or controlled.

(2) The authors use an AAV-CaMKII-eGFP as a control (Sham) throughout the dataset; however, in the trace fear conditioning experiments, there are no AAV-CaMKII-ChR2-eYFP or AAV-CaMKII-eNpHR3.0-eYFP controls without optogenetic stimulation. Therefore, it is unclear the extent to which viral expression of optogenetic actuators impacts behavior. Additionally, the authors only provided optogenetic stimulation during contextual fear recall and tone fear recall. Additional experiments disrupting each pathway during trace conditioning would have provided additional insight into the role of each pathway in the initial encoding of fear memories.

(3) The location and extent of viral expression across animals were not systematically quantified.

Overall, however, these weaknesses do not significantly detract from the main conclusions of the paper. The authors' data convincingly demonstrates that disruption of the trisynaptic circuit bidirectionally modulates both fear- and anxiety-like behaviors while disruption of the temporammonic pathway has no effect on anxiety-like behaviors but disrupts fear-related behaviors. It is interesting to note, however, that the TA activation had no effect on tone-related fear conditioning, suggesting a potential specialized role of the temporammonic pathway specifically in contextual fear memory.

Author response:

Public Reviews:

Reviewer #1 (Public review):

Summary:

The hippocampus, especially the ventral subregion, has been related to emotional processing. However, the specific circuitry involved deserves further investigation. By using a bidirectional optogenetic modulation, Kambali et al. have investigated the role of different inputs to vCA1 (i.e., from vCA3 and entorhinal cortex) in anxiety- and fear-related responses. The major findings of this work suggested that both inputs to vCA1 control fear-related responses, whereas only the projection between vCA3 and vCA1 controls anxiety-related behavior. Overall, the authors used an advanced methodological approach, which allows them to modulate specific brain circuits, to study specific hippocampal projections, providing some new information regarding the hippocampal function in anxiety and fear.

Strengths:

(1) The manuscript is well written, clear and has a detailed and specific discussion.

(2) Results from each optogenetic manipulation are clear in different anxiety- and fear-related tasks, demonstrating the robustness of the findings.

(3) The overall conclusions are very interesting and might be relevant for the field of mental health disorders accompanied by anxiety- and fear-related alterations.

Weaknesses:

(1) The major differences in basal behavioral performance in the different paradigms between the two optogenetic modulations prevent the achievement of strong conclusive results.

The two projections of ventral CA1 were studied independently in different cohorts of animals tested at different times during the study. This difference in timing may have contributed to variations in the basal behavioral performance between the two projections. Importantly we found that within each cohort – control and optogenetic manipulation, the basal performance within each set of experiments (i.e., corresponding to projections) is highly consistent, e.g., basal cued and contextual freezing responses and responses to OFF conditions in Vogel conflict test. Moreover, the ANOVA statistics conducted across the baseline and ON conditions for each task revealed robust significant effects of bidirectional optogenetic modulation for each cohort. In case of the fear responses, a point to note is that the freezing levels in SHAM controls differ between projections but are consistent between two types of assessments (tone and context) within each projection. We will mention these limitations in the revised manuscript.

(2) Data presentation and representative figures need a major revision.

The figures will be rearranged according to the projections. The anxiety-related figures and fear response related figures will be grouped for each projection to improve clarity and readability. The revised manuscript will include representative heat maps for each behavioral task for both projections in addition to population quantification data.

(3) No analysis has been performed to analyze potential sex differences in behavioral domains where sex is important.

This assessment was not done in the original submission. We will perform statistical analysis for male and female mice separately and if the results are sex-dependent, we will present separate figures. Otherwise, the combined data presentation will be followed.

Reviewer #2 (Public review):

Summary:

This paper uses an optogenetic approach to either activate or inhibit separate neural pathways projecting to the ventral CA1 hippocampal subregion, from either CA3 or the entorhinal cortex. The authors report that manipulation of the vCA3→vCA1 pathway affected behavioural performance on a number of tasks: elevated plus maze, open field, Vogel conflict test and freezing behaviour to both context and a trace CS cue. In contrast, optogenetic manipulation of neural activity in the EC→vCA1 pathway only affected behaviour on the trace CS/context fear memory test but had no effect on the elevated plus maze, open field or Vogel conflict test. The authors suggest different roles for these two ventral hippocampal pathways in fear versus anxiety.

Strengths:

This is an interesting study addressing an important question in a highly topical subject area. The experiments are well conducted and have generated interesting and important data.

Weaknesses:

While I am broadly sympathetic to the overall narrative of the paper, I have some questions/comments around the specific interpretation of the results presented. In my view, the authors' claims may not be completely supported by their data, but the data are interesting nonetheless.

In terms of the framework presented by the authors for interpreting their data, many would argue that freezing (or at least reduced activity/behavioural inhibition) to the context provides a readout of conditioned anxiety rather than fear. In this sense, the context is a signal of potential threat (i.e. the context becomes associated with both shock and with the absence of shock) and thus generates anxiety rather than fear. Likewise, the trace CS cue could be considered as an ambiguous predictor of shock in that the shock doesn't occur straight away.

In contrast, a punctate CS cue which co-terminates with shock would be a reliable signal of imminent threat and thus generates a fear response. Thus, it might be argued that all of the assays adopted by the authors are readouts of anxiety (albeit comprising tests of both conditioned and unconditioned anxiety).

We agree with the reviewer that context and trace fear conditioning do not represent an “imminent” threat as severe as would likely be internalized in delay fear conditioning. However, the goal of the study was to probe hippocampal dependent processes (contextual and trace fear conditioning are strongly modulated by the hippocampus while delay conditioning is not). Consistent with several other studies, we believe the conditional nature of the task (context and trace are invariably linked to shock) provides support for a “non-ambiguous” relationship that is conducive for measuring the assessment of fear-based behavior.

Several studies show clear differences in the involvement of amygdala and hippocampus in delay vs. trace fear conditioning. Inactivating amygdala led to deficits in contextual and delay conditioning but had no effect on trace conditioning. In contrast, inactivating hippocampus led to deficits in trace and contextual but not delay fear conditioning. These findings suggest that a temporal gap between the CS and US can generate amygdala-independent but hippocampal-dependent fear conditioning (Raybuck J. D., Lattal K. M 2011, PMID: 21283812). Lesions of the entorhinal cortex impair the acquisition of trace fear conditioning but not the acquisition of delay fear conditioning (Raybuck J. D., Lattal K. M 2011, PMID: 21283812) . Further, using single unit recording during fear retention tests after delay or trace fear conditioning, the study showed that entorhinal neurons specifically respond after trace but not after delay fear conditioning (Kong et al 2023, PMID: 36919333). These findings demonstrate that trace fear conditioning and delay fear conditioning may involve overlapping but largely different neuronal circuits. A knockdown of the expression of the α5-subunit–containing GABA_𝐴 receptors in the CA1 region (α5CA1KO mice) leads to improved spatial learning and enhanced trace fear conditioning memory, actually to the level of delay fear conditioning, suggesting that α5GABA_𝐴Rs in CA1 pyramidal neurons normally constrain hippocampus-dependent memory processes and that trace fear conditioning in the absence of a5-GABA_𝐴 receptors in CA1 has the same effect size as delay fear conditioning (Engin et al 2020, PMID: 32934095), supporting the view that trace fear conditioning is not “ambiguous”.

For example, from the authors' perspective, it is not clear a priori why the Vogel conflict test is considered anxiety, but contextual freezing is considered fear? Indeed, in the Discussion, the authors mention another study in which the data from the Vogel conflict test align with fear assays rather than anxiety tests. Can the authors elaborate on their distinction? I appreciate that, in practice, it might be difficult to distinguish between fear and anxiety at the behavioral level in rodents (although opposing effects of fear and anxiety on pain responses might be one option). At the very least, this issue merits further discussion.

We will make this distinction clearer in the revisions. Briefly, behavioral actions in the Vogel conflict test are generally considered to be most pertinent to general anxiety disorders in humans and anxiolytics have high predictive validity in animals in this task. In particular, the robust actions of benzodiazepines and 5-HT_1A partial agonists parallel their clinical efficacy in patients (McMillan and Brocco, 2003, PMID: 12600703).

Our previous study (Engin et al 2016, PMID: 26971710) used global diazepam-induced neuronal inhibition and identified that positive modulation of α2-GABA_𝐴Rs in dentate gyrus granule cells and CA3 pyramidal neurons is required to reduce anxiety-like behaviors while inhibition of positive modulation of α2-GABA_𝐴Rs in CA1 pyramidal neurons is required to reduce fear-related behaviors. The effects were absent when α2-GABA_𝐴Rs was knocked out in the respective subregions. These results indicate that these intrahippocampal subregions can modulate fear and anxiety-like behaviors independently of the amygdala. In the previous study we used conditional α2-GABA_𝐴R knockouts in hippocampal subregions and subjected these mice to systemic diazepam. In these experiments, diazepam still acts on α1-, α3- and α5-_𝐴Rs in the hippocampal subregions and cell types in which when α2-GABA_𝐴Rs are lacking. Therefore, for example when α2CA1KO mice were administered diazepam, diazepam still led to inhibition of pyramidal neurons in CA3 and DG via α1-, α2-, α3- and α5- GABA_𝐴Rs, and in addition, diazepam also inhibited α1-, α3- and α5- GABA_𝐴Rs in CA1 itself. Diazepam also acted on GABA_𝐴Rs in amygdala or other brain regions. These are fundamentally different experimental conditions compared to the optogenetic experiment described in this paper. Moreover, in contrast to the current paper, the previous work did not examine projections but used global diazepam-induced neuronal inhibition as a baseline. Moreover, whereas the previous paper examined whether a specific neuronal cell type was required for anxiolytic-like or fear-like actions, the current manuscript examined whether activation or inhibition of neuronal projections is sufficient to modulate anxiety- and fear-related behaviors. Overall, one cannot easily compare the results in the Vogel conflict test in both papers.

Another question is whether rather than representing a qualitative difference between the contributions of the vCA3→vCA1 and EC→vCA1 pathways to different aspects of fear/anxiety behaviours, the different results reflect a quantitative difference between the magnitude of effects in vCA1 that are generated from optogenetic manipulation of the two pathways, coupled with the possibility that behaviour on the trace CS/context fear memory task is more sensitive to manipulation than the "anxiety tests". The possibility that vCA3→vCA1 stimulation is more effective is potentially supported by the c-fos measurements in vCA1. vCA3→vCA1 stimulation produced a much bigger vCA1 c-fos response (approx. 350% c-fos cell activation; see Figure 1E) compared to activation of the EC→vCA1 pathway (approx. 170% c-fos cell activation; see Figure 4E).

Furthermore, in some studies, there seem to be quite large differences between the laser OFF conditions for the different groups (which presumably one would not expect to be different). For example, compare laser OFF for the Inhibition group for time in open arms of EPM in Figure 5C (> 40%) versus laser OFF for the Inhibition group for time in open arms of EPM in Fig. 2C (< 20%). This could potentially result in ceiling effects, such that it is very hard to see a further increase in time in the open arms from a level already above 40% when the laser is then switched on. This could complicate the interpretation of the laser ON condition.

The magnitude of activation as evidenced by c-fos measurements differs between the two projections. This might reflect different levels of modulations of CA1 neuronal activity. The fact that the two projections were studied at different time points (see response to reviewer 1) may also have contributed to the difference. The revised manuscript will include a formal discussion about magnitude of modulation that could contribute to differential sensitivity for the modulation of anxiety-like behaviors. However, the inputs from these two projections systems target different regions of CA1 pyramidal neurons and each pathway has distinct roles in other processes (sensory versus memory-based completion) – thus a dissociation may also be present for other types of behavior as well including the modulation of anxiety-like behaviors.

While it is possible that ceiling effects could impact our interpretation, we believe ceiling effects would only impact one direction of the optogenetic manipulation and there was no effect of activation (Fig. 5C) or bidirectional modulation of anxiety-related behavior in the novel open field test (Fig. 5F) which has levels of behavior comparable to Figure 2F.

Likewise, there is a big difference between the behavioral performance of the two SHAM groups in Figure 3 (compare SHAM in 3 B, C and SHAM in 3 D, E). How is this explained? Could this generate a ceiling effect? This may also merit some discussion. More details on the SHAM procedure(s) in the main manuscript may also be helpful.

With respect to contextual fear, ceiling effects are not a major factor as we still see enhanced freezing in the activation condition. With tone fear, we cannot formally exclude a ceiling effect, and this will be addressed as a potential confound in the manuscript.

According to Figure 3A, the test of freezing response to the trace Tone CS is conducted in a different context from the conditioning context. The data presented in Figure 3 for tone fear are the levels of freezing during the presentation of this cue in different contexts. It would be important to present both pre-CS and CS freezing levels here to determine how much of the freezing is actually driven by the punctate tone CS. The pre-CS freezing levels in this different context would also provide a nice control for the contextual fear conditioning.

We agree and will analyze and report the pre-CS freezing data in the revision.

Reviewer #3 (Public review):

Summary:

In their paper entitled "Ventral hippocampal temporoammonic and Schaffer collateral pathways differential control fear- and anxiety-related behaviors" the authors use a bidirectional optogenetic approach to elucidate the role of temporammonic (TA) and Schaffer collateral (SC) inputs to the ventral hippocampus (CA1) in modulating both fear and anxiety-related behaviors. While fear and anxiety behaviors are often considered on a continuous spectrum, identifying neural pathways that are differentially activated represents an important open question in the field. The authors find that optogenetic stimulation or inhibition of the Schaffer Collateral pathway in the ventral hippocampus (CA3-CA1) bidirectionally modulates both fear-related and anxiety-related behavioral paradigms. More specifically, optogenetic excitation of the CA3-CA1 pathway using ChR2-expressing viral constructs increases anxiety-like behaviors in numerous behavioral paradigms (elevated plus maze, open field, Vogel conflict test). Conversely, optogenetic inhibition using halorhodopsin reduced anxiety-like behaviours. To examine fear behaviors, the authors examined contextual and trace fear conditioning. Similar to their results with anxiety-like behaviors, the authors observed bidirectional fear modulation following optogenetic stimulation of the vCA3-vCA1 pathway. The authors next examined the temporammonic pathway originating from the lateral entorhinal cortex to vCA1. Unlike with SC stimulation, stimulation of the TA pathway had no effect on anxiety-like behaviors but did bidirectionally modulate contextual fear conditioning. Together, these results differentiate the SC and TA pathways in the ventral hippocampus as distinct regulators of affective behavior.

Strengths:

The paper has numerous technical strengths, including dissecting the role of both excitation and inhibition of both pathways and the use of behavioral measures of anxiety and fear. This balanced and internally controlled design allows readers to evaluate the effects of both pathways in a single study, thereby reducing technical complications from experiments being completed across laboratories and experimental conditions.

Weaknesses:

There are a few limitations of the study, however, which bear discussion.

(1) The authors use halorhodopsin to achieve optogenetic inhibition. Halorhodopsin is generally considered a first-generation optogenetic actuator, as it is a Cl- pump rather than an ion channel. This limits the degree of inhibition (i.e. by preventing shunting inhibition) and can result in altered chloride gradients in the period immediately following optogenetic stimulation. This is of particular concern in this paper as the stimulation parameters and behavioral analysis are not temporally correlated, therefore confounds of disrupted chloride cannot be experimentally accounted for or controlled.

Choice of halorhodopsin was in part influenced by a report that spontaneous archaerhodopsin activation was paradoxically associated with increased spontaneous release of neurotransmitter from presynaptic terminals, whereas activation of chloride-reducing halorhodopsin triggered neurotransmitter release upon light onset (Mahn et al., PMID: 26950004), suggesting that halorhodospin may be advantageous in studies inhibiting presynaptic nerve terminals. Halorhodpsin has been used in several studies to effectively silence activity and had substantial influence on behavioral in our studies that was inversely proportional to ChR2 stimulation. While perhaps not optimal out of an abundance of caution, we chose it over Archaerhodopsin based on the cited literature.

(2) The authors use an AAV-CaMKII-eGFP as a control (Sham) throughout the dataset; however, in the trace fear conditioning experiments, there are no AAV-CaMKII-ChR2-eYFP or AAV-CaMKII-eNpHR3.0-eYFP controls without optogenetic stimulation. Therefore, it is unclear the extent to which viral expression of optogenetic actuators impacts behavior. Additionally, the authors only provided optogenetic stimulation during contextual fear recall and tone fear recall. Additional experiments disrupting each pathway during trace conditioning would have provided additional insight into the role of each pathway in the initial encoding of fear memories.

Thank you for your observation. We have used a SHAM control that was injected with the AAV vector without any opsins. In fear conditioning experiments we performed optogenetic manipulations only during the fear response either with context or cue recall. This aligned well with our hypothesis to test whether the intrahippocampal projections play any role in fear response modulation. Investigating the role of each pathway during acquisition of trace and/or contextual fear conditioning is also highly relevant; however, evaluating these projections in fear memory formation was beyond the scope of this study. The observation that we can bidirectionally modulate fear responses with light is consistent with (although it does not prove) a light-specific modulation. In any case, even if there were baseline effects without light, they would still be suggestive of the effects observed being mediated by the optogenetic actuators.

(3) The location and extent of viral expression across animals were not systematically quantified.Overall, however, these weaknesses do not significantly detract from the main conclusions of the paper. The authors' data convincingly demonstrates that disruption of the trisynaptic circuit bidirectionally modulates both fear- and anxiety-like behaviors while disruption of the temporammonic pathway has no effect on anxiety-like behaviors but disrupts fear-related behaviors. It is interesting to note, however, that the TA activation had no effect on tone-related fear conditioning, suggesting a potential specialized role of the temporammonic pathway specifically in contextual fear memory.

Thank you for your thoughtful description of the present study. It is true that TA pathway is distinct from vCA3 to vCA1 pathway in various ways, one being the synapse formation of these two projections are at different locations or layers on vCA1 neurons i.e., the TA pathway synapses on the stratum lacunosum-moleculare (LMol) layer while the vCA3 to vCA1 pathway synapses at stratum radiatum (Rad), close to the CA1 pyramidal cell layer, which is in line with differential functions of the two projections They modulate the pyramidal cell activity in a different way, with TA pathway synapses being distinct from vCA3 to vCA1 synapses on the pyramidal cell layer, which may result in different computational properties of the two projections. Additionally, TA projections are modulated by dopamine while projections from vCA3 are not, but the projections from vCA3 receive inputs from various sources including collaterals, and entorhinal via dentate gyrus. These distinct features of the two projections may contribute to differential modulation of vCA1 activity. We note that cue-related fear is not affected by the TA activation, however even in this case, the TA pathway activation by channelrhodopsin or inhibition by halorhodopsin results in a decrease or an increase of the contextual fear response, respectively.

eLife Assessment

This study provides valuable insights into the regulation of myogenic differentiation by identifying Leiomodin 1 as a modulator of proteome dynamics during myogenic differentiation. The combination of quantitative proteomics with functional perturbation experiments offers solid evidence supporting the idea that SIRT1 influences perturbations of myogenic differentiation upon LMOD1 inactivation. These findings advance our understanding of muscle differentiation and will be of interest to researchers studying muscle development and related pathologies

Reviewer #1 (Public review):

The main significance of this work is characterizing the function of a new gene Lmod1 in muscle stem cell biology. The study suggests an intriguing regulatory mechanism by which Sirt1 sequesters Lmod1 in a specific temporal window during myogenesis.

Comments on revisions:

The authors have satisfactorily addressed my inquires. Thank you.

Reviewer #2 (Public review):

Summary:

In this manuscript, the authors identify Leiomodin-1 (LMOD1) as a key regulator of early myogenic differentiation, demonstrating its interaction with SIRT1 to influence SIRT1's cellular localization and gene expression. The authors propose that LMOD1 translocates SIRT1 from the nucleus to the cytoplasm to permit the expression of myogenic differentiating genes such as MYOD or Myogenin.

Strengths:

A major strength of this work lies in the robust temporal resolution achieved through a time-course mass spectrometry analysis of in vitro muscle differentiation. This provides novel insights into the dynamic process of myogenic differentiation, often under explored in terms of temporal progression. The authors provide a strong mechanistic case for how LMOD1 exerts its role on muscle differentiation which opens avenues to modulate.

Weaknesses:

In the revised manuscript, the authors begin to translate their in vitro findings to an in vivo context by examining SIRT1 expression across a regeneration time course (Fig. 4I). They observe an increase in SIRT1 expression concomitant with LMOD1, supporting a potential role for SIRT1 in myogenic differentiation. Future studies will be required to provide deeper mechanistic insight into SIRT1 function in vivo.

Discussion:

Overall, the study emphasizes the importance of understanding the temporal dynamics of molecular players during myogenic differentiation and provides valuable proteomic data that will benefit the field. Future studies should explore whether LMOD1 modulates the nuclear-cytoplasmic shuttling of other transcription factors during muscle development and how these processes are mechanistically achieved. Investigating whether LMOD1 can be therapeutically targeted to enhance muscle regeneration in contexts such as exercise, aging, and disease will be critical for translational applications. Additionally, elucidating the interplay among LMOD1, LMOD2, and LMOD3 could uncover broader implications for actin cytoskeletal regulation in muscle biology. The authors have nicely updated their discussion.

Reviewer #3 (Public review):

Summary:

In this manuscript, the investigators identified LMOD1 as one of a subset of cytoskeletal proteins that levels increase in early stages of myogenic differentiation. Lmod1 is understudied in striated muscle and in particular in myogenic differentiation. Thus, this is an important study. It is also a very thorough study, with perhaps even too much data presented. Importantly, the investigators observed that LMOD1 appears to be important for skeletal regeneration, myogenic differentiation and that it interacts with SIRT1. Both primary myoblast differentiation and skeletal muscle regeneration were studied. Rescue experiments confirmed these observations: SIRT1 can rescue perturbations of myogenic differentiation as a result of LMOD1 knockdown.

Strengths:

Particular strengths include: an important topic, the use of primary skeletal cultures, the use of both cell culture and in vivo approaches, careful biomarker analysis of primary mouse myoblast differentiation, the use of two methods to probe the function of the Lmod1/SIRT1 pathway via using depletion approaches and inhibitors, and the generation of six independent myoblast cultures. Results support their conclusions.

Weaknesses:

(1) Figure 1. Images of cells in Figure 1A are too small to be meaningful (especially in comparison to the other data presented in this figure). Perhaps make graphs smaller?

(2) Line 148 "We found LMOD2 to be the most abundant Lmod in whole skeletal muscle." This is confusing since most, if not all, prior studies have shown that Lmod3 is the predominant isoform in skeletal muscle. The two papers that are cited are incorrectly cited. Clarification to resolve this discrepancy is needed.

(3) Figure 2. Immunofluorescence (IF) panels are too small to be meaningful. Perhaps the graphs could be made smaller and more space allocated for the IF panels? This issue is apparent for just about all IF panels - they are simply too small to be meaningful. Additionally, in many of the immunofluorescence figures, the colors that were used make it difficult to discern the stained cellular structures. For example, in Figure S1, orange and purple are used - they do not stand out as well as other colors that are more commonly used.

(4) There is huge variability in many experiments presented - as such, more samples appear to be required to allow for meaningful data to be obtained. For example, Figure S2. Many experimental groups, only have 3 samples - this is highly problematic - I would estimate that 5-6 would be the minimum.

(5) Ponceau S staining is often used as a loading control in this manuscript for western blots. The area/molecular weight range actually used should be specified. Not clear why in some experiments GAPDH staining is used, in other experiments Ponceau S staining is used, and in some, both are used. In some experiments the variability of total protein loaded from lane-to-lane is disconcerting. For example, in Figure S4C there appears to be more than normal variability. Can the protein assay be redone and the samples run again?

(6) Figure S3 - Lmod3 is included in the figure but no mention of it occurs in the title of the figure and/or legend.

(7) Abstract, line 25. "overexpression accelerates and improves the formation of myotubes". This is a confusing sentence. How is it improving the formation? A little more information about how they are different than developing myotubes in normal/healthy muscle would be helpful

(8) Impossible from IF figures presented to determine where Lmod1 localizes in the myocytes. Information on its subcellular localization is important. Does it localize with Lmod2 and Lmod3 at thin filament pointed ends?

Comments on revisions:

Many comments have been adequately addressed. However, some concerns remain.

Former Concern #2. The issue with the lack of detection of LMOD3 in their muscle samples is troublesome and has not been adequately resolved in the revised manuscript. It is a fact that most, if not all, studies on Lmod3 report that it is the most abundant isoform in skeletal muscle. This issue should be discussed in the manuscript. It is recognized that a different assay was utilized in this paper. The papers that are cited continue to remain incorrect. Specifically:

Tsukada et al., reports abundance of LMOD2 in cardiac muscle, not in skeletal muscle.

Nworu et al., 2015 reports on LMOD3 in skeletal muscle.

Kiss et al.,2020. While this paper reveals an important function for Lmod2 in thin filament length regulation, it is clearly shows many examples of high expression of Lmod3 in various skeletal muscles isolated from mice.

Former Concern #3. With respect to small sample numbers. Hopefully a statistical editor is available to comment. While this reviewer is happy that other assays were used to verify their data, the problem still remains that many experimental groups only have 3 samples (with high variability).

Former Concern #3. Many immunofluorescence panels are hard to evaluate because of their small size.

Author response:

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

eLife Assessment

This valuable study offers insights into the role of Leiomodin-1 (LMOD1) in muscle stem cell biology, advancing our understanding of myogenic differentiation and indicating LMOD1 as a regulator of muscle regeneration, aging, and exercise adaptation. The integration of in vitro and in vivo approaches, complemented by proteomic and imaging methodologies, is solid. However, certain aspects require further attention to improve the clarity, impact, and overall significance of the work, particularly in substantiating the in vivo relevance. This work will provide a starting point that will be of value to medical biologists and biochemists working on LMOD and its variants in muscle biology.

Thank you for the positive feedback on our manuscript and the constructive criticism provided by the reviewers that helped us improve our manuscript.

Public Reviews:

Reviewer #1 (Public review):

This manuscript by Ori and colleagues investigates the role of Lmod1 in muscle stem cell activation and differentiation. The study begins with a time-course mass spectrometry analysis of primary muscle stem cells, identifying Lmod1 as a pro-myogenic candidate (Figure 1). While the initial approach is robust, the subsequent characterization lacks depth and clarity. Although the data suggest that Lmod1 promotes myogenesis, the underlying mechanisms remain vague, and key experiments are missing. Please find my comments below.

We thank the reviewer for the positive feedback on our manuscript and the helpful comments, which helped improve it.

(1) The authors mainly rely on coarse and less-established readouts such as myotube length and spherical Myh-positive cells. More comprehensive and standard analyses, such as co-staining for Pax7, MyoD, and Myogenin, would allow quantification of quiescent, activated, and differentiating stem cells in knockdown and overexpression experiments. The exact stage at which Lmod1 functions (stem cell, progenitor, or post-fusion) is unclear due to the limited depth of the analysis. Performing similar experiments on cultured single EDL fibers would add valuable insights.

We thank the reviewer for this comment. In addition to performing standard measurements such as staining for Myogenin and Myosin Heavy Chain (Figure S2H), we focused on morphological readouts, such as myotube formation, because LMOD1 is an actin cytoskeleton-associated protein. Therefore, we reasoned its function would be most directly reflected in structural changes during differentiation, rather than solely in early transcriptional markers. 

Regarding the use of standard markers, we have already performed co-staining for Myogenin and Myosin Heavy Chain (MHC), which effectively quantifies early myogenic committed (Myogenin+/MHC-) and terminally differentiating (Myogenin+/MHC+) cells (Figure S2H). We did not include Pax7 as our primary culture system consists of already activated myoblasts, where Pax7 is not a reliable marker of quiescence. Our data also suggest that Lmod1 is important in regulating differentiation with comparably only mild effects on proliferation (S2D-E), therefore, we focused on this stage of myogenesis.

Our focus on differentiation over activation is further supported by multiple lines of evidence. First, analysis of publicly available transcriptome datasets reveals that Lmod1 mRNA levels actually decrease upon Muscle Stem Cell (MuSC) activation, suggesting its primary role is not during this initial phase. We added this data for clarification to Figure S1B. This aligns perfectly with our in vivo data from cardiotoxin-induced muscle regeneration, where abundance of LMOD1 protein peaks at days 4-7 post-injury — a time point coinciding with new myofiber formation and maturation — rather than during the initial activation and proliferation phase (days 1-3) (Figure 4I).

Given this strong evidence pointing to a primary role for LMOD1 during the later stages of differentiation, we believe our current analyses are the most relevant. While single EDL fiber cultures are valuable for studying the quiescence-to-activation transition, they would not provide significant additional insight into the specific differentiation-centric mechanism we are investigating here. We are confident that our chosen readouts appropriately address Lmod1's function in the differentiation of myoblasts and formation of myotubes.

(2) In supplementary Figure 2E, the distinction between Hoechst-positive cells and total cell counts is unclear. The authors should clarify why Hoechst-positive cells increase and relabel "reserve cells," as the term is confusing without reading the legend.

We thank the reviewer for pointing out the confusion regarding the naming of the cell populations and the increase in Hoechst-positive cells. We have now modified this and revised the terminology used in Figure S2E to improve clarity. Specifically, we have relabeled "reserve cells" as "non-proliferating myoblasts (Ki67-/Hoechst+)" to describe these cells more accurately without requiring the legend for interpretation. Regarding the increase in Hoechst-positive cells, we observed a slight (26%) but significant decrease in the number of proliferating myoblasts (Ki67+/Hoechst+) (Figures S2D and S2E). The relative increase in non-proliferating (Ki67-/Hoechst+) cells is a consequence of the significant reduction in the number of proliferating cells (Ki67+/Hoechst+) cells. Importantly, the total cell count (sum of Ki67-/Hoechst+) and (Ki67+/Hoechst+) remained stable. This has been clarified in the revised figure legend and main text as follows:

“This was accompanied by a proportional increase in non-proliferating myoblasts (Ki67-/Hoechst+), while the total Hoechst-positive cell count (Ki67+/Hoechst+ and Ki67-/Hoechst+) remained unchanged (Figure S2E).”

(3) The specificity of Lmod1 and Sirt1 immunostaining needs validation using siRNA-treated samples, especially as these data form the basis of the mechanistic conclusions.

We have validated the specificity of the LMOD1 antibody using multiple approaches. Specifically, we performed immunofluorescence and immunoblotting on Lmod1 siRNA-transfected samples, where we observed a significant reduction in the Lmod1 protein signal compared to control conditions (see manuscript data from Figure S2G).

Additionally, LMOD1 overexpression experiments demonstrated a corresponding increase in the signal for LMOD1 using immunofluorescence analyses, confirming the specificity of the antibody for detecting LMOD1.

For the reviewers’ interest, we add Author response image 1:

Author response image 1.

Specificity of antibodies detecting LMOD1. Representative immunofluorescence images of LMOD1 in primary myoblast cultures following siLmod1 knockdown, LMOD1 overexpression, or controls transfected with a non-targeting siRNA (siCtrl) after one day of differentiation. LMOD1 (purple), SIRT1 (yellow), and nuclei (Hoechst, blue). Scale bar: 10 µm.

For the SIRT1 antibody used in our immunostaining, the specificity was validated by transfecting primary myoblasts with siRNA targeting Sirt1 and performing immunoblot analyses (Figure S5A). These showed a significant reduction in SIRT1 protein levels, confirming both the effectiveness of the siRNA and, critically, the antibody's ability to specifically recognize and detect SIRT1 protein. Furthermore, the same SIRT1 antibody was utilized in our nuclear-cytoplasmic fractionation experiments (Figure S4C), and its ability to detect SIRT1 in the expected subcellular compartments further supports its specific binding to SIRT1. While direct immunofluorescence on Sirt1 siRNA-transfected samples was not performed, the robust demonstration of the antibody's specificity for Sirt1 protein via immunoblotting (i.e., correct molecular weight band, significantly reduced by Sirt1 siRNA) and its distribution in subcellular fractions, which is fully consistent with the localization immunostaining performed at the same time points (compare Figure S4C and 5A), provide strong evidence on the antibody’s specificity, also in immunofluorescence experiments.

(4) The authors must test the effect of Lmod1 siRNA on Sirt1 localization, as only overexpression experiments are shown

We carefully considered performing this experiment. However, the knockdown of Lmod1 significantly impairs myogenic differentiation, a crucial cellular process that itself can influence protein localization. Consequently, if SIRT1 localization would be altered following knockdown of Lmod1, it would be challenging to disentangle whether this was a direct result of LMOD1 absence impacting SIRT1 trafficking or an indirect consequence of the cells failing to differentiate properly. This would make it difficult to draw clear conclusions regarding a direct causal link between LMOD1 and SIRT1 localization from such an experiment. Therefore, we focused on overexpression experiments, where we could demonstrate that altering LMOD1 levels is sufficient to affect SIRT1 localization. Our nuclear-cytoplasmic fractionation experiments clearly show that LMOD1 overexpression leads to changes in SIRT1 distribution (Figure 5H-K). These findings provide evidence that LMOD1 can directly modulate SIRT1 localization, supporting our mechanistic conclusions.

(5) In Figure S3, the biotin signal in LMOD2 samples appears weak. The authors need to address whether comparing LMOD1 and LMOD2 is valid given the apparent difference in reaction efficiency. It would also help to highlight where Sirt1 falls on the volcano plot in S3B.

We agree that the overall biotin signal on the streptavidin blot for the LMOD2-BirA* sample appears weaker than for LMOD1-BirA*. To provide a more direct comparison of the bait proteins themselves, we have now added a bar graph to the revised Figure S3D, which quantifies the relative abundance of LMOD1 and LMOD2 bait proteins in the pull down experiments. This analysis shows that the levels of LMOD1-BirA* and LMOD2-BirA* were comparable in our BioID samples. Furthermore, the validity of the LMOD2 BioID experiment is strongly supported by the identification of several known LMOD1 and LMOD2 interaction partners. As shown in the dataset, well-established interactors such as TMOD1, TPM3, and TMOD3 were identified, with some even showing stronger enrichment with LMOD2 than with LMOD1. This confirms that the biotinylation reaction was efficient enough to capture proximal proteins for both baits.

Regarding SIRT1, we have now highlighted in yellow its position on the volcano plot in the revised Figure S3E. As can be seen, SIRT1 was identified in the LMOD1-BirA sample and showed enrichment. We believe these clarifications, along with the additional expression data and the successful identification of known interactors, confirm the validity of our comparative BioID analysis.

(6) The immunostaining data suggest that Lmod1 remains cytoplasmic throughout differentiation, whereas Sirt1 shows transient cytoplasmic localization at day 1 of differentiation. The authors should explain why Sirt1 is not constantly sequestered if Lmod1's cytoplasmic localization is consistent. It is also unclear whether day 1 is the key time point for Lmod1 function, as its precise role during myogenesis remains ambiguous.

We thank the reviewer for this comment. We have no data explaining why SIRT1 is not constantly sequestered while LMOD1 remains consistently cytoplasmic. We can only speculate that the transient cytoplasmic localization of SIRT1 may be linked to the availability and functional role of LMOD1 throughout the differentiation process. While LMOD1 is present at low levels in proliferating primary myoblasts, its expression increases upon the initiation of differentiation (Figure 2A). Initially, during the early stages of differentiation, LMOD1 may not be required for actin nucleation as the major remodeling of the cytoskeleton has not yet begun. During this phase, LMOD1 might have the capacity to sequester SIRT1 in the cytoplasm.

However, as differentiation progresses and morphological changes take place, LMOD1 may switch its functional role to actin nucleation, thereby releasing SIRT1. This transition could explain why SIRT1 is free to localize transiently to the cytoplasm, particularly at day 1, when cytoskeletal remodeling is beginning but not yet fully established.

Additionally, as LMOD1 and SIRT1 are known to colocalize in the nucleus, they may exit the nucleus together. Once in the cytoplasm, LMOD1 may become engaged in actin nucleation, allowing SIRT1 to function independently, which could explain the transient nature of SIRT1’s cytoplasmic localization.

We have acknowledged this gap in our understanding in the discussion of the revised manuscript:

“Our immunostaining data show that while LMOD1 is consistently cytoplasmic, its partner SIRT1 is only transiently localized in the cytoplasm. This suggests that their interaction is dynamically regulated. We hypothesize that the function of LMOD1 is determined by the changing availability of its binding partners during differentiation. During the initial phase, LMOD1 may primarily function to sequester SIRT1, a key regulator of myogenic genes. As differentiation proceeds, the increased expression of cytoskeletal components, such as its canonical partners TMODs and TPMs, likely shifts the function of LMOD1 towards its role in actin nucleation. This molecular switch, potentially driven by a change in the interactome of LMOD1, could then result in the release of SIRT1 from the cytoplasm. Such a mechanism may coordinate transcriptional regulation with cytoskeletal remodeling during myoblast differentiation.”

(7) The introduction does not sufficiently establish the motivation or knowledge gap this work aims to address. Instead, it reads like a narration of disparate topics in a single paragraph. The authors should clarify the statement in line 150, "since this protein has been...,".

We thank the reviewer for requesting clarification regarding our focus on LMOD1 (Introduction and Line 150 in the original submission). In the revised manuscript, we shortened the introduction and more clearly emphasized the motivation of our study:

“Although these mechanisms contribute to remodeling the cellular architecture of MuSCs, a comprehensive understanding of the temporal dynamics of proteome remodeling during differentiation remains lacking. To address this knowledge gap, we performed an unbiased proteomic analysis of the early stages of myogenic differentiation to identify previously unrecognized proteins involved in this process and to examine how they functionally interact with established regulatory pathways.”

Our decision to focus on LMOD1 was driven by its significant upregulation in our temporal proteome dataset, together with its previously uncharacterized role in primary myoblasts. Furthermore, to strengthen the interpretation of LMOD1’s role, particularly in the context of aging, we have integrated a new analysis of published transcriptomic datasets. This can be found in the main text as follows:

“Surprisingly, we detected LMOD1 in freshly isolated muscle stem cells (MuSCs), but not LMOD2. Additionally, we observed that the protein levels of LMOD1 increased in MuSCs isolated from older mice (Figure 2C and Figure S1B). We further analyzed published transcriptomic data sets that describe changes between young and old MuSCs in both quiescent and activated states in young and old animals (Liu et al. 2013; Lukjanenko et al. 2016). In these analyzed transcriptomic data sets, Lmod1 was found to be significantly downregulated during the activation of MuSCs in both young and old mice (see Figure S1B).

To assess the in vivo relevance of our finding, we queried two proteomic datasets of freshly isolated MuSCs and four different skeletal muscles (gastrocnemius, G; soleus, S; tibialis anterior, TA; extensor digitorum longus, EDL) (Schüler et al. 2021). We found LMOD2 to be the most abundant leiomodin protein in whole skeletal muscle, consistent with data from (Tsukada et al. 2010; Nworu et al. 2015; Kiss et al. 2020), while the overall abundance of LMOD1 was lower since this protein has been mainly associated with smooth muscle cells (Nanda and Miano 2012; Conley et al. 2001; Nanda et al. 2018) (Figure 2B).”

Overall, while the identification of Lmod1 as a pro-myogenic factor is convincing, the mechanistic insights are insufficient, and the manuscript would benefit from addressing these concerns.

We thank the reviewer for their constructive criticism. In the revised manuscript, we have strengthened our mechanistic insights and the validation of our findings by implementing the suggestions of the reviewers and including new experimental data to address their concerns.

Reviewer #2 (Public review):

Summary:

In this manuscript, the authors identify Leiomodin-1 (LMOD1) as a key regulator of early myogenic differentiation, demonstrating its interaction with SIRT1 to influence SIRT1's cellular localization and gene expression. The authors propose that LMOD1 translocates SIRT1 from the nucleus to the cytoplasm to permit the expression of myogenic differentiation genes such as MYOD or Myogenin.

Strengths:

A major strength of this work lies in the robust temporal resolution achieved through a time-course mass spectrometry analysis of in vitro muscle differentiation. This provides novel insights into the dynamic process of myogenic differentiation, often under-explored in terms of temporal progression. The authors provide a strong mechanistic case for how LMOD1 exerts its role in muscle differentiation which opens avenues to modulate.

We thank the reviewer for the positive feedback on our manuscript and the insightful comments which helped to improve the manuscript!

Weaknesses:

One limitation of the study is the in vivo data. Although the authors do translate their findings in vivo for LMOD1 localization and expression, the cross-sectional imaging is not highly convincing. Longitudinal cuts or isolated fibers could have been more useful specimens to answer these questions. Moreover, the authors do not assess their in vitro SIRT1 findings in vivo. A few key experiments in regenerating or aged mice would strengthen the mechanistic insight of the findings.

We agree that longitudinal cuts and isolated fibers can provide excellent morphological detail for specific questions. However, for our primary objective in this study, which was to assess the temporal expression and localization of LMOD1 across the tissue during the regeneration process, we decided that cross-sectional analysis provided the most robust and reliable overview. Cross-sectional imaging effectively captures the spatial distribution of LMOD1 across multiple myofibers and their surrounding microenvironment, simultaneously assessing the whole cross-sectional area. By using this approach, we were able to evaluate the broader tissue architecture and cellular context, which was essential for understanding the dynamic changes occurring during regeneration. We were also able to investigate all myofibers of a muscle, and not only a small proportion, which we would analyze with longitudinal sections and isolated myofibers. Therefore, we continued using cross-sections for further analyses.

We fully agree with the reviewer that validating our in vitro SIRT1 findings in an in vivo context is an essential next step. To address this, we performed additional analyses on our existing regenerating muscle samples and incorporated new immunostainings for SIRT1 and PAX7 into the regeneration time-course (now shown in revised Figure 4I), providing further in vivo support for our proposed mechanism. We focused specifically on cross-sections collected at day 5 post-injury, a time point selected based on the peak in LMOD1 expression, to assess whether SIRT1 levels increase in parallel with LMOD1 during regeneration. Notably, SIRT1 abundance is elevated at day 5 post-injury, underscoring its involvement in early myogenic differentiation. This conclusion is further supported by the localization of SIRT1 within mononucleated cells and newly formed myofibers at this stage of regeneration.

Finally, we agree that further mechanistic studies in vivo would be highly valuable. While we were able to address SIRT1 dynamics in our regeneration model as suggested, an aged mouse cohort was unfortunately not available to us for this kind of study. Furthermore, more extensive in vivo experiments, such as those involving genetic manipulation, were beyond the scope of the current study, partly due to constraints related to animal welfare regulations and our approved experimental protocols.

Discussion:

Overall, the study emphasizes the importance of understanding the temporal dynamics of molecular players during myogenic differentiation and provides valuable proteomic data that will benefit the field. Future studies should explore whether LMOD1 modulates the nuclear-cytoplasmic shuttling of other transcription factors during muscle development and how these processes are mechanistically achieved. Investigating whether LMOD1 can be therapeutically targeted to enhance muscle regeneration in contexts such as exercise, aging, and disease will be critical for translational applications. Additionally, elucidating the interplay among LMOD1, LMOD2, and LMOD3 could uncover broader implications for actin cytoskeletal regulation in muscle biology.

We thank the reviewer for this excellent suggestion for future analyses. We have included these important considerations and future avenues in the Discussion of the revised manuscript:

“Our immunostaining data show that while LMOD1 is consistently cytoplasmic, its partner SIRT1 is only transiently localized in the cytoplasm. This suggests that their interaction is dynamically regulated. We hypothesize that the function of LMOD1 is determined by the changing availability of its binding partners during differentiation. During the initial phase, LMOD1 may primarily function to sequester SIRT1, a key regulator of myogenic genes. As differentiation proceeds, the increased expression of cytoskeletal components, such as its canonical partners TMODs and TPMs, likely shifts the function of LMOD1 towards its role in actin nucleation. This molecular switch, potentially driven by a change in the interactome of LMOD1, could then result in the release of SIRT1 from the cytoplasm. Such a mechanism may coordinate transcriptional regulation with cytoskeletal remodeling during myoblast differentiation.”

“Moreover, delineating the functional specialization and potential redundancy among leiomodin proteins represents an important next step. Our data indicate that LMOD1 primarily regulates early myogenic differentiation (Figure 3). In contrast, the lack of an early functional phenotype upon LMOD2 depletion, together with its upregulation at later stages (Figure S2A), suggests a temporal shift in regulatory control. Accordingly, a systematic comparative analysis of LMOD1, LMOD2, and LMOD3 will be required to elucidate their distinct roles in actin cytoskeleton regulation across the myogenic program, particularly with respect to myofibril maturation and maintenance.”

Recommendations for the authors:

Reviewer #2 (Recommendations for the authors):

Major Changes:

(1) In Vivo Data on SIRT1:

The inclusion of in vivo data on SIRT1 localization and expression would significantly strengthen the manuscript. Similar staining techniques used for LMOD1 could be applied to SIRT1. Additionally, imaging muscle specimens such as longitudinal sections or isolated myofibers would provide clearer insights into SIRT1's spatial distribution and improve upon the less convincing cross-sectional images currently presented (Figure 2).

We fully agree that providing in vivo data on SIRT1 localization and expression is a crucial step to support our in vitro findings. Following the reviewer's suggestion, we have performed new experiments on muscle regeneration samples using the analyses of cross-sections as done for the analysis of LMOD1 localization. Specifically, we performed immunostaining for SIRT1 on cross-sections from muscle samples collected at day 5 post-injury, a time point selected based on the observed peak in LMOD1 expression. These new data (now included in revised Figure 4I) allowed us to assess whether SIRT1 levels increase during regeneration in parallel with an increase in LMOD1 abundance.

Regarding the suggestion to use longitudinal sections or isolated myofibers, we agree that these preparations offer excellent answers for certain questions. For the primary goal of our study, to assess the temporal expression changes across the entire regenerating tissue at different time points, we found that cross-sections provided the most comprehensive and robust overview and therefore did not use longitudinal sections or isolated myofibers. 

Performing additional animal experiments to obtain these specific preparations was beyond the scope of the current study and subject to constraints from our approved animal welfare protocols.

(2) Morphology of siLmod1 Cells:

The morphology of siLmod1-treated cells in vitro (Figure 3) raises concerns. Assessing cell viability or cell death in these experiments would help ensure that differences are not due to dead or unhealthy cells being quantified. There is also a notable discrepancy between the control panels in Figures 3C and 3H compared to the experimental conditions in 3F and 3K, particularly in terms of cell length and morphology. These inconsistencies should be addressed or clarified.

We acknowledge the visual discrepancies in cell morphology noted by the reviewer (e.g., between Figures 3C/3H and 3F/3K). These differences can be attributed to biological variability between primary myoblast cultures isolated from different mice. Such variability includes differences in myogenic potential and the fact that cells are not synchronized, leading to variations in differentiation efficiency, baseline morphology, and cell length across cultures (Cornelison 2008; Vaughan and Lamia 2019). To account for this, we decided to use n=6 biological replicates, i.e., primary myoblast cultures isolated from 6 different mice, for immunofluorescence analysis, ensuring robust quantitative data. Furthermore, we confirmed that this phenotype was not an artifact of culture conditions, as we consistently observed the same effect of Lmod1 knockdown independently of the passage number of the myoblasts or the donor mouse.

To address the concerns that morphological changes in siLmod1-treated cells might reflect cell death, we performed a TUNEL assay (transfection at day 1, analysis at day 3 of differentiation). This revealed no significant increase in TUNEL-positive (apoptotic) cells in siLmod1- (or siSirt1-) transfected samples versus siCtrl-transfected cells. These new data have been added to the revised manuscript as Supplementary Figure S2I. The TUNEL data indicate that the observed morphological changes upon knockdown of Lmod1 are not due to induced cell death. Supported by these results, our interpretation is that knockdown of Lmod1 impairs or arrests differentiation rather than causing cell death. Furthermore, our quantification of different cell populations showed shifts indicative of impaired differentiation (e.g., accumulation of cells at earlier stages) without exhibiting significant loss in cell numbers. For example, the numbers of myogenin+/MHC- and myogenin+/MHC+ cell populations, and differentiated myotubes, were not significantly reduced after transfection with siLmod1. A slight, not significant trend towards fewer non-proliferating myoblasts/reserve cells characterized by the expression of Myogenin-/MHC-Hoechst+ (Figure S2H) was noted. Overall, cells appeared to be 'stuck' in differentiation, consistent with the role of Lmod1 in impairing differentiation but not causing cell death. We have further clarified this aspect in the revised manuscript.

(3) LMOD1 and SIRT1 Interaction in Myogenic Cells:

Strengthening the connection between LMOD1 and SIRT1 within the myogenic system would enhance the manuscript. Could proximity ligation assays (PLA) be performed in myogenic cells, as was done in HEK293T cells? Additionally, investigating whether SIRT1 remains in the nucleus upon LMOD1 knockdown using siRNA would provide mechanistic insight into their interaction during myogenic differentiation.

We would like to clarify that the Proximity Ligation Assays (PLA) shown in Figure 4H were indeed performed in primary myoblasts, confirming the LMOD1-SIRT1 interaction directly in a myogenic context. We have modified the text to clarify that primary myoblasts were used for the PLA assays.

Minor Points:

(1) Was Lmod1 knockdown confirmed in vivo?

To target Lmod1 in Muscle Stem Cells (MuSCs) in vivo, we utilized self-delivering Accell siRNAs. This delivery system has been previously validated and shown to be highly effective for targeting MuSCs in regenerating muscle (Bentzinger et al., Cell Stem Cell, 2013).

While this is an established method for delivery, confirming knockdown specifically within the rare MuSC population is technically challenging using bulk tissue analysis, as the target signal is diluted by numerous other cell types. 

Therefore, to ensure the efficacy of our specific siRNA, we performed in vitro validation. For the reviewers' interest, we add Author response image 2 showing the efficiency of the respective siRNAs:

Author response image 2.

Knockdown efficiency of siRNAs targeting Lmod1 and Lmod2 following using the same self-delivering siRNA in proliferating primary myoblasts as used in in vivo experiments. Self-delivering Accell siRNA was added to primary myoblasts cultured in low serum media for 48 hours. Relative mRNA expression levels of Lmod1 and Lmod2 were measured after self-delivering Accell siRNA transfection targeting either Lmod1 (siLmod1) or Lmod2 (siLmod2). Expression levels were compared to control siRNA-transfected cells (siCtrl) and normalized to Gapdh expression.

Based on the documented efficacy of this delivery system from prior literature and our own validation of the specific siRNAs used here, we are confident in the knockdown efficiency of the respective siRNAs. We decided not to perform additional animal experiments due to animal welfare considerations.

(2) Some of the western blot bands do not appear to match the expected patterns for the tested proteins compared to controls (e.g., Figure S2J, S4C). Ensure that these are accurately labeled and include the entire membrane for transparency and reproducibility.

Regarding Figure S2J, we agree that the presentation could be confusing to the reader. The blot shows LMOD1 and LMOD2 knockdown, while the bar plot quantifies only the change in LMOD2 levels. We have now revised the figure legend to explicitly state this. We hope this makes the presentation of our data clearer.

For Figure S4C, we believe the concern about 'patterns' relates to loading variability. In this experiment, we manually counted the nuclei before lysis to ensure that each nuclear fraction started with an equal amount of material. We then loaded the cytoplasmic fractions in proportion to these counts. The purity of the fractions was additionally confirmed using nuclear (H4) and cytoplasmic (ALDOA) markers. As stated in the figure, the nuclear/cytoplasmic ratio of LMOD1 or SIRT1 was normalized across the entire lane of the Ponceau S staining, which we have now clarified in the relevant figure legends.

Finally, regarding transparency, the presented immunoblot images are representative crops, which is standard practice for clarity. We are committed to reproducibility and will provide full, uncropped scans of all blots in the final version of the manuscript, in line with eLife publishing guidelines. 

(3) Figure S1B appears to reuse images from Figure 2D (rotated). Verify that this is acceptable for the journal's guidelines, and if necessary, provide additional justification or clarification.

We acknowledge that the image presented in Figure S1B was accidentally reused as a representative example in Figure 2D. To address this and prevent any potential redundancy or confusion, we have revised Figure S1B by replacing the duplicated image with a different, representative example from our dataset. The updated figure now contains unique image data, and we believe this revision fully resolves the concern.

(4) Ensure consistent scale bars across images, particularly in Figures 3C and 3H, where discrepancies might affect interpretation.

We thank the reviewer for pointing this out, we have now standardized all scale bars throughout the manuscript to ensure consistency. All immunofluorescence images of cultured cells (including Fig 3C and 3H) now have a 50 µm scale bar, and all tissue cross-sections have a 100 µm scale bar. This change has been implemented in the revised figures.

Reviewer #3 (Public review):

Summary:

In this manuscript, the investigators identified LMOD1 as one of a subset of cytoskeletal proteins whose levels increase in the early stages of myogenic differentiation. Lmod1 is understudied in striated muscle and in particular in myogenic differentiation. Thus, this is an important study. It is also a very thorough study - with perhaps even too much data presented. Importantly, the investigators observed that LMOD1 appears to be important for skeletal regeneration, and myogenic differentiation and that it interacts with SIRT1. Both primary myoblast differentiation and skeletal muscle regeneration were studied. Rescue experiments confirmed these observations: SIRT1 can rescue perturbations of myogenic differentiation as a result of LMOD1 knockdown.

Strengths:

Particular strengths include: important topic, the use of primary skeletal cultures, the use of both cell culture and in vivo approaches, careful biomarker analysis of primary mouse myoblast differentiation, the use of two methods to probe the function of the Lmod1/SIRT1 pathway via using depletion approaches and inhibitors, and generation of six independent myoblast cultures. Results support their conclusions.

We thank the reviewer for the positive assessment of our work and the helpful comments for improving our manuscript.

Weaknesses:

(1) Figure 1. Images of cells in Figure 1A are too small to be meaningful (especially in comparison to the other data presented in this figure). Perhaps the authors could make graphs smaller?

We have adjusted the size of the images across all figure panels to ensure better visibility and clarity. We hope these adjustments improve the presentation of the data.

(2) Line 148 "We found LMOD2 to be the most abundant Lmod in the whole skeletal muscle." This is confusing since most, if not all, prior studies have shown that Lmod3 is the predominant isoform in skeletal muscle. The two papers that are cited are incorrectly cited. Clarification to resolve this discrepancy is needed.

We acknowledge that LMOD2 and LMOD3 are predominantly expressed in skeletal and cardiac muscles (Tsukada et al. 2010; Nworu et al. 2015), www.proteinatlas.org) and LMOD3’s transcription is directly regulated by MRTF/SRF and MEF2 to coordinate sarcomeric assembly (Cenik et al. 2015). However, our statement refers specifically to the analysis of the proteomic datasets from freshly isolated MuSCs and four distinct skeletal muscles (G, S, TA, EDL) generated by Schüler et al. 2021. Crucially, LMOD3 was not detected in the quantitative mass spectrometry data for the EDL, G, S, or TA muscle samples analyzed in this specific study. In the context of this particular dataset, LMOD2 was the most highly abundant Leiomodin isoform detected in the whole skeletal muscle samples. This finding suggests a differential expression and function between LMOD isoforms depending on the muscle type and/or developmental/regenerative state. We have revised and corrected this clarification in the manuscript, including correcting the initial citations.

(3) Figure 2. Immunoflorescence (IF) panels are too small to be meaningful. Perhaps the graphs could be made smaller and more space allocated for the IF panels? This issue is apparent for just about all IF panels - they are simply too small to be meaningful. Additionally, in many of the immunofluorescence figures, the colors that were used make it difficult to discern the stained cellular structures. For example, in Figure S1, orange and purple are used - they do not stand out as well as other colors that are more commonly used.

We agree that the IF panels were too small for optimal interpretation and have adjusted them in Figure 2 and throughout the manuscript. Regarding the color choices, we appreciate the reviewer's comments. Our initial selection (e.g., orange and purple in Figure S1) was intended to enhance accessibility for individuals with common color vision deficiencies, including red-green color blindness. However, we acknowledge the reviewer's point that these combinations provided insufficient contrast for discerning cellular structures. Therefore, we have revised the color schemes to use green, red, and blue, which should offer improved contrast.

(4) There is huge variability in many experiments presented - as such, more samples appear to be required to allow for meaningful data to be obtained. For example, Figure S2. Many experimental groups, only have 3 samples - this is highly problematic - I would estimate that 5-6 would be the minimum.

We thank the reviewer for the comment regarding experimental variability and sample size. In our study, n=3 biological replicates, i.e., independent primary cell cultures obtained from different mice, were primarily used for immunoblots. We acknowledge that variability can be observed between distinct primary cell cultures due to factors such as inherent differences in myogenic potential, cell cycle state (as cells were not synchronized), and passage number. Importantly, despite this inter-sample variation, the investigated phenotypes showed consistent trends across biological replicates. Rather than increasing the number of replicates for immunoblots, we opted for validating our key findings using independent approaches with a higher number of replicates. For instance, qRT-PCR analyses (to confirm knockdown efficiency) and immunofluorescence analyses were mostly performed using five to six independent myoblast cultures (biological replicates).

(5) Ponceau S staining is often used as a loading control in this manuscript for western blots. The area/molecular weight range actually used should be specified. Not clear why in some experiments GAPDH staining is used, in other experiments Ponceau S staining is used, and in some, both are used. In some experiments, the variability of total protein loaded from lane to lane is disconcerting. For example, in Figure S4C there appears to be more than normal variability. Can the protein assay be redone and samples run again?

We have clarified in the relevant figure legends that Ponceau S normalization, when used, was based on the quantification of the entire lane. Our standard loading control is GAPDH. We used Ponceau S for normalization only when GAPDH was deemed unsuitable, e.g., in nuclear-cytoplasmic fractionation experiments where GAPDH is not present in all fractions.

Concerning the variability observed in Figure S4C, we manually counted the nuclei before lysis to ensure that each nuclear fraction started with an equal amount of material. We then loaded the cytoplasmic fractions in proportion to these counts. The purity of the fractions was additionally confirmed using nuclear (H4) and cytoplasmic (ALDOA) markers. The nuclear/cytoplasmic ratio of LMOD1 or SIRT1 was normalized across the entire lane of the Ponceau S staining, which we have now clarified in the relevant figure legends.

(6) Figure S3 - Lmod3 is included in the figure but no mention of it occurs in the title of the figure and/or legend.

We wish to clarify that the protein identified in Figure S3 is TMOD3 (Tropomodulin 3), not LMOD3. TMOD3 is a known pointed-end capping protein regulating the actin filament nucleation process together with LMODs (Fowler and Dominguez 2017; Boczkowska et al. 2015), so its presence in our dataset was expected and helps validate our results.

(7) Abstract, line 25. "overexpression accelerates and improves the formation of myotubes". This is a confusing sentence. How is it improving the formation? A little more information about how they are different than developing myotubes in normal/healthy muscles would be helpful.

We thank the reviewer for the comment. To clarify, we have revised the sentence in line 25 to "improves the initiation of myotube formation." This change reflects our observation that overexpression of LMOD1 leads to a more rapid onset of myotube formation, as evidenced by earlier expression of differentiation markers and accelerated fusion of myoblasts into myotubes compared to GFP overexpression myoblast cell line. These findings suggest that LMOD1 overexpression enhances the efficiency of the early stages of differentiation and fusion, thereby contributing to improved initiation of myotube formation.

(8) It is impossible from the IF figures presented to determine where Lmod1 localizes in the myocytes. Information on its subcellular localization is important. Does it localize with Lmod2 and Lmod3 at thin filament pointed ends?

Several publications suggest that LMODs are involved in actin nucleation and interact with TMODs at the thin filament pointed ends (Boczkowska et al. 2015; Fowler and Dominguez 2017; Fowler, Greenfield, and Moyer 2003; Tsukada et al. 2010; Rao, Madasu, and Dominguez 2014). We performed F-actin (Phalloidin) staining together with LMOD1 staining and observed possible co-localization (see Author response image 3). Specifically, we noted an accumulation of LMOD1 at the ends of myocytes, indicating that LMOD1 might play a role in the elongation and guidance of myotube differentiation. For the reviewer’s interest, we include Author response image 3 as it was not part of the original manuscript. While performing subcellular localization stainings, we added the F-actin/Phalloidin staining to explore potential interactions, but this aspect was not further investigated in the current study.

Author response image 3.

Co-staining of LMOD1 and Phalloidin in differentiating myocytes.Example image showing immunofluorescence staining of LMOD1 (purple) and F-actin (green; Phalloidin) in differentiating primary myocytes. LMOD1 appears to accumulate at the ends of elongated myocytes and co-localizes with actin structures (highlighted in boxes), suggesting a potential role in myotube elongation and guidance during differentiation.

Our study focused on a distinct role for LMOD1, independent from its function in actin filament nucleation, and we therefore did not pursue further co-localization staining with LMOD2 or LMOD3. We recognize the potential importance of exploring these interactions and their relevance to thin filament organization in skeletal muscle. However, although this was beyond the scope of our current work, we will investigate this aspect in the future.

References

Boczkowska, Malgorzata, Grzegorz Rebowski, Elena Kremneva, Pekka Lappalainen, and Roberto Dominguez. 2015. “How Leiomodin and Tropomodulin Use a Common Fold for Different Actin Assembly Functions.” Nature Communications 6 (1): 8314.

Cenik, Bercin K., Ankit Garg, John R. McAnally, John M. Shelton, James A. Richardson, Rhonda Bassel-Duby, Eric N. Olson, and Ning Liu. 2015. “Severe Myopathy in Mice Lacking the MEF2/SRF-Dependent Gene Leiomodin-3.” The Journal of Clinical Investigation 125 (4): 1569–78.

Cornelison, D. D. W. 2008. “Context Matters: In Vivo and in Vitro Influences on Muscle Satellite Cell Activity.” Journal of Cellular Biochemistry 105 (3): 663–69.

Fowler, Velia M., and Roberto Dominguez. 2017. “Tropomodulins and Leiomodins: Actin Pointed End Caps and Nucleators in Muscles.” Biophysical Journal 112 (9): 1742–60.

Fowler, Velia M., Norma J. Greenfield, and Jeannette Moyer. 2003. “Tropomodulin Contains Two Actin Filament Pointed End-Capping Domains.” The Journal of Biological Chemistry 278 (41): 40000–9.

Liu, Ling, Tom H. Cheung, Gregory W. Charville, Bernadette Marie Ceniza Hurgo, Tripp Leavitt, Johnathan Shih, Anne Brunet, and Thomas A. Rando. 2013. “Chromatin Modifications as Determinants of Muscle Stem Cell Quiescence and Chronological Aging.” Cell Reports 4 (1): 189–204.

Lukjanenko, Laura, M. Juliane Jung, Nagabhooshan Hegde, Claire Perruisseau-Carrier, Eugenia Migliavacca, Michelle Rozo, Sonia Karaz, et al. 2016. “Loss of Fibronectin from the Aged Stem Cell Niche Affects the Regenerative Capacity of Skeletal Muscle in Mice.” Nature Medicine 22 (8): 897–905.

Nworu, Chinedu U., Robert Kraft, Daniel C. Schnurr, Carol C. Gregorio, and Paul A. Krieg. 2015. “Leiomodin 3 and Tropomodulin 4 Have Overlapping Functions during Skeletal Myofibrillogenesis.” Journal of Cell Science 128 (2): 239–50.

Rao, Jampani Nageswara, Yadaiah Madasu, and Roberto Dominguez. 2014. “Mechanism of Actin Filament Pointed-End Capping by Tropomodulin.” Science 345 (6195): 463–67.

Schüler, Svenja C., Joanna M. Kirkpatrick, Manuel Schmidt, Deolinda Santinha, Philipp Koch, Simone Di Sanzo, Emilio Cirri, Martin Hemberg, Alessandro Ori, and Julia von Maltzahn. 2021. “Extensive Remodeling of the Extracellular Matrix during Aging Contributes to Age-Dependent Impairments of Muscle Stem Cell Functionality.” Cell Reports 35 (10): 109223.

Tsukada, Takehiro, Christopher T. Pappas, Natalia Moroz, Parker B. Antin, Alla S. Kostyukova, and Carol C. Gregorio. 2010. “Leiomodin-2 Is an Antagonist of Tropomodulin-1 at the Pointed End of the Thin Filaments in Cardiac Muscle.” Journal of Cell Science 123 (Pt 18): 3136–45.

Vaughan, Megan, and Katja A. Lamia. 2019. “Isolation and Differentiation of Primary Myoblasts from Mouse Skeletal Muscle Explants.” Journal of Visualized Experiments: JoVE, no. 152 (October). https://doi.org/10.3791/60310.

eLife Assessment

This elegant study presents a valuable approach to probing the structural features of the full-length human Hv1 channel as a purified protein, supported by rigorous biochemical assays and spectral FRET analysis, which will interest biophysicists and physiologists studying Hv1 and other ion channels. Overall, the work introduces an interesting labeling strategy and provides methodological observations that are of value in investigating hHV1. However, the analysis appears incomplete, requiring additional structural interpretation and mechanistic insight.

Reviewer #1 (Public review):

In this study, the noncanonical amino acid acridon-2-ylalanine (Acd) was inserted at various positions within the human Hv1 protein using a genetic code expansion approach. The purified mutants with incorporated fluorophore were shown to be functional using a proton flux assay in proteoliposomes. FRET between native tryptophan and tyrosine residues and Acd were quantified using spectral FRET analysis. Predicted FRET efficiencies calculated from an AlphaFold model of the Hv1 dimer were compared to the corresponding experimental values. Spectral FRET analysis was also used to test whether structural rearrangements caused by Zn2+, a well-known Hv1 inhibitor, could be detected. The experimental data provide a good validation of the approach, but further expansion of the analysis will be necessary to differentiate between intra- and intersubunit structural features.

Interestingly, the observed rearrangements induced by Zn2+ were not limited to the protein region proximal to the extracellular binding site but extended to the intracellular side of the channel. This finding agrees with previous studies showing that some extracellular Hv1 inhibitors, such as Zn2+ or AGAP/W38F, can cause long-range structural changes propagating to the intracellular vestibule of the channel (De La Rosa et al. J. Gen. Physiol. 2018, and Tang et al. Brit J. Pharm 2020). The authors should consider adding these references.

Since one of the main goals of this work was to validate Acd incorporation and the spectral FRET analysis approach to detect conformational changes in hHv1 in preparation for future studies, the authors should consider removing one subunit from their dimer model, recalculating FRET efficiencies for the monomer, and comparing the predicted values to the experimental FRET data. This comparison could support the idea that the reported FRET measurements can inform not only on intrasubunit structural features but also on subunit organization.

Reviewer #2 (Public review):

This manuscript by Carmona, Zagotta, and Gordon is generally well-written. It presents a crude and incomplete structural analysis of the voltage-gated proton channel based on measured FRET distances. The primary experimental approach is Förster Resonance Energy Transfer (FRET), using a fluorescent probe attached to a noncanonical amino acid. This strategy is advantageous because the noncanonical amino acid likely occupies less space than conventional labels, allowing more effective incorporation into the channel structure.

Fourteen individual positions within the channel were mutated for site-specific labeling, twelve of which yielded functional protein expression. These twelve labeling sites span discrete regions of the channel, including P1, P2, S0, S1, S2, S3, S4, and the dimer-connecting coiled-coil domain. FRET measurements are achieved using acridon-2-ylalanine (Acd) as the acceptor, with four tryptophan or four tyrosine residues per monomer serving as donors. In addition to estimating distances from FRET efficiency, the authors analyze full FRET spectra and investigate fluorescence lifetimes on the nanosecond timescale.

Despite these strengths, the manuscript does not provide a clear explanation of how channel structure changes during gating. While a discrepancy between AlphaFold structural predictions and the experimental measurements is noted, it remains unclear whether this mismatch arises from limitations of the model or from the experimental approach. No further structural analysis is presented to resolve this issue or to clarify the conformational states of the protein.

The manuscript successfully demonstrates that Acd can be incorporated at specific positions without abolishing channel function, and it is noteworthy that the reconstituted proteins function as voltage-activated proton channels in liposomes. The authors also report reversible zinc inhibition of the channel, suggesting that zinc induces structural changes in certain channel regions that can be reversed by EDTA chelation. However, this observation is not explored in sufficient depth to yield meaningful mechanistic insight.

Overall, while the study introduces an interesting labeling strategy and provides valuable methodological observations, the analysis appears incomplete. Additional structural interpretation and mechanistic insight are needed.

Major Points

(1) Tryptophan and tyrosine exhibit similar quantum yields, but their extinction coefficients differ substantially. Is this difference accounted for in your FRET analysis? Please clarify whether this would result in a stronger weighting of tryptophan compared to tyrosine.

(2) Is the fluorescence of acridon-2-ylalanine (Acd) pH-dependent? If so, could local pH variations within the channel environment influence the probe's photophysical properties and affect the measurements?

(3) Several constructs (e.g., K125Tag, Y134Tag, I217Tag, and Q233Tag) display two bands on SDS-PAGE rather than a single band. Could this indicate incomplete translation or premature termination at the introduced tag site? Please clarify.

(4) In Figure 5F, the comparison between predicted FRET values and experimentally determined ratio values appears largely uninformative. The discussion on page 9 suggests either an inaccurate structural model or insufficient quantification of protein dynamics. If the underlying cause cannot be distinguished, how do the authors propose to improve the structural model of hHV1 or better describe its conformational dynamics?

(5) Cu²⁺, Ru²⁺, and Ni²⁺ are presented as suitable FRET acceptors for Acd. Would Zn²⁺ also be expected to function as an acceptor in this context? If so, could structural information be derived from zinc binding independently of Trp/Tyr?

(6) The investigated structure is most likely dimeric. Previous studies report that zinc stabilizes interactions between hHV1 monomers more strongly than in the native dimeric state. Could this provide an explanation for the observed zinc-dependent effects? Additionally, do the detergent micelles used in this study predominantly contain monomers or dimers?

(7) hHV1 normally inserts into a phospholipid bilayer, as used in the reconstitution experiments. In contrast, detergent micelles may form monolayers rather than bilayers. Could the authors clarify the nature of the micelles used and discuss whether the protein is expected to adopt the same fold in a monolayer environment as in a bilayer?

eLife Assessment

This study addresses a fundamental question in glycobiology by elucidating how a single-site processive enzyme orchestrates the alternating addition of sugars to synthesize complex polysaccharides such as hyaluronan. The findings are compelling, providing a clear mechanistic framework supported by strong experimental validation. Major strengths include the integration of high-resolution structural data with rigorous biochemical analyses, resulting in a well-supported model of hyaluronan assembly.

Reviewer #1 (Public review):

Summary:

This revised manuscript describes critical intermediate reaction steps of a HA synthase at the molecular level; specifically, they examine the 2nd step, polymerization, adding GlcA to GlcNAc to form the initial disaccharide of the repeating HA structure. Unlike the vast majority of known glycosyltransferases, the viral HAS (a convenient proxy extrapolated to resemble the vertebrate forms) uses a single pocket to catalyze both monosaccharide transfer steps. The authors work illustrates the interactions needed to bind & proof-read the UDP-GlcA using direct and '2nd layer' amino acid residues. This step also allows the HAS to distinguish the two UDP-sugars; this is very important as the enzymes are not known or observed to make homopolymers of only GlcA or GlcNAc, but only make the HA disaccharide repeats GlcNAc-GlcA.

Strengths:

Techniques & analysis; overview of HA synthase mechanisms

Weaknesses:

None

Comments on revisions:

Previous clarity issues in the original submission were all resolved. Again, this is a very well done body of work!!

Reviewer #2 (Public review):

Summary:

The paper by Stephens and co-workers provides important mechanistic insight into how hyaluronan synthase (HAS) coordinates alternating GlcNAc and GlcA incorporation using a single Type-I catalytic centre. Through cryo-EM structures capturing both "proofreading" and fully "inserted" binding poses of UDP-GlcA, combined with detailed biochemical analysis, the authors show how the enzyme selectively recognizes the GlcA carboxylate, stabilizes substrates through conformational gating, and requires a priming GlcNAc for productive turnover.

These findings clarify how one active site can manage two chemically distinct donor sugars while simultaneously coupling catalysis to polymer translocation.

The work also reports a DDM-bound, detergent-inhibited conformation that possibly illuminates features of the acceptor pocket, although this appears to be a purification artefact (it is indeed inhibitory) rather than a relevant biological state.

Overall, the study convincingly establishes a unified catalytic mechanism for Type-I HAS enzymes and represents a significant advance in understanding HA biosynthesis at the molecular level.

Strengths:

There are many strengths.

This is a multi-disciplinary study with very high-quality cryo-EM and enzyme kinetics (backed up with orthogonal methods of product analysis) to justify the conclusions discussed above.

Comments on revisions:

The suggestions made in the initial comments have all been responded to very well.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript describes critical intermediate reaction steps of a HA synthase at the molecular level; specifically, it examines the 2nd step, polymerization, adding GlcA to GlcNAc to form the initial disaccharide of the repeating HA structure. Unlike the vast majority of known glycosyltransferases, the viral HAS (a convenient proxy extrapolated to resemble the vertebrate forms) uses a single pocket to catalyze both monosaccharide transfer steps. The authors' work illustrates the interactions needed to bind & proof-read the UDP-GlcA using direct and '2nd layer' amino acid residues. This step also allows the HAS to distinguish the two UDP-sugars; this is very important as the enzymes are not known or observed to make homopolymers of only GlcA or GlcNAc, but only make the HA disaccharide repeats GlcNAc-GlcA.

Strengths:

Overall, the strengths of this paper lie in its techniques & analysis.

The authors make significant leaps forward towards understanding this process using a variety of tools and comparisons of wild-type & mutant enzymes. The work is well presented overall with respect to the text and illustrations (especially the 3D representations), and the robustness of the analyses & statistics is also noteworthy.

Furthermore, the authors make some strides towards creating novel sugar polymers using alternative primers & work with detergent binding to the HAS. The authors tested a wide variety of monosaccharides and several disaccharides for primer activity and observed that GlcA could be added to cellobiose and chitobiose, which are moderately close structural analogs to HA disaccharides. Did the authors also test the readily available HA tetramer (HA4, [GlcA-GlcNAc]2) as a primer in their system? This is a highly recommended experiment; if it works, then this molecule may also be useful for cryo-EM studies of CvHAS as well.

The reviewer requested testing whether an HA tetratsaccharide could also serve as an glycosyl transfer acceptor for HAS. The commerically available HA tetrasaccharide (HA4) is terminated at its non-reducing end by GlcA, therein we proceeded to measure its effect on UDP-GlcNAc turnover kientics. Titration of HA4 failed to elicit any detectable change in UDP-GlcNAc turnover rate, indicating no priming. This is now mentioned in the main text and the data is shown in Fig. S9.

Weaknesses:

In the past, another report describing the failed attempt of elongating short primers (HA4 & chitin oligosaccharides larger than the cello- or chitobiose that have activity in this report) with a vertebrate HAS, XlHAS1, an enzyme that seems to behave like the CvHAS ( https://pubmed.ncbi.nlm.nih.gov/10473619/); this work should probably be cited and briefly discussed. It may be that the longer primers in the 1999 paper and/or the different construct or isolation specifics (detergent extract vs crude) were not conducive to the extension reaction, as the authors extracted recombinant enzyme.

We apologize for the oversight. This reference is now cited (ref. 18) together with the description of the failed elongation of HA4 by CvHAS.

There are a few areas that should be addressed for clarity and correctness, especially defining the class of HAS studied here (Class I-NR) as the results may (Class I-R) or may not (Class II) align (see comment (a) below), but overall, a very nicely done body of work that will significantly enhance understanding in the field.

Done as requested

Reviewer #2 (Public review):

Summary:

The paper by Stephens and co-workers provides important mechanistic insight into how hyaluronan synthase (HAS) coordinates alternating GlcNAc and GlcA incorporation using a single Type-I catalytic centre. Through cryo-EM structures capturing both "proofreading" and fully "inserted" binding poses of UDP-GlcA, combined with detailed biochemical analysis, the authors show how the enzyme selectively recognizes the GlcA carboxylate, stabilizes substrates through conformational gating, and requires a priming GlcNAc for productive turnover.

These findings clarify how one active site can manage two chemically distinct donor sugars while simultaneously coupling catalysis to polymer translocation.

The work also reports a DDM-bound, detergent-inhibited conformation that possibly illuminates features of the acceptor pocket, although this appears to be a purification artefact (it is indeed inhibitory) rather than a relevant biological state.

Overall, the study convincingly establishes a unified catalytic mechanism for Type-I HAS enzymes and represents a significant advance in understanding HA biosynthesis at the molecular level.

Strengths:

There are many strengths.

This is a multi-disciplinary study with very high-quality cryo-EM and enzyme kinetics (backed up with orthogonal methods of product analysis) to justify the conclusions discussed above.

Weaknesses:

There are few weaknesses.

The abstract and introduction assume a lot of detailed prior knowledge about hyaluronan synthases, and in doing so, risk lessening the readership pool.

A lot of discussion focuses on detergents (whose presence is totally inhibitory) and transfer to non-biological acceptors (at high concentrations). This risks weakening the manuscript.

The abstract and parts of the introduction have been revised to address the reviewer’s concerns.

Reviewer #1 (Recommendations for the authors):

(1) As noted above, please state in title, abstract & introduction that this work is focused on a "Class I-NR HAS" (as described in Ref. #4), and NOT all HAS families...this is truly essential to note as someone working with the Pasteurella HAS version (Class II) would be totally misled & at this point, no one knows the Streptococcus HAS (Class-IR) mechanistic details which could be different due to its inverse molecular directionality of elongation compared to the CvHAS Class I-NR enzyme.

Done as requested.

(2) Page 6 - for the usefulness of the HAS mutants as being folded correctly, it was stated these mutants are suitable since they all 'purify' similarly...the use of the more proper term should probably be 'chromatograph', similarly suggesting similar hydrodynamic radii without massive folding issues.

This has been revised to state that they all exhibited comparable size exclusion chromatography profiles.

“All mutants share similar size exclusion chromatography profiles with the WT enzyme, suggesting that the substitutions do not cause a folding defect (Fig. S3).”

(3) Page 7 - please check these sentences (& rest of paragraph?) as the meaning is not clear. "First, UDP-GlcNAc was titrated in the presence of excess UDP-GlcA, resulting in a response similar to the acceptor-free condition (Fig. 2C). However, the maximum reaction velocity at 20 mM UDP-GlcNAc was approximately 25% lower than that measured in the presence of UDP-GlcNAc only (Fig. 2C)."

The paragraph has been revised to avoid confusion.

(4) In Methods, please use an italicized 'g' for the centrifugation steps globally.

Changed as requested

(5) Please note the source/vendor for the HA standards on gels.

Done

(6) Page 35 - TLC section.

(a) 'n-butanol' (with italic n) is the most widespread chemical name (not butan-1-ol).

Done

(b) Also, for all of the TLC images, the origin and the solvent front should be marked.

Changed as suggested.

Reviewer #2 (Recommendations for the authors):

A number of minor issues should be addressed.

(1) Abstract

Two comments on the Abstract, which I found surprisingly weak given the quality of the work, and lacking a key detail.

A major conceptual contribution of this work is the demonstration of how a single Type-I catalytic centre discriminates, positions, and transfers two chemically distinct substrates in an alternating pattern. This distinguishes HAS from dual-active-site (Type-II) glycosyltransferases and is important for understanding HA polymerization.

However, this central point is not clearly articulated in the abstract. I suggest explicitly stating that HAS performs both GlcNAc and GlcA transfer reactions within a single catalytic site, and that the proofreading/inserted poses illuminate how this multifunctionality is achieved.

The abstract currently ends with the observation of a DDM-bound, detergent-inhibited state. While this is interesting, it absolutely does not represent the central conceptual advance of the study and gives the abstract an artefactual ending.

I strongly recommend revising the final sentences to emphasize the broader mechanistic insight and not an "artefact" (indeed, the enzyme is inactive in the presence of this detergent; it is thus a very unusual way to conclude an abstract).

That is, finish with the wider implications of how HAS coordinates alternating substrate use, proofreading, and polymer translocation. Ending on the main mechanistic or biological significance would make the abstract considerably stronger and more aligned with the main message of the paper.

The abstract has been revised thoroughly to reflect the important insights gained on CvHAS’ catalytic function and HA biogenesis in general.

(2) Introduction

The distinction between single active-centre enzymes, which transfer both sugars alternately, and twin catalytic domain enzymes that each perform one addition is surely central to the whole paper. But it is not discussed. Surely this has to be covered. There is a lot of work in this space, including, but not limited to:

https://doi.org/10.1093/glycob/cwg085

https://doi.org/10.1093/glycob/10.9.883

https://doi.org/10.1093/glycob/cwad075 (includes this author team)

Originally back to https://doi.org/10.1021/bi990270y

If the authors instead assume such a level of knowledge for the reader, then surely they are writing for a specialist audience, not consistent with the wider readership ambitions of eLife?

The Introduction has been revised as suggested by the reviewer, providing necessary background to frame our description of the Chlorella virus HAS. We made a deliberate effort to put new insights into a broader context.

(3) Results and Discussion

DDM "was observed for >50% of the analysed particles". I struggled with this. I couldn't understand how the authors selected particles that did or did not contain DDM. The main body text states: "To our surprise, careful sorting of the UDP-GlcA supplemented cryo EM dataset revealed a CvHAS subpopulation that was not bound to the substrate, but, instead, a DDM molecule near the active site (Fig 3A and S7). This was observed for >50% of the analyzed particles."

That reads like there is one sample with two populations. But the figures and the methods section suggest differently: they suggest two samples with different data-collection regimes. That does not match the main text. Could this be clarified?

Yes, that wasn’t explained well. We clarified the text to stress that the DDM-bound sample came from a dataset that was intended to resolve an UDP-GlcA-bound state, but instead revealed the inhibition by DDM.

Also in this space, in the modern world, "nominal magnification" has no real meaning, and calibrated pixel size would be more appropriate. Can this be given, please?

The relevant Methods section now states: “imaging of … was performed at a calibrated pixel size of 0.652 Å”.

The discovery of DDM in the active site is surprising. But it is an inhibitory artefact. Is this section pushed a little too hard? Also, "The coordination of DDM's maltoside moiety, an αlinked glucose disaccharide, is consistent with priming by cellobiose and chitobiose." I'm not sure why an α-linked maltose is consistent with the binding of a β-linked cellobiose. That makes no sense. There will be no other enzymes where starch and cellulose oligos are mutually accepted. Consider rewriting.

We like to stress the DDM coordination because it could lead to the development of compounds that can really function as inhibitors, either for HAS or other related enzymes. In the observed DDM binding pose, the alpha-linkage is not recognized. Instead, the reducing end glucosyl unit stacks against Trp342 while the non-reducing unit extends into the catalytic pocket. Hence, a similar binding pose is conceivable for cellobiose and potentially also for chitobiose. The relevant section has been reworded.

eLife Assessment

This important study introduces an approach to discovering antibiotic resistance determinants by leveraging diverse susceptibility profiles among related mycobacterial species, with particular relevance to high-level resistance against natural product-derived antibiotics. The research provides convincing evidence for the role of ADP-ribosylation enzymes in rifamycin resistance among mycobacteria, whilst also demonstrating that antibiotic susceptibility is not correlated with growth rate or intracellular compound concentration. The revision is substantively improved, but some broader claims still require additional experimental support. This work lays a significant foundation for understanding the complexity of antibiotic resistance mechanisms in mycobacteria and opens new avenues for future antimicrobial research.

Reviewer #1 (Public review):

This work analyzes innate resistance to drugs in mycobacteria by comparing minimum inhibitory concentrations (MICs) across a diverse panel of mycobacterial species. The results show that MICs are poorly correlated with growth rate while phylogeny associated with horizontal gene transfer underlies the observed differences in MIC, an important demonstration. A further investigation into the driver for the vast differences in susceptibility profiles shows that for three drugs the MIC is not correlated with intrabacterial drug concentrations where intrabacterial drug concentration is comprised of cytosolic and cell wall associated drug. This is a striking observation. The authors delve into the mechanisms that drive resistance to rifamycins and confirm that resistance is driven by ADP-ribosyltransferases of which two variant groups exist, one of which is kinetically faster and apparently is superior at modifying more hydrophobic rifamycins. The relative role of the two ADP-ribosyltransferases in conferring resistance especially in the species with both orthologs is not fully understood since the modified drug can possibly be further modified and transcriptional downregulation experiments performed in this work do not provide genetic evidence of perturbation of mRNA levels of the respective open reading frames.

Comments on revisions:

Demonstration of the level of transcriptional downregulation of the two Arr orthologs would have been a nice demonstration of (1) the utility of CRISPRi in other mycobacteria, (2) that the difference in rifabutin susceptibility during knockdown of Arr-1 vs Arr-X can fully be ascribed to the role of Arr-X in modifying the drug.

Reviewer #3 (Public review):

This manuscript presents a macroevolutionary approach to identification of novel high-level antibiotic resistance determinants that takes advantage of the natural genetic diversity within a genus (mycobacteria, in this case) by comparing antibiotic resistance profiles across related bacterial species and then using computational, molecular, and cellular approaches to identify and characterize the distinguishing mechanisms of resistance. The approach is contrasted with "microevolutionary" approaches based on comparing resistant and susceptible strains of the same species and approaches based on ecological sampling that may not include clinically relevant pathogens or related species. The potential for new discoveries with the macroevolution-inspired approach is evident in the diversity of drug susceptibility profiles revealed amongst the selected mycobacterial species and the identification and characterization of a new group of rifamycin-modifying ADP-ribosyltransferase (Arr) orthologs of previously described mycobacterial Arr enzymes. Additional findings that intra-bacterial antibiotic accumulation does not always predict potency within this genus, that M. marinum is a better proxy for M. tuberculosis drug susceptibility than the commonly used saprophyte M. smegmatis, and that susceptibility to semi-synthetic antibiotic classes is generally less variable than susceptibility to antibiotics more directly derived from natural products strengthen the claim that the macroevolutionary lens is valuable for elucidating general principles of susceptibility within a genus.

There are some limitations to the work. The argument for the novelty of the approach could be better articulated. While the opportunities for new discoveries presented by identification of discrepant susceptibility results between related species is evident, it is less clear how the macroevolutionary approach is further leveraged for the discovery of truly novel resistance mechanisms. The example of the discovery of Arr-X enzymes presented here relied upon foundational knowledge of previously characterized Arr orthologs. There is less clarity about what the pipeline would look like for discovery of previously unknown determinants when one is agnostic to putative mechanisms. From the point at which interspecies differences in susceptibility are noted, does the framework still remain distinct from other discovery frameworks and approaches?

While the experimentation and analyses performed are generally well designed and rigorous, there are a few instances in which broad claims are based on inferences from sample sets or data sets that are, at present, too limited to provide robust support. For example, the claim that rifampicin modification, and precisely ADP-ribosylation, is the dominant mechanism of resistance to rifampicin in mycobacteria is still a bit premature or at least an over-generalization, as other enzymatic modification mechanisms and other mechanisms such as helR-mediated dissociation of rifampicin-stalled RNA polymerases, efflux, etc were not examined. CRISPR interference was used in a demonstrative example to support this assertion, but would need to be applied more systematically to be more conclusive. The general claim that intra-bacterial antibiotic accumulation does not predict potency in mycobacteria may be another over-generalization based on the limited set of drugs and species studied.

Comments on revisions:

Discussion, lines 321-323: "We found that resistance to these antibiotics in mycobacteria do not correlate with by uptake/efflux mechanisms in the species tested..." is an over-generalization and conflicts with the following statement on lines 199-201: "for BDQ we could observe some correlation between antibiotic potency and [BDQ]IB which could be indicative of efflux playing a role in antibiotic efficacy. Given that the current statement in the Discussion only applies to 2 of 3 drugs tested, a more specific or nuanced interpretation seems warranted.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

This work shows that resistance profiles to a variety of drugs are variable between different mycobacterial species and are not correlated with growth rate or intrabacterial compound concentration (at least for linezolid, bedaquiline, and Rifampicin). Note that intrabacterial compound concentration does not distinguish between cytosolic and periplasmic/cell wall-associated drugs. The susceptibility profiles for a wide range of mycobacteria tested under the same conditions against 15 commonly used antimycobacterial drugs provide the first recorded cross-species comparison which will be a valuable resource for the scientific community. To understand the reasons for the high Rifampicin resistance seen in many mycobacteria, the authors confirm the presence of the arr gene known to encode a Rif ribosyltransferase involved in Rif resistance in M. smegmatis in the resistant mycobacteria after confirming the absence of on-target mutations in the RpoB RRDR. Metabolomic analyses confirm the presence of ribosylated Rif in some of the naturally resistant mycobacteria which may not be entirely surprising but an important confirmation. Presumably M. branderi is highly resistant despite lacking the arr homolog due to the rpoB S45N mutation. M. flavescens has an MIC similar to that of M. smegmatis, despite having both Arr-1 and Arr-X. Various Arr-1 and Arr-X proteins are expressed and characterized for catalytic activity which shows that Arr-X is a faster enzyme,, especially with respect to more hydrophobic rifamycins. M. flavescens has similar MIC values to Rifapentine and Rifabutin to M. smegmatis. Thus, the Arr-1 versus Arr-X comparison does not provide a complete explanation for the underlying reasons driving natural Rif resistance in mycobacteria. Downregulation of Arr-X expression in M. conceptionense confers increased sensitivity to Rifabutin confirming its role as a rifamycin-inactivating enzyme.

Overall, the comparison of cross-species susceptibility profiles is novel; the demonstration that MIC is not correlated with intracellular drug concentration is important but not sufficiently interrogated, the demonstration that Arr-X is also a Rif ADP-ribosyltransferase is a good confirmation and shows that it is more efficient than Arr-1 on hydrophobic rifamycins is interesting but maybe not entirely surprising. The manuscript seems to have two parts that are related, but the rifamycin modification aspect of the work is not strongly linked to the first part since it interrogates the modification of one drug but not the common cause of natural resistance for other drugs.

Reviewer #2 (Public review):

Summary:

The authors use a variety of methods to investigate the mechanisms of innate drug resistance in mycobacteria. They end up focusing on two primary determinants - drug accumulation, which correlates rather poorly with resistance for many species, and, for the rifamycins, ADP-ribosyltransferases. The latter enzymes do appear to account for a good deal of resistance, though it is difficult to extrapolate quantitatively what their relative contributions are.

Overall, they make excellent use of biochemical methods to support their conclusions. Though they set out to draw very broad lessons, much of the focus ends up being on rifamycins. This is still a very interesting set of conclusions.

Strengths:

(1) A very interesting approach and set of questions.

(2) Outstanding technical approaches to measuring intracellular drug concentrations and chemical modification of rifamycins.

(3) Excellent characterization of variant rifamycin ADP-ribosyltransferases

Weaknesses:

(1) Figure 3c/d: These panels show the same experiment done twice, yet they display substantially different results in certain cases. For instance, M. smegmatis appears to show an order of magnitude lower RIF accumulation in panel d compared to M. flavescens, despite them displaying equal accumulation in panel c. The authors should provide justification for this variation, particularly as quantitative intra-species comparisons are central to the conclusions of this figure.

The data in panels 3c and 3d are from different sets of experiments. The reviewer is correct with regards to M. smegmatis. The data indeed is ~ 1 order of magnitude different. However, the data for other species is very similar. The reviewer may also have noticed that the error bars are also larger in 3d, compared to 3c, indicating a greater variation between independent experiments use in 3d. We do not have a good explanation for this, other than the experiments shown in 3d were associated with greater biological variability.

(2) There are several technical concerns with Figure 3 that affect how to interpret the work. According to the methods, the authors did not appear to normalize to an internal standard, only to an external antibiotic standard (which may account for some of the technical variation alluded to above).

We agree that using a labeled drug as an internal standard (IS) would be ideal. However, the experiment initially followed an untargeted metabolomics approach, which later shifted to relative drug quantification. At that stage, normalizing with IS was impractical because proper implementation would require multiple IS across the chromatographic range. Therefore, we opted for total ion current (TIC) normalization, which accounts for variability in overall metabolite abundance—even though the experimental setup was already adjusted for each bacterial species’ growth rate. Additionally, we prepared external standard curves for each drug to enable quantification, and the amount of drug added to each plate was considered when reporting these values.

Second, the authors used different concentrations of drug for each species to try to match the species' MICs. I appreciate the authors' thinking on this, but I think for an uptake experiment it would be more appropriate to treat with the same concentration of drug since uptake is likely saturable at higher drug concentrations. In the current setup, for the species with higher MIC, they have to be able to uptake substantially more antibiotics than the species with low MIC in order to end up with the same normalized uptake value in Figure 3d. It would be helpful to repeat this experiment with a single drug concentration in the media for all species and test whether that gives the same results seen here.

We respectfully disagree with the reviewer. Experiments such as the one proposed by the review work well when MIC values are a few fold apart, for strains of the same species, but have not been tested when MIC values are 100-1000-fold apart, with different species. Furthermore, what would be the interpretation of compound uptake at 1000-fold the MIC for one species and MIC level for another? By using antibiotic concentrations at the respective MIC for each species we are at least under conditions where we know the biological effect of the antibiotic across species is the same, based on its potency.

(3) Figure 4f: This panel seems to argue against the idea that the efficacy of RIF ribosylation is what's driving drug susceptibility. M. flavescens is similarly resistant to RIF as M. smegmatis, yet M. flavescens has dramatically lower riboslyation of RIF. This is perhaps not surprising, as the authors appropriately highlight the number of different rif-modifying enzymes that have been identified that likely also contribute to drug resistance. However, I do think this means that the authors can't make the claim that the resistance they observe is caused by rifamycin modification, so those claims in the text and figure legend should be altered unless the authors can provide further evidence to support them. This experiment also has results that are inconsistent with what appears to be an identical experiment performed in Supplemental Figure 5b. The authors should provide context for why these results differ.

In regard to enzyme efficiency, the apparent rate of all Arr-1 is relatively similar in converting RIF into ADP-Ribosyl-Rif between species. However, Arr-X is much more efficient when compared to Arr-1 in both M. flavescents and M. conceptionense. This is indicated by the apparent rate measured and displayed on figure 5c.

Proteomics data shows that there is upregulation of Arr-1 and Arr-X upon rifampicin treatment in M. flavescens and M. conceptionense. However, the same experiment was not performed in Arr-1 KD. Therefore, we can’t verify through this approach if the activity observed in vivo directly correlates with a higher expression of Arr-X alone. Of note, likely both enzymes contribute to resistance to rifamycins, as per our results with the Arr-X KD and sensitization of M. conceptionense to RIF.

Author response image 1.

It is also worth mentioning that there are other enzymes in the pathway of RIF ribosylation and their efficiency is unknown (Author response image 2). Therefore ADP-Ribosyl-RIF It is not an “end-metabolite” and maybe not the sole determinant of RIF resistance via ADP-ribosylation. Downstream enzymes can also account for the difference observed between M. flavescens and M. smegmatis.

Author response image 2.

It is correct that the Rifampicin MIC for M. flavescens is the same as M. smegmatis.

(4) Fig 4f/5c: M. flavescens has both Arr-1 and Arr-X, yet it appears to not have ribosylated RIF. This result seems to undermine the authors' reliance on the enzyme assay shown in Fig 5c - in that assay, M. flavescens Arr-X is very capable of modifying rifampicin, yet that doesn't appear to translate to the in vivo setting. This is of importance because the authors use this enzyme assay to argue that Arr-X is a fundamentally more powerful RIF resistance mechanism than Arr-1 and that it has specificity for rifabutin. However, the result in Figure 4f would argue that the enzyme assay results cannot be directly translated to in vivo contexts. For the authors to claim that Arr-X is most potent at modifying rifabutin, they could test their CRISPRi knockdowns of Arr-X and Arr-1 under treatment with each of the rifamycins they use in the enzyme assay. The authors mentioned that they didn't do this because all the strains are resistant to those compounds; however, if Arr-X is important for drug resistance, it would be reasonable to expect to see sensitization of the bacteria to those compounds upon knockdown.

The reviewer is reading Fig. 4f incorrectly, probably because it is plotted in a linear scale instead of logarithmic scale. Ribosylated Rif is present in M. flavescens, just at lower levels than M. conceptionense and M. smegmatis. In species where there is no Arr-1 or Arr-3, ribosylated RIF is not detected at all (e.g. M. tuberculosis), i.e., concentration is zero. Therefore, any detection of ribosylated RIF can be considered significant. In addition, as mentioned before, ADP-ribosylation of RIF is not the final product of the reaction and further studies need to be undertaken to understand subsequent reactions.

(5) Figure 5d: The authors use this CRISRPi experiment to claim that ArrX from M. conceptionanse is more potent at inactivating rifabutin than Arr-1. This claim depends on there being equal degrees of knockdown of Arr-1 and Arr-X, so the authors should validate the degree of knockdown they get. This is particularly important because, to my knowledge, nobody has used this system in M. conceptionanse before.

We agree with the reviewer that a qPCR should have been performed to define the extent of interference in the strain. generated Unfortunately, at this time a qPCR was not performed in the strains tested to confirm the extent of down regulation. Although it is the best practice to validate the strain KD, there is no indication that the effect observed is due to unspecific downregulation. The genetic environment in which Arr-X is positioned is different from Arr-1 and the targeting oligonucleotides are specific and would not promiscuously bind to Arr-1. Said that, this is indeed a fault in our setup.

(6) The authors' arguments about Arr-X and Arr-1 would be strengthened by showing by LC/MS that Arr-X knockdown in M. conceptionense results in more loss of ribosyl-rifabutin than knockdown of Arr-1.

We agree with the reviewer that performing the LC-MS analysis of the Arr-x knockdown would have strengthened the argument of our paper. Unfortunately, this experiment was not performed.

Reviewer #3 (Public review):

This manuscript presents a macroevolutionary approach to the identification of novel high-level antibiotic resistance determinants that takes advantage of the natural genetic diversity within a genus (mycobacteria, in this case) by comparing antibiotic resistance profiles across related bacterial species and then using computational, molecular, and cellular approaches to identify and characterize the distinguishing mechanisms of resistance. The approach is contrasted with "microevolutionary" approaches based on comparing resistant and susceptible strains of the same species and approaches based on ecological sampling that may not include clinically relevant pathogens or related species. The potential for new discoveries with the macroevolution-inspired approach is evident in the diversity of drug susceptibility profiles revealed amongst the selected mycobacterial species and the identification and characterization of a new group of rifamycin-modifying ADP-ribosyltransferase (Arr) orthologs of previously described mycobacterial Arr enzymes. Additional findings that intra-bacterial antibiotic accumulation does not always predict potency within this genus, that M. marinum is a better proxy for M. tuberculosis drug susceptibility than the commonly used saprophyte M. smegmatis, and that susceptibility to semi-synthetic antibiotic classes is generally less variable than susceptibility to antibiotics more directly derived from natural products strengthen the claim that the macroevolutionary lens is valuable for elucidating general principles of susceptibility within a genus.

There are some limitations to the work. The argument for the novelty of the approach could be better articulated. While the opportunities for new discoveries presented by the identification of discrepant susceptibility results between related species are evident, it is less clear how the macroevolutionary approach is further leveraged for the discovery of truly novel resistance determinants. The example of the discovery of Arr-X enzymes presented here relied upon foundational knowledge of previously characterized Arr orthologs. There is little clarity on what the pipeline for identifying more novel resistance determinants would look like. In other words, what does the macroevolutionary perspective contribute to discovery from the point of finding interspecies differences in susceptibility? Does the framework still remain distinct from other discovery frameworks and approaches? If so, how?

Thanks for pointing this out, as this is a critical feature of our study and method. Our approach relies on inter-species comparative genomics and phenotypes, and therefore, it is distinct from inter-strains comparison. This difference is dramatic, and it becomes clearer when we are comparing the core genome of M. tuberculosis (one species) 92% with the core genome of the genus, circa of 1%. While we focus on rifamycin in this manuscript, future manuscripts will investigate many of the other dozens of “inconsistencies” observed between the genetic makeup of different mycobacterial species and there actual performance in the presence of different antibiotics.

While the experimentation and analyses performed appear well-designed and rigorous, there are a few instances in which broad claims are based on inferences from sample sets or data sets that are too limited to provide robust support. For example, the claim that rifampicin modification, and precisely ADP-ribosylation, is the dominant mechanism of resistance to rifampicin in mycobacteria may be a bit premature or an over-generalization, as other enzymatic modification mechanisms and other mechanisms such as helR-mediated dissociation of rifampicin-stalled RNA polymerases, efflux, etc were not examined nor were CRISPRi knockdown experiments conducted beyond an experiment to tease out the role of Arr-X and Arr-1 in one strain. The general claim that intra-bacterial antibiotic accumulation does not predict potency in mycobacteria may be another over-generalization based on the limited number of drugs and species studied, but perhaps the intended assertion was that antibiotic accumulation ALONE does not predict potency.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Major comments

(1) The metabolomics is done using mycobacteria grown on filters. Initially, mycobacterial cells are grown on the filters for 5 doublings before being transferred to drug-containing (or free) agar for one doubling. Is this based on calculated doubling time in liquid culture or a true determination of the fact that the biomass increases to what would amount to 5 doublings?

The doubling time used is the one determined in liquid media. Although it is possible that the growth kinetics in solid media is slightly different from liquid (±10%), this experimental design is well established for M. tuberculosis (since Proc Natl Acad Sci U S A. 2010 May 25;107(21):9819-24.) and M. smegmatis (unpublished). Therefore, we used the growth rate as a proxy for having the same biomass of cells for each species tested. A maximum difference of 10% was observed between M. tuberculosis growth in liquid and in solid media, however, cells grow exponentially for much longer in filters. This makes filter-based experiments more reliable, as few growth phase-derived differences are present.

(2) The demonstration that intrabacterial drug concentrations vary between mycobacterial species in a manner not related to MIC for at least LZD and RIF, is an important finding. However, intrabacterial does not mean cytoplasmic since a considerable fraction could be present in the periplasmic/cell wall layers. Ideally, this would need to be determined but would of course be a massive undertaking since the method needs validation & optimization for each mycobacterial species. Nevertheless, this has to be mentioned. In addition, three drugs are limiting. Measuring additional drug concentrations in these 5 mycobacteria would at least establish some confirmation about the extent of this lack of correlation. Thus, could the authors measure concentrations of additional drugs with intracellular targets?

Testing additional drugs can be beneficial and would be an expansion of our paper, which will definitely be on future plans for further studies focusing on other antibiotics described here. It would also provide new insights into other possible mechanisms of resistance in mycobacterial species. However, in this study we aimed to first determine the antibiotic response profile in different mycobacterial species, and once we identified interesting resistance phenotypes that could not be readily explained by known mechanisms of resistance, we narrowed it down to certain drugs and species that would potentially provide insights into new mechanisms of antibiotic resistance. Finally, exploring drug concentration across multiple bacterial compartments is a dauting task and it has not been done extensively with any species, not to mention with multiple species, many of which are still lacking any study of their actual cell envelope.

(3) CRISPRi was used to reduce transcription in M. conceptionense. What was the level of gene downregulation?

As mentioned previously, a setback from our setup is that the level of KD was not measured at this instance.

Minor comments:

(1) The introduction mentions the fast and slow-growing mycobacteria which are classified based on the time that it takes to observe colonies on solid agar. However, in liquid medium, there is less correlation between the reported growth on agar and doubling time in liquid (Figure 1b, Figure 2d). This could be mentioned in the results section. In Figure 2d, the filled circles represent fast-growers but this does not hold well for liquid culture and it might make more sense to not distinguish between fast- and slow-growers in these graphs. A small complication would also be the fact that the doubling time represents growth in a liquid medium with Tyloxapol as a detergent whereas the MIC and metabolomics are done on solid agar with no detergent. The metabolomics is done after a doubling but for those where agar growth and liquid growth have large discrepancies in growth rate, there could be some differences.

Apologies for this misunderstanding. Fast- and slow-growth phenotypes are determined in Lowenstein-Jensen (LJ) agar, not in 7H10 agar (used in our study and most studies of mycobacteria). Furthermore, this is a qualitative definition, not a quantitative one. Therefore, our measurements do not need to correlate with fast- and slow-growth phenotypes, unless we had used that one specific medium. Furthermore, in liquid medium, we determined growth rate directly, which is never done with LJ medium.

In addition to adding the same amount of cells to each filter, we also perform TIC normalization, which should account for how rich the samples were – and therefore how much material we had. Therefore, we do not observe discrepancies due to differences in growth rate and the presence/absence of detergent in the media.

It is also worth mentioning that this experimental set up has been well established in many M. tuberculosis labs that study metabolism. Importantly, the use of detergent drastically affects mass spectrometry, and therefore cannot be used.

(2) Figure 1g in the text should be Figure 1f.

Apologies, it has been fixed.

(3) Figure S1 would be ideal to have in (supplementary) table format.

This data is now being provided in a table format.

(4) Table S1 - ethambutol misspelt.

Spelling has been corrected.

(5) MIC for species such as M. abscessus could depend on medium (7H9-based medium can give different MIC values than CAMH).

Indeed, different media can significantly change MIC values, and this is true for many bacterial species, if not all. For this study we used only species that could be grown in 7H9 broth containing 10 % ADC, 0.05% glycerol 0.05% tyloxapol and 7H10 plates containing 10% OADC and 0.05% glycerol. MIC₉₉ was determined in the latter as we found more efficient and robust to do our tests it in solid media. The goal of our experiment was not to the determined the “true” MIC for the antibiotics tested, as this value does not exist. It was to find lack of correlations between relative values and the presence of genes that can account for it.

(6) The statement "the experiment was performed at a concentration of antibiotic equal to its MIC" initially seems confusing. It was not equal to the MIC but performed at 6-fold the respective MIC of the species in question. Maybe re-phrasing this would help.

Apologies for this oversight. It has been corrected.

(7) Note that some mutations outside the RRDR (eg. V170F and I491F) can also cause Rif resistance.

Author response image 3.

A Rainbow diagram of RpoB X-Ray structure coloured according to sequence conservation. Dark purple indicates high conservation, whereas dark orange indicates low conservation. RIF (showed in magenta) is bound to RpoB. Zoomed view displays that the RIF-binding pocket is considerably conserved. B RpoB protein sequence has an 81bp region called Rifampicin Resistance Determining Region (RRDR) that is known to be important for RIF binding and is where most mutations occur in drug-resistant TB. Sequence alignment displays that the RRDR region is conserved with the exception of M. branderi, which has an Asn instead of a Ser residue in position 456 (numbering is related to the M. tuberculosis sequence), highlighted in bold.

Attached we have a structural alignment of RpoB of the species highlighted on this paper. Although there is variability within the sequences, which is also displayed in Author response image 3 with the conservation analysis, the residues that have been implicated with resistance (including V170 and I491) are conserved. Alignment sent on .fasta file that can be opened in jalview.

(8) Discuss how the RpoB S450N mutation in M. branderi confers the observed level of resistance.

That’s a great point, thank you. Now it reads as:

“The rifampicin (RIF) binding pocket is generally conserved, but Mycobacterium branderi has an S450N mutation in the RRDR region. While this specific mutation hasn't been found in clinical isolates, it's located at the binding site and may confer resistance (273). Although both serine (S) and asparagine (N) have similar side chains, related mutations like S450Q have been linked to resistance (156). Thus, M. branderi may be RIF-resistant due to this mutation. In contrast, M. conceptionense, M. flavescens, and M. smegmatis show no target sequence differences that explain their resistance”

(9) The statement that the three tested NTM are sensitive to rifabutin ("resistant to all rifamycins except for rifabutin") needs to be interpreted considering what sensitivity means. The MIC is still high (1.6-3.1 ug/mL) when compared to that of Mtb. The 2-fold differences in MIC between M. smegmatis and M. conceptionense do not really prove or disprove the role of Arr-X in rifabutin resistance.

We fixed the sentence to be more careful with the language on the text. We agree, but it is worth mentioning that generally with bacteria there is a regulation by the CLSI. Each bacterial species has a range that is considered sensitive or resistant, but these are not available for the species used in this study. In general, bacteria with MIC values above 8 µg/mL are considered resistant to rifampin (J Antibiot 2014 67:625).

(10) Figure 1d: It's hard to quantify the sensitivity of the plates. Can this be done by MIC? Was only rifabutin tested or also rifampicin?

The initial experiments described on the paper were all performed using Rifampicin only. Then, the MIC for the remaining rifamycins was determined for M. smegmatis, M. flavescens and M. conceptionense, and can be perused on “Supplementary table 4”. Figure 5d is to illustrate the effect of the KD in M. conceptionense sensitivity to rifabutin.

(11) Is there data to show the ADP-ribosylation of rifabutin in M. conceptionense and the CRISPRi strains?

Unfortunately, we did not perform LC-MS analysis on M. conceptionense CRISPRi strains exposed to rifabutin to measure potential ADP-ribosylation.

Reviewer #2 (Recommendations for the authors):

(1) It would be useful if the authors would complete Figure 1A by determining growth rates for the remaining 18 strains that they currently omitted.

These growth rates were obtained using roller bottles and in at least 3 independent experiments, unfortunately the throughput is far ideal. The goal of the experiment was to highlight difference in growth rate, beyond fast- and slow-growth, which we did. Adding the remaining values would not change this conclusion. Growth rate variation in 7H9 is significant and the point is made in our figure.

(2) The authors should justify their choice of species used in Figures 3-4. It would be useful to know, for instance, if the authors chose these species in an unbiased fashion, or if they were chosen because the authors had already determined that they possess rifamycin-modifying enzymes of interest. In that case, they wouldn't necessarily be a representative sample to use for the correlation analysis of antibiotic uptake and potency in Figure 3.

They were chosen because of their resistance profile for BDQ, LZD and RIF. This has been addressed in the text, which now reads “Given the antibiotic response profiles observed, we selected BDQ, LZD and RIF to explore the molecular causes of these dramatic changes in antibiotic potency observed across the Mycobacterium genus.”

(3) Figure 4b: The data in this panel appear inconsistent - for instance, M. houstonense appears to grow at 10X Mtb MIC, but fails to grow at 1X Mtb MIC. Repeating this experiment would better establish the validity of the authors' claims about the relative susceptibility of these strains to RIF.

The figures got rotated when exported from illustrator. Corrected figure is uploaded, and original plate photos are also uploaded for clarity.

(4) Figure 4e: Does Arr-X get upregulated in these proteomic datasets? The authors' argument that proteomic upregulation correlates with important drug resistance genes would imply that it might be, so that would be useful information to provide.

Arr-X is slightly upregulated, but not statistically significant – this could be due to the native expression of Arr-1. Data is displayed in a previous answer.

(5) I wasn't able to find the supplementary tables that the authors allude to - not sure if that was a file mixup, but those tables would be useful for interpreting the manuscript.

We are sorry that you couldn’t access the table. It must be a file corruption issues, as the other reviewers were able to. We will make sure that all tables are available and accessible.

(6) For LC/MS, the authors use peak height instead of peak area, which they argue correlates better with the amount of drug in cells because of the poor peak shape they observed for linezolid. This is not standard practice, so the authors should provide evidence to support this claim by running an LC/MS standard curve, then showing the correlation between peak height and amount of compound added as well as the correlation between peak area and compound.

Thank you for pointing that out, accuracy calculated and displayed. Both peak area and height can be used, but indeed area is standard practice.

(7) The authors should provide methods information about the LC column and the gradient settings used for LC-MS, as well as the settings of the MS.

The full method has been added to the paper.

Reviewer #3 (Recommendations for the authors):

I have only minor comments aside from the information in the Public Review:

(1) Results, section on Intra-bacterial antibiotic accumulation, line 8: "experiment was performed at a concentration of antibiotic PROPORTIONAL to its MIC" would be more accurate?

Agreed and adjusted according to Reviewer’s suggestion.

(2) Results, section on A minor role for pre-existing target modification, last sentence: the mere presence of RIF-ribosylating enzymes does not, in and of itself indicate that "RIF modification, and precisely ADP-ribosylation, is the dominant mechanism of resistance to RIF in mycobacteria", as other mechanisms and other forms of modifying enzymes are known to confer rifamycin resistance, with redundancy (e.g., other rifampicin-modifying enzymes, or helR-mediated dissociation of rifampicin-stalled RNA polymerases from DNA). It would be more appropriate to suggest the results presented to this point indicate RIF modification is common among mycobacteria. The evidence from the CRISPRi knockdown of Arrs shown in Fig 5d is the kind of evidence that suggests ribosylation as a dominant mechanism, at least against rifabutin in this particular species.

Absolutely, there are other possible modifying enzymes that could be encoded by these mycobacterial species. There is a possibility that M. flavescens and M. smegmatis encode for a putative helR (attached alignment) but further experiments would need to be carried out to confirm its ability to displace RIF in the RNAP. Interestingly, the presence of both Arr and HelR has been studied in M. abscessus and those mechanisms of resistance are independent from each other (Molecular Cell 2022 82(17):3166-3177.e5).

(3) Discussion, 2nd sentence needs grammatical editing.

Rephrased and it reads “Using our mycobacterial library, we identified for the first time high- and ultra-high-level intrinsic resistance (3) to many of the antibiotics tested. Of note, the resistant phenotype is naturally occurring and not a result of mutations due to exposure to the antibiotic in the clinic – which is the more traditional approach for probing mechanisms of antibiotic resistance. Our observations revealed that resistance profiles are highly variable across the genus and do not follow phylogeny, implicating HGT as the key mechanism for acquisition of resistance determinants and evolution of antibiotic resistance in mycobacteria (42).”

(4) Discussion, page 7, first line: the inclusion of LZD and BDQ in this statement seems at odds with Figure 2c and the statements in the first paragraph of page 5 highlighting these as examples of drugs to which most mycobacteria are susceptible.

Indeed, many of the species are susceptible, however the MIC₉₉ levels observed have never been reported before, and therefore we found it to be an interesting finding to highlight. From a treatment perspective, knowing which species are sensitive to which drugs is of course the most useful outcome of our study.

(5) The next sentence..."We found that resistance to these antibiotics in mycobacteria cannot be explained by uptake/efflux mechanisms..." is a bit of an over-generalization and conflicts with the evidence presented earlier that efflux could be playing a role in BDQ resistance and the published evidence establishing a clinically significant role for efflux-mediated BDQ resistance in M. tuberculosis, M. avium complex and M. abscessus complex.

We rephrased it to make it more specific to our findings. It reads “We found that resistance to these antibiotics in mycobacteria do not correlate with by uptake/efflux mechanisms in the species tested and it does not correlate with growth rate. Identification of mycobacterial species highly resistant to BDQ and LZD is worrisome as most of this species, if not all, have never been exposed to these drugs.”

(6) Methods, section on In vitro activity assay of Arr enzymes, line 1: reference(s) should be provided for previously reported methods.

Reference now added.

(7) Figure 2d: the low end of the susceptibility range is not well defined.

In this figure the susceptibility is not defined as the lowest area of the graph, but the lower concentrations are indeed harder to be defined. Hopefully supplementary figure 1 and the additional table containing the MIC can be informative to address this comment.

(8) Figures 3c,d: the presentation of the relative antibiotic concentrations could be harmonized between the graphs in 3c and those in 3d to enable a more ready comparison.

We disagree. The goal of these different panels is exactly to illustrate two distinct points. C gives the relative concentration of antibiotic, while D correlates relative concentration with MIC99. The use of log scale in D further clarifies that there is no correlation between intracellular antibiotic concentration and potency (MIC). This information is not present in C.

(9) Figure 4f and Supplementary Figure 5b: it is difficult to understand the limited amount of ribsosyl-RIF in M. flavescens in Fig 4f relative to Supplementary Figure 5b (esp. when considering M. smeg as a common comparator); and, further, to understand the seeming lack of correlation between RIF susceptibility, ribosylation and Arr number and catalytic efficiency for these two strains without considering additional resistance mechanisms.

In reality the difference between figure 4f and Supplementary figure 5b is mainly due to M. smegmatis – that has an apparent lower production of ribosyl-RIF in the experiment described in the supplementary figure. The values for M. flavescens are relatively similar. In addition, the ADP-Ribosyl-RIF is not the final metabolite of the pathway.

In regards of having the entire picture, it is true that we were unable to completely unravel and correlate MIC value, expression of Arr-1, expression of Arr-3, efficiency of each enzyme, production of ADP-Ribosyl-RIF and the presence of other possible mechanisms of resistance and this is indeed a setback in our study, and of most studies ever published, which usually focus on one resistant determinant.

eLife Assessment

This valuable study used genetic and pharmacological manipulations of insulin/IGF signaling in renal glomerular podocytes to address the role of insulin/IGF axis in podocytes. Solid data are presented to demonstrate that co-inhibition of insulin/IGF signaling in podocytes led to aberrant splicing of mRNAs, which could contribute to the loss of podocytes in vitro and in vivo in mice. As it stands, the study lacks the assessment of developmental phenotype of podocytes in the mouse model.

[Editors' note: this paper was reviewed by Review Commons.]

Reviewer #1 (Public review):

Summary:

In this manuscript, the role of the insulin receptor and the insulin growth factor receptor was investigated in podocytes. Mice, where both receptors were deleted, developed glomerular dysfunction and developed proteinuria and glomerulrosclerosis over several months. Because of concerns about incomplete KO, the authors generated and studied podocyte cell lines where both receptors were deleted. Loss of both receptors was highly deleterious with greater than 50% cell death. To elucidate the mechanism of cell death, the authors performed global proteomics and found that spliceosome proteins were downregulated. They confirmed this directly by using long-read sequencing. These results suggest a novel role for insulin and IGF1R signaling in RNA splicing in podocytes.

This is primarily a descriptive study and no technical concerns are raised. The mechanism of how insulin and IGF1 signaling regulates splicing is not directly addressed but implicates potentially the phosphorylation downstream of these receptors. In the revised manuscript, it is shown that the mouse KO is incomplete potentially explaining the slow onset of renal insufficiency. Direct measurement of GFR and serial serum creatinines might also enhance our understanding of progression of disease, proteinuria is a strong sign of renal injury. An attempt to rescue the phenotype by overexpression of SF3B4 would also be useful but may be masked by defects in other spliceosome genes. As insulin and IGF are regulators of metabolism, some assessment of metabolic parameters would be an optional add-on.

Significance:

With the GLP1 agonists providing renal protection, there is great interest in understanding the role of insulin and other incretins in kidney cell biology. It is already known that Insulin and IGFR signaling play important roles in other cells of the kidney. So, there is great interest in understanding these pathways in podocytes. The major advance is that these two pathways appear to have a role in RNA metabolism.

Comments on revised version:

I'm satisfied with the revised manuscript and the responses to my previous concerns.

Reviewer #2 (Public review):

Summary:

In this manuscript, submitted to Review Commons (journal agnostic), Coward and colleagues report on the role of insulin/IGF axis in podocyte gene transcription. They knocked out both the insulin and IGFR1 mice. Dual KO mice manifested a severe phenotype, with albuminuria, glomerulosclerosis, renal failure and death at 4-24 weeks.

Long read RNA sequencing was used to assess splicing events. Podocyte transcripts manifesting intron retention were identified. Dual knock-out podocytes manifested more transcripts with intron retention (18%) compared wild-type controls (18%), with an overlap between experiments of ~30%.

Transcript productivity was also assessed using FLAIR-mark-intron-retention software. Intron retention w seen in 18% of ciDKO podocyte transcripts compared to 14% of wild-type podocyte transcripts (P=0.004), with an overlap between experiments of ~30% (indicating the variability of results with this method). Interestingly, ciDKO podocytes showed downregulation of proteins involved in spliceosome function and RNA processing, as suggested by LC/MS and confirmed by Western blot.

Pladienolide (a spliceosome inhibitor) was cytotoxic to HeLa cells and to mouse podocytes but no toxicity was seen in murine glomerular endothelial cells.

The manuscript is generally clear and well-written. Mouse work was approved in advance. The four figures are generally well-designed, bars/superimposed dot-plots.

Methods are generally well described.

Comments on revised version:

Coward and colleagues have done an excellent job of responding to all the reviewer comments.

Reviewer #4 (Public review):

Summary and background:

This report entitled "The insulin/IGF axis is critically important (for) controlling gene transcription in the podocyte" from Hurcombe et al is based on a mouse double knockdown of the IR and IGF1R and a parallel cultured mouse podocyte model. Insulin/IGF signaling system in mammals evolved as three gene reduplicated peptides (insulin, IGF-1, and IGF-2) and their two receptors IR and IGF1R that cross-react to variable extents with the peptides, are ubiquitously expressed, and signal through parallel pathways. The major downstream effect of insulin is to regulate glucose uptake and metabolism, while that of the IGF pathways is to regulate growth and cell cycling in part through mTORC1. The GH-IGF-1-IGF1R pathway regulates post-natal growth. IGF-2 signaling is thought to play a major role in regulating intrauterine growth and development, although IGF-2 is also present at high levels in post-natal life. Thus, one would anticipate that reducing IR/IGF1R signaling in any cell would slow growth and cell cycling by reducing growth factor and metabolic mTORC1-mediated and other processes including the splicing of RNA for protein synthesis.

Mouse IR/IGF1R double knockdown model:

A double knockdown mouse model was generated by interbreeding mice with different genetic backgrounds carrying floxed sites for IR and IGF-1R to produce mixed background offspring with both floxed IR and IGF-1R genes. These mice were crossed so that the podocin promoter driven-Cre (that comes on at about embryonic day 12 bas podocytes are developing) would delete IR and IGF-1R genes. Since podocin is believed to be an absolutely podocyte-specific protein, this podocin promoter this is predicted to specifically knock down the IR and IGF1R genes only in podocytes. The weight and growth of double KO offspring was not different from controls, but some proportion of the double knockdown mice subsequently developed proteinuria by 6 months and 20% died, although no specific data is provided to identify the cause of the deaths since eGFR was not decreased. Surviving mice were evaluated at 6 months of age. The efficacy of knockdown was not demonstrated in the mouse model itself, although a temperature-sensitive cell line developed from these double knockdown mice showed that expression of IR and IGF-1R proteins in the Cre-treated cell line were both reduced by about 50% (no statistical analysis of this result provided). In the knockout mice, proteinuria was significantly increased by 6 months, but not at earlier time points. Histologic analysis showed proteinaceous casts, glomerulosclerosis and interstitial fibrosis. Podocyte number was stated to be reduced by about 30% in double knockdown mice, although the method by which this was evaluated seems to have been by counting WT1 positive nuclei in glomerular cross-sections, an approach that is well-known not to be a reliable way of assessing true podocyte number. No information is provided about podocyte size, density or glomerular volume.

Comment: If IR/IGF1R deletion plays a significant role in normal podocyte function sufficient to cause proteinuria and glomerulosclerosis then the effect of reduced IR and IGF1R protein expression on podocyte function would have been expected to produce a phenotype before 6 months. A more likely scenario to explain the overall result is that deleting the IR and IGF1R genes at about embryonic day12 impacted podocyte development to a variable extent such that some mice developed fewer podocytes per glomerulus than other mice. As mice grow and their glomeruli and glomerular capillary area increases, those mice with fewer podocytes would not be able to completely cover the filtration surface with foot processes and would develop proteinuria and glomerulosclerosis. If reduced podocyte number per glomerulus is the proximate cause of the observed proteinuria, then modulation of the body and kidney growth rate by calorie restriction to slow growth (lower circulating IGF-1 levels) would be expected to be protective, while a high protein high calorie diet (higher circulating IGF-1 levels) or uni-nephrectomy to increase kidney growth rate would be expected to enhance proteinuria and glomerulosclerosis.

The model as used may be more representative of a variable degree of podocyte depletion than an effect of impaired IR/IGF1R signaling. Therefore, although the phenotype may be ultimately attributable to the IR/IGF1R gene deletions the proteinuria and glomerulosclerotic phenotype itself was probably a consequence of defective podocyte development. Examining podocyte number, size, density and glomerular volume at earlier time points (4 weeks) would help to answer this question. Therefore, a more appropriate title would be "The insulin/IGF axis is critically important (for) normal podocyte development and deployment". In this context the effect of the knockdowns on splicing would make more sense.

Cell culture studies. A cell line was generated using a temperature sensitive SV40 system that has been previously reported from this laboratory. A detailed analysis is provided to show that double knockout cells exhibited abnormal spliceosome activity. This forms the basis for the conclusion that "The insulin/IGF axis is critically important (for) controlling gene transcription in the podocyte". There are several concerns that weaken this conclusion.

(1) In the double knockdown cell culture system about 30% of cells were "lost" by 3 days and about 70% of cells were "lost" by 5days. The studies were done at the 3 day time point. It is not clear whether "lost" cells were in the process of dying, stress-induced detachment, or just growing more slowly than control due to reduced IR and IGF-1R signaling. These processes could have impacted splicing in a non-specific way independent of IR/IGF1R signaling itself.

(2) Can a single cell line derived from the double floxed mice be relied on to provide an unbiased picture of the effect of deleting IR and IGF-1R? Presumably, the transfection and selection process will select for cells that survive thereby including unknown biases, possibly related to spliceosome function. Is a single cell line adequate? These investigators have extensive experience with this type of analysis, but this question is not addressed in the discussion.

(3) To determine whether the effect is specific to reduced IR/IGFR signaling the deletion of IR and IGF-1R could be corrected by transfecting full length IR and IGF-1R cDNAs into the cells to restore normal IR/IGF1R signaling. If transfected cells with intact IR and IGF-1R expression and activity returns spliceosome activity to normal this would be evidence that receptors themselves play some role in spliceosome activity, as opposed to the downstream effect on growth limitation/stress on the cells.

(4) Other ways of testing whether the splicing effect is specifically due to reduced IR/IGF-1R signaling would be to (a) block IR and IGF1R receptors using available inhibitors, (b) remove or reduce insulin, IGF-1 and IGF-2 levels in the culture medium, (c) use low glucose and amino acid culture medium to slow growth rate independent of receptor function, (d) or block intra-cellular signaling via the IR and IGF-1R receptors through mTORC1 inhibition using rapamycin or other signaling targets.

(5) It would be useful to determine whether the cultured cells stressed in other ways (e.g. ischemia, toxins, etc.) also results in the same splicing abnormalities.

Author response:

The following is the authors’ response to the original reviews

Many thanks for your helpful and constructive comments for our work examining the effect of inhibiting both the insulin receptor (IR) and IGF1 receptor (IGF1R) in the podocyte. We are pleased to submit an updated manuscript addressing your concerns.

(1) A major concern was a lack of mechanistic insight into how deletion (or knock-down) of both receptors caused the spliceosomal phenotype (Reviewer 1 and Reviewer 3).

We now think this is due to the lack of a network of insulin/IGF phospho-signalling events to a variety of spliceosomal proteins and kinases. The reasons for this are as follows:

A. Since submitting our paper Turewicz et al have published a comprehensive phospho-proteomic paper examining the effects of 100nM insulin on human primary myotubes (DOI: 10.1038/s41467-025-56335-6). They discovered that multiple post-translational phosphorylation events occur in a variety of spliceosomal proteins at differing time points (1 minute to 60 minutes). Furthermore, they show that mRNA splicing is rapidly modified in response to insulin stimulation in their cells. This follows elegant work from Bastista et al who studied diabetic and non-diabetic iPSC derived human myositis and also detected a spliceosome phosphorylation signature (DOI: 10.1016/j.cmet.2020.08.007).

B. We have examined phospho-proteosome changes that occur in wild -type podocytes (expressing both the IR and IGF1R) compared to double (IR and IGF1R) knockout cells using phosho-proteomics. We have done this 3 days after inducing receptor knockdown, before major cell loss, and have stimulated the cells with either 10nM insulin or 100mg IGF1.

Interestingly, we detected several post-translational modifications (PTM) in our data set that are also present in Turewicz’s studies. Of note, 100nM insulin (as used by Turewicz) will signal through both the insulin and IGF1 receptor (and hybrid Insulin/IGF1 receptors) which is relevant to our studies.

Our work shows a cascade of phospho- signalling events affecting multiple components of the spliceosomal complex and evidence of kinase modulation (phosphorylation) (New Figure 7 and supplementary Figure 5). Also new results section in paper (lines 391-425 in track changes version). We acknowledge that we only studied a single time point after stimulation (10 minutes) and could have missed other PTM in the spliceosomal complex and other kinases. This is mentioned in our new limitations of study section (lines 595-606). This will be a focus of future work. We did not find major PTM differences when stimulating with either insulin or IGF1 in our studies and suspect that the doses of insulin (10nM) and IGF1 (100mg) used are still able to signal through cognate receptors.

Furthermore, we have examined the relative contributions of the insulin and IGF1 receptor in detail in the model (addressed in point 13 below).

(2) The phenotype of the mouse is only superficially addressed. The main issues are that the completeness of the mouse KO is never assessed nor is the completeness of the KO in cell lines. The absence of this data is a significant weakness. (Reviewer 1)

We apologise for not making this clear, but we did assess the level of receptor knockdown in both the animal and cell models. The in vivo model showed variable and non-complete levels of insulin receptor and IGF1 receptor podocyte knock down (shown in supplementary Figure 1C). This is why we made the in vitro floxed podocyte cell lines in which we could robustly knockdown both the IR and IGF1R. We show this using Western blotting (shown in Figure 2A). We agree that calling the models knockout is misleading and have changed all to knock down (KD) now.

(3) The mouse experiments would be improved if the serum creatinine’s were measured to provide some idea how severe the kidney injury is. (Reviewer 1)

There is variability in creatinine levels which is not uncommon in transgenic mouse models (probably partly due to variability in receptor knock down levels with cre-lox system). This is part of rationale of developing the robust double receptor knockout cell models where we robustly knocked out both receptors by >80%. We have added measured creatinine levels in a subset of mice in supplementary data (New Supplementary Figure 1E) and mention this in the text (lines 285-286). As some mice died we expect they may have developed acute kidney injury, but we did not serially measure the creatinine’s in every mouse over time. We could have assessed the GFR in a more sensitive way to look at differences. However, we consider the highly significant levels of albuminuria and histological damage observed in our models show a significant kidney phenotype.

(4) An attempt to rescue the phenotype by overexpression of SF3B4 would also be useful. If this didn't work, an explanation in the text would suffice. (Reviewer 1).

We did consider doing this but on reflection think it is very unlikely to rescue the phenotype as an array of different spliceosomal proteins quantitatively changed and were differentially phosphorylated / dephosphorylated throughout the complex (as we hope our revised work illustrates now). We think a single protein rescue is highly unlikely to work. We hope this is an appropriate explanation for this action. We have mentioned this in the text now in our discussion (lines 601-602).

(5) As insulin and IGF are regulators of metabolism, some assessment of metabolic parameters would be an optional add-on. (Reviewer 1).

Thank you for this suggestion. We did not extensively examine the metabolism of the mice however we did perform blood glucose measurement and weight which are included in the paper (Figure 1A and Figure 1B).

(6) The authors should caveat the cell experiments by discussing the ramifications of studying the 50% of the cells that survive vs the ones that died. (Reviewer 1).

We appreciate this and this was the rationale behind cells being studied after 3 days differentiation for total and phospho-proteomics before significant cell loss to avoid the issue of studying the 50% of cells that survive (which happened at 7 days). We have made this clearer in the manuscript. We also have added the data showing less cell death at 3 days in the cell model (New Supp Figure 2B).

(7) It would be helpful to say that tissue scoring was performed by an investigator masked to sample identity. (Reviewer 2)

We did this and have added to manuscript (line 113).

(8) Data are presented as mean/SEM. In general, mean/SD or median/IQR are preferred to allow the reader to evaluate the spread of the data. There may be exceptions where only SEM is reasonable. (Reviewer 2)

All graphs have now been changed to SD rather than SEM.

(9) It would be useful to for the reader to be told the number of over-lapping genes (with similar expression between mouse groups) and the results of a statistical test comparing WT and KO mice. The overlap of intron retention events between experimental repeats was about 30% in both knock-out podocytes. This seems low and I am curious to know whether this is typical for this method; a reference could be helpful. (Reviewer 2)

This is an excellent question. We had 30% overlap as the parameters used for analysis were very stringent. We suspect we could get more than 30% by being less stringent, which still be considered as similar events if requested. Our methods were based on FLAIR analysis (PMID: 32188845). We have added this reference to the manuscript (Line 242 & 680).

(10) With the GLP1 agonists providing renal protection, there is great interest in understanding the role of insulin and other incretins in kidney cell biology. It is already known that Insulin and IGFR signaling play important roles in other cells of the kidney. So, there is great interest in understanding these pathways in podocytes. The major advance is that these two pathways appear to have a role in RNA metabolism, the major limitations are the lack of information regarding the completeness of the KO's. If, for example, they can determine that in the mice, the KO is complete, that the GFR is relatively normal, then the phenotype they describe is relatively mild. (Reviewer 1)

Thank you. The receptor knock-out (KO) in the mice is highly unlikely to be complete (Please see comments above and Supplementary Figure 1C). There are many examples of “KO” animal models targeting other tissues showing that complete KO of these receptors seems difficult to achieve, particularly in reference to the IGF1 receptor. In the brain, which also contains terminally differentiated cells, barely 50% of IGF1R knockdown was achieved in the target cells (PMID:28595357). In ovarian granulosa cells (PMID:28407051) -several tissue specific drivers tried but couldn't achieve any better than 80%. The paper states that 10% of IGF1R is sufficient for function in these cells so they conclude that their knockdown animals are probably still responding to IGF1. Finally, in our recent IGF1R podocyte knockdown model we found Cre levels were important for excision of a single homozygous floxed gene (PMID: 38706850) hence we were not surprised that trying to excise two homozygous floxed genes (insulin receptor and IGF1 receptor) was challenging. This was the rationale for making the double receptor knockout cell lines to understand processes / biology in more detail. As stated earlier, we have changed our description of the mice and cell lines from knock-out to knock-down throughout the revised manuscript as this is more accurate.

(11) For the in vivo studies, the only information given is for mice at 24 weeks of age. There needs to be a full-time course of when the albuminuria was first seen and the rate of development. Also, GFR was not measured. Since the podocin-Cre utilized was not inducible, there should be a determination of whether there was a developmental defect in glomeruli or podocytes. Were there any differences in wither prenatal post-natal development or number of glomeruli? (Reviewer 3)

We have added further urinary Albumin:creatinine ratio (uACR) data at 12, 16 and 20 weeks to manuscript. We do not think there was a major developmental phenotype as albuminuria did not become significantly different until several months of age (new Supp Figure 1B). We did consider using a doxycycline inducible model but we know the excision efficiency is much less than the constitutive podocin-cre driven model Author response image 1. This would likely give a very mild (if any) phenotype when attempting to knockout both receptors and not reveal the biology adequately. We acknowledge the weaknesses of the animal model and this was the rationale for generating the cell models.

(12) Although the in vitro studies are of interest, there are no studies to determine if this is the underlying mechanism for the in vivo abnormalities seen in the mice. Cultured podocytes may not necessarily reflect what is occurring in podocytes in vivo. (Reviewer 3)

This is a good point. We have now immune-stained the DKD and WT mice for Sf3b4 (a spliceosomal change in our in vitro proteomics) and also find a significant reduction in this protein in podocytes of the DKD mice (New Figure 3F).

(13) Given that both receptors are deleted in the podocyte cell line, it is not clear if the spliceosome defect requires deletion of both receptors or if there is redundancy in the effect. The studies need to be repeated in podocyte cell lines with either IR or IGFR single deletions. (Reviewer 3)

We have now performed proteomics and phospho-proteomics in all 4 cell types (Wild-type, Insulin receptor knock down, IGF1R knockdown and double knockdown) at 3 days (New Figure 8 and supplementary Figure 6. Also new results section lines 425 to 450). This shows that both receptors contribute to the pathways (and hence there is a high level of compensation built into the system). For total proteins we detected that spliceosomal tri-snRNP was only reduced when both receptors were lacking but other proteins / pathways had an incremental effect of losing the insulin or IGF1 receptor. Likewise, the spliceosomal phospho-signaling events can go through either the insulin or igf1 receptors predominantly or through both. We think this reflects the complexity of this system and how evolutioatily it has developed in mammals to protect against its loss.

Finally in revision we have rewritten the discussion with a “limitations of the study” section and hopefully in an easier to read fashion for the readership.

Author response image 1.

(A) mT/mG reporter mouse crossed to constitutional podocin Cre heterozygous mouse. Illustrates podocyte specificity for Cre driver and excision Of reporter Figure shows GFP expression in Cre producing cells (top panel scale bar=250vm; bottom panel scale bar=50pm). Cre expression causes GFP to be switched on. (B) mT/mG reporter mouse crossed to podocin RtTA— tet-o-cre heterozygous mouse shows podocyte specificity for driver and approximately 60% excision. (top and bottom panels scale bar=250pm; middle panel scale bar=50pm). Doxycycline required for expression showing not leaky.

eLife Assessment

In this valuable study, through carefully executed and rigorously controlled experiments, the authors challenged a previously reported role of the Death Receptor 6 (DR6/Tnfrsf21) in Wallerian degeneration (WD). Using two DR6 knockout mouse lines and multiple WD assays, both in vitro and in vivo, the authors provided convincing evidence that loss of DR6 in mice does not protect peripheral axons from WD after injury, at least in the specific contexts of the mice and analyses performed in this study. Due to the lack of certain specific parameters from previous studies (sex, age, mouse strains etc.), the exact reasons underlying the observed inconsistencies between current and previous reports on the protective effects of DR6 remains to be determined. Overall, this is a carefully executed study providing invaluable information toward understanding DR6's role (or lack thereof) in axon degeneration.

Reviewer #1 (Public review):

Summary:

The authors show that genetic deletion of the orphan tumor necrosis factor receptor DR6 in mice does not protect peripheral axons against degeneration after axotomy. Similarly, Schwann cells in DR6 mutant mice react to axotomy similarly to wild type controls. These negative results are important because previous work has indicated that loss or inhibition of DR6 is protective in disease models and also against Wallerian degeneration of axons following injury. This carefully executed counterexample is important for the field to consider.

Strengths:

A strength of the paper is the use of two independent mouse strains that knockout DR6 in slightly different ways. The authors confirm that DR6 mRNA is absent in these models (western blots for DR6 protein are less convincingly null, but given the absence of mRNA, this is likely an issue of antibody specificity). One of the DR6 knockout strains used is the same strain used in a previous paper examining the effects of DR6 on Wallerian degeneration.

The authors use a series of established assays to evaluate axon degeneration, including light and electron microscopy on nerve histological samples and cultured dorsal root ganglion neurons in which axons are mechanically severed and degeneration is scored in time lapse microscopy. These assays consistently show a lack of effect of loss of DR6 on Wallerian degeneration in both mouse strains examined.

Additional strengths are that the authors examine both the axonal response and the Schwann cell response to axotomy and use both in vivo and in vitro assays.

Therefore, these experiments, the author's data support their conclusion that loss of DR6 does not protect against Wallerian degeneration.

Weaknesses:

A weakness of this paper is that no effort is made to determine why the results presented here may differ from previous studies. A notable possibility is that the original mouse strain that showed 5 of 13 mice being protected from Wallerian degeneration was studies on a segregating C57BL/6.129S background.

Finally, it is important to note that previously reported effects of DR6 inhibition, such as protection of cultured cortical neurons from beta-amyloid toxicity, are not necessarily the same as Wallerian degeneration of axons distal to an injury studied here. The negative results presented here showing that loss of DR6 is not protective against Wallerian degeneration induced by injury are important given the interest in DR6 as a therapeutic target. However, care should be taken in attempting to extrapolate these results to other disease contexts such as ALS or Alzheimer's disease.

Reviewer #3 (Public review):

Summary:

The authors revisit the role of DR6 in axon degeneration following physical injury (Wallerian degeneration), examining both its effects on axons and its role in regulating the Schwann cell response to injury. Surprisingly, and in contrast to previous studies, they find that DR6 deletion does not delay the rate of axon degeneration after injury, suggesting that DR6 is not a mediator of this process.

Overall, this is a valuable study. As the authors note, the current literature on DR6 is inconsistent, and these results provide useful new data and clarification. This work will help other researchers interpret their own data and re-evaluate studies related to DR6 and axon degeneration.

Strengths:

(1) The use of two independent DR6 knockout mouse models strengthens the conclusions, particularly when reporting the absence of a phenotype.

(2) The focus on early time points after injury addresses a key limitation of previous studies. This approach reduces the risk of missing subtle protective phenotypes and avoids confounding results with regenerating axons at later time points after axotomy.

Comments on revisions:

I thank the authors for their thorough responses to my previous comments. The revisions have addressed the points raised and have improved the clarity and overall quality of the manuscript. I appreciate the effort taken to strengthen the presentation of the work.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary:

The authors show that genetic deletion of the orphan tumor necrosis factor receptor DR6 in mice does not protect peripheral axons against degeneration after axotomy. Similarly, Schwann cells in DR6 mutant mice react to axotomy similarly to wild-type controls. These negative results are important because previous work has indicated that loss or inhibition of DR6 is protective in disease models and also against Wallerian degeneration of axons following injury. This carefully executed counterexample is important for the field to consider.

Strengths:

A strength of the paper is the use of two independent mouse strains that knock out DR6 in slightly different ways. The authors confirm that DR6 mRNA is absent in these models (western blots for DR6 protein are less convincingly null, but given the absence of mRNA, this is likely an issue of antibody specificity). One of the DR6 knockout strains used is the same strain used in a previous paper examining the effects of DR6 on Wallerian degeneration.

The authors use a series of established assays to evaluate axon degeneration, including light and electron microscopy on nerve histological samples and cultured dorsal root ganglion neurons in which axons are mechanically severed and degeneration is scored in time-lapse microscopy. These assays consistently show a lack of effect of loss of DR6 on Wallerian degeneration in both mouse strains examined.

Therefore, in the specific context of these experiments, the author's data support their conclusion that loss of DR6 does not protect against Wallerian degeneration.

Weaknesses:

(1) The major weaknesses of this paper include the tone of correcting previously erroneous results and the lack of reporting on important details around animal experiments that would help determine whether the results here really are discordant with previous studies, and if so, why.

The authors do not report the genetic strain background of the mice used, the sex distributions of their experimental cohorts, or the age of the mice at the time the experiments were performed. All of these are important variables.

(Response 1) We thank the reviewer for emphasizing the importance of reporting the sex, age, and genetic background of the experimental animals used in our axon protection analyses. We have incorporated this information into the revised manuscript wherever available. The sole exception concerns the genetic background of the conditional DR6 mice generated by Genentech, which remains unknown. The original publication describing these mice (Tam et al., 2012, Dev Cell, PMID 22340501) did not report this information, and we were unable to obtain it directly from Genentech. Details regarding the genetic background of the Wld^S and aPhr1 mutant mice are provided in their respective original publications, which are cited in our manuscript. Because the Gamage et al. study from the Deppmann laboratory did not report the sex or age of the animals used, we were unable to assess whether these variables might contribute to the differences observed between the two studies. Moreover, we are not aware of published evidence identifying sex or age as modifiers of structural axon preservation in axotomized peripheral nerve stumps in mouse models of delayed Wallerian degeneration. Furthermore, in the original publications describing the phenotypes of transgenic Nmnat2 and Wld^S mice, as well as Sarm1 or Phr1 knockout mice, sex and age of the animals used in the Wallerian degeneration assays were not reported (PMIDs 23995269, 12106171, 22678360, 23665224). Although, to our knowledge, no large-scale systematic studies have been conducted, over the last 15 years we have never observed any sex-based differences in Wallerian degeneration phenotypes in these mutants exhibiting prominent axon protection. This topic was discussed informally at conferences, and we are also not aware of other investigators having observed such effects.

In response to the reviewer’s comment regarding “tone”, we made sure that our data and interpretations are presented in a professional, balanced, and objective manner, including a detailed discussion of potential alternative explanations for the discrepant findings.

(2) The DR6 knockout strain reported in Gamage et al. (2017) was on a C57BL/6.129S segregating background. Gamage et al. reported that loss of DR6 protected axons from Wallerian degeneration for up to 4 weeks, but importantly, only in 38.5% (5 out of 13) mice they examined. In the present paper, the authors speculate on possible causes for differences between the lack of effect seen here and the effects reported in Gamage et al., including possible spontaneous background mutations, epigenetic changes, genetic modifiers, neuroinflammation, and environmental differences. A likely explanation of the incomplete penetrance reported by Gamage et al. is the segregating genetic background and the presence of modifier loci between C57BL/6 and 129S. The authors do not report the genetic background of the mice used in this study, other than to note that the knockout strain was provided by the group in Gamage et al. However, if, for example, that mutation has been made congenic on C57BL/6 in the intervening years, this would be important to know. One could also argue that the results presented here are consistent with 8 out of 13 mice presented in Gamage et al.

(Response 2) As noted above, we now provide information on the genetic background of the mice in the revised manuscript, where available. We have not backcrossed the constitutive DR6 knockout mice obtained from the Deppmann laboratory (Gamage et al.) to a C57BL/6 background; our colony was maintained primarily through intercrosses of heterozygous animals. Similarly, the conditional DR6 mutant mice used in this study were also not backcrossed to C57BL/6 mice.

We respectfully hold a different view regarding the reviewer’s final point. We understand it is not appropriate to infer consistency between two datasets by disregarding the subset of results that do not align. By the same logic, it would be flawed to draw conclusions from the Gamage et al. study based solely on the single Wld^S mouse out of five that did not show axon preservation after nerve injury. Selectively omitting conflicting data does not provide a valid basis for establishing phenotype concordance across studies.

To further strengthen our study, we note that we performed additional analyses on three more nerve samples from constitutive DR6 null mice during the revision process and have incorporated the resulting data in Fig. 1.

(3) Age is also an important variable. The protective effects of the spontaneous WldS mutation decrease with age, for example. It is unclear whether the possible protective effects of DR6 also change with age; perhaps this could explain the variable response seen in Gamage et al. and the lack of response seen here.

(Response 3) As discussed above, we now provide the age information for the mice used for the Wallerian degeneration assessments in the respective figure legends. To our knowledge, there are no prior reports suggesting that age is a significant determinant of structural axon preservation in the indicated mutants. Electrophysiological function and neuromuscular junction preservation decrease with age in axotomized Wld^S mice (e.g., PMIDs 12231635, 19158292, 15654865), but these parameters are not subject of our study, and we have not studied them. Unfortunately, a direct comparison of ages between our DR6 mutant mice and those used in Gamage et al. (2017) is not possible, as the earlier study from the Deppmann laboratory did not report this information.

(4) It is unclear if sex is a factor, but this is part of why it should be reported.

(Response 4) We now report the requested sex information for our axon preservation analyses during nerve injury-induced Wallerian degeneration in the DR6 mouse models in Figs. 1 and 2.

(5) The authors also state that they do not see differences in the Schwann cell response to injury in the absence of DR6 that were reported in Gamage et al., but this is not an accurate comparison. In Gamage et al., they examined Schwann cells around axons that were protected from degeneration 2 and 4 weeks post-injury. Those axons had much thinner myelin, in contrast to axons protected by WldS or loss of Sarm1, where the myelin thickness remained relatively normal. Thus, Gamage et al. concluded that the protection of axons from degeneration and the preservation of Schwann cell myelin thickness are separate processes. Here, since no axon protection was seen, the same analysis cannot be done, and we can only say that when axons degenerate, the Schwann cells respond the same whether DR6 is expressed or not.

(Response 5) We appreciate the reviewer’s detailed comments. Our intention was not to directly compare our findings with those of Gamage et al. regarding the myelin behavior at these time points (because we never observed axon protection), but rather to note that we did not observe any DR6-dependent alterations in Schwann cell responses under conditions where axons undergo normal Wallerian degeneration. As the reviewer correctly points out, Gamage et al. analyzed Schwann cell myelin surrounding axons that were protected from degeneration for extended periods, a context fundamentally different from the complete lack of axon protection observed in our DR6-deficient models. Therefore, the specific dissociation between axon preservation and myelin maintenance claimed by Gamage et al. cannot be evaluated in our study. A statement to make this point clearer has been incorporated in the revised manuscript.

We fully agree with the reviewer’s concluding point: in our experiments, once axons degenerate, Schwann cell responses proceed similarly regardless of DR6 expression. This agreement reinforces one of the central conclusions of our work.

(6) The authors also take issue with Colombo et al. (2018), where it was reported that there is an increase in axon diameter and a change in the g-ratio (axon diameter to fiber diameter - the axon + myelin) in peripheral nerves in DR6 knockout mice. This change resulted in a small population of abnormally large axons that had thinner myelin than one would expect for their size. The change in g-ratio was specific to these axons and driven by the increased axon diameter, not decreased myelin thickness, although those two factors are normally loosely correlated. Here, the authors report no changes in axon size or g-ratio, but this could also be due to how the distribution of axon sizes was binned for analysis, and looking at individual data points in supplemental figure 3A, there are axons in the DR6 knockout mice that are larger than any axons in wild type. Thus, this discrepancy may be down to specifics and how statistics were performed or how histograms were binned, but it is unclear if the results presented here are dramatically at odds with the results in Colombo et al. (2018).

(Response 6) Several points raised by the reviewer appear to reflect differences in interpretation of the findings reported in Colombo et al. (2018). That study did not report altered myelination in DR6 null mice at stages when myelination is largely complete (P21). Instead, modest changes were observed at P1, which were reduced by P7, and P21 mutants were reported to be indistinguishable from controls. No analyses of peripheral nerves in older animals were presented, and the authors concluded in the discussion that myelination in young adult DR6 null mice appears normal. In contrast, our analysis of constitutive DR6 null mice at P1 does not reproduce the increase in the number of myelinated fibers per unit area reported by Colombo et al. We obtained similar results in the independent conditional DR6 knockout mouse line. Differences in nerve tissue processing, embedding, staining, or in the microscopic imaging and quantification of thinly myelinated axons in P1 sciatic nerve cross-sections may have contributed to the observed discrepancy. However, because the relevant methodological details were not described in Colombo et al., the underlying reasons for these differences cannot be determined and remain speculative.

(7) Finally, it is important to note that previously reported effects of DR6 inhibition, such as protection of cultured cortical neurons from beta-amyloid toxicity, are not necessarily the same as Wallerian degeneration of axons distal to an injury studied here. The negative results presented here, showing that loss of DR6 is not protective against Wallerian degeneration induced by injury, are important given the interest in DR6 as a therapeutic target, but they are specific to these mice and this mechanism of induced axon degeneration. The extent to which these findings contradict previous work is difficult to assess due to the lack of detail in describing the mouse experiments, and care should be taken in attempting to extrapolate these results to other disease contexts, such as ALS or Alzheimer's disease.

(Response 7) We agree with the reviewer’s point and emphasize that our manuscript carefully differentiates our data regarding the function of DR6 in Wallerian degeneration from the potential involvement of DR6 in other forms of axon degeneration. Our findings do not conflict with previous work on DR6 in the context of in vitro beta-amyloid and prion toxicity as well as in vitro models of ALS and multiple sclerosis. We believe these distinctions are explicitly and appropriately articulated throughout the entire manuscript and in more detail in the discussion section.

Reviewer #1 (Recommendations for the authors):

(1) The authors should include additional information about the mice used, including strain background for both the DR6 mice and the Cre transgenes crossed into the DR6 conditional knockout, the age of the mice when the nerve crush experiments were performed, and the sex distributions of the experimental cohorts. This information is critical for reproducibility in animal experiments, and that point is compounded here, where the major focus of this paper is taking issue with the reproducibility of previous work.

(Response 8) This information has been included in the revision. See above responses.

(2) In the abstract, reference 5 is cited as a study on the response to Schwann cells to injury in a DR6 background, but this probably should be reference 10.

(Response 9) This typo has been corrected.

(3) "Site-by-site comparison" in line 201 should be side-by-side?

(Response 10) This typo has been corrected.

(4) The paper contains a lot of self-evaluative wording, "surprising contrast," "compelling evidence," "robust results." Whether those adjectives apply should be for the reader to decide, and a drier, more objective tone in the presentation would improve the paper.

(Response 11) We agree that excessive self-evaluative wording can weaken objectivity. In the manuscript, such phrasing is used sparingly and intentionally to highlight differences from previously published studies, guide the reader, and convey scholarly judgment. We do not consider this limited use to be counterproductive. The adjectives “surprising,” “compelling,” and “robust” each appear only one to three times across the entire manuscript, and the specific phrase “robust results” does not appear at all.

(5) In Figure 2A, DR6-/-, there is no significant difference, but there is also a lot of variability, and one could argue the authors are seeing axon protection comparable to WldS in 40% of their samples (2/5), which is very similar to Gamage et al.

(Response 12) We respectfully disagree with this reasoning as it relies on selectively emphasizing only a subset of the data. Please also see our response #2 for more detailed discussion.

(6) Overall, the data presented here are convincing and support the conclusions drawn, but the paper needs to focus more on the negative results at hand and less on bashing previous studies, particularly when the results presented here do definitively show that the previous studies were incorrect and plausible explanations for differences in outcome exist.

(Response 13) We have carefully revisited the wording of the manuscript and are confident that our emphasis remains on the central negative finding that DR6 does not regulate axon degeneration and Schwann cell injury responses during Wallerian degeneration. We do not believe the manuscript “bashes” previous studies; nonetheless, we thoroughly re-examined all relevant sections to ensure that our language is neutral, accurate, and non-inflammatory. We believe the current phrasing presents our interpretations in an appropriately balanced, objective, and professional manner.

Reviewer #2 (Public review):

Summary:

This manuscript by Beirowski, Huang, and Babetto revisits the proposed role of Death Receptor 6 (DR6/Tnfrsf21) in Wallerian degeneration (WD). A prior study (Gamage et al., 2017) suggested that DR6 deletion delays axon degeneration and alters Schwann cell responses following peripheral nerve injury. Here, the authors comprehensively test this claim using two DR6 knockout mouse models (the line used in the earlier report plus a CMV-Cre derived floxed ko line) and multiple WD assays in vivo and in vitro, aligned with three positive controls, Sarm1 WldS and Phr1/Mycbp2 mutants. Contrary to the prior findings, they find no evidence that DR6 deletion affects axon degeneration kinetics or Schwann cell dynamics (assessed by cJun expression or [intact+degenerating] myelin abundance after injury) during WD. Importantly, in DRG explant assays, neurites from DR6-deficient mice degenerated at rates indistinguishable from controls. The authors conclude that DR6 is dispensable for WD, and that previously reported protective effects may have been due to confounding factors such as genetic background or spontaneous mutations.

Strengths:

The authors employ two independently generated DR6 knockout models, one overlapping with the previously published study, and confirm loss of DR6 expression by qPCR and Western blotting. Multiple complementary readouts of WD are applied (structural, ultrastructural, molecular, and functional), providing a robust test of the hypothesis.

Comparisons are drawn with established positive controls (WldS, SARM1, Phr1/Mycbp2 mutants), reinforcing the validity of the assays.

By directly addressing an influential but inconsistent prior report, the manuscript clarifies the role of DR6 and prevents potential misdirection of therapeutic strategies aimed at modulating WD in the PNS. The discussion thoughtfully considers possible explanations for the earlier results, including colony-specific second-site mutations that could explain the incomplete penetrance of the earlier reported phenotype of only 36%.

Weaknesses:

(1) The study focuses on peripheral nerves. The manuscript frequently refers to CNS studies to argue for consistency with their findings. It would be more accurate to frame PNS/CNS similarities as reminiscences rather than as consistencies (e.g., line 205ff in the Discussion).

(Response 14) Axon protection in all key genetic models of delayed axon degeneration, including Wld^S, SARM1, Phr1/Mycbp2 mutants, has been demonstrated in both the peripheral and central nervous systems. This observation supports the view that core molecular mechanisms regulating axon degeneration are conserved across neuronal populations throughout the entire nervous system. We have scrutinized the wording in our manuscript and are not aware that we frequently refer to CNS studies in regards to axon degeneration. Nevertheless, we have replaced the term “consistent” to avoid potential ambiguity when we discuss the earlier study showing normal Wallerian degeneration in the optic nerves from DR6 knockout mice.

(2) The DRG explant assays are convincing, though the slight acceleration of degeneration in the DR6 floxed/Cre condition is intriguing (Figure 4E). Could the authors clarify whether this is statistically robust or biologically meaningful?

(Response 15) We thank the reviewer for noting this aspect of our in vitro data in Fig. 4. The difference observed in the DR6 floxed/Cre condition is statistically significant at the 6h time point following disconnection, as indicated by the p value shown in Fig. 4E. However, a similarly statistically significant acceleration of axon degeneration was not observed in DRG axotomy experiments using constitutive DR6 knockout preparations, although a trend toward more rapid axon breakdown is apparent at 6 h post-axotomy (Fig. 4B). These observations may suggest reduced stability of DR6-deficient axons in this specific neuron-only in vitro context. Further investigation would be required to determine the biological significance of this effect. In contrast, our in vitro quantitative analyses of the initiation and early phases of Wallerian degeneration (Fig. 2) revealed no evidence of accelerated axon disintegration in the DR6 mutant mouse models, highlighting potential differences between in vitro and in vitro systems.

(3) In the summary (line 43), the authors refer to Hu et al. (2013) (reference 5) as the study that previously reported AxD delay and SC response alteration after injury. However, this study did not investigate the PNS, and I believe the authors intended to reference Gamage et al. (2017) (reference 10) at this point.

(Response 16) Thanks for pointing this out. We have corrected this typo in the revised manuscript.

(4) In line 74ff of the results section, the authors claim that developmental myelination is not altered in DR6 mutants at postnatal day 1. However, the variability in Figure S2 appears substantial, and the group size seems underpowered to support this claim. Colombo et al. (2018) (reference 11) reported accelerated myelination at P1, but this study likewise appears underpowered. Possible reasons for these discrepancies and the large variability could be that only a defined cross-sectional area was quantified, rather than the entire nerve cross-section.

(Response 17) We confirm that the quantification of thinly myelinated axons was performed on entire sciatic nerves from P1 mouse pups, as described in the methods section in our original manuscript. The data shown in Fig. S2 were obtained from 5-9 pups per experimental group. Sample sizes were determined based on a priori power analyses using pilot data, which indicated that a minimum of five biological replicates was sufficient to detect statistically significant differences with acceptable confidence. Comparable sample sizes have been used in our previous studies and by other groups to assess early postnatal myelination (e.g., PMIDs 21949390, 28484008). Several published studies have reported analyses using 3-4 animals per group (e.g., PMIDs 28484008, 25310982, 29367382). For comparison, the study by Colombo et al. used 3-8 pups for the analysis presented in their Fig. 3. We note that the apparent variability in Fig. S2 may be accentuated by the scaling of the y-axis, which was chosen to ensure that individual data points are clearly resolved and visible.

(5) The authors stress the data of Gamage et al. (2017) on altered SC responses in DR6 mutants after injury. They employed cJun quantification to show that SC reprogramming after injury is not altered in DR6 mutants. This approach is valid and the conclusion trustworthy. Here, the addition of data showing the combined abundance of intact and degenerated myelin does not add much insight. However, Gamage et al. (2017) reported altered myelin thickness in a subset of axons at 14 days after injury, which is considerably later than the time points analyzed in the present study. While, in the Reviewer's view, the thin myelin observed by Gamage et al. in fact resembles remyelination, the authors may wish to highlight the difference in the time points analyzed.

(Response 18) We consider the additional quantification of the area occupied by intact myelin and myelin debris to provide complementary information that supports the c-Jun-based conclusion that Schwann cell injury responses are normal in DR6-deficient nerves following lesion. We agree with this reviewer that the thin myelin observed by Gamage et al. resembles remyelination, raising the possibility that axon regeneration occurred into the distal nerve stump at the studied 14d post-injury time point (see their Fig. 3). This may have been interpreted as axon protection in this study. In our study, it was impossible to examine such myelin effects since axon protection was never observed in any of the DR6 mutant models at any of the time point we investigated. We have incorporated appropriate additional text to highlight this difference. See also response #5 above.

Reviewer #3 (Public review):

Summary:

The authors revisit the role of DR6 in axon degeneration following physical injury (Wallerian degeneration), examining both its effects on axons and its role in regulating the Schwann cell response to injury. Surprisingly, and in contrast to previous studies, they find that DR6 deletion does not delay the rate of axon degeneration after injury, suggesting that DR6 is not a mediator of this process.

Overall, this is a valuable study. As the authors note, the current literature on DR6 is inconsistent, and these results provide useful new data and clarification. This work will help other researchers interpret their own data and re-evaluate studies related to DR6 and axon degeneration.

Strengths:

(1) The use of two independent DR6 knockout mouse models strengthens the conclusions, particularly when reporting the absence of a phenotype.

(2) The focus on early time points after injury addresses a key limitation of previous studies. This approach reduces the risk of missing subtle protective phenotypes and avoids confounding results with regenerating axons at later time points after axotomy.

Weaknesses:

(1) The study would benefit from including an additional experimental paradigm in which DR6 deficiency is expected to have a protective effect, to increase confidence in the experimental models, and to better contextualize the findings within different pathways of axon degeneration. For example, DR6 deletion has been shown in more than one study to be partially axon protective in the NGF deprivation model in DRGs in vitro. Incorporating such an experiment could be straightforward and would strengthen the paper, especially if some of the neuroprotective effects previously reported are confirmed.

(Response 19) We thank the reviewer for these suggestions. We would like to highlight that our study addresses the role of DR6 in Wallerian degeneration, whereas in vitro NGF deprivation has been used to model developmental axon pruning. Previous work indicates fundamental biological differences between these regressive pathways regulating the stereotyped removal of axon segments. We feel that studying this alternative form of axon degeneration is beyond the scope of the current work and could be addressed in a separate manuscript. Although additional tests will be needed, we note that our preliminary data using samples from both DR6 knockout mouse models suggest no axon protection after NGF-deprivation in DRG neuron preparations in our hands (deprivation of the growth factor and administration of anti-NGF antibody).

(2) The quality of some figures could be improved, particularly the EM images in Figure 2. As presented, they make it difficult to discern subtle differences.

(Response 20) We have pseudocolored intact (turquoise) and degenerated (magenta) myelinated fibers on the high-resolution semithin micrographs (not electron micrographs) in the new Fig. 2 to make the distinction between the two fiber categories clearer.

Reviewer #3 (Recommendations for the authors):

(1) Line 121: The authors mention toluidine blue staining, but it does not appear to be shown in Figure S5.

(Response 21) This appears to be a misunderstanding. Fig. S5A shows the ultrastructure of dedifferentiated Schwann cells in transmission electron micrographs, while Figs. S5B and C show quantification of the area occupied by myelin sheaths and myelin debris profiles on osmium tetroxide and toluidine blue stained nerve sections from the two DR6 mutant models, based on semithin light microscopy. These are two different aspects of the analysis. The text has been modified in the revised manuscript to make the distinction clearer.

(2) Line 175: The authors should add NMNAT2 to the list of enzymes implicated in the regulation of Wallerian degeneration in mammals.

(Response 22) Nmnat2 and a literature reference (Milde et al., 2013) has been incorporated in the discussion of the revised manuscript to address this point.

(3) Line 201: Please correct the typo "site-by-site" to "side-by-side."

(Response 23) This typo has been corrected.

eLife Assessment

This fundamental work significantly advances our understanding of how contact-dependent antagonism enables keystone bacteria to establish and maintain their niche over time. The evidence obtained is convincing, supporting most of the conclusions drawn. This work will be of significant interest to the microbiome research community.

Reviewer #1 (Public review):

Summary:

In this study, the authors investigate the physiological role of the Type VI secretion system (T6SS) in a naturally evolved gut microbiome derived from wild mice (the WildR microbiome). Focusing on Bacteroides acidifaciens, the authors use newly developed genetic tools and strain-replacement strategies to test how T6SS-mediated antagonism influences colonization, persistence, and fitness within a complex gut community. They further show that the T6SS resides on an integrative and conjugative element (ICE), is distributed among select community members, and can be horizontally transferred, with context-dependent effects on colonization and persistence. The authors conclude that the T6SS stabilizes strain presence in the gut microbiome while imposing ecological and physiological constraints that shape its value across contexts.

This study is likely to have a significant impact on the microbiome field by moving experimental tests of T6SS function out of simplified systems and into a naturally co-evolved gut community. The WildR system, together with the strain replacement strategy, ICE-seq approach, and genetic toolkit, represents a powerful and reusable platform for future mechanistic studies of microbial antagonism and mobile genetic elements in vivo.

The datasets, including isolate genomes, metagenomes, and ICE distribution maps, will be a valuable community resource, particularly for researchers interested in strain-resolved dynamics, horizontal gene transfer, and ecological context dependence. Even where mechanistic resolution is incomplete, the work provides a strong experimental foundation upon which such questions can be directly addressed.

Overall, this study occupies a space between system building and mechanistic dissection. The authors demonstrate that the T6SS influences persistence and community structure in vivo, but the physiological basis of these effects remains unresolved. Interpreting the results as evidence of fitness costs or selective advantage, therefore, requires caution, as multiple ecological and host-mediated processes could produce similar abundance trajectories.

Placing the findings within the broader literature on microbial antagonism, particularly work emphasizing measurable costs, benefits, and tradeoffs, would help readers better contextualize what is directly demonstrated here versus what remains an open question. Viewed in this light, the principal contribution of the study is to show that such questions can now be addressed experimentally in a realistic gut ecosystem.

Strengths:

A major strength of this study is that it directly interrogates the physiological role of the T6SS in a naturally evolved gut microbiome, rather than relying on simplified pairwise or in vitro systems. By working within the WildR community, the authors advance beyond descriptive surveys of T6SS prevalence and address function in an ecologically relevant context.

The authors provide clear genetic evidence that Bacteroides acidifaciens uses a T6SS to antagonize co-resident Bacteroidales, and that loss of T6SS function specifically compromises long-term persistence without affecting initial colonization. This temporal separation is well designed and supports the conclusion that the T6SS contributes to maintenance rather than establishment within the community.

Another strength is the identification of the T6SS on an integrative and conjugative element (ICE) and the demonstration that this element is distributed among, and exchanged between, community members. The use of ICE-seq to track distribution and transfer provides strong support for horizontal mobility and adds mechanistic depth to the study.

Finally, the transfer of the T6SS-ICE into Phocaeicola vulgatus and the observation of context-dependent colonization benefits followed by decline is a compelling result that moves the study beyond simple "T6SS is beneficial" narratives and highlights ecological contingency.

Weaknesses:

Despite these strengths, there is a mismatch between the precision of the claims and the precision of the measurements, particularly regarding fitness costs, physiological burden, and the mechanistic role of the T6SS.

First, while the authors conclude that the T6SS "stabilizes strain presence" and that its value is constrained by fitness costs, these costs are not directly measured. Persistence, abundance trajectories, and eventual loss are informative outcomes, but they do not uniquely identify fitness tradeoffs. Decline could arise from multiple non-exclusive mechanisms, including community restructuring, host-mediated effects, incompatibilities of the ICE in new hosts, or ecological retaliation, none of which are disentangled here.

Second, the manuscript frames the T6SS as having a defined physiological role, yet the data do not resolve which physiological processes are under selection. The experiments demonstrate that T6SS activity affects persistence, but they do not distinguish whether this occurs via direct killing, resource release, niche modification, or higher-order community effects. As a result, "physiological role" remains underspecified and risks being conflated with ecological outcome.

Third, although the authors emphasize context dependence, the study offers limited quantitative insight into what aspects of context matter. Differences between native and recipient hosts, or between early and late colonization phases, are described but not mechanistically interrogated, making it difficult to generalize beyond the specific cases examined.

Fourth is the lack of engagement with recent experimental literature demonstrating functional roles of the T6SS beyond simple interference competition. While the authors focus on persistence and competitive outcomes, they do not adequately situate their findings within recent work demonstrating that T6SS-mediated antagonism can serve additional physiological functions, including resource acquisition and DNA uptake, thereby linking killing to measurable benefits and tradeoffs. The absence of this literature makes it difficult to place the authors' conclusions about physiological role and fitness cost within the current conceptual framework of the field. Without this context, the physiological interpretation of the results remains incomplete, and alternative functional explanations for the observed dynamics are underexplored.

A further limitation concerns the taxonomic scope of the functional analysis. The authors state that the role of the T6SS in the murine environment is functionally investigated using genetically tractable Bacteroides species, citing the lack of genetic tools for Mucispirillum schaedleri. While this is a reasonable, practical choice, it means that a substantial fraction of T6SS-encoding species in the WildR community are not experimentally interrogated. Consequently, conclusions about the role of the T6SS in the murine gut necessarily reflect the subset of taxa that are genetically accessible and may not fully capture community-level or niche-specific functions of T6SS activity. Given that M. schaedleri is represented as a metagenome-assembled genome, its isolation and genetic manipulation would be technically challenging. Nonetheless, explicitly acknowledging this limitation and slightly tempering claims of generality would strengthen the manuscript.

Finally, several interpretations would benefit from more cautious language. In particular, claims invoking fitness costs, selective advantage, or physiological burden should be explicitly framed as inferences from persistence dynamics, rather than as direct measurements, unless supported by additional quantitative fitness or growth assays.

Reviewer #2 (Public review):

Summary:

In this study, the authors set out to determine how a contact-dependent bacterial antagonistic system contributes to the ability of specific bacterial strains to persist within a complex, native gut community derived from wild animals. Rather than focusing on simplified or artificial models, the authors aimed to examine this system in a biologically realistic setting that captures the ecological complexity of the gut environment. To achieve this, they combined controlled laboratory experiments with animal colonization studies and sequencing-based tracking approaches that allow individual strains and mobile genetic elements to be followed over time.

Strengths:

A major strength of the work is the integration of multiple complementary approaches to address the same biological question. The use of defined but complex communities, together with in vivo experiments, provides a strong ecological context for interpreting the results. The data consistently show that the antagonistic system is not required for initial establishment but plays a critical role in long-term strain persistence. This insight that moves beyond traditional invasion-based views of microbial competition. The observation that transferable genetic elements can confer only temporary advantages, and may impose longer-term costs depending on community context, adds important nuance to current understanding of microbial fitness.

Weaknesses:

Overall, there is not a lack of evidence, but a deliberate trade-off between ecological realism and mechanistic resolution, which leaves some causal pathways open to interpretation.

Reviewer #3 (Public review):

Summary:

Shen et al. investigate the contribution of the type VI secretion system of Bacteroidales in the gut microbiome assembly and targeting of closely related species. They demonstrate that B. acidifaciens relies on T6SS-mediated antagonism to prevent displacement by co-resident Bacteroidales and other members of the microbiome, allowing B. acidifaciens to persist in the gut.

Strengths:

Using a gnotobiotic model colonized with a wild-mouse microbiome is a significant strength of this study. This approach allows tracking of microbiome changes over time and directly examining targeting by Bacteroidales carrying T6SS in a more natural setting. The development of ICE-seq for mapping the distribution of the T6SS in the microbiome is remarkable, enabling the study of how this bacterial weapon is transferred between microbiome members without requiring long-read metagenomics methods.

Weaknesses:

Some conclusions are based on only four mice per condition. The author should consider increasing the sample size.

Overall, the authors successfully achieved their objectives, and their experimental design and results support their findings. As mentioned in the discussion, it would be important to investigate the role of the T6SS in resilience to disturbances in the microbiome, such as antibiotics, diet, or pathogen invasion. This work represents a step forward in understanding how contact-dependent competition influences the gut microbiome in relevant ecological contexts.

Author response:

We appreciate that the reviewers provided an overall positive assessment of our manuscript and offered constructive suggestions for improvement. All three reviewers noted that a key strength of our study is the implementation of a gut microbiome model for the characterization of interbacterial antagonism pathways such as the type VI secretion system (T6SS) that approaches natural complexity. They note our work represents a significant advance in microbiome research, and generates resources that will be of use to many researchers in the field. Two of the reviewers point out that the complexity of our model limits the nature of measurements we can make, and suggest we temper the strength of the some of the conclusions we draw. As noted in more detail below, in our revised manuscript, we will be more precise in the wording we use to characterize our findings, and we will be more explicit about what the measurements we are able to make allow us to conclude about the physiological role of the T6SS in the gut microbiome.

Reviewer #1 (Public review):

Summary:

In this study, the authors investigate the physiological role of the Type VI secretion system (T6SS) in a naturally evolved gut microbiome derived from wild mice (the WildR microbiome). Focusing on Bacteroides acidifaciens, the authors use newly developed genetic tools and strain-replacement strategies to test how T6SS-mediated antagonism influences colonization, persistence, and fitness within a complex gut community. They further show that the T6SS resides on an integrative and conjugative element (ICE), is distributed among select community members, and can be horizontally transferred, with context-dependent effects on colonization and persistence. The authors conclude that the T6SS stabilizes strain presence in the gut microbiome while imposing ecological and physiological constraints that shape its value across contexts.

This study is likely to have a significant impact on the microbiome field by moving experimental tests of T6SS function out of simplified systems and into a naturally co-evolved gut community. The WildR system, together with the strain replacement strategy, ICE-seq approach, and genetic toolkit, represents a powerful and reusable platform for future mechanistic studies of microbial antagonism and mobile genetic elements in vivo.

The datasets, including isolate genomes, metagenomes, and ICE distribution maps, will be a valuable community resource, particularly for researchers interested in strain-resolved dynamics, horizontal gene transfer, and ecological context dependence. Even where mechanistic resolution is incomplete, the work provides a strong experimental foundation upon which such questions can be directly addressed.

Overall, this study occupies a space between system building and mechanistic dissection. The authors demonstrate that the T6SS influences persistence and community structure in vivo, but the physiological basis of these effects remains unresolved. Interpreting the results as evidence of fitness costs or selective advantage, therefore, requires caution, as multiple ecological and host-mediated processes could produce similar abundance trajectories.

Placing the findings within the broader literature on microbial antagonism, particularly work emphasizing measurable costs, benefits, and tradeoffs, would help readers better contextualize what is directly demonstrated here versus what remains an open question. Viewed in this light, the principal contribution of the study is to show that such questions can now be addressed experimentally in a realistic gut ecosystem.

We thank the reviewer for this thoughtful summary of our study. We were glad to read they conclude our work will have a significant impact on the microbiome field and that the resources we have developed will be of value to the community.

Strengths:

A major strength of this study is that it directly interrogates the physiological role of the T6SS in a naturally evolved gut microbiome, rather than relying on simplified pairwise or in vitro systems. By working within the WildR community, the authors advance beyond descriptive surveys of T6SS prevalence and address function in an ecologically relevant context.

The authors provide clear genetic evidence that Bacteroides acidifaciens uses a T6SS to antagonize co-resident Bacteroidales, and that loss of T6SS function specifically compromises long-term persistence without affecting initial colonization. This temporal separation is well designed and supports the conclusion that the T6SS contributes to maintenance rather than establishment within the community.

Another strength is the identification of the T6SS on an integrative and conjugative element (ICE) and the demonstration that this element is distributed among, and exchanged between, community members. The use of ICE-seq to track distribution and transfer provides strong support for horizontal mobility and adds mechanistic depth to the study.

Finally, the transfer of the T6SS-ICE into Phocaeicola vulgatus and the observation of context-dependent colonization benefits followed by decline is a compelling result that moves the study beyond simple "T6SS is beneficial" narratives and highlights ecological contingency.

We appreciate this detailed and nuanced characterization of the strengths of our study.

Weaknesses:

Despite these strengths, there is a mismatch between the precision of the claims and the precision of the measurements, particularly regarding fitness costs, physiological burden, and the mechanistic role of the T6SS.

We acknowledge that in some places, our manuscript could benefit from greater precision in the language we use when linking the outcomes we observe in our study to their potential underlying causes. Specific revisions we propose to address this concern are described below.

First, while the authors conclude that the T6SS "stabilizes strain presence" and that its value is constrained by fitness costs, these costs are not directly measured. Persistence, abundance trajectories, and eventual loss are informative outcomes, but they do not uniquely identify fitness tradeoffs. Decline could arise from multiple non-exclusive mechanisms, including community restructuring, host-mediated effects, incompatibilities of the ICE in new hosts, or ecological retaliation, none of which are disentangled here.

We agree that multiple mechanisms could explain why P. vulgatus carrying a T6SS-encoding ICE declines over time. Our use of the term “fitness cost” to describe this trend was not meant to imply any particular underlying mechanism, but was rather our attempt to characterize the phenotypic outcome we observed in simplified terms. We note that ecological context is an important determinant of the fitness cost or benefit of any given trait, and our study sheds light on the importance of the presence of the WildR community and the mouse intestinal environment to the fitness contribution of the ICE to P. vulgatus. Nonetheless, to avoid implying an overly simplistic interpretation of our results, we propose to modify the language used in the manuscript when describing the contribution of the T6SS to species persistence in WildR-colonized mice.

Second, the manuscript frames the T6SS as having a defined physiological role, yet the data do not resolve which physiological processes are under selection. The experiments demonstrate that T6SS activity affects persistence, but they do not distinguish whether this occurs via direct killing, resource release, niche modification, or higher-order community effects. As a result, "physiological role" remains underspecified and risks being conflated with ecological outcome.

We acknowledge that our study does not fully resolve the physiological processes under selection that mediate role of the T6SS in maintaining B. acidifaciens populations in WildR-colonized mice. Indeed, several of the outcomes of T6SS activity the reviewer lists, such as target cell killing and nutrient release, are inextricably linked and thus inherently difficult to disentangle. We note that we did attempt to measure higher-order community effects of T6SS activity with metagenomic sequencing, but acknowledge that this approach may not have been sufficiently sensitive to detect small community shifts mediated by a relatively low-abundance species. To address the concern that our current framing implies more of a mechanistic understanding that our study achieves, we propose to substitute “ecological” for “physiological” where appropriate when summarizing our key findings.

Third, although the authors emphasize context dependence, the study offers limited quantitative insight into what aspects of context matter. Differences between native and recipient hosts, or between early and late colonization phases, are described but not mechanistically interrogated, making it difficult to generalize beyond the specific cases examined.

We are not entirely clear what the reviewer means by “differences between native and recipient hosts”, but we agree that additional quantitative studies will be needed to address the generalizability of our findings. Future studies are also needed to address the mechanistic basis for the difference in the benefit conferred by the T6SS that we observed between P. vulgatus and B. acidifaciens.

Fourth is the lack of engagement with recent experimental literature demonstrating functional roles of the T6SS beyond simple interference competition. While the authors focus on persistence and competitive outcomes, they do not adequately situate their findings within recent work demonstrating that T6SS-mediated antagonism can serve additional physiological functions, including resource acquisition and DNA uptake, thereby linking killing to measurable benefits and tradeoffs. The absence of this literature makes it difficult to place the authors' conclusions about physiological role and fitness cost within the current conceptual framework of the field. Without this context, the physiological interpretation of the results remains incomplete, and alternative functional explanations for the observed dynamics are underexplored.

We thank the reviewer for specifically highlighting the potential pertinence of this literature to our study. Indeed, we did not cite studies indicating a link between T6SS activity and the uptake of DNA and other resources released by targeted cells. As we note above, the release of intracellular contents from target cells is an inevitable consequence of the delivery of lytic effectors. Thus, distinguishing between fitness benefits conferred from the elimination of competitor species and those arising from scavenging the nutrients released during this process is not straightforward. Measuring the benefits deriving from the uptake of certain released molecules, such as DNA, was not immediately feasible in the system employed in this study and instead we focused on the direct lytic consequences of the effectors delivered via the T6SS. We will revise our Discussion to include reference to how downstream consequences of T6SS activity on target cells could impact the community, and thus the adaptive role of the T6SS in the microbiome.

A further limitation concerns the taxonomic scope of the functional analysis. The authors state that the role of the T6SS in the murine environment is functionally investigated using genetically tractable Bacteroides species, citing the lack of genetic tools for Mucispirillum schaedleri. While this is a reasonable, practical choice, it means that a substantial fraction of T6SS-encoding species in the WildR community are not experimentally interrogated. Consequently, conclusions about the role of the T6SS in the murine gut necessarily reflect the subset of taxa that are genetically accessible and may not fully capture community-level or niche-specific functions of T6SS activity. Given that M. schaedleri is represented as a metagenome-assembled genome, its isolation and genetic manipulation would be technically challenging. Nonetheless, explicitly acknowledging this limitation and slightly tempering claims of generality would strengthen the manuscript.

The reviewer points out that studying the T6SS activity in M. schadleri would potentially expand the generality of our claims. We agree that having an isolate of this species along with genetic tools for its manipulation would allow us to probe the importance of the T6SS in the gut microbiome more broadly. At the suggestion of the reviewer, we will add explicit mention for the need to develop such tools, an endeavor that lies outside of the scope of the current study.

Finally, several interpretations would benefit from more cautious language. In particular, claims invoking fitness costs, selective advantage, or physiological burden should be explicitly framed as inferences from persistence dynamics, rather than as direct measurements, unless supported by additional quantitative fitness or growth assays.

We agree with the reviewer that invoking fitness costs, selective advantages or physiological burdens should be done cautiously, and in our revised manuscript we will carefully re-evalute our usage of those terms. However, we would also argue invoking fitness costs and benefits when describe strain persistence dynamics in mice has substantial precedent in the literature ((Feng et al. 2020, Brown et al. 2021, Park et al. 2022, Segura Munoz et al. 2022), to list a handful of representative examples published by different groups). It is unclear to us what additional in vivo growth measurements could be taken to substantiate our claim that the T6SS provides a fitness benefit to B. acidifaciens during prolonged gut colonization, or that carrying the ICE imposes a fitness cost on P. vulgatus during long-term colonization. Our in vitro experiments evaluating the competitiveness conferred by T6SS activity provide a measure of insight into its fitness benefits, but as our in vivo strain persistence data and the work of many others show, in vitro measurements do not necessarily capture in vivo parameters.

Reviewer #2 (Public review):

Summary:

In this study, the authors set out to determine how a contact-dependent bacterial antagonistic system contributes to the ability of specific bacterial strains to persist within a complex, native gut community derived from wild animals. Rather than focusing on simplified or artificial models, the authors aimed to examine this system in a biologically realistic setting that captures the ecological complexity of the gut environment. To achieve this, they combined controlled laboratory experiments with animal colonization studies and sequencing-based tracking approaches that allow individual strains and mobile genetic elements to be followed over time.

Strengths:

A major strength of the work is the integration of multiple complementary approaches to address the same biological question. The use of defined but complex communities, together with in vivo experiments, provides a strong ecological context for interpreting the results. The data consistently show that the antagonistic system is not required for initial establishment but plays a critical role in long-term strain persistence. This insight that moves beyond traditional invasion-based views of microbial competition. The observation that transferable genetic elements can confer only temporary advantages, and may impose longer-term costs depending on community context, adds important nuance to current understanding of microbial fitness.

We thank the reviewer for the positive feedback and are glad they agree our study provides new insight into the role of interbacterial antagonism in natural communities.

Weaknesses:

Overall, there is not a lack of evidence, but a deliberate trade-off between ecological realism and mechanistic resolution, which leaves some causal pathways open to interpretation.

The reviewer makes a good point that the complexity of the experimental system we employ precludes some lines of experimentation that would yield more mechanistic information. As the reviewer notes, we were aware of the tradeoff between mechanistic resolution and ecological realism when selecting our experimental system. Our deliberate choice to favor biological complexity over mechanistic clarity in this study stemmed from our perception that a major gap in understanding of the T6SS and other antagonism pathways lies in defining their ecological function in complex microbial communities.

Reviewer #3 (Public review):

Summary:

Shen et al. investigate the contribution of the type VI secretion system of Bacteroidales in the gut microbiome assembly and targeting of closely related species. They demonstrate that B. acidifaciens relies on T6SS-mediated antagonism to prevent displacement by co-resident Bacteroidales and other members of the microbiome, allowing B. acidifaciens to persist in the gut.

Strengths:

Using a gnotobiotic model colonized with a wild-mouse microbiome is a significant strength of this study. This approach allows tracking of microbiome changes over time and directly examining targeting by Bacteroidales carrying T6SS in a more natural setting. The development of ICE-seq for mapping the distribution of the T6SS in the microbiome is remarkable, enabling the study of how this bacterial weapon is transferred between microbiome members without requiring long-read metagenomics methods.

We thank the reviewer for their enthusiasm toward our study.

Weaknesses:

Some conclusions are based on only four mice per condition. The author should consider increasing the sample size.

We agree that in some experiments it would be beneficial to increase the sample size from four mice. However, the experiments we performed for this study are time and resource intensive. Additionally, the experiments on which we base our primary conclusions were all independently replicated with similar results. Given these factors, we determined that the extra confidence that might be afforded by increasing our sample size did not merit the delay in publication and investment in resources that would be required.

Overall, the authors successfully achieved their objectives, and their experimental design and results support their findings. As mentioned in the discussion, it would be important to investigate the role of the T6SS in resilience to disturbances in the microbiome, such as antibiotics, diet, or pathogen invasion. This work represents a step forward in understanding how contact-dependent competition influences the gut microbiome in relevant ecological contexts.

We agree that investigating the role of the T6SS during perturbations of the microbiome is a key next step for this work and thank the reviewer for highlighting this important future direction.

References

Brown, E. M., H. Arellano-Santoyo, E. R. Temple, Z. A. Costliow, M. Pichaud, A. B. Hall, K. Liu, M. A. Durney, X. Gu, D. R. Plichta, C. A. Clish, J. A. Porter, H. Vlamakis and R. J. Xavier (2021). "Gut microbiome ADP-ribosyltransferases are widespread phage-encoded fitness factors." Cell Host Microbe 29(9): 1351-1365 e1311.

Feng, L., A. S. Raman, M. C. Hibberd, J. Cheng, N. W. Griffin, Y. Peng, S. A. Leyn, D. A. Rodionov, A. L. Osterman and J. I. Gordon (2020). "Identifying determinants of bacterial fitness in a model of human gut microbial succession." Proc Natl Acad Sci U S A 117(5): 2622-2633.

Park, S. Y., C. Rao, K. Z. Coyte, G. A. Kuziel, Y. Zhang, W. Huang, E. A. Franzosa, J. K. Weng, C. Huttenhower and S. Rakoff-Nahoum (2022). "Strain-level fitness in the gut microbiome is an emergent property of glycans and a single metabolite." Cell 185(3): 513-529 e521.

Segura Munoz, R. R., S. Mantz, I. Martinez, F. Li, R. J. Schmaltz, N. A. Pudlo, K. Urs, E. C. Martens, J. Walter and A. E. Ramer-Tait (2022). "Experimental evaluation of ecological principles to understand and modulate the outcome of bacterial strain competition in gut microbiomes." ISME J 16(6): 1594-1604.

eLife Assessment

This study reports an important and novel finding that TENT5A, an enzyme involved in fine-tuning poly(A) tail length on selected mRNAs, is required for proper enamel mineralization in mice. The evidence supporting the authors' conclusion that reduced expression of enamel matrix proteins (EMPs) in TENT5A-deficient mice results from shortened poly(A) tails remains incomplete, as TENT5A may possess additional functions independent of post-transcriptional regulation that are not addressed in the current study.

Reviewer #1 (Public review):

Summary:

The authors aim to determine whether TENT5A, a post-transcriptional regulator previously implicated in bone formation, also plays a role in enamel development. Using a mouse model lacking TENT5A, they report hypomineralized enamel with structural defects, accompanied by reduced expression, altered poly(A) tail length, and impaired secretion of enamel matrix proteins, particularly amelogenin. By combining ultrastructural imaging, transcriptomics, direct RNA sequencing, and protein localization analyses, the study proposes that TENT5A promotes cytoplasmic polyadenylation and translation of a subset of extracellular matrix transcripts required for enamel biomineralization.

Strengths:

A major strength of this work is its conceptual novelty. To my knowledge, this is the first study to demonstrate that a non-canonical poly(A) polymerase plays a direct role in enamel development, extending post-transcriptional regulation by cytoplasmic polyadenylation from bone to enamel, a biologically distinct and non-regenerative mineralized tissue. The identification of amelogenin as a dominant, tissue-specific target provides a new perspective on how enamel matrix production is regulated beyond transcriptional control.

In addition, the study is supported by a comprehensive and complementary set of approaches linking molecular changes to tissue-level phenotypes. The use of direct RNA sequencing provides strong evidence for selective regulation of poly(A) tail length in specific transcripts rather than global effects on mRNA metabolism, and the phenotypic analyses convincingly connect altered post-transcriptional regulation to defects in enamel structure and mineralization.

Weaknesses:

Although the data support a role for TENT5A in stabilizing and promoting translation of amelogenin and related transcripts, the mechanism underlying substrate specificity remains unresolved. Poly(A) tail length alone does not explain why certain transcripts are regulated while others are not, and the proposed involvement of protein partners or RNA processing steps remains speculative. This limitation should be more clearly framed as an open question rather than an emerging mechanism.

A further limitation is the lack of direct human genetic or clinical evidence linking TENT5A to enamel defects. In humans, loss-of-function variants in TENT5A are known to cause a recessive form of osteogenesis imperfecta, but TENT5A has not been associated with amelogenesis imperfecta or other enamel phenotypes. This limits immediate translational interpretation of the mouse enamel phenotype and highlights the need for future human genetic or clinical studies.

Finally, the manuscript does not address whether other members of the TENT5 family are expressed in ameloblasts or could compensate for the loss of TENT5A, leaving open questions about redundancy and specificity within this family.

Reviewer #2 (Public review):

Summary:

The manuscript by Aranaz-Novaliches describes a study of Tent5a knockout (KO) mice. The authors demonstrate a severe enamel phenotype in these mice, characterized by hypoplastic enamel with markedly disturbed organization of enamel rods. Additionally, they report that Amelx expression is reduced in the mutant compared to wild type (WT) at both mRNA and protein levels. The authors also examine the distribution and co-localization of Amelx and Ambn in ameloblasts and the enamel matrix. These findings are novel and provide important insights into the role of polyadenylation in regulating enamel matrix protein translation and its downstream effects on protein trafficking, secretion, and enamel formation. However, I have multiple concerns regarding the data and its analysis that need to be addressed.

Specific comments:

(1) Introduction

The structure of the introduction is unconventional. The first sentence of the third paragraph states that the goal of this study is to investigate the role of TENT5A in enamel formation, but the rest of the paragraph focuses on enamel in general. The following paragraph claims that the authors discovered the effects of Tent5a deficiency on enamel formation for the first time, yet most of the paragraph discusses enamel proteins and amelogenesis. The choice of references is problematic. The authors cite Sire et al. (2007), which focuses on the origin and evolution of enamel mineralisation genes, a poor fit for this context. A more appropriate source would be a recent review, e.g., Lacruz R et al., Physiol Rev. 2017;97(3):939-993. Ambn constitutes ~5% of the enamel matrix, not 10%. Reference 16 (Martin) is not ideal for murine enamel; more detailed studies exist, e.g., Smith CE et al., J Anat. 2019;234(2):274-290. References on protein-protein interactions (17-19) are also off: Wald et al. studied Ambn-Ambn and Amelx-Amelx interactions separately; Fang et al. focused on Amelx self-assembly only; Kawasaki and Weiss addressed gene evolution. The authors should cite work from Moradian-Oldak's lab, which clearly demonstrates Amelx-Ambn interactions. The last paragraph contains confusing statements, e.g., "TENT5a localized in rER promotes the expression of AmelX and other secreted protein transcripts." Also, the manuscript does not convincingly show disruption of self-assembly beyond overall enamel disorganization.

(2) Results

(a) microCT

Quantitative microCT analyses of WT and KO enamel are needed. At a minimum, enamel thickness and density should be measured from at least three biological replicates per genotype. Severe malocclusion in KO mice is not discussed. The mandibular incisor appears abraded, while the maxillary incisor is overgrown. Is maxillary enamel as affected as mandibular? The age of the mice is not specified. High-resolution scans of isolated mandibular incisors described in Materials and Methods should be included.

(b) SEM

The term "disorganized crystal structure" is incorrect - SEM cannot reveal crystal structure. This requires electron/X-ray diffraction or vibrational spectroscopy. Likely, the authors meant disorganized rods and interrod enamel. The phrase "weak HAP composition" is unclear. Can the increase in interprismatic matrix volume and reduction in rod diameter be quantified? Since rods are secreted by distal Tomes' processes and interrod by proximal Tomes' processes, an imbalance may indicate alterations in the ameloblast secretory apparatus. TEM studies of demineralized incisors are recommended to assess ameloblast ultrastructure.

(c) EMP expression

There is a discrepancy between WB images and data in Figure S2a. In Figure 2b, Amelx band is stronger than Ambn (expected, as Amelx is ~20× more abundant), but in Figure S2a, Ambn appears higher. How was protein intensity in Fig. S2a calculated? Optical density? Was normalization applied? Co-localization in Figure 2d was performed on LS8 cells, which lack a true ameloblast phenotype. Amelx expression in LS8 cells is ~2% of actin (Sarkar et al., 2014), whereas in murine incisors, it is ~600× higher than actin (Bui et al., 2023). Ambn signal is weaker than Amelx, which may affect co-localization results.

(d) Splicing products in Figure 2e

All isoforms except one contain exon 4. The major functional splice product of Amelx lacks exon 4 (Haruyama et al. J Oral Biosci. 2011;53(3):257-266), and there are some indications that the presence of exon 4 can lead to enamel defects. Can it be that the observed phenotype is due to the presence of exon 4?

(e) Co-localization studies

The presented co-localization studies do not demonstrate self-assembly defects; they reflect enamel microstructural defects observed by SEM. Self-assembly occurs at the nanoscale and cannot be assessed by light microscopy except with advanced optical methods. Conclusions based on single images are weak. The authors should perform experiments at least on three biological replicates per genotype, quantify results (e.g., total gray values per ROI of equal pixel size), and use co-localization metrics such as Mander's coefficient. Claims about alternative secretory pathways require much stronger evidence.

The authors should avoid implying that mRNA is inside the ER lumen. It is likely associated with the outer rER surface, which is expected. The resolution of the methods used is insufficient to confirm ER lumen localization.

Reviewer #3 (Public review):

Summary:

It is well established that poly(A) tails at the 3' end of mRNA are critical for mRNA stability, providing another layer of gene regulation. TENT5A is one of the non-canonical poly(A) polymerases that add an extra poly(A) tail. This manuscript demonstrates that the Tent5A mutation leads to mineralization abnormalities in the tooth, shorter poly(A) tails in amelogenin mRNA and some other selected mRNAs, and provides a list of TENT5A interacting proteins.

Strengths:

(1) The authors show in vivo genetic evidence that Tent5a is critical for normal tooth mineralization.

(2) The authors show that the length of the poly(A) tail in amelogenin (AmelX) is 13 bases shorter in Tent5a mutants but not in other mRNAs, such as ameloblastin (Ambn).

(3) Differentially expressed genes (DEGs) in Tent5A mutant tissues (cervical loop) are identified, and some of them show different lengths of poly(A) tails.

(4) TENT5A interacting proteins are identified. Together with the DEGs, these datasets will provide valuable research tools to the community.

Weaknesses:

(1) There is no direct evidence to support the main conclusion; the length of the poly(A) tail is critical for normal tooth mineralization.

(2) The RNAseq data to identify TENT5A substrate is based on the assumption that shorter poly(A) tailed RNA is less stable. However, there are multiple reasons for the differential expression of RNA in Tent5A mutant tissues.

(3) Several TENT5A-interacting proteins have been identified, but, beyond their colocalization with a target mRNA, no mechanistic studies have been conducted.

Author response:

We thank the editors and reviewers for their careful and constructive evaluation of our manuscript. We appreciate the recognition of the conceptual novelty and in vivo relevance of our findings. We have carefully considered all comments and outline below the major revisions and additional analyses we will undertake. For clarity, we address the reviewers’ comments in thematic sections.

Cell-autonomous contribution of Tent5a to phenotype

We agree that the use of a complete knockout model raises the possibility of indirect or non-cell-autonomous effects on tooth development, particularly given the observed dentin alterations. To address this point directly, we are generating and analyzing an ameloblast-specific conditional model we have already on shelf (Ambn-Cre; Tent5a^flox/flox) to determine whether the enamel phenotype arises from cell-autonomous loss of TENT5A in the secretory epithelium. This approach will allow us to distinguish epithelial-intrinsic effects from potential secondary contributions of odontoblasts or mesenchymal tissues. Results from this model will be incorporated into the revised manuscript.

Mechanistic basis and substrate specificity

We agree that the mechanism underlying substrate selectivity of TENT5A requires further clarification. We have performed multiple classical RNA–protein interaction assays, including CLIP-based approaches, without identifying a clear sequence-specific recognition motif. In the revised manuscript, we will present substrate specificity as an open mechanistic question rather than implying a defined recognition mechanism.

To strengthen this aspect, we will extend our analysis to include combined immunoprecipitation strategies and investigation of potential ribosome-associated or co-translational interactions of TENT5A.

In addition, we will further validate selected high-confidence TENT5A interactors identified in our dataset in context of putative changes in AmelX-polyA tail length.

Poly(A) tail length and functional causality

We acknowledge that shortening of the poly(A) tail alone does not formally establish causality. However, our data consistently show that TENT5A-dependent shortening of poly(A) tails correlates with reduced mRNA and protein levels of key enamel matrix components. In the revised manuscript, we will clarify this mechanistic framework more explicitly, integrating poly(A) length, transcript abundance, and protein-level data in a structured manner, while clearly distinguishing correlation from formal proof of causality.

We will also perform additional functional assays, including mRNA stability measurements in vitro in cells with genetic ablation of Tent5a, to further test the link between poly(A) shortening and reduced AmelX protein levels.

Quantitative microCT and enamel morphology

We will include quantitative microCT analyses of enamel thickness and mineral density from multiple biological replicates per genotype (n ≥ 3). Sample numbers will be explicitly stated throughout. Additional high-resolution scans of isolated incisors will be provided. We will also quantify occlusal angle and include whole-skull reconstructions to document malocclusion. Maxillary enamel will be analyzed and quantified alongside mandibular enamel.

SEM terminology will be corrected (e.g., replacing “crystal structure” with “rod/interrod organization”), and structural parameters such as rod diameter and interprismatic matrix proportion will be quantitatively assessed.

We agree that ultrastructural analysis of ameloblast secretory morphology is important. We have experience with TEM analysis of demineralized incisors and will perform additional ultrastructural examination to assess the integrity of Tomes’ processes and the secretory apparatus in Tent5a-deficient ameloblasts. These data will allow us to distinguish between primary alterations in secretory morphology and downstream effects on matrix organization.

Amelx splice variants

We will re-analyze our RNA-seq data with specific attention to exon 4-containing isoforms and clarify the distribution of splice variants in WT and KO samples. These findings will be explicitly discussed in the context of prior literature.

Co-localization and self-assembly claims

We agree that conventional light microscopy cannot directly resolve nanoscale self-assembly events. In Figure 3, our intention was to demonstrate differential subcellular distribution and partial segregation of AMELX and AMBN within secretory compartments, rather than to claim direct visualization of molecular self-assembly. In the revised manuscript, we will clarify this distinction, moderate the terminology accordingly, and provide explicit quantitative co-localization analyses across multiple biological replicates.

TENT5 family paralogs

To address potential redundancy within the TENT5 family, we will analyze published single-cell RNA-seq datasets (Sharir et al., 2019; Krivanek et al., 2020) to assess expression of TENT5 paralogs in ameloblasts. These findings will be validated using targeted transcriptional analyses.

Human clinical relevance

We appreciate the suggestion to examine potential human enamel phenotypes. We will pursue retrospective analysis of clinical and imaging data from patients carrying TENT5A variants through our collaborations with rare disease networks and specialized centers in Europe and the United States. Any relevant findings will be incorporated into the revised manuscript.

Tissue sampling clarification

We apologize for imprecise terminology regarding transcriptomic sampling. The analyzed tissue corresponds to the proximal incisor region up to the mineralization stage. We will include a schematic and clarify nomenclature throughout the manuscript.

Language and data clarity

The manuscript will be thoroughly revised for clarity, consistency of terminology, figure referencing, and accuracy of citations. We will explicitly clarify the methodology used for protein quantification, including normalization strategy and densitometric analysis, to address inconsistencies noted in the supplementary data. We will also expand the discussion to address the biological relevance of moderate poly(A) shortening, referencing established literature demonstrating that even subtle changes in tail length can significantly influence translational efficiency.

Although AMELX is the most abundant enamel matrix protein and exhibits a consistent TENT5A-dependent poly(A) shortening phenotype, our data demonstrate that multiple secreted proteins are similarly affected. We will revise the text to clearly articulate that the enamel phenotype likely reflects the combined contribution of multiple TENT5A-regulated secretory factors rather than a single-gene effect.

We believe these revisions will substantially strengthen the mechanistic, quantitative, and conceptual framework of the study and provide a clearer foundation for interpreting TENT5A-dependent regulation of enamel biomineralization.

eLife Assessment

This important study characterizes a cascade of neural processes triggered by memory-based prediction errors. The study uses an impressive collection of approaches and methods to characterize and measure cognitive control, arousal, and memory changes as a function of memory-based violations. The analyses are technically sophisticated and rigorous and, taken together, provide solid evidence that there are multiple processes accompanying prediction errors, and that they differentially relate to successful encoding. The manuscript would be much improved by the addition of a discussion or visual schematic that integrates the numerous findings together into a more coherent model.

Reviewer #1 (Public review):

Summary:

This manuscript describes a multi-modal study of associative learning and memory in humans that combines scalp EEG, pupillometry and behavioral analysis to explore the construct of mnemonic prediction errors (MPEs), in terms of their relationship to attention and cognitive control. Across two pooled studies, participants performed associative memory tasks in which they learned the relationship between a cue word (action verb) and a subsequent picture (animate or inanimate) with a strong vs. weak (4 or 1 repetitions) encoding manipulation. At test, participants were encouraged to generate a prediction following the cue word to determine whether the subsequently presented picture was a match or a mismatch. The timecourse of pupillary responses during match decisions was decomposed using temporal principal components analysis, which identified 6 distinct and overlapping processes. Some of the components (PC3/PC4) exhibited sensitivity to both the strength and mismatch conditions, as well as behavior (both RT and accuracy) and retrieval success on the subsequent trial. Furthermore, relationships were also observed between pupillary responses (specifically for PC4) and both frontal theta and posterior alpha power measures obtained from scalp EEG in Experiment 2, as well as for frontal theta and subsequent learning from mismatch stimuli (assessed using subsequent memory findings from a surprise recognition test). The authors suggest the findings indicate that MPEs elicit changes in attention, arousal and cognitive control which impact subsequent learning.

Strengths:

This manuscript has many strengths, including a clever study design, thoughtful integration of multiple neurocognitive measures, and a set of rigorous and technically sophisticated analyses, which reveal a large set of relationships among the measures and behavior. The findings demonstrating brain/physiology-behavior relationships are particularly important, in that they point to potential functional consequences of MPES.

Weaknesses:

The technical proficiency and complexity of the study and analysis also present a clear limitation and challenge for interpretation. As a reader, even those who are quite knowledgeable about the methods, constructs, and questions being addressed will often struggle (as this reviewer did) to keep the large set of findings in mind and gain an understanding of how they all fit together.

Indeed, it seems like there are many threads running together in the paper, which makes it challenging to find the through-line of the key findings, or to understand how they might relate to some pre-existing hypotheses, rather than merely interesting patterns detected in the data. In the Introduction and Discussion, it seems as if the key question is to understand the pathways by which MPEs impact cognition, but this is a rather broad topic, so it is not clear exactly what the authors are aiming at with this question and study design.

As an example, authors operationalize frontal theta power as an index of cognitive control demand, and one of the pathways by which MPEs impact cognition. But this point becomes somewhat circular, since it is not clear how or why the Mismatch x Strength interaction in frontal theta reflects that demand. It would have been better to set this pattern up in the Introduction as a theoretically driven hypothesis, since it currently appears more like a post-hoc interpretation. This is mirrored by how the issue is first brought up in the Introduction, where it states somewhat vaguely: "whether MPEs are followed by an increase in frontal theta... warrants closer examination". Later in the results, there are findings relating frontal theta to pupil dilation, posterior alpha suppression and then subsequent memory. It was hard to understand how all the findings might be linked together functionally or conceptually. Are the authors potentially postulating a mediating or mechanistic pathway, in which the MPE leads to increased cognitive control (frontal theta), which then leads to enhanced subsequent memory of those events? If this is the case, then maybe a formal path analysis would be the best way to test or state this hypothesis. It would also be useful to specify more clearly how the pupil components and alpha suppression factor into this mediating path, since it was not clear.

Relatedly, the authors suggest that internal attention and arousal also play relevant roles in this pathway, but these are also not clear. In some cases, it is stated as if this is a distinct pathway from the cognitive control one, since there is a focus in the results on the independence of frontal theta and posterior alpha, but elsewhere they seem to be treated as two aspects, or distinct steps, within a single pathway. Again, these different threads of the findings were quite challenging for the reader to follow. Pathway analyses, such as with multiple mediation or moderated mediation, could be a useful way to address this question. For example, it seems as if readiness-to-remember is another behavioral outcome (like subsequent memory) that could be used in the search for mediators.

At the minimum, it would be quite helpful to have diagrammatic figures that specify the hypothesized and observed relationships between independent variables (Strength, Mismatch), physiological indices (pupil dilation components, frontal theta, posterior alpha) and key outcome measures (accuracy, RT, next-trial retrieval success, subsequent memory), so that the reader can refer back to them as each component of the analyses is conducted.

Minor Points:

Many figures had x-axes showing a pupil component or EEG power metric broken down by quartile or quintile. Yet nowhere is it ever explained why this graphical (or analytic?) approach is used and what it reflects, or how it is decided which break down to use (quartile/quintile). If the data are analyzed as a correlation, why is a scatterplot not shown instead?

It was surprising that, unlike readiness-to-remember, which was analyzed via logistic regression and odds-ratio, subsequent memory was not analyzed in the same fashion (i.e., as a binary outcome variable predicted by frontal theta), rather than in a reverse chronological one (subsequent memory predicting frontal theta). Historically, it was the case that subsequent memory was analyzed in this manner, but that was before the era in which trial-level linear mixed-effect models were in wide usage, as they are implemented in this study. Thus, the choice seems like a wasted opportunity or a step backwards analytically.

Reviewer #2 (Public review):

Summary:

The authors studied cognitive control and attention in response to mnemonic prediction errors (MPEs): situations in which the external reality violates internal memory-based predictions. The behavioral task first established strong versus weak predictions, and then either confirmed or violated these predictions. The authors examined markers of cognitive control (frontal theta) and attention (posterior alpha suppression, pupil response) while strong and weak predictions were confirmed or violated. They found increased cognitive control (frontal theta) for strong MPEs, which correlated with subsequent memory. Markers of attention (alpha suppression, pupil response) also accompanied strong MPEs but did not correlate with subsequent memory. Pupil response was investigated using an interesting approach that decomposes the response into different components, finding that different components respond earlier or later and show different correlations with MPEs and their strength. The authors also investigated how EEG, reaction time, and pupil responses correlated with one another, providing further insight into the mechanism underlying the response to MPEs. Together, the study points toward multiple control and attention mechanisms involved in MPE response and memory.

Strengths:

The study has a clear behavioral paradigm with multiple measures - behavioral, EEG, and pupillometry that offer an investigation into different aspects of MPE response and memory.

The study is also very comprehensive in looking at multiple phases in processing MPEs: the prediction phase (prior to the violation), the response to MPEs, and subsequent memory of MPEs, all within one study. Specifically, the link between neural mechanisms and subsequent memory is a major advancement, as most prior studies did not include this component. Mechanisms underlying subsequent memory of MPEs are theoretically important, as a primary function of MPEs is to promote learning and memory. As the authors mention, the different neural and pupillary signals are not robustly correlated, suggesting multiple mechanisms underlying MPE detections, which is interesting, offers avenues for future research, and can facilitate a better theory of how MPEs are processed in the brain. Finally, the decomposition of pupil response into different components and their correlation with behavior (RT during match/MPE detection) is interesting.

Weaknesses:

The methods are rigorous, and the claims are mostly supported by the data, but there are a few weaknesses or places that could be improved:

(1) The authors conduct PCA analysis to identify different components of the pupillary response to MPE and relate them to behavior. Specifically, the authors identify components PC3 and PC4, which they interpret as related to MPE. However, some parts of the interpretation could be clearer or better justified:

(a) The authors refer to PC4 as "post-decision cognitive processing". But, given that RT was between .5-.7s, and PC3 peaked after more than 1s, wouldn't it be cautious to interpret PC3 as post-decision as well?

(b) MPEs overall elicit longer RTs in this study, suggesting that long RT is a behavioral marker of MPE. Nonetheless, the authors argue on p. 12: "Altogether, these findings indicate that when stronger mnemonic predictions (as indexed by shorter RTs) were violated." And, PC3 is correlated with shorter RTs for mismatches, meaning that behaviorally, these trials were more similar to matches. Thus, how do the authors interpret shorter versus longer RTs for MPEs, and what processes do these RT reflect?

(2) The brain to pupil relationship (p. 13-14): If I understand correctly, this was done on a trial-by-trial basis, but the high temporal resolution allows doing the analysis in a time-resolved manner - does brain activity at a certain time point preceding/following the pupil response correlate with the pupil response? It might be that cognitive control influences attention mechanisms or vice versa (because there is some overlap in the response). Although not testing causality, this temporally resolved correlation would be an interesting way to start probing how signals might influence each other.

(3) The relationships the authors find between brain measures and pupil components were largely not specific to mismatches/matches. However, are they specific to this task? I think it would benefit the paper to show that these relationships are potentially specific to making match/mismatch memory decisions, versus, e.g., any stimulus processing. For example, the authors could run the same analyses locked to stimuli in the study phase, anticipating a different pattern, if indeed these findings are specific to the associative memory task.

(4) During memory retrieval (i.e., before the probe), the authors find that frontal theta, a marker of cognitive control, was associated on a trial-by-trial basis with more posterior alpha (i.e., less alpha suppression, potentially reflecting less attention), and that this association was stronger for weaker predictions. The authors interpreted this as weaker predictions necessitating more cognitive control, and that more cognitive control was recruited specifically in trials where retrieval included less content (memory reinstatement) to attend to. Generally, cognitive control is recruited to facilitate memory retrieval. If so, one possible interpretation is that this correlation reflects cognitive control effort that has failed to produce enough memory reinstatement. The other possibility is that this correlation reflects more specific retrieval of the correct probe, without retrieval of interfering items (i.e., overall less content). I believe that the former explanation predicts that this correlation would be associated with longer RTs (more difficult decisions), while the latter predicts shorter RTs (easier decisions due to successful retrieval), at least for matches.

(5) In section 3, the authors found a positive relationship between alpha during memory retrieval and PC3 during MPE. If I understood correctly, this means that less attention during retrieval (less suppression) is correlated with a stronger PC3 response. How do the authors interpret this? Maybe along the same lines as in (5), specifically retrieving the correct information (i.e., less retrieved content to attend to) means a stronger prediction, leading to a stronger MPE, and a stronger MPE response, as reflected by PC3?

(6) The results with subsequent memory are important and address a major gap in the field that largely did not relate neural effects of MPE to subsequent memory. However, one major limitation of the study is that the authors did not test memory for matches. I understand the logic of avoiding testing matches. Because matches were repeated more times in the study, it's not a fair comparison, and could change participants' overall criterion for old/new decisions. However, one possibility would have been to test only the weak prediction; this could have given some specificity to the neural subsequent memory findings.

(7) The authors nicely characterized the different PC of pupillary MPE response. But, with respect to subsequent memory, they only present pupil size. Unless there is some methodological reason that prevents testing subsequent memory on the PC, I think this will be very informative about the potential mechanisms underlying memory of MPE.

(8) This paper includes many interesting findings, and I am not sure how they all come together into a cohesive mechanistic understanding of MPE response and subsequent memory. I think the paper would benefit from either a conceptual mechanism figure or, in the Discussion, have a summary of a proposed mechanism integrating the findings together.

(9) Relatedly, the section "Immediate, strength-sensitive neurocognitive impacts of MPEs" does not link the arguments to specific data points, so it's hard to follow which data specifically the authors are interpreting.

(10) If I understand correctly, the authors did not find improved memory for strong compared to weak MPE. First, I think this behavioral result should be incorporated in the main paper and in the interpretation of the results. Second, given that the neural effects the authors tested either correlated with memory for strong MPE or did not show a relationship with memory, what neural/pupil response could explain memory for weak MPE?

eLife Assessment

This computational study constitutes an extension to prior work on biophysical calcium-based synaptic plasticity rules with metaplasticity, investigating how single neurons can learn to perform non-linear pattern classification. This important work presents a significantly simpler solution to the studied problem with potentially broad applicability, there is however incomplete evidence to support the core conclusions.

Reviewer #1 (Public review):

Summary:

This computational modelling study addresses the important question of how neurons can learn non-linear functions using biologically realistic plasticity mechanisms. The study extends the previous related work on metaplasticity by Khodadadi et al. (2025), using the same detailed biophysical model and basic study design, while significantly simplifying the synaptic plasticity rule by removing non-linearities, reducing the number of free parameters, and limiting plasticity to only excitatory synapses. The rule itself is supervised by the presence or absence of a binary dopamine reward signal, and gated by separate calcium-sensitive thresholds for potentiation and depression. The author shows that, when paired with a strong form of dendritic non-linearity called a "plateau potential" and appropriate pre-existing dendritic clustering of features, this simpler learning mechanism can solve a non-linear classification task similar to the classic XOR logic operator, with equal or better performance than the previous publication. The primary claims of this publication are that metaplasticity is required for learning non-linear feature classification, and that simultaneous dynamics in two separate thresholds (for potentiation and depression) are critical in this process. By systematically studying the properties of a biophysically plausible supervised learning rule, this paper adds interesting insights into the mechanics of learning complex computations in single neurons.

Strengths:

The simplified form of the learning rule makes it easier to understand and study than previous metaplasticity rules, and makes the conclusions more generalizable, while preserving biological realism. Since similar biophysical mechanisms and dynamics exist in many different cell types across the whole brain, the proposed rule could easily be integrated into a wide range of computational models specializing in brain regions beyond the striatum (which is the focus of this study), making it of broad interest to computational neuroscientists. The general approach of systematically fixing or modifying each variable while observing the effects and interactions with other variables is sound and brings great clarity to understanding the dynamic properties and mechanics of the proposed learning rule.

Weaknesses:

General notes

(1) The credibility of the main claims is mainly limited by the very narrow range of model parameters that was explored, including several seemingly arbitrary choices that were not adequately justified or explored.

(2) The choice to use a morphologically detailed biophysical model, rather than a simpler multi-compartment model, adds a great deal of complexity that further increases uncertainty as to whether the conclusions can generalize beyond the specific choices of model and morphology studied in this paper.

(3) The requirement for pre-existing synaptic clustering, while not implausible, greatly limits the flexibility of this rule to solve non-linear problems more generally.

(4) In order to claim that two thresholds are truly necessary, the author would have to show that other well-known rules with a single threshold (e.g., BCM) cannot solve this problem. No such direct head-to-head comparisons are made, raising the question of whether the same task could be achieved without having two separate plasticity thresholds.

Specific notes

(1) Regarding the limited hyperparameter search:

(a) On page 5, the author introduces the upper LTP threshold Theta_LTP. It is not clear why this upper threshold is necessary when the weights are already bounded by w_max. Since w_max is just another hyperparameter, why not set it to a lower value if the goal is to avoid excessively strong synapses? The values of w_max and Theta_LTP appear to have been chosen arbitrarily, but this question could be resolved by doing a proper hyperparameter search over w_max in the absence of an upper Theta_LTP.

(b) The author does not explore the effect of having separate learning rates for theta_LTP and theta_LTD, which could also improve learning performance in the NFBP. A more comprehensive exploration of these parameters would make the inclusion of theta_max (and the specific value chosen) a lot less arbitrary.

(c) Figure 4 Supplements 3-4: The author shows results for a hyperparameter search of the learning rule parameters, which is important to see. However, the parameter search is very limited: only 3 parameter values were tried, and there is no explanation or rationale for choosing these specific parameters. In particular, the metaplasticity learning rates do not even span one order of magnitude. If the author wants to claim that the learning rule is insensitive to this parameter, it should be explored over a much broader range of values (e.g., something like the range [0.1-10]).

(2) Regarding the similarity to BCM, the author would ideally directly implement the BCM learning rule in their model, but at the least the author could have shown whether a slight variant of their rule presented here can be effective: for example having a single (plastic, not fixed) Ca-dependent threshold that applies to both LTP and LTD, with a single learning rate parameter.

(3) This paper is extremely similar (and essentially an extension) to the work of Khodadadi et al. (2025). Yet this paper is not mentioned at all in the introduction, and the relation between these papers is not made clear until the discussion, leaving me initially puzzled as to what problems this paper addresses that have not already been extensively solved. The introduction could be reworked to make this connection clearer while pointing out the main differences in approach (e.g., the important distinction between "boosting" nonlinearities and plateau potentials).

(4) The introduction is missing some citations of other recent work that has addressed single-neuron non-linear computation and learning, such as Gidon et al (2020); Jones & Kording (2021).

(5) Figure 1: The figure prominently features mGluR next to the CaV channel, but there is no mention of mGluR in the introduction. The introduction should be updated to include this.

(6) Could the author explain why there is a non-monotonic increase/decrease in the [Ca]_L in Figure 2B_4? Perhaps my confusion comes from not understanding what a single line represents. Does each line represent the [Ca] in a single spine (and if so, which spine), or is each line an average of all the spines in a given stim condition?

(7) Row 124 (page 4): L-type Ca microdomains (in which ions don't diffuse and therefore don't interact with Ca_NMDA) is a critical assumption of this model. The references for this appear only in the discussion, so when reading this paper, I found myself a bit confused about why the same ion is treated as two completely independent variables with separate dynamics. Highlighting the assumption (with citations) a bit more clearly in the results section when describing the rule would help with understanding.

(8) Row 149 (page 5): The current formulation of the update rule is not actually multiplicative. The fact that the update is weight-dependent alone does not make it a multiplicative rule, and judging by equation (1) it appears to simply be an additive rule with a weight regularization term that guarantees weight bounds. For example, a similar weight-dependent update is also a core component of BTSP (Milstein et al. 2021; Galloni et al. 2025), which is another well-known *additive* rule. An actual multiplicative rule implies that the update itself is applied via a multiplication, i.e. w_new = w_old * delta_w

For an example of a genuinely multiplicative rule, see: Cornford et al. 2024, "Brain-like learning with exponentiated gradients"). Multiplicative rules have very different properties to additive rules, since larger weights tend to grow quickly while small weights shrink towards 0.

(9) Equation 1 (page 5): Shouldn't the depression term be written as: (w_min - w)? This term would be negative if w is larger than w_min, leading to LTD. As it is written now, a large w and small w_min would just cause further potentiation instead of depression.

(10) In the introduction, the teaching signal is described in binary terms (DA peak, or DA pause), but in Equation 1, it actually appears to take on 3 different values. Could the author clarify what the difference is between a "DA pause" and the "no DA" condition? The way I read it, pause = absence of DA = no DA

(11) Figure 3: In these experimental simulations, DA feedback comes in 400ms after the stimulus. The author could motivate this choice a bit better and explain the significance of this delay. Clearly, the equations have a delta_t term, but as far as the learning algorithm is concerned, it seems like learning would be more effective at delta_t=0. Is the choice of 400ms mainly motivated by experimental observations? On a related note, is it meaningful that the 200ms delta_t before the next stimulus is shorter than the 400ms pause from the first stimulus? Wouldn't the DA that arrives shortly before a stimulus also have an effect on the learning rule?

(12) Figure 4C: How is it possible that the theta_LTP value goes higher than the upper threshold (dashed line)? Equation 3 implies that it should always be lower.

(13) Row 429 (page 11): The statement that "without metaplasticity the NFBP cannot be solved" is overly general and not supported by the evidence presented. There exist many papers in which people solve similar non-linear feature learning problems with Hebbian or other bio-plausible rules that don't have metaplasticity. A more accurate statement that can be made here is that the specific rule presented in this paper requires metaplasticity.

(14) The methods section does not make any mention of publicly available code or a GitHub repository. The author should add a link to the code and put some effort into improving the documentation so that others can more easily assess the code and reproduce the simulations.

Reviewer #2 (Public review):

Summary:

The manuscript proposes interesting synaptic plasticity rules grounded in experimental data. Its main features are:

(1) plasticity depends on local calcium concentration driven by presynaptic activity and is independent of somatic action potentials,

(2) the rules incorporate metaplasticity, and (3) they demonstrate how a single neuron could address the feature-binding problem at the dendritic level.

The work extends a previous study (https://doi.org/10.7554/eLife.97274.2), to which the author also contributed.

The author models two calcium thresholds (LTP/LTD) from two different calcium sources (NMDA/VGCC), and these thresholds are flexible (metaplasticity rule, similar to BCM), which is claimed to be necessary for successful learning of both FBP and NFBP (linear and nonlinear feature binding problem with 1 or 2 patterns). The role of each threshold seems to be opposite and complementary. One extra condition has been added: an upper threshold for LTP. This threshold serves to stop synaptic strengthening once synapses are strong enough to evoke a plateau. With that, synapses are not strengthened to the maximal value, avoiding strong supralinear integration for irrelevant patterns.

Strengths:

The current model implements not only local synaptic plasticity but also metaplasticity and solves the FBP at the dendrite level. Another strong aspect of the model is that metaplasticity in the LTD threshold protects strengthened synapses from weakening. In this way, as the author mentioned, metaplasticity is able to protect learned patterns from being forgotten or weakened and prevent irrelevant patterns from being stored. This is a nice modelling example of metaplasticity being helpful in preventing the catastrophic interference or forgetting (as has been explicitly discussed in a recent article https://doi.org/10.1016/j.tins.2022.06.002 ). The author might want to briefly mention or emphasize this aspect of the model, which might be interesting also for the AI community.

Weaknesses:

(1) What is novel in the current paper as compared to Khodadadi et al. eLife 2025? That is not completely clear and should be made clearer. Is it only a minor difference related to the fact that the new learning rule has metaplasticity in both calcium thresholds and is simpler? This seems to be just an incremental increase in knowledge/methods. Can the author defend his paper against this point from the „devil's advocate"? How is the conclusion of the author in the abstract that „metaplasticity in both thresholds is necessary" reconcilable with his previous publication (Khodadadi et al. eLife 2025), in which only metaplasticity in one threshold was successful in solving the nonlinear feature binding problem?

(2) As far as I can judge without testing the model, metaplasticity causes thresholds to monotonically increase during systematic pattern presentation, which stabilizes weights and allows pattern separation. Due to the closed-loop nature of the current implementation, where metaplasticity only happens if plasticity happens, this also effectively locks patterns in place. However, flexible learning is an essential mechanism for survival. Imagine a mutation event takes place and bananas suddenly become red and/or strawberries turn yellow. It seems that the current model would be unable to adapt to these new patterns even if rewards were to be shifted. While out of the scope of the study, due to its importance, I feel that pattern shifting/relearning should at least be briefly discussed. How could the model be improved to allow relearning?

eLife Assessment

Based on the perceived low efficacy of current therapies targeted to FGFR2 in gastric cancer (GC), the authors investigate an approach which combines SHP2 inhibition with existing FGFR2 inhibitors. The data were largely collected and analysed using solid and validated methodology. There is some useful data regarding combination therapy in a new clinical cohort, which supports previous studies that have reported the potential of targeting RTKs together with phosphatases.

Reviewer #1 (Public review):

The manuscript entitled "Blocking SHP2 1 benefits FGFR2 inhibitor and overcomes its resistance in 2 FGFR2-amplified gastric cancer" by Zhang, et al., reports that FGFR2 was amplification in 6.2% (10/161) of gastric cancer samples and that dual blocking SHP2 and FGFR2 enhanced the effects of FGFR2 inhibitor (FGFR2i) in FGFR2-amplified GC both in vitro and in vivo via suppressing RAS/ERK and PI3K/AKT pathways. Furthermore, the authors also showed that SHP2 blockade suppressed PD-1 expression and promoted IFN-γ secretion of CD8+ 46 T cells, enhancing the cytotoxic functions of T cells. Thus, the authors concluded that dual blocking SHP2 and FGFR2 is a compelling strategy for treatment of FGFR2-amplified gastric cancer. Although the finding is interesting, the finding that FGFR2 is amplified in gastric cancer and that FGFR inhibitors have some effect on treating gastric cancer is not novel. The data quality is not high, the effects of double inhibitions are not significant. It appears that the conclusions are largely overstated, the supporting data is weak and not compelling.

The data in Figure 1 is not novel; similar data have been reported elsewhere.

It is unclear why the two panels in fig 2a and 2b can not be integrated into one panel, which will make it easier to compare the activities.

The synergetic effects of azd4547 and shp099 are not significant in Fig 2e and 2f, as well as in Fig. 3g and Fig. 4f

Data in Fig. 5 is weak and can be removed. It is unclear why FGFR inhibitor has some activities toward t cells since t cells do not express FGFR.

Reviewer #2 (Public review):

Summary:

This manuscript reports the application of a combined targeted therapeutic approach to gastric cancer treatment. The RTK, FGFR2 and the phosphatase, SHP2 are targeted with existing drugs; AZD457 and SHP099, respectively. Having shown increased mRNA levels of FGFR2 and SHP2 in a patient population and highlighted the issue of resistance to single therapies the combination of inhibitors is shown to reduce cancer-related signalling in two gastric cell lines. The efficacy of the dual therapy is further demonstrated in a single patient case study and mouse xenograft models. Finally, the rationale for SHP2 inhibition is shown to be linked to immune response.

Strengths:

The data is generally well presented, and the study invokes a novel patient data set which could have wider value. The study provides additional evidence to support the combined therapeutic approach of RTK and phosphatase inhibition.

Weaknesses:

Combined therapy approaches targeting RTKs and SHP2 have been widely reported. Indeed, SHP099 in combination with FGFR inhibitors has been shown to overcome adaptive resistance in FGFR-driven cancers. Furthermore, the inhibition of SHP2 has been documented to have important implications in both targeting proliferative signalling as well as immune response. Thus, it is difficult to see novelty or a significant scientific advance in this manuscript. Although the data is generally well presented, there is inconsistency in the interpretation of the experimental outcomes from ex vivo, patient and mouse systems investigated. In addition, the study provides only minor or circumstantial understanding of the dual mechanism.

Using data from a 161 patient cohort FGFR2 was identified as displaying amplification of FGFR2 in ~6% with concomitant elevation of mRNA of patients which correlated with PTPN11 (SHP2) mRNA expression. The broader context of this data is of value and could add a different patient demographic to other data on gastric cancer. However, there is no detail on patient stratification or prior therapeutic intervention.

Comments on revisions: This has been attended to in the revised version

In SNU16 and KATOIII cells the combined therapy is shown to be effective and appears to be correlated with increase apoptotic effects (i.e. not immune response).

Fig 2E suggests that the combined therapy in SNU16 cells is little better than FGFR2-directed AZD457 inhibitor alone, particularly at the higher dose.

The individual patient case study described via Fig 3 suggests efficacy of the combined therapy (at very high dosage), however the cell biopsies only show reduced phosphorylation of ERK, but not AKT. This is at odds with the ex vivo cell-based assays. Thus, it is not clear how relevant this study is.

The mouse xenograft study shows a convincing reduction in tumor mass/volume and a clear reduction in pAKT, whilst pERK remains largely unaffected by the combined therapeutic approach. This is in conflict with the previous data which seems to show the opposite effect.

Comments on revisions: The authors have clarified this point

In all, the impact of the dual therapy is unclear with respect to the two pathways mediated by ERK and AKT.

Finally, the authors demonstrate the impact of SHP2 on PD-1 expression and propose that the SHP099/AZD4547 combination therapy significantly induces the production of IFN-γ in CD8+ T cells. This part of the study is unconvincing and would benefit from an investigation of the tumor micro-environment to assess T cell infiltration.

Reviewer #3 (Public review):

Summary:

Fibroblast growth factor receptor 2 (FGFR2) is a receptor tyrosine kinase that can be amplified in gastric cancer and serves as a potential therapeutic target for this patient population. However, targeting FGFR2 has shown limited efficacy. Thus, this study seeks to identify additional molecules that can be effectively targeted in FGFR2 amplified gastric cancer, with a focus on Src homology region 2-containing protein tyrosine phosphatase 2 (SHP2). The authors first demonstrate that 6% of gastric cancer patients in a cohort of human patient samples exhibit FGFR2 amplification. Furthermore, they demonstrate that FGFR2 mRNA expression is positively correlated with PTPN11 gene expression (which is the gene that encodes the SHP2 protein). Using human gastric cancer cell lines with amplified FGFR2, the authors then test the effects of combining the FGFR inhibitor AZD4547 with the SHP2 inhibitor SHP099 on tumor cell death and signaling molecules. They demonstrate that combining the two inhibitors is more effective at tumor cell killing and reducing activation of downstream signaling pathways than either inhibitor alone. In further studies, the authors obtained gastric cancer cells with FGFR2 amplification from a patient that was treated with FGFR2 inhibitor. While this patient initially showed a partial response, the patient ultimately progressed, demonstrating resistance to FGFR2 inhibition. Following isolation of tumor cells from the patient's ascites, the authors demonstrate that these cells are sensitive to the combination treatment of AZD4547 and SHP099. Further studies were performed using a xenograft model using athymic nude mice in which the combination of SHP099 and AZD4547 were found to reduce tumor growth more significantly than either treatment alone. Finally, the authors demonstrate using an in vitro culture model that this combination treatment enhances T cell mediated cytotoxicity. The authors conclude that targeting FGFR2 and SHP2 represents a potential combination strategy in gastric patients with FGFR2 amplification.

Strengths:

The authors demonstrate that FGFR2 amplification positively correlates with PTPN11 in human gastric cancer samples, providing a rationale for combination therapies. Furthermore, convincing data are provided demonstrating that targeting both FGFR and SHP2 is more effective than targeting either pathway alone using in vitro and in vivo models. The use of cells derived from a gastric cancer patient that progressed following treatment with an FGFR inhibitor is also a strength. The findings from this study support the conclusion that SHP2 inhibitors enhance the efficacy of FGFR-targeted therapies in cancer patients. This study also suggests that targeting SHP2 may also be an effective strategy for targeting cancers that are resistant to FGFR-targeted therapies.

Weaknesses:

The main caveat with these studies is the lack of an immune competent model with which to test the finding that this combination therapy enhances T cell cytotoxicity in vivo.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

The data in Figure 1 is not novel, similar data has been reported elsewhere.

We are grateful for the critical evaluation of our finding. Although there have been a few researches indicating the prevalence of FGFR2-amplified GC patients, our research provided a novel dataset of 161 GC patients using next-generation sequencing (NGS) in China, further emphasizing the high frequency of FGFR2 amplification in gastric cancer patients. Moreover, the proportion of FGFR2-amplified GC patients in our center (6.2%) is relatively higher than that of TCGA cohort (5%).

We have transferred the original Figure 1C and 1D to the supplementary figures, and constructed a novel pie chart for Nanjing Drum Tower Hospital cohort to compare with the TCGA cohort.

It is unclear why the two panels in Fig 2a and 2b can not be integrated into one panel, which will make it easier to compare the activities.

Thanks for pointing this out. In the first figure of Figure 2a and 2b, we performed gradient concentration CCK8 detection on the cytotoxicity of SHP099 against tumor cells. In the second figure, we selected 10 μm (IC50) as the fixed concentration of SHP099 for combined efficacy testing with gradient concentration of AZD4547. Moreover, the units of the horizontal axis in both figure 2a and 2b cannot be unified. Therefore, we believe that the two figures in figures 2a and 2b are not suitable for merging into one figure.

For the convenience of observation, we integrated the first panel of figure 2a and 2b into one panel, and integrated the second panel in the same way.

The synergetic effects of azd4547 and shp099 are not significant in Fig 2e and 2f, as well as in Fig. 3g and fig. 4f

In Fig 2e and 2f, we not only analyzed the synergetic effects of 3 nM (a relatively lower dose) AZD4547 and 10 μm SHP099, but also 10 nM (a relatively higher dose) AZD4547 and 10 μm SHP099. The synergetic effects of different dosage combinations should be compared correctly. From our perspective, the combination treatment led to a stronger inhibition of phospho-FGFR, phospho-SHP2 and FGFR2-initiated downstream signaling molecules, especially in KATOIII.

For ease of comparison, we circled 10 μm SHP099, 10nM AZD4547 and 10nM AZD4547+10 μm SHP099 in red.

Author response image 1.

Author response image 2.

We also circled 10μM SHP099, 3nM AZD4547 and 3nM AZD4547+10 μm SHP099 in blue.

Author response image 3.

Author response image 4.

For ease of comparison, we also conducted grayscale value analysis and normalization using image J.

Author response image 5.

Author response image 6.

Author response image 7.

Author response image 8.

In Fig. 3g, the combination therapy exhibited relatively stronger inhibitory effects on phospho-ERK, phospho-AKT and phospho-mTOR.

For ease of comparison, we conducted grayscale value analysis and normalization using image J.

The unclear effect of combination therapy may be due to the presence of impurities other than tumor cells in patient’s ascites.

Author response image 9.

In Fig. 4f, it was obvious that phospho-AKT and phospho-mTOR were further suppressed in combination group.

For ease of comparison, we conducted grayscale value analysis and normalization using image J.

Author response image 10.

Therefore, in our opinions, our data could relatively sufficiently confirm the synergetic effects of AZD4547 and SHP099.

Data in Fig. 5 is weak and can be removed. It is unclear why FGFR inhibitor has some activities toward t cells since t cells do not express FGFR.

The activation effect of SHP099 on T cells has been validated in many articles. In a previous study published in Cancer Immunology Research, it was pointed out that the combination of FGFR2 inhibitor erdafitinib and PD-1 antibody can activate T cells and downregulate T cell surface exhaustion related factors (including PD-1) in vivo Therefore, the anti-tumor immune effect of FGFR2 inhibitor cannot be ignored. Although T cells do not express FGFR, FGFR2 inhibitors may still affect PD-1 expression on the surface of T cells in some other ways, which requires further research. We have deleted fig.5D in our article. We believe that the combination of FGFR2 inhibitor and SHP2 inhibitor not only has a direct killing effect on tumor cells, but also promotes anti-tumor immunity by activating T cells. Therefore, we believe that the in vitro data in Figure 5 is also meaningful.

Reviewer #2 (Public review):

Strengths:

The data is generally well presented and the study invokes a novel patient data set which could have wider value. The study provides additional evidence to support the combined therapeutic approach of RTK and phosphatase inhibition.

We sincerely thank the reviewer for the critical evaluation and appreciation of our findings.

Weaknesses:

Combined therapy approaches targeting RTKs and SHP2 have been widely reported. Indeed, SHP099 in combination with FGFR inhibitors has been shown to overcome adaptive resistance in FGFR-driven cancers. Furthermore, the inhibition of SHP2 has been documented to have important implications in both targeting proliferative signalling as well as immune response. Thus, it is difficult to see novelty or a significant scientific advance in this manuscript. Although the data is generally well presented, there is inconsistency in the interpretation of the experimental outcomes from ex vivo, patient and mouse systems investigated. In addition, the study provides only minor or circumstantial understanding of the dual mechanism.

We acknowledge that our research on the mechanism of dual inhibition is not deep enough. There remain more in-depth mechanisms of the combination of SHP2 inhibitor and RTK inhibitors needed to be explored, and it would be the main direction of our future study.

Using data from a 161 patient cohort FGFR2 was identified as displaying amplification of FGFR2 in ~6% with concomitant elevation of mRNA of patients which correlated with PTPN11 (SHP2) mRNA expression. The broader context of this data is of value and could add a different patient demographic to other data on gastric cancer. However, there is no detail on patient stratification or prior therapeutic intervention.

Thanks for pointing this out and we have added a table on patients’ stratification such as age, gender and so on. Unfortunately, data on patients’ prior therapeutic intervention weren’t collected.

In SNU16 and KATOIII cells the combined therapy is shown to be effective and appears to be correlated with increased apoptotic effects (i.e. not immune response).

Fig 2E suggests that the combined therapy in SNU16 cells is a little better than FGFR2-directed AZD457 inhibitor alone, particularly at the higher dose.

The individual patient case study described via Fig 3 suggests efficacy of the combined therapy (at very high dosage), however, the cell biopsies only show reduced phosphorylation of ERK, but not AKT. This is at odds with the ex vivo cell-based assays. Thus, it is not clear how relevant this study is.

The mouse xenograft study shows a convincing reduction in tumor mass/volume and clear reduction in pAKT, whilst pERK remains largely unaffected by the combined therapeutic approach. This is in conflict with the previous data which seems to show the opposite effect. In all, the impact of the dual therapy is unclear with respect to the two pathways mediated by ERK and AKT.