eLife Assessment

This study demonstrates that endothelial toll-like receptor 4 is a central regulator of leptomeningeal inflammation in the context of neonatal E. coli meningitis. The data are derived from cell type-specific gene knockout in mice as well as from cultured endothelial cells, and are generally solid, with only minor weaknesses in analysis and interpretation. This work is important as it advances our understanding of host cellular processes and molecular pathways underlying meningitis pathogenesis.

Reviewer #1 (Public review):

Summary:

In this manuscript, Seegren and colleagues demonstrate that in a mouse model of neonatal E. coli meningitis, loss of endothelial toll-like receptor 4 (TLR4) leads to a marked decrease in transcriptional dysregulation across multiple leptomeningeal cell types, a decrease in vascular permeability, and a decrease in macrophage abundance. In contrast, loss of macrophage TLR4 had less pronounced effects. Using cultured wild-type and TLR4-knockout endothelial cells, the authors further demonstrate that TLR4-NF-κB signaling leads to reversible internalization of the tight junction protein claudin-5, establishing a potential mechanism of increased vascular permeability. Finally, the authors use RNA-sequencing of wild-type and TLR4-knockout endothelial cells to define the TLR4-dependent cell-autonomous transcriptional response to E. coli.

Strengths:

(1) The authors address an important, well-motivated hypothesis related to the cellular and molecular mechanisms of leptomeningeal inflammation.

(2) The authors use model systems (mouse conditional knockouts and cultured endothelial cells) that are appropriate to address their hypotheses. The data are of high quality.

Weaknesses:

(1) The authors perform single-nucleus RNA-seq on dissected leptomeninges from control and E. coli-infected mice across three genotypes (WT, Tlr4MKO, and Tlr4ECKO). A major discovery from this experiment, as summarized by the authors, is: "Tlr4ECKO mice exhibited a global attenuation of infection-induced transcriptional responses across all major leptomeningeal cell types, as judged by the positions of cell clusters in the UMAP." This conclusion could be considerably strengthened by improving the qualitative and quantitative analysis.

(2) The authors interpret E. coli infection-induced increases in leptomeningeal sulfo-NHS-biotin as evidence of compromised BBB integrity (i.e., extravasation from the vasculature) (Results, page 7), but another possible route in this context is sulfo-NHS-biotin entry from the dura across a compromised arachnoid barrier. The complete rescue in Tlr4ECKOs is strongly suggestive that the vascular route dominates, but it would strengthen the work if the authors could assess arachnoid barrier fidelity (e.g. via immunohistochemistry). At a minimum, authors should mention that the sulfo-NHS-biotin signal in this context may represent both vascular and arachnoid barrier extravasation.

(3) The authors state that "deletion of TLR4 prevented both NF-κB nuclear translocation and Cldn5 internalization in response to E. coli (Figure 4A-D)" (Results, page 9). In Figures 4C and D, however, there is no indicator of a statistical test directly comparing the two genotypes. A comparison of within-genotype P-values should not be used to support a genotype difference (PMID: 34726155).

(4) In the first paragraph of the Results, the authors summarize the meningeal layers as (1) pia, (2) subarachnoid space, (3) arachnoid, and (4) dura, and then state "The second and third layers constitute the leptomeninges." This definition of leptomeninges seems to omit the pia, which is widely considered part of the leptomeninges (PMID: 37776854).

(5) The Cdh5-CreER/+;Tlr4 fl/- mouse lacks TLR4 in all endothelial cells (i.e., in peripheral organs as well as CNS/leptomeninges), and, as the authors note, the periphery is exposed to E. coli. It would be helpful if the authors could comment in the Discussion on the possibility that peripheral effects (e.g., peripheral endothelial cytokine production, changes to blood composition as a result of changes to peripheral endothelial permeability) may contribute to the observed leptomeningeal phenotypes.

Reviewer #2 (Public review):

Summary:

The authors use a postnatal mouse model of E. coli bacterial meningitis and a mouse brain endothelioma cell line combined with cell-type-specific gene deletion to study the function of endothelial TLR4, a cell surface receptor that recognizes gram positive bacterial wall components, in the local leptomeningeal (LPM) response with a focus on endothelial barrier breakdown mediated by TLR4. Single-cell transcriptional profiling and imaging studies using whole-mount preps of the LPM support that LPM endothelial, CD206+ local macrophage and LPM fibroblast and arachnoid barrier cell inflammatory response and is abrogated in endothelial-specific KO of TLR4, pointing to a role for endothelial TLR4 in local LPM response. Culture studies using Bend3.1 cells (a mouse brain endothelioma cell line) support a direct role for TLR4 in the bacteria-mediated inflammatory response and in internalization of Cldn5 via the endosomal-lysosomal pathway, resulting in loss of barrier integrity

Strengths:

The local LPM cell response in meningitis and the role of specific LPM cells in inflammation and CNS barrier breakdown have not been extensively studied, despite ample evidence for primary immune response in the meninges in human patients and in animal models. The authors employ a robust, multi-model approach using both in vivo and in vitro models with cell-type-specific knockout to study the function of TLR4 in brain endothelial cell response. The authors nicely combine functional barrier assays with IF for junctional localization in their experimental design, and they delve into potential mechanisms of Cldn5 internalization using markers of endosomal-lysosomal pathway localization. The authors also describe a new type of barrier assay using a streptavidin-coated plate upon which barrier-forming cell cultures can be placted, this could be a very useful alternative or complement to other size-selective barrier assays and presumably could work for other barrier forming cells types, likely epithelial cells.

Weaknesses:

(1) There are no measures of bacterial burden in peripheral organs, blood, in the LPM or brain in the TLR4 endothelial cKO mice. Lack of TLR4 in endothelial cells could prevent bacterial 'access' into the LPM and brain, essentially preventing meningitis and leading to a lack of inflammatory responses in the LPM-located cells simply because there is no bacteria present. Bacteremia may also be reduced, as might inflammatory responses in peripheral organs with TLR4-deficient peripheral endothelium. Bacterial counts and inflammatory measures in peripheral organs and blood are important to better understand the mechanism(s) underlying the reduced inflammatory profile in LPM cells and no LPM endothelial breakdown in the Tlr4 endothelial cKO mice. In other words, does deleting TLR4 in EC protect against the development of meningitis by somehow blocking bacteria access to the LPM (this would be supported by low or no CFU counts in infected Tlr4 endothelial cKO) or is it what the authors appear to propose in Figure 1J that TLF4 in EC is the only cell responding to the bacteria to trigger the immune cascade in the LPM? More data is needed to resolve this, as this is a major claim of the paper.

(2) The authors look at the underlying cortical response (cerebral vasculature for ICAM and immune cells) but do not use markers that could identify microglia (Iba1), the primary resident immune cell (CD206 is not useful, at this stage, in perivascular macrophages that are extremely sparse in the postnatal brain). This would be important to better study the impact on CNS resident immune cell morphological activation.

(3) The authors suggest that Cldn5 junctional localization is selectively disrupted upon bacterial exposure, mediated by TLR4 - they suggest this based on studying PECAM, GLUT-1, ZO-1 and B-catenin (all normally junction or cell surface located in cultured Bend3.1) in relationship to Cldn5 localization (normally high) - it is possibly these are also impact by bacteria exposure (maybe through different mechanisms?) - a better measure would be to use the similar cyto/PM measure they do for Cldn5 in Fig. 4D and to evaluate this or to use intensity measurements.

(4) The discussion could benefit from delving more into the prior literature on E.coli-mediated breakdown of junctions in cultured human microvascular brain endothelial cell model and critical host-pathogen interactions of the bacteria with ECs (PMID: 14593586), and how this might involve TLR4.

(5) It would be important to discuss how their results relate to earlier studies on TLR4-/- and TLR2-/- global knockout mice and protection vs vulnerability to development of meningitis (see PMCID: PMC3524395) - this paper showed that TLR4 global KO mice have increased susceptibility to die from meningitis and have much higher CFU counts in the CNS. In this manuscript and their prior work (Wang et al., 2023), this group shown that both global TLR4-/- mutants and their EC-specific KO have reduced barrier permeability, but we don't have any information about CFU or susceptibility to death from meningitis in their models.

Reviewer #3 (Public review):

Summary:

This study investigates the molecular underpinnings of immune responses in the leptomeninges in neonatal bacterial meningitis. Bacterial meningitis is a major disease burden, particularly for neonates, and it has previously been noted that the meningeal immune environment in infants is permissive to opportunistic infection (Kim et al., Sci Immunol, 2023). There is less known about the contribution of the stromal compartment to meningeal immune responses. Seegren et al. interrogate the role of leptomeningeal endothelium in host defence in E. coli infected neonatal mice using mouse genetic tools to delete the LPS receptor Tlr4 from either endothelial cells (using Cdh5-CreER) or macrophages (using LysM-Cre). The authors use snRNAseq, cleared cortical mounts, and in vitro work to define the impact of E. coli infection on leptomeningeal endothelial cells. This study uses a range of innovative techniques to probe the role of the stromal compartment in meningitis.

Strengths:

This study makes excellent use of cleared cortical mounts to examine the biology of the leptomeninges, in particular, changes to the endothelium, with unprecedented detail. In combination with high-quality sequencing data provide new insights into the impact of meningitis on the leptomeninges. The data presented by the authors is of very high quality.

Weaknesses:

The weaknesses of the study were in terms of interpretation and perhaps study design.

(1) Most importantly, the authors need to provide additional validation of their conditional knockout models. The authors need to confirm that the Cdh5-CreER does not impact leptomeningeal fibroblasts and to confirm gene deletion in macrophages.

(2) The authors could also strengthen the paper by providing data on the impact of these conditional knockout models on the course of meningitis and bacterial burden.

(3) Finally, it is perhaps not surprising that Tlr4 is required for meningitis responses with E. coli. However, it is unclear if these findings can be generalised to other, more common, meningitis infections (streptococcal/pneumococcal).

(4) There are additional minor issues; for instance, the arachnoid fibroblast 2 population appears to closely resemble dural border cells.

(5) The cell line model (bEnd.3) is a relatively low-fidelity model of BBB endothelial cells, and this should be acknowledged.

With these caveats, it is difficult to be certain that the endothelium alone is the driver of meningeal immune responses in meningitis, and what the impact of these is.

eLife Assessment

This fundamental study advances our understanding of how dietary patterns shape cancer immunity by identifying a link between a Mediterranean-mimicking diet, gut bacteria, and a metabolite that enhances anti-tumor immune responses. The evidence supporting the main conclusions is solid, based on carefully controlled diet experiments, measurements of gut-derived molecules, and functional immune analyses across multiple models, together with supportive observations in human data. The work will be of broad interest to biologists working on microbiota and cancer. However, there are several issues that the authors should address to improve the manuscript.

Reviewer #1 (Public review):

Summary:

In brief, this manuscript addresses a very interesting topic, namely, the impact of the Mediterranean diet on the development of cancer. Using one mouse model and three tumor cell lines, the data show that a Mediterranean diet is sufficient to promote an anti-tumor response mediated by the microbiota, metabolites, and the immune system. Mechanistically, the Mediterranean diet promotes the expansion of Bacteroides thetaiotaomicron (B. theta for short), which converts tryptophan into 3-IAA. Both B. theta and the metabolite are sufficient to phenocopy the effect of the Mediterranean diet on cancer growth in vivo. The manuscript also shows that this effect is mediated by CD8 T cells and suggests, by way of in vitro assays, that 3-IAA sustains the functionality of CD8 T cells, preserving their exhaustion and blocking the ISR pathway.

Strengths:

The conclusions of this manuscript are potentially interesting and of potential clinical relevance.

Weaknesses:

For a full technical evaluation of the strength of the data, I am missing important technical and experimental details (e.g., number of independent experiments, statistics), and found some legends with potential labelling inaccuracies.

Reviewer #2 (Public review):

Summary:

The authors aimed to investigate the mechanistic link between a Mediterranean-mimicking diet (MedDiet)-specifically the synergy between high fiber and fish oil-and its ability to suppress tumor growth. They successfully identify that this dietary combination alters the gut microbiome to favor the expansion of Bacteroides thetaiotaomicron. This bacterium metabolizes dietary tryptophan into indole-3-acetic acid (3-IAA), which then acts systemically to prevent CD8+ T-cell exhaustion.

Strengths:

The study integrates controlled dietary interventions, microbiome perturbation, metabolite profiling, and immune functional analyses into a coherent and well-organized framework, making the overall logic of the work easy to follow. The dietary design is carefully controlled, allowing clear interpretation of which broad dietary features are associated with the observed antitumor effects. The immune dependence of the phenotype is addressed using appropriate experimental approaches, and the results broadly support a role for gut microbiota-derived metabolites in shaping immune cell function. In addition, analyses of human datasets provide important context and enhance the potential relevance and usefulness of the findings for a broader research community.

Weaknesses:

While the manuscript provides strong support for a role of the microbial metabolite indole-3-acetic acid and downstream stress signaling in shaping immune cell function, the upstream mechanism by which this metabolite exerts its effects remains unresolved. In particular, the specific molecular sensor or binding target through which the metabolite acts has not been identified, and this uncertainty limits mechanistic precision. Framing this point more explicitly as an open question would help align the interpretation with the current data.

In addition, at several points, the presentation may imply that a single microbial species is uniquely responsible for the observed effects. However, the experimental evidence more directly demonstrates sufficiency under the tested conditions rather than necessity. A clearer distinction between "sufficient" and "necessary" claims would help readers better assess the generality of the findings and their applicability to more complex microbial communities.

The interpretation of the human data also warrants some caution. The diet-associated score applied to human datasets is derived from gene-expression signatures identified in mouse models and therefore represents an indirect proxy rather than a direct measure of dietary intake. Although the score correlates with clinical outcomes, it does not establish that patient survival is driven by consumption of specific dietary components such as fiber and fish oil.

eLife Assessment

This study presents a valuable finding that depletion of the rRNA methyltransferase METTL5 enhances anti-tumor immunity through a novel mechanism involving neoantigen generation from non-canonical translation. The evidence supporting the central conclusions is solid, with comprehensive multi-omics data including ribosome profiling, immunopeptidomics, TCR sequencing, and multiple in vivo tumor models demonstrating synergy with immune checkpoint blockade.

Reviewer #1 (Public review):

Summary:

In this manuscript, Zhang et al. demonstrate that depletion of the 18S rRNA m6A methyltransferase Mettl5 compromises translation fidelity and consequently increases neoantigen generation, thereby uncovering an unexpected role for Mettl5 in tumor immunity. Mettl5-KO tumors exhibit enhanced CD8⁺ T-cell infiltration and show improved responses to immune checkpoint blockade. Mechanistically, loss of Mettl5 perturbs the local structure of 18S rRNA and disrupts the ribosome's ability to perform accurate translation. Subsequent ribosome profiling and mass spectrometry analyses provide compelling evidence that Mettl5 functions as a previously unrecognized regulator of translation to participate in tumor immune evasion.

Strengths:

This study presents a comprehensive set of experimental data supporting a mechanistic link between rRNA modification, translation fidelity, and neoantigen generation. The observed synergistic effect of Mettl5 depletion and anti-PD-1 therapy highlights the potential translational relevance of targeting rRNA modifications in cancer immunotherapy.

Weaknesses:

(1) In light of the principal function of Mettl5, which is to methylate 18S rRNA within the small ribosomal subunit, the authors focus primarily on translation fidelity, largely associated with elongation, but provide limited exploration of potential effects on translation initiation. Loss of Mettl5 may alter the initiation landscape, potentially promoting alternative or noncanonical initiation events (e.g., initiation at CUG codons), which could also contribute to the observed neoantigen repertoire changes. Further investigation into initiation-level alterations would strengthen the mechanistic interpretation.

(2) Given the broad involvement of rRNA methyltransferases in ribosome function, the authors should incorporate a parallel analysis using another enzyme (e.g., Zcchc4 or Nsun5) as a negative control. Such an experiment is essential to demonstrate that the tumor immunity phenotype observed is specific to Mettl5 rather than a general consequence of perturbing rRNA modification.

Reviewer #2 (Public review):

Summary:

This study demonstrates that METTL5-mediated rRNA m⁶A1832 modification regulates tumor neoantigen generation by maintaining translational fidelity. Loss of METTL5 in tumor cells promotes immune cell infiltration into the tumor microenvironment and enhances the therapeutic efficacy of anti-PD-1 treatment, identifying a novel and potentially important target for cancer immunotherapy.

Strengths:

In murine tumor models, the authors found that Mettl5 depletion increases CD8⁺T cell infiltration and T cell receptor (TCR) repertoire diversity, and revealed a novel mechanism by which reduced ribosomal translation fidelity enhances non-canonical translation, thereby promoting the production of tumor neoantigens.

Weaknesses:

(1) While Mettl5 knockout enhances T-cell infiltration into tumors, it remains unclear whether loss of Mettl5 affects the expression of chemokines involved in immune cell recruitment.

(2) Although the authors report a significant reduction in tumor cell growth as well as tumor volume and weight, direct evidence demonstrating T-cell-mediated cytotoxicity is lacking.

eLife Assessment

The importance of uterine natural killer (NK) cells in reproductive success has been demonstrated in mice and humans; however, it is still unclear how uterine NK cells are developed. In this valuable manuscript, the authors provide convincing evidence that TGF-b signaling in NK cells supports normal pregnancy in mice by the conversion of conventional NK cells into uterine tissue-resident NK cells. There are some concerns about the paper, particularly around Figures 1A, 1C, and 2E.

Reviewer #1 (Public review):

This is an excellent paper from Dr. Yokoyama and colleagues. The experiments are technically demanding, given the very low cell numbers and the challenges of working with implantation sites at gestational days 6.5, 10.5, and 14.5. Overall, the impact of TGF-β receptor II deficiency in the NK lineage on uterine trNK cell numbers and litter size is convincing, and the authors' conclusions are well supported by the data. Less convincing, however, is the claim that the decrease in trNK cells is compensated by an increase in cNK cells; rather, the absence of TGF-β receptor II appears to result in an overall reduction of NK/ILC1 cells.

Major Points:

(1) Figure 1A and B

Although a trend is evident, it does not appear that the absolute number of cNK cells at day 14 is significantly changed from day 6.5?

(2) Figure 2E

The authors state, "This reduction of uterine trNK cells was accompanied by a concomitant increase in the absolute number and frequency of CD49b+Eomes+ cNK cells within the pregnant uterus of TGF-βRIINcr1Δ dams (Figure 2 D, E). The number of cNK cells appears relatively low (visually ~1,000-1,300), and although the difference is statistically significant, its physiological relevance is unclear. More importantly, this modest increase does not correlate with the marked decrease in trNK and ILC1 populations, as cNK cells do not appear to accumulate. In my opinion, the conclusion "Collectively, these findings indicate that a TGF-β-driven differentiation pathway directs the conversion of peripheral cNK cells into uterine trNK cells during murine pregnancy" should be slightly toned down.

(3) Figures 2-4

It is unclear whether the littermate controls are floxed mice or floxhet-Ncr1iCre mice? This distinction is important, as Ncr1iCre expression itself could potentially lead to a phenotype.

Reviewer #2 (Public review):

In their manuscript "TGF-β drives the conversion of conventional NK cells into uterine tissue-resident NK cells to support murine pregnancy", Yokoyama and colleagues investigate the role of Tgfbr2 expression by NK cells in the formation of tissue-resident uterine NK cells and subsequent importance in murine pregnancy. By transferring congenic splenic conventional NK cells into pregnant mice, they show conversion of circulating NK cells into uterine ivCD45 negative tissue-resident NK cells. When interfering with the formation of uterine trNK cells, spiral artery remodelling was impaired, fetal resorption rates were increased, and litter sizes were reduced.

Generally, this is a research topic of high interest, yet the manuscript is lacking detailed mechanistic insights, and some questions remain open. At the current state, the data represent an interesting characterisation of the Tgfbr2-fl/fl Ncr1-Cre mice in pregnancy, but considering (a) the recent publication by the group (Reference 17) on the role of Eomes+ cNK cells during pregnancy, (b) the previously described role of Tgfbr2 and autocrine TGFb expression for uterine NK cell differentiation in virgin mice (also cited by the authors), and (c) the well-known relevance of uterine NK cells during pregnancy, additional experiments addressing the specific role of Tgfb during pregnancy would help to improve novelty and significance of the manuscript. To this end, the following aspects should be discussed and, where applicable, experimentally addressed by the authors:

(1) The authors suggest cNK extravasation and local differentiation into iv- trNK.

Can it be estimated how much this process contributes to the trNK pool vs. a potential local proliferation of already existing trNK? How do absolute numbers of CD49a+ Eomes+ trNK change during pregnancies? (In Figure 1A, the cell numbers of CD49a+ Eomes+ trNK seem to go down dramatically between gd 6.5 and 14.5). The plot in 1B could also include absolute numbers of ILC1s and trNKs. Would recruited cNK cells compensate for a potential loss of CD49a+ Eomes+ trNK?

(2) Figure 1C: 2.5

Mio cNK cells have been transferred, but only very few cells can be detected within the uterus (concatenated FACS plot shown). What may represent the limit to generate uterine trNK out of cNK? Is the niche supporting cNK-trNK differentiation limited? Is it only a specific subset of (splenic) cNK capable of differentiating into trNK? Is gd 0.5 the optimal timepoint for the transfer? Is there continuous recruitment of cNK into the uterus and differentiation into trNK, or is it enhanced at specific timepoints of pregnancy? Could there be local proliferation of cNK-derived trNK? This could be studied by proliferation dye dilution of WT cNK cells in this transfer-setup.

(3) The authors should consider inducible Tgfbr2 deletion (e.g. with Tamoxifen-inducible Cre) to enable development of the uterine NK compartment in virgin mice and only ablate trNK differentiation during pregnancy. This could help to estimate the turnover of cNK into trNK, or to understand if constant cNK recruitment is required to form the uterine trNK compartment during pregnancy.

(4) Did the authors consider transfer of Tgfbr2-floxed Ncr1-Cre cNK in the same setup as in Fig. 1C? This experiment could confirm the requirement of Tgfbr-dependent signalling for cNK to trNK conversion during pregnancy versus effects of Tgfb signals on trNK numbers in the uterus at steady state (before pregnancy).

(5) Figures 2D/E

The authors should state that ILC1s are reduced in the virgin uterus of female Tgfbr2-floxed or Tgfb1-floxed Ncr1-Cre mice and cite the relevant work (the Ref #29 discussed in this context did not show that?). It would be helpful to include an analysis of all three uterine ILC subsets in steady state. This could help to answer the question if the cNK cell changes are pregnancy-specific or a general phenomenon in Tgfbr2-floxed Ncr1-Cre mice.

(6) Figure 2E

Please phrase more carefully about the "concomitant increase" of cNKs, since this increase is much less pronounced compared to the very strong reduction (absence) of trNKs in Tgfbr2-floxed Ncr1-Cre mice. Do the authors suggest that cNKs are halted at this stage and cannot differentiate into trNK, based on these data?

(7) Figure 3/4

Can the reduced litter size and the abnormal spiral artery formation be rescued by transfer of WT cNK into Tgfbr2-floxed Ncr1-Cre mice?

eLife Assessment

This important study reports that the human posterior inferotemporal cortex (hPIT) functions as an attentional priority map, integrating both top-down and bottom-up attentional signals rather than serving solely as an object-processing region. The experiments and analyses are well conducted and provide compelling evidence that hPIT bridges dorsal and ventral attention networks and is robustly modulated by attention across diverse visual tasks. The study will be relevant for researchers investigating visual attention, high-level visual cortex, and the neural mechanisms that integrate endogenous and exogenous attentional control.

Reviewer #1 (Public review):

The manuscript titled "The distinct role of human PIT in attention control" by Huang et al. investigates the role of the human posterior inferotemporal cortex (hPIT) in spatial attention. Using fMRI experiments and resting-state connectivity analyses, the authors present compelling evidence that hPIT is not merely an object-processing area, but also functions as an attentional priority map, integrating both top-down and bottom-up attentional processes. This challenges the traditional view that attentional control is localized primarily in frontoparietal networks.

The manuscript is strong and of high potential interest to the cognitive neuroscience community. Below, I raise questions and suggestions to help with the reliability, methodology, and interpretation of the findings.

(1) The authors argue that hPIT satisfies the criteria for a priority map, but a clearer justification would strengthen this claim. For example, how does hPIT meet all four widely recognized criteria, such as spatial selectivity, attentional modulation, feature invariance, and input integration, when compared to classical regions such as LIP or FEF? A more systematic summary of how hPIT meets these benchmarks would be helpful. Additionally, to what extent are the observed attentional modulations in hPIT independent of general task difficulty or behavioral performance?

(2) The authors report that hPIT modulation is invariant to stimulus category, but there appear to be subtle category-related effects in the data. Were the face, scene, and scrambled images matched not only in terms of luminance and spatial frequency, but also in terms of factors such as semantic familiarity and emotional salience? This may influence attentional engagement and bias interpretation.

(3) The result that attentional load modulates hPIT is important and adds depth to the main conclusions. However, some clarifications would help with the interpretation. For example, were there observable individual differences in the strength of attentional modulation? How consistent were these effects across participants?

(4) The resting-state data reveal strong connections between hPIT and both dorsal and ventral attention networks. However, the analysis is correlational. Are there any complementary insights from task-based functional connectivity or latency analyses that support a directional flow of information involving hPIT? In addition, do the authors interpret hPIT primarily as a convergence hub receiving input from both DAN and VAN, or as a potential control node capable of influencing activity in these networks? Also, were there any notable differences between hemispheres in either the connectivity patterns or attentional modulation?

(5) A few additional questions arise regarding the anatomical characteristics of hPIT: How consistent were its location and size across participants? Were there any cases where hPIT could not be reliably defined? Given the proximity of hPIT to FFA and LOp, how was overlap avoided in ROI definition? Were the functional boundaries confirmed using independent contrasts?

Comments on revisions:

The authors have successfully addressed my previous questions and concerns. The public comments above reflect my views on the initial submission and, in my opinion, will remain helpful for general readers. Given this, I do not have additional public comments and will keep my previous public review unchanged.

Reviewer #2 (Public review):

Summary

This study investigates the role of the human posterior inferotemporal cortex (hPIT) in attentional control, proposing that hPIT serves as an attentional priority map that integrates both top-down (endogenous) and bottom-up (exogenous) attentional processes. The authors conducted three types of fMRI experiments and collected resting-state data from 15 participants. In Experiment 1, using three different spatial attention tasks, they identified the hPIT region and demonstrated that this area is modulated by attention across tasks. In Experiment 2, by manipulating the presence or absence of visual stimuli, they showed that hPIT exhibits strong attentional modulation in both conditions, suggesting its involvement in both bottom-up and top-down attention. Experiment 3 examined the sensitivity of hPIT to stimulus features and attentional load, revealing that hPIT is insensitive to stimulus category but responsive to task load - further supporting its role as an attentional priority map. Finally, resting-state functional connectivity analyses showed that hPIT is connected to both dorsal and ventral attention networks, suggesting its potential role as a bridge between the two systems. These findings extend prior work on monkey PITd and provide new insights into the integration of endogenous and exogenous attention.

Strength

(1) The study is innovative in its use of specially designed spatial attention tasks to localize and validate hPIT, and in exploring the region's role in integrating both endogenous and exogenous attention, as prior works focus primarily on its involvement in endogenous attention.

(2) The authors provided very comprehensive experiment designs with clear figures and detailed descriptions.

(3) A broad range of analyses was conducted to support the hypothesis that hPIT functions as an attentional priority map -- including experiments of attentional modulation under both top-down and bottom-up conditions, sensitivity to stimulus features and task load, and resting-state functional connectivity. These analyses showed consistent results.

(4) Multiple appropriate statistical analyses - including t-tests, ANOVAs, and post-hoc tests-were conducted, and the results are clearly reported.

Comments on revisions:

The authors have addressed our comments in their revised manuscript and in their response to the reviewers. We don't have any further suggestions or comments.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

The manuscript titled "The distinct role of human PIT in attention control" by Huang et al. investigates the role of the human posterior inferotemporal cortex (hPIT) in spatial attention. Using fMRI experiments and resting-state connectivity analyses, the authors present compelling evidence that hPIT is not merely an object-processing area, but also functions as an attentional priority map, integrating both top-down and bottom-up attentional processes. This challenges the traditional view that attentional control is localized primarily in frontoparietal networks.

The manuscript is strong and of high potential interest to the cognitive neuroscience community. Below, I raise questions and suggestions to help with the reliability, methodology, and interpretation of the findings.

Thank you for a nice summary of the key points of our study. Below you will find our reply to your questions.

(1) The authors argue that hPIT satisfies the criteria for a priority map, but a clearer justification would strengthen this claim. For example, how does hPIT meet all four widely recognized criteria, such as spatial selectivity, attentional modulation, feature invariance, and input integration, when compared to classical regions such as LIP or FEF? A more systematic summary of how hPIT meets these benchmarks would be helpful. Additionally, to what extent are the observed attentional modulations in hPIT independent of general task difficulty or behavioral performance?

Great suggestions! For the first suggestion, we have included a clearer justification in the discussion part of manuscript (line 405-406). For the second one, all participants received task practice prior to scanning, and task accuracy exceeded 90%, suggesting the tasks were not overly demanding. Although ceiling effects limit the interpretability of behavioral-performance correlations, we argue that higher task demands would likely require greater attentional effort, leading to stronger modulation in hPIT, which aligns with our findings.

(2) The authors report that hPIT modulation is invariant to stimulus category, but there appear to be subtle category-related effects in the data. Were the face, scene, and scrambled images matched not only in terms of luminance and spatial frequency, but also in terms of factors such as semantic familiarity and emotional salience? This may influence attentional engagement and bias interpretation.

The response of hPIT is not sensitive to stimulus category, but attentional modulation in hPIT is slightly stronger to faces than scenes and scrambled images. Although faces used in the task had neutral expressions and the scene pictures were also neutral, we acknowledge that we indeed cannot exclusively eliminate the possibility that potential semantic familiarity or emotional salience may contribute to the subtle category-related effects in the results of experiment 3. This limitation has been noted in the discussion part of manuscript (line 440-442).

(3) The result that attentional load modulates hPIT is important and adds depth to the main conclusions. However, some clarifications would help with the interpretation. For example, were there observable individual differences in the strength of attentional modulation? How consistent were these effects across participants?

Yes, individual differences exist. In the manuscript, we have included individual subject data points in the figure 6B. No data exceeded three standard deviations from the group mean, suggesting that the attentional modulation effects were generally consistent across participants.

(4) The resting-state data reveal strong connections between hPIT and both dorsal and ventral attention networks. However, the analysis is correlational. Are there any complementary insights from task-based functional connectivity or latency analyses that support a directional flow of information involving hPIT? In addition, do the authors interpret hPIT primarily as a convergence hub receiving input from both DAN and VAN, or as a potential control node capable of influencing activity in these networks? Also, were there any notable differences between hemispheres in either the connectivity patterns or attentional modulation?

Though it’s hard to generate directional flow of information from fMRI due to the low temporal resolution. We agree that besides resting-state connection, task-based functional connectivity analyses would have the potential to provide additional information about whether hPIT serves as a convergence node or a control hub. We have conducted task-based functional connectivity analyses, specifically PPI, using data from experiment 2, which revealed task-modulated right hPIT connectivity with FFA, LOp, and TPJ, suggesting hPIT may allocate attentional resources to object-processing regions following priority map generation (line 378-383). Given the limited number of significant PPI results and the inherent constraints of fMRI in capturing fast or transient attention-related interactions, the present data do not allow us to determine the role of hPIT. Future studies combining effective connectivity or causal perturbation methods (e.g., DCM, TMS-fMRI) would be ideal to test whether hPIT acts as a control node influencing activity within DAN and VAN.

We also observed modest hemispheric asymmetries in connectivity—for instance, both left and right hPIT showed stronger connectivity with right-hemisphere attention nodes. This has been described in the results part of manuscript (line 373-377).

(5) A few additional questions arise regarding the anatomical characteristics of hPIT: How consistent were its location and size across participants? Were there any cases where hPIT could not be reliably defined? Given the proximity of hPIT to FFA and LOp, how was overlap avoided in ROI definition? Were the functional boundaries confirmed using independent contrasts?

We can see a relatively consistent size and location of hPIT across subjects in Supplementary Figure 1, where the voxel size and location for individual subjects reported. The consistency also demonstrated by figure 4C.

We avoided overlap with the FFA and LOp by manually delineating the hPIT which is defined by conjunction maps across three tasks and by avoiding overlapping voxels. The FFA was defined using an independent contrast (Exp3 contrast [face-scene]) and the Lop location was defined by anatomical parcellation (Glasser et al., 2016).

Reviewer #2 (Public review):

Summary

This study investigates the role of the human posterior inferotemporal cortex (hPIT) in attentional control, proposing that hPIT serves as an attentional priority map that integrates both top-down (endogenous) and bottom-up (exogenous) attentional processes. The authors conducted three types of fMRI experiments and collected resting-state data from 15 participants. In Experiment 1, using three different spatial attention tasks, they identified the hPIT region and demonstrated that this area is modulated by attention across tasks. In Experiment 2, by manipulating the presence or absence of visual stimuli, they showed that hPIT exhibits strong attentional modulation in both conditions, suggesting its involvement in both bottom-up and top-down attention. Experiment 3 examined the sensitivity of hPIT to stimulus features and attentional load, revealing that hPIT is insensitive to stimulus category but responsive to task load - further supporting its role as an attentional priority map. Finally, resting-state functional connectivity analyses showed that hPIT is connected to both dorsal and ventral attention networks, suggesting its potential role as a bridge between the two systems. These findings extend prior work on monkey PITd and provide new insights into the integration of endogenous and exogenous attention.

Strengths

(1) The study is innovative in its use of specially designed spatial attention tasks to localize and validate hPIT, and in exploring the region's role in integrating both endogenous and exogenous attention, as prior works focus primarily on its involvement in endogenous attention.

(2) The authors provided very comprehensive experiment designs with clear figures and detailed descriptions.

(3) A broad range of analyses was conducted to support the hypothesis that hPIT functions as an attentional priority map -- including experiments of attentional modulation under both top-down and bottom-up conditions, sensitivity to stimulus features and task load, and resting-state functional connectivity. These analyses showed consistent results.

(4) Multiple appropriate statistical analyses - including t-tests, ANOVAs, and post-hoc tests - were conducted, and the results are clearly reported.

Thank you for a nice summary of the key points and strengths of our study.

Weaknesses

(1) The sample size is relatively small (n = 15), and inter-subject variability is big in Figures 5 and 6, as seen in the spread of individual data points and error bars. The analysis of attention-modulated voxel map intersections appears to be influenced by multiple outliers.

We agree that the sample size (n = 15) is not ideal, and we acknowledge that some data points in Figures 5 and 6 appear to be potential outliers. However, according to conventional outlier detection criteria, all data points fell within three standard deviations of the group mean and were therefore retained for analysis.

Moreover, the attention-modulated voxel intersection map shown in Figure 4C is insensitive to outliers, because the intersection plotted is based on the number of subjects

(2) The authors acknowledge important limitations, including the lack of exploration of feature-based attention and the temporal constraints inherent to fMRI.

Yes, we have mentioned these limitations in the discussion.

(3) Prior research has established that regions such as the prefrontal cortex (PFC) and posterior parietal cortex (PPC) are involved in both endogenous and exogenous attention and have been proposed as attentional priority maps. It remains unclear what is uniquely contributed by hPIT, how it functionally interacts with these classical attentional hubs, and whether its role is complementary or redundant. The study would benefit from more direct comparisons with these regions.

In this study, we define the ROI base on intersection across three different types of spatial attention tasks, which is a stricter criterion. And the results didn’t reveal spatial attentional modulation across tasks besides PITd. This could be due to the lack of lateralized responses in PFC/PPC. To evaluate whether a region qualifies as a priority map, we applied four widely accepted criteria (as mentioned in introduction). While dorsal and ventral attention network (DAN and VAN) regions can be considered supportive components of the priority map system, our findings suggest that among the regions tested, only hPIT fully meets all criteria. In Experiment 2, we included regions such as VFC (as part of PFC) and IPS (as part of PPC), and our findings suggest these areas are more involved in top-down attention. In the revision, we have performed additional analysis on PPC (IPS) and PFC (FEF, VFC), shown in Figure S2.

(4) The functional connectivity analysis is only performed on resting-state data, and this approach does not capture context-dependent interactions. Task-based data analysis can provide stronger evidence.

We acknowledge that resting-state FC is limited in assessing task-specific communication. To further investigate the role of hPIT, we have conducted PPI analysis, which revealed task-modulated right hPIT connectivity in attention allocation (line 378-383).

(5) The study does not report whether attentional modulation in hPIT is consistent across the two hemispheres. A comparison of hemispheric effects could provide important insight into lateralization and inter-individual variability, especially given the bilateral localization of hPIT.

We thank the reviewer for this suggestion. hPIT was localized bilaterally using the same intersection-based method in Experiment 1. We have now performed additional analysis and found hemispheric differences in hPIT attentional modulation (Experiment 2). Besides, we also found in Experiment 3, the difference of load modulation (averaged across stimulus categories) in left and right hPIT was not significant. These results have been reported in the results part of manuscript (line 347-351).

eLife Assessment

This study reports an mRNA-based strategy for restoring sperm motility in a mouse model of monogenic male infertility. The work is technically innovative and potentially valuable, as it demonstrates feasibility of in vivo testicular mRNA delivery without genomic integration of foreign DNA. However, although partial recovery of sperm motility is supported, the evidence for meaningful restoration of fertility remains incomplete, with weak IVF outcomes and difficult-to-interpret ICSI results. In addition, mechanistic questions regarding the persistence of mRNA and the specificity of germ-cell targeting remain insufficiently resolved, limiting the strength of the authors' conclusions.

Reviewer #4 (Public Review):

I maintain that the images in Figure 12 (new Figure 14) do not support the authors' interpretation that 2-cell embryos resulted from in vitro fertilization (IVF) of Amrc-/- rescued sperm. They are clearly not normal 2-cell embryos and instead look very much like fragmented eggs that can be seen occasionally following in vitro fertilization procedures even when that is done with wild type eggs and sperm. The only portion of current Figure 14 that has normal looking 2-cell embryos is in panel 14A4, where wild type B6D2 sperm were used. Even in that panel, there are some fragmented eggs that the authors identify as 2-cell embryos.

The authors offer the explanation that CD1 eggs fertilized by B6D2F1 hybrid male sperm do not develop beyond the 2-cell stage, citing a 2008 paper published in Biology of Reproduction by Fernandez-Goonzalez et al. I read through that paper very carefully and even had a colleague read through it in case I missed something, but that paper says nothing at all about strain incompatibilities, much less 2-cell arrest due to them. The only crosses done in that paper are CD1 eggs x CD1 sperm and B6D2 eggs x B6D2 sperm, all by intracytoplasmic sperm injection and not standard in vitro fertilization. [Note that the paper does mention performing in vitro fertilization but says nothing about how it was done or what mouse strains were used.] I even searched the literature for information regarding incompatibility between these strains and could find nothing relevant. But even if the authors are correct and there happens to be a strain incompatibility and 2-cell arrest is expected, what the authors are calling 2-cell embryos are clearly not.

A second explanation offered by the authors is that they used collagenase to remove the cumulus cells and that this may have affected the appearance of the embryos. This technique is actually used to remove both the cumulus cells and the zona pellucida and has been described as a gentler way to do so than other standard methods (hyaluronidase treatment followed by acid Tyrodes to remove the zona pellucida) (Yamatoya et al., Reprod Med Biol 2011, DOI 10.1007/s12522-011-0075-8). I think it is highly relevant to the current study that the method they used to remove cumulus cells also dissolves the zona, either partially or completely. Given that many of the eggs, fragmented eggs, and 2-cell embryos (from the WT sperm) in Figure 14A are lacking a zona pellucida, it seems very likely that many of the eggs were either zona-free or had partial zona dissolution from the start. In fact, the authors state in the Methods section that "Cumulus-free and zona-free eggs were collected..." for how IVF was done. Partial zona dissolution is standard in some protocols for performing IVF using frozen mouse sperm, which usually have much lower motility and overall efficacy than fresh sperm. In any case, it would improve transparency if the manuscript made clear somewhere other than buried in the Methods that the IVF procedure was done on eggs with partially or fully removed zonas, to allow proper interpretation.

In the rebuttal, the authors go on to state: "To provide additional functional evidence, we complemented the IVF experiments with ICSI using rescued Armc2-/- sperm and B6D2 oocytes, which allowed embryos to develop to the blastocyst stage. In these experiments, 25% of injected oocytes reached the blastocyst stage with rescued sperm compared to 13% for untreated Armc2-/- sperm (Supplementary Fig. 9) These results support the functional competence of rescued sperm and demonstrate partial recovery of fertilization ability following Armc2 mRNA electroporation."

Their conclusion that the data support partial recovery of fertilization ability following Armc2 mRNA electroporation in my opinion has no basis. This experiment was done only once, and no information is provided regarding how many eggs underwent ICSI or how many reached the blastocyst stage. The authors claim that the rescued sperm were better than the Armc2-/- sperm in producing blastocysts, but this is based on a simple percentage report of 25% vs 13% without any statistical analysis, even on the results from the single experiment presented.

Overall, the paper shows rescue of some sperm motility by the new method they use, and the new title is therefore appropriate. The authors have also dealt reasonably with many of the original concerns regarding documenting that their methodology was effective in producing protein (at least the GFP marker) in spermatogenic cells. In my view the authors have, however, not shown any indication of functional recovery over what is already known for the knockout sperm, that ICSI can support blastocyst stage embryo development. They also have not, in my view, justified the claims at the end of the abstract "These motile sperm were able to produce embryos by IVF..." and that "...mRNA electroporation can restore...partially fertilizing ability..."

Reviewer #5 (Public Review) :

While the study presents an innovative and potentially impactful mRNA-based approach for addressing monogenic causes of male infertility, several significant weaknesses limit the strength of the authors' central conclusions.

First, the functional evidence for true fertility restoration remains incomplete. Although the authors convincingly demonstrate partial recovery of sperm motility, the downstream reproductive outcomes, particularly for IVF, are weak. Importantly, these concerns are shared by all three reviewers and the former Reviewing Editor, and to my eye they are both thoughtfully articulated and well warranted. The ICSI data show modest improvement, but this rescue is difficult to interpret.

In parallel, significant mechanistic questions persist regarding the unusually prolonged persistence of naked mRNA and reporter protein expression in germ cells, which is not fully reconciled with established mRNA and protein half-life biology and is supported largely by inference rather than by direct decay measurements.

Finally, although the authors have conducted additional cellular analyses, concerns about the extent and specificity of germ-cell targeting versus Sertoli-cell expression remain unresolved. Together, these issues do not negate the technical novelty of the work, but they do constrain the confidence with which the current dataset can support the authors' strongest therapeutic claims.

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

The authors assess the effectiveness of electroporating mRNA into male germ cells to rescue the expression of proteins required for spermatogenesis progression in individuals where these proteins are mutated or depleted. To set up the methodology, they first evaluated the expression of reporter proteins in wild-type mice, which showed expression in germ cells for over two weeks. Then, they attempted to recover fertility in a model of late spermatogenesis arrest that produces immotile sperm. By electroporating the mutated protein, the authors recovered the motility of ~5% of the sperm; although the sperm regenerated was not able to produce offspring using IVF, the embryos reached the 2-cell state (in contrast to controls that did not progress past the zygote state).

This is a comprehensive evaluation of the mRNA methodology with multiple strengths. First, the authors show that naked synthetic RNA, purchased from a commercial source or generated in the laboratory with simple methods, is enough to express exogenous proteins in testicular germ cells. The authors compared RNA to DNA electroporation and found that germ cells are efficiently electroporated with RNA, but not DNA. The differences between these constructs were evaluated using in vivo imaging to track the reporter signal in individual animals through time. To understand how the reporter proteins affect the results of the experiments, the authors used different reporters: two fluorescent (eGFP and mCherry) and one bioluminescent (Luciferase). Although they observed differences among reporters, in every case expression lasted for at least two weeks. The authors used a relevant system to study the therapeutic potential of RNA electroporation. The ARMC2-deficient animals have impaired sperm motility phenotype that affects only the later stages of spermatogenesis. The authors showed that sperm motility was recovered to ~5%, which is remarkable due to the small fraction of germ cells electroporated with RNA with the current protocol. The sperm motility parameters were thoroughly assessed by CASA. The 3D reconstruction of an electroporated testis using state-of-the-art methods to show the electroporated regions is compelling.

The main weakness of the manuscript is that although the authors manage to recover motility in a small fraction of the sperm population, it is unclear whether the increased sperm quality is substantial to improve assisted reproduction outcomes. The authors found that the rescued sperm could be used to obtain 2-cell embryos via IVF, but no evidence for more advanced stages of embryo differentiation was provided. The motile rescued sperm was also successfully used to generate blastocyst by ICSI, but the statistical significance of the rate of blastocyst production compared to non-rescued sperm remains unclear. The title is thus an overstatement since fertility was never restored for IVF, and the mutant sperm was already able to produce blastocysts without the electroporation intervention.

Overall, the authors clearly show that electroporating mRNA can improve spermatogenesis as demonstrated by the generation of motile sperm in the ARMC2 KO mouse model.

We thank the reviewer for this thoughtful and constructive comment. We agree that our study demonstrates a partial functional recovery of spermatogenesis rather than a complete restoration of fertility. Our main objective was to establish and validate a proof-of-concept approach showing that mRNA electroporation can rescue the expression of a missing or mutated protein in post-meiotic germ cells and result in the production of motile sperm.

To address the reviewer’s concern, we have the title and discussion to more accurately reflect the scope of our findings. The new title reads:

“Sperm motility in mice with oligo-astheno-teratozoospermia restored by in vivo injection and electroporation of naked mRNA”

In the manuscript, we now emphasize that while motility recovery was significant, complete fertility restoration was not achieved. We have also clarified that:

The 5% recovery in motile sperm represents a substantial improvement considering the small population of germ cells reached by the current electroporation method.

The 2-cell embryo formation observed after IVF serves as a strong indication of partial functional recovery

Finally, we now explicitly state in the Discussion that this approach should be considered a therapeutic proof-of-concept, demonstrating feasibility and potential, rather than a fully curative intervention.

Reviewer #2 (Public review):

The authors inject, into the rete testes, mRNA and plasmids encoding mRNAs for GFP and then ARMC2 (into infertile Armc2 KO mice) in a gene therapy approach to express exogenous proteins in male germ cells. They do show GFP epifluorescence and ARMC2 protein in KO tissues, although the evidence presented is weak. Overall, the data do not necessarily make sense given the biology of spermatogenesis and more rigorous testing of this model is required to fully support the conclusions, that gene therapy can be used to rescue male infertility.

In this revision, the authors attempt to respond to the critiques from the first round of reviews. While they did address many of the minor concerns, there are still a number to be addressed. With that said, the data still do not support the conclusions of the manuscript.

We thank the reviewer for their careful and detailed assessment of our manuscript. We appreciate the concerns raised regarding mRNA stability, GFP localization, and the interpretation of spermatogenesis stages, and we have addressed these points in the manuscript and in the responses below.

(1) The authors have not satisfactorily provided an explanation for how a naked mRNA can persist and direct expression of GFP or luciferase for ~3 weeks. The most stable mRNAs in mammalian cells have half-lives of ~24-60 hours. The stability of the injected mRNAs should be evaluated and reported using cell lines. GFP protein's half-life is ~26 hours, and luciferase protein's half-life is ~2 hours.

We thank the reviewer for this important comment. The stability of mRNA-GFP was assessed by RT-QPCR in HEK cells and seminiferous tubule cells (Fig. 5). mRNA-GFP was detected for up to 60 hours in HEK cells and for up to two weeks in seminiferous tubule cells (Fig. 5A). Together, these results suggest that the long-lasting fluorescence observed in our experiments reflects a combination of transcript stability, efficient translation within germ cells and the slow protein turnover that is typical of the spermatogenic lineage.

(2) There is no convincing data shown in Figs. 1-8 that the GFP is even expressed in germ cells, which is obviously a prerequisite for the Armc2 KO rescue experiment shown in the later figures! In fact, to this reviewer the GFP appears to be in Sertoli cell cytoplasm, which spans the epithelium and surrounds germ cells - thus, it can be oft-confused with germ cells. In addition, if it is in germ cells, then the authors should be able to show, on subsequent days, that it is present in clones of germ cells that are maturing. Due to intracellular bridges, a molecule like GFP has been shown to diffuse readily and rapidly (in a matter of minutes) between adjacent germ cells. To clarify, the authors must generate single cell suspensions and immunostain for GFP using any of a number of excellent commercially-available antibodies to verify it is present in germ cells. It should also be present in sperm, if it is indeed in the germline.

We thank the reviewer for this insightful comment. To directly address the concern, we performed additional experiments to assess GFP expression in germ cells following in vivo mRNA delivery. GFP-encoding mRNA was injected and electroporated into the testes on day 0. On day 1, testes were collected, enzymatically dissociated, and the resulting seminiferous tubule cell suspensions were cultured for 12 hours. Live cells were then analyzed by fluorescence microscopy (Fig. 10).

We observed GFP expression in various germ cell types, including pachytene spermatocytes (53,4 %) (Fig 10 A-), round spermatids (25 %) (Fig 10B-E) and in elongated spermatids (11,4%) (Fig 10 C-E). The identification of these cell types was based on DAPI nuclear staining patterns, cell size fig 10 F, non-adherent characteristics, and the use of an enzymatic dissociation protocol.

Fluorescence imaging revealed strong cytoplasmic GFP signals in each of these populations, confirming efficient transfection and translation of the delivered mRNA. These results demonstrate that the in vivo injection and electroporation protocol enables effective mRNA transfection across multiple stages of spermatogenesis. These results confirm that the injected mRNA is efficiently translated in germ cells at various stages of spermatogenesis. Together, these data validate the germ cell-specific nature of the GFP signal, supporting the Armc2 KO rescue experiments.

As mentioned previously, we assessed the stability of mRNA-GFP using RT-QPCR in HEK cells and seminiferous tubule cells (see Fig. 5). mRNA-GFP was detected for up to 60 hours in HEK cells and for up to two weeks in seminiferous tubule cells. Together, these results suggest that the long-lasting fluorescence observed in our experiments reflects a combination of transcript stability and local translation within germ cells, as well as the slow protein turnover typical of the spermatogenic lineage.

Other comments:

70-1 This is an incorrect interpretation of the findings from Ref 5 - that review stated there were ~2,000 testis-enriched genes, but that does not mean "the whole process involves around two thousand of genes"

We thank the reviewer for this helpful comment. We agree that our previous phrasing was imprecise. We have revised the sentence to clarify that approximately 2,000 genes show testis-enriched expression, rather than implying that the entire spermatogenic process is limited to these genes. The corrected sentence now reads:

“Spermatogenesis involves the coordinated expression of a large number of genes, with approximately 2,000 showing testis-enriched expression, about 60% of which are expressed exclusively in the testes”

74 would specify 'male':

we have now specified it as you suggested.

79-84 Are the concerns with ICSI due to the procedure itself, or the fact that it's often used when there is likely to be a genetic issue with the male whose sperm was used? This should be clarified if possible, using references from the literature, as this reviewer imagines this could be a rather contentious issue with clinicians who routinely use this procedure, even in cases where IVF would very likely have worked:

We thank the reviewer for this important comment. Concerns about ICSI outcomes indeed reflect two partly overlapping causes: the procedure itself (direct sperm injection and associated laboratory manipulations) and the clinical/genetic background of couples undergoing ICSI (especially men with severe male-factor infertility). Large reviews and meta-analyses report a small increase in some perinatal and congenital risks after ART/ICSI, but these studies conclude that it is difficult to fully disentangle procedural effects from parental factors. Importantly, genetic or epigenetic abnormalities in the male (which motivate use of ICSI) likely contribute to adverse outcomes in offspring, while some studies also suggest that ICSI-specific manipulations may alter epigenetic marks in embryos. For these reasons professional bodies recommend reserving ICSI for appropriate male-factor indications rather than as routine insemination for non-male-factor cases

We have revised the text accordingly to clarify this distinction:

“ICSI can efficiently overcome the problems faced.  Nevertheless, concerns persist regarding the potential risks associated with this technique, including blastogenesis defect, cardiovascular defect, gastrointestinal defect, musculoskeletal defect, orofacial defect, leukemia, central nervous system tumors, and solid tumors [1-4]. Statistical analyses of birth records have demonstrated an elevated risk of birth defects, with a 30-40 % increased  likelihood in cases involving ICSI [1-4], and a prevalence of birth defects between 1 % and 4 % [3]. It is important to note, however, that the origin of these risks remains debated. Several large epidemiological and mechanistic studies indicate that both the procedure itself (direct microinjection and in vitro manipulation) and the underlying genetic or epigenetic abnormalities often present in men requiring ICSI contribute to the observed outcomes [1, 3] [5, 6] . To overcome these drawbacks, a number of experimental strategies have been proposed to bypass ARTs and restore spermatogenesis and fertility, including gene therapy [7-10].”

199 Codon optimization improvement of mRNA stability needs a reference;

We have added the references accordingly: [11-15]

In one study using yeast transcripts, optimization improved RNA stability on the order of minutes (e.g., from ~5 minutes to ~17 minutes); is there some evidence that it could be increased dramatically to days or weeks?

We agree with the reviewer that codon optimization can enhance mRNA stability, but available evidence indicates that this effect is moderate. In Saccharomyces cerevisiae, Presnyak et al. (2015) [16] showed that codon optimization increased mRNA half-life from approximately 5 minutes to ~17 minutes, representing a several-fold improvement rather than a shift to days or weeks. Similar codon-dependent stabilization has been observed in mammalian systems, where transcripts enriched in optimal codons exhibit longer half-lives and enhanced translation efficiency [11]; [17]). However, these studies consistently report effects on the scale of minutes to hours. In mammalian cells, the prolonged stability of therapeutic or vaccine mRNAs—lasting for days—is primarily achieved through additional features such as optimized untranslated regions, chemical nucleotide modifications (e.g., N¹-methylpseudouridine), and protective delivery systems, rather than codon usage alone ([18]; [19]).

Other molecular optimizations that improve in vivo mRNA stability and translation include a poly(A) tail, which binds poly(A)-binding proteins to protect the transcript from 3′ exonuclease degradation and promotes ribosome recycling, and a CleanCap structure at the 5′ end, which mimics the natural Cap 1 configuration, protects against 5′ exonuclease attack, and enhances translational initiation [11-15]. Together, these modifications act synergistically to stabilize the transcript and support efficient translation.

472-3 The reported half-life of EGFP is ~36 hours - so, if the mRNA is unstable (and not measured, but certainly could be estimated by qRT-PCR detection of the transcript on subsequent days after injection) and EGFP is comparatively more stable (but still hours), how does EGFP persist for 21 days after injection of naked mRNA??

We thank the reviewer for this important comment. The stability of mRNA-GFP was assessed by RT-QPCR in HEK cells and seminiferous tubule cells (Fig. 5). mRNA-GFP was detected for up to 60 hours in HEK cells and for up to two weeks in seminiferous tubule cells (Fig. 5). Together, these results suggest that the long-lasting fluorescence observed in our experiments reflects a combination of transcript stability, efficient translation within germ cells and the slow protein turnover that is typical of the spermatogenic lineage.

Curious why the authors were unable to get anti-GFP to work in immunostaining?

We appreciate the reviewer’s question. We attempted to detect GFP using several commercially available anti-GFP antibodies under various standard immunostaining conditions. However, in our hands, these antibodies consistently produced either no signal or high background staining, making the results unreliable. We therefore relied on direct detection of GFP fluorescence, which provides a more accurate and specific readout of protein expression in our system.

In Fig. 3-4, the GFP signals are unremarkable, in that they cannot be fairly attributed to any structure or cell type - they just look like blobs; and why, in Fig. 4D-E, why does the GFP signal appear stronger at 21 days than 15 days? And why is it completely gone by 28 days? This data is unconvincing.

We would like to thank the reviewer for their comments. Figure 3–4 provides a global overview of GFP expression on the surface of the testis. The entire testis was imaged using an inverted epifluorescence microscope, and the GFP signal represents a composite of multiple seminiferous tubules across the tissue surface. Due to this whole-organ imaging approach, it is not possible to resolve individual structures such as the basement membrane or lumen, which is why the signals may appear as diffuse “blobs.”

Regarding the time-course in Figure 4D–E, the apparent increase in GFP signal at 21 days compared with 15 days likely reflects accumulation and translation of the delivered mRNA in germ cells over time, whereas the absence of signal at 28 days corresponds to the natural turnover and degradation of GFP protein and mRNA in the tissue. We hope this explanation clarifies the observed patterns of fluorescence.

If the authors did a single cell suspension, what types or percentage of cells would be GFP+? Since germ cells are not adherent in culture, a simple experiment could be done whereby a single cell suspension could be made, cultured for 4-6 hours, and non-adherent cells "shaken off" and imaged vs adherent cells. Cells could also be fixed and immunostained for GFP, which has worked in many other labs using anti-GFP.

We thank the reviewer for this insightful comment. To directly address the concern, we performed additional experiments to assess GFP expression in germ cells following in vivo mRNA delivery. GFP-encoding mRNA was injected and electroporated into the testes on day 0. On day 1, testes were collected, enzymatically dissociated, and the resulting seminiferous tubule cell suspensions were cultured for 12 hours. Live cells were then analyzed by fluorescence microscopy (Fig. 10).

We observed GFP expression in various germ cell types, including pachytene spermatocytes (53,4 %) (Fig 10 A-), round spermatids (25 %) (Fig 10B-E) and in elongated spermatids (11,4%) (Fig 10 C-E). The identification of these cell types was based on DAPI nuclear staining patterns, cell size fig 10 F, non-adherent characteristics, and the use of an enzymatic dissociation protocol.

Fluorescence imaging revealed strong cytoplasmic GFP signals in each of these populations, confirming efficient transfection and translation of the delivered mRNA. These results demonstrate that the in vivo injection and electroporation protocol enables effective mRNA transfection across multiple stages of spermatogenesis.

These results confirm that the injected mRNA is efficiently translated in germ cells at various stages of spermatogenesis. Together, these data validate the germ cell-specific nature of the GFP signal, supporting the Armc2 KO rescue experiments.

As mentioned previously, we assessed the stability of mRNA-GFP using RT-QPCR in HEK cells and seminiferous tubule cells (see Fig. 5). mRNA-GFP was detected for up to 60 hours in HEK cells and for up to two weeks in seminiferous tubule cells. Together, these results suggest that the long-lasting fluorescence observed in our experiments reflects a combination of transcript stability and local translation within germ cells, as well as the slow protein turnover typical of the spermatogenic lineage.

In Fig. 5, what is the half-life of luciferase? From this reviewer's search of the literature, it appears to be ~2-3 h in mammalian cells. With this said, how do the authors envision detectable protein for up to 20 days from a naked mRNA? The stability of the injected mRNAs should be shown in a mammalian cell line - perhaps this mRNA has an incredibly long half-life, which might help explain these results. However, even the most stable endogenous mRNAs (e.g., globin) are ~24-60 hrs.

We did not directly assess the stability of luciferase mRNA, but we evaluated the persistence of GFP mRNA, which was synthesized and optimized using the same sequence optimization and chemical modification strategy as the luciferase mRNA. In these experiments, mRNA-GFP was detectable in seminiferous tubule cells for up to two weeks after injection. We therefore expect a similar stability profile for the luciferase mRNA. These findings suggest that the prolonged fluorescence or bioluminescence observed in our study likely reflects a combination of factors, including enhanced transcript stability, local translation within germ cells, and the inherently slow protein turnover characteristic of the spermatogenic lineage.

527-8 The Sertoli cell cytoplasm is not just present along the basement membrane as stated, but also projects all the way to the lumina:

we clarified the sentence " Sertoli cells have an oval to elongated nucleus and the cytoplasm presents a complex shape (“tombstone” pattern) along the basement membrane, with long projections that extend toward the lumen."

529-30 This is incorrect, as round spermatids are never "localized between the spermatocytes and elongated spermatids" - if elongated spermatids are present, rounds are not - they are never coincident in the same testis section:

We thank the reviewer for this important comment and for drawing attention to the detailed staging of the seminiferous epithelium. We agree that the spatial organization of germ cells varies depending on the stage of spermatogenesis. While round spermatids (steps 1–8) and elongated spermatids (steps 9–16) are typically associated with distinct stages, transitional stages of the seminiferous epithelium can contain both cell types in close proximity, reflecting the continuous and overlapping nature of spermatid differentiation (Meistrich, 2013, Methods Mol. Biol. 927:299–307). We have revised the text to clarify this point, indicating that the relative positioning of germ cell types depends on the stage of the seminiferous cycle rather than implying their constant coexistence within the same tubule section.

Fig. 7. To this reviewer, all of the GFP appears to be in Sertoli cell cytoplasm In Figs 1-8 there is no convincing evidence presented that GFP is expressed in germ cells! In fact, it appears to be in Sertoli cells.

We thank the reviewer for their observation. As previously mentioned, we have included an additional experiment specifically demonstrating GFP expression in germ cells (fig 10). This new data provides clear evidence that the GFP signal is not restricted to Sertoli cells and confirms successful uptake and translation of GFP mRNA in germ cells.

Fig. 9 - alpha-tubuline?

We corrected the figure.

Fig. 11 - how was sperm morphology/motility not rescued on "days 3, 6, 10, 15, or 28 after surgery", but it was in some at 21 and 35? How does this make sense, given the known kinetics of male germ cell development??

We note the reviewer’s concern regarding the timing of motile sperm appearance. Variability among treated mice is expected because transfection efficiency differed between spermatogonia and spermatids. Full spermiogenesis requires ~15 days, and epididymal transit adds ~8 days, consistent with motile sperm appearing around 21 days post-injection in some mice.

And at least one of the sperm in the KO in Fig. B5 looks relatively normal, and the flagellum may be out-of-focus in the image? With only a few sperm for reviewers to see, how can we know these represent the population?

We thank the reviewer for their comment. Upon closer examination of the image, the flagellum of the spermatozoon in question is clearly abnormally short and this is not due to being out of focus. Furthermore, the supplementary figure shows that the KO consistently lacks normal spermatozoa. These defects are consistent with previous findings from our laboratory [22], confirming that the observed phenotype is representative of the KO population rather than an isolated occurrence.

Reviewer #3 (Public review):

Summary:

The authors used a novel technique to treat male infertility. In a proof-of-concept study, the authors were able to rescue the phenotype of a knockout mouse model with immotile sperm using this technique. This could also be a promising treatment option for infertile men.

Strengths:

In their proof-of-concept study, the authors were able to show that the novel technique rescues the infertility phenotype of Armc2 knockout spermatozoa. In the current version of the manuscript, the authors have added data on in vitro fertilisation experiments with Armc2 mRNA-rescued sperm. The authors show that Armc2 mRNA-rescued sperm can successfully fertilise oocytes that develop to the blastocyst stage. This adds another level of reliability to the data.

Weaknesses:

Some minor weaknesses identified in my previous report have already been fixed. The technique is new and may not yet be fully established for all issues. Nevertheless, the data presented in this manuscript opens the way for several approaches to immotile spermatozoa to ensure successful fertilisation of oocytes and subsequent appropriate embryo development.

[Editors' note: The images in Figure 12 do not support the authors' interpretation that 2-cell embryos resulted from in vitro fertilization. Instead, the cells shown appear to be fragmented, unfertilized eggs. Combined with the lack of further development, it seems highly unlikely that fertilization was successful.]

We thank the reviewer for their careful evaluation and constructive feedback. We appreciate the acknowledgment of the strengths of our study, particularly the proof-of-concept demonstration that Armc2-mRNA electroporation can rescue sperm motility in Armc2 knockout mice.

Regarding the concern raised by the editor about Figure 12, we would like to clarify two technical points. First, the IVF experiments were performed using CD1 oocytes and B6D2 sperm. Due to strain-specific incompatibilities, fertilization of CD1 oocytes by B6D2 sperm typically does not progress beyond the two-cell stage (Fernández-González [23] et al., 2008, Biology of Reproduction). Therefore, the observation of two-cell embryos represents the expected limit of development in this cross and serves as a strong indication of successful fertilization, even though further development is not possible. Second, the oocytes used in these experiments were treated with collagenase to remove cumulus cells. This enzymatic treatment can sometimes affect the morphology of early embryos, which may explain why the two-cell embryos in Figure 12 appear less regular or somewhat fragmented. We also included a control showing embryos from B6D2 sperm with the same collagenase treatment on CD1 oocytes, which yielded similar appearances (Fig14 A4).

To provide additional functional evidence, we complemented the IVF experiments with ICSI using rescued Armc2^–/– sperm and B6D2 oocytes, which allowed embryos to develop to the blastocyst stage. In these experiments, 25% of injected oocytes reached the blastocyst stage with rescued sperm compared to 13% for untreated Armc2–/– sperm (Supplementary Fig. 9) These results support the functional competence of rescued sperm and demonstrate partial recovery of fertilization ability following Armc2 mRNA electroporation.

We have clarified these points in the revised Results and Discussion sections to emphasize that the IVF data indicate partial functional recovery of rescued sperm rather than full fertility restoration. These clarifications address the editor’s concern while accurately representing the technical limitations of the strain combination used in our experiments.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Fig 12 and Supplementary Fig 9 are mislabeled in the text and rebuttal.

We thank the reviewer for pointing this out. We have carefully checked the manuscript and the rebuttal text, and corrected all references to Figure 12 and Supplementary Figure 9 to ensure they are accurately labeled and consistent throughout the text.

Reviewer #3 (Recommendations for the authors):

The contribution of the newly added authors should be clarified. All other aspects of inadequacy raised in my previous report have been adequately addressed.

No further comments.

We thank the reviewer for noting this. The contributions of the newly added authors have been clarified in the Author Contributions section of the revised manuscript. All other points raised in the previous review have been addressed as indicated.

References

(1) Hansen, M., et al., Assisted reproductive technologies and the risk of birth defects--a systematic review. Hum Reprod, 2005. 20(2): p. 328-38.

(2) Halliday, J.L., et al., Increased risk of blastogenesis birth defects, arising in the first 4 weeks of pregnancy, after assisted reproductive technologies. Hum Reprod, 2010. 25(1): p. 59-65.

(3) Davies, M.J., et al., Reproductive technologies and the risk of birth defects. N Engl J Med, 2012. 366(19): p. 1803-13.

(4) Kurinczuk, J.J., M. Hansen, and C. Bower, The risk of birth defects in children born after assisted reproductive technologies. Curr Opin Obstet Gynecol, 2004. 16(3): p. 201-9.

(5) Graham, M.E., et al., Assisted reproductive technology: Short- and long-term outcomes. Dev Med Child Neurol, 2023. 65(1): p. 38-49.

(6) Palermo, G.D., et al., Intracytoplasmic sperm injection: state of the art in humans. Reproduction, 2017. 154(6): p. F93-f110.

(7) Usmani, A., et al., A non-surgical approach for male germ cell mediated gene transmission through transgenesis. Sci Rep, 2013. 3: p. 3430.

(8) Raina, A., et al., Testis mediated gene transfer: in vitro transfection in goat testis by electroporation. Gene, 2015. 554(1): p. 96-100.

(9) Michaelis, M., A. Sobczak, and J.M. Weitzel, In vivo microinjection and electroporation of mouse testis. J Vis Exp, 2014(90).

(10) Wang, L., et al., Testis electroporation coupled with autophagy inhibitor to treat non-obstructive azoospermia. Mol Ther Nucleic Acids, 2022. 30: p. 451-464.

(11) Wu, Q., et al., Translation affects mRNA stability in a codon-dependent manner in human cells. eLife, 2019. 8: p. e45396.

(12) Gallie, D.R., The cap and poly(A) tail function synergistically to regulate mRNA translational efficiency. Genes & Development, 1991. 5(11): p. 2108-2116.

(13) Henderson, J.M., et al., Cap 1 messenger RNA synthesis with co-transcriptional CleanCap® analog improves protein expression in mammalian cells. Nucleic Acids Research, 2021. 49(8): p. e42.

(14) Stepinski, J., et al., Synthesis and properties of mRNAs containing novel “anti-reverse” cap analogs. RNA, 2001. 7(10): p. 1486-1495.

(15) Sachs, A.B., P. Sarnow, and M.W. Hentze, Starting at the beginning, middle, and end: translation initiation in eukaryotes. Cell, 1997. 89(6): p. 831-838.

(16) Presnyak, V., et al., Codon optimality is a major determinant of mRNA stability. Cell, 2015. 160(6): p. 1111-24.

(17) Cao, D., et al., Unlock the sustained therapeutic efficacy of mRNA. J Control Release, 2025. 383: p. 113837.

(18) Karikó, K., et al., Incorporation of pseudouridine into mRNA yields superior nonimmunogenic vector with increased translational capacity and biological stability. Mol Ther, 2008. 16(11): p. 1833-40.

(19) Pardi, N., et al., mRNA vaccines — a new era in vaccinology. Nature Reviews Drug Discovery, 2018. 17(4): p. 261-279.

(20) Meistrich, M.L. and R.A. Hess, Assessment of Spermatogenesis Through Staging of Seminiferous Tubules, in Spermatogenesis: Methods and Protocols, D.T. Carrell and K.I. Aston, Editors. 2013, Humana Press: Totowa, NJ. p. 299-307.

(21) Au - Mäkelä, J.-A., et al., JoVE, 2020(164): p. e61800.

(22) Coutton, C., et al., Bi-allelic Mutations in ARMC2 Lead to Severe Astheno-Teratozoospermia Due to Sperm Flagellum Malformations in Humans and Mice. Am J Hum Genet, 2019. 104(2): p. 331-340.

(23) Fernández-Gonzalez, R., et al., Long-term effects of mouse intracytoplasmic sperm injection with DNA-fragmented sperm on health and behavior of adult offspring. Biol Reprod, 2008. 78(4): p. 761-72.

eLife Assessment

This is an important study that advances our understanding of the transition from quadrupedal to bipedal gait in a neuromechanical model of the Japanese macaque. The method and results are solid; the neuromusculoskeletal model successfully reproduces experimental data, and the stability analysis based on an inverted pendulum model effectively explains the effects of different transition strategies. However, the study would benefit from a more comprehensive sensitivity analysis. The findings are highly relevant for researchers in motor control, comparative physiology/biomechanics, and robotics.

Reviewer #1 (Public review):

Summary:

The article investigates how the Japanese macaque makes gait transitions between quadruped and biped gaits. It presents a compelling neuromechanical simulation that replicates the transition and an interesting analysis based on an inverted pendulum that can explain why some transition strategies are successful and others are not.

Strengths:

I enjoyed reading this article. I think it presents an interesting study and elegant modeling approaches (musculoskeletal + inverted pendulum). The study is well conducted, and the results are interesting. I particularly liked how the success of gait transitions could be predicted based on the inverted pendulum and its saddle node stability. I think it makes a useful and interesting contribution to the state of the art.

Weaknesses:

The article is already in great shape, but could be improved a bit by:

(1) Strengthening the comparison to animal data. In particular, videos of the real animal should be included + snapshots of their gaits (quadruped, biped, and transitions).

(2) Exploring and testing a broader range of conditions. I think it would be very interesting to test gaits and gait transitions on up and down slopes (both with the musculoskeletal model and with the inverted pendulum model). This could be used to make predictions on how the real animal adapts to those conditions. Ideally, this should be tested on the animal as well. I think this could increase (even more) the impact of this work.

(3) Better explaining several aspects of the PSO optimization.

(4) (Ideally) performing a sensitivity analysis on the optimized parameters (e.g. variations of +-5, 10, 20%) in order to determine their respective importance and how much their instantiated values have influenced the results.

(5) Running a spell checker, as there are quite a few typos.

Reviewer #2 (Public review):

Summary:

This article presents a neuromusculoskeletal (NMS) model of the Japanese Macaque. This model is added with a neural feedforward controller based on CPG and synergy that allows for reproducing quadrupedal and bipedal gait as well as the transition between quadrupedal and bipedal gait. The model and controller were validated using experimental data. Results were also compared to an inverted pendulum model to show that the transition between quadrupedal and bipedal in macaque is using this kind of representation for transition and stability. Overall, the article is very interesting, but it sometimes lacks clarity.

Strengths:

The results of the model present impressive results for quadrupedal, bipedal, and transition, validated by experimental data. NMS controllers based on feedforward controllers are very difficult to fine-tune.

Weaknesses:

(1) The movement regulator is not clear and should be better explained. At first, it seems that it is just a new CPG/synergy (feedforward) added, but in the methods, it seems to be a feedback controller.

(2) It is also not clear what is meant by discretizing the weight for the trigger limb from 0 to 1 (page 8).

(3) The controller is mainly using a feedforward controller, allowing only anticipatory movement. Animals are also using a reflex-based feedback controller. A controller with feedback/reflex could reduce failed attempts in training and better represent the transition.

(4) There are small typos throughout the article that should be corrected.

Reviewer #3 (Public review):

Summary:

The purpose of this study was to test the hypothesis that the inverted pendulum mechanism contributes to the gait transition from quadrupedal to bipedal gait in Japanese macaques. The author uses a neuromusculoskeletal model to generate different motor tasks by varying motor command parameters during forward dynamics simulations. After simulations were done, the authors used dynamical system analysis of the inverted pendulum model to reveal the underlying principles of gait transition control. The authors showed that successful gait transition from quadrupedal to bipedal gait mostly depends on increased step length of a hindlimb.

Strengths:

This study is important not only for understanding gait transition, but also to understand stability control of bipedal gaits. Another advantage of this study is that it allows us to estimate the effect of one control mechanism and find its effect and limits. In animal studies, we also have a combination of compensatory stability control mechanisms.

Weaknesses:

Any simulation is not perfect, so discrepancies from experimental data are expected. A 2D model is used, but the advantage of using a 3D model is not clear, and it is much more complicated.

Author response:

eLife Assessment

This study provides a valuable contribution to understanding how negative affect influences food-choice decision making in bulimia nervosa, using a mechanistic approach with a drift diffusion model (DDM) to examine the weighting of tastiness and healthiness attributes. The solid evidence is supported by a robust crossover design and rigorous statistical methods, although concerns about the interpretation of group differences across neutral and negative conditions limit the interpretability of the results.

We are grateful for this improved assessment. Below, we provide detailed responses that we believe address the noted concerns about interpreting group differences across conditions. If these clarifications resolve the interpretability concerns, we would be grateful if the editors would consider updating the eLife assessment accordingly.

Public Reviews:

Reviewer #1 (Public review):

Summary:

Using a computational modeling approach based on the Drift and Diffusion Model (DDM) introduced by Ratcliff and McKoon in 2008, the article by Shevlin and colleagues investigates whether there are differences between neutral and negative emotional states in:

(1) The timings of the integration in food choices of the perceived healthiness and tastiness of food options in individuals with bulimia nervosa (BN) and healthy participants

(2)The weighting of the perceived healthiness and tastiness of these options.

Strengths:

By looking at the mechanistic part of the decision process, the approach has potential to improve the understanding of pathological food choices.

Weaknesses:

I thank the author for reviewing their manuscript.

However, I still have major concerns.

The authors say that they removed any causal claims in their revised version of the manuscript. The sentence before the last one of the abstract still says "bias for high-fat foods predicted more frequent subjective binge episodes over three months". This is a causal claim that I already highlighted in my previous review, specifically for that sentence (see my second sentence of my major point 2 of my previous review).

We appreciate the Reviewer's continued attention to causal language. We acknowledge that our use of the term 'predicted', though intended to refer to statistical prediction in a regression model, could be misinterpreted as implying causation. We have therefore revised this sentence to read: 'bias for high-fat foods was associated with more frequent subjective binge episodes over three months’.

I also noticed that a comment that I added was not sent to the authors. In this comment I was highlighting that in Figure 2 of Galibri et al., I was uncertain about a difference between neutral and negative inductions of the average negative rating after the induction in the BN group (i.e. comparing the negative rating after negative induction in BN to the negative rating after neutral induction in BN). Figure 2 of Galibri et al. looks to me that:

(1) The BN participants were more negative before the induction when they came to the neutral session than when they came to the negative session.

(2) The BN participants looked almost negatively similar (taking into account the error bars reported) after the induction in both sessions

These observations are of high importance because they may support the fact that BN patients were likely in a similar negative state to run the food decision task in both conditions (negative and neutral). Therefore, the lack of difference in food choices in BN patients is unsurprising and nothing could be concluded from the DDM analyses. Moreover, the strong negative ratings of BN patients in the neutral condition as compared to healthy participants together with almost similar negative ratings after the two inductions contradict the authors' last sentence of their abstract.

I appreciate that the authors reproduced an analysis of their initial paper regarding the negative ratings (i.e. Table S1). It partly answers my aforementioned point but does not address the fact that BN may have been in a similar negative state in both conditions (neutral and negative) when running the food decision task: if BN patients were similarly negative after both induction (neutral and negative), nothing can be concluded from their differences in their results obtained from the DDM. As the authors put it, "not all loss-ofcontrol eating occurs in the context of negative state", I add that far from all negative states lead to a loss-of-control eating in BN patients. This grounds all my aforementioned remarks and my remarks of my first review.

A solution for that is to run a paired t-test in BN patients only comparing the score after the induction in the two conditions (neutral and negative) reported in Figure 2 of their initial article.

We appreciate the reviewer’s concern. We understand how the visual representation in Figure 2, which displays between-subject error bars, might suggest similar post-induction affect levels. However, the within-subject paired comparison (which appropriately accounts for individual differences in baseline affect) reveals a significant difference, which we detail below.

While BN participants did report higher baseline negative affect than the HC group prior to the mood inductions, this does not negate the effectiveness of the manipulation. The critical comparison is the within-subject change from pre- to post-induction (detailed below) which shows that negative affect was significantly higher after the negative induction than the neutral induction.

As we reported in the Supplementary Information (Table S1), our initial analyses of self-reported affect ratings used a linear mixed-effects model with group (HC = 0, BN = 1), condition (Neutral = 0, Negative = 1), and time (pre-induction = 0, post-induction = 1) as fixed effects, including all interactions, and random intercepts for participants. This approach accounts for individual differences in baseline affect.

However, to address the reviewer's concerns, we conducted two simple effects analyses using estimated marginal means. As the reviewer suggested, we directly compared post-induction affect between conditions within the BN group (described in the second analysis below). In the first analysis, we examined the diagnosis × time interaction within each condition separately. In the Negative condition, individuals with BN demonstrated a substantial increase in negative affect from pre- to post-induction (mean difference = 20.36, t = 4.84, p < 0.0001, Cohen’s d = 0.97). In the second analysis, we examined the condition × time interaction within each group separately. Among the BN group, we found that reported affect was significantly higher following the negative mood induction than after the neutral affect induction (mean difference = -17.40, t = -4.13, p = 0.0003, Cohen’s d = 0.83). This difference in post-induction negative affect between conditions within the BN group represents a meaningful and statistically robust difference in affective states. These within-group effects confirm that the negative mood induction was (1) effective in the BN group and (2) produced significantly greater negative affect than the neutral mood induction.

These findings confirm that participants completed the food decision task under meaningfully different affective states, supporting the interpretability of the subsequent DDM analyses. We now report these analyses in the Supplementary Information.

I appreciate the analysis that the authors added with the restrictive subscale of the EDE-Q.

That this analysis does not show any association with the parameters of interest does not show that there is a difference in the link between self reported restrictions and self reported binges. Only such a difference would allow us to claim that the results the authors report may be related to binges.

We thank the reviewer for raising this important point about specificity. To address this concern, we examined the correlation between self-reported binge frequency (both subjective binge episodes and objective binge episodes over the past three months) and EDE-Q Restraint subscale in our BN sample.

The correlation between these measures were modest and non-significant (subjective binge frequency: Spearman’s p = 0.21, p = 0.306; objective binge frequency: Spearman’s p = 0.05, p = 0.806), indicating that both binge frequency measures and dietary restraint were relatively independent dimensions of eating pathology in our sample. This dissociation supports the specificity of our findings: the fact that our DDM parameters were associated with binge frequency but not with dietary restraint suggests that the affect-induced changes in decisionmaking we observed are specifically related to binge-eating behavior rather than reflecting a correlate of dietary restraint. We now report this analysis in the Supplementary Information.

I appreciate the wording of the answer of the authors to my third point: "the results suggest that individuals whose task behavior is more reactive to negative affect tend to be the most symptomatic, but the results do not allow us to determine whether this reactivity causes the symptoms". This sentence is crystal clear and sums very well the limits of the associations the authors report with binge eating frequency. However, I do not see this sentence in the manuscript. I think the manuscript would benefit substantially from adding it.

We thank the reviewer for the suggestion. We have added the following sentences that convey this information to the end of the third paragraph of the discussion:

“These results suggest that individuals whose task behavior is more reactive to negative affect tend to be the most symptomatic. However, our correlational design does not allow us to determine whether this reactivity causes the symptoms.”

Statistical analyses:

If I understood well the mixed models performed, analyses of supplementary tables S1 and S27 to S32 are considering all measures as independent which means that the considered score of each condition (neutral vs negative) and each time (before vs after induction) which have been rated by the same participants are independent. Such type of analyses does not take into account the potential correlation between the 4 scores of a given participant. As a consequence, results may lead to false positives that a linear mixed model does not address. The appropriate analysis would be to run adapted statistical tests pairing the data without running any mixed model.

We appreciate the reviewer's attention to the statistical approach. However, we respectfully note that mixed-effects models do account for within-subject correlations, contrary to the reviewer’s interpretation.

The linear mixed-effects model we employed explicitly accounts for the correlation among repeated measures from the same participant through the random intercept term. This random effect structure models the non-independence of observations within participants, allowing for correlated errors within individuals while assuming independence between individuals. This is a standard and appropriate approach for analyzing repeated-measures data (Bates et al., 2015).

The mixed-effects model is, in fact, more appropriate than separate paired t-tests for our design because it:

(1) Simultaneously models all fixed effects (group, condition, time) and their interactions in a single unified framework;

(2) Properly partitions variance into within-subject and between-subject components;

(3) Provides greater statistical power and more precise estimates by using all available data simultaneously; and

(4) Allows for direct testing of three-way interactions that cannot be assessed through pairwise comparisons alone.

Paired tests (e.g., t-tests), as the reviewer suggests, would require multiple separate analyses and would not allow us to test our primary hypotheses about group × condition × time interactions. The mixed-effects approach provides a more comprehensive and statistically rigorous analysis of our repeated-measures design. To clarify this even further in the manuscript, we have added the following in our methods when describing our model, “participant-level random intercepts were included to account for within-subject correlations across repeated measurements.”

Notes:

It is not because specific methods like correlating self reported measures over long periods with almost instantaneous behaviors (like tasks) have been used extensively in studies that these methods are adapted to answer a given scientific question. Measures aggregated over long periods miss the variations in instantaneous behaviors over these periods.

We acknowledge the reviewer’s concern about the temporal mismatch between our session-level task measures and the 3-month aggregated symptom reports. This is a valid limitation of crosssectional designs, and we agree that examining how task performance fluctuates in relation to real-time symptom variation would provide richer insights into the potential dynamics of these relationships.

We agree that we cannot capture how daily changes in task performance relate to momentary symptom occurrence. In response to previous rounds of helpful reviews, we added this limitation to the Discussion section, noting that future research employing ecological momentary assessment (EMA) or daily diary methods could examine whether the decision-making processes we identified also fluctuate in relation to real-time symptom occurrence.

We note that our finding that affect-induced changes in decision-making parameters were associated with subjective binge frequency suggests that this laboratory-measured reactivity may reflect a stable individual difference that manifests across contexts and time periods. While our current study provides initial evidence that individual differences in affect-related decisionmaking are associated with symptom severity, we acknowledge that longitudinal designs with repeated assessments would strengthen causal and temporal inferences.

Reviewer #2 (Public review):

Summary:

Binge eating is often preceded by heightened negative affect, but the specific processes underlying this link are not well-understood. The purpose of this manuscript was to examine whether affect state (neutral or negative mood) impacts food choice decisionmaking processes that may increase the likelihood of binge eating in individuals with bulimia nervosa (BN). The researchers used a randomized crossover design in women with BN (n=25) and controls (n=21), in which participants underwent a negative or neutral mood induction prior to completing a food-choice task. The researchers found that despite no differences in food choices in the negative and neutral conditions, women with BN demonstrated a stronger bias toward considering the 'tastiness' before the 'healthiness' of the food after the negative mood induction.

Strengths:

The topic is important and clinically relevant, and the methods are sound. The use of computational modeling to understand nuances in decision-making processes and how that might relate to eating disorder symptom severity is a strength of the study.

Weaknesses:

Sample size was relatively small, and participants were all women with BN, which limits generalizability of findings to the larger population of individuals who engage in binge eating. It is likely that the negative affect manipulation was weak and may not have been potent enough to change behavior. These limitations are adequately noted in the discussion.

We are grateful to Reviewer #2 for their careful and supportive review of our manuscript. We appreciate their recognition that computational modeling can reveal nuanced alterations in decision-making processes that may not be apparent in overt behavioral choices. Their balanced assessment of both the strengths and limitations of our work has been helpful in contextualizing our findings appropriately. We have carefully considered their comments regarding sample size and the potential limitations of our mood induction procedure, both of which we discuss in detail in the manuscript's limitations section.

Reviewer #3 (Public review):

Summary:

The study uses the food choice task, a well-established method in eating disorder research, particularly in anorexia nervosa. However, it introduces a novel analytical approach-the diffusion decision model-to deconstruct food choices and assess the influence of negative affect on how and when tastiness and healthiness are considered in decision-making among individuals with bulimia nervosa and healthy controls.

Strengths:

The introduction provides a comprehensive review of the literature, and the study design appears robust. It incorporates separate sessions for neutral and negative affect conditions and counterbalances tastiness and healthiness ratings. The statistical methods are rigorous, employing multiple testing corrections.

A key finding-that negative affect induction biases individuals with bulimia nervosa toward prioritizing tastiness over healthiness-offers an intriguing perspective on how negative affect may drive binge eating behaviors.

Weaknesses:

A notable limitation is the absence of a sample size calculation, which, combined with the relatively small sample, may have contributed to null findings. Additionally, while the affect induction method is validated, it is less effective than alternatives such as image or film-based stimuli (Dana et al., 2020), potentially influencing the results.

We are grateful to Reviewer #3 for their thoughtful evaluation of our work. We appreciate their recognition that the diffusion decision model provides a novel analytical lens for understanding how negative affect influences the dynamics of food-related decision-making in bulimia nervosa. Their balanced assessment of both the methodological strengths of our design (counterbalancing, rigorous statistical corrections) and its limitations (sample size, mood induction efficacy) has been valuable in ensuring we appropriately contextualize our findings and their implications. Specifically, we have taken their comments regarding sample size and the relative efficacy of different mood induction methods seriously, and we address these important methodological considerations in our discussion of the study's limitations.

Recommendations for the authors:

Reviewer #2 (Recommendations for the authors):

The authors have addressed my previous comments, and I do not have any additional suggestions for improvement.

We thank the reviewer for their time, effort, and insightful feedback.

Reviewer #3 (Recommendations for the authors):

The authors have adequately addressed my feedback. I have no further comments.

We thank the reviewer for their time, effort, and insightful feedback.

eLife Assessment

This study makes a valuable contribution to understanding how negative affect shapes food-choice decision making in bulimia nervosa by leveraging a mechanistic drift diffusion model to quantify the weighting of tastiness and healthiness attributes. The evidence is solid, supported by a randomized crossover design and generally appropriate statistical analyses. However, the interpretability of the findings is limited by ambiguities in the affect manipulation, particularly regarding whether neutral and negative inductions yielded reliably distinct affective states at the time of task performance in the bulimia nervosa group. Consequently, session-related differences in model parameters cannot be unequivocally attributed to negative affect rather than to uncontrolled state or contextual factors, and clearer separation of affective conditions alongside analyses aligned with the paired data structure would strengthen the conclusions.

Reviewer #1 (Public review):

Summary:

Using a computational modeling approach based on the Drift and Diffusion Model (DDM) introduced by Ratcliff and McKoon in 2008, the article by Shevlin and colleagues investigates whether there are differences between neutral and negative emotional states in:

(1) The timings of the integration in food choices of the perceived healthiness and tastiness of food options in individuals with bulimia nervosa (BN) and healthy participants (2) The weighting of the perceived healthiness and tastiness of these options.

Strengths:

By looking at the mechanistic part of the decision process, the approach has potential to improve the understanding of pathological food choices.

Weaknesses:

I thank the authors for revising their manuscript.

I still notice that the authors did not go through their manuscript to look for wordings refering to a prediction interpretation of their results while I already highlighted the inappropriateness of this wording in my two first rounds of reviews: e.g. there is still "we used zero-inflated negative binomial models to predict the three-month frequency" and I can find other statements like this. The design of their study does not allow such claims.

The authors answered my major concern regarding the experimental induction towards a negative or a neutral state before running the food decision task. My concern is: BN patients already seemed to be already in a high negative state before undergoing the neutral induction, while these patients are in a lower negative state before undergoing the negative induction. It is therefore not surprising that patients seem to report a similar level of negative state after the two inductions (according to the figure of the authors' previous article). Of note is that the additional analysis the authors ran within the BN group only provides a significant result: this result shows that there has been an induction but does not rule out that patients were in the exact same magnitude of negative state to perform the task as the figure in their previously published article suggests it. The major issue is to show that:

(1) As compared to the neutral induction, there has been a higher variation in negative state after as compared to before the negative induction.

(2) The magnitude of the negative state after the negative induction is higher than the magnitude of the negative state after the neutral induction.

The first point shows that the induction worked. The second point shows that the participants are in two distinct states. Without showing the second point, it may be possible that one induction increases the negative state of participants to the same level as the one of the second induction that has not increased anything.

Within this context, how is it possible to associate, in patients, a difference in the DDM between the two sessions to a negative state (which is one of the main focus of the article) rather than to another parameter that has not been captured? A similar situation would be in an experiment studying the consequence of stress, a stressfull induction over relaxed participants attending the lab has high chances to raise the level of stress of those participants to the same level as the one that the same participants would experience after a neutral induction when these participants attend the lab with an already high level of stress. In that case, would it be approrpiate to claim that a difference at a task performed after the induction would be related to stress while the participants would be at the same level of stress when performing the task despite the fact that the induction worked ?

In the experiment performed by the authors, the additional analysis to perform would be a paired sample t-test (or the appropriate non-parametric test) to check whether the magnitude of negative state of BN patients was different between the negative and neutral conditions after the induction only. If not, associating the difference at the DDM with negative states in BN is highly misleading.

I read carefully the authors' answer related to mixed models: they claim that mixed models take into account correlations within their repeated data. The specification of the structure of the covariance matrix allows to control only partly for that. I notice that the authors did not specify the structure of that matrix: the article they refer to to justify the appropriatness of their analyses is not adapted. The specification of the structure of the covariance matrix needs to address, in a mixed model, the difference in handling 4 repeated data per participants that cannot be paired as compared to 4 repeated data that can be paired (two per session with one before and one after the neutral or negative priming sessions, if I count right). Of note is that a covariance structure that is left free of constraint for the fit of the model does not capture appropriately the pairing of the data: it has all chances to capture the covariance in a different way. And a covariance structure that has constraints has more chances to lead to a model that cannot be estimated because of an absence of convergence of the algorithms.

By the way, a single two-sample t-test (or a Mann-Whitney test if appropriate), and not a set of multiple paired-sample t-test as the authors suggest, would answer the goal of the authors to test for what they call the three-way interaction in their comment. This test would be performed between the two groups of participants (BN/controls) with the computation for each participant separately: (assessment after neutral induction-assessment before neutral induction)-(assessment after negative induction-assessment before negative induction). This analysis answers points 1, 2 and 4 they raise together with my point of controlling for the paired data. I would have agreed with their choice of a mixed model if they had an unbalanced dataset within each participant.

Reviewer #2 (Public review):

Summary:

Binge eating is often preceded by heightened negative affect, but the specific processes underlying this link are not well-understood. The purpose of this manuscript was to examine whether affect state (neutral or negative mood) impacts food choice decision-making processes that may increase likelihood of binge eating in individuals with bulimia nervosa (BN). The researchers used a randomized crossover design in women with BN (n=25) and controls (n=21), in which participants underwent a negative or neutral mood induction prior to completing a food-choice task. The researchers found that despite no differences in food choices in the negative and neutral conditions, women with BN demonstrated a stronger bias toward considering the 'tastiness' before the 'healthiness' of the food after the negative mood induction.

Strengths:

The topic is important and clinically relevant and methods are sound. The use of computational modeling to understand nuances in decision-making processes and how that might relate to eating disorder symptom severity is a strength of the study.

Weaknesses:

Sample size was relatively small, and participants were all women with BN, which limits generalizability of findings to the larger population of individuals who engage in binge eating. It is likely that the negative affect manipulation was weak and may not have been potent enough to change behavior. These limitations are adequately noted in the discussion.

eLife Assessment

Hoverflies are known for their sexually dimorphic visual systems and exquisite flight behaviors. This valuable study reports how two types of visual descending neurons differ between males and females in their motion- and speed-dependent responses, yet surprisingly, the behavior they control lacks any sexual dimorphism. The results convincingly support these findings, which will be of interest for studies of visuomotor transformations and network-level brain organization.

Reviewer #1 (Public review):

Summary:

Hoverflies are known for a striking sexual dimorphism in eye morphology and early visual system physiology. Surprisingly, the male and female flight behaviors show only subtle differences. Nicholas et al. investigate the sensori-motor transformation of sexually dimorphic visual information to flight steering commands via descending neurons. The authors combined intra- and extracellular recordings, neuroanatomy, and behavioral analysis. They convincingly demonstrate that descending neurons show sexual dimorphisms - in particular at high optic flow velocities - while wing steering responses seem relatively monomorphic. The study highlights a very interesting discrepancy between neuronal and behavioral response properties.

More specifically, the authors focused on two types of descending neurons that receive inputs from well-characterized wide-field sensitive tangential cells: OFS DN1, which receives inputs from so-called HS cells, and OFS DN2, which receives input from a set of VS cells. Their likely counterparts in Drosophila connect to the neck, wing, and haltere neuropils. The authors characterized the visual response properties of these two neuronal classes in both male and female hoverflies and identified several interesting differences. They then presented the same set of stimuli, tracked wing beat amplitude, and analyzed the sum and the difference of right and left wing beat amplitude as a readout of lift or thrust, and yaw turning, respectively. Behavioral responses showed little to no sexual dimorphism, despite the observed neuronal differences.

Strengths:

I find the question very interesting and the results both convincing and intriguing. A fundamental goal in neuroscience is to link neuronal responses and behavior. The current study highlights that the transformations - even at the level of descending neurons to motoneurons - are complex and less straightforward than one might expect.

Weaknesses:

The authors investigated two types of descending neurons, but it was not clear to me how many other descending neurons are thought to be involved in wing steering responses to wide-field motion. I would suggest providing a more in-depth overview of what is known about hoverflies and Drosophila, since the conclusions drawn from the study would be different if these two types were the only descending neurons involved, as opposed to representing a subset of the neurons conveying visual information to the wing neuropil.

Both neuronal classes have counterparts in Drosophila that also innervate neck motor regions. The authors filled the hoverfly DNs in intracellular recordings to characterize their arborization in the ventral nerve cord. In my opinion, these anatomical data could be further exploited and discussed a bit more: is the innervation in hoverflies also consistent with connecting to the neck and haltere motor regions? Are there any obvious differences and similarities to the Drosophila neurons mentioned by the authors? If the arborization also supports a role in neck movements, the authors could discuss whether they would expect any sexual dimorphism in head movements.

Reviewer #2 (Public review):

Summary:

Many fly species exhibit male-specific visual behaviors during courtship, while little is known about the circuit underlying the dimorphic visuomotor transformations. Nicholas et al focus on two types of visual descending neurons (DNs) in hoverflies, a species in which only males exhibit high-speed pursuit of conspecifics. They combined electrophysiology and behavior analysis to identify these DNs and characterize their response to a variety of visual stimuli in both male and female flies. The results show that the neurons in both sexes have similar receptive fields but exhibit speed-dependent dimorphic responses to different optic flow stimuli.

Strengths:

Hoverflies, though not a common model system, show very interesting dimorphic behaviors and provide a unique and valuable entry point to explore the brain organization behind sexual dimorphism. The findings here are not only interesting on their own right but will also likely inspire those working in other systems, particularly Drosophila.

The authors employed rigorous morphology, electrophysiology, and behavior methods to deliver a comprehensive characterization of the neurons in question. The precision of the measurements allowed for identifying a subtle and nuanced neuronal dimorphism and set a standard for future work in this area.

Weaknesses:

Cell-typing using receptive field preferred directions (RFPDs): if I understood correctly, this classification method mostly relies on the LPDs near the center of the receptive field (median within the contour in Fig.1). I have two concerns here. First, this method is great if we are certain there are only two types of visual DNs as described in the manuscript. But how certain is this? Given the importance of vision in flight control, I would expect many DNs that transmit optic flow information to the motor center. I'd also like to point out that there are other lobula plate tangential cells (LPTCs) than HS and VS cells, which are much less studied and could potentially contribute to dimorphic behaviors. Second, this method feels somewhat impoverished given the richness of the data. The authors have nicely mapped out the directional tuning for almost the entire visual field. Instead of reducing this measurement to 2 values (center and direction), I was wondering if there is a better method to fully utilize the data at hand to get a better characterization of these DNs. As the authors are aware, local features alone can be ambiguous in characterizing optic flows. What's more, taking into account more global features can be useful for discovering potentially new cell types.

Line 131, it wasn't clear to me why full-screen stimuli were used for comparison here, instead of the full receptive field maps. Male flies exhibit sexual dimorphic behaviors only during courtship, which would suggest that small-sized visual stimuli (mimicking an intruder or female conspecific) would be better suited to elicit dimorphic neuronal responses. A similar comment applies to the later results as well. Based on the receptive field mapping in Figure 1, I'm under the impression that these 2 DN types are more suited to detect wide-field optic flows, those induced by self-motion as mentioned in the manuscript. The results are still very interesting, but it's good to make this point clear early on to help set appropriate expectations. Conversely, this would also suggest that there are other visual DN types that are responsible for the courtship-related sexually dimorphic behaviors.

Author response:

eLife Assessment

Hoverflies are known for their sexually dimorphic visual systems and exquisite flight behaviors. This valuable study reports how two types of visual descending neurons differ between males and females in their motion- and speed-dependent responses, yet surprisingly, the behavior they control lacks any sexual dimorphism. The results convincingly support these findings, which will be of interest for studies of visuomotor transformations and network-level brain organization.

This statement perfectly recapitulates our findings.

Public Reviews:

Reviewer #1 (Public review):  

Summary: 

Hoverflies are known for a striking sexual dimorphism in eye morphology and early visual system physiology. Surprisingly, the male and female flight behaviors show only subtle differences. Nicholas et al. investigate the sensori-motor transformation of sexually dimorphic visual information to flight steering commands via descending neurons. The authors combined intra- and extracellular recordings, neuroanatomy, and behavioral analysis. They convincingly demonstrate that descending neurons show sexual dimorphisms - in particular at high optic flow velocities - while wing steering responses seem relatively monomorphic. The study highlights a very interesting discrepancy between neuronal and behavioral response properties.

Thank you for this summary. Most of the statement perfectly recapitulates the main findings of our paper. However, we want to emphasize that some hoverfly flight behaviors are strongly sexually dimorphic, especially those related to courtship and mating. Indeed, only male hoverflies pursue targets at high speed, chase away territorial intruders, and pursue females for mating. However, other flight behaviours, such as those related to optomotor responses and flights between flowers when feeding, are not sexually dimorphic. We will amend the Introduction to make the difference between flight behaviors clear.

More specifically, the authors focused on two types of descending neurons that receive inputs from well-characterized wide-field sensitive tangential cells: OFS DN1, which receives inputs from so-called HS cells, and OFS DN2, which receives input from a set of VS cells. Their likely counterparts in Drosophila connect to the neck, wing, and haltere neuropils. The authors characterized the visual response properties of these two neuronal classes in both male and female hoverflies and identified several interesting differences. They then presented the same set of stimuli, tracked wing beat amplitude, and analyzed the sum and the difference of right and left wing beat amplitude as a readout of lift or thrust, and yaw turning, respectively. Behavioral responses showed little to no sexual dimorphism, despite the observed neuronal differences.

Thank you for this very nice summary of our work. We want to clarify that LPTC input to DN1 and DN2 has not been shown directly in hoverflies using e.g. dye coupling, or dual recordings. Instead, the presumed HS and VS input is inferred from morphological and physiological DN evidence, and comparisons to similar data in Drosophila and blowflies. We will amend the Introduction to clarify this. The rest of the paragraph perfectly recapitulates the main findings of our paper.

Strengths:

I find the question very interesting and the results both convincing and intriguing. A fundamental goal in neuroscience is to link neuronal responses and behavior. The current study highlights that the transformations - even at the level of descending neurons to motoneurons - are complex and less straightforward than one might expect.

Thank you.

Weaknesses:

The authors investigated two types of descending neurons, but it was not clear to me how many other descending neurons are thought to be involved in wing steering responses to wide-field motion. I would suggest providing a more in-depth overview of what is known about hoverflies and Drosophila, since the conclusions drawn from the study would be different if these two types were the only descending neurons involved, as opposed to representing a subset of the neurons conveying visual information to the wing neuropil.

This is a great point. There are around 1000 fly DNs, of which many could respond to widefield motion, without being specifically tuned to widefield motion. For example, many looming sensitive neurons also respond to widefield motion, and could therefore be involved in the WBA movements that we measured here. In addition, there are many multimodal neurons that could be involved in optomotor responses in free flight, but these may not have been stimulated when we only provided visual input. Furthermore, many visual neurons are modulated by proprioceptive feedback, which is lacking in immobilized physiology preps. Finally, in blowflies, up to 5 optic flow sensitive DNs have been identified morphologically, and in Drosophila 3 have been identified morphologically and physiologically. In summary, it is more than likely that other neurons project visual widefield motion information to the wing neuropil. We will amend our Introduction and Discussion to make this important point clear to the readers.

Both neuronal classes have counterparts in Drosophila that also innervate neck motor regions. The authors filled the hoverfly DNs in intracellular recordings to characterize their arborization in the ventral nerve cord. In my opinion, these anatomical data could be further exploited and discussed a bit more: is the innervation in hoverflies also consistent with connecting to the neck and haltere motor regions? Are there any obvious differences and similarities to the Drosophila neurons mentioned by the authors? If the arborization also supports a role in neck movements, the authors could discuss whether they would expect any sexual dimorphism in head movements.

These are all great points. We did not see any clear arborizations to the frontal nerve, where we would expect to find the neck motor neurons (NMNs). In addition, while we did see fine arborizations throughout the length of the thoracic ganglion, we saw no strong outputs projecting directly to the haltere nerve (HN). In the revised version of the MS we will modify figure 4 (morphological characterization) to clarify.

There are important differences between the morphology of DN1 and DN2 in hoverflies and DNHS1 and DNOVS2 in Drosophila, in terms of their projections in the thoracic ganglion. For example, In Drosophila DNOVS2, there are several fine branches along the length of the neuron in the thoracic ganglia. Similarly, we found fine branches in Eristalis tenax DN2, however, in addition, we found a wide branch projecting to the area of the thoracic ganglion where the prothoracic and pterothoracic nerves likely get their inputs (Figure 4), suggesting that the neuron could contribute to controlling the wings and/or the forelegs (which is why we quantified the WBA). In Drosophila DNHS1, there is a similar fat branch to the prothoracic and pterothoracic nerves, which we also found in Eristalis tenax OFS DN1 (Figure 4). Indeed, while Drosophila DNHS1 and DNOVS2 have quite strikingly different morphology, DN1 and DN2 in Eristalis looked quite similar. We will modify the Results section to make this clear.

In addition, to investigate this further, in the revised version of the MS we will include analysis of the movement of different body parts (including the head) to investigate the presence of any potential sexual dimorphism. Unfortunately, however, this will not include the halteres, as they cannot be seen well in the videos.

Reviewer #2 (Public review):

Summary:

Many fly species exhibit male-specific visual behaviors during courtship, while little is known about the circuit underlying the dimorphic visuomotor transformations. Nicholas et al focus on two types of visual descending neurons (DNs) in hoverflies, a species in which only males exhibit high-speed pursuit of conspecifics. They combined electrophysiology and behavior analysis to identify these DNs and characterize their response to a variety of visual stimuli in both male and female flies. The results show that the neurons in both sexes have similar receptive fields but exhibit speed-dependent dimorphic responses to different optic flow stimuli.

This statement perfectly recapitulates the main findings of our paper. However, as mentioned above, while hoverfly flight behaviors related to courtship and mating are strongly sexually dimorphic, other flight behaviours, such as those related to optomotor responses and flights between flowers when feeding, are not. We will amend the Introduction to make the difference between flight behaviors clear.

Strengths:

Hoverflies, though not a common model system, show very interesting dimorphic behaviors and provide a unique and valuable entry point to explore the brain organization behind sexual dimorphism. The findings here are not only interesting on their own right but will also likely inspire those working in other systems, particularly Drosophila.

Thank you.

The authors employed rigorous morphology, electrophysiology, and behavior methods to deliver a comprehensive characterization of the neurons in question. The precision of the measurements allowed for identifying a subtle and nuanced neuronal dimorphism and set a standard for future work in this area.

Thank you.

Weaknesses:

Cell-typing using receptive field preferred directions (RFPDs): if I understood correctly, this classification method mostly relies on the LPDs near the center of the receptive field (median within the contour in Fig.1). I have two concerns here. First, this method is great if we are certain there are only two types of visual DNs as described in the manuscript. But how certain is this? Given the importance of vision in flight control, I would expect many DNs that transmit optic flow information to the motor center. I'd also like to point out that there are other lobula plate tangential cells (LPTCs) than HS and VS cells, which are much less studied and could potentially contribute to dimorphic behaviors.

This is very true, and an important point. As mentioned above, in blowflies, up to 5 optic flow sensitive DNs have been identified morphologically, however, if these correspond to 5 different physiological types remain unclear. In both blowflies and Drosophila 3 have been identified morphologically and physiologically (DNHS1, DNOVS1, DNOVS2). Importantly, in both blowflies and fruitflies DNOVS1 gives graded responses, and no action potentials, meaning that we would not be able to record from it using extracellular electrophysiology.

We previously used clustering techniques to show that in Eristalis, we can reliably distinguish two types of optic flow sensitive DNs from extracellular electrophysiological data, based on a range of receptive field parameters, and we think that these correspond to DNHS1 and DNOVS2 in Drosophila (Nicholas et al, J Comp Physiol A, 2020, cited in paper). As mentioned above in response to Reviewer 1, this does not mean that there are no other neurons that could respond to widefield optic flow, and which might be involved in the WBA we recorded in the paper. However, the point of this paper was not to conclusively show that there are only two optic flow sensitive descending neurons. The point was to say that there are two quite distinct optic flow sensitive neurons that have similar receptive fields in males and females, while the responses to widefield motion show differences between males and females.

We will modify the Introduction and Discussion to make these important points clear to the Reader, including the discussion of the 45-60 LPTCs that exist in the lobula plate, and what their role might be.

Second, this method feels somewhat impoverished given the richness of the data. The authors have nicely mapped out the directional tuning for almost the entire visual field. Instead of reducing this measurement to 2 values (center and direction), I was wondering if there is a better method to fully utilize the data at hand to get a better characterization of these DNs. As the authors are aware, local features alone can be ambiguous in characterizing optic flows. What's more, taking into account more global features can be useful for discovering potentially new cell types.

This is a great point, and we did an extensive analysis of other receptive field properties in this study (shown in supp fig 1). In addition, and as mentioned above, we have published a clustering analysis across receptive field properties of these neurons (Nicholas et al, J Comp Physiol A, 2020, cited in paper). The point that we attempted to make in this paper was that by using two strikingly simple metrics, we can reliably distinguish which of the two neuron types we are recording from (if we accept that there are two main types that we are likely to record from) simply based on location and overall directional preference. This makes automated analysis very easy and straightforward. Indeed, we now use this routinely to ID what neuron we are recording from, rather than making a human-based assumption.

However, we agree that further in depth analysis is warranted. Therefore, to address this, we will provide additional receptive field analysis and clustering in the revised version of the MS. In addition, we want to highlight that all data is uploaded to DataDryad for anyone interested in doing additional in-depth analyses.

Line 131, it wasn't clear to me why full-screen stimuli were used for comparison here, instead of the full receptive field maps. Male flies exhibit sexual dimorphic behaviors only during courtship, which would suggest that small-sized visual stimuli (mimicking an intruder or female conspecific) would be better suited to elicit dimorphic neuronal responses. A similar comment applies to the later results as well. Based on the receptive field mapping in Figure 1, I'm under the impression that these 2 DN types are more suited to detect wide-field optic flows, those induced by self-motion as mentioned in the manuscript. The results are still very interesting, but it's good to make this point clear early on to help set appropriate expectations. Conversely, this would also suggest that there are other visual DN types that are responsible for the courtship-related sexually dimorphic behaviors.

Thank you for mentioning these important points. Our reasoning for using full-screen stimuli for the analysis on line 131 was that since we used the small sinusoidal gratings for mapping the receptive fields, and to subsequently classify the neurons, it would be unfair to use the same data to investigate potential sexual dimorphism. I.e., we selected neurons that fulfilled certain criteria, and then we cannot rightfully use the same criteria to determine differences. This was not explicitly mentioned in the paper, so we will modify the text to make this clear to the Reader.

However, in Supp Figure 1d/e we show that there are no striking receptive field differences between males and females in terms of receptive field center nor directional preference. In Supp Figure 1f we show that there is no difference between male and female receptive field height and width. We will modify the text to draw the Reader’s attention to this figure, and also mention the additional analysis done in response to the comment above.

As a side note, I personally expected at least DNHS1 to have a smaller receptive field in males, as the hoverfly HSN is strikingly sexually dimorphic (Nordström et al, Curr Biol 2008), and also very sensitive to small objects. However, while optic flow sensitive DNs do respond to small objects (see e.g. the J Comp Physiol paper mentioned above) we did not detect any obvious sexual dimorphism in receptive field properties. Indeed, we think that a different subset of DNs control target pursuit behavior (target selective DNs (TSDNs)). This will be addressed in the modified version of the paper.

eLife Assessment

This useful study examines whether the sugar trehalose, coordinates energy supply with the gene programs that build muscle in the cotton bollworm (Helicoverpa armigera). The evidence for this currently is incomplete. The central claim - that trehalose specifically regulates an E2F/Dp-driven myogenic program - is not supported by the specificity of the data: perturbations and sequencing are systemic, alternative explanations such as general energy or amino-acid scarcity remain plausible, and mechanistic anchors are also limited. The work will interest researchers in insect metabolism and development; focused, tissue-resolved measurements together with stronger mechanistic controls would substantially strengthen the conclusions.

Reviewer #1 (Public review):

Summary:

In this work by Mohite et al., they have used transcriptomic and metabolic profiling of H. armigera, muscle development, and S. frugiperda to link energy trehalose metabolism and muscle development. They further used several different bioinformatics tools for network analysis to converge upon transcriptional control as a potential mechanism of metabolite-regulated transcriptional programming for muscle development. The authors have also done rescue experiments where trehalose was provided externally by feeding, which rescues the phenotype. Though the study is exciting, there are several concerns and gaps that lead to the current results as purely speculative. It is difficult to perform any genetic experiments in non-model insects; the authors seem to suggest a similar mechanism could also be applicable in systems like Drosophila; it might be possible to perform experiments to fill some missing mechanistic details.

A few specific comments below:

The authors used N-(phenylthio) phthalimide (NPP), a trehalose-6-phosphate phosphatase (TPP) inhibitor. They also find several genes, including enzymes of trehalose metabolism, that change. Further, several myogenic genes are downregulated in bulk RNA sequencing. The major caveat of this experiment is that the NPP treatment leads to reduced muscle development, and so the proportion of the samples from the muscles in bulk RNA sequencing will be relatively lower, which might have led to the results. So, a confirmatory experiment has to be performed where the muscle tissues are dissected and sequenced, or some of the interesting targets could be validated by qRT-PCR. Further to overcome the off-target effects of NPP, trehalose rescue experiments could be useful.

Even the reduction in the levels of ADP, NAD, NADH, and NMN, all of which are essential for efficient energy production and utilization, could be due to the loss of muscles, which perform predominantly metabolic functions due to their mitochondria-rich environment. So it becomes difficult to judge if the levels of these energy molecules' reduction are due to a cause or effect.

The authors have used this transcriptomic data for pathway enrichment analysis, which led to the E2F family of transcription factors and a reduction in the level of when trehalose metabolism is perturbed. EMSA experiments, though, confirm a possibility of the E2F interaction with the HaTPS/TPP promoter, but it lacks proper controls and competition to test the actual specificity of this interaction. Several transcription factors have DNA-binding domains and could bind any given DNA weakly, and the specificity is ideally known only from competitive and non-competitive inhibition studies.

The work seems to have connected the trehalose metabolism with gene expression changes, though this is an interesting idea, there are no experiments that are conclusive in the current version of the manuscript. If the authors can search for domains in the E2F family of transcription factors that can bind to the metabolite, then, if not, a chip-seq is essential to conclusively suggest the role of E2F in regulating gene expression tuned by the metabolites.

Some of the above concerns are partially addressed in experiments where silencing of E2F/Dp shows similar phenotypes as with NPP and dsRNA. It is also notable that silencing any key transcription factor can have several indirect effects, and delayed pupation and lethality could not be definitely linked to trehalose-dependent regulation.

Trehalose rescue experiments that rescue phenotype and gene expression are interesting. But is it possible that the fed trehalose is metabolized in the gut and might not reach the target tissue? In which case, the role of trehalose in directly regulating transcription factors becomes questionable. So, a confirmatory experiment is needed to demonstrate that the fed trehalose reaches the target tissues. This could possibly be done by measuring the trehalose levels in muscles post-rescue feeding. Also, rescue experiments need to be done with appropriate control sugars.

No experiments are performed with non-target control dsRNA. All the experiments are done with an empty vector. But an appropriate control should be a non-target control.

Reviewer #2 (Public review):

Summary:

This study shows that the knockdown of the effects of TPS/TPP in Helicoverpa armigera and Spodoptera frugiperda can be rescued by trehalose treatment. This suggests that trehalose metabolism is necessary for development in the tissues that NPP and dsRNA can reach.

Strengths:

This study examines an important metabolic process beyond model organisms, providing a new perspective on our understanding of species-specific metabolism equilibria, whether conserved or divergent.

Weaknesses:

While the effects observed may be truly conserved across Lepidopterans and may be muscle-specific, the study largely relies on one species and perturbation methods that are not muscle-specific. The technical limitations arising from investigations outside model systems, where solid methods are available, limit the specificity of inferences that may be drawn from the data.

Reviewer #3 (Public review):

The hypothesis is that Trehalose metabolism regulates transcriptional control of muscle development in lepidopteran insects.

The manuscript investigates the role of Trehalose metabolism in muscle development. Through sequencing and subsequent bioinformatics analysis of insects with perturbed trehalose metabolism (knockdown of TPS/TPP), the authors have identified transcription factor E2F, which was validated through RT-PCR. Their hypothesis is that trehalose metabolism regulates E2F, which then controls the myogenic genes. Counterintuitive to this hypothesis, the investigators perform EMSAs with the E2F protein and promoter of the TPP gene and show binding. Their knockdown experiments with Dp, the binding partner of E2F, show direct effect on several trehalose metabolism genes. Similar results are demonstrated in the trehalose feeding experiment, where feeding trehalose leads to partial rescue of the phenotype observed as a result of Dp knockdown. This seems contradictory to their hypothesis. Even more intriguing is a similar observation between paramyosin, a structural muscle protein, and E2F/Dp - they show that paramyosin regulates E2F/Dp and E2F/Dp regulated paramyosin. The only plausible way to explain the results is the existence of a feed-forward loop between TPP-E2F/Dp and paramyosin-E2F/Dp. But the authors have mentioned nothing in this line. Additionally, I think trehalose metabolism impacts amino acid content in insects, and that will have a direct bearing on muscle development. The sequencing analysis and follow-up GSEA studies have demonstrated enrichment of several amino acid biosynthetic genes. Yet authors make no efforts to measure amino acid levels or correlate them with muscle development. Any study aiming to link trehalose metabolism and muscle development and not considering the above points will be incomplete.

The result section of the manuscript is quite concise, to my understanding (especially the initial few sections), which misses out on mentioning details that would help readers understand the paper better. While technical details of the methods should be in the Materials and Methods section, the overall experimental strategy for the experiments performed should be explained in adequate detail in the results section itself or in figure legends. I would request authors to include more details in the results section. As an extension of the comment above, many times, abbreviations have been used without introducing them. A thorough check of the manuscript is required regarding this.

The Spodoptera experiments appear ad hoc and are insufficient to support conservation beyond Helicoverpa. To substantiate this claim, please add a coherent, minimal set of Spodoptera experiments and present them in a dedicated subsection. Alternatively, consider removing these data and limiting the conclusions (and title) to H. armigera.

In order to check the effects of E2F/Dp, a dsRNA-mediated knockdown of Dp was performed. Why was the E2F protein, a primary target of the study, not chosen as a candidate? The authors should either provide justification for this or perform the suggested experiments to come to a conclusion. I would like to point out that such experiments were performed in Drosophila.

Silencing of HaDp resulted in a significant decrease in HaE2F expression. I find this observation intriguing. DP is the cofactor of E2F, and they both heterodimerise and sit on the promoter of target genes to regulate them. I would request authors to revisit this result, as it contradicts the general understanding of how E2F/Dp functions in other organisms. If Dp indeed controls E2F expression, then further experiments should be conducted to come to a conclusion convincingly. Additionally, these results would need thorough discussion with citations of similar results observed for other transcription factor-cofactor complexes.

I consider the overall bioinformatics analysis to remain very poorly described. What is specifically lacking is clear statements about why a particular dry lab experiments were conducted.

In my judgement, the EMSA analysis presented is technically poor in quality. It lacks positive and negative controls, does not show mutation analysis or super shifts. Also, it lacks any competition assays that are important to prove the binding beyond doubt. I am not sure why protein is not detected at all in lower concentrations. Overall, the EMSA assays need to be redone; I find the current results to be unacceptable.

GSEA studies clearly indicate enrichment of the amino acid synthesis gene in TPP knockdown samples. This supports the plausible theory that a lack of Trehalose means a lack of enough nutrients, therefore less of that is converted to amino acids, and therefore muscle development is compromised. Yet the authors make no effort to measure amino acid levels. While nutrients can be sensed through signalling pathways leading to shut shutdown of myogenic genes, a simple and direct correlation between less raw material and deformed muscle might also be possible.

The authors are encouraged to stick to one color palette while demonstrating sequencing results. Choosing a different color palette for representing results from the same sequencing analysis confuses readers.

Expression of genes, as understood from sequencing analysis in Figure 1D, Figure 2F, and Figure 3D, appears to be binary in nature. This result is extremely surprising given that the qRT-PCR of these genes have revealed a checker and graded expression.

In several graphs, non-significant results have been interpreted as significant in the results section. In a few other cases, the reported changes are minimal, and the statistical support is unclear; please recheck the analyses and include exact statistics. In the results section, fold changes observed should be discussed, as well as the statistical significance of the observed change.

Finally, I would add that trehaolse metabolism regulates cell cycle genes, and muscle development genes establish correlation and causation. The authors should ensure that any comments they make are backed by evidence.

Author response:

eLife Assessment

This useful study examines whether the sugar trehalose, coordinates energy supply with the gene programs that build muscle in the cotton bollworm (Helicoverpa armigera). The evidence for this currently is incomplete. The central claim - that trehalose specifically regulates an E2F/Dp-driven myogenic program - is not supported by the specificity of the data: perturbations and sequencing are systemic, alternative explanations such as general energy or amino-acid scarcity remain plausible, and mechanistic anchors are also limited. The work will interest researchers in insect metabolism and development; focused, tissue-resolved measurements together with stronger mechanistic controls would substantially strengthen the conclusions.

We thank the reviewer for the thoughtful and constructive evaluation of our work and for recognizing its potential relevance to researchers working on insect metabolism and development. We fully agree that our current evidence is preliminary and that the mechanistic link between trehalose and the E2F/Dp‑driven myogenic program needs to be strengthened.

Our intention was to present trehalose-E2F/Dp coupling as a working model emerging from our data, rather than as a fully established pathway. We agree that systemic manipulations of trehalose and whole‑larval RNA‑seq cannot fully differentiate global metabolic stress from specific effects on myogenic programs. In the revision, we plan to include additional metabolic readouts (e.g., ATP/AMP ratio, key amino acids where available) to better discuss the overall energetic and nutritional state. We will reanalyze our RNA‑seq data to more clearly distinguish broad stress/metabolic signatures from cell‑cycle/myogenic signatures. Furthermore, we will reframe our discussion to explicitly state that we cannot completely rule out a contribution of general energy or amino‑acid scarcity at this stage.

We acknowledge that, with our current experiments, the specificity for an E2F/Dp‑driven program is inferred mainly from enrichment of E2F targets among differentially expressed genes, and expression changes in canonical E2F partners and downstream cell‑cycle/myogenic regulators. To address this more rigorously, we are performing targeted qRT-PCR for a panel of well‑characterized E2F/Dp target genes and myogenic markers in larval muscle versus non‑muscle tissues, following trehalose perturbation. Where technically feasible, testing whether partial knockdown of HaE2F or HaDp modifies the effect of trehalose manipulation on selected myogenic markers. These data, even if limited, will help to provide a more direct functional link, and we will include them in the manuscript if completed in time. In parallel, we will soften statements that imply a fully established, trehalose‑specific regulation of E2F/Dp and instead present this as a strong candidate pathway suggested by the current data.

We fully agree that tissue‑resolved analyses are essential to move from systemic correlations to causality in muscle. We are in the process of standardizing larval muscle dissections and isolating thoracic/abdominal body wall muscle for trehalose, glycogen, and expression assays. Comparing expression of key metabolic and myogenic genes in muscle versus fat body and midgut, under trehalose manipulation. These tissue‑resolved data will directly address whether the transcriptional changes we report are preferentially localized to muscle.

We are grateful for the reviewer’s critical but encouraging comments. We will moderate our central claims, also explicitly consider and discuss alternative explanations. Further, we will add tissue‑resolved and more focused mechanistic data as far as possible within the current revision. We believe these changes will substantially strengthen the manuscript and better align our conclusions with the evidence we presently have.

Public Reviews:

Reviewer #1 (Public review):

Summary:

In this work by Mohite et al., they have used transcriptomic and metabolic profiling of H. armigera, muscle development, and S. frugiperda to link energy trehalose metabolism and muscle development. They further used several different bioinformatics tools for network analysis to converge upon transcriptional control as a potential mechanism of metabolite-regulated transcriptional programming for muscle development. The authors have also done rescue experiments where trehalose was provided externally by feeding, which rescues the phenotype. Though the study is exciting, there are several concerns and gaps that lead to the current results as purely speculative. It is difficult to perform any genetic experiments in non-model insects; the authors seem to suggest a similar mechanism could also be applicable in systems like Drosophila; it might be possible to perform experiments to fill some missing mechanistic details.

A few specific comments below:

The authors used N-(phenylthio) phthalimide (NPP), a trehalose-6-phosphate phosphatase (TPP) inhibitor. They also find several genes, including enzymes of trehalose metabolism, that change. Further, several myogenic genes are downregulated in bulk RNA sequencing. The major caveat of this experiment is that the NPP treatment leads to reduced muscle development, and so the proportion of the samples from the muscles in bulk RNA sequencing will be relatively lower, which might have led to the results. So, a confirmatory experiment has to be performed where the muscle tissues are dissected and sequenced, or some of the interesting targets could be validated by qRT-PCR. Further to overcome the off-target effects of NPP, trehalose rescue experiments could be useful.

Thank you for this valuable comment. We will validate the gene expression data using qRT-PCR on muscle tissue samples from both treated and control groups. This will help determine whether the gene expression patterns observed in the RNA-seq data are muscle-specific or systemic.

Even the reduction in the levels of ADP, NAD, NADH, and NMN, all of which are essential for efficient energy production and utilization, could be due to the loss of muscles, which perform predominantly metabolic functions due to their mitochondria-rich environment. So it becomes difficult to judge if the levels of these energy molecules' reduction are due to a cause or effect.

We thank the reviewer for this thoughtful comment and agree that reduced levels of ADP, NAD, NADH, and NMN could arise either from a disturbance of energy metabolism or from loss of mitochondria‑rich muscles. Our current data cannot fully separate these two possibilities. Still, several studies support the interpretation that perturbing trehalose metabolism causes a primary systemic energy deficit that is coupled to mitochondrial function, not merely a passive consequence of tissue loss.

For example:

(1) Our previous study in H. armigera showed that chemical inhibition of trehalose synthesis results in depletion of trehalose, glucose, glucose‑6‑phosphate, and suppression of the TCA cycle, indicating reduced energy levels and dysregulated fatty‑acid oxidation (Tellis et al., 2023).

(2) Chang et al. (2022) showed that trehalose catabolism and mitochondrial ATP production are mechanistically linked. HaTreh1 localizes to mitochondria and physically interacts with ATP synthase subunit α. 20‑hydroxyecdysone increases HaTreh1 expression, enhances its binding to ATP synthase, and elevates ATP content, while knockdown of HaTreh1 or HaATPs‑α reduces ATP levels.

(3) Similarly, our previous study inhibition of Treh activity in H. armigera generates an “energy‑deficient condition” characterized by deregulation of carbohydrate, protein, fatty‑acid, and mitochondria‑related pathways, and a concomitant reduction in key energy metabolites (Tellis et al., 2024).

(4) The starvation study in H. armigera has shown that reduced hemolymph trehalose is associated with respiratory depression and large‑scale reprogramming of glycolysis and fatty‑acid metabolism (Jiang et al., 2019).

These findings support a direct coupling between trehalose availability and systemic energy/redox state. Therefore, the coordinated decrease in ADP, NAD, NADH, and NMN following TPS/TPP silencing is consistent with a primary disturbance of systemic energy and mitochondrial metabolism rather than exclusively a secondary consequence of muscle loss. We agree, however, that the present whole‑larva metabolite measurements do not allow a quantitative partitioning between changes due to altered muscle mass and those due to intrinsic metabolic impairment at the cellular level. Thus, tissue-specific quantification of these metabolites would allow us to directly test whether altered energy metabolites are a cause or consequence of muscle loss.

References:

(1) Tellis, M. B., Mohite, S. D., Nair, V. S., Chaudhari, B. Y., Ahmed, S., Kotkar, H. M., & Joshi, R. S. (2024). Inhibition of Trehalose Synthesis in Lepidoptera Reduces Larval Fitness. Advanced Biology, 8(2), 2300404.

(2) Chang, Y., Zhang, B., Du, M., Geng, Z., Wei, J., Guan, R., An, S. and Zhao, W., 2022. The vital hormone 20-hydroxyecdysone controls ATP production by upregulating the binding of trehalase 1 with ATP synthase subunit α in Helicoverpa armigera. Journal of Biological Chemistry, 298(2).

(3) Tellis, M., Mohite, S. and Joshi, R., 2024. Trehalase inhibition in Helicoverpa armigera activates machinery for alternate energy acquisition. Journal of Biosciences, 49(3), p.74.

(4) Jiang, T., Ma, L., Liu, X.Y., Xiao, H.J. and Zhang, W.N., 2019. Effects of starvation on respiratory metabolism and energy metabolism in the cotton bollworm Helicoverpa armigera (Hübner)(Lepidoptera: Noctuidae). Journal of Insect Physiology, 119, p.103951.

The authors have used this transcriptomic data for pathway enrichment analysis, which led to the E2F family of transcription factors and a reduction in the level of when trehalose metabolism is perturbed. EMSA experiments, though, confirm a possibility of the E2F interaction with the HaTPS/TPP promoter, but it lacks proper controls and competition to test the actual specificity of this interaction. Several transcription factors have DNA-binding domains and could bind any given DNA weakly, and the specificity is ideally known only from competitive and non-competitive inhibition studies.

We thank the reviewer for this important comment and fully agree that EMSA alone, without appropriate competition and control reactions, cannot establish the specificity or functional relevance of a transcription factor-DNA interaction. In our study, we found the E2F family from GRN analysis of the RNA seq data obtained upon HaTPS/TPP silencing, suggesting a potential regulatory connection. After that, we predicted E2F binding sites on the promoter of HaTPS/TPP. The EMSA experiments were intended as preliminary evidence that E2F can associate with the HaTPS/TPP promoter in vitro. We will clarify this in the manuscript by softening our conclusion to indicate that our data support a “possible E2F-HaTPS/TPP interaction”. We also perform EMSA with specific and non‑specific competitors to confirm the E2F binding to the HaTPS/TPP promoter.

The work seems to have connected the trehalose metabolism with gene expression changes, though this is an interesting idea, there are no experiments that are conclusive in the current version of the manuscript. If the authors can search for domains in the E2F family of transcription factors that can bind to the metabolite, then, if not, a chip-seq is essential to conclusively suggest the role of E2F in regulating gene expression tuned by the metabolites.

A previous study in D. melanogaster, Zappia et al., (2016) showed vital role of E2F in skeletal muscle required for animal viability. They have shown that Dp knockdown resulted in reduced expression of genes encoding structural and contractile proteins, such as Myosin heavy chain (Mhc), fln, Tropomyosin 1 (Tm1), Tropomyosin 2 (Tm2), Myosin light chain 2 (Mlc2), sarcomere length short (sals) and Act88F, and myogenic regulators, such as held out wings (how), Limpet (Lmpt), Myocyte enhancer factor 2 (Mef2) and spalt major (salm). Also, ChiP-qRT-PCR showed upstream regions of myogenic genes, such as how, fln, Lmpt, sals, Tm1 and Mef2, were specifically enriched with E2f1, E2f2, and Dp antibodies in comparison with a nonspecific antibody. Further, Zappia et al. (2019) reported a chip-seq dataset that suggests that E2F/Dp directly activates the expression of glycolytic and mitochondrial genes during muscle development. Zappia et al., (2023) showed the regulation of one of the glycolytic genes, Phosphoglycerate kinase (Pgk) by E2F during Drosophila development.

However, the regulation of trehalose metabolic genes by E2F/Dp and vice versa was not studied previously. So here in our study, we tried to understand the correlation of trehalose metabolism and E2F/Dp in the muscle development of H. armigera.

References:

(1) Zappia, M.P. and Frolov, M.V., 2016. E2F function in muscle growth is necessary and sufficient for viability in Drosophila. Nature Communications, 7(1), p.10509.

(2) Zappia, M.P., Rogers, A., Islam, A.B. and Frolov, M.V., 2019. Rbf activates the myogenic transcriptional program to promote skeletal muscle differentiation. Cell reports, 26(3), pp.702-719.

(3) Zappia, M. P., Kwon, Y.-J., Westacott, A., Liseth, I., Lee, H. M., Islam, A. B., Kim, J., & Frolov, M. V. (2023a). E2F regulation of the Phosphoglycerate kinase gene is functionally important in Drosophila development. Proceedings of the National Academy of Sciences, 120(15), e2220770120.

Some of the above concerns are partially addressed in experiments where silencing of E2F/Dp shows similar phenotypes as with NPP and dsRNA. It is also notable that silencing any key transcription factor can have several indirect effects, and delayed pupation and lethality could not be definitely linked to trehalose-dependent regulation.

Yes. It’s true that silencing of any key transcription factor can have several indirect effects. Our intention was not to argue that delayed pupation and lethality are exclusively due to trehalose-dependent regulation, but that E2F/Dp and HaTPS/TPP silencing showed a consistent set of phenotypes and molecular changes, such as (i) transcriptomic enrichment of E2F targets upon trehalose perturbation, (ii) reduced HaTPS/TPP expression following E2F/Dp silencing, (iii) reduced myogenic gene expression that parallels the phenotypes observed with HaTPS/TPP silencing and (iv) restoration of E2F and Dp expression in E2F/Dp‑silenced insects upon trehalose feeding in the rescue assay. Together, these findings support a functional association between E2F/Dp and trehalose homeostasis. At the same time, we fully acknowledge that these results do not exclude additional, trehalose‑independent roles of E2F/Dp in development.

Trehalose rescue experiments that rescue phenotype and gene expression are interesting. But is it possible that the fed trehalose is metabolized in the gut and might not reach the target tissue? In which case, the role of trehalose in directly regulating transcription factors becomes questionable. So, a confirmatory experiment is needed to demonstrate that the fed trehalose reaches the target tissues. This could possibly be done by measuring the trehalose levels in muscles post-rescue feeding. Also, rescue experiments need to be done with appropriate control sugars.

Yes, it’s possible that, to some extent, trehalose is metabolized in the gut. Even though trehalase is present in the insect gut, some of the trehalose will be absorbed via trehalose transporters on the gut lining. Trehalose feeding was not rescued in insects fed with the control diet (empty vector and dsHaTPP), which contains chickpea powder, which is composed of an ample amount of amino acids and carbohydrates. Insects fed exclusively on a trehalose-containing diet are rescued, but not on a control diet that contains other carbohydrates. We agree that direct measurement of trehalose in target tissues will provide important confirmation. In the manuscript, we will measure trehalose levels in muscle, gut, and haemolymph after trehalose feeding.

No experiments are performed with non-target control dsRNA. All the experiments are done with an empty vector. But an appropriate control should be a non-target control.

Yes, there was no experiment with non-target dsRNA. Earlier, we have optimized a protocol for dsRNA delivery and its effectiveness in target knockdown (concentration, time) experiment, and published several research articles using a similar protocol:

(1) Chaudhari, B.Y., Nichit, V.J., Barvkar, V.T. and Joshi, R.S., 2025. Mechanistic insights in the role of trehalose transporter in metabolic homeostasis in response to dietary trehalose. G3: Genes, Genomes, Genetics, p. jkaf303.

(2) Barbole, R.S., Sharma, S., Patil, Y., Giri, A.P. and Joshi, R.S., 2024. Chitinase inhibition induces transcriptional dysregulation altering ecdysteroid-mediated control of Spodoptera frugiperda development. Iscience, 27(3).

(3) Patil, Y.P., Wagh, D.S., Barvkar, V.T., Gawari, S.K., Pisalwar, P.D., Ahmed, S. and Joshi, R.S., 2025. Altered Octopamine synthesis impairs tyrosine metabolism affecting Helicoverpa armigera vitality. Pesticide Biochemistry and Physiology, 208, p.106323.

(4) Tellis, M.B., Chaudhari, B.Y., Deshpande, S.V., Nikam, S.V., Barvkar, V.T., Kotkar, H.M. and Joshi, R.S., 2023. Trehalose transporter-like gene diversity and dynamics enhances stress response and recovery in Helicoverpa armigera. Gene, 862, p.147259.

(5) Joshi, K.S., Barvkar, V.T., Hadapad, A.B., Hire, R.S. and Joshi, R.S., 2025. LDH-dsRNA nanocarrier-mediated spray-induced silencing of juvenile hormone degradation pathway genes for targeted control of Helicoverpa armigera. International Journal of Biological Macromolecules, p.148673.

The same vector backbone and preparation procedures were used for both control and experimental constructs, allowing us to specifically compare the effects of the target dsRNA. The phenotypes and gene expression changes we observed were specific to the target genes and were not seen in the empty vector controls, suggesting that the effects are not due to nonspecific responses of dsRNA delivery or vector components.<br /> We acknowledge your suggestions, and in future studies, we will keep non-target dsRNA as a control in silencing assays.

Reviewer #2 (Public review):

Summary:

This study shows that the knockdown of the effects of TPS/TPP in Helicoverpa armigera and Spodoptera frugiperda can be rescued by trehalose treatment. This suggests that trehalose metabolism is necessary for development in the tissues that NPP and dsRNA can reach.

Strengths:

This study examines an important metabolic process beyond model organisms, providing a new perspective on our understanding of species-specific metabolism equilibria, whether conserved or divergent.

Weaknesses:

While the effects observed may be truly conserved across Lepidopterans and may be muscle-specific, the study largely relies on one species and perturbation methods that are not muscle-specific. The technical limitations arising from investigations outside model systems, where solid methods are available, limit the specificity of inferences that may be drawn from the data.

Thank you for this potting out this experimental weakness. We will validate the gene expression data using qRT-PCR on muscle tissue samples from both treated and control groups. We will also perform metabolite analysis with muscle samples. This will help to determine whether the observed gene expression patterns and metabolite changes are muscle-specific or systemic.

Reviewer #3 (Public review):

The hypothesis is that Trehalose metabolism regulates transcriptional control of muscle development in lepidopteran insects.

The manuscript investigates the role of Trehalose metabolism in muscle development. Through sequencing and subsequent bioinformatics analysis of insects with perturbed trehalose metabolism (knockdown of TPS/TPP), the authors have identified transcription factor E2F, which was validated through RT-PCR. Their hypothesis is that trehalose metabolism regulates E2F, which then controls the myogenic genes. Counterintuitive to this hypothesis, the investigators perform EMSAs with the E2F protein and promoter of the TPP gene and show binding. Their knockdown experiments with Dp, the binding partner of E2F, show direct effect on several trehalose metabolism genes. Similar results are demonstrated in the trehalose feeding experiment, where feeding trehalose leads to partial rescue of the phenotype observed as a result of Dp knockdown. This seems contradictory to their hypothesis. Even more intriguing is a similar observation between paramyosin, a structural muscle protein, and E2F/Dp - they show that paramyosin regulates E2F/Dp and E2F/Dp regulated paramyosin. The only plausible way to explain the results is the existence of a feed-forward loop between TPP-E2F/Dp and paramyosin-E2F/Dp. But the authors have mentioned nothing in this line. Additionally, I think trehalose metabolism impacts amino acid content in insects, and that will have a direct bearing on muscle development. The sequencing analysis and follow-up GSEA studies have demonstrated enrichment of several amino acid biosynthetic genes. Yet authors make no efforts to measure amino acid levels or correlate them with muscle development. Any study aiming to link trehalose metabolism and muscle development and not considering the above points will be incomplete.

We appreciate the reviewer’s efforts in the careful evaluation of this manuscript and constructive comments. From our and earlier data we found it was difficult to consider linear pathway “trehalose → E2F → muscle,” but rather a regulatory module in which trehalose metabolism and E2F/Dp form an interdependent circuit controlling myogenic genes. E2F/Dp binds and activates trehalose metabolism genes (TPS/TPP, Treh1) and myogenic structural genes, consistent with EMSA (TPS/TPP-E2F) and predicted binding sites of E2F on metabolic genes, Treh1, Pgk, and myogenic genes such as Act88F, Prm, Tm1, Fln, etc. At the same time, perturbing trehalose synthesis reduces E2F/Dp expression and myogenic gene expression, and trehalose feeding partially restores all three. This bidirectional influence is similar to E2F‑dependent control of carbohydrate metabolism and systemic sugar homeostasis described in D. melanogaster, where E2F/Dp both regulates metabolic genes and is itself constrained by metabolic state (Zappia et al., 2023a; Zappia et al., 2021).

The reciprocal regulation between Prm and E2F/Dp is indeed intriguing. Rather than a paradox, we interpret this as evidence that E2F/Dp couples metabolic genes and structural muscle genes within a shared module, and that key sarcomeric components (such as paramyosin) feed back on this transcriptional program. Similar cross‑talk between E2F‑controlled metabolic programs and tissue function has been documented in D. melanogaster muscle and fat body, where E2F loss in one tissue elicits systemic changes in the other (Zappia et al., 2021). For further confirmation of E2F-regulated Prm, we will perform EMSA on the Prm promoter with appropriate controls.

We fully agree that amino‑acid metabolism is a critical missing piece. In the manuscript, we will quantify the amino acid levels and include the results: “Amino acids display differential levels showing cysteine, leucine, histidine, valine, and proline showed significant reductions, while isoleucine and lysine showed non-significant reductions upon trehalose metabolism perturbation. These results are consistent with previous reports published by Tellis et al. (2024) and Shi et al. (2016)”. We will reframe our conclusions more cautiously as establishing a trehalose-E2F/Dp-muscle development, while stating that “definitive causal links via amino‑acid metabolism remain to be demonstrated”.

Reference:

(1) Zappia, M. P., Kwon, Y.-J., Westacott, A., Liseth, I., Lee, H. M., Islam, A. B., Kim, J., & Frolov, M. V. (2023a). E2F regulation of the Phosphoglycerate kinase gene is functionally important in Drosophila development. Proceedings of the National Academy of Sciences, 120(15), e2220770120.

(2) Zappia, M.P., Guarner, A., Kellie-Smith, N., Rogers, A., Morris, R., Nicolay, B., Boukhali, M., Haas, W., Dyson, N.J. and Frolov, M.V., 2021. E2F/Dp inactivation in fat body cells triggers systemic metabolic changes. elife, 10, p.e67753.

(3)Tellis, M., Mohite, S. and Joshi, R., 2024. Trehalase inhibition in Helicoverpa armigera activates machinery for alternate energy acquisition. Journal of Biosciences, 49(3), p.74.

(4) Shi, J.F., Xu, Q.Y., Sun, Q.K., Meng, Q.W., Mu, L.L., Guo, W.C. and Li, G.Q., 2016. Physiological roles of trehalose in Leptinotarsa larvae revealed by RNA interference of trehalose-6-phosphate synthase and trehalase genes. Insect Biochemistry and Molecular Biology, 77, pp.52-68.

Author response image 1.

The result section of the manuscript is quite concise, to my understanding (especially the initial few sections), which misses out on mentioning details that would help readers understand the paper better. While technical details of the methods should be in the Materials and Methods section, the overall experimental strategy for the experiments performed should be explained in adequate detail in the results section itself or in figure legends. I would request authors to include more details in the results section. As an extension of the comment above, many times, abbreviations have been used without introducing them. A thorough check of the manuscript is required regarding this.

Thank you very much for pointing out this issue. We will revise the manuscript content according to these suggestions.

The Spodoptera experiments appear ad hoc and are insufficient to support conservation beyond Helicoverpa. To substantiate this claim, please add a coherent, minimal set of Spodoptera experiments and present them in a dedicated subsection. Alternatively, consider removing these data and limiting the conclusions (and title) to H. armigera.

We thank the reviewer for this helpful comment. We agree that, in this current version of the manuscript, the S. frugiperda experiments are not sufficiently systematic to support strong claims about conservation beyond H. armigera. Our primary focus in this study is indeed on H. armigera, and the addition of the S. frugiperda data was intended only as preliminary, supportive evidence rather than a central component of our conclusions. To avoid over‑interpretation and to keep the manuscript focused and coherent, we will remove all S. frugiperda data from the revised version, including the corresponding text and figures. We will also adjust the title, abstract, and conclusion to clearly state that our findings are limited to H. armigera.

In order to check the effects of E2F/Dp, a dsRNA-mediated knockdown of Dp was performed. Why was the E2F protein, a primary target of the study, not chosen as a candidate? The authors should either provide justification for this or perform the suggested experiments to come to a conclusion. I would like to point out that such experiments were performed in Drosophila.

Thank you for this thoughtful comment and the specific suggestion. We agree that directly targeting E2F would, in principle, be an informative complementary approach. In our study, however, we prioritized Dp knockdown for two main reasons. First, E2F is a large family, and E2F-Dp functions as an obligate heterodimer. Previous work in D. melanogaster has shown that depletion of Dp is sufficient to disrupt E2F-dependent transcription broadly, often with more efficient loss of complex activity than targeting individual E2F isoforms (Zappia et al., 2021; Zappia et al., 2016). Second, in our preliminary trials, we performed a dsRNA feeding assay with dsHaE2F, dsHaDp, and combined dsHaE2F plus dsHaDp. In that assay, we did not achieve silencing of E2F in dsRNA targeting HaE2F (dsHaE2F). So here, as E2F is a large family, other E2F isoforms may be compensating for the silencing effect of targeted HaE2F. However, HaE2F showed significantly reduced expression upon dsHaDp and combined dsHaE2F plus dsHaDp feeding (Figure A), whereas HaDp showed a significant reduction in its expression in all three conditions (Figure B).  As we observed reduced expression of both HaE2F and HaDp upon combined feeding of dsHaE2F and dsHaDp, we further performed a rescue assay by exogenous feeding of trehalose. We observed the significant upregulation of HaE2F, HaDp, trehalose metabolic genes (HaTPS/TPP and HaTreh1), and myogenic genes (HaPrm and HaTm2) (Figure C). For these reasons, we focused on Dp silencing as a more reliable way to impair E2F/Dp complex function in H. armigera.

Author response image 2.

References:

(1) Zappia, M.P. and Frolov, M.V., 2016. E2F function in muscle growth is necessary and sufficient for viability in Drosophila. Nature Communications, 7(1), p.10509.

(2) Zappia, M.P., Guarner, A., Kellie-Smith, N., Rogers, A., Morris, R., Nicolay, B., Boukhali, M., Haas, W., Dyson, N.J. and Frolov, M.V., 2021. E2F/Dp inactivation in fat body cells triggers systemic metabolic changes. elife, 10, p.e67753.

Silencing of HaDp resulted in a significant decrease in HaE2F expression. I find this observation intriguing. DP is the cofactor of E2F, and they both heterodimerise and sit on the promoter of target genes to regulate them. I would request authors to revisit this result, as it contradicts the general understanding of how E2F/Dp functions in other organisms. If Dp indeed controls E2F expression, then further experiments should be conducted to come to a conclusion convincingly. Additionally, these results would need thorough discussion with citations of similar results observed for other transcription factor-cofactor complexes.

Thank you for highlighting this point and for prompting us to examine these data more carefully. Silencing HaDp leading to reduced HaE2F mRNA is indeed unexpected if one only considers the canonical view of E2F/Dp as a heterodimer that co-occupies target promoters without strongly regulating each other’s expression. However, several lines of work suggest that transcription factor-cofactor networks frequently include feedback loops in which cofactors influence the expression of their partner TFs. First, in multiple systems, transcription factors and their cofactors are known to regulate each other’s transcription, forming positive or negative feedback loops. For example, in hematopoietic cells, the transcription factor Foxp3 controls the expression of many of its own cofactors, and some of these cofactors in turn facilitate or stabilize Foxp3 expression, forming an interconnected regulatory network rather than a simple one‑way interaction (Rudra et al., 2012). Second, E2F/Dp complexes exhibit non‑canonical regulatory mechanisms and can regulate broad sets of targets, including other transcriptional regulators. Several studies show that E2F/Dp proteins not only control classical cell‑cycle genes but also participate in diverse processes such as DNA damage signaling, mitochondrial function, and differentiation (Guarner et al., 2017; Ambrus et al., 2013; Sánchez-Camargo et al., 2021). In D. melanogaster, complete loss of dDP alters the expression of direct targets E2F/DP, including dATM (Guarner et al., 2017).

All these reports indicate that the E2F-Dp complex sits at the top of multi‑layer regulatory hierarchies. Such architectures make it plausible that Dp silencing in H. armigera could modulate HaE2F expression in a non-canonical way.

References:

(1) Rudra, D., DeRoos, P., Chaudhry, A., Niec, R.E., Arvey, A., Samstein, R.M., Leslie, C., Shaffer, S.A., Goodlett, D.R. and Rudensky, A.Y., 2012. Transcription factor Foxp3 and its protein partners form a complex regulatory network. Nature immunology, 13(10), pp.1010-1019.

(2) Guarner, A., Morris, R., Korenjak, M., Boukhali, M., Zappia, M.P., Van Rechem, C., Whetstine, J.R., Ramaswamy, S., Zou, L., Frolov, M.V. and Haas, W., 2017. E2F/DP prevents cell-cycle progression in endocycling fat body cells by suppressing dATM expression. Developmental cell, 43(6), pp.689-703.

(3) Ambrus, A.M., Islam, A.B., Holmes, K.B., Moon, N.S., Lopez-Bigas, N., Benevolenskaya, E.V. and Frolov, M.V., 2013. Loss of dE2F compromises mitochondrial function. Developmental cell, 27(4), pp.438-451.

(4) Sánchez-Camargo, V.A., Romero-Rodríguez, S. and Vázquez-Ramos, J.M., 2021. Non-canonical functions of the E2F/DP pathway with emphasis in plants. Phyton, 90(2), p.307.

I consider the overall bioinformatics analysis to remain very poorly described. What is specifically lacking is clear statements about why a particular dry lab experiments were conducted.

We again thank the reviewer for advising us to give a biological context/motivation for every bioinformatics analysis performed. The bioinformatics analyses devised here, try to explain the systems-level perturbations of HaTPS/TPP silencing to explain the observed phenotype and to discover transcription factors potentially modulating the HaTPS/TPP induced gene regulatory changes.

(1) Gene set enrichment analyses:

Differential gene expression analyses of the bulk RNA sequencing data followed by qRT-PCR confirmed the transcriptional changes in myogenic genes and gene expression alterations in metabolic and cell cycle-related genes. These perturbations merely confirmed the effect induced by HaTPS/TPP silencing in obviously expected genes. We wanted to see whether using an “unbiased” system-level statistical analyses like gene set enrichment analyses (GSEA), can reveal both expected and novel biological processes that underlie HaTPS/TPP silencing. GSEA results revealed large-scale transcriptional changes in 11 enriched processes, including amino acid metabolism, energy metabolism, developmental regulatory processes, and motor protein activity. GSEA not only divulged overall transcriptionally enriched pathways but also identified the genes undergoing synchronized pathway-level transcriptional change upon HaTPS/TPP silencing.

(2) Gene regulatory network analysis:

Although GSEA uncovered potential pathway-level changes, we were also interested in identifying the gene regulatory network associated with such large-scale process-level transcriptional perturbations. Interestingly, the biological processes undergoing perturbations were also heterogeneous (e.g., motor protein activity, energy metabolism, amino acid metabolism, etc.). We hypothesized that the inference of a causal gene regulatory network associated with the genes associated with GSEA-enriched biological processes should predict core/master transcription factors that might synchronously regulate metabolic and non-metabolic processes related to HaTPS/TPP silencing, thereby providing a broad understanding of the perturbed phenotype. The gene regulatory network analysis statistically inferred an “active” gene regulatory network corresponding to the GSEA-enriched KEGG gene sets. Ranking the transcription factors (TFs) based on the number of outgoing connections (outdegree centrality) within the active gene regulatory network, E2F family TFs were identified to be top-ranking, highly connected transcription factors associated with the transcriptionally enriched processes. This suggests that E2F family TFs are central to controlling the flow of regulatory information within this network. Intriguingly, E2F has been previously implicated in muscle development in insects (Zappia et al., 2016). Further extracting the regulated targets of E2F family TFs within this network revealed the mechanistic connection with the 11 enriched processes. This GRN analysis was crucial in discovering and prioritizing E2F TFs as central transcription factors mediating HaTPS/TPP silencing effects, which was not apparent using trivial analyses like differential gene expression analysis.

As per the reviewer’s suggestions, we will add these outlined points in the text of the manuscript (Results section) to further give context and clarity to the bioinformatics analyses conducted in this study.

In my judgement, the EMSA analysis presented is technically poor in quality. It lacks positive and negative controls, does not show mutation analysis or super shifts. Also, it lacks any competition assays that are important to prove the binding beyond doubt. I am not sure why protein is not detected at all in lower concentrations. Overall, the EMSA assays need to be redone; I find the current results to be unacceptable.

Thank you for pointing out this issue. We will reperform the EMSA analysis with appropriate controls.  Although the gel image was not clear, there was a light band of protein (indicated by the white square) observed in well No. 8, where we used 8 μg of E2F protein and 75 ng of HaTPS/TPP promoter, upon gel stained with SYPRO Ruby protein stain, suggesting weak HaTPS/TPP-E2F complex formation.

GSEA studies clearly indicate enrichment of the amino acid synthesis gene in TPP knockdown samples. This supports the plausible theory that a lack of Trehalose means a lack of enough nutrients, therefore less of that is converted to amino acids, and therefore muscle development is compromised. Yet the authors make no effort to measure amino acid levels. While nutrients can be sensed through signalling pathways leading to shut shutdown of myogenic genes, a simple and direct correlation between less raw material and deformed muscle might also be possible.

We quantified amino acid levels as per the suggestion, and we observed differential levels of amino acids upon trehalose metabolism perturbation.

However, we observed that insect were failed to rescue when fed a control chickpea-based artificial diet that contained nutrients required for normal growth and development. Based on this observation, we conclude that trehalose deficiency is the only possible cause for the defect in muscle development.

The authors are encouraged to stick to one color palette while demonstrating sequencing results. Choosing a different color palette for representing results from the same sequencing analysis confuses readers.

Thank you for the comment. We will revise the color palette as per the suggestion.

Expression of genes, as understood from sequencing analysis in Figure 1D, Figure 2F, and Figure 3D, appears to be binary in nature. This result is extremely surprising given that the qRT-PCR of these genes have revealed a checker and graded expression.

Thank you for pointing out this issue. We will revise the scale range for these figures to get more insights about gene expression levels and include figures as per the suggestion.

In several graphs, non-significant results have been interpreted as significant in the results section. In a few other cases, the reported changes are minimal, and the statistical support is unclear; please recheck the analyses and include exact statistics. In the results section, fold changes observed should be discussed, as well as the statistical significance of the observed change.

We will revise the analyses and include exact statistics as per the suggestion.

Finally, I would add that trehalose metabolism regulates cell cycle genes, and muscle development genes establish correlation and causation. The authors should ensure that any comments they make are backed by evidence.

We thank the reviewer for this insightful comment.  Although direct evidence in insects is currently lacking, multiple independent studies in yeast, plants and mammalian systems support a regulatory link between trehalose metabolism and the cell cycle. In budding yeast Saccharomyces cerevisiae, neutral Treh (Nth1) is directly phosphorylated and activated by the major cyclin‑dependent kinase Cdk1 at G1/S, routing stored trehalose into glycolysis to fuel DNA replication and mitosis (Ewald et al., 2016). CDK‑dependent regulation of trehalase activity has also been reported in plants, where CDC28‑mediated phosphorylation channels glucose into biosynthetic pathways necessary for cell proliferation (Lara-núñez et al., 2025). Furthermore, budding yeast cells accumulate trehalose and glycogen upon entry into quiescence and subsequently mobilize these stores to generate a metabolic “finishing kick” that supports re‑entry into the cell cycle (Silljé et al., 1999; Shi et al., 2010). Exogenous trehalose that perturbs the trehalose cycle impairs glycolysis, reduces ATP, and delays cell cycle progression in S. cerevisiae, highlighting a dose‑ and context‑dependent control of growth versus arrest (Zhang, Zhang and Li, 2020). In mammalian systems, trehalose similarly modulates proliferation-differentiation decisions. In rat airway smooth muscle cells, low trehalose concentrations promote autophagy, whereas higher doses induce S/G2–M arrest, downregulate Cyclin A1/B1, and trigger apoptosis, indicating a shift from controlled growth to cell elimination at higher exposure (Xiao et al., 2021). In human iPSC‑derived neural stem/progenitor cells, low‑dose trehalose enhances neuronal differentiation and VEGF secretion, while higher doses are cytotoxic, again highlighting a tunable impact on cell‑fate outcomes (Roose et al., 2025). In wheat, exogenous trehalose under heat stress reduces growth, lowers auxin, gibberellin, abscisic acid and cytokinin levels, and represses CycD2 and CDC2 expression, suggesting that trehalose signalling integrates with hormone pathways and core cell‑cycle regulators to restrain proliferation during stress (Luo, Liu, and Li, 2021). Together, these studies showed the importance of trehalose metabolism in cell‑cycle regulation to decide whether cells and tissues proliferate, differentiate, or remain quiescent.

With respect to muscle development, previous work has implicated glycolytic metabolism in myogenesis and muscle growth. Tixier et al. (2013) showed that loss of key glycolytic genes results in abnormally thin muscles, while Bawa et al. (2020) demonstrated that loss of TRIM32 decreases glycolytic flux and reduces muscle tissue size. These findings indicate that carbohydrate and energy metabolism pathways are important determinants of muscle structure and growth. However, there are no previous studies about the role of trehalose metabolism in muscle development, other than as an energy source, so here we specifically set out to establish the involvement of trehalose metabolism in muscle development.

References:

(1) Ewald, J.C. et al. (2016) “The yeast cyclin-dependent kinase routes carbon fluxes to fuel cell cycle progression,” Molecular cell, 62(4), pp. 532–545.

(2) Lara-núñez, A. et al. (2025) “The Cyclin-Dependent Kinase activity modulates the central carbon metabolism in maize during germination,” (January), pp. 1–16.

(3) Silljé, H.H.W. et al. (1999) “Function of trehalose and glycogen in cell cycle progression and cell viability in Saccharomyces cerevisiae,” Journal of bacteriology, 181(2), pp. 396–400.

(4) Shi, L. et al. (2010) “Trehalose Is a Key Determinant of the Quiescent Metabolic State That Fuels Cell Cycle Progression upon Return to Growth,” 21, pp. 1982–1990.

(5) Zhang, X., Zhang, Y. and Li, H. (2020) “Regulation of trehalose, a typical stress protectant, on central metabolisms, cell growth and division of Saccharomyces cerevisiae CEN. PK113-7D,” Food Microbiology, 89, p. 103459.

(6) Xiao, B. et al. (2021) “Trehalose inhibits proliferation while activates apoptosis and autophagy in rat airway smooth muscle cells,” Acta Histochemica, 123(8), p. 151810.

(7) Roose, S.K. et al. (2025) “Trehalose enhances neuronal differentiation with VEGF secretion in human iPSC-derived neural stem / progenitor cells,” Regenerative Therapy, 30, pp. 268–277.

(8) Luo, Y., Liu, X. and Li, W. (2021) “Exogenously-supplied trehalose inhibits the growth of wheat seedlings under high temperature by affecting plant hormone levels and cell cycle processes,” Plant Signaling & Behavior, 16(6).

(9) Tixier, V., Bataillé, L., Etard, C., Jagla, T., Weger, M., DaPonte, J.P., Strähle, U., Dickmeis, T. and Jagla, K., 2013. Glycolysis supports embryonic muscle growth by promoting myoblast fusion. Proceedings of the National Academy of Sciences, 110(47), pp.18982-18987.

(10) Bawa, S., Brooks, D.S., Neville, K.E., Tipping, M., Sagar, M.A., Kollhoff, J.A., Chawla, G., Geisbrecht, B.V., Tennessen, J.M., Eliceiri, K.W. and Geisbrecht, E.R., 2020. Drosophila TRIM32 cooperates with glycolytic enzymes to promote cell growth. elife, 9, p.e52358.

Finally, we appreciate the meticulous review of this manuscript and constructive comments. We will perform the recommended experiments, data analysis, and revise the manuscript accordingly.

eLife Assessment

This valuable study reports results showing how different neurons in the dysgranular retrosplenial cortex code spatial orientation. Specifically, the paper reports that some neurons maintain tuning for a single head direction across multi-compartmental environments, while other neurons are tuned to different head directions that reflect the geometry within each compartment. The study was viewed as likely to expand the field's understanding of directional tuning of neurons, but incomplete evidence was provided to support the conclusions.

Reviewer #1 (Public review):

Summary:

The dysgranular retrosplenial cortex (RSD) and hippocampus both encode information related to an animal's navigation through space. Here, the authors study the different ways in which these two brain regions represent spatial information when animals navigate through interconnected rooms. Most importantly, they find that the RSD contains a small fraction of neurons that encode properties of interconnected rooms by firing in different head directions within each room. This direction is shifted by 180 degrees in 2-room environments, and by 90 degrees in 4-room environments. While it cannot be definitively proven that this encoding is not just related to the presence of exits (doors) in each room, this is a noteworthy finding and will motivate further study in more complex and well-controlled environments to understand this coding scheme in the RSD. The recordings and analyses used to identify these multi-directional cells are mostly solid. Additional conclusions regarding the rotational symmetry across rooms seen in the RSD neurons that do not encode direction (representing the majority of RSD neurons) remain incomplete, given the evidence presented thus far. The differences between RSD and hippocampus encoding of space are clear and consistent with prior observations.

Strengths:

(1) Use of tetrode recordings from the RSD to identify multi-direction cells that only encode one direction in each room, but shift the preferred direction by either 180 or 90 degrees depending on the number of rooms in the environment.

(2) Solid controls to show that this multi-direction encoding is stable over time and across some environmental manipulations.

(3) Convincing evidence that these multi-direction cells can co-exist with single-direction head direction cells in the RSD (as both cell types can be simultaneously recorded).

(4) Convincing evidence for clear differences between directional and spatial encoding in the RSD versus hippocampus, consistent with prior observations.

Weaknesses:

(1) The paper mostly uses the term "retrosplenial cortex", but it is important to clarify that the study is only focused on the dysgranular retrosplenial cortex (RSD; Brodmann Area 30) and not the granular retrosplenial cortex (Brodmann Area 29). These are two distinct regions (despite the similar names), each with distinct connectivity and distinct behavioral encoding and function, so it is important to clarify in the abstract and title that the present study is solely about the RSD to prevent confusion in the literature.

(2) The proportion of each observed cell type is not clearly stated, although it is clear that the multi-directional cells are in the minority. Having the proportion of well-isolated neurons in distinct sessions that encode each type of information (e.g., multi vs single direction encoding) would greatly aid the interpretation of the result and help the field know how common each cell type is in the RSD.

(3) The authors state that "MDCs [multi-directional cells] never exhibited multidirectional activity within a single room" - but many of the single room examples from the 4-room environment (shown in Figures 2E and 2F) reveal multi-peaked directional encoding. This suggests that the multi-direction encoding may be more compatible with encoding some property of the number of exits rather than relative room orientations.

(4) The spatial rotation analyses of non-directional cell analyses are considered incomplete. This is impacted by the slower speed at the doors and hence altered firing rates (as evidenced in spatial rate plots). The population rate is not relevant as the correlational analyses are done on a single cell level. Since some cells fire more with increasing speed and others fire less, that will necessarily result in a population rate map that minimizes firing rate differences near the doorway, where the animals move more slowly. But on a single cell level, that reduced speed is having a big effect, as evidenced by individual rate map examples, and the rooms will need to be rotated to obtain a higher correlation by overlapping the doorway regions. This does not necessarily say anything about spatial coding across the two or four interconnected rooms being rotationally symmetric, and it would appear difficult to draw any conclusions related to spatial encoding from those analyses.

Reviewer #2 (Public review):

Summary:

Laurent et al. perform in vivo electrophysiological recordings in the retrosplenial cortex of rats foraging in multi-compartment environments with either identical or unique visual features. The authors characterize two types of directional signals in the area that they have previously reported: classic head direction cells anchored to the global allocentric reference frame and multi-direction cells (MDCs), which have a rotationally preserved directional field anchored to local compartments. The primary finding of this work is that MDCs seem sensitive to local environmental geometry rather than visual context. They also show that MDC tuning persists in the absence of hippocampal place field repetition, further dissociating the RSC local directional signal from the broader allocentric representation of space. A novel observation is that RSC non-directional spatial signals are anchored to the local environment, which could and should be explored further. While the data is solid and the analyses are mostly appropriate, the primary findings are incremental, and more interesting novel claims are not explored in detail or not explicitly tested.

Strengths:

The environmental manipulations clearly demonstrate that tuning is not modulated by complex visual information.

The finding that RSC two-dimensional spatial responses are stable and anchored to environmental features is novel and can be further explored in future work.

Weaknesses:

The observation that BDCs and MDCs are insensitive to visual context builds upon the author's previous work (and replicates aspects of Zhang et al., 2022) but leaves many open questions that are not addressed with the current set of experiments. Specifically, what exactly are MDCs anchoring to? The primary theory is that they anchor to environmental geometry, but there are no explicit experimental manipulations to test this theory. It is important to note that 2- and 4-compartment environments share many features, including the same cardinal axes, making any differences/similarities in these two conditions difficult to interpret.

The main finding presented with respect to BDC/MDs tuning is that they are not sensitive to visual context as manipulated by distinct visual patterns on the wall and floor in multicompartment environments. One could argue that the individual rooms are, in actuality, quite similar in low-level visual features - each possesses a large white background square visual feature on a single wall with a fixed relationship to the door(s). How can the authors rule out that i) BDC/MDC responses are modulated by these low-level features rather than geometry and/or ii) that the rats are not paying attention to any visual features at all? There is no task requiring them to indicate which room they are in. Furthermore, the doorways themselves are prominent visual features that are present in each context. It would be interesting to see if MDC/BDC tuning persisted in a square room where the number of doorways was manipulated to rule out this possibility.

A strong possibility is that the rotational symmetry of both MDCs and non-directional spatial neurons is related to i) door-related firing, 2) stereotyped movement, and 3) stereotyped directional sampling. In Supplemental Figure 8, the authors begin to address this by comparing a 'population ratemap' to a 'population speed map.' I do not think this is sufficient and is difficult to interpret. Instead, the authors should assess whether MDC and BDCs fire more at doorways and what the overlap is with the speed-modulated cells they report. Moreover, they should assess whether the spatial speed profile itself is rotationally symmetric within each session. It would also be useful to look at the confluence of the variables simultaneously using some form of regression analysis. The authors could generate a directional predictor that captures the main response property of these cells and see if it accounts for greater variability in spiking than speed or x,y position. Finally, rotationally symmetric directional sampling biases could arise from the doors being present on the same two walls in each room. The authors should assess whether MDC tuning is still present if directional sampling is randomly downsampled to match directional observations in each compartment.

Recent work has demonstrated that neurons with egocentric corner or boundary tuning are observed in RSC. The authors do not address whether egocentric tuning contributes to MDC signals. An explicit analysis of the relationship and potential overlap of MDC and egocentric populations is warranted.

Many of the MDCs presented in the main figures are not especially compelling. This includes alterations to MDC tuning in Figure 2, which is a key datapoint. The authors should show significantly more (if not all) examples of MDCs in each environment. It would similarly be useful to see all/more examples of non-directional spatially tuned neurons with rotationally symmetric firing patterns.

"One might hypothesize that specific environmental cues, such as door orientation or landmark positioning, drive these tuning shifts. However, our results argue against this interpretation. In four-room environments, each room had multiple entry points, yet MDCs never exhibited multidirectional activity within a single room."

I do not understand the logic here. Can the authors unpack this? Also, it is clear that some of the example cells have more than one peak in individual compartments. How is this quantified?

Reviewer #3 (Public review):

Summary:

The authors examine firing of dysgranular retrosplenial cortex (dRSC) neurons in relation to head orientation and location for rats exploring open-field environments. One environment utilized was a square arena with high walls that is split into two rectangular spaces connected by a doorway. Another environment is a square arena split into quadrants connected by doors near the center. For each, the different sub-spaces of the environments are either identical in terms of visual and tactile cues or different. For head direction neurons, the authors present one population where each neuron maintains a single tuning direction for the two or four sub-compartments of the two environments. A second population exhibits what is termed multi-directional firing, wherein neurons exhibit (overall) two or four head direction peaks in firing. For such neurons, firing in each of the sub-compartments is associated with only a single preferred direction, but the directions across compartments are shown to be at 180-degree (two-compartment environment) or 90-degree offsets. The offsets evidence tuning to the "same" orientation for the sub-compartments that are, in the global reference frame, oriented at 180 or 90 degree offsets. The results are similar whether or not the sub-compartments have the same or different tactile and visual cues. Thus, the first population is said to be global in its head direction tuning, while the second relates to each local environment in a way that is systematic across sub-compartments. Spatially-specific activity of another population of non-direction-tuned RSC neurons is examined, and comparisons of sub-compartment spatial firing maps suggest that spatial tuning in RSC also repeats across compartments when the firing maps for the compartments are rotated to match each other (as in physical space). Finally, a population of hippocampal "place" cells exhibited different location mapping across sub-compartments. The findings are interpreted to indicate that RSC can simultaneously map orientation in both local and global reference frames, possibly forming a mechanism whereby the sub-compartments' shared geometry (given by the boundary shapes and the door locations) can be related to each other and to the global space they share.

Strengths:

This paper addresses an interesting problem and expands how the field will think about directional tuning.

Weaknesses:

It is not clear that the experimental design allows for a clear interpretation of the data. Rates for preferred turning are low, as are ratemap correlations for spatially-tuned neurons.

(1) It is concerning that the neurons with head direction tuning have fairly low peak firing rates (mean close to 5 Hz), where prior studies examining head direction tuning in dRSC found head direction-tuned neurons with peak rates more than an order of magnitude higher (100 Hz or more). Under circumstances where neurons are tuned well to variables other than head direction (for example, angular velocity of movement), weak head direction tuning may be observed if those other variables are not sampled equally across head directions. The manuscript contains no rigorous control for this possibility. One place to start to address this issue would be to map out variables such as angular velocity by head orientation, and to test whether such relationships also carry 90 and 180 degree offsets.

(2) There is some question as to whether dRSC neurons (spatial or directional) following the sub-compartment "geometry" is appropriate in terms of interpreting the data. In the condition with sub-compartments carrying different tactile and visual cues, it seems that such cues pertain only to the floor of the environments. The distal visual space of the boundaries appears to be identical. One is left to wonder whether distinguishing environments according to boundary wall visual cues would lead to different results. The CA1 data does not help to rule this possibility out. A second reason to doubt the "shared geometry" interpretation is that there is no condition where sub-compartment geometry is varied. It is also the case that the sub-compartment doorways may stand as the only salient distal visual cue linking the environments. Local sensory cues and geometry seem not so disentangled in this study, but this is a major claim in the abstract.

(3) There is some concern with the interpretation that the spatial tuning of some dRSC neurons repeats in rotated form across sub-compartments. The firing rate map correlations are very low on average (~0.2), and far lower than the population of CA1 having repeating fields across the same vs different visual/tactile cue conditions. The authors should define the chance level of ratemap correlation by shuffling neuron identities. Apologies if this is indeed the current approach, but it seems not to be (I was left a bit lost by the description in the methods). For any population of hippocampal place cells, the cross-neuron correlations of firing rate maps are typically not zero, and correlations at 0.2 would normally be evidence for remapping.

(4) A somewhat picky point here that is not meant to claim that multi-compartment studies are not useful - the introduction states that real-world environments typically consist of multi-compartment rooms. This is certainly not true for rodents and is only sometimes true in humans.

(5) The discussion lacks a consideration of how such dRSC output might impact the target structures of dRSC.

(6) The discussion speaks to the idea that multi-directional neurons may aid in transitioning between contexts (sub-compartments). But it is notable that none of the multidirectional neurons have multi-directional tuning in all sub-compartments, but such firing was seen in the 2017 Nature Neuroscience study by Jacob/Jeffery. The discussion should address this difference and perhaps posit a means by which the firing of global and local head direction neurons can be related to each other to yield navigation that depends on both scales.

(7) The authors should provide the size of the smoothing function for spatial firing rate maps.

(8) The authors should devise a measure to define directional tuning in 4 directions (with 90-degree offsets).

(9) Figures 2D and 2H - The offsets in preferred tuning across sub-compartments are rather variable.

eLife Assessment

This important study substantially advances the imaging toolbox available to neuroscientists by presenting a tunable Bessel (tBessel-TPFM) platform that enables high-speed volumetric two-photon imaging. The evidence supporting the novel methodology is convincing, with rigorous benchmarking and demonstrations of a wide range of neuroimaging applications covering vascular dynamics, neurovascular coupling, optogenetic perturbation, and microglial responses. The work will be of broad interest to neuroscientists and imaging system tool developers.

Reviewer #1 (Public review):

Summary:

This manuscript presents a tunable Bessel-beam two-photon fluorescence microscopy (tBessel-TPFM) platform that enables high-speed volumetric imaging with stable axial focus. The work is technically strong and broadly significant, as it substantially improves the flexibility and practicality of Bessel-beam-based two-photon microscopy. The demonstrations are generally strong and bridge a wide range of neuroimaging applications, namely vascular dynamics, neurovascular coupling, optogenetic perturbation, and microglial responses. These convincingly show that the approach enables biological measurements that are difficult or impractical with existing methods.

The evidence supporting the technical and biological claims is generally strong. The optical design is carefully motivated, clearly described, and validated through a combination of simulations and experimental characterization. The biological applications are diverse and well chosen to highlight the strengths of the proposed method, and the data are of high quality, with appropriate controls and comparative measurements where relevant.

Strengths:

(1) The optical innovation addresses a well-recognized limitation of existing Bessel-TPFM implementations, namely axial focus drift during tuning, and does so using a relatively simple, light-efficient, and cost-effective design.

(2) The manuscript provides convincing experimental evidence for this being a versatile platform to map flow dynamics across diverse vessel sizes and orientations in both healthy and pathological states.

(3) Biological demonstrations are comprehensive and span multiple domains such as hemodynamics, neurovascular coupling, and neuroimmune responses.

(4) Quantitative analyses of blood flow across vessel sizes and orientations, including kilohertz line scanning, are particularly compelling and clearly beyond the reach of standard Gaussian TPFM.

(5) Particular advantages are that higher blood slow speeds become measurable up to 23mm/sec (20x more than conventional frame scanning), and that simultaneous (Bessel-)imaging and (Gaussian-)perturbation are possible because of the stable axial focus.

Weaknesses:

(1) At present, the paper does not properly position the new Bessel-beam method against previous work, and fails to compare it to alternative fast volumetric imaging methods without Bessel beams.

(2) The cost-effectiveness of the proposed method is not well described or supported by evidence; it would be useful to include more detail or remove this claim.

(3) Some biological conclusions, e.g., regarding novel features of microglial dynamics (i.e., the observed two-wave responses and coordinated extension-retraction), are based on relatively limited sample size and would benefit from clearer discussion of variability across animals and fields of view.

(4) The use of neural network-based denoising for microglial imaging is reasonable but introduces potential concerns about trustworthiness; additional clarification of validation or failure modes would strengthen confidence in these results.

To conclude, most of the authors' claims are well supported by the data. The central conclusion, namely that tBessel-TPFM provides tunable volumetric imaging enabling experiments not feasible with existing two-photon approaches, is justified. Some biological interpretations would benefit from a more cautious framing, but they do not undermine the main technical and methodological contributions of the study. This is a strong and technically rigorous manuscript that makes a substantial methodological advance with clear relevance to neuroscience and intravital imaging. Minor clarifications and a slightly more measured discussion of certain biological findings are recommended.

Reviewer #2 (Public review):

Summary:

The authors describe a tunable Bessel beam two-photon microscope (tBessel-TPFM) designed to overcome a common limitation of Bessel-based volumetric imaging: axial shifts of the effective focus during Bessel beam parameter tuning. Their optical design allows independent control of axial beam length and resolution while keeping the axial center fixed. This is extensively validated through simulations and experiments.

Strengths:

A major strength of the work is the breadth of validation combined with the level of technical detail provided. The authors carefully characterize the optical performance of the system and clearly explain the design choices and underlying derivations, which will make it easier for others to understand and implement. The authors demonstrate the utility of the method across several in vivo applications, including neurovascular imaging, blood flow measurements, optogenetic stimulation, and microglial dynamics.

Weaknesses:

In the in vivo demonstrations, the authors employ different Bessel beam configurations across experiments, but the beam parameters are not dynamically tuned during live imaging. A video example showing continuous or interactive tuning of the Bessel beam within a single in vivo imaging sequence would further highlight the practical advantages of this platform and strengthen the case for its potential applications. In addition, while excitation powers are reported, the manuscript does not place these values in the broader context of known photodamage thresholds for two-photon microscopy, which would be helpful to the readers. Denoising/image restoration are applied in one of the in vivo examples, but it is unclear why this step was used specifically for this dataset and whether it was necessary to achieve adequate SNR or primarily included as an additional demonstration.

Reviewer #3 (Public review):

Summary:

The manuscript presents an elegant and cost-effective approach for generating a tunable Bessel beam on a conventional two-photon microscope. The authors assemble a compact optical module comprising three axicons and a series of lenses that permits rapid adjustment of both lateral resolution and axial extent without modifying the focal plane. This flexibility enables the system to be readily adapted to a variety of biological preparations. As a proof of concept, the authors employ the device to record blood flow velocities in cortical microcapillaries, arterioles, and venules, thereby directly visualizing vasodilatation and vasoconstriction dynamics and permitting quantitative analysis of neurovascular coupling across cortical layers in awake mice.

The authors demonstrate that the tunability of the Bessel beam can be exploited to match the numerical aperture to the vessel type: a high NA configuration, albeit slower scan, is optimal for resolving flow in capillaries, whereas a low NA setting provides faster acquisition suitable for arterioles and venules. By implementing a one-dimensional line scan with the Bessel beam, they achieve an imaging speed that is twentyfold faster than conventional frame-by-frame scanning, which proves sufficient to capture hemodynamic transients before and after an induced ischemic stroke.

In addition to pure observation, the authors integrate a co-propagating Gaussian line to the system, allowing simultaneous imaging and photostimulation within the same focal plane. This capability addresses a common limitation of other Bessel beam implementations, in which the observation and perturbation planes often become misaligned when the Bessel beam is altered. The manuscript also emphasizes the advantage of Bessel beam excitation for calcium imaging after a perturbation, because it captures neuronal activity in planes both above and below the nominal focal plane, signals that would be missed with a standard Gaussian focus. Finally, the authors apply the technique to investigate the neuroimmune response following targeted microglial ablation; they report that adjacent microglia extend processes toward the injury site while retracting processes in the opposite direction.

Overall, the work offers a technically straightforward yet powerful extension to existing two-photon platforms, providing high-speed, volumetric imaging and stimulation capabilities that are well-suited to a broad range of neurovascular and neuroimmune studies. The experimental validation is quite thorough, and the presented data convincingly illustrates the benefits of the approach.

Strengths:

The authors present a truly clever and inexpensive optical module that can be integrated into almost any two-photon microscope, providing a tunable Bessel beam with a minimal modification of the existing system. The experimental data and accompanying quantitative analysis convincingly demonstrate that the system can reveal physiological events, such as capillary flow, calcium transients across multiple axial planes, and microglial process dynamics, that are difficult or impossible to capture with a conventional Gaussian beam. The breadth of experiments chosen for the manuscript illustrates the practical utility of the device and supports the authors' conclusions that it extends the functional repertoire of standard two-photon microscopy.

Weaknesses:

The manuscript would benefit from a more detailed contextualisation of the claimed speed advantage. Although the authors mention other techniques in the introduction, they do not provide any direct comparison with other state-of-the-art high-speed two-photon approaches such as light beads microscopy (Demas et al., Nat. Methods 2021), temporal multiplexing schemes (Weisenburger et al., Cell 2019), or random access microscopy (Villette et al., Cell 2019). A brief comparison of imaging speed, spatial resolution, and instrumental complexity would enable readers to assess the relative merits of the present method.

A second limitation that warrants discussion is the inherent trade off between volumetric coverage and image specificity. Because the Bessel beam excites fluorescence throughout an extended axial range, the detector inevitably integrates signal from a three dimensional volume into a two dimensional image. In densely labelled tissue, this can lead to significant signal crosstalk, reducing contrast and complicating quantitative interpretation. A brief analysis of how labeling density affects the fidelity of flow or calcium measurements, or suggestions for mitigating crosstalk (e.g., computational deconvolution, adaptive excitation shaping, or combinatorial sparse labeling), would broaden the applicability of the technique.

eLife Assessment

The study investigates, from multiple angles, the still-debated function of insect rhodopsin-7 (Rh7). The authors present compelling results for its ancient phylogenetic origin across pan-arthropods, a non-visual role based on expression analyses in the fly brain, an unusual G-protein signalling pathway, and - using behavioural genetics - that Rh7 affects how Drosophila melanogaster interprets and responds to light-dark transitions. Through this, the work provides fundamental new insights into the evolution and function of non-visual opsins.

Reviewer #1 (Public review):

Summary:

The study investigates the Drosophila non-visual light receptor rhodopsin7 with regard to its role in light information processing and resulting consequences for behavioral patterns and circadian clock function. Using behavioral, in situ staining, and receptor activation assays together with different fly mutants, the authors show that rhodopsin7 is an important determinant of activity under and response to darkness, which likely signals via a pathway distinct from other, visual Drosophila rhodopsins. Based on phylogenetic analysis, the authors further discuss a potentially conserved functional role of non-visual photoreceptors like rhodopsin7 and the mammalian melanopsin light information processing and circadian clock modulation.

Strengths:

The manuscript follows a very clear structure with all investigations logically building onto each other. Background information and methodology are provided in appropriate detail so that readers can fully understand why and how experiments were conducted. It is further praiseworthy that the authors provide the details that allow also non-experts in the field to fully understand their approaches. Experimental work was conducted in a highly standardized manner, and also considered potential "side-aspects" like the consequences of temperature cycles and changed photoperiods. The detailed and clear description of the obtained results makes them very convincing, with (almost) all observable patterns being addressed.

By highlighting the evolutionary old phylogenetic position of rhodopsin7 and its conservation across numerous clades, the authors provide strong reasoning for the relevance of their work, also pointing out the similarities to the mammalian melanopsin. The postulated hypothesis regarding protein structure and functioning, as well as the role in light information processing and behavioral and circadian clock modulation are well based on the authors' observations, and speculative aspects are correctly pointed out.

Weaknesses:

Where the manuscript still has potential for improvement is the discussion, which in its current form does seem slightly self-contained and does not fully integrate the findings of previous studies on Drosophila rhodopsin7. As the introduction specifically points out that previous findings have been contradictory, this seems like a missed opportunity. Further details on this are provided in the recommendations below.

Similarly, the manuscript currently lacks a discussion of the possible relevance of rhodopsin7 (and other non-visual light receptors in other organisms) in the context of a species' environment and lifestyle, i.e., what is the relevance/benefit of having rhodopsin7 in the fly's everyday life? While this clearly involves speculation, when done carefully, it can elevate the paper's relevance from a primarily academic to a societal one.

An additional point concerns the title and abstract, which postulate rhodopsin7 roles in contrast vision as well as motion and brightness perception. Contrast remains poorly defined in the text, leaving it ambiguous whether it refers to bright/dark contrasts, e.g., along edges, or the temporal contrast that results from dark pulses (startle response). While the latter seems to apply here, the former is likely more intuitive. Thus, this aspect should be rephrased (also in the title) or properly clarified early on. Regarding motion detection, this is backed up by the optomotor response results, but the findings stand somewhat isolated from the other results, lacking a clear connection aside from general visual processing. Lastly, brightness perception is mentioned in the abstract, but never again, possibly due to inconsistent phrasing throughout the manuscript.

Reviewer #2 (Public review):

Summary:

This is a very interesting paper bringing new and important information about the poorly understood rhodopsin 7 photoreceptive molecule. The very ancient origin of the gene is revealed in addition to data supporting a signaling pathway that is different from the one known for the canonical rhodopsins. Precise expression data, particularly in the optic lobe of the fly, as well as clear behavioral phenotypes in responses to light changes, make this study a strong contribution to the understanding of the still-debated function of rhodopsin 7.

Specific comments

(1) Title and abstract: Contribution of Rh7 to circadian clock regulation

(a) It is not that clear to me what rhodopsin does in terms of circadian regulation (even though its function might be circadianly regulated). The clear role in the light/dark distribution of activity might not be circadian per se, but mostly light/dark-driven, and there is no evidence here for a role in the entrainment of the clock.

(b) The authors should cite Lazopulo, which nicely shows that Rh7 has an important role in peripheral neurons to allow flies to escape from blue light (see below).

(2) Figure 2 C

The finding showing that Galphaz but not Galphaq can trigger signaling from light-excited Rh7 is a very intriguing finding to better understand Rh7 function. Since Galphaz is related to Gi/o, it would be interesting to test those, for example, by expressing RNAi with Rh7-gal4 and testing the Light-dark or light-off response behavior.

(3) Figures 3-4

The change in the locomotor activity distribution between light and dark in LD conditions provides a nice assay for Rh7 function. Since Lazopulo et al. (2019) have shown that wild-type but not Rh7 mutants do escape from blue light, it would be important to compare and discuss these LD behavior data with the Lazopulo results. Precisely, is this nighttime preference linked to blue light?

The expression data are really nice and show that Rh7 is mostly a non-retinal photoreceptor. However, the paper would be strongly reinforced by correlating this with the LD behavior. The LD phenotype should be tested in flies with Rh7 expression rescued under Rh7gal4 control (as done for the startle response). This is important to show whether the expression pattern is likely responsible for the described Rh7 function in LD. If L5 and or M11 drivers are available, they should be used to rescue Rh7? Since expression in some clock neurons is shown, the rescue experiment should also be done with a clock neuron driver.

In the same line, can the LD phenotype (or startle response phenotype of Figure 4) be restored by expressing Rh7 under ppk control, as shown for the blue light avoidance phenotype by Lazopulo et al?

Finally, the Rh7 "darkfly" rescued flies should be tested in LD.

Reviewer #3 (Public review):

Summary:

While our knowledge regarding visual opsins is largely very good, a lot more uncertainty exists around the role of non-visual opsins. Using the power of the Drosophila melanogaster model system, Kirsh et al. investigate the role of the non-visual opsin Rhodopsin7 (Rh7). Expression analysis, based on Rh7-Gal4>UAS-GFP and HRC in situ staining, reveals strong expression in the optic lobes and somewhat weaker, but nevertheless extensive expression in the brain. An investigation of motor activity reveals that loss of function leads to an altered day and night rhythm, specifically decreasing activity during the dark phase. These flies were also less sensitive, but still responsive to a light-induced startle response and showed deficiencies in the optomotor response. To further investigate how Rh7 may modulate these responses, inspired by the Dark line of flies (which were kept in the dark for ~1400 generations) and which has accumulated C-terminal related losses, the authors conducted rescues with an intact and a C-terminal-deficient Rh7 and were able to pinpoint that region as an important driver of related behavioral shifts. These findings are particularly intriguing as Rh7 represents an ancient opsin with phylogenetic and mechanistic parallels to mammalian melanopsin.

Strengths:

The paper is well-written and contains high-quality data with appropriate sample sizes, and the conclusions are well supported.

Weaknesses:

No weaknesses were identified by this reviewer, but the following recommendations are made:

(1) The authors should clarify exactly what tissues were taken for the comparative qPCR. This is particularly interesting in terms of the retina. Since Rh7 appears not to be expressed within the photoreceptor cells of the retina, this raises the important question as to which cells it is expressed in. To address this important question, it would also be helpful to include an expression analysis of the retina itself (by extending the RH7-GFP expression patterns and/or adding HCR in situ of the ommatidia array). The cell types of the retina are very well classified, and some evidence already exists for Rh7 expression in support cells (e.g., Charlton-Perkins et al., (2017); PMID: 28562601). This study has a unique opportunity to investigate this further by adding these critical data for a more complete picture of Rh7.

(2) Mammalian opsins should be included in the phylogenetic analysis illustrated in Figure 2A and indicate their position on the tree. This will allow readers to better put the authors' statements regarding the intermediate position of Rh7 into perspective. In addition, note that the distinction between red and deep red is easy to miss regarding the Rh7 cluster. Perhaps the authors could use a more distinct colour scheme, for example, orange and deep red.

(3) More details should be provided on the optomotor response experiments. Specifically, specifications of the frequencies used for the optomotor response are needed. Results show a relatively large level of variation, which may be due to different angular perspectives that flies may have had while viewing the stimulus. If possible, provide videos as examples, as they will make it clearer to viewers how much flies could move around in the setup (from the methods, it seems they could move within the 2.2 of the 3 cm diameter of the arena, which would lead to substantial differences in the visual angle of the viewed grating.

eLife Assessment

This useful study raises interesting questions but provides inadequate evidence of an association between atovaquone-proguanil use (as well as toxoplasmosis seropositivity) and reduced Alzheimer's dementia risk. The findings are intriguing but they are correlative and hypothesis-generating with the strong possibility of residual confounding.

[Note: The final version has been published in Brain, Behavior, and Immunity: https://doi.org/10.1016/j.bbi.2026.106473]

Author response:

[Note: The final version has been published in Brain, Behavior, and Immunity: https://doi.org/10.1016/j.bbi.2026.106473]

eLife Assessment

Rhis useful study raises interesting questions but provides inadequate evidence of an association between atovaquone-proguanil use (as well as toxoplasmosis seropositivity) and reduced Alzheimer's dementia risk. The findings are intriguing but they are correlative and hypothesis-generating with the strong possibility of residual confounding.

We thank the editors and reviewers for characterizing our work as useful and for the opportunity to publish a Reviewed Preprint with a corresponding response. However, the statements in the Assessment characterizing the evidence as ‘inadequate’ and asserting a ‘strong possibility of residual confounding’ are factually incorrect as applied to our data and incompatible with the empirical findings presented in the manuscript. We have notified the editors of this factual inaccuracy. As the Assessment will be published as originally written, we provide clarification here to ensure an accurate scientific record for readers of the Reviewed Preprint.

Our study shows that the association between atovaquone–proguanil (A/P) exposure and reduced dementia risk, first identified in a rigorously matched national cohort in Israel, is robustly reproduced across three independently constructed age-stratified cohorts in the U.S. TriNetX network (with exposure at ages 50–59, 60–69, and 70–79). In each cohort, individuals exposed to A/P were compared with rigorously matched individuals who received another medication at the same age and were then followed over a decade for incident dementia. Cases and controls were matched on all major established dementia risk factors: age, sex, race/ethnicity, diabetes, hypertension, obesity, and smoking status.

Across all three strata, each containing more than 10,000 exposed individuals with an equal number of matched controls, we observed substantial and consistent reductions in cumulative dementia incidence (HR 0.34–0.51), extremely low P-values (10^–16 to 10^–40), and continuously widening divergence of Kaplan–Meier curves over the follow-up period. To more rigorously exclude the possibility of unmeasured baseline differences in health status, we additionally performed, for the purpose of this response, comparative analyses of key indicators of frailty and clinical utilization, including emergency and inpatient encounters, as well as the prevalence of mild cognitive impairment prior to medication exposure (values provided below in response to Reviewer #2, Weakness 1). These analyses provide clear evidence showing no pattern suggestive of exposed individuals being medically or cognitively healthier at baseline.

Taken together, these findings constitute a rigorously matched and independently replicated association across two national health systems, using TriNetX, the most widely cited real-world evidence platform in published cohort studies. Replication across three age strata, each with >10,000 exposed individuals, followed for a decade, and matched on all major known risk factors for dementia, meets the accepted epidemiologic definition of strong and reproducible evidence.

Although we disagree with elements of the editorial Assessment that appear inconsistent with the empirical findings, we will proceed with publication of the current manuscript as a Reviewed Preprint in order to ensure timely dissemination of findings with meaningful implications for public health and dementia prevention. In this initial public version, the point-by-point responses below provide concise explanations addressing the critiques underlying the Assessment. A revised manuscript, incorporating expanded baseline comparisons across each TriNetX age stratum, additional stringent exclusions, and an expanded discussion that will address the remarks presented in this review, will be submitted shortly.

Reviewer #1 (Public review):

Summary:

This useful study provides incomplete evidence of an association between atovaquone-proguanil use (as well as toxoplasmosis seropositivity) and reduced Alzheimer's dementia risk. The study reinforces findings that VZ vaccine lowers AD risk and suggests that this vaccine may be an effect modifier of A-P's protective effect. Strengths of the study include two extremely large cohorts, including a massive validation cohort in the US. Statistical analyses are sound, and the effect sizes are significant and meaningful. The CI curves are certainly impressive.

Weaknesses include the inability to control for potentially important confounding variables. In my view, the findings are intriguing but remain correlative / hypothesis generating rather than causative. Significant mechanistic work needs to be done to link interventions which limit the impact of Toxoplasmosis and VZV reactivation on AD.

We thank the reviewer for describing our study as useful and for highlighting several of its strengths, including the very large cohorts, sound statistical analyses, meaningful effect sizes, and the impressive CI curves. We also appreciate the reviewer’s recognition that our findings reinforce prior evidence linking VZV vaccination to reduced AD risk.

Regarding the statement that the evidence remains incomplete due to “inability to control for potentially important confounding variables,” we refer to our introductory explanation above. As noted there, our analyses meet the accepted criteria for reproducible epidemiological evidence, and the assumption of uncontrolled confounding is contradicted by rigorous matching and by additional baseline evaluations. We fully agree that mechanistic work is warranted, and our epidemiologic findings strongly motivate such efforts.

We address the reviewer’s specific comments in detail below.

(1) Most of the individuals in the study received A-P for malaria prophylaxis as it is not first line for Toxo treatment. Many (probably most) of these individuals were likely to be Toxo negative (~15% seropositive in the US), thereby eliminating a potential benefit of the drug in most people in the cohort. Finally, A-P is not a first line treatment for Toxo because of lower efficacy.

We agree that individuals in our cohort received Atovaquone-Proguanil (A-P) for malaria prophylaxis rather than for treatment of toxoplasmosis. However, this does not contradict our interpretation. Because latent CNS colonization by T. gondii is not currently considered clinically actionable, asymptomatic carriers are not offered treatment, and therefore would only receive an anti-Toxoplasma regimen unintentionally, through a medication prescribed for another indication such as malaria prophylaxis. Importantly, atovaquone is an established therapy for toxoplasmosis, including CNS disease, with documented efficacy and CNS penetration in current treatment guidelines. It is therefore reasonable to assume that, during the multi-week course typically administered for malaria prophylaxis, A-P would exert significant anti-Toxoplasma activity in individuals with latent CNS infection, potentially reducing or eliminating parasite burden even though the medication was not prescribed for that purpose.

The reviewer notes that only ~15% of individuals in the U.S. are Toxoplasma-seropositive, based on surveys performed primarily in young adults of reproductive age (serologic testing is most commonly obtained in women during prenatal care). However, seropositivity increases cumulatively over the lifespan, and few reliable estimates exist for the age groups in which Alzheimer’s disease and dementia occur. Even if we accept the lower estimate of ~15% latent colonization in older adults, this proportion is still smaller than the lifetime cumulative incidence of dementia in the general population.

Therefore, if latent toxoplasmosis contributes causally to dementia risk, and A-P is capable of eliminating latent Toxoplasma in the subset of individuals who harbor it, then a multi-week course of treatment—such as the one routinely taken for malaria prophylaxis—would be expected to produce a substantial reduction in dementia incidence at the population level, of the same order of magnitude reported here. A protective effect concentrated in a minority of exposed individuals is fully compatible with, and can mechanistically explain, the large overall reduction in risk that we observe.

Finally, the reviewer notes that A-P is not a first-line treatment for toxoplasmosis due to assumed lower efficacy. This point does not undermine our results. Even a second-line agent, when administered over several weeks—as is routinely done for malaria prophylaxis—is expected to exert substantial anti-Toxoplasma activity. The long duration of exposure in large populations receiving A-P for travel provides a unique natural experiment that does not exist for other anti-Toxoplasma medications, which, when prescribed for their non-Toxoplasma indications, are not taken more than a few days. Thus, the widespread use of A-P for malaria prophylaxis allows a unique opportunity to evaluate long-term outcomes following inadvertent anti-Toxoplasma treatment.

Moreover, “first line” recommendations in clinical guidelines refer to treatment of acute toxoplasmosis in immunosuppressed individuals, where tachyzoites are actively replicating. These guidelines do not consider efficacy against latent CNS colonization, which is dominated by bradyzoites, a biologically distinct form, in immunocompetent individuals. Therefore, the guideline hierarchy is not informative regarding which medication is more effective at clearing latent brain infection, the stage we consider most relevant to dementia risk.

(2) A-P exposure may be a marker of subtle demographic features not captured in the dataset such as wealth allowing for global travel and/or genetic predisposition to AD. This raises my suspicion of correlative rather than casual relationships between A-P exposure and AD reduction. The size of the cohort does not eliminate this issue, but rather narrows confidence intervals around potentially misleading odds ratios which have not been adjusted for the multitude of other variables driving incident AD.

We agree that prior to matching, A-P exposure may be associated with demographic features such as health or to travel internationally. However, this does not apply after matching. In all age-stratified analyses, exposed and control individuals were rigorously matched on all major risk factors known to influence dementia risk, including age, sex, race/ethnicity, smoking status, hypertension, diabetes, and obesity. Owing to the extremely large pool of individuals in TriNetX (~120M), our matching was performed stringently, producing exposed and unexposed cohorts that are near-identical with respect to the established determinants of dementia risk.

The reviewer correctly identifies that large cohorts alone do not eliminate confounding; however, confounding must still be biologically and epidemiologically plausible. Any hypothetical confounder capable of producing a 50–70% reduction in dementia incidence over a decade would need to: (1) produce a very large protective effect against dementia; (2) be strongly associated with A-P exposure; and (3) remain entirely uncorrelated with age, sex, race/ethnicity, smoking, diabetes, hypertension and obesity, which have been rigorously matched. No such factor has been proposed. The suggestion that an unspecified ‘subtle demographic feature’ could produce effects of this magnitude remains hypothetical, and no such factor has been described in the dementia risk literature.

If a specific evidence-supported confounder is proposed that meets these criteria, we would be pleased to test it empirically in our cohorts. In the absence of such a proposal, the interpretation that the association is merely “correlative rather than causal” remains speculative and does not negate the strength of a replicated, rigorously matched, long-term association across large cohorts in two national health systems.

(3) The relationship between herpes virus reactivation and Toxo reactivation seems speculative.

We respectfully disagree with the characterization of the herpesvirus–Toxoplasma interaction as speculative. The mechanism we describe is biologically valid, based on established virology and parasitology literature showing that latent T. gondii infection can reactivate from its bradyzoite state under inflammatory or immune-modifying conditions, including viral triggers. A published clinical report has documented CNS co-reactivation of T. gondii and a herpesvirus, explicitly noting that HHV-6 reactivation can promote Toxoplasma reactivation in neural tissue (Chaupis et al., Int J Infect Dis, 2016).

Moreover, this mechanism is the only currently evidence-supported explanation that simultaneously and parsimoniously accounts for all of the epidemiologic observations in our study:

(1) Substantially higher cumulative incidence of dementia in individuals with positive Toxoplasma serology, indicating that latent infection is a risk factor for subsequent cognitive decline;

(2) Strong protective association following A-P exposure, a medication with established activity against Toxoplasma gondii, including in the CNS;

(3) Independent protection conferred by VZV vaccination, observed consistently for two vaccines with distinct formulations (one live attenuated, one recombinant protein), whose only shared property is suppression of VZV reactivation;

(4) Greater protective effect of A-P among individuals who were not vaccinated against VZV, consistent with a model in which dementia risk requires both herpesvirus reactivation and persistent latent Toxoplasma infection—such that reducing either factor alone (via VZV vaccination or anti-Toxoplasma suppression) substantially lowers risk.

Taken together, these observations are difficult to reconcile under any alternative hypothesis.  

To date, we are unaware of any other biologically coherent mechanism that can explain all four findings simultaneously. We would welcome any alternative explanation capable of accounting for these converging epidemiologic signals, as such a proposal could meaningfully advance the scientific discussion. In the absence of a competing explanation, the interaction between latent toxoplasmosis and herpesvirus reactivation remains the most parsimonious hypothesis supported by current knowledge.

Finally, while observational studies are inherently limited in their ability to provide causal inference, the mechanism we propose is biologically grounded and experimentally testable. Our results provide a strong rationale for mechanistic studies and clinical trials, and warrant publication precisely because they generate a verifiable hypothesis that can now be evaluated directly.

(4) A direct effect on A-P on AD lesions independent on infection is not considered as a hypothesis. Given the limitations above and effects on metabolic pathways, it probably should be. The Toxo hypothesis would be more convincing if the authors could demonstrate an enhanced effect of the drug in Toxo positive individuals without no effect in Toxo negative individuals.

A direct effect of A-P on AD established lesions is indeed possible, and this hypothesis would be of significant therapeutic interest. However, we did not consider it within the scope of our epidemiologic analyses because all cohorts explicitly excluded individuals with existing dementia. Under these conditions, proposing a disease-modifying effect on established Alzheimer’s lesions based on our data would itself be speculative. Evaluating such a mechanism would be better answered by mechanistic or interventional studies rather than inference from populations without baseline disease.

We also agree that demonstrating a stronger protective effect among Toxoplasma-positive individuals would be informative. Unfortunately, this “natural experiment” cannot be performed using the available data: Toxoplasma serology is rarely ordered in older adults, and A-P exposure is itself uncommon, resulting in a cohort overlap far too small to yield valid statistical inference (n≈25 in TriNetX).

Thus, while both proposed hypotheses are scientifically attractive and merit further study, neither can be resolved using currently available real-world clinical data. Our findings provide the rationale to investigate both hypotheses experimentally, and we hope our report will motivate such studies.

Reviewer #2 (Public review):

Summary:

This manuscript examines the association between atovaquone/proguanil use, zoster vaccination, toxoplasmosis serostatus and Alzheimer's Disease, using 2 databases of claims data. The manuscript is well written and concise. The major concerns about the manuscript center around the indications of atovaquone/proguanil use, which would not typically be active against toxoplasmosis at doses given, and the lack of control for potential confounders in the analysis.

Strengths:

(1) Use of 2 databases of claims data.

(2) Unbiased review of medications associated with AD, which identified zoster vaccination associated with decreased risk of AD, replicating findings from other studies.

We thank the reviewer for the thoughtful assessment and for noting key strengths of our work, including (1) the use of two large national databases, and (2) the unbiased discovery approach that replicated the widely reported association between zoster vaccination and reduced Alzheimer’s disease (AD) risk. We agree that these features highlight the validity and reproducibility of the analytic framework.

Below we respond to the reviewer’s perceived weaknesses.

Weaknesses:

(1) Given that atovaquone/proguanil is likely to be given to a healthy population who is able to travel, concern that there are unmeasured confounders driving the association.

We agree that, prior to matching, A-P exposure may correlate with demographic or health-related differences (e.g., ability to travel). However, this potential bias was explicitly controlled for in the study design. Across all three age-stratified TriNetX cohorts, exposed and unexposed individuals were rigorously matched on all major established dementia risk factors: age, sex, race/ethnicity, smoking status, obesity, diabetes mellitus, and hypertension. Comparative analyses confirm that these risk factors are equivalently distributed at baseline.

As noted in our response to Reviewer #1, for any hypothetical unmeasured confounder to explain the results, it would need to satisfy three conditions simultaneously:

(1) Be capable of producing a 50–70% reduction in dementia incidence sustained over a decade and across three distinct age strata (ages 50–79);

(2) Be strongly associated with likelihood of receiving A-P;

(3) Remain entirely uncorrelated with age, sex, race/ethnicity, smoking, diabetes, hypertension, or obesity, all of which were rigorously matched and balanced at baseline.

No such factor has been proposed in the literature or by the reviewer. Thus, the concern remains hypothetical and unsupported by any measurable demographic or biological mechanism.

Importantly, empirical evidence contradicts the notion of a “healthy traveler” bias:

Emergency and inpatient encounter rates prior to exposure were comparable between A-P users and controls. Across the three age-stratified cohorts, emergency visits were similar or slightly higher among A-P users (EMER: 19.6% vs 16.4%, 19.9% vs 14.2%, 22.0% vs 14.8%), and inpatient encounters were effectively equivalent (IMP: 14.8% vs 15.2%, 17.7% vs 17.6%, 22.1% vs 22.2%). These patterns directly contradict the suggestion that A-P users were a healthier or less medically burdened population at baseline.

Prevalence of mild cognitive impairment was not lower among A-P users and was, in fact, slightly higher in the oldest cohort. Across the three age groups, baseline diagnoses of mild cognitive impairment (MCI) were comparable or slightly higher among exposed individuals (0.1% vs 0.1%, 0.3% vs 0.2%, 1.1% vs 0.6%). These data contradict the suggestion that A-P users had superior baseline cognition.

The strongest protective association occurred in the youngest stratum (age 50–59; HR 0.34). At this age, when nearly all individuals are sufficiently healthy to travel internationally, A-P uptake is the least likely to confound health status. A frailty-based “healthy traveler” hypothesis would instead predict the opposite pattern, with older adults showing the greatest apparent benefit, since health limitations are more likely to restrict travel in later life. In contrast, the protective association weakens with increasing age, empirically contradicting any explanation based on differential travel capacity.

In conclusion, the empirical evidence directly contradicts the existence of a ‘healthy traveler’ effect.

(2) The dose of atovaquone in atovaquone/proguanil is unlikely to be adequate suppression of toxo (much less for treatment/elimination of toxo), raising questions about the mechanism.

A few important points should address the reviewer’s concern:

In our cohorts, A-P was prescribed for malaria prophylaxis, as correctly noted. In this setting, it is taken for the entire duration of travel, plus several days before and after, typically resulting in many weeks of continuous exposure. This creates an unintentional but scientifically valuable natural experiment, in which a CNS-penetrating anti-Toxoplasma agent is administered for long durations.

Atovaquone is an established treatment for CNS toxoplasmosis, has strong CNS penetration, and is included in current clinical guidelines for acute toxoplasmosis in immunocompromised patients, although at higher doses. Because latent, asymptomatic CNS colonization is not treated in clinical practice, there are currently no data establishing the dose required to eliminate bradyzoite-stage Toxoplasma in immunocompetent individuals.

Our observations concern atovaquone–proguanil (A-P), a fixed-dose combination of atovaquone with proguanil, a DHFR inhibitor targeting a key metabolic pathway shared by malaria parasites and T. gondii. The combination has well-established synergistic effects in malaria prophylaxis and the same mechanism would be expected to enhance anti-Toxoplasma activity. This fixed-dose regimen has never been formally evaluated for toxoplasmosis treatment at prolonged durations or against latent bradyzoite infection.

Our hypothesis does not require or imply complete eradication of Toxoplasma. A clinically meaningful reduction in latent cyst burden among the subset of colonized individuals may be sufficient to alter long-term disease trajectories. Thus, a population-level decrease in dementia incidence does not require universal clearance of infection, but only partial suppression or reduction of parasite load in susceptible individuals, which is entirely compatible with the known pharmacology and duration of A-P exposure.

(3) Unmeasured bias in the small number of people who had toxoplasma serology in the TriNetX cohort.

The relatively small number of older adults with Toxoplasma serology stems from current clinical practice: serologic testing is mostly performed in women during reproductive years due to risks in pregnancy, whereas in older adults a positive result has no clinical consequence and therefore testing is rarely ordered.

Importantly, the seropositive and seronegative groups were drawn from the same underlying population of individuals who underwent serology testing, and the only difference between groups is the test result itself. Because the decision to order a test is made prior to and independent of the result, there is no plausible rationale by which the serology outcome (positive or negative) would introduce a bias favoring either group beyond the result of the test itself.

Furthermore, the two groups were here also rigorously matched on all major dementia risk factors, including age, sex, race/ethnicity, smoking, diabetes, hypertension, and BMI, and these characteristics are similarly distributed between groups. A small sample size does not imply bias; it simply reduces statistical power. Despite this limitation, the observed association (HR = 2.43, p = 0.001) remains strongly significant.

Finally, this result is consistent with multiple published studies reporting higher rates of Toxoplasma seropositivity among individuals with Alzheimer’s disease, dementia, and even mild cognitive impairment, such that our finding reinforces a broader and independently observed epidemiologic pattern. Importantly, in our cohort the serology testing clearly preceded dementia diagnosis, which supports the plausibility of a causal rather than merely correlative relationship between latent toxoplasmosis and cognitive decline.

To conclude our provisional response, we thank the editor and reviewers for raising points that will be further addressed and expanded upon in the discussion of the forthcoming revision. We welcome transparent scientific dialogue and acknowledge that, as with all observational research, residual confounding cannot be eliminated with absolute certainty. However, we disagree with the overall Assessment and emphasize that our findings—reproduced independently across two national health systems and three age-stratified cohorts, each rigorously matched on all major determinants of dementia risk, meet, and in many respects exceed, current standards for high-quality observational evidence.

Assigning the results to “residual confounding” requires more than speculation: it requires identification of a confounding factor that is (1) anchored in established dementia risk literature, (2) empirically plausible, and (3) quantitatively capable of generating a sustained ~50 percent reduction in dementia incidence over a decade. No such factor has been identified to date. We note that the assertion of “residual confounding” has not been supported by a specific, quantitatively plausible mechanism. A hypothetical bias that is both extremely large in effect and uncorrelated with all major risk factors is not statistically or biologically credible.

The explanation we propose, reduction in dementia risk through elimination of latent Toxoplasma gondii, is biologically grounded, directly supported by independent epidemiologic literature, and uniquely capable of accounting for all convergent observations in our data. No alternative hypothesis has been put forward that can plausibly explain these findings.

A revised version of the manuscript will be submitted shortly, incorporating expanded baseline analyses, with the strictest possible exclusion criteria (including congenital, vascular, chromosomal, and neurodegenerative disorders such as Parkinson’s disease), and complete tabulated comparisons. These data will further reinforce that the observed protective associations are not attributable to any measurable confounding. We also plan to enhance the discussion in order to address the points raised by the reviewers.

In light of the expanded analyses, any reservations expressed in the initial Assessment can now be re-evaluated on the basis of the empirical evidence. The findings reported in our study meet, and in several respects exceed, current epidemiologic standards for high-quality observational research, clearly warrant publication, and provide a robust scientific foundation for future mechanistic and interventional studies to determine whether elimination of latent toxoplasmosis can prevent or treat dementia.

Reviewer #1 (Public review):

Summary:

This useful study provides incomplete evidence of an association between atovaquone-proguanil use (as well as toxoplasmosis seropositivity) and reduced Alzheimer's dementia risk. The study reinforces findings that VZ vaccine lowers AD risk and suggests that this vaccine may be an effect modifier of A-P's protective effect. Strengths of the study include two extremely large cohorts, including a massive validation cohort in the US. Statistical analyses are sound, and the effect sizes are significant and meaningful. The CI curves are certainly impressive.

Weaknesses include the inability to control for potentially important confounding variables. In my view, the findings are intriguing but remain correlative / hypothesis generating rather than causative. Significant mechanistic work needs to be done to link interventions which limit the impact of Toxoplasmosis and VZV reactivation on AD.

Weaknesses:

Major:

(1) Most of the individuals in the study received A-P for malaria prophylaxis as it is not first line for Toxo treatment. Many (probably most) of these individuals were likely to be Toxo negative (~15% seropositive in the US), thereby eliminating a potential benefit of the drug in most people in the cohort. Finally, A-P is not a first line treatment for Toxo because of lower efficacy.

(2) A-P exposure may be a marker of subtle demographic features not captured in the dataset such as wealth allowing for global travel and/or genetic predisposition to AD. This raises my suspicion of correlative rather than casual relationships between A-P exposure and AD reduction. The size of the cohort does not eliminate this issue, but rather narrows confidence intervals around potentially misleading odds ratios which have not been adjusted for the multitude of other variables driving incident AD.

(3) The relationship between herpes virus reactivation and Toxo reactivation seems speculative.

(4) A direct effect on A-P on AD lesions independent on infection is not considered as a hypothesis. Given the limitations above and effects on metabolic pathways, it probably should be. The Toxo hypothesis would be more convincing if the authors could demonstrate an enhanced effect of the drug in Toxo positive individuals without no effect in Toxo negative individuals.

Minor:

(5) "Clinically meaningful" should be eliminated from the discussion given that this is correlative evidence.

Reviewer #2 (Public review):

Summary:

This manuscript examines the association between atovaquone/proguanil use, zoster vaccination, toxoplasmosis serostatus and Alzheimer's Disease, using 2 databases of claims data. The manuscript is well written and concise. The major concerns about the manuscript center around the indications of atovaquone/proguanil use, which would not typically be active against toxoplasmosis at doses given, and the lack of control for potential confounders in the analysis.

Strengths:

(1) Use of 2 databases of claims data.

(2) Unbiased review of medications associated with AD, which identified zoster vaccination associated with decreased risk of AD, replicating findings from other studies.

Weaknesses:

(1) Given that atovaquone/proguanil is likely to be given to a healthy population who is able to travel, concern that there are unmeasured confounders driving the association.

(2) The dose of atovaquone in atovaquone/proguanil is unlikely to be adequate suppression of toxo (much less for treatment/elimination of toxo), raising questions about the mechanism.

(3) Unmeasured bias in the small number of people who had toxoplasma serology in the TriNetX cohort.

eLife Assessment

This important work establishes a connection between PRMT1 and SFPQ by identifying common phenotypes downstream of their inactivation. In the resubmission, authors now include NMD as a contributor to aberrant gene expression underpinning craniofacial development. The complementary experiments help strengthen some solid conclusions. This paper describes an interesting mechanism for the regulation of RNA levels, which is of interest to the readers of eLife.

Reviewer #1 (Public review):

The current manuscript investigates a regulatory axis containing Prmt1, which methylates RNA binding proteins and alters intron splicing outcomes and expression of matrix genes. Authors test the effects of deficient Prmt1, Sfpq, and various other factors, using a combination of bioinformatic analyses and wet-lab validation approaches. Authors show that intron retention often triggers NMD, contributing to aberrant gene expression regulation and craniofacial development. The revised manuscript introduces several complementary experiments that help to strengthen conclusions. For example, authors directly investigate NMD-mediated transcript turnover to better understand how retention contributes to expression changes in genes of interest, and they assess several additional factors downstream of Prmt1 to justify a centralized interested in the PRMT1/SFPQ axis.

Weaknesses:

However, some points remain unaddressed or unexplored, which could bolster conclusions. For example, the transcriptome data from knockdown experiments indicate robust exon skipping, suggesting that analysis of these patterns in parallel with intron retention could provide additional insights into the responsive gene programs. Given that SFPQ is known to have multiple regulatory roles, a more thorough investigation of its possible mechanisms of action during craniofacial development would allow for definitive conclusions about the isolated impact of SFPQ-dependent splicing. Although authors employ CUT&Tag analysis of Pol II binding at the promoters and across the gene body, at the current scope, no change in Pol II association (i.e., absence of transcriptional repression) does not directly indicate a lack of transcriptional regulation by other means (pause release, elongation rate or processivity, transcription termination, etc.). Without a more thorough investigation of these mechanisms, this confounds definitive claims about their relative contributions to the gene expression landscape.

Reviewer #2 (Public review):

Summary:

The manuscript by Lima et al examines the role of Prmt1 and SFPQ in craniofacial development. Specifically, the authors test the idea that Prmt1 directly methylates specific proteins that results in intron retention in matrix proteins. The protein SFPQ is methylated by Prmt1 and functions downstream to mediate Prmt1 activity. The genes with retained introns activate the NMD pathway to reduce the RNA levels. This paper describes an interesting mechanism for the regulation of RNA levels during development.

Strengths:

The phenotypes support what the authors claim that Prmt1 is involved in craniofacial development and splicing. They use of state of the art sequencing to determine the specific genes that have intron retention and changes in gene expression is a strength.

Weaknesses:

The results now support the conclusions;however, it is still unclear how direct the relationship is between Prmt1 and SFPQ.

Author response:

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

Public Reviews:

Reviewer #1 ( Public review):

The strength of the current study lies in their establishing the molecular mechanism through which PRMT1 could alter craniofacial development through regulation of the transcriptome, but the data presented to support the claim that a PRMT1-SFPQ axis directly regulates intron retention of the relevant gene networks should be robust and with multiple forms of clear validation. For example, elevated intron retention findings are based on the intron retention index, and according to the manuscript, are assessed considering the relative expression of exons and introns from a given transcript. However, delineating between intron retention and other forms of alternative splicing (i.e., cryptic splice site recognition) requires a more comprehensive consideration of the intron splicing defects that could be represented in data. A certain threshold of intron read coverage (i.e., the percent of an intron that is covered by mapped reads) is needed to ascertain if those that are proximal to exons could represent alternative introns ends rather than full intron retention events. In other words, intron retention is a type of alternative splicing that can be difficult to analyze in isolation given the confounding influence of cryptic splicing and cryptic exon inclusion. If other forms of alternative splicing were assessed and not detected, more confident retention calls can be made.

This manuscript is a mechanistic exploration that follows previous work we published on the role of Prmt1 in craniofacial development, in which genetic deletion of Prmt1 in CNCCs leads to cleft palate and mandibular hypoplasia (PMID: 29986157).

As the reviewer pointed out, a certain threshold of intron read coverage is needed to assess intron retention events. We employed IRTools to assess the collective changes of intron retention between cell-states associated with certain biological function or pathway. IRTools incorporated considerations for intron read coverage by checking the evenness of read distribution in an intron. Specifically, every constitutive intronic regions (CIR) is divided into 10 equally sized bins and the proportion of reads that map to each bin is calculated. CIRs are then ranked according to their imbalance in bin-wise reads distribution, represented by the proportion of reads in its most populated bin. Those among top 1% are considered to contain potentially false IR events and excluded. We further addressed this question by developing another measure of intron retention, intron retention coefficient (IRC), which assesses IR events using the junction reads (Supplemental Figure-S8). Junction reads that straddle two exons are called exon-exon junction reads (spliced reads), and those that straddle an exon and a neighboring intron are called exon-intron junction reads (retained reads). The IRC of an intron is defined as the fraction of junction reads that are exon-intron junction reads: IRC = exon-intron read-count / (exon-exon read-count + exon-intron read-count), where exon-intron read-count = (5’ exon-intron read-count + 3’ exon-intron read-count) / 2. The IRC of a gene is defined as the exon-intron fraction of all junction reads overlapping or over the constitutive introns of this gene. In the calculation of the IRC, only exon-intron junction reads that cover the junction point and overlap both of each side for at least 8 bps were counted, and only exon-exon junction reads that jump over the relevant junction points and overlap each of the respective exons for at least 8 bps were counted. In this process, evenness of the proportion of exon-intron junction reads that are 5’ or 3’ exon-intron junction reads are taken into account. As shown in the Supplemental Figure S7A and S7B, IRC analysis generated consistent results with those obtained from using IRI (Figure 3A and 3I).

In addition, as the reviewer pointed out, intron retention can be difficult to analyze in isolation. We followed the reviewer’s suggestion that “If other forms of alternative splicing were assessed and not detected, more confident retention calls can be made“ and analyzed other forms of alternative splicing for all ECM and GAG genes with significant IRI increase (genes highlighted in Figure-3A and 3I) using rMATS (Supplemental Figure-S9). Among these genes, only 5 genes (Cthcr1, Mmp23, Adamts10, Ccdc80 and Col25a1) showed statistically significant changes in skipped exon, 1 gene (Bmp7) showed significant changes in mutually exclusive exons, and none showed significant changes in alternative 5’ or 3’ splicing. SE and MXE changes detected were marginal (Supplemental figure S8), while the majority of matrix genes with significant intron retention didn’t exhibit other forms of alternative splicing, further supporting the confidence of intron retention calls.

While data presented to support the PRMT1-SFPQ activation axis is quite compelling, that this is directly responsible for the elevated intron retention remains enigmatic. First, in characterizing their PRMT1 knockout model, it is unclear whether the elevated intron retention events directly correspond to downregulated genes.

In the revised manuscript, we demonstrate IR-triggered NMD as a mechanism for transcript decay and downregulation of matrix genes. When IR-triggered NMD was blocked by chemical inhibitor NMDI14, the intron-retaining transcripts showed significant accumulation (new Figure-4). NMD is the RNA surveillance system to degrade aberrant RNAs. Intron retention-triggered NMD in cancer has both promotive and suppressive roles and NMD inhibitors has been tested for cancer therapy including immunotherapy. During embryonic development, the functional significance of NMD machinery is suggested by human genetic findings and mouse genetic models. NMD is driven by a protein complex composed of SMG and UPF proteins. Smg6, Upf1, Upf2 and Upf3a knockout mouse die at early embryonic stages (E5.5-E9.5), and Smg1 gene trap mutant mice die at E12.5 (PMID: 29272451). SMG9 mutation in human patients causes malformation in the face, hand, heart and brain (PMID: 27018474).

We show that in CNCCs NMD functions both as a physiological mechanism and invoked by molecular insult. Blocking NMD in CNCCs caused significant accumulation of intron-retaining Adamts2, Alpl, Eln, Matn2, Loxl1 and Bgn transcripts, suggesting a basal role for NMD to degrade intron-retaining transcripts (Figure-4Ba-4Bf). We further demonstrated the accumulation of Adamts2 and Fbln5 using semi-quantitative PCR with the detection of a longer product from Adamts2 intron 19 and Fbln5 intron 7 (Figure-4Ca-4Ch). In CNCCs and ST2 cells, NMD is further invoked by Prmt1 and Sfpq deficiency. In Prmt1 deficient CNCCs, NMD blockage led to higher accumulation of intron-retaining Adamts2 and Alpl transcripts, suggesting that Prmt1 deficiency triggers NMD to reduce intron-containing transcripts (Figure-4Aa, 4Ab). In Sfpq-depleted ST2 cells, blocking NMD caused accumulation of intron-retaining transcripts Col4a2, St6galnac3 and Ptk7 (Figure-9B, 9C).

Moreover, intron splicing is a well-documented node for gene regulation during embryogenesis and in other proliferation models, and craniofacial defects are known to be associated with 'spliceosomopathies'. However, reproduction of this phenotype does not suggest that the targets of interest are inherently splicing factors, and a more robust assessment is needed to determine the exact nature of alternative splicing in this system. Because there are several known splicing factors downstream of PRMT1 and presented in the supplemental data, the specific attribution of retention to SFPQ would be additionally served by separating its splicing footprint from that of other factors that are primed to cause alternative splicing.

We have previously shown that a group of splicing factors depends on Prmt1 for arginine methylation, including SFPQ (PMID: 31451547). We tested additional splicing factors that are highly expressed in CNCCs and depends on PRMT1 for arginine methylation: SRSF1, EWSR1, TAF15, TRA2B and G3BP1 (Figure-5, 6 and 10). Among these factors, EWSR1 and TRA2B are both methylated in CNCCs and depend on PRMT1 for methylation (Fig. 5 and Supplemental Figure-S3B, S3C). We weren’t able to assess TAF15 methylation because of lack of efficient antibody for the PLA assay. We also demonstrated that their protein expression or subcellular localization was not altered by Prmt1 deletion in CNCCs, unlike SFPQ (Supplemental Figure-S4). To define their splicing footprint, we performed siRNA-mediated knockdown in ST2 cells, followed by RNA-seq and IRI analysis to define differentially regulated genes and introns, which revealed distinct biological pathways regulated by SFPQ, EWSR1, TRA2B and TAF15, but minimal roles of EWSR1, TRA2B and TAF15 on intron retention when compared to SFPQ (Fig. 10F-10S, Supplemental Figure S7A-S7F, Supplemental Tables S4-S6). ECM genes are significantly downregulated by all four splicing factors (Fig. 10F-10I), but EWSR1, TRA2B and TAF15 function through IR-independent mechanisms, such as exon skipping, as exemplified by Postn (Fig. 10J-10S).

Clarifying the relationship between SFPQ and splicing regulation is important given that the observed splicing defects are incongruous with published data presented by Takeuchi et al., (2018) regarding SFPQ control of neuronal apoptosis in mice. In this system, SFPQ was more specifically attributed to the regulation of transcription elongation over long introns and its knockout did not result in significant splicing changes. Thus, to establish the specificity for the SFPQ in regulating these retention events, authors would need to show that the same phenotype is not achieved by mis-regulation of other splicing factors. That the authors chose SFPQ based on its binding profile is understandable but potentially confounding given its mechanism of action in transcription of long introns (Takeuchi 2018). Because mechanisms and rates of transcription can influence splicing and exon definition interactions, the role of SFPQ as a transcription elongation factor versus a splicing factor is inadequately disentangled by authors.

To test whether SFPQ acts as a transcription elongation factor, we performed Pol II Cut&Tag in ST2 cells and demonstrated that depletion of SFPQ only caused marginal changes in either the promoter region or gene body of ECM genes, suggesting that the role of SFPQ as a transcriptional activator or elongation factor is minimal (Fig. 7G, 7H). This finding is distinct from SFPQ function in neurons (PMID: 29719248), suggesting that the activation or recruitment of SFPQ in transcriptional regulation may involve tissue-specific factors in neurons.

Reviewer #2 (Public review):

Summary:

The manuscript by Lima et al examines the role of Prmt1 and SFPQ in craniofacial development. Specifically, the authors test the idea that Prmt1 directly methylates specific proteins that results in intron retention in matrix proteins. The protein SFPQ is methylated by Prmt1 and functions downstream to mediate Prmt1 activity. The genes with retained introns activate the NMD pathway to reduce the RNA levels. This paper describes an interesting mechanism for the regulation of RNA levels during development.

Strengths:

The phenotypes support what the authors claim that Prmt1 is involved in craniofacial development and splicing. The use of state-of-the-art sequencing to determine the specific genes that have intron retention and changes in gene expression is a strength.

Weaknesses:

Some of the data seems to contradict the conclusions. And it is unclear how direct the relationships are between Prmt1 and SFPQ.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

First, the claims regarding the effect of PRMT1 loss on splicing are unclear by the section title. In other words, does loss PRMT1 change the incidence of baseline alternative splicing events, or does it introduce new retention events that are responsible for underwriting the craniofacial phenotype? Consistent with this idea, the narrative could benefit from more cellular and/or histological validations of the transcriptomic defects discovered in the RNAseq, which could help contextualize the bioinformatics data with the developmental defects. Moreover, the conclusions drawn about intron retention could be clarified in terms of how applicable the mechanism is likely to be outside of this tissue-specific set of responsive introns.

Loss of Prmt1 did not cause a global shift in intron retention, as shown in Supplemental Figure S2. Instead, Prmt1 deletion caused increase of intron retention specifically in genes enriched in cartilage development, glycosaminoglycan biology, dendrite and axon, and decreased intron retention in mitochondria and metabolism genes (Table. S1). We also tested matrix protein expression by histology to confirm that transcriptomic defects revealed at the RNA level resulted in lower protein production. The new data are in Figure 3E-3H.

Additionally, invoking NMD to align splicing and differential gene expression data understandable but lacking sufficient controls to be conclusive, such as positive control genes to confirm inhibition of NMD.

To validate the blockage of NMD, glutathione peroxidase 1 (Gpx1) intron 1, a well-documented substrate for NMD, is tested as positive control (Fig 4Ac, 4Ad, 9B).

Additionally, it should be clarified whether NMD is a basal mechanism for the regulation of these introns or whether it is an induced mechanism that is invoked by the molecular insult.

In CNCCs, NMD functions both as a physiological mechanism and invoked by molecular insult. Please refer to responses to Reviewer 1’s public review for detailed explanations.

Further, authors present data downstream of two siRNAs for the same gene target, but it remains unclear how siRNAs for the same gene target produce different effects. It may be helpful for authors to clarify how many of the transcriptomic defects are shared versus unique between the siRNAs.

To address this question, we used bioinformatic analysis of the whole genome data to the similarity in changes caused by the two SFPQ-targeting siRNAs. As shown in the new Fig. 7Ba & 7Bb, transcriptomic and intron changes are consistent between the two siRNAs, suggesting that genes targeted by the two siRNA predominantly overlap. This overlap is illustrated by scatter plot analysis of RNAseq DEG and IRI data from each siRNA against SFPQ.

Finally, we stress the importance of presenting the full conceptual basis for SFPQ's potential role in splicing and gene expression. It is significant to note that SFPQ has been previously studied as a splicing factor and was instead determined to function in support of the transcription elongation rather than in splicing. Thus, if authors are confident that the SFPQ manifests directly in splicing changes they encumber the burden of proof to show that its role in transcription, nor another splicing factor, are driving splicing changes.

We demonstrated that depletion of SFPQ only caused marginal changes in either the promoter region or gene body of ECM genes, suggesting that the role of SFPQ as a transcriptional activator or elongation factor is minimal (Fig. 7G, 7H). Please refer to responses to Reviewer 1’s public review for detailed explanations.

Reviewer #2 (Recommendations for the authors):

(1) It is not clear why the authors focused on intron retention targets vs the other possibilities. Skipped Exon is much higher in terms of the number of changes, please clarify. For the intron retention how is this quantified? The traces are nice, but it is hard to tell which part is retained at this magnification. Also, because the focus is on extracellular matrix (ECM) and NMD it would be nice to show some of those targets here. In the tbx1 trace, some are up and some are down. What does that mean for the gene expression?

We have investigated SE initially and found that genes with significant changes in Prmt1 CKO CNCCs fall into diverse functional pathways. Among them, a few genes are critical for skeletal formation, including Postn and Fn, and the function of their exon skipping has been documented. For example, the two exons that are skipped in Postn, Exon17 and 21, have been shown to regulate craniofacial skeleton shape and mandibular condyle hypertrophic zone thickness using transgenic mouse models (PMID: 36859617). As illustrated by Figure 10, the skipped exon of Postn is regulated by multiple splicing factors that may perform overlapping functions in vivo.

Intron retention of each gene is quantified by the ratio of the overall read density of its constitutive intronic regions (CIRs) to the overall read density of its constitutive exonic regions (CERs) and defined as the intron retention index (IRI). In the first section of Response to Reviewer 1’s comments, we explained additional bioinformatic analysis that was performed to address reviewers’ questions, support the confidence of intron event calls and rule out the possibility of other alternative splicing mechanisms, such as by SE, MXE, A5SS or A3SS (Supplemental Figure S5, S6, Table S7).

(2) RNA-Sequencing of Prmt1 mutants nicely shows gene expression changes, including in ECM and GAG genes. While validation of the sequencing results is not necessarily required, it would be very interesting to show the expression in situ. In addition, the heat map shows both downregulated but also upregulated transcripts. This is expected since this protein regulates many genes. However, the volcano plot shows a significant number of genes upregulated. It would be interesting to show what the upregulated genes are. And what is the proposed mechanism for Prmt1 regulation of upregulated genes?

Validation for the transcriptomic changes is shown in Fig. 3E-3H using immunostaining.

As for upregulated genes in Prmt1 mutant, top pathways include cytokine-mediated signaling pathway, signal transduction by p53 signaling pathway and cell morphogenesis (Figure 2E), which are consistent with our previous reports that Prmt1 deletion induces cytokine production in oral epithelium and leads to p53 accumulation in embryonic epicardium (PMID: 32521264, 29420098). Besides these pathways, Prmt1 deletion also caused upregulation of genes involved in adult behavior, postsynaptic organization and apoptotic process, which is consistent with findings from other labs on PRMT1 function in neuronal and cancer cells (PMID: 34619150, 33127433).

(3) Specific transcripts were shown to have elevated intron retention involved in the ECM and GAG pathway. However in Figure 3D it seems to show the opposite with intronic expression decreased and exonic increases and intronic decrease. This is very important to the final conclusion of the paper. In addition, is there a direct relationship between increased intron and downregulation of this specific gene expression? It seems a bit correlational as it could also be an indirect mechanism. One way to test this is to do in vitro translation with and without the specific intron to test if it results in lower expression.

We apologize for the mis-labeling in previous version of Figure 3D, which is now corrected. We also tried to test the direct relationship between intron and downregulation of matrix genes such as Adamts2 using in vitro experiments, however, the introns of matrix genes with high retention tends to be long, many 10 to 50kb in length, making it challenging to generate mini-gene constructs for molecular analysis. We used a different approach and demonstrated that inhibition of NMD with a chemical inhibitor NMDI14 caused dramatic accumulation of the Adamts2, Alpl, Eln, Matn2, Loxl1 and Bgn transcripts, suggesting that retained introns triggered NMD to regulate gene expression and this mechanism acts as a physiological level in CNCCs (Fig. 4). We also blocked NMD in control and Prmt1 null CNCCs, where NMD blockage led to higher accumulation of Adamts2 and Alpl transcripts, suggesting that upon Prmt1 deficiency, NMD is further utilized to degrade intron-containing transcripts (Fig. 4). Similarly, in Sfpq-depleted ST2 cells, blocking NMD caused accumulation of intron-retaining transcripts Col4a2, St6galnac3 and Ptk7 (Fig. 9A, 9B).

(4) While Figure 4 nicely shows the methylation of SFPQ is reduced in Prmt1 CKO cells, it is unclear which reside this methylation occurs. Also the overall expression of SFPQ is also down so it is possible that the methylation is indirect ie Prmt1 regulates some other methyltransferase that regulates SFPQ. Or that because the overall level of SFPQ is down, there is no protein to methylate. How do the authors differentiate between these possibilities?

Previously, arginine methylation of SFPQ has been characterized using in vitro reaction and cell lines with biochemical assays by Snijders., et al in 2015 (PMID: 25605962). Among all PRMTs that catalyze asymmetric arginine dimethylation (ADMA), SFPQ is methylated by only PRMT1 and PRMT3, with PRMT1 showing higher efficiency while PRMT3 showing a lower efficiency. However, PRMT3 is mainly cytosolic. Its expression in CNCCs is about 100-fold lower than PRMT1 (Fig. 1). Based on these knowledges, PRMT1 is the primary arginine methyltransferase for SFPQ, a nuclear protein in CNCCs. We and others have shown in a previous publication that SFPQ methylation on arginine 7 and 9 depends on PRMT1 (PMID: 31451547).

To investigate SFPQ protein degradation in CNCCs, we used MG132 to block proteasomal degradation and observed a partial rescue of SFPQ protein degradation in Prmt1 mutant embryos, suggesting that SFPQ is degraded through proteasomal-mediated mechanism. To address the relationship between SFPQ methylation and protein expression, we assessed arginine methylation of SFPQ that accumulated after MG132 treatment. The accumulated SFPQ was not methylated, confirming the absence of methylation even when SFPQ protein expression is restored.

Snijders., et al, also shown that citrullination induced by PADI4 regulate SFPQ stability (Snijders 2015). We considered this possibility and assessed the expression levels of PADIs. In E13.5 and E15.5 CNCCs, PADI1-4 mRNA expression levels are very low (TPM<5), suggesting that PADIs may not regulate SFPQ stability in CNCCs. A detailed mechanism as to how PRMT1-mediated SFPQ methylation controls stability awaits further investigation.

(5) For the Sfpq deleted experiment, it seems that the two knockdowns are not similar in the gene targets and GO terms different except Wnt signaling. This makes this data difficult to interpret. The genes identified as intron retention are different than the ones identified in Prmt1 deletion and not reduced as much. How does this fit in with the Prmt1 story? If working through Sfpq, it assumes that the targets will be similar and more the 8% would be in common.

To address the first concern, we used bioinformatic analysis of the whole genome data to the similarity in changes caused by the two SFPQ-targeting siRNAs. As shown in the new Fig. 7Ba & 7Bb, transcriptomic and intron changes are consistent between the two siRNAs, suggesting that genes targeted by the two siRNA predominantly overlap. This overlap is illustrated by scatter plot analysis of RNAseq DEG and IRI data from each siRNA against SFPQ.

We have previously identified a group of splicing factors that depends on PRMT1 for arginine methylation, including SFPQ (PMID: 31451547). In the new data in Figures 5, 6 and 10, we tested an additional five PRMT1-dependent splicing factors that are highly expressed in CNCCs: SRSF1, EWSR1, TAF15, TRA2B and G3BP1 (Fig. 5, 6 and 10). Among these factors, SRSF1 and G3BP1 are predominantly expressed in the cytosol of NCCs at E13.5. As splicing activity in the nucleus is needed for pre-mRNA splicing, we excluded these two and focused on the other three proteins. EWSR1 and TRA2B are both methylated in CNCCs and depend on PRMT1 for methylation (Fig. 5). We weren’t able to assess TAF15 methylation because of lack of efficient antibody for the PLA assay. We also demonstrated that their protein expression or subcellular localization was not altered by Prmt1 deletion in CNCCs, unlike SFPQ (Fig. S2). To define their splicing footprint, we performed siRNA-mediated knockdown in ST2 cells, followed by RNA-seq and IRI analysis to define differentially regulated genes and introns, which revealed distinct biological pathways regulated by SFPQ, EWSR1, TRA2B and TAF15, but minimal roles of EWSR1, TRA2B and TAF15 on intron retention when compared to SFPQ (Fig. 10F-10I, Supplemental Figure S7A-S7F). ECM genes are significantly downregulated by all four splicing factors (Fig. 10J-10M), but EWSR1, TRA2B and TAF15 regulate transcription or exon skipping instead of IR, as exemplified by Alpl and Postn (Fig. 10N-10T).

(6) The addition of an NMD mechanism is interesting but not surprising that when inhibiting the pathway broadly, there is an increase in gene expression in the mesoderm cell line. How specific is this to craniofacial development?

NMD is driven by a protein complex composed of SMG and UPF proteins. We show in the revised manuscript that NMD is both a physiological mechanism in CNCCs and triggered by genetic disturbance (Fig. 4). These data are in line with human patient reports where SMG9 mutation in human causes malformation in the face, hand, heart and brain (PMID: 27018474). Mouse genetic studies also demonstrated roles of NMD components during embryonic development.Smg6, Upf1, Upf2 and Upf3a knockout mouse die at early embryonic stages (E5.5-E9.5), and Smg1 gene trap mutant mice die at E12.5 (Han 2018). Additionally, intron retention-triggered NMD in cancer has both promotive and suppressive roles and NMD inhibitors has been tested for cancer therapy and recently cancer immunotherapy. Our findings highlight matrix genes as one of the key targets for NMD during craniofacial development.

Minor:

(1) The supplemental figures are difficult to understand. In the first upload there are many figures and tables, some excel files that are separate uploads and some not. Please upload as separate files so it is clear. And also put them in order that they are in the manuscript.

(2) For the heat map in figure 2B, it would be good to show all the genes or none at all. It seems a bit like cherry-picking to highly only a few. And they are not labeled where they are located in the graph. Are these the top lines if so please label.

(3) Gene names in Figure 3A are difficult to read. I would also not consider BMP7 an ECM gene.

(4) A summary diagram of the interactions proposed will help to make this more understandable.

The supplemental figures are reorganized and uploaded as separate word and excel documents. For Heat map in Fig. 2B, we have removed the gene names. For Fig. 3A, only the most significantly changed gene are labeled in red dots with names. We didn’t label all the genes because of the large number of genes. For the new Figure 3B, we have replaced BMP7. A schematic summary is also added to Supplemental Fig. S9 to illustrate the PRMT1-SFPQ pathway.

eLife Assessment

This important study provides one mechanism that can explain the rapid diversification of poison-antidote pairs in fission yeast: recombination between existing pairs. The evidence is largely solid, but the study needs to tune down its claims (as it is not shown that the novel poison-antidote can serve as a meiotic driver), and to address small experimental requests. The work is of interest to scientists studying genetic incompatibilities.

Reviewer #1 (Public review):

Summary

The authors determine the phylogenetic relation of the roughly two dozen wtf elements of 21 S. pombe isolates and show that none of them in the original S. pombe are essential for robust mitotic growth. It would be interesting to test their meiotic function by simply crossing each deletion mutant with the parent and analyzing spores for non-Mendelian inheritance. If this has been reported already, that information should be added to the MS. If not, I suggest the authors do these simple experiments and add this information.

Strengths:

The most interesting data (Fig. 4) show that one recombinant (wtfC4) between wtf18 and wtf23 produces in mitotic growth a poison counteracted by its own antidote but not by the parental antidotes. Again, it would be interesting to test this recombinant in a more natural setting - meiosis between it and each of the parents.

Weaknesses:

Some minor rewriting is needed.

Comments on Revision:

(1) The parameter for "maximum growth rate" in Figure 2D needs to be defined and put on the graph.

(2) On page 8, line 182, the authors should consider testing the hybrid wtf in meiosis using strain 975 of Leupold, which is h+, or another standard h+ strain. I don't think the antidote allele is needed; rather, it seems to me it would counter the lethality of the poison protein and should be omitted to test drive of the hybrid wtf. This is a simple experiment and would add considerably to the paper.

Reviewer #3 (Public review):

Summary:

In this manuscript, Wang and colleagues explore factors contributing to the diversification of wtf meiotic drivers. wtf genes are autonomous, single-gene poison-antidote meiotic drivers that encode both a spore-killing poison (short isoform) and an antidote to the poison (long isoform) through alternative transcriptional initiation. There are dozens of wtf drivers present in the genomes of various yeast species, yet the evolutionary forces driving their diversification remain largely unknown. This manuscript is written in a straightforward and effective manner, and the analyses and experiments are easy to follow and interpret. While I find the research question interesting and the experiments persuasive, they do not provide any deeper mechanistic understanding of this gene family.

Revision update:

Having read the response to the reviewers, I believe the major issues have been addressed. However, I would strongly suggest toning down the claim regarding the chimeric WTF element in the abstract, which currently reads

"As proof-of-principle, we generate a novel meiotic driver through artificial recombination between wtf drivers, and its encoded poison cannot be detoxified by the antidotes encoded by their parental wtf genes but can be detoxified by its own antidote."

As the author reports in their response, despite various attempts, it was not possible to show that this chimeric WTF element was indeed capable of meiotic drive in a natural context (not transgenic overexpression experiment). thus the authors should not claim they generated "a novel meiotic driver"

Strengths:

(1) The authors present a comprehensive compendium and analysis of the evolutionary relationships among wtf genes across 21 strains of S. pombe

(2) The authors found that a synthetic chimeric wtf gene, combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves like a meiotic driver that could only be rescued by the chimeric antidote but neither of the parental antidotes. This is a very interesting observation that could account for their inception and diversification.

Weaknesses:

(1) Deletion strains

The authors separately deleted all 25 Wtf genes in the S. pombe ference strain. Next, the authors performed spot assay to evaluate the effect of wtf gene knockout on the yeast growth. They report no difference to the WT and conclude that the wtf genes might be largely neutral to the fitness of their carriers in the asexual life cycle at least in normal growth condition.

The authors could have conducted additional quantitative growth assays in yeast, such as growth curves or competition assays, which would have allowed them to detect subtle fitness effects that cannot be quantified with a spot assay. Furthermore, the authors do not rule out simpler explanations, such as genetic redundancy. This could have been addressed by crossing mutants of closely related paralogs or editing multiple wtf genes in the same genetic background.

Another concern is the lack of detailed information about the 25 knockout strains used in the study. There is no information provided on how these strains were generated or, more importantly, validated. Many of these wtf genes have close paralogs and are flanked by repetitive regions, which could complicate the generation of such deletion strains. As currently presented, these results would be difficult to replicate in other labs due to insufficient methodological details

Revision update:

The authors measured the fitness of the deletion strains using growth curves (Fig. 2C and D) and no significant differences were found, further supporting their claims. The requested information (details on the generation of the deletion strains) is now available in the methods section.

(2) Lack of controls

The authors found that a synthetic chimeric wtf gene, constructed by combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves as a meiotic driver that can be rescued only by its corresponding chimeric antidote, but not by either of the parental antidotes (Figure 4F). In contrast, three other chimeric wtf genes did not display this property (Figure 4C-E). No additional experiments were conducted to explain these differences, and basic control experiments, such as verifying the expression of the chimeric constructs, were not performed to rule out trivial explanations. This should be at the very least discussed. Also, it would have been better to test additional chimeras.

Revision update:

The authors report that the expression of the construct was measured. However, they do not make reference to any specific figure or section of the main text. It would be very useful if the authors explicitly referenced where exactly changes were made (this is true for all changed made)

(3) Statistical analyses

In line 130 the authors state that: "Given complex phylogenetic mixing observed among wtf genes (Figure 1E), we tested whether recombination occurred. We detected signals of recombination in the 25 wtf genes of the S. pombe reference genome (p = 0) and in the wtf genes of the 21 S. pombe strains (p = 0) using pairwise homoplasy index (HPI) test. "<br /> Reporting a p-value of 0 is not appropriate. Please report exact P-values.

Revision update:

This has been addressed.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary

The authors determine the phylogenetic relation of the roughly two dozen wtf elements of 21 S. pombe isolates and show that none of them in the original S. pombe are essential for robust mitotic growth. It would be interesting to test their meiotic function by simply crossing each deletion mutant with the parent and analyzing spores for non-Mendelian inheritance. If this has been reported already, that information should be added to the manuscript. If not, I suggest the authors do these simple experiments and add this information.

Thanks for the great summary! All the wtf genes have been tested for meiotic drive phenotypes previously by Bravo Nunez et al. (2020; http://doi.org/10.1371/journal.pgen.1008350). The reference was cited in our original manuscript, and we added the details in the revised manuscript.  

Strengths:

The most interesting data (Figure 4) show that one recombinant (wtfC4) between wtf18 and wtf23 produces in mitotic growth a poison counteracted by its own antidote but not by the parental antidotes. Again, it would be interesting to test this recombinant in a more natural setting - meiosis between it and each of the parents.

Thanks for this insightful comment! As suggested, we have tried to test this recombinant in a more natural setting. We created a recombinant strain (wtfC4) based on the laboratory strain 972h-. Specifically, we replaced the last exon of the original wtf23 gene with the last exon of wtf18. However, we encountered a challenge: since strain 972h- has only one mating type and cannot undergo meiosis on its own, we had to mate the recombinant strain with a BN0 h⁺ strain that only carries the wtf23^antidote. Unfortunately, despite of tens of attempts over nearly a year, we did not observe meiotic driver phenotype as expected. This might be due to issues with the proper splicing and expression of the potential poison and antidote proteins or due to the genetic background. Similarly, the drive activity of wtf13 has been shown to be specifically suppressed in certain backgrounds.

Weaknesses:

In the opinion of this reviewer, some minor rewriting is needed.

We did the rewriting as this reviewer suggested.

Reviewer #2 (Public review):

Summary:

This important study provides a mechanism that can explain the rapid diversification of poison-antidote pairs (wtf genes) in fission yeast: recombination between existing genes.

Thanks!

Strengths:

The authors analyzed the diversity of wtf in S. pombe strains, and found pervasive copy number variations. They further detected signals of recurrent recombination in wtf genes. To address whether recombination can generate novel wtf genes, the authors performed artificial recombination between existing wft genes, and showed that indeed a new wtf can be generated: the poison cannot be detoxified by the antidotes encoded by parental wtf genes but can be detoxified by own antidote.

Thanks for the great summary!

Weaknesses:

The study can benefit from demonstrating that the novel poison-antidote constructed by the authors can serve as a meiotic driver.

Thanks for this insightful comment! As suggested, we have tried to test this recombinant in a more natural setting. We created a recombinant strain (wtfC4) based on the laboratory strain 972h-. Specifically, we replaced the last exon of the original wtf23 gene with the last exon of wtf18. However, we encountered a challenge: since strain 972h- has only one mating type and cannot undergo meiosis on its own, we had to mate the recombinant strain with a BN0 h⁺ strain that only carries the wtf23^antidote. Unfortunately, despite of tens of attempts over nearly a year, we did not observe meiotic driver phenotype as expected. This might be due to issues with the proper splicing and expression of the potential poison and antidote proteins or due to the genetic background. Similarly, the drive activity of wtf13 has been shown to be specifically suppressed in certain backgrounds.

Reviewer #3 (Public review):

Summary:

In this manuscript, Wang and colleagues explore factors contributing to the diversification of wtf meiotic drivers. wtf genes are autonomous, single-gene poison-antidote meiotic drivers that encode both a spore-killing poison (short isoform) and an antidote to the poison (long isoform) through alternative transcriptional initiation. There are dozens of wtf drivers present in the genomes of various yeast species, yet the evolutionary forces driving their diversification remain largely unknown. This manuscript is written in a straightforward and effective manner, and the analyses and experiments are easy to follow and interpret. While I find the research question interesting and the experiments persuasive, they do not provide any deeper mechanistic understanding of this gene family.

Thanks! Please see the following for our point-to-point response.

Strengths:

(1) The authors present a comprehensive compendium and analysis of the evolutionary relationships among wtf genes across 21 strains of S. pombe.

(2) The authors found that a synthetic chimeric wtf gene, combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves like a meiotic driver that could only be rescued by the chimeric antidote but neither of the parental antidotes. This is a very interesting observation that could account for their inception and diversification.

Thanks for the great summary!

Weaknesses:

(1) Deletion strains

The authors separately deleted all 25 Wtf genes in the S. pombe ference strain. Next, the authors performed a spot assay to evaluate the effect of wtf gene knockout on the yeast growth. They report no difference to the WT and conclude that the wtf genes might be largely neutral to the fitness of their carriers in the asexual life cycle at least in normal growth conditions.

The authors could have conducted additional quantitative growth assays in yeast, such as growth curves or competition assays, which would have allowed them to detect subtle fitness effects that cannot be quantified with a spot assay. Furthermore, the authors do not rule out simpler explanations, such as genetic redundancy. This could have been addressed by crossing mutants of closely related paralogs or editing multiple wtf genes in the same genetic background.

Another concern is the lack of detailed information about the 25 knockout strains used in the study. There is no information provided on how these strains were generated or, more importantly, validated. Many of these wtf genes have close paralogs and are flanked by repetitive regions, which could complicate the generation of such deletion strains. As currently presented, these results would be difficult to replicate in other labs due to insufficient methodological details

We generated growth curves for all the 25 wtf deletion strains. We provided the details for wtf gene knockout. However, for 25 wtf genes, there are too many combinations for editing two genes, and it is technically challenging to knock out multiple wtf together. Nevertheless, our results suggest single wtf genes have little effect on the host fitness under normal condition.

(2) Lack of controls

The authors found that a synthetic chimeric wtf gene, constructed by combining exons 1-5 of wtf23 and exon 6 of wtf18, behaves as a meiotic driver that can be rescued only by its corresponding chimeric antidote, but not by either of the parental antidotes (Figure 4F). In contrast, three other chimeric wtf genes did not display this property (Figure 4C-E). No additional experiments were conducted to explain these differences, and basic control experiments, such as verifying the expression of the chimeric constructs, were not performed to rule out trivial explanations. This should be at the very least discussed. Also, it would have been better to test additional chimeras.

We verified the expression of the chimeric genes. The last exon of wtf18 is too small (128bp) to do more meaningful chimeras.

(3) Statistical analyses

In line 130 the authors state that: "Given complex phylogenetic mixing observed among wtf genes (Figure 1E), we tested whether recombination occurred. We detected signals of recombination in the 25 wtf genes of the S. pombe reference genome (p = 0) and in the wtf genes of the 21 S. pombe strains (p = 0) using pairwise homoplasy index (HPI) test." Reporting a p-value of 0 is not appropriate. Exact P-values should be reported. 

Due to software limitations, the PHI test reports p-values of 0.0 for extremely significant results. We have therefore reported them as <0.0001 in the revised manuscript.

Recommendations for the authors:

Reviewing Editor Comments:

Regarding the synthetic chimeric wtf gene constructed by combining exons of wtf23 and wtf18, the authors did not explicitly test whether it acts as a meiotic driver in the natural context of a cross. Instead, they examined this possibility only through transgenic overexpression experiments. Given that this is arguably the most important claim of the paper, it is critical that the authors perform, report, and discuss such an experiment in a natural context, regardless of the outcome. It is not necessary to test other recombinants or other wtf loci.

Thanks for this insightful comment! As suggested, we have tried to test this recombinant in a more natural setting. We created a recombinant strain (wtfC4) based on the laboratory strain 972h-. Specifically, we replaced the last exon of the original wtf23 gene with the last exon of wtf18. However, we encountered a challenge: since strain 972h- has only one mating type and cannot undergo meiosis on its own, we had to mate the recombinant strain with a BN0 h⁺ strain that only carries the wtf23^antidote. Unfortunately, despite of tens of attempts over nearly a year, we did not observe meiotic driver phenotype as expected. This might be due to issues with the proper splicing and expression of the potential poison and antidote proteins or due to the genetic background. Similarly, the drive activity of wtf13 has been shown to be specifically suppressed in certain backgrounds.

Reviewer #1 (Recommendations for the authors):

The paper is very well written, but some minor points should be corrected or checked.

(1) Line 95: Why "Putative"? Is it not clear what a wtf pseudogene is?

“Putative” was removed.

(2) Line 105: Does "known functional" mean they are active (i.e., have been tested and shown to be active)? If so, a reference should be added.

We used “known meiotic divers”, and added reference here.

(3) Line 135: "no recombination signal was tested". Do the authors mean no signal was inferred? 

We changed “tested” to “detected”.

(4) Line 147: References for "known functional meiotic drivers (wtf23) and artificially generated meiotic driver (wtf18)" should be given. A statement of how wtf18 was "artificially generated" is essential so the reader knows how that element differs from the wtfC4 generated here.

Reference for wtf23. As for wtf18, we have specified in the follow text, namely “we artificially introduced an in-frame ATG codon right before the start of exon 2, generating wtf18poison/-0M.”

(5) Lines 154 and 424 say an ATG codon was introduced "right before the start of exon 2," but Figure 4B shows it before exon 1.

We thank the reviewer. The introduced ATG is the second start codon in the long transcript and the first in the short transcript. The right panel of Figure 4B shows the short transcript, so the text and figure are consistent.

(6) Line 159: The wtf18 mutant with this additional ATG codon should be tested in meiosis, to see if "putative" is correct.

Thanks. As wtfC4, we came with technical challenges to show the driver phenotype in a natural setting, and thus removed this statement.

(7) Line 181: change "driver" to "drive".

Driver is correct.

(8) Line 184: insert to read "wtf genes tested". Also, what is the basis for proposing that "the last exon might be crucial for antidote function"?

“Tested” added, and removed the statement.

(9) Line 198: change to read "detects only large differences".

Done as suggested.

(10) Line 204: change "removed" to "removal".

Done as suggested.

(11) Lines 242 and 243: Are "Splittree4" and "SplitsTree4" different, or is this a misprint?

Corrected!

(12) Lines 274-5 and 412 -3 would read better as "strains were diluted in five 10-fold steps” and “...μL of each dilution spotted on” “…to assay for…"

Done as suggested.

(13) Line 284 says "No new data were generated." This is clearly wrong. Perhaps the authors mean there are no supplementary data files.

Corrected!

(14) Line 406: Change "is" to "are".

Corrected!

(15) Line 413: Surely, they were spotted onto YE agar medium, not liquid medium.

Corrected!

(16) Figure 3C: Define "Rho" and the scale used.

The definition of Rho has been added to the Methods section in the revised manuscript.

Reviewer #2 (Recommendations for the authors):

The evidence is largely solid, but the study can benefit from demonstrating that the novel poison-antidote constructed by the authors can serve as a meiotic driver.

As suggested, we have tried to test this recombinant in a more natural setting. We created a recombinant strain (wtfC4) based on the laboratory 972h-. Specifically, we replaced the last exon of the original wtf23 gene with the last exon of wt18f. However, we encountered a challenge: since 972h- is a mating-type strain and cannot undergo meiosis on its own, we had to mate the recombinant strain with a BN0 h⁺ strain that carries the wtf23^antidote. Unfortunately, despite of tens of attempts over nearly a year, we did not observe meiotic driver phenotype as expected. This might be due to issues with the proper splicing and expression of the potential poison and antidote proteins.

Reviewer #3 (Recommendations for the authors):

I strongly recommend the authors provide all the details concerning the generation of the knock-out strains, including specific primers used (for both the deletion and validation), the result of these validations, and the specific genotype (and ID) of the strains generated.

These details are now included in the Materials and Methods section and in Supplementary.

Please also provide exact P-values (see point 3).

Due to software limitations, the PHI test reports p-values of 0.0 for extremely significant results. We have therefore reported them as <0.0001 in the revised manuscript.

eLife Assessment

This valuable study uses tools of population and functional genomics to examine long non-coding RNAs (lncRNAs) in the context of human evolution. Analyses of computationally predicted human-specific lncRNAs and their genomic targets lead to the development of hypotheses regarding the potential roles of these genetic elements in human biology. Compared to previous versions, the conclusions regarding evolutionary acceleration and adaptation have become more solid by more fully taking data and literature on human/chimpanzee genetics and functional genomics into account.

Joint Public Review:

While DNA sequence divergence, differential expression and differential methylation analysis have been conducted between humans and the great apes to study changes that "make us human", the role of lncRNAs and their impact on the human genome and biology has not been fully explored. In this study the authors computationally predict HSlncRNAs as well as their DNA Binding sites using a method they have developed previously and then examine these predicted regions with different types of enrichment analyses. Broadly the analysis are straightforward and after identifying these regions/HSlncRNAs they examined their effects using different external datasets.

Comments on the latest version from Reviewer #2:

I think this is as good as it is going to get, and I do appreciate that the authors are still engaging in good faith after all these rounds of revision, so I am happy to stop here! I do think the paper is significantly improved from the last time around, and the conclusions have been tempered significantly.

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #2 (Public review):

In this valuable manuscript, Lin et al attempt to examine the role of long non coding RNAs (lncRNAs) in human evolution, through a set of population genetics and functional genomics analyses that leverage existing datasets and tools. Although the methods are incomplete and at times inadequate, the results nonetheless point towards a possible contribution of long non coding RNAs to shaping humans, and suggest clear directions for future, more rigorous study.

Comments on revisions:

I thank the authors for their revision and changes in response to previous rounds of comments. As before, I appreciate the changes made in response to my comments, and I think everyone is approaching this in the spirit of arriving at the best possible manuscript, but we still have some deep disagreements on the nature of the relevant statistical approach and defining adequate controls. I highlight a couple of places that I think are particularly relevant, but note that given the authors disagree with my interpretation, they should feel free to not respond!

(1) On the subject of the 0.034 threshold, I had previously stated: "I do not agree with the rationale for this claim, and do not agree that it supports the cutoff of 0.034 used below."

In their reply to me, the authors state:

"What we need is a gene number, which (a) indicates genes that effectively differentiate humans from chimpanzees, (b) can be used to set a DBS sequence distance cutoff. Since this study is the first to systematically examine DBSs in humans and chimpanzees, we must estimate this gene number based on studies that identify differentially expressed genes in humans and chimpanzees. We choose Song et al. 2021 (Song et al. Genetic studies of human-chimpanzee divergence using stem cell fusions. PNAS 2021), which identified 5984 differentially expressed genes, including 4377 genes whose differential expression is due to trans-acting differences between humans and chimpanzees. To the best of our knowledge, this is the only published data on trans-acting differences between humans and chimpanzees, and most HS lncRNAs and their DBSs/targets have trans-acting relationships (see Supplementary Table 2). Based on these numbers, we chose a DBS sequence distance cutoff of 0.034, which corresponds to 4248 genes (the top 20%), slightly fewer than 4377."

I have some notes here. First, Agoglia et al, Nature, 2021, also examined the nature of cis vs trans regulatory differences between human and chimps using a very similar set up to Song et al; their Supplementary Table 4 enables the discovery of genes with cis vs trans effects although admittedly this is less straightforward than the Song et al data. Second, I can't actually tell how the 4377 number is arrived at. From Song et al, "Of 4,671 genes with regulatory changes between human-only and chimpanzee-only iPSC lines, 44.4% (2,073 genes) were regulated primarily in cis, 31.4% (1,465 genes) were regulated primarily in trans, and the remaining 1,133 genes were regulated both in cis and in trans (Fig. 2C). This final category was further broken down into a cis+trans category (cis- and transregulatory changes acting in the same direction) and a cis-trans category (cis- and trans-regulatory changes acting in opposite directions)." Even when combining trans-only and cis&trans genes that gives 2,598 genes with evidence for some trans regulation. I cannot find 4,377 in the main text of the Song et al paper.

Elsewhere in their response, the authors respond to my comment that 0.034 is an arbitrary threshold by repeating the analyses using a cutoff of 0.035. I appreciate the sentiment here, but I would not expect this to make any great difference, given how similar those numbers are! A better approach, and what I had in mind when I mentioned this, would be to test multiple thresholds, ranging from, eg,0.05 to 0.01 <DBS dist =0.01 -> 0.034 -> 0.05> at some well-defined step size.

(1) We sincerely thank the reviewer for this critical point. Our initial purpose, based on DBS distances from the human genome to chimpanzee genome and archaic genomes, was that genes with large DBS distances may have contributed more to human evolution. However, our ORA (overrepresentation analysis) explored only genes with large DBS distances (the legend of old Figure 2 was “1256 target genes whose DBSs have the largest distances from modern humans to chimpanzees and Altai Neanderthals are enriched in different Biological Processes GO terms”), with the use of the cutoff (threshold) of 0.034 for defining large distance. The cutoff is not totally unreasonable (as our new results and the following sensitivity analysis indicate), but this approach was indirect and flawed.

(2) We have now performed ORA using two methods. The first uses only DBS distances. Instead of using a cutoff, we now sort genes by DBS distance (human-chimpanzee distances and human-Altai Neanderthal distance, respectively, see Supplementary Table 5) and use the top 25% and bottom 25% of genes to perform ORA. This directly examines whether DBS distances along indicate that genes with large DBS distances contribute more to human evolution than genes with small DBS distances. The second also explores the ASE genes (allele-specific expression, genes undergoing human/chimpanzee-specific regulation in the tetraploid human–chimpanzee hybrid iPS) reported by Agoglia et al. 2021. We select the top 50% and bottom 50% of genes with large and small DBS distances, intersect them with ASE genes from Agoglia et al. 2021 (their Supplementary Table 4), and apply ORA to the intersections. Both the results are that: (a) more GO terms are obtained from genes with large DBS distances, (b) more human evolution-related GO terms are obtained from genes with large DBS distances (Supplementary Table 5,6,7; Figure 2; Supplementary Fig. 15). These results directly suggest that genes with large DBS distances contribute more to human evolution than genes with small DBS distances, which is a key theme of the study.

(3) Regarding Song et al 2021, the statement of “we differentiated…allotetraploid (H1C1a, H1C1b, H2C2a, H2C2b) lines into ectoderm, mesoderm, and endoderm” made us assume that their differentiated hybrid cell lines cover more tissue types than those of Agoglia et al. 2021. Now, upon re-examining Supplementary Table 5 of Song et al. and Supplementary Table 4 of Agoglia et al. 2021, we find that the latter more clearly indicates significant ASE genes (p-adj<0.01 and |LFC>0.5| in GRCh38 and PanTro5).

(4) We have also performed two additional analyses in response to the suggestion of “test multiple thresholds, ranging from, eg, 0.05 to 0.01 <DBS dist =0.01 -> 0.034 -> 0.05> at some well-defined step size”. First, we performed a multi-threshold sensitivity analysis using a spectrum of cutoffs (0.03, 0.034, 0.04, 0.05), and tracked the number of genes identified and the enrichment significance of key GO terms (e.g., "neuron projection development," "behavior") across these thresholds. The result confirms that while the absolute number of genes varies with the cutoffs, the core biological conclusion (specifically, the significant enrichment of target genes in neurodevelopmental and cognitive functions) remains stable and significant. For instance, "behavior" maintains strong statistical significance (FDR<0.01) in both the human-chimpanzee and human-Altai Neanderthal comparisons across all tested cutoffs, and "Neuron projection development" also remains significant across three (0.03, 0.034, 0.04) of the four cutoffs in the Altai comparison. This pattern suggests that our core findings regarding neurodevelopmental functions are robust across a range of cutoffs. Nevertheless, we did not extend the analysis to smaller cutoffs (e.g., 0.01 or 0.02) because such values would identify an excessively large number of genes (>10000) for ORA, which would render the GOterm enrichment analysis less meaningful due to a loss of specificity.

Second, we have performed an additional validation to directly evaluate whether the 0.034 cutoff itself represents a stringent and biologically meaningful value. We sought to empirically determine how often a DBS sequence distance of 0.034 or greater might occur by chance in promoter regions, thereby testing its significance as a marker of potential evolutionary divergence. We randomly sampled 10,000 windows from annotated promoter regions across the hg38 genome, each with a size matching the average length of DBSs (147 bp). We then calculated the per-base sequence distances for these random windows between modern humans and chimpanzees, as well as between modern humans and the three archaic humans (Altai, Denisovan, Vindija). The analysis reveals that a distance of ≥0.034 is a rare event in random promoter sequences: for Human-Chimp, Human-Altai, HumanDenisovan, and Human-Vindija, 5.49% (549/10000), 0.31% (31/10000), 4.47% (447/10000), and0.03% (3/10000) of random windows reach this distance. This empirical evidence suggests that 0.034 is a sufficiently strong cutoff for defining large DBS distance, it would occur very unlikely in a random genomic background (P<0.1 for Chimpanzee and P<0.05 for the archaic humans), and DBSs exceeding this cutoff are significantly enriched for sequences that have undergone substantial evolutionary change instead of being random neutral variations.  

(5) We present new Figure 2, Supplementary Table 5,6,7, and Supplementary Fig. 15. We have substantially revised section 2.3, related sections in Results, Supplementary Note 3, and Supplementary Table 8. We have removed related descriptions and explanations in the main text and Supplementary Notes. The results of the above two analyses are presented here as two Author response images.

Author response table 1.

Sensitivity analysis of GO-term enrichment across different DBS sequence distance cutoffs. The table shows the numbers of target genes identified and the false discovery rates (FDR) for the enrichment of three selected GO terms at four different distance cutoffs. Note that, unlike in the old Figure 2, the results for chimpanzees and Altai Neanderthals are not directly comparable here, as the numbers of target genes used for the enrichment analysis differ between them at each cutoff.

Author response image 1.

Distribution of per-base sequence distances for DBS size-matched random genomic windows in Ensembl-annotated promoter regions, calculated between modern humans and (A) chimpanzee, (B) Altai Neanderthal, (C) Denisovan, and (D) Vindija Neanderthal genomes.

(2) The authors have introduced a new TFBS section, as a control for their lncRNAs - this is welcome, though again I would ask for caution when interpreting results. For instance, in their reply to me the authors state: "The number of HS TFs and HS lncRNAs (5 vs 66) <HS TF vs all HS lncRNAs> alone lends strong evidence suggesting that HS lncRNAs have contributed more significantly to human evolution than HS TFs (note that 5 is the union of three intersections between <many2zero + one2zero> and the three <human TF list>)."

But this assumes the denominator is the same! There are 35899 lncRNAs according to the current GENCOVE build; 66/35899 = 0.0018, so, 0.18% of lncRNAs are HS. The authors compare this to 5 TFs. There are 19433 protein coding genes in the current GENCOVE build, which naively (5/19433) gives a big depletion (0.026%) relative to the lnc number. However, this assumes all protein coding genes are TFs, which is not the case. A quick search suggests that ~2000 protein coding genes are TFs (see, eg, https://pubmed.ncbi.nlm.nih.gov/34755879/); which gives an enrichment (although I doubt it is a statistically significant one!) of HS TFs over HS lncRNAs (5/2000 = 0.0025). Hence my emphasis on needing to be sure the controls are robust and valid throughout!

We thank the reviewer for this comment. While 5 vs 66 reveals a difference, a direct comparison is too simplified. The real take-home message of the new TFBS section is not the numbers but the distributions of HS TFs’ targets and HS lncRNAs’ targets across GTEx organs and tissues (Figure 3 and Supplementary Figures 24, 25) - correlated HS lncRNA-target transcript pairs are highly enriched in brain regions, but correlated HS TF-target transcript pairs are distributed broadly across GTEx tissues and organs. We have now removed the simple comparison of “5 vs 66” and more carefully explained our comparison in section 2.6.

(3) In my original review I said: line 187: "Notably, 97.81% of the 105141 strong DBSs have counterparts in chimpanzees, suggesting that these DBSs are similar to HARs in evolution and have undergone human-specific evolution." I do not see any support for the inference here. Identifying HARs and acceleration relies on a far more thorough methodology than what's being presented here. Even generously, pairwise comparison between two taxa only cannot polarise the direction of differences; inferring human-specific change requires outgroups beyond chimpanzee.

In their reply to me, the authors state:

Here, we actually made an analogy but not an inference; therefore, we used such words as "suggesting" and "similar" instead of using more confirmatory words. We have revised the latter half sentence, saying "raising the possibility that these sequences have evolved considerably during human evolution".

Is the aim here to draw attention to the ~2.2% of DBS that do not have a counterpart? In that case, it would be better to rewrite the sentence to emphasise those, not the ones that are shared between the two species? I do appreciate the revised wording, though.

(1) Our original phrasing may be misleading, and we agree entirely that “pairwise comparison between two taxa only cannot polarise the direction of differences; inferring human-specific change requires outgroups beyond chimpanzee”. As explained in that reply, we know and think that DBSs and HARs are two different classes of sequences, and indeed, identifying HARs and acceleration relies on a far more thorough methodology. Yet, three factors prompted us to compare them. First, both suggest the importance of sequences outside genes. Second, both are quite “old” sequences and have undergone considerable evolution recently (although the references are different). Third, both have contributed greatly to human brain evolution.  

(2) Here, our stress is 97.81% but not 2.2%, and we have made this analogy more clearly and cautiously. Relevant revisions have been made in the Results, Discussion, and Methods sections.   

(3) We also have further determined whether the 2.2% DBSs are human-specific gains by analyzing them using the UCSC Multiz Alignments of 100 Vertebrates. The result confirms that all 2248 DBSs are present in the human genome but are absent from the chimpanzee genome and all other aligned vertebrate genomes. We add this result into the manuscript.

(4) Finally, Line 408: "Ensembl-annotated transcripts (release 79)" Release 79 is dated to March 2015, which is quite a few releases and genome builds ago. Is this a typo? Both the human and the chimpanzee genome have been significantly improved since then!

(1) We thank the reviewer for this comment, which prompts us to provide further explanation and additional data. First, we began predicting HS lncRNAs’ DBSs when Ensembl release 79 was available, but did not re-predict DBSs when new Ensembl releases were published because (a) these new Ensembl releases are based also on hg38, (b) we did not find any fault in the LongTarget program during our use, nor received any one from users, (c) predicting lncRNAs’ DBSs using the LongTarget program is highly time-consuming.  

(2) Second, to assess the influence of newer Ensembl releases, we compared the promoters annotated in release 79 and in release 115. We found that the vast majority (87.3%) of promoters newly annotated in release 115 belong to non-coding genes. Thus, using release 115 may predict more DBSs in non-coding genes, but downstream analyses based on protein-coding genes would be essentially the same (meaning that all figures and tables would be the same).

(3) Third, a key element of this study is GTEx data analysis, and these data were also published years ago.  

(4) Finally, some lncRNA genes have new gene symbols in new Ensembl releases. To allow researchers to use our data conveniently, we have added a new column titled "Gene symbol (Ensembl release115)" to Supplementary Tables 2A and 2B.  

Summary:

Major changes based on Reviewer’s comments:

(1) The following revisions are made to address the comment on “the 0.034 threshold”: (a) Section 2.3, section 2.4, Supplementary Note 3, and related contents in Discussion and Methods are revised, (b) new Figure 2, Supplementary Figure 15, new Supplementary Table 5,6,7, (c) Table 2 and Supplementary Table 8 are revised.

(2) To address the comment on “new TFBS section”, section 2.6 and section 4.13 are revised.  

(3) To address the comment on “97.81% and 2.2% of DBSs”, section 2.3 is revised.

(4) The following revisions are made to address the comment on “release 79”: (a) the old Supplementary Table 2, 3 are merged to Supplementary Table 2AB, and the new column "Gene symbol (Ensembl release115)" is added to Supplementary Table 2AB, (b) accordingly, Supplementary Table 4,5 are renamed to Supplementary Table 3,4.

Additional revisions:

(1) Section 2.5 “Young weak DBSs may have greatly promoted recent human evolution” is moved into Supplementary Note 3 (which now has the subtitle “Target genes with specific DBS features are enriched in specific functions”), because this section is short and lacking sufficient cross-validation.

(2) Considerable minor revisions of sentences have been made.

(3) Since there are many supplementary figures, the main text now cites only Supplementary Notes, as the reader can easily access supplementary figures in Supplementary Notes.

eLife Assessment

This important study provides evidence supporting the hypothesis that postnatal visual experience shapes the patterns of functional connectivity between extrastriate visual cortex and frontal regions, by comparing neonates, blind and sighted adults using resting-state fMRI. The evidence supporting the main claim is convincing, and the authors' interpretations are appropriately calibrated in the discussion. Nevertheless, the study design and methodology are inherently limited to resolve the underlying mechanisms driving connectivity changes during neurodevelopment (experience-related plasticity vs post-natal experience-independent maturation). This study will be of broad interest to neuroscientists and neuroimaging researchers studying vision, plasticity and brain development.

Reviewer #1 (Public review):

Summary:

The present study evaluates the role of visual experience in shaping functional correlations between human extrastriate visual cortex and frontal regions. The authors used fMRI to assess "resting-state" temporal correlations in three groups: sighted adults, congenitally blind adults, and neonates. Previous research has already demonstrated differences in functional correlations between visual and frontal regions in sighted compared to early blind individuals. The novel contribution of the current study lies in the inclusion of an infant dataset, which allows for an assessment of the developmental origins of these differences.

The main results of the study reveal that correlations between prefrontal and visual regions are more prominent in the blind and infant groups, with the blind group exhibiting greater lateralization. Conversely, correlations between visual and somato-motor cortices are more prominent in sighted adults. Based on these data, the authors conclude that visual experience shapes these cortical networks through activity-dependent plasticity. This study provides novel insights into the impact of visual experience on the development of temporal correlations in the brain.

Strengths:

The dissociations in functional correlations observed among the sighted adult, congenitally blind, and neonate groups provide strong support for the main conclusion regarding postnatal experience-driven shaping of visual-frontal connectivity.

The neonatal data offers a unique and valuable developmental anchor for interpreting divergence between blind and sighted adults. This is a major advance over prior studies limited to adult comparisons.

Convergence with prior findings in the blind and sighted adult groups reinforces the reliability and external validity of the present results.

The split-half reliability analysis in the infant and adult data increases confidence in the robustness of the reported group differences.

Weaknesses:

The methodology cannot determine whether group differences in correlations reflect direct changes in communication between visual and frontal regions or indirect effects mediated by other structures.

The cross-sectional design cannot reveal the timecourse over which visual experience shapes connectivity between infancy and adulthood.

Whether the infant resting-state patterns imply similar functional capacity to blind adults (e.g., cross-modal task responses) remains untested.

Comments on revisions:

The authors have done a fantastic job addressing my remaining questions.

Reviewer #2 (Public review):

Summary:

Tian et al. explore the developmental origins of cortical reorganization in blindness. Previous work has found that a set of regions in the occipital cortex show different functional responses and patterns of functional correlations in blind vs. sighted adults. Here, Tian et al. explore how this organisation arises over development, asking whether the infant brain looks more like the blind adult pattern, or more like the sighted adult pattern. Their analyses reveal that the answer depends on the particular networks investigated. Some functional connections in infants look more like blind than sighted adults; other functional connections look more like sighted than blind adults; and others fall somewhere in the middle, or show an altogether different pattern in infants compared with both sighted and blind adults.

Strengths:

The paper addresses very important questions about the "starting state" in the developing visual cortex, and how cortical networks are shaped by experience. Another clear strength lies in the unequivocal nature of many results. Many results have very large effect sizes, critical interactions between regions and groups are tested and found, and infant analyses are replicated in split halves of the data.

Weaknesses:

While potential roles of experience (e.g., visual, cross-modal) are discussed in detail, little consideration is given to the role of experience-independent maturation. The infants scanned are extremely young, only 2 weeks old. It is possible that the sighted adult pattern may still emerge later in infancy or childhood, regardless of infant visual experience. If so, the blind adult pattern may depend on blindness-related experience only (which may or may not reflect "visual" experience per se). In short, it is not clear that the age range studied is a clear-cut "starting point" for development, after which all change can be attributed to experience.

Reviewer #3 (Public review):

Summary

This study aimed to investigate whether the differences observed in the organization of visual brain networks between blind and sighted adults result from a reorganization of an early functional architecture due to blindness, or whether the early architecture is immature at birth and requires visual experience to develop functional connections. This question was investigated through the comparison of 3 groups of subjects with resting-state functional MRI (rs-fMRI). Based on convincing analyses, the study suggests that: 1) secondary visual cortices showed higher connectivity to prefrontal cortical regions (PFC) than to non-visual sensory areas (S1/M1 and A1) in infants like in blind adults, in contrast to sighted adults; 2) the V1 connectivity pattern of infants lies between that of sighted adults (showing stronger functional connectivity with non-visual sensory areas than with PFC) and that of blind adults (showing stronger functional connectivity with PFC than with non-visual sensory areas); 3) the laterality of the connectivity patterns of infants resembled those of sighted adults more than those of blind adults, but infants showed a less differentiated fronto-occipital connectivity pattern than adults.

Strengths

- The question investigated in this article is important for understanding the mechanisms of plasticity during typical and impaired development, and the approach considered, which compares different groups of subjects including, neonates/infants and blind adults, is highly original.

- Overall, the presented analyses are solid and well-detailed, and the results and discussion are convincing.

Weaknesses

- While it is informative to compare the "initial" state (close to birth) and the "final" states in blind and sighted adults to study the impact of post-natal and visual experience, this study does not analyze the chronology of this development and when the specialization of functional connections is completed. This would require investigating the evolution of functional connectivity of the visual system as a function of visual experience and thus as a function of age, at least during toddlerhood given the early and intense maturation of the visual system after birth. This could be achieved by analyzing different developmental periods using open databases such as the Baby Connectome Project.

- The rationale for grouping full-term neonates and preterm infants (scanned at term-equivalent age) is not understandable when seeking to perform comparisons with adults. Even if the study results do not show differences between full-terms and preterms in terms of functional connectivity differences between regions and of connectivity patterns, preterms group had different neurodevelopment and post-natal (including visual) experiences (even a few weeks might have an impact). And actually they show reduced connectivity strength systematically for all regions compared with full-terms (Sup Fig 7). Considering a more homogeneous group of neonates would have strengthened the study design.

- The rationale for presenting results on the connectivity of secondary visual cortices before the one of primary cortices (V1) could be clarified.

- The authors acknowledge the methodological difficulties for defining regions of interest (ROIs) in infants in a similar way as adults. Since the brain development is not homogeneous and synchronous across brain regions (in particular with the frontal and parietal lobes showing a delayed growth), this poses major problems for registration. This raises the question of whether the study findings could be biased by differences in ROI positioning across groups.

Comments on revisions:

The authors have addressed my specific recommendations, but some weaknesses in the study remain, particularly the inclusion of preterm infants alongside full-term neonates.

Author response:

The following is the authors’ response to the previous reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

The present study evaluates the role of visual experience in shaping functional correlations between human extrastriate visual cortex and frontal regions. The authors used fMRI to assess "resting-state" temporal correlations in three groups: sighted adults, congenitally blind adults, and neonates. Previous research has already demonstrated differences in functional correlations between visual and frontal regions in sighted compared to early blind individuals. The novel contribution of the current study lies in the inclusion of an infant dataset, which allows for an assessment of the developmental origins of these differences.

The main results of the study reveal that correlations between prefrontal and visual regions are more prominent in the blind and infant groups, with the blind group exhibiting greater lateralization. Conversely, correlations between visual and somato-motor cortices are more prominent in sighted adults. Based on these data, the authors conclude that visual experience plays an instructive role in shaping these cortical networks. This study provides valuable insights into the impact of visual experience on the development of functional connectivity in the brain.

Strengths:

The dissociations in functional correlations observed among the sighted adult, congenitally blind, and neonate groups provide strong support for the main conclusion regarding postnatal experience-driven shaping of visual-frontal connectivity.

The inclusion of neonates offers a unique and valuable developmental anchor for interpreting divergence between blind and sighted adults. This is a major advance over prior studies limited to adult comparisons.

Convergence with prior findings in the blind and sighted adult groups reinforces the reliability and external validity of the present results.

The split-half reliability analysis in the infant data increases confidence in the robustness of the reported group differences.

Weaknesses:

The manuscript risks overstating a mechanistic distinction between sighted and blind development by framing visual experience as "instructive" and blindness as "reorganizing." Similarly, the binary framing of visual experience and blindness as independent may oversimplify shared plasticity mechanisms.

The interpretation of changes in temporal correlations as altered neural communication does not adequately consider how shifts in shared variance across networks may influence these measures without reflecting true biological reorganization.

The discussion does not substantively engage with the longstanding debate over whether sensory experience plays an instructive or permissive role in cortical development.

The relationship between resting-state and task-based findings in blindness remains unclear.

Reviewer #2 (Public review):

Summary:

Tian et al. explore the developmental origins of cortical reorganization in blindness. Previous work has found that a set of regions in the occipital cortex show different functional responses and patterns of functional correlations in blind vs. sighted adults. Here, Tian et al. explore how this organization arises over development. Is the "starting state" more like the blind pattern, or more like the adult pattern? Their analyses reveal that the answer depends on the particular networks investigated. Some functional connections in infants look more like blind than sighted adults; other functional connections look more like sighted than blind adults; and others fall somewhere in the middle, or show an altogether different pattern in infants compared with both sighted and blind adults.

Strengths:

The paper addresses very important questions about the starting state in the developing visual cortex, and how cortical networks are shaped by experience. Another clear strength lies in the unequivocal nature of many results. Many results have very large effect sizes, critical interactions between regions and groups are tested and found, and infant analyses are replicated in split halves of the data.

Weaknesses:

While potential roles of experience (e.g., visual, cross-modal) are discussed in detail, little consideration is given to the role of experience-independent maturation. The infants scanned are extremely young, only 2 weeks old. It is possible then that the sighted adult pattern may still emerge later in infancy or childhood, regardless of infant visual experience. If so, the blind adult pattern may depend on blindness-related experience only (which may or may not reflect "visual" experience per se). In short, it is not clear that birth, or the first couple weeks of life, are a clear cut "starting point" for development, after which all change can be attributed to experience.

Reviewer #3 (Public review):

Summary

This study aimed to investigate whether the differences observed in the organization of visual brain networks between blind and sighted adults result from a reorganization of an early functional architecture due to blindness, or whether the early architecture is immature at birth and requires visual experience to develop functional connections. This question was investigated through the comparison of 3 groups of subjects with resting-state functional MRI (rs-fMRI). Based on convincing analyses, the study suggests that: 1) secondary visual cortices showed higher connectivity to prefrontal cortical regions (PFC) than to non-visual sensory areas (S1/M1 and A1) in infants like in blind adults, in contrast to sighted adults; 2) the V1 connectivity pattern of infants lies between that of sighted adults (showing stronger functional connectivity with non-visual sensory areas than with PFC) and that of blind adults (showing stronger functional connectivity with PFC than with non-visual sensory areas); 3) the laterality of the connectivity patterns of infants resembled those of sighted adults more than those of blind adults, but infants showed a less differentiated fronto-occipital connectivity pattern than adults.

Strengths

- The question investigated in this article is important for understanding the mechanisms of plasticity during typical and impaired development, and the approach considered, which compares different groups of subjects including, neonates/infants and blind adults, is highly original.

- Overall, the presented analyses are solid and well detailed, and the results and discussion are convincing.

Weaknesses

- While it is informative to compare the "initial" state (close to birth) and the "final" states in blind and sighted adults to study the impact of post-natal and visual experience, this study does not analyze the chronology of this development and when the specialization of functional connections is completed. This would require investigating the evolution of functional connectivity of the visual system as a function of visual experience and thus as a function of age, at least during toddlerhood given the early and intense maturation of the visual system after birth. This could be achieved by analyzing different developmental periods using open databases such as the Baby Connectome Project.

- The rationale for grouping full-term neonates and preterm infants (scanned at term-equivalent age) is not understandable when seeking to perform comparisons with adults. Even if the study results do not show differences between full-terms and preterms in terms of functional connectivity differences between regions and of connectivity patterns, preterms group had different neurodevelopment and post-natal (including visual) experiences (even a few weeks might have an impact). And actually they show reduced connectivity strength systematically for all regions compared with full-terms (Sup Fig 7). Considering a more homogeneous group of neonates would have strengthen the study design.

- The rationale for presenting results on the connectivity of secondary visual cortices before the one of primary cortices (V1) could be clarified.

- The authors acknowledge the methodological difficulties for defining regions of interest (ROIs) in infants in a similar way as adults. Since the brain development is not homogeneous and synchronous across brain regions (in particular with the frontal and parietal lobes showing a delayed growth), this poses major problems for registration. This raises the question of whether the study findings could be biased by differences in ROI positioning across groups.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

The authors are appropriately cautious in many parts of the discussion and include several helpful control analyses. Nonetheless, additional clarification of key assumptions and potential confounds would strengthen the paper.

(1) The current framing labels vision as "instructive" and blindness as "reorganizing," but it is unclear why these two experiential factors are characterized differently. Both involve activity-dependent changes to functional architecture from a shared immature scaffold. Labeling them differently risks conflating divergent outcomes with distinct underlying mechanisms. Just because visual and blind adults show different patterns of functional connectivity does not mean they reflect separate processes. While the discussion briefly acknowledges the possibility of shared plasticity mechanisms, much of the framing across the manuscript, including in the abstract and introduction, implies a dichotomy. A clearer articulation of the criteria used to assign these labels, or reconsideration of whether such a distinction is warranted, would improve conceptual clarity. The current framing appears analogous to saying that "heat causes expansion" and "cold causes contraction" as if these were separate mechanisms, when they are actually two directions of change along a single factor: temperature. A more parsimonious framework, such as activity-dependent reweighting of pre-existing connectivity, may better capture the nature of plasticity at play in both sighted and blind development.

Following the reviewer’s suggestion, we have revised the manuscript to clarify that both vision and blindness can be understood as manifestations of a common framework of experience-driven plasticity. We removed all mention of reorganization and clarify and modified the wording throughout.

Specifically:

Abstract: “Are infant visual cortices functionally like those of sighted adults, with blindness leading to functional change? We find that, on the contrary that secondary visual cortices of infants are functionally more like those of blind adults: stronger coupling with PFC than with nonvisual sensory-motor networks, suggesting that visual experience modifies elements of the sighted-adult long-range functional connectivity profile. Infant primary visual cortices are in-between blind and sighted adults i.e., more balanced PFC and sensory-motor connectivity than either adult group. The lateralization of occipital-to-frontal connectivity in infants resembles the sighted adults, consistent with the idea that blindness leads to functional change. These results suggest that both vision and blindness modify functional connectivity through experience-driven (i.e., activity-dependent) plasticity.” (Page 1, Line 13)

Introduction: We replaced “blindness leads to functional reorganization” with “blindness modifies this functional connectivity” (Page 2, Line 52), and the following sentence has also been modified to: “lifetime visual experience shapes connectivity toward the sighted-adult pattern” (Page 2, Line 54) For the lateralization patterns, we now describe them as “blindness-related modification” rather than “reorganization”, to keep the interpretation descriptive rather than mechanistic. (Page 4, Line 114),

(2) In interpreting the functional correlation differences, the discussion should more explicitly consider how statistical interdependence between areas could influence the observed results. For example, an increase in shared variance between visual and motor areas, such as might result from visually guided action, could result in a reduction in the apparent strength of visual-prefrontal temporal correlation (at the resolution of fMRI) without any true biological change in communication between visual-prefrontal cortex. This possibility is not ruled out by reporting groupwise patterns of relative connectivity. A more cautious systems-level framing could help clarify the distinction between neural plasticity and statistical redistribution of variance.

We thank the reviewer for raising this important point. We agree that resting-state fMRI provides a measure of statistical synchrony in BOLD signals rather than direct causal interactions between regions. This a fundamental limitation of resting state fMRI, which we now note in the Discussion section. Such changes in correlation are consistent with a variety of underlying biological mechanisms. Online task is one factor that influences cross-region correlations. In the current study, both blind and sighted groups were measured while blindfolded and were not performing visually guided actions during the resting state fMRI scans. It is possible that past visual-guided action experience changes the resting state correlations of sighted participants. Indeed, this is one interesting hypothesis.

In the revised Discussion, we now explicitly note this limitation and clarify that differences in FC do not by themselves establish whether or how underlying neurophysiological mechanisms are changed. We also emphasize that future work will need to investigate whether FC changes are accompanied by alterations in structural connectivity and to probe causal interactions and mechanistic underpinnings as follows：

“Resting-state functional connectivity captures synchrony in BOLD signal fluctuations rather than causal interactions and differences in functional connectivity cannot on their own reveal how underlying neurophysiological mechanisms are modified.” (page 13,line 342)

“Future studies will be needed to determine whether these functional changes are accompanied by alterations in structural connectivity, and to probe causal interactions and mechanistic underpinnings.” (page 13,line 350)

(3) The mechanistic interpretation of group differences in visual-motor coupling would benefit from stronger network-level justification. Direct connections between these areas are sparse in primates. If effects reflect indirect polysynaptic interactions or shared thalamic input, as the authors suggest, one might expect corresponding group differences in intermediate regions (e.g., parietal cortex, thalamus) that mediate these interactions. Is there any evidence for this in the data?

We thank the reviewer for raising this point. We agree and as noted above, resting state fMRI cannot distinguish between direct causal interactions between two regions and ones that a mediating region is involved. This is a fundamental limitation of resting state fMRI. The current study further focused on testing a specific hypothesis motivated by previously observed group differences between blind and sighted adults and our analyses focused on ROI-to-ROI connectivity between occipital, frontal, and sensory-motor cortices, and did not include these additional regions. In prior work, we and others, have looked at effects in parietal cortices (Abboud & Cohen, 2019; Bedny et al., 2009; Deen et al., 2015; Kanjlia et al., 2016, 2021; Sen et al., 2022). In blindness, parietal networks show increased correlations with some visual areas, rather than decreased. Regarding the thalamus, there is less clear evidence and there is some ongoing work trying to address this question. A couple of studies suggest that there is indeed increased connectivity between some parts of the thalamus and visual cortex in blindness. Although the anatomical information is limited, some of the work suggests that this increase is with higher-cognitive nuclei of the thalamus (Bedny et al., 2011; Liu et al., 2007).

We agree that this is an important direction for future work. To acknowledge this point, we have revised the manuscript to highlight the potential role of cortical and subcortical hub regions in mediating connectivity changes. The text has been modified as follows:

“Connectivity changes between two areas could be mediated by ‘third-party’ hub regions. For example, posterior parietal cortex serves as a cortical hub for multisensory integration and visuo-motor coordination and could mediate occipital-to-sensory-motor communication (Rolls et al., 2023; Sereno & Huang, 2014). Subcortical structures such as the thalamus could also play a mediating role (Vega-Zuniga et al., 2025).” (page 13,line 345)

(4) The discussion would benefit from deeper engagement with prior work on experience-dependent plasticity, particularly the longstanding distinction between instructive and permissive roles of experience. While the authors briefly define these concepts and reference their historical use, a more explicit consideration of how their findings relate to this broader literature would help clarify whether such distinctions are necessary or appropriate.

We thank the reviewer for this thoughtful suggestion to engage more explicitly with the longstanding literature on instructive versus permissive roles of experience. However, most of this literature comes from animal models, where experimental manipulations of the anatomical structure, of experience itself (e.g., controlled rearing studies) and sometimes of neural activity patterns allow clear tests of these mechanisms. Such manipulations are not feasible in humans. The terminology in the animal literature does not directly map onto the methods and data available in the present study or in other work with humans. For this reason, the current data does not allow us to fully engage with the debates in the animal literature and doing risks overinterpreting our findings.

Nevertheless, we agree that once the instructive/permissive framework has been introduced, it is important to clarify how our results relate to it, rather than only providing definitions. We have therefore added the following text to the discussion:

“In humans, such manipulations are not feasible, leaving us to study only the consequences of the presence or absence of vision. Under an instructive account, visual and multisensory experience could strengthen coupling between visual and other non-visual sensory-motor cortices through coordinated activity, thereby establishing the sighted-adult connectivity pattern. In the absence of visual input, by contrast, the lack of such coordinated activity may prevent these couplings from being established. Alternatively, vision may act permissively, indirectly enabling maturational processes that shift connectivity toward the sighted-adult configuration.” (page 14,line 362)

(5) The revised discussion acknowledges the divergence between resting-state and task-based findings, but does not fully frame the theoretical implications of this discrepancy. Although this study cannot resolve the issue with its own data, a more integrative discussion could help clarify whether these measures reflect distinct functional states, developmental trajectories, or mechanisms of plasticity. Without such framing, readers are left without clear guidance on how to reconcile the present results with prior work on cross-modal recruitment in blindness.

We thank the reviewer for this thoughtful comment. We agree that know how resting-state evidence relates to task-based evidence is a fundamentally important issue. We now discuss this more in the Introduction as well as in the Discussion.

There is a sizable literature of both task-based and resting state studies. Some of prior studies have measured resting state and task-based data within the same participants and found relationships (Kanjlia et al., 2016, 2021; Lane et al., 2015). We now clarify this in the introduction. These studies find that within visual cortices of blind people, the task-based profile of a cortical area is related to its resting state connectivity pattern (Abboud & Cohen, 2019; Deen et al., 2015; Kanjlia et al., 2016, 2021). This suggests that these two measures are related. However, the timecourse of this relationship, the developmental trajectory and mechanism of plasticity is not known. We note this now in the introduction on page 2. Primarily this is because there is very little relevant developmental evidence. For example, in the current study we find that the resting state profile of secondary visual networks in infants is similar to that of blind adults. However, we do not know whether the visual cortices of infants show task-based cross modal responses. To our knowledge nobody has tested this question. We agree with the reviewer that raising this question in the paper is better than not commenting on the relationship at all.

To address the reviewer’s comment, we have expanded the discussion to situate our results within a developmental framework, highlighting how early intrinsic connectivity may scaffold alternative trajectories shaped by either visual experience or blindness. The revised text now reads as follows:

“Conversely, for people who remain blind throughout life, visual-PFC connectivity could enable recruitment of visual cortices for higher-order non-visual functions, such as language and executive control (Bedny et al., 2011; Kanjlia et al., 2021). Our results suggest that blind adults may build on connectivity patterns already present in infancy: like blind adults, sighted infants show stronger occipital–PFC than occipital–sensory–motor coupling. Repeated engagement of occipital networks during higher cognitive tasks in early development could intern enhance connectivity and specialization of visual networks for non-visual higher-order functions.

Some prior studies have measured resting-state and task-based functional profiles in the same participants. These studies find that within visual cortices of blind people, the task-based profile of a cortical area is related to its resting state connectivity pattern (citations.) This suggests that these two measures are related. However, the timecourse of this relationship, the developmental trajectory and mechanism of plasticity is not known. Primarily this is because there is very little relevant developmental evidence. For example, in the current study we find that the resting state profile of secondary visual networks in infants is similar to that of blind adults. However, we do not know whether the visual cortices of infants show enhanced task-based cross modal responses, relative to sighted adults and how this compares to responses observed in blind adults. Future work with infants and children would be able to address this question.

In the current study, the clearest evidence for functional change driven by blindness was observed for laterality. Connectivity lateralization in sighted infants resembles that of sighted adults, in both V1 and secondary visual cortices. Relative to both sighted infants and sighted adults, blind adults show more lateralized connectivity patterns between occipital and prefrontal cortices. Previous studies suggest that in people born blind occipital and non-occipital language responses are co-lateralized (Lane et al., 2017; Tian et al., 2023). We speculate that habitual activation of visual cortices by higher-cognitive tasks, such as language, which are themselves highly lateralized, contributes to this biased connectivity pattern of occipital cortex in blindness. Taken together, these results suggest a developmental framework in which intrinsic connectivity present in infancy provides a scaffold that is subsequently shaped and reinforced by experience-dependent recruitment, through either visual experience or the lifelong absence of vision in blindness. Longitudinal work across successive developmental stages will be crucial to test how the alternative trajectories shaped by visual experience versus blindness unfold over development.” (page 14-15)

(6) The split-half reliability analysis is a valuable control. Additional details would clarify what these noise ceilings reflect. Were the rsFC patterns for each ROI calculated only for the ROIs included in the current study or was a broader assessment across the whole brain performed? It also would be helpful to report whether reliability differed for individual ROIs within and between groups. Even if global reliability is matched, selective differences could influence group comparisons. Several infants in the dhcp dataset were scanned twice. Were any second scans included in the current analyses? Comparing first versus second scans directly could strengthen the claim that several weeks of visual experience are insufficient to shift connectivity toward a sighted adult profile.

Thanks to the reviewer’s comments on the reliability of the current study.

In the present study, the noise ceiling was computed from the reliability of the ROI-wise FC profiles used across all analyses. Reliability was estimated using a split-half procedure: each rs-fMRI time series was divided into two equal halves, FC among all ROIs included in the study was computed separately for each half, and the noise ceiling for each ROI was defined as the Pearson correlation between its two FC profiles. Then we averaged these ROI-wise noise ceilings to evaluate group-level reliability, which exceeded 0.70 in all three groups and found no significant difference across groups. This provides an estimate of the upper bound on explainable variance for the exact FC features subjected to statistical testing (Lage-Castellanos et al., 2019). A brief description has been added to the manuscript (page 19, line 518).

Regarding the reviewer’s question about the scope of rsFC features used in the noise-ceiling analysis: we computed noise ceilings only for the ROIs included in the present study, because all analyses in this work were conducted at the ROI–ROI level and did not involve voxelwise whole-brain FC. Thus, the noise-ceiling estimates correspond directly to the full set of FC features on which all statistical comparisons were based.

As suggested by the reviewer, we examined noise ceilings for each ROI separately. All ROIs showed high absolute reliability (noise ceiling > 0.80) across the three groups, indicating that the ROI-wise FC estimates are generally robust across participants. Although many ROIs exhibited statistically significant group differences in noise ceiling (one-way ANOVA, p < 0.05), the effect sizes were small to moderate (partial η² < 0.14). These differences indicate that reliability may vary modestly across groups at the ROI level, and we cannot fully determine whether such variability contributes to the observed different FC patterns across groups. We have included this point in the revised manuscript (page 19, line 525), along with the full statistical results for the ROI-wise noise ceilings in the Supplementary Table S2.

Last, we fully agree that longitudinal comparisons across multiple time points can provide important insights into how early visual experience shapes connectivity. At the same time, in the present dataset, the first scan occurred at a preterm age and the second at term-equivalent age. The differences between the first and second scans would reflect not only additional weeks of visual input, but also differences in prematurity status and overall neurodevelopmental maturity, which would make the interpretation of such comparisons difficult in the context of our current aims. We have clarified in the revised manuscript that only term-equivalent (second) scans were included. We see careful longitudinal work as an important avenue for addressing this question more directly.

(7) The signal dropout assessment in the infant dataset is a valuable quality control step. Applying the same metric to the adult datasets would help harmonize preprocessing across groups and increase confidence in group-level comparisons.

Thank you for this valuable suggestion. Following your comment, we applied the same signal dropout assessment to the adult datasets. One participant in the sighted adult group and two participants in the blind adult group showed signal dropout in one ROI each. The corresponding results are now included in the Supplementary Materials (Figure S13). The findings remain unchanged after this additional control analysis. We also add the relevant content in the Method part as follows:

“The same signal dropout assessment was also applied to the blind and sighted adults to ensure consistent quality control across groups. One participant in the sighted adult group and two participants in the blind adult group exhibited signal dropout in one ROI each. Excluding these participants did not alter the group-level results (see Figure S13).” (page 16, line 449)

Minor:

(8) The authors added accurate anatomical descriptions to the methods but a less precise characterization remains in the introduction: "Anatomically, these regions correspond roughly to the location of areas such as motion area V5/MT+, the lateral occipital complex (LO), V3a and V4v in sighted people."

We thank the reviewer for this helpful comment. We have revised the Introduction to provide a fuller anatomical description, consistent with the Methods. The text now reads:

“Anatomically, these regions in sighted people approximately correspond to the locations of motion-sensitive V5/MT+ and the lateral occipital complex (LO), as well as ventral portions of occipito-temporal cortex including V4v and dorsal portions including V3a. The occipital ROI also extends ventrally into the middle portion of the ventral temporal lobe and dorsally into the intraparietal sulcus and superior parietal lobule.” (page 3, line 88)

(9)Typo: "lager effect" should be "larger effect."

Secondary visual cortices showed a significant within > between difference in both groups, with a lager effect in the blind group (post-hoc tests, Bonferroni-corrected paired: t-test: sighted adults within hemisphere > between hemisphere: t (49) = 7.441, p = 0.012; blind adults within hemisphere > between hemisphere: t (29) = 10.735, p < 0.001; V1: F(1, 78) =87.211, p < 0.001).

We thank the reviewer for catching this typo. We have corrected “lager effect” to “larger effect” in the revised manuscript. (page 9, line 214)

Reviewer #2 (Recommendations for the authors):

All of my other concerns were adequately addressed.

We thank the reviewer for their positive evaluation, and we are glad that our revisions have addressed their concerns.

Reviewer #3 (Recommendations for the authors):

In my view, qualifying infants as "sighted" is confusing and unnecessary: why not simplifying and homogenizing the wording along the manuscript and figures?

We thank the reviewer for this suggestion. We agree and have revised the manuscript to use consistent wording, avoiding the qualification of infants as “sighted.”

l188, I don't understand the sentence "By contrast, in sighted adults, this cross-hemisphere difference is weak or absent."

We thank the reviewer for noting that this sentence was unclear. We have revised the text to provide a more precise explanation. The text now reads:

“By contrast, in sighted adults this lateralized pattern is weaker: visual areas in each hemisphere show only a modest preference for ipsilateral prefrontal cortices, and connectivity with the contralateral PFC remains comparatively strong.” (page 8, line 207)

l193: "Secondary visual cortices showed a significant within > between difference in both groups, with a lager effect in the blind group": providing effect sizes for the 2 groups would strengthen this result (+ note the typo laRger).<br /> - Figure S7, S11: Please add titles of y-axes.

Thank you for this helpful suggestion. We have corrected the typo and added the effect sizes for both groups in the revised text. The revised sentence now reads as follows:

“Secondary visual cortices showed a significant within > between difference in both groups, with a larger effect in the blind group (post-hoc tests, Bonferroni-corrected paired: t-test: sighted adults within hemisphere > between hemisphere: t (49) = 7.441, p = 0.012, cohen’d = 0.817; blind adults within hemisphere > between hemisphere: t (29) = 10.735, p < 0.001, cohen’d = 1.96).” (page 9, line 214)

Titles of the y-axes have also been added to Figures S7 and S11.

eLife Assessment

This important work describes wing mechanosensory neurons in detail, extending our understanding of sensorimotor processing in the fruit fly. The evidence presented convincingly supports the authors' identification of these neurons and leverages state-of-the-art methods to generate a near-complete map of wing mechanosensory circuitry. Overall, this study provides new hypotheses and invaluable tools for investigating proprioceptive motor control of the wing in Drosophila.

Reviewer #1 (Public review):

Summary:

Lesser et al provide a comprehensive description of Drosophila wing proprioceptive sensory neurons at the electron microscopy resolution. This "tour-de-force", provides a strong foundation for future structural and functional research aimed at understanding wing motor control in Drosophila with implications to understanding wing control across other insects.

Strengths:

(1) Authors leverage previous research that described many of the fly wing proprioceptors, and combine this knowledge with EM connectome data such that they now provide a near-complete morphological description of all wing proprioceptors.

(2) Authors cleverly leverage genetic tools and EM connectome data to tie the location of proprioceptors on the wings with axonal projections in the connectome. This enables them to both align with previous literature as well as make some novel claims.

(3) In addition to providing a full description of wing proprioceptors, authors also identified a novel population of sensors on the wing tegula that make direct connections with the B1 wing motor neurons implicating the role of tegula in wing movements that was previously underappreciated.

(4) Despite being the most comprehensive description so far, it is reassuring that authors clearly state the missing elements in the discussion.

Weaknesses:

(1) Authors do their main analysis on data from FANC connectome but provide corresponding IDs for sensory neurons in the MANC connectome. I wonder how the connectivity matrix compares across FANC and MANC if the authors perform similar analysis as they have done in Fig. 2. This could be a valuable addition and potentially also pick up any sexual dimorphism.

(2) Authors speculate about presence of gap junctions based on density of mitochondria. I'm not convinced about this given mitochondrial densities could reflect other things that correlate with energy demands in sub-compartments.

Overall, I consider this an exceptional analysis which will be extremely valuable to the community.

Reviewer #2 (Public review):

Summary:

Lesser et al. present an atlas of Drosophila wing sensory neurons. They proofread the axons of all sensory neurons in the wing nerve of an existing electron microscopy dataset, the female adult fly nerve cord (FANC) connectome. These reconstructed sensory axons were linked with light microscopy images of full-scale morphology to identify their origin in the periphery of the wing and encoded sensory modalities. The authors described the morphology and postsynaptic targets of proprioceptive neurons as well as previously unknown sensory neurons.

Strengths:

The authors present a valuable catalogue of wing sensory neurons, including previously undescribed sensory axons in the Drosophila wing. By providing both connectivity information with linked genetic drive lines, this research facilitates future work on the wing motor-sensory network and applications relating to Drosophila flight. The findings were linked to previous research as well as their putative role in the proprioceptive and nerve cord circuitry, providing testable hypotheses for future studies.

Weaknesses:

With future use as an atlas, it should be noted that the evidence is based on sensory neurons on only one side of the nerve cord. Fruit flies have stereotyped left/right hemispheres in the brain and left/right hemisegments in the nerve cord. Comparison of left and right neurons of the nervous system can give a sense of how robust the morphological and connectivity findings are. Unfortunately, this dataset has damage to the right side, making such comparisons unreliable.

Reviewer #3 (Public review):

Summary:

The authors aim to identify the peripheral end organ origin in the fly's wing of all sensory neurons in the Anterior Dorsal Mesothoracic nerve. They reconstruct the neurons and their downstream partners in an electron microscopy volume of a female ventral nerve cord, analyse the resulting connectome and identify their origin with review of the literature and imaging of genetic driver lines. While some of the neurons were already known through previous work, the authors expand on the identification and create a near complete map of the wing mechanosensory neurons at synapse resolution.

Strengths:

The authors elegantly combine electron microscopy neuron morphology, connectomics and light microscopy methods to bridge the gap between fly wing sensory neuron anatomy and ventral nerve cord morphology. Further, they use EM ultrastructural observations to make predictions on the signaling modality of some of the sensory neurons and thus their function in flight.

The work is as comprehensive as state of the art methods allow to create a near complete map of the wing mechanosensory neurons. This work will be of importance to the field of fly connectomics and modelling of fly behavior as well as a useful resource to the Drosophila research community.

Through this comprehensive mapping of neurons to the connectome the authors create a lot of hypotheses on neuronal function partially already confirmed with the literature and partially to be tested in the future. The authors achieved their aim of mapping the periphery of the fly's wing to axonal projections in the ventral nerve cord, beautifully laying out their results to support their mapping.

The authors identify the neurons in a previously published connectome of a male fly ventral nerve cord to enable cross-individual analysis of connections and find no indication of sexual dimorphism at the sensory neuron level. Further, together with their companion paper Dhawan et al., 2025 describing the haltere sensory neurons in the same EM dataset, they cover the entire mechanosensory space involved in Drosophila flight.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary:

Lesser et al provide a comprehensive description of Drosophila wing proprioceptive sensory neurons at the electron microscopy resolution. This “tour-de-force” provides a strong foundation for future structural and functional research aimed at understanding wing motor control in Drosophila with implications for understanding wing control across other insects.

Strengths:

(1) The authors leverage previous research that described many of the fly wing proprioceptors, and combine this knowledge with EM connectome data such that they now provide a near-complete morphological description of all wing proprioceptors.

(2) The authors cleverly leverage genetic tools and EM connectome data to tie the location of proprioceptors on the wings with axonal projections in the connectome. This enables them to both align with previous literature as well as make some novel claims.

(3) In addition to providing a full description of wing proprioceptors, the authors also identified a novel population of sensors on the wing tegula that make direct connections with the B1 wing motor neurons, implicating the role of the tegula in wing movements that was previously underappreciated.

(4) Despite being the most comprehensive description so far, it is reassuring that the authors clearly state the missing elements in the discussion.

Weaknesses:

(1) The authors do their main analysis on data from the FANC connectome but provide corresponding IDs for sensory neurons in the MANC connectome. I wonder how the connectivity matrix compares across FANC and MANC if the authors perform a similar analysis to the one they have done in Figure 2. This could be a valuable addition and potentially also pick up any sexual dimorphism.

We agree that systematic comparisons will provide valuable insights as more connectome datasets become available. However, the primary goal of this study was to link central axon morphology with peripheral structures in the wing. We deliberately omitted more detailed and quantitative analyses of the downstream VNC circuitry, apart from providing a global view of the connectivity matrix and using it to cluster the sensory axon types. A more detailed and systematic comparison of wing sensorimotor circuit connectivity across different connectome datasets (FANC, MANC, BANC, IMAC) is the subject of ongoing work in our lab, which we feel is beyond the scope of this study. Here, we chose to match the wing proprioceptors to axons in MANC to demonstrate their stereotypy across individuals and to make them more accessible to other researchers. We found no obvious sexual dimorphism at the level of wing sensory neurons. We now note this in the Discussion.

(2) The authors speculate about the presence of gap junctions based on the density of mitochondria. I’m not convinced about this, given that mitochondrial densities could reflect other things that correlate with energy demands in sub-compartments.

We have moved speculation about mitochondria and gap junctions to the Discussion.

(3) I’m intrigued by how the tegula CO is negative for iav. I wonder if authors tried other CO labeling genes like nompc. And what does this mean for the nature of this CO. Some more discussion on this anomaly would be helpful.

Based on this suggestion, we have added an image showing that tegula CO neurons are labeled by nompC-Gal4.

(4) The authors conclude there are no proprioceptive neurons in sclerite pterale C based on Chat-Gal4 expression analysis. It would be much more rigorous if authors also tried a pan-neuronal driver like nsyb/elav or other neurotransmitter drivers (Vglut, GAD, etc) to really rule this out. (I hope I didn’t miss this somewhere.)

To address this, we imaged OK371-GFP, which labels glutamatergic neurons, in the wing and wing hinge. We saw expression in the wing, as others have reported (Neukomm et. al., 2014), but we saw no expression at the wing hinge. Apart from a handful of glutamatergic gustatory neurons in the leg, we are not aware of any other sensory neurons in the fly that are not labeled by Chat-Gal4.

Overall, I consider this an exceptional analysis that will be extremely valuable to the community.

We sincerely appreciate the reviewer’s positive feedback.

Reviewer #2 (Public review):

Summary:

Lesser et al. present an atlas of Drosophila wing sensory neurons. They proofread the axons of all sensory neurons in the wing nerve of an existing electron microscopy dataset, the female adult fly nerve cord (FANC) connectome. These reconstructed sensory axons were linked with light microscopy images of full-scale morphology to identify their origin in the periphery of the wing and encoded sensory modalities. The authors described the morphology and postsynaptic targets of proprioceptive neurons as well as previously unknown sensory neurons.

Strengths:

The authors present a valuable catalogue of wing sensory neurons, including previously undescribed sensory axons in the Drosophila wing. By providing both connectivity information with linked genetic drive lines, this research facilitates future work on the wing motor-sensory network and applications relating to Drosophila flight. The findings were linked to previous research as well as their putative role in the proprioceptive and nerve cord circuitry, providing testable hypotheses for future studies.

Weaknesses:

(1) With future use as an atlas, it should be noted that the evidence is based on sensory neurons on only one side of the nerve cord. Fruit flies have stereotyped left/right hemispheres in the brain and left/right hemisegments in the nerve cord. The comparison of left and right neurons of the nervous system can give a sense of how robust the morphological and connectivity findings are. Here, the authors have not compared the left and right side sensory axons from the wing nerve, leaving potential for developmental variability across samples and left/right hemisegments.

The right ADMN nerve in the FANC dataset is partially severed, making left/right comparisons unreliable (see Azevedo 2024, Extended Data Figure 4). We have updated the text to explain this within the Methods section of the paper.

(2) Not all links between the EM reconstructions and driver lines are convincing. To strengthen these, for all EM-LM matches in Figures 3-7, rotated views of the driver line (matching the rotated EM views) should be shown to provide a clearer comparison of the data. In particular, Figure 3G and Figure 7B are not very convincing based on the images shown. MCFO imaging of the driver lines in Figure 3G and 7B would make this position stronger if a clone that matches the EM reconstruction could be identified.

Many of the z-stack images in the paper are from the Janelia FlyLight collection, and unfortunately their imaging parameters were not optimized for orthogonal views. Rotated views are blurry and not especially helpful for comparison to EM reconstruction. We now point out in the text that interested readers can access the z-stacks from FlyLight to see the dorsal-ventral projections.

Regarding Figure 3G and 7B, we have added markers to the image with corresponding descriptions in the legend to guide the reader through the image of the busy driver line. Although these lines label many cells in the VNC as a whole, they sparsely label cells in the ADMN, making them nonetheless useful for identifying peripheral sensory neurons.

(3) Figure 7B looks like the driver line might have stochastic expression in the sensory neuron, which further reduces confidence in the result shown in Figure 7C. Is this expression pattern in the wing consistently seen? Many split-GAL4s have stochastic expressions. The evidence would be strengthened if the authors presented multiple examples (~4-5) of each driver line’s expression pattern in the supplement.

Figure 7B shows sparse labeling of the driver line using the MCFO technique, as specified in the legend. Its unilateral expression is therefore not due to stochastic expression of the Gal4 line. We have added the “MFCO” label to the image to clarify.

(4) Certain claims in this work lack quantitative evidence. On line 128, for instance, “Overall, our comprehensive reconstruction revealed many morphological subgroups with overlapping postsynaptic partners, suggesting a high degree of integration within wing sensorimotor circuits.” If a claim of subgroups having shared postsynaptic partners is being made, there should have been quantitative evidence. For example, cosine similar amongst members of each group compared to the cosine similarity of shuffled/randomised sets of axons from different groups. The heat map of cosine similarity in Figure 2B alone is not sufficient.

We agree that illustrating the extent of shared postsynaptic partners across subgroups strengthens this point. We added a visualization showing pairwise similarity scores for within- and between-cluster neuron pairs (Figure 2B inset). We also performed a permutation test to determine that within-cluster similarity is significantly higher than between clusters, and we report the test in the results as well as the figure legend. This analysis provides a more quantitative summary of the qualitative trends in connectivity that are summarized in Figure 2B.

(5) Similarly, claims about putative electrical connections to b1 motor neurons are very speculative. The authors state that “their terminals contain very densely packed mitochondria compared to other cells”, without providing a quantitative comparison to other sensory axons. There is also no quantitative comparison to the one example of another putative electrical connection from the literature. Further, it should be noted that this connection from Trimarchi and Murphey, 1997, is also stated as putative on line 167, which further weakens this evidence. Quantification would strongly strengthen this position. Identification of an example of high mitochondrial density at a confirmed electrical connection would be even better. In the related discussion section “A potential metabolic specialization for flight circuitry”, it should be more clearly noted that the dense mitochondria could be unrelated to a putative electrical connection. If the authors have an alternative hypothesis about the mitochondria density, this should be stated as well.

We agree with the reviewer that the link between mitochondrial density and metabolic specialization is purely speculative in this context. Based on reviewer feedback, we have moved all mention of the relationship between mitochondrial density and gap junction coupling to the Discussion. We acknowledge that this may seem like a somewhat random and not quantitatively supported observation. However, we found the coincidence striking and worthy of mention, though it is only tangentially relevant to the rest of the paper. From conversations with colleagues, we have also heard that this relationship is consistent with as yet unpublished work in other model organisms (e.g., zebrafish, mouse).

The electrical coupling to b1 motor neurons is well-established (Fayyazuddin and Dickinson, 1999), and we have updated the text to state this more clearly. However, we agree that whether the specific neurons we have identified based on their anatomy are the same ones functionally identified through whole-nerve recordings remains unknown.

(6) It would be appropriate to cite previous work using a similar strategy to match sensory axons to their cell bodies/dendrites at the periphery using driver lines and connectomics (see Figure 5 for example in the following paper: https://doi.org/10.7554/eLife.40247 ).

At this point, there are now dozens of papers that match the axons of sensory neurons to their cell bodies/dendrites in the periphery by comparing light microscopy and connectomics. When we dug in, we found examples in C. elegans, Ciona intestinalis, zebrafish, and mouse, all published prior to the study cited above. For basically every animal for which scientists have acquired EM volumes of neural tissue, they have used other anatomical labeling methods to determine cell types inside and outside the imaged volume. In summary, we found it difficult to establish a single primary citation for this approach. In lieu of this, we have added a citation to an earlier review by a pioneer in EM connectomics that discusses the general approach of matching cells across different labeling/imaging modalities (Meinertzhagen et al., 2009).

The methods section is very sparse. For the sake of replicability, all sections should be expanded upon.

We have expanded the methods section, and also a STAR methods table.

Reviewer #3 (Public review):

Summary:

The authors aim to identify the peripheral end-organ origin in the fly’s wing of all sensory neurons in the anterior dorsomedial nerve. They reconstruct the neurons and their downstream partners in an electron microscopy volume of a female ventral nerve cord, analyse the resulting connectome, and identify their origin with a review of the literature and imaging of genetic driver lines. While some of the neurons were already known through previous work, the authors expand on the identification and create a near-complete map of the wing mechanosensory neurons at synapse resolution.

Strengths:

The authors elegantly combine electron microscopy, neuron morphology, connectomics, and light microscopy methods to bridge the gap between fly wing sensory neuron anatomy and ventral nerve cord morphology. Further, they use EM ultrastructural observations to make predictions on the signaling modality of some of the sensory neurons and thus their function in flight.

The work is as comprehensive as state-of-the-art methods allow to create a near-complete mapof the wing mechanosensory neurons. This work will be of importance to the field of fly connectomics and modelling of fly behavior, as well as a useful resource to the Drosophila research community.

Through this comprehensive mapping of neurons to the connectome, the authors create a lot of hypotheses on neuronal function, partially already confirmed with the literature and partially to be tested in the future. The authors achieved their aim of mapping the periphery of the fly’s wing to axonal projections in the ventral nerve cord, beautifully laying out their results to support their mapping.

The authors identify the neurons in a previously published connectome of a male fly ventral nerve cord to enable cross-individual analysis of connections. Further, together with their companion paper, Dhawan et al. 2025, describing the haltere sensory neurons in the same EM dataset, they cover the entire mechanosensory space involved in Drosophila flight.

Weaknesses:

The connectomic data are only available upon request; the inclusion of a connectivity table of the reconstructed neurons would aid analysis reproducibility and cross-dataset comparisons.

We have added a connectivity table as well as analysis scripts in the github repository for the paper (https://github.com/EllenLesser/Lesser_eLife_2025).

Recommendations for the authors:

Reviewer #2 (Recommendations for the authors):

The methods section should be expanded in every aspect. Most pressing sections are:

(1) Data and Code availability: All code should be included as a Zenodo database, the suggestion to ask authors for code upon request is inappropriate.

We have added all code to a public github repository, which is now linked in the Methods section.

(2) Samples: Standard cornmeal and molasses medium should have a reference, as many institutes use different recipes.

The recipe used by the University of Washington fly kitchen is based on the Bloomington standard Cornmeal, Molasses and Yeast Medium recipe, which can be found at https://bdsc.indiana.edu/information/recipes/molassesfood.html. The UW recipe is slightly modified for different antifungal ingredients and includes tegosept, propionic acid, and phosophoric acid.

(3) Table 3: Driver lines labelling wing sensory neurons: The genetic driver lines should have associated Bloomington stock centre numbers. Additionally, relevant information for effector lines used should be included in the methods.

We now include the Bloomington stock numbers and more information on effector lines in the STAR methods table.

Minor corrections:

(1) Lines 119-120: “Notably, many of the axons do not form crisp cluster boundaries, suggesting that multimodal sensory information is integrated at early stages of sensory processing.” We do not follow the logic of this statement and suspect it is a bit too speculative.

We removed this sentence from the manuscript.

(2) Figure 1: The ADMN is missing in the schematics and would be helpful to depict for non-experts. Is this what is highlighted in Figure 1D?

Yes, and we now label 1D as the ADMN wing nerve.

(3) Figure 1B: Which driver lines are being depicted here? Looking at Table 3 does not clarify. It should be specified at least in the figure legend.

As stated in the legend, we include a table of all of the driver lines we screened and which sensory structures they label.

(4) Figure 1C: There are some minor placement issues with the text in the schematic. There is an arrow very close to the “CO” on the top right, which makes the “O” look like the symbol for male. “ax ii” is a bit too close to the wing hinge

We updated the figure to address this issue.

(5) Figure 1D: The outlined grey masks are not clear. The use of colour would be very useful for the reader to help understand what the authors are referring to here

We now use color for the masks.

(6) Figure 2A: It is unclear if the descending neuron and non-motor efferent neuron are not shown because they are under the described threshold, or to simplify the plot. They should be included in the plot if over the threshold.

We have updated the legend to specify that the exclusion of the descending and non-motor efferent neurons are to visually simplify the plot. We include % of sensory output to each of these neurons in the legend, and they are included in the connectivity matrix data in the public  GitHub repository associated with the paper, included in the Methods.

(7) Figure 2B: What clustering is used specifically? The method says it’s from Scikit-learn, but there are many types of clustering available in this package.

We now include the specific clustering type used in the Methods section, which is agglomerative clustering.

(8) Figure 3A: What does the green box behind the plot represent?

The green box represents the tegula CO axons, which we now specify in the legend.

(9) Figure 3C: the “C” is clipped at the top.

We updated the figure to address this issue.

(10) Figure 4A: the main text says a “group of four axons” (line 203) while the figure says 5 axons.

We updated the text to address this issue.

(11) Line 360: “We found that the campaniform sensilla on the tegula provide the most direct feedback onto wing steering motor neurons”. We struggled to find where this was directly shown, because several sensory axon types directly synapse onto motor neurons.

We now specify in the text that this finding is shown in Figure 3.

Reviewer #3 (Recommendations for the authors):

I would like to congratulate the authors on their beautiful, easy-to-read, and easy-to-comprehend manuscript, with clear figures and nice visualizations. This work provides a valuable resource that will contribute to the interpretability of connectomic data and further to connectome-based modeling of fly behavior.

We sincerely appreciate the reviewer’s positive feedback.

eLife Assessment

This important work examines the effects of side-wall confinement on chemotaxis of swimming bacteria in a shallow microfluidic channel. The authors present convincing experimental evidence, combined with geometric analysis and numerical simulations of simplified models, showing that chemotaxis is enhanced when the distance between the side walls is comparable to the intrinsic radius of chiral circular swimming near open surfaces. This study should be of interest to scientists specializing in bacteria-surface interactions.

Reviewer #1 (Public review):

The authors show experimentally that, in 2D, bacteria swim up a chemotactic gradient much more effectively when they are in the presence of lateral walls. Systematic experiments identify an optimum for chemotaxis for a channel width of ~8µm, a value close to the average radius of the circle trajectories of the unconfined bacteria in 2D. These chiral circles impose that the bacteria swim preferentially along the right-side wall, which indeed yields chemotaxis in the presence of a chemotactic gradient. These observations are backed by numerical simulations and a geometrical analysis.

Reviewer #3 (Public review):

This paper addresses, through experiment and simulation, the combined effects of bacterial circular swimming near no-slip surfaces and chemotaxis in simple linear gradients. The authors have constructed a microfluidic device in which a gradient of L-aspartate is established, to which bacteria respond while swimming while confined in channels of different widths. There is a clear effect that the chemotactic drift velocity reaches a maximum in channel widths of about 8 microns, similar in size to the circular orbits that would prevail in the absence of side walls. Numerical studies of simplified models confirm this connection.

The experimental aspects of this study are well executed. The design of the microfluidic system is clever in that it allows a kind of "multiplexing" in which all the different channel widths are available to a given sample of bacteria.<br /> The authors have included a useful intuitive explanation of their results via a geometric model of the trajectories. In future work it would be interesting to analyze further the voluminous data on the trajectories of cells by formulating the mathematical problem in terms of a suitable Fokker-Planck equation for the probability distribution of swimming directions. In particular, this might help understand how incipient circular trajectories are interrupted by collisions with the walls and how this relates to enhanced chemotaxis.

The authors argue that these findings may have relevance to a number of physiological and ecological contexts. As these would be characterized by significant heterogeneity in pore sizes and geometries, further work will be necessary to translate the present results to those situations.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

This article deals with the chemotactic behavior of E coli bacteria in thin channels (a situation close to 2D). It combines experiments and simulations.

The authors show experimentally that, in 2D, bacteria swim up a chemotactic gradient much more effectively when they are in the presence of lateral walls. Systematic experiments identify an optimum for chemotaxis for a channel width of ~8µm, close to the average radius of the circle trajectories of the unconfined bacteria in 2D. It is known that these circles are chiral and impose that the bacteria swim preferentially along the right-side wall when there is no chemotactic gradient. In the presence of a chemotactic gradient, this larger proportion of bacteria swimming on the right wall yields chemotaxis. This effect is backed by numerical simulations and a geometrical analysis.

If the conclusions drawn from the experiments presented in this article seem clear and interesting, I find that the key elements of the mechanism of this wall-directed chemotaxis are not sufficiently emphasized. Moreover, the paper would be clearer with more details on the hypotheses and the essential ingredients of the analyses.

We thank the reviewer for these constructive suggestions. We agree that emphasizing the underlying mechanism is crucial for the clarity of our findings. In the revised manuscript, we have now explicitly highlighted the critical roles of chiral circular motion and the alignment effect following side-wall collisions in both the Abstract (lines 25-27) and the Discussion (lines 391-393). Furthermore, we have added a new analysis of bacterial trajectories post-collision (Fig. S2), which demonstrates that cells predominantly align with and swim along the sidewalls. We have also clarified the assumptions in our numerical simulations, specifically how the radius of circular trajectories and the alignment effect are incorporated into the equations of motion. Please refer to our detailed responses in the "Recommendations for the authors" section for further specifics.

Reviewer #2 (Public review):

Summary:

In this study, the authors investigated the chemotaxis of E. coli swimming close to the bottom surface in gradients of attractant in channels of increasingly smaller width but fixed height = 30 µm and length ~160 µm. In relatively large channels, they find that on average the cells drift in response to the gradient, despite cells close to the surface away from the walls being known to not be chemotactic because they swim in circles.

They find that this average drift is due to the cell localization close to the side walls, where they slide along the wall. Whereas the bacteria away from the walls have no chemotaxis (as shown before), the ones on the left side wall go down-gradient on average, but the ones on the right-side wall go up-gradient faster, hence the average drift. They then study the effect of reducing channel width. They find that chemotaxis is higher in channels with a width of about 8 µm, which approximately corresponds to the radius of the circular swimming R. This higher chemotactic drift is concomitant to an increased density of cells on the RSW. They do simulations and modeling to suggest that the disruption of circular swimming upon collision with the wall increases the density of cells on the RSW, with a maximal effect at w = ~ 2/3 R, which is a good match for their experiments.

Strengths:

The overall result that confinement at the edge stabilises bacterial motion and allows chemotaxis is very interesting although not entirely unexpected. It is also important for understanding bacterial motility and chemotaxis under ecologically relevant conditions, where bacteria frequently swim under confinement (although its relevance for controlling infections could be questioned). The experimental part of the study is nicely supported by the model.

Weaknesses:

Several points of this study, in particular the interpretation of the width effect, need better clarification:

(1) Context:

There are a number of highly relevant previous publications that should have been acknowledged and discussed in relation to the current work:

https://pubs.rsc.org/en/content/articlehtml/2023/sm/d3sm00286a

https://link.springer.com/article/10.1140/epje/s10189-024-00450-7

https://doi.org/10.1016/j.bpj.2022.04.008

https://doi.org/10.1073/pnas.1816315116

https://www.pnas.org/doi/full/10.1073/pnas.0907542106

https://doi.org/10.1038/s41467-020-15711-0

http://doi.org/10.1038/s41467-020-15711-0

http://doi.org/10.1039/c5sm00939a

We appreciate the reviewer bringing these important publications to our attention. We have now cited and discussed these works in the Introduction (lines 55-62 and 76-85) to better contextualize our study regarding bacterial motility and chemotaxis in confined geometries.

(2) Experimental setup:

a) The channels are built with asymmetric entrances (Figure 1), which could trigger a ratchet effect (because bacteria swim in circle) that could bias the rate at which cells enter into the channel, and which side they follow preferentially, especially for the narrow channel. Since the channel is short (160 µm), that would reflect on the statistics of cell distribution. Controls with straight entrances or with a reversed symmetry of the channel need to be performed to ensure that the reported results are not affected by this asymmetry.

We appreciate the reviewer's insight regarding the potential ratchet effect caused by asymmetric entrances. To rule this out, we fabricated a control device with straight entrances and repeated the measurements. As shown in Figure S3, the chemotactic drift velocity follows the same trend as observed in the original setup, confirming an optimal width of ~9 mm. These results demonstrate that the entrance geometry does not bias the reported statistics. We have updated the manuscript text at lines 233-235.

b) The authors say the motile bacteria accumulate mostly at the bottom surface. This is strange, for a small height of 30 µm, the bacteria should be more-or-less evenly spread between the top and bottom surface. How can this be explained?

We apologize for not explaining this clearly in the text. As shown by Wei et al., Phys. Rev. Lett. 135, 188401 (2025), significant surface accumulation occurs in channels with heights exceeding 20 µm. In our specific experimental setup, we did not use Percoll to counteract gravity. Therefore, the bacteria accumulated mostly at the bottom surface under the combined influence of gravity and hydrodynamic attraction. This bottom-surface localization is supported by our observation that the bacterial trajectories were predominantly clockwise (characteristic of the bottom surface) rather than counter-clockwise (characteristic of the top surface). We have added this explanation to Line 141.

c) At the edge, some of the bacteria could escape up in the third dimension (http://doi.org/10.1039/c5sm00939a). What is the magnitude of this phenomenon in the current setup? Does it have an effect?

We thank the reviewer for raising this important point regarding 3D escape. We have quantified this phenomenon and found the escape rate from the edge into the third dimension to be 0.127 s^-1. This corresponds to a mean residence time that allows a cell moving at 20 mm/s to travel approximately 157.5 mm along the edge. Since this distance is comparable to the full length of our lanes (~160 mm), most cells traverse the entire edge without escaping. Furthermore, our analysis is based on the average drift of the surface trajectories per unit of time; this metric is independent of the absolute number of cells present. Therefore, the escape phenomenon does not significantly impact our conclusions. We have added a statement clarifying this at line 154.

d) What is the cell density in the device? Should we expect cell-cell interactions to play a role here? If not, I would suggest to de-emphasize the connection to chemotaxis in the swarming paper in the introduction and discussion, which doesn't feel very relevant here, and rather focus on the other papers mentioned in point 1.

The cell density in our experiments was approximately 1.3×10^-3 μm^-2. Given this low density, we do not expect cell-cell interactions to play a role in the observed behaviors.

Regarding the connection to swarming chemotaxis: We agree that our low-density setup differs from a high-density swarm; however, we believe the comparison remains relevant for two reasons. First, it provides a necessary contrast to studies showing surface inhibition of chemotaxis. Second, while we eliminate cell-cell interactions, we isolate the geometric aspect of swarming. In a swarm, cells move within narrow lanes created by their neighbors. Our device mimics this specific physical confinement by replacing neighboring cells with PDMS sidewalls. This allows us to decouple the effects of physical confinement from cell-cell interactions. We have added the text (Line 370) to clarify this rationale and have incorporated the additional references in introduction as suggested in point 1.

e) We are not entirely convinced by the interpretation of the results in narrow channels. What is the causal relationship between the increased density on the RSW and the higher chemotactic drift? The authors seem to attribute higher drift to this increased RSW density, which emerges due to the geometric reasons. But if there is no initial bias, the same geometric argument would induce the same increased density of down-gradient swimmers on the LSW, and so, no imbalance between RSW and LSW density. Could it be the opposite that the increased RSW density results from chemotaxis (and maybe reinforces it), not the other way around? Confinement could then deplete one wall due to the proximity of the other, and/or modify the swimming pattern - 8 µm is very close to the size of the body + flagellum. To clarify this point, we suggest measuring the bacterial distributions in the absence of a gradient for all channel widths as a control.

We thank the reviewer for this insightful comment regarding the causal relationship between cell density and chemotactic drift. We apologize if the initial explanation was unclear.

Regarding the no-gradient control: Without an attractant gradient (and no initial bias), there is no breaking of symmetry and the labels of "LSW" and "RSW" are arbitrary. Therefore, there will be no asymmetry in the bacterial distributions on both sides (within experimental fluctuations) in the absence of a gradient for any channel width.

Regarding the causality and density imbalance: We agree that the increased RSW density is a result of chemotaxis, which is then reinforced by the lane geometry especially at narrow lane width. The mechanism relies on the coupling of chemotactic bias with surface circularity. The angle ranges that lead to RSW-UG accumulation (Fig. 6A-C) coincide with the up-gradient direction. Because these cells experience suppressed tumbling (longer runs), they can maintain the steady circular trajectories required to reach and align with the RSW. Conversely, while pure geometric analysis suggests a similar potential for LSW-DG accumulation, these trajectories coincide with the down-gradient direction. These cells experience enhanced tumbling, which distorts the circular trajectories. This prevents them from effectively reaching the LSW and also increases the probability of them leaving the wall. Therefore, the causality is indeed a positive feedback loop: the attractant gradient creates an initial bias that allows the RSW-UG fraction to form stable trajectories; the optimal lane width (matching the swimming radius) then maximizes this capture efficiency, further enriching the RSW fraction and enhancing the overall drift.

We have added clarifications regarding these points in the revised manuscript (the last paragraph of “Results”).

(3) Simulations:

The simulations treat the wall interaction very crudely. We would suggest treating it as a mechanical object that exerts elastic or "hard sphere" forces and torques on the bacteria for more realistic modeling.

We appreciate the reviewer's suggestion to incorporate more detailed mechanical interactions, such as elastic or hard-sphere forces, for the wall collisions. While we agree that a full hydrodynamic or mechanical model would offer higher fidelity, our experimental observations suggest that a simplified kinematic approach is sufficient for the specific phenomena studied here.

As shown in the new Fig. S2, our analysis of cell trajectories in the 44-µm-wide channels reveals that cells colliding with the sidewalls tend to align with the surface almost instantaneously. The timescale required for this alignment is negligible compared to the typical wall residence time (see also Ref. 6). Consequently, to maintain computational efficiency without sacrificing the essential physics of the accumulation effect, we employed a coarse-grained phenomenological model where a bacterium immediately aligns parallel to the wall upon contact, similar to approaches used previously (Ref. 43). We have added relevant text to the manuscript on lines 168-171.

Notably, the simulations have a constant (chemotaxis independent) rate of wall escape by tumbling. We would expect that reduced tumbling due to up-gradient motility induces a longer dwell time at the wall.

We apologize for the confusion. The chemotaxis effect is indeed fully integrated into our simulation. Specifically, the simulated cells sense the chemical gradient and adjust their motor CW bias (B) accordingly. This adjustment directly modulates the tumble rate (k), calculated as k \= B/0.31 s^-1. Consequently, the wall escape rate is not constant but varies with the chemotactic response. We also imposed a maximum detention time limit which, when combined with the variable tumble rate, results in an average wall residence time of approximately 2 s, consistent with our experimental observations (Fig. S6B). We have clarified these details in the final section of 'Materials and Methods'.

Reviewer #3 (Public review):

This paper addresses through experiment and simulation the combined effects of bacterial circular swimming near no-slip surfaces and chemotaxis in simple linear gradients. The authors have constructed a microfluidic device in which a gradient of L-aspartate is established to which bacteria respond while swimming while confined in channels of different widths. There is a clear effect that the chemotactic drift velocity reaches a maximum in channel widths of about 8 microns, similar in size to the circular orbits that would prevail in the absence of side walls. Numerical studies of simplified models confirm this connection.

The experimental aspects of this study are well executed. The design of the microfluidic system is clever in that it allows a kind of "multiplexing" in which all the different channel widths are available to a given sample of bacteria.

While the data analysis is reasonably convincing, I think that the authors could make much better use of what must be voluminous data on the trajectories of cells by formulating the mathematical problem in terms of a suitable Fokker-Planck equation for the probability distribution of swimming directions. In particular, I would like to see much more analysis of how incipient circular trajectories are interrupted by collisions with the walls and how this relates to enhanced chemotaxis. In essence, there needs to be a much clearer control analysis of trajectories without sidewalls to understand the mechanism in their presence.

We thank the reviewer for this insightful suggestion. We agree that understanding how circular trajectories are interrupted by wall collisions is central to explaining the enhanced chemotaxis. While we did not explicitly formulate a Fokker-Planck equation, we have addressed the reviewer's core point by employing two complementary mathematical approaches that model the probability distribution of swimming directions and wall interactions:

(1) Stochastic simulations (Langevin approach): As detailed in the "Simulation of E. coli chemotaxis within lane confinements" subsection of “Results” and Figure 5, we modeled cells as self-propelled particles performing random walks. This model explicitly accounts for the "interruption" of circular trajectories by incorporating a constant angular velocity (circular swimming) and an alignment effect upon collision with sidewalls. These simulations successfully reproduced the experimental trends, confirming that the interplay between circular radius and lane width determines the optimal drift velocity.

(2) Geometric probability analysis: To provide the "intuitive understanding", we included a specific Geometrical Analysis section (the last subsection of “Results”) and Figure 6. This analysis mathematically formulates the problem by calculating the exact proportion of swimming angles that allow a cell to transition from a circular trajectory in the bulk to an up-gradient trajectory along the Right Sidewall (RSW). By integrating over the possible swimming directions, we derived the probability of wall interception as a function of lane width (w) and swimming radius (r). This analysis reveals that the interruption of circular paths is most favorable for chemotaxis when w » (0.7-0.8)´r.

(3) Control analysis: regarding the "control analysis of trajectories without sidewalls," we utilized the cells in the Middle Area (MA) of the wide lanes as an internal control. As shown in Fig. 2B and 4A, these cells exhibit typical surface-associated circular swimming (Fig. 3B) but generate zero net drift. This serves as the baseline "no sidewall" condition, demonstrating that the chemotactic enhancement is strictly driven by the rectification of circular swimming into wall-aligned motion at the boundaries.

The authors argue that these findings may have relevance to a number of physiological and ecological contexts. Yet, each of these would be characterized by significant heterogeneity in pore sizes and geometries, and thus it is very unclear whether or how the findings in this work would carry over to those situations.

We thank the reviewer for this important observation regarding environmental heterogeneity. We agree that we should be cautious about directly extrapolating to complex ecological contexts without qualification. We have revised the last sentence of the abstract to adopt a more measured tone: "Our results may offer insights into bacterial navigation in complex biological environments such as host tissues and biofilms, providing a preliminary step toward exploring microbial ecology in confined habitats and potential strategies for controlling bacterial infections."

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Key elements of the mechanism of wall-directed chemotaxis are not sufficiently emphasized:

For instance, the chirality of the trajectories is an essential part of the analysis but is mentioned only briefly in the introduction. In the geometrical analysis, I understand that one of the critical parameters is the angle at which bacteria "collide" with the walls. But, again, this remains largely implicit in the discussion. This comes to the point that these ideas are not even mentioned in the abstract which doesn't provide any hint of a mechanism. An analysis of the actual trajectories of the cells after they hit the walls, as a function of their initial angle would be helpful in comparison with the simulations and the geometrical analysis.

We appreciate the reviewer's insightful comment regarding the need to better emphasize the mechanism of wall-directed chemotaxis. We agree that the chirality of trajectories and the geometry of wall collisions are central to our analysis and were previously under-emphasized.

To address this, we have made the following revisions:

(1) We have revised the Abstract (lines 25-27) and the Discussion (lines 391-393) to explicitly highlight the crucial role of chiral circular motion and the alignment effect following sidewall collisions.

(2) We further analyzed bacterial trajectories at different collision angles. Typical examples are shown in Supplementary Fig. S2. We observed that cells tend to align with and swim along the sidewalls regardless of their initial collision angles. This finding is now described in the main text at lines 168-171.

The motion of the bacteria is modelled as run-and-tumble at several places in the manuscript, and in particular in the simulations. Yet, the trajectories of the bacteria seem to be smooth in this almost 2D geometry, except of course when they directly interact with the walls (I hardly see tumbles in the MA region in Figure 1B). Can the authors elaborate on the assumptions made in the numerical simulations? In particular, how is the radius of the trajectories included in these equations of motion (line 514)?

We apologize for the lack of clarity regarding the bacterial motion model. It has been established that while bacteria do tumble near solid surfaces, they exhibit a smaller reorientation angle compared to bulk fluids; in fact, the most probable reorientation angle on a surface is zero (Ref. 41). Consequently, tumbles are often difficult to distinguish from runs with the naked eye. Additionally, the trajectories in Figure 1B are plotted on a 44 mm ´ 150 mm canvas with unequal coordinate scales, which may further obscure the visual distinctness of tumbling events.

Regarding the equations of motion: We modeled the bacteria as self-propelled particles governed by the internal chemotaxis pathway, alternating between run and tumble states. As noted in the equations on lines 286 & 578, we incorporated the circular motion by introducing a constant angular velocity, −ν₀/r, during the run state. Here, ν₀ represents the swimming speed, r denotes the radius of circular swimming, and the negative sign indicates clockwise chirality. Furthermore, to model the hydrodynamic interaction with the boundaries, we assumed that when a cell collides with a sidewall, its velocity vector instantly aligns parallel to that wall.

The comparison of Figure 5B (simulations) with Figure 4B (experiments) does not strike me as so "similar". Why are the points at small widths so noisy (Figure 5AB)? Figure 5C is cut at these widths, it should be plotted over the entire scale.

We acknowledge that the agreement between simulation and experiment is less robust in the narrowest channels. The discrepancy and "noise" at small widths in Figure 5 arise from the limitations of the self-propelled particle model in highly confined geometries. Specifically, our simulation treats bacteria as point particles and does not explicitly calculate the physical exclusion (steric effects) caused by the finite size of the flagella and cell body.

In the experimental setup, steric constraints within narrow channels (comparable to the cell size) restrict the cells' ability to turn freely, effectively stabilizing their motion. However, because our model allows particles to reorient more freely than actual cells would in such confined spaces, it produces fluctuations and an overestimation of the drift velocity at small widths. If these confinement effects were fully incorporated, the cell density mismatch between the left and right sidewalls would be reduced, leading to lower drift velocities that match the experimental data more closely.

Regarding Figure 5C: Since the "active particle" assumption loses physical validity in channels narrower than the scale of the bacterium, the simulation results in this regime are not representative of biological reality. Plotting these non-physical points would distort the analysis. Therefore, we have maintained the truncation of Figure 5C at 4 mm to ensure the data presented is physically meaningful. We have added a clear discussion of these model limitations to the manuscript at lines 310-314.

These important precisions should be added to the text or in a supplementary section. A validated mechanism describing in detail the impact of the walls on the cell trajectories would greatly improve the conclusions.

We thank the reviewer for the suggestions. As noted in the responses above, we have incorporated the details concerning the simulation assumptions and the model limitations at narrow widths into the revised manuscript. We have performed further analysis of the collision trajectories between bacteria and the sidewalls. As illustrated in the new Fig. S2, the data confirms that cells tend to align with and swim along the sidewalls following a collision, regardless of the initial impact angle.

Reviewer #2 (Recommendations for the authors):

Minor points

(1) Related to swimming in 3D: The authors should specify the depth of field of the objective in their setup.

We thank the reviewer for pointing this out. We have calculated the depth of field (DOF) of our objective to be approximately 3.7 µm. This estimate is based on the standard formula:

where l = 610 nm (emission wavelength), n = 1.0 (refractive index), NA = 0.45 (numeric aperture), M = 20 (magnification), and e = 6.5 µm (camera resolution). We have added this specification to the "Microscopy and Data Acquisition" section of “Materials and Methods”.

(2) Related to the interpretation of the width effect: We think plotting the cell enrichment, ie the probabilities P in Figure 4B normalized to the expected value if cells were homogeneously distributed ((3µm)/w for the side walls, (w - 6µm)/w for the middle) would help understand the strength of the wall 'siphoning' effect.

We thank the reviewer for the suggestion. We have calculated the cell enrichment by normalizing the observed probabilities against the expected values for a homogeneous distribution, as suggested. The resulting relationship between cell enrichment and lane width is presented in Figure S4.

Related to simulations:

(1) Showing vd for the 3 regions in Figure S5 would be helpful also to understand the underlying mechanism.

We thank the reviewer for the suggestion. The V_d values for the three regions are shown in Fig. S5.

(2) Figure 5B vs 4B: There is a mismatch in the right vs left side density at w=6µm in the simulations that is not here in the experiments. What could explain this difference?

We appreciate the reviewer pointing this out. The mismatch in the simulations is due to the simplified treatment of cells as self-propelled particles, which overlooks the physical volume of the cell body and flagella. In narrow channels (w\=6 mm), these physical constraints would restrict the cells' ability to change direction freely - a factor not fully captured in the simulation. Accounting for these steric effects would trap cells more effectively against the walls, reducing the density asymmetry between the LSW and RSW and lowering the drift velocity. This would bring the simulation results closer to the experimental observations. We have added a discussion of these limitations and effects to the revised manuscript (lines 310-314).

(3) The simulations essentially assume that the density of motile cells is homogeneous and equal at both x=0 and x=L open ends of the channel. Is it the case in the experiments, even with the gradient, and the walls creating some cell transport?

We thank the reviewer for pointing this out. The simulation assumption is consistent with our experimental observations. Our data were recorded within 160-μm-long lanes located in the center of the wider (400 μm) cell channel. In this central region, the cells maintain a continuous flux. Furthermore, experiments were performed within 8 min of flow, limiting the time for significant cell density gradients to establish. As illustrated in Author response image 11, the inhomogeneity in the measured cell density distribution is insignificant across the length of the observation window, indicating that the walls and gradient do not create significant heterogeneity at the boundaries of the region of interest.

Author response image 1.

The cell density distribution along the gradient field from the data of 44-μm-wide lane.

(4) Line 506: There is something strange with the definition of the bias. B cannot be the tumbling bias if k=B/0.31 s^-1 and the tumble-to-run rate is 5/s, because then the tumbling bias is B/0.31 / (B/0.31 + 5). Please clarify.

We apologize for the confusion caused by the notation. In our model, B represents the CW bias of the individual flagellar motor, not the macroscopic tumbling bias of the cell. We assume the run-to-tumble rate is equivalent to the motor CCW-to-CW switching rate (k). Previous studies have shown that this rate increases linearly with the motor CW bias according to k=B/t, where t is a characteristic time (Ref. 50).

Based on experimental data for wildtype cells, the average run time in the near-surface region is ~2.0 s (corresponding to a run-to-tumble rate of ~0.5 s^-1) (Ref. 11), and the steady-state wildtype CW bias is ~0.15. Using these values, we determined t ~ 0.31 s. Consequently, the switching rate is defined as k=B/0.31 s^-1. Since the tumble duration is constant (0.2 s) (Ref. 51), the tumble-to-run rate is fixed at 5 s^-1. We have clarified these definitions and parameter values in lines 569-573.

Other minor comments:

(1) Line 20 and lines 34-35: We think that the connection to infection is questionable here and should be toned down.

Thank you for the suggestion. We have revised Line 20 to read: “Understanding bacterial behavior in confined environments is helpful to elucidating microbial ecology and developing strategies to manage bacterial infections.” Additionally, we modified lines 34-35 to state: “Our results may offer insights into bacterial navigation in complex biological environments such as host tissues and biofilms, providing a preliminary step toward exploring microbial ecology in confined habitats and potential strategies for controlling bacterial infections.”

(2) Line 49: Consider highlighting the change in the sense of rotation at the air-liquid interface.

Thank you for the suggestion. We have now highlighted the difference in chirality between trajectories at the air-liquid interface and those at the liquid-solid interface. The text has been updated to read: “For example, E. coli swim clockwise when observed from above a solid surface, whereas Caulobacter crescentus move in tight, counter-clockwise circles when viewed from the liquid side.”

(3) Lines 58-59: The sentence should be better formulated, explaining what is CheY-P and that its concentration changes because of a change in phosphorylation (P).

Thank you for the suggestion. We have reformulated this section to explicitly define CheY-P and explain how its concentration is regulated through phosphorylation. The revised text reads: “The transmembrane chemoreceptors detect attractants or repellents and transmit signals into the cell by modulating the autophosphorylation of the histidine kinase CheA. Attractant binding suppresses CheA autophosphorylation, while repellent binding promotes it. This modulation alters the concentration of the phosphorylated response regulator protein, CheY-P.”

(4) Lines 63-64: CheR CheB do a bit more than "facilitating" adaptation, they mediate it. The notation CheB(p) may be confusing, since "-P" was used above for CheY.

Thank you for pointing this out. We have corrected the notation and strengthened the description of the enzymes' roles. The revised text is: “The adaptation enzymes CheR and CheB methylate and demethylate the receptors, respectively, mediating sensory adaptation.”

(5) Line 130: there must be a typo in the formula.

We have replaced the ambiguous lag time variable in Fig. 1C with _n_Δt to ensure mathematical consistency.

(6) Additionally, \Delta t is both the time between the frame here and the lag time in Figure 1.

Thank you for highlighting this ambiguity. We have updated the notation to distinguish these two values. The lag time in Figure 1 is now explicitly denoted as _n_Δt, while Δt remains the time interval between individual frames.

(7) Line 162: "Consistent with previous reports," a reference to said reports is missing.

Thank you for pointing this out. We have now added the reference (Ref. 41) to support this statement.

(8) Figure 1B: Are these tracks in the presence of a gradient? Same as used in panel C? This needs to be explained.

Response: Thank you for this question. We confirm that the tracks shown in Figure 1B were indeed recorded in the presence of a gradient and represent a subset of the data used in Figure 1C. We have clarified this in the figure legend as follows: "Thirty bacterial trajectories selected from the data of the 44-mm-wide lane in gradient assays. These represent a subset of the trajectories analyzed in panel C."

(9) Simulations: the equation for x(t) should also be given for completeness.

Thank you for the suggestion. For completeness, we have added the position updating equations for the run state to the Materials and Methods section (lines 579-580). The equations are defined as:

(10) Figure S2: For the swimming directions that are more unstable due to the surface friction torque, RSW-DG, and LSW-UG, one would have expected that the Up-gradient motion is more persistent than the down gradient one. It seems to be the opposite. Is it significant, and what could be the reason for this?

We apologize for the lack of clarity in our original explanation. While we would generally expect up-gradient motion to be more persistent than down-gradient motion in bulk fluid, our measurements near the surface show a different trend due to the specific contributions of run and tumble states to the escape rate. Cells swimming up-gradient (UG) in the LSW experience higher probability of running. Consequently, they are subjected to the destabilizing surface friction torque for a greater proportion of time compared to cells swimming down-gradient (DG) in the RSW. This can be explained mathematically. The escape rates for RSW-DG and LSW-UG can be expressed as:

Where B⁺ and B⁻ represent the tumble bias (probability of tumbling) when swimming up-gradient and down-gradient, respectively, and k_T and k_R denote the escape rates during a tumble and a run, respectively. Due to the chemotactic response, 0≤ B⁺< B⁻ ≤1. Crucially, our system is characterized by k_R>k_T (the escape rate is higher during a run than a tumble). Therefore, the lower tumble bias during up-gradient swimming (B⁺< B⁻) increases the weight of the run-state escape term((1−B⁺)k_R), leading to a higher overall escape rate for LSW-UG compared to RSW-DG. We have added an intuitive understanding of k_R>k_T in the Supplemental text.

eLife Assessment

This fundamental work reveals that the accessibility of the unstructured C-terminal tail of α-tubulin differs with the state of the microtubule lattice. Accessibility increases with the expansion of the lattice induced by GTP and certain MAPs, which can then dictate the subsequent interactions between MAPs and microtubules, and post-translational modifications of tubulin tails. The evidence supporting the conclusion is compelling, although the characterisation of the probes does not answer whether they directly affect the lattice or expose the C-terminal tail of α-tubulin. The probes can be used as tools in the future to study differences in microtubule lattice regulation under different conditions both in vitro and in vivo. This work will be of great interest to the cytoskeleton field.

Reviewer #1 (Public review):

Summary:

This is a careful and comprehensive study demonstrating that effector-dependent conformational switching of the MT lattice from compacted to expanded deploys the alpha tubulin C-terminal tails so as to enhance their ability to bind interactors.

Strengths:

The authors use 3 different sensors for the exposure of the alpha CTTs. They show that all 3 sensors report exposure of the alpha CTTs when the lattice is expanded by GMPCPP, or KIF1C, or a hydrolysis-deficient tubulin. They demonstrate that expansion-dependent exposure of the alpha CTTs works in tissue culture cells as well as in vitro.

Appraisal:

The authors have gone to considerable lengths to test their hypothesis that microtubule expansion favours deployment of the alpha tubulin C-terminal tail, allowing its interactors, including detyrosinase enzymes, to bind. There is a real prospect that this will change thinking in the field. One very interesting possibility, touched on by the authors, is that the requirement for MAP7 to engage kinesin with the MT might include a direct effect of MAP7 on lattice expansion.

Impact:

The possibility that the interactions of MAPS and motors with a particular MT or region feed forward to determine its future interaction patterns is made much more real. Genuinely exciting.

Reviewer #2 (Public review):

The unstructured α- and β-tubulin C-terminal tails (CTTs), which differ between tubulin isoforms, extend from the surface of the microtubule, are post-translationally modified, and help regulate the function of MAPs and motors. Their dynamics and extent of interactions with the microtubule lattice are not well understood. Hotta et al. explore this using a set of three distinct probes that bind to the CTTs of tyrosinated (native) α-tubulin. Under normal cellular conditions, these probes associate with microtubules only to a limited extent, but this binding can be enhanced by various manipulations thought to alter the tubulin lattice conformation (expanded or compact). These include small-molecule treatment (Taxol), changes in nucleotide state, and the binding of microtubule-associated proteins and motors. Overall, the authors conclude that microtubule lattice "expanders" promote probe binding, suggesting that the CTT is generally more accessible under these conditions. Consistent with this, detyrosination is enhanced. Mechanistically, molecular dynamics simulations indicate that the CTT may interact with the microtubule lattice at several sites, and that these interactions are affected by the tubulin nucleotide state.

Strengths and weaknesses:

Key strengths of the work include the use of three distinct probes that yield broadly consistent findings, and a wide variety of experimental manipulations (drugs, motors, MAPs) that collectively support the authors' conclusions, alongside a careful quantitative approach.

The challenges of studying the dynamics of a short, intrinsically disordered protein region within the complex environment of the cellular microtubule lattice, amid numerous other binders and regulators, should not be understated. While it is very plausible that the probes report on CTT accessibility as proposed, the possibility of confounding factors (e.g., effects on MAP or motor binding) cannot be ruled out. Sensitivity to the expression level clearly introduces additional complications. Likewise, for each individual "expander" or "compactor" manipulation, one must consider indirect consequences (e.g., masking of binding sites) in addition to direct effects on the lattice; however, this risk is mitigated by the collective observations all pointing in the same direction.

The discussion does a good job of placing the findings in context and acknowledging relevant caveats and limitations. Overall, this study introduces an interesting and provocative concept, well supported by experimental data, and provides a strong foundation for future work. This will be a valuable contribution to the field.

Reviewer #3 (Public review):

Summary:

In this study, the authors investigate how the structural state of the microtubule lattice influences the accessibility of the α-tubulin C-terminal tail (CTT). By developing and applying new biosensors, they reveal that the tyrosinated CTT is largely inaccessible under normal conditions but becomes more accessible upon changes to the tubulin conformational state induced by taxol treatment, MAP expression, or GTP-hydrolysis-deficient tubulin. The combination of live imaging, biochemical assays, and simulations suggests that the lattice conformation regulates the exposure of the CTT, providing a potential mechanism for modulating interactions with microtubule-associated proteins. The work addresses a highly topical question in the microtubule field and proposes a new conceptual link between lattice spacing and tail accessibility for tubulin post-translational modification. Future work is required to distinguish CTT exposure in the microtubule lattice is sensitive to additional factors present in vivo but not in vitro.

Strengths:

(1) The study targets a highly relevant and emerging topic-the structural plasticity of the microtubule lattice and its regulatory implications.

(2) The biosensor design represents a methodological advance, enabling direct visualization of CTT accessibility in living cells.

(3) Integration of imaging, biochemical assays, and simulations provides a multi-scale perspective on lattice regulation.

(4) The conceptual framework proposed lattice conformation as a determinant of post-translational modification accessibility is novel and potentially impactful for understanding microtubule regulation.

[Editors' note: the authors have responded to the reviewers and this version was assessed by the editors.]

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This is a careful and comprehensive study demonstrating that effector-dependent conformational switching of the MT lattice from compacted to expanded deploys the alpha tubulin C-terminal tails so as to enhance their ability to bind interactors.

Strengths:

The authors use 3 different sensors for the exposure of the alpha CTTs. They show that all 3 sensors report exposure of the alpha CTTs when the lattice is expanded by GMPCPP, or KIF1C, or a hydrolysis-deficient tubulin. They demonstrate that expansion-dependent exposure of the alpha CTTs works in tissue culture cells as well as in vitro.

Weaknesses:

There is no information on the status of the beta tubulin CTTs. The study is done with mixed isotype microtubules, both in cells and in vitro. It remains unclear whether all the alpha tubulins in a mixed isotype microtubule lattice behave equivalently, or whether the effect is tubulin isotype-dependent. It remains unclear whether local binding of effectors can locally expand the lattice and locally expose the alpha CTTs.

Appraisal:

The authors have gone to considerable lengths to test their hypothesis that microtubule expansion favours deployment of the alpha tubulin C-terminal tail, allowing its interactors, including detyrosinase enzymes, to bind. There is a real prospect that this will change thinking in the field. One very interesting possibility, touched on by the authors, is that the requirement for MAP7 to engage kinesin with the MT might include a direct effect of MAP7 on lattice expansion.

Impact:

The possibility that the interactions of MAPS and motors with a particular MT or region feed forward to determine its future interaction patterns is made much more real. Genuinely exciting.

We thank the reviewer for their positive response to our work. We agree that it will be important to determine if the bCTT is subject to regulation similar to the aCTT. However, this will first require the development of sensors that report on the accessibility of the bCTT, which is a significant undertaking for future work.

We also agree that it will be important to examine whether all tubulin isotypes behave equivalently in terms of exposure of the aCTT in response to conformational switching of the microtubule lattice.

We thank the reviewer for the comment about local expansion of the microtubule lattice. We believe that Figure 3 does show that local binding of effectors can locally expand the lattice and locally expose the alpha-CTTs. We have added text to clarify this.

Reviewer #2 (Public review):

The unstructured α- and β-tubulin C-terminal tails (CTTs), which differ between tubulin isoforms, extend from the surface of the microtubule, are post-translationally modified, and help regulate the function of MAPs and motors. Their dynamics and extent of interactions with the microtubule lattice are not well understood. Hotta et al. explore this using a set of three distinct probes that bind to the CTTs of tyrosinated (native) α-tubulin. Under normal cellular conditions, these probes associate with microtubules only to a limited extent, but this binding can be enhanced by various manipulations thought to alter the tubulin lattice conformation (expanded or compact). These include small-molecule treatment (Taxol), changes in nucleotide state, and the binding of microtubule-associated proteins and motors. Overall, the authors conclude that microtubule lattice "expanders" promote probe binding, suggesting that the CTT is generally more accessible under these conditions. Consistent with this, detyrosination is enhanced. Mechanistically, molecular dynamics simulations indicate that the CTT may interact with the microtubule lattice at several sites, and that these interactions are affected by the tubulin nucleotide state.

Strengths:

Key strengths of the work include the use of three distinct probes that yield broadly consistent findings, and a wide variety of experimental manipulations (drugs, motors, MAPs) that collectively support the authors' conclusions, alongside a careful quantitative approach.

Weaknesses:

The challenges of studying the dynamics of a short, intrinsically disordered protein region within the complex environment of the cellular microtubule lattice, amid numerous other binders and regulators, should not be understated. While it is very plausible that the probes report on CTT accessibility as proposed, the possibility of confounding factors (e.g., effects on MAP or motor binding) cannot be ruled out. Sensitivity to the expression level clearly introduces additional complications. Likewise, for each individual "expander" or "compactor" manipulation, one must consider indirect consequences (e.g., masking of binding sites) in addition to direct effects on the lattice; however, this risk is mitigated by the collective observations all pointing in the same direction.

The discussion does a good job of placing the findings in context and acknowledging relevant caveats and limitations. Overall, this study introduces an interesting and provocative concept, well supported by experimental data, and provides a strong foundation for future work. This will be a valuable contribution to the field.

We thank the reviewer for their positive response to our work. We are encouraged that the reviewer feels that the Discussion section does a good job of putting the findings, challenges, and possibility of confounding factors and indirect effects in context. 

Reviewer #3 (Public review):

Summary:

In this study, the authors investigate how the structural state of the microtubule lattice influences the accessibility of the α-tubulin C-terminal tail (CTT). By developing and applying new biosensors, they reveal that the tyrosinated CTT is largely inaccessible under normal conditions but becomes more accessible upon changes to the tubulin conformational state induced by taxol treatment, MAP expression, or GTP-hydrolysis-deficient tubulin. The combination of live imaging, biochemical assays, and simulations suggests that the lattice conformation regulates the exposure of the CTT, providing a potential mechanism for modulating interactions with microtubule-associated proteins. The work addresses a highly topical question in the microtubule field and proposes a new conceptual link between lattice spacing and tail accessibility for tubulin post-translational modification.

Strengths:

(1) The study targets a highly relevant and emerging topic-the structural plasticity of the microtubule lattice and its regulatory implications.

(2) The biosensor design represents a methodological advance, enabling direct visualization of CTT accessibility in living cells.

(3) Integration of imaging, biochemical assays, and simulations provides a multi-scale perspective on lattice regulation.

(4) The conceptual framework proposed lattice conformation as a determinant of post-translational modification accessibility is novel and potentially impactful for understanding microtubule regulation.

Weaknesses:

There are a number of weaknesses in the paper, many of which can be addressed textually. Some of the supporting evidence is preliminary and would benefit from additional experimental validation and clearer presentation before the conclusions can be considered fully supported. In particular, the authors should directly test in vitro whether Taxol addition can induce lattice exchange (see comments below).

We thank the reviewer for their positive response to our work. We have altered the text and provided additional experimental validation as requested (see below).

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) The resolution of the figures is insufficient.

(2) The provision of scale bars is inconsistent and insufficient.

(3) Figure 1E, the scale bar looks like an MT.

(4) Figure 2C, what does the grey bar indicate?

(5) Figure 2E, missing scale bar.

(6) Figure 3 C, D, significance brackets misaligned.

(7) Figure 3E, consider using the same alpha-beta tubulin / MT graphic as in Figure 1B.

(8) Figure 5E, show cell boundaries for consistency?

(9) Figure 6D, stray box above the y-axis.

(11) Figure S3A, scale bar wrong unit again.

(12) S3B "fixed" and mount missing scale bar in the inset.

(13) S4 scale bars without scale, inconsistency in scale bars throughout all the figures.

We apologize for issues with the figures. We have corrected all of the issues indicated by the reviewer.

(10) Figure 6F, surprising that 300 mM KCL washes out rigor binding kinesin

We thank the reviewer for this important point. To address the reviewer’s concern, we have added a new supplementary figure (new Figure 6 – Figure Supplement 1) which shows that the washing step removes strongly-bound (apo) KIF5C(1-560)-Halo⁵⁵⁴ protein from the microtubules. In addition, we have made a correction to the Materials and Methods section noting that ATP was added in addition to the KCl in the wash buffer. We apologize for omitting this detail in the original submission. We also added text noting that the wash out step was based on Shima et al., 2018 where the observation chamber was washed with either 1 mM ATP and 300 mM K-Pipes or with 10 mM ATP and 500 mM K-Pipes buffer. In our case, the chamber was washed with 3 mM ATP and 300 mM KCl. It is likely that the addition of ATP facilitates the detachment of strongly-bound KIF5C.

(14) Supplementary movie, please identify alpha and beta tubules for clarity. Please identify residues lighting up in interaction sites 1,2 & 3.

Thank you for the suggestions. We have made the requested changes to the movie.

Reviewer #2 (Recommendations for the authors):

There appear to have been some minor issues (perhaps with .pdf conversion) that leave some text and images pixelated in the .pdf provided, alongside some slightly jarring text and image positioning (e.g., Figure 5E panels). The authors should carefully look at the figures to ensure that they are presented in the clearest way possible.

We apologize for these issues with the figures. We have reviewed the figures carefully to ensure that they are presented in the clearest way possible.

The authors might consider providing a more definitive structural description of compact vs expanded lattice, highlighting what specific parameters are generally thought to change and by what magnitude. Do these differ between taxol-mediated expansion or the effects of MAPs?

Thank you for the suggestion. We have added additional information to the Introduction section.

Reviewer #3 (Recommendations for the authors):

(1) Figure 1 should include a schematic overview of all constructs used in the study. A clear illustration showing the probe design, including the origin and function of each component (e.g., tags, domains), would improve clarity.

Thank you for the suggestion. We have added new illustrations to Figure 1 showing the origin and design (including domains and tags) of each probe.

(2) Add Western blot data for the 4×CAP-Gly construct to Figure 1C for completeness.

We thank the reviewer for this suggestion. We carried out a far-western blot using the purified 4xCAPGly-mEGFP protein to probe GST-Y, GST-DY, and GST-DC2 proteins (new Figure 1 – Figure Supplement 1C). We note that some bleed-through signal can be seen in the lanes containing GST-ΔY and GST-ΔC2 protein due to the imaging requirements and exposure needed to visualize the 4xCAPGly-mEGFP protein. Nevertheless, the blot shows that the purified CAPGly sensor specifically recognizes the native (tyrosinated) CTT sequence of TUBA1A.

(3) Essential background information on the CAP-Gly domain, SXIP motif, and EB proteins is missing from the Introduction. These concepts appear abruptly in the Results and should be properly introduced.

Thank you for the suggestion. We have added additional information to the Introduction section about the CAP-Gly domain. However, we feel that introducing the SXIP motif and EB proteins at this point would detract from the flow of the Introduction and we have elected to retain this information in the Results section when we detail development of the 4xCAPGly probe.

(4) In Figure 2E, it remains possible that the CAP-Gly domain displacement simply follows the displacement of EB proteins. An experiment comparing EB protein localization upon Taxol treatment would clarify this relationship.

We thank the reviewer for raising this important point. To address the reviewer’s concern, we utilized HeLa cells stably expressing EB3-GFP. We performed live-cell imaging before and after Taxol addition (new Figure 2 – Figure Supplement 1C). EB3-EGFP was lost from the microtubule plus ends within minutes and did not localize to the now-expanded lattice.

(5) Statements such as "significantly increased" (e.g., line 195) should be replaced with quantitative information (e.g., "1.5-fold increase").

We have made the suggested changes to the text.

(6) Phrases like "became accessible" should be revised to "became more accessible," as the observed changes are relative, not absolute. The current wording implies a binary shift, whereas the data show a modest (~1.5-fold) increase.

We have made the suggested changes to the text.

(7) Similarly, at line 209, the terms "minimally accessible" versus "accessible" should be rephrased to reflect the small relative change observed; saturation of accessibility is not demonstrated.

We have made the suggested changes to the text.

(8) Statements that MAP7 "expands the lattice" (line 222) should be made cautiously; to my knowledge, that has not been clearly established in the literature.

We thank the reviewer for this important comment. We have added text indicating that MAP7’s ability to induce or presence an expanded lattice has not been clearly established.

(9) In Figures 3 and 4, the overexpression of MAP7 results in a strikingly peripheral microtubule network. Why is there this unusual morphology?

The reviewer raises an interesting question. We are not sure why the overexpression of MAP7 results in a strikingly peripheral microtubule network but we suspect this is unique to the HeLa cells we are using. We have observed a more uniform MAP7 localization in other cell types [e.g. COS-7 cells (Tymanskyj et al. 2018), consistent with the literature [e.g. BEAS-2B cells (Shen and Ori-McKenney 2024), HeLa cells (Hooikaas et al. 2019)].

(10) In Supplementary Figure 5C, the Western blot of detyrosination levels is inconsistent with the text. Untreated cells appear to have higher detyrosination than both wild-type and E254A-overexpressing cells. Do you have any explanation?

We thank the reviewer for this important comment. We do not have an explanation at this point but plan to revisit this experiment. Unfortunately, the authors who carried out this work recently moved to a new institution and it will be several months before they are able to get the cell lines going and repeat the experiment. We thus elected to remove what was Supp Fig 5C until we can revisit the results. We believe that the important results are in what is now Figure 5 - Figure Supplement 1A,B which shows that the expression levels of the WT and E254E proteins are similar to each other.

(11) The image analysis method in Figures 5B and 5D requires clarification. It appears that "density" was calculated from skeletonized probe length over total area, potentially using a strict intensity threshold. It looks like low-intensity binding has been excluded; otherwise, the density would be the same from the images. If so, this should be stated explicitly. A more appropriate analysis might skeletonize and integrate total fluorescence intensity relative to the overall microtubule network.

We have added additional information to the Materials and Methods section to clarify the image analysis. We appreciate the reviewer’s valuable feedback and the suggestion to use the integrated total fluorescence intensity, which is a theoretically sound approach. While we agree that integrated intensity is a valid metric for specific applications, its appropriate use depends on two main preconditions:

(1) Consistent microscopy image acquisition conditions.

(2) Consistent probe expression levels across all cells and experiments.

We successfully maintained consistent image acquisition conditions (e.g., exposure time) throughout the experiment. However, despite generating a stably-expressing sensor cell lines to minimize variation, there remains an inherent, biological variability in probe expression levels between individual cells. Integrated intensity is highly susceptible to this cell-to-cell variability. Relying on it would lead to a systematic error where differences in the total amount of expressed probe would be mistaken for differences in Y-aCTT accessibility.

The density metric (skeletonized probe length / total cell area) was deliberately chosen as it serves as a geometric measure rather than an intensity-based normalization. The density metric quantifies the proportion of the microtubule network that is occupied by Y-aCTT-labeled structures, independent of fluorescence intensity. Thus, the density metric provides a more robust and interpretable measure of Y-aCTT accessibility under the variable expression conditions inherent to our experimental system. Therefore, we believe that this geometric approach represents the most appropriate analysis for our image dataset.

(12) In Figure 5D, the fold-change data are difficult to interpret due to the compressed scale. Replotting is recommended. The text should also discuss the relative fold changes between E254A and Taxol conditions, Figure 2H.

We appreciate the reviewer's insightful comment. We agree that the presence of significant outliers led to a compressed Y-axis scale in Figure 5D, obscuring the clear difference between the WT-tubulin and E254A-tubulin groups. As suggested, we have replotted Figure 5D using a broken Y-axis to effectively expand the relevant lower range of the data while still accurately representing all data points, including the outliers. We believe that the revised graph significantly enhances the clarity and interpretability of these results. For Figure 2, we have added the relative fold changes to the text as requested.

(13) Figure 6. The authors should directly test in vitro whether Taxol addition can induce lattice exchange, for example, by adding Taxol to GDP-microtubules and monitoring probe binding. Including such an assay would provide critical mechanistic evidence and substantially strengthen the conclusions. I was waiting for this experiment since Figure 2.

We thank the reviewer for this suggestion. As suggested, we generated GDP-MTs from HeLa tubulin and added it to two flow chambers. We then flowed in the YL1/2^Fab-EGFP probe into the chambers in the presence of DMSO (vehicle control) or Taxol. Static images were taken and the fluorescence intensity of the probe on microtubules in each chamber was quantified. There was a slight but not statistically significant difference in probe binding between control and Taxol-treated GDP-MTs (Author response image 1). While disappointing, these results underscore our conclusion (Discussion section) that microtubule assembly in vitro may not produce a lattice state resembling that in cells, either due to differences in protofilament number and/or buffer conditions and/or the lack of MAPs during polymerization.

Author response image 1.

References

Hooikaas, P. J., Martin, M., Muhlethaler, T., Kuijntjes, G. J., Peeters, C. A. E., Katrukha, E. A., Ferrari, L., Stucchi, R., Verhagen, D. G. F., van Riel, W. E., Grigoriev, I., Altelaar, A. F. M., Hoogenraad, C. C., Rudiger, S. G. D., Steinmetz, M. O., Kapitein, L. C. and Akhmanova, A. (2019). MAP7 family proteins regulate kinesin-1 recruitment and activation. J Cell Biol, 218, 1298-1318.

Shen, Y. and Ori-McKenney, K. M. (2024). Microtubule-associated protein MAP7 promotes tubulin posttranslational modifications and cargo transport to enable osmotic adaptation. Dev Cell, 59, 1553-1570.

Tymanskyj, S. R., Yang, B. H., Verhey, K. J. and Ma, L. (2018). MAP7 regulates axon morphogenesis by recruiting kinesin-1 to microtubules and modulating organelle transport. Elife, 7.

eLife Assessment

This fundamentally significant study provides the first genome-wide characterization of H3K115 acetylation and identifies a striking and previously unappreciated association of this globular-domain histone modification with fragile nucleosomes at CpG island promoters, active enhancers, and CTCF binding sites. While the work is largely descriptive and correlative in nature the evidence is compelling. The authors present multiple, orthogonal genomic and biochemical analyses that consistently support their central conclusions.

Reviewer #1 (Public review):

Summary

The authors set out to define the genomic distribution and potential functional associations of acetylation of histone H3 lysine 115 (H3K115ac), a poorly characterized modification located in the globular domain of histone H3. Using native ChIP-seq and complementary genomic approaches in mouse embryonic stem cells and during differentiation to neural progenitor cells, they report that H3K115ac is enriched at CpG island promoters, active enhancers, and CTCF binding sites, where it preferentially localizes to regions containing fragile or subnucleosomal particles. These observations suggest that H3K115ac marks destabilized nucleosomes at key regulatory elements and may serve as an informative indicator of chromatin accessibility and regulatory activity.

Strengths

A major strength of this study is its focus on a histone post-translational modification in the globular domain, an area that has received far less attention than histone tail modifications despite strong evidence from structural and in vitro studies that such marks can directly influence nucleosome stability. The authors employ a wide range of complementary genomic analyses-including paired-end ChIP-seq, fragment size-resolved analyses, contour (V-) plots, and sucrose gradient fractionation-to consistently support the association of H3K115ac with fragile nucleosomes across promoters, enhancers, and architectural elements. The revised manuscript is careful in its interpretation and provides a coherent and internally consistent picture of how H3K115ac differs from other acetylation marks such as H3K27ac and H3K122ac. The datasets generated will be valuable to the chromatin community as a resource for further exploration of nucleosome dynamics at regulatory elements.

Weaknesses

The conclusions are largely correlative. While the authors provide strong evidence for the localization of H3K115ac to fragile nucleosomes, the work does not directly test whether this modification causally contributes to nucleosome destabilization or regulatory function in vivo. Questions regarding the enzymes responsible for depositing or removing H3K115ac and its direct functional consequences therefore remain open.

Overall assessment and impact

Overall, the authors largely achieve their stated aims by providing a detailed and well-supported characterization of H3K115ac distribution in mammalian chromatin and its association with fragile nucleosomes at regulatory elements. While mechanistic insight remains to be established, the study introduces a compelling new perspective on globular-domain histone acetylation and highlights H3K115ac as a potentially useful marker for identifying functionally important regulatory regions. The work is likely to stimulate further mechanistic studies and will be of broad interest to researchers studying chromatin structure, transcriptional regulation, and genome organization.

Reviewer #2 (Public review):

Summary:

Kumar et al. aimed to assess the role of the understudied H3K115 acetylation mark, which is located in the nucleosomal core. To this end, the authors performed ChIP-seq experiments of H3K115ac in mouse embryonic stem cells as well as during differentiation into neuronal progenitor cells. Subsequent bioinformatic analyses revealed an association of H3K115ac with fragile nucleosomes at CpG island promoters, as well as with enhancers and CTCF binding sites. This is an interesting study, which provides important novel insights into the potential function of H3K115ac. However, the study is mainly descriptive, and functional experiments are missing.

Strengths:

(1) The authors present the first genome-wide profiling of H3K115ac and link this poorly characterized modification to fragile nucleosomes, CpG island promoters, enhancers, and CTCF binding sites.

(2) The study provides a valuable descriptive resource and raises intriguing hypotheses about the role of H3K115ac in chromatin regulation.

(3) The breadth of the bioinformatic analyses adds to the value of the dataset

Comments on revisions:

The authors sufficiently addressed my concerns.

Reviewer #3 (Public review):

Summary:

Kumar et al. examine the H3K115 epigenetic mark located on the lateral surface of the histone core domain and present evidence that it may serve as a marker enriched at transcription start sites (TSSs) of active CpG island promoters and at polycomb-repressed promoters. They also note enrichment of the H3K115ac mark is found on fragile nucleosomes within nucleosome-depleted regions, on active enhancers and CTCF bound sites. They propose that these observations suggest that H3K115ac contributes to nucleosome destabilization and so may servers a marker of functionally important regulatory elements in mammalian genomes.

Strengths:

The authors present novel observations suggesting that acetylation of a histone residue in a core (versus on a histone tail) domain may serve a functional role in promoting transcription in CPG islands and polycomb-repressed promoters. They present a solid amount of confirmatory in silico data using appropriate methodology that supports the idea that H3K115ac mark may function to destabilize nucleosomes and contribute to regulating ESC differentiation. These findings are quite novel.

Weaknesses:

Additional experiments to confirm specificity of the antibodies used have been done, improving confidence in the study.

Author response:

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

Public Reviews:

Reviewer #1 (Public reviews):

(1) The absence of replicate paired-end datasets limits confidence in peak localization.

The reviewer was under the impression that that we did not perform biological replicates of our ChIP-seq experiments. All ChIP-seq (and ATAC-seq) experiments were performed with biological replicates and the Pearson’s correlations (all >0.9) between replicates were provided in Supplementary Table 1. We had indicated this in the text and methods but will try to make this even clearer.

(2) The analyses are primarily correlative, making it difficult to fully assess robustness or to support strong mechanistic conclusions.

Histone modifications are difficult to alter genetically because of the high copy number of histone genes and inhibition of HATs/HDACs in general leads to alterations in other histone modifications. It is an inherent challenge in establishing causality of histone modifications, especially histone acetylation marks.

(3) Some claims (e.g., specificity for CpG islands, "dynamic" regulation during differentiation) are not fully supported by the analyses as presented.

We have modified the text in response to this point. The new text reads: “Non-CGI promoters have lower overall levels of transcription compared to CGI promoters, and for this promoter class H3K115ac enrichment detected by ChIP is only really seen for the highest quartile of transcription (4SU) quartile of expression (Figure 1G). CGI promoters on the other hand, exhibit significant levels of detected H3K115ac even for the lowest quartile of expression. These results suggest a special link between CGI promoters and H3K115ac”.

(4) Overall, the study introduces an intriguing new angle on globular PTMs, but additional rigor and mechanistic evidence are needed to substantiate the conclusions.

We agree that the paper does not provide mechanistic details or solid causality of H3K115ac. We have only emphasized the potential role of H3K115ac in nucleosome fragility based on our in vivo data and previously published in-vitro experiments (Manohar et.al., 2009, Chatterjee et. al., 2015). We do provide the evidence that H3K115ac is enriched on subnucleosomal particles via sucrose gradient sedimentation of MNase-digested chromatin (Figure 3C-D).

Reviewer #2 (Public review):

(1) I am not fully convinced about the specificity of the antibody. Although the experiment in Figure S1A shows a specific binding to H3K115ac-modified peptides compared to unmodified peptides, the authors do not show any experiment that shows that the antibody does not bind to unrelated proteins. Thus, a Western of a nuclear extract or the chromatin fraction would be critical to show. Also, peptide competition using the H3K115ac peptide to block the antibody may be good to further support the specificity of the antibody. Also, I don't understand the experiment in Figure S1B. What does it tell us when the H3K115ac histone mark itself is missing? The KLF4 promoter does not appear to be a suitable positive control, given that hundreds of proteins/histone modifications are likely present at this region. It is important to clearly demonstrate that the antibody exclusively recognizes H3K115ac, given that the conclusion of the manuscript strongly depends on the reliability of the obtained ChIP-Seq data.

ChIP-qPCR in S1B includes competition from native chromatin and shows high specificity to its target. We have provided antibody validation in three ways:

- Western blot with dot-blot of synthetic peptides (Figure S1A).

- Western blots with Whole cell extracts (Figure 4D).

- ChIP-qPCR on native chromatin spiked with a cocktail of synthetic mono-nucleosomes, each carrying a single acetylation and a specific barcode (SNAP-ChIP K-AcylStat Panel).

We could not include H3K115ac marked nucleosomes as they are not available in the panel. Figure S1B shows that the H3K115ac antibody exhibits negligible binding to known K-acyl marks, comparable to an unmodified nucleosome. Because of the absence of a H3K115ac modified barcoded nucleosome, we used the KLF4 promoter from mESCs as a positive control, in agreement with ChIP-seq signal shown in the genome browser profile (Figure 1E), the KLF4 promoter shows a significantly higher signal than the gene body.

(2) The association of H3K115ac with fragile nucleosomes is based on MNase-sensitivity and fragment length, which are indirect methods and can have technical bias. Experiments that support that the H3K115ac modified nucleosomes are indeed more fragile are missing.

We have performed ChIP-seq on MNase digested mESC chromatin fractionated on sucrose gradients and this shows that H3K115ac is enriched in fractions containing sub-nucleosomal and fragile nucleosomes but depleted in fractions containing stable nucleosomes (Figure 3D).

(3) The comparison of H3K115ac with H3K122ac and H3K64ac relies on publicly available datasets. Since the authors argue that these marks are distinct, data generated under identical experimental conditions would be more convincing. At a minimum, the limitations of using external datasets should be discussed.

H3K64ac and H3K122ac datasets were generated by us in a previous publication (Pradeepa et. al., 2016) using same native MNase ChIP protocol as used here. The ChIP-seq datasets for H3K122ac and H3K27ac are processed in an identical manner, with the same computational pipelines, to the H3K115ac data sets generated in this paper.

(4) The enrichment of H3K115ac at enhancers and CTCF binding sites is notable but remains descriptive. It would be interesting to clarify whether H3K115ac actively influences transcription factor/CTCF binding or is a downstream correlate.

We agree with the reviewer’s comment, but we have not claimed causality.

(5) No information is provided about how H3K115ac may be deposited/removed. Without this information, it is difficult to place this modification into established chromatin regulatory pathways.

Due to broad target specificity, redundancies and crosstalk among different classes of HATs and HDACs, it is not tractable to answer this question in the current manuscript.

Reviewer #3 (Public reviews):

Reviewer 3 is mistaken in thinking our ChIP experiments are performed under cross-linked conditions. As clearly stated in the main text and methods, all our ChIP-seq for histone modifications is done on native MNase-digested chromatin – with no cross-linking. This includes the spike-in experiment shown in Fig S1B to test H3K115ac antibody specificity against the bar-coded SNAP-ChIP® K-AcylStat Panel from Epicypher. We could not include H3K115ac bar-coded nucleosomes in that experiment since they are not available in the panel.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) I have two primary concerns that resound through the entire paper:

(a) Overall, the manuscript is making strong claims based on entirely correlative datasets. No quantitative analyses are performed to demonstrate co-occupancy/localization. Please see more detailed descriptions below.

Our responses to specific points are provided against each comment below.

(b) Lack of paired-end replicates for H3K115ac ChIP-seq. While the reviewer token for the deposited data was not made accessible to me, looking at Supplementary Table 1, it appears there are two H3K115ac ChIP-seq datasets. One is paired-end and is single-read. So are peaks called with only one replicate of PE? Or are inaccurate peaks called with SR datasets? Either way, this is not a rigorous way to evaluate H3K115ac localization.

We are sorry that this reviewer was not able to access the data – the token for the GEO accession was provided for reviewers at the journal’s request. All ChIP-seq (and ATAC-seq) experiments (paired and single-end) were performed with two biological replicates and the Pearson’s correlations (all >0.9) between replicates were provided in Supplementary Table 1. This was indicated in both the main text and in the methods. In the revised manuscript we have tried to make this even clearer and have put the relevant Pearsons coefficient (r) into the text at the appropriate places. For the reviewer’s information, here is the complete list of data samples in the GEO Accession:

Author response image 1.

While I agree that H3K115ac occupancy is high at +CGIs, the authors downplay that H3K122ac and H3K27ac is also more highly enriched at these locations (page 7, last sentence of first paragraph). I imagine this is all due to the more highly transcribed nature of these genes. Sub-stratifying the K27ac and K122ac by transcription (as in Figure 1G) would help to demonstrate a unique nature of H3K115ac. But even better would be to do an analysis that plots H3K115ac enrichment vs transcription for every individual gene rather than aggregate analyses that are biased by single locations. For example, make an XY scatterplot of RNAPII occupancy or 4SU-seq signal vs H3K115ac level, where each point represents a single gene. Because the interpretation that it is CGI-based and not transcription is confounded with the fact that -CGI are more lowly transcribed. So, looking at Figure 1G, even the -CGI occupancy of H3K115ac is correlated with transcription, but it is just more lowly transcribed.

We thank the reviewer for these suggestions but point out that Figure 1G shows H3K115ac signal for CGI+ and CGI– TSS that are matched for expressions levels (quartiles of 4SU-seq). Fig 1F shows that H3k115ac is much more of a discriminator between CGI+ and – than H3K27ac or H3K122ac.

(2) H3K115ac, H3K27ac, and H3K122ac are all more enriched (in aggregate) at +CGI locations (Fig 1F); so do these locations just have more positioned nucleosomes? More H3.3? So that these PTMs are just more enriched due to the opportunity?

Positioned nucleosomes are generally found downstream of the TSS of active CpG island promoters, so what the reviewer suggests may well account for the relative enrichment of H327ac and H3K122ac at CGI+ vs CGI- promoters in Fig.1F. But H3K115ac localisation is distinct, with the peak at the nucleosome-depleted region not the +1 nucleosome. This is also confirmed by the contour plots in Fig 3. Our observation is also not explained by an enrichment of H3.3 at CGI promoters, since we show that H3K115ac is not specific to H3.3 (Fig 4D).

(3) The authors note in paragraph 2 of page 7 that "H3K115ac does not scale linearly with gene expression..." but the authors never show a quantification of this; stratification in four clusters is not able to make a linear correlation. Furthermore, in the second line of page 7, the authors state that the levels do generally correlate with transcription. To claim it is a specific CGI link and not transcription is tricky, but I encourage the authors to consider more quantifiable ways, rather than correlations, to demonstrate this point, if it is observed.

We thank the reviewer for this comment, and taking it into consideration, we have decided to re-phrase this paragraph. The new text reads: “Non-CGI promoters have lower overall levels of transcription compared to CGI promoters, and for this promoter class H3K115ac enrichment detected by ChIP is only really seen for the highest quartile of transcription (4SU) quartile of expression (Figure 1G). CGI promoters on the other hand, exhibit significant levels of detected H3K115ac even for the lowest quartile of expression. These results suggest a special link between CGI promoters and H3K115ac”.

(4) The authors claim on page 7 that "on average, transcription increased from TSS that also gained H3K115ac but to a modest extent, compared with the more substantial loss of H3K115ac from downregulated TSS". However, both upregulated and downregulated are significant; the difference in magnitude could simply be due to more highly or more lowly transcribed locations, meaning that fold change could be more robustly detected. I caution the authors to substantiate claims like this rather than stating a correlation.

We thank the reviewer for this comment which relates to the data in Fig 2A. It is Fig. 2B shows that the association of H3K115ac loss with downregulation is statistically stronger than H3K115ac gain with upregulation, but only for CGI promoters. With regard to the text on the original pg 7 that is referred to, we have now reworded this to read “Average levels of transcription increased from TSS that also gained H3K115ac, and there was loss of H3K115ac from downregulated TSS (Figure 2A).”

(5) For Figure 2C, the authors argue that H3K115ac correlate with bivalent locations. So this is all qualitative and aggregate localization; please quantitatively demonstrate this claim.

Figure S2D provides statistics for this (observed/expected and Fishers exact test).

(6) The authors claim in Figure 2 that H3115ac is dynamic during differentiation (title of Figure 2). However, there are locations that gain and lose, or maintain H3K115ac. In fact, the most discussed locations are H3K115ac with no change (2C); which means it is NOT dynamic during differentiation. So what is the message for the role during differentiation? From Supplemental Table 1, it appears there is a single ChIP experiment for H3K115ac in NPC, and it is a single read. So this is also a difficult claim with one replicate. Related to this, in S2A, the authors show K115ac where there is no change in transcription; so what is the role of H3K115ac at TSSs relevant to differentiation - it is at both locations changed and unchanged in transcription, but H3K115ac levels itself do not change at these subsets. So, how is this dynamic? This is very confusing, and clearer analyses and descriptions are necessary to deconvolute these data.

We apologise for the misleading title for Figure 2. This has now been amended to “Changes in H3K115ac during differentiation”. The message of this figure is that whilst changes in H3K115ac at TSS are small (panels A-C), at enhancers the changes are much more dramatic (panel D). The reviewer is incorrect about the number of replicates for NPCs – there are two biological replicates (see response to point 1b).

(7) The authors go on to examine H3K115ac enrichment on fragile nucleosomes through sucrose gradient sedimentation. A control for H3K27ac or H3K122ac would be nice for comparison.

We do not have the material available to perform these experiments

(8) When discussing Figures 3 and SF3, the authors mention performing a different MNase for a second ChIP. Showing the MNase distribution for both the more highly digested and the lowly digested would be nice. a) Related to the above, the authors show input in SF3E to argue that the difference in H3K115ac vs H3K27ac is not due to the library, but they do not show the MNase digestion patterns, which is more important for this argument.

Input libraries (first two graphs of FigS3E) are the MNase-digested chromatin. Comparison of nucleotide frequencies from millions of reads is more robust method than the fragment length patterns.

(9) The authors move on to examine H3K115ac at enhancers. Just out of curiosity, given what was found at promoters, is H3K115ac enriched at +CGI enhancers? And what is the correlation with enhancer transcription?

This is an interesting point, but the number of enhancers associated with CGI is not very high and so we did not focus on this. We have not analysed a correlation with eRNAs in this paper.

(10) The authors state on page 14 that the most frequent changes in H3K115ac during differentiation are at these enhancers. So do these changes connect with differentiation-specific genes, and/or genes that have altered transcription during differentiation? Just trying to understand the functional role.

Given the challenges of connecting enhancers with target genes, we have not addressed this question quantitatively. However, we draw the reviewer’s attention to the Genome Browser shots in Figures 2D and S2C, which show clear gain of H3K115ac (and ATAC-seq peaks) at intra and intergenic regions close to genes whose transcription is activated during the differentiation to NPCs.

(11) Related, at the end of page 14, the authors state that the changes in H3K115ac correlate with changes in ATAC-seq; I imagine this dynamic is not unique for H3K115ac and this is observed for other PTMs (H3K27ac), so assessing and clarifying this, to again get to the specific interest of H3K115ac, would be ideal.

We have not claimed that chromatin accessibility is unique to H3K115ac. It is the location of H3K115ac which is found inside the ATAC-seq peak region while H3K27ac is found only upstream/downstream of the ATAC peak that is so striking. This is apparent in Fig 4C.

(12) The authors examine levels of H3K115ac in H3.3 KO cell lines via western blot (Figure 4D), but no replicates and/or quantification are shown.

We now provide a biological replicate for the Western Blot (new FigS4H) together with an image of the whole gel for the data in Fig 4D

(13) In Figure S4 and at the end of page 17, the authors are arguing that there is a link to pioneer TF complexes, based on Oct4 binding. First, while Oct4 has pioneering activity, not all Oct4 sites (or motifs) are pioneering; this has been established. So if you want to use Oct4, substratifying by pioneer vs no pioneer is necessary. Second, demonstrating this is unique to pioneer and not to non-pioneer TFs would be an important control.

In response to the reviewer’s comment, we have removed the term “pioneer” from the manuscript.

(14) Minor point: Figure 4 A and B, there are some formatting issues with the scale bars.

We thank the reviewer for pointing this out, and the errors have been corrected in the revised figure.

(15) Minor point is that it should be clear when single replicates of data are used and when PE/SR sequences are combined or which one is used in each analysis, as this was hard to discern when reading the paper and figure legends.

We have clearly stated in the text that, after Figure2, we repeated all experiments in paired-end mode. All processing steps are defined separately for single end and paired end datasets in the method section. Details of biological replicates are provided in Sup. Table 1. These concerns are also addressed in our response to Reviewer’s public comment-1.

(16) Minor point: it is surprising that different MNase and different units were used in the ChIP vs sucrose sedimentation. Could the authors clarify why?

Chromatin prep for sucrose gradients were done on a much larger scale than for ChIP-seq and required different setups to obtain the right level of MNase digestion.

(17) The authors note that fragile nucleosomes contain H2A.Z and H3.3, but they never perform an analysis of available data to demonstrate a correlation (or better a quantifiable correlation) between H3K115ac occupancy and these marks at the locations they identify H3K115ac.

Since have shown (Fig. 4) that depletion of H3.3 does not affect overall levels of H3K115ac, we do not think there is value in further quantitative correlative analyses of H3K115ac and variant histones.

(18) Minor point: What is the overlap in peaks for H3K115ac, H3K122ac, and H3K27ac (Figure 1C)?

Nearly all H3K115ac peaks overlap with H3K122ac and/or H3K27ac. Its most distinct properties are its association with CGI promoters, fragile nucleosomes and its unique localisation within the NDRs, three points that the manuscript is focussed on.

Reviewer #3 (Recommendations for the authors):

(1) The western blot results in Figure 4D probing for H3, H3.3, and H3K115ac use Ponceau S staining, presumably of an area of the membrane where histones might be expected to migrate, as a measure of loading. However, the Ponceau S bands appear uniformly weaker in the H3.3KO lanes, yet despite this, blotting with H3.3 antibody detects a band in H3.3 knockout ESCs, suggesting that the antibody does not have a high degree of specificity. Again, a blocking experiment with appropriate peptides would instill more confidence in the specificity of these reagents, and/or the authors could provide independent validation of the knockout model to differentiate between a partial knockout or antibody cross-reactivity (e.g., by Sanger sequencing).

In a revised Fig. S4H we now show the whole gel corresponding to this blot but including co-staining with an antibody for H4 to provide a better loading control. We also provide a biological replicate of this Western blot in the lower panel of Fig. S4H.

(2) The manuscript would benefit from in vitro follow-up and validation, but if the authors intend to keep the manuscript primarily in silico, I suggest dedicating a few lines in each section to explain the plots, their axes, and their purpose, as well as to assist with interpretation, rather than directly discussing the results. This would make the manuscript more accessible and understandable for a broader audience in the field of epigenetics.

In the revised version, we have tried to improve the text to make the data more accessible to a broad audience.

eLife Assessment

In this important study, the authors conducted extensive sets of computational and investigations of the mechanism of cholesterol transport in the smoothened (SMO) protein. The computational component integrated multiple state-of-the-art approaches such as adaptive sampling, free energy simulations, and Markov state modeling, providing compelling support for the proposed mechanistic model, which is further validated with solid experimental mutagenesis data.

Reviewer #1 (Public review):

Summary:

This manuscript uses primarily simulation tools to probe the pathway of cholesterol transport with the smoothened (SMO) protein. The pathway to the protein and within SMO is clearly discovered and interactions deemed important are tested experimentally to validate the model predictions.

Strengths:

The authors have clearly demonstrated how cholesterol might go from the membrane through SMO for the inner and outer leaflets of a symmetrical membrane model. The free energy profiles, structural conformations and cholesterol-residue interactions are clearly described.

Weaknesses:

None. I find the revised manuscript strong and the work should be published.

Reviewer #2 (Public review):

Summary:

In this work, the authors applied a range of computational methods to probe the translocation of cholesterol through the Smoothened receptor. They test whether cholesterol is more likely to enter the receptor straight from the outer leaflet of membrane or via a binding pathway in the inner leaflet first. Their data reveal that both pathways are plausible but that the free energy barriers of pathway 1 is lower suggesting this route is preferable. They also probe the pathway of cholesterol transport from the transmembrane region to the cysteine-rich domain (CRD).

Strengths:

A wide range of computational techniques are used, including potential of mean force calculations, adaptative sampling, dimensionality reduction using tICA, and MSM modelling. These are all applied in a rigorous manner and the data are very convincing. The computational work is an exemplar of a well-carried out study.

Their computational predictions are experimentally supported using mutagenesis, with an excellent agreement between their PMF and mRNA fold change data.

The data are described clearly and coherently, with excellent use of figures. They combine their findings into a mechanism for cholesterol transport, which on the whole seems sound.

Their methods are described well, and much of their analysis methods have been made available via GitHub, which is an additional strength.

Reviewer #3 (Public review):

This manuscript presents a study combining molecular dynamics simulations and Hedgehog (Hh) pathway assays to investigate cholesterol translocation pathways to Smoothened (SMO), a G protein-coupled receptor central to Hedgehog signal transduction. The authors identify and characterize two putative cholesterol access routes to the transmembrane domain (TMD) of SMO and propose a model whereby cholesterol traverses through the TMD to the cysteine-rich domain (CRD), which is presented as the primary site of SMO activation.

The MD simulations and biochemical experiments are carefully executed and provide useful data.

Comments on revisions:

I appreciate the authors' detailed response and the substantial revisions made to the manuscript. The changes addressing Comments 3.1-3.5 have significantly improved the balance and framing of the work, and my primary concerns regarding overstatement and selective interpretation have been satisfactorily addressed.

The authors' rebuttal to my initial review includes extended argumentation regarding specific interpretations of prior studies and broader models of SMO regulation. These issues represent longstanding differences in interpretation that have already been discussed extensively in the literature and are not essential to evaluating the quality or conclusions of the present study.

For readers seeking a comprehensive and balanced overview of cholesterol-dependent SMO activation that integrates both CRD- and TMD-centered models, I would point to recent review articles (e.g., Zhang and Beachy, Nat Rev Mol Cell Biol2023). I do not feel it is productive to rehash these debates further in the context of this review, and I have no additional substantive concerns with the revised manuscript.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

Summary:

This manuscript uses primarily simulation tools to probe the pathway of cholesterol transport with the smoothened (SMO) protein. The pathway to the protein and within SMO is clearly discovered, and interactions deemed important are tested experimentally to validate the model predictions.

Strengths:

The authors have clearly demonstrated how cholesterol might go from the membrane through SMO for the inner and outer leaflets of a symmetrical membrane model. The free energy profiles, structural conformations, and cholesterol-residue interactions are clearly described.

We thank the reviewer for their kind words.

(1) Membrane Model: The authors decided to use a rather simple symmetric membrane with just cholesterol, POPC, and PSM at the same concentration for the inner and outer leaflets. This is not representative of asymmetry known to exist in plasma membranes (SM only in the outer leaflet and more cholesterol in this leaflet). This may also be important to the free energy pathway into SMO. Moreover, PE and anionic lipids are present in the inner leaflet and are ignored. While I am not requesting new simulations, I would suggest that the authors should clearly state that their model does not consider lipid concentration leaflet asymmetry, which might play an important role.

We thank the reviewer for their comment. Membrane asymmetry is inherent in endogenous systems; we acknowledge that as a limitation of our current model. We have addressed the comment by adding this limitation to our discussion in the manuscript.

Added lines: (End of paragraph 6, Results subsection 2):

“One possibility that might alter the thermodynamic barriers is native membrane asymmetry, particularly the anionic lipid-rich inner leaflet. This presents as a limitation of our current model.”

(2) Statistical comparison of barriers: The barriers for pathways 1 and 2 are compared in the text, suggesting that pathway 2 has a slightly higher barrier than pathway 1. However, are these statistically different? If so, the authors should state the p-value. If not, then the text in the manuscript should not state that one pathway is preferred over the other.

We thank the reviewer for their comment. We have added statistical t-tests for the barriers.

Changes made: (Paragraph 6, Results subsection 2)

“However, we also observe that pathway 1 shows a lower thermodynamic barrier (5.8 ± 0.7 kcal/mol v/s 6.5 ± 0.8 kcal/mol, p = 0.0013)”

(3) Barrier of cholesterol (reasoning): The authors on page 7 argue that there is an enthalpy barrier between the membrane and SMO due to the change in environment. However, cholesterol lies in the membrane with its hydroxyl interacting with the hydrophilic part of the membrane and the other parts in the hydrophobic part. How is the SMO surface any different? It has both characteristics and is likely balanced similarly to uptake cholesterol. Unless this can be better quantified, I would suggest that this logic be removed.

We thank the reviewer for this suggestion. We have removed the line to avoid confusion.

Reviewer #2 (Public review):

Summary:

In this work, the authors applied a range of computational methods to probe the translocation of cholesterol through the Smoothened receptor. They test whether cholesterol is more likely to enter the receptor straight from the outer leaflet of the membrane or via a binding pathway in the inner leaflet first. Their data reveal that both pathways are plausible but that the free energy barriers of pathway 1 are lower, suggesting this route is preferable. They also probe the pathway of cholesterol transport from the transmembrane region to the cysteine-rich domain (CRD).

Strengths:

(1) A wide range of computational techniques is used, including potential of mean force calculations, adaptive sampling, dimensionality reduction using tICA, and MSM modelling. These are all applied rigorously, and the data are very convincing. The computational work is an exemplar of a well-carried out study.

(2) The computational predictions are experimentally supported using mutagenesis, with an excellent agreement between their PMF and mRNA fold change data.

(3) The data are described clearly and coherently, with excellent use of figures. They combine their findings into a mechanism for cholesterol transport, which on the whole seems sound.

(4) The methods are described well, and many of their analysis methods have been made available via GitHub, which is an additional strength.

Weaknesses:

(1) Some of the data could be presented a little more clearly. In particular, Figure 7 needs additional annotation to be interpretable. Can the position of the cholesterol be shown on the graph so that we can see the diameter change more clearly?

We thank the reviewer for this suggestion. We have added the cholesterol positions as requested.

Changes made: (Caption, Figure 7)

“The tunnel profile during cholesterol translocation in SMO. (a) Free energy plot of the zcoordinate v/s the tunnel diameter when cholesterol is present in the core TMD. The tunnel shows a spike in the radius in the TMD domain, indicating the presence of a cholesterol-accommodating cavity. (b) Representative figure for the tunnel when a cholesterol molecule is in the TMD. (c) Same as (a), when cholesterol is at the TMD-CRD interface. (e) same as (b), when cholesterol is at the TMD-CRD interface. (e) same as (a), when cholesterol is at the CRD binding site. (f) same as (b), when cholesterol is at the CRD binding site. Tunnel diameters shown as spheres. Cholesterol positions marked on plots using dotted lines. All snapshots presented are frames taken from MD simulations.”

(2) In Figure 3C, it doesn’t look like the Met is constricting the tunnel at all. What residue is constricting the tunnel here? Can we see the Ala and Met panels from the same angle to compare the landscapes? Or does the mutation significantly change the tunnel? Why not A283 to a bulkier residue? Finally, the legend says that the figure shows that cholesterol can still pass this residue, but it doesn’t really show this. Perhaps if the HOLE graph was plotted, we could see the narrowest point of the tunnel and compare it to the size of cholesterol.

We thank the reviewer for this suggestion. A283 was mutated to methionine as it presents with a longer heavy tail containing sulfur. We have plotted the tunnel radii for both WT and A283M mutants and added them as a supplemental figure. As shown in the figure, the presence of methionine doesn’t completely block the tunnel, but occludes it, thereby increasing the barrier for cholesterol transport slightly.

Changes made: (End of Results subsection 1)

“When we calculated the PMF for cholesterol entry, A^2.60fM mutant showed restricted tunnel but it did not fully block the tunnel (Figure 3—figure Supplement 3).”

(3) The PMF axis in 3b and d confused me for a bit. Looking at the Supplementary data, it’s clear that, e.g., the F455I change increases the energy barrier for chol entering the receptor. But in 3d this is shown as a -ve change, i.e., favourable. This seems the wrong way around for me. Either switch the sign or make this clearer in the legend, please.

We thank the reviewer for this suggestion. We measured ∆PMF as PMF_WT PMF_mutant, hence the negative values. We have added additional text to the legend to clarify this.

Changes made: (Caption, Figure 3)

“(b) ∆Gli1 mRNA fold change (high SHH vs untreated) and ∆ PMF (difference of peak PMF , calculated as PMF_WT - PMF_mutant) plotted for the mutants in Pathway 1. (c) Example mutant A^2_._60fM shows that cholesterol can enter SMO through Pathway 1 even on a bulky mutation. (d) Same as (b) but for Pathway 2 (e) Example mutant L^5.62fA shows that cholesterol can enter SMO through Pathway 2 due to lesser steric hindrance. All snapshots presented are frames taken from MD simulations.”

Changes made: (Caption, Figure 6)

“(b) ∆Gli1 mRNA fold change (high SHH vs untreated) and ∆ PMF (difference of peak PMF, calculated as PMF_WT - PMF_mutant) plotted for mutants along the TMD-CRD pathway. (c, d) Example mutants Y^LDA and F^5.65fA show that cholesterol is unable to translocate through this pathway because of the loss of crucial hydrophobic contacts provided by Y207 and F484 and along the solvent-exposed pathway.”

(4) The impact of G280V is put down to a decrease in flexibility, but it could also be a steric hindrance. This should be discussed.

We thank the reviewer for this suggestion. We have added it as a possible mechanism of the decrease in activity of SMO.

Changes made: (Paragraph 5, Results subsection 1)

“We mutated G280^2.57f  to valine - G^2.57fV to test whether reducing the flexibility of TM2 prevents cholesterol entry into the TMD. Consequently, the activity of mSMO showed a decrease. However, this decrease could also be attributed to steric hindrance added by the presence of a bulky propyl group in valine.”

(5) Are the reported energy barriers of the two pathways (5.8plus minus0.7 and 6.5plus minus0.8 kcal/mol) significantly and/or substantially different enough to favour one over the other? This could be discussed in the manuscript.

We thank the reviewer for this suggestion. We have added statistical t-tests for the barriers.

Changes made: (Paragraph 6, Results subsection 2)

“However, we also observe that pathway 1 shows a lower thermodynamic barrier (5.8 ± 0.7 kcal/mol v/s 6.5 ± 0.8 kcal/mol, p = 0.001)”

(6) Are the energy barriers consistent with a passive diffusion-driven process? It feels like, without a source of free energy input (e.g., ion or ATP), these barriers would be difficult to overcome. This could be discussed.

We thank the reviewer for this suggestion. We have added a discussion to further clarify this point.

Discussion: (Paragraph 6, Results subsection 2)

“These values are comparable to ATP-Binding Cassette (ABC) transporters of membrane lipids, which use ATP hydrolysis (-7.54 ± 0.3 kcal/mol) (Meurer et al., 2017) to drive lipid transport from the membrane to an extracellular acceptor. Some of these transporters share the same mechanism as SMO, where the lipid from the inner leaflet is flipped and transported to the extracellular acceptor protein (Tarling et al., 2013). Additionally, for secondary active transporters that do not use ATP for the transport of substrates, a thermodynamic barrier of 5-6 kcal/mol has been reported in literature. (Chan et al., 2022; Selvam et al., 2019; McComas et al., 2023; Thangapandian et al., 2025).”

(7) Regarding the kinetics from MSM, it is stated that the values seen here are similar to MFS transporters, but this then references another MSM study. A comparison to experimental values would support this section a lot.

We thank the reviewer for this suggestion. We have added a discussion discussing millisecond-scale timescales measured for MFS transporters.

Changes made: (Paragraph 2, Results subsection 5)

“These timescales are comparable to the substrate transport timescales of Major Facilitator Superfamily (MFS) transporters (Chan et al., 2022). Furthermore, several experimental studies have also resolved the millisecond-scale kinetics of MFS transporters (Blodgett and Carruthers, 2005; Körner et al., 2024; Bazzone et al., 2022; Smirnova et al., 2014; Zhu et al., 2019), further corroborating the results from our study.”

Reviewer #2 (Recommendations for the authors):

(1) The heatmaps in Figures 2a and 4a are great. On these, an arrow denotes what looks like a minimum energy path. Is it possible to see this plotted, as this might show the height of the energy barriers more clearly?

We thank the reviewer for this suggestion. We have computed the minimum energy paths for both pathways and presented them in a supplementary figure.

Added lines: (Paragraph 4, Results subsection 1):

For further clarity, we have plotted the minimum energy path taken by cholesterol as it translocates along this pathway (Figure 2—figure Supplement 3)a,b)

Added lines: (Paragraph 4, Results subsection 2):

For further clarity, we have plotted the minimum energy path taken by cholesterol as it translocates along this pathway (Figure 2—figure Supplement 3)c,d)

(2) The tiCA data in S15 is first referred to on line 137, but the technique isn’t introduced until line 222. This makes understanding the data a little confusing. Reordering this might improve readability.

We thank the reviewer for this suggestion. We have reordered the text to make it clearer.

Changes made: (Paragraph 2, Results subsection 1) This provides evidence for multiple stable poses along the pathway as observed in the multiple stable poses of cholesterol in Cryo-EM structures of SMO bound to sterols (Deshpande et al., 2019; Qi et al., 2019b, 2020). A reliable estimate of the barriers comes from using the time-lagged Independent Components (tICs), which project the entire dataset along the slowest kinetic degrees of freedom. Overall, the highest barrier along Pathway 1 is ∼ 5.8 ± 0.7 kcal/mol, and it is associated with the entry of cholesterol into the TMD (Figure 2—Figure Supplement 2).

Changes made: (Paragraph 3, Results subsection 2)

“On plotting the first two components of tICs, (Figure 2—Figure Supplement 2), we observe that the energetic barrier between η and θ is ∼6.5 ± 0.8 kcal/mol.”

(3) Missing bracket on line 577.

We thank the reviewer for this suggestion. The typo has been fixed.

(4) Line 577: Fig. S2nd?

We thank the reviewer for this suggestion. This typo has been fixed.

Reviewer #3 (Public review):

Summary:

This manuscript presents a study combining molecular dynamics simulations and Hedgehog (Hh) pathway assays to investigate cholesterol translocation pathways to Smoothened (SMO), a G protein-coupled receptor central to Hedgehog signal transduction. The authors identify and characterize two putative cholesterol access routes to the transmembrane domain (TMD) of SMO and propose a model whereby cholesterol traverses through the TMD to the cysteine-rich domain (CRD), which is presented as the primary site of SMO activation. The MD simulations and biochemical experiments are carefully executed and provide useful data.

Weaknesses:

However, the manuscript is significantly weakened by a narrow and selective interpretation of the literature, overstatement of certain conclusions, and a lack of appropriate engagement with alternative models that are well-supported by published data-including data from prior work by several of the coauthors of this manuscript. In its current form, the manuscript gives a biased impression of the field and overemphasizes the role of the CRD in cholesterol-mediated SMO activation. Below, I provide specific points where revisions are needed to ensure a more accurate and comprehensive treatment of the biology.

(1) Overstatement of the CRD as the Orthosteric Site of SMO Activation

The manuscript repeatedly implies or states that the CRD is the orthosteric site of SMO activation, without adequate acknowledgment of alternative models. To give just a few examples (of many in this manuscript):

(a) “PTCH is proposed to modulate the Hh signal by decreasing the ability of membrane cholesterol to access SMO’s extracellular cysteine-rich domain (CRD)” (p. 3).

(b) “In recent years, there has been a vigorous debate on the orthosteric site of SMO” (p. 3).

(c) “cholesterol must travel through the SMO TMD to reach the orthosteric site in the CRD” (p. 4).

(d) “we observe cholesterol moving along TM6 to the TMD-CRD interface (common pathway, Fig. 1d) to access the orthosteric binding site in the CRD” (p. 6).

While the second quote in this list at least acknowledges a debate, the surrounding text suggests that this debate has been entirely resolved in favor of the CRD model. This is misleading and not reflective of the views of other investigators in the field (see, for example, a recent comprehensive review from Zhang and Beachy, Nature Reviews Molecular and Cell Biology 2023, which makes the point that both the CRD and 7TM sites are critical for cholesterol activation of SMO as well as PTCH-mediated regulation of SMO-cholesterol interactions).

In contrast, a large body of literature supports a dual-site model in which both the CRD and the TMD are bona fide cholesterol-binding sites essential for SMO activation. Examples include:

(a) Byrne et al., Nature 2016: point mutation of the CRD cholesterol binding site impairs-but does not abolish-SMO activation by cholesterol (SMO D99A, Y134F, and combination mutants - Fig 3 of the 2016 study).

(b) Myers et al., Dev Cell 2013 and PNAS 2017: CRD deletion mutants retain responsiveness to PTCH regulation and cholesterol mimetics (similar Hh responsiveness of a CRD deletion mutant is also observed in Fig. 4 Byrne et al, Nature 2016).

(c) Deshpande et al., Nature 2019: mutation of residues in the TMD cholesterol binding site blocks SMO activation entirely, strongly implicating the TMD as a required site, in contrast to the partial effects of mutating or deleting the CRD site.

Qi et al., Nature 2019, and Deshpande et al., Nature 2019, both reported cholesterol binding at the TMD site based on high-resolution structural data. Oddly, Deshpande et al., Nature 2019, is not cited in the discussion of TMD binding on p. 3, despite being one of the first papers to describe cholesterol in the TMD site and its necessity for activation (the authors only cite it regarding activation of SMO by synthetic small molecules).

Kinnebrew et al., Sci Adv 2022 report that CRD deletion abolished PTCH regulation, which is seemingly at odds with several studies above (e.g., Byrne et al, Nature 2016; Myers et al, Dev Cell 2013); but this difference may reflect the use of an N-terminal GFP fusion to SMO in the Kinnebrew et al 2022, which could alter SMO activation properties by sterically hindering activation at the TMD site by cholesterol (but not synthetic SMO agonists like SAG); in contrast, the earlier work by Byrne et al is not subject to this caveat because it used an untagged, unmodified form of SMO.

Although overexpression of PTCH1 and SMO (wild-type or mutant) has been noted as a caveat in studies of CRD-independent SMO activation by cholesterol, this reviewer points out that several of the studies listed above include experiments with endogenous PTCH1 and low-level SMO expression, demonstrating that SMO can clearly undergo activation by cholesterol (as well as regulation by PTCH1) in a manner that does not require the CRD.

Recommendation: The authors should revise the manuscript to provide a more balanced overview of the field and explicitly acknowledge that the CRD is not the sole activation site. Instead, a dual-site model is more consistent with available structural, mutational, and functional data. In addition, the authors should reframe their interpretation of their MD studies to reflect this broader and more accurate view of how cholesterol binds and activates SMO.

We thank the reviewer for this comprehensive overview of the existing literature. We agree that cholesterol binding to both the TMD and CRD sites is required for full activation of SMO. As described below in responses to comments, we have made changes to the manuscript to make this point clear. For instance, in the revised manuscript, we refrain from calling the CRD cholesterol binding site the “orthosteric site”. Instead, we highlight that the goal of the manuscript is not to resolve the debate over whether the TMD or CRD site is more important for PTCH1 regulation by SMO but rather to use molecular dynamics to understand the fascinating question of how cholesterol in the membrane can reach the CRD, located at a significant distance above the outer leaflet of the membrane. We believe that this is an important goal since there is an abundance of evidence that supports the view that PTCH1 inhibits SMO by reducing cholesterol access to the CRD. This evidence is now summarized succinctly in the introduction:

Changes made: (Paragraph 4, Introduction)

“While cholesterol binding to both the TMD and CRD sites is required for full SMO activation, our work focuses on how cholesterol gains access to the CRD site, perched above the outer leaflet of the membrane (Luchetti et al., 2016; Kinnebrew et al., 2022). Multiple lines of evidence suggest that PTCH1-regulated cholesterol binding to the CRD plays an instructive role in SMO regulation both in cells and animals. Mutations in residues predicted to make hydrogen bonds with the hydroxyl group of cholesterol bound to the CRD reduced both the potency and efficacy of SHH in cellular signaling assays (Kinnebrew et al., 2022; Byrne et al., 2016) and, more importantly, eliminated HH signaling in mouse embryos (Xiao et al., 2017). Experiments using both covalent and photocrosslinkable sterol probes in live cells directly show that PTCH1 activity reduces sterol access to the CRD (Kinnebrew et al., 2022; Xiao et al., 2017). Notably, our simulations evaluate a path of cholesterol translocation that includes both the TMD and CRD sites: cholesterol first enters the 7-transmembrane domain bundle from the membrane; it then engages the TMD site before continuing along a conduit to the CRD site. Thus, we analyze translocation energetics and residue-level contacts along a path that includes both the TMD and the CRD.”

However, Reviewer 3 makes several comments below that are biased, inaccurate, or selective. We feel it is important to address these so readers can approach the literature from a balanced perspective. Indeed, the eLife review forum provides an ideal venue to present contrasting views on a scientific model. We encourage the editors to publish both Reviewer 3’s comments and our response in full so readers can read the original papers and reach their own conclusions. It is important to note these issues are not relevant to the quality of the computational and experimental data presented in this paper.

We have now removed the term “orthosteric” to describe the CRD site throughout the paper and clearly state in the introduction that “both the CRD and TMD sites are required for SMO activation” but that our focus is on how cholesterol moves from the membrane to the CRD site. There is no doubt that cholesterol binding to the CRD plays a key role in SMO activation– our focus on this path is justified and does not devalue the importance of the TMD site. Our prior models (see Figure 7 of Kinnebrew 2022 explicitly include contributions of both sites).

Now we respond to some of the concerns outlined, individually:

(1) Byrne et al., Nature 2016: point mutation of the CRD cholesterol binding site impairs-but does not abolish-SMO activation by cholesterol (SMO D99A, Y134F, and combination mutants - Fig 3 of the 2016 study)

The fact that a point mutation dramatically diminishes (but does not abolish signaling) does not mean that the CRD cholesterol binding site is not important for SMO regulation. Indeed, the reviewer fails to mention that Song et. al. (Molecular Cell, 2017) found that a SMO protein carrying a subtle mutation at D99 (D95/99N, a residue that makes a hydrogen bond with the cholesterol hydroxyl) completely abolishes SMO signaling in mouse embryos. Thus, the CRD site is critical for SMO activation in an intact animal, justifying our focus on evaluating the path of cholesterol translocation to the CRD site.

(2) Myers et al., Dev Cell 2013 and PNAS 2017: CRD deletion mutants retain responsiveness to PTCH regulation and cholesterol mimetics (similar Hh responsiveness of a CRD deletion mutant is also observed in Fig 4 Byrne et al, Nature 2016).

The Reviewer fails to note that CRD-deleted versions of SMO have markedly (>10-fold) higher basal (i.e. ligand-independent) activity compared to full-length SMO. The response to SHH is minimal (∼2-fold), compared to >50-100-fold with full-length SMO. Thus, CRD-deleted SMO is likely in a non-native conformation. Local changes in cholesterol accessibility caused by PTCH1 inactivation or cholesterol loading can cause small fluctuations in delta-CRD activity, but this cannot be used to infer meaningful insights about how native, full-length SMO (with >10-fold lower basal activity) is regulated. We encourage the reviewer to read our previous paper (Kinnebrew et. al. 2022), which presents a unified view of how the TMD and CRD sites together regulate SMO activation.

A more physiological experiment, reported in Kinnebrew et. al. 2022, tested mutations in residues that make hydrogen bonds with cholesterol at the CRD and TMD sites in the context of full-length SMO. These mutants were stably expressed at moderate levels in Smo^−/− cells. Mutations at the CRD site reduced the fold-increase in signaling output in response to SHH, as would be expected for a PTCH1-regulated site. In contrast, analogous mutations in the TMD site reduced the magnitude of both basal and maximal signaling, without affecting the fold-change in response to SHH. In signaling assays, the key parameter in evaluating the impact of a mutation is whether it impacts the change in output in response to a signal (in this case PTCH1 inactivation by SHH). A mutation in SMO that affects PTCH1 regulation is expected to decrease the fold-change in signaling in response to SHH, a criterion that is fulfilled by mutations in the CRD site. Accordingly, mutations in the CRD site abolish SMO signaling in mouse embryos (Xiao et al., 2017).

(3) Deshpande et al., Nature 2019: mutation of residues in the TMD cholesterol binding site blocks SMO activation entirely, strongly implicating the TMD as a required site, in contrast to the partial effects of mutating or deleting the CRD site.

Introduction of bulky mutations at the TMD site (V333F) that abolish SMO activity were first reported by Byrne et. al. 2016 and were used to markedly increase the stability of SMO for protein expression. These mutations indeed stabilize the inactive state of SMO, increasing protein abundance and completely preventing its localization at primary cilia. SMO variants carrying such bulky mutations cannot be used to infer the importance of the TMD site since they do not distinguish between the following possibilities: (1) SMO is inactive because the sterol cannot bind, or (2) SMO is inactive because it is locked in an inactive conformation, or (3) SMO is inactive because it cannot localize to primary cilia (where it must be localized to activate downstream signaling).

As described in Response 3.3, a better evaluation of the importance of the TMD site is the use of mutations in residues that make hydrogen bonds with the hydroxyl group of TMD cholesterol. These mutations do not markedly increase protein stability or prevent ciliary localization (Kinnebrew 2022, Fig.S2). While a TMD site mutation decreases the magnitude of maximal (and basal) SMO signaling, it does not impact the fold-increase in signal output in response to Hh ligands (the key parameter that should be used to evaluate PTCH1 activity).

(4) Qi et al., Nature 2019, and Deshpande et al., Nature 2019, both reported cholesterol binding at the TMD site based on high-resolution structural data. Oddly, Deshpande et al., Nature 2019 not cited in the discussion of TMD binding on p. 3, despite being one of the first papers to describe cholesterol in the TMD site and its necessity for activation (the authors only cite it regarding activation of SMO by synthetic small molecules)

The reference has now been added at this location in the manuscript.

(5) Kinnebrew et al., Sci Adv 2022 report that CRD deletion abolished PTCH regulation, which is seemingly at odds with several studies above (e.g., Byrne et al, Nature 2016; Myers et al, Dev Cell 2013); but this difference may reflect the use of an N-terminal GFP fusion to SMO in the Kinnebrew et al 2022, which could alter SMO activation properties by sterically hindering activation at the TMD site by cholesterol (but not synthetic SMO agonists like SAG); in contrast, the earlier work by Byrne et al is not subject to this caveat because it used an untagged, unmodified form of SMO.

The reviewer fails to note that CRD deleted versions of SMO have markedly (>10-fold) higher basal activity than full-length SMO. The response to SHH is minimal (∼2fold), compared to >50-fold with full-length SMO. Thus, CRD-deleted SMO is likely in a non-native conformation. Local changes in cholesterol accessibility caused by PTCH1 inactivation or cholesterol loading can cause small fluctuations in delta-CRD activity, but this cannot be used to infer meaningful insights about how native, full-length SMO (with >10-fold lower basal activity) is regulated. Please see Response 3.3 for further details.

Reviewer 3 presents an incomplete picture of the extensive experiments reported in Kinnebrew et. al. to establish the functionality of YFP-tagged delta-CRD SMO. Most importantly, a TMDselective sterol analog (KK174) can fully activate YFP-tagged delta-CRD, showing conclusively that the YFP fusion does not block sterol access to the TMD site. The fact that this protein is nearly unresponsive to SHH highlights the critical role of the CRD-bound cholesterol in SMO regulation by PTCH1. Indeed, the YFP-tagged, CRD-deleted SMO was made purposefully to test the requirement of the CRD in a construct that had normal basal activity. Again, this data justifies the value of investigating the path of cholesterol movement from the membrane via the TMD site to the CRD.

(6) Although overexpression of PTCH1 and SMO (wild-type or mutant) has been noted as a caveat in studies of CRD-independent SMO activation by cholesterol, this reviewer points out that several of the studies listed above include experiments with endogenous PTCH1 and low-level SMO expression, demonstrating that SMO can clearly undergo activation by cholesterol (as well as regulation by PTCH1) in a manner that does not require the CRD.

This comment is inaccurate. The data presented in Deshpande et. al. (and prior work in Myers et. al.) used transient transfection to overexpress SMO in Smo^−/− cells. At the individual cell level transient transfection produces expression levels that are markedly higher (10-1000-fold) than stable expression (in addition to being more variable). Most scientists would agree that stable expression (as used in Kinnebrew 2022) at a moderate expression level is a better system to compare mutant phenotypes, assess basal and activated signaling, and provide an accurate measure of the fold-change in signal output in response to SHH. Notably, introduction of a mutation in the CRD cholesterol binding site at the endogenous mouse Smo locus (an even better experiment than stable expression) leads to complete loss of SMO activity (PMID 28344083). This result again justifies our investigation of the pathway of cholesterol movement from the membrane to the CRD site.

We have changed the initial discussion and reflect a more general outlook.

Changes made: (Paragraph 1, Introduction)

“PTCH modulates the availability of accessible cholesterol at the primary cilium and thereby regulates SMO, with models invoking effects on both the CRD and 7TM pockets.”

Changes made: (Results subsection 3, paragraph 1)

“According to the dual-site model, to reach the binding site in the CRD (ζ), cholesterol translocate along the TMD-CRD interface from the TM binding site (α∗) is required.”

Added lines: (Paragraph 5, Results subsection 3):

“The computational investigation showed here covers the dual-site model, where cholesterol reaches the CRD site via binding to the TM binding site first. In comparison to the CRD site, the TM site is more stable by ∼ 2 kcal/mol (Figure 2—Figure Supplement 3b, d).”

Added lines: (Paragraph 2, Conclusions):

“Here we have explored the role the CRD-site plays in SMO activation. In addition, through simulating the CRD site-dependent SMO activation hypothesis, we have also simulated the TMD site-dependent activation. We show that the overall stability of cholesterol is higher than the CRD site by ∼ 2 kcal/mol.”

(2) Bias in Presentation of Translocation Pathways

The manuscript presents the model of cholesterol translocation through SMO to the CRD as the predominant (if not sole) mechanism of activation. Statements such as: "Cholesterol traverses SMO to ultimately reach the CRD binding site" (p. 6) suggest an exclusivity that is not supported by prior literature in the field. Indeed, the authors’ own MD data presented here demonstrate more stable cholesterol binding at the TMD than at the CRD (p 17), and binding of cholesterol to the TMD site is essential for SMO activation. As such, it is appropriate to acknowledge that cholesterol may activate SMO by translocating through the TM5/6 tunnel, then binding to the TMD site, as this is a likely route of SMO activation in addition to the CRD translocation route they highlight in their discussion.

The authors describe two possible translocation pathways (Pathway 1: TM2/3 entry to TMD; Pathway 2: TM5/6 entry and direct CRD transfer), but do not sufficiently acknowledge that their own empirical data support Pathway 2 as more relevant. Indeed, because their experimental data suggest Pathway 2 is more strongly linked to SMO activation, this pathway should be weighted more heavily in the authors’ discussion. In addition, Pathway 2 is linked to cholesterol binding to both the TMD and CRD sites (the former because the TMD binding site is at the terminus of the hydrophobic tunnel, the latter via the translocation pathway described in the present manuscript), so it is appropriate that Pathway 2 figures more prominently than Pathway 1 in the authors’ discussion.

The authors also claim that "there is no experimental structure with cholesterol in the inner leaflet region of SMO TMD" (p 16). However, a structural study of apo-SMO from the Manglik and Cheng labs (Zhang et al., Nat Comm, 2022) identified a cholesterol molecule docked at the TM5/6 interface and also proposed a "squeezing" mechanism by which cholesterol could enter the TM5/6 pocket from the membrane. The authors do not consider this SMO conformation in their models, nor do they discuss the possibility that conformational dynamics at the TM5/6 interface could facilitate cholesterol flipping and translocation into the hydrophobic conduit, despite both possibilities having precedent in the 2022 empirical cryoEM structural analysis.

Recommendation: The authors should avoid oversimplifying the SMO cholesterol activation process, either by tempering these claims or broadening their discussion to better reflect the complexity and multiplicity of cholesterol access and activation routes for SMO. They should also consider the 2022 apo-SMO cryoEM structure in their analysis of the TM5/6 translocation pathway.

We thank the reviewer for this comprehensive overview of the existing literature and parts we have missed to include in the discussion. We agree with the reviewer, since our data shows that both pathways are probable. Through our manuscript, we have avoided using a competitive approach (that one pathway dominates over the other). Instead, we have evaluated both pathways independently and presented a comparative rather than competitive overview of both pathways from our observations. While we agree that experimental evidence suggests the inner leaflet pathway is possible, we cannot discount the observations made in previous studies that support the outer leaflet pathway, particularly Hedger et al. (2019), Bansal et al. (2023), and Kinnebrew et al. (2021). Therefore, considering the reviewer’s comments have made the following changes:

(1) Added lines: (Paragraph 3, Conclusions):

“We show that the barriers associated with the pathway starting from the outer leaflet are lower by ∼0.7 kcal, (p=0.0013). We also provide evidence that cholesterol can enter SMO via both leaflets, considering that multiple computational and experimental studies have found cholesterol entry sites and activation modulation via the outer leaflet, between TM2TM3. This is countered by evidence from multiple experimental and computational studies corroborating entry via the inner leaflet, between TM5-TM6, including this study. Overall, we posit that cholesterol translocation from either pathway is feasible.”

(2)nChanges made: (Paragraph 6, Results subsection 2)

“Based on our experimental and computational data, we conclude that cholesterol translocation can happen via either pathway. This is supported on the basis of the following observations: mutations along pathway 2 affect SMO activity more significantly, and the presence of a direct conduit that connects the inner leaflet to the TMD binding site. In addition, a resolved structure of SMO in the presence of cholesterol shows a cholesterol situated at the entry point from the membrane into the protein between TM5 and TM6, in the inner leaflet. However, we also observe that pathway 1 shows a lower thermodynamic barrier (5.8 ± 0.7 kcal/mol vs. 6.5 ± 0.8 kcal/mol, p \= 0.0013). Additionally, PTCH1 controls cholesterol accessibility in the outer leaflet. This shows that there is a possibility for transport from both leaflets. One possibility that might alter the thermodynamic barriers is native membrane asymmetry, particularly the anionic lipid-rich inner leaflet. This presents as a limitation of our current model.”

(3)nChanges made: (Paragraph 1, Results subsection 2)

“In a structure resolved in 2022, cholesterol was observed at the interface between the protein and the membrane, in the inner leaflet, between TMs 5 and 6. However, cholesterol in the inner leaflet has a downward orientation, with the polar hydroxyl group pointing intracellularly (η). A striking observation is that this cholesterol binding site pose was never used as a starting point for simulations and was discovered independent of the pose described in Zhang et al. (2022) (Figure 4—Figure Supplement 1).”

(3) Alternative Possibility: Direct Membrane Access to CRD

The possibility that the CRD extracts cholesterol directly from the membrane outer leaflet is not considered. While the crystal structures place the CRD in a stable pose above the membrane, multiple cryo-EM studies suggest that the CRD is dynamic and adopts a variety of conformations, raising the possibility that the stability of the CRD in the crystal structures is a result of crystal packing and that the CRD may be far more dynamic under more physiological conditions.

Recommendation: The authors should explicitly acknowledge and evaluate this potential mechanism and, if feasible, assess its plausibility through MD simulations.

We thank the reviewer for the suggestion. We have addressed this comment by calculating the distance from the lipid headgroups for each lipid in the membrane to the cholesterol binding site. We show that in our study, we do not observe any bending of the CRD over the membrane, precluding any cholesterol from being extracted from the membrane directly.

Added lines: (Paragraph 3, Conclusions):

“An alternative possibility states that the flexibility associated with the CRD would allow it to directly access the membrane, and consequently, cholesterol. In the extensive simulations reported in this study, the binding site of cholesterol in the CRD remains at least 20 Å away from the nearest lipid head group in the membrane, suggesting that such direct extraction and the bending of the CRD do not occur within the timescales sampled (Appendix 2 – Figure 6).

The mechanistic details of this process are still unexplored and form the basis of future work.”

(4) Inconsistent Framing of Study Scope and Limitations

The discussion contains some contradictory and misleading language. For example, the authors state that "In this study we only focused on the cholesterol movement from the membrane to the CRD binding site," and then several sentences later state that "We outline the entire translocation mechanism from a kinetic and thermodynamic perspective." These statements are at odds. The former appropriately (albeit briefly) notes the limited scope of the modeling, while the latter overstates the generality of the findings.

In addition, the authors’ narrow focus on the CRD site constitutes a major caveat to the entire work. It should be acknowledged much earlier in the manuscript, preferably in the introduction, rather than mentioned as an aside in the penultimate paragraph of the conclusion.

Recommendation: The authors should clarify the scope of the study and expand the discussion of its limitations. They should explicitly acknowledge that the study models one of several cholesterol access routes and that the findings do not rule out alternative pathways.

We thank the reviewer for the suggestion. We have addressed this comment by explicitly mentioning the scope of the study.

Changes made: (Paragraph 3, Conclusions)

“We outline the entire translocation mechanism from a kinetic and thermodynamic perspective for one of the leading hypotheses for the activation mechanism of SMO.”

(5) Summary:

This study has the potential to make a useful contribution to our understanding of cholesterol translocation and SMO activation. However, in its current form, the manuscript presents an overly narrow and, at times, misleading view of the literature and biological models; as such, it is not nearly as impactful as it could be. I strongly encourage the authors to revise the manuscript to include:

(1) A more balanced discussion of the CRD vs. TMD binding sites.

(2) Acknowledgment of alternative cholesterol access pathways.

(3) More comprehensive citation of prior structural and functional studies.

(4) Clarification of assumptions and scope.

Of note, the above suggestions require little to no additional MD simulations or experimental studies, but would significantly enhance the rigor and impact of the work.

We thank the reviewer for the suggestions. We have taken into account the literature and diverse viewpoints. We have changed the initial discussion and reflected a more general outlook. In the revised version of the manuscript, we have refrained from referring to the CRD site as the orthosteric site. Instead, we refer to it as the CRD sterol-binding site. To better represent the dual-site model, we add further discussion in the Introduction. Through our manuscript, we have avoided using a competitive approach (that one pathway dominates over the other). Instead, we have evaluated both pathways independently and presented a comparative rather than competitive overview of both pathways from our observations. We explicitly mention the scope of the study.

eLife Assessment

This study addresses a key, long-standing question about how visual feature selectivity is organized in mid-level visual cortex, using an ambitious combination of large-scale neural recordings and image synthesis. It provides important insights into the complexity of single-neuron selectivity and suggests a structured organization across cortical depth. While the evidence is generally solid and technically impressive, several key claims would be strengthened by additional controls, particularly regarding the sources of similarity across neurons and the dependence of the results on modeling choices.

Reviewer #1 (Public review):

Willeke et al. hypothesize that macaque V4, like other visual areas, may exhibit a topographic functional organization. One challenge to studying the functional (tuning) organization of V4 is that neurons in V4 are selective for complex visual stimuli that are hard to parameterize. Thus, the authors leverage an approach comprising digital twins and most exciting stimuli (MEIs) that they have pioneered. This data-driven, deep-learning framework can effectively handle the difficulty of parametrizing relevant stimuli. They verify that the model-synthesized MEIs indeed drive V4 neurons more effectively than matched natural image controls. They then performed psychophysics experiments (on humans) along with the application of contrastive learning to illustrate that anatomically neighboring neurons often care about similar stimuli. Importantly, the weaknesses of the approach are clearly appreciated and discussed.

Comments:

(1) The correlation between predictions and data is 0.43. I'd agree with the authors that this is "reliable" and would recommend that they discuss how the fact that performance is not saturated influences the results.

(2) Modeling V4 using a CNN and claiming that the identified functional groups look like those found in artificial vision systems may be a bit circular.

(3) No architecture other than ResNet-50 was tested. This might be a major drawback, since the MEIs could very well be reflections of the architecture and also the statistics of the dataset, rather than intrinsic biological properties. Do the authors find the same result with different architectures as the basis of the goal-driven model?

(4) The closed-loop analysis seems to be using a much smaller sample of the recorded neurons - "resulting in n=55 neurons for the analysis of the closed-loop paradigm".

(5) A discussion on adversarial machine learning and the adversarial training that was used is lacking.

Reviewer #2 (Public review):

This is an ambitious and technically powerful study, investigating a long-standing question about the functional organization of area V4. The project combined large-scale single-unit electrophysiology in macaque V4 with deep learning-based activation maximization to characterize neuronal tuning in natural image space. The authors built predictive encoding models for V4 neurons and used these models to synthesize most exciting images (MEIs), which are subsequently validated in vivo using a closed-loop experimental paradigm.

Overall, the manuscript advances three main claims:

(1) Individual V4 neurons showed complex and highly structured selectivity for naturalistic visual features, including textures, curvatures, repeating patterns, and apparently eye-like motifs.

(2) Neurons recorded along the same linear probe penetration tended to have more similar MEIs than neurons recorded at different cortical locations (this similarity was supported by human psychophysics and by distances in a learned, contrastive image embedding space).

(3) MEIs clustered into a limited number of functional groups that resembled feature visualizations observed in deep convolutional neural networks.

Strengths:

(1) The study is important in that it is the first to apply activation maximization to neurons sampled at such fine spatial resolution. The authors used 32-channel linear silicon probes, spanning approximately 2 mm of cortical depth, with inter-contact spacing of roughly 60 µm. This enabled fine sampling across most of the cortical thickness of V4, substantially finer resolution than prior Utah-array or surface-biased approaches.

(2) A key strength is the direct in vivo validation of model-derived synthetic images by stimulating the same neurons used to build the models, a critical step often absent in other neural network-based encoding studies.

(3) More broadly, the study highlights the value of probing neuronal selectivity with rich, naturalistic stimulus spaces rather than relying exclusively on oversimplified stimuli such as Gabors.

Weaknesses:

(1) A central claim is that neurons sampled within the same penetration shared MEI tuning properties compared to neurons sampled in different penetrations because of functional organization. I am concerned about technical correlations in activity due to technical or methodology-related approaches (for example, shared reference or grounding) instead of functional organization alone. These recordings were obtained with linear silicon probes, and there have been observations that neuronal activity along this type of probe (including neuropixels probes) may be correlated above what prior work showed, using manually advanced single electrodes. For example, Fujita et al. (1992) showed finer micro-domains and systematic changes in selectivity along a cortical penetration, and it is not clear if that is true or detectable here. I think that the manuscript would be strengthened by a more thorough and explicit characterization of lower-level response correlations (at the neuronal electrophysiology level) prior to starting with fitting models. In particular, the authors could examine noise correlations along the electrode shaft (using the repeated test images, for example), as well as signal correlations in tuning, both within and across sessions. It would also be helpful to clarify whether these correlations depended on penetration day, recording chamber hole (how many were used?), or spatial separation between penetrations, and whether repeated use of the same hole yielded stable or changing correlations. Illustrations of the peristimulus time histogram changes across the shaft and across penetrations would also help. All of this would help us understand if the reports of clustering were technically inevitable due to the technique.

(2) It is difficult to understand a story of visual cortex neurons without more information about their receptive field locations and widths, particularly given that the stimulus was full-screen. I understand that there was a sparse random dot stimulus used to find the population RF, so it should be possible to visualize the individual and population RFs. Also, the investigators inferred the locations of the important patches using a masking algorithm, but where were those masks relative to the retinal image, and how distributed were they as a function of the shaft location? This would help us understand how similar each contact was.

(3) A major claim is that V4 MEIs formed groups that were comparable to those produced by artificial vision systems, "suggesting potential shared encoding strategies." The issue is that the "shared encoding strategy" might be the authors' use of this same class of models in the first place. It would be useful to know if different functional groups arise as a function of other encoding neural network models, beyond the robust-trained ResNet-50. I am unsure to what extent the reported clustering, depth-wise similarity, and correspondence to artificial features depended on architectural and training bias. It would substantially strengthen the manuscript to test whether a similar organizational structure would emerge using alternative encoding models, such as attention-based vision transformers, self-supervised visual representations, or other non-convolutional architectures. Another important point of contrast would be to examine the functional groups encoded by the ResNet architecture before its activations were fit to V4 neuronal activity: put simply, is ResNet just re-stating what it already knows?

(4) Several comparisons to prior work are presented largely at a qualitative level, without quantitative support. For example, the authors state that their MEIs are consistent with known tuning properties of macaque V4, such as selectivity for shape, curvature, and texture. However, this claim is not supported by explicit image analyses or metrics that would substantiate these correspondences beyond appeal to visual inspection. Incorporating quantitative analyses, for instance, measures of curvature, texture statistics, or comparisons to established stimulus sets, would strengthen these links to prior literature and clarify the relationship between the synthesized MEIs and previously characterized V4 tuning properties.

Author response:

We thank the reviewers for their careful reading and constructive feedback. We were glad to see that they recognized both the technical scope of the study and its contribution as the first to apply activation maximization with such fine spatial sampling. Their appreciation for the critical in vivo validation of model-derived stimuli is very encouraging.

The reviewers raised several important points that we plan to address in the revised manuscript. These center on:

Model Architecture and Potential Circularity:

Both reviewers raised the concern that using a CNN-based model could introduce circularity when comparing V4 functional groups to artificial vision systems, and questioned whether similar results would emerge with alternative architectures. We believe that the in vivo verification provides a critical control for this concern: the MEIs synthesized by our model were empirically validated to elicit significantly higher responses than matched natural image controls, demonstrating that the model captures genuine biological tuning properties rather than architectural artifacts. This means that even if these features emerged from the particular architectural choice, the biological neurons seem to prefer the same features. We will clarify this point in the respective section in the revised manuscript.

Recording locations and spike sorting contamination:

Reviewer #2 raised concerns about potential correlation artefacts along the silicon probe. Unfortunately, assessing functional correlations across sessions proved challenging because neurons recorded at different penetration sites had non-overlapping receptive fields, precluding direct comparison of responses to identical stimuli across recording sites. We will make this limitation explicit in the manuscript. Furthermore, we maintain conservative standards for spike sorting to minimize the risk of multi-unit activity (MUA) "smearing" across unit definitions. Our primary analyses are restricted to well-isolated single units that meet all isolation metrics. Due to our low-impedance ground placed on the bone, shared-reference contamination as a source of tuning similarity is also mitigated.

Quantitative Comparisons to Prior Literature:

Reviewer #2 also noted that our comparisons between MEIs and known V4 tuning properties (e.g., shape, curvature, texture selectivity) were presented qualitatively, and suggested that explicit image analyses or metrics would strengthen these links to prior literature. We will revise the text to more carefully frame these comparisons as qualitative observations consistent with prior findings.

Alternative Similarity Metrics:

We will expand our justification for the Böhm et al. contrastive embedding approach in the Methods section. However, we believe that a systematic comparison of multiple clustering and similarity methods is beyond the scope of the current study.

In the revised manuscript, we will address these points primarily through clarifications and expanded discussion. Specifically, we will: (1) strengthen our discussion of model architecture choice emphasizing that in vivo verification serves as a critical control against architectural artifacts; (2) clarify the stringent matching criteria underlying our closed-loop sample size and its consistency with the larger population analyses; (3) explicitly describe the recording geometry, including the use of multiple grid holes, and explain why direct functional comparisons across penetrations were precluded by non-overlapping receptive fields; (4) better characterize the spatial relationship between receptive fields and MEI masks; (5) reframe comparisons to prior V4 literature as qualitative observations rather than quantitative validations; and (6) expand our justification for the contrastive embedding approach. We believe these revisions will improve the clarity and rigor of the manuscript while appropriately scoping the claims to what the current data support.

eLife Assessment

This valuable study introduces miRTarDS, a novel computational framework that predicts microRNA-target interactions based on a publicly available pretrained Sentence-BERT language model and downstream classification analysis. The strength of the evidence is incomplete, as the evaluation framework relies on unreliable ground-truth and false sets. Furthermore, the analysis fails to compare miRTarDS against existing state-of-the-art biomedical language models.

Reviewer #1 (Public review):

The author presents a new method for microRNA target prediction based on (1) a publicly available pretrained Sentence-BERT language model that the author fine-tunes using MeSH information and (2) downstream classification analysis for microRNA target prediction. In particular, the author's approach, named "miRTarDS", attempts to solve the microRNA target prediction problem by utilizing disease information (i.e., semantic similarity scores) from their language model. The author then compares the prediction performance with other sequence- and disease-based methods and attempts to show that miRTarDS is superior or at least comparable to existing methods. The author's general approach to this microRNA target prediction problem seems promising, but fails to demonstrate concrete computational evidence that miRTarDS outperforms other existing methods. The author's claim that disease information-based language models are sufficient is unfounded. The manuscript requires substantial rewriting and reorganization for readers with a strong background in biomedical research.

A major issue related to the author's claim of computational advance of miRTarDS: The author does not introduce existing biomedical-specific language models, and does not compare them against miRTarDS's fine-tuned model. The performance of miRTarDS is largely dependent on the semantic embedding of disease terms. The author shows in Figure 5 that MeSH-based fine-tuning leads to a substantial improvement in MeSH-based correlation compared to the publicly available pretrained SBERT model "multi-qa-MiniLM-L6-cos-v1" without sacrificing a large amount of BIOSSES-based correlation. However, the author does not compare the performance of MeSH- and BIOSSES-based correlation with existing language models such as ChatGPT, BioBERT, PubMedBERT, and more. Also, the substantial improvement in MeSH-based correlation is a mere indication that the MeSH-based fine-tuning strategy was reasonable and not that it's superior to the publicly available pretrained SBERT model "multi-qa-MiniLM-L6-cos-v1".

Another major issue is in the author's claim that disease-information from miRTarDS's language model is "sufficient" for accurate microRNA target prediction. Available microRNA targets with experimental evidence are largely biased for those with disease implications that have been reported in the biomedical literature. It's possible that their language model is biased by existing literature that has also been used to build microRNA target databases. Therefore, it is important that the author provides strong evidence that excludes the possibility of data leakage circularity. Similar concerns are prevalent across the manuscript, and so I highly recommend that the author reassess the evaluation frameworks and account for inflated performance, biased conclusions, and self-confirming results.

Last but not least, the manuscript requires a deeper and careful description and computational encoding of microRNA biology. I'd advise the author to include an expert in microRNA biology to improve the quality of this manuscript. For example, the author uses the pre-miRNA notation and replaces the mature miRNA notation to maintain computational encoding consistency across databases. However, the mature microRNA notation "the '-3p' or '-5p' is critical as the 3p and 5p mature microRNAs have different seed sequences and thus different mRNA targets. The 3p mature microRNA would most likely not target an mRNA targeted by the 5p mature microRNA.

Reviewer #2 (Public review):

Summary:

This study introduces a novel knowledge-driven approach, miRTarDS, which enables microRNA-Target Interaction (MTI) prediction by leveraging the disease association degree between a miRNA and its target gene. The core hypothesis is that this single feature is sufficient to distinguish experimentally validated functional MTIs from computationally predicted MTIs in a binary classification setting. To quantify the disease association, the authors fine-tuned a Sentence-BERT (SBERT) model to generate embeddings of disease descriptions and compute their semantic similarity. Using only this disease association feature, miRTarDS achieved an F1 score of 0.88 on the test set.

Strengths:

The primary strength is the innovative use of the disease association degree as an independent feature for MTI classification. In addition, this study successfully adapts and fine-tunes the Sentence-BERT (SBERT) model to quantify the semantic similarity between biomedical texts (disease descriptions). This approach establishes a critical pathway for integrating powerful language models and the vast growth in clinical/disease data into biochemical discovery, like MTI prediction.

Weaknesses:

The main weakness lies in its definition of the ground-truth dataset, which serves as a foundation for methodological evaluation. The study defines the Negative Set as computationally predicted MTIs that lack experimental evidence. However, the absence of experimental validation does not equate to non-functionality. Similarly, the miRAW sets are classified by whether the target and miRNA could form a stable duplex structure according to RNA structure prediction. This definition is biologically irrelevant, as duplex stability does not fully encapsulate the complex in vivo binding of miRNAs within the AGO protein complex.

Author response:

Public Reviews:

Reviewer #1 (Public review):

The author presents a new method for microRNA target prediction based on (1) a publicly available pretrained Sentence-BERT language model that the author fine-tunes using MeSH information and (2) downstream classification analysis for microRNA target prediction. In particular, the author's approach, named "miRTarDS", attempts to solve the microRNA target prediction problem by utilizing disease information (i.e., semantic similarity scores) from their language model. The author then compares the prediction performance with other sequence- and disease-based methods and attempts to show that miRTarDS is superior or at least comparable to existing methods. The author's general approach to this microRNA target prediction problem seems promising, but fails to demonstrate concrete computational evidence that miRTarDS outperforms other existing methods. The author's claim that disease information-based language models are sufficient is unfounded. The manuscript requires substantial rewriting and reorganization for readers with a strong background in biomedical research.

We appreciate the reviewer’s careful examination of modeling, benchmarking, and interpretation, and we are particularly encouraged that they found the proposed method promising. We will make corresponding revisions to the manuscript based on the reviewer’s comments.

A major issue related to the author's claim of computational advance of miRTarDS: The author does not introduce existing biomedical-specific language models, and does not compare them against miRTarDS's fine-tuned model. The performance of miRTarDS is largely dependent on the semantic embedding of disease terms. The author shows in Figure 5 that MeSH-based fine-tuning leads to a substantial improvement in MeSH-based correlation compared to the publicly available pretrained SBERT model "multi-qa-MiniLM-L6-cos-v1" without sacrificing a large amount of BIOSSES-based correlation. However, the author does not compare the performance of MeSH- and BIOSSES-based correlation with existing language models such as ChatGPT, BioBERT, PubMedBERT, and more. Also, the substantial improvement in MeSH-based correlation is a mere indication that the MeSH-based fine-tuning strategy was reasonable and not that it's superior to the publicly available pretrained SBERT model "multi-qa-MiniLM-L6-cos-v1".

We thank the reviewer for the constructive suggestions regarding the benchmarking of language models. We acknowledge that the performance of miRTarDS largely depends on the semantic embeddings of disease terms. So, in the revisions, I will: 1) conduct a literature review to introduce existing biomedical-specific language models, and 2) perform a horizontal comparison between our fine-tuned model and these existing models, to more comprehensively evaluate the model’s capabilities.

Another major issue is in the author's claim that disease-information from miRTarDS's language model is "sufficient" for accurate microRNA target prediction. Available microRNA targets with experimental evidence are largely biased for those with disease implications that have been reported in the biomedical literature. It's possible that their language model is biased by existing literature that has also been used to build microRNA target databases. Therefore, it is important that the author provides strong evidence that excludes the possibility of data leakage circularity. Similar concerns are prevalent across the manuscript, and so I highly recommend that the author reassess the evaluation frameworks and account for inflated performance, biased conclusions, and self-confirming results.

We thank the reviewer for the comment. We recognize that existing experimentally validated microRNA targets may be biased toward those reported in biomedical literature as disease‑related. To mitigate this bias, we attempted to extract predicted microRNA targets that share a very similar number of miRNA- and gene‑ disease entries as the experimentally validated microRNA targets using the K‑Nearest Neighbors (KNN) method. Then applied Positive‑Unlabeled (PU) Learning to classify the two groups. PU‑Learning is designed to address scenarios where only a subset of the training data is explicitly labeled as positive, while the remaining data are unlabeled—with the unlabeled set containing both potential positives and true negatives—which is highly suitable for the application context of this manuscript [1]. Preliminary results show that after applying the new data extraction and classification approach, model performance drops to around F1=0.73 (the MISIM method also shows a decline, with F1 around 0.58; detailed code is available on GitHub). The specific reasons for this require further investigation.

Last but not least, the manuscript requires a deeper and careful description and computational encoding of microRNA biology. I'd advise the author to include an expert in microRNA biology to improve the quality of this manuscript. For example, the author uses the pre-miRNA notation and replaces the mature miRNA notation to maintain computational encoding consistency across databases. However, the mature microRNA notation "the '-3p' or '-5p' is critical as the 3p and 5p mature microRNAs have different seed sequences and thus different mRNA targets. The 3p mature microRNA would most likely not target an mRNA targeted by the 5p mature microRNA.

We thank the reviewer for the critique and suggestion. We fully agree with the reviewer that the distinction between the 3p and 5p mature strands is critical for determining mRNA targeting, as they possess distinct seed sequences. In our study, we relied on the miRNA–disease associations provided by the HMDD database, which annotates interactions at the pre-miRNA level: “… the enriched functions of each mature miRNA are aggregated to the corresponding miRNA precursor.” [2] Furthermore, existing literature suggests that the pre-miRNA level can be appropriate and informative for disease association analyses: “Compared with the mature miRNA method, the pre-miRNA method is more useful for studying disease association.” [3] We also find that, in some cases, both strands cooperate to regulate the same or complementary pathways [4]. We acknowledge the reviewer’s point as an important consideration for future revision. We plan to consult or collaborate with biologists to enhance the quality of the manuscript in biology.

Reviewer #2 (Public review):

This study introduces a novel knowledge-driven approach, miRTarDS, which enables microRNA-Target Interaction (MTI) prediction by leveraging the disease association degree between a miRNA and its target gene. The core hypothesis is that this single feature is sufficient to distinguish experimentally validated functional MTIs from computationally predicted MTIs in a binary classification setting. To quantify the disease association, the authors fine-tuned a Sentence-BERT (SBERT) model to generate embeddings of disease descriptions and compute their semantic similarity. Using only this disease association feature, miRTarDS achieved an F1 score of 0.88 on the test set.

We thank the reviewers for their positive feedback, especially for their recognition of the novelty of this manuscript.

Strengths:

The primary strength is the innovative use of the disease association degree as an independent feature for MTI classification. In addition, this study successfully adapts and fine-tunes the Sentence-BERT (SBERT) model to quantify the semantic similarity between biomedical texts (disease descriptions). This approach establishes a critical pathway for integrating powerful language models and the vast growth in clinical/disease data into biochemical discovery, like MTI prediction.

We would like to thank the reviewer again for their positive feedback. We appreciate their recognition of the novelty of our work, as well as their acknowledgment that the proposed method paves the way for integrating language models with clinical/disease data into biochemical discovery.

Weaknesses:

The main weakness lies in its definition of the ground-truth dataset, which serves as a foundation for methodological evaluation. The study defines the Negative Set as computationally predicted MTIs that lack experimental evidence. However, the absence of experimental validation does not equate to non-functionality. Similarly, the miRAW sets are classified by whether the target and miRNA could form a stable duplex structure according to RNA structure prediction. This definition is biologically irrelevant, as duplex stability does not fully encapsulate the complex in vivo binding of miRNAs within the AGO protein complex.

We thank the reviewers for their constructive feedback. We have realized that treating predicted MTI as a negative class may pose some issues. Therefore, we have decided to adopt Positive Unlabeled (PU) Learning in subsequent updates. This classification method can be applied to datasets such as ours, which contain only positive classes and lack negative ones [1]. We used the miRAW dataset to enable a horizontal comparison of our method with traditional sequence-based prediction approaches. We acknowledge that miRAW may overlook some biological insights, and we plan to optimize the construction of test datasets in the future. Some preliminary explorations have already been conducted, and the relevant code is available on GitHub.

Furthermore, we will make the following revisions: 1) We will clearly specify the version of miRBase and incorporate more miRNA-related databases. 2) Conduct a further literature review on miRNA biological mechanisms to enhance the quality of the manuscript in biology. 3) Perform a more comprehensive evaluation of the model’s performance. 4) Attempt to identify some representative MTIs that have been overlooked by existing prediction tools but can be predicted by our proposed method.

References

(1) Li, F., Dong, S., Leier, A., Han, M., Guo, X., Xu, J., ... & Song, J. (2022). Positive-unlabeled learning in bioinformatics and computational biology: a brief review. Briefings in Bioinformatics, 23(1), bbab461.

(2) Huang, Z., Shi, J., Gao, Y., Cui, C., Zhang, S., Li, J., ... & Cui, Q. (2019). HMDD v3. 0: a database for experimentally supported human microRNA–disease associations. Nucleic acids research, 47(D1), D1013-D1017.

(3) Wang, H., & Ho, C. (2023). The human pre-miRNA distance distribution for exploring disease association. International Journal of Molecular Sciences, 24(2), 1009.

(4) Mitra, R., Adams, C. M., Jiang, W., Greenawalt, E., & Eischen, C. M. (2020). Pan-cancer analysis reveals cooperativity of both strands of microRNA that regulate tumorigenesis and patient survival. Nature Communications, 11(1), 968.

eLife Assessment

There is a perennial question in the field of birdsong: the contribution of the cerebellum to singing and processing song-related information. This study provides a valuable first step into this discussion, using electrophysiology of cerebellar neurons during a battery of assays, including singing and song playback. While the electrophysiological dataset here is novel and could shed light on key aspects of the neural control of vocal behavior, the evidence provided for the conclusions reached by the authors is currently incomplete.

Reviewer #1 (Public review):

In this study, Ursu, Centeno, and Leblois record from the cerebellum of zebra finches and analyze neurons for auditory and song-related activity. The paper covers a lot of ground, ranging from lesions of the deep nuclei to song and white noise playback inside and outside of singing, and some level of survey of response types across cerebellar lobules, to provide foundational information on cerebellar relationships with song. There are a number of interesting observations in the study, to me most notably, the lack of responsivity of song-related activity in lobule IV to distorted auditory feedback. This observation is interesting in light of the perennial idea that the cerebellum may participate in rapid error corrections in other somatic control domains. If such a role were relevant for song, it stands to reason that some alteration of activity could be found there. Of course, on the other hand, zebra finches do not show rapid corrections during DAF, so perhaps the null result does not resolve much. Nevertheless, these data are important steps forward in establishing the involvement or lack of involvement in a broader set of brain structures beyond the song control system typically studied. While the study presents some interesting and important inroads, in my opinion, there was a general lack of 'polish' to the study that led to ambiguity in the report and confusing displays. This detracted from rigorous reporting of the findings.

Reviewer #2 (Public review):

In this paper, the authors investigate the role of the cerebellum in song production in the zebra finch. First, they replicate prior studies to show that lesions of the lateral deep cerebellar nuclei (latDCN, primarily lobules IV-VII and IX) result in shorter duration syllables and song motifs than sham controls. The authors then record neural activity from the cerebellum during both passive auditory exposure in anesthetized birds and in freely singing animals. The authors claim that across multiple lobules, the cerebellum receives "non-selective" auditory inputs locked to syllable boundaries (based on acute recordings) and that cerebellar neurons display song-locked responses that are unaffected by auditory feedback perturbations (in chronic recordings). Moreover, the authors emphasized the distinct properties of lobule IV, which they argue is tightly locked to the onset and offset of syllables, and conclude that the cerebellum might contribute to the duration of song elements.

This paper presents novel and useful descriptions of song-related neural activity in the cerebellum. However, there are multiple serious issues. First, there are major issues with the design and presentation of the analysis of the electrophysiological data; based on these, it is unclear whether the authors are justified in some of their conclusions about neural tuning or are entitled to any of their claims about the specific tuning or function of neurons in particular lobules. Second, because the authors' conceptual framework seems to ignore possible non-auditory inputs to the cerebellum, their results on (minimal) effects of auditory manipulation during singing are over-interpreted with respect to providing evidence of a forward model. Third, the paper's central assertion - that the songbird cerebellum may contribute to the duration of vocal events during song - was firmly established by a prior lesion study (Radic et al., 2024). Although the authors do cite this prior study with respect to longer-term postlesion changes after cerebellar lesions, this paper also showed a large change in syllable duration immediately after cerebellar lesion (Figure 5 in Radic et al). The electrophysiological results in the present paper could provide valuable insights into the neural mechanisms underlying this already-described role of the songbird cerebellum; however, given the other concerns above, it is not clear that the authors have done so.

Author response:

Public Reviews:

Reviewer #1 (Public review):

In this study, Ursu, Centeno, and Leblois record from the cerebellum of zebra finches and analyze neurons for auditory and song-related activity. The paper covers a lot of ground, ranging from lesions of the deep nuclei to song and white noise playback inside and outside of singing, and some level of survey of response types across cerebellar lobules, to provide foundational information on cerebellar relationships with song. There are a number of interesting observations in the study, to me most notably, the lack of responsivity of song-related activity in lobule IV to distorted auditory feedback. This observation is interesting in light of the perennial idea that the cerebellum may participate in rapid error corrections in other somatic control domains. If such a role were relevant for song, it stands to reason that some alteration of activity could be found there. Of course, on the other hand, zebra finches do not show rapid corrections during DAF, so perhaps the null result does not resolve much. Nevertheless, these data are important steps forward in establishing the involvement or lack of involvement in a broader set of brain structures beyond the song control system typically studied. While the study presents some interesting and important inroads, in my opinion, there was a general lack of 'polish' to the study that led to ambiguity in the report and confusing displays. This detracted from rigorous reporting of the findings.

We thank reviewer #1 for his comments. We will clarify the possible misleading or ambiguous claims and interpretations in the present manuscript and polish the presentation of the results. We will also modify the discussion to better place or results within the current knowledge on cerebellum and songbirds, and in particular address the link between our findings and the low sensitivity to auditory feedback in zebra finches.

Reviewer #2 (Public review):

In this paper, the authors investigate the role of the cerebellum in song production in the zebra finch. First, they replicate prior studies to show that lesions of the lateral deep cerebellar nuclei (latDCN, primarily lobules IV-VII and IX) result in shorter duration syllables and song motifs than sham controls. The authors then record neural activity from the cerebellum during both passive auditory exposure in anesthetized birds and in freely singing animals. The authors claim that across multiple lobules, the cerebellum receives "non-selective" auditory inputs locked to syllable boundaries (based on acute recordings) and that cerebellar neurons display song-locked responses that are unaffected by auditory feedback perturbations (in chronic recordings). Moreover, the authors emphasized the distinct properties of lobule IV, which they argue is tightly locked to the onset and offset of syllables, and conclude that the cerebellum might contribute to the duration of song elements.

This paper presents novel and useful descriptions of song-related neural activity in the cerebellum. However, there are multiple serious issues. First, there are major issues with the design and presentation of the analysis of the electrophysiological data; based on these, it is unclear whether the authors are justified in some of their conclusions about neural tuning or are entitled to any of their claims about the specific tuning or function of neurons in particular lobules. Second, because the authors' conceptual framework seems to ignore possible non-auditory inputs to the cerebellum, their results on (minimal) effects of auditory manipulation during singing are over-interpreted with respect to providing evidence of a forward model. Third, the paper's central assertion - that the songbird cerebellum may contribute to the duration of vocal events during song - was firmly established by a prior lesion study (Radic et al., 2024). Although the authors do cite this prior study with respect to longer-term postlesion changes after cerebellar lesions, this paper also showed a large change in syllable duration immediately after cerebellar lesion (Figure 5 in Radic et al). The electrophysiological results in the present paper could provide valuable insights into the neural mechanisms underlying this already-described role of the songbird cerebellum; however, given the other concerns above, it is not clear that the authors have done so.

We thank reviewer #2 for these comments. We will improve the presentation of the results, in particular our cell-type classification of the electrophysiology recordings based on latest literature and  the statistics of the tuning differences between lobules. We will also modify the discussion regarding singing related internal models and consider non-auditory feedback. Finally, we will clarify the position of our work within the existing songbird literature and clarify what are the specific contributions of this work. We fully agree that prior studies have already shown the behavioural effects of lesions, as already clearly mentioned in introduction and discussion, and rather aimed at reproducing partially these results before diving into neural mechanisms. We will clarify this point in our revision.

eLife Assessment

This study makes a valuable contribution to our understanding of how the lipase regulator ABHD5 may control lipase activity through interactions with lipid droplets and cellular membranes. By combining multiscale molecular dynamics simulations with experimental approaches, the authors provide novel molecular insights into this membrane-protein interaction and present evidence suggesting that the regulatory mechanism depends on protein conformational changes and local membrane remodeling. While much of the evidence supporting the main conclusions is convincing, several aspects of the analysis, interpretation, and discussion remain incomplete. Overall, this work will be of interest to structural and molecular biologists working on lipid metabolism and membrane biophysics.

Reviewer #1 (Public review):

Summary:

In this study, the authors investigated the detailed structural mechanism of activation of ABHD5 upon interaction with lipid structures (bilayer and LD). The authors used an elaborate multiscale computational workflow, incorporating coarse-grained, all-atom, and enhanced-sampling molecular dynamics simulations, to propose a structural mechanism for the interaction and activation of ABHD5, as well as its specific interaction with TAG in LD. The authors then corroborated these observations with experimental studies involving hydrogen-deuterium exchange coupled with mass spectrometry of wild-type ABHD5 to assess the structural and conformational changes in ABHD5 upon binding, as well as mutagenesis with cell-based and in vitro assays monitoring membrane association, defining specific interactions that infer ABHD5 to localize LD.

Strengths:

The manuscript is well-written, and the data are reported in high-quality figures. The experimental design and data analysis are rigorous and support the conclusion. One major strength is the multiscale computational work that reveals a mechanism for the insertion of ABHD5 into lipid bilayers and LD involving the insertion of the N-term portion and the lid helix motif. The design of the computational workflow was very elaborate, and the undertaking was quite extensive, with multiple strategies to (GC, all-atom MD and GaMD). The authors then elegantly generate a hypothesis from these observations to experimentally corroborate the proposed mechanism. Particularly, the HDX-MS data support the engagement of the two regions upon binding, and the fluorescence microscopy data show the role of specific residues in localization/specificity to LD.

Weaknesses:

The following limitation is noted. Central to this manuscript is the model, as observed computationally, that initial lipid interaction by the N-term insertion is followed by the insertion of lid-helix in the membrane, which undergoes a conformational switch in the process. However, HDX-MS reveals that, in the unbound form, the lid helix region displays a bimodal isotopic envelope, revealing two species, one with low uptake, suggesting a structured species and one with high uptake, suggesting a less structured species. It is unclear from the manuscript whether the authors think the bimodality fits EX1 regime kinetics or not. Regardless, the model of unbound ABHD5 shows a lid-helix region devoid of secondary structure (Figure 5A), which is more consistent with the unprotected species. The authors also mention that previous modeling had pointed to the high flexibility of the insertion domain. Upon binding, the lid-helix region seems to be ordered from computational observations and loses bimodality by HDX-MS with a deuterium uptake consistent with the protected species of the bimodal envelop in the unbound form. The authors fall short of interpreting or even discussing what the bimodality of the lid-helix represents in the unbound form. What does the protected species in the bimodal envelope represent? Is it a transition representing lid-helix formation and unfolding? Does it imply that interaction and insertion into the lipid structures are governed by conformational selection? This issue should be at the very least acknowledged and discussed, or optimally investigated by performing more integrative studies of the HDX-MS data with the extensive computational data at hand, using existing protection factor calculations or HDX-guided ensemble refinement methods.

Reviewer #2 (Public review):

Summary:

The manuscript describes a combined computational and experimental approach to investigate the ABHD5 binding to and insertion into membranes.

Strengths:

Mutational experiments support computational findings obtained on ABHD5 membrane insertion with enhanced-sampling atomistic simulations.

Weaknesses:

While the addressed problem is interesting, I have several concerns, which fall into two categories:

(A) I see statements throughout the manuscript, e.g. on PNPLA activation, that are not supported by the results.

(B) The presentation of the computational and experimental results lacks in part clarity and detail.

Comments and questions on (A):

(1) I think the following statements in the abstract, which go beyond ABHD5 membrane binding, are not supported by the presented data:

the addition "to control lipolytic activation" in the 3rd sentence of the abstract.

further below ".... transforming ABHD5 into an active and membrane-localized regulator".

(2) The authors state in the Introduction (page numbers and line numbers are missing to be more specific):

"We hypothesize that binding of ABHD5 alters the nanoscale chemical and biophysical properties of the LD monolayer, which, combined with direct protein-protein interactions, enables PNPLA paralogs to access membrane-restricted substrates. This regulatory mechanism represents a paradigm shift from conventional enzyme-substrate interactions to sophisticated allosteric control systems that operate at membrane interfaces."

This hypothesis and the suggested paradigm shift are not supported by the data. Protein-protein interactions are not considered. What is meant by "sophisticated allosteric control"?

(3) The authors state in the Results section:

"We hypothesize that this TAG nanodomain is critical for ABHD5-activated TAG hydrolysis by PNPLA2." In previous pages, the authors state the location of the nanodomain: "TAG nanodomain under ABHD5".

If the nanodomain is located under ABHD5, how can it be accessible to PNPLA2? To my understanding, ABHD5 then sterically blocks access of PNPLA2 to the TAG nandomain.

(4) Another statement: "Our findings suggest that ABHD5-mediated membrane remodeling regulates lipolysis in part by regulating PNPLA2 access to its TAG substrate."

I don't see how the reported results support this statement (see point 3 above).

Comments and questions on (B):

(1) The authors state that the GaMD simulations started "from varying conformations observed during CGMD".

What is missing is a clear description of the CGMD simulation conformations, and the CG simulations as a whole, prior to the results section on GaMD. The authors use standard secondary and tertiary constraints in the Martini CG simulations. Do the authors observe some (constrained) conformational changes of ABHD5 already in the CG simulations (depending on the strength of the constraints)? Or do the conformational changes occur exclusively in the GaMD simulations? Both are fine, but this needs to be described.

(2) The authors write: "Three replicas of GaMD were performed."

Do these replicas lead to similar, or statistically identical, membrane-bound ABHD5 conformations? Is this information, i.e. a statistical analysis of differences in the replica runs, already included in the manuscript?

(3) The authors state on the hydrogen exchange results:

"HDX-MS provided orthogonal experimental evidence for the dynamics of the lid. In solution, a peptide (residues 200-226) spanning the lid helix displayed a bimodal isotopic distribution (Fig. S4), indicating the coexistence of different conformations. Upon LD binding, this distribution shifted to a single, low-exchange peak, demonstrating stabilization of the membrane-bound conformation with reduced solvent accessibility. These experimental observations corroborate our MD simulations."

I find this far too short to be understandable. Also, there are no computational results of ABHD5 in solution that show a bimodal conformational distribution of the lid helix, which is observed in the hydrogen exchange experiments. Which aspects of the MD simulations are corroborated?

eLife Assessment

This important study introduces NoSeMaze, a semi-naturalistic platform for continuous, high-dimensional tracking of social and cognitive behaviors in group-housed mice, and uses it to show that individual social rank is stable across changing social contexts. By integrating automated dominance measures, proactive social behaviors, and reinforcement-learning-based profiles, the authors demonstrate a novel framework for examining how stable individual differences shape social structure. The strength of evidence is solid, advancing our understanding that social hierarchy can be viewed as a trait-like dimension of individuality. Yet, the interpretation of dominance in this paradigm and its broader ecological meaning remains somewhat ambiguous. This work will be of broad relevance for behavioral neuroscience and social behavior research.

Reviewer #1 (Public review):

Summary:

The goal of the study was to address the question of the degree to which social position in a group is a stable trait that persists across conditions. Reinwald et al. use a custom-built cage system with automated tracking and continuous testing for social dominance that does not require intervention by the experimenter. Remixing of individuals from different groups revealed that social position was rather stable and not really predictable from other measures that were taken. The authors conclude that social position is multifaceted but dependent on characteristics like personality traits.

Strengths:

(1) Reductionistic, highly controlled setting that allows for the control of many confounding variables.

(2) Very interesting and important question.

(3) Confirms the emergence of inter-individual behavior-driven differences in inbred mice in a shared environment.

(4) Innovative paradigm and experimental setup.

(5) Fresh perspective on an old question that makes the best use of modern technology.

(6) Intelligent use of behavioral and cognitive covariables to generate a non-social context.

(7) Bold and almost provocative conclusion, inviting discussion and further elaboration.

Weaknesses:

(1) Reductionistic, highly controlled setting that blends out much of the complexity of social behavior in a community.

(2) The motivation to enter the test tube is not "trait" (or at least not solely a trait) but the basic need to reach food and water; chasing behavior would be less dependent on this stimulus.

(3) Dominance is only one aspect of sociality, social structure is reduced to rank. The information that might lie in the chasing behavior is not optimally used to explain social behavior beyond the rank measure.

(4) Focus on rank bears the risk of overgeneralization for readers not familiar with the context.

(5) Conclusion only valid for the reductionistic setting, in which environment, social and non-social changes only within narrow limits, and in which the mouse population does not face challenges

(6) Animals are not naive at the beginning of the experiment, but are already several weeks old.

In summary, this is a wonderful study, but not one that is easy to interpret. The bold conclusion is valid only within the constraints of the study, but nevertheless points in an important direction. The paradigm is clever and could be used for many interesting follow-ups.

To define social position as a personality trait will elicit strong opposition and much debate; the nuances of the paper might be lost on many readers and call for the (re)-consideration of many concepts that are touched. I find this attitude a strength of the paper, but the approach bears the risk of misunderstanding.

Reviewer #2 (Public review):

Summary:

This manuscript presents the "NoSeMaze", a novel automated platform for studying social behavior and cognitive performance in group-housed male mice. The authors report that mice form robust, transitive dominance hierarchies in this environment and that individual social rank remains largely stable across multiple group compositions. They further demonstrate that social dominance and aggressive behaviors, like chasing, are partially dissociable and that dominance traits are independent of non-social cognitive performance. The study includes a genetic manipulation of oxytocin receptor expression in the anterior olfactory nucleus, which showed only transient effects on social rank.

Strengths:

(1) Innovative Methodology:<br /> The NoSeMaze platform is a technically elegant and conceptually well-integrated system that enables fully automated, long-term monitoring of both social and cognitive behaviors in large groups of group-housed mice. It combines tube-test-like dominance contests, voluntary chase-escape interactions, and an embedded operant olfactory discrimination task within a single, ethologically relevant environment. This modular design allows for high-throughput, minimally invasive behavioral assessment without the need for repeated handling or artificial isolation.

(2) Experimental Scale and Rigor:<br /> The study includes 79 male mice and over 4,000 mouse-days of observation across multiple group reshufflings. The use of RFID-based identification, automated data logging, and longitudinal design enables robust quantification of individual trait stability and group-level social structure.

(3) Multidimensional Behavioral Profiling:<br /> The integration of social (tube dominance, proactive chasing), physical (body weight), and cognitive (olfactory learning task) measures offers a rich, multi-dimensional profile of each individual mouse. The authors' finding that social dominance traits and non-social cognitive performance are largely uncorrelated reinforces emerging models of orthogonal behavioral trait axes or "animal personalities".

(4) Clarity and Data Analysis:<br /> The analytical framework is well-suited to the study's complexity, with appropriate use of dominance metrics, mixed-effects models, and permutation tests. The analyses are clearly explained, statistically rigorous, and supported by transparent supplementary materials.

Weaknesses:

(1) Conceptual Novelty and Prior Work:<br /> While the study is carefully executed and methodologically innovative, several of its core findings reaffirm concepts already established in the literature. The emergence of stable, transitive social hierarchies, the persistence of individual differences in social behavior, and the presence of non-despotic social structures have all been previously reported in mice, including under semi-naturalistic conditions (e.g., Fan et al., 2019; Forkosh et al., 2019). Although this work extends those findings with greater behavioral resolution and scale, the manuscript would benefit from a clearer articulation of what is genuinely novel at the conceptual level, beyond the technological advance.

(2) Role of OXTR Deletion:<br /> The inclusion of the OXTR manipulation feels somewhat disconnected from the manuscript's central aims. The effects were minimal and transient, and the authors defer full interpretation to a separate study.

(3) Scope Limitations (Sex and Age):<br /> The study is limited to male mice, and although this is acknowledged, the title and overall framing imply broader generalizability. This sex-specific focus represents a common but problematic bias. Additionally, results from the older mouse cohort are under-discussed; if age had no effect, this should be explicitly stated.

(4) Ambiguity of Dominance as a Construct:<br /> While the study robustly quantifies social rank and hierarchy structure, the broader functional meaning of "dominance" remains unclear. As in prior work (e.g., Varholick et al., 2019), dominance rank here shows only weak associations with physical attributes (e.g., body weight), cognitive strategy, or neuromodulatory manipulation (OXTR deletion). This recurring pattern, where rank metrics are reliably established yet poorly predictive of other behavioral or biological traits, raises important questions about what such measures actually capture. In particular, it challenges the assumption that outcomes in paradigms like the tube test or chase frequency necessarily reflect dominance per se, rather than other constructs.

Reviewer #3 (Public review):

Reinwald et al. present the NoSeMaze, a semi-natural behavioral system designed to track social behaviors alongside reinforcement-learning in large groups of mice. Accumulating more than 4,000 days of behavioral monitoring, the authors demonstrate that social rank (determined by tube competitions) is a stable trait across shuffled cohorts and correlated with active chasing behaviors. The system also provides a solid platform for long-term measurements of reinforcement learning, including flexibility, response adaptation, and impulsiveness. Yet, the authors show that social ranking and chasing are mostly independent of these cognitive traits, and both seem mostly independent of oxytocin signaling in the AON.

Strengths:

(1) The neuroethological approach for automated tracking of several mice under semi-natural conditions is still rare in social behavioral research and should be encouraged.

(2) The assessment of dominance by two independent measures, i.e., spontaneous tube competitions and proactive chasing, is innovative and valuable.

(3) The integration of a long-term reinforcement-learning module into the semi-natural system provides novel opportunities to combine cognitive traits into social personality assessments.

(4) The open-source system provides a valuable resource for the scientific community.

Limitations:

(1) Apparent ambiguity and inconsistency in age structure and cohort participation across rounds, raising concerns about uncontrolled confounds.

(2) Chasing behavior appears more stable than tube-test competitions (Figure 4D vs. Figure 3D), which challenges the authors' decision to treat tube competitions as the primary basis for hierarchy determination.

Major concerns:

(1) Unclear and inconsistent handling of age groups and repeated sampling. The manuscript repeatedly refers to "younger" and "older" adults, but it is unclear whether age was ever controlled for or included in models. Some mice completed only one round, others 2-5 rounds, without explanation of the criteria or balancing.

(2) Stability of chasing appears stronger than the stability of tube competitions. Figure 4D shows highly consistent chasing behavior across weeks, while Figure 3D shows weaker and more variable correlations for tube-based David scores. This is also evident from Figure 5A-B,D. Thus, it appears that chasing, which serves to quantify dominance in similar semi-natural setups, may be a more reliable and behaviorally meaningful measure of dominance than the incidental tube competitions.

(3) Unbalanced participation across rounds compromises stability analyses. Stability analyses (e.g., ICCs, round-to-round correlations) assume comparable sampling across individuals. However, some mice contribute 1 round, others 2, 3, 4, and even 5 rounds. This imbalance may inflate stability estimates or confound group reshuffling effects, and the rationale for variable participation is not explained.

eLife Assessment

This valuable study highlights the challenge of identifying the role of immune imprinting in influenza immunity. The manuscript provides solid evidence that statistical support for imprinting depends heavily on model choice and can be found in the absence of imprinting due to age-related processes. However, the results are incomplete in that the impact of incorrectly modeling imprinting is not clear. The work will be of interest to researchers who study adaptive immunity in any system where imprinting may be observed.

Reviewer #1 (Public review):

Summary:

This manuscript uses serological data to quantify the effects of imprinting on subsequent influenza antibody responses. While this is an admirable goal, the HI dataset sounds impressive, and the authors developed a number of models, the manuscript came off as very dense and technical. One of the biggest pitfalls is that it is not easy to understand the lessons learned. The two Results section headers make clear statements - there was an imprinting signal in the HI titers, but much of this signal could also be seen in an imprinting-free simulation - and then the Discussion states a number of limitations. This is fine, but it leaves the reader wondering exactly how large an error would be introduced by ignoring imprinting effects altogether; alternatively, if imprinting is purposefully added, what would the expected effect size be? The comments below will provide some concrete steps to help clarify these points.

Major comments:

(1) Lines 107-133: The first Results section is a dense slog of information, and the reader is never given a good overview of what the imprinting coefficients exactly are. As the paper currently stands, if you do not start by reading the Methods, you will take away very little. I suggest adding a schematic for any of your models, showing what HI titers would be expected with/without imprinting effects. or age effects, or both, to tie in your modeling coefficients with quantities that all readers are familiar with.

(1.1) Clarify what the imprinting coefficient (y-axis in Figure 1A) looks like in this schematic.

(1.2) Another aspect that I missed: In addition to stating which models were best by BIC, what is the absolute effect size in the HI titers? During my initial reading, I had hoped that Figure 3 would answer this question, but it turned out to be just an overview of the dataset. I strongly suggest having such a figure to show the imprinting effect inferred by different models. What would the expected effect be if you kept someone's birth year constant but tuned their age? What if you kept their age at collection constant but tuned their birth year?

(1.3) It would also help to explain in your schematic what the x-axis labels (H1, H2, H1/H3) would look like in these scenarios, and what imprinting relative to H3 means.

(2) As mentioned above, it was hard to understand the takeaway messages, such as:

(2.1) A similar question would be: If you model antibody titers without imprinting, how far off would you be from the actual measurements (2x off, 4x off...)? If you add the imprinting effect, how much closer do you get?

(2.2) Are there specific age groups that require imprinting to be taken into consideration, since otherwise HI analyses will be markedly off?

(2.3) Are there age groups where imprinting can be safely ignored?

(3) HI titers against multiple H1 and H3 variants were measured, but it is unclear how these are used, and why titers against a single variant each season would not have worked equally well.

Reviewer #2 (Public review):

Summary:

In this study, the authors were testing the hypothesis that hemagglutination inhibition antibody titers, measured later in life, might be higher against influenza viruses that belong to the same hemagglutinin classification group as the influenza virus that a person was likely first exposed to early in life. This is one conceptualization of a phenomenon termed immune imprinting, which may explain previously observed differences in susceptibility to severe influenza infection between cohorts that were likely first exposed to different hemagglutinin groups. The results of the analysis provide some support for this analysis. However, support for the hypothesis is not consistently observed across sensitivity analyses, and a simulation study finds that antibody patterns consistent with immune imprinting may arise due to other factors in the absence of true imprinting effects. Therefore, overall support for the hypothesis is weak. Nonetheless, this study is important in that it provides guidance and has developed an analytic methodology for additional studies in this area of research. These findings and methods may also be useful for other infectious diseases for which patterns consistent with immune imprinting have been observed.

Strengths:

The strengths of this study include the relatively large cohort data source with broad age representation, rigorous statistical methods, and the use of sensitivity and simulation analyses to assess the robustness of the results.

Weaknesses:

The model outcome includes antibody titers measured against many different viruses, and the imprinting parameter was defined at the subtype level. This may obscure specific imprinting effects related to finer structural similarities between first and subsequent virus exposures. This analysis focuses only on one component of the immune response to influenza; immune imprinting may also involve other immune mechanisms. The analysis was carried out in a Chinese cohort, and vaccination status of the cohort is not discussed; the results may not be generalizable to other populations, particularly if vaccination patterns differ.

eLife Assessment

This study makes an important contribution by revealing how saccades selectively disrupt spatial working memory while sparing other object features, and by demonstrating how this mechanism is altered in aging and neurodegeneration. The findings are supported by convincing evidence derived from well-controlled eye-tracking experiments and systematic generative model comparisons. Together, the work provides a computationally grounded framework that is of importance for understanding trans-saccadic memory and its clinical relevance.

Reviewer #1 (Public review):

Summary:

This study employed a saccade-shifting sequential working memory paradigm, manipulating whether a saccade occurred after each memory array to directly compare retinotopic and transsaccadic working memory for both spatial location and color. Across four participant groups (young and older healthy adults, and patients with Parkinson's disease and Alzheimer's disease), the authors found a consistent saccade-related cost specifically for spatial memory - but not for color - regardless of differences in memory precision. Using computational modeling, they demonstrate that data from healthy participants are best explained by a complex saccade-based updating model that incorporates distractor interference. Applying this model to the patient groups further elucidates the sources of spatial memory deficits in PD and AD. The authors then extend the model to explain copying deficits in these patient groups, providing evidence for the ecological validity of the proposed saccade-updating retinotopic mechanism.

Strengths:

Overall, the manuscript is well written, and the experimental design is both novel and appropriate for addressing the authors' key research questions. I found the study to be particularly comprehensive: it first characterizes saccade-related costs in healthy young adults, then replicates these findings in healthy older adults, demonstrating how this "remapping" cost in spatial working memory is age-independent. After establishing and validating the best-fitting model using data from both healthy groups, the authors apply this model to clinical populations to identify potential mechanisms underlying their spatial memory impairments. The computational modeling results offer a clearer framework for interpreting ambiguities between allocentric and retinotopic spatial representations, providing valuable insight into how the brain represents and updates visual information across saccades. Moreover, the findings from the older adult and patient groups highlight factors that may contribute to spatial working memory deficits in aging and neurological disease, underscoring the broader translational significance of this work.

Comments on revisions:

The authors have addressed my earlier concerns.

Reviewer #2 (Public review):

Summary:

Zhao et al investigate how object location and colour are degraded across saccadic eye movements. They employ an eye-tracking task that requires participants to remember two sequentially presented items and subsequently report the colour and position of either one of these. Through counterbalancing of the presence or absence of saccades across items, the authors endeavour to dissect the impact of saccades independently on item location or colour. These behavioural findings form the basis of generative models designed to test competing, nested accounts of how stored information is stored and updated across saccades.

Strengths:

The combination of eye-tracking and generative modelling is a strength of the paper, which opens new perspectives into the impact of Alzheimer's and Parkinson's disease on the performance of visuospatial cognitive tests. The finding that the model parameters covary with clinical performance on the ROCF test is a nice example of a "computational assay" of disease.

Comments on revisions:

I thank the authors for their detailed responses and revisions arising from my feedback on the original manuscript. The revised manuscript adequately addresses all of my concerns.

Reviewer #3 (Public review):

Summary:

The manuscript introduces a visual paradigm aimed at studying tran-saccadic memory.

The authors observe how memory of object location is selectively impaired across eye movements, whereas object colour memory is relatively immune to intervening eye movements.<br /> Results are reported for young and elderly healthy controls, as well as PD and AD participants.

A computational model is introduced to account for these results, indicating how early differences in memory encoding and decay (but not tran-saccadic updating per se) can account for the observed differences between healthy controls and clinical groups.

In the revised manuscript, the authors have addressed most of my initial concerns. The dataset is generally compelling, as it includes healthy younger and older adults as well as clinical populations. In addition, the authors propose an interesting modelling approach designed to isolate and characterize the key components underlying the observed patterns of results.

It is important to acknowledge potential limitations of the modelling approach, particularly the differences in the number of parameters across the tested models. As models with more parameters typically achieve better fit, this issue warrants careful consideration. The authors have substantially addressed this point in their rebuttal.

Concerns regarding the specificity of the findings were also raised and have been adequately discussed in the authors' response. Specifically, they clarified the selective impact of saccade-related costs on spatial working memory updating across eye movements-without affecting feature‑based memory (e.g., color) -as well as the specificity of the updating effects observed with the Rey-Osterrieth Complex Figure.

Author response:

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

Reviewer #1 (Public review):

Thank you so much for your comprehensive and insightful assessment of our manuscript. We appreciate your recognition of the novelty of our experimental design and the utility of our computational framework for interpreting visual remapping across the lifespan and in clinical populations. We are very grateful for your suggestions regarding the narrative flow, which have helped us to improve the manuscript's focus and coherence. Our responses to your specific concerns are detailed below.

(1) Relevance of the figure-copy results (pp. 13-15). Is it necessary to include the figure-copy task results within the main text? The manuscript already presents a clear and coherent narrative without this section. The figure-copy task represents a substantial shift from the LOCUS paradigm to an entirely different task that does not measure the same construct. Moreover, the ROCF findings are not fully consistent with the LOCUS results, which introduces confusion and weakens the manuscript's coherence. While I understand the authors' intention to assess the ecological validity of their model, this section does not effectively strengthen the manuscript and may be better removed or placed in the Supplementary Materials.

We thank the reviewer  for their perspective regarding the narrative flow and the transition between the LOCUS paradigm and the ROCF results. However, we remain keen to retain these findings in the main text, as they provide critical ecological and clinical validation for the computational mechanisms identified in our study.

We think these results strengthen the manuscript for the following main reasons:

(1) The ROCF we used is a standard neuropsychological tool for identifying constructional apraxia. Our results bridge the gap between basic cognitive neuroscience and clinical application by demonstrating that specific remapping parameters—rather than general memory precision—predict real-world deficits in patients.

(2) The finding that our winning model explains approximately 62% of the variance in ROCF copy scores across all diagnostic groups further indicates that these parameters from the LOCUS task represent core computational phenotypes that underpin complex, real-life visuospatial construction (copying drawings).

(3) Previous research has often observed only a weak or indirect link between drawing ability and traditional working memory measures, such as digit span (Senese et al., 2020). This was previously attributed to “deictic” strategies—like frequent eye and hand movements—that minimise the need to hold large amounts of information in memory (Ballard et al., 1995; Cohen, 2005; Draschkow et al., 2021). While our study was not exclusively designed to catalogue all cognitive contributions to drawing, the findings provide significant and novel evidence indicating that transsaccadic integration is a critical driver of constructional (copying drawing) ability. By demonstrating this link, the results provide evidence to stimulate a new direction for future research, shifting the focus from general memory capacity toward the precision of spatial updating across eye movements.

In summary, by including the ROCF results in the main text, we provide evidence for a functional role for spatial remapping that extends beyond perceptual stability into the domain of complex visuomotor control. We have expanded on these points throughout the revised manuscript:

In the Introduction: p.2:

“The clinical relevance of these spatial mechanisms is underscored by significant disruptions to visuospatial processing and constructional apraxia—a deficit in copying and drawing figures—observed in neurodegenerative conditions such as Alzheimer's disease (AD) and Parkinson's disease (PD).[20,21] This raises a crucial question: do clinical impairments in complex visuomotor tasks stem from specific failures in transsaccadic remapping? If so, the computational parameters that define normal spatial updating should also provide a mechanistic account of these clinical deficits, differentiating them from general age-related decline.”

p.3: "Finally, by linking these mechanistic parameters to a standard clinical measure of constructional ability (the Rey-Osterrieth Complex Figure task), we demonstrate that transsaccadic updating represents a core computational phenotype underpinning real-world visuospatial construction in both health and neurodegeneration.

In the Results:

“To assess whether the mechanistic parameters derived from the LOCUS task represent core phenotypes of real-world visuospatial abilities, we also instructed all participants to complete the Rey-Osterrieth Complex Figure copy task (ROCF; Figure 7A) on an Android tablet using a digital pen (see examples in Figure 7B; all Copy data are available in the open dataset: https://osf.io/95ecp/). The ROCF is a gold-standard neuropsychological tool for identifying constructional apraxia.[29] Historically, drawing performance has shown only weak or indirect correlations with traditional working memory measures.[30] This disconnect has been attributed to active visual-sampling strategies—frequent eye movements that treat the environment as an external memory buffer, minimising the necessity of holding large volumes of information in internal working memory.[3–5]

We hypothesised that drawing accuracy is primarily constrained by the precision of spatial updating across frequent saccades rather than raw memory capacity. To evaluate the ecological validity of the identified saccade-updating mechanism, we modelled individual ROCF copy scores across all four groups using the estimated (maximum a posteriori) parameters from the winning “Dual (Saccade) + Interference” model (Model 7; Figure 8) as regressors in a Bayesian linear model. Prior to inclusion, each regressor was normalised by dividing by the square root of its variance.

This model successfully explained 61.99% of the variance in ROCF copy scores, indicating that these computational parameters are strong predictors of real-word constructional ability (Figure 8A). … This highlights the critical role of accurate remapping based on saccadic information; even if the core saccadic update mechanism is preserved across groups (as shown in previous analyses), the precision of this updating process is crucial for complex visuospatial tasks. Moreover, worse ROCF copy performance is associated particularly with higher initial angular encoding error. This indicates that imprecision in the initial registration of angular spatial information contributes to difficulties in accurately reproducing complex visual stimuli.”

In the Discussion:

“Importantly, our computational framework establishes a direct mechanistic link between trassaccadic updating and real-world constructional ability. Specifically, higher saccade and angular encoding errors contribute to poorer ROCF copy scores. By mapping these mechanistic estimates onto clinical scores, we found that the parameters derived from our winning model explain approximately 62% of the variance in constructional performance across groups. These findings suggest that the computational parameters identified in the LOCUS task represent core phenotypes of visuospatial ability, providing a mechanistic bridge between basic cognitive theory and clinical presentation.

This relationship provides novel insights into the cognitive processes underlying drawing, specifically highlighting the role of transsaccadic working memoty.ry. Previous research has primarily focused on the roles of fine motor control and eye-hand coordination in this skill.[4,50–55] This is partly because of consistent failure to find a strong relation between traditional memory measures and copying abili [4,31] For instance, common measures of working memory, such as digit span and Corsi block tasks, do not directly predict ROCF copying performance.[31,56] Furthermore, in patients with constructional apraxia, these memory performance measures often remain relatively preserved despite significant drawing impairments.[56–58] In the literature, this lack of association has often been attributed to “deictic” visual-sampling strategies, characterised by frequent eye movements that treat the environment as an external memory buffer, thereby minimising the need to maintain a detailed internal representation.[4,59] In a real-world copying task, the ROCF requires a high volume of saccades, making it uniquely sensitive to the precision of the dynamic remapping signals identified here. Recent eye-tracking evidence confirms that patients with AD exhibit significantly more saccades and longer fixations during figure copying compared to controls, potentially as a compensatory response to trassaccadic working memory constraints.[56] This high-frequency sampling—averaging between 150 and 260 saccades for AD patients compared to approximately 100 for healthy controls—renders the task highly dependent on the precision of dynamic remapping signals.[56] To ensure this relationship was not driven by a general "g-factor" or non-spatial memory impairment, we further investigated the role of broader cognitive performance using the ACE-III Memory subscale. We found that the relationship between transsaccadic working memory and ROCF performance remains highly significant, even after controlling for age, education, and ACE-III Memory subscore. This suggests that transsaccadic updating may represent a discrete computational phenotype required for visuomotor control, rather than a non-specific proxy for global cognitive decline.

In other words, even when visual information is readily available in the world, the act of copying depends critically on working memory across saccades. This reveals a fundamental computational trade-off: while active sampling strategies (characterised with frequent eye-hand movements) effectively reduce the load on capacity-limited working memory, they simultaneously increase the demand for precise spatial updating across eye movements. By treating the external world as an "outside" memory buffer, the brain minimises the volume of information it must hold internally, but it becomes entirely dependent on the reliability with which that information is remapped after each eye movement. This perspective aligns with, rather contradicts, the traditional view of active sampling, which posits that individuals adapt their gaze and memory strategies based on specific task demands.[3,60] Furthermore, this perspective provides a mechanistic framework for understanding constructional apraxia; in these clinical populations, the impairment may not lie in a reduced memory "span," but rather in the cumulative noise introduced by the constant spatial remapping required during the copying process.[58,61]

Beyond constructional ability, these findings suggest that the primary evolutionary utility of high-resolution spatial remapping lies in the service of action rather than perception. While spatial remapping is often invoked to explain perceptual stability,[11–13,15] the necessity of high-resolution transsaccadic memory for basic visual perception is debated.[13,62–64] A prevailing view suggests that detailed internal models are unnecessary for perception, given the continuous availability of visual information in the external world.[13,44] Our findings support an alternative perspective, aligning with the proposal that high-resolution transsaccadic memory primarily serves action rather than perception.[13] This is consistent with the need for precise localisation in eye-hand coordination tasks such as pointing or grasping.[65] Even when unaware of intrasaccadic target displacements, individuals rapidly adjust their reaching movements, suggesting direct access of the motor system to remapping signals.66 Further support comes from evidence that pointing to remembered locations is biased by changes in eye position,[67] and that remapping neurons reside within the dorsal “action” visual pathway, rather than the ventral “perception” visual pathway.[13,68,69] By demonstrating a strong link between transsaccadic working memory and drawing (a complex fine motor skill), our findings suggest that precise visual working memory across eye movements plays an important role in complex fine motor control.”

(2) Model fitting across age groups (p. 9).

It is unclear whether it is appropriate to fit healthy young and healthy elderly participants' data to the same model simultaneously. If the goal of the model fitting is to account for behavioral performance across all conditions, combining these groups may be problematic, as the groups differ significantly in overall performance despite showing similar remapping costs. This suggests that model performance might differ meaningfully between age groups. For example, in Figure 4A, participants 22-42 (presumably the elderly group) show the best fit for the Dual (Saccade) model, implying that the Interference component may contribute less to explaining elderly performance.

Furthermore, although the most complex model emerges as the best-fitting model, the manuscript should explain how model complexity is penalized or balanced in the model comparison procedure. Additionally, are Fixation Decay and Saccade Update necessarily alternative mechanisms? Could both contribute simultaneously to spatial memory representation? A model that includes both mechanisms-e.g., Dual (Fixation) + Dual (Saccade) + Interference-could be tested to determine whether it outperforms Model 7 to rule out the sole contribution of complexity.

We thank you for the opportunity to expand upon and clarify our modelling approach. Our decision to use a common generative model for both young and older adults was grounded in the empirical finding that there was no significant interaction between age group and saccade condition for either location or colour memory. While older adults demonstrated lower baseline precision, the specific "saccade cost" remained remarkably consistent across cohorts. This was the justification we proceeded on to use of a common model to assess quantitative differences in parameter estimates while maintaining a consistent mechanistic framework for comparison.

Moreover, our winning model nests simpler models as special cases, providing the flexibility to naturally accommodate groups where certain components—such as interference—might play a reduced role. This ultimately confirms that the mechanisms for age-related memory deficits in this task reflect more general decline rather than a qualitative failure of the saccadic remapping process.

This approach is further supported by the properties of the Bayesian model selection (BMS) procedure we used, which inherently penalises the inclusion of unnecessary parameters. Unlike maximum likelihood methods, BMS compares marginal likelihoods, representing the evidence for a model integrated over its entire parameter space. This follows the principle of Bayesian Occam’s Razor, where a model is only favoured if the improvement in fit justifies the additional parameter space; redundant parameters instead "dilute" the probability mass and lower the model evidence.

Consequently, we contend that a hybrid model combining fixation and saccade mechanisms is unnecessary, as we have already adjudicated between alternative mechanisms of equal complexity. Specifically, Model 6 (Dual Fixation + Interference) and Model 7 (Dual Saccade + Interference) possess an identical number of parameters. The fact that Model 7 emerged as the clear winner—providing substantial evidence against Model 6 with a Bayes Factor of 6.11—demonstrates that our model selection is driven by the specific mechanistic account of the data rather than a simple preference for complexity.

We have revised the Results and Discussion sections of the manuscript to state these points more explicitly for readers and have included references to established literature regarding the robustness of marginal likelihoods in guarding against overfitting.

In the Results,

“By fitting these models to the trial-by-trial response data from all healthy participants (N=42), we adjudicated between competing mechanisms to determine which best explained participant performance (Figure 4). We used random-effects Bayesian model selection to identify the most plausible generative model. This process relies on the marginal likelihood (model evidence), which inherently balances model fit against complexity—a principle often referred to as Occam’s razor.[25–27] The analysis yielded a strong result: the “Dual (Saccade) + Interference” model (Model 7 in Table 1) emerged as the winning model, providing substantial evidence against the next best alternative with a Bayes Factor of 6.11.”

In the Discussion:

“Our framework employs Variational Laplace, a method used to recover computational phenotypes in clinical populations like those with substance use disorders,[34,35] and the models we fit using this procedure feature time-dependent parameterisation of variance—conceptually similar to the widely-used Hierarchical Gaussian Filter.[36–39] Importantly, the risk of overfitting is mitigated by the Bayesian Model Selection framework; by utilising the marginal likelihood for model comparison, the procedure inherently penalises excessive model complexity and promotes generalisability.[25–27,40] This generalisability was further evidenced by the model's ability to predict performance on the independent ROCF task, confirming that these parameters represent robust mechanistic phenotypes rather than idiosyncratic fits to the initial dataset.”

Minor point: On p. 9, line 336, Figure 4A does not appear to include the red dashed vertical line that is mentioned as separating the age groups.

Thank you for pointing out this inconsistency. We apologise for the oversight; upon further review, we concluded that the red dashed vertical line was unnecessary for the clear presentation of the data. We have therefore removed the line from Figure 4A and deleted the corresponding sentence in the figure caption.

(3) Clarification of conceptual terminology.

Some conceptual distinctions are unclear. For example, the relationship between "retinal memory" and "transsaccadic memory," as well as between "allocentric map" and "retinotopic representation," is not fully explained. Are these constructs related or distinct? Additionally, the manuscript uses terms such as "allocentric map," "retinotopic representation," and "reference frame" interchangeably, which creates ambiguity. It would be helpful for the authors to clarify the relationships among these terms and apply them consistently.

Thank you for pointing this out. We have revised the manuscript to ensure that these terms are applied with greater precision and consistency. Our revisions standardise the terminology based on the following distinctions:

Reference frames: We distinguish between the eye-centred reference frame (coordinate systems that shift with gaze) and the world-centred reference frame (coordinate systems anchored to the environment).

Retinotopic representation vs. allocentric map: We clarify that retinotopic representations are encoded within an eye-centred reference frame and are updated with every ocular movement. Conversely, the allocentric map is anchored to stable environmental features, remaining invariant to the observer’s gaze direction or position.

Retinotopic memory vs. transsaccadic memory: We have removed the term "retinal memory" to avoid ambiguity. We now consistently use retinotopic memory to describe the persistence of visual information in eye-centred coordinates within a single fixation. In contrast, transsaccadic memory refers to the higher-level integration of visual information across saccades, which involves the active updating or remapping of representations to maintain stability.

To incorporate these clarifications, we have implemented the following changes:

In the Introduction, the second paragraph has been entirely rewritten to establish these definitions at the outset, providing a clearer theoretical framework for the study.

“Central to this enquiry is the nature of the coordinate system used for the brain's internal spatial representation. Does the brain maintain a single, world-centred (allocentric) map, or does it rely on a dynamic, eye-centred (retinotopic) representation?[11,13,15,16] In the latter system, retinotopic memory preserves spatial information within a fixation, whereas transsaccadic memory describes the active process of updating these representations across eye movements to achieve spatiotopic stability—the perception of a stable world despite eye movements.[11,16–18] If spatial stability is indeed reconstructed through such remapping, the mechanism remains unresolved: do we retain memories of absolute fixation locations, or do we reconstruct these positions from noisy memories of the intervening saccade vectors? We can test these hypotheses by analysing when and where memory errors occur. Assuming that memory precision declines over time,[19] the resulting error distributions should reveal the specific variables that are represented and updated across each saccade.”

In the Results, the opening section of the Results has been reorganised to align with this terminology. We have ensured that the hypotheses and behavioural data—specifically the definition of "saccade cost"—are introduced using this consistent conceptual vocabulary to improve the overall coherence of the narrative.

(4) Rationale for the selective disruption hypothesis (p. 4, lines 153-154). The authors hypothesize that "saccades would selectively disrupt location memory while leaving colour memory intact." Providing theoretical or empirical justification for this prediction would strengthen the argument.

We have revised the Results to state the hypothesis more explicitly and expanded the Discussion to provide a robust theoretical and empirical rationale:

In the Results,

“This design allowed us to isolate and quantify the unique impact of saccades on spatial memory, enabling us to test competing hypotheses regarding spatial representation. If spatial memory were solely underpinned by an allocentric mechanism, precision should remain comparable across all conditions as the representation would be world-centred and unaffected by eye movements. Thus, performance in the no-saccade condition should be comparable to the two-saccade condition. Conversely, if spatial memory relies on a retinotopic representation requiring active updating across eye movements, the two-saccade condition was anticipated to be the most challenging due to cumulative decay in the memory traces used for stimulus reconstruction after each saccade.[22] Critically, we hypothesised that this saccade cost would be specific to the spatial domain; while location requires active remapping via noisy oculomotor signals, non-spatial features like colour are not inherently tied to coordinate transformations and should therefore remain stable (see more in Discussion below).

Meanwhile, the no-saccade condition was expected to yield the most accurate localisation, relying solely on retinotopic information (retinotopic working memory). These predictions were confirmed in young healthy adults (N = 21, mean age = 24.1 years, ranged between 19 and 34). A repeated measures ANOVA revealed a significant main effect of saccades on location memory (F(2.2,43.9)=33.2, p<0.001, partial η²=0.62), indicating substantial impairment after eye movements (Figure 2A). In contrast, colour memory remained remarkably stable across all saccade conditions (Figure 2B; F(2.2, 44.7) = 0.68, p=0.53, partial η² =0.03).

This “saccade cost”—the loss of memory precision following an eye movement—indicates that spatial representations require active updating across saccades rather than being maintained in a static, world-centred reference frame.

Critically, our comparison between spatial and colour memory does not rely on the absolute magnitude of errors, which are measured in different units (degrees of visual angle vs. radians). Instead, we assessed the relative impact of the same saccadic demand on each feature within the same trial. While location recall showed a robust saccade cost, colour recall remained statistically unchanged. To ensure this null effect was not due to a lack of measurement sensitivity, we examined the recency effect; recall performance for the second item was predicted to be better than for the first stimulus in each condition.[23,24] As expected, colour memory for Item 2 was significantly more accurate than for Item 1 (F(1,20) = 6.52, p = 0.02, partial η² = 0.25), demonstrating that the task was sufficiently sensitive to detect standard working memory fluctuations despite the absence of a saccade-induced deficit.”

In the Discussion, we now write that on p.18:

“A clear finding was the specificity of the saccade cost to spatial features; it was not observed for non-spatial features like colour, even in neurodegenerative conditions. This discrepancy challenges notions of fixed visual working memory capacity unaffected by saccades.16,44–46 The differential impact on spatial versus non-spatial features in transsaccadic memory aligns with the established "what" and "where" pathways in visual processing.32,33 For objects to remain unified, object features must be bound to stable representations of location across saccades.19 One possibility is that remapping updates both features and location through a shared mechanism, predicting equal saccadic interference for both colour and location in the present study.

However, our findings suggest otherwise. One potential concern is whether this dissociation simply reflects the inherent spatial noise introduced by fixational eye movements (FEMs), such as microssacades and drifts.47 Because locations are stored in a retinotopic frame, fixational instability necessarily shifts retinal coordinates over time. However, the "saccade cost" here was defined as the error increase relative to a no-saccade baseline of equal duration; because both conditions are subject to the same fixational drift, any FEM-induced noise is effectively subtracted out. Thus, despite the ballistic and non-Gaussian nature of FEMs,48 they cannot account for the fact the saccade cost in the spatial memory, but total absence in the colour domain. Another possibility is that this dissociation reflects differences in baseline task difficulty or dynamic range. Yet, the presence of a robust recency effect in colour memory (Figure 2B) confirms that our paradigm was sensitive to memory-dependent variance and was not limited by floor or ceiling effects.

The fact that identical eye movements—executed simultaneously and with identical vectors—systematically degraded spatial precision while sparing colour suggests a feature-specific susceptibility to transsaccadic remapping. This supports the view that the computational process of updating an object’s location involves a vector-subtraction mechanism—incorporating noisy oculomotor commands (efference copies)—that introduces specific spatial variance. Because this remapping is a coordinate transformation, the resulting sensorimotor noise does not functionally propagate to non-spatial feature representations. Consequently, features like colour may be preserved or automatically remapped without the precision loss associated with spatial updating.11,49 Our paradigm thus provides a refined tool to investigate the architecture of transsaccadic working memory across distinct object features.”

(5) Relationship between saccade cost and individual memory performance (p. 4, last paragraph).

The authors report that larger saccades were associated with greater spatial memory disruption. It would be informative to examine whether individual differences in the magnitude of saccade cost correlate with participants' overall/baseline memory performance (e.g. their memory precision in the no-saccade condition). Such analyses might offer insights into how memory capacity/ability relates to resilience against saccade-induced updating.

We have now conducted the correlation analysis to determine whether baseline memory capacity (no-saccade condition) predicts resilience to saccade-induced updating. The results indicate that these two factors are independent.

To clarify the nature of the saccade-induced impairment, we have updated the text as follows:

p.4: “This “saccade cost”—the loss of memory precision following an eye movement—indicates that spatial representations require active updating across saccades rather than being maintained in a static, world-centred reference frame.”

p.5: “Further analysis examined whether individual differences in baseline memory precision (no-saccade condition) predicted resilience to saccadic disruption. Crucially, individual saccade costs (defined as the precision loss relative to baseline) did not correlate with baseline precision (rho = 0.20, p = 0.20). This suggests that the noise introduced by transsaccadic remapping acts as an independent, additive source of variance that is not modulated by an individual’s underlying memory capacity. These findings imply a functional dissociation between the mechanisms responsible for maintaining a representation and those involved in its coordinate transformation.”

(6) Model fitting for the healthy elderly group to reveal memory-deficit factors (pp. 11-12). The manuscript discusses model-based insights into components that contribute to spatial memory deficits in AD and PD, but does not discuss components that contribute to spatial memory deficits in the healthy elderly group. Given that the EC group also shows impairments in certain parameters, explaining and discussing these outcomes of the EC group could provide additional insights into age-related memory decline, which would strengthen the study's broader conclusions.

This is a very good point. We rewrote the corresponding results section (p.12-13):

“Modelling reveals the sources of spatial memory deficits in healthy aging and neurodegeneration - To understand the source of the observed deficits, we applied the winning ‘Dual (Saccade) + Interference’ model the data from all participants (YC, EC, AD, and PD). By fitting the model to the entire dataset, we obtained estimates of the parameters for each individual, which then formed the basis for our group-level analysis. To formally test for group differences, we used Parametric Empirical Bayes (PEB), a hierarchical Bayesian approach that compares parameter estimates across groups while accounting for the uncertainty of each estimate [28]. This allowed us to identify which specific cognitive mechanisms, as formalised by the model parameters, were affected by age and disease.

The Bayesian inversion used here allows us to quantify the posterior mode and variance for each parameter and the covariance for each parameter. From these, we can compute the probabilities that pairs of parameters differ from one another, which we report as P(A>B)—meaning the posterior probability that the parameter for group A was greater than that for group B.

We first examined the specific parameters differentiating healthy elderly (EC) from young controls (YC) to isolate the factors contributing to non-pathological, age-related decline. The analysis revealed that healthy ageing is primarily characterised by a significant increase in Radial Decay (P(EC > YC) = 0.995), a heightened susceptibility to Interference (P(EC > YC) = 1.000), and a reduction in initial Angular Encoding precision (P(YC < EC) = 0.002; Figure 6). These results suggest that normal ageing degrades the fidelity of the initial memory trace and its resilience over time, while the core computational process of updating information across saccades remains intact.

Beyond these baseline ageing effects, our clinical cohorts exhibited more severe and condition-dependent impairments. Radial decay showed a clear, graded impairment: AD patients had a greater decay rate than PD patients (P(AD > PD) = 1.000), who in turn were more impaired than the EC group (P(PD > EC) = 0.996). A similar graded pattern was observed for Interference, where AD patients were most susceptible (P(AD > PD) = 0.999), while the PD and EC groups did not significantly differ (P(PD > EC) = 0.532).

Patients with AD also showed a tendency towards greater angular decay than controls (P(AD > EC) = 0.772), although this fell below the 95% probability threshold. This effect was influenced by a lower decay rate in the PD group compared to the EC group (P(PD < EC) = 0.037). In contrast, group differences in encoding were less pronounced. While YC exhibited significantly higher precision than all other groups, AD patients showed significantly higher angular encoding error than PD patients (P(AD > PD) = 0.985), though neither group differed significantly from the EC group.

Crucially, parameters related to the saccade itself—saccade encoding and saccade decay—did not differentiate the groups. This indicates that neither healthy ageing nor the early stages of AD and PD significantly impair the fundamental machinery for transsaccadic remapping. Instead, the visuospatial deficits in these conditions arise from specific mechanistic failures: a faster decay of radial position information and increased susceptibility to interference, both of which are present in healthy ageing but significantly amplified by neurodegeneration.”

In the Discussion, we added:

“Although saccade updating was an essential component of the winning model, its two key parameters—initial encoding error and decay rate during maintenance—did not significantly differ across groups. This indicates that the core computational process of updating spatial information based on eye movements is largely preserved in healthy aging and neurodegeneration.

Instead, group differences were driven by deficits in angular encoding error (precision of initial angle from fixation), angular decay, radial decay (decay in memory of distance from fixation), and interference susceptibility. This implies a functional and neuroanatomical dissociation: while the ventral stream (the “what” pathway) shows an age-related decline in the quality and stability of stored representations, the dorsal-stream (the “where” pathway) parietal-frontal circuits responsible for coordinate transformations remain functionally robust.[31–34] These spatial updating mechanisms appear resilient to the normal ageing trajectory and only break down when challenged by the specific pathological processes seen in Alzheimer’s or Parkinson’s disease.”

(7) Presentation of saccade conditions in Figure 5 (p. 11). In Figure 5, it may be clearer to group the four saccade conditions together within each patient group. Since the main point is that saccadic interference on spatial memory remains robust across patient groups, grouping conditions by patient type rather than intermixing conditions would emphasize this interpretation.

There are several valid ways to present these plots, but we chose this format because it allows for a direct visual comparison of the post-hoc group differences within each specific task demand. This arrangement clearly illustrates the graded impairment from young controls through to patients with Alzheimer’s disease across every condition. This structure also directly mirrors our two-way ANOVA, which identified significant main effects for both Group and Condition, but crucially, no significant Group x Condition interaction. We felt that grouping the data by participant group would force readers to look across four separate clusters to compare the slopes, making the stability of the saccadic remapping mechanism much harder to grasp at a glance.

Reviewer #1 (Recommendations for the authors):

(1) Formatting of statistical parameters.

The formatting of statistical symbols should be consistent throughout the manuscript. Some instances of F, p, and t are italicized, while others are not. All statistical symbols should be italicized.

Thank you for pointing this out. We have audited the manuscript. While we have revised the text to address these instances throughout the Results and Methods sections, any remaining minor formatting inconsistencies will be corrected during the final typesetting stage.

(2) Minor typographical issues.

(a) Line 532: "are" should be "be."

(b) Line 654: "cantered" should be "centered."

(c) Line 213: In "(p(bonf) < 0.001, |t| {greater than or equal to} 5.94)," the t value should be reported with its degrees of freedom, and t should be reported before p. The same applies to line 215.

Thank you for your careful reading. All corrected.

Reviewer #2 (Public review):

We thank you for your positive feedback regarding our eye-tracking methodology and computational approach. We appreciate your critical insights into the feature-specific disruption hypothesis and the task structure. We have substantially revised the results and discussion about the saccadic interference on colour memory. Below we will answer your suggestions point-by-point:

Reviewer #2 (Recommendations for the authors):

(1) The study treats colour and location errors as comparable when arguing that saccades selectively disrupt spatial but not colour memory. However, these measures are defined in entirely different units (degrees of visual angle vs radians on a colour wheel) and are not psychophysically or statistically calibrated. Baseline task difficulty, noise level, or dynamic range do not appear to be calibrated or matched across features. As a result, the null effect of saccades on colour could reflect lower sensitivity or ceiling effects rather than implicit feature-specific robustness.

We agree that direct comparisons of absolute error magnitudes across different dimensions are not appropriate. Our argument for feature-specific disruption relies not on the scale of errors, but on the presence or absence of a saccade cost within identical trials. In our within-subject design, the same saccade vectors produced a systematic increase in location error while leaving colour error statistically unchanged. To address sensitivity, we observed that colour memory was sufficiently precise to show a significant recency effect (p = 0.02). To further quantify the evidence for the null effect, we performed Bayesian repeated measures ANOVAs, which yielded a BF10 = 0.22. This provides substantial evidence that saccades do not disrupt colour precision, regardless of baseline sensitivity.

We have substantially revised this in Results, Methods and Discussion:

In the Results:

“This design allowed us to isolate and quantify the unique impact of saccades on spatial memory, enabling us to test competing hypotheses regarding spatial representation. If spatial memory were solely underpinned by an allocentric mechanism, precision should remain comparable across all conditions as the representation would be world-centred and unaffected by eye movements. Thus, performance in the no-saccade condition should be comparable to the two-saccade condition. Conversely, if spatial memory relies on a retinotopic representation requiring active updating across eye movements, the two-saccade condition was anticipated to be the most challenging due to cumulative decay in the memory traces used for stimulus reconstruction after each saccade.[22] Critically, we hypothesised that this saccade cost would be specific to the spatial domain; while location requires active remapping via noisy oculomotor signals, non-spatial features like colour are not inherently tied to coordinate transformations and should therefore remain stable (see more in Discussion below).

Meanwhile, the no-saccade condition was expected to yield the most accurate localisation, relying solely on retinotopic information (retinotopic working memory). These predictions were confirmed in young healthy adults (N = 21, mean age = 24.1 years, ranged between 19 and 34). A repeated measures ANOVA revealed a significant main effect of saccades on location memory (F(2.2,43.9)=33.2, p<0.001, partial η²=0.62), indicating substantial impairment after eye movements (Figure 2A). In contrast, colour memory remained remarkably stable across all saccade conditions (Figure 2B; F(2.2, 44.7) = 0.68, p=0.53, partial η² =0.03).

This “saccade cost”—the loss of memory precision following an eye movement—indicates that spatial representations require active updating across saccades rather than being maintained in a static, world-centred reference frame.

Critically, our comparison between spatial and colour memory does not rely on the absolute magnitude of errors, which are measured in different units (degrees of visual angle vs. radians). Instead, we assessed the relative impact of the same saccadic demand on each feature within the same trial. While location recall showed a robust saccade cost, colour recall remained statistically unchanged. To ensure this null effect was not due to a lack of measurement sensitivity, we examined the recency effect; recall performance for the second item was predicted to be better than for the first stimulus in each condition.[23,24] As expected, colour memory for Item 2 was significantly more accurate than for Item 1 (F(1,20) = 6.52, p = 0.02, partial η² = 0.25), demonstrating that the task was sufficiently sensitive to detect standard working memory fluctuations despite the absence of a saccade-induced deficit.”

In the Methods, at the beginning of “Statistical Analysis”, we added

“Because location and colour recall involve different scales and units, all analyses were performed independently for each feature to avoid cross-dimensional magnitude comparisons.” (p25)

In the Discussion, we added:

“A potential concern is whether the observed dissociation between colour and location reflects differences in baseline task difficulty or dynamic range. Yet, the presence of a robust recency effect in colour memory (Figure 2B) confirms that our paradigm was sensitive to memory-dependent variance and was not limited by floor or ceiling effects.”

(2) Colour and then location are probed serially, without a counter-balanced order. This fixed response order could introduce a systematic bias because location recall is consistently subject to longer memory retention intervals and cognitive interference from the colour decision. The observed dissociation-saccades impair location but not colour, and may therefore reflect task structure rather than implicit feature-specific differences in trans-saccadic memory.

Thank you for the insightful observation regarding our fixed response order. We acknowledge that that a counterbalanced design is typically preferred to mitigate potential order effects. However, we chose this consistent sequence to ensure the task remained accessible for cognitively impaired patients (i.e., the Alzheimer’s disease (AD) and Parkinson’s disease (PD) cohorts). Conducting an eye-tracking memory task with cognitively impaired patients is challenging, as they may struggle with task engagement or forget complex instructions. During the design phase, we prioritised a consistent structure to reduce the cognitive load and task-switching demands that typically challenge these cohorts.

Critically, because the saccade cost is a relative measure calculated by comparing conditions with identical timings, any bias from the fixed order is present in both the baseline and saccade trials. The disruption we report is therefore a specific effect of eye movements that goes beyond the noise introduced by the retention interval or the preceding colour report.

We added the following text in the Methods – experimental procedure (p.22):

“Recall was performed in a fixed order, with colour reported before location. This sequence was primarily chosen to minimise cognitive load and task-switching demands for the two neurological patient cohorts, ensuring the paradigm remained accessible for individuals with AD and PD. While this order results in a slightly longer retention interval for location recall, the saccade cost was identified by comparing location error across experimental conditions with similar timings but varying saccadic demands.”

(3) Relatedly, because spatial representations are retinotopic, fixational eye movements (FEMs - microsaccades and drift) displace the retinal coordinates of encoded positions, increasing apparent spatial noise with time delays. Colour memory, however, is feature-based and unaffected by small retinal translations. Thus, any between-condition or between-group differences in FEMs could selectively inflate location error and the associated model parameters (encoding noise, decay, interference), while leaving colour error unchanged. Note that FEMs tend to be slightly ballistic [1,2], hence not well modelled with a Gaussian blur.

This is a very insightful point. We have now addressed this in detail within the discussion:

“However, our findings suggest otherwise. One potential concern is whether this dissociation simply reflects the inherent spatial noise introduced by fixational eye movements (FEMs), such as microssacades and drifts.[46] Because locations are stored in a retinotopic frame, fixational instability necessarily shifts retinal coordinates over time. However, the "saccade cost" here was defined as the error increase relative to a no-saccade baseline of equal duration; because both conditions are subject to the same fixational drift, any FEM-induced noise is effectively subtracted out. Thus, despite the ballistic and non-Gaussian nature of FEMs,n [47] they cannot account for the fact the saccade cost in the spatial memory, but total absence in the colour domain. Another possibility is that this dissociation reflects differences in baseline task difficulty or dynamic range. Yet, the presence of a robust recency effect in colour memory (Figure 2B) confirms that our paradigm was sensitive to memory-dependent variance and was not limited by floor or ceiling effects.”

(4) There is no in silico demonstration that the modelling framework can recover the true generating model from synthetic data or recover accurate parameters under realistic noise levels, which can be challenging in generative models with a hierarchical structure (as per [3], for example). Figure 8b shows that the parameters possess substantial posterior covariance, which raises concerns as to whether they can be reliably disambiguate.

Many thanks for this comment. We have added a simple recovery analysis as detailed below but are also keen to ensure we fully answer your question—which has more to do with empirical rather than simulated data—and make clear the rationale for this analysis in this instance.

We added this in Supplementary Materials:

“Model validation and recovery analysis

The following section provides a detailed technical assessment of the model inversion scheme, focusing on the discriminability of the model space and the identifiability of individual parameters.

Recovery analyses of this sort are typically used prior to collecting data to allow one to determine whether, in principle, the data are useful in disambiguating between hypotheses. In this sense, they have a role analogous to a classical power calculation. However, their utility is limited when used post-hoc when data have already been collected, as the question of whether the models can be disambiguated becomes one of whether non-trivial Bayes factors can be identified from those data.

The reason for including a recovery analysis here is not to identify whether the model inversion scheme identifies a ‘true’ model. The concept of ‘true generative models’ commits to a strong philosophical position which is at odds with the ‘all models are wrong, but some are useful’ perspective held by many in statistics, e.g., (So, 2017). Of note, one can always confound a model recovery scheme by generating the same data in a simple way, and in (one of an infinite number of) more complex ways. A good model inversion scheme will always recover the simple model and therefore would appear to select the ‘wrong’ model in a recovery analysis. However, it is still the best explanation for the data. For these reasons, we do not necessarily expect ‘good’ recoverability in all parameter ranges. This is further confounded by the relationship between the models we have proposed—e.g., an interference model with very low interference will look almost identical to a model with no interference. The important question here is whether they can be disambiguated with real data.

Instead, the value of a post-hoc recovery analysis here is to evaluate whether there was a sensible choice of model space—i.e., that it was not a priori guaranteed that a single model (and, specifically, the model we found to be the best explanation for the data) would explain the results of all others. To address this, for each model, we simulated 16 datasets, each of which relied upon parameters sampled from the model priors, which included examples of each of the experimental conditions. We then fit each of these datasets to each of the 7 models to construct the confusion matrix shown in the lower panel of Supplementary Figure 3, by accumulating evidence over each of the 16 participants generated according to each ‘true’ model (columns) for each of the possible explanatory models (rows). This shows that no one model, for the parameter ranges sampled here, explains all other datasets. Interestingly, our ‘winning’ model in the empirical analysis is not the best explanation for any of the datasets simulated (including its own). This is reassuring, in that it implies this model winning was not a foregone conclusion and is driven by the data—not just the choice of model space.”

Your point about the posterior covariance is well founded. As we describe in Supplementary Materials, this is an inherent feature of inverse problems (analogous to EEG source localisation). However, the fact that our posterior densities move significantly away from the prior expectations demonstrates that the data are indeed informative. By adopting a Bayesian framework, we are able to explicitly quantify this uncertainty rather than ignoring it, providing a more transparent account of parameter identifiability. We have added the following in the same section of Supplementary Materials:

“This problem is an inverse problem—inferring parameters from a non-linear model. We therefore expect a degree of posterior covariance between parameters and, consequently, that they cannot be disambiguated with complete certainty. While some degree of posterior covariance is inherent to inverse models—including established methods like EEG source localisation—the fact that many of the parameters are estimated with posterior densities that do not include their prior expectations implies the data are informative about these.

The advantage of the Bayesian approach we have adopted here is that we can explicitly quantify posterior covariance between these parameters, and therefore the degree to which they can be disambiguated. While the posterior covariance matrices from empirical data are the relevant measure here, we can better understand the behaviour of the model inversion scheme in relation to the specific models used using the model recovery analysis reported in Supplementary figure 3.

The middle panel of the figure is key, along with the correlation coefficients reported in the figure caption. Here, we see at least a weak positive correlation (in some cases much stronger) for almost all parameters and limited movement from prior expectations for those parameters that are less convincingly recovered. This reinforces that the ability of the scheme to recover parameters is best assessed in terms of the degree of movement of posterior from prior values following fitting to empirical data.”

(5) The authors employ Bayes factors (BFs) to disambiguate models, but BFs would also strengthen the claims that location, but not colour, is impacted by saccades. Despite colour being a circular variable, colour error is analysed using ANOVA on linearised differences (radians). The authors should also arguably use circular statistics, such as the von Mises distribution, for the analysis of colour.

Regarding the use of circular statistics, you are correct that such error distributions are not suitable for ANOVA, and it is better to use circular statistics. However, for the present dataset, we used the mean absolute angular error per condition (ranging from 0 to π radians), which represents the shortest distance on the colour wheel between the target and the response.

This approach effectively linearises the measure by removing the 2π wrap-around boundary. because the observed errors were relatively small and did not cluster near the π boundary—even in the patient cohorts (Figure 5B)—the "wrap-around" effect of circular space is negligible. Moreover, by analysing the mean error across trials for each condition, rather than trial-wise data, we invoke the Central Limit Theorem. This ensures that the distribution of these means is approximately normal, satisfying the fundamental assumptions of ANOVA. Due to these reasons, we adopted simpler linear models. We confirmed that the data did not violate the assumptions of linear statistics. In this low-noise regime, linear and circular models converge on the same conclusions. This has been revised in Methods:

“For colour memory, we calculated the absolute angular error, defined as the shortest distance on the colour wheel between the target and the reported colour (range 0 to π radians). For the primary statistical analyses, we utilised the mean absolute error per condition for each participant. By analysing these condition-wise means rather than trial-wise raw data, we invoke the Central Limit Theorem, which ensures that the sampling distribution of these means approximates normality. Because the absolute errors in this paradigm were relatively small and did not approach the π boundary (Figure 5B) even in the clinical cohorts, the data were treated as a continuous measure in our linear ANOVAs and regression models. Moreover, because location and colour recall involve different scales and units, all analyses were performed independently for each feature to avoid cross-dimensional magnitude comparisons.”

We have also now integrated Bayesian repeated measures ANOVA throughout the manuscript. The Results section for the young healthy adults now reads (p. 4):

“A repeated measures ANOVA revealed a significant main effect of saccades on location memory (F(3, 20) = 51.52, p < 0.001, partial η²=0.72), with Bayesian analysis providing decisive evidence for the inclusion of the saccade factor (BF_incl = 3.52 x 10^13, P(incl|data) = 1.00). In contrast, colour memory remained remarkably stable across all saccade conditions (F(3, 20) = 0.57, p = 0.64, partial η² =0.03). This null effect was supported by Bayesian analysis, which provided moderate evidence in favour of the null hypothesis (BF₀₁ = 8.46, P(excl|data) = 0.89), indicating that the data were more than eight times more likely under the null model than a model including saccade-related impairment.”

For elderly healthy adults:

“In contrast, colour memory remained unaffected by saccade demands (F(3, 20) = 0.57, p = 0.65, partial η² =0.03), again supported by the Bayesian analysis: BF₀₁ = 8.68, P(excl|data) = 0.90.”

For patient cohorts:

“Bayesian repeated measures ANOVAs further supported this dissociation, providing moderate evidence for the null hypothesis in the AD group (BF₀₁ = 3.35, P(excl|data) = 0.77) and weak evidence in the PD group (BF₀₁ = 2.23, P(excl|data) = 0.69). This indicates that even in populations with established neurodegeneration, the detrimental impact of eye movements is specific to the spatial domain.”

Related description is also updated in Methods – Statistical Analysis.

Minor:

(1) The modelling is described as computational but is arguably better characterised as a heuristic generative model at Marr's algorithmic level. It does not derive from normative computational principles or describe an implementation in neural circuits.

We appreciate your perspective on the classification of our model within Marr’s hierarchy. We agree that our framework is best characterised as an algorithmic-level generative model. Our objective was to identify the mechanistic principles governing transsaccadic updating rather than to provide a normative derivation or a specific circuit-level implementation.

To ensure readers do not over-interpret the term ‘computational’, we have added a clarifying statement in the Discussion acknowledging the algorithmic nature of the model. Interestingly, we note that a model predicated on this form of spatial diffusion implies a neural field representation with a spatial connectivity kernel whose limit approximates the second derivative of a Dirac delta function. While a formal neural field implementation is beyond the scope of the present work, our algorithmic results provide the necessary constraints for such future biophysical models.

p.20: “While we describe the present framework as 'computational', it is more precisely characterised as an algorithmic-level generative model within Marr’s hierarchy. Our focus was on defining the rules of spatial integration and the sources of eye-movement-induced noise, rather than deriving these processes from normative principles or defining their specific neural implementation.”

(2) I did not find a description of the recruitment and characterization of the AD and PD patients.

Apologies for this omission. We have now included a detailed description of participant recruitment and clinical characterisation in the Methods section and also updated Table 2:

“A total of 87 participants completed the study: 21 young healthy adults (YC), 21 older healthy adults (EC), 23 patients with Parkinson’s disease (PD), and 22 patients with Alzheimer’s disease (AD). Their demographic and clinical details are summarised in Table 2. Initially, 90 participants were recruited (22 YC, 21 EC, 25 PD, 22 AD); however, three individuals (1 YC and 2 PD) were excluded from all analyses due to technical issues during data acquisition.

All participants were recruited locally in Oxford, UK. None were professional artists, had a history of psychiatric illness, or were taking psychoactive medications (excluding standard dopamine replacement therapy for PD patients). Young participants were recruited via the University of Oxford Department of Experimental Psychology recruitment system. Older healthy volunteers (all >50 years of age) were recruited from the Oxford Dementia and Ageing Research (OxDARE) database.

Patients with PD were recruited from specialist clinics in Oxfordshire. All had a clinical diagnosis of idiopathic Parkinson's disease and no history of other major neurological or psychiatric conditions. While specific dosages of dopamine replacement therapy (e.g., levodopa equivalent doses) were not systematically recorded, all patients were tested while on their regular medication regimen ('ON' state).

Patients with PD were recruited from clinics in the Oxfordshire area. All had a clinical diagnosis of idiopathic Parkinson’s disease and no history of other major neurological or psychiatric illnesses. While all patients were tested in their regular medication ‘ON’ state, the specific pharmacological profiles—including the exact types of medication (e.g., levodopa, dopamine agonists, or combinations) and dosages—were not systematically recorded. The disease duration and PD severity were also un-recorded for this study.

Patients with AD were recruited from the Cognitive Disorders Clinic at the John Radcliffe Hospital, Oxford, UK. All AD participants presented with a progressive, multidomain, predominantly amnestic cognitive impairment. Clinical diagnoses were supported by structural MRI and FDG-PET imaging consistent with a clinical diagnosis of AD dementia (e.g., temporo-parietal atrophy and hypometabolism).69 All neuroimaging was reviewed independently by two senior neurologists (S.T. and M.H.).

Global cognitive function was assessed using the Addenbrooke’s Cognitive Examination-III (ACE-III).70 All healthy participants scored above the standard cut-off of 88, with the exception of one elderly participant who scored 85. In the PD group, two participants scored below the cut-off (85 and 79). In the AD group, six participants scored above 88; these individuals were included based on robust clinical and radiological evidence of AD pathology rather than their ACE-III score alone.”

(3) YA and OA patients appear to differ in gender distribution.

We acknowledge the difference in gender distribution between the young (71.4% female) and older adult (57.1% female) cohorts. However, we do not anticipate that gender influences the fundamental computational mechanisms of retinotopic maintenance or transsaccadic remapping. These processes represent low-level visuospatial functions for which there is no established evidence of gender-specific differences in precision or coordinate transformation. We have ensured that the gender distribution for each cohort is clearly listed in the demographics table (Table 2) for full transparency.

Thank you very much for very insightful feedback!

Reviewer #3 (Public review):

Thank you for the positive feedback regarding our inclusion of clinical groups and the identification of computational phenotypes that differentiate these cohorts.

To address your concerns about the model, we have clarified our use of Bayesian Model Selection, which inherently penalises model complexity to ensure that our results are not driven solely by the number of parameters. We will also provide further evidence regarding model generalisability to address the concern of overfitting.

Regarding the link with the ROCF, we have revised the manuscript to better highlight the specific relationship between our transsaccadic parameters and the ROCF data and better motivate the inclusion of these results in the main text.

Below is our response to your suggestions point-by-point:

(1) The models tested differ in terms of the number of parameters. In general, a larger number of parameters leads to a better goodness of fit. It is not clear how the difference in the number of parameters between the models was taken into account. It is not clear whether the modelling results could be influenced by overfitting (it is not clear how well the model can generalize to new observations).

To ensure our results were not driven by the number of parameters, we utilised random-effects Bayesian Model Selection (BMS) to adjudicate between our candidate models. Unlike maximum likelihood methods, BMS relies on the marginal likelihood (model evidence), which inherently balances model fit against parsimony—a principle known as the Occam’s Razor (Rasmussen and Ghahramani, 2000). In this framework, a model is only preferred if the improvement in fit justifies the additional parameter space; redundant parameters actually lower model evidence by diluting the probability mass. We would be happy to point toward literature that discusses how these marginal likelihood approximations provide a more robust guard against overfitting than standard metrics like BIC or AIC (MacKay, 2003; Murray and Ghahramani, 2005; Penny, 2012).

The fact that the "Dual (Saccade) + Interference" model (Model 7) emerged as the winner—with a Bayes Factor of 6.11 against the next best alternative—demonstrates that its complexity was statistically justified by its superior account of the trial-by-trial data.

Furthermore, to address the risk of overfitting, we established the generalisability of these parameters by using them to predict performance on an independent clinical task. These parameters successfully explained ~62% of the variance in ROCF copy scores—a very distinct, real-world task--confirming that they represent robust computational phenotypes rather than idiosyncratic fits to the initial dataset.

In the Results (p10):

“We used random-effects Bayesian model selection to identify the most plausible generative model. This process relies on the marginal likelihood (model evidence), which inherently balances model fit against complexity—a principle often referred to as Occam’s razor.[25–27]”

In the Discussion (p17):

“Importantly, the risk of overfitting is mitigated by the Bayesian Model Selection framework; by utilising the marginal likelihood for model comparison, the procedure inherently penalises excessive model complexity and promotes generalisability.[25–27,42] This generalisability was further evidenced by the model's ability to predict performance on the independent ROCF task, confirming that these parameters represent robust mechanistic phenotypes rather than idiosyncratic fits to the initial dataset.”

(2) Results specificity: it is not clear how specific the modelling results are with respect to constructional ability (measured via the Rey-Osterrieth Complex Figure test). As with any cognitive test, performance can also be influenced by general, non-specific abilities that contribute broadly to test success.

We agree that constructional performance is influenced by both specific mechanistic constraints and general cognitive abilities. To isolate the unique contribution of transsaccadic updating, we therefore performed a partial correlation analysis across the entire sample. We examined the relationship between location error in the two-saccades condition (our primary behavioural measure of transsaccadic memory) and ROCF copy scores. Even after partialling out the effects of global cognitive status (ACE-III total score), age, and years of education, the correlation remained highly significant (rho = -0.39, p < 0.001).

This suggests that our model captures a specific computational phenotype—the precision of spatial updating during active visual sampling—rather than acting as a proxy for non-specific cognitive decline. This mechanistic link explains why traditional working memory measures (e.g., digit span or Corsi blocks) frequently fail to predict drawing performance; unlike those tasks, figure copying requires thousands of saccades, making it uniquely sensitive to the precision of the dynamic remapping signals identified by our modelling framework.

We added the following text in the Discussion (p19):

“We also found that the relationship between transsaccadic working memory and ROCF performance remains highly significant (rho = -0.39, p < 0.001), even after controlling for age, education, and global cognitive status (ACE-III total score). Consequently, transsaccadic updating may represent a discrete computational phenotype required for visuomotor control, rather than a non-specific proxy for global cognitive decline.[57]”

Reviewer #3 (Recommendations for the authors):

(1) The authors mention in the introduction the following: "One key hypothesis is that we use working memory across visual fixations to update perception dynamically", citing the following manuscript:

Harrison, W. J., Stead, I., Wallis, T. S. A., Bex, P. J. & Mattingley, J. B. A computational 906 account of transsaccadic attentional allocation based on visual gain fields. Proc. Natl. 907 Acad. Sci. U.S.A. 121, e2316608121 (2024).

However, the manuscript above does not refer explicitly to the involvement of working memory in transaccadic integration of object location in space. Rather, it takes advantage of recent evidence showing how the true location of a visual object is represented in the activity of neurons in primary visual cortex ( A. P. Morris, B. Krekelberg, A stable visual world in primate primary visual cortex. Curr. Biol. 29, 1471-1480.e6 (2019) ). The model hypothesizes that true locations of objects are readily available, and then allocates attention in real-world coordinates, allowing efficient coordination of attention and saccadic eye movements.

Thank you for clarification. As suggested, we have now included the citation of Morris & Krekelberg (2019) to acknowledge the evidence for stable object locations within the primary visual cortex.

(2) The authors in the introduction and the title use the terms 'transaccadic memory' and 'spatial working memory'. However, it is not clear whether these can be used interchangeably or are reflecting different constructs.

Classical measures of visuo-spatial working memory are derived from the Corsi task (or similar), where the location of multiple objects is displayed and subsequently remembered. In such tasks, eye movements and saccades are not generally considered, only memory performance, representing the visuo-spatial span.

Transaccadic memory tasks are instead explicitly measuring the performance on remembered object locations of features across explicit eye movements, usually using a very limited number of objects (1 or 2, as is the case for the current manuscript).

While the two constructs share some features, it is not clear whether they represent the same underlying ability or not, especially because in transaccadic tasks, participants are required to perform one or more saccades, thus representing a dual-task case.

I think the relationship between 'transaccadic memory' and 'spatial working memory' should be clarified in the manuscript.

Thank you. Yes, we have added this within the Methods - Measurement of saccade cost to clarify that spatial working memory is the broad cognitive construct responsible for short-term maintenance, whereas transsaccadic memory is the specific, dynamic process of remapping representations to maintain stability across eye movements.

In Methods (p.22):

“Within this framework, it is important to distinguish between the broad construct of spatial working memory and the specific process of transsaccadic memory. While spatial working memory refers to the general ability to maintain spatial information over short intervals, transsaccadic memory describes the dynamic updating of these representations—termed remapping—to ensure stability across eye movements. Unlike classical 'static' measures of spatial working memory, such as the Corsi block task which focuses on memory span, transsaccadic memory tasks explicitly require the integration of stored visual information with motor signals from intervening saccades. Our paradigm treats transsaccadic updating as a core computational process within spatial working memory, where eye-centred representations are actively reconstructed based on noisy memories of the intervening saccade vectors.”

(3) In Figure 1, the second row indicates the presentation of item 2. Indeed, in the condition 'saccade-after-item-1', the target in the second row of Figure 1 is displaced, as expected. This clarifies the direction and amplitude of the first saccade requested. However, from Figure 1, it is hard to understand the amplitude and direction of the second requested saccade. I think the figure should be updated, giving a full description of the direction and amplitude of the second saccade as well ('saccade-after-item-2' and 'two-saccades' conditions).

We agree that making the figure legend more self-contained is beneficial for the reader. While the specific physical parameters and the trial sequence for each condition are detailed in the Results and Methods sections, we have now updated the legend for Figure 1 to explicitly define these details. Specifically, we have clarified that the colour wheel itself served as the target for the second instructed saccade (i.e., the movement from the second fixation cross to the colour wheel location). We have also included the quantitative constraint that all saccade vectors were at least 8.5 degrees of visual angle in amplitude. Given the limited space within a figure legend, we hope these concise additions provide the transparency requested without interrupting the conceptual flow of the diagram.

Updated Figure 1 legend:

“Participants were asked to fixate a white cross, wherever it appeared. They had to remember the colour and location of a sequence of two briefly presented coloured squares (Item 1 and 2), each appearing within a white square frame. They then fixated a colour wheel wherever it appeared on the screen, which served as the target for the second instructed saccade (i.e., a movement from the second fixation cross to the colour wheel location). This cued recall of a specific square (Item 1 or Item 2 labelled within the colour wheel). Participants selected the remembered colour on the colour wheel which led to a square of that colour appearing on the screen. They then dragged this square to its remembered location on the screen. Saccadic demands were manipulated by varying the locations of the second frame and the colour wheel, resulting in four conditions in their reliance on retinotopic versus transsaccadic memory: (1) No-Saccade condition providing a baseline measure of within-fixation precision as no eye movements were required. (2) Saccade After Item 1; (3) Saccade After Item 2; (4) Saccades after both items (Two Saccades condition). In all conditions requiring eye movements, saccade vectors were constrained to a minimum amplitude of 8.5° (degrees of visual angle). While the No-Saccade condition isolates retinotopic working memory, conditions (2) to (4) collectively quantify the impact of varying saccadic demands and timings on the maintenance of spatial information, thereby assessing the efficacy of the transsaccadic updating process.”

(4) The authors write: "Eye tracking analysis confirmed high compliance: participants correctly maintained fixation or executed saccades as instructed on the vast majority of trials (83% {plus minus} 14%). Non-compliant trials were excluded 136 from further analysis." 14% of excluded trials are a substantial fraction of trials, given the task requirements. Is this proportion of excluded trials different between experimental groups, and are experimental groups contributing equally to this proportion?

We thank the reviewer for pointing this out, and we apologise for the confusion. The 83% trial number was actually across all four cohorts, and all conditions, and it was actually above 90% for YC, EC and even AD, but dropped to 60 ish in PD group.

We now have conducted a full analysis of compliant trial counts using a mixed ANOVA (4 saccade conditions x 4 cohorts). This analysis revealed a main effect of group (F(3, 80) = 8.06, p < 0.001), which was driven by lower compliance in the PD cohort (mean approx. 25.4 trials per condition) compared to the AD, EC, and YC cohorts (means ranging from 35.8 to 38.9 trials per condition). Crucially, however, the interaction between group and condition was not statistically significant (p = 0.151). This indicates that the relative impact of saccade demands on trial retention was consistent across all four groups.

Because our primary behavioural measure—the saccade cost—is a within-subject comparison of impairment across conditions, these differences in absolute trial numbers do not introduce a systematic bias into our findings. Furthermore, even with the higher attrition in the PD group, we retained a sufficient number of high-quality trials (minimum mean of ~23 trials in the most demanding condition) to support robust trial-by-trial parameter estimation and valid statistical inference. We have updated the Results and Methods to reflect these details.

In Results (p4):

“To mitigate potential confounds, we monitored eye position throughout the experiment. Eye-tracking analysis confirmed high compliance in healthy adults, who followed instructions on the vast majority of trials (Younger Adults: 97.2 ± 5.2 %; Older Adults: 91.3 ± 20.4 %). The mean difference between these groups was negligible, representing just 1.25 trials per condition, and was not statistically significant (t(80) = 0.16, p = 1.000; see more in Methods – Eyetracking data analysis). Non-compliant trials were excluded from all further analyses.”

In Methods (p27):

“Eye-tracking analysis confirmed high compliance overall, with participants correctly maintaining fixation or executing saccades on the vast majority of trials (83% across all participants). A mixed ANOVA revealed a main effect of group on trial retention (F(3, 80) = 8.06, p < 0.001, partial η² = 0.23), primarily due to lower compliance in the PD cohort (YC: 97±4%; EC: 91±10%; AD: 95±5%; PD: 63±38%). Importantly, there was no significant interaction between group and saccade condition (F(3.36, 80) = 1.78, p = 0.15, partial η² = 0.008), suggesting that trial attrition was not disproportionately affected by specific task demands in any group.

We acknowledge that this reduced trial count in the PD group represents a limitation for across-cohort comparison. However, the absolute number of compliant trials in PD group (mean approx. 25 per condition) remained sufficient for robust trial-by-trial parameter estimation. Furthermore, the lack of a significant group-by-condition interaction confirms that the results reported for this cohort remain valid and that our primary finding of a selective spatial memory deficit is robust to these differences in data retention.”

(5) Modelling

(a) Degrees of freedom, cross-validation, number of parameters.

I appreciate the effort in introducing and testing different models. Models of increase in complexity and are based on different assumptions about the main drivers and mechanisms underlying the dependent variable. The models differ in the number of parameters. How are the differences in the number of parameters between models taken into account in the modelling analysis? Is there a cost associated with the extra parameters included in the more complex models?

(b) Cross-validation and overfitting.

Overfitting can occur when a model learns the training data but cannot generalize to novel datasets. Cross-validation is one approach that can be used to avoid overfitting. Was cross-validation (or other approaches) implemented in the fitting procedure against overfitting? Otherwise, the inference that can be derived from the modelled parameters can be limited.

To address your concerns regarding model complexity and overfitting, we would like to clarify our use of Bayesian Model Selection (BMS). Unlike frequentist methods that often rely on cross-validation to assess generalisability, we used random-effects BMS based on the marginal likelihood (model evidence). This approach inherently implements Bayesian Occam’s Razor by integrating out the parameters. Under this framework, the use of the marginal likelihood for model selection provides a mathematically equivalent safeguard to frequentist cross-validation, as it evaluates the model's ability to generalise across the entire parameter space rather than just finding a maximum likelihood fit for the training data. Thus, models are penalised not just for the absolute number of parameters, but for their overall functional flexibility. A more complex model is only preferred if the improvement in model fit is substantial enough to outweigh this inherent penalty. The emergence of Model 7 as the winner (Bayes Factor = 6.11 against the next best alternative) confirms that its additional complexity is statistically justified.

Furthermore, in this study we provided an external validation of these recovered parameters by demonstrating that they explain 62% of the variance in an independent, real-world, clinical task (ROCF copy). This empirical evidence confirms that our model captures robust mechanistic phenotypes rather than idiosyncratic noise. We have updated the Results and Discussion to explicitly state these.

In Results: (p10)

“We used random-effects Bayesian model selection to identify the most plausible generative model. This process relies on the marginal likelihood (model evidence), which inherently balances model fit against complexity—a principle often referred to as Occam’s razor.[26–28]”

In Discussion: (p17)

“Importantly, the risk of overfitting is mitigated by the Bayesian Model Selection framework; by utilising the marginal likelihood for model comparison, the procedure inherently penalises excessive model complexity and promotes generalisability.[26–28,43] This generalisability was further evidenced by the model's ability to predict performance on the independent ROCF task, confirming that these parameters represent robust mechanistic phenotypes rather than idiosyncratic fits to the initial dataset.”

(6) n. of participants.

(a) The authors write the following: "A total of healthy volunteers (21 young adults, mean age = 24.1 years; 21 older adults, mean age = 72.4 years) participated in this study. Their demographics are shown in Table 1. All participants were recruited locally in Oxford." However, Table 1 reports the data from more than 80 participants, divided into 4 groups. Details about the PD and AD groups are missing. Please clarify.

We apologize for this lack of clarity in the text. We have rewrote and expand the “Participants” section and corrected Table 2 in the Methods section to reflect the correct number of participants.

In Methods (p20):

“A total of 87 participants completed the study: 21 young healthy adults (YC), 21 older healthy adults (EC), 23 patients with Parkinson’s disease (PD), and 22 patients with Alzheimer’s disease (AD). Their demographic and clinical details are summarised in Table 2. Initially, 90 participants were recruited (22 YC, 21 EC, 25 PD, 22 AD); however, three individuals (1 YC and 2 PD) were excluded from all analyses due to technical issues during data acquisition.

All participants were recruited locally in Oxford, UK. None were professional artists, had a history of psychiatric illness, or were taking psychoactive medications (excluding standard dopamine replacement therapy for PD patients). Young participants were recruited via the University of Oxford Department of Experimental Psychology recruitment system. Older healthy volunteers (all >50 years of age) were recruited from the Oxford Dementia and Ageing Research (OxDARE) database.

Patients with PD were recruited from specialist clinics in Oxfordshire. All had a clinical diagnosis of idiopathic Parkinson's disease and no history of other major neurological or psychiatric conditions. While specific dosages of dopamine replacement therapy (e.g., levodopa equivalent doses) were not systematically recorded, all patients were tested while on their regular medication regimen ('ON' state).

Patients with PD were recruited from clinics in the Oxfordshire area. All had a clinical diagnosis of idiopathic Parkinson’s disease and no history of other major neurological or psychiatric illnesses. While all patients were tested in their regular medication ‘ON’ state, the specific pharmacological profiles—including the exact types of medication (e.g., levodopa, dopamine agonists, or combinations) and dosages—were not systematically recorded. The disease duration and PD severity were also un-recorded for this study.

Patients with AD were recruited from the Cognitive Disorders Clinic at the John Radcliffe Hospital, Oxford, UK. All AD participants presented with a progressive, multidomain, predominantly amnestic cognitive impairment. Clinical diagnoses were supported by structural MRI and FDG-PET imaging consistent with a clinical diagnosis of AD dementia (e.g., temporo-parietal atrophy and hypometabolism).[70] All neuroimaging was reviewed independently by two senior neurologists (S.T. and M.H.).

Global cognitive function was assessed using the Addenbrooke’s Cognitive Examination-III (ACE-III).[71] All healthy participants scored above the standard cut-off of 88, with the exception of one elderly participant who scored 85. In the PD group, two participants scored below the cut-off (85 and 79). In the AD group, six participants scored above 88; these individuals were included based on robust clinical and radiological evidence of AD pathology rather than their ACE-III score alone.”

(b) As modelling results rely heavily on the quality of eye movements and eye traces, I believe it is necessary to report details about eye movement calibration quality and eye traces quality for the 4 experimental groups, as noisier data could be expected from naïve and possibly older participants, especially in case of clinical conditions. Potential differences in quality between groups should be discussed in light of the results obtained and whether these could contribute to the observed patterns.

Thank you for pointing this out. We have revised the Methods about how calibration was done:

(p27) “Prior to the experiment, a standard nine-point calibration and validation procedure was performed. Participants were instructed to fixate a small black circle with a white centre (0.5 degrees) as it appeared sequentially at nine points forming a 3 x 3 grid across the screen. Calibration was accepted only if the mean validation error was below 0.5 degrees and the maximum error at any single point was below 1.0 degree. If these criteria were not met, or if the experimenter noticed significant gaze drift between blocks, the calibration procedure was repeated. This calibration ensured high spatial accuracy across the entire display area, facilitating the precise monitoring of fixations on item frames and saccadic movements to the response colour wheel.”

Moreover, as detailed in our response to Point 4, while the PD group exhibited lower compliance, there was no interaction between group and saccade condition for compliance (p = 0.151). This confirms that any noise or trial attrition was distributed evenly across experimental conditions. Consequently, the observed "saccade cost" (the difference in error between conditions) is not an artefact of unequal noise but represents a genuine mechanistic impairment in spatial updating. We have updated the Methods to clarify this distinction.

Furthermore, our Bayesian framework explicitly estimates precision (random noise) as a distinct parameter from updating cost (saccade cost). This allows the model to partition the variance: even if a clinical group is "noisier" overall, this is captured by the precision parameter, ensuring it does not inflate the specific estimate of saccade-driven memory impairment.

(7) Figure 5. I suggest reporting these results using boxplots instead of barplots, as the former gives a better overview of the distributions.

We appreciate the suggestion to use boxplots to better illustrate data distributions. However, we have chosen to retain the current bar plot format due to the visual and statistical complexity of our 4 x 4 x 2 experimental design. Figure 5 represents 16 distinct distributions across four groups and four conditions for both location and colour measures; employing boxplots/violins for this density of data would significantly increase visual clutter and make the figure difficult to parse.

Furthermore, the primary objective of this figure is to reflect the statistical analysis and illustrate group differences in overall performance and highlight the specific finding that patients with AD were significantly more impaired across all conditions compared to YC, EC, and PD groups. Our statistical focus remains on the mean effects—specifically the significant main effect of group (F(3, 318) = 59.71, p < 0.001) and the critical null-interaction between group and condition (p = 0.90). The error measure most relevant to these comparisons is the standard error of the mean (SEM), rather than the interquartile range (IQR). We think that bar plots provide the most straightforward and scannable representation of these mean differences and the consistent pattern of decay across cohorts for the final manuscript layout.

To address the reviewer’s request for distributional transparency, we have provided a version of Figure 5 using grouped boxplots in the supplementary material (Supplementary figure 2). We note, however, that the spread of raw data points in these plots does not directly reflect the variance associated with our within-subject statistical comparisons.

(8) Results specificity, trans-saccadic integration and ROCF. The authors demonstrate that the derived model parameters account for a significant amount of variability in ROCF performance across the experimental groups tested (Figure 8A). However, it remains unclear how specific the modelling results are with respect to the ROCF.

The ROCF is generally interpreted as a measure of constructional ability. Nevertheless, as with any cognitive test, performance can also be influenced by more general, non-specific abilities that contribute broadly to test success. To more clearly link the specificity between modelling results and constructional ability, it would be helpful to include a test measure for which the model parameters would not be expected to explain performance, for example, a verbal working memory task.

I am not necessarily suggesting that new data should be collected. However, I believe that the issue of specificity should be acknowledged and discussed as a potential limitation in the current context.

We appreciate this important point regarding the discriminant validity of our findings. We agree that cognitive performance in clinical populations is often influenced by a general "g-factor" or non-specific executive decline. However, we chose the ROCF Copy task specifically because it is a hallmark clinical measure of constructional ability that effectively serves as a real-world transsaccadic task, requiring participants to integrate spatial information across hundreds of saccades between the model figure and the drawing surface.

To address the reviewer’s concern regarding specificity, we leveraged the fact that all participants completed the ACE-III, which includes a dedicated verbal memory component (the ACE Memory subscale). We conducted a partial correlation analysis and found that the relationship between transsaccadic working memory and ROCF copy performance remains highly significant (rho = -0.46, p < 0.001), even after controlling for age, education, and the ACE-III Memory subscale score. This suggests that the link between transsaccadic updating and constructional ability is mechanistically specific rather than a byproduct of global cognitive impairment. We have substantially revised the Discussion to highlight this link and the supporting statistical evidence.

We first updated the last paragraph of Introduction:

“Finally, by linking these mechanistic parameters to a standard clinical measure of constructional ability (the Rey-Osterrieth Complex Figure task), we demonstrate that transsaccadic updating represents a core computational phenotype underpinning real-world visuospatial construction in both health and neurodegeneration.”

The new section in Discussion highlighting the ROCF copy link:

“Importantly, our computational framework establishes a direct mechanistic link between trassaccadic updating and real-world constructional ability. Specifically, higher saccade and angular encoding errors contribute to poorer ROCF copy scores. By mapping these mechanistic estimates onto clinical scores, we found that the parameters derived from our winning model explain approximately 62% of the variance in constructional performance across groups. These findings suggest that the computational parameters identified in the LOCUS task represent core phenotypes of visuospatial ability, providing a mechanistic bridge between basic cognitive theory and clinical presentation.

This relationship provides novel insights into the cognitive processes underlying drawing, specifically highlighting the role of transsaccadic working memory. Previous research has primarily focused on the roles of fine motor control and eye-hand coordination in this skill.[4,50–55] This is partly because of consistent failure to find a strong relation between traditional memory measures and copying ability.[4,31] For instance, common measures of working memory, such as digit span and Corsi block tasks, do not directly predict ROCF copying performance.[31,56] Furthermore, in patients with constructional apraxia, these memory performance often remain relatively preserved despite significant drawing impairments.[56–58] In literature, this lack of association has often been attributed to “deictic” visual-sampling strategies, characterised by frequent eye movements that treat the environment as an external memory buffer, thereby minimising the need to maintain a detailed internal representation.[4,59] In a real-world copying task, the ROCF requires a high volume of saccades, making it uniquely sensitive to the precision of the dynamic remapping signals identified here. Recent eye-tracking evidence confirms that patients with AD exhibit significantly more saccades and longer fixations during figure copying compared to controls, potentially as a compensatory response to trassaccadic working memory constraints.[56] This high-frequency sampling—averaging between 150 and 260 saccades for AD patients compared to approximately 100 for healthy controls—renders the task highly dependent on the precision of dynamic remapping signals.[56] We also found that the relationship between transsaccadic working memory and ROCF performance remains highly significant (rho = -0.46, p < 0.001), even after controlling for age, education, and ACE-III Memory subscore. Consequently, transsaccadic updating may represent a discrete computational phenotype required for visuomotor control, rather than a non-specific proxy for global cognitive decline.[58]

In other words, even when visual information is readily available in the world, the act of drawing performance depends critically on working memory across saccades. This reveals a fundamental computational trade-off: while active sampling strategies (characterised with frequent eye-hand movements) effectively reduce the load on capacity-limited working memory, they simultaneously increase the demand for precise spatial updating across eye movements. By treating the external world as an "outside" memory buffer, the brain minimises the volume of information it must hold internally, but it becomes entirely dependent on the reliability with which that information is remapped after each eye movement. This perspective aligns with, rather contradicts, the traditional view of active sampling, which posits that individuals adapt their gaze and memory strategies based on specific task demands.[3,60] Furthermore, this perspective provides a mechanistic framework for understanding constructional apraxia; in these clinical populations, the impairment may not lie in a reduced memory "span," but rather in the cumulative noise introduced by the constant spatial remapping required during the copying process.[58,61]

Beyond constructional ability, these findings suggest that the primary evolutionary utility of high-resolution spatial remapping lies in the service of action rather than perception. While spatial remapping is often invoked to explain perceptual stability,[11–13,15] the necessity of high-resolution transsaccadic memory for basic visual perception is debated.[13,62–64] A prevailing view suggests that detailed internal models are unnecessary for perception, given the continuous availability of visual information in the external world.[13,44] Our findings support an alternative perspective, aligning with the proposal that high-resolution transsaccadic memory primarily serves action rather than perception.[13] This is consistent with the need for precise localisation in eye-hand coordination tasks such as pointing or grasping.[65] Even when unaware of intrasaccadic target displacements, individuals rapidly adjust their reaching movements, suggesting direct access of the motor system to remapping signals.[66] Further support comes from evidence that pointing to remembered locations is biased by changes in eye position,[67] and that remapping neurons reside within the dorsal “action” visual pathway, rather than the ventral “perception” visual pathway.[13,68,69] By demonstrating a strong link between transsaccadic working memory and drawing (a complex fine motor skill), our findings suggest that precise visual working memory across eye movements plays an important role in complex fine motor control.”

We are deeply grateful to the reviewers for their meticulous reading of our manuscript and for the constructive feedback provided throughout this process. Your insights have significantly enhanced the clarity and rigour of our work.

In addition to the changes requested by the reviewers, we wish to acknowledge a reporting error identified during the revision process. In the original Results section, the repeated measures ANOVA statistics for YC included Greenhouse-Geisser corrections, and the between-subjects degrees of freedom were incorrectly reported as within-subjects residuals. Upon re-evaluation of the data, we confirmed that the assumption of sphericity was not violated; therefore, we have removed the unnecessary Greenhouse-Geisser corrections and corrected the degrees of freedom throughout the Results and Methods sections. We have ensured that these statistical updates are reflected accurately in the revised manuscript and that they do not alter the significance or interpretation of any of our primary findings.

We hope that these revisions address all the concerns raised and provide a more robust account of our findings. We look forward to your further assessment of our work.

eLife Assessment

Studying the biological roles of polyphosphates in metazoans has been a longstanding challenge to the field given that the polyP synthase has yet to be discovered in metazoans. This important study capitalizes on the sophisticated genetics available in the Drosophila system and uses a combination of methodologies to start to tease apart how polyphosphate participates in Drosophila development and in the clotting of Drosophila hemolymph. The data validating one of these tools (cyto-FLYX ) are solid and well-documented and they will open up a field of research into the functional roles of polyP in a metazoan model. The other tools for tissue specific knockdown of polyP (Mito-FLYX, ER-FLYX, and Nuc-FLYX) have not yet been validated but will be invaluable to the field when they are.

Reviewer #2 (Public review):

Summary:

The authors of this paper note that although polyphosphate (polyP) is found throughout biology, the biological roles of polyP have been under-explored, especially in multicellular organisms. The authors created transgenic Drosophila that expressed a yeast enzyme that degrades polyP, targeting the enzyme to different subcellular compartments (cytosol, mitochondria, ER, and nucleus, terming these altered flies Cyto-FLYX, Mito-FLYX, etc.). The authors show the localization of polyP in various wild-type fruit fly cell types and demonstrate that the targeting vectors did indeed result in expression of the polyP degrading enzyme in the cells of the flies. They then go on to examine the effects of polyP depletion using just one of these targeting systems (the Cyto-FLYX). The primary findings from depletion of cytosolic polyP levels in these flies is that it accelerates eclosion and also appears to participate in hemolymph clotting. Perhaps surprisingly, the flies seemed otherwise healthy and appeared to have little other noticeable defects. The authors use transcriptomics to try to identify pathways altered by the cyto-FLYX construct degrading cytosolic polyP, and it seems likely that their findings in this regard will provide avenues for future investigation. And finally, although the authors found that eclosion is accelerated in pupae of Drosophila expressing the Cyto-FLYX construct, the reason why this happens remains unexplained.

Strengths:

The authors capitalize on the work of other investigators who had previously shown that expression of recombinant yeast exopolyphosphatase could be targeted to specific subcellular compartments to locally deplete polyP, and they also use a recombinant polyP binding protein (PPBD) developed by others to localize polyP. They combine this with the considerable power of Drosophila genetics to explore the roles of polyP by depleting it in specific compartments and cell types to tease out novel biological roles for polyP in a whole organism. This is a substantial advance.

Weaknesses:

Page 4 of Results (paragraph 1): I'm a bit concerned about the specificity of PPBD as a probe for polyP. The authors show that the fusion partner (GST) isn't responsible for the signal, but I don't think they directly demonstrate that PPBD is binding only to polyP. Could it also bind to other anionic substances? A useful control might be to digest the permeabilized cells and tissues with polyphosphatase prior to PPBD staining, and show that the staining is lost.

In the hemolymph clotting experiments, the authors collected 2 ul of hemolymph and then added 1 ul of their test substance (water or a polyP solution). They state that they added either 0.8 or 1.6 nmol polyP in these experiments (the description in the Results differs from that of the Methods). I calculate this will give a polyP concentration of 0.3 or 0.6 mM. This is an extraordinarily high polyP concentration, and is much in excess of the polyP concentrations used in most of the experiments testing the effects of polyP on clotting of mammalian plasma. Why did the authors choose this high polyP concentration? Did they try lower concentrations? It seems possible that too high a polyP concentration would actually have less clotting activity than the optimal polyP concentration.

In the revised version of the manuscript, the authors have productively responded to the previous criticisms. Their new data show stronger controls regarding the specificity of PPBD with regard to its interaction with polyP. The authors also have repeated their hemolymph clotting experiments with lower polyP concentrations, which are likely to be more physiological.

Reviewer #3 (Public review):

Summary:

Sarkar, Bhandari, Jaiswal and colleagues establish a suite of quantitative and genetic tools to use Drosophila melanogaster as a model metazoan organism to study polyphosphate (polyP) biology. By adapting biochemical approaches for use in D. melanogaster, they identify a window of increased polyP levels during development. Using genetic tools, they find that depleting polyP from the cytoplasm alters the timing of metamorphosis, accelerationg eclosion. By adapting subcellular imaging approaches for D. melanogaster, they observe polyP in the nucleolus of several cell types. They further demonstrate that polyP localizes to cytoplasmic puncta in hemocytes, and further that depleting polyP from the cytoplasm of hemocytes impairs hemolymph clotting. Together, these findings establish D. melanogaster as a tractable system for advancing our understanding of polyP in metazoans.

Strengths:

• The FLYX system, combining cell type and compartment-specific expression of ScPpx1, provides a powerful tool for the polyP community.

• The finding that cytoplasmic polyP levels change during development and affect the timing of metamorphosis is an exciting first step in understanding the role of polyP in metazoan development, and possible polyP-related diseases.

• Given the significant existing body of work implicating polyP in the human blood clotting cascade, this study provides compelling evidence that polyP has an ancient role in clotting in metazoans.

Limitations:

• While the authors demonstrate that HA-ScPpx1 protein localizes to the target organelles in the various FLYX constructs, the capacity of these constructs to deplete polyP from the different cellular compartments is not shown. This is an important control to both demonstrate that the GTS-PPBD labeling protocol works, and also to establish the efficacy of compartment-specific depletion. While not necessary to do for all the constructs, it would be helpful to do this for the cyto-FLYX and nuc-FLYX.

• The cell biological data in this study clearly indicates that polyP is enriched in the nucleolus in multiple cell types, consistent with recent findings from other labs, and also that polyP affects gene expression during development. Given that the authors also generate the Nuc-FLYX construct to deplete polyP from the nucleus, it is surprising that they test how depleting cytoplasmic but not nuclear polyP affects development. However, providing these tools is a service to the community, and testing the phenotypic consequences of all the FLYX constructs may arguably be beyond the scope of this first study.

Editors' note: The authors have satisfactorily responded to our most major concerns related to the specificity of PPDB and the physiological levels of polyPs in the clotting experiments. We also recognise the limitations related to the depletion of polyP in other tissues and hope that these data will be made available soon.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Polymers of orthophosphate of varying lengths are abundant in prokaryotes and some eukaryotes, where they regulate many cellular functions. Though they exist in metazoans, few tools exist to study their function. This study documents the development of tools to extract, measure, and deplete inorganic polyphosphates in *Drosophila*. Using these tools, the authors show:

(1) That polyP levels are negligible in embryos and larvae of all stages while they are feeding. They remain high in pupae but their levels drop in adults.

(2) That many cells in tissues such as the salivary glands, oocytes, haemocytes, imaginal discs, optic lobe, muscle, and crop, have polyP that is either cytoplasmic or nuclear (within the nucleolus).

(3) That polyP is necessary in plasmatocytes for blood clotting in Drosophila.

(4) That ployP controls the timing of eclosion.

The tools developed in the study are innovative, well-designed, tested, and well-documented. I enjoyed reading about them and I appreciate that the authors have gone looking for the functional role of polyP in flies, which hasn't been demonstrated before. The documentation of polyP in cells is convincing as its role in plasmatocytes in clotting.

We sincerely thank the reviewer for their encouraging assessment and for recognizing both the innovation of the FLYX toolkit and the functional insights it enables. Their remarks underscore the importance of establishing Drosophila as a tractable model for polyP biology, and we are grateful for their constructive feedback, which further strengthened the manuscript.

Its control of eclosion timing, however, could result from non-specific effects of expressing an exogenous protein in all cells of an animal.

We now explicitly state this limitation in the revised manuscript (p.16, l.347–349). The issue is that no catalytic-dead ScPpX1 is available as a control in the field. We plan to generate such mutants through systematic structural and functional studies and will update the FLYX toolkit once they are developed and validated. Importantly, the accelerated eclosion phenotype is reproducible and correlates with endogenous polyP dynamics.

The RNAseq experiments and their associated analyses on polyP-depleted animals and controls have not been discussed in sufficient detail.  In its current form, the data look to be extremely variable between replicates and I'm therefore unsure of how the differentially regulated genes were identified.

We thank the reviewer for pointing out the lack of clarity. We have expanded our RNAseq analysis in the revised manuscript (p.20, l.430–434). Because of inter-sample variation (PC2 = 19.10%, Fig. S7B), we employed Gene Set Enrichment Analysis (GSEA) rather than strict DEG cutoffs. This method is widely used when the goal is to capture pathway-level changes under variability (1). We now also highlight this limitation explicitly (p.20, l.430–432) and provide an additional table with gene-specific fold change (See Supplementary Table for RNA Sequencing Sheet 1). Please note that we have moved RNAseq data to Supplementary Fig. 7 and 8 as suggested in the review.

It is interesting that no kinases and phosphatases have been identified in flies. Is it possible that flies are utilising the polyP from their gut microbiota? It would be interesting to see if these signatures go away in axenic animals.

This is an interesting possibility. Several observations argue that polyP is synthesized by fly tissues: (i) polyP levels remain very low during feeding stages but build up in wandering third instar larvae after feeding ceases; (ii) PPBD staining is absent from the gut except the crop (Fig. S3O–P); (ii) In C. elegans, intestinal polyP was unaffected when worms were fed polyP-deficient bacteria (2); (iv) depletion of polyP from plasmatocytes alone impairs hemolymph clotting, which would not be expected if gut-derived polyP were the major source and may have contributed to polyP in hemolymph. Nevertheless, we agree that microbiota-derived polyP may contribute, and we plan systematic testing in axenic flies in future work.

Reviewer #2 (Public review):

Summary:

The authors of this paper note that although polyphosphate (polyP) is found throughout biology, the biological roles of polyP have been under-explored, especially in multicellular organisms. The authors created transgenic Drosophila that expressed a yeast enzyme that degrades polyP, targeting the enzyme to different subcellular compartments (cytosol, mitochondria, ER, and nucleus, terming these altered flies Cyto-FLYX, Mito-FLYX, etc.). The authors show the localization of polyP in various wild-type fruit fly cell types and demonstrate that the targeting vectors did indeed result in the expression of the polyP degrading enzyme in the cells of the flies. They then go on to examine the effects of polyP depletion using just one of these targeting systems (the Cyto-FLYX). The primary findings from the depletion of cytosolic polyP levels in these flies are that it accelerates eclosion and also appears to participate in hemolymph clotting. Perhaps surprisingly, the flies seemed otherwise healthy and appeared to have little other noticeable defects. The authors use transcriptomics to try to identify pathways altered by the cyto-FLYX construct degrading cytosolic polyP, and it seems likely that their findings in this regard will provide avenues for future investigation. And finally, although the authors found that eclosion is accelerated in the pupae of Drosophila expressing the Cyto-FLYX construct, the reason why this happens remains unexplained.

Strengths:

The authors capitalize on the work of other investigators who had previously shown that expression of recombinant yeast exopolyphosphatase could be targeted to specific subcellular compartments to locally deplete polyP, and they also use a recombinant polyP-binding protein (PPBD) developed by others to localize polyP. They combine this with the considerable power of Drosophila genetics to explore the roles of polyP by depleting it in specific compartments and cell types to tease out novel biological roles for polyP in a whole organism. This is a substantial advance.

We are grateful to the reviewer for their thorough and thoughtful evaluation. Their balanced summary of our work, recognition of the strengths of our genetic tools, and constructive suggestions have been invaluable in clarifying our experiments and strengthening the conclusions.

Weaknesses:

Page 4 of the Results (paragraph 1): I'm a bit concerned about the specificity of PPBD as a probe for polyP. The authors show that the fusion partner (GST) isn't responsible for the signal, but I don't think they directly demonstrate that PPBD is binding only to polyP. Could it also bind to other anionic substances? A useful control might be to digest the permeabilized cells and tissues with polyphosphatase prior to PPBD staining and show that the staining is lost.

To address this concern, we have done two sets of experiments:

(1) We generated a PPBD mutant (GST-PPBD^Mut). We establish that GST-PPBD binds to polyP-2X FITC, whereas GST-PPBD^Mut and GST do not bind polyP₁₀₀-2X FITC using Microscale Thermophoresis. We found that, unlike the punctate staining pattern of GST-PPBD (wild-type), GST-PPBD^Mut does not stain hemocytes. This data has been added to the revised manuscript (Fig. 2B-D, p.8, l.151–165).

(2) A study in C.elegans by Quarles et.al has performed a similar experiment, suggested by the reviewer. In that study, treating permeabilized tissues with polyphosphatase prior to PPBD staining resulted in a decrease of PPBD-GFP signal from the tissues (2). We also performed the same experiment where we subjected hemocytes to GST-PPBD staining with prior incubation of fixed and permeabilised hemocytes with ScPpX1 and heat-inactivated ScPpX1 protein. We find that both staining intensity and the number of punctae are higher in hemocytes left untreated and in those treated with heat-inactivated ScPpX1. The hemocytes pre-treated with ScPpX1 showed reduced staining intensity and number of punctae. This data has been added to the revised manuscript (Fig. 2E-G, p.8, l.166-172).

Further, Saito et al. reported that PPBD binds to polyP in vitro, as well as in yeast and mammalian cells, with a high affinity of ~45µM for longer polyP chains (35 mer and above) (3). They also show that the affinity of PPBD with RNA and DNA is very low. Furthermore, PPBD could detect differences in polyP labeling in yeasts grown under different physiological conditions that alter polyP levels (3). Taken together, published work and our results suggest that PPBD specifically labels polyP.

In the hemolymph clotting experiments, the authors collected 2 ul of hemolymph and then added 1 ul of their test substance (water or a polyP solution). They state that they added either 0.8 or 1.6 nmol polyP in these experiments (the description in the Results differs from that of the Methods). I calculate this will give a polyP concentration of 0.3 or 0.6 mM. This is an extraordinarily high polyP concentration and is much in excess of the polyP concentrations used in most of the experiments testing the effects of polyP on clotting of mammalian plasma. Why did the authors choose this high polyP concentration? Did they try lower concentrations? It seems possible that too high a polyP concentration would actually have less clotting activity than the optimal polyP concentration.

We repeated the assays using 125 µM polyP, consistent with concentrations employed in mammalian plasma studies (4,5). Even at this lower, physiologically relevant concentration, polyP significantly enhanced clot fibre formation (Included as Fig. S5F–I, p.12, l.241–243). This reconfirms the conclusion that polyP promotes hemolymph clotting.

Author response image 1.

Reviewer #3 (Public review):

Summary:

Sarkar, Bhandari, Jaiswal, and colleagues establish a suite of quantitative and genetic tools to use Drosophila melanogaster as a model metazoan organism to study polyphosphate (polyP) biology. By adapting biochemical approaches for use in D. melanogaster, they identify a window of increased polyP levels during development. Using genetic tools, they find that depleting polyP from the cytoplasm alters the timing of metamorphosis, accelerating eclosion. By adapting subcellular imaging approaches for D. melanogaster, they observe polyP in the nucleolus of several cell types. They further demonstrate that polyP localizes to cytoplasmic puncta in hemocytes, and further that depleting polyP from the cytoplasm of hemocytes impairs hemolymph clotting. Together, these findings establish D. melanogaster as a tractable system for advancing our understanding of polyP in metazoans.

Strengths:

(1) The FLYX system, combining cell type and compartment-specific expression of ScPpx1, provides a powerful tool for the polyP community.

(2) The finding that cytoplasmic polyP levels change during development and affect the timing of metamorphosis is an exciting first step in understanding the role of polyP in metazoan development, and possible polyP-related diseases.

(3) Given the significant existing body of work implicating polyP in the human blood clotting cascade, this study provides compelling evidence that polyP has an ancient role in clotting in metazoans.

We sincerely thank the reviewer for their generous and insightful comments. Their recognition of both the technical strengths of the FLYX system and the broader biological implications reinforces our confidence that this work will serve as a useful foundation for the community.

Limitations:

(1) While the authors demonstrate that HA-ScPpx1 protein localizes to the target organelles in the various FLYX constructs, the capacity of these constructs to deplete polyP from the different cellular compartments is not shown. This is an important control to both demonstrate that the GTS-PPBD labeling protocol works, and also to establish the efficacy of compartment-specific depletion. While not necessary to do this for all the constructs, it would be helpful to do this for the cyto-FLYX and nuc-FLYX.

We confirmed polyP depletion in Cyto-FLYX using the malachite green assay (Fig. 3D, p.10, l.212–214). The efficacy of ScPpX1 has also been earlier demonstrated in mammalian mitochondria (6). Our preliminary data from Mito-ScPpX1 expressed ubiquitously with Tubulin-Gal4 showed a reduction in polyP levels when estimated from whole flies (See Author response image 2 below, ongoing investigation). In an independent study focusing on mitochondrial polyP depletion, we are characterizing these lines in detail  and plan to check the amount of polyP contributed to the cellular pool by mitochondria using subcellular fractionation. Direct phenotypic and polyP depletion analyses of Nuc-FLYX and ER-FLYX are also being carried out, but are in preliminary stages. That there is a difference in levels of polyP in various tissues and that we get a very little subscellular fraction for polyP analysis have been a few challenging issues. This analysis requires detailed, independent, and careful analysis, and thus, we refrain from adding this data to the current manuscript.

Author response image 2.

Regarding the specificity, Saito et.al. reported that PPBD binds to polyP in vitro, as well as in yeast and mammalian cells with a high affinity of ~45µM for longer polyP chains (35 mer and above) (3). They also show that the affinity of PPBD with RNA and DNA is very low. Further, PPBD could reveal differences in polyP labeling with yeasts grown in different physiological conditions that can alter polyP levels. Now in the manuscript, we included following data to show specificity of PPBD:

To address this concern we have done two sets of experiments:

We generated a PPBD mutant (GST-PPBD^Mut). Using Microscale Thermophoresis, we establish that GST-PPBD binds to polyP₁₀₀-2X-FITC, whereas, GST-PPBD^Mut and GST do not bind polyP₁₀₀-2X-FITC at all. We found that unlike the punctate staining pattern of GST-PPBD (wild-type), GST-PPBD^Mut does not stain hemocytes. This data has been added to the revised manuscript (Fig. 2B-D, p.8, l.151–165).

A study in C.elegans by Quarles et.al has performed a similar experiment suggested by the reviewer. In that study, treating permeabilized tissues with polyphosphatase prior to PPBD staining resulted in decrease of PPBD-GFP signal from the tissues (2). We also performed the same experiment where we subjected hemocytes to GST-PPBD staining with prior incubation of fixed and permeabilised hemocytes with ScPpX1 and heat inactivated ScPpX1 protein. We find that both intensity of staining and number of punctae are higher in hemocytes that were left untreated and the one where heat inactivated ScPpX1 was added. The hemocytes pre-treated with ScPpX1 showed reduced staining intensity and number of punctae. This data has been added to the revised manuscript (Fig. 2E-G, p.8, l.166-172).

(2) The cell biological data in this study clearly indicates that polyP is enriched in the nucleolus in multiple cell types, consistent with recent findings from other labs, and also that polyP affects gene expression during development. Given that the authors also generate the Nuc-FLYX construct to deplete polyP from the nucleus, it is surprising that they test how depleting cytoplasmic but not nuclear polyP affects development. However, providing these tools is a service to the community, and testing the phenotypic consequences of all the FLYX constructs may arguably be beyond the scope of this first study.

We agree this is an important avenue. In this first study, we focused on establishing the toolkit and reporting phenotypes with Cyto-FLYX. We are systematically assaying phenotypes from all FLYX constructs, including Nuc-FLYX, in ongoing studies

Recommendations for the authors:

Reviewing Editor Comment:

The reviewers appreciated the general quality of the rigour and work presented in this manuscript. We also had a few recommendations for the authors. These are listed here and the details related to them can be found in the individual reviews below.

(1) We suggest including an appropriate control to show that PPBD binds polyP specifically.

We have updated the response section as follows:

(a) Highlighted previous literature that showed the specificity of PPBD.

(b) We show that the punctate staining observed by PPBD is not demonstrated by the mutant PPBD (PPBD^Mut) in which amino acids that are responsible for polyP binding are mutated.

(c) We show that PPBD^Mut does not bind to polyP using Microscale Thermophoresis.

(d) We show that treatment of fixed and permeabilised hemocytes with ScPpX1 reduces the PPBD staining intensity and number of punctae, as compared to tissues left untreated or treated with heat-inactivated ScPpX1.

We have included these in our updated revised manuscript (Fig. 2B-G, p.8, l.151–157)

(2) The high concentration of PolyP in the clotting assay might be impeding clotting. The authors may want to consider lowering this in their assays.

We have addressed this concern in our revised manuscript. We have performed the clotting assays with lower polyP concentrations (concentrations previously used in clotting experiments with human blood and polyP). Data is included in Fig. S5F–I, p.12, l.241–243.

(3) The RNAseq study: can the authors please describe this better and possibly mine it for the regulation of genes that affect eclosion?

In our revised manuscript, we have included a broader discussion about the RNAseq analysis done in the article in both the ‘results’ and the ‘discussion’ sections, where we have rewritten the narrative from the perspective of accelerated eclosion. (p.15 l.310-335, p. 20, l.431-446).

(4) Have the authors considered the possibility that the gut microbiota might be contributing to some of their measurements and assays? It would be good to address this upfront - either experimentally, in the discussion, or (ideally) both.

This is an exciting possibility. Several observations argue that fly tissues synthesize polyP: (i) polyP levels remain very low during feeding stages but build up in wandering third instar larvae after feeding ceases; (ii) PPBD staining is absent from the gut except the crop (Fig. S3O–P); (iii) in C. elegans, intestinal polyP was unaffected when worms were fed polyP-deficient bacteria (2); (iv) depletion of polyP from plasmatocytes alone impairs hemolymph clotting, which would not be expected if gut-derived polyP were the major source and may have contributed to polyP in hemolymph. Nevertheless, microbiota-derived polyP may contribute, and we plan systematic testing in axenic flies in future work.

Reviewer #1 (Recommendations for the authors):

(1) While the authors have shown that the depletion tool results in a general reduction of polyP levels in Figure 3D, it would have been nice to show this via IHC. Particularly since the depletion depends on the strength of the Gal4, it is possible that the phenotypes are being under-estimated because the depletions are weak.

We agree that different Gal4 lines have different strengths and will therefore affect polyP levels and the strength of the phenotype differently.

We performed PPBD staining on hemocytes expressing ScPPX; however, we observed very intense, uniform staining throughout the cells, which was unexpected. It seems like PPBD is recognizing overexpressed ScPpX1. Indeed, in an unpublished study by Manisha Mallick (Bhandari lab), it was found that His-ScPpX1 specifically interacts with GST-PPBD in a protein interaction assay (See Author response image 3). Due to these issues, we refrained from IHC/PPBD-based validation.

Author response image 3.

(2) The subcellular tools for depletion are neat! I wonder why the authors didn't test them. For example in the salivary gland for nuclear depletion?

We have addressed this question in the reviewer responses. We are systematically assaying phenotypes from all FLYX constructs, including Mito-FLYX, and Nuc-FLYX, in ongoing independent investigations. As discussed in #1, a possible interaction of ScPpX and PPBD is making this test a bit more challenging, and hence, they each require a detailed investigation.

(a) Does the absence of clotting defects using Lz-gal4 suggest that PolyP is more crucial in the plasmatocytoes and for the initial clotting process? And that it is dispensible/less important in the crystal cells and for the later clotting process. Or is it that the crystal cells just don't have as much polyP? The image (2E-H) certainly looks like it.

In hemolymph, the primary clot formation is a result of the clotting factors secreted from the fat bodies and the plasmatocytes. The crystal cells are responsible for the release of factors aiding in successfully hardening the soft clot initially formed. Reports suggest that clotting and melanization of the clot are independent of each other (7). Since Crystal cells do not contribute to clot fibre formation, the absence of clotting defects using LzGAL4-CytoFLYX is not surprising. Alternatively, PolyP may be secreted from all hemocytes and contribute to clotting; however, the crystal cells make up only 5% hemocytes, and hence polyP depletion in those cells may have a negligible effect on blood clotting.

Crystal cells do show PPBD staining. Whether polyP is significantly lower in levels in the crystal cells as compared to the plasmatocytes needs more systematic investigation. Image (2E-H) is a representative image of the presence of polyP in crystal cells and can not be considered to compare polyP levels in the crystal cells vs Plasmatocytes.

(b) The RNAseq analyses and data could be better presented. If the data are indeed variable and the differentially expressed genes of low confidence, I might remove that data entirely. I don't think it'll take away from the rest of the work.

We understand this concern and, therefore, in the revised manuscript, we have included a broader discussion about the RNAseq analysis done in the article in both the ‘results’ and the ‘discussion’ sections, where we have rewritten the narrative from the perspective of accelerated eclosion. (p.15 l.310-335, p. 20, l.431-446). We have also stated the limitations of such studies.

(c) I would re-phrase the first sentence of the results section.

We have re-phrased it in the revised manuscript.

Reviewer #2 (Recommendations for the authors):

(1) The authors created several different versions of the FLYX system that would be targeted to different subcellular compartments. They mostly report on the effects of cytosolic targeting, but some of the constructs targeted the polyphosphatase to mitochondria or the nucleus.

They report that the targeting worked, but I didn't see any results on the effects of those constructs on fly viability, development, etc.

There is a growing literature of investigators targeting polyphosphatase to mitochondria and showing how depleting mitochondrial polyP alters mitochondrial function. What was the effect of the Nuc-FLYX and Mito-FLYX constructs on the flies?

Also, the authors should probably cite the papers of others on the effects of depleting mitochondrial polyP in other eukaryotic cells in the context of discussing their findings in flies.

We have addressed this question in the reviewer responses. We did not see any obvious developmental or viability defects with any of the FLYX lines, and only after careful investigation did we come across the clotting defects in the CytoFLYX. We are currently systematically assaying phenotypes from all FLYX constructs, including Mito-FLYX and Nuc-FLYX, in independent ongoing investigations.

We have discussed the heterologous expression of mitochondrial polyphosphatase in mammalian cells to justify the need for developing Mito-FLYX (p. 10, l. 197-200). In the discussion section, we also discuss the presence and roles of polyP in the nucleus and how Nuc-FLYX can help study such phenomena (p. 19, l. 399-407).

(2) The authors should number the pages of their manuscript to make it easier for reviewers to refer to specific pages.

We have numbered our lines and pages in the revised manuscript.

(3) Abstract: the abbreviation, "polyP", is not defined in the abstract. The first word in the abstract is "polyphosphate", so it should be defined there.

We have corrected it in the revised version.

(4) The authors repeatedly use the phrase, "orange hot", to describe one of the colors in their micrographs, but I don't know how this differs from "orange".

‘OrangeHot’ is the name of the LUT used in the ImageJ analysis and hence referred to as the colour

(5) First page of the Introduction: the phrase, "feeding polyP to αβ expression Alzheimer's model of Caenorhabditis elegans" is awkward (it literally means feeding polyP to the model instead of the worms).

We have revised it. (p.3, l.55-57).

(6) Page 2 of the Introduction: The authors should cite this paper when they state that NUDT3 is a polyphosphatase: https://pubmed.ncbi.nlm.nih.gov/34788624/

We have cited the paper in the revised version of the manuscript. (p.4, l. 68-70)

(7) Page 2 of Results: The authors report the polyP content in the third instar larva (misspelled as "larval") to five significant digits ("419.30"). Their data do not support more than three significant digits, though.

We have corrected it in the revised manuscript.

(8) Page 3 of Results (paragraph 1): When discussing the polyP levels in various larval stages, the authors are extracting total polyP from the larvae. It seems that at least some of the polyP may come from gut microbes. This should probably be mentioned.

This is an interesting possibility. Several observations argue that polyP is synthesized by fly tissues: (i) polyP levels remain very low during feeding stages but build up in wandering third instar larvae after feeding ceases; (ii) PPBD staining is absent from the gut except the crop (Fig. S3O–P); (ii) In C. elegans, intestinal polyP was unaffected when worms were fed polyP-deficient bacteria (2); (iv) depletion of polyP from plasmatocytes alone impairs hemolymph clotting, which would not be expected if gut-derived polyP were the major source and may have contributed to polyP in hemolymph. We mention this limitation in the revised manuscript (p.19-20, l. 425-433).

(9) Page 3 of Results (paragraph 2): stating that the 4% paraformaldehyde works "best" is imprecise. What do the authors mean by "best"?

We have addressed this comment in the revised manuscript and corrected it as 4% paraformaldehyde being better among the three methods we used to fix tissues, which also included methanol and Bouin’s fixative  (p.8, l. 152-154).

(10) Page 4 of Results (paragraph 2, last line of the page): The scientific literature is vast, so one can never be sure that one knows of all the papers out there, even on a topic as relatively limited as polyP. Therefore, I would recommend qualifying the statement "...this is the first comprehensive tissue staining report...". It would be more accurate (and safer) to say something like, "to our knowledge, this is the first..." There is a similar statement with the word "first" on the next page regarding the FLYX library.

We have addressed this concern and corrected it accordingly in the revised version of the manuscript (p.9, l. 192-193)

Reviewer #3 (Recommendations for the authors):

(1) The authors should include in their discussion a comparison of cell biological observations using the polyP binding domain of E. coli Ppx (GST-PPBD) to fluorescently label polyP in cells and tissues with recent work using a similar approach in C. elegans (Quarles et al., PMID:39413779).

In the revised manuscript, we have cited the work of Quarles et al. and have added a comparison of observations (p.19,l.408-410). In the discussion, we have also focused on multiple other studies about how polyP presence in different subcellular compartments, like the nucleus, can be assayed and studied with the tools developed in this study.

(2) The gene expression studies of time-matched Cyto-FLYX vs WT larvae is very intriguing. Given the authors' findings that non-feeding third instar Cyto-FLYX larvae are developmentally ahead of WT larvae, can the observed trends be explained by known changes in gene expression that occur during eclosion? This is mentioned in the results section in the context of genes linked to neurons, but a broader discussion of which pathway changes observed can be explained by the developmental stage difference between the WT and FLYX larvae would be helpful in the discussion.

We have included a broader discussion about the RNAseq analysis done in the article in both the ‘results’ and the ‘discussion’ sections, where we have rewritten the narrative from the perspective of accelerated eclosion. (p.15 l.310-335, p. 20, l.431-446). We have also stated the limitations of such studies.

(3) The sentence describing NUDT3 is not referenced.

We have addressed this comment and have cited the paper of NUDT3 in the revised version of the manuscript.(p.4, l. 68-70)

(4) In the first sentence of the results section, the meaning/validity of the statement "The polyP levels have decreased as evolution progressed" is not clear. It might be more straightforward to give an estimate of the total pmoles polyP/mg protein difference between bacteria/yeast and metazoans.

In the revised manuscript, we have given an estimate of the polyP content across various species across evolution to uphold the statement that polyP levels have decreased as evolution progressed (p. 5, l. 87-91).

(5) The description of the malachite green assay in the results section describes it as "calorimetric" but this should read "colorimetric?"

We have corrected it in the revised manuscript.

References

(1) Chicco D, Agapito G. Nine quick tips for pathway enrichment analysis. PLoS Comput Biol. 2022 Aug 11;18(8):e1010348.

(2) Quarles E, Petreanu L, Narain A, Jain A, Rai A, Wang J, et al. Cryosectioning and immunofluorescence of C. elegans reveals endogenous polyphosphate in intestinal endo-lysosomal organelles. Cell Rep Methods. 2024 Oct 8;100879.

(3) Saito K, Ohtomo R, Kuga-Uetake Y, Aono T, Saito M. Direct labeling of polyphosphate at the ultrastructural level in Saccharomyces cerevisiae by using the affinity of the polyphosphate binding domain of Escherichia coli exopolyphosphatase. Appl Environ Microbiol. 2005 Oct;71(10):5692–701.

(4) Smith SA, Mutch NJ, Baskar D, Rohloff P, Docampo R, Morrissey JH. Polyphosphate modulates blood coagulation and fibrinolysis. Proc Natl Acad Sci USA. 2006 Jan 24;103(4):903–8.

(5) Smith SA, Choi SH, Davis-Harrison R, Huyck J, Boettcher J, Rienstra CM, et al. Polyphosphate exerts differential effects on blood clotting, depending on polymer size. Blood. 2010 Nov 18;116(20):4353–9.

(6) Abramov AY, Fraley C, Diao CT, Winkfein R, Colicos MA, Duchen MR, et al. Targeted polyphosphatase expression alters mitochondrial metabolism and inhibits calcium-dependent cell death. Proc Natl Acad Sci USA. 2007 Nov 13;104(46):18091–6.

(7) Schmid MR, Dziedziech A, Arefin B, Kienzle T, Wang Z, Akhter M, et al. Insect hemolymph coagulation: Kinetics of classically and non-classically secreted clotting factors. Insect Biochem Mol Biol. 2019 Jun;109:63–71.

(8) Jian Guan, Rebecca Lee Hurto, Akash Rai, Christopher A. Azaldegui, Luis A. Ortiz-Rodríguez, Julie S. Biteen, Lydia Freddolino, Ursula Jakob. HP-Bodies – Ancestral Condensates that Regulate RNA Turnover and Protein Translation in Bacteria. bioRxiv 2025.02.06.636932; doi: https://doi.org/10.1101/2025.02.06.636932.

(9) Lonetti A, Szijgyarto Z, Bosch D, Loss O, Azevedo C, Saiardi A. Identification of an evolutionarily conserved family of inorganic polyphosphate endopolyphosphatases. J Biol Chem. 2011 Sep 16;286(37):31966–74.

eLife Assessment

This study presents a valuable advance in reconstructing naturalistic speech from intracranial ECoG data using a dual-pathway model. The evidence supporting the claims of the authors is solid. This work will be of interest to cognitive neuroscientists and computer scientists/engineers working on speech reconstruction from neural data.

Reviewer #1 (Public review):

Summary:

This paper introduces a dual-pathway model for reconstructing naturalistic speech from intracranial ECoG data. It integrates an acoustic pathway (LSTM + HiFi-GAN for spectral detail) and a linguistic pathway (Transformer + Parler-TTS for linguistic content). Output from the two components are later merged via CosyVoice2.0 voice cloning. Using only 20 minutes of ECoG data per participant, the model achieves high acoustic fidelity and linguistic intelligibility.

Strengths:

(1) The proposed dual-pathway framework effectively integrates the strengths of neural-to-acoustic and neural-to-text decoding and aligns well with established neurobiological models of dual-stream processing in speech and language.

(2) The integrated approach achieves robust speech reconstruction using only 20 minutes of ECoG data per subject, demonstrating the efficiency of the proposed method.

(3) The use of multiple evaluation metrics (MOS, mel-spectrogram R², WER, PER) spanning acoustic, linguistic (phoneme and word), and perceptual dimensions, together with comparisons against noise-degraded baselines, adds strong quantitative rigor to the study.

Comments on revisions:

I thank the authors for their thorough efforts in addressing my previous concerns. I believe this revised version is significantly strengthened, and I have no further concerns.

Reviewer #2 (Public review):

Summary:

The study by Li et al. proposes a dual-path framework that concurrently decodes acoustic and linguistic representations from ECoG recordings. By integrating advanced pre-trained AI models, the approach preserves both acoustic richness and linguistic intelligibility, and achieves a WER of 18.9% with a short (~20-minute) recording.

Overall, the study offers an advanced and promising framework for speech decoding. The method appears sound, and the results are clear and convincing. My main concerns are the need for additional control analyses and for more comparisons with existing models.

Strengths:

• This speech-decoding framework employs several advanced pre-trained DNN models, reaching superior performance (WER of 18.9%) with relatively short (~20-minute) neural recording.

• The dual-pathway design is elegant, and the study clearly demonstrates its necessity: The acoustic pathway enhances spectral fidelity while the linguistic pathway improves linguistic intelligibility.

Comments on revisions:

The authors have thoughtfully addressed my previous concerns about the weaknesses. I have no further concerns.

Author response:

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

Public Reviews:

Reviewer #1 (Public review):

Summary

This paper introduces a dual-pathway model for reconstructing naturalistic speech from intracranial ECoG data. It integrates an acoustic pathway (LSTM + HiFi-GAN for spectral detail) and a linguistic pathway (Transformer + Parler-TTS for linguistic content). Output from the two components is later merged via CosyVoice2.0 voice cloning. Using only 20 minutes of ECoG data per participant, the model achieves high acoustic fidelity and linguistic intelligibility.

Strengths

(1) The proposed dual-pathway framework effectively integrates the strengths of neural-to-acoustic and neural-to-text decoding and aligns well with established neurobiological models of dual-stream processing in speech and language.

(2) The integrated approach achieves robust speech reconstruction using only 20 minutes of ECoG data per subject, demonstrating the efficiency of the proposed method.

(3) The use of multiple evaluation metrics (MOS, mel-spectrogram R², WER, PER) spanning acoustic, linguistic (phoneme and word), and perceptual dimensions, together with comparisons against noisedegraded baselines, adds strong quantitative rigor to the study.

We thank Reviewer #1 for the supportive comments. In addition, we appreciate Reviewer #1’s thoughtful comments and feedback. By addressing these comments, we believe we have greatly improved the clarity of our claims and methodology. Below we list our point-to-point responses addressing concerns raised by Reviewer #1.

Weaknesses:

(1) It is unclear how much the acoustic pathway contributes to the final reconstruction results, based on Figures 3B-E and 4E. Including results from Baseline 2 + CosyVoice and Baseline 3 + CosyVoice could help clarify this contribution.

We sincerely appreciate the inquiry from Reviewer 1. We thank the reviewer for this suggestion. However, we believe that directly applying CosyVoice to the outputs of Baseline 2 or Baseline 3 in isolation is not methodologically feasible and would not correctly elucidate the contribution of the auditory pathway and might lead to misinterpretation.

The role of CosyVoice 2.0 in our framework is specifically voice cloning and fusion, not standalone enhancement. It is designed to integrate information from two pathways. Its operation requires two key inputs:

(1) A voice reference speech that provides the target speaker's timbre and prosodic characteristics. In our final pipeline, this is provided by the denoised output of the acoustic pathway (Baseline 2).

(2) A target word sequence that specifies the linguistic content to be spoken. This is obtained by transcribing the output of the linguistic pathway (Baseline 3) using Whisper ASR. Therefore, the standalone outputs of Baseline 2 and Baseline 3 are the purest demonstrations of what each pathway contributes before fusion. The significant improvement in WER/PER and MOS in the final output (compared to Baseline 2) and the significant improvement in melspectrogram R² (compared to Baseline 3) together demonstrate the complementary contributions of the two pathways. The fusion via CosyVoice is the mechanism that allows these contributions to be combined. We have added a clearer explanation of CosyVoice's role and the rationale for not testing it on individual baselines in the revised manuscript (Results section: "The fine-tuned voice cloner further enhances...").

Edits:

Page 11, Lines 277-282:

“ Voice cloning is used to bridge the gap between acoustic fidelity and linguistic intelligibility in speech reconstruction. This approach strategically combines the strengths of complementary pathways: the acoustic pathway preserves speaker-specific spectral characteristics while the linguistic pathway maintains lexical and phonetic precision. By integrating these components through neural voice cloning, we achieve balanced reconstruction that overcomes the limitations inherent in isolated systems. CosyVoice 2.0, the voice cloner module serves specifically as a voice cloning and fusion engine, requiring two inputs: (1) a voice reference speech (provided by the denoised output of the acoustic pathway) to specify the target speaker's identity, and (2) a target word sequence (transcribed from the output of the linguistic pathway) to specify the linguistic content. The standalone baseline outputs of the two pathways can be integrated in this way.”

(2) As noted in the limitations, the reconstruction results heavily rely on pre-trained generative models. However, no comparison is provided with state-of-the-art multimodal LLMs such as Qwen3-Omni, which can process auditory and textual information simultaneously. The rationale for using separate models (Wav2Vec for speech and TTS for text) instead of a single unified generative framework should be clearly justified. In addition, the adaptor employs an LSTM architecture for speech but a Transformer for text, which may introduce confounds in the performance comparison. Is there any theoretical or empirical motivation for adopting recurrent networks for auditory processing and Transformer-based models for textual processing?

We thank the reviewer for the insightful suggestion regarding multimodal large language models (LLMs) such as Qwen3-Omni. It is important to clarify the distinction between general-purpose interactive multimodal models and models specifically designed for high-fidelity voice cloning and speech synthesis.

As for the comparison with the state-of-the-art multimodal LLMs:

Qwen3-Omni and GLM-4-Voice are powerful conversational agents capable of processing multiple modalities including text, speech, image, and video, as described in its documentation (see: https://help.aliyun.com/zh/model-studio/qwen-tts-realtime and https://docs.bigmodel.cn/cn/guide/models/sound-and-video/glm-4-voice). However, it is primarily optimized for interactive dialogue and multimodal understanding rather than for precise, speaker-adaptive speech reconstruction from neural signals. In contrast, CosyVoice 2.0, developed by the same team at Alibaba, is specifically designed for voice cloning and text-to-speech synthesis (see: https://help.aliyun.com/zh/model-studio/text-to-speech). It incorporates advanced speaker adaptation and acoustic modeling capabilities that are essential for reconstructing naturalistic speech from limited neural data. Therefore, our choice of CosyVoice for the final synthesis stage aligns with the goal of integrating acoustic fidelity and linguistic intelligibility, which is central to our study.

For the selection of LSTM and Transformer in the two pathways:

The goal of the acoustic adaptor is to reconstruct fine-grained spectrotemporal details (formants, harmonic structures, prosodic contours) with millisecond-to-centisecond precision. These features rely heavily on local temporal dynamics and short-to-medium range dependencies (e.g., within and between phonemes/syllables). In our ablation studies (to be added in the supplementary), we found that Transformer-based adaptors, which inherently emphasize global sentence-level context through self-attention, tended to oversmooth the reconstructed acoustic features, losing critical fine-temporal details essential for naturalness. In contrast, the recurrent nature of LSTMs, with their inherent temporal state propagation, proved more effective at modeling these local sequential dependencies without excessive smoothing, leading to higher mel-spectrogram fidelity. This aligns with the neurobiological observation that early auditory cortex processes sound with precise temporal fidelity. Moreover, from an engineering perspective, LSTM-based decoders have been empirically shown to perform well in sequential prediction tasks with limited data, as evidenced in prior work on sequence modeling and neural decoding (1).

The goal of the linguistic adaptor is to decode abstract, discrete word tokens. This task benefits from modeling long-range contextual dependencies across a sentence to resolve lexical ambiguity and syntactic structure (e.g., subject-verb agreement). The self-attention mechanism of Transformers is exceptionally well-suited for capturing these global relationships, as evidenced by their dominance in NLP. Our experiments confirmed that a Transformer adaptor outperformed an LSTM-based one in word token prediction accuracy.

While a unified multimodal LLM could in principle handle both modalities, such models often face challenges in modality imbalance and task specialization. Audio and text modalities have distinct temporal scales, feature distributions, and learning dynamics. By decoupling them into separate pathways with specialized adaptors, we ensure that each modality is processed by an architecture optimized for its inherent structure. This divide-and-conquer strategy avoids the risk of one modality dominating or interfering with the learning of the other, leading to more stable training and better final performance, especially important when adapting to limited neural data.

Edits:

Page 9, Lines 214-223:

“The acoustic pathway, implemented through a bi-directional LSTM neural adaptor architecture (Fig. 1B), specializes in reconstructing fundamental acoustic properties of speech. This module directly processes neural recordings to generate precise time-frequency representations, focusing on preserving speaker-specific spectral characteristics like formant structures, harmonic patterns, and spectral envelope details. Quantitative evaluation confirms its core competency: achieving a mel-spectrogram R² of 0.793 ± 0.016 (Fig. 3B) demonstrates remarkable fidelity in reconstructing acoustic microstructure. This performance level is statistically indistinguishable from original speech degraded by 0dB additive noise (0.771 ± 0.014, p = 0.242, one-sided t-test). We chose a bidirectional LSTM architecture for this adaptor because its recurrent nature is particularly suited to modeling the fine-grained, short- to medium-range temporal dependencies (e.g., within and between phonemes and syllables) that are critical for acoustic fidelity. An ablation study comparing LSTM against Transformerbased adaptors for this task confirmed that LSTMs yielded superior mel-spectrogram reconstruction fidelity (higher R²), as detailed in Table S1, likely by avoiding the oversmoothing of spectrotemporal details sometimes induced by the strong global context modeling of Transformers”.

“To confirm that the acoustic pathway’s output is causally dependent on the neural signal rather than the generative prior of the HiFi-GAN, we performed a control analysis in which portions of the input ECoG recording were replaced with Gaussian noise. When either the first half, second half, or the entirety of the neural input was replaced by noise, the melspectrogram R² of the reconstructed speech dropped markedly, corresponding to the corrupted segment (Fig. S5). This demonstrates that the reconstruction is temporally locked to the specific neural input and that the model does not ‘hallucinate’ spectrotemporal structure from noise. These results validate that the acoustic pathway performs genuine, input-sensitive neural decoding”.

Edits:

Page 10, Lines 272-277:

“We employed a Transformer-based Seq2Seq architecture for this adaptor to effectively capture the long-range contextual dependencies across a sentence, which are essential for resolving lexical ambiguity and syntactic structure during word token decoding. This choice was validated by an ablation study (Table S2), indicating that the Transformer adaptor outperformed an LSTM-based counterpart in word prediction accuracy”

(3) The model is trained on approximately 20 minutes of data per participant, which raises concerns about potential overfitting. It would be helpful if the authors could analyze whether test sentences with higher or lower reconstruction performance include words that were also present in the training set.

Thank you for raising the important concern regarding potential overfitting given the limited size of our training dataset (~20 minutes per participant). To address this point directly, we performed a detailed lexical overlap analysis between the training and test sets.

The test set contains 219 unique words. Among these:

127 words (58.0%) appeared in the training set (primarily high-frequency, common words).

92 words (42.0%) were entirely novel and did not appear in the training set. We further examined whether trials with the best reconstruction (WER = 0) relied more on training vocabulary. Among these top-performing trials, 55.0% of words appeared in the training set. In contrast, the worst-performing trials showed 51.9% overlap in words in the training set. No significant difference was observed, suggesting that performance is not driven by simple lexical memorization.

The presence of a substantial proportion of novel words (42%) in the test set, combined with the lack of performance advantage for overlapping content, provides strong evidence that our model is generalizing linguistic and acoustic patterns rather than merely memorizing the training vocabulary. High reconstruction performance on unseen words would be improbable under severe overfitting.

Therefore, we conclude that while some lexical overlap exists (as expected in natural language), the model’s performance is driven by its ability to decode generalized neural representations, effectively mitigating the overfitting risk highlighted by the reviewer.

(4) The phoneme confusion matrix in Figure 4A does not appear to align with human phoneme confusion patterns. For instance, /s/ and /z/ differ only in voicing, yet the model does not seem to confuse these phonemes. Does this imply that the model and the human brain operate differently at the mechanistic level?

We thank the reviewer for this detailed observation regarding the difference between our model's phoneme confusion patterns and typical human perceptual confusions (e.g., the lack of /s/-/z/ confusion).

The reviewer is correct in inferring a mechanistic difference. This divergence is primarily attributable to the Parler-TTS model acting as a powerful linguistic prior. Our linguistic pathway decodes word tokens, which Parler-TTS then converts to speech. Trained on massive corpora to produce canonical pronunciations, Parler-TTS effectively performs an implicit "error correction." For instance, if the neural decoding is ambiguous between the words "sip" and "zip," the TTS model's strong prior for lexical and syntactic context will likely resolve it to the correct word, thereby suppressing purely acoustic confusions like voicing.

This has important implications for interpreting our model's errors and its relationship to brain function. The phoneme errors in our final output reflect a combination of neural decoding errors and the generative biases of the TTS model, which is optimized for intelligibility rather than mimicking raw human misperception. This does imply our model operates differently from the human auditory periphery. The human brain may first generate a percept with acoustic confusions, which higher-level language regions then disambiguate. Our model effectively bypasses the "confused percept" stage by directly leveraging a pre-trained, high-level language model for disambiguation. This is a design feature contributing to its high intelligibility, not necessarily a flaw. This observation raises a fascinating question: Could a model that more faithfully simulates the hierarchical processing of the human brain (including early acoustic confusions) provide a better fit to neural data at different processing stages? Future work could further address this question.

Edits:

add another paragraph in Discussion (Page 14, Lines 397-398):

“The phoneme confusion pattern observed in our model output (Fig. 4A) differs from classic human auditory confusion matrices. We attribute this divergence primarily to the influence of the Parler-TTS model, which serves as a strong linguistic prior in our pipeline. This component is trained to generate canonical speech from text tokens. When the upstream neural decoding produces an ambiguous or erroneous token sequence, the TTS model’s internal language model likely performs an implicit ‘error correction,’ favoring linguistically probable words and pronunciations. This underscores that our model’s errors arise from a complex interaction between neural decoding fidelity and the generative biases of the synthesis stage”

(5) In general, is the motivation for adopting the dual-pathway model to better align with the organization of the human brain, or to achieve improved engineering performance? If the goal is primarily engineeringoriented, the authors should compare their approach with a pretrained multimodal LLM rather than relying on the dual-pathway architecture. Conversely, if the design aims to mirror human brain function, additional analysis, such as detailed comparisons of phoneme confusion matrices, should be included to demonstrate that the model exhibits brain-like performance patterns.

Our primary motivation is engineering improvement, to overcome the fundamental trade-off between acoustic fidelity and linguistic intelligibility that has limited previous neural speech decoding work. The design is inspired by the related works of the convergent representation of speech and language perception (2). However, we do not claim that our LSTM and Transformer adaptors precisely simulate the specific neural computations of the human ventral and dorsal streams. The goal was to build a high-performance, data-efficient decoder. We will clarify this point in the Introduction and Discussion, stating that while the architecture is loosely inspired by previous neuroscience results, its primary validation is its engineering performance in achieving state-of-the-art reconstruction quality with minimal data.

Edits:

Page 14, Line 358-373:

“In this study, we present a dual-path framework that synergistically decodes both acoustic and linguistic speech representations from ECoG signals, followed by a fine-tuned zero-shot text-to-speech network to re-synthesize natural speech with unprecedented fidelity and intelligibility. Crucially, by integrating large pre-trained generative models into our acoustic reconstruction pipeline and applying voice cloning technology, our approach preserves acoustic richness while significantly enhancing linguistic intelligibility beyond conventional methods. Our dual-pathway architecture, while inspired by converging neuroscience insights on speech and language perception, was principally designed and validated as an engineering solution. The primary goal to build a practical decoder that achieves state-of-theart reconstruction quality with minimal data. The framework's success is therefore ultimately judged by its performance metrics, high intelligibility (WER, PER), acoustic fidelity (melspectrogram R²), and perceptual quality (MOS), which directly address the core engineering challenge we set out to solve. Using merely 20 minutes of ECoG recordings, our model achieved superior performance with a WER of 18.9% ± 3.3% and PER of 12.0% ± 2.5% (Fig. 2D, E). This integrated architecture, combining pre-trained acoustic (Wav2Vec2.0 and HiFiGAN) and linguistic (Parler-TTS) models through lightweight neural adaptors, enables efficient mapping of ECoG signals to dual latent spaces. Such methodology substantially reduces the need for extensive neural training data while achieving breakthrough word clarity under severe data constraints. The results demonstrate the feasibility of transferring the knowledge embedded in speech-data pre-trained artificial intelligence (AI) models into neural signal decoding, paving the way for more advanced brain-computer interfaces and neuroprosthetics”.

Reviewer #2 (Public review):

Summary:

The study by Li et al. proposes a dual-path framework that concurrently decodes acoustic and linguistic representations from ECoG recordings. By integrating advanced pre-trained AI models, the approach preserves both acoustic richness and linguistic intelligibility, and achieves a WER of 18.9% with a short (~20-minute) recording.

Overall, the study offers an advanced and promising framework for speech decoding. The method appears sound, and the results are clear and convincing. My main concerns are the need for additional control analyses and for more comparisons with existing models.

Strengths:

(1) This speech-decoding framework employs several advanced pre-trained DNN models, reaching superior performance (WER of 18.9%) with relatively short (~20-minute) neural recording.

(2) The dual-pathway design is elegant, and the study clearly demonstrates its necessity: The acoustic pathway enhances spectral fidelity while the linguistic pathway improves linguistic intelligibility.

We thank Reviewer #2 for supportive comments. In addition, we appreciate Reviewer #2’s thoughtful comments and feedback. By addressing these comments, we believe we have greatly improved the clarity of our claims and methodology. Below we list our point-to-point responses addressing concerns raised by Reviewer #2.

Weaknesses:

The DNNs used were pre-trained on large corpora, including TIMIT, which is also the source of the experimental stimuli. More generally, as DNNs are powerful at generating speech, additional evidence is needed to show that decoding performance is driven by neural signals rather than by the DNNs' generative capacity.

Thank you for raising this crucial point regarding the potential for pre-trained DNNs to generate speech independently of the neural input. We fully agree that it is essential to disentangle the contribution of the neural signals from the generative priors of the models. To address this directly, we have conducted two targeted control analyses, as you suggested, and have integrated the results into the revised manuscript (see Fig. S5 and the corresponding description in the Results section):

(1) Random noise input: We fed Gaussian noise (matched in dimensionality and temporal structure to real ECoG recordings) into the trained adaptors. The outputs were acoustically unstructured and linguistically incoherent, confirming that the generative models alone cannot produce meaningful speech without valid neural input.

(2) Partial sentence input (real + noise): For the acoustic pathway, we systematically replaced portions of the ECoG input with noise. The reconstruction quality (mel-spectrogram R²) dropped significantly in the corrupted segments, demonstrating that the decoding is temporally locked to the neural signal and does not “hallucinate” speech from noise.

These results provide strong evidence that our model’s performance is causally dependent on and sensitive to the specific neural input, validating that it performs genuine neural decoding rather than merely leveraging the generative capacity of the pre-trained DNNs.

The detailed edits are in the “recommendations” below. (See recommendations (1) and (2))

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) Clarify the results shown in Figure 4E. The integrated approach appears to perform comparably to Baseline 3 in phoneme class clarity. However, Baseline 3 represents the output of the linguistic pathway alone, which is expected to encode information primarily at the word level.

We appreciate the reviewer's observation and agree that clarification is needed. The phoneme class clarity (PCC) metric shown in Figure 4E measures whether mis-decoded phonemes are more likely to be confused within their own class (vowel-vowel or consonantconsonant) rather than across classes (vowel-consonant). A higher PCC indicates that the model's errors tend to be phonologically similar sounds (e.g., one vowel mistaken for another), which is a reasonable property for intelligibility.

We would like to clarify the nature of Baseline 3. As stated in the manuscript (Results section: "The linguistic pathway reconstructs high-intelligibility, higher-level linguistic information"), Baseline 3 is the output of our linguistic pathway. This pathway operates as follows: the ECoG signals are mapped to word tokens via the Transformer adaptor, and these tokens are then synthesized into speech by the frozen Parler-TTS model. Crucially, the input to Parler-TTS is a sequence of word tokens.

It is important to distinguish between the levels of performance measured: Word Error Rate (WER) reflects accuracy at the lexical level (whole words). The linguistic pathway achieves a low WER by design, as it directly decodes word sequences. Phoneme Error Rate (PER) reflects accuracy at the sublexical phonetic level (phonemes). A low WER generally implies a low PER, because robust word recognition requires reliable phoneme-level representations within the TTS model's prior. This explains why Baseline 3 also exhibits a low PER. However, acoustic fidelity (captured by metrics like mel-spectrogram R²) requires the preservation of fine-grained spectrotemporal details such as pitch, timbre, prosody, and formant structures, information that is not directly encoded at the lexical level and is therefore not a strength of the purely linguistic pathway.

While Parler-TTS internally models sub-word/phonetic information to generate the acoustic waveform, the primary linguistic information driving the synthesis is at the lexical (word) level. The generated speech from Baseline 3 therefore contains reconstructed phonemic sequences derived from the decoded word tokens, not from direct phoneme-level decoding of ECoG.

Therefore, the comparable PCC between our final integrated model and Baseline 3 (linguistic pathway) suggests that the phoneme-level error patterns (i.e., the tendency to confuse within-class phonemes) in our final output are largely inherited from the high-quality linguistic prior embedded in the pre-trained TTS model (Parler-TTS). The integrated framework successfully preserves this desirable property from the linguistic pathway while augmenting it with speaker-specific acoustic details from the acoustic pathway, thereby achieving both high intelligibility (low WER/PER) and high acoustic fidelity (high melspectrogram R²).

We will revise the caption of Figure 4E and the corresponding text in the Results section to make this interpretation explicit.

Edits:

Page 12, Lines 317-322:

“In addition to the confusion matrices, we categorized the phonemes into vowels and consonants to assess the phoneme class clarity. We defined "phoneme class clarity" (PCC) as the proportion of errors where a phoneme was misclassified within the same class versus being misclassified into a different class. The purpose of introducing PCC is to demonstrate that most of the misidentified phonemes belong to the same category (confusion between vowels or consonants), rather than directly comparing the absolute accuracy of phoneme recognition. For instance, a vowel being mistaken for another vowel would be considered a within-class error, whereas a vowel being mistaken for a consonant would be classified as a between-class error” 

(2) Add results from Baseline 2 + CosyVoice and Baseline 3 + CosyVoice to clarify the contribution of the auditory pathway.

Thank you for the suggestion. We appreciate the opportunity to clarify the role of CosyVoice in our framework.

As explained in our response to point (1), CosyVoice 2.0 is designed as a fusion module that requires two inputs: 1) a voice reference (from the acoustic pathway) to specify speaker identity, and 2) a word sequence (from the linguistic pathway) to specify linguistic content. Because it is not a standalone enhancer, applying CosyVoice to a single pathway output (e.g., Baseline 2 or 3 alone) is not quite feasible and would not reflect its intended function and could lead to misinterpretation of each pathway’s contribution.

Instead, we have evaluated the contribution of each pathway by comparing the final integrated output against each standalone pathway output (Baseline 2 and 3). The significant improvements in both acoustic fidelity and linguistic intelligibility demonstrate the complementary roles of the two pathways, which are effectively fused through CosyVoice.

(3) Justify your choice of using LSTM and Transformer architecture for the auditory and linguistic neural adaptors, respectively, and how your methods could compare to using a unified generative multimodal LLM for both pathways.

Thank you for revisiting this important point. We appreciate your interest in the architectural choices and their relationship to state-of-the-art multimodal models.

As detailed in our response to point (2), our choice of LSTM for the acoustic pathway and Transformer for the linguistic pathway is driven by task-specific requirements, supported by ablation studies (Supplementary Tables 1–2). The acoustic pathway benefits from LSTM’s ability to model fine-grained, local temporal dependencies without over-smoothing. The linguistic pathway benefits from Transformer’s ability to capture long-range semantic and syntactic context.

Regarding comparison with unified multimodal LLMs (e.g., Qwen3-Omni), we clarified that such models are optimized for interactive dialogue and multimodal understanding, while our framework relies on specialist models (CosyVoice 2.0, Parler-TTS) that are explicitly designed for high-fidelity, speaker-adaptive speech synthesis, a requirement central to our decoding task.

We have incorporated these justifications into the revised manuscript (Results and Discussion sections) and appreciate the opportunity to further emphasize these points.

Edits:

Page 9, Lines 214-223:

“The acoustic pathway, implemented through a bi-directional LSTM neural adaptor architecture (Fig. 1B), specializes in reconstructing fundamental acoustic properties of speech. This module directly processes neural recordings to generate precise time-frequency representations, focusing on preserving speaker-specific spectral characteristics like formant structures, harmonic patterns, and spectral envelope details. Quantitative evaluation confirms its core competency: achieving a mel-spectrogram R² of 0.793 ± 0.016 (Fig. 3B) demonstrates remarkable fidelity in reconstructing acoustic microstructure. This performance level is statistically indistinguishable from original speech degraded by 0dB additive noise (0.771 ± 0.014, p = 0.242, one-sided t-test). We chose a bidirectional LSTM architecture for this adaptor because its recurrent nature is particularly suited to modeling the fine-grained, short- to medium-range temporal dependencies (e.g., within and between phonemes and syllables) that are critical for acoustic fidelity. An ablation study comparing LSTM against Transformerbased adaptors for this task confirmed that LSTMs yielded superior mel-spectrogram reconstruction fidelity (higher R²), as detailed in Table S1, likely by avoiding the oversmoothing of spectrotemporal details sometimes induced by the strong global context modeling of Transformers”.

“To confirm that the acoustic pathway’s output is causally dependent on the neural signal rather than the generative prior of the HiFi-GAN, we performed a control analysis in which portions of the input ECoG recording were replaced with Gaussian noise. When either the first half, second half, or the entirety of the neural input was replaced by noise, the melspectrogram R² of the reconstructed speech dropped markedly, corresponding to the corrupted segment (Fig. S5). This demonstrates that the reconstruction is temporally locked to the specific neural input and that the model does not ‘hallucinate’ spectrotemporal structure from noise. These results validate that the acoustic pathway performs genuine, input-sensitive neural decoding”.

Page 10, Lines 272-277:

“We employed a Transformer-based Seq2Seq architecture for this adaptor to effectively capture the long-range contextual dependencies across a sentence, which are essential for resolving lexical ambiguity and syntactic structure during word token decoding. This choice was validated by an ablation study (Table S2), indicating that the Transformer adaptor outperformed an LSTM-based counterpart in word prediction accuracy”.

(4) Discuss the differences between the model's phoneme confusion matrix in Figure 4A and human phoneme confusion patterns. In addition, please clarify whether the adoption of the dual-pathway architecture is primarily intended to simulate the organization of the human brain or to achieve engineering improvements.

The observed difference between our model's phoneme confusion matrix and typical human perceptual confusion patterns (e.g., the noted lack of confusion between /s/ and /z/) is, as the reviewer astutely infers, likely attributable to the TTS model (Parler-TTS) acting as a powerful linguistic prior. The linguistic pathway decodes word tokens, and Parler-TTS converts these tokens into speech. Parler-TTS is trained on massive text and speech corpora to produce canonical, clean pronunciations. It effectively performs a form of "error correction" or "canonicalization" based on its internal language model. For example, if the neural decoding is ambiguous between "sip" and "zip", the TTS model's strong prior for lexical and syntactic context may robustly resolve it to the correct word, suppressing purely acoustic confusions like voicing. Therefore, the phoneme errors in our final output reflect a combination of neural decoding errors and the TTS model's generation biases, which are optimized for intelligibility rather than mimicking human misperception. We will add this explanation to the paragraph discussing Figure 4A.

Our primary motivation is engineering improvement, to overcome the fundamental tradeoff between acoustic fidelity and linguistic intelligibility that has limited previous neural speech decoding work. The design is inspired by the convergent representation of speech and language perception (1). However, we do not claim that our LSTM and Transformer adaptors precisely simulate the specific neural computations of the human ventral and dorsal streams. The goal was to build a high-performance, data-efficient decoder. We will clarify this point in the Introduction and Discussion, stating that while the architecture is loosely inspired by previous neuroscience results, its primary validation is its engineering performance in achieving state-of-the-art reconstruction quality with minimal data.

Edits:

Pages 2-3, Lines 74-85:

“Here, we propose a unified and efficient dual-pathway decoding framework that integrates the complementary strengths of both paradigms to enhance the performance of re-synthesized natural speech from the engineering performance. Our method maps intracranial electrocorticography (ECoG) signals into the latent spaces of pre-trained speech and language models via two lightweight neural adaptors: an acoustic pathway, which captures low-level spectral features for naturalistic speech synthesis, and a linguistic pathway, which extracts high-level linguistic tokens for semantic intelligibility. These pathways are fused using a finetuned text-to-speech (TTS) generator with voice cloning, producing re-synthesized speech that retains both the acoustic spectrotemporal details, such as the speaker’s timbre and prosody, and the message linguistic content. The adaptors rely on near-linear mappings and require only 20 minutes of neural data per participant for training, while the generative modules are pre-trained on large unlabeled corpora and require no neural supervision”.

Page 14, Lines 358-373:

“In this study, we present a dual-path framework that synergistically decodes both acoustic and linguistic speech representations from ECoG signals, followed by a fine-tuned zero-shot text-to-speech network to re-synthesize natural speech with unprecedented fidelity and intelligibility. Crucially, by integrating large pre-trained generative models into our acoustic reconstruction pipeline and applying voice cloning technology, our approach preserves acoustic richness while significantly enhancing linguistic intelligibility beyond conventional methods. Our dual-pathway architecture, while inspired by converging neuroscience insights on speech and language perception, was principally designed and validated as an engineering solution. The primary goal to build a practical decoder that achieves state-of-the-art reconstruction quality with minimal data. The framework's success is therefore ultimately judged by its performance metrics, high intelligibility (WER, PER), acoustic fidelity (mel-spectrogram R²), and perceptual quality (MOS), which directly address the core engineering challenge we set out to solve. Using merely 20 minutes of ECoG recordings, our model achieved superior performance with a WER of 18.9% ± 3.3% and PER of 12.0% ± 2.5% (Fig. 2D, E). This integrated architecture, combining pre-trained acoustic (Wav2Vec2.0 and HiFi-GAN) and linguistic (Parler-TTS) models through lightweight neural adaptors, enables efficient mapping of ECoG signals to dual latent spaces. Such methodology substantially reduces the need for extensive neural training data while achieving breakthrough word clarity under severe data constraints. The results demonstrate the feasibility of transferring the knowledge embedded in speech-data pre-trained artificial intelligence (AI) models into neural signal decoding, paving the way for more advanced brain-computer interfaces and neuroprosthetics”.

Reviewer #2 (Recommendations for the authors):

(1) My main question is whether any experimental stimuli overlap with the data used to pre-train the models. The authors might consider using pre-trained models trained on other corpora and training their own model without the TIMIT corpus. Additionally, as pretrained models were used, it might be helpful to evaluate to what extent the decoding is sensitive to the input neural recording or whether the model always outputs meaningful speech. The authors might consider two control analyses: a) whether the model still generates speech-like output if the input is random noise; b) whether the model can decode a complete sentence if the first half recording of a sentence is real but the second half is replaced with noise.

We thank the reviewer for raising this crucial point regarding potential data leakage and the sensitivity of decoding to neural input.

We confirm that the pre-training phase of our core models (Wav2Vec2.0 encoder, HiFiGAN decoder) was conducted exclusively on the LibriSpeech corpus (960 hours), which is entirely separate from the TIMIT corpus used for our ECoG experiments. The subsequent fine-tuning of the CosyVoice 2.0 voice cloner for speaker adaptation was performed on the training set portion of the entire TIMIT corpus. Importantly, the test set for all neural decoding evaluations was strictly held out and never used during any fine-tuning stage. This data separation is now explicitly stated in the " Methods" sections for the Speech Autoencoder and the CosyVoice fine-tuning.

Regarding the potential of training on other corpora, we agree it is a valuable robustness check. Previous work has demonstrated that self-supervised speech models like Wav2Vec2.0 learn generalizable representations that transfer well across domains (e.g., Millet et al., NeurIPS 2022). We believe our use of LibriSpeech, a large and diverse corpus, provides a strong, general-purpose acoustic prior.

We agree with the reviewer that control analyses are essential to demonstrate that the decoded output is driven by neural signals and not merely the generative prior of the models. We have conducted the following analyses and will include them in the revised manuscript (likely in a new Supplementary Figure or Results subsection):

(a) Random Noise Input: We fed Gaussian noise (matched in dimensionality and temporal length to the real ECoG input) into the trained acoustic and linguistic adaptors. The outputs were evaluated. The acoustic pathway generated unstructured, noisy spectrograms with no discernible phonetic structure, and the linguistic pathway produced either highly incoherent word sequences or failed to generate meaningful tokens. The fusion via CosyVoice produced unintelligible babble. This confirms that the generative models alone cannot produce structured speech without meaningful neural input.

(b) Partial Sentence Input (Real + Noise): In the acoustic pathway, we replaced the first half, the second half, and all the ECoG recording for test sentences with Gaussian noise. The melspectrogram R² showed a clear degradation in the reconstructed speech corresponding to the noisy segment. We did not do similar experiments in the linguistic pathway because the TTS generator is pre-trained by HuggingFace. We did not train any parameters of Parler-TTS. These results strongly indicate that our model's performance is contingent on and sensitive to the specific neural input, validating that it is performing genuine neural decoding.

Edits:

Page 19, Lines 533-538:

“The parameters in Wav2Vec2.0 were frozen within this training phase. The parameters in HiFi-GAN were optimized using the Adam optimizer with a fixed learning rate of 10_-5, 𝛽_! = 0.9, 𝛽₂ = 0.999. We trained this Autoencoder in LibriSpeech, a 960-hour English speech corpus with a sampling rate of 16kHz, which is entirely separate from the TIMIT corpus used for our ECoG experiments. We spent 12 days in parallel training on 6 Nvidia GeForce RTX3090 GPUs. The maximum training epoch was 2000. The optimization did not stop until the validation loss no longer decreased”.

Edits:

Page9, Lines214-223:

“The acoustic pathway, implemented through a bi-directional LSTM neural adaptor architecture (Fig. 1B), specializes in reconstructing fundamental acoustic properties of speech. This module directly processes neural recordings to generate precise time-frequency representations, focusing on preserving speaker-specific spectral characteristics like formant structures, harmonic patterns, and spectral envelope details. Quantitative evaluation confirms its core competency: achieving a mel-spectrogram R² of 0.793 ± 0.016 (Fig. 3B) demonstrates remarkable fidelity in reconstructing acoustic microstructure. This performance level is statistically indistinguishable from original speech degraded by 0dB additive noise (0.771 ± 0.014, p = 0.242, one-sided t-test). We chose a bidirectional LSTM architecture for this adaptor because its recurrent nature is particularly suited to modeling the fine-grained, short- to medium-range temporal dependencies (e.g., within and between phonemes and syllables) that are critical for acoustic fidelity. An ablation study comparing LSTM against Transformer-based adaptors for this task confirmed that LSTMs yielded superior mel-spectrogram reconstruction fidelity (higher R²), as detailed in Table S1, likely by avoiding the oversmoothing of spectrotemporal details sometimes induced by the strong global context modeling of Transformers”.

“To confirm that the acoustic pathway’s output is causally dependent on the neural signal rather than the generative prior of the HiFi-GAN, we performed a control analysis in which portions of the input ECoG recording were replaced with Gaussian noise. When either the first half, second half, or the entirety of the neural input was replaced by noise, the melspectrogram R² of the reconstructed speech dropped markedly, corresponding to the corrupted segment (Fig. S5). This demonstrates that the reconstruction is temporally locked to the specific neural input and that the model does not ‘hallucinate’ spectrotemporal structure from noise. These results validate that the acoustic pathway performs genuine, input-sensitive neural decoding”

(2) For BCI applications, the decoding speed matters. Please report the model's inference speed. Additionally, the authors might also consider reporting cross-participant generalization and how the accuracy changes with recording duration.

We thank the reviewer for these practical and important suggestions. 

Inference Speed: You are absolutely right. On our hardware (single NVIDIA GeForce RTX 3090 GPU), the current pipeline has an inference time that is longer than the duration of the target speech segment. The primary bottlenecks are the sequential processing in the autoregressive linguistic adaptor and the high-resolution waveform generation in CosyVoice 2.0. This latency currently limits real-time application. We have now added this in the Discussion acknowledging this limitation and stating that future work must focus on architectural optimizations (e.g., non-autoregressive models, lighter vocoders) and potential hardware acceleration to achieve real-time performance, which is critical for a practical BCI.

Cross-Participant Generalization: We agree that this is a key question for scalability. Our framework already addresses part of the cross-participant generalization challenge through the use of pre-trained generative modules (HiFi-GAN, Parler-TTS, CosyVoice 2.0), which are pretrained on large corpora and shared across all participants. Only a small fraction of the model, the lightweight neural adaptors, is subject-specific and requires a small amount of supervised fine-tuning (~20 minutes per participant). This design significantly reduces the per-subject calibration burden. As the reviewer implies, the ultimate goal would be pure zero-shot generalization. A promising future direction is to further improve cross-participant alignment by learning a shared neural feature encoder (e.g., using contrastive or self-supervised learning on aggregated ECoG data) before the personalized adaptors. We have added a paragraph in the Discussion outlining this as a major next step to enhance the framework’s practicality and further reduce calibration time.

Accuracy vs. Recording Duration: Thank you for this insightful suggestion. To systematically evaluate the impact of training data volume on performance, we have conducted additional experiments using progressively smaller subsets of the full training set (i.e., 25%, 50%, and 75%). When we used more than 50% of the training data, performance degrades gracefully rather than catastrophically with less data, which is promising for potential clinical scenarios where data collection may be limited. We add another figure (Fig. S4) to demonstrate this.

Edits:

Pages 15-16, Lines 427-452:

“There are several limitations in our study. The quality of the re-synthesized speech heavily relies on the performance of the generative model, indicating that future work should focus on refining and enhancing these models. Currently, our study utilized English speech sentences as input stimuli, and the performance of the system in other languages remains to be evaluated. Regarding signal modality and experimental methods, the clinical setting restricts us to collecting data during brief periods of awake neurosurgeries, which limits the amount of usable neural activity recordings. Overcoming this time constraint could facilitate the acquisition of larger datasets, thereby contributing to the re-synthesis of higher-quality natural speech. Furthermore, the inference speed of the current pipeline presents a challenge for real-time applications. On our hardware (a single NVIDIA GeForce RTX 3090 GPU), synthesizing speech from neural data takes approximately two to three times longer than the duration of the target speech segment itself. This latency is primarily attributed to the sequential processing in the autoregressive linguistic adaptor and the computationally intensive high-fidelity waveform generation in the vocoder (CosyVoice 2.0). While the current study focuses on offline reconstruction accuracy, achieving real-time or faster-than-real-time inference is a critical engineering goal for viable speech BCI prosthetics. Future work must therefore prioritize architectural optimizations, such as exploring non-autoregressive decoding strategies and more efficient neural vocoders, alongside potential hardware acceleration. Additionally, exploring non-invasive methods represents another frontier; with the accumulation of more data and the development of more powerful generative models, it may become feasible to achieve effective non-invasive neural decoding for speech resynthesis. Moreover, while our framework adopts specialized architectures (LSTM and Transformer) for distinct decoding tasks, an alternative approach is to employ a unified multimodal large language model (LLM) capable of joint acoustic-linguistic processing. Finally, the current framework requires training participant-specific adaptors, which limits its immediate applicability for new users. A critical next step is to develop methods that learn a shared, cross-participant neural feature encoder, for instance, by applying contrastive or selfsupervised learning techniques to larger aggregated ECoG datasets. Such an encoder could extract subject-invariant neural representations of speech, serving as a robust initialization before lightweight, personalized fine-tuning. This approach would dramatically reduce the amount of per-subject calibration data and time required, enhancing the practicality and scalability of the decoding framework for real-world BCI applications”

“In summary, our dual-path framework achieves high speech reconstruction quality by strategically integrating language models for lexical precision and voice cloning for vocal identity preservation, yielding a 37.4% improvement in MOS scores over conventional methods. This approach enables high-fidelity, sentence-level speech synthesis directly from cortical recordings while maintaining speaker-specific vocal characteristics. Despite current constraints in generative model dependency and intraoperative data collection, our work establishes a new foundation for neural decoding development. Future efforts should prioritize: (1) refining few-shot adaptation techniques, (2) developing non-invasive implementations, (3) expanding to dynamic dialogue contexts, and (4) cross-subject applications. The convergence of neurophysiological data with multimodal foundation models promises transformative advances, not only revolutionizing speech BCIs but potentially extending to cognitive prosthetics for memory augmentation and emotional communication. Ultimately, this paradigm will deepen our understanding of neural speech processing while creating clinically viable communication solutions for those with severe speech impairments”

Edits: 

add another section in Methods: Page 22, Line 681:

“Ablation study on training data volume”.

“To assess the impact of training data quantity on decoding performance, we conducted an additional ablation experiment. For each participant, we created subsets of the full training set corresponding to 25%, 50%, and 75% of the original data by random sampling while preserving the temporal continuity of speech segments. Personalized acoustic and linguistic adaptors were then independently trained from scratch on each subset, following the identical architecture and optimization procedures described above. All other components of the pipeline, including the frozen pre-trained generators (HiFi-GAN, Parler-TTS) and the CosyVoice 2.0 voice cloner, remained unchanged. Performance metrics (mel-spectrogram R², WER, PER) were evaluated on the same held-out test set for all data conditions. The results (Fig. S4) demonstrate that when more than 50% of the training data is utilized, performance degrades gracefully rather than catastrophically, which is a promising indicator for clinical applications with limited data collection time”.

(3) I appreciate that the author compared their model with the MLP, but more comparisons with previous models could be beneficial. Even simply summarizing some measures of earlier models, such as neural recording duration, WER, PER, etc., is ok.

Thank you for this suggestion. We agree that a broader comparison contextualizes our contribution. We also acknowledge that given the differences in tasks, signal modality, and amount of data, it’s hard to draw a direct comparison. The main goal of this table is to summarize major studies, their methods and results for reference. We have now added a new Supplementary Table that summarizes key metrics from several recent and relevant studies in neural speech decoding. The table includes:

- Neural modality (e.g., ECoG, sEEG, Utah array)

- Approximate amount of neural data used per subject for decoder training

- Primary task (perception vs. production)

-Decoding framework

-Reported Word Error Rate (WER) or similar intelligibility metrics (e.g., Character Error Rate)

-Reported acoustic fidelity metrics (if available, e.g., spectral correlation)

This table includes works such as Anumanchipalli et al., Nature 2019; Akbari et al., Sci Rep 2019; Willett et al., Nature 2023; and other contemporary studies. The table clearly shows that our dual-path framework achieves a highly competitive WER (~18.9%) using an exceptionally short neural recording duration (~20 minutes), highlighting its data efficiency. We will refer to this table in the revised manuscript.

Edits:

Page 14, Lines 374-376:

“Our framework establishes a framework for speech decoding by outperforming prior acousticonly or linguistic-only approaches (Table S3) through integrated pretraining-powered acoustic and linguistic decoding”

Minor:

(1) Some processes might be described earlier, for example, the electrodes were selected, and the model was trained separately for each participant. That information was only described in the Method section now.

Thank you for catching these. We have revised the manuscript accordingly.

Edits:

Page4, Lines 89-95:

“Our proposed framework for reconstructing speech from intracranial neural recordings is designed around two complementary decoding pathways: an acoustic pathway focused on preserving low-level spectral and prosodic detail, and a linguistic pathway focused on decoding high-level textual and semantic content. For every participant, our adaptor is independently trained, and we select speech-responsive electrodes (selection details are provided in the Methods section) to tailor the model to individual neural patterns. These two streams are ultimately fused to synthesize speech that is both natural-sounding and intelligible, capturing the full richness of spoken language. Fig. 1 provides a schematic overview of this dual-pathway architecture”

(2) Line 224-228 Figure 2 should be Figure 3

Thank you for catching these. We have revised the manuscript accordingly. The information about participant-specific training and electrode selection is now briefly mentioned in the "Results" overview (section: "The acoustic and linguistic performance..."), with details still in the Methods. The figure reference error has been corrected.

Edits:

Page7, Lines 224-228:

“However, exclusive reliance on acoustic reconstruction reveals fundamental limitations. Despite excellent spectral fidelity, the pathway produces critically impaired linguistic intelligibility. At the word level, intelligibility remains unacceptably low (WER = 74.6 ± 5.5%, Fig. 3D), while MOS and phoneme-level precision fares only marginally better (MOS = 2.878 ± 0.205, Fig. 3C; PER = 28.1 ± 2.2%, Fig. 3E)”.

(3) For Figure 3C, why does the MOS seem to be higher for baseline 3 than for ground truth? Is this significant?

This is a detailed observation. Baseline 3 achieves a mean opinion score of 4.822 ± 0.086 (Fig. 3C), significantly surpassing even the original human speech (4.234 ± 0.097, p = 6.674×10⁻33). We believe this trend arises because the TIMIT corpus, recorded decades ago, contains inherent acoustic noise and relatively lower fidelity compared to modern speech corpus. In contrast, the Parler-TTS model used in Baseline 3 is trained on massive, highquality, clean speech datasets. Therefore, it synthesizes speech that listeners may subjectively perceive as "cleaner" or more pleasant, even if it lacks the original speaker's voice. Crucially, as the reviewer implies, our final integrated output does not aim to maximize MOS at the cost of speaker identity; it successfully balances this subjective quality with high intelligibility and restored acoustic fidelity. We will add a brief note explaining this possible reason in the caption of Figure 3C.

Edits:

Page9, Lines 235-245:

“The linguistic pathway reconstructs high-intelligibility, higher-level linguistic information”

“The linguistic pathway, instantiated through a pre-trained TTS generator (Fig. 1B), excels in reconstructing abstract linguistic representations. This module operates at the phonological and lexical levels, converting discrete word tokens into continuous speech signals while preserving prosodic contours, syllable boundaries, and phonetic sequences. It achieves a mean opinion score of 4.822 ± 0.086 (Fig. 3C) - significantly surpassing even the original human speech (4.234 ± 0.097, p = 6.674×10⁻33) in that the TIMIT corpus, recorded decades ago, contains inherent acoustic noise and relatively lower fidelity compared to modern speech corpus.  Complementing this perceptual quality, objective intelligibility metrics confirm outstanding performance: WER reaches 17.7 ± 3.2%, with PER at 11.0 ± 2.3%”.

Reference

(1) Chen M X, Firat O, Bapna A, et al. The best of both worlds: Combining recent advances in neural machine translation[C]//Proceedings of the 56th annual meeting of the association for computational linguistics (Volume 1: Long papers). 2018: 76-86

(2) P. Chen et al. Do Self-Supervised Speech and Language Models Extract Similar Representations as Human Brain? 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2024). 2225–2229 (2024).

(3) H. Akbari, B. Khalighinejad, J. L. Herrero, A. D. Mehta, N. Mesgarani, Towards reconstructing intelligible speech from the human auditory cortex. Scientific reports 9, 874 (2019).

(4) S. Komeiji et al., Transformer-Based Estimation of Spoken Sentences Using Electrocorticography. Int Conf Acoust Spee, 1311-1315 (2022).

(5) L. Bellier et al., Music can be reconstructed from human auditory cortex activity using nonlinear decoding models. Plos Biology 21,  (2023).

(6) F. R. Willett et al., A high-performance speech neuroprosthesis. Nature 620,  (2023).

(7) S. L. Metzger et al., A high-performance neuroprosthesis for speech decoding and avatar control. Nature 620, 1037-1046 (2023).

(8) J. W. Li et al., Neural2speech: A Transfer Learning Framework for NeuralDriven Speech Reconstruction. Int Conf Acoust Spee, 2200-2204 (2024).

(9) X. P. Chen et al., A neural speech decoding framework leveraging deep learning and speech synthesis. Nat Mach Intell 6,  (2024).

(10) M. Wairagkar et al., An instantaneous voice-synthesis neuroprosthesis. Nature,  (2025).

eLife Assessment

In this important study, the authors engineered and characterised novel genetically encoded calcium indicators (GECIs) and an analytical tool (CaFire) capable of reporting and quantifying various sub-synaptic events, including miniature synaptic events, with a speed and sensitivity approaching that of intracellular electrophysiological recordings. They present compelling data validating this toolset, which will be of interest to neurobiologists studying synaptic calcium dynamics in various model systems.

Reviewer #1 (Public review):

Summary:

Chen et al. engineered and characterized a suite of next-generation GECIs for the Drosophila NMJ that allow for the visualization of calcium dynamics within the presynaptic compartment, at presynaptic active zones, and in the postsynaptic compartment. These GECIs include ratiometric presynaptic Scar8m (targeted to synaptic vesicles), ratiometric active zone localized Bar8f (targeted to the scaffold molecule BRP), and postsynaptic SynapGCaMP8m. The authors demonstrate that these new indicators are a large improvement on the widely used GCaMP6 and GCaMP7 series GECIs, with increased speed and sensitivity. They show that presynaptic Scar8m accurately captures presynaptic calcium dynamics with superior sensitivity to the GCaMP6 and GCaMP7 series and with similar kinetics to chemical dyes. The active-zone targeted Bar8f sensor was assessed for the ability to detect release-site specific nanodomain changes, but the authors concluded that this sensor is still too slow to accurately do so. Lastly, the use of postsynaptic SynapGCaMP8m was shown to enable the detection of quantal events with similar resolution to electrophysiological recordings. Finally, the authors developed a Python-based analysis software, CaFire, that enables automated quantification of evoked and spontaneous calcium signals. These tools will greatly expand our ability to detect activity at individual synapses without the need for chemical dyes or electrophysiology.

Strengths:

In this study, the authors rigorously compare their newly engineered GECIs to those previously used at the Drosophila NMJ, highlighting improvements in localization, speed, and sensitivity. These comparisons appropriately substantiate the authors claim that their GECIs are superior to the ones currently in use.

The authors demonstrate the ability of Scar8m to capture subtle changes in presynaptic calcium resulting from differences between MN-Ib and MN-Is terminals and from the induction of presynaptic homeostatic potentiation (PHP), rivaling the sensitivity of chemical dyes.

The improved postsynaptic SynapGCaMP8m is shown to approach the resolution of electrophysiology in resolving quantal events.

The authors created a publicly available pipeline that streamlines and standardizes analysis of calcium imaging data.

Reviewer #2 (Public review):

Summary:

Calcium ions play a key role in synaptic transmission and plasticity. To improve calcium measurements at synaptic terminals, previous studies have targeted genetically encoded calcium indicators (GECIs) to pre- and postsynaptic locations. Here, Chen et al. improve these constructs by incorporating the latest GCaMP8 sensors and a stable red fluorescent protein to enable ratiometric measurements. Extensive characterization in the Drosophila neuromuscular junction demonstrates favorable performance of these new constructs relative to previous genetically encoded and synthetic calcium indicators in reporting synaptic calcium events. In addition, they develop a new analysis platform, 'CaFire', to facilitate automated quantification. Impressively, by positioning postsynaptic GCaMP8m near glutamate receptors, the authors show that their sensors can report miniature synaptic events with speed and sensitivity approaching that of intracellular electrophysiological recordings. These new sensors and the analysis platform provide a valuable tool for resolving synaptic events using all-optical methods.

Strength:

The authors present rigorous characterization of their sensors using well-established assays. They employ immunostaining and super-resolution STED microscopy to confirm correct subcellular targeting. Additionally, they quantify response amplitude, rise and decay kinetics, and provide side-by-side comparisons with earlier-generation GECIs and synthetic dyes. Importantly, they show that the new sensors can reproduce known differences in evoked Ca²⁺ responses between distinct nerve terminals. Finally, they present what appears to be the first simultaneous calcium imaging and intracellular mEPSP recording to directly assess the sensitivity of different sensors in detecting individual miniature synaptic events.

The revised version contains extensive new data and clarification that fully addressed my previous concerns. In particular, I appreciate the side-by-side comparison with synthetic calcium indicator OGB-1 and the cytosolic version of GCaMP8m (now presented in Figure 3), which compellingly supports the favorable performance of their new sensors.

Weakness:

I have only one remaining suggestion about the precision of terminology, which I do think is important. The authors clarified in the revision that they "define SNR operationally as the fractional fluorescence change (ΔF/F).", and basically present ΔF/F values whenever they mentioned about SNR. However, if the intention is to present ΔF/F comparisons, I would strongly suggest replacing all mentions of "SNR" in the manuscript with "ΔF/F" or "fractional/relative fluorescence change".

SNR and ΔF/F are fundamentally different quantities, each with a well-defined and distinct meaning: SNR measures fluorescence change relative to baseline fluctuations (noise), whereas ΔF/F measures fluorescence change relative to baseline fluorescence (F₀). While larger ΔF/F values often correlate with improved detectability, SNR also depends on additional factors such as indicator brightness, light collection efficiency, camera noise, and the stability of the experimental preparation. Referring to ΔF/F as SNR can therefore be misleading and may cause confusion for readers, particularly those from quantitative imaging backgrounds. Clarifying the terminology by consistently using ΔF/F would improve conceptual accuracy without requiring any reanalysis of the data.

Reviewer #3 (Public review):

Genetically encoded calcium indicators (GECIs) are essential tools in neurobiology and physiology. Technological constraints in targeting and kinetics of previous versions of GECIs have limited their application at the subcellular level. Chen et al. present a set of novel tools that overcome many of these limitations. Through systematic testing in the Drosophila NMJ, they demonstrate improved targeting of GCaMP variants to synaptic compartments and report enhanced brightness and temporal fidelity using members of the GCaMP8 series. These advancements are likely to facilitate more precise investigation of synaptic physiology. This manuscript could be improved by further testing the GECIs across physiologically relevant ranges of activity, including at high frequency and over long imaging sessions. Moreover, while the authors provide a custom software package (CaFire) for Ca2+ imaging analysis, comparisons to existing tools and more guidance for broader usability are needed.

In this revised version, Chen et al. answered most of our concerns. The tools developed here will be useful for the community.

Author response:

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

Reviewer #1

Chen et al. engineered and characterized a suite of next-generation GECIs for the Drosophila NMJ that allow for the visualization of calcium dynamics within the presynaptic compartment, at presynaptic active zones, and in the postsynaptic compartment. These GECIs include ratiometric presynaptic Scar8m (targeted to synaptic vesicles), ratiometric active zone localized Bar8f (targeted to the scaffold molecule BRP), and postsynaptic SynapGCaMP8m. The authors demonstrate that these new indicators are a large improvement on the widely used GCaMP6 and GCaMP7 series GECIs, with increased speed and sensitivity. They show that presynaptic Scar8m accurately captures presynaptic calcium dynamics with superior sensitivity to the GCaMP6 and GCaMP7 series and with similar kinetics to chemical dyes. The active-zone targeted Bar8f sensor was assessed for the ability to detect release-site-specific nanodomain changes, but the authors concluded that this sensor is still too slow to accurately do so. Lastly, the use of postsynaptic SynapGCaMP8m was shown to enable the detection of quantal events with similar resolution to electrophysiological recordings. Finally, the authors developed a Python-based analysis software, CaFire, that enables automated quantification of evoked and spontaneous calcium signals. These tools will greatly expand our ability to detect activity at individual synapses without the need for chemical dyes or electrophysiology.

We thank this Reviewer for the overall positive assessment of our manuscript and for the incisive comments.

(1) The role of Excel in the pipeline could be more clearly explained. Lines 182-187 could be better worded to indicate that CaFire provides analysis downstream of intensity detection in ImageJ. Moreover, the data type of the exported data, such as .csv or .xlsx, should be indicated instead of 'export to graphical program such as Microsoft Excel'.

We thank the Reviewer for these comments, many of which were shared by the other reviewers. In response, we have now 1) more clearly explained the role of Excel in the CaFire pipeline (lines 677-681), 2) revised the wording in lines 676-679 to indicate that CaFire provides analysis downsteam of intensity detection in ImageJ, and 3) Clarified the exported data type to Excel (lines 677-681). These efforts have improved the clarity and readability of the CaFire analysis pipeline.

(2) In Figure 2A, the 'Excel' step should either be deleted or included as 'data validation' as ImageJ exports don't require MS Excel or any specific software to be analysed. (Also, the graphic used to depict Excel software in Figure 2A is confusing.)

We thank the reviewer for this helpful suggestion. In the Fig. 2A, we have changed the Excel portion and clarified the processing steps in the revised methods. Specifically, we now indicate that ROIs are first selected in Fiji/ImageJ and analyzed to obtain time-series data containing both the time information and the corresponding imaging mean intensity values. These data are then exported to a spreadsheet file (e.g., Excel), which is used to organize the output before being imported into CaFire for subsequent analysis. These changes can be found in the Fig. 2A and methods (lines 676-681).

(3) Figure 2B should include the 'Partition Specification' window (as shown on the GitHub) as well as the threshold selection to give the readers a better understanding of how the tool works.

We absolutely agree with this comment, and have made the suggested changes to the Fig. 2B. In particular, we have replaced the software interface panels and now include windows illustrating the Load File, Peak Detection, and Partition functions. These updated screenshots provide a clearer view of how CaFire is used to load the data, detect events, and perform partition specification for subsequent analysis. We agree these changes will give the readers a better understanding of how the tool works, and we thank the reviewer for this comment.

(4) The presentation of data is well organized throughout the paper. However, in Figure 6C, it is unclear how the heatmaps represent the spatiotemporal fluorescence dynamics of each indicator. Does the signal correspond to a line drawn across the ROI shown in Figure 6B? If so, this should be indicated.

We apologize that the heatmaps were unclear in Fig panel 6C (Fig. 7C in the Current revision). Each heatmap is derived from a one-pixel-wide vertical line within a miniature-event ROI. These heatmaps correspond to the fluorescence change in the indicated SynapGCaMP variant of individual quantal events and their traces shown in Fig. 7C, with a representative image of the baseline and peak fluorescence shown in Fig. 7B. Specifically, we have added the following to the revised Fig. 7C legend:

The corresponding heatmaps below were generated from a single vertical line extracted from a representative miniature-event ROI, and visualize the spatiotemporal fluorescence dynamics (ΔF/F) along that line over time.

(5) In Figure 6D, the addition of non-matched electrophysiology recordings is confusing. Maybe add "at different time points" to the end of the 6D legend, or consider removing the electrophysiology trace from Figure 6D and referring the reader to the traces in Figure 7A for comparison (considering the same point is made more rigorously in Figure 7).

This is a good point, one shared with another reviewer. We apologize this was not clear, and have now revised this part of the figure to remove the electrophysiological traces in what is now Fig. 7 while keeping the paired ones still in what is now Fig. 8A as suggested by the reviewer. We agree this helps to clarify the quantal calcium transients.

(6) In GitHub, an example ImageJ Script for analyzing the images and creating the inputs for CaFire would be helpful to ensure formatting compatibility, especially given potential variability when exporting intensity information for two channels. In the Usage Guide, more information would be helpful, such as how to select ∆R/R, ideally with screenshots of the application being used to analyze example data for both single-channel and two-channel images.

We agree that additional details added to the GitHub would be helpful for users of CaFire. In response, we have now added the following improvements to the GitHub site: 

- ImageJ operation screenshots

Step-by-step illustrations of ROI drawing and Multi Measure extraction.

- Example Excel file with time and intensity values

Demonstrates the required data format for CaFire import, including proper headers.

- CaFire loading screenshots for single-channel and dual-channel imaging

Shows how to import GCaMP into Channel 1 and mScarlet into Channel 2.

- Peak Detection and Partition setting screenshots

Visual examples of automatic peak detection, manual correction, and trace partitioning.

- Instructions for ROI Extraction and CaFire Analysis

A written guide describing the full workflow from ROI selection to CaFire data export.

These changes have improved the usability and accessibility of CaFire, and we thank the reviewer for these points.

Reviewer #2

Calcium ions play a key role in synaptic transmission and plasticity. To improve calcium measurements at synaptic terminals, previous studies have targeted genetically encoded calcium indicators (GECIs) to pre- and postsynaptic locations. Here, Chen et al. improve these constructs by incorporating the latest GCaMP8 sensors and a stable red fluorescent protein to enable ratiometric measurements. In addition, they develop a new analysis platform, 'CaFire', to facilitate automated quantification. Using these tools, the authors demonstrate favorable properties of their sensors relative to earlier constructs. Impressively, by positioning postsynaptic GCaMP8m near glutamate receptors, they show that their sensors can report miniature synaptic events with speed and sensitivity approaching that of intracellular electrophysiological recordings. These new sensors and the analysis platform provide a valuable tool for resolving synaptic events using all-optical methods.

We thank the Reviewer for their overall positive evaluation and comments.

Major comments:

(1) While the authors rigorously compared the response amplitude, rise, and decay kinetics of several sensors, key parameters like brightness and photobleaching rates are not reported. I feel that including this information is important as synaptically tethered sensors, compared to freely diffusible cytosolic indicators, can be especially prone to photobleaching, particularly under the high-intensity illumination and high-magnification conditions required for synaptic imaging. Quantifying baseline brightness and photobleaching rates would add valuable information for researchers intending to adopt these tools, especially in the context of prolonged or high-speed imaging experiments.

This is a good point made by the reviewer, and one we agree will be useful for researchers to be aware. First, it is important to note that the photobleaching and brightness of the sensors will vary depending on the nature of the user’s imaging equipment, which can vary significantly between widefield microscopes (with various LED or halogen light sources for illumination), laser scanning systems (e.g., line scans with confocal systems), or area scanning systems using resonant scanners (as we use in our current study). Under the same imaging settings, GCaMP8f and 8m exhibit comparable baseline fluorescence, whereas GCaMP6f and 6s are noticeably dimmer; because our aim is to assess each reagent’s potential under optimal conditions, we routinely adjust excitation/camera parameters before acquisition to place baseline fluorescence in an appropriate dynamic range. As an important addition to this study, motivated by the reviewer’s comments above, we now directly compare neuronal cytosolic GCaMP8m expression with our Scar8m sensor, showing higher sensitivity with Scar8m (now shown in the new Fig. 3F-H).

Regarding photobleaching, GCaMP signals are generally stable, while mScarlet is more prone to bleaching: in presynaptic area scanned confocal recordings, the mScarlet channel drops by ~15% over 15 secs, whereas GCaMP6s/8f/8m show no obvious bleaching over the same window (lines 549-553). In contrast, presynaptic widefield imaging using an LED system (CCD), GCaMP8f shows ~8% loss over 15 secs (lines 610-611). Similarly, for postsynaptic SynapGCaMP6f/8f/8m, confocal resonant area scans show no obvious bleaching over 60 secs, while widefield shows ~2–5% bleaching over 60 secs (lines 634-638). Finally, in active-zone/BRP calcium imaging (confocal), mScarlet again bleaches by ~15% over 15 s, while GCaMP8f/8m show no obvious bleaching. The mScarlet-channel bleaching can be corrected in Huygens SVI (Bleaching correction or via the Deconvolution Wizard), whereas we avoid applying bleaching correction to the green GCaMP channel when no clear decay is present to prevent introducing artifacts. This information is now added to the methods (lines 548-553).

(2) In several places, the authors compare the performance of their sensors with synthetic calcium dyes, but these comparisons are based on literature values rather than on side-by-side measurements in the same preparation. Given differences in imaging conditions across studies (e.g., illumination, camera sensitivity, and noise), parameters like indicator brightness, SNR, and photobleaching are difficult to compare meaningfully. Additionally, the limited frame rate used in the present study may preclude accurate assessment of rise times relative to fast chemical dyes. These issues weaken the claim made in the abstract that "...a ratiometric presynaptic GCaMP8m sensor accurately captures .. Ca²⁺ changes with superior sensitivity and similar kinetics compared to chemical dyes." The authors should clearly acknowledge these limitations and soften their conclusions. A direct comparison in the same system, if feasible, would greatly strengthen the manuscript.

We absolutely agree with these points made the reviewer, and have made a concerted effort to address them through the following:

We have now directly compared presynaptic calcium responses on the same imaging system using the chemical dye Oregon Green Bapta-1 (OGB-1), one of the primary synthetic calcium indicators used in our field. These experiments reveal that Scar8f exhibits markedly faster kinetics and an improved signal-to-noise ratio compared to OGB-1, with higher peak fluorescence responses (Scar8f: 0.32, OGB-1: 0.23). The rise time constants of the two indicators are comparable (both ~3 msecs), whereas the decay of Scar8f is faster than that of OGB-1 (Scar8f: ~40, OGB-1: ~60), indicating more rapid signal recovery. These results now directly demonstrate the superiority of the new GCaMP8 sensors we have engineered over conventional synthetic dyes, and are now presented in the new Fig. 3A-E of the manuscript.

We agree with the reviewer that, in the original submission, the relatively slow resonant area scans (~115 fps) limited the temporal resolution of our rise time measurements. To address this, we have re-measured the rise time using higher frame-rate line scans (kHz). For Scar8f, the rise time constant was 6.736 msec at ~115 fps resonant area scanned, but shortened to 2.893 msec when imaged at ~303 fps, indicating that the original protocol underestimated the true kinetics. In addition, for Bar8m, area scans at ~118 fps yielded a rise time constant of 9.019 msec, whereas line scans at ~1085 fps reduced the rise time constant to 3.230 msec. These new measurements are now incorporated into the manuscript ( Figs. 3,4, and 6) to more accurately reflect the fast kinetics of these indicators.

(3) The authors state that their indicators can now achieve measurements previously attainable with chemical dyes and electrophysiology. I encourage the authors to also consider how their tools might enable new measurements beyond what these traditional techniques allow. For example, while electrophysiology can detect summed mEPSPs across synapses, imaging could go a step further by spatially resolving the synaptic origin of individual mEPSP events. One could, for instance, image MN-Ib and MN-Is simultaneously without silencing either input, and detect mEPSP events specific to each synapse. This would enable synapse-specific mapping of quantal events - something electrophysiology alone cannot provide. Demonstrating even a proof-of-principle along these lines could highlight the unique advantages of the new tools by showing that they not only match previous methods but also enable new types of measurements.

These are excellent points raised by the reviewer. In response, we have done the following: 

We have now included a supplemental video as “proof-of-principle” data showing simultaneous imaging of SynapGCaMP8m quantal events at both MN-Is and -Ib, demonstrating that synapse-specific spatial mapping of quantal events can be obtained with this tool (see new Supplemental Video 1). 

We have also included an additional discussion of the potential and limitations of these tools for new measurements beyond conventional approaches. This discussion is now presented in lines 419-421 in the manuscript.

(4) For ratiometric measurements, it is important to estimate and subtract background signals in each channel. Without this correction, the computed ratio may be skewed, as background adds an offset to both channels and can distort the ratio. However, it is not clear from the Methods section whether, or how, background fluorescence was measured and subtracted.

This is a good point, and we agree more clarification about how ratiometric measurements were made is needed. In response, we have now added the following to the Methods section (lines 548-568):

Time-lapse videos were stabilized and bleach-corrected prior to analysis, which visibly reduced frame-toframe motion and intensity drift. In the presynaptic and active-zone mScarlet channel, a bleaching factor of ~1.15 was observed during the 15 sec recording. This bleaching can be corrected using the “Bleaching correction” tool in Huygens SVI. For presynaptic and active-zone GCaMP signals, there was minimal bleaching over these short imaging periods. Therefore, the bleaching correction step for GCaMP was skipped. Both GCaMP and mScarlet channels were processed using the default settings in the Huygens SVI “Deconvolution Wizard” (with the exception of the bleaching correction option). Deconvolution was performed using the CMLE algorithm with the Huygens default stopping criterion and a maximum of 30 iterations, such that the algorithm either converged earlier or, if convergence was not reached, was terminated at this 30iteration limit; no other iteration settings were used across the GCaMP series. ROIs were drawn on the processed images using Fiji ImageJ software, and mean fluorescence time courses were extracted for the GCaMP and mScarlet channels, yielding F_GCaMP(t) and F_mScarlet(t). F(t)s were imported into CaFire with GCaMP assigned to Channel #1 (signal; required) and mScarlet to Channel #2 (baseline/reference; optional). If desired, the mScarlet signal could be smoothed in CaFire using a user-specified moving-average window to reduce high-frequency noise. In CaFire’s ΔR/R mode, the per-frame ratio was computed as R(t)=F_GCaMP(t) and F_mScarlet(t); a baseline ratio R0 was estimated from the pre-stimulus period, and the final response was reported as ΔR/R(t)=[R(t)−R0]/R0, which normalizes GCaMP signals to the co-expressed mScarlet reference and thereby reduces variability arising from differences in sensor expression level or illumination across AZs.

(5) At line 212, the authors claim "... GCaMP8m showing 345.7% higher SNR over GCaMP6s....(Fig. 3D and E) ", yet the cited figure panels do not present any SNR quantification. Figures 3D and E only show response amplitudes and kinetics, which are distinct from SNR. The methods section also does not describe details for how SNR was defined or computed.

This is another good point. We define SNR operationally as the fractional fluorescence change (ΔF/F). Traces were processed with CaFire, which estimates a per-frame baseline F₀(t) with a user-configurable sliding window and percentile. In the Load File panel, users can specify both the length of the moving baseline window and the desired percentile; the default settings are a 50-point window and the 30th percentile, representing a 101-point window centered on each time point (previous 50 to next 50 samples) and took the lower 30% of values within that window to estimate F₀(t). The signal was then computed as ΔF/F=[F(t)−F0(t)]/F0(t). This ΔF/F value is what we report as SNR throughout the manuscript and is now discussed explicitly in the revised methods (lines 686-693).

(6) Lines 285-287 "As expected, summed ΔF values scaled strongly and positively with AZ size (Fig. 5F), reflecting a greater number of Cav2 channels at larger AZs". I am not sure about this conclusion. A positive correlation between summed ΔF values and AZ size could simply reflect more GCaMP molecules in larger AZs, which would give rise to larger total fluorescence change even at a given level of calcium increase.

The reviewer makes a good point, one that we agree should be clarified. The reviewer is indeed correct that larger active zones should have more abundant BRP protein, which in turn will lead to a higher abundance of the Bar8f sensor, which should lead to a higher GCaMP response simply by having more of this sensor. However, the inclusion of the ratiometric mScarlet protein should normalize the response accurately, correcting for this confound, in which the higher abundance of GCaMP should be offset (normalized) by the equally (stoichiometric) higher abundance of mScarlet. Therefore, when the ∆R/R is calculated, the differences in GCaMP abundance at each AZ should be corrected for the ratiometric analysis. We now use an improved BRP::mScarlet3::GCaMP8m (Bar8m) and compute ΔR/R with R(t)=F_GCaMP8m/F_mScarlet3. ROIs were drawn over individual AZs (Fig. 6B). CaFire estimated R0 with a sliding 101-point window using the lowest 10% of values, and responses were reported as ΔR/R=[R−R0]/R0. Area-scan examples (118 fps) show robust ΔR/R transients (peaks ≈1.90 and 3.28; tau rise ≈9.0–9.3 ms; Fig. 6C, middle).

We have now made these points more clearly in the manuscript (lines 700-704) and moved the Bar8f intensity vs active zone size data to Table S1. Together, these revisions improve the indicator-abundance confound (via mScarlet normalization). 

(6) Lines 313-314: "SynapGCaMP quantal signals appeared to qualitatively reflect the same events measured with electrophysiological recordings (Fig. 6D)." This statement is quite confusing. In Figure 6D, the corresponding calcium and ephys traces look completely different and appear to reflect distinct sets of events. It was only after reading Figure 7 that I realized the traces shown in Figure 6D might not have been recorded simultaneously. The authors should clarify this point.

Yes, we absolutely agree with this point, one shared by Reviewer 1. In response, we have removed the electrophysiological traces in Fig. 6 to clarify that just the calcium responses are shown, and save the direct comparison for the Fig. 7 data (now revised Fig. 8).

(8) Lines 310-313: "SynapGCaMP8m .... striking an optimal balance between speed and sensitivity", and Lines 314-316: "We conclude that SynapGCaMP8m is an optimal indicator to measure quantal transmission events at the synapse." Statements like these are subjective. In the authors' own comparison, GCaMP8m is significantly slower than GCaMP8f (at least in terms of decay time), despite having a moderately higher response amplitude. It is therefore unclear why GCaMP8m is considered 'optimal'. The authors should clarify this point or explain their rationale for prioritizing response amplitude over speed in the context of their application.

This is another good point that we agree with, as the “optimal” sensor will of course depend on the user’s objectives. Hence, we used the term “an optimal sensor” to indicate it is what we believed to be the best one for our own uses. However, this point should be clarified and better discussed. In response, we have revised the relevant sections of the manuscript to better define why we chose the 8m sensors to strike an optimal balance of speed and sensitivity for our uses, and go on to discuss situations in which other sensor variants might be better suited. These are now presented in lines 223-236 in the revised manuscript, and we thank the reviewer for making these comments, which have improved our study.

Minor comments

(1)  Please include the following information in the Methods section:

(a) For Figures 3 and 4, specify how action potentials were evoked. What type of electrodes were used, where were they placed, and what amount of current or voltage was applied?

We apologize for neglecting to include this information in the original submission. We have now added this information to the revised Methods section (lines 537-543).

(b) For imaging experiments, provide information on the filter sets used for each imaging channel, and describe how acquisition was alternated or synchronized between the green and red channels in ratiometric measurements. Additionally, please report the typical illumination intensity (in mW/mm²) for each experimental condition.

We thank the reviewer for this helpful comment. We have now added detailed information about the imaging configuration to the Methods (lines 512-528) with the following:

Ca2+ imaging was conducted using a Nikon A1R resonant scanning confocal microscope equipped with a 60x/1.0 NA water-immersion objective (refractive index 1.33). GCaMP signals were acquired using the FITC/GFP channel (488-nm laser excitation; emission collected with a 525/50-nm band-pass filter), and mScarlet/mCherry signals were acquired using the TRITC/mCherry channel (561-nm laser excitation; emission collected with a 595/50-nm band-pass filter). ROIs focused on terminal boutons of MN-Ib or -Is motor neurons. For both channels, the confocal pinhole was set to a fixed diameter of 117.5 µm (approximately three Airy units under these conditions), which increases signal collection while maintaining adequate optical sectioning. Images were acquired as 256 × 64 pixel frames (two 12-bit channels) using bidirectional resonant scanning at a frame rate of ~118 frames/s; the scan zoom in NIS-Elements was adjusted so that this field of view encompassed the entire neuromuscular junction and was kept constant across experiments. In ratiometric recordings, the 488-nm (GCaMP) and 561-nm (mScarlet) channels were acquired in a sequential dual-channel mode using the same bidirectional resonant scan settings: for each time point, a frame was first collected in the green channel and then immediately in the red channel, introducing a small, fixed frame-to-frame temporal offset while preserving matched spatial sampling of the two channels.

Directly measuring the absolute laser power at the specimen plane (and thus reporting illumination intensity in mW/mm²) is technically challenging on this resonant-scanning system, because it would require inserting a power sensor into the beam path and perturbing the optical alignment; consequently, we are unable to provide reliable absolute mW/mm² values. Instead, we now report all relevant acquisition parameters (objective, numerical aperture, refractive index, pinhole size, scan format, frame rate, and fixed laser/detector settings) and note that laser powers were kept constant within each experimental series and chosen to minimize bleaching and phototoxicity while maintaining an adequate signal-to-noise ratio. We have now added the details requested in the revised Methods section (lines 512-535), including information about the filter sets, acquisition settings, and typical illumination intensity.

(2) Please clarify what the thin versus thick traces represent in Figures 3D, 3F, 4C, and 4E. Are the thin traces individual trials from the same experiment, or from different experiments/animals? Does the thick trace represent the mean/median across those trials, a fitted curve, or a representative example?

We apologize this was not more clear in the original submission. Thin traces are individual stimulus-evoked trials (“sweeps”) acquired sequentially from the same muscle/NMJ in a single preparation; the panel is shown as a representative example of recordings collected across animals. The thick colored trace is the trialaveraged waveform (arithmetic mean) of those thin traces after alignment to stimulus onset and baseline subtraction (no additional smoothing beyond what is stated in Methods). The thick black curve over the decay phase is a single-exponential fit used to estimate τ. Specifically, we fit the decay segment by linear regression on the natural-log–transformed baseline-subtracted signal, which is equivalent to fitting y = y_peak·e^−t/τdecay over the decay window (revised Fig.4D and Fig.5C legends).

(3) Please clarify what the reported sample size (n) represents. Does it indicate the number of experimental repeats, the number of boutons or PSDs, or the number of animals?

Again, we apologize this was not clear. (n) refers to the number of animals (biological replicates), which is reported in Supplementary Table 1. All imaging was performed at muscle 6, abdominal segment A3. Per preparation, we imaged 1-2 NMJs in total, with each imaging targeting 2–3 terminal boutons at the target NMJ and acquired 2–3 imaging stacks choosing different terminal boutons per NMJ. For the standard stimulation protocol, we delivered 1 Hz stimulation for 1ms and captured 14 stimuli in a 15s time series imaging (lines 730-736).

Reviewer #3

Genetically encoded calcium indicators (GECIs) are essential tools in neurobiology and physiology. Technological constraints in targeting and kinetics of previous versions of GECIs have limited their application at the subcellular level. Chen et al. present a set of novel tools that overcome many of these limitations. Through systematic testing in the Drosophila NMJ, they demonstrate improved targeting of GCaMP variants to synaptic compartments and report enhanced brightness and temporal fidelity using members of the GCaMP8 series. These advancements are likely to facilitate more precise investigation of synaptic physiology.

This is a comprehensive and detailed manuscript that introduces and validates new GECI tools optimized for the study of neurotransmission and neuronal excitability. These tools are likely to be highly impactful across neuroscience subfields. The authors are commended for publicly sharing their imaging software.

This manuscript could be improved by further testing the GECIs across physiologically relevant ranges of activity, including at high frequency and over long imaging sessions. The authors provide a custom software package (CaFire) for Ca2+ imaging analysis; however, to improve clarity and utility for future users, we recommend providing references to existing Ca2+ imaging tools for context and elaborating on some conceptual and methodological aspects, with more guidance for broader usability. These enhancements would strengthen this already strong manuscript.

We thank the Reviewer for their overall positive evaluation and comments. 

Major comments:

(1) Evaluation of the performance of new GECI variants using physiologically relevant stimuli and frequency. The authors took initial steps towards this goal, but it would be helpful to determine the performance of the different GECIs at higher electrical stimulation frequencies (at least as high as 20 Hz) and for longer (10 seconds) (Newman et al, 2017). This will help scientists choose the right GECI for studies testing the reliability of synaptic transmission, which generally requires prolonged highfrequency stimulation.

We appreciate this point by the reviewer and agree it would be of interest to evaluate sensor performance with higher frequency stimulation and for a longer duration. In response, we performed a variety of stimulation protocols at high intensities and times, but found the data to be difficult to separate individual responses given the decay kinetics of all calcium sensors. Hence, we elected not to include these in the revised manuscript. However, we have now included an evaluation of the sensors with 20 Hz electrical stimulation for ~1 sec using a direct comparison of Scar8f with OGB-1. These data are now presented in a new Fig. 3D,E and discussed in the manuscript (lines 396-403).

(2) CaFire.

The authors mention, in line 182: 'Current approaches to analyze synaptic Ca2+ imaging data either repurpose software designed to analyze electrophysiological data or use custom software developed by groups for their own specific needs.' References should be provided. CaImAn comes to mind (Giovannucci et al., 2019, eLife), but we think there are other software programs aimed at analyzing Ca2+ imaging data that would permit such analysis.

Thank you for the thoughtful question. At this stage, we’re unable to provide a direct comparison with existing analysis workflows. In surveying prior studies that analyze Drosophila NMJ Ca²⁺ imaging traces, we found that most groups preprocess images in Fiji/ImageJ and then rely on their own custom-made MATLAB or Python scripts for downstream analysis (see Blum et al. 2021; Xing and Wu 2018). Because these pipelines vary widely across labs, a standardized head-to-head evaluation isn’t currently feasible. With CaFire, our goal is to offer a simple, accessible tool that does not require coding experience and minimizes variability introduced by custom scripts. We designed CaFire to lower the barrier to entry, promote reproducibility, and make quantal event analysis more consistent across users. We have added references to the sentence mentioned above.

Regarding existing software that the reviewer mentioned – CaImAn (Giovannucci et al. 2019): We evaluated CaImAn, which is a powerful framework designed for large-scale, multicellular calcium imaging (e.g., motion correction, denoising, and automated cell/ROI extraction). However, it is not optimized for the per-event kinetics central to our project - such as extracting rise and decay times for individual quantal events at single synapses. Achieving this level of granularity would typically require additional custom Python scripting and parameter tuning within CaImAn’s code-centric interface. This runs counter to CaFire’s design goals of a nocode, task-focused workflow that enables users to analyze miniature events quickly and consistently without specialized programming expertise.

Regarding Igor Pro (WaveMetrics), (Müller et al. 2012): Igor Pro is another platform that can be used to analyze calcium imaging signals. However, it is commercial (paid) software and generally requires substantial custom scripting to fit the specific analyses we need. In practice, it does not offer a simple, open-source, point-and-click path to per-event kinetic quantification, which is what CaFire is designed to provide.

The authors should be commended for making their software publicly available, but there are some questions:

How does CaFire compare to existing tools?

As mentioned above, we have not been able to adapt the custom scripts used by various labs for our purposes, including software developed in MatLab (Blum et al. 2021), Python (Xing and Wu 2018), and Igor (Müller et al. 2012). Some in the field do use semi-publically available software, including Nikon Elements (Chen and Huang 2017) and CaImAn (Giovannucci et al. 2019). However, these platforms are not optimized for the per-event kinetics central to our project - such as extracting rise and decay times for individual quantal events at single synapses. We have added more details about CaFire, mainly focusing on the workflow and measurements, highlighting the superiority of CaFire, showing that CaFire provides a no-code, standardized pipeline with automated miniature-event detection and per-event metrics (e.g., amplitude, rise time τ, decay time τ), optional ΔR/R support, and auto-partition feature. Collectively, these features make CaFire simpler to operate without programming expertise, more transparent and reproducible across users, and better aligned with the event-level kinetics required for this project.

Very few details about the Huygens deconvolution algorithms and input settings were provided in the methods or text (outside of MLE algorithm used in STED images, which was not Ca2+ imaging). Was it blind deconvolution? Did the team distill the point-spread function for the fluorophores? Were both channels processed for ratiometric imaging? Were the same settings used for each channel? Importantly, please include SVI Huygens in the 'Software and Algorithms' Section of the methods.

We thank the reviewer for raising this important point. We have now expanded the Methods to describe our use of Huygens in more detail and have added SVI Huygens Professional (Scientific Volume Imaging, Hilversum, The Netherlands) to the “Software and Algorithms” section. For Ca²⁺ imaging data, time-lapse stacks were processed in the Huygens Deconvolution Wizard using the standard estimation algorithm (CMLE). This is not a blind deconvolution procedure. Instead, Huygens computes a theoretical point-spread function (PSF) from the full acquisition metadata (objective NA, refractive index, voxel size/sampling, pinhole, excitation/emission wavelengths, etc.); if refractive index values are provided and there is a mismatch, the PSF is adjusted to account for spherical aberration. We did not experimentally distill PSFs from bead measurements, as Huygens’ theoretical PSFs are sufficient for our data.

Both green (GCaMP) and red (mScarlet) channels were processed for ratiometric imaging using the same workflow (stabilization, optional bleaching correction, and deconvolution within Huygens). For each channel, the PSF, background, and SNR were estimated automatically by the same built-in algorithms, so the underlying procedures were identical even though the numerical values differ between channels because of their distinct wavelengths and noise characteristics. Importantly, Huygens normalizes each PSF to unit total intensity, such that the deconvolution itself does not add or remove signal and therefore preserves intensity ratios between channels; only background subtraction and bleaching correction can change absolute fluorescence values. For the mScarlet channel, where we observed modest bleaching (~1.10 over 15 sec), we applied Huygens’ bleaching correction and visually verified that similar structures maintained comparable intensities after correction. For presynaptic GCaMP signals, bleaching over these short recordings was negligible, so we omitted the bleaching-correction step to avoid introducing multiplicative artifacts. This workflow ensures that ratiometric ΔR/R measurements are based on consistently processed, intensity-conserving deconvolved images in both channels.

The number of deconvolution iterations could have had an effect when comparing GCAMP series; please provide an average number of iterations used for at least one experiment. For example, Figure 3, Syt::GCAMP6s, Scar8f & Scar8m, and, if applicable, the maximum number of permissible iterations.

We thank the reviewer for this comment. For all Ca²⁺ imaging datasets, deconvolution in Huygens was performed using the recommended default settings of the CMLE algorithm with a maximum of 30 iterations. The stopping criterion was left at the Huygens default, so the algorithm either converged earlier or, if convergence was not reached, terminated at this 30-iteration limit. No other iteration settings were used across the GCaMP series (lines 555-559).

Please clarify if the 'Express' settings in Huygens changed algorithms or shifted input parameters.

We appreciate the reviewer’s question regarding the Huygens “Express” settings. For clarity, we note that all Ca²⁺ imaging data reported in this manuscript were deconvolved using the “Deconvolution Wizard”, not the “Deconvolution Express” mode. In the Wizard, we explicitly selected the CMLE algorithm (or GMLE in a few STED-related cases as recommended by SVI), using the recommended maximum of 30 iterations, and other recommended settings while allowing Huygens to auto-estimate background and SNR for each channel.Bleaching correction was toggled manually per channel (applied to mScarlet when bleaching was evident, omitted for GCaMP when bleaching was negligible), as described in the revised Methods (lines 553-559).

By contrast, the Deconvolution Express tool in Huygens is a fully automated front-end that can internally adjust both the choice of deconvolution algorithm (e.g., CMLE vs. GMLE/QMLE) and key input parameters such as SNR, number of iterations, and quality threshold based on the selected “smart profile” and the image metadata. In preliminary tests on our datasets, Express sometimes produced results that were either overly smoothed or showed subtle artifacts, so we did not use it for any data included in this study. Instead, we relied exclusively on the Wizard with explicitly controlled settings to ensure consistency and transparency across all GCaMP series and ratiometric analyses.

We suggest including a sample data set, perhaps in Excel, so that future users can beta test on and organize their data in a similar fashion.

We agree that this would be useful, a point shared by R1 above. In response, we have added a sample data set to the GitHub site and included sample ImageJ data along with screenshots to explain the analysis in more detail. These improvements are discussed in the manuscript (lines 705-708).

(3) While the challenges of AZ imaging are mentioned, it is not discussed how the authors tackled each one. What is defined as an active zone? Active zones are usually identified under electron microscopy. Arguably, the limitation of GCaMP-based sensors targeted to individual AZs, being unable to resolve local Ca2+ changes at individual boutons reliably, might be incorrect. This could be a limitation of the optical setup being used here. Please discuss further. What sensor performance do we need to achieve this performance level, and/or what optical setup would we need to resolve such signals?

We appreciate the reviewer’s thoughtful comments and agree that the technical challenges of active zone (AZ) Ca²⁺ imaging merit further clarification. We defined AZs, as is the convention in our field, as individual BRP puncta at NMJs. These BRP puncta co-colocalize with individual puncta of other AZ components, including CAC, RBP, Unc13, etc. ROIs were drawn tightly over individual BRP puncta and only clearly separable spots were included.

To tackle the specific obstacles of AZ imaging (small signal volume, high AZ density, and limited photon budget at high frame rates), we implemented both improved sensors and optimized analysis (Fig. 6). First, we introduced a ratiometric AZ-targeted indicator, BRP::mScarlet3::GCaMP8m (Bar8m), and computed ΔR/R with ΔR/R with R(t)=F_GCaMP8m/F_mScarlet3. ROIs were drawn over individual AZs (Fig. 6B). Under our standard resonant area-scan conditions (~118 fps), Bar8m produces robust ΔR/R transients at individual AZs (example peaks ≈ 3.28; τ_rise≈9.0 ms; Fig. 6C, middle), indicating that single-AZ signals can be detected reproducibly when AZs are optically resolvable.

Second, we increased temporal resolution using high-speed Galvano line-scan imaging (~1058 fps), which markedly sharpened the apparent kinetics (τ_rise≈3.23 ms) and revealed greater between-AZ variability (Fig. 6C, right; 6D–E). Population analyses show that line scans yield much faster rise times than area scans (Fig. 6D) and a dramatically higher fraction of significantly different AZ pairs (8.28% and 4.14% in 8f and 8m areascan vs 78.62% in 8m line-scan, lines 721-725), uncovering pronounced AZ-to-AZ heterogeneity in Ca²⁺ signals. Together, these revisions demonstrate that under our current confocal configuration, AZ-targeted GCaMP8m can indeed resolve local Ca²⁺ changes at individual, optically isolated boutons.

We have revised the Discussion to clarify that our original statement about the limitations of AZ-targeted GCaMPs refers specifically to this combination of sensor and optical setup, rather than an absolute limitation of AZ-level Ca²⁺ imaging. In our view, further improvements in baseline brightness and dynamic range (ΔF/F or ΔR/R per action potential), combined with sub-millisecond kinetics and minimal buffering, together with optical configurations that provide smaller effective PSFs and higher photon collection (e.g., higher-NA objectives, optimized 2-photon or fast line-scan modalities, and potentially super-resolution approaches applied to AZ-localized indicators), are likely to be required to achieve routine, high-fidelity Ca²⁺ measurements at every individual AZ within a neuromuscular junction.

(4) In Figure 5: Only GCAMP8f (Bar8f fusion protein) is tested here. Consider including testing with GCAMP8m. This is particularly relevant given that GCAMP8m was a more successful GECI for subcellular post-synaptic imaging in Figure 6.

We appreciate this point and request by Reviewer 3. The main limitation for detecting local calcium changes at AZs is the speed of the calcium sensor, and hence we used the fastest available (GCaMP8f) to test the Bar8f sensor. While replacing GCaMP8f with GCaMP8m would indeed be predicted to enhance sensitivity (SNR), since GCaMP8m does not have faster kinetics relative to GCaMP8f, it is unlikely to be a more successful GECI for visualizing local calcium differences at AZs. 

That being said, we agree that the Bar8m tool, including the improved mScarlet3 indicator, would likely be of interest and use to the field. Fortunately, we had engineered the Bar8m sensor while this manuscript was in review, and just recently received transgenic flies. We have evaluated this sensor, as requested by the reviewer, and included our findings in Fig. 1 and 6. In short, while the sensitivity is indeed enhanced in Bar8m compared to Bar8f, the kinetics remain insufficient to capture local AZ signals. These findings are discussed in the revised manuscript (lines 424-442, 719-730), and we appreciate the reviewer for raising these important points.

In earlier experiments, Bar8f yielded relatively weak fluorescence, so we traded frame rate for image quality during resonant area scans (~60 fps). After switching to Bar8m, the signal was bright enough to restore our standard 118 fps area-scan setting. Nevertheless, even with dual-channel resonant area scans and ratiometric (GCaMP/mScarlet) analysis, AZ-to-AZ heterogeneity remained difficult to resolve. Because Ca²⁺ influx at individual active zones evolves on sub-millisecond timescales, we adopted a high-speed singlechannel Galvano line-scan (~1 kHz) to capture these rapid transients. We first acquired a brief area image to localize AZ puncta, then positioned the line-scan ROI through the center of the selected AZ. This configuration provided the temporal resolution needed to uncover heterogeneity that was under-sampled in area-scan data. Consistent with this, Bar8m line-scan data showed markedly higher AZ heterogeneity (significant AZ-pair rate ~79%, vs. ~8% for Bar8f area scans and ~4% for Bar8m area scans), highlighting Bar8m’s suitability for quantifying AZ diversity. We have updated the text, Methods, and figure legend accordingly (tell reviewer where to find everything).

(5) Figure 5D and associated datasets: Why was Interquartile Range (IQR) testing used instead of ZScoring? Generally, IQR is used when the data is heavily skewed or is not normally distributed. Normality was tested using the D'Agostino & Pearson omnibus normality test and found that normality was not violated. Please explain your reasoning for the approach in statistical testing. Correlation coefficients in Figures 5 E & F should also be reported on the graph, not just the table. In Supplementary Table 1. The sub-table between 4D-F and 5E-F, which describes the IQR, should be labeled as such and contain identifiers in the rows describing which quartile is described. The table description should be below. We would recommend a brief table description for each sub-table.

Thank you for this helpful suggestion. We have updated the analysis in two complementary ways. First, we now perform paired two-tailed t-tests between every two AZs within the same preparation (pairwise AZ–AZ comparisons of peak responses). At α<0.05, the fraction of significant AZ pairs is ~79% for Bar8m line-scan data versus ~8% for Bar8f area-scan data, indicating markedly greater AZ-to-AZ diversity when measured at high temporal resolution. Second, for visually marking the outlying AZs, we re-computed the IQR (Q1–Q3) based on the individual values collected from each AZs(15 data points per AZ, 30 AZs for each genotype), and marked AZs whose mean response falls above Q3 or below Q1; IQR is used here solely as a robust dispersion reference rather than for hypothesis testing. Both analyses support the same observation: Bar8m line-scan data reveal substantially higher AZ heterogeneity than Bar8f and Bar8m area-scan data. We have revised the Methods, figure panels, and legends accordingly (t-test details; explicit “IQR (Q1–Q3)” labeling; significant AZ-pair rates reported on the plots) (lines 719-730).

(6) Figure 6 and associated data. The authors mention: ' SynapGCaMP quantal signals appeared to qualitatively reflect the same events measured with electrophysiological recordings (Fig. 6D).' If that was the case, shouldn't the ephys and optical signal show some sort of correlation? The data presented in Figure 6D show no such correlation. Where do these signals come from? It is important to show the ROIs on a reference image.

We apologize this was not clear, as similar points were raised by R1 and R2. We were just showing separate (uncorrelated) sample traces of electrophysiological and calcium imaging data. Given how confusing this presentation turned out to be, and the fact that we show the correlated ephys and calcium imaging events in Fig. 7, we have elected to remove the uncorrelated electrophysiological events in Fig. 6 to just focus on the calcium imaging events (now Figures 7 and 8).

Figure 7B: Were Ca2+ transients not associated with mEPSPs ever detected? What is the rate of such events?

This is an astute question. Yes indeed, during simultaneous calcium imaging and current clamp electrophysiology recordings, we occasionally observed GCaMP transients without a detectable mEPSP in the electrophysiological trace. This may reflect the detection limit of electrophysiology for very small minis; with our noise level and the technical limitation of the recording rig, events < ~0.2 mV cannot be reliably detected, whereas the optical signal from the same quantal event might still be detected. The fraction of calcium-only events was ~1–10% of all optical miniature events, depending on genotype (higher in lines with smaller average minis). These calcium-only detections were low-amplitude and clustered near the optical threshold (lines 361-365).

Minor comments

(1) It should be mentioned in the text or figure legend whether images in Figure 1 were deconvolved, particularly since image pre-processing is only discussed in Figure 2 and after.

We thank the reviewer for pointing this out. Yes, the confocal images shown in Figure 1 were also deconvolved in Huygens using the CMLE-based workflow described in the revised Methods. We applied deconvolution to improve contrast, reduce out-of-focus blur, and better resolve the morphology of presynaptic boutons, active zones, and postsynaptic structures, so that the localization of each sensor is more clearly visualized. We have now explicitly stated in the Fig. 1 legend and Methods (lines 575-577) that these images were deconvolved prior to display. 

(2) The abbreviation, SNR, signal-to-noise ratio, is not defined in the text.

We have corrected this error and thank the reviewer for pointing this out.

(3) Please comment on the availability of fly stocks and molecular constructs.

We have clarified that all fly stocks and molecular constructs will be shared upon request (lines 747-750). We are also in the process of depositing the new Scar8f/m, Bar8f/m, and SynapGCaMP sensors to the Bloomington Drosophila Stock Center for public dissemination.

(4) Please add detection wavelengths and filter cube information for live imaging experiments for both confocal and widefield.

We thank the reviewer for this helpful suggestion. We have now added the detection wavelengths and filter cube configurations for both confocal and widefield live imaging to the Methods.

For confocal imaging, GCaMP signals were acquired on a Nikon A1R system using the FITC/GFP channel (488-nm laser excitation; emission collected with a 525/50-nm band-pass filter), and mScarlet signals were acquired using the TRITC/mCherry channel (561-nm laser excitation; emission collected with a 595/50-nm band-pass filter). Both channels were detected with GaAsP detectors under the same pinhole and scan settings described above (lines 512-517).

For widefield imaging, GCaMP was recorded using a GFP filter cube (LED excitation ~470/40 nm; emission ~525/50 nm), which is now explicitly described in the revised Methods section (lines 632-633).

(5) Please include a mini frequency analysis in Supplemental Figure S1.

We apologize for not including this information in the original submission. This is now included in the Supplemental Figure S1.

(6) In Figure S1B, consider flipping the order of EPSP (currently middle) and mEPSP (currently left), to easily guide the reader through the quantification of Figure S1A (EPSPs, top traces & mEPSPs, bottom traces).

We agree these modifications would improve readability and clarity. We have now re-ordered the electrophysiological quantifications in Fig. S1B as requested by the reviewer.

(7) Figure 6C: Consider labeling with sensor name instead of GFP.

We agree here as well, and have removed “GFP” and instead added the GCaMP variant to the heatmap in Fig. 7C.

(8) Figure 6E, 7B, 7E: Main statistical differences highlighting sensor performance should be represented on the figures for clarity.

We did not show these differences in the original submission in an effort to keep the figures “clean” and for clarity, putting the detailed statistical significance in Table S1. However, we agree with the reviewer that it would be easier to see these in the Fig. 6E and 7B,E graphs. This information has now been added the Figs. 7 and 8.

(9) Please report if the significance tested between the ephys mini (WT vs IIB-/-, WT vs IIA-/-, IIB-/- vs IIA-/-) is the same as for Ca2+ mini (WT vs IIB-/-, WT vs IIA-/-, IIB-/- vs IIA-/-). These should also exhibit a very high correlation (mEPSP (mV) vs Ca2+ mini deltaF/F). These tests would significantly strengthen the final statement of "SynapGCaMP8m can capture physiologically relevant differences in quantal events with similar sensitivity as electrophysiology."

We agree that adding the more detailed statistical analysis requested by the reviewer would strengthen the evidence for the resolution of quantal calcium imaging using SynapGCaMP8m. We have included the statistical significance between the ephys and calcium minis in Fig. 8 and included the following in the revised methods (lines 358-361), the Fig. 8 legend and Table S1:

Using two-sample Kolmogorov–Smirnov (K–S) tests, we found that SynapGCaMP8m Ca²⁺ minis (ΔF/F, Fig. 8E) differ significantly across all genotype pairs (WT vs IIB^-/-, WT vs IIA^-/-, IIB^-/- vs IIA^-/-; all p < 0.0001). The genotype rank order of the group means (±SEM) is IIB^-/- > WT > IIA^-/- (0.967 ± 0.036; 0.713 ± 0.021; 0.427 ± 0.017; n=69, 65, 59). For electrophysiological minis (mEPSP amplitude, Fig. 8F), K–S tests likewise show significant differences for the same comparisons (all p < 0.0001) with D statistics of 0.1854, 0.3647, and 0.4043 (WT vs IIB^-/-, WT vs IIA^-/-, IIB^-/- vs IIA^-/-, respectively). Group means (±SEM) again follow IIB^-/- > WT > IIA^-/- (0.824 ± 0.017 mV; 0.636 ± 0.015 mV; 0.383 ± 0.007 mV; n=41 each). These K–S results demonstrate identical significance and rank order across modalities, supporting our conclusion that SynapGCaMP8m resolves physiologically relevant quantal differences with sensitivity comparable to electrophysiology.

References

Blum, Ian D., Mehmet F. Keleş, El-Sayed Baz, Emily Han, Kristen Park, Skylar Luu, Habon Issa, Matt Brown, Margaret C. W. Ho, Masashi Tabuchi, Sha Liu, and Mark N. Wu. 2021. 'Astroglial Calcium Signaling Encodes Sleep Need in Drosophila', Current Biology, 31: 150-62.e7.

Chen, Y., and L. M. Huang. 2017. 'A simple and fast method to image calcium activity of neurons from intact dorsal root ganglia using fluorescent chemical Ca(2+) indicators', Mol Pain, 13: 1744806917748051.

Giovannucci, Andrea, Johannes Friedrich, Pat Gunn, Jérémie Kalfon, Brandon L. Brown, Sue Ann Koay, Jiannis Taxidis, Farzaneh Najafi, Jeffrey L. Gauthier, Pengcheng Zhou, Baljit S. Khakh, David W. Tank, Dmitri B. Chklovskii, and Eftychios A. Pnevmatikakis. 2019. 'CaImAn an open source tool for scalable calcium imaging data analysis', eLife, 8: e38173.

Müller, M., K. S. Liu, S. J. Sigrist, and G. W. Davis. 2012. 'RIM controls homeostatic plasticity through modulation of the readily-releasable vesicle pool', J Neurosci, 32: 16574-85.

Wu, Yifan, Keimpe Wierda, Katlijn Vints, Yu-Chun Huang, Valerie Uytterhoeven, Sahil Loomba, Fran Laenen, Marieke Hoekstra, Miranda C. Dyson, Sheng Huang, Chengji Piao, Jiawen Chen, Sambashiva Banala, Chien-Chun Chen, El-Sayed Baz, Luke Lavis, Dion Dickman, Natalia V. Gounko, Stephan Sigrist, Patrik Verstreken, and Sha Liu. 2025. 'Presynaptic Release Probability Determines the Need for Sleep', bioRxiv: 2025.10.16.682770.

Xing, Xiaomin, and Chun-Fang Wu. 2018. 'Unraveling Synaptic GCaMP Signals: Differential Excitability and Clearance Mechanisms Underlying Distinct Ca²⁺ Dynamics in Tonic and Phasic Excitatory, and Aminergic Modulatory Motor Terminals in Drosophila', eneuro, 5: ENEURO.0362-17.2018.

eLife Assessment

This important study combines real-time key point tracking with transdermal activation of sensory neurons as a general technique to explore how somatosensory stimulation impacts behavior in freely moving mice. After addressing concerns about classification of the behavioral responses to nociceptor stimulation, the authors now convincingly demonstrate a state-dependence in the behavioral response following nociceptor activation, highlighting how their real-time optogenetic stimulation capabilities can yield new insights into complex sensory processing. This work is a technological advancement that will be of interest to a broad readership, in particular labs studying somatosensation, enabling rigorous investigation of behaviors that were previously difficult or impossible to study.

Reviewer #1 (Public review):

Summary:

This study presents a system for delivering precisely controlled cutaneous stimuli to freely moving mice by coupling markerless real-time tracking to transdermal optogenetic stimulation, using the tracking signal to direct a laser via galvanometer mirrors. The principal claims are that the system achieves sub-mm targeting accuracy with a latency of <100 ms. Due to the nature of mouse gait, this enables accurate targeting of forepaws even when mice are moving.

Strengths:

The study is of high quality and the evidence for the claims is convincing. There is increasing focus in neurobiology in studying neural function in freely moving animals, engaged in natural behaviour. However, a substantial challenge is how to deliver controlled stimuli to sense organs under such conditions. The system presented here constitutes notable progress towards such experiments in the somatosensory system and is, in my view, a highly significant development that will be of interest to a broad readership.

My comments on the original submission have been fully addressed.