Reviewer #2 (Public Review):

Summary:

This study examined the role of a prefrontal cortex cell type in active avoidance behavior. The authors conduct a series of behavioral experiments incorporating fiber photometry and optogenetic silencing. The results indicate that prefrontal parvalbumin (PV) neurons play a permissive role in performing signaled active avoidance learning, for which details are sorely lacking. Notably, infralimbic parvalbumin activity resolves incompatible defensive responses to threat by suppressing conditional freezing in order to permit active instrumental controlling responses. The overall findings provide a significant contribution to our understanding of mechanisms that support aversively motivated instrumental learning and may provide insight into both stress vulnerability and resilience processes.

Strengths:

The writing and presentation of data is clear. The authors use a number of temporally-relevant methods and analyses that identify a novel prefrontal mechanism in resolving the conflict between competing actions (freezing vs escape avoidance). The authors conduct an extensive number of experiments to demonstrate that the uncovered prefrontal mechanism is selective for the initiation of avoidance under threat circumstances, not reward settings or general features of movement.

Weaknesses:

The study exclusively focuses on parvalbumin cells, thus questions remain whether the present findings are specific to parvalbumin or applicable to other prefrontal interneuron subtypes. The exact mechanisms that coordinate infralimbic parvalbumin cell activity and threat avoidance behavior are not explored.

Reviewer #3 (Public Review):

Summary:

Here the authors study the role of parvalbumin (PV) expressing neurons in the ventromedial prefrontal cortex (vMPFC) of mice in active avoidance behavior using fiber photometry and optogenetic inhibition.

Strengths:

The methods are appropriate, the experiments are well done, and the results are all consistent with the conceptual model in which vmPFC PV neurons inhibit freezing to enable avoidance movements. There are good controls to rule out a role for cue offset in triggering changes in PV neuron activity, or for a nonspecific role of vmPFC PV neurons in movement initiation.

Weaknesses:

Although potential mechanisms, i.e., the impact of PV neuron activity on the broader circuit, are discussed, they are not directly examined here. There is some discordance between changes in neural activity and behavior: in Figure 4C, the relationship between PV neuron activity and movement emerges almost immediately during learning, but successful active avoidance emerges much more gradually. Again, this is discussed and plausible explanations for this discrepancy are provided.

Reviewer #1 (Public Review):

Summary:

Peterson et al., present a series of experiments in which the Pavlovian performance (i.e. time spent at a food cup/port) of male and female rats is assessed in various tasks in which context/cue/outcome relationships are altered. The authors find no sex differences in context-irrelevant tasks, and no such differences in tasks in which the context signals that different cues will earn different outcomes. They do find sex differences, however, when a single outcome is given and context cues must be used to ascertain which cue will be rewarded with that outcome (Ctx-dep O1 task). Specifically, they find that males acquired the task faster, but that once acquired, performance of the task was more resilient in female rats against exposures to a stressor. Finally, they show that these sex differences are reflected in differential rates of c-fos expression in all three subregions of rat OFC, medial, lateral and ventral, in the sense that it is higher in females than males, and only in the animals subject to the Ctx-dep O1 task in which sex differences were observed.

Strengths:

• Well written<br /> • Experiments elegantly designed<br /> • Robust statistics<br /> • Behaviour is the main feature of this manuscript, rather than any flashy techniques or fashionable lab methodologies, and luckily the behaviour is done really well.<br /> • For the most part I think the conclusions were well supported, although I do have some slightly different interpretations to the authors in places.

Weaknesses:

The authors have done an excellent job of addressing all previous weaknesses. I have no further comments.

Reviewer #2 (Public Review):

Summary:

A bidirectional occasion-setting design is used to examine sex differences in the contextual modulation of reward-related behaviour. It is shown that females are slower to acquire contextual control over cue-evoked reward seeking. However, once established, the contextual control over behaviour was more robust in female rats (i.e., less within-session variability and greater resistance to stress) and this was also associated with increased OFC activation.

Strengths:

The authors use sophisticated behavioural paradigms to study the hierarchical contextual modulation of behaviour. The behavioural controls are particularly impressive and do, to some extent, support the specificity of the conclusions. The analyses of the behavioural data are also elegant, thoughtful, and rigorous.

Weaknesses:

The authors have addressed the major weaknesses that I identified in a previous review.

Reviewer #3 (Public Review):

Summary:

This manuscript reports an experiment that compared groups of rats acquisition and performance of a Pavlovian bi-conditional discrimination, in which the presence of one cue, A, signals that the presentation of one CS, X, will be followed by a reinforcer and a second CS, Y, will be nonreinforced. Periods of cue A alternated with periods of cue B, which signaled the opposite relationship, cue X is nonreinforced and cue Y is reinforced. This is a conditional discrimination problem in which the rats learned to approach the food cup in the presence of each CS conditional on the presence of the third background cue. The comparison groups consisted of the same conditional discrimination with the exception that each CS was paired with a different reinforcer. This makes the problem easier to solve as the background is now priming a differential outcome. A third group received simple discrimination training of X reinforced and Y nonreinforced in cues A and B, and the final group were trained with X and Y reinforced on half the trials (no discrimination). The results were clear that the latter two discrimination learning procedures resulted in rapid learning in comparison to the first. Rats required about 3 times as many 4-session blocks to acquire the bi-conditional discrimination than the other two discrimination groups. Within the biconditional discrimination group, female and male rats spent the same amount of time in the food cup during the rewarded CS, but females spent more time in the food cup during CS- than males. The authors interpret this as a deficit in discrimination performance in females on this task and use a measure that exaggerates the difference in CS+ and CS_ responding (a discrimination ratio) to support their point. When tested after acute restraint stress, the male rats spent less time in the food cup during the reinforced CS in comparison to the female rats, but did not lose discrimination performance entirely. The was also some evidence of more fos positive cells in the orbitofrontal cortex in females. Overall, I think the authors were successful in documenting performance on the biconditional discrimination task, showing that it is more difficult to perform than other discriminations is valuable and consistent with the proposal that accurate performance requires encoding of conditional information (which the authors refer to as "context"). There is evidence that female rats spend more time in the food cup during CS-, but this I hesitate to agree that this is an important sex difference. There is no cost to spending more time in the food cup during CS- and they spend much less time there than during CS+. Males and females also did not differ in their CS+ responding, suggesting similar levels of learning, A number of factors could contribute to more food cup time in CS-, such as smaller body size and more locomotor activity. The number of food cup entries during CS+ and CS- was not reported here. Nevertheless, I think the manuscript will make a useful contribution to the field and hopefully lead readers to follow up on these types of tasks. One area for development would be to test the associative properties of the cues controlling the conditional discrimination, can they be shown to have the properties of Pavlovian occasion setting stimuli? Such work would strengthen the justification/rationale for using the term "context" and "occasion setter" to refer to these stimuli in this task in the way the authors do in this paper.

Strengths:

Nicely designed and conducted experiment.<br /> Documents performance difference by sex.

Weaknesses:

Overstatement of sex differences.<br /> Inconsistent, confusing, and possibly misleading use of terms to describe/imply the underlying processes contributing to performance.

Reviewer #1 (Public Review):

Summary:

This interesting study investigates the neurobiological mechanisms underlying the stable operation and maintenance of functionally appropriate rhythmic motor patterns during changing environmental conditions - temperature in this study in the crab Cancer borealis stomatogastric neural pattern generating network producing the pyloric motor rhythm, which is naturally subjected to temperature perturbations over a substantial range. This study is relevant to the general problem that some rhythmic motor systems adjust to changing environmental conditions and state changes by increasing the cycle frequency in a smooth monotonic fashion while maintaining the relative timing of different network activity pattern phases that determine proper motor coordination. How this is achieved mechanistically in complex dynamic motor networks is not understood, particularly how the frequency and phase adjustments are achieved as conditions change while avoiding operational instabilities on different time scales. The authors specifically studied the contributions of the hyperpolarization-activated inward current (Ih), which is involved in rhythm control, to the adjustments of frequency and phases in the pyloric rhythmic pattern as the temperature was altered from 11 degrees C to 21 degrees C. They present strong evidence that this current is a critical biophysical feature in the ability of this system to adjust transiently and persistently to temperature perturbations appropriately. After blocking Ih in the pyloric network with cesium, the network was unable to reliably produce its characteristic rapid and smooth increase in the frequency of the triphasic rhythmic motor pattern in response to increasing temperature or its typical steady-state increase in frequency over this Q10 temperature range.

Strengths:

(1) The authors addressed this problem by technically rigorous experiments in the crab Cancer borealis stomatogastric ganglion (STG) in vitro, which readily allows for neuronal activity recording in a behaviorally and architecturally defined rhythmic neural circuit in conjunction with the application of blockers of Ih and synaptic receptors to disrupt circuit interactions. This approach is an effective way to experimentally investigate how complex rhythmic networks, at least in poikilotherms, mechanistically adjust to environmental perturbations such as temperature.

(2) While previous work demonstrated that Ih increases in pyloric neurons as temperature increases, the authors here establish that this increase is necessary for normal responses of STG neural activity to temperature, which consist of a smooth monotonic increase in the frequency of rhythmic activity with increasing temperature.

(3) The data shows that blocking Ih with cesium causes the frequency to transiently decrease ("jags") when the temperature increases and then increases after the temperature stabilizes at a steady state, revealing a non-monotonic frequency response to temperature perturbations.

(4) The authors dissect some of the underlying neuronal and circuit dynamics, presenting evidence that after blocking Ih, the non-monotonic jags in the frequency response are mediated by intrinsic properties of pacemaker neurons, while in the steady state, Ih determined the overall frequency change (i.e., temperature sensitivity) through network interactions.

(5) The authors' results highlight the existence of more complex dynamic responses to increasing temperature for the first time, suggesting a longer timescale process than previously recognized that may result from interactions between multiple channels and/or ion channel kinetics.

Weaknesses:

The involvement of Ih in achieving the frequency and phase adjustments as conditions change and allowing smooth transitions to avoid operational instabilities in other complex rhythmic motor netReviewer #2 (Public Review):

Summary:

Using the crustacean stomatogastric nervous system (STNS), the authors present an interesting study wherein the contribution of the Ih current to temperature-induced changes in the frequency of a rhythmically active neural circuit is evaluated. Ih is a hyperpolarization-activated cation current that depolarizes neurons. Under normal conditions, increasing the temperature of the STNS increases the frequency of the spontaneously active pyloric rhythm. Notably, under normal conditions, as temperature systematically increases, the concomitant increase in pyloric frequency is smooth (i.e., monotonic). By contrast, blocking Ih with extracellular cesium produces temperature-induced pyloric frequency changes that follow a characteristic sawtooth response (i.e., non-monotonic). That is, in cesium, increasing temperature initially results in a transient drop in pyloric frequency that then stabilizes at a higher frequency. Thus, the authors conclude that Ih establishes a mechanism that ensures smooth changes in neural network frequency during environmental disturbances, a feature that likely bestows advantages to the animal's function.

The study describes several surprising and interesting findings. In general, the study's primary observation of the cesium-induced sawtooth response is remarkable. To my knowledge, this type of response has not yet been described in neurobiological systems, and I suspect that the unexpected response will be of interest to many readers.

At first glance, I had some concerns regarding the use of extracellular cesium to understand network phenomena. Yes, extracellular cesium blocks Ih. But extracellular cesium has also been shown to block astrocytic potassium channels, at least in mammalian systems (i.e., K-IR, PMID: 10601465), and such a blockade can elevate extracellular potassium. I was heartened to see that the authors acknowledge the non-specificity of cesium (lines 320-325) and I agree with the authors' contention that "a first approximation most of the effects seen here can likely be attributed to Cs+ block of Ih". Upon reflecting on the potential confound, I was also reassured to see that extracellular cesium alone does not increase pyloric frequency, an effect that might be expected if cesium indirectly raises [K+]outside. I suggest including that point in the discussion.

In summary, the authors present a solid investigation of a surprising biological phenomenon. In general, my comments are fairly minor. This is an interesting study.

Strengths:

A major strength of the study is the identification of an ionic conductance that mediates stable, monotonic changes in oscillatory frequency that accompany changes in the environment (i.e., temperature).

Weaknesses:

A potential experimental concern stems from the use of extracellular cesium to attribute network effects specifically to Ih. Previous work has shown that extracellular cesium also blocks inward-rectifier potassium channels expressed by astrocytes, and that such blockade may also elevate extracellular potassium, an action that generally depolarizes neurons. Notably, the authors address this potential concern in the discussion.works, for example, in homeotherms, is not established, so the present results may have limited general extrapolations.

Reviewer #2 (Public Review):

Summary:

Using the crustacean stomatogastric nervous system (STNS), the authors present an interesting study wherein the contribution of the Ih current to temperature-induced changes in the frequency of a rhythmically active neural circuit is evaluated. Ih is a hyperpolarization-activated cation current that depolarizes neurons. Under normal conditions, increasing the temperature of the STNS increases the frequency of the spontaneously active pyloric rhythm. Notably, under normal conditions, as temperature systematically increases, the concomitant increase in pyloric frequency is smooth (i.e., monotonic). By contrast, blocking Ih with extracellular cesium produces temperature-induced pyloric frequency changes that follow a characteristic sawtooth response (i.e., non-monotonic). That is, in cesium, increasing temperature initially results in a transient drop in pyloric frequency that then stabilizes at a higher frequency. Thus, the authors conclude that Ih establishes a mechanism that ensures smooth changes in neural network frequency during environmental disturbances, a feature that likely bestows advantages to the animal's function.

The study describes several surprising and interesting findings. In general, the study's primary observation of the cesium-induced sawtooth response is remarkable. To my knowledge, this type of response has not yet been described in neurobiological systems, and I suspect that the unexpected response will be of interest to many readers.

At first glance, I had some concerns regarding the use of extracellular cesium to understand network phenomena. Yes, extracellular cesium blocks Ih. But extracellular cesium has also been shown to block astrocytic potassium channels, at least in mammalian systems (i.e., K-IR, PMID: 10601465), and such a blockade can elevate extracellular potassium. I was heartened to see that the authors acknowledge the non-specificity of cesium (lines 320-325) and I agree with the authors' contention that "a first approximation most of the effects seen here can likely be attributed to Cs+ block of Ih". Upon reflecting on the potential confound, I was also reassured to see that extracellular cesium alone does not increase pyloric frequency, an effect that might be expected if cesium indirectly raises [K+]outside. I suggest including that point in the discussion.

In summary, the authors present a solid investigation of a surprising biological phenomenon. In general, my comments are fairly minor. This is an interesting study.

Strengths:

A major strength of the study is the identification of an ionic conductance that mediates stable, monotonic changes in oscillatory frequency that accompany changes in the environment (i.e., temperature).

Weaknesses:

A potential experimental concern stems from the use of extracellular cesium to attribute network effects specifically to Ih. Previous work has shown that extracellular cesium also blocks inward-rectifier potassium channels expressed by astrocytes, and that such blockade may also elevate extracellular potassium, an action that generally depolarizes neurons. Notably, the authors address this potential concern in the discussion.

Reviewer #3 (Public Review):

Summary:

This paper presents a systematic analylsis of the role of the hyperpolarization-activated inward current (the h current) in the response of the pyloric rhythm of the stomatogastric ganglion (STG) of the crab. In a detailed set of experiments, they analyze the effect of blocking h current with bath infusion of the h current blocker cesium (perfused as CsCl). They show interesting and reproducible effects that blockade of h current results in a period of frequency decrease after an upward step in temperature, followed by a slow increase in frequency.<br /> This contrasts with the normal temperature response that shows an increase in frequency with an increase in temperature without a downward "jag" in the frequency response. This is an important paper for showing the role of h current in stabilizing network dynamics in response to perturbations such as a temperature change.

The major effects are shown very clearly and convincingly in a range of experiments with combined intracellular recording from neurons during changes in temperature.

They also provide additional detailed analyses of the effect of picrotoxin on these changes, showing that most of the effects except for the loss of frequency increase, appear to indicate that these effects are due to the role of h current in the pacemaker neurons PD.

Weaknesses :

I know the Marder lab has detailed models of the pyloric rhythm. I am not saying they have to add modeling to this already extensive and detailed paper, but it would be useful to know how much of these temperature effects have been modeled successfully and which ones have never been shown in the models.

They describe the ionic mechanism for the decrease and increase in frequency as a difference in temperature sensitivity of different components of the A current, but it seems like it is also a function of the time course of the response to change in temperature (i.e. the different components could have the same final effect of temperature but show a different time course of the change). They could mention any known data about the mechanism for how temperature is altering these channel kinetics and whether this indicates a change in time course of response to the same temperature, or a difference in actual steady-state temperature sensitivity.

Reviewer #1 (Public Review):

This work successfully identified and validated TRLs in hepatic metastatic uveal melanoma, providing new horizons for enhanced immunotherapy. Uveal melanoma is a highly metastatic cancer that, unlike cutaneous melanoma, has a limited effect on immune checkpoint responses, and thus there is a lack of formal clinical treatment for metastatic UM. In this manuscript, the authors described the immune microenvironmental profile of hepatic metastatic uveal melanoma by sc-RNAseq, TCR-seq, and PDX models. Firstly, they identified and defined the phenotypes of tumor-reactive T lymphocytes (TRLs). Moreover, they validated the activity of TILs by in vivo PDX modeling as well as in vitro co-culture of 3D tumorsphere cultures and autologous TILs. Additionally, the authors found that TRLs are mainly derived from depleted and late-activated T cells, which recognize melanoma antigens and tumor-specific antigens. Most importantly, they identified TRLs-associated phenotypes, which provide new avenues for targeting expanded T cells to improve cellular and immune checkpoint immunotherapy.

Comments on revised manuscript

The revised manuscript has addressed all my concerns.

Reviewer #2 (Public Review):

Summary:

The study's goal is to characterize and validate tumor-reactive T cells in liver metastases of uveal melanoma (UM), which could contribute to enhancing immunotherapy for these patients. The authors used single-cell RNA and TCR sequencing to find potential tumor-reactive T cells and then used patient-derived xenograft (PDX) models and tumor sphere cultures for functional analysis. They discovered that tumor-reactive T cells exist in activated/exhausted T cell subsets and in cytotoxic effector cells. Functional experiments with isolated TILs show that they are capable of killing UM cells in vivo and ex vivo.

Strengths:

The study highlights the potential of using single-cell sequencing and functional analysis to identify T cells that can be useful for cell therapy and marker selection in UM treatment. This is important and novel as conventional immune checkpoint therapies are not highly effective in treating UM. Additionally, the study's strength lies in its validation of findings through functional assays, which underscores the clinical relevance of the research.

Weaknesses:

The manuscript may pose challenges for individuals with limited knowledge of single-cell analysis and immunology markers, making it less accessible to a broader audience.

Reviewer #1 (Public Review):

Summary:

This is large-scale genomics and transcriptomics study of the epidemic community-acquired methicillin-resistant S. aureus clone USA300, designed to identify core genome mutations that drove the emergence of the clone. It used publicly available datasets and a combination of genome-wide association studies (GWAS) and independent principal-component analysis (ICA) of RNA-seq profiles to compare USA300 versus non-USA300 within clonal complex 8. By overlapping the analyses the authors identified a 38bp deletion upstream of the iron-scavenging surface-protein gene isdH that was both significantly associated with the USA300 lineage and with a decreased transcription of the gene.

Strengths:

Several genomic studies have investigated genomic factors driving the emergence of successful S. aureus clones, in particular USA300. These studies have often focussed on acquisition of key accessory genes or have focussed on a small number of strains. This study makes a smart use of publicly available repositories to leverage the sample size of the analysis and identify new genomics markers of USA300 success.

The approach of combining large-scale genomics and transcriptomics analysis is powerful, as it allows to make some inferences on the impact of the mutations. This is particular important for mutations in intergenic regions, whose functional impact is often uncertain.

The statistical genomics approaches are elegant and state-of-the-art and can be easily applied to other contexts or pathogens.

Weaknesses:

The main weakness of this work is that these data don't allow a casual inference on the role of isdH in driving the emergence of USA300. It is of course impossible to prove which mutation or gene drove the success of the clone, however, experimental data would have strengthen the conclusions of the authors in my opinion.

Another limitation of this approach is that the approach taken here doesn't allow to make any conclusions on the adaptive role of the isdH mutation. In other words, it is still possible that the mutation is just a marker of USA300 success, due to other factors such as PVL, ACMI or the SCCmecIVa. This is because by its nature this analysis is heavy influenced by population structure. Usually, GWAS is applied to find genetic loci that are associated with a phenotype and are independent of the underlying population structure. Here, authors are using GWAS to find loci that are associated with a lineage. In other words, they are simply running a univariate analysis (likely a logistic regression) between genetic loci and the lineage without any correction for population structure, since population structure is the outcome. Therefore, this approach can't be applied to most phenotype-genotype studies where correction for population structure is critical.

Finally, the approach used is complex and not easily reproduced to another dataset. Although I like DBGWAS and find the network analysis elegant, I would be interested in seeing how a simpler GWAS tool like Pyseer would perform.

Joint Public Review:

The present study explored the principles that allow cells to maintain complex subcellular proteinaceous structures despite the limited lifetimes of the individual protein components. This is particularly critical in the case of neurons, where the size and protein composition of synapses define synaptic strength and encode memory.

PSD95 is an abundant synapse protein that acts as a scaffold in the recruitment of transmitter receptors and other signaling proteins and is required for proper memory formation. The authors used super-resolution microscopy to study PSD95 super-complexes isolated from the brains of mice expressing tagged PSD variants (Halo-Tag, mEos, GFP). Their results show compellingly that a large fraction (~25%) of super-complexes contains two PSD95 copies about 13 nm apart, that there is substantial turnover of PSD95 proteins in super-complexes over a period of seven days, and that ~5-20% of the super-complexes contain new and old PSD95 molecules. This percentage is higher in synaptic fractions as compared to total brain lysates, and highest in isocortex samples (~20%). These important findings support the notion put forward by Crick that sequential subunit replacement gives synaptic super-complexes long lifetimes and thus aids in memory maintenance. Overall, this is very interesting, providing key insights into how synaptic protein complexes are formed and maintained. On the other hand, the actual role of these PSD95 super-complexes in long-term memory storage remains unknown.

Strengths

(1) The study employed an appropriate and validated methodology.

(2) Large numbers of PSD95 super-complexes from three different mouse models were imaged and analyzed, providing adequately powered sample sizes.

(3) State-of-the-art super-resolution imaging techniques (PALM and MINFLUX) were used, providing a robust, high-quality, cross-validated analysis of PSD95 protein complexes that is useful for the community.

(4) The result that PSD95 proteins in dimeric complexes are on average 12.7 nm apart is useful and has implications for studies on the nanoscale organization of PSD95 at synapses.

(5) The finding that postsynaptic protein complexes can continue to exist while individual components are being renewed is important for our understanding of synapse maintenance and stability.

(6) The data on the turnover rate of PSD95 in super-complexes from different brain regions provide a first indication of potentially meaningful differences in the lifetime of super-complexes between brain regions.

Weaknesses

(1) The manuscript emphasizes the hypothesis that stable super-complexes, maintained through sequential replacement of subunits, might underlie the long-term storage of memory. While an interesting idea, this notion requires considerably more research. The presented experimental data are indeed consistent with this notion, but there is no evidence that these complexes are causally related to memory storage.

(2) Much of the presented work is performed on biochemically isolated protein complexes. The biochemical isolation procedures rely on physical disruption and detergents that are known to alter the composition and structure of complexes in certain cases. Thus, it remains unclear how the protein complexes described in this study relate to PSD95 complexes in intact synapses.

(3) Because not all GFP molecules mature and fold correctly in vitro and the PSD95-mEos mice used were heterozygous, the interpretation of the corresponding quantifications is not straightforward.

(4) It was not tested whether different numbers of PSD95 molecules per super-complex might contribute to different retention times of PSD95, e.g. in synaptic vs. total-forebrain super-complexes.

(5) The conclusion that the population of 'mixed' synapses is higher in the isocortex than in other brain regions is not supported by statistical analysis.

(6) The validity of conclusions regarding PSD95 degradation based on relative changes in the occurrence of SiR-Halo-positive puncta is limited.

Reviewer #1 (Public Review):

Summary:

Bowler et al. present a thoroughly tested system for modularized behavioral control of navigation-based experiments, particularly suited for pairing with 2-photon imaging but applicable to a variety of techniques. This system, which they name behaviorMate, represents a valuable contribution to the field. As the authors note, behavioral control paradigms vary widely across laboratories in terms of hardware and software utilized and often require specialized technical knowledge to make changes to these systems. Having a standardized, easy-to-implement, and flexible system that can be used by many groups is therefore highly desirable. This work will be of interest to systems neuroscientists looking to integrate flexible head-fixed behavioral control with neural data acquisition.

Strengths:

The present manuscript provides compelling evidence of the functionality and applicability of behaviorMate. The authors report benchmark tests for real-time update speed between the animal's movement and the behavioral control, on both the treadmill-based and virtual reality (VR) setups. Further, they nicely demonstrate and quantify reliable hippocampal place cell coding in both setups, using synchronized 2-photon imaging. This place cell characterization also provides a concrete comparison between the place cell properties observed in treadmill-based navigation vs. visual VR in a single study, which itself is a helpful contribution to the field.

Documentation for installing and operating behaviorMate is available via the authors' lab website and linked in the manuscript.

Weaknesses:

The following comments are mostly minor suggestions intended to add clarity to the paper and provide context for its significance.

(1) As VRMate (a component of behaviorMate) is written using Unity, what is the main advantage of using behaviorMate/VRMate compared to using Unity alone paired with Arduinos (e.g. Campbell et al. 2018), or compared to using an existing toolbox to interface with Unity (e.g. Alsbury-Nealy et al. 2022, DOI: 10.3758/s13428-021-01664-9)? For instance, one disadvantage of using Unity alone is that it requires programming in C# to code the task logic. It was not entirely clear whether VRMate circumvents this disadvantage somehow -- does it allow customization of task logic and scenery in the GUI? Does VRMate add other features and/or usability compared to Unity alone? It would be helpful if the authors could expand on this topic briefly.

(2) The section on "context lists", lines 163-186, seemed to describe an important component of the system, but this section was challenging to follow and readers may find the terminology confusing. Perhaps this section could benefit from an accompanying figure or flow chart, if these terms are important to understand.

(2a) Relatedly, "context" is used to refer to both when the animal enters a particular state in the task like a reward zone ("reward context", line 447) and also to describe a set of characteristics of an environment (Figure 3G), akin to how "context" is often used in the navigation literature. To avoid confusion, one possibility would be to use "environment" instead of "context" in Figure 3G, and/or consider using a word like "state" instead of "context" when referring to the activation of different stimuli.

(3) Given the authors' goal of providing a system that is easily synchronizable with neural data acquisition, especially with 2-photon imaging, I wonder if they could expand on the following features:

(3a) The authors mention that behaviorMate can send a TTL to trigger scanning on the 2P scope (line 202), which is a very useful feature. Can it also easily generate a TTL for each frame of the VR display and/or each sample of the animal's movement? Such TTLs can be critical for synchronizing the imaging with behavior and accounting for variability in the VR frame rate or sampling rate.

(3b) Is there a limit to the number of I/O ports on the system? This might be worth explicitly mentioning.

(3c) In the VR version, if each display is run by a separate Android computer, is there any risk of clock drift between displays? Or is this circumvented by centralized control of the rendering onset via the "real-time computer"?

Reviewer #2 (Public Review):

Summary:

The authors present behaviorMate, an open-source behavior recording and control system including a central GUI and compatible treadmill and display components. Notably, the system utilizes the "Intranet of things" scheme and the components communicate through a local network, making the system modular, which in turn allows user to easily configure the setup to suit their experimental needs. Overall, behaviorMate is a valuable resource for researchers performing head-fixed imaging studies, as the commercial alternatives are often expensive and inflexible to modify.

Strengths and Weaknesses:

The manuscript presents two major utilities of behaviorMate: (1) as an open-source alternative to commercial behavior apparatus for head-fixed imaging studies, and (2) as a set of generic schema and communication protocols that allows the users to incorporate arbitrary recording and stimulation devices during a head-fixed imaging experiment. I found the first point well-supported and demonstrated in the manuscript. Indeed, the documentation, BOM, CAD files, circuit design, source, and compiled software, along with the manuscript, create an invaluable resource for neuroscience researchers looking to set up a budget-friendly VR and head-fixed imaging rig. Some features of behaviorMate, including the computer vision-based calibration of the treadmill, and the decentralized, Android-based display devices, are very innovative approaches and can be quite useful in practical settings. However, regarding the second point, my concern is that there is not adequate documentation and design flexibility to allow the users to incorporate arbitrary hardware into the system. In particular:

(1) The central controlling logic is coupled with GUI and an event loop, without a documented plugin system. It's not clear whether arbitrary code can be executed together with the GUI, hence it's not clear how much the functionality of the GUI can be easily extended without substantial change to the source code of the GUI. For example, if the user wants to perform custom real-time analysis on the behavior data (potentially for closed-loop stimulation), it's not clear how to easily incorporate the analysis into the main GUI/control program.

(2) The JSON messaging protocol lacks API documentation. It's not clear what the exact syntax is, supported key/value pairs, and expected response/behavior of the JSON messages. Hence, it's not clear how to develop new hardware that can communicate with the behaviorMate system.

(3) It seems the existing control hardware and the JSON messaging only support GPIO/TTL types of input/output, which limits the applicability of the system to more complicated sensor/controller hardware. The authors mentioned that hardware like Arduino natively supports serial protocols like I2C or SPI, but it's not clear how they are handled and translated to JSON messages.

Additionally, because it's unclear how easy to incorporate arbitrary hardware with behaviorMate, the "Intranet of things" approach seems to lose attraction. Since currently, the manuscript focuses mainly on a specific set of hardware designed for a specific type of experiment, it's not clear what are the advantages of implementing communication over a local network as opposed to the typical connections using USB.

In summary, the manuscript presents a well-developed open-source system for head-fixed imaging experiments with innovative features. The project is a very valuable resource to the neuroscience community. However, some claims in the manuscript regarding the extensibility of the system and protocol may require further development and demonstration.

Reviewer #3 (Public Review):

In this work, the authors present an open-source system called behaviourMate for acquiring data related to animal behavior. The temporal alignment of recorded parameters across various devices is highlighted as crucial to avoid delays caused by electronics dependencies. This system not only addresses this issue but also offers an adaptable solution for VR setups. Given the significance of well-designed open-source platforms, this paper holds importance.

Advantages of behaviorMate:

The cost-effectiveness of the system provided.

The reliability of PCBs compared to custom-made systems.

Open-source nature for easy setup.

Plug & Play feature requiring no coding experience for optimizing experiment performance (only text-based Json files, 'context List' required for editing).

Points to clarify:

While using UDP for data transmission can enhance speed, it is thought that it lacks reliability. Are there error-checking mechanisms in place to ensure reliable communication, given its criticality alongside speed?

Considering this year's price policy changes in Unity, could this impact the system's operations?

Also, does the Arduino offer sufficient precision for ephys recording, particularly with a 10ms check?

Could you clarify the purpose of the Sync Pulse? In line 291, it suggests additional cues (potentially represented by the Sync Pulse) are needed to align the treadmill screens, which appear to be directed towards the Real-Time computer. Given that event alignment occurs in the GPIO, the connection of the Sync Pulse to the Real-Time Controller in Figure 1 seems confusing. Additionally, why is there a separate circuit for the treadmill that connects to the UI computer instead of the GPIO? It might be beneficial to elaborate on the rationale behind this decision in line 260. Moreover, should scenarios involving pupil and body camera recordings connect to the Analog input in the PCB or the real-time computer for optimal data handling and processing?

Given that all references, as far as I can see, come from the same lab, are there other labs capable of implementing this system at a similar optimal level?

Reviewer #1 (Public Review):

Summary:

Qin and colleagues analysed data from the Human Connectome Project on four right-handed subgroups with different gyrification patterns in Heschl's gyrus. Based on these groups, the authors highlight the structure-function relationship of planum temporale asymmetry in lateralised language processing at the group level and next at the individual level. In particular, the authors propose that especially microstructural asymmetries are related to functional auditory language asymmetries in the planum temporale.

Strengths:

The study is interesting because of an ongoing and long-standing debate about the relationship between structural and functional brain asymmetries, and in particular whether structural brain asymmetries can be seen as markers of functional language brain lateralisation.

In this debate, the relationship between Heschl's gyrus asymmetry and planum temporale asymmetry is rare and therefore valuable here. A large sample size and inter-rater reliability support the findings.

Weaknesses:

In this case of multiple brain measures, it would be important to provide the reader with some sort of effect size (e.g. Cohen's d) to help interpret the results. In addition, the authors highlight the microstructural results in spite of the macrostructural results. However, the macrostructural surface results are also strong. I would suggest either reducing the emphasis on micro vs macrostructural results or adding information to justify the microstructural importance.

Reviewer #2 (Public Review):

Summary:

The authors assessed the link between structural and functional lateralization in area PT, one of the brain areas that is known to exhibit strong structural lateralization, and which is known to be implicated in speech processing. Importantly, they included the sulcal configuration of Heschl's gyrus (HG), presenting either as a single or duplicated HG, in their analysis. They found several significant associations between microstructural indices and task-based functional lateralization, some of which depended on the sulcal configuration.

Strengths:

A clear strength is the large sample size (n=907), an openly available database, and the fact that HG morphology was manually classified in each individual. This allows for robust statistical testing of the effects across morphological categories, which is not often seen in the literature.

Weaknesses:

- Unfortunately, no left-handers were included in the study. It would have been a valuable addition to the literature, to study the effect of handedness on the observed associations, as many previous studies on this topic were not adequately powered. The fact that only right-handers were studied should be pointed out clearly in the introduction or even the abstract.

- The tasks to quantify functional lateralization were not specifically designed to pick up lateralization. In the interest of the sample size, it is understandable that the authors used the available HCP-task-battery results, however, it would have been feasible to access another dataset for validation. A targeted subset of results, concerning for example the relationship between sulcal morphology and task-based functional lateralization, could be re-assessed using other open-access fMRI datasets.

- The study is mainly descriptive and the general discussion of the findings in the larger context of brain lateralization comes a bit short. For example, are the observed effects in line with what we know from other 'language-relevant' areas? What could be the putative mechanisms that give rise to functional lateralization based on the microstructural markers observed? And which mechanisms might be underlying the formation of a duplicated HG?

Reviewer #1 (Public Review):

Summary:

In this study, the authors trained a variational autoencoder (VAE) to create a high-dimensional "voice latent space" (VLS) using extensive voice samples, and analyzed how this space corresponds to brain activity through fMRI studies focusing on the temporal voice areas (TVAs). Their analyses included encoding and decoding techniques, as well as representational similarity analysis (RSA), which showed that the VLS could effectively map onto and predict brain activity patterns, allowing for the reconstruction of voice stimuli that preserve key aspects of speaker identity.

Strengths:

This paper is well-written and easy to follow. Most of the methods and results were clearly described. The authors combined a variety of analytical methods in neuroimaging studies, including encoding, decoding, and RSA. In addition to commonly used DNN encoding analysis, the authors performed DNN decoding and resynthesized the stimuli using VAE decoders. Furthermore, in addition to machine learning classifiers, the authors also included human behavioral tests to evaluate the reconstruction performance.

Weaknesses:

This manuscript presents a variational autoencoder (VAE) to evaluate voice identity representations from brain recordings. However, the study's scope is limited by testing only one model, leaving unclear how generalizable or impactful the findings are. The preservation of identity-related information in the voice latent space (VLS) is expected, given the VAE model's design to reconstruct original vocal stimuli. Nonetheless, the study lacks a deeper investigation into what specific aspects of auditory coding these latent dimensions represent. The results in Figure 1c-e merely tested a very limited set of speech features. Moreover, there is no analysis of how these features and the whole VAE model perform in standard speech tasks like speech recognition or phoneme recognition. It is not clear what kind of computations the VAE model presented in this work is capable of. Inclusion of comparisons with state-of-the-art unsupervised or self-supervised speech models known for their alignment with auditory cortical responses, such as Wav2Vec2, HuBERT, and Whisper, would strengthen the validation of the VAE model and provide insights into its relative capabilities and limitations.

The claim that the VLS outperforms a linear model (LIN) in decoding tasks does not significantly advance our understanding of the underlying brain representations. Given the complexity of auditory processing, it is unsurprising that a nonlinear model would outperform a simpler linear counterpart. The study could be improved by incorporating a comparative analysis with alternative models that differ in architecture, computational strategies, or training methods. Such comparisons could elucidate specific features or capabilities of the VLS, offering a more nuanced understanding of its effectiveness and the computational principles it embodies. This approach would allow the authors to test specific hypotheses about how different aspects of the model contribute to its performance, providing a clearer picture of the shared coding in VLS and the brain.

The manuscript overlooks some crucial alternative explanations for the discriminant representation of vocal identity. For instance, the discriminant representation of vocal identity can be either a higher-level abstract representation or a lower-level coding of pitch height. Prior studies using fMRI and ECoG have identified both types of representation within the superior temporal gyrus (STG) (e.g., Tang et al., Science 2017; Feng et al., NeuroImage 2021). Additionally, the methodology does not clarify whether the stimuli from different speakers contained identical speech content. If the speech content varied across speakers, the approach of averaging trials to obtain a mean vector for each speaker-the "identity-based analysis"-may not adequately control for confounding acoustic-phonetic features. Notably, the principal component 2 (PC2) in Figure 1b appears to correlate with absolute pitch height, suggesting that some aspects of the model's effectiveness might be attributed to simpler acoustic properties rather than complex identity-specific information.

Methodologically, there are issues that warrant attention. In characterizing the autoencoder latent space, the authors initialized logistic regression classifiers 100 times and calculated the t-statistics using degrees of freedom (df) of 99. Given that logistic regression is a convex optimization problem typically converging to a global optimum, these multiple initializations of the classifier were likely not entirely independent. Consequently, the reported degrees of freedom and the effect size estimates might not accurately reflect the true variability and independence of the classifier outcomes. A more careful evaluation of these aspects is necessary to ensure the statistical robustness of the results.

Reviewer #2 (Public Review):

Summary:

Lamothe et al. collected fMRI responses to many voice stimuli in 3 subjects. The authors trained two different autoencoders on voice audio samples and predicted latent space embeddings from the fMRI responses, allowing the voice spectrograms to be reconstructed. The degree to which reconstructions from different auditory ROIs correctly represented speaker identity, gender, or age was assessed by machine classification and human listener evaluations. Complementing this, the representational content was also assessed using representational similarity analysis. The results broadly concur with the notion that temporal voice areas are sensitive to different types of categorical voice information.

Strengths:

The single-subject approach that allows thousands of responses to unique stimuli to be recorded and analyzed is powerful. The idea of using this approach to probe cortical voice representations is strong and the experiment is technically solid.

Weaknesses:

The paper could benefit from more discussion of the assumptions behind the reconstruction analyses and the conclusions it allows. The authors write that reconstruction of a stimulus from brain responses represents 'a robust test of the adequacy of models of brain activity' (L138). I concur that stimulus reconstruction is useful for evaluating the nature of representations, but the notion that they can test the adequacy of the specific autoencoder presented here as a model of brain activity should be discussed at more length. Natural sounds are correlated in many feature dimensions and can therefore be summarized in several ways, and similar information can be read out from different model representations. Models trained to reconstruct natural stimuli can exploit many correlated features and it is quite possible that very different models based on different features can be used for similar reconstructions. Reconstructability does not by itself imply that the model is an accurate brain model. Non-linear networks trained on natural stimuli are arguably not tested in the same rigorous manner as models built to explicitly account for computations (they can generate predictions and experiments can be designed to test those predictions). While it is true that there is increasing evidence that neural network embeddings can predict brain data well, it is still a matter of debate whether good predictability by itself qualifies DNNs as 'plausible computational models for investigating brain processes' (L72). This concern is amplified in the context of decoding and naturalistic stimuli where many correlated features can be represented in many ways. It is unclear how much the results hinge on the specificities of the specific autoencoder architectures used. For instance, it would be useful to know the motivations for why the specific VAE used here should constitute a good model for probing neural voice representations.

Relatedly, it is not clear how VAEs as generative models are motivated as computational models of voice representations in the brain. The task of voice areas in the brain is not to generate voice stimuli but to discriminate and extract information. The task of reconstructing an input spectrogram is perhaps useful for probing information content, but discriminative models, e.g., trained on the task of discriminating voices, would seem more obvious candidates. Why not include discriminatively trained models for comparison?

The autoencoder learns a mapping from latent space to well-formed voice spectrograms. Regularized regression then learns a mapping between this latent space and activity space. All reconstructions might sound 'natural', which simply means that the autoencoder works. It would be good to have a stronger test of how close the reconstructions are to the original stimulus. For instance, is the reconstruction the closest stimulus to the original in latent space coordinates out of using the experimental stimuli, or where does it rank? How do small changes in beta amplitudes impact the reconstruction? The effective dimensionality of the activity space could be estimated, e.g. by PCA of the voice samples' contrast maps, and it could then be estimated how the main directions in the activity space map to differences in latent space. It would be good to get a better grasp of the granularity of information that can be decoded/ reconstructed.

What can we make of the apparent trend that LIN is higher than VLS for identity classification (at least VLS does not outperform LIN)? A general argument of the paper seems to be that VLS is a better model of voice representations compared to LIN as a 'control' model. Then we would expect VLS to perform better on identity classification. The age and gender of a voice can likely be classified from many acoustic features that may not require dedicated voice processing.

The RDM results reported are significant only for some subjects and in some ROIs. This presumably means that results are not significant in the other subjects. Yet, the authors assert general conclusions (e.g. the VLS better explains RDM in TVA than LIN). An assumption typically made in single-subject studies (with large amounts of data in individual subjects) is that the effects observed and reported in papers are robust in individual subjects. More than one subject is usually included to hint that this is the case. This is an intriguing approach. However, reports of effects that are statistically significant in some subjects and some ROIs are difficult to interpret. This, in my view, runs contrary to the logic and leverage of the single-subject approach. Reporting results that are only significant in 1 out of 3 subjects and inferring general conclusions from this seems less convincing.

The first main finding is stated as being that '128 dimensions are sufficient to explain a sizeable portion of the brain activity' (L379). What qualifies this? From my understanding, only models of that dimensionality were tested. They explain a sizeable portion of brain activity, but it is difficult to follow what 'sizable' is without baseline models that estimate a prediction floor and ceiling. For instance, would autoencoders that reconstruct any spectrogram (not just voice) also predict a sizable portion of the measured activity? What happens to reconstruction results as the dimensionality is varied?

A second main finding is stated as being that the 'VLS outperforms the LIN space' (L381). It seems correct that the VAE yields more natural-sounding reconstructions, but this is a technical feature of the chosen autoencoding approach. That the VLS yields a 'more brain-like representational space' I assume refers to the RDM results where the RDM correlations were mainly significant in one subject. For classification, the performance of features from the reconstructions (age/ gender/ identity) gives results that seem more mixed, and it seems difficult to draw a general conclusion about the VLS being better. It is not clear that this general claim is well supported.

It is not clear why the RDM was not formed based on the 'stimulus GLM' betas. The 'identity GLM' is already biased towards identity and it would be stronger to show associations at the stimulus level.

Multiple comparisons were performed across ROIs, models, subjects, and features in the classification analyses, but it is not clear how correction for these multiple comparisons was implemented in the statistical tests on classification accuracies.

Risks of overfitting and bias are a recurrent challenge in stimulus reconstruction with fMRI. It would be good with more control analyses to ensure that this was not the case. For instance, how were the repeated test stimuli presented? Were they intermingled with the other stimuli used for training or presented in separate runs? If intermingled, then the training and test data would have been preprocessed together, which could compromise the test set. The reconstructions could be performed on responses from independent runs, preprocessed separately, as a control. This should include all preprocessing, for instance, estimating stimulus/identity GLMs on separately processed run pairs rather than across all runs. Also, it would be good to avoid detrending before GLM denoising (or at least testing its effects) as these can interact.

Reviewer #3 (Public Review):

Summary:

In this manuscript, Lamothe et al. sought to identify the neural substrates of voice identity in the human brain by correlating fMRI recordings with the latent space of a variational autoencoder (VAE) trained on voice spectrograms. They used encoding and decoding models, and showed that the "voice" latent space (VLS) of the VAE performs, in general, (slightly) better than a linear autoencoder's latent space. Additionally, they showed dissociations in the encoding of voice identity across the temporal voice areas.

Strengths:

- The geometry of the neural representations of voice identity has not been studied so far. Previous studies on the content of speech and faces in vision suggest that such geometry could exist. This study demonstrates this point systematically, leveraging a specifically trained variational autoencoder.

- The size of the voice dataset and the length of the fMRI recordings ensure that the findings are robust.

Weaknesses:

- Overall, the VLS is often only marginally better than the linear model across analysis, raising the question of whether the observed performance improvements are due to the higher number of parameters trained in the VAE, rather than the non-linearity itself. A fair comparison would necessitate that the number of parameters be maintained consistently across both models, at least as an additional verification step.

- The encoding and RSM results are quite different. This is unexpected, as similar embedding geometries between the VLS and the brain activations should be reflected by higher correlation values of the encoding model.

- The consistency across participants is not particularly high, for instance, S1 seemed to have demonstrated excellent performances, while S2 showed poor performance.

- An important control analysis would be to compare the decoding results with those obtained by a decoder operating directly on the latent spaces, in order to further highlight the interest of the non-linear transformations of the decoder model. Currently, it is unclear whether the non-linearity of the decoder improves the decoding performance, considering the poor resemblance between the VLS and brain-reconstructed spectrograms.

Reviewer #1 (Public Review):

Lu & Golomb combined EEG, artificial neural networks, and multivariate pattern analyses to examine how different visual variables are processed in the brain. The conclusions of the paper are mostly well supported, but some aspects of methods and data analysis would benefit from clarification and potential extensions.

The authors find that not only real-world size is represented in the brain (which was known), but both retinal size and real-world depth are represented, at different time points or latencies, which may reflect different stages of processing. Prior work has not been able to answer the question of real-world depth due to the stimuli used. The authors made this possible by assessing real-world depth and testing it with appropriate methodology, accounting for retinal and real-world size. The methodological approach combining behavior, RSA, and ANNs is creative and well thought out to appropriately assess the research questions, and the findings may be very compelling if backed up with some clarifications and further analyses.

The work will be of interest to experimental and computational vision scientists, as well as the broader computational cognitive neuroscience community as the methodology is of interest and the code is or will be made available. The work is important as it is currently not clear what the correspondence between many deep neural network models and the brain is, and this work pushes our knowledge forward on this front. Furthermore, the availability of methods and data will be useful for the scientific community.

Some analyses are incomplete, which would be improved if the authors showed analyses with other layers of the networks and various additional partial correlation analyses.

Clarity

(1) Partial correlations methods incomplete - it is not clear what is being partialled out in each analysis. It is possible to guess sometimes, but it is not entirely clear for each analysis. This is important as it is difficult to assess if the partial correlations are sensible/correct in each case. Also, the Figure 1 caption is short and unclear.

For example, ANN-EEG partial correlations - "Finally, we directly compared the timepoint-by-timepoint EEG neural RDMs and the ANN RDMs (Figure 3F). The early layer representations of both ResNet and CLIP were significantly correlated with early representations in the human brain" What is being partialled out? Figure 3F says partial correlation

Issues / open questions

(2) Semantic representations vs hypothesized (hyp) RDMs (real-world size, etc) - are the representations explained by variables in hyp RDMs or are there semantic representations over and above these? E.g., For ANN correlation with the brain, you could partial out hyp RDMs - and assess whether there is still semantic information left over, or is the variance explained by the hyp RDMs?

(3) Why only early and late layers? I can see how it's clearer to present the EEG results. However, the many layers in these networks are an opportunity - we can see how simple/complex linear/non-linear the transformation is over layers in these models. It would be very interesting and informative to see if the correlations do in fact linearly increase from early to later layers, or if the story is a bit more complex. If not in the main text, then at least in the supplement.

(4) Peak latency analysis - Estimating peaks per ppt is presumably noisy, so it seems important to show how reliable this is. One option is to find the bootstrapped mean latencies per subject.

(5) "Due to our calculations being at the object level, if there were more than one of the same objects in an image, we cropped the most complete one to get a more accurate retinal size. " Did EEG experimenters make sure everyone sat the same distance from the screen? and remain the same distance? This would also affect real-world depth measures.

Reviewer #2 (Public Review):

Summary:

This paper aims to test if neural representations of images of objects in the human brain contain a 'pure' dimension of real-world size that is independent of retinal size or perceived depth. To this end, they apply representational similarity analysis on EEG responses in 10 human subjects to a set of 200 images from a publicly available database (THINGS-EEG2), correlating pairwise distinctions in evoked activity between images with pairwise differences in human ratings of real-world size (from THINGS+). By partialling out correlations with metrics of retinal size and perceived depth from the resulting EEG correlation time courses, the paper claims to identify an independent representation of real-world size starting at 170 ms in the EEG signal. Further comparisons with artificial neural networks and language embeddings lead the authors to claim this correlation reflects a relatively 'high-level' and 'stable' neural representation.

Strengths:

- The paper features insightful figures/illustrations and clear figures.

- The limitations of prior work motivating the current study are clearly explained and seem reasonable (although the rationale for why using 'ecological' stimuli with backgrounds matters when studying real-world size could be made clearer; one could also argue the opposite, that to get a 'pure' representation of the real-world size of an 'object concept', one should actually show objects in isolation).

- The partial correlation analysis convincingly demonstrates how correlations between feature spaces can affect their correlations with EEG responses (and how taking into account these correlations can disentangle them better).

- The RSA analysis and associated statistical methods appear solid.

Weaknesses:

- The claim of methodological novelty is overblown. Comparing image metrics, behavioral measurements, and ANN activations against EEG using RSA is a commonly used approach to study neural object representations. The dataset size (200 test images from THINGS) is not particularly large, and neither is comparing pre-trained DNNs and language models, or using partial correlations.

- The claims also seem too broad given the fairly small set of RDMs that are used here (3 size metrics, 4 ANN layers, 1 Word2Vec RDM): there are many aspects of object processing not studied here, so it's not correct to say this study provides a 'detailed and clear characterization of the object processing process'.

- The paper lacks an analysis demonstrating the validity of the real-world depth measure, which is here computed from the other two metrics by simply dividing them. The rationale and logic of this metric is not clearly explained. Is it intended to reflect the hypothesized egocentric distance to the object in the image if the person had in fact been 'inside' the image? How do we know this is valid? It would be helpful if the authors provided a validation of this metric.

- Given that there is only 1 image/concept here, the factor of real-world size may be confounded with other things, such as semantic category (e.g. buildings vs. tools). While the comparison of the real-world size metric appears to be effectively disentangled from retinal size and (the author's metric of) depth here, there are still many other object properties that are likely correlated with real-world size and therefore will confound identifying a 'pure' representation of real-world size in EEG. This could be addressed by adding more hypothesis RDMs reflecting different aspects of the images that may correlate with real-world size.

- The choice of ANNs lacks a clear motivation. Why these two particular networks? Why pick only 2 somewhat arbitrary layers? If the goal is to identify more semantic representations using CLIP, the comparison between CLIP and vision-only ResNet should be done with models trained on the same training datasets (to exclude the effect of training dataset size & quality; cf Wang et al., 2023). This is necessary to substantiate the claims on page 19 which attributed the differences between models in terms of their EEG correlations to one of them being a 'visual model' vs. 'visual-semantic model'.

- The first part of the claim on page 22 based on Figure 4 'The above results reveal that real-world size emerges with later peak neural latencies and in the later layers of ANNs, regardless of image background information' is not valid since no EEG results for images without backgrounds are shown (only ANNs).

Appraisal of claims:

While the method shows useful and interesting patterns of results can be obtained by combining contrasting behavioral/image metrics, the lack of additional control models makes the evidence for the claimed unconfounded representation of real-world size in EEG responses incomplete.

Discussion of likely impact:

The paper is likely to impact the field by showcasing how using partial correlations in RSA is useful, rather than providing conclusive evidence regarding neural representations of objects and their sizes.

Additional context important to consider when interpreting this work:

- Page 20, the authors point out similarities of peak correlations between models ('Interestingly, the peaks of significant time windows for the EEG × HYP RSA also correspond with the peaks of the EEG × ANN RSA timecourse (Figure 3D,F)'. Although not explicitly stated, this seems to imply that they infer from this that the ANN-EEG correlation might be driven by their representation of the hypothesized feature spaces. However this does not follow: in EEG-image metric model comparisons it is very typical to see multiple peaks, for any type of model, this simply reflects specific time points in EEG at which visual inputs (images) yield distinctive EEG amplitudes (perhaps due to stereotypical waves of neural processing?), but one cannot infer the information being processed is the same. To investigate this, one could for example conduct variance partitioning or commonality analysis to see if there is variance at these specific time-points that is shared by a specific combination of the hypothesis and ANN feature spaces.

- Page 22 mentions 'The significant time-window (90-300ms) of similarity between Word2Vec RDM and EEG RDMs (Figure 5B) contained the significant time-window of EEG x real-world size representational similarity (Figure 3B)'. This is not particularly meaningful given that the Word2Vec correlation is significant for the entire EEG epoch (from the time-point of the signal 'arriving' in visual cortex around ~90 ms) and is thus much less temporally specific than the real-world size EEG correlation. Again a stronger test of whether Word2Vec indeed captures neural representations of real-world size could be to identify EEG time-points at which there are unique Word2Vec correlations that are not explained by either ResNet or CLIP, and see if those time-points share variance with the real-world size hypothesized RDM.

Reviewer #3 (Public Review):

The authors used an open EEG dataset of observers viewing real-world objects. Each object had a real-world size value (from human rankings), a retinal size value (measured from each image), and a scene depth value (inferred from the above). The authors combined the EEG and object measurements with extant, pre-trained models (a deep convolutional neural network, a multimodal ANN, and Word2vec) to assess the time course of processing object size (retinal and real-world) and depth. They found that depth was processed first, followed by retinal size, and then real-world size. The depth time course roughly corresponded to the visual ANNs, while the real-world size time course roughly corresponded to the more semantic models.

The time course result for the three object attributes is very clear and a novel contribution to the literature. However, the motivations for the ANNs could be better developed, the manuscript could better link to existing theories and literature, and the ANN analysis could be modernized. I have some suggestions for improving specific methods.

(1) Manuscript motivations<br /> The authors motivate the paper in several places by asking " whether biological and artificial systems represent object real-world size". This seems odd for a couple of reasons. Firstly, the brain must represent real-world size somehow, given that we can reason about this question. Second, given the large behavioral and fMRI literature on the topic, combined with the growing ANN literature, this seems like a foregone conclusion and undermines the novelty of this contribution.

While the introduction further promises to "also investigate possible mechanisms of object real-world size representations.", I was left wishing for more in this department. The authors report correlations between neural activity and object attributes, as well as between neural activity and ANNs. It would be nice to link the results to theories of object processing (e.g., a feedforward sweep, such as DiCarlo and colleagues have suggested, versus a reverse hierarchy, such as suggested by Hochstein, among others). What is semantic about real-world size, and where might this information come from? (Although you may have to expand beyond the posterior electrodes to do this analysis).

Finally, several places in the manuscript tout the "novel computational approach". This seems odd because the computational framework and pipeline have been the most common approach in cognitive computational neuroscience in the past 5-10 years.

(2) Suggestion: modernize the approach<br /> I was surprised that the computational models used in this manuscript were all 8-10 years old. Specifically, because there are now deep nets that more explicitly model the human brain (e.g., Cornet) as well as more sophisticated models of semantics (e.g., LLMs), I was left hoping that the authors had used more state-of-the-art models in the work. Moreover, the use of a single dCNN, a single multi-modal model, and a single word embedding model makes it difficult to generalize about visual, multimodal, and semantic features in general.

(3) Methodological considerations<br /> a) Validity of the real-world size measurement<br /> I was concerned about a few aspects of the real-world size rankings. First, I am trying to understand why the scale goes from 100-519. This seems very arbitrary; please clarify. Second, are we to assume that this scale is linear? Is this appropriate when real-world object size is best expressed on a log scale? Third, the authors provide "sand" as an example of the smallest real-world object. This is tricky because sand is more "stuff" than "thing", so I imagine it leaves observers wondering whether the experimenter intends a grain of sand or a sandy scene region. What is the variability in real-world size ratings? Might the variability also provide additional insights in this experiment?<br /> b) This work has no noise ceiling to establish how strong the model fits are, relative to the intrinsic noise of the data. I strongly suggest that these are included.

Reviewer #1 (Public Review):

Summary:

Kv2 subfamily potassium channels contribute to delayed rectifier currents in virtually all mammalian neurons and are encoded by two distinct types of subunits: Kv2 alpha subunits that have the capacity to form homomeric channels (Kv2.1 and Kv2.2), and KvS or silent subunits (Kv5,6,8.9) that can assemble with Kv2.1 or Kv2.2 to form heteromeric channels with novel biophysical properties. Many neurons express both types of subunits and therefore have the capacity to make both homomeric Kv2 channels and heteromeric Kv2/KvS channels. Determining the contributions of each of these channel types to native potassium currents has been very difficult because the differences in biophysical properties are modest and there are no Kv2/KvS-specific pharmacological tools. The authors set out to design a strategy to separate Kv2 and Kv2/KvS currents in native neurons based on their observation that Kv2/KvS channels have little sensitivity to the Kv2 pore blocker RY785 but are blocked by the Kv2 VSD blocker GxTx. They clearly demonstrate that Kv2/KvS currents can be differentiated from Kv2 currents in native neurons using a two-step strategy to first selectively block Kv2 with RY785, and then block both with GxTx. The manuscript is beautifully written; takes a very complex problem and strategy and breaks it down so both channel experts and the broad neuroscience community can understand it.

Strengths:

The compounds the authors use are highly selective and unlikely to have significant confounding cross-reactivity to other channel types. The authors provide strong evidence that all Kv2/KvS channels are resistant to RY785. This is a strength of the strategy - it can likely identify Kv2/KvS channels containing any of the 10 mammalian KvS subunits and thus be used as a general reagent on all types of neurons. The limitation then of course is that it can't differentiate the subtypes, but at this stage, the field really just needs to know how much Kv2/KvS channels contribute to native currents and this strategy provides a sound way to do so.

Weaknesses:

The authors are very clear about the limitations of their strategy, the most important of which is that they can't differentiate different subunit combinations of Kv2/KvS heteromers. This study is meant to be a start to understanding the roles of Kv2/KvS channels in vivo. As such, this is a minor weakness, far outweighed by the potential of the strategy to move the field through a roadblock that has existed since its inception.

The study accomplishes exactly what it set out to do: provide a means to determine the relative contributions of homomeric Kv2 and heteromeric Kv2/KvS channels to native delayed rectifier K+ currents in neurons. It also does a fabulous job laying out the case for why this is important to do.

Reviewer #2 (Public Review):

Summary:

Silent Kv subunits and the channels containing these Kv subunits (Kv2/KvS heteromers) are in the process of discovery. It is believed that these channels fine-tune the voltage-activated K+ currents that repolarize the membrane potential during action potentials, with a direct effect on cell excitability, mostly by determining action potentials firing frequency.

Strengths:

What makes silent Kv subunits even more important is that, by being expressed in specific tissues and cell types, different silent Kv subunits may have the ability to fine-tune the delayed rectifying voltage-activated K+ currents that are one of the currents that crucially determine cell excitability in these cells. The present manuscript introduces a pharmacological method to dissect the voltage-activated K+ currents mediated by Kv2/KvS heteromers as a means of starting to unveil their importance, together with Kv2-only channels, to the cells where they are expressed.

Weaknesses:

While the method is effective in quantifying these currents in any isolated cell under an electric voltage clamp, it is ineffective as a modulating maneuver to perhaps address these currents in an in vivo experimental setting. This is an important point but is not a claim made by the authors. There are other caveats with the methods and data:

(i) The need for a 'cocktail' of blockers to supposedly isolate Kv2 homomers and Kv2/KvS heteromers' currents from others may introduce errors in the quantification Kv2/KvS heteromers-mediated K+ currents and that is due to possible blockers off targets.

(ii) During the electrophysiology experiments, the authors use a holding potential that is not as negative as it is needed for the recording of the full population of the Kv2/KvS channels. Depolarized holding potentials lead to a certain level of inactivation of the channels, that vary according to the KvS involved/present in that specific population of channels. As a reminder, some KvS promote inactivation and others prevent inactivation. Therefore, the data must be interpreted as such.

(iii) The analysis of conductance activation by using tail currents is only accurate when dealing with non-inactivating conductances. Also, in dealing with a heterogenous population of Kv2/KvS heteromers, heterogenous K+ conductance deactivation kinetics is a must. Indeed, different KvS may significantly relate to different deactivation kinetics as well.

(iv) Silent Kv subunits may be retained in the ER, in heterologous systems like CHO cells. This aspect may subestimate their expression in these systems. Nevertheless, the authors show similar data in CHO cells and in primary neurons.

(v) The hallmark of silent Kv subunits is their effect on the time inactivation of K+ currents. As such, data should be shown throughout, preferably, from this perspective, but it was only done so in Figure 4G.

(vi) Functional characterization of currents only, as suggested by the authors as a bona fide of Kv2 and Kv2/KvS currents, should not be solely trusted to classify the currents and their channel mediators.

Reviewer #1 (Public Review):

Summary:

In their current study, Cummings et al have approached this fundamental biochemical problem using a combination of purified enzyme-substrate reactions, MS/MS, and microscopy in vitro to provide key insights into the hierarchy of generating polyglycylation in cilia and flagella. They first establish that TTLL8 is a monoglycylase, with the potential to add multiple mono glycine residues on both α- and β-tubulin. They then go on to establish that monoglycylation is essential for TTLL10 binding and catalytic activity, which progressively reduces as the level of polyglycylation increases. This provides an interesting mechanism of how the level of polyglycylation is regulated in the absence of a deglycylase. Finally, the authors also establish that for efficient TTLL10 activity, it is not just monoglycylation, but also polyglutamylation that is necessary, giving a key insight into how both these modifications interact with each other to ensure there is a balanced level of PTMs on the axonemes for efficient cilia function.

Strengths:

The manuscript is well-written, and experiments are succinctly planned and outlined. The experiments were used to provide the conclusions to what the authors were hypothesising and provide some new novel possible mechanistic insights into the whole process of regulation of tubulin glycylation in motile cilia.

Weaknesses:

The initial part of the manuscript where the authors discuss about the requirement of monoglycylation by TTLL8 is not new. This was established back in 2009 when Rogowski et al (2009) showed that polyglycylation of tubulin by TTLL10 occurs only when co-expressed in cells with TTLL3 or TTLL8. So, this part of the study adds very little new information to what was known.

The study also fails to discuss the involvement of the other monoglycylase, TTLL3 in the entire study, which is a weakness as in vivo, in cells, both the monoglycylases act in concert and so, may play a role in regulating the activity of TTLL10.

Reviewer #2 (Public Review):

In their manuscript, Cummings et al. focus on the enzymatic activities of TTLL3, TTLL8, and TTLL10, which catalyze the glycylation of tubulin, a crucial posttranslational modification for cilia maintenance and motility. The experiments are beautifully performed, with meticulous attention to detail and the inclusion of appropriate controls, ensuring the reliability of the findings. The authors utilized in vitro reconstitution to demonstrate that TTLL8 functions exclusively as a glycyl initiase, adding monoglycines at multiple positions on both α- and β-tubulin tails. In contrast, TTLL10 acts solely as a tubulin glycyl elongase, extending existing glycine chains. A notable finding is the differential substrate recognition between TTLL glycylases and TTLL glutamylases, highlighting a broader substrate promiscuity in glycylases compared to the more selective glutamylases. This observation aligns with the greater diversification observed among glutamylases. The study reveals a hierarchical mechanism of enzyme recruitment to microtubules, where TTLL10 binding necessitates prior monoglycylation by TTLL8. This binding is progressively inhibited by increasing polyglycine chain length, suggesting a self-regulatory mechanism for polyglycine chain length control. Furthermore, TTLL10 recruitment is enhanced by TTLL6-mediated polyglutamylation, illustrating a complex interplay between different tubulin modifications. In addition, they uncover that polyglutamylation stimulates TTLL10 recruitment without necessarily increasing glycylation on the same tubulin dimer, due to the potential for TTLLs to interact with neighboring tubulin dimers. This mechanism could lead to an enrichment of glycylation on the same microtubule, contributing to the complexity of the tubulin code. The article also addresses a significant challenge in the field: the difficulty of generating microtubules with controlled posttranslational modifications for in vitro studies. By identifying the specific modification sites and the interplay between TTLL activities, the authors provide a valuable tool for creating differentially glycylated microtubules. This advancement will facilitate further studies on the effects of glycylation on microtubule-associated proteins and the broader implications of the tubulin code. In summary, this study substantially contributes to our knowledge of posttranslational enzymes and their regulation, offering new insights into the biochemical mechanisms underlying microtubule modifications. The rigorous experimental approach and the novel findings presented make this a pivotal addition to the field of cellular and molecular biology.

Reviewer #1 (Public Review):

Summary:

DMS-MaP is a sequencing-based method for assessing RNA folding by detecting methyl adducts on unpaired A and C residues created by treatment with dimethylsulfate (DMS). DMS also creates methyl adducts on the N7 position of G, which could be sensitive to tertiary interactions with that atom, but N7-methyl adducts cannot be detected directly by sequencing. In this work, the authors adopt a previously developed method for converting N7-methyl-G to an abasic site to make it detectable by sequencing and then show that the ability of DMS to form an N7-methyl-G adduct is sensitive to RNA structural context. In particular, they look at the G-quadruplex structure motif, which is dense with N7-G interactions, is biologically important, and lacks conclusive methods for in-cell structural analysis.

Strengths:

- The authors clearly show that established methods for detecting N7-methyl-G adducts can be used to detect those adducts from DMS and that the formation of those adducts is sensitive to structural context, particularly G-quadruplexes.

- The authors assess the N7-methyl-G signal through a wide range of useful probing analyses, including standard folding, adduct correlations, mutate-and-map, and single-read clustering.

- The authors show encouraging preliminary results toward the detection of G-quadruplexes in cells using their method. Reliable detection of RNA G-quadruplexes in cells is a major limitation for the field and this result could lead to a significant advance.

- Overall, the work shows convincingly that N7-methyl-G adducts from DMS provide valuable structural information and that established data analyses can be adapted to incorporate the information.

Weaknesses:

- Most of the validation work is done on the spinach aptamer and it is the only RNA tested that has a known 3D structure. Although it is a useful model for validating this method, it does not provide a comprehensive view of what results to expect across varied RNA structures.

- It's not clear from this work what the predictive power of BASH-MaP would be when trying to identify G-quadruplexes in RNA sequences of unknown structure. Although clusters of G's with low reactivity and correlated mutations seem to be a strong signal for G-quadruplexes, no effort was made to test a range of G-rich sequences that are known to form G-quadruplexes or not. Having this information would be critical for assessing the ability of BASH-MaP to identify G-quadruplexes in cells.

- Although the authors present interesting results from various types of analysis, they do not appear to have developed a mature analysis pipeline for the community to use. I would be inclined to develop my own pipeline if I were to use this method.

- There are various aspects of the DAGGER analysis that don't make sense to me:<br /> (1) Folding of the RNA based on individual reads does not represent single-molecule folding since each read contains only a small fraction of the possible adducts that could have formed on that molecule. As a result, each fold will largely be driven by the naive folding algorithm. I recommend a method like DREEM that clusters reads into profiles representing different conformations.<br /> (2) How reliable is it to force open clusters of low-reactivity G's across RNA's that don't already have known G-quadruplexes?<br /> (3) By forcing a G-quadruplex open it will be treated as a loop by the folding algorithm, so the energetics won't be accurate.<br /> (4) It's not clear how signals on "normal" G's are treated. In Figure 5C some are wiped to 0 but others are kept as 1.

Reviewer #2 (Public Review):

Summary:

The manuscript introduces BASH MaP and DAGGER, innovative tools for analyzing RNA tertiary structures, specifically focusing on the G-quadruplexes. Traditional methods have struggled to detect and analyze these structures due to their reliance on interactions on the Hoogsteen face of guanine, which are not readily observable through conventional probing that targets Watson-Crick interactions. BASH MaP employs dimethyl sulfate and potassium borohydride to enhance the detection of N7-methylguanosine by converting it into an abasic site, thereby enabling its identification through misincorporation during reverse transcription. This method provides higher precision in identifying G-quadruplexes and offers deeper insights into RNA's structural dynamics and alternative conformations in both vitro and cellular contexts. Overall, the study is well-executed, demonstrating robust signal detection of N7-Gs with some compelling positive controls, thorough analysis, and beautifully presented figures.

Strengths:

The manuscript introduces a new method to detect G-quadruplexes (G-qs) that simplifies and potentially enhances the robustness and quantification compared to previous methods relying on reverse transcription truncations. The authors provide a strong positive control, demonstrating a 70% misincorporation at endogenous N7-G within the 18S rRNA, which illustrates BASH MaP's high signal-to-noise ratio. The data concerning the detection of positive control G-qs is particularly compelling.

Weaknesses:

Figure 3E shows considerable variability in the correlations among guanosines, suggesting that the methods may struggle with specificity in determining guanosine participation within and between different quadruplexes. There is no estimation of the methods false positive discovery rate.

Reviewer #3 (Public Review):

Summary:

In this study, the authors aim to develop an experimental/computational pipeline to assess the modification status of an RNA following treatment with dimethylsulfate (DMS). Building upon the more common DMS Map method, which predominantly assesses the modification status of the Watson-Crick-Franklin face of A's and C's, the authors insert a chemical processing step in the workflow prior to deep sequencing that enables detection of methylation at the N7 position of guanosine residues. This approach, termed BASH MaP, provides a more complete assessment of the true modification status of an RNA following DMS treatment and this new information provides a powerful set of constraints for assessing the secondary structure and conformational state of an RNA. In developing this work, the authors use Spinach as a model RNA. Spinach is a fluorogenic RNA that binds and activates the fluorescence of a small molecule ligand. Crystal structures of this RNA with ligand bound show that it contains a G-quadruplex motif. In applying BASH MaP to Spinach, the authors also perform the more standard DMS MaP for comparison. They show that the BASH MaP workflow appears to retain the information yielded by DMS MaP while providing new information about guanosine modifications. In Spinach, the G-quadruplex G's have the least reactive N7 positions, consistent with the engagement of N7 in hydrogen bonding interactions at G's involved in quadruplex formation. Moreover, because the inclusion of data corresponding to G increases the number of misincorporations per transcript, BASH MaP is more amenable to analysis of co-occurring misincorporations through statistical analysis, especially in combination with site-specific mutations. These co-occurring misincorporations provide information regarding what nucleotides are structurally coupled within an RNA conformation. By deploying a likelihood-ratio statistical test on BASH MaP data, the authors can identify Gs in G-quadruplexes, deconvolute G-G correlation networks, base-triple interactions and even stacking interactions. Further, the authors develop a pipeline to use the BASH MaP-derived G-modification data to assist in the prediction of RNA secondary structure and identify alternative conformations adopted by a particular RNA. This seems to help with the prediction of secondary structure for Spinach RNA.

Strengths:

The BASH Map procedure and downstream data analysis pipeline more fully identify the complement of methylations to be identified from the DMS treatment of RNA, thereby enriching the information content. This in turn allows for more robust computational/statistical analysis, which likely will lead to more accurate structure predictions. This seems to be the case for the Spinach RNA.

Weaknesses:

The authors demonstrate that their method can detect G-quadruplexes in Spinach and some other RNAs both in vitro and in cells. However, the performance of BASH MaP and associated computational analysis in the context of other RNAs remains to be determined.

Reviewer #1 (Public Review):

Summary:

The authors performed experimental evolution of MreB mutants that have a slow-growing round phenotype and studied the subsequent evolutionary trajectory using analysis tools from molecular biology. It was remarkable and interesting that they found that the original phenotype was not restored (most common in these studies) but that the round phenotype was maintained.

Strengths:

The finding that the round phenotype was maintained during evolution rather than that the original phenotype, rod-shaped cells, was recovered is interesting. The paper extensively investigates what happens during adaptation with various different techniques. Also, the extensive discussion of the findings at the end of the paper is well thought through and insightful.

Weaknesses:<br /> I find there are three general weaknesses:

(1) Although the paper states in the abstract that it emphasizes "new knowledge to be gained" it remains unclear what this concretely is. On page 4 they state 3 three research questions, these could be more extensively discussed in the abstract. Also, these questions read more like genetics questions while the paper is a lot about cell biological findings.

(2) it is not clear to me from the text what we already know about the restoration of MreB loss from suppressors studies (in the literature). Are there suppressor screens in the literature and which part of the findings is consistent with suppressor screens and which parts are new knowledge?

(3) The clarity of the figures, captions, and data quantification need to be improved.

Reviewer #2 (Public Review):

Yulo et al. show that deletion of MreB causes reduced fitness in P. fluorescens SBW25 and that this reduction in fitness may be primarily caused by alterations in cell volume. To understand the effect of cell volume on proliferation, they performed an evolution experiment through which they predominantly obtained mutations in pbp1A that decreased cell volume and increased viability. Furthermore, they provide evidence to propose that the pbp1A mutants may have decreased PG cross-linking which might have helped in restoring the fitness by rectifying the disorganised PG synthesis caused by the absence of MreB. Overall this is an interesting study.

Queries:

Do the small cells of mreB null background indeed have have no DNA? It is not apparent from the DAPI images presented in Supplementary Figure 17. A more detailed analysis will help to support this claim.

What happens to viability and cell morphology when pbp1A is removed in the mreB null background? If it is actually a decrease in pbp1A activity that leads to the rescue, then pbp1A- mreB- cells should have better viability, reduced cell volume and organised PG synthesis. Especially as the PG cross-linking is almost at the same level as the T362 or D484 mutant.

What is the status of PG cross-linking in ΔmreB Δpflu4921-4925 (Line 7)?

What is the morphology of the cells in Line 2 and Line 5? It may be interesting to see if PG cross-linking and cell wall synthesis is also altered in the cells from these lines.

The data presented in 4B should be quantified with appropriate input controls.

What are the statistical analyses used in 4A and what is the significance value?

A more rigorous statistical analysis indicating the number of replicates should be done throughout.

Reviewer #3 (Public Review):

This paper addresses an understudied problem in microbiology: the evolution of bacterial cell shape. Bacterial cells can take a range of forms, among the most common being rods and spheres. The consensus view is that rods are the ancestral form and spheres the derived form. The molecular machinery governing these different shapes is fairly well understood but the evolutionary drivers responsible for the transition between rods and spheres are not. Enter Yulo et al.'s work. The authors start by noting that deletion of a highly conserved gene called MreB in the Gram-negative bacterium Pseudomonas fluorescens reduces fitness but does not kill the cell (as happens in other species like E. coli and B. subtilis) and causes cells to become spherical rather than their normal rod shape. They then ask whether evolution for 1000 generations restores the rod shape of these cells when propagated in a rich, benign medium.

The answer is no. The evolved lineages recovered fitness by the end of the experiment, growing just as well as the unevolved rod-shaped ancestor, but remained spherical. The authors provide an impressively detailed investigation of the genetic and molecular changes that evolved. Their leading results are:

(1) The loss of fitness associated with MreB deletion causes high variation in cell volume among sibling cells after cell division.

(2) Fitness recovery is largely driven by a single, loss-of-function point mutation that evolves within the first ~250 generations that reduces the variability in cell volume among siblings.

(3) The main route to restoring fitness and reducing variability involves loss of function mutations causing a reduction of TPase and peptidoglycan cross-linking, leading to a disorganized cell wall architecture characteristic of spherical cells.

The inferences made in this paper are on the whole well supported by the data. The authors provide a uniquely comprehensive account of how a key genetic change leads to gains in fitness and the spectrum of phenotypes that are impacted and provide insight into the molecular mechanisms underlying models of cell shape.

Suggested improvements and clarifications include:

(1) A schematic of the molecular interactions governing cell wall formation could be useful in the introduction to help orient readers less familiar with the current state of knowledge and key molecular players.

(2) More detail on the bioinformatics approaches to assembling genomes and identifying the key compensatory mutations are needed, particularly in the methods section. This whole subject remains something of an art, with many different tools used. Specifying these tools, and the parameter settings used, will improve transparency and reproducibility, should it be needed.

(3) Corrections for multiple comparisons should be used and reported whenever more than one construct or strain is compared to the common ancestor, as in Supplementary Figure 19A (relative PG density of different constructs versus the SBW25 ancestor).

(4) The authors refrain from making strong claims about the nature of selection on cell shape, perhaps because their main interest is the molecular mechanisms responsible. However, I think more can be said on the evolutionary side, along two lines. First, they have good evidence that cell volume is a trait under strong stabilizing selection, with cells of intermediate volume having the highest fitness. This is notable because there are rather few examples of stabilizing selection where the underlying mechanisms responsible are so well characterized. Second, this paper succeeds in providing an explanation for how spherical cells can readily evolve from a rod-shaped ancestor but leaves open how rods evolved in the first place. Can the authors speculate as to how the complex, coordinated system leading to rods first evolved? Or why not all cells have lost rod shape and become spherical, if it is so easy to achieve? These are important evolutionary questions that remain unaddressed. The manuscript could be improved by at least flagging these as unanswered questions deserving of further attention.

The value of this paper stems both from the insight it provides on the underlying molecular model for cell shape and from what it reveals about some key features of the evolutionary process. The paper, as it currently stands, provides more on which to chew for the molecular side than the evolutionary side. It provides valuable insights into the molecular architecture of how cells grow and what governs their shape. The evolutionary phenomena emphasized by the authors - the importance of loss-of-function mutations in driving rapid compensatory fitness gains and that multiple genetic and molecular routes to high fitness are often available, even in the relatively short time frame of a few hundred generations - are well-understood phenomena and so arguably of less broad interest. The more compelling evolutionary questions concern the nature and cause of stabilizing selection (in this case cell volume) and the evolution of complexity. The paper misses an opportunity to highlight the former and, while claiming to shed light on the latter, provides rather little useful insight.

Joint Public Review:

An outside expert evaluated your responses to the original reviewers and offered the following comments:

The main criticism was whether deleterious variants were appropriately classified in the work. The authors use two different methods to characterize the effect of alleles to satisfy these comments. The result is somewhat complex. The authors do replicate the effect of dominance on fixation and segregation of deleterious alleles by classifying polymorphisms as synonymous or synonymous with SNPeff. This is not entirely surprising as it is approximately equivalent to classifying based on fold degeneracy (but it includes sites that have other than 0 or 4 fold degeneracy). However, the authors do not mention in the text that their observation of increased segregating deleterious mutations in recessive alleles was only statistically significant in A. halleri (for both analyses). Using SIFT, the authors only find an effect of dominance in A. lyrata. So in reality, while the trends are the same across the analyses, the statistical significance of the effects of dominance was not consistent.

Reviewer 2 had several more detailed criticisms of the manuscript. The first was that the authors should explore the dominance of linked deleterious mutations themselves. I agree that this would be interesting, but it is very difficult to accomplish, and I agree with the author's reluctance to do much more here. The reviewer also criticized the authors simulation approach. The authors provided their simulation script as requested, but declined to do additional simulations under varied selection coefficients. I felt this was a minimally adequate response to the reviewers concerns, but the authors could have reasonably conducted a few additional simulations under varied selection coefficients.

I think that the scope of the findings described in the assessment was reasonable. This is interesting work, but despite the author's arguments, the system is somewhat unique if for no other reason than that balancing selection at S-loci is uniquely strong

Reviewer #1 (Public Review):

The authors provided a detailed analysis of the real-time structural changes in actin filaments resulting from cofilin binding, using High-Speed Atomic Force Microscopy (HSAFM). The cofilin family controls the lifespan of actin filaments in cells by severing the filament and promoting depolymerization. Understanding the effects of cofilin on actin filament structure is critical. It is widely acknowledged that cofilin binding significantly shortens the pitch of the actin helix. The authors previously reported (1) that this shortening extends to the unbound region of the actin filament on the pointed end side of the cofilin binding cluster. In this study, the authors presented substantially improved AFM images and provide detailed accounts of the dynamics observed. It was found that a minimal cofilin-binding cluster, consisting of 2-4 molecules, could induce changes in the helical parameters over one or more actin crossover repeats. Adjacent to the cofilin-binding clusters, the actin crossovers were observed to shorten within seconds, and this shortening was limited to one side of the cluster. Additionally, the phosphate binding to the actin filament was observed to stabilize the helical twist, suggesting a mechanism in which cofilin preferentially binds to ADP-bound actin filaments. These findings significantly advance our understanding of actin filament dynamics which is essential for a wide of cellular processes.

However, two insufficient parts exist. Readers should be aware of possible errors in the Mean Axial Distance (MAD) analysis and the limitations of discussions about the actin subunit structure.

The authors have presented findings that the MAD within actin filaments exhibits a significant dependency on the helical twist. However, difficulty in determining each subunit interval from the AFM image might affect the analysis. For example, the observation of three peaks in HHP6 of Figure Supplement 6C, corresponding to 4.5 pairs, showed peak intervals of 5, 11.8, 8.7, and 5.7 nm (measured from the figure). The second region (11.8 nm) appears excessively long. If one peak is hidden in the second region, the MAD becomes 5.5 nm.

The authors also suggest a strong link between the C-form (cofilin binding form of actin found in cofilactin) and the formation of regions of the short pitch helix outside the cofilin binding cluster. However, the AFM observation did not provide any evidence about the actin form in these regions because of measurement limitations. Additionally, Oda et al. (2) have demonstrated that the C-form is highly unstable in the absence of cofilin binding, casting doubt on the possibility of the C-form propagating without cofilin binding. The "C-actin-like structure" in the paper is not necessarily related to the C-form actin. It might be one of the G-forms (monomeric actin forms) or another unknown form.

(1) K. X. Ngo et al., a, Cofilin-induced unidirectional cooperative conformational changes in actin filaments revealed by high-speed atomic force microscopy. eLife 4, (2015).<br /> (2) T. Oda et al., Structural Polymorphism of Actin. Journal of molecular biology 431, 3217-3228 (2019).

Reviewer #2 (Public Review):

Summary:

This study by Ngo et al. uses mostly high-speed AFM to estimate conformational changes within actin filaments, as they get decorated by cofilin. The authors build on their earlier study (Ngo et al. eLife 2015) where they used the same technique to monitor the expansion of cofilin clusters on actin filaments, and the propagation of the associated conformational changes in the filament (reduction of the helical pitch). Here, they propose a higher-resolution description of the binding of cofilin to actin filaments.

Strengths:

The high speed AFM technique used here is quite original to address this question, compared to more classical light and electron microscopy techniques. It can certainly bring valuable information as it provides a high spatial resolution while monitoring live events. Also, in this paper, a nice effort was made to make the 3D structures and conformational changes clear and understandable.

Weaknesses:

In spite of the authors' response to my earlier comments, I still have concerns regarding the AFM technique. In particular, regarding the interactions of the filaments with the surface, which I still find unclear and potentially problematic.

The filaments appear densely packed on the surface, and even clearly in register in some images (if not most images, e.g., Figs 3AD, 4BC, 5A, 8AC). I understand that there are practical reasons for this, but isn't there a risk that this could affect the result? Maybe I did not understand the authors' response well enough, but I did not see a clear control that would alleviate my concern.

The properties of the lipid layer and its interaction with the actin filaments are still unclear to me. A poor control of these interactions is a problem if one aims to measure conformational changes at high resolution. The strength of the interaction appears tuned by the ratio of lipids put on the surface to change its electrostatic charge. A strong attachment likely does more than suppress torsional motion (as claimed in Fig 8A). It may also hinder cofilin binding in several ways (lower availability of binding sites on the filament facing the surface, electrostatic interactions between cofilin and the surface, etc.). Here again, I was not fully reassured by the authors' response.

The identification of cofilactin regions relies on the additional height of the "peaks", due to the presence of cofilin. It thus seems that cofilin is detected every half helical pitch (HHP), and I still don't understand how the authors can make reliable claims regarding the presence or absence of cofilin between these peaks.

Reviewer #2 (Public Review):

Summary:<br /> The manuscript by Kelbert et al. presents results on the involvement of the yeast transcription factor Sfp1 in the stabilisation of transcripts whose synthesis it stimulates. Sfp1 is known to affect the synthesis of a number of important cellular transcripts, such as many of those that code for ribosomal proteins. The hypothesis that a transcription factor can remain bound to the nascent transcript and affect its cytoplasmic half-life is attractive, but the methods used to demonstrate the half-life effects and the association of Sfp1 with cytoplasmic transcripts remain to be fully validated, as explained in my comments on the results below:

Comments on methodology and results:<br /> (1) A two-hybrid-based assay for protein-protein interactions identified Sfp1, a transcription factor known for its effects on ribosomal protein gene expression, as interacting with Rpb4, a subunit of RNA polymerase II. Classical two-hybrid experiments depend on the presence of the tested proteins in the nucleus of yeast cells, suggesting that the observed interaction occurs in the nucleus. Unfortunately, the two-hybrid method cannot determine whether the interaction is direct or mediated by nucleic acids.

(2) Inactivation of nup49, a component of the nuclear pore complex, resulted in the redistribution of GFP-Sfp1 into the cytoplasm at the temperature non-permissive for the nup49-313 strain, suggesting that GFP-Sfp1 is a nucleo-cytoplasmic shuttling protein. This observation confirmed the dynamic nature of the nucleo-cytoplasmic distribution of Sfp1. For example, a similar redistribution to the cytoplasm was previously reported following rapamycin treatment and under starvation (Marion et al., PNAS 2004). In conjunction with the observation of an interaction with Rpb4, the authors observed slower nuclear import kinetics for GFP-Sfp1 in the absence of Rpb4 when cells were transferred to a glucose-containing medium after a period of starvation. Since the redistribution of GFP-Sfp1 was abolished in an rpb1-1/nup49-313 double mutant, the authors concluded that Sfp1 localisation to the cytoplasm depends on transcription. The double mutant yeast cells may show a variety of non-specific effects at the restrictive temperature, and whether transcription is required for Sfp1 cytoplasmic localisation remains incompletely demonstrated.

(3) Under starvation conditions, which led to the presence of Sfp1 in the cytoplasm and have previously been correlated with a decrease in the transcription of Sfp1 target genes, the authors observed that a plasmid-based expressed GFP-Sfp1 accumulated in cytoplasmic foci. These foci were also labelled by P-body markers such as Dcp2 and Lsm1. The quality of the microscopic images provided does not allow to determine whether Rpb4-RFP colocalises with GFP-Sfp1.

(4) To understand to which RNA Sfp1 might bind, the authors used an N-terminally tagged fusion protein in a cross-linking and purification experiment. This method identified 264 transcripts for which the CRAC signal was considered positive and which mostly correspond to abundant mRNAs, including 74 ribosomal protein mRNAs or metabolic enzyme-abundant mRNAs such as PGK1. The authors did not provide evidence for the specificity of the observed CRAC signal, in particular, what would be the background of a similar experiment performed without UV cross-linking. In a validation experiment, the presence of several mRNAs in a purified SFP1 fraction was measured at levels that reflect the relative levels of RNA in a total RNA extract. Negative controls showing that abundant mRNAs not found in the CRAC experiment were clearly depleted from the purified fraction with Sfp1 would be crucial to assessing the specificity of the observed protein-RNA interactions. The CRAC-selected mRNAs were enriched for genes whose expression was previously shown to be upregulated upon Sfp1 overexpression (Albert et al., 2019). The presence of unspliced RPL30 pre-mRNA in the Sfp1 purification was interpreted as a sign of co-transcriptional assembly of Sfp1 into mRNA, but in the absence of valid negative controls, this hypothesis would require further experimental validation.

(5) To address the important question of whether co-transcriptional assembly of Spf1 with transcripts could alter their stability, the authors first used a reporter system in which the RPL30 transcription unit is transferred to vectors under different transcriptional contexts, as previously described by the Choder laboratory (Bregman et al. 2011). While RPL30 expressed under an ACT1 promoter was barely detectable, the highest levels of RNA were observed in the context of the native upstream RPL30 sequence when Rap1 binding sites were also present. Sfp1 showed better association with reporter mRNAs containing Rap1 binding sites in the promoter region. However, removal of the Rap1 binding sites from the reporter vector also led to a drastic decrease in reporter mRNA levels. Whether the fraction of co-purified RNA is nuclear and co-transcriptional or not cannot be inferred from these results.

(6) To complement the biochemical data presented in the first part of the manuscript, the authors turned to the deletion or rapid depletion of SFP1 and used labelling experiments to assess changes in the rate of synthesis, abundance, and decay of mRNAs under these conditions. An important observation was that in the absence of Sfp1, mRNAs encoding ribosomal protein genes not only had a reduced synthesis rate but also an increased degradation rate. This important observation needs careful validation, as genomic run-on experiments were used to measure half-lives, and this particular method was found to give results that correlated poorly with other measures of half-life in yeast (e.g. Chappelboim et al., 2022 for a comparison). Similarly, the use of thiolutin to block transcription as a method of assessing mRNA half-life has been reported to be problematic, as thiolutin can specifically inhibit the degradation of ribosomal protein mRNA (Pelechano & Perez-Ortin, 2008). Specific repressible reporters, such as those used by Baudrimont et al. (2017), would need to be tested to validate the effect of Sfp1 on the half-life of specific mRNAs. Also, it would be very difficult to infer from the images presented whether the rate of deadenylation is altered by Sfp1.

(7) The effects of SFP1 on transcription were investigated by chromatin purification with Rpb3, a subunit of RNA polymerase, and the results were compared with synthesis rates determined by genomic run-on experiments. The decrease in polII presence on transcripts in the absence of SFP1 was not accompanied by a marked decrease in transcript output, suggesting an effect of Sfp1 in ensuring robust transcription and avoiding RNA polymerase backtracking. To further investigate the phenotypes associated with the depletion or absence of Sfp1, the authors examined the presence of Rpb4 along transcription units compared to Rpb3. One effect of spf1 deficiency was that this ratio, which decreased from the start of transcription towards the end of transcripts, increased slightly. The results presented are largely correlative and could arise from the focus on very specific types of mRNAs, such as those of ribosomal protein genes, which are sensitive to stress and are targeted by very active RNA degradation mechanisms activated, for example, under heat stress (Bresson et al., 2020).

Strengths:<br /> - Diversity of experimental approaches used<br /> - Validation of large-scale results with appropriate reporters

Weaknesses:<br /> - Choice of evaluation method to test mRNA half-life<br /> - Lack of controls for the CRAC results

Reviewer #3 (Public Review):

This work probes the control of the hox operon in the cyanobacterium Synechocystis, where this operon directs the synthesis of a bidirectional hydrogenase that functions to produce hydrogen. In assessing the control of the hox system, the authors focused on the relative contributions of cyAbrB2, alongside SigE (and to a lesser extent, SigA and cyAbrB1) under both aerobic and microoxic conditions. In mapping the binding sites of these different proteins, they discovered that cyAbrB2 bound many sites throughout the chromosome, repressed many of its target genes, and preferentially bound regions that were (relatively) rich in AT-residues. These characteristics led the authors to consider that cyAbrB2 may function as a nucleoid-associated protein (NAP) in Synechocystis, given the functional similarities with other NAPs like H-NS. They assessed the local chromosome conformation in both wild type and cyabrB2 mutant strains at multiple sites within a 40 kb window on either side of the hox locus, using a region within the hox operon as bait. They concluded that cyAbrB2 functions as a nucleoid associated protein that influences the activity of SigE through its modulation of chromosome architecture.

The authors approached their experiments carefully, and the data were generally very clearly presented. At the same time, the overall work contains many lines of inquiry and different protein investigations that in some ways made it more challenging to identify the overall take-away message(s).

Based on the data presented, the authors make a strong case for cyAbrB2 as a nucleoid-associated protein, given the multiple ways in which is seems to function similarly to the well-studied Escherichia coli H-NS protein. They now provide additional commentary that relates cyAbrB2 with other nucleoid-associated proteins.

Previous work had revealed a role for SigE in the control of hox cluster expression, which nicely justified its inclusion (and focus) in this study. The focus on cyAbrB2 is also well-justified, given previous reports of its control of hox expression; however, it shares binding sites with an essential homologue cyAbrB1. Interestingly, while the B1 protein appears to bind similar sites, instead of repressing hox expression, it is known as an activator of this operon. If the information on cyAbrB1 is retained in the manuscript, it would be important to consider how cyAbrB1 activity might influence the results described here (although the authors could also consider removing the cyAbrB1 information to help improve the focus of the manuscript).

Reviewer #1 (Public Review):

This manuscript presents a model in which combined action of the transporter-like protein DISP and the sheddases ADAM10/17 promote shedding of a mono-cholesteroylated Sonic Hedgehog (SHH) species following cleavage of palmitate from the dually lipidated precursor ligand. The authors propose that this leads to transfer of the cholesterol-modified SHH to HDL for solubilization. The minimal requirement for SHH release by this mechanism is proposed to be the covalently linked cholesterol modification because DISP could promote transfer of a cholesteroylated mCherry reporter protein to serum HDL. The authors used an in vitro system to demonstrate dependency on DISP/SCUBE2 for release of the cholesterol modified ligand. These results confirm previously published results from other groups (PMC3387659 and PMC3682496).

A strength of the work is the use of a bicistronic SHH-Hhat system to consistently generate dually-lipidated ligand to determine the quantity and lipidation status of SHH released into cell culture media.

Key shortcomings include the unusual normalization strategies used for many experiments and the lack of quantification/statistical analyses for several experiments. Due to these omissions, it is difficult to conclude that the data justify the conclusions. The significance of the data provided is overstated because many of the presented experiments confirm/support previously published work. The study provides a modest advance in understanding of the complex issue of SHH membrane extraction.

Reviewer #2 (Public Review):

Ehring et al. analyze contributions of Dispatched, Scube2, serum lipoproteins and Sonic Hedgehog lipid modifications to the generation of different Shh release forms. Hedgehog proteins are anchored in cellular membranes by N-terminal palmitate and C-terminal cholesterol modifications, yet spread through tissues and are released into the circulation. How Hedgehog proteins can be released, and in which form, remains controversial. The authors systematically dissect contributions of several previously identified factors, and present evidence that Disp, Scube2 and lipoproteins concertedly act to release a novel Shh variant that is cholesterol-modified but not palmitoylated. The results provide new insights into the function of Disp and Scube2 in Hedgehog release. The findings concerning the function of lipoproteins and cholesterol in Hedgehog release are largely confirmatory (PMID 23554573, 20685986). However, in light of the multitude of competing models for Hedgehog release, the present study is a valuable contribution that provides further insights into the relevance of lipoproteins in this process.

A novel and surprising finding of the present study is the differential removal of Shh N- or C-terminal lipid anchors depending on the presence of HDL and/or Disp. In particular, the identification of a non-palmitoylated but cholesterol-modified Shh variant that associates with lipoproteins is potentially important. The authors use RP-HPLC and defined controls to assess the properties of processed Shh forms, but their precise molecular identity remains to be defined. A caveat is the strong reliance on over-expression of Shh in a single cell line. The authors detect Shh variants that are released independently of Disp and Scube2 in secretion assays, which however are excluded from interpretation as experimental artifacts. Thus, it would be important to demonstrate key findings in cells that secrete Shh endogenously.

Reviewer #2 (Public Review):

In this work, Sarkar et al. investigated the potential ability of adenosine triphosphate (ATP) as a solubilizer of protein aggregates by combining MD simulations and ThT/TEM experiments. They explored how ATP influences the conformational behaviors of Trp-cage and β-amyloid Aβ40 proteins. Currently, there are no experiments in the literature supporting their simulation results of ATP on Trp-cage. The simulation protocol employed for the Aβ40 monomer system is conventional MD simulation, while REMD simulation (an enhanced sampling method) is used for the Aβ monomer + ATP system. It is not clear whether the conformational difference is caused by ATP or by the different simulation methods used. ThT/TEM experiments should be performed on Aβ40 fibrils rather than on Aβ(16-22) aggregates. Moreover, to elucidate their experimental results that ATP can dissolve preformed Aβ fibrils, the authors need to study the influence of ATP on Aβ fibrils instead of on Aβ dimer in their MD simulations. The novelty of this study is limited. The role of ATP in inhibiting Aβ fibril formation and dissolving preformed Aβ fibrils has been reported in previous experimental and computational studies (Journal of Alzheimer's Disease, 2014, 41: 561; Science 2017, 2017, 356, 753-756 J. Phys. Chem. B 2019, 123, 9922−9933; Scientific Reports, 2024, 14: 8134). However, most of those papers are not discussed in this manuscript. Additionally, some details of MD simulations and data analysis are missing in the manuscript, including the initial structures of all the simulations, the method for free energy calculation, the dielectric constant used, etc.

Reviewer #1 (Public Review):

Summary:

This work combines molecular dynamics (MD) simulations along with experimental elucidation of the efficacy of ATP as a biological hydrotrope. While ATP is broadly known as the energy currency, it has also been suggested to modulate the stability of biomolecules and their aggregation propensity. In the computational part of the work, the authors demonstrate that ATP increases the population of the more expanded conformations (higher radius of gyration) in both a soluble folded mini-protein Trp-cage and an intrinsically disordered protein (IDP) Aβ40. Furthermore, ATP is shown to destabilise the pre-formed fibrillar structures using both simulation and experimental data (ThT assay and TEM images). They have also suggested that the biological hydrotrope ATP has significantly higher efficacy as compared to the commonly used chemical hydrotrope sodium xylene sulfonate (NaXS).

Strengths:

This work presents a comprehensive and compelling investigation of the effect of ATP on the conformational population of two types of proteins: globular/folded and IDP. The role of ATP as an "aggregate solubilizer" of pre-formed fibrils has been demonstrated using both simulation and experiments. They also elucidate the mechanism of action of ATP as a multi-purpose solubilizer in a protein-specific manner. Depending on the protein, it can interact through electrostatic interactions (for predominantly charged IDPs like Aβ40), or primarily van der Waals' interactions through (for Trp-Cage).

Weaknesses:

The data presented by the authors are sound and adequately support the conclusions drawn by the authors. However, there are a few points that could be discussed or elucidated further to broaden the scope of the conclusions drawn in this work as discussed below:

(i) The concentration of ATP used in the simulations is significantly higher (500 mM) as compared to those used in the experiments (6-20 mM) or cellular cytoplasm (~5 mM as mentioned by the authors). Since the authors mention already known concentration dependence of the effect of ATP, it is worth clarifying the possible limitations and implications of the high ATP concentrations in the simulations. It seems ATP can stabilise the proteins at low concentrations, but the current work does not address this possible effect. It would be interesting to see whether the effect of ATP on globular proteins and IDPs remains similar even at lower ATP concentrations.

(ii) The authors make a somewhat ambitious statement that the role of ATP as a solubilizer of pre-formed fibrils could be used as a therapeutic strategy in protein aggregation-related diseases. However, it is not clear how it would be so since ATP is a promiscuous substrate in several biochemical processes and any additional administration of ATP beyond normal cellular concentration (~5 mM) could be detrimental.

(iii) A natural question arises about what is so special about ATP as a solubilizer. The authors have also asked this question but in a limited scope of comparing to a commonly used chemical hydrotrope NaXS. However, a bigger question would be what kind of chemical/physical features make ATP special? For example, (i) if the amphiphilic property is important, what about some standard surfactants? (ii) how would ATP compare to other nucleotides like ADP or GTP? It might be useful to explore such questions in the future to further establish the special role of ATP in this regard.

(iv) In Figure 2F, it seems that in the presence of 0.5 M ATP, the Rg increases (as expected), but the number of native contacts remains almost similar. The reduction in the number of native contacts at higher ATP concentrations is not as dramatic as the increase in Rg. This is somewhat counterintuitive and should be looked into. Normally one would expect a monotonous reduction in the number of native contacts as the protein unfolds (increase in Rg).

Reviewer #3 (Public Review):

Summary:

Since its first experimental report in 2017 (Patel et al. Science 2017), there have been several studies on the phenomenon in which ATP functions as a biological hydrotrope of protein aggregates. In this manuscript, by conducting molecular dynamics simulations of three different proteins, Trp-cage, Abeta40 monomer, and Abeta40 dimer at a high concentration of ATP (0.1, 0.5 M), Sarkar et al. find that the amphiphilic nature of ATP, arising from its molecular structure consisting of phosphate group (PG), sugar ring, and aromatic base, enables it to interact with proteins in a protein-specific manner and prevents their aggregation and solubilize if they aggregate. The authors also point out that in comparison with NaXS, which is the traditional chemical hydrotrope, ATP is more efficient in solubilizing protein aggregates because of its amphiphilic nature.

Trp-cage, featured with a hydrophobic core in its native state, is denatured at high ATP concentration. The authors show that the aromatic base group (purine group) of ATP is responsible for inducing the denaturation of helical motifs in the native state.

For Abeta40, which can be classified as an IDP with charged residues, it is shown that ATP disrupts the salt bridge (D23-K28) required for the stability of beta-turn formation.

By showing that ATP can disassemble preformed protein oligomers (Abeta40 dimer), the authors argue that ATP is "potent enough to disassemble existing protein droplets, maintaining proper cellular homeostasis," and enhancing solubility.

Overall, the message of the paper is clear and straightforward to follow. I did not follow all the literature, but I see in the literature search, that there are several studies on this subject. (J. Am. Chem. Soc. 2021, 143, 31, 11982-11993; J. Phys. Chem. B 2022, 126, 42, 8486-8494; J. Phys. Chem. B 2021, 125, 28, 7717-7731; J. Phys. Chem. B 2020, 124, 1, 210-223).

If this study is indeed the first one to test using MD simulations whether ATP is a solubilizer of protein aggregates, it may deserve some attention from the community. But, the authors should definitely discuss the content of existing studies, and make it explicit what is new in this study.

Strengths:

The authors showed that due to its amphiphilic nature, ATP can interact with different proteins in a protein-specific manner, a. finding more general and specific than merely calling ATP a biological hydrotrope.

Weaknesses:

(1) My only major concern is that the simulations were performed at unusually high ATP concentrations (100 and 500 mM of ATP), whereas the real cellular concentration of ATP is 1-5 mM. Even if ATP is a good solubilizer of protein aggregates, the actual concentration should matter. I was wondering if there is a previous report on a titration curve of protein aggregates against ATP, and what is the transition mid-point of ATP-induced solubility of protein aggregates.

For instance, urea or GdmCl have long been known as the non-specific denaturants of proteins, and it has been well experimented that their transition mid-point of protein unfolding is ~(1 - 6) M depending on the proteins.

(2) The sentence "... a clear shift of relative population of Abeta40 conformational subensemble towards a basin with higher Rg and lower number of contacts in the presence of ATP" is not a precise description of Figures 4A and 4B. It is not clear from the figures whether the Rg of Abeta40 is increased when Abeta40 is subject to ATP. The authors should give a more precise description of what is observed in the result from their simulations or consider a better-order parameter to describe the change in molecular structure. In addition, the disruption of beta-sheet from Figure 4E to 4F is not very clear. The authors may want to use an arrow to indicate the region of the contact map associated with this change.

Although the full atomistic simulations were carried out, the analyses demonstrated in this study are a bit rudimentary and coarse-grained (e.g, Rg is a rather poor order parameter to discuss dynamics involved in proteins). The authors could go beyond and say more about how ATP interacts with proteins and disrupts the stable configurations.

(3) Although the amphiphilic character of ATP is highlighted, a similar comment can be made as to GTP. Is GTP, whose cellular concentration is ~0.5 mM, also a good solubilizer of protein aggregates? If not, why? Please comment.

Reviewer #1 (Public Review):

Summary:<br /> This paper presents a cognitive model of out-of-distribution generalisation, where the representational basis is grid-cell codes. In particular, the authors consider the tasks of analogies, addition, and multiplication, and the out-of-distribution tests are shifting or scaling the input domain. The authors utilise grid cell codes, which are multi-scale as well as translationally invariant due to their periodicity. To allow for domain adaptation, the authors use DPP-A which is, in this context, a mechanism of adapting to input scale changes. The authors present simulation results demonstrating that this model can perform out-of-distribution generalisation to input translations and re-scaling, whereas other models fail.

Strengths:<br /> This paper makes the point it sets out to - that there are some underlying representational bases, like grid cells, that when combined with a domain adaptation mechanism, like DPP-A, can facilitate out-of-generalisation. I don't have any issues with the technical details.

Weaknesses:<br /> The paper does leave open the bigger questions of 1) how one learns a suitable representation basis in the first place, 2) how to have a domain adaptation mechanism that works in more general settings other than adapting to scale. Overall, I'm left wondering whether this model is really quite bespoke or whether there is something really general here. My comments below are trying to understand how general this approach is.

COMMENTS<br /> This work relies on being able to map inputs into an appropriate representational space. The inputs were integers so it's easy enough to map them to grid locations. But how does this transfer to making analogies in other spaces? Do the inputs need to be mapped (potentially non-linearly) into a space where everything is linear? In general, what are the properties of the embedding space that allows the grid code to be suitable? It would be helpful to know just how much leg work an embedding model would have to do.

It's natural that grid cells are great for domain shifts of translation, rescaling, and rotation, because they themselves are multi-scaled and are invariant to translations and rotations. But grid codes aren't going to be great for other types of domain shifts. Are the authors saying that to make analogies grid cells are all you need? If not then what else? And how does this representation get learned? Are there lots of these invariant codes hanging around? And if so how does the appropriate one get chosen for each situation? Some discussion of the points is necessary as otherwise, the model seems somewhat narrow in scope.

For effective adaptation of scale, the authors needed to use DPP-A. Being that they are relating to brains using grid codes, what processes are implementing DPP-A? Presumably, a computational module that serves the role of DPP-A could be meta-learned? I.e. if they change their task set-up so it gets to see domain shifts in its training data an LSTM or transformer could learn to do this. The presented model comparisons feel a bit of a straw man.

I couldn't see it explained exactly how R works.

Reviewer #2 (Public Review):

Summary:<br /> This paper presents a model of out-of-distribution (OOD) generalization that focuses on modeling an analogy task, in which translation or scaling is tested with training in one part of the space and testing in other areas of the space progressively more distant from the training location. Similar tests were performed on arithmetic including addition and multiplication, and similarly impressive results appear for addition but not multiplication. The authors show that a grid cell coding scheme helps performance on these analogy and arithmetic tasks, but the most dramatic increase in performance is provided by a complex algorithm for distributional point-process attention (DPP-A) based on maximizing the determinant of the covariance matrix of the grid embeddings.

Strengths:<br /> The results appear quite impressive. The results for generalization appear quite dramatic when compared to other coding schemes (i.e. one-hot) or when compared to the performance when ablating the DPP-A component but retaining the same inference modules using LSTM or transformers. This appears to be an important result in terms of generalization of results in an analogy space.

Weaknesses:<br /> There are a number of ways that its impact and connection to grid cells could be enhanced. From the neuroscience perspective, the major comments concern making a clearer and stronger connection to the actual literature on grid cells and grid cell modeling, and discussing the relationship of the complex DPP-A algorithm to biological circuits.

Major comments:<br /> 1. They should provide more citations to other groups that have explored analogy using this type of task. Currently, they only cite one paper (Webb et al., 2020) by their own group in their footnote 1 which used the same representation of behavioral tasks for generalization of analogy. It would be useful if they could cite other papers using this simplified representation of analogy and also show the best performance of other algorithms from other groups in their figures, so that there is a sense of how their results compare to the best previous algorithm by other groups in the field (or they can identify which of their comparison algorithms corresponds to the best of previously published work).

2. While the grid code they use is very standard and based on grid cell researchers (Bicanski and Burgess, 2019), the rest of the algorithm doesn't have a clear claim on biological plausibility. It has become somewhat standard in the field to ignore the problem of how the brain could biologically implement the latest complex algorithm, but it would be useful if they at least mention the problem (or difficulty) of implementing DPP-A in a biological network. In particular, does maximizing the determinant of the covariance matrix of the grid code correspond to something that could be tested experimentally?

3. Related to major comment 2., it would be very exciting if they could show what the grid code looks like after the attentional modulation inner product xT w has been implemented. This could be highly useful for experimental researchers trying to connect these theoretical simulation results to data. This would be most intuitive to grid cell researchers if it is plotted in the same format as actual biological experimental data - specifically which grid cell codes get strengthened the most (beyond just the highest frequencies).

4. To enhance the connection to biological systems, they should cite more of the experimental and modeling work on grid cell coding (for example on page 2 where they mention relational coding by grid cells). Currently, they tend to cite studies of grid cell relational representations that are very indirect in their relationship to grid cell recordings (i.e. indirect fMRI measures by Constaninescu et al., 2016 or the very abstract models by Whittington et al., 2020). They should cite more papers on actual neurophysiological recordings of grid cells that suggest relational/metric representations, and they should cite more of the previous modeling papers that have addressed relational representations. This could include work on using grid cell relational coding to guide spatial behavior (e.g. Erdem and Hasselmo, 2014; Bush, Barry, Manson, Burges, 2015). This could also include other papers on the grid cell code beyond the paper by Wei et al., 2015 - they could also cite work on the efficiency of coding by Sreenivasan and Fiete and by Mathis, Herz, and Stemmler.

Reviewer #4 (Public Review):

In this manuscript by Sha et al. the authors test the role of TNFa in modulating tumor regression/recurrence under therapeutic pressure from castration (or enzalutamide) in both in vitro and in vivo models of prostate cancer. Using the PTEN-null genetic mouse model, they compare the effect of a TNFα ligand trap, etanercept, at various points pre- and post-castration. Their most interesting findings from this experiment were that etanercept given 3 days prior to castration prevented tumor regression, which is a common phenotype seen in these models after castration, but etanercept given 1 day prior to castration prevented prostate cancer recurrence after castration. They go on to perform RNA sequencing on tumors isolated from either sham or castrate mice from two time points post-castration to study acute and delayed transcriptional responses to androgen deprivation. They found enrichment of gene sets containing TNF-targets which initially decrease post-castration but are elevated by 35 days, the time at which tumors recur. The authors conduct a similar set of experiments using human prostate cancer cell lines treated with the androgen receptor inhibitor enzalutamide and observe that drug treatment leads to cells with basal stem-like features that express high levels of TNF. They noticed that CCL2 levels correlate with changes in TNF levels raising the possibility that CCL2 might be a critical downstream effector for disease recurrence. To this end, they treated PTEN-null and hi-MYC castrated mice with a CCR2-antagonist (CCR2a) because CCR2 is one receptor of CCL2 and monitors tumor growth dynamics. Interestingly, upon treatment with CCR2a, tumors did not recur according to their measurements. They go on to demonstrate that the tumors pre-treated with CCR2a had reduced levels of putative TAMs and increased CTLs in the context of TNF or CCR2 inhibition providing a cellular context associated with disease regression. Lastly, they perform single-cell RNA sequencing to further characterize the tumor microenvironment post-castration and report that the ratio of CTLs to TAMs is lower in a recurrent tumor.

While the concepts behind the study have merit, the data are incomplete and do not fully support the authors' conclusions. The author's definition of recurrence is subjective given that the amount of disease regression after castration is both variable (Figure 8) and relatively limited, particularly in the PTEN loss model. Critical controls are missing. For example, both drug experiments were completed without treating non-castrate plus drug controls which raises the question of how specific these findings are to castration resistance. No validation was performed to ensure that either the TNF ligand trap or the CCR2 agonist was acting on target. The single-cell sequencing experiments were done without replicates which raises concern about its interpretation. At a conceptual level, the authors say that a major cause of disease recurrence in the immunosuppressive TME, but provide little functional data that macrophages and T cells are directly responsible for this phenotype. Statistical analyses were performed on only select experiments. In summary, further work is recommended to support the conclusions of this story.

Reviewer #1 (Public Review):

Summary:<br /> Sha K et al aimed at identifying the mechanism of response and resistance to castration in the Pten knockout GEM model. They found elevated levels of TNF overexpressed in castrated tumors associated with an expansion of basal-like stem cells during recurrence, which they show occurring in prostate cancer cells in culture upon enzalutamide treatment. Further, the authors carry on a timed dependent analysis of the role of TNF in regression and recurrence to show that TNF regulates both processes. Similarly, CCL2, which the authors had proposed as a chemokine secreted upon TNF induction following enzalutamide treatment, is also shown to be elevated during recurrence and associated with the remodeling of an immunosuppressive microenvironment through depletion of T cells and recruitment of TAMs.

Strengths:

The paper exploits a well-established GEM model to interrogate mechanisms of response to standard-of-care treatment. This is of utmost importance since prostate cancer recurrence after ADT or ARSi marks the onset of an incurable disease stage for which limited treatments exist. The work is relevant in the confirmation that recurrent prostate cancer is mostly an immunologically "cold" tumor with an immunosuppressive immune microenvironment

Weaknesses:

While the data is consistent and the conclusions are mostly supported and justified, the findings overall are incremental and of limited novelty. The role of TNF and NF-kB signaling in tumor progression and the role of the CCL2-CCR2 in shaping the immunosuppressive microenvironment are well established.

On the other hand, it is unclear why the authors decided to focus on the basal compartment when there is a wealth of literature suggesting that luminal cells are if not exclusively, surely one of the cells of origin of prostate cancer and responsible for recurrence upon antiandrogen treatment. As a result, most of the later shown data has to be taken with caution as it is not known if the same phenomena occur in the luminal compartment.

Reviewer #2 (Public Review):

Summary:

In this study, Sha and Zhang et al. reported that androgen deprivation therapy (ADT) induces a switch to a basal-stemness status, driven by the TNF-CCL2-CCR2 axis. Their results also reveal that enhanced CCL2 coincides with increased macrophages and decreased CD8 T cells, suggesting that ADT resistance may be related to the TNF/CCL2/CCR2-dependent immunosuppressive tumor microenvironment (TME). Overall, this is a very interesting study with a significant amount of data.

Strengths:

The strengths of the study include various clinically relevant models, cutting-edge technology (such as single-cell RNA-seq), translational potential (TNF and CCR2 inhibitors), and novel insights connecting stemness lineage switch to an immunosuppressive TME. Thus, I believe this work would be of significant interest to the field of prostate cancer and journal readership.

Weaknesses:

(1) One of the key conclusions/findings of this study is the ADT-induced basal-stemness lineage switch driving ADT resistance. However, most of the presented evidence supporting this conclusion only selects a couple of marker genes. What exacerbates this issue is that different basal-stemness markers were often selected with different results. For example, Figure S1A uses CD166/EZH2 as markers, while Figure S1B uses ITGb1/EZH2. In contrast, Figure 1D uses Sca1/CD49, and Figure 2B-C uses CD49/CD166. Since many basal-stemness lineage gene signatures have been previously established, the study should examine various basal-stemness gene signatures rather than a couple of selected markers. Moreover, why were none of the stemness/basal-gene signatures significantly changed in the GO enrichment analysis in Figure 6A/B?

(2) A related weakness is the lack of functional results supporting the stemness lineage switch. Although the authors present colony formation assay results, these could be influenced simply by promoted cell proliferation, which is not a convincing indicator of stemness. To support this key conclusion, widely accepted stemness assays, such as the prostasphere formation assay (in vitro) and Extreme Limiting Dilution Analysis (ELDA) xenograft assay (in vivo), should be carried out.

(3) Another significant concern is that this study uses concurrency to demonstrate a causal relationship in many key results, which is entirely different. For example, Figure S4A and S4B only show increased CCL2 and TNF secretion simultaneously, which cannot support that CCL2 is dependent on TNF. Similarly, Figure 5A only shows that CCL2 increased coincidently with a rise in TNF, which cannot support a causal relationship. To support the causal relationship of this conclusion, it is necessary to show that TNF-KO/KD would abolish the increased CCL2 secretion.

(4) Some of the selective data presentations are not explained and are difficult to understand. For example, why does CD49 staining in Figure S3A have data for all four time points, while CD166 in Figure S3D only has data for the last time point (day 21)? Similarly, although several TNF_UP gene signatures were highlighted in Figure 4B, several TNF_DN signatures were also enriched in the same table, such as RUAN_RESPONSE_TO_TNF_DN. What is the explanation for these contrasting results?

Reviewer #3 (Public Review):

Summary:

The current manuscript evaluates the role of TNF in promoting AR targeted therapy regression and subsequent resistance through CCL2 and TAMs. The current evidence supports a correlative role for TNF in promoting cancer cell progression following AR inhibition. Weaknesses include a lack of descriptive methodology of the pre-clinical GEM model experiments and it is not well-defined which cell types are impacted in this pre-clinical model which will be quite heterogenous with regards to cancer, normal, and microenvironment cells.

Strengths:

(1) Appropriate use of pre-clinical models and GEM models to address the scientific questions.

(2) Novel finding of TNF and interplay of TAMs in promoting cancer cell progression following AR inhibition.

(3) Potential for developing novel therapeutic strategies to overcome resistance to AR blockade.

Weaknesses:

(1) There is a lack of description regarding the GEM model experiments - the age at which mice experiments are started.

(2) Tumor volume measurements are provided but in this context, there is no discussion on how the mixed cancer and normal epithelial and microenvironment is impacted by AR therapy which could lead to the subtle changes in tumor volume.

(3) There are no readouts for target inhibition across the therapeutic pre-clinical trials or dosing time courses.

(4) The terminology of regression and resistance appears arbitrary. The data seems to demonstrate a persistence of significant disease that progresses, rather than a robust response with minimal residual disease that recurs within the primary tumor.

(5) It is unclear if the increase in basal-like stem cells is from normal basal cells or cancer cells with a basal stem-like property.

6) In the Hi-MYC model, MYC expression is regulated by AR inhibition and is profoundly ARi responsive at early time points.

Reviewer #1 (Public Review):

Summary:

The authors use an interesting expression system called a retron to express single-stranded DNA aptamers. Expressing DNA as a single-stranded sequence is very hard - DNA is naturally double-stranded. However, the successful demonstration by the authors of expressing Lettuce, which is a fluorogenic DNA aptamer, allowed visual demonstration of both expression and folding. This method will likely be the main method for expressing and testing DNA aptamers of all kinds, including fluorogenic aptamers like Lettuce and future variants/alternatives.

Strengths:

This has an overall simplicity which will lead to ready adoption. I am very excited about this work. People will be able to express other fluorogenic aptamers or DNA aptamers tagged with Lettuce with this system.

Weaknesses:

Several things are not addressed/shown:

(1) How stable are these DNA in cells? Half-life?

(2) What concentration do they achieve in cells/copy numbers? This is important since it relates to the total fluorescence output and, if the aptamer is meant to bind a protein, it will reveal if the copy number is sufficient to stoichiometrically bind target proteins. Perhaps the gels could have standards with known amounts in order to get exact amounts of aptamer expression per cell?

(3) Microscopic images of the fluorescent E. coli - why are these not shown (unless I missed them)? It would be good to see that cells are fluorescent rather than just showing flow sorting data.

(4) I would appreciate a better Figure 1 to show all the intermediate steps in the RNA processing, the subsequent beginning of the RT step, and then the final production of the ssDNA. I did not understand all the processing steps that lead to the final product, and the role of the 2'OH.

(5) I would like a better understanding or a protocol for choosing insertion sites into MSD for other aptamers - people will need simple instructions.

(6) Can the gels be stained with DFHBI/other dyes to see the Lettuce as has been done for fluorogenic RNAs?

(7) Sometimes FLAPs are called fluorogenic RNA aptamers - it might be good to mention both terms initially since some people use fluorogenic aptamer as their search term.

(8) What E coli strains are compatible with this retron system?

(9) What steps would be needed to use in mammalian cells?

(10) Is the conjugated RNA stable and does it degrade to leave just the DNA aptamer?

Reviewer #2 (Public Review):

Summary:

This manuscript explores a DNA fluorescent light-up aptamer (FLAP) with the specific goal of comparing activity in vitro to that in bacterial cells. In order to achieve expression in bacteria, the authors devise an expression strategy based on retrons and test four different constructs with the aptamer inserted at different points in the retron scaffold. They only observe binding for one scaffold in vitro, but achieve fluorescence enhancement for all four scaffolds in bacterial cells. These results demonstrate that aptamer performance can be very different in these two contexts.

Strengths:

-Given the importance of FLAPs for use in cellular imaging and the fact that these are typically evolved in vitro, understanding the difference in performance between a buffer and a cellular environment is an important research question.

-The return strategy utilized by the authors is thoughtful and well-described.

-The observation that some aptamers fail to show binding in vitro but do show enhancement in cells is interesting and surprising.

Weaknesses:

-This study hints toward an interesting observation, but would benefit from greater depth to more fully understand this phenomenon. Particularly challenging is that FLAP performance is measured in vitro by affinity and in cells by enhancement, and these may not be directly proportional. For example, it may be that some constructs have much lower affinity but a greater enhancement and this is the explanation for the seemingly different performance.

-The authors only test enhancement at one concentration of fluorophore in cells (and this experimental detail is difficult to find and would be helpful to include in the figure legend). This limits the conclusions that can be drawn from the data and limits utility for other researchers aiming to use these constructs.

-The FLAP that is used seems to have a relatively low fluorescence enhancement of only 2-3 fold in cells. It would be interesting to know if this is also the case in vitro. This is lower than typical FLAPs and it would be helpful for the authors to comment on what level of enhancement is needed for the FLAP to be of practical use for cellular imaging.

Reviewer #1 (Public Review):

Summary:

The authors have developed a valuable method based on a fully cell-free system to express a channel protein and integrate it into a membrane vesicle in order to characterize it biophysically. The study presents a useful alternative to study channels that are not amenable to being studied by more traditional methods.

Strengths:

The evidence supporting the claims of the authors is solid and convincing. The method will be of interest to researchers working on ionic channels, allowing them to study a wide range of ion channel functions such as those involved in transport, interaction with lipids, or pharmacology.

Weaknesses:

The inclusion of a mechanistic interpretation of how the channel protein folds into a protomer or a tetramer to become functional in the membrane would strengthen the study.

Reviewer #2 (Public Review):

It is challenging to study the biophysical properties of organelle channels using conventional electrophysiology. The conventional reconstitution methods require multiple steps and can be contaminated by endogenous ionophores from the host cell lines after purification. To overcome this challenge, in this manuscript, Larmore et al. described a fully synthetic method to assay the functional properties of the TRPP channel family. The TRPP channels are an important organelle ion channel family that natively traffic to primary cilia and ER organelles. The authors utilized cell-free protein expression and reconstitution of the synthetic channel protein into giant unilamellar vesicles (GUV), the single channel properties can be measured using voltage-clamp electrophysiology. Using this innovative method, the authors characterized their membrane integration, orientation, and conductance, comparing the results to those of endogenous channels. The manuscript is well-written and may present broad interest to the ion channel community studying organelle ion channels. Particularly because of the challenges of patching native cilia cells, the functional characterization is highly concentrated in very few labs. This method may provide an alternative approach to investigate other channels resistant to biophysical analysis and pharmacological characterization.

Reviewer #1 (Public Review):

In this important study, Huffer et al posit that non-cold sensing members of the TRPM subfamily of ion channels (e.g., TRPM2, TRPM4, TRPM5) contain a binding pocket for icilin which overlaps with the one found in the cold-activated TRPM8 channel.

The authors identify the residues involved in icilin binding by analyzing the existing TRPM8-icilin complex structures and then use their previously published approach of structure-based sequence comparison to compare the icilin binding residues in TRPM8 to other TRPM channels. This approach uncovered that the residues are conserved in a number of TRPM members: TRPM2, TRPM4, and TRPM5. The authors focus on TRPM4, with the rationale that it has the simplest activation properties (a single Ca2+-binding site). Electrophysiological studies show that icilin by itself does not activate TRPM4, but it strongly potentiates the Ca2+ activation of TRPM4, and introducing the A867G mutation (the mutation that renders avian TRPM8 sensitive to icilin) further increases the potentiating effects of the compound. Conversely, the mutation of a residue that likely directly interacts with icilin in the binding pocket, R901H, results in channels whose Ca2+ sensitivity is not potentiated by icilin.

The data indicate that, just like in TRPV channels, the binding pockets and allosteric networks might be conserved in the TRPM subfamily.

The data are convincing, and the authors employ good experimental controls.

Reviewer #2 (Public Review):

Summary:

The authors set out to study whether the cooling agent binding site in TRPM8, which is located between the S1-S4 and the TRP domain, is conserved within the TRPM family of ion channels. They specifically chose the TRPM4 channel as the model system, which is directly activated by intracellular Ca2+. Using electrophysiology, the authors characterized and compared the Ca2+ sensitivity and the voltage dependence of TRPM4 channels in the absence and presence of synthetic cooling agonist icilin. They also analyzed the mutational effects of residues (A867G and R901H; equivalent mutations in TRPM8 were shown involved in icilin sensitivity) on Ca2+ sensitivity and voltage-dependence of TRPM4 in the absence and presence of Ca2+. Based on the results as well as structure/sequence alignment, the authors concluded that icilin likely binds to the same pocket in TRPM4 and suggested that this cooling agonist binding pocket is conserved in TRPM channels.

Strengths:

The authors gave a very thorough introduction to the TRPM channels. They have nicely characterized the Ca2+ sensitivity and the voltage-dependence of TRPM4 channels and demonstrated icilin potentiates the Ca2+ sensitivity and diminishes the outward rectification of TRPM4. These results indicate icilin modulates TRPM4 activation by Ca2+.

Weaknesses:

The reviewer has a few concerns. First, icilin alone (at 25µM) and in the absence of Ca2+ does not activate the TRPM4 channel. Have the authors titrated a wide range of icilin concentrations (without Ca2+ present) for TRPM4 activation? It raises the question that whether icilin is indeed an agonist for TRPM4 channel. This has not been tested so it is unclear. One may argue that icilin needs Ca2+ as a co-factor for channel activation just like in TRPM8 channel. This leads to the second concern, which is a complication in the experimental design and data interpretation. TRPM4 itself requires Ca2+ for activation to begin with, thus it is hard to dissect whether the current observed here for TRPM4 is activated by Ca2+ or by icilin plus its cofactor Ca2+. This is the difference between TRPM8 and TRPM4, as TRPM8 itself is not activated by Ca2+, thus TRPM8 activation is through icilin and Ca2+ acts as a prerequisite for icilin activation.

The results presented in this study are only sufficient to show that icilin modulates the Ca2+-dependent activation of TRPM4 and icilin at best may act as an allosteric modulator for TRPM4 function. One cannot conclude from the current work that icilin is an agonist or even specifically a cooling agonist for TRPM4. Icilin is a cooling agonist for TRPM8, but it does not mean that if icilin modulates TRPM4 activity then it serves as a cooling agonist for TRPM4.

For the mutation data on A867G, Figure 4A-B, left panels, it looks like A867G has stronger Ca2+ sensitivity compared to the WT in the absence of icilin and the onset of current activation is faster than the WT, or this is simply due to the scale of the data figure are different between A867G and the WT. Overall the mutagenesis data are weak to support the conclusion that icilin binds to the S1-S4 pocket. The authors need to mutate more residues that are involved in direct interaction with icilin based on the available structural information, including but limited to residues equivalent to Y745 and H845 in human TRPM8.

The authors set out to study the conservation of the cooling agonist binding site in TRPM family, but only tested a synthetic cooling agonist icilin on TRPM4. In order to draw a broad conclusion as the title and the discussion have claimed, the authors need to more cooling compounds, including the most well-known natural cooling agonist menthol, and other cooling agonists such as WS-12 and/or C3, and test their effects on several TRPM channels, not just TRPM4. With the current data, the authors need to significantly tone down the claim of a conserved cooling agonist binding pocket in the TRPM family.

On page 11, the authors suggest based on the current data, that TRPM2 and TRPM5 may also be sensitive to cooling agonists because the key residues are conserved. TRPM2 is the closest homolog to TRPM8 but is menthol-insensitive. There are studies that attempted to convert menthol sensitivity to TRPM2, for example, Bandell 2006 attempted to introduce S2 and TRP domains from TRPM8 into TRPM2 but failed to make TRPM2 a menthol-sensitive channel. The sequence conservation or structural similarity is not sufficient for the authors to suggest a shared cooling agonist sensitivity or even a common binding site in the TRPM2 and TRPM5 channels. Again, as pointed out above, the authors need to establish the actual activation of other TRPM channels by these agonists first, before proceeding to functionally probe whether other TRPM channels adopt a conserved agonist binding site.

Taken together, this current work presents data to show the modulatory effects of icilin on the Ca2+ dependent activation and voltage dependence of the TRPM4 channel.

Reviewer #3 (Public Review):

Summary:

The family of transient receptor potential (TRP) channels are tetrameric cation selective channels that are modulated by a variety of stimuli, most notably temperature. In particular, the Transient receptor potential Melastatin subfamily member 8 (TRPM8) is activated by noxious cold and other cooling agents such as menthol and icilin and participates in cold somatosensation in humans. The abundance of TRP channel structural data that has been published in the past decade demonstrates clear architectural conservation within the ion channel family. This suggests the potential for unifying mechanisms of gating despite their varied modes of regulation, which are not yet understood. To address this question, the authors examine the 264 structures of TRP channels determined to date and observe a potential binding pocket for icilin in multiple members of the Melastatin subfamily, TRPM2, TRPM4, and TRPM5. Interestingly, none of the other Melastatin subfamily members had been shown to be sensitive to icilin apart from TRPM8. Each of these channels is activated by intracellular calcium (Ca2+) and a Ca2+ binding site neighbors the predicted pocket for icilin binding in all cryo-EM structures. The authors examined whether icilin could modulate the activation of TRPM4 in the presence of intracellular Ca2+. The addition of icilin enhances Ca2+-dependent activation of TRPM4, promotes channel opening at negative membrane potentials, and improves the kinetics of opening. Furthermore, mutagenesis of TRPM4 residues within the putative icilin binding pocket predicted to enhance or diminish TRPM4 activity elicit these behaviors. Overall, this study furthers our understanding of the Melastatin subfamily of TRP channel gating and demonstrates that a conserved binding pocket observed between TRPM4 and TRPM8 channel structures can function similarly to regulate channel gating.

Strengths:

This is a simple and elegant study capitalizing on a vast amount of high-resolution structural information from the TRP channel of ion channels to identify a conserved binding pocket that was previously unknown in the Melastatin subfamily, which is interrogated by the authors through careful electrophysiology and mutagenesis studies.

Weaknesses:

No weaknesses were identified by this reviewer.

Reviewer #1 (Public Review):

In their manuscript "PDGFRRa signaling regulates Srsf3 transcript binding to affect PI3K signaling and endosomal trafficking" Forman and colleagues use iMEPM cells to characterize the effects of PDGF signaling on alternative splicing. They first perform RNA-seq using a one-hour stimulation with Pdgf-AA in control and Srsf3 knockdown cells. While Srsf3 manipulation results in a sizeable number of DE genes, PDGF does not. They then turn to examine alternative splicing, due to findings from this lab. They find that both PDGF and Srsf3 contribute much more to splicing than transcription. They find that the vast majority of PDGF-mediated alternative splicing depends upon Srsf3 activity and that skipped exons are the most common events with PDGF stimulation typically promoting exon skipping in the presence of Srsf3. They used eCLIP to identify RNA regions bound to Srsf3. Under both PDGF conditions, the majority of peaks were in exons with +PDGF having a substantially greater number of these peaks. Interestingly, they find differential enrichment of sequence motifs and GC content in stimulated versus unstimulated cells. They examine 2 transcripts encoding PI3K pathway (enriched in their GO analysis) members: Becn1 and Wdr81. They then go on to examine PDGFRRa and Rab5, an endosomal marker, colocalization. They propose a model in which Srsf3 functions downstream of PDGFRRa signaling to, in part, regulate PDGFRa trafficking to the endosome. The findings are novel and shed light on the mechanisms of PDGF signaling and will be broadly of interest. This lab previously identified the importance of PDGF naling on alternative splicing. The combination of RNA-seq and eCLIP is an exceptional way to comprehensively analyze this effect. The results will be of great utility to those studying PDGF signaling or neural crest biology. There are some concerns that should be considered, however.

(1) It took some time to make sense of the number of DE genes across the results section and Figure 1. The authors give the total number of DE genes across Srsf3 control and loss conditions as 1,629 with 1,042 of them overlapping across Pdgf treatment. If the authors would add verbiage to the point that this leaves 1,108 unique genes in the dataset, then the numbers in Figure 1D would instantly make sense. The same applies to PDGF in Figure 1F and the Venn diagrams in Figure 2.

(2) The percentage of skipped exons in the +PSI on the righthand side of Figure 2F is not readable.

(3) It would be useful to have more information regarding the motif enrichment in Figure 3. What is the extent of enrichment? The authors should also provide a more complete list of enriched motifs, perhaps as a supplement.

(4) It is unclear what subset of transcripts represent the "overlapping datasets" on lines 280-315. The authors state that there are 149 unique overlapping transcripts, but the Venn diagram shows 270. Also, it seems that the most interesting transcripts are the 233 that show alternative splicing and are bound by Srsf3. Would the results shown in Figure 5 change if the authors focused on these transcripts?

(5) In general, there is little validation of the sequencing results, performing qPCR on Arhgap12 and Cep55. The authors should additionally validate the PI3K pathway members that they analyze. Related, is Becn1 expression downregulated in the absence of Srsf3, as would be predicted if it is undergoing NMD?

(6) What is the alternative splicing event for Acap3?

(7) The insets in Figure 6 C"-H" are useful but difficult to see due to their small size. Perhaps these could be made as their own figure panels.

(8) In Figure 6A, it is not clear which groups have statistically significant differences. A clearer visualization system should be used.

(9) Similarly in Figure 6B, is 15 vs 60 minutes in the shSrsf3 group the only significant difference? Is there a difference between scramble and shSrsf3 at 15 minutes? Is there a difference between 0 and 15 minutes for either group?

Reviewer #2 (Public Review):

Summary:

This manuscript builds upon the work of a previous study published by the group (Dennison, 2021) to further elucidate the coregulatory axis of Srsf3 and PDGFRa on craniofacial development. The authors in this study investigated the molecular mechanisms by which PDGFRa signaling activates the RNA-binding protein Srsf3 to regulate alternative splicing (AS) and gene expression (GE) necessary for craniofacial development. PDGFRa signaling-mediated Srsf3 phosphorylation drives its translocation into the nucleus and affects binding affinity to different proteins and RNA, but the exact molecular mechanisms were not known. The authors performed RNA sequencing on immortalized mouse embryonic mesenchyme (MEPM) cells treated with shRNA targeting 3' UTR of Srsf3 or scramble shRNA (to probe AS and DE events that are Srsf3 dependent) and with and without PDGF-AA ligand treatment (to probe AS and DE events that are PDGFRa signaling dependent). They found that PDGFRa signaling has more effect on AS than on DE. A matching eCLIP-seq experiment was performed to investigate how Srsf3 binding sites change with and without PDGFRa signaling.

Strengths:

(1) The work builds well upon the previous data and the authors employ a variety of appropriate techniques to answer their research questions.

(2) The authors show that Srsf3 binding pattern within the transcript as well as binding motifs change significantly upon PDGFRa signaling, providing a mechanistic explanation for the significant changes in AS.

(3) By combining RNA-seq and eCLIP datasets together, the authors identified a list of genes that are directly bound by Srsf3 and undergo changes in GE and/or AS. Two examples are Becn1 and Wdr81, which are involved in early endosomal trafficking.

Weaknesses:

(1) The authors identify two genes whose AS are directly regulated by Srsf3 and involved in endosomal trafficking; however, they do not validate the differential AS results and whether changes in these genes can affect endosomal trafficking. In Figure 6, they show that PDGFRa signaling is involved in endosome size and Rab5 colocalization, but do not show how Srsf3 and the two genes are involved.

(2) The proposed model does not account for other proteins mediating the activation of Srsf3 after Akt phosphorylation. How do we know this is a direct effect (and not a secondary or tertiary effect)?

Reviewer #1 (Public Review):

Summary:

Liver cancer shows a higher incidence in males than females with incompletely understood causes. This study utilized a mouse model that lacks the bile acid feedback mechanisms (FXR/SHP DKO mice) to study how dysregulation of bile acid homeostasis and a high circulating bile acid may underlie the gender-dependent prevalence and prognosis of HCC. By transcriptomics analysis comparing male and female mice, unique sets of gene signatures were identified and correlated with HCC outcomes in human patients. The study showed that the ovariectomy procedure increased HCC incidence in female FXR/SHP DKO mice that were otherwise resistant to age-dependent HCC development and that removing bile acids by blocking intestine bile acid absorption reduced HCC progression in FXR/SHP DKO mice. Based on these findings, the authors suggest that gender-dependent bile acid metabolism may play a role in the male-dominant HCC incidence, and that reducing bile acid levels and signaling may be beneficial in HCC treatment.

Strengths:

(1) Chronic liver diseases often preceed the development of liver and bile duct cancer. Advanced chronic liver diseases are often associated with dysregulation of bile acid homeostasis and cholestasis. This study takes advantage of a unique FXR/SHP DKO model that develops high organ bile acid exposure and spontaneous age-dependent HCC development in males but not females to identify unique HCC-associated gene signatures. The study showed that the unique gene signature in female DKO mice that had lower HCC incidence also correlated with lower-grade HCC and better survival in human HCC patients.

(2) The study also suggests that differentially regulated bile acid signaling or gender-dependent response to altered bile acids may contribute to gender-dependent susceptibility to HCC development and/or progression.

Weaknesses:

(1) HCC shows heterogeneity, and it is unclear what tissues (tumor or normal) were used from the DKO mice and human HCC gene expression dataset to obtain the gene signature, and how the authors reconcile these gene signatures with HCC prognosis.

(2) The authors identified a unique set of gene expression signatures that are linked to HCC patient outcomes, but analysis of these gene sets to understand the causes of cancer promotion is still lacking. The studies of urea cycle metabolism and estrogen signaling were preliminary and inconclusive. These mechanistic aspects may be followed up in revision or future studies.

(3) While high levels of bile acids are convincingly shown to promote HCC progression, their role in HCC initiation is not established. The DKO model may be limited to conditions of extremely high levels of organ bile acid exposure. The DKO mice do not model the human population of HCC patients with various etiology and shared liver pathology (i.e. cirrhosis). Therefore, high circulating bile acids may not fully explain the male prevalence of HCC incidence.

(4) The authors showed lower circulating bile acids and increased fecal bile acid excretion in female mice and hypothesized that this may be a mechanism underlying the lower bile acid exposure that contributed to lower HCC incidence in female DKO mice. Additional analysis of organ bile acids within the enterohepatic circulation may be performed because a more accurate interpretation of the circulating bile acids and fecal bile acids can be made in reference to organ bile acids and total bile acid pool changes in these mice.

Reviewer #2 (Public Review):

Summary:

The manuscript of Patton et al. shows that in mice in which both FXR and SHP are knocked out, the sex difference in liver cancer risk is recapitulated. Authors show that the protection against tumor development seen in female mice is dependent upon ovarian hormone secretion and higher fecal bile acid excretion in females compared to males. The female liver-specific gene signature correlates with low-grade tumors and better survival in human HCC patients.

The combination of the use of the double knockout mice together with ovariectomy in female mice and using a bile acid raisin in male mice to underscore their conclusion is strong. However, there are also some shortcomings, that should be addressed.

Strengths:

(1) Using computational modelling, Patton and colleagues correlate mouse DKO transcriptome data to the clinical outcomes of HCC patients using HCC transcriptome datasets.

(2) The dependence of female protection on ovarian hormones and increased fecal bile acid excretion is nicely shown by combining ovariectomy and bile acid raisin with the use of double knockout mice.

Weaknesses:

(1) The translational value to human HCC is not so strong yet. Authors show that there is a correlation between the female-selective gene signature and low-grade tumors and better survival in HCC patients overall. However, these data do not show whether this signature is more highly correlated with female tumor burden and survival. In other words, whether the mechanisms of female protection may be similar between humans and mice. In that respect, it would also be good to elaborate on whether women have higher fecal BA excretion and lower serum BA concentration.

(2) The authors should perform a thorough spelling and grammar check.

(3) There are quite some errors and inaccuracies in the result section, figures, and legends. The authors should correct this.

Reviewer #1 (Public Review):

Summary:

Wilson's Disease (WD) is an inherited rare pathological condition due to a mutation in ATP7B that alters mitochondrial structure and dysfunction. Additionally, WD results in dysregulated copper metabolism in patients. These metabolic abnormalities affect the functions of the liver and can result in cholecystitis. Understanding the immune component and its contribution to WD and cholecystitis has been challenging. In this work, the authors have performed single-cell RNA sequencing of mesenchymal tissue from three WD patients and three liver hemangioma patients.

Strengths:

The authors describe the transcriptomic alterations in myeloid and lymphoid compartments.

Weaknesses:

In brief, this manuscript lacks a clear focus, and the writing needs vast improvement. Figures lack details (or are misrepresented), the results section only catalogs observations, and the discussion needs to focus on their findings' mechanistic and functional relevance. The major weakness of this manuscript is that the authors do not provide a mechanistic link between the absence of ATP7B and NK cells' impaired/altered functions. While the work is of high clinical relevance, there are various areas that could be improved.

Reviewer #2 (Public Review):

Summary:

Wilson's disease is a rare genetic disorder caused by mutations in the ATP7B gene. Previous studies have documented that ATP7B mutations can disrupt copper metabolism, affecting brain and liver function. In this paper, the authors performed a retrospective clinical study and found that Wilson's disease has a high incidence of cholecystitis. Single-cell RNA-seq analysis revealed changes in the immune microenvironment, including the activation of immune responses and the exhaustion of natural killer cells.

Strengths:

A key finding of this study is that the predominant ATP7B gene mutation in the Chinese population is the 2333G>T (p. R778L) mutation. The authors reported associations between Wilson's disease and cholecystitis, as well as the exhaustion of natural killer cells.

Weaknesses:

The underlying mechanisms linking ATP7B mutations to cholecystitis and natural killer cell exhaustion remain unclear. Specifically, it is not yet determined whether copper metabolism alterations directly cause cholecystitis and natural killer cell exhaustion, or if these effects are secondary to liver dysfunction.

Reviewer #1 (Public Review):

Summary:<br /> This manuscript aimed to investigate the emergence of emotional sensitivity and its relationship with gestational age. Using an oddball paradigm and event-related potentials, the authors conducted an experiment in 120 healthy neonates with a gestational age range of 35 to 40 weeks. A significant developmental milestone was identified at 37 weeks gestational age, marking a crucial juncture in neonatal emotional responsiveness.

Strengths:<br /> This study has several strengths, by providing profound insights into the early development of social-emotional functioning and unveiling the role of gestational age in shaping neonatal perceptual abilities. The methodology of this study demonstrates rigor and well-controlled experimental design, particularly involving matched control sounds, which enhances the reliability of the research. Their findings not only contribute to the field of neurodevelopment, but also showcase potential clinical applications, especially in the context of autism screening and early intervention for neurodevelopmental disorders.

Comments on the revised version:

After reviewing the authors' response letter and the revised manuscript, I believe they have done a commendable job in addressing my comments.<br /> Additionally, I concur with the concerns raised by Reviewer #2 regarding several potential confounding factors that require better control in their experimental design. These include the differences in physical properties between vocal and nonvocal stimuli, as well as the infant's exposure to the speech/auditory environment. These concerns should be thoroughly and explicitly discussed in the manuscript, ensuring a clearer understanding for the readers.

Reviewer #2 (Public Review):

This is an important and very interesting report on a change in newborns' neural abilities to distinguish auditory signals as a function of the gestational age (GA) of the infant at birth (from 35 weeks GA to 40 weeks GA). The authors tested neural discrimination of sounds that were labeled 'happy' vs 'neutral' by listeners that represent two categories of sound, either human voices or auditory signals that mimic only certain properties of the human vocal signals. The finding is that a change occurs in neural discrimination of the happy and neutral auditory signals for infants born at or after 37 weeks of gestation, and not prior (at 35 or 36 weeks of gestation), and only for discrimination of the human vocal signals; no change occurs in discrimination of the nonhuman signals over the 35- to 40-week gestational ages tested. The neural evidence of discrimination of the vocal happy-neutral distinction and the absence of the discrimination of the control signals is convincing. The authors interpret this as a 'landmark' in infants' ability to detect changes in emotional vocal signals, and remark on the potential value of the test as a marker of the infants' interest in emotional signals, underscoring the fact that children at risk for autism spectrum disorder may not show the discrimination. Although the finding is novel and interesting, additional discussion is essential so that readers understand two potential caveats affecting this interpretation.

Comments on the revised version:

The revised manuscript does discuss the limitations of the control stimuli, as well as the limitations with regard to conclusions that can be drawn from this data set. I therefore expected the authors to temper a bit their recommendation that this could be a 'screening' signal for autism because these data are not sufficiently strong to make that recommendation. Also, in the same vein, perhaps the title might be adjusted somewhat to suggest less certainty, for example, by using the word "change" rather than "milestone"'? The data are of interest, but the limitations are genuine limitations.

Reviewer #1 (Public Review):

Summary:

In this manuscript, Tung and colleagues identify Calreticulin as a repressor of ATF6 signaling using a crispr screen and characterize the functional interaction between ATF6 and CALR.

Strengths:

The manuscript is well written and interesting with an innovative experimental design which provides some new mechanistic insight into ATF6 regulation as well as crosstalk with the IRE1 pathway. The methods used were fit for purpose and reasonable conclusions were drawn from the data presented.

Comments on latest version:

The authors did a good job at addressing my comments even though they found several aspects to exceed the scope of the work. The manuscript is clearer now and the model pushed by the authors is better supported by the data. One point I am curious about the authors' opinion would be about the status of ATF6alpha activation in pathological cells in which CALR is mutated (e.g., myeloproliferative neoplasms), although this neither challenges the conclusions of the manuscript and my positive opinion of the work.

Reviewer #3 (Public Review):

In this study, Ruan et al. investigate the role of the IQCH gene in spermatogenesis, focusing on its interaction with calmodulin and its regulation of RNA-binding proteins. The authors examined sperm from a male infertility patient with an inherited IQCH mutation as well as Iqch CRISPR knockout mice. The authors found that both human and mouse sperm exhibited structural and morphogenetic defects in multiple structures, leading to reduced fertility in Ichq-knockout male mice. Molecular analyses such as mass spectrometry and immunoprecipitation indicated that RNA-binding proteins are likely targets of IQCH, with the authors focusing on the RNA-binding protein HNRPAB as a critical regulator of testicular mRNAs. The authors used in vitro cell culture models to demonstrate an interaction between IQCH and calmodulin, in addition to showing that this interaction via the IQ motif of IQCH is required for IQCH's function in promoting HNRPAB expression. In sum, the authors concluded that IQCH promotes male fertility by binding to calmodulin and controlling HNRPAB expression to regulate the expression of essential mRNAs for spermatogenesis. These findings provide new insight into molecular mechanisms underlying spermatogenesis and how important factors for sperm morphogenesis and function are regulated.

The strengths of the study include the use of mouse and human samples, which demonstrate a likely relevance of the mouse model to humans; the use of multiple biochemical techniques to address the molecular mechanisms involved; the development of a new CRISPR mouse model; ample controls; and clearly displayed results. Assays are done rigorously and in a quantitative manner. Overall, the claims made by the authors in this manuscript are well-supported by the data provided.

Reviewer #3 (Public Review):

Summary:

This study used prolonged stimulation of a limb to examine possible plasticity in somatosensory evoked potentials induced by the stimulation. They also studied the extent that the blood brain barrier (BBB) was opened by the prolonged stimulation and whether that played a role in the plasticity. They found that there was potentiation of the amplitude and area under the curve of the evoked potential after prolonged stimulation and this was long-lasting (>5 hrs). They also implicated extravasation of serum albumin, caveolae-mediated transcytosis, and TGFb signalling, as well as neuronal activity and upregulation of PSD95. Transcriptomics was done and implicated plasticity related genes in the changes after prolonged stimulation, but not proteins associated with the BBB or inflammation. Next, they address the application to humans using a squeeze ball task. They imaged the brain and suggested that the hand activity led to an increased permeability of the vessels, suggesting modulation of the BBB.

Strengths:

The strengths of the paper are the novelty of the idea that stimulation of the limb can induce cortical plasticity in a normal condition, and it involves the opening of the BBB with albumin entry. In addition, there are many datasets, both rat and human data.

Weaknesses:

The explanation of why prolonged stimulation in the rat was considered relevant to normal conditions is still somewhat weak. The authors argue that the stimulation frequency they used is similar to rhythmic whisker movement. That is a good argument. However, the intensity they used, 2 mA is in the range they say can elicit a seizure if stimulation is 50 Hz. So that weakens the argument.

The authors made a lot of the requested changes but some questions were not addressed or the explanations were so brief that the confusion remained. Please go over the revisions again and make sure sentences are complete, jargon is explained, and arguments/justifications are clear. It will help the reader greatly.

The authors responded to the previous comments of Reviewer 2 regarding experimental design and variability of washout periods. It would be useful to incorporate the response into the paper so the readers know why the authors think the variability was not an important factor in the results.

Comments on the revised version:

The manuscript is improved.

Reviewer #1 (Public Review):

Wang, He et al have constructed a comprehensive single nucleus atlas for the gills of the deep sea Bathymodioline mussels, which possess intracellular symbionts that provide a key source of carbon and allow them to live in these extreme environments. They provide annotations of the different cell states within the gills, shedding light on how multiple cell types cooperate to give rise to the emergent functions of the composite tissues and the gills as a whole. They pay special attention to characterizing the bacteriocyte cell populations and identifying sets of genes that may play a role in their interaction with the symbiotes.

Wang, He et al sample mussels from 3 different environments: animals from their native methane rich environment, animals transplanted to a methane-poor environment to induce starvation and animals that have been starved in the methane-poor environment and then moved back to the methane-rich environment. They demonstrated that starvation had the biggest impact on bacteriocyte transcriptomes. They hypothesize that the up-regulation of genes associated with lysosomal digestion leads to the digestion of the intracellular symbiont during starvation, while the non-starved and reacclimated groups more readily harvest the nutrients from symbiotes without destroying them. Further work exploring the differences in symbiote populations between ecological conditions will further elucidate the dynamic relationship between host and symbiote. This will help disentangle specific changes in transcriptomic state that are due to their changing interactions with the symbiotes from changes associated with other environmental factors.

This paper makes available a high quality dataset that is of interest to many disciplines of biology. The unique qualities of this non-model organism and collection of conditions sampled make it of special interest to those studying deep sea adaptation, the impact of environmental perturbation on Bathymodioline mussels populations, and intracellular symbiotes. The authors also use a diverse array of tools to explore and validate their data.

Reviewer #2 (Public Review):

Wang, He et al. shed insight into the molecular mechanisms of deep-sea chemosymbiosis at the single-cell level. They do so by producing a comprehensive cell atlas of the gill of Gigantidas platifrons, a chemosymbiotic mussel that dominates the deep-sea ecosystem. They uncover novel cell types and find that the gene expression of bacteriocytes, the symbiont-hosting cells, supports two hypotheses of host-symbiont interactions: the "farming" pathway, where symbionts are directly digested, and the "milking" pathway, where nutrients released by the symbionts are used by the host. They perform an in situ transplantation experiment in the deep sea and reveal transitional changes in gene expression that support a model where starvation stress induces bacteriocytes to "farm" their symbionts, while recovery leads to the restoration of the "farming" and "milking" pathways.

A major strength of this study includes the successful application of advanced single nucleus techniques to a non-model, deep sea organism that remains challenging to sample. I also applaud the authors for performing an in situ transplantation experiment in a deep sea environment. From gene expression profiles, the authors deftly provide a rich functional description of G. platifrons cell types that is well-contextualized within the unique biology of chemosymbiosis. These findings offer significant insight into the molecular mechanisms of deep-sea host-symbiont ecology, and will serve as a valuable resource for future studies into the striking biology of G. platifrons.

The authors' conclusions are generally well-supported by their results. However, I recognize that the difficulty of obtaining deep-sea specimens may have impacted experimental design and no replicates were sampled.

It is notable that the Fanmao cells were much more sparsely sampled. It appears that fewer cells were sequenced, resulting in the Starvation and Reconstitution conditions having 2-3x more cells after doublet filtering. These discrepancies also are reflected in the proportion of cells that survived QC, suggesting a distinction in quality or approach. However, the authors provide clear and sufficient evidence via bootstrapping that batch effects between the three samples are negligible. While batch effect does not appear to have affected gene expression profiles, the proportion of cell types may remain sensitive to sampling techniques, and thus interpretation of Fig. S12 must be approached with caution.

Reviewer #3 (Public Review):

Wang et al. explored the unique biology of the deep-sea mussel Gigantidas platifrons to understand fundamental principles of animal-symbiont relationships. They used single-nucleus RNA sequencing and validation and visualization of many of the important cellular and molecular players that allow these organisms to survive in the deep-sea. They demonstrate that a diversity of cell types that support the structure and function of the gill including bacteriocytes, specialized epithelial cells that host sulfur-oxidizing or methane-oxidizing symbionts as well as a suite of other cell types including supportive cells, ciliary, and smooth muscle cells. By performing experiments of transplanting mussels from one habitat which is rich in methane to methane-limited environments, the authors showed that starved mussels may consume endosymbionts versus in methane-rich environments upregulated genes involved in glutamate synthesis. These data add to the growing body of literature that organisms control their endosymbionts in response to environmental change.

The conclusions of the data are well supported. The authors adapted a technique that would have been technically impossible in their field environment by preserving the tissue and then performing nuclear isolation after the fact. The use of single-nucleus sequencing opens the possibility of new cellular and molecular biology that is not possible to study in the field. Additionally, the in-situ data (both WISH and FISH) are high-quality and easy to interpret. The use of cell-type-specific markers along with a symbiont-specific probe was effective. Finally, the SEM and TEM were used convincingly for specific purposes in the case of showing the cilia that may support water movement.

The one particular area for future exploration surrounds the concept of a proliferative progenitor population within the gills. The authors recover molecular markers for these putative populations and additional future work will uncover if these are indeed proliferative cells that contribute to symbiont colonization.

Overall the significance of this work is identifying the relationship between symbionts and bacteriocytes and how these host bacteriocytes modulate their gene expression in response to environmental change. It will be interesting to see how similar or different these data are across animal phyla. For instance, the work of symbiosis in cnidarians may converge on similar principles of there may be independent ways in which organisms have been able to solve these problems.

Reviewer #1 (Public Review):

Summary:

The manuscript by Dubicka and co-workers on calcification in miliolid foraminifera presents an interesting piece of work. The study uses confocal and electron microscopy to show that the traditional picture of calcification in porcelaneous foraminifera is incorrect.

Strengths:<br /> The authors present high-quality images and an original approach to a relatively solid (so I thought) model of calcification.

Weaknesses:

There are several major shortcomings. Despite the interesting subject and the wonderful images, the conclusions of this manuscript are simply not supported at all by the results. The fluorescent images may not have any relation to the process of calcification and should therefore not be part of this manuscript. The SEM images, however, do point to an outdated idea of miliolid calcification. I think the manuscript would be much stronger with the focus on the SEM images and with the speculation of the physiological processes greatly reduced.

Comments on revised version:

I continue to disagree. As the authors acknowledge: 'may be a hint indicating ACC...', but it may also be something else. This is really something else than showing ACC is involved in foraminiferal calcification. I still think the reasoning is shaky and below, I will clarify why the fluorescence may well not be related to ACC and in fact, some or even most of the vesicles may not play the role that the authors suggest. Even if they do, the conclusions are not supported by the data presented here. Unfortunately, I found some of the other answers to my question not satisfactory either.

Reviewer #2 (Public Review):

Summary:

Dubicka et al. in their paper entitled " Biocalcification in porcelaneous foraminifera" suggest that in contrast to the traditionally claimed two different modes of test calcification by rotallid and porcelaneous miliolid formaminifera, both groups produce calcareous tests via the intravesicular mineral precursors (Mg-rich amorphous calcium carbonate). These precursors are proposed to be supplied by endocytosed seawater and deposited in situ as mesocrystals formed at the site of new wall formation within the organic matrix. The authors did not observe the calcification of the needles within the transported vesicles, which challenges the previous model of miliolid mineralization. Although the authors argue that these two groups of foraminifera utilize the same calcification mechanism, they also suggest that these calcification pathways evolved independently in the Paleozoic.

Comments on the revised version

In my reply to the author's rebuttal letter, I will focus on one key point. The main observation supporting the author's conclusion, as expressed in the abstract, is:

"We found that both groups [i.e., rotaliids and miliolids, the latter documented in the reviewed paper] produced calcareous shells via the intravesicular formation of unstable mineral precursors (Mg-rich amorphous calcium carbonates) supplied by endocytosed seawater and deposited at the site of new wall formation within the organic matrix. Precipitation of high-Mg calcitic mesocrystals took place in situ and formed a dense, chaotic meshwork of needle-like crystallites."

In my review, I pointed out that there is no support for the existence of an intracellular, vesicular intermediate amorphous phase.

The authors replied:

"We used laser line 405 nm and multiphoton excitation to detect ACCs. These wavelengths (partly) permeate the shell to excite ACCs autofluorescence. The autofluorescence of the shells is present as well but not clearly visible in movie S4 as the fluorescence of ACCs is stronger. This may be related to the plane/section of the cell which is shown. The laser permeates the shell above the ACCs (short distance) but to excite the shell CaCO3 around foraminifera in the same three-dimensional section where ACCs are shown, the light must pass a thick CaCO3 area due to the three-dimensional structure of the foraminiferan shell. Therefore, the laser light intensity is reduced. In a revised version, a movie/image with reduced threshold is shown."

This reply does not address the reviewer's concerns. Detection of ACC with 405 nm excitation is not sufficient; many organic components can fluoresce under violet light excitation. For example, Delvene et al. (2002) (https://doi.org/10.18261/let.55.4.7) showed that "the Pleistocene and Jurassic microborings emit in the blue-yellow spectral region (420-600 nm) with a laser excitation of 405 nm, which coincides with the emission due to NADPH [nicotinamide adenine dinucleotide], FAD [flavin adenine dinucleotide], and riboflavin pigments characteristic of some cyanobacteria." Traditionally, in geological or biogenic calcium carbonate samples, Raman spectroscopic characterization of ACC and its magnesium content can be used (e.g., Wang, D., Hamm, L. M., Bodnar, R. J. & Dove, P. M. Raman spectroscopic characterization of the magnesium content in amorphous calcium carbonates. J. Raman Spectrosc. 43, 543-548 (2012); Perrin, J. et al. Raman characterization of synthetic magnesian calcites. Am. Mineral. 101, 2525-2538 (2016)). However, in biological, living-cell systems, Mehta et al. (2022) (doi: 10.1016/j.saa.2022.121262) successfully used FTIR spectroscopy to identify ACC by two characteristic FTIR vibrations at ca. 860 cm-1 and ca. 306 cm-1. Other methods such as STXM analyses at the C K-edge (Monteil et al. 2021, doi: 10.1038/s41396-020-00747-3) are also available. Because the core of the authors' interpretation (i.e., detection of ACC in vesicles) is not supported by hard evidence, the claim that the study represents a "paradigm shift" is far-fetched and the whole model is based on speculations. If the authors are able to unequivocally confirm the presence of ACC within the vesicles and its subsequent transformation into calcitic needles, the other problems noted in the paper will be relatively trivial.

Reviewer #2 (Public Review):

Summary:

In their manuscript, Daniel Spari et al. explored the dual role of ATP in exacerbating sepsis, revealing that ATP from both host and bacteria significantly impacts immune responses and disease progression.

Strengths:

The study meticulously examines the complex relationship between ATP release and bacterial growth, membrane integrity, and how bacterial ATP potentially dampens inflammatory responses, thereby impairing survival in sepsis models. Additionally, this compelling paper implies a concept that bacterial OMVs act as vehicles for the systemic distribution of ATP, influencing neutrophil activity and exacerbating sepsis severity.

Weaknesses:

(1) The researchers extracted and cultivated abdominal fluid on LB agar plates, then randomly picked 25 colonies for analysis. However, they didn't conduct 16S sequencing on the fluid itself. It's worth noting that the bacterial species present may vary depending on the individual patients. It would be beneficial if the authors could specify whether they've verified the existence of unculturable species capable of secreting high levels of Extracellular ATP.

(2) Do mice lacking commensal bacteria show a lack of Extracellular ATP following cecal ligation puncture?

(3) The authors isolated various bacteria from abdominal fluid, encompassing both Gram-negative and Gram-positive types. Nevertheless, their emphasis appeared to be primarily on the Gram-negative E. coli. It would be beneficial to ascertain whether the mechanisms of Extracellular ATP release differ between Gram-positive and Gram-negative bacteria. This is particularly relevant given that the Gram-positive bacterium E. faecalis, also isolated from the abdominal fluid, is recognized for its propensity to release substantial amounts of Extracellular ATP.

(4) The authors observed changes in the levels of LPM, SPM, and neutrophils in vivo. However, it remains uncertain whether the proliferation or migration of these cells is modulated or inhibited by ATP receptors like P2Y receptors. This aspect requires further investigation to establish a convincing connection.

(5) Additionally, is it possible that the observed in vivo changes could be triggered by bacterial components other than Extracellular ATP? In this research field, a comprehensive collection of inhibitors is available, so it is desirable to utilize them to demonstrate clearer results.

(6) Have the authors considered the role of host-derived Extracellular ATP in the context of inflammation?

(7) The authors mention that Extracellular ATP is rapidly hydrolyzed by ectonucleotases in vivo. Are the changes of immune cells within the peritoneal cavity caused by Extracellular ATP released from bacterial death or by OMVs?

(8) In the manuscript, the sample size (n) for the data consistently remains at 2. I would suggest expanding the sample size to enhance the robustness and rigor of the results.

Reviewer #1 (Public Review):

Summary:

Extracellular ATP represents a danger-associated molecular pattern associated to tissue damage and can act also in an autocrine fashion in macrophages to promote proinflammatory responses, as observed in a previous paper by the authors in abdominal sepsis. The present study addresses an important aspect possibly conditioning the outcome of sepsis that is the release of ATP by bacteria. The authors show that sepsis-associated bacteria do in fact release ATP in a growth dependent and strain-specific manner. However, whether this bacterial derived ATP play a role in the pathogenesis of abdominal sepsis has not been determined. To address this question, a number of mutant strains of E. coli has been used first to correlate bacterial ATP release with growth and then, with outer membrane integrity and bacterial death. By using E. coli transformants expressing the ATP-degrading enzyme apyrase in the periplasmic space, the paper nicely shows that abdominal sepsis by these transformants results in significantly improved survival. This effect was associated to the reduction of small peritoneal macrophages and CX3CR1+ monocytes, and increase in neutrophils. To extrapolate the function of bacterial ATP from the systemic response to microorganisms, the authors exploited bacterial OMVs either loaded or not with ATP to investigate the systemic effects devoid of living microorganisms. This approach showed that ATP-loaded OMVs induced degranulation of neutrophils after lysosomal uptake, suggesting this mechanism could contribute to sepsis severity.

Strengths:

The most compelling part of the study is the analysis of E. coli mutants to address different aspects of bacterial release of ATP that could be pathogenically relevant during systemic dissemination of bacteria in the host.

Weaknesses:

As pointed out in the limitations of the study whether ATP-loaded OMVs could provide a mechanistic proof of the pathogenetic role of bacteria-derived ATP independently of live microorganisms in sepsis is interesting but not definitively convincing. It could be useful to see whether degranulation of neutrophils is differently induced also by apyrase-expressing vs control E. coli transformants. Also, the increase of neutrophils in bacterial ATP-depleted abdominal sepsis, which has better outcome than "ATP-proficient" sepsis, seems difficult to correlate to the hypothesized tissue damage induced by ATP delivered via non-infectious OMVs. Is neutrophils count affected by ATP delivered via OMVs? Probably a comparison of cytokine profiles in the abdominal fluids of E. coli and OMV treated animals could be helpful in defining the different responses induced by OMV-delivered vs bacterial-released ATP.

The analyses performed on OMV treated versus E. coli infected mice are not immediately related and difficult to combine when trying to draw a pathogenetic hypothesis for bacterial ATP in sepsis.

It's not clear why lung neutrophils were used for RNAseq.

Reviewer #1 (Public Review):

Summary:

The authors utilize fluid-structure interaction analyses to simulate fluid flow within and around the Cambrian cnidarian Quadrapyrgites to reconstruct feeding/respiration dynamics. Based on vorticity and velocity flow patterns, the authors suggest that the polyp expansion and contraction ultimately develop vortices around the organism that are like what modern jellyfish employ for movement and feeding. Lastly, the authors suggest that this behavior is likely a prerequisite transitional form to swimming medusae.

Strengths:

While fluid-structure-interaction analyses are common in engineering, physics, and biomedical fields, they are underutilized in the biological and paleobiological sciences. Zhang et al. provide a strong approach to integrating active feeding dynamics into fluid flow simulations of ancient life. Based on their data, it is entirely likely the described vortices would have been produced by benthic cnidarians feeding/respiring under similar mechanisms. However, some of the broader conclusions require additional justification.

Weaknesses:

(1) The claim that olivooid-type feeding was most likely a prerequisite transitional form to jet-propelled swimming needs much more support or needs to be tailored to olivooids. This suggests that such behavior is absent (or must be convergent) before olivooids, which is at odds with the increasing quantities of pelagic life (whose modes of swimming are admittedly unconstrained) documented from Cambrian and Neoproterozoic deposits. Even among just medusozoans, ancestral state reconstruction suggests that they would have been swimming during the Neoproterozoic (Kayal et al., 2018; BMC Evolutionary Biology) with no knowledge of the mechanics due to absent preservation.<br /> (2) While the lack of ambient flow made these simulations computationally easier, these organisms likely did not live in stagnant waters even within the benthic boundary layer. The absence of ambient unidirectional laminar current or oscillating current (such as would be found naturally) biases the results.<br /> (3) There is no explanation for how this work could be a breakthrough in simulation gregarious feeding as is stated in the manuscript.

Despite these weaknesses the authors dynamic fluid simulations convincingly reconstruct the feeding/respiration dynamics of the Cambrian Quadrapyrgites, though the large claims of transitionary stages for this behavior are not adequately justified. Regardless, the approach the authors use will be informative for future studies attempting to simulate similar feeding and respiration dynamics.

Reviewer #2 (Public Review):

Summary:

The authors seek to elucidate the early evolution of cnidarians through computer modeling of fluid flow in the oral region of very small, putative medusozoan polyps. They propose that the evolutionary advent of the free-swimming medusoid life stage was preceded by a sessile benthic life stage equipped with circular muscles that originally functioned to facilitate feeding and that later became co-opted for locomotion through jet propulsion.

Strengths:

Assumptions of the modeling exercise laid out clearly; interpretations of the results of the model runs in terms of functional morphology plausible. An intriguing investigation that should stimulate further discussion and research.

Weaknesses:

Speculation on the origin of the medusoid life stage in cnidarians heavily dependent on prior assumptions concerning the soft part anatomy and material properties of the skeleton of the modeled fossil organism that may be open to alternative interpretations. Logically, of course, the hypothesis that cnidarian medusae originated from benthic polyps must be evaluated along with the alternative hypotheses that the medusa came first and that the ancestral cnidarian exhibited both life stages.

Reviewer #1 (Public Review):

Gating of Kv10 channels is unique because it involves coupling between non-domain swapped voltage sensing domains, a domain-swapped cytoplasmic ring assembly formed by the N- and C-termini, and the pore domain. Recent structural data suggests that activation of the voltage sensing domain relieves a steric hindrance to pore opening, but the contribution of the cytoplasmic domain to gating is still not well understood. This aspect is of particular importance because proteins like calmodulin interact with the cytoplasmic domain to regulate channel activity. The effects of calmodulin (CaM) in WT and mutant channels with disrupted cytoplasmic gating ring assemblies are contradictory, resulting in inhibition or activation, respectively. The underlying mechanism for these discrepancies is not understood. In the present manuscript, Reham Abdelaziz and collaborators use electrophysiology, biochemistry and mathematical modeling to describe how mutations and deletions that disrupt inter-subunit interactions at the cytoplasmic gating ring assembly affect Kv10.1 channel gating and modulation by CaM. In the revised manuscript, additional information is provided to allow readers to identify within the Kv10.1 channel structure the location of E600R, one of the key channel mutants analyzed in this study. However, the mechanistic role of the cytoplasmic domains that this study focuses on, as well as the location of the ΔPASCap deletion and other perturbations investigated in the study remain difficult to visualize without additional graphical information. This can make it challenging for readers to connect the findings presented in the study with a structural mechanism of channel function.

The authors focused mainly on two structural perturbations that disrupt interactions within the cytoplasmic domain, the E600R mutant and the ΔPASCap deletion. By expressing mutants in oocytes and recording currents using Two Electrode Voltage-Clamp (TEV), it is found that both ΔPASCap and E600R mutants have biphasic conductance-voltage (G-V) relations and exhibit activation and deactivation kinetics with multiple voltage-dependent components. Importantly, the mutant-specific component in the G-V relations is observed at negative voltages where WT channels remain closed. The authors argue that the biphasic behavior in the G-V relations is unlikely to result from two different populations of channels in the oocytes, because they found that the relative amplitude between the two components in the G-V relations was highly reproducible across individual oocytes that otherwise tend to show high variability in expression levels. Instead, the G-V relations for all mutant channels could be well described by an equation that considers two open states O1 and O2, and a transition between them; O1 appeared to be unaffected by any of the structural manipulations tested (i.e. E600R, ΔPASCap, and other deletions) whereas the parameters for O2 and the transition between the two open states were different between constructs. The O1 state is not observed in WT channels and is hypothesized to be associated with voltage sensor activation. O2 represents the open state that is normally observed in WT channels and is speculated to be associated with conformational changes within the cytoplasmic gating ring that follow voltage sensor activation, which could explain why the mutations and deletions disrupting cytoplasmic interactions affect primarily O2.

Severing the covalent link between the voltage sensor and pore reduced O1 occupancy in one of the deletion constructs. Although this observation is consistent with the hypothesis that voltage-sensor activation drives entry into O1, this result is not conclusive. Structural as well as functional data has established that the coupling of the voltage sensor and pore does not entirely rely on the S4-S5 covalent linker between the sensor and the pore, and thus the severed construct could still retain coupling through other mechanisms, which is consistent with the prominent voltage dependence that is observed. If both states O1 and O2 require voltage sensor activation, it is unclear why the severed construct would affect state O1 primarily, as suggested in the manuscript, as opposed to decreasing occupancy of both open states. In line with this argument, the presence of Mg2+ in the extracellular solution affected both O1 and O2. This finding suggests that entry into both O1 and O2 requires voltage-sensor activation because Mg2+ ions are known to stabilize the voltage sensor in its most deactivated conformations.

Activation towards and closure from O1 is slow, whereas channels close rapidly from O2. A rapid alternating pulse protocol was used to take advantage of the difference in activation and deactivation kinetics between the two open components in the mutants and thus drive an increasing number of channels towards state O1. Currents activated by the alternating protocol reached larger amplitudes than those elicited by a long depolarization to the same voltage. This finding is interpreted as an indication that O1 has a larger macroscopic conductance than O2. In the revised manuscript, the authors performed single-channel recordings to determine why O1 and O2 have different macroscopic conductance. The results show that at voltages where the state O1 predominates, channels exhibited longer open times and overall higher open probability, whereas at more depolarized voltages where occupancy of O2 increases, channels exhibited more flickery gating behavior and decreased open probability. These results are informative but not conclusive because additional details about how experiments were conducted, and group data analysis are missing. Importantly, results showing inhibition of single ΔPASCap channels by a Kv10-specific inhibitor are mentioned but not shown or quantitated - these data are essential to establish that the new O1 conductance indeed represents Kv10 channel activity.

It is shown that conditioning pulses to very negative voltages result in mutant channel currents that are larger and activate more slowly than those elicited at the same voltage but starting from less negative conditioning pulses. In voltage-activated curves, O1 occupancy is shown to be favored by increasingly negative conditioning voltages. This is interpreted as indicating that O1 is primarily accessed from deeply closed states in which voltage sensors are in their most deactivated position. Consistently, a mutation that destabilizes these deactivated states is shown to largely suppress the first component in voltage-activation curves for both ΔPASCap and E600R channels.

The authors then address the role of the hidden O1 state in channel regulation by calmodulation. Stimulating calcium entry into oocytes with ionomycin and thapsigarging, assumed to enhance CaM-dependent modulation, resulted in preferential potentiation of the first component in ΔPASCap and E600R channels. This potentiation was attenuated by including an additional mutation that disfavors deeply closed states. Together, these results are interpreted as an indication that calcium-CaM preferentially stabilizes deeply closed states from which O1 can be readily accessed in mutant channels, thus favoring current activation. In WT channels lacking a conducting O1 state, CaM stabilizes deeply closed states and is therefore inhibitory. It is found that the potentiation of ΔPASCap and E600R by CaM is more strongly attenuated by mutations in the channel that are assumed to disrupt interaction with the C-terminal lobe of CaM than mutations assumed to affect interaction with the N-terminal lobe. These results are intriguing but difficult to interpret in mechanistic terms. The strong effect that calcium-CaM had on the occupancy of the O1 state in the mutants raises the possibility that O1 can be only observed in channels that are constitutively associated with CaM. To address this, a biochemical pull-down assay was carried out to establish that only a small fraction of channels are associated with CaM under baseline conditions. These CaM experiments are potentially very interesting and could have wide physiological relevance. However, the approach utilized to activate CaM is indirect and could result in additional non-specific effects on the oocytes that could affect the results.

Finally, a mathematical model is proposed consisting of two layers involving two activation steps for the voltage sensor, and one conformational change in the cytoplasmic gating ring - completion of both sets of conformational changes is required to access state O2, but accessing state O1 only requires completion of the first voltage-sensor activation step in the four subunits. The model qualitatively reproduces most major findings on the mutants. Although the model used is highly symmetric and appears simple, the mathematical form used for the rate constants in the model adds a layer of complexity to the model that makes mechanistic interpretations difficult. In addition, many transitions that from a mechanistic standpoint should not depend on voltage were assigned a voltage dependence in the model. These limitations diminish the overall usefulness of the model which is prominently presented in the manuscript. The most important mechanistic assumptions in the model are not addressed experimentally, such as the proposition that entry into O1 depends on the opening of the transmembrane pore gate, whereas entry into O2 involves gating ring transitions - it is unclear why O2 would require further gating ring transitions to conduct ions given that the gating ring can already support permeation by O1 without any additional conformational changes.

Reviewer #3 (Public Review):

In the present manuscript, Abdelaziz and colleagues interrogate the gating mechanisms of Kv10.1, an important voltage-gated K+ channel in cell cycle and cancer physiology. At the molecular level, Kv10.1 is regulated by voltage and Ca-CaM. Structures solved using Cryo-EM for Kv10.1 as well as other members of the KCNH family (Kv11 and Kv12) show channels that do not contain a structured S4-S5 linker imposing therefore a non-domain swapped architecture in the transmembrane region. However, the cytoplasmatic N- and C- terminal domains interact in a domain swapped manner forming a gating ring. The N-terminal domain (PAS domain) of one subunit is located close to the intracellular side of the voltage sensor domain and interacts with the C-terminal domain (CNBHD domain) of the neighbor subunit. Mutations in the intracellular domains has a profound effect in the channel gating. The complex network of interactions between the voltage-sensor and the intracellular domains makes the PAS domain a particularly interesting domain of the channel to study as responsible for the coupling between the voltage sensor domains and the intracellular gating ring.

The coupling between the voltage-sensor domain and the gating ring is not fully understood and the authors aim to shed light into the details of this mechanism. In order to do that, they use well established techniques such as site-directed mutagenesis, electrophysiology, biochemistry and mathematical modeling. In the present work, the authors propose a two open state model that arises from functional experiments after introducing a deletion on the PAS domain (ΔPAS Cap) or a point mutation (E600R) in the CNBHD domain. The authors measure a bi-phasic G-V curve with these mutations and assign each phase as two different open states, one of them not visible on the WT and only unveiled after introducing the mutations. The hypothesis proposed by the authors could change the current paradigm in the current understanding for Kv10.1 and it is quite extraordinary; therefore, it requires extraordinary evidence to support it.

STRENGTHS: The authors use adequate techniques such as electrophysiology and site-directed mutagenesis to address the gating changes introduced by the molecular manipulations. They also use appropriate mathematical modeling to build a Markov model and identify the mechanism behind the gating changes.

WEAKNESSES: The results presented by the authors do not fully support their conclusions since they could have alternative explanations. The authors base their primary hypothesis on the bi-phasic behavior of a calculated G-V curve that do not match the tail behavior, the experimental conditions used in the present manuscript introduce uncertainties, weakening their conclusions and complicating the interpretation of the results. Therefore, their experimental conditions need to be revisited

I have some concerns related to the following points:

(1) Biphasic gating behavior<br /> The authors use the TEVC technique in oocytes extracted surgically from Xenopus Leavis frogs. The method is well established and is adequate to address ion channel behavior. The experiments are performed in chloride-based solutions which present a handicap when measuring outward rectifying currents at very depolarizing potentials due to the presence of calcium activated chloride channel expressed endogenously in the oocytes; these channels will open and rectify chloride intracellularly adding to the outward rectifying traces during the test pulse.<br /> The authors calculate their G-V curves from the test pulse steady-state current instead of using the tail currents. The conductance measurements are normally taken from the 'tail current' because tails are measured at a fix voltage hence maintaining the driving force constant. Calculating the conductance from the traces should not be a problem, however, in the present manuscript, the traces and the tail currents do not agree. The tail traces shown in Fig1E do not show an increasing current amplitude in the voltage range from +50mV to +120mV, they seem to have reached a 'saturation state', suggesting that the traces from the test pulse contain an inward chloride current contamination. In addition, this second component identified by the authors as a second open state appears after +50mV and seems to never saturate. The normalization to the maximum current level during the test pulse, exaggerates this second component on the calculated G-V curve. It's worth noticing that the ΔPASCap mutant experiments on Fig 5 in Mes based solutions do not show that second component on the G-V.

Because these results are the foundation for their two open state hypotheses, I will strongly suggest the authors to repeat all their Chloride-based experiments in Mes-based solutions to eliminate the undesired chloride contribution to the mutants current and clarify the contribution of the mutations to the Kv10.1 gating.

(2) Two step gating mechanism.<br /> The authors interpret the results obtained with the ΔPASCap and the E600R as two step gating mechanisms containing two open states (O1 and O2) and assign them to the voltage sensor movement and gating ring rotation respectively. It is not clear, however how the authors assign the two open states.<br /> The results show how the first component is conserved amongst mutations; however, the second one is not. The authors attribute the second component, hence the second open state to the movement of the gating ring. This scenario seems unlikely since there is a clear voltage-dependence of the second component that will suggest an implication of a voltage-sensing current.

The split channel experiment is interesting but needs more explanation. I assume the authors expressed the 2 parts of the split channel (1-341 and 342-end), however Tomczak et al showed in 2017 how the split presents a constitutively activated function with inward currents that are not visible here, this point needs clarification.

Moreover, the authors assume that the mutations introduced uncover a new open state, however the traces presented for the mutations suggest that other explanations are possible. Other gating mechanisms like inactivation from the closed state, can be introduced by the mutations. The traces presented for ΔPASCap but specially E600R present clear 'hooked tails', a direct indicator of a populations of inactive channels during the test pulse that recover from inactivation upon repolarization (Tristani-Firouzi M, Sanguinetti MC. J Physiol. 1998). The results presented by the authors can be alternatively explained with a change in the equilibrium between the close to inactivated/recovery from inactivation to the open state. Finally, the authors state that they do not detect "cumulative inactivation after repeated depolarization" but that is considering inactivation only from the open state and ignoring the possibility of the existence of close state inactivation or, that like in hERG, that the channel inactivates faster that what it activates (Smith PL, Yellen G. J Gen Physiol. 2002).

(3) Single channel conductance.<br /> The single channels experiments are a great way to assess the different conductance of single channel openings, unfortunately the authors cannot measure accurately different conductances for the two proposed open states. The Markov Model built by the authors, disagrees with their interpretation of the experimental results assigning the exact same conductance to the two modeled open states. To interpret the mutant data, it is needed to add data with the WT for comparison and in presence of specific blockers.

Reviewer #1 (Public Review):

This manuscript by Negi et al. investigates the effects of different ubiquitin and ubiquitin-like modifications on the stability of substrate proteins, seeking to provide mechanistic insights into known effects of these modifications on cellular protein abundance. The authors focus on comparative studies of two modifications, ubiquitin and FAT10 (a protein with two ubiquitin-like domains), on a panel of substrate proteins; prior work had established that FAT10-conjugated proteins had lower stability to proteosomal degradation than Ub-modified counterparts.

Strengths of the work include its integration of data across diverse approaches, including molecular dynamics simulations, solution NMR spectroscopy, and in vitro and cellular stability assays. From these, the authors provide provocative mechanistic insight into the lower stability of FAT10 on its own, and in FAT10-mediated destabilization of substrate proteins in computational and experimental findings. Notably, such destabilization impacts both the tag and tagged proteins, raising some provocative questions about mechanism. The data here are generally compelling, albeit with minor concerns on presentation in parts. Conclusions from this work will be interesting to scientists in several fields, particularly those interested in cellular proteostasis and in vitro protein design / long-range communication.

The most substantial weakness of this work from my perspective is the specificity of these destabilization effects. In particular, technical challenges of producing bona fide Ub- or FAT10-conjugated substrates with native linkages limits the ability to conduct in vitro studies on exactly the same molecules as being studied in cellular environments. Given some discussion in the manuscript about the importance of linkage location on the specificity of certain tag/substrate interactions, this raises an understandable but unfortunate caveat that needs to be considered more fully both in general and in light of data from other fields (e.g. single molecule pulling) showing site-dependence of comparable effects. I note that these concerns do not impact the caliber of the conclusions themselves, but perhaps suggest area for caution as to their potential impact at this time.

Reviewer #2 (Public Review):

"Plasticity of the proteasome-targeting signal Fat10 enhances substrate degradation" is a nice study where the authors have shown the differences between two protein degradation tags namely, FAT10 and ubiquitin. Even though these tags are closely related in terms of folds, they have differential efficiency in degrading the substrates covalently attached to them. The authors have utilised extensive MD simulations combined with biophysics and cell biology to show the structural dynamics these tags provide for proteasomal degradation.

Reviewer #1 (Public Review):

This work focuses on the trade-off between precision and robustness in morphogen gradients of Hedgehog signaling. It presents a framework for how hedgehog signaling rises to precise responses and robust responses. This Framework is based on the characteristics of the hedgehog signaling pathway and specifically on the characteristics of the dynamical and stationary gradients that it forms in the Drosophila wing disc. On the one hand, the manuscript takes into account known results showing that the Hedgehog stationary gradient is robust due to a self-enhanced degradation (via activation of the Patched receptor). On the other hand, it uses the concept of dynamic interpretation of the gradient introduced by the leading author of this manuscript. According to this interpretation, different targets may be responding to a single signaling threshold and what differentiates the targets is whether they respond to the transient gradient, which extends over more cells, or if they respond to the stationary gradient. The Framework presented in this manuscript takes this prior knowledge and builds on it. The Framework proposes that the response from different targets will not be equally robust. Specifically, if the target responds to the stationary gradient, it will be a target with a robust response. Conversely, if the target responds to the gradient while it is being built, then it will be less robust but more precise. This framework is analyzed using mathematical models. Finally, experimental data that partially corroborate this framework are presented, focusing on the col and Dpp targets, which, according to previous results, read the stationary and transient gradients, respectively. To changes in Hh levels, the col pattern is more robust than the Dpp pattern. Furthermore, it is shown that this robustness decreases if the Patched receptor is not regulated. Hence, these experimental results confirm that the robustness is target-specific, as predicted by the models. The precision of the Dpp pattern is not tested experimentally.

Reviewer #2 (Public Review):

This paper presents a modeling analysis of a diffusing morphogen (hh) that patterns the wing disk by controlling the expression of dpp and col. Two modes of gene expression control/interpretation are analyzed and presented, one is a response using a steady state threshold (col), which could be robust (defined as a small spatial shift of the gene expression when hh dosage changes) by a ptch mediated negative feedback mechanism; the other is the "overshoot" where an earlier hh gradient profile pre-steady state is read at a threshold to activate the gene (dpp), which is less robust to dosage changes but has better boundary features. Experimental measurements of pattern widths of col and dpp were performed under different hh dosage to test the models. How these different modes were achieved by each gene was unclear.

The reviewer found this study presents at best incremental advances to the field. It doesn't provide substantial progress conceptually or experimentally from Eldar et al., 2003, Adleman et al., 2022 and particularly Nahmad and Stathopoulos, 2009. The experimental data and interpretation appear to lack the rigor needed to challenge the model predictions.

The authors pitched the difference between dpp and col in their response to hh dosage change as a tradeoff between robustness and precision. Specifically, the robustness refers to positioning and the precision refers to sharpness, which are somewhat arbitrary - as robustness could also refer to maintaining the sharpness of a expression boundary and precision can also refer to the position. Particularly for dpp, whose developmental significance of stripe position and sharpness is not analyzed (disc growth, pSmad, etc, for example - does a sharper but more mislocated dpp domain help the tissue?). The relationship between positioning and sharpness of a pattern in a morphogen system has been extensively discussed by many authors on a theoretcial level. The authors' theoretical analysis is clear and simple but not new. Experimental evidence indicates that dpp and col are regulated very differently by hh, particularly in terms of timing of response (Nahmad and Stathopoulos, 2009). No comparison of the GRNs from hh to these two genes was made or experimentally tested. It is difficult to conclude that their behaviors in response to hh dosage change are indeed from the hh gradient profile. It is also difficult to speculate if either of these genes (particularly dpp) is facing a true biological tradeoff or tuning back and forth between positioning and sharpness during evolution.

Methods 4.5: To measure widths of gene expression patterns, the authors used a background subtraction, followed by normalization and then thresholded the boundary at 0.2 - this approach firstly is oversimplifying the profile of the expression gradient/profile which could be informative in model testing (e.g., sharpness of dpp?). Secondly, the sequence of the analysis steps may introduce larger errors to lower signal-to-noise images where the subtraction narrows the pattern more than those with higher signal-to-noise (e.g., the 18 degree vs 25 degree images, Fig.6A), this would result in errors in the width measurements. Importantly, disk size and wing size controls are not reported.

Reviewer #1 (Public Review):

Kerkoerle and colleagues present a very interesting comparative fMRI study in humans and monkeys, assessing neural responses to surprise reactions at the reversal of a previously learned association. The implicit nature of this task, assessing how this information is represented without requiring explicit decision making, is an elegant design. The paper reports that both humans and monkeys show neural responses across a range of areas when presented with incongruous stimulus pairs. Monkeys also show a surprise response when the stimuli are presented in the reversed direction. However, humans show no such surprise response based on this reversal, suggesting that they encode the relationship reversibly and bidirectionally, unlike the monkeys. This has been suggested as a hallmark of symbolic representation, that might be absent in nonhuman animals.

I find this experiment and the results quite compelling, and the data do support the hypothesis that humans are somewhat unique in their tendency to form reversible, symbolic associations. I think that an important strength of the results is that the critical finding is the presence of an interaction between congruity and canonicity in macaques, which does not appear in humans. These results go a long way to allay concerns I have about the comparison of many human participants to a very small number of macaques.

The results do appear to show that macaques show the predicted interaction effect (even despite the sample size), while humans do not. I think this is quite convincing. (Although had the results turned out differently (for example an effect in humans that was absent in macaques), I think this difference in sample size would be considerably more concerning.)

I would also note that while I agree with the authors conclusions, it is also notable to me that the congruity effect observed in humans (red vs blue lines in Fig. 2B) appears to be far more pronounced than any effect observed in the macaques (Fig. 3C-3). Again, this does not challenge the core finding of this paper but does suggest methodological or possibly motivational/attentional differences between the humans and the monkeys (or, for example, that the monkeys had learned the associations less strongly and clearly than the humans). The authors now discuss this more fully.

This is a strong paper with elegant methods and makes a worthwhile contribution to our understanding of the neural systems supporting symbolic representations in humans, as opposed to other animals.

Reviewer #2 (Public Review):

In their article titled, van Kerkoerle et al address the timely question of whether non-human primates (rhesus macaques) possess the ability for reverse symbolic inference as observed in humans. Through an fMRI experiment in both humans and monkeys, they analyzed the bold signal in both species while observing audio-visual and visual-visual stimuli pairs that had been previously learned in a particular direction. Remarkably, the findings pertaining to humans revealed that a broad brain network exhibited increased activity in response to surprises occurring in both the learned and reverse directions. Conversely, in monkeys, the study uncovered that the brain activity within sensory areas only responded to the learned direction but failed to exhibit any discernible response to the reverse direction. These compelling results indicate that the capacity for reversible symbolic inference may be specific to humans, even though it remains to be tested in other species.

In general, the manuscript is skillfully crafted and highly accessible to readers. The experimental design exhibits originality, and the analyses are tailored to effectively address the central question at hand. Although the first experiment raised a number of methodological inquiries, the subsequent second experiment thoroughly addresses these concerns and effectively replicates the initial findings, thereby significantly strengthening the overall study. Overall, this article is of high quality and brings new insight into human cognition.

The main limitation of the studies is the sample size of the non-human primate group (n=2 and n=3). Nevertheless, this limitation is carefully addressed and discussed in the manuscript.

Reviewer #3 (Public Review):

Original review

This study investigates the hypothesis that humans (but not non-human primates) spontaneously learn reversible temporal associations (i.e., learning a B-A association after only being exposed to A-B sequences), which the authors consider to be a foundational property of symbolic cognition. To do so, they expose humans and macaques to 2-item sequences (in a visual-auditory experiment, pairs of images and spoken nonwords, and in a visual-visual experiment, pairs of images and abstract geometric shapes) in a fixed temporal order, then measure the brain response during a test phase to congruent vs. incongruent pairs (relative to the trained associations) in canonical vs. reversed order (relative to the presentation order used in training). The advantage of neuroimaging for this question is that it removes the need for a behavioral test, which non-human primates can fail for reasons unrelated to the cognitive construct being investigated. In humans, the researchers find statistically indistinguishable incongruity effects in both directions (supporting a spontaneous reversible association), whereas in monkeys they only find incongruity effects in the canonical direction (supporting an association but a lack of spontaneous reversal). Although the precise pattern of activation varies by experiment type (visual-auditory vs. visual-visual) in both species, the authors point out that some of the regions involved are also those that are most anatomically different between humans and other primates. The authors interpret their findings to support the hypothesis that reversible associations, and by extension symbolic cognition, is uniquely human.

This study is a valuable complement to prior behavioral work on this question. However, I have some concerns about methods and framing.

Methods - Design issues:

(1) The authors originally planned to use the same training/testing protocol for both species but the monkeys did not learn anything, so they dramatically increased the amount of training and evaluation. By my calculation from the methods section, humans were trained on 96 trials and tested on 176, whereas the monkeys got an additional 3,840 training trials and 1,408 testing trials. The authors are explicit that they continued training the monkeys until they got a congruity effect. On the one hand, it is commendable that they are honest about this in their write-up, given that this detail could easily be framed as deliberate after the fact. On the other hand, it is still a form of p-hacking, given that it's critical for their result that the monkeys learn the canonical association (otherwise, the critical comparison to the non-canonical association is meaningless).

(2) Between-species comparisons are challenging. In addition to having differences in their DNA, human participants have spent many years living in a very different culture than that of NHPs, including years of formal education. As a result, attributing the observed differences to biology is challenging. One approach that has been adopted in some past studies is to examine either young children or adults from cultures that don't have formal educational structures. This is not the approach the authors take. This major confound needs to minimally be explicitly acknowledged up front.

(3) Humans have big advantages in processing and discriminating spoken stimuli and associating them to visual stimuli (after all, this is what words are in spoken human languages). Experiment 2 ameliorates these concerns to some degree, but still it is difficult to attribute the failure of NHPs to show reversible associations in Experiment 1 to cognitive differences rather than the relative importance of sound string to meaning associations in the human vs. NHP experiences.

(4) More minor: The localizer task (math sentences vs. other sentences) makes sense for math but seems to make less sense for language: why would a language region respond more to sentences that don't describe math vs. ones that do?

Methods - Analysis issues:

(5) The analyses appear to "double dip" by using the same data to define the clusters and to statistically test the average cluster activation (Kriegeskorte et al., 2009). The resulting effect sizes are therefore likely inflated, and the p-values are anticonservative.

FRAMING:

(6) The framing ("Brain mechanisms of reversible symbolic reference: A potential singularity of the human brain") is bigger than the finding (monkeys don't spontaneously reverse a temporal association but humans do). The title and discussion are full of buzzy terms ("brain mechanisms", "symbolic", and "singularity") that are only connected to the experiments by a debatable chain of assumptions.

First, this study shows relatively little about brain "mechanisms" of reversible symbolic associations, which implies insights about how these associations are learned, recognized, and represented. But we're only given standard fMRI analyses that are quite inconsistent across similar experimental paradigms, with purely suggestive connections between these spatial patterns and prior work on comparative brain anatomy.

Second, it's not clear what the relationship is between symbolic cognition and a propensity to spontaneously reverse a temporal association. Certainly if there are inter-species differences in learning preferences this is important to know about, but why is this construed as a difference in the presence or absence of symbols? Because the associations aren't used in any downstream computation, there is not even any way for participants to know which is the sign and which is the signified: these are merely labels imposed by the researchers on a sequential task.

Third, the word "singularity" is both problematically ambiguous and not well supported by the results. "Singularity" is a highly loaded word that the authors are simply using to mean "that which is uniquely human". Rather than picking a term with diverse technical meanings across fields and then trying to restrict the definition, it would be better to use a different term. Furthermore, even under the stated definition, this study performed a single pairwise comparison between humans and one other species (macaques), so it is a stretch to then conclude (or insinuate) that the "singularity" has been found (see also pt. 2 above).

(7) Related to pt. 6, there is circularity in the framing whereby the authors say they are setting out to find out what is uniquely human, hypothesizing that the uniquely human thing is symbols, and then selecting a defining trait of symbols (spontaneous reversible association) *because* it seems to be uniquely human (see e.g., "Several studies previously found behavioral evidence for a uniquely human ability to spontaneously reverse a learned association (Imai et al., 2021; Kojima, 1984; Lipkens et al., 1988; Medam et al., 2016; Sidman et al., 1982), and such reversibility was therefore proposed as a defining feature of symbol representation reference (Deacon, 1998; Kabdebon and Dehaene-Lambertz, 2019; Nieder, 2009).", line 335). They can't have it both ways. Either "symbol" is an independently motivated construct whose presence can be independently tested in humans and other species, or it is by fiat synonymous with the "singularity". This circularity can be broken by a more modest framing that focuses on the core research question (e.g., "What is uniquely human? One possibility is spontaneous reversal of temporal associations.") and then connects (speculatively) to the bigger conceptual landscape in the discussion ("Spontaneous reversal of temporal associations may be a core ability underlying the acquisition of mental symbols").

Comments on revised version:

I thank the authors for engaging constructively with my comments. I'm convinced by the responses to my original points 1, 2, 3, and 4. I'm also partially convinced by the response to point 6 (with qualifications discussed below). I do want to clear the record on points 1 and 6 (about which the authors expressed offense at aspects of my original comments), and to press on points 5 and 7.

(1) It's very helpful to know that the plan was always to extend training in Expt 1. The rationale is now clear in the methods, although I'd encourage the authors to also emphasize this if space permits in the vicinity of lines 211-216, which still read as if the extended training was a post hoc decision ("the canonical congruity effect... was not significant... after 3 days of exposure... Thus... monkeys were further exposed..."). The authors have objected to my original use of "p hacking", which I agree was too strong (my apologies). My intention was only to point out that *if it were the case that training duration was conditional on the monkeys' success at learning the canonical association* (which the authors have now clarified was not the case), then this would be steering the study post hoc to achieve a desired outcome. I recognize the authors' point that the canonical direction was a sanity check, not the effect of interest (reversed association), but it's still true that they needed to achieve this sanity check in order for the absence of a reversed effect to be meaningful. This was the source of my original concern. This point is only clarificational (no action is recommended).

(5) The authors have said they don't understand my concern about "double-dipping" in the statistical analyses, so I will attempt to clarify. First, I should stress that this concern applies only to the whole-brain results (Tables 1-4), not the fROI results. As the authors point out, this was indeed unclear, and I apologize. My concern about Tables 1-4 is that they seem to be derived using the classical technique of thresholding contrasts at some significance level to define clusters and then reporting cluster statistics (in this case, t-values) derived from *the same contrast in the same activation maps*. If this is not what was done (i.e., if orthogonal data and/or contrasts were used to define clusters and quantify contrasts within clusters, as in the fROI analyses), then this point is moot (and clarification in the paper would be helpful). But if this is what was done, then this procedure is known to be distortionary (e.g., Kriegeskorte et al 2009, "Nonindependent selective analysis is incorrect and should not be acceptable in neuroscientific publications").

(6) The authors have objected to my use of the term "insinuate" as pejorative. I don't share this impression (and insult was certainly not my intent) but I'm happy to concede that a less loaded term (e.g., "suggest") would have been a better choice. I apologize. In any case, I stand by my intended original concern that a key idea in this piece (that reversible symbolic inference is a singularity of the human brain) is being advanced rhetorically rather than empirically, by repeatedly supplying it to readers (albeit with qualifiers like "potential") as an interpretive lens through which to view empirical results that only directly support a more modest claim (that macaques spontaneously reverse sequential associations less readily than humans do). To be clear, it is good that the authors don't make this stronger claim outright, and it is fine to motivate a more modest research question (e.g., do species differ in spontaneous reversal of associations) on the grounds that it is a stepping stone to a bigger one (what is the singularity). But by placing the bigger framing front and center in this way, there's a risk that this paper will be received by the community as establishing a conclusion that it does not actually establish.

(7) The authors have said they don't understand the circularity I'm alleging. Having read the revision, I believe the issue is still there, so I'll make another attempt. The problem is most clearly apparent in the Discussion text quoted in my original comment (lines 347-350 of the revision, emphasis mine): "Several studies previously found behavioural evidence for a *uniquely human* ability to spontaneously reverse a learned association (Imai et al., 2021; Kojima, 1984; Lipkens et al., 1988; Medam et al., 2016; Sidman et al., 1982), and such reversibility was *therefore* proposed as a defining feature of symbol representation reference (Deacon, 1998; Kabdebon and Dehaene-Lambertz, 2019; Nieder, 2009)." In other words, reversal of associations is selected as a defining feature of symbols and targeted by this study *because* it is thought to be uniquely human. This is fine, but it prohibits you from then advocating the hypothesis that symbolic cognition is the singularity (lines 49-52), because "symbol" is being defined such that this is necessarily the case. To minimally paraphrase what I perceive to be the circular logic in the framing, the argument seems to go: "What is uniquely human? Symbols. What are symbols? That which is uniquely human." In my original comment, I suggested a reframing that would fix this issue, namely: "What is uniquely human? Spontaneous reversal of temporal associations." The authors say they don't see the difference between this framing and their own, so I'll try to clarify: the difference is that it sidesteps the notion of "symbol", and in so doing removes the circular definitions of "symbol" and "singularity" in terms of each other. This suggestion was given not as a prescription but as an example to show that the issue can be remedied by revisions to the framing without doing damage to the empirical claims. If the authors prefer a different remedy that avoids circular definitions of terms, that's fine.

Reviewer #1 (Public Review):

More than ten years ago, it was shown that activity in the primary visual cortex of mice substantially increases when mice are running compared to when they are sitting still. This finding 'revolutionised' our thinking about visual cortex, turning away from it being a passive image processor and highlighting the influence of non-visual factors. The current study now for the first time repeats this experiment in marmosets. The authors find that in contrast to mice, marmoset V1 activity is slightly suppressed during running, and they relate this to differences in gain modulations of V1 activity between the two species.

Strengths

- Replication in primates of the original finding in mice partly took so long, because of the inherent difficulties with recording from the brain of a running primate. In fact one recent, highly related study on macaques looked at spontaneous limb movements as the macaque was sitting. The treadmill for the marmosets in the current study is a very elegant solution to the problem of running in primates. It allows for true replication of the 'running vs stationary' experiment and undoubtedly opens up many possibilities for other experiments recording from a head-fixed but active marmoset.<br /> - In addition to their own data in marmoset, the authors run their analyses on a publicly available data set in mouse. This allows them to directly compare mouse and marmoset findings, which significantly strengthens their conclusions.<br /> - Marmoset vision is fundamentally different from mouse vision as they have a fovea and make goal-directed eye movements. In this revised version of their paper, the authors acknowledge this and investigate the possible effect of eye movements and pupil size on the differences they find between running and stationary. They conclude that eye input does not explain all these differences.

Significance

The paper provides interesting new evidence to the ongoing discussion about the influence of non-visual factors in general, and running in particular, on visual cortex activity. As such, it helps to pull this discussion out of the rodent field mainly and into the field of primate research. The bigger question of *why* there are differences between rodents and primates remains still unanswered, but the authors do their best to provide possible explanations. The elegant experimental set-up of the marmoset on a treadmill will certainly add new findings to this issue also in the years to come.

Reviewer #2 (Public Review):

This work aims at answering whether activity in primate visual cortex is modulated by locomotion, as was reported for mouse visual cortex. The finding that the activity in mouse visual cortex is modulated by running has changed the concept of primary sensory cortical areas. However, it was an open question whether this modulation generalizes to primates.

To answer this fundamental question the authors established a novel paradigm in which a head-fixed marmoset was able to run on a treadmill while watching a visual stimulus on a display. In addition, eye movements and running speed were monitored continuously and extracellular neuronal activity in primary visual cortex recorded using high-channel-count electrode arrays. This paradigm uniquely permitted to investigate whether locomotion modulates sensory evoked activity in visual cortex of marmoset. Moreover, to directly compare the responses in marmoset visual cortex to responses in mouse visual cortex the authors made use of a publicly-available mouse dataset from the Allen Institute. In this dataset the mouse was also running on a treadmill and observing a set of visual stimuli on a display. The authors took extra care to have the marmoset and mouse paradigms as comparable as possible.

To characterize the visually driven activity the authors present a series of moving gratings and estimate receptive fields with sparse noise. To estimate the gain modulation by running the authors split the dataset into epochs of running and non-running which allowed them to estimate the visually evoked firing rates in both behavioral states.

Strengths:

The novel paradigm of head-fixed marmosets running on a treadmill while being presented with a visual stimulus is unique and ideally tailored to answering the question that the authors aimed to answer. Moreover, the authors took extra care to ensure that the paradigm in marmoset matched as closely as possible to the conditions in the mouse experiments such that the results can be directly compared. To directly compare their data the authors re-analyzed publicly available data from visual cortex of mice recorded at the Allen Institute. Such a direct comparison, and reuse of existing datasets, is another strong aspect of the work. Finally, the presented new marmoset dataset appears to be of high quality, the comparison between mouse and marmoset visual cortex is well done and the results and interpretation straightforward.

Weaknesses:

It is known that the locomotion gain modulation varies with layer in mouse visual cortex, with neurons in the infragranular layers expressing a diversity of modulations (Erisken et al. 2014 Current Biology). However, for the marmoset dataset the layer information was unfortunately not recorded, leaving this point open for future studies.

Nonetheless, the aim of comparing the locomotion induced modulation of activity in primate and mouse primary visual cortex was convincingly achieved by the authors. The results shown in the figures support the conclusion that locomotion modulates the activity in primate and mouse visual cortex differently. While mice show a profound gain increase, neurons in primate visual cortex show little modulation or even a reduction in response strength.

This work will have a strong impact on the field of visual neuroscience but also on neuroscience in general. It revives the debate of whether results obtained in the mouse model system can be simply generalized to other mammalian model systems, such as non-human primates. Based on the presented results, the comparison between the mouse and primate visual cortex is not as straightforward as previously assumed. This will likely trigger more comparative studies between mice and primates in the future, which is important and absolutely needed to advance our understanding of the mammalian brain.

Moreover, the reported finding that neurons in primary visual cortex of marmosets do not increase their activity during running is intriguing, as it makes you wonder why neurons in the mouse visual cortex do so. The authors discuss a few ideas in the paper which can be addressed in future experiments. In this regard it is worth noting that the authors report an interesting difference between the foveal and peripheral part of the visual cortex in marmoset. It will be interesting to investigate these differences in more detail in future studies. Likewise, while running might be an important behavioral state for mice, other behavioral states might be more relevant for marmosets and do modulate the activity of primate visual cortex more profoundly. Future work could leverage the opportunities that the marmoset model system offers to reveal new insights about behavioral related modulation in the primate brain.

Reviewer #3 (Public Review):

Prior studies have shown that locomotion (e.g., running) modulates mouse V1 activity to a similar extent as visual stimuli. However, it's unclear if these findings hold in species with more specialized and advanced visual systems such as nonhuman primates. In this work, Liska et al. leverage population and single neuron analyses to investigate potential differences and similarities in how running modulates V1 activity in marmosets and mice. Specifically, they discovered that although a shared gain model could describe well the trial-to-trial variations of population-level neural activity for both species, locomotion more strongly modulated V1 population activity in mice. Furthermore, they found that at the level of individual units, marmoset V1 neurons, unlike mice V1 neurons, experience suppression of their activity during running.

A major strength of this work is the introduction and completion of primate electrophysiology recordings during locomotion. Data of this kind were previously limited, and this work moves the field forward in terms of data collection in a domain previously inaccessible in primates. Another core strength of this work is that it adds to a limited collection of cross-species data collection and analysis of neural activity at the single-unit and population level, attempting to standardize analysis and data collection to be able to make inferences across species. In particular, the findings on how the primate peripheral and foveal V1 representations functionally relate to and differ from the mice V1 representations speak to the power of these cross-species comparisons.

However, there are still some lingering potential extensions to this work, largely acknowledged by the authors. One of these extensions involves more detailed eye movement analysis within species, such as microsaccades in marmosets and the potential impact on marmoset V1 activity. In the mouse data, similar eye-related analyses were not possible, in part due to instability in the eye recordings of many mouse sessions that made it challenging to replicate partnered analyses for the marmosets. We agree with the authors' assessment that these analyses can be targeted in future work and still believe that the marmoset eye-movement findings provide novel insights that will inform future cross-species comparisons of the visual system. Furthermore, another important issue not fully explored is the possible effects of the reward scheme during marmoset locomotion on V1 activity. The authors note that, unlike their mice counterparts, the marmosets were encouraged to run via liquid rewards, given after subjects traversed a specific distance. While the authors discuss the changes in arousal present when marmosets were running, there are still some unanswered questions on how their reward scheme may affect biomarkers (e.g., pupil sizes) and marmoset V1 activity.

Overall, the methods and data support the work's main claims. Single neuron and population level approaches demonstrate that the activity of V1 in mice and marmoset are categorically different. Since primate V1 is so diverse and differs from mouse V1, this presents important limitations on direct inferences from mouse V1 to primate V1. This work is a great step forward in the field, especially with the novel methodology of collecting neural activity from running primates.

Reviewer #3 (Public Review):

Summary:

The way an unavailable (distractor) alternative impacts decision quality is of great theoretical importance. Previous work, led by some of the authors of this study, had converged on a nuanced conclusion wherein the distractor can both improve (positive distractor effect) and reduce (negative distractor effect) decision quality, contingent upon the difficulty of the decision problem. In very recent work, Cao and Tsetsos (2022) reanalyzed all relevant previous datasets and showed that once distractor trials are referenced to binary trials (in which the distractor alternative is not shown to participants), distractor effects are absent. Cao and Tsetsos further showed that human participants heavily relied on additive (and not multiplicative) integration of rewards and probabilities.

The present study by Wong et al. puts forward a novel thesis according to which interindividual differences in the way of combining reward attributes underlie the absence of detectable distractor effect at the group level. They re-analysed the 144 human participants and classified participants into a "multiplicative integration" group and an "additive integration" group based on a model parameter, the "integration coefficient", that interpolates between the multiplicative utility and the additive utility in a mixture model. They report that participants in the "multiplicative" group show a negative distractor effect while participants in the "additive" group show a positive distractor effect. These findings are extensively discussed in relation to the potential underlying neural mechanisms.

Strengths:

- The study is forward looking, integrating previous findings well, and offering a novel proposal on how different integration strategies can lead to different choice biases.<br /> - The authors did an excellent job in connecting their thesis with previous neural findings. This is a very encompassing perspective that is likely to motivate new studies towards better understanding of how humans and other animals integrate information in decisions under risk and uncertainty.<br /> - Despite that some aspects of the paper are very technical, methodological details are well explained and the paper is very well written.

Weaknesses:

- The authors quantify the distractor variable as "DV - HV", i.e., the relative distractor variable. Conclusions mostly hold when the distractor is quantified in absolute terms (as "DV", see also Cao & Tsetsos, 2023). However, it is not entirely clear why the impact of the distractor alternative is not identical when the distractor variable is quantified in absolute vs. relative terms. Although understanding this nuanced point seems to extend beyond the scope of the paper, it could provide valuable decision-theoretic (and mechanistic) insights.<br /> - The central finding of this study is that participants who integrate reward attributes multiplicatively show a positive distractor effect while participants who integrate additively show a negative distractor effect. This is a very interesting and intriguing observation. However, it does not explain why the integration strategy covaries with the direction of the distractor effect. As the authors acknowledge, the composite model is not explanatory. Although beyond the scope of this paper, it would be valuable to provide a mechanistic explanation of this covariation pattern.

Reviewer #1 (Public Review):

Summary:

The current study provided a follow-up analysis using published datasets focused on the individual variability of both the distraction effect (size and direction) and the attribute integration style, as well as the association between the two. The authors tried to answer the question of whether the multiplicative attribute integration style concurs with a more pronounced and positively oriented distraction effect.

Strengths:

The analysis extensively examined the impacts of various factors on decision accuracy, with particular focus on using two-option trials as control trials, following the approach established by Cao & Tsetsos (2022). The statistical significance results were clearly reported.

The authors meticulously conducted supplementary examinations, incorporating the additional term HV+LV into GLM3. Furthermore, they replaced the utility function from the expected value model with values from the composite model.

Weaknesses:

The authors did a great job addressing the weaknesses I raised in the previous round of review, except on the generalizability of the current result in the larger context of multi-attribute decision-making. It is not really a weakness of the manuscript but more of a limitation of the studied topic, so I want to keep this comment for public readers.

The reward magnitude and probability information are displayed using rectangular bars of different colors and orientations. Would that bias subjects to choose an additive rule instead of the multiplicative rule? Also, could the conclusion be extended to other decision contexts such as quality and price, where a multiplicative rule is hard to formulate?

Overall, the authors have achieved their aims after clarifying that the study was trying to establish a correlation between the integration style and attraction effect. This result may be useful to inspire neuroimaging or neuromodulation studies that investigate multi-attribute decision making.

Reviewer #2 (Public Review):

This paper addresses the empirical demonstration of "distractor effects" in multi-attribute decision-making. It continues a debate in the literature on the presence (or not) of these effects, which domains they arise in, and their heterogeneity across subjects. The domain of the study is in a particular type of multi-attribute decision-making: choices over risky lotteries. The paper reports a re-analysis of lottery data from multiple experiments run previously by the authors and other labs involved in the debate.

Methodologically, the analysis assumes a number of simple forms for how attributes are aggregated (adaptively, or multiplicatively, or both) and then applies a "reduced form" logistic regression to the choices with a number of interaction terms intended to control for various features of the choice set. One of these interactions, modulated by ternary/binary treatment, is interpreted as a "distractor effect."

The claimed contribution of the re-analysis is to demonstrate correlation in the strength/sign of this treatment effect with another estimated parameter: the relative mixture of additive/multiplicative preferences.

Major Issues

(1) How to Interpret GLM 1 and 2

This paper, and others before it, have used a binary logistic regression with a number of interaction terms to attempt to control for various features of the choice set and how they influence choice. It is important to recognize that this modelling approach is not derived from a theoretical claim about the form of the computational model that guides decision-making in this task, nor an explicit test for a distractor effect. This can be seen most clearly in the equations after line 321 and its corresponding log-likelihood after 354, which contain no parameter or test for "distractor effects". Rather the computational model assumes a binary choice probability, and then shoehorns the test for distractor effects via a binary/ternary treatment interaction in a separate regression (GLM 1 and 2). This approach has already led to multiple misinterpretations in the literature (see Cao & Tsetsos, 2022; Webb et al., 2020). One of these misinterpretations occurred in the datasets the authors study, in which the lottery stimuli contained a confound with the interaction that Chau et al., (2014) were interpreting as a distractor effect (GLM 1). Cao & Tsetsos (2022) demonstrated that the interaction was significant in binary choice data from the study, therefore it can not be caused by a third alternative. This paper attempts to address this issue with a further interaction with the binary/ternary treatment (GLM 2). Therefore the difference in the interaction across the two conditions is claimed to now be the distractor effect. The validity of this claim brings us to what exactly is meant by a "distractor effect."

The paper begins by noting that "Rationally, choices ought to be unaffected by distractors" (line 33). This is not true. There are many normative models which allow for the value of alternatives (even low-valued "distractors") to influence choices, including a simple random utility model. Since Luce (1959), it has been known that the axiom of "Independence of Irrelevant Alternatives" (that the probability ratio between any two alternatives not depend on a third) is an extremely strong axiom, and only a sufficiency axiom for a random utility representation (Block and Marschak, 1959). It is not a necessary condition of a utility representation, and if this is our definition of rational (which is highly debatable), not necessary for it either. Countless empirical studies have demonstrated that IIA is falsified, and a large number of models can address it, including a simple random utility model with independent normal errors (i.e. a multivariate Probit model). In fact, it is only the multinomial Logit model that imposes IIA. It is also why so much attention is paid to the asymmetric dominance effect, which is a violation of a necessary condition for random utility (the Regularity axiom).

So what do the authors even mean by a "distractor effect." It is true that the form of IIA violations (i.e. their path through the probability simplex as the low-option varies) tells us something about the computational model underlying choice (after all, different models will predict different patterns). But we do not know how the interaction terms in the binary logit regression relate to the pattern of the violations because there is no formal theory that relates them. Any test for relative value coding is a joint test of the computational model and the form of the stochastic component (Webb et al,. 2020). These interaction terms may simply be picking up substitution patterns that can be easily reconciled with some form of random utility. While we can not check all forms of random utility in these datasets (because the class of such models is large), this paper doesn't even rule any of these models out.

(2) How to Interpret the Composite (Mixture) model?

On the other side of the correlation is the results from the mixture model for how decision-makers aggregate attributes. The authors report that most subjects are best represented by a mixture between additive and multiplicative aggregation models. The authors justify this with the proposal that these values are computed in different brain regions and then aggregated (which is reasonable, though raises the question of "where" if not the mPFC). But an equally reasonable interpretation is that the improved fit of the mixture model simply reflects a misspecification of two extreme aggregation process (additive and EV), so the log-likelihood is maximized at some point in between them.

One possibility is a model with utility curvature. How much of this result is just due to curvature in valuation? There are many reasonable theories for why we should expect curvature in utility for human subjects (for example, limited perception: Robson, 2001, Khaw, Li Woodford, 2019; Netzer et al., 2022) and of course many empirical demonstrations of risk aversion for small stakes lotteries. The mixture model, on the other hand, has parametric flexibility.

There is also a large literature on testing expected utility jointly with stochastic choice, and the impact of these assumptions on parameter interpretation (Loomes & Sugden, 1998; Apesteguia & Ballester, 2018; Webb, 2019). This relates back to the point above: the mixture may reflect the joint assumption of how choice departs from deterministic EV.

(3) So then how should we interpret the correlation that the authors report?

On one side we have the impact of the binary/ternary treatment which demonstrates some impact of the low value alternative on a binary choice probability. This may reflect some deep flaw in existing theories of choice, or it may simply reflect some departure from purely deterministic expected value maximization that existing theories can address. We have no theory to connect it to, so we cannot tell. On the other side of the correlation with have the mixture between additive and multiplicative preferences over risk. This result may reflect two distinct neural processes at work, or it may simply reflect a misspecification of the manner in which humans perceive and aggregate attributes of a lottery (or even just the stimuli in this experiment) by these two extreme candidates (additive vs. EV). Again, this would entail some departure from purely deterministic expected value maximization that existing theories can address.

It is entirely possible that the authors are reporting a result that points to the more exciting of these two possibilities. But it is also possible (and perhaps more likely) that the correlation is more mundane. The paper does not guide us to theories that predict such a correlation, nor reject any existing ones. In my opinion, we should be striving for theoretically-driven analyses of datasets, where the interpretation of results is clearer.

(4) Finally, the results from these experiments might not have external validity for two reasons. First, the normative criterion for multi-attribute decision-making differs depending on whether the attributes are lotteries or nor (i.e. multiplicative vs additive). Whether it does so for humans is a matter of debate. Therefore if the result is unique to lotteries, it might not be robust for multi-attribute choice more generally. The paper largely glosses over this difference and mixes literature from both domains. Second, the lottery information was presented visually and there is literature suggesting this form of presentation might differ from numerical attributes. Which is more ecologically valid is also a matter of debate.

Minor Issues:

The definition of EV as a normative choice baseline is problematic. The analysis requires that EV is the normative choice model (this is why the HV-LV gap is analyzed and the distractor effect defined in relation to it). But if the binary/ternary interaction effect can be accounted for by curvature of a value function, this should also change the definition of which lottery is HV or LV for that subject!

Comments on latest version: the authors did respond to some of my comments with discussion points in the paper.

References

Apesteguia, J. & Ballester, M. Monotone stochastic choice models: The case of risk and time preferences. Journal of Political Economy (2018).

Block, H. D. & Marschak, J. Random Orderings and Stochastic Theories of Responses. Cowles Foundation Discussion Papers (1959).

Khaw, M. W., Li, Z. & Woodford, M. Cognitive Imprecision and Small-Stakes Risk Aversion. Rev. Econ. Stud. 88, 1979-2013 (2020).

Loomes, G. & Sugden, R. Testing Different Stochastic Specifications of Risky Choice. Economica 65, 581-598 (1998).

Luce, R. D. Indvidual Choice Behaviour. (John Wiley and Sons, Inc., 1959).

Netzer, N., Robson, A. J., Steiner, J. & Kocourek, P. Endogenous Risk Attitudes. SSRN Electron. J. (2022) doi:10.2139/ssrn.4024773.

Robson, A. J. Why would nature give individuals utility functions? Journal of Political Economy 109, 900-914 (2001).

Webb, R. The (Neural) Dynamics of Stochastic Choice. Manage Sci 65, 230-255 (2019).

Reviewer #2 (Public Review):

This paper examined how the activity of neurons in the entopeduncular nucleus (EPN) of mice relates to kinematics, value, and reward. The authors recorded neural activity during an auditory-cued two-alternative choice task, allowing them to examine how neuronal firing relates to specific movements like licking or paw movements, as well as how contextual factors like task stage or proximity to a goal influence the coding of kinematic and spatiotemporal features. The data shows that the firing of individual neurons is linked to kinematic features such as lick or step cycles. However, the majority of neurons exhibited activity related to both movement types, suggesting that EPN neuronal activity does not merely reflect muscle-level representations. This contradicts what would be expected from traditional action selection or action specification models of the basal ganglia.

The authors also show that spatiotemporal variables account for more variability compared to kinematic features alone. Using demixed Principal Component Analysis, they reveal that at the population level, the three principal components explaining the most variance were related to specific temporal or spatial features of the task, such as ramping activity as mice approached reward ports, rather than trial outcome or specific actions. Notably, this activity was present in neurons whose firing was also modulated by kinematic features, demonstrating that individual EPN neurons integrate multiple features. A weakness is that what the spatiotemporal activity reflects is not well specified. The authors suggest some may relate to action value due to greater modulation when approaching a reward port, but acknowledge action value is not well parametrized or separated from variables like reward expectation.

A key goal was to determine whether activity related to expected value and reward delivery arose from a distinct population of EPN neurons or was also present in neurons modulated by kinematic and spatiotemporal features. In contrast to previous studies (Hong & Hikosaka 2008 and Stephenson-Jones et al., 2016), the current data reveals that individual neurons can exhibit modulation by both reward and kinematic parameters. Two potential differences may explain this discrepancy: First, the previous studies used head-fixed recordings, where it may have been easier to isolate movement versus reward-related responses. Second, those studies observed prominent phasic responses to the delivery or omission of expected rewards - responses largely absent in the current paper. This absence suggests a possibility that neurons exhibiting such phasic "reward" responses were not sampled, which is plausible since in both primates and rodents, these neurons tend to be located in restricted topographic regions. Alternatively, in the head-fixed recordings, kinematic/spatial coding may have gone undetected due to the forced immobility.

Overall, this paper offers needed insight into how the basal ganglia output encodes behavior. The EPN recordings from freely moving mice clearly demonstrate that individual neurons integrate reward, kinematic, and spatiotemporal features, challenging traditional models. However, the specific relationship between spatiotemporal activity and factors like action value remains unclear.

Reviewer #1 (Public Review):

The authors in this paper investigate the nature of the activity in the rodent EPN during a simple freely moving cue-reward association task. Given that primate literature suggests movement coding whereas other primate and rodent studies suggest mainly reward outcome coding in the EPNs, it is important to try to tease apart the two views. Through careful analysis of behavior kinematics, position, and neural activity in the EPNs, the authors reveal an interesting and complex relationship between the EPN and mouse behavior.

Strengths:

(1) The authors use a novel freely moving task to study EPN activity, which displays rich movement trajectories and kinematics. Given that previous studies have mostly looked at reward coding during head-fixed behavior, this study adds a valuable dataset to the literature.

(2) The neural analysis is rich and thorough. Both single neuron level and population level (i.e. PCA) analysis are employed to reveal what EPN encodes.

Weaknesses:

(1) One major weakness in this paper is the way the authors define the EPN neurons. Without a clear method of delineating EPN vs other surrounding regions, it is not convincing enough to call these neurons EPNs solely from looking at the electrode cannula track from Figure 2B. Indeed, EPN is a very small nucleus and previous studies like Stephenson-Jones et al (2016) have used opto-tagging of Vglut2 neurons to precisely label EPN single neurons. Wallace et al (2017) have also shown the existence of SOM and PV-positive neurons in the EPN. By not using transgenic lines and cell-type specific approaches to label these EPN neurons, the authors miss the opportunity to claim that the neurons recorded in this study do indeed come from EPN. The authors should at least consider showing an analysis of neurons slightly above or below EPN and show that these neurons display different waveforms or firing patterns.

(2) The authors fail to replicate the main finding about EPN neurons which is that they encode outcome in a negative manner. Both Stephenson-Jones et al (2016) and Hong and Hikosaka (2008) show a reward response during the outcome period where firing goes down during reward and up during neutral or aversive outcome. However, Figure 2 G top panel shows that the mean population is higher during correct trials and lower during incorrect trials. This could be interesting given that the authors might try recording from another part of EPN that has not been studied before. However, without convincing evidence that the neurons recorded are from EPN in the first place (point 1), it is hard to interpret these results and reconcile them with previous studies.

3) The authors say that: 'reward and kinematic doing are not mutually exclusive, challenging the notion of distinct pathways and movement processing'. However, it is not clear whether the data presented in this work supports this statement. First, the authors have not attempted to record from the entire EPN. Thus it is possible that the coding might be more segregated in other parts of EPN. Second, EPNs have previously been shown to display positive firing for negative outcomes and vice versa, something which the authors do not find here. It is possible that those neurons might not encode kinematic and movement variables. Thus, the authors should point out in the main text the possibility that the EPN activity recorded might be missing some parts of the whole EPN.

4). The authors use an IR beam system to record licks and make a strong claim about the nature of lick encoding in the EPN. However, the authors should note that IR beam system is not the most accurate way of detecting licks given that any object blocking the path (paw or jaw-dropping) will be detected as lick events. Capacitance based, closed-loop detection, or video capturing is better suited to detect individual licks. Given that the authors are interested in kinematics of licking, this is important. The authors should either point this out in the main text or verify in the system if the IR beam is correctly detecting licks using a combination of those methods.

Reviewer #1 (Public Review):

Summary:

Rook et al examined the role of BMP signaling in cerebellum development, using chick as a model alongside human tissue samples. They first examined p-SMADs and found differences between the species, with human samples retaining high p-SMAD after foliation, while in chick, BMP signaling appears to decrease following foliation. To understand the role of BMP during early development, they then used early chick embryos to modulate BMP, using either a constitutively active BMP regulator to increase BMP signaling or overexpressing the negative intracellular BMP regulator to decrease BMP signaling. After validating the constructs in ovo, the authors then examined GNP morphology and migration. They then determined whether the effects were cell autonomous.

Strengths:

The experiments were well-designed and well-controlled. The figures were extremely clear and convincing, and the accompanying drawings help orient the reader to easily understand the experimental set up. These studies also help clarify the role of BMP at different stages of cerebellum development, suggesting early BMP signaling is required for dorsalization, not rhombic lip induction, and that later BMP signaling is needed to regulate the timing of migration and maturation of granule neurons.

Weaknesses:

While these studies certainly hint that BMP modulation may affect tumor growth, this was not explicitly tested here. Future studies are required to generalize the functional role of BMP signaling in normal cerebellum development to malignant growth.

Reviewer #2 (Public Review):

Summary:

This is a fundamental and elegant study showing the role of BMP signaling in cerebellar development. This is an important question because there are multiple diseases, including aggressive childhood cancers, which involve granule cell precursors. Thus understanding of the factors that govern the formation of the granule cell layer is important both from a basic science and a disease perspective.

Overall, the manuscript is clear and well-written. The figures are extremely clear, wonderfully informative, and overall quite beautiful.

Figures 1-3 show the experimental design and report how BMP activity is altered over development in both the chick and the human developing cerebellum. Both data is very impressive and convincing.

They then go on to modulate BMP activity in the developing chick, using a complex electroporation paradigm that allows them to label cells with GFP as well as with cell-specific reporters of BMP activity levels. They bidirectionally modulate BMP levels and then can look at both cell-specific and non-specific alterations in the formation of the external and internal granule cell layer, across different developmental timepoints. These are really elegant and rigorous experiments, as they look at both sagittal and transverse sections to collect this data. This makes the data extremely compelling. With these rigorous techniques, they show that BMP signaling serves more than one function across development: it is involved in the initial tangential migration from the rhombic lip, but at a later time, both up- and down-regulation of BMP activity reduces density of amplifying cells in the external granule cell layer.

Strengths:

Overall, I think the paper is interesting and important and the data is strong. The use of both chick and human tissue strengthens the findings. They are extremely rigorous, analyzing data from multiple planes at multiple ages, which also really strengthens their findings. The dual electroporation approach is extremely elegant, providing beautiful visual representations of their findings.

Weaknesses:

I find no significant weaknesses.

Reviewer #1 (Public Review):

Summary:

Zhu et al. set out to better understand the neural mechanisms underlying Drosophila larval escape behavior. The escape behavior comprises several sequenced movements, including a lateral roll motion followed by fast crawling. The authors specifically were looking to identify neurons important for the roll-to-crawl transition.

Strengths:

This paper is clearly written, and the experiments are logical and complementary. They support the author's main claim that SeIN128 is a type of descending neuron that is both necessary and sufficient to modulate the termination of rolling. In general, the rigor is high.

Weaknesses:

-This manuscript is narrowly focused on Drosophila larval escape behavior. It would be more accessible to a broader audience if this work were put into a larger context of descending control.

Reviewer #2 (Public Review):

Summary:

The authors have addressed the majority of my comments, and I believe the revised manuscript has improved significantly.

The escape behavior of Drosophila larvae includes rolling followed by fast crawling, but the neural mechanism of this sequence was unclear. The authors determined the function of SeIN128, a group of descending neurons that terminate rolling and shorten crawling latency. SeIN128 receives inputs from Basin-2 and A00c neurons, which facilitate rolling, and makes reciprocal inhibitory synapses onto Basin-2 and A00c. SeIN128 shows a delayed activity peak upon Basins or A00c stimulation. Gad staining indicates that SeIN128 neurons are GABAergic, and blocking of SeIN128 function caused increased rolling probability and prolonged rolling. RNAi knockdown of GABA receptors in Basins suggests that several GABA receptors, especially GABA-A-R, mediate the SeIN128 to Basins inhibition. Among Basins subtypes, both Basin-2 and Basin-4 facilitate rolling but SeIN128 specifically terminates rolling elicited by Basin-2 activation. Overall, SeIN128 forms a feedback inhibition ensemble with Basin-2 and A00c that terminates rolling and shifts the animal to crawling.

Overall, this study discovered a neural mechanism that serves as a switch from rolling to fast crawling behaviors in Drosophila larvae. It addressed important open questions of how neural circuits determine the sequence of locomotor behaviors and how animals switch from one behavior to another. Its results support the conclusions and are backed up with proper control experiments.

Strengths:

- The question (i.e., the neural circuitry of action selection) addressed by this study is important.<br /> - Larval and adult Drosophila is a powerful model system in neuroscience study, with rich genetic tools, diverse behaviors, and well-studied nervous systems. This study makes good use of them.<br /> - The experiments, analyses, and results are rigorous and support the major claims. This study combined multiple innovative approaches, such as automated, machine-learning-based behavioral assays, EM reconstruction of larval CNS neurons, and genetic manipulation of specific neurons. A wide range of control experiments enhanced the credibility of the results.<br /> - The graphical representations are clear and mindfully arranged.

Weaknesses:

I believe "Corkscrew-like rolling" is not an accurate term for larval rolling. The neuromuscular basis of rolling was recently studied by Cooney et. al., showing that rolling is the circumferential propagation of muscle activity where all segments contract similarly and synchronously. So using another term instead of "Corkscrew-like rolling" may help.

Reviewer #3 (Public Review):

Summary:

Combining the behavioral assays with optogenetics, imaging, and connectome approaches, this meticulous study characterizes the underlying neuronal mechanisms of escape behavior in Drosophila larvae. The authors identify the neurons and provide convincing evidence to support their function in the roll-to-crawl locomotor transition.

Strengths:

It is a very thorough characterization of locomotor sequences in terms of underlying neural circuits. The findings shed light on investigating the analogous behaviors in other systems.

Weaknesses:

None. The authors have revised the article to improve the presentation and clarity.

Reviewer #1 (Public Review):

Summary:

Campbell et al investigated the effects of light on the human brain, in particular the subcortical part hypothalamus during auditory cognitive tasks. The mechanisms and neuronal circuits underlying light effects in non-image forming responses are so far mostly studied in rodents but are not easily translated in humans. Therefore, this is a fundamental study aiming to establish the impact light illuminance has on the subcortical structures using the high-resolution 7T fMRI. The authors found that parts of the hypothalamus are differently responding to illuminance. In particular, they found that the activity of the posterior hypothalamus increases while the activity of the anterior and ventral parts of the hypothalamus decreases under high illuminance. The authors also report that the performance of the 2-back executive task was significantly better in higher illuminance conditions. However, it seems that the activity of the posterior hypothalamus subpart is negatively related to the performance of the executive task, implying that it is unlikely that this part of the hypothalamus is directly involved in the positive impact of light on performance observed. Interestingly, the activity of the posterior hypothalamus was, however, associated with an increased behavioural response to emotional stimuli. This suggests that the role of this posterior part of the hypothalamus is not as simple regarding light effects on cognitive and emotional responses. This study is a fundamental step towards our better understanding of the mechanisms underlying light effects on cognition and consequently optimising lighting standards.

Strengths:

While it is still impossible to distinguish individual hypothalamic nuclei, even with the high-resolution fMRI, the authors split the hypothalamus into five areas encompassing five groups of hypothalamic nuclei. This allowed them to reveal that different parts of the hypothalamus respond differently to an increase in illuminance. They found that higher illuminance increased the activity of the posterior part of the hypothalamus encompassing the MB and parts of the LH and TMN, while decreasing the activity of the anterior parts encompassing the SCN and another part of TMN. These findings are somewhat in line with studies in animals. It was shown that parts of the hypothalamus such as SCN, LH, and PVN receive direct retinal input in particular from ipRGCs. Also, acute chemogenetic activation of ipRGCs was shown to induce activation of LH and also increased arousal in mice.

Weaknesses:

While the light characteristics are well documented and EDI calculated for all of the photoreceptors, it is not very clear why these irradiances and spectra were chosen. It would be helpful if the authors explained the logic behind the four chosen light conditions tested. Also, the lights chosen have cone-opic EDI values in a high correlation with the melanopic EDI, therefore we can't distinguish if the effects seen here are driven by melanopsin and/or other photoreceptors. In order to provide a more mechanistic insight into the light-driven effects on cognition ideally one would use silent substitution approach to distinguish between different photoreceptors. This may be something to consider when designing the follow-up studies.

Reviewer #2 (Public Review):

Summary

The interplay between environmental factors and cognitive performance has been a focal point of neuroscientific research, with illuminance emerging as a significant variable of interest. The hypothalamus, a brain region integral to regulating circadian rhythms, sleep, and alertness, has been posited to mediate the effects of light exposure on cognitive functions. Previous studies have highlighted the role of the hypothalamus in orchestrating bodily responses to light, implicating specific neural pathways such as the orexin and histamine systems, which are crucial for maintaining wakefulness and processing environmental cues. Despite advancements in our understanding, the specific mechanisms through which varying levels of light exposure influence hypothalamic activity and, in turn, cognitive performance, remain inadequately explored. This gap in knowledge underscores the need for high-resolution investigations that can dissect the nuanced impacts of illuminance on different hypothalamic regions. Utilizing state-of-the-art 7 Tesla functional magnetic resonance imaging (fMRI), the present study aims to elucidate the differential effects of light on hypothalamic dynamics and establish a link between regional hypothalamic activity and cognitive outcomes in healthy young adults. By shedding light on these complex interactions, this research endeavours to contribute to the foundational knowledge necessary for developing innovative therapeutic strategies aimed at enhancing cognitive function through environmental modulation.

Strengths:

(1) Considerable Sample Size and Detailed Analysis: The study leverages a robust sample size and conducts a thorough analysis of hypothalamic dynamics, which enhances the reliability and depth of the findings.<br /> (2) Use of High-Resolution Imaging: Utilizing 7 Tesla fMRI to analyze brain activity during cognitive tasks offers high-resolution insights into the differential effects of illuminance on hypothalamic activity, showcasing the methodological rigour of the study.<br /> (3) Novel Insights into Illuminance Effects: The manuscript reveals new understandings of how different regions of the hypothalamus respond to varying illuminance levels, contributing valuable knowledge to the field.<br /> (4) Exploration of Potential Therapeutic Applications: Discussing the potential therapeutic applications of light modulation based on the findings suggests practical implications and future research directions.

The current version of the manuscript addresses previous weaknesses, including details about the illuminance levels, light spectral characteristics used in the MRI study, and light patterns during behavioural tasks. The authors effectively tackle open questions in the field and provide solid evidence that enhances our understanding of the mechanisms underlying the effects of light on cognition.

Reviewer #3 (Public Review):

Summary:

Campbell and colleagues use a combination of high-resolution fMRI, cognitive tasks and different intensities of light illumination to test the hypothesis that the intensity of illumination differentially impacts hypothalamic substructures that, in turn, promote alterations in arousal that affect cognitive and affective performance. The authors find evidence in support of a posterior-to-anterior gradient of increased blood flow in the hypothalamus during task performance that they later relate to performance on two different tasks. The results provide an enticing link between light levels, hypothalamic activity and cognitive/affective function, however clarification of some methodological choices will help to improve confidence in the findings.

Strengths:

* The authors' focus on the hypothalamus and its relationship to light intensity is an important and understudied question in neuroscience.

Weaknesses:

* I found it challenging to relate the authors hypotheses, which I found to be quite compelling, to the apparatus used to test the hypotheses - namely, the use of orange light vs. different light intensities; and the specific choice of the executive and emotional tasks, which differed in key features (e.g., block-related vs. event-related designs) that were orthogonal to the psychological constructs being challenged in each task.

* Given the small size of the hypothalamus and the irregular size of the hypothalamic parcels, I wondered whether a more data-driven examination of the hypothalamic time series would have provided a more parsimonious test of their hypothesis.

Reviewer #1 (Public Review):

Summary:

The current study aims to quantify associations between regular use of proton-pump inhibitors (PPI) - defined as using PPI most days of the week during the last 4 weeks at one cross-section in time - with several respiratory outcomes (6 in total: risk of influenza, pneumonia, COVID-19, other respiratory tract infections, as well as COVID-19 severity and mortality) up to several years later in time.

Strengths:

Several sensitivity analyses were performed, including i) estimation of the e-value to assess how strong unmeasured confounders should be to explain observed effects, ii) comparison with another drug with a similar indication to potentially reduce (but not eliminate) confounding by indication, iii)

Weaknesses:

While the original submission had several weaknesses, the authors have appropriately addressed all issues raised. There are inevitable weaknesses remaining, but these are appropriately highlighted in the discussion. Remaining weaknesses that remain - but are highlighted in the discussion - include the fact that the main exposure of interest is only measured at one time-point whereas outcomes are assessed over a long time period, the inclusion of prevalent users leading to potential bias (e.g. those experiencing bad outcomes already stopping because of side-effects before inclusion in the study), and the possibility of unmeasured confounding explaining observations (e.g. severity of underlying comorbidities leading to PPI prescriptions combined with the absence of information about comorbidity severity), and potential selection bias.

Reviewer #1 (Public Review):

Summary:

Horn and colleagues present data suggesting that the targeting of GREM1 has little impact on a mouse model of metabolic dysfunction-associated steatohepatitis. Importantly, they also challenge existing data on the detection of GREM1 by ELISA in serum or plasma by demonstrating that high-affinity binding of GREM1 to heparin would lead to localisation of GREM1 in the ECM or at the plasma membrane of cells.

Strengths:

This is an impressive tour-de-force study around the potential of targeting GREM1 in MASH.

This paper will challenge many existing papers in the field around our ability to detect GREM1 in circulation, at least using antibody-mediated detection.

Well-controlled, detailed studies like this are critically important in order to challenge less vigorous studies in the literature.

The impressive volume of high-level, well-controlled data using an impressive range of in vitro biochemical techniques, rodent models, and human liver slices.

Weaknesses: only minor.

(1) The authors clearly show that heparin can limit the diffusion of GREM1 into the circulation-however, in a setting where GREM1 is produced in excess (e.g. cancer), could this "saturate" the available heparin and allow GREM1 to "escape" into the circulation?

(2) Secondly, has the author considered that GREM1 be circulating bound to a chaperone protein like albumin which would reduce its reactivity with GREM1 detection antibodies?

(3) Statistics-there is no mention of blinding of samples-I assume this was done prior to analysis?

(4) Line 211-I suggest adding the Figure reference at the end of this sentence to direct the reader to the relevant data.

(5) Figure 1E Y-axis units are a little hard to interpret-can integers be used?

(6) Did the authors attempt to detect GREM1 protein by IHC? There are published methods for this using the R&D Systems mouse antibody (PMID 31384391).

(7) Did the authors ever observe GREM1 internalisation using their Atto-532 labelled GREM1?

(8) Did the authors complete GREM1 ISH in the rat CDAA-HFD model? Was GREM1 upregulated, and if so, where?

(9) Supplementary Figure 4C - why does the GFP level decrease in the GREM1 transgenic compared to control the GFP mouse? No such change is observed in Supplementary Figure 4E.

Reviewer #2 (Public Review):

It is controversial whether liver gremlin-1 expression correlates with liver fibrosis in metabolic dysfunction-associated steatohepatitis (MASH). Horn et al. developed an anti-Gremlin-1 antibody in-house and tested its ability to neutralize gremlin-1 and treat liver fibrosis. This article has the advantage of testing its hypothesis with different animal and human liver fibrosis models and using a variety of research methodologies.

The experimental design and results support the conclusion that the anti-gremlin-1 antibody had no therapeutic effect on treating liver fibrosis, so there are no other suggestions for new experiments:

(1) The authors used RNAscope in situ hybridization to establish the correlation between Gremlin-1 expression and NMSH livers or cell lines.

(2) A luminescent oxygen channelling immunoassay was used to measure circulating Gremlin-1 concentration. They found that Gremlin-1 binds to heparin very efficiently, preventing Gremlin-1 from entering circulation, and restricting Gremlin-1's ability to mediate organ cross-communication.

(3) The authors developed a suitable NMSH rat model which is a choline-deficient, L-amino acid defined high fat 1% cholesterol diet (CDAA-HFD) fed rat model of NMSH, and created a selective anti-Gremlin-1 antibody which is heparin-displacing 0030:HD antibody. They also used human cirrhotic precision-cut liver slices to test their hypotheses. They demonstrated that neutralization of Gremlin-1 activity with monoclonal therapeutic antibodies does not reduce liver inflammation or liver fibrosis.

One concern is that several reagents and assays are made in-house without external validation. Also, will those in-house reagents and assays be available to the science community?

Overall this manuscript provides useful information that gremlin-1 has a limited role in liver fibrosis pathogenesis and treatment.

Reviewer #1 (Public Review):

In their manuscript, the authors propose a learning scheme to enable spiking neurons to learn the appearance probability of inputs to the network. To this end, the neurons rely on error-based plasticity rules for feedforward and recurrent connections. The authors show that this enables the networks to spontaneously sample assembly activations according to the occurrence probability of the input patterns they respond to. They also show that the learning scheme could explain biases in decision-making, as observed in monkey experiments. While the task of neural sampling has been solved before in other models, the novelty here is the proposal that the main drivers of sampling are within-assembly connections, and not between-assembly (Markov chains) connections as in previous models. This could provide a new understanding of how spontaneous activity in the cortex is shaped by synaptic plasticity.

The manuscript is well written and the results are presented in a clear and understandable way. The main results are convincing, concerning the spontaneous firing rate dependence of assemblies on input probability, as well as the replication of biases in the decision-making experiment. Nevertheless, the manuscript and model leave open several important questions. The main problem is the unclarity, both in theory and intuitively, of how the sampling exactly works. This also makes it difficult to assess the claims of novelty the authors make, as it is not clear how their work relates to previous models of neural sampling.

Regarding the unclarity of the sampling mechanism, the authors state that within-assembly excitatory connections are responsible for activating the neurons according to stimulus probability. However, the intuition for this process is not made clear anywhere in the manuscript. How do the recurrent connections lead to the observed effect of sampling? How exactly do assemblies form from feedforward plasticity? This intuitive unclarity is accompanied by a lack of formal justification for the plasticity rules. The authors refer to a previous publication from the same lab, but it is difficult to connect these previous results and derivations to the current manuscript. The manuscript should include a clear derivation of the learning rules, as well as an (ideally formal) intuition of how this leads to the sampling dynamics in the simulation.

Some of the model details should furthermore be cleared up. First, recurrent connections transmit signals instantaneously, which is implausible. Is this required, would the network dynamics change significantly if, e.g., excitation arrives slightly delayed? Second, why is the homeostasis on h required for replay? The authors show that without it the probabilities of sampling are not matched, but it is not clear why, nor how homeostasis prevents this. Third, G and M have the same plasticity rule except for G being confined to positive values, but there is no formal justification given for this quite unusual rule. The authors should clearly justify (ideally formally) the introduction of these inhibitory weights G, which is also where the manuscript deviates from their previous 2020 work. My feeling is that inhibitory weights have to be constrained in the current model because they have a different goal (decorrelation, not prediction) and thus should operate with a completely different plasticity mechanism. The current manuscript doesn't address this, as there is no overall formal justification for the learning algorithm.

Finally, the authors should make the relation to previous models of sampling and error-based plasticity more clear. Since there is no formal derivation of the sampling dynamics, it is difficult to assess how they differ exactly from previous (Markov-based) approaches, which should be made more precise. Especially, it would be important to have concrete (ideally experimentally testable) predictions on how these two ideas differ. As a side note, especially in the introduction (line 90), this unclarity about the sampling made it difficult to understand the contrast to Markovian transition models.

There are also several related models that have not been mentioned and should be discussed. In 663 ff. the authors discuss the contributions of their model which they claim are novel, but in Kappel et al (STDP Installs in Winner-Take-All Circuits an Online Approximation to Hidden Markov Model Learning) similar elements seem to exist as well, and the difference should be clarified. There is also a range of other models with lateral inhibition that make use of error-based plasticity (most recently reviewed in Mikulasch et al, Where is the error? Hierarchical predictive coding through dendritic error computation), and it should be discussed how the proposed model differs from these.

Reviewer #2 (Public Review):

Summary:

The paper considers a recurrent network with neurons driven by external input. During the external stimulation predictive synaptic plasticity adapts the forward and recurrent weights. It is shown that after the presentation of constant stimuli, the network spontaneously samples the states imposed by these stimuli. The probability of sampling stimulus x^(i) is proportional to the relative frequency of presenting stimulus x^(i) among all stimuli i=1,..., 5.

Methods:

Neuronal dynamics:

For the main simulation (Figure 3), the network had 500 neurons, and 5 non-overlapping stimuli with each activating 100 different neurons where presented. The voltage u of the neurons is driven by the forward weights W via input rates x, the inhibitory recurrent weights G, are restricted to have non-negative weights (Dale's law), and the other recurrent weights M had no sign-restrictions. Neurons were spiking with an instantaneous Poisson firing rate, and each spike-triggered an exponentially decaying postsynaptic voltage deflection. Neglecting time constants of the postsynaptic responses, the expected postsynaptic voltage reads (in vectorial form) as

u = W x + (M - G) f (Eq. 5)

where f =; phi(u) represents the instantaneous Poisson rate, and phi a sigmoidal nonlinearity. The rate f is only an approximation (symbolized by =;) of phi(u) since an additional regularization variable h enters (taken up in Point 4 below). The initialisation of W and M is Gaussian with mean 0 and variance 1/sqrt(N), N the number of neurons in the network. The initial entries of G are all set to 1/sqrt(N).

Predictive synaptic plasticity:

The 3 types of synapses were each adapted so that they individually predict the postsynaptic firing rate f, in matrix form

ΔW ≈ (f - phi( W x ) ) x^T<br /> ΔM ≈ (f - phi( M f ) ) f^T<br /> ΔG ≈ (f - phi( M f ) ) f^T but confined to non-negative values of G (Dale's law).

The ^T tells us to take the transpose, and the ≈ again refers to the fact that the ϕ entering in the learning rule is not exactly the ϕ determining the rate, only up to the regularization (see Point 4).

Main formal result:

As the authors explain, the forward weight W and the unconstrained weight M develop such that, in expectations,

f =; phi( W x ) =; phi( M f ) =; phi( G f ) ,

consistent with the above plasticity rules. Some elements of M remain negative. In this final state, the network displays the behaviour as explained in the summary.

Major issues:

Point 1: Conceptual inconsistency

The main results seem to arise from unilaterally applying Dale's law only to the inhibitory recurrent synapses G, but not to the excitatory recurrent synapses M.

In fact, if the same non-negativity restriction were also imposed on M (as it is on G), then their learning rules would become identical, likely leading to M=G. But in this case, the network becomes purely forward, u = W x, and no spontaneous recall would arise. Of course, this should be checked in simulations.

Because Dale's law was only applied to G, however, M and G cannot become equal, and the remaining differences seem to cause the effect.

Predictive learning rules are certainly powerful, and it is reasonable to consider the same type of error-correcting predictive learning rule, for instance for different dendritic branches that both should predict the somatic activity. Or one may postulate the same type of error-correcting predictive plasticity for inhibitory and excitatory synapses, but then the presynaptic neurons should not be identical, as it is assumed here. Both these types of error-correcting and error-forming learning rules for same-branches and inhibitory/excitatory inputs have been considered already (but with inhibitory input being itself restricted to local input, for instance).

Point 2: Main result as an artefact of an inconsistently applied Dale's law?

The main result shows that the probability of a spontaneous recall for the 5 non-overlapping stimuli is proportional to the relative time the stimulus was presented. This is roughly explained as follows: each stimulus pushes the activity from 0 up towards f =; phi( W x ) by the learning rule (roughly). Because the mean weights W are initialized to 0, a stimulus that is presented longer will have more time to push W up so that positive firing rates are reached (assuming x is non-negative). The recurrent weights M learn to reproduce these firing rates too, while the plasticity in G tries to prevent that (by its negative sign, but with the restriction to non-negative values). Stimuli that are presented more often, on average, will have more time to reach the positive target and hence will form a stronger and wider attractor. In spontaneous recall, the size of the attractor reflects the time of the stimulus presentation. This mechanism so far is fine, but the only problem is that it is based on restricting G, but not M, to non-negative values.

Point 3: Comparison of rates between stimulation and recall.

The firing rates with external stimulations will be considerably larger than during replay (unless the rates are saturated).

This is a prediction that should be tested in simulations. In fact, since the voltage roughly reads as<br /> u = W x + (M - G) f,<br /> and the learning rules are such that eventually M =; G, the recurrences roughly cancel and the voltage is mainly driven by the external input x. In the state of spontaneous activity without external drive, one has<br /> u = (M - G) f ,<br /> and this should generate considerably smaller instantaneous rates f =; phi(u) than in the case of the feedforward drive (unless f is in both cases at the upper or lower ceiling of phi). This is a prediction that can also be tested.

Because the figures mostly show activity ratios or normalized activities, it was not possible for me to check this hypothesis with the current figures. So please show non-normalized activities for comparing stimulation and recall for the same patterns.

Point 4: Unclear definition of the variable h.<br /> The formal definition of h = hi is given by (suppressing here the neuron index i and the h-index of tau)

tau dh/dt = -h if h>u, (Eq. 10)<br /> h = u otherwise.

But if it is only Equation 10 (nothing else is said), h will always become equal to u, or will vanish, i.e. either h=u or h=0 after some initial transient. In fact, as soon as h>u, h is decaying to 0 according to the first line. If u is >0, then it stops at u=h according to the second line. No reason to change h=u further. If u<=0 while h>u, then h is converging to 0 according to the first line and will stay there. I guess the authors had issues with the recurrent spiking simulations and tried to fix this with some regularization. However as presented, it does not become clear how their regulation works.

BTW: In Eq. 11 the authors set the gain beta to beta = beta0/h which could become infinite and, putatively more problematic, negative, depending on the value of h. Maybe some remark would convince a reader that no issues emerge from this.

Added from discussions with the editor and the other reviewers:

Thanks for alerting me to this Supplementary Figure 8. Yes, it looks like the authors did apply there Dale's law for both the excitatory and inhibitory synapses. Yet, they also introduced two types of inhibitory pathways converging both to the excitatory and inhibitory neurons. For me, this is a confirmation that applying Dale's law to both excitatory and inhibitory synapses, with identical learning rules as explained in the main part of the paper, does not work.

Adding such two pathways is a strong change from the original model as introduced before, and based on which all the Figures in the main text are based. Supplementary Figure 8 should come with an analysis of why a single inhibitory pathway does not work. I guess I gave the reason in my Points 1-3. Some form of symmetry breaking between the recurrent excitation and recurrent inhibition is required so that, eventually, the recurrent excitatory connection will dominate.

Making the inhibitory plasticity less expressive by applying Dale's law to only those inhibitory synapses seems to be the answer chosen in the Figures of the main text (but then the criticism of unilaterally applying Dale's law).

Applying Dale's law to both types of synapses, but dividing the labor of inhibition into two strictly separate and asymmetric pathways, and hence asymmetric development of excitatory and inhibitory weights, seems to be another option. However, introducing such two separate inhibitory pathways, just to rescue the fact that Dale's law is applied to both types of synapses, is a bold assumption. Is there some biological evidence of such two pathways in the inhibitory, but not the excitatory connections? And what is the computational reasoning to have such a separation, apart from some form of symmetry breaking between excitation and inhibition? I guess, simpler solutions could be found, for instance by breaking the symmetry between the plasticity rules for the excitatory and inhibitory neurons. All these questions, in my view, need to be addressed to give some insights into why the simulations do work.

Overall, Supplementary Figure 8 seems to me too important to be deferred to the Supplement. The reasoning behind the two inhibitory pathways should appear more prominently in the main text. Without this, important questions remain. For instance, when thinking in a rate-based framework, the two inhibitory pathways twice try to explain the somatic firing rate away. Doesn't this lead to a too strong inhibition? Can some steady state with a positive firing rate caused by the recurrence, in the absence of an external drive, be proven? The argument must include the separation into Path 1 and Path 2. So far, this reasoning has not been entered.

In fact, it might be that, in a spiking implementation, some sparse spikes will survive. I wonder whether at least some of these spikes survive because of the other rescuing construction with the dynamic variable h (Equation 10, which is not transparent, and that is not taken up in the reasoning either, see my Point 4).

Perhaps it is helpful for the authors to add this text in the reply to them.

Reviewer #3 (Public Review):

Summary:

The work shows how learned assembly structure and its influence on replay during spontaneous activity can reflect the statistics of stimulus input. In particular, stimuli that are more frequent during training elicit stronger wiring and more frequent activation during replay. Past works (Litwin-Kumar and Doiron, 2014; Zenke et al., 2015) have not addressed this specific question, as classic homeostatic mechanisms forced activity to be similar across all assemblies. Here, the authors use a dynamic gain and threshold mechanism to circumnavigate this issue and link this mechanism to cellular monitoring of membrane potential history.

Strengths:

(1) This is an interesting advance, and the authors link this to experimental work in sensory learning in environments with non-uniform stimulus probabilities.

(2) The authors consider their mechanism in a variety of models of increasing complexity (simple stimuli, complex stimuli; ignoring Dale's law, incorporating Dale's law).

(3) Links a cellular mechanism of internal gain control (their variable h) to assembly formation and the non-uniformity of spontaneous replay activity. Offers a promise of relating cellular and synaptic plasticity mechanisms under a common goal of assembly formation.

Weaknesses:

(1) However, while the manuscript does show that assembly wiring does follow stimulus likelihood, it is not clear how the assembly-specific statistics of h reflect these likelihoods. I find this to be a key issue.

(2) The authors' model does take advantage of the sigmoidal transfer function, and after learning an assembly is either fully active or nearly fully silent (Figure 2a). This somewhat artificial saturation may be the reason that classic homeostasis is not required since runaway activity is not as damaging to network activity.

(3) Classic mechanisms of homeostatic regulation (synaptic scaling, inhibitory plasticity) try to ensure that firing rates match a target rate (on average). If the target rate is the same for all neurons then having elevated firing rates for one assembly compared to others during spontaneous activity would be difficult. If these homeostatic mechanisms were incorporated, how would they permit the elevated firing rates for assemblies that represent more likely stimuli?

Reviewer #1 (Public Review):

Summary:

The authors used a novel multi-dimensional experience sampling (mDES) approach to identify data-driven patterns of experience samples that they use to interrogate fMRI data collected during naturalistic movie-watching data. They identify a set of multi-sensory features of a set of movies that delineate low-dimensional gradients of BOLD fMRI signal patterns that have previously been linked to fundamental axes of cortical organization.

Strengths:

The novel solution to challenges associated with experience sampling offers potential access to aspects of experience that have been challenging to assess. While inventive, I worry that the reliability of the mDES approach is currently under-investigated, making it challenging to interpret the import of the later analyses, which are themselves strong and compelling.

Weaknesses:

The lack of direct interrogation of individual differences/reliability of the mDES scores warrants some pause.

Reviewer #2 (Public Review):

Summary:

The present study explores how thoughts map onto brain activity, a notoriously challenging question because of the dynamic, subjective, and abstract nature of thoughts. To tackle this question, the authors collected continuous thought ratings from participants watching a movie, and additionally made use of an open-source fMRI dataset recorded during movie watching as well as five established gradients of brain variation as identified in resting state data. Using a voxel-space approach, the results show that episodic knowledge, verbal detail, and sensory engagement of thoughts commonly modulate the activation of the visual and auditory cortex, while intrusive distraction modulates the frontoparietal network. Additionally, sensory engagement is mapped onto a gradient from the primary to the association cortex, while episodic knowledge is mapped onto a gradient from the dorsal attention network to the visual cortex. Building on the association between behavioral performance and neural activation, the authors conclude that sensory coupling to external input and frontoparietal executive control is key to comprehension in naturalistic settings.

The manuscript stands out for its methodological advancements in quantifying thoughts over time and its aim to study the implementation of thoughts in the brain during naturalistic movie watching. However, the conceptualization of thoughts remains vague, its distinction from other concepts like attention is unclear, and interindividual differences are not sufficiently addressed, limiting the study's insights into brain function.

Strengths:

(1) The study raises a question that has been difficult to study in naturalistic settings so far but is key to understanding human cognition, namely how thoughts map onto brain activation.

(2) The thought ratings introduce a novel method for continuously tracking thoughts, promising utility beyond this study.

(3) The authors substantiated the effects of thinking from multiple perspectives, using diverse data types, metrics, and analyses.

(4) The figures are highly informative, accessible, and consistent, aiding comprehension.

Weaknesses:

(1) The dimensions of thought seem to distinguish between sensory and executive processing states. However, it is unclear if this effect primarily pertains to thinking. I could imagine highly intrusive distractions in movie segments to correlate with stagnating plot development, little change in scenery, or incomprehensible events. Put differently, it may primarily be the properties of the movies that evoke different processing modes, but these properties are not accounted for. For example, I'm wondering whether a simple measure of engagement with stimulus materials could explain the effects just as much. How can the effects of thinking be distinguished from the perceptual and semantic properties of the movie, as well as attentional effects? Is the measure used here capturing thought processes beyond what other factors could explain?

(2) I'm skeptical about taking human thought ratings at face value. Intrusive distraction might imply disengagement from stimulus materials, but it could also be an intended effect of the movie to trigger higher-level, abstract thinking. Can a label like intrusive distraction be misleading without considering the actual thought and movie content?

(3) A jittered sampling approach is used to acquire thought ratings every 15 seconds. Are ratings for the same time point averaged across participants? If so, how consistent are ratings among participants? High consistency would suggest thoughts are mainly stimulus-evoked. Low consistency would question the validity of applying ratings from one (group of) participant(s) to brain-related analyses of another participant.

(4) Using three different movies to conclude that different genres evoke different thought patterns (e.g., line 277) seems like an overinterpretation with only one instance per genre.

(5) I see no indication that results were cross-validated, and no effect sizes are reported, leaving the robustness and strength of effects unknown.

Reviewer #3 (Public Review):

This study attempted to investigate the relationship between processing in the human brain during movie watching and corresponding thought processes. This is a highly interesting question, as movie watching presents a semi-constrained task, combining naturally occurring thoughts and common processing of sensory inputs across participants. This task is inherently difficult because in order to know what participants are thinking at any given moment, one has to interrupt the same thought process which is the object of study.

This study attempts to deal with this issue by aggregating staggered experience sampling data across participants in one behavioral study and using the population-level thought patterns to model brain activity in different participants in an open-access fMRI dataset.

The behavioral data consist of 120 participants who watched 3 11-minute movie clips. Participants responded to the mDES questionnaire: 16 visual scales characterizing ongoing thought 5 times, two minutes apart, in each clip. The 16 items are first reduced to 4 factors using PCA, and their levels are compared across the different movies. The factors are "episodic knowledge", "intrusive distraction", "verbal detail", and "sensory engagement". The factors differ between the clips, and distraction is negatively correlated with movie comprehension, and sensory engagement is positively correlated with comprehension.

The components are aggregated across participants (transforming single-subject mDES answers into PCA space and concatenating responses of different participants), and are used as regressors in a GLM analysis. This analysis identifies brain regions corresponding to the components. The resulting brain maps reveal activations that are consistent with the proposed mental processes (e.g. negative loading for intrusion in the frontoparietal network, and positive loadings for visual and auditory cortices for sensory engagement).

Then, the coordinates for brain regions that were significant for more than one component are entered into a paper search in neurosynth. It is not clear what this analysis demonstrates beyond the fact that sensory engagement contains both visual and auditory components.

The next analysis projected group-averaged brain activation onto gradients (based on previous work) and used gradient timecourses to predict the behavioral report timecourses. This revealed that high activations in gradient 1 (sensory→association) predicted high sensory engagement, and that "episodic knowledge" thought patterns were predicted by increased visual cortex activations. Then, permutation tests were performed to see whether these thought pattern-related activations corresponded to well-defined regions on a given cluster.

This paper is framed as presenting a new paradigm but it does little to discuss what this paradigm serves, what its limitations are, and how it should have been tested. I assume that the novelty is in using experience sampling from 1 sample to model the responses of a second sample.

What are the considerations for treating high-order thought patterns that occur during film viewing as stable enough to be used across participants? What would be the limitations of this method? (Do all people reading this paper think comparable thoughts reading through the sections?)

How does this approach differ from collaborative filtering, (for example as presented in Chang et al., 2021)?

In conclusion, this study tackles a highly interesting subject and does it creatively and expertly. It fails to discuss and establish the utility and appropriateness of its proposed method.

Luke J. Chang et al. ,Endogenous variation in ventromedial prefrontal cortex state dynamics during naturalistic viewing reflects affective experience.Sci. Adv.7,eabf7129(2021).DOI:10.1126/sciadv.abf7129

Reviewer #1 (Public Review):

Summary:

In this work, Noorman and colleagues test the predictions of the "four-stage model" of consciousness by combining psychophysics and scalp EEG in humans. The study relies on an elegant experimental design to investigate the respective impact of attentional and perceptual blindness on visual processing.

The study is very well summarised, the text is clear and the methods seem sound. Overall, a very solid piece of work. I haven't identified any major weaknesses. Below I raise a few questions of interpretation that may possibly be the subject of a revision of the text.

(1) The perceptual performance on Fig1D appears to show huge variation across participants, with some participants at chance levels and others with performance > 90% in the attentional blink and/or masked conditions. This seems to reveal that the procedure to match performance across participants was not very successful. Could this impact the results? The authors highlight the fact that they did not resort to post-selection or exclusion of participants, but at the same time do not discuss this equally important point.

(2) In the analysis on collinearity and illusion-specific processing, the authors conclude that the absence of a significant effect of training set demonstrates collinearity-only processing. I don't think that this conclusion is warranted: as the illusory and non-illusory share the same shape, so more elaborate object processing could also be occuring. Please discuss.

(3) Discussion, lines 426-429: It is stated that the results align with the notion that processes of perceptual segmentation and organization represent the mechanism of conscious experience. My interpretation of the results is that they show the contrary: for the same visibility level in the attentional blind or masking conditions, these processes can be implicated or not, which suggests a role during unconscious processing instead.

(4). The two paradigms developed here could be used jointly to highlight non-idiosyncratic NCCs, i.e. EEG markers of visibility or confidence that generalise regardless of the method used. Have the authors attempted to train the classifier on one method and apply it to another (e.g. AB to masking and vice versa)? What perceptual level is assumed to transfer?

(5). How can the results be integrated with the attentional literature showing that attentional filters can be applied early in the processing hierarchy?

Reviewer #2 (Public Review):

Summary:

This is a very elegant and important EEG study that unifies within a single set of behaviorally equated experimental conditions conscious access (and therefore also conscious access failures) during visual masking and attentional blink (AB) paradigms in humans. By a systematic and clever use of multivariate pattern classifiers across conditions, they could dissect, confirm, and extend a key distinction (initially framed within the GNWT framework) between 'subliminal' and 'pre-conscious' unconscious levels of processing. In particular, the authors could provide strong evidence to distinguish here within the same paradigm these two levels of unconscious processing that precede conscious access : (i) an early (< 80ms) bottom-up and local (in brain) stage of perceptual processing ('local contrast processing') that was preserved in both unconscious conditions, (ii) a later stage and more integrated processing (200-250ms) that was impaired by masking but preserved during AB. On the basis of preexisting studies and theoretical arguments, they suggest that this later stage could correspond to lateral and local recurrent feedback processes. Then, the late conscious access stage appeared as a P3b-like event.

Strengths:

The methodology and analyses are strong and valid. This work adds an important piece in the current scientific debate about levels of unconscious processing and specificities of conscious access in relation to feed-forward, lateral, and late brain-scale top-down recurrent processing.

Weaknesses:

- The authors could improve clarity of the rich set of decoding analyses across conditions.<br /> - They could also enrich their Introduction and Discussion sections by taking into account the importance of conscious influences on some unconscious cognitive processes (revision of traditional concept of 'automaticity'), that may introduce some complexity in Results interpretation<br /> - They should discuss the rich literature reporting high-level unconscious processing in masking paradigms (culminating in semantic processing of digits, words or even small group of words, and pictures) in the light of their proposal (deeper unconscious processing during AB than during masking).

Reviewer #3 (Public Review):

Summary:

This work aims to investigate how perceptual and attentional processes affect conscious access in humans. By using multivariate decoding analysis of electroencephalography (EEG) data, the authors explored the neural temporal dynamics of visual processing across different levels of complexity (local contrast, collinearity, and illusory perception). This is achieved by comparing the decidability of an illusory percept in matched conditions of perceptual (i.e., degrading the strength of sensory input using visual masking) and attentional impairment (i.e., impairing top-down attention using attentional blink, AB). The decoding results reveal three distinct temporal responses associated with the three levels of visual processing. Interestingly, the early stage of local contrast processing remains unaffected by both masking and AB. However, the later stage of collinearity and illusory percept processing are impaired by the perceptual manipulation but remain unaffected by the attentional manipulation. These findings contribute to the understanding of the unique neural dynamics of perceptual and attentional functions and how they interact with the different stages of conscious access.

Strengths:

The study investigates perceptual and attentional impairments across multiple levels of visual processing in a single experiment. Local contrast, collinearity, and illusory perception were manipulated using different configurations of the same visual stimuli. This clever design allows for the investigation of different levels of visual processing under similar low-level conditions.

Moreover, behavioural performance was matched between perceptual and attentional manipulations. One of the main problems when comparing perceptual and attentional manipulations on conscious access is that they tend to impact performance at different levels, with perceptual manipulations like masking producing larger effects. The study utilizes a staircasing procedure to find the optimal contrast of the mask stimuli to produce a performance impairment to the illusory perception comparable to the attentional condition, both in terms of perceptual performance (i.e., indicating whether the target contained the Kanizsa illusion) and metacognition (i.e., confidence in the response).

The results show a clear dissociation between the three levels of visual processing in terms of temporal dynamics. Local contrast was represented at an early stage (~80 ms), while collinearity and illusory perception were associated with later stages (~200-250 ms). Furthermore, the results provide clear evidence in support of a dissociation between the effects of perceptual and attentional processes on conscious access: while the former affected both neuronal correlates of collinearity and illusory perception, the latter did not have any effect on the processing of the more complex visual features involved in the illusion perception.

Weaknesses:

The design of the study and the results presented are very similar to those in Fahrenfort et al. (2017), reducing its novelty. Similar to the current study, Fahrenfort et al. (2017) tested the idea that if both masking and AB impact perceptual integration, they should affect the neural markers of perceptual integration in a similar way. They found that behavioural performance (hit/false alarm rate) was affected by both masking and AB, even though only the latter was significant in the unmasked condition. An early classification peak was instead only affected by masking. However, a late classification peak showed a pattern similar to the behavioural results, with classification affected by both masking and AB.

The interpretation of the results mainly centres on the theoretical framework of the recurrent processing theory of consciousness (Lamme, 2020), which lead to the assumption that local contrast, collinearity, and the illusory perception reflect feedforward, local recurrent, and global recurrent connections, respectively. It should be mentioned, however, that this theoretical prediction is not directly tested in the study. Moreover, the evidence for the dissociation between illusion and collinearity in terms of lateral and feedback connections seems at least limited. For instance, Kok et al. (2016) found that, whereas bottom-up stimulation activated all cortical layers, feedback activity induced by illusory figures led to a selective activation of the deep layers. Lee & Nguyen (2001), instead, found that V1 neurons respond to illusory contours of the Kanizsa figures, particularly in the superficial layers. They all mention feedback connections, but none seem to point to lateral connections.

Moreover, the evidence in favour of primarily lateral connections driving collinearity seems mixed as well. On one hand, Liang et al. (2017) showed that feedback and lateral connections closely interact to mediate image grouping and segmentation. On the other hand, Stettler et al. (2002) showed that, whereas the intrinsic connections link similarly oriented domains in V1, V2 to V1 feedback displays no such specificity. Furthermore, the other studies mentioned in the manuscript did not investigate feedback connections but only lateral ones, making it difficult to draw any clear conclusions.

Reviewer #1 (Public Review):

Summary:

It is suggested that for each limb the RG (rhythm generator) can operate in three different regimes: a non-oscillating state-machine regime, and in a flexordriven and a classical half-center oscillatory regime. This means that the field can move away from the old concept that there is only room for the classic half-center organization

Strengths:

A major benefit of the present paper is that a bridge was made between various CPG concepts ( "a potential contradiction between the classical half-center and flexor-driven concepts of spinal RG operation"). Another important step forward is the proposal about the neural control of slow gait ("at slow speeds ({less than or equal to} 0.35 m/s), the spinal network operates in a state regime and requires external inputs for phase transitions, which can come from limb sensory feedback and/or volitional inputs (e.g. from the motor cortex").

Weaknesses:

Some references are missing.

Reviewer #2 (Public Review):

Summary:

The biologically realistic model of the locomotor circuits developed by this group continues to define the state of the art for understanding spinal genesis of locomotion. Here the authors have achieved a new level of analysis of this model to generate surprising and potentially transformative new insights. They show that these circuits can operate in three very distinct states and that, in the intact cord, these states come into successive operation as the speed of locomotion increases. Equally important, they show that in spinal injury the model is "stuck" in the low speed "state machine" behavior.

Strengths:

There are many strengths for the simulation results presented here. The model itself has been closely tuned to match a huge range of experimental data and this has a high degree of plausibility. The novel insight presented here, with the three different states, constitutes a truly major advance in the understanding of neural genesis of locomotion in spinal circuits. The authors systematically consider how the states of the model relate to presently available data from animal studies. Equally important, they provide a number of intriguing and testable predictions. It is likely that these insights are the most important achieved in the past 10 years. It is highly likely proposed multi-state behavior will have a transformative effect on this field.

Weaknesses:

I have no major weaknesses. A moderate concern is that the authors should consider some basic sensitivity analyses to determine if the 3 state behavior is especially sensitive to any of the major circuit parameters - e.g. connection strengths in the oscillators or?

Reviewer #3 (Public Review):

Summary:

This work probes the control of walking in cats at different speeds and different states (split-belt and regular treadmill walking). Since the time of Sherrington there has been ongoing debate on this issue. The authors provide modeling data showing that they could reproduce data from cats walking on a specialized treadmill allowing for regular and split-belt walking. The data suggest that a non-oscillating state-machine regime best explains slow walking - where phase transitions are handled by external inputs into the spinal network. They then show at higher speeds a flexor-driven and then a classical half-center regime dominates. In spinal animals, it appears that a non-oscillating state-machine regime best explains the experimental data. The model is adapted from their previous work, and raises interesting questions regarding the operation of spinal networks, that, at low speeds, challenge assumptions regarding central pattern generator function. This is an interesting study. I have a few issues with the general validity of the treadmill data at low speeds, which I suspect can be clarified by the authors.

Strengths:

The study has several strengths. Firstly the detailed model has been well established by the authors and provides details that relate to experimental data such as commissural interneurons (V0c and V0d), along with V3 and V2a interneuron data. Sensory input along with descending drive is also modelled and moreover the model reproduces many experimental data findings. Moreover, the idea that sensory feedback is more crucial at lower speeds, also is confirmed by presynaptic inhibition increasing with descending drive. The inclusion of experimental data from split-belt treadmills, and the ability of the model to reproduce findings here is a definite plus.

Weaknesses:

Conceptually, this is a very useful study which provides interesting modeling data regarding the idea that the network can operate in different regimes, especially at lower speeds. The modelling data speaks for itself, but on the other hand, sensory feedback also provides generalized excitation of neurons which in turn project to the CPG. That is they are not considered part of the CPG proper. In these scenarios, it is possible that an appropriate excitatory drive could be provided to the network itself to move it beyond the state-machine state - into an oscillatory state. Did the authors consider that possibility? This is important since work using L-DOPA, for example, in cats or pharmacological activation of isolated spinal cord circuits, shows the CPG capable of producing locomotion without sensory or descending input.

Reviewer #1 (Public Review):

Summary:

This paper provides a resource for researchers studying the marine annelid Platynereis dumerilii. It is only the third whole-body connectome to be assembled and thus provides a comparison with those less complex animals: the nematode Caenorhabditis elegans and the tunicate Ciona intestinialis. The paper catalogs all cells in the body, not just neurons, and details how sensory neurons, interneurons, motor neurons, and effector organs are connected. From this, the authors are able to extract information about the organization of different aspects of the nervous system. These include the extent of recurrent connectivity, unimodal and multimodal sensory processing, and long-range and short-range connectivity.

Several interesting conclusions are drawn, including the concept that circuit evolution might have proceeded by duplication and diversion of cell types, much as it has been posited that gene evolution has occurred. It also informs the understanding of the evolution of segmental body plans in annelids by mapping and comparing cells in each segment.

Strengths:

This paper contains a wealth of data. The raw dataset is available. The codes and scripts are provided to allow interested readers to utilize this dataset.

The analysis is painstakingly meticulous. The diagrams are organized to orient the reader to the complexities of this overwhelming analysis

Weaknesses:

The strength of the paper is also its weakness. It contains so much data and analysis that it is burdensome to read and understand. There are 16 multi-panel data figures in the main text, and \another 38 supplemental figures, and 5 videos.

The impact of the paper is diminished by its size and depth. The paper could be broken up into smaller thematic papers that would be more accessible to researchers interested in particular topics. For example, there could be a single paper on the mushroom body and another paper on the segmental organization.

Reviewer #2 (Public Review):

Summary:

The stated ambition of the authors in this manuscript is to thoroughly analyze the complete neural connectome of the three-day larva of the marine annelid Platynereis. This manuscript follows several previous publications by the same group on the same volume of serial EM data, addressing several specialized functional circuits, and supersedes a previous preprint published in 2020. To this end, the authors have annotated the whole cell complement of the larva, including non-neural cells, with the collaborative tool CATMAID, traced the whole neurite extensions of neural cells, and annotated all synapses. The connectome has been algorithmically analyzed to extract the principal modules, adding several new, so far unexplored neural circuits to the list.

Strengths:

This remarkable study adds a third species to the list of animals in which the full connectome and functional modules have been analyzed, alongside C. elegans and Ciona intestinalis. It represents a leap in phylogeny, with Platynereis being a representative of the lophotrochozoans. Also, Platynereis has considerably more neurons than the latter species. The study provides a complete picture of the set of neural modules that are necessary for the survival of an autonomous marine larva with an active lifestyle.

The analysis is particularly impressive for revealing the complete innervation of the entire set of effector cells in the Platynereis larva, including muscle fibers, glands, pigment cells, ciliated cells, and helping understand the overall control of the organism's behavior through multiple sensory pathway integrations. It also reveals layers of neuronal intercalation in sensory-effector pathways that allow further integration even in a larva with limited behavioral complexity. The structure of the developing mushroom bodies, proposed ancestral bilaterian brain sensory integrative units, is detailed, as well as a complex mechanosensory module specific to a swimming larva.

A key new aspect of this connectome study is the thorough analysis of segmental cell types and intersegmental connectivity. Metameric organization is widespread in bilaterians and is nowhere clearer than in annelids. This metameric organization is even proposed by some authors to be an ancestral trait of bilaterians. Here, the authors show that homologous cell types and connectivity are shared not only by all segments of the animal but also by its non-segmental terminal parts (anterior prostomium and posterior pygidium). They suggest, in turn, that the entire body of the annelid may be formed of ancestral metameric units, an idea proposed before but here strongly supported by a list of homologous cell types. This is the most thorough evidence obtained so far for this provocative and stimulating evolutionary hypothesis.

Reviewer #1 (Public Review):

Summary:

The authors state the study's goal clearly: "The goal of our study was to understand to what extent animal individuality is influenced by situational changes in the environment, i.e., how much of an animal's individuality remains after one or more environmental features change." They use visually guided behavioral features to examine the extent of correlation over time and in a variety of contexts. They develop new behavioral instrumentation and software to measure behavior in Buridan's paradigm (and variations thereof), the Y-maze, and a flight simulator. Using these assays, they examine the correlations between conditions for a panel of locomotion parameters. They propose that inter-assay correlations will determine the persistence of locomotion individuality.

Strengths:

The OED defines individuality as "the sum of the attributes which distinguish a person or thing from others of the same kind," a definition mirrored by other dictionaries and the scientific literature on the topic. The concept of behavioral individuality can be characterized as:<br /> (1) a large set of behavioral attributes,<br /> (2) with inter-individual variability, that are<br /> (3) stable over time.

A previous study examined walking parameters in Buridan's paradigm, finding that several parameters were variable between individuals, and that these showed stability over separate days and up to 4 weeks (DOI: 10.1126/science.aaw718). The present study replicates some of those findings and extends the experiments from temporal stability to examining the correlation of locomotion features between different contexts.

The major strength of the study is using a range of different behavioral assays to examine the correlations of several different behavior parameters. It shows clearly that the inter-individual variability of some parameters is at least partially preserved between some contexts, and not preserved between others. The development of high-throughput behavior assays and sharing the information on how to make the assays is a commendable contribution.

Weaknesses:

The definition of individuality considers a comprehensive or large set of attributes, but the authors consider only a handful. In Supplemental Fig. S8, the authors show a large correlation matrix of many behavioral parameters, but these are illegible and are only mentioned briefly in Results. Why were five or so parameters selected from the full set? How were these selected? Do the correlation trends hold true across all parameters? For assays in which only a subset of parameters can be directly compared, were all of these included in the analysis, or only a subset?

The correlation analysis is used to establish stability between assays. For temporal re-testing, "stability" is certainly the appropriate word, but between contexts, it implies that there could be 'instability'. Rather, instead of the 'instability' of a single brain process, a different behavior in a different context could arise from engaging largely (or entirely?) distinct context-dependent internal processes, and have nothing to do with process stability per se. For inter-context similarities, perhaps a better word would be "consistency".

The parameters are considered one by one, not in aggregate. This focuses on the stability/consistency of the variability of a single parameter at a time, rather than holistic individuality. It would appear that an appropriate measure of individuality stability (or individuality consistency) that accounts for the high-dimensional nature of individuality would somehow summarize correlations across all parameters. Why was a multivariate approach (e.g. multiple regression/correlation) not used? Treating the data with a multivariate or averaged approach would allow the authors to directly address 'individuality stability', along with the analyses of single-parameter variability stability.

The correlation coefficients are sometimes quite low, though highly significant, and are deemed to indicate stability. For example, in Figure 4C top left, the % of time walked at 23{degree sign}C and 32{degree sign}C are correlated by 0.263, which corresponds to an R2 of 0.069 i.e. just 7% of the 32{degree sign}C variance is predictable by the 23{degree sign}C variance. Is it fair to say that a 7% determination indicates parameter stability? Another example: "Vector strength was the most correlated attention parameter... correlations ranged... to -0.197," which implies that 96% (1 - R2) of Y-maze variance is not predicted by Buridan variance. At what level does an r value not represent stability?

The authors describe a dissociation between inter-group differences and inter-individual variation stability, i.e. sometimes large mean differences between contexts, but significant correlation between individual test and retest data. Given that correlation is sensitive to slope, this might be expected to underestimate the variability stability (or consistency). Is there a way to adjust for the group differences before examining the correlation? For example, would it be possible to transform the values to in-group ranks prior to correlation analysis?

What is gained by classifying the five parameters into exploration, attention, and anxiety? To what extent have these classifications been validated, both in general and with regard to these specific parameters? Is the increased walking speed at higher temperatures necessarily due to an increased 'explorative' nature, or could it be attributed to increased metabolism, dehydration stress, or a heat-pain response? To what extent are these categories subjective?

The legends are quite brief and do not link to descriptions of specific experiments. For example, Figure 4a depicts a graphical overview of the procedure, but I could not find a detailed description of this experiment's protocol.

Using the current single-correlation analysis approach, the aims would benefit from re-wording to appropriately address single-parameter variability stability/consistency (as distinct from holistic individuality). Alternatively, the analysis could be adjusted to address the multivariate nature of individuality, so that the claims and the analysis are in concordance with each other.

The study presents a bounty of new technology to study visually guided behaviors. The GitHub link to the software was not available. To verify the successful transfer of open hardware and open-software, a report would demonstrate transfer by collaboration with one or more other laboratories, which the present manuscript does not appear to do. Nevertheless, making the technology available to readers is commendable.

The study discusses a number of interesting, stimulating ideas about inter-individual variability, and presents intriguing data that speaks to those ideas, albeit with the issues outlined above.

While the current work does not present any mechanistic analysis of inter-individual variability, the implementation of high-throughput assays sets up the field to more systematically investigate fly visual behaviors, their variability, and their underlying mechanisms.

Reviewer #2 (Public Review):

Summary:

The authors repeatedly measured the behavior of individual flies across several environmental situations in custom-made behavioral phenotyping rigs.

Strengths:

The study uses several different behavioral phenotyping devices to quantify individual behavior in a number of different situations and over time. It seems to be a very impressive amount of data. The authors also make all their behavioral phenotyping rig design and tracking software available, which I think is great and I'm sure other folks will be interested in using and adapting it to their own needs.

Weaknesses/Limitations:

I think an important limitation is that while the authors measured the flies under different environmental scenarios (i.e. with different lighting and temperature) they didn't really alter the "context" of the environment. At least within behavioral ecology, context would refer to the potential functionality of the expressed behaviors so for example, an anti-predator context, a mating context, or foraging. Here, the authors seem to really just be measuring aspects of locomotion under benign (relatively low-risk perception) contexts. This is not a flaw of the study, but rather a limitation to how strongly the authors can really say that this demonstrates that individuality is generalized across many different contexts. It's quite possible that rank order of locomotor (or other) behaviors may shift when the flies are in a mating or risky context.

The analytical framework in terms of statistical methods is lacking. It appears as though the authors used correlations across time/situations to estimate individual variation; however, far more sophisticated and elegant methods exist. The paper would be a lot stronger, and my guess is, much more streamlined if the authors employ hierarchical mixed models to analyse these data these models could capture and estimate differences in individual behavior across time and situations simultaneously. Along with this, it's currently unclear whether and how any statistical inference was performed. Right now, it appears as though any results describing how individuality changes across situations are largely descriptive (i.e. a visual comparison of the strengths of the correlation coefficients?).

Another pretty major weakness is that right now, I can't find any explicit mention of how many flies were used and whether they were re-used across situations. Some sort of overall schematic showing exactly how many measurements were made in which rigs and with which flies would be very beneficial.

I don't necessarily doubt the robustness of the results and my guess is that the author's interpretations would remain the same, but a more appropriate modeling framework could certainly improve their statistical inference and likely highlight some other cool patterns as these methods could better estimate stability and covariance in individual intercepts (and potentially slopes) across time and situation.

Reviewer #3 (Public Review):

This manuscript is a continuation of past work by the last author where they looked at stochasticity in developmental processes leading to inter-individual behavioural differences. In that work, the focus was on a specific behaviour under specific conditions while probing the neural basis of the variability. In this work, the authors set out to describe in detail how stable the individuality of animal behaviours is in the context of various external and internal influences. They identify a few behaviours to monitor (read outs of attention, exploration, and 'anxiety'); some external stimuli (temperature, contrast, nature of visual cues, and spatial environment); and two internal states (walking and flying).

They then use high-throughput behavioural arenas - most of which they have built and made plans available for others to replicate - to quantify and compare combinations of these behaviours, stimuli, and internal states. This detailed analysis reveals that:

(1) Many individualistic behaviours remain stable over the course of many days.<br /> (2) That some of these (walking speed) remain stable over changing visual cues. Others (walking speed and centrophobicity) remain stable at different temperatures.<br /> (3) All the behaviours they tested failed to remain stable over the spatially varying environment (arena shape).<br /> (4) Only angular velocity (a readout of attention) remains stable across varying internal states (walking and flying).

Thus, the authors conclude that there is a hierarchy in the influence of external stimuli and internal states on the stability of individual behaviours.

The manuscript is a technical feat with the authors having built many new high-throughput assays. The number of animals is large and many variables have been tested - different types of behavioural paradigms, flying vs walking, varying visual stimuli, and different temperatures among others.

Reviewer #1 (Public Review):

Summary:

This work aims to understand the role of thalamus POm in dorsal lateral striatum (DLS) projection in learning a sensorimotor associative task. The authors first confirm that POm forms "en passant" synapses with some of the DLS neuronal subtypes. They then perform a go/no-go associative task that consists of the mouse learning to discriminate between two different textures and to associate one of them with an action. During this task, they either record the activity of the POm to DLS axons using endoscopy or silence their activity. They report that POm axons in the DLS are activated around the sensory stimulus but that the activity is not modulated by the reward. Last, they showed that silencing the POm axons at the level of DLS slows down learning the task.

The authors show convincing evidence of projections from POm to DLS and that POm inputs to DLS code for whisking whatever the outcome of the task is. However, their results do not allow us to conclude if more neurons are recruited during the learning process or if the already activated fibres get activated more strongly. Last, because POm fibres in the DLS are also projecting to S1, silencing the POm fibres in the DLS could have affected inputs in S1 as well and therefore, the slowdown in acquiring the task is not necessarily specific to the POm to DLS pathway.

Strengths:

One of the main strengths of the paper is to go from slice electrophysiology to behaviour to get an in-depth characterization of one pathway. The authors did a comprehensive description of the POm projections to the DLS using transgenic mice to unambiguously identify the DLS neuronal population. They also used a carefully designed sensorimotor association task, and they exploited the results in depth.

It is a very nice effort to have measured the activity of the axons in the DLS not only after the mice have learned the task but throughout the learning process. It shows the progressive increase of activity of POm axons in the DLS, which could imply that there is a progressive strengthening of the pathway. The results show convincingly that POm axons in the DLS are not activated by the outcome of the task but by the whisker activity, and that this activity on average increases with learning.

Weaknesses:

One of the main targets of the striatum from thalamic input are the cholinergic neurons that weren't investigated here, is there information that could be provided?

It is interesting to know that the POm projects to all neuronal types in the DLS, but this information is not used further down the manuscript so the only take-home message of Figure 1 is that the axons that they image or silence in the DLS are indeed connected to DLS neurons and not just passing fibres. In this line, are these axons the same as the ones projecting to S1? If this is the case, why would we expect a different behaviour of the axon activity at the DLS level compared to S1?

The authors used endoscopy to measure the POm axons in the DLS activity, which makes it impossible to know if the progressive increase of POm response is due to an increase of activity from each individual neuron or if new neurons are progressively recruited in the process.

The picture presented in Figure 4 of the stimulation site is slightly concerning as there are hardly any fibres in neocortical layer 1 while there seems to be quite a lot of them in layer 4, suggesting that the animal here was injected in the VB. This is especially striking as the implantation and projection sites presented in Figures 1 and 2 are very clean and consistent with POm injection.

Reviewer #2 (Public Review):

Summary:

Yonk and colleagues show that the posterior medial thalamus (POm), which is interconnected with sensory and motor systems, projects directly to major categories of neurons in the striatum, including direct and indirect pathway MSNs, and PV interneurons. Activity in POm-striatal neurons during a sensory-based learning task indicates a relationship between reward expectation and arousal. Inhibition of these neurons slows reaction to stimuli and overall learning. This circuit is positioned to feed salient event activation to the striatum to set the stage for effective learning and action selection.

Strengths:

The results are well presented and offer interesting insight into an understudied thalamostriatal circuit. In general, this work is important as part of a general need for an increased understanding of thalamostriatal circuits in complex learning and action selection processes, which have generally received less attention than corticostriatal systems.

Weaknesses:

There could be a stronger connection between the connectivity part of the data - showing that POm neurons context D1, D2, and PV neurons in the striatum but with some different properties - and the functional side of the project. One wonders whether the POm neurons projecting to these subtypes or striatal neurons have unique signaling properties related to learning, or if there is a uniform, bulk signal sent to the striatum. This is not a weakness per se, as it's reasonable for these questions to be answered in future papers.

All the in vivo activity-related conclusions stem from data from just 5 mice, which is a relatively small sample set. Optogenetic groups are also on the small side.

Reviewer #3 (Public Review):

Yonk and colleagues investigate the role of the thalamostriatal pathway. Specifically, they studied the interaction of the posterior thalamic nucleus (PO) and the dorsolateral striatum in the mouse. First, they characterize connectivity by recording DLS neurons in in-vitro slices and optogenetically activating PO terminals. PO is observed to establish depressing synapses onto D1 and D2 spiny neurons as well as PV neurons. Second, the image PO axons are imaged by fiber photometry in mice trained to discriminate textures. Initially, no trial-locked activity is observed, but as the mice learn PO develops responses timed to the audio cue that marks the start of the trial and precedes touch. PO does appear to encode the tactile stimulus type or outcome. Optogenetic suppression of PO terminals in striatum slow task acquisition. The authors conclude that PO provides a "behaviorally relevant arousal-related signal" and that this signal "primes" striatal circuitry for sensory processing.

A great strength of this paper is its timeliness. Thalamostriatal processing has received almost no attention in the past, and the field has become very interested in the possible functions of PO. Additionally, the experiments exploit multiple cutting-edge techniques.

There seem to be some technical/analytical weaknesses. The in vitro experiments appear to have some contamination of nearby thalamic nuclei by the virus delivering the opsin, which could change the interpretation. Some of the statistical analyses of these data also appear inappropriate. The correlative analysis of Pom activity in vivo, licking, and pupil could be more convincingly done.

The bigger weakness is conceptual - why should striatal circuitry need "priming" by the thalamus in order to process sensory stimuli? Why would such circuitry even be necessary? Why is a sensory signal from the cortex insufficient? Why should the animal more slowly learn the task? How does this fit with existing ideas of striatal plasticity? It is unclear from the experiments that the thalamostriatal pathway exists for priming sensory processing. In fact, the optogenetic suppression of the thalamostriatal pathway seems to speak against that idea.

Reviewer #1 (Public Review):

Summary:

The novel advance by Wang et al is in the demonstration that, relative to a standard extinction procedure, the retrieval-extinction procedure more effectively suppresses responses to a conditioned threat stimulus when testing occurs just minutes after extinction. The authors provide some solid evidence to show that this "short-term" suppression of responding involves engagement of the dorsolateral prefrontal cortex.

Strengths:

Overall, the study is well-designed and the results are potentially interesting. There are, however, a few issues in the way that it is introduced and discussed. Some of the issues concern clarity of expression/communication. However, others relate to a theory that could be used to help the reader understand why the results should have come out the way that they did. More specific comments and questions are presented below.

Weaknesses:

INTRODUCTION & THEORY

(1) Can the authors please clarify why the first trial of extinction in a standard protocol does NOT produce the retrieval-extinction effect? Particularly as the results section states: "Importantly, such a short-term effect is also retrieval dependent, suggesting the labile state of memory is necessary for the short-term memory update to take effect (Fig. 1e)." The importance of this point comes through at several places in the paper:

1A. "In the current study, fear recovery was tested 30 minutes after extinction training, whereas the effect of memory reconsolidation was generally evident only several hours later and possibly with the help of sleep, leaving open the possibility of a different cognitive mechanism for the short-term fear dementia related to the retrieval-extinction procedure." ***What does this mean? The two groups in study 1 experienced a different interval between the first and second CS extinction trials; and the results varied with this interval: a longer interval (10 min) ultimately resulted in less reinstatement of fear than a shorter interval. Even if the different pattern of results in these two groups was shown/known to imply two different processes, there is absolutely no reason to reference any sort of cognitive mechanism or dementia - that is quite far removed from the details of the present study.

1B. "Importantly, such a short-term effect is also retrieval dependent, suggesting the labile state of memory is necessary for the short-term memory update to take effect (Fig. 1e)." ***As above, what is "the short-term memory update"? At this point in the text, it would be appropriate for the authors to discuss why the retrieval-extinction procedure produces less recovery than a standard extinction procedure as the two protocols only differ in the interval between the first and second extinction trials. References to a "short-term memory update" process do not help the reader to understand what is happening in the protocol.

(2) "Indeed, through a series of experiments, we identified a short-term fear amnesia effect following memory retrieval, in addition to the fear reconsolidation effect that appeared much later."<br /> ***The only reason for supposing two effects is because of the differences in responding to the CS2, which was subjected to STANDARD extinction, in the short- and long-term tests. More needs to be said about how and why the performance of CS2 is affected in the short-term test and recovers in the long-term test. That is, if the loss of performance to CS1 and CS2 is going to be attributed to some type of memory updating process across the retrieval-extinction procedure, one needs to explain the selective recovery of performance to CS2 when the extinction-to-testing interval extends to 24 hours. Instead of explaining this recovery, the authors note that performance to CS1 remains low when the extinction-to-testing interval is 24 hours and invoke something to do with memory reconsolidation as an explanation for their results: that is, they imply (I think) that reconsolidation of the CS1-US memory is disrupted across the 24-hour interval between extinction and testing even though CS1 evokes negligible responding just minutes after extinction.

(3) The discussion of memory suppression is potentially interesting but, in its present form, raises more questions than it answers. That is, memory suppression is invoked to explain a particular pattern of results but I, as the reader, have no sense of why a fear memory would be better suppressed shortly after the retrieval-extinction protocol compared to the standard extinction protocol; and why this suppression is NOT specific to the cue that had been subjected to the retrieval-extinction protocol.

3A. Relatedly, how does the retrieval-induced forgetting (which is referred to at various points throughout the paper) relate to the retrieval-extinction effect? The appeal to retrieval-induced forgetting as an apparent justification for aspects of the present study reinforces points 2 and 3 above. It is not uninteresting but needs some clarification/elaboration.

(4) Given the reports by Chalkia, van Oudenhove & Beckers (2020) and Chalkia et al (2020), some qualification needs to be inserted in relation to reference 6. That is, reference 6 is used to support the statement that "during the reconsolidation window, old fear memory can be updated via extinction training following fear memory retrieval". This needs a qualifying statement like "[but see Chalkia et al (2020a and 2020b) for failures to reproduce the results of 6]."

https://pubmed.ncbi.nlm.nih.gov/32580869/<br /> https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7115860/

CLARIFICATIONS, ELABORATIONS, EDITS

(5) The Abstract was not easy to follow:

5A. What does it mean to ask: "whether memory retrieval facilitates update mechanisms other than memory reconsolidation"? That is, in what sense could or would memory retrieval be thought to facilitate a memory update mechanism?

5B. "First, we demonstrate that memory reactivation prevents the return of fear shortly after extinction training in contrast to the memory reconsolidation effect which takes several hours to emerge and such a short-term amnesia effect is cue independent (Study 1, N = 57 adults)."<br /> ***The phrasing here could be improved for clarity: "First, we demonstrate that the retrieval-extinction protocol prevents the return of fear shortly after extinction training (i.e., when testing occurs just min after the end of extinction)." Also, cue-dependence of the retrieval-extinction effect was assessed in study 2.

5C. "Furthermore, memory reactivation also triggers fear memory reconsolidation and produces cue-specific amnesia at a longer and separable timescale (Study 2, N = 79 adults)." ***In study 2, the retrieval-extinction protocol produced a cue-specific disruption in responding when testing occurred 24 hours after the end of extinction. This result is interesting but cannot be easily inferred from the statement that begins "Furthermore..." That is, the results should be described in terms of the combined effects of retrieval and extinction, not in terms of memory reactivation alone; and the statement about memory reconsolidation is unnecessary. One can simply state that the retrieval-extinction protocol produced a cue-specific disruption in responding when testing occurred 24 hours after the end of extinction.

5D. "...we directly manipulated brain activities in the dorsolateral prefrontal cortex and found that both memory retrieval and intact prefrontal cortex functions were necessary for the short-term fear amnesia."<br /> ***This could be edited to better describe what was shown: E.g., "...we directly manipulated brain activities in the dorsolateral prefrontal cortex and found that intact prefrontal cortex functions were necessary for the short-term fear amnesia after the retrieval-extinction protocol."

5E. "The temporal scale and cue-specificity results of the short-term fear amnesia are clearly dissociable from the amnesia related to memory reconsolidation, and suggest that memory retrieval and extinction training trigger distinct underlying memory update mechanisms."<br /> ***The pattern of results when testing occurred just minutes after the retrieval-extinction protocol was different from that obtained when testing occurred 24 hours after the protocol. Describing this in terms of temporal scale is unnecessary, and suggesting that memory retrieval and extinction trigger different memory update mechanisms is not obviously warranted. The results of interest are due to the combined effects of retrieval+extinction and there is no sense in which different memory update mechanisms should be identified with retrieval (mechanism 1) and extinction (mechanism 2).

5F. "These findings raise the possibility of concerted memory modulation processes related to memory retrieval..."<br /> ***What does this mean?

(6) "...suggesting that the fear memory might be amenable to a more immediate effect, in addition to what the memory reconsolidation theory prescribes..."<br /> ***What does it mean to say that the fear memory might be amenable to a more immediate effect?

(7) "Parallel to the behavioral manifestation of long- and short-term memory deficits, concurrent neural evidence supporting memory reconsolidation theory emphasizes the long-term effect of memory retrieval by hypothesizing that synapse degradation and de novo protein synthesis are required for reconsolidation."<br /> ***This sentence needs to be edited for clarity.

(8) "previous behavioral manipulations engendering the short-term declarative memory effect..."<br /> ***What is the declarative memory effect? It should be defined.

(9) "The declarative amnesia effect emerges much earlier due to the online functional activity modulation..."<br /> ***Even if the declarative memory amnesia effect had been defined, the reference to online functional activity modulation is not clear.

(10) "However, it remains unclear whether memory retrieval might also precipitate a short-term amnesia effect for the fear memory, in addition to the long-term prevention orchestrated by memory consolidation."<br /> ***I found this sentence difficult to understand on my first pass through the paper. I think it is because of the phrasing of memory retrieval. That is, memory retrieval does NOT precipitate any type of short-term amnesia for the fear memory: it is the retrieval-extinction protocol that produces something like short-term amnesia. Perhaps this sentence should also be edited for clarity.

I will also note that the usage of "short-term" at this point in the paper is quite confusing: Does the retrieval-extinction protocol produce a short-term amnesia effect, which would be evidenced by some recovery of responding to the CS when tested after a sufficiently long delay? I don't believe that this is the intended meaning of "short-term" as used throughout the majority of the paper, right?

(11) "To fully comprehend the temporal dynamics of the memory retrieval effect..."<br /> ***What memory retrieval effect? This needs some elaboration.

(12) "We hypothesize that the labile state triggered by the memory retrieval may facilitate different memory update mechanisms following extinction training, and these mechanisms can be further disentangled through the lens of temporal dynamics and cue-specificities."<br /> ***What does this mean? The first part of the sentence is confusing around the usage of the term "facilitate"; and the second part of the sentence that references a "lens of temporal dynamics and cue-specificities" is mysterious. Indeed, as all rats received the same retrieval-extinction exposures in Study 2, it is not clear how or why any differences between the groups are attributed to "different memory update mechanisms following extinction".

(13) "In the first study, we aimed to test whether there is a short-term amnesia effect of fear memory retrieval following the fear retrieval-extinction paradigm."<br /> ***Again, the language is confusing. The phrase, "a short-term amnesia effect" implies that the amnesia itself is temporary; but I don't think that this implication is intended. The problem is specifically in the use of the phrase "a short-term amnesia effect of fear memory retrieval." To the extent that short-term amnesia is evident in the data, it is not due to retrieval per se but, rather, the retrieval-extinction protocol.

(14) The authors repeatedly describe the case where there was a 24-hour interval between extinction and testing as consistent with previous research on fear memory reconsolidation. Which research exactly? That is, in studies where a CS re-exposure was combined with a drug injection, responding to the CS was disrupted in a final test of retrieval from long-term memory which typically occurred 24 hours after the treatment. Is that what the authors are referring to as consistent? If so, which aspect of the results are consistent with those previous findings? Perhaps the authors mean to say that, in the case where there was a 24-hour interval between extinction and testing, the results obtained here are consistent with previous research that has used the retrieval-extinction protocol. This would clarify the intended meaning greatly.

DATA

(15) Points about data:

15A. The eight participants who were discontinued after Day 1 in study 1 were all from the no-reminder group. Can the authors please comment on how participants were allocated to the two groups in this experiment so that the reader can better understand why the distribution of non-responders was non-random (as it appears to be)?

15B. Similarly, in study 2, of the 37 participants that were discontinued after Day 2, 19 were from Group 30 min, and 5 were from Group 6 hours. Can the authors comment on how likely these numbers are to have been by chance alone? I presume that they reflect something about the way that participants were allocated to groups, but I could be wrong.

15C. "Post hoc t-tests showed that fear memories were resilient after regular extinction training, as demonstrated by the significant difference between fear recovery indexes of the CS+ and CS- for the no-reminder group (t26 = 7.441, P < 0.001; Fig. 1e), while subjects in the reminder group showed no difference of fear recovery between CS+ and CS- (t29 = 0.797, P = 0.432, Fig. 1e)."<br /> ***Is the fear recovery index shown in Figure 1E based on the results of the 1st test trial only? How can there have been a "significant difference between fear recovery indexes of the CS+ and CS- for the no-reminder group" when the difference in responding to the CS+ and CS- is used to calculate the fear recovery index shown in 1E? What are the t-tests comparing exactly, and what correction is used to account for the fact that they are applied post-hoc?

15D. "Finally, there is no statistical difference between the differential fear recovery indexes between CS+ in the reminder and no reminder groups (t55 = -2.022, P = 0.048; Fig. 1c, also see Supplemental Material for direct test for the test phase)."<br /> ***Is this statement correct - i.e., that there is no statistically significant difference in fear recovery to the CS+ in the reminder and no reminder groups? I'm sure that the authors would like to claim that there IS such a difference; but if such a difference is claimed, one would be concerned by the fact that it is coming through in an uncorrected t-test, which is the third one of its kind in this paragraph. What correction (for the Type 1 error rate) is used to account for the fact that the t-tests are applied post-hoc? And if no correction, why not?

15E. In study 2, why is responding to the CS- so high on the first test trial in Group 30 min? Is the change in responding to the CS- from the last extinction trial to the first test trial different across the three groups in this study? Inspection of the figure suggests that it is higher in Group 30 min relative to Groups 6 hours and 24 hours. If this is confirmed by the analysis, it has implications for the fear recovery index which is partly based on responses to the CS-. If not for differences in the CS- responses, Groups 30 minutes and 6 hours are otherwise identical.

15F. Was the 6-hour group tested at a different time of day compared to the 30-minute and 24-hour groups; and could this have influenced the SCRs in this group?

15G. Why is the range of scores in "thought control ability" different in the 30-minute group compared to the 6-hour and 24-hour groups? I am not just asking about the scale on the x-axis: I am asking why the actual distribution of the scores in thought control ability is wider for the 30-minute group?

(16) During testing in each experiment, how were the various stimuli presented? That is, was the presentation order for the CS+ and CS- pseudorandom according to some constraint, as it had been in extinction? This information should be added to the method section.

(17) "These results are consistent with previous research which suggested that people with better capability to resist intrusive thoughts also performed better in motivated dementia in both declarative and associative memories."<br /> ***Which parts of the present results are consistent with such prior results? It is not clear from the descriptions provided here why thought control ability should be related to the present findings or, indeed, past ones in other domains. This should be elaborated to make the connections clear.

Reviewer #2 (Public Review):

Summary

The study investigated whether memory retrieval followed soon by extinction training results in a short-term memory deficit when tested - with a reinstatement test that results in recovery from extinction - soon after extinction training. Experiment 1 documents this phenomenon using a between-subjects design. Experiment 2 used a within-subject control and saw that the effect was also observed in a control condition. In addition, it also revealed that if testing is conducted 6 hours after extinction, there is no effect of retrieval prior to extinction as there is recovery from extinction independently of retrieval prior to extinction. A third group also revealed that retrieval followed by extinction attenuates reinstatement when the test is conducted 24 hours later, consistent with previous literature. Finally, Experiment 3 used continuous theta-burst stimulation of the dorsolateral prefrontal cortex and assessed whether inhibition of that region (vs a control region) reversed the short-term effect revealed in Experiments 1 and 2. The results of the control groups in Experiment 3 replicated the previous findings (short-term effect), and the experimental group revealed that these can be reversed by inhibition of the dorsolateral prefrontal cortex.

Strengths

The work is performed using standard procedures (fear conditioning and continuous theta-burst stimulation) and there is some justification for the sample sizes. The results replicate previous findings - some of which have been difficult to replicate and this needs to be acknowledged - and suggest that the effect can also be observed in a short-term reinstatement test.

The study establishes links between memory reconsolidation and retrieval-induced forgetting (or memory suppression) literature. The explanations that have been developed for these are distinct and the current results integrate these, by revealing that the DLPFC activity involved in retrieval-extinction short-term effect. There is thus some novelty in the present results, but numerous questions remain unaddressed.

Weakness

The fear acquisition data is converted to a differential fear SCR and this is what is analysed (early vs late). However, the figure shows the raw SCR values for CS+ and CS- and therefore it is unclear whether the acquisition was successful (despite there being an "early" vs "late" effect - no descriptives are provided).

In Experiment 1 (Test results) it is unclear whether the main conclusion stems from a comparison of the test data relative to the last extinction trial ("we defined the fear recovery index as the SCR difference between the first test trial and the last extinction trial for a specific CS") or the difference relative to the CS- ("differential fear recovery index between CS+ and CS-"). It would help the reader assess the data if Figure 1e presents all the indexes (both CS+ and CS-). In addition, there is one sentence that I could not understand "there is no statistical difference between the differential fear recovery indexes between CS+ in the reminder and no reminder groups (P=0.048)". The p-value suggests that there is a difference, yet it is not clear what is being compared here. Critically, any index taken as a difference relative to the CS- can indicate recovery of fear to the CS+ or absence of discrimination relative to the CS-, so ideally the authors would want to directly compare responses to the CS+ in the reminder and no-reminder groups. The latter issue is particularly relevant in Experiment 2, in which the CS- seems to vary between groups during the test and this can obscure the interpretation of the result.

In Experiment 1, the findings suggest that there is a benefit of retrieval followed by extinction in a short-term reinstatement test. In Experiment 2, the same effect is observed on a cue that did not undergo retrieval before extinction (CS2+), a result that is interpreted as resulting from cue-independence, rather than a failure to replicate in a within-subjects design the observations of Experiment 1 (between-subjects). Although retrieval-induced forgetting is cue-independent (the effect on items that are suppressed [Rp-] can be observed with an independent probe), it is not clear that the current findings are similar. Here, both cues have been extinguished and therefore been equally exposed during the critical stage.

The findings in Experiment 2 suggest that the amnesia reported in Experiment 1 is transient, in that no effect is observed when the test is delayed by 6 hours. The phenomena whereby reactivated memories transition to extinguished memories as a function of the amount of exposure (or number of trials) is completely different from the phenomena observed here. In the former, the manipulation has to do with the number of trials (or the total amount of time) that the cues are exposed to. In the current study, the authors did not manipulate the number of trials but instead the retention interval between extinction and test. The finding reported here is closer to a "Kamin effect", that is the forgetting of learned information which is observed with intervals of intermediate length (Baum, 1968). Because the Kamin effect has been inferred to result from retrieval failure, it is unclear how this can be explained here. There needs to be much more clarity on the explanations to substantiate the conclusions.

There are many results (Ryan et al., 2015) that challenge the framework that the authors base their predictions on (consolidation and reconsolidation theory), therefore these need to be acknowledged. Similarly, there are reports that failed to observe the retrieval-extinction phenomenon (Chalkia et al., 2020), and the work presented here is written as if the phenomenon under consideration is robust and replicable. This needs to be acknowledged.

The parallels between the current findings and the memory suppression literature are speculated in the general discussion, and there is the conclusion that "the retrieval-extinction procedure might facilitate a spontaneous memory suppression process". Because one of the basic tenets of the memory suppression literature is that it reflects an "active suppression" process, there is no reason to believe that in the current paradigm, the same phenomenon is in place, but instead, it is "automatic". In other words, the conclusions make strong parallels with the memory suppression (and cognitive control) literature, yet the phenomena that they observed are thought to be passive (or spontaneous/automatic).<br /> Ultimately, it is unclear why 10 mins between the reminder and extinction learning will "automatically" suppress fear memories. Further down in the discussion, it is argued that "For example, in the well-known retrieval-induced forgetting (RIF) phenomenon, the recall of a stored memory can impair the retention of related long-term memory and this forgetting effect emerges as early as 20 minutes after the retrieval procedure, suggesting memory suppression or inhibition can occur in a more spontaneous and automatic manner". I did not follow with the time delay between manipulation and test (20 mins) would speak about whether the process is controlled or automatic.

Among the many conclusions, one is that the current study uncovers the "mechanism" underlying the short-term effects of retrieval extinction. There is little in the current report that uncovers the mechanism, even in the most psychological sense of the mechanism, so this needs to be clarified. The same applies to the use of "adaptive".

Whilst I could access the data on the OFS site, I could not make sense of the Matlab files as there is no signposting indicating what data is being shown in the files. Thus, as it stands, there is no way of independently replicating the analyses reported.

The supplemental material shows figures with all participants, but only some statistical analyses are provided, and sometimes these are different from those reported in the main manuscript. For example, the test data in Experiment 1 is analysed with a two-way ANOVA with the main effects of group (reminder vs no-reminder) and time (last trial of extinction vs first trial of the test) in the main report. The analyses with all participants in the sup mat used a mixed two-way ANOVA with a group (reminder vs no reminder) and CS (CS+ vs CS-). This makes it difficult to assess the robustness of the results when including all participants. In addition, in the supplementary materials, there are no figures and analyses for Experiment 3.

One of the overarching conclusions is that the "mechanisms" underlying reconsolidation (long term) and memory suppression (short term) phenomena are distinct, but memory suppression phenomena can also be observed after a 7-day retention interval (Storm et al., 2012), which then questions the conclusions achieved by the current study.

References:

Baum, M. (1968). Reversal learning of an avoidance response and the Kamin effect. Journal of Comparative and Physiological Psychology, 66(2), 495.<br /> Chalkia, A., Schroyens, N., Leng, L., Vanhasbroeck, N., Zenses, A. K., Van Oudenhove, L., & Beckers, T. (2020). No persistent attenuation of fear memories in humans: A registered replication of the reactivation-extinction effect. Cortex, 129, 496-509.<br /> Ryan, T. J., Roy, D. S., Pignatelli, M., Arons, A., & Tonegawa, S. (2015). Engram cells retain memory under retrograde amnesia. Science, 348(6238), 1007-1013.<br /> Storm, B. C., Bjork, E. L., & Bjork, R. A. (2012). On the durability of retrieval-induced forgetting. Journal of Cognitive Psychology, 24(5), 617-629.

Reviewer #3 (Public Review):

SUMMARY

Wang et al. have addressed how acquired fear and extinction memories evolve over time. Using a retrieval-extinction procedure in healthy humans, they have investigated the recovery of fear memories 30-60 minutes., 6 hours, and 24 hours after the retrieval-extinction phase. They have addressed this research question through 3 different experiments which included manipulations of the reminder cue, the time interval, and brain activity. Together, the studies suggest that early on after retrieval-extinction (30-60 min. later), retrieval-extinction may lead to an attenuation of fear recovery (after reinstatement) for all fear cues, as well as the non-reminded ones. Study 3 moreover suggests that this effect may depend on normal dlPFC function. In addition, the paper also contains data in line with prior findings suggesting that a 6-hour interval does not benefit from the reminder cue, and that a 24-hour interval does, and specifically for the reminded fear cue. The latter findings are seen as evidence of fear memory reconsolidation.

STRENGTHS

(1) The paper combines three related human fear conditioning studies, each with decent sample sizes. The authors are transparent about the fact that they excluded many participants and about which conditions they belonged to.

(2) The effect that this paper investigates (short-term fear memory after a retrieval-extinction procedure) has not been studied extensively, thus making it a relevant topic.

(3) The application of brain stimulation as a means to study causal relationships is interesting and goes beyond the purely behavioral or pharmacological interventions that are often used in human fear conditioning research. Also, the use of an active control stimulation is a strength of the study.

WEAKNESSES

(1) The entire study hinges on the idea that there is memory 'suppression' if (1) the CS+ was reminded before extinction and (2) the reinstatement and memory test takes place 30 minutes later (in Studies 1 & 2). However, the evidence supporting this suppression idea is not very strong. In brief, in Study 1, the effect seems to only just reach significance, with a medium effect size at best, and, moreover, it is unclear if this is the correct analysis (which is a bit doubtful, when looking at Figure 1D and E). In Study 2, there was no optimal control condition without reminder and with the same 30-min interval (which is problematic, because we can assume generalization between CS1+ and CS2+, as pointed out by the authors, and because generalization effects are known to be time-dependent). Study 3 is more convincing, but entails additional changes in comparison with Studies 1 and 2, i.e., applications of cTBS and an interval of 1 hour instead of 30 minutes (the reason for this change was not explained). So, although the findings of the 3 studies do not contradict each other and are coherent, they do not all provide strong evidence for the effect of interest on their own.

Related to the comment above, I encourage the authors to double-check if this statement is correct: "Also, our results remain robust even with the "non-learners" included in the analysis (Fig. S1 in the Supplemental Material)". The critical analysis for Study 1 is a between-group comparison of the CS+ and CS- during the last extinction trial versus the first test trial. This result only just reached significance with the selected sample (p = .048), and Figures 1D and E even seem to suggest otherwise. I doubt that the analysis would reach significance when including the "non-learners" - assuming that this is what is shown in Supplemental Figure 1 (which shows the data from "all responded participants").

Also related to the comment above, I think that the statement "suggesting a cue-independent short-term amnesia effect" in Study 2 is not correct and should read: "suggesting extinction of fear to the CS1+ and CS2+", given that the response to the CS+'s is similar to the response to the CS-, as was the case at the end of extinction. Also the next statement "This result indicates that the short-term amnesia effect observed in Study 2 is not reminder-cue specific and can generalize to the non-reminded cues" is not fully supported by the data, given the lack of an appropriate control group in this study (a group without reinstatement). The comparison with the effect found in Study 1 is difficult because the effect found there was relatively small (and may have to be double-checked, see remarks above), and it was obtained with a different procedure using a single CS+. The comparison with the 6-h and 24-h groups of Study 2 is not helpful as a control condition for this specific question (i.e., is there reinstatement of fear for any of the CS+'s) because of the large procedural difference with regard to the intervals between extinction and reinstatement (test).

(2) It is unclear which analysis is presented in Figure 3. According to the main text, it either shows the "differential fear recovery index between CS+ and CS-" or "the fear recovery index of both CS1+ and CS2+". The authors should clarify what they are analyzing and showing, and clarify to which analyses the ** and NS refer in the graphs. I would also prefer the X-axes and particularly the Y-axes of Fig. 3a-b-c to be the same. The image is a bit misleading now. The same remarks apply to Figure 5.

(3) In general, I think the paper would benefit from being more careful and nuanced in how the literature and findings are represented. First of all, the authors may be more careful when using the term 'reconsolidation'. In the current version, it is put forward as an established and clearly delineated concept, but that is not the case. It would be useful if the authors could change the text in order to make it clear that the reconsolidation framework is a theory, rather than something that is set in stone (see e.g., Elsey et al., 2018 (https://doi.org/10.1037/bul0000152), Schroyens et al., 2022 (https://doi.org/10.3758/s13423-022-02173-2)).

In addition, the authors may want to reconsider if they want to cite Schiller et al., 2010 (https://doi.org/10.1038/nature08637), given that the main findings of this paper, nor the analyses could be replicated (see, Chalkia et al., 2020 (https://doi.org/10.1016/j.cortex.2020.04.017; https://doi.org/10.1016/j.cortex.2020.03.031).

Relatedly, it should be clarified that Figure 6 is largely speculative, rather than a proven model as it is currently presented. This is true for all panels, but particularly for panel c, given that the current study does not provide any evidence regarding the proposed reconsolidation mechanism.

Lastly, throughout the paper, the authors equate skin conductance responses (SCR) with fear memory. It should at least be acknowledged that SCR is just one aspect of a fear response, and that it is unclear whether any of this would translate to verbal or behavioral effects. Such effects would be particularly important for any clinical application, which the authors put forward as the ultimate goal of the research.

(4) The Discussion quite narrowly focuses on a specific 'mechanism' that the authors have in mind. Although it is good that the Discussion is to the point, it may be worthwhile to entertain other options or (partial) explanations for the findings. For example, have the authors considered that there may be an important role for attention? When testing very soon after the extinction procedure (and thus after the reminder), attentional processes may play an important role (more so than with longer intervals). The retrieval procedure could perhaps induce heightened attention to the reminded CS+ (which could be further enhanced by dlPFC stimulation)?

(5) There is room for improvement in terms of language, clarity of the writing, and (presentation of the) statistical analyses, for all of which I have provided detailed feedback in the 'Recommendations for the authors' section. Idem for the data availability; they are currently not publicly available, in contrast with what is stated in the paper. In addition, it would be helpful if the authors would provide additional explanation or justification for some of the methodological choices (e.g., the 18-s interval and why stimulate 8 minutes after the reminder cue, the choice of stimulation parameters), and comment on reasons for (and implications of) the large amount of excluded participants (>25%).

Finally, I think several statements made in the paper are overly strong in light of the existing literature (or the evidence obtained here) or imply causal relationships that were not directly tested.

Reviewer #1 (Public Review):

Summary:

This is a large cohort of ischemic stroke patients from a single centre. The author successfully set up predictive models for PTS.

Strengths:

The design and implementation of the trial are acceptable, and the results are credible. It may provide evidence of seizure prevention in the field of stroke treatment.

Weaknesses:

The methodology needs further consideration. The Discussion needs extensive rewriting.

Reviewer #2 (Public Review):

Summary

The authors present multiple machine-learning methodologies to predict post-stroke epilepsy (PSE) from admission clinical data.

Strengths

The Statistical Approach section is very well written. The approaches used in this section are very sensible for the data in question.

Weaknesses

There are many typos and unclear statements throughout the paper.

There are some issues with SHAP interpretation. SHAP in its default form, does not provide robust statistical guarantees of effect size. There is a claim that "SHAP analysis showed that white blood cell count had the greatest impact among the routine blood test parameters". This is a difficult claim to make.

The Data Collection section is very poorly written, and the methodology is not clear.

There is no information about hyperparameter selection for models or whether a hyperparameter search was performed. Given this, it is difficult to conclude whether one machine learning model performs better than others on this task.

The inclusion and exclusion criteria are unclear - how many patients were excluded and for what reasons?

There is no sensitivity analysis of the SMOTE methodology: How many synthetic data points were created, and how does the number of synthetic data points affect classification accuracy?

Did the authors achieve their aims? Do the results support their conclusions?

The paper does not clarify the features' temporal origins. If some features were not recorded on admission to the hospital but were recorded after PSE occurred, there would be temporal leakage.

The authors claim that their models can predict PSE. To believe this claim, seeing more information on out-of-distribution generalisation performance would be helpful. There is limited reporting on the external validation cohort relative to the reporting on train and test data.

For greater certainty on all reported results, it would be most appropriate to perform n-fold cross-validation, and report mean scores and confidence intervals across the cross-validation splits

The likely impact of the work on the field

If this model works as claimed, it will be useful for predicting PSE. This has some direct clinical utility.

Analysis of features contributing to PSE may provide clinical researchers with ideas for further research on the underlying aetiology of PSE.

Additional context that might help readers

The authors show force plots and decision plots from SHAP values. These plots are non-trivial to interpret, and the authors should include an explanation of how to interpret them.

Reviewer #3 (Public Review):

Summary:

The authors report the performance of a series of machine learning models inferred from a large-scale dataset and externally validated with an independent cohort of patients, to predict the risk of post-stroke epilepsy. Some of the reported models have very good explicative and predictive performances.

Strengths:

The models have been derived from real-world large-scale data.

Performances of the best-performing models are very good according to the external validation results.

Early prediction of the risk of post-stroke epilepsy would be of high interest to implement early therapeutic interventions that could improve prognosis.

Weaknesses:

There are issues with the readability of the paper. Many abbreviations are not introduced properly and sometimes are written inconsistently. A lot of relevant references are omitted. The methodological descriptions are extremely brief and, sometimes, incomplete.

The dataset is not disclosed, and neither is the code (although the code is made available upon request). For the sake of reproducibility, unless any bioethical concerns impede it, it would be good to have these data disclosed.

Although the external validation is appreciated, cross-validation to check the robustness of the models would also be welcome.

Reviewer #1 (Public Review):

This study uses MEG to test for a neural signature of the trial history effect known as 'serial dependence.' This is a behavioral phenomenon whereby stimuli are judged to be more similar than they really are, in feature space, to stimuli that were relevant in the recent past (i.e., the preceding trials). This attractive bias is prevalent across stimulus classes and modalities, but a neural source has been elusive. This topic has generated great interest in recent years, and I believe this study makes a unique contribution to the field. The paper is overall clear and compelling, and makes effective use of data visualizations to illustrate the findings. Below, I list several points where I believe further detail would be important to interpreting the results. I also make suggestions for additional analyses that I believe would enrich understanding but are inessential to the main conclusions.

(1) In the introduction, I think the study motivation could be strengthened, to clarify the importance of identifying a neural signature here. It is clear that previous studies have focused mainly on behavior, and that the handful of neuroscience investigations have found only indirect signatures. But what would the type of signature being sought here tell us? How would it advance understanding of the underlying processes, the function of serial dependence, or the theoretical debates around the phenomenon?

(1a) As one specific point of clarification, on p. 5, lines 91-92, a previous study (St. John-Saaltink et al.) is described as part of the current study motivation, stating that "as the current and previous orientations were either identical or orthogonal to each other, it remained unclear whether this neural bias reflected an attraction or repulsion in relation to the past." I think this statement could be more explicit as to why/how these previous findings are ambiguous. The St. John-Saaltink study stands as one of very few that may be considered to show evidence of an early attractive effect in neural activity, so it would help to clarify what sort of advance the current study represents beyond that.

(1b) The study motivation might also consider the findings of Ranieri et al (2022, J. Neurosci) Fornaciai, Togoli, & Bueti (2023, J. Neurosci), and Luo & Collins (2023, J. Neurosci) who all test various neural signatures of serial dependence.

(2) Regarding the methods and results, it would help if the initial description of the reconstruction approach, in the main text, gave more context about what data is going into reconstruction (e.g., which sensors), a more conceptual overview of what the 'reconstruction' entails, and what the fidelity metric indexes. To me, all of that is important to interpreting the figures and results. For instance, when I first read, it was unclear to me what it meant to "reconstruct the direction of S1 during the S2 epoch" (p. 10, line 199)? As in, I couldn't tell how the data/model knows which item it is reconstructing, as opposed to just reporting whatever directional information is present in the signal.

(2a) Relatedly, what does "reconstruction strength" reflect in Figure 2a? Is this different than the fidelity metric? Does fidelity reflect the strength of the particular relevant direction, or does it just mean that there is a high level of any direction information in the signal?

(3) Then in the Methods, it would help to provide further detail still about the IEM training/testing procedure. For instance, it's not entirely clear to me whether all the analyses use the same model (i.e., all trained on stimulus encoding) or whether each epoch and timepoint is trained on the corresponding epoch and timepoint from the other session. This speaks to whether the reconstructions reflect a shared stimulus code across different conditions vs. that stimulus information about various previous and current trial items can be extracted if the model is tailored accordingly. Specifically, when you say "aim of the reconstruction" (p. 31, line 699), does that simply mean the reconstruction was centered in that direction (that the same data would go into reconstructing S1 or S2 in a given epoch, and what would differentiate between them is whether the reconstruction was centered to the S1 or S2 direction value)? Or were S1 and S2 trained and tested separately for the same epoch? And was training and testing all within the same time point (i.e., train on delay, test on delay), or train on the encoding of a given item, then test the fidelity of that stimulus code under various conditions?

(3a) I think training and testing were done separately for each epoch and timepoint, but this could have important implications for interpreting the results. Namely if the models are trained and tested on different time points, and reference directions, then some will be inherently noisier than others (e.g., delay period more so than encoding), and potentially more (or differently) susceptible to bias. For instance, the S1 and S2 epochs show no attractive bias, but they may also be based on more high-fidelity training sets (i.e., encoding), and therefore less susceptible to the bias that is evident in the retrocue epoch.

(4) I believe the work would benefit from a further effort to reconcile these results with previous findings (i.e., those that showed repulsion, like Sheehan & Serences), potentially through additional analyses. The discussion attributes the difference in findings to the "combination of a retro-cue paradigm with the high temporal resolution of MEG," but it's unclear how that explains why various others observed repulsion (thought to happen quite early) that is not seen at any stage here. In my view, the temporal (as well as spatial) resolution of MEG could be further exploited here to better capture the early vs. late stages of processing. For instance, by separately examining earlier vs. later time points (instead of averaging across all of them), or by identifying and analyzing data in the sensors that might capture early vs. late stages of processing. Indeed, the S1 and S2 reconstructions show subtle repulsion, which might be magnified at earlier time points but then shift (toward attraction) at later time points, thereby counteracting any effect. Likewise, the S1 reconstruction becomes biased during the S2 epoch, consistent with previous observations that the SD effects grow across a WM delay. Maybe both S1 and S2 would show an attractive bias emerging during the later (delay) portion of their corresponding epoch? As is, the data nicely show that an attractive bias can be detected in the retrocue period activity, but they could still yield further specificity about when and where that bias emerges.

(5) A few other potentially interesting (but inessential considerations): A benchmark property of serial dependence is its feature-specificity, in that the attractive bias occurs only between current and previous stimuli that are within a certain range of similarity to each other in feature space. I would be very curious to see if the neural reconstructions manifest this principle - for instance, if one were to plot the trialwise reconstruction deviation from 0, across the full space of current-previous trial distances, as in the behavioral data. Likewise, something that is not captured by the DoG fitting approach, but which this dataset may be in a position to inform, is the commonly observed (but little understood) repulsive effect that appears when current and previous stimuli are quite distinct from each other. As in, Figure 1b shows an attractive bias for direction differences around 30 degrees, but a repulsive one for differences around 170 degrees - is there a corresponding neural signature for this component of the behavior?

Reviewer #2 (Public Review):

Summary:

The study aims to probe the neural correlates of visual serial dependence - the phenomenon that estimates of a visual feature (here motion direction) are attracted towards the recent history of encoded and reported stimuli. The authors utilize an established retro-cue working memory task together with magnetoencephalography, which allows to probe neural representations of motion direction during encoding and retrieval (retro-cue) periods of each trial. The main finding is that neural representations of motion direction are not systematically biased during the encoding of motion stimuli, but are attracted towards the motion direction of the previous trial's target during the retrieval (retro-cue period), just prior to the behavioral response. By demonstrating a neural signature of attractive biases in working memory representations, which align with attractive behavioral biases, this study highlights the importance of post-encoding memory processes in visual serial dependence.

Strengths:

The main strength of the study is its elegant use of a retro-cue working memory task together with high temporal resolution MEG, enabling to probe neural representations related to stimulus encoding and working memory. The behavioral task elicits robust behavioral serial dependence and replicates previous behavioral findings by the same research group. The careful neural decoding analysis benefits from a large number of trials per participant, considering the slow-paced nature of the working memory paradigm. This is crucial in a paradigm with considerable trial-by-trial behavioral variability (serial dependence biases are typically small, relative to the overall variability in response errors). While the current study is broadly consistent with previous studies showing that attractive biases in neural responses are absent during stimulus encoding (previous studies reported repulsive biases), to my knowledge it is the first study showing attractive biases in current stimulus representations during working memory. The study also connects to previous literature showing reactivations of previous stimulus representations, although the link between reactivations and biases remains somewhat vague in the current manuscript. Together, the study reveals an interesting avenue for future studies investigating the neural basis of visual serial dependence.

Weaknesses:

The main weakness of the current manuscript is that the authors could have done more analyses to address the concern that their neural decoding results are driven by signals related to eye movements. The authors show that participants' gaze position systematically depended on the current stimuli's motion directions, which together with previous studies on eye movement-related confounds in neural decoding justifies such a concern. The authors seek to rule out this confound by showing that the consistency of stimulus-dependent gaze position does not correlate with (a) the neural reconstruction fidelity and (b) the repulsive shift in reconstructed motion direction. However, both of these controls do not directly address the concern. If I understand correctly the metric quantifying the consistency of stimulus-dependent gaze position (Figure S3a) only considers gaze angle and not gaze amplitude. Furthermore, it does not consider gaze position as a function of continuous motion direction, but instead treats motion directions as categorical variables. Therefore, assuming an eye movement confound, it is unclear whether the gaze consistency metric should strongly correlate with neural reconstruction fidelity, or whether there are other features of eye movements (e.g., amplitude differences across participants, and tuning of gaze in the continuous space of motion directions) which would impact the relationship with neural decoding. Moreover, it is unclear whether the consistency metric, which does not consider history dependencies in eye movements, should correlate with attractive history biases in neural decoding. It would be more straightforward if the authors would attempt to (a) directly decode stimulus motion direction from x-y gaze coordinates and relate this decoding performance to neural reconstruction fidelity, and (b) investigate whether gaze coordinates themselves are history-dependent and are attracted to the average gaze position associated with the previous trials' target stimulus. If the authors could show that (b) is not the case, I would be much more convinced that their main finding is not driven by eye movement confounds.

I am not convinced by the across-participant correlation between attractive biases in neural representations and attractive behavioral biases in estimation reports. One would expect a correlation with the behavioral bias amplitude, which is not borne out. Instead, there is a correlation with behavioral bias width, but no explanation of how bias width should relate to the bias in neural representations. The authors could be more explicit in their arguments about how these metrics would be functionally related, and why there is no correlation with behavioral bias amplitude.

The sample size (n = 10) is definitely at the lower end of sample sizes in this field. The authors collected two sessions per participant, which partly alleviates the concern. However, given that serial dependencies can be very variable across participants, I believe that future studies should aim for larger sample sizes.

It would have been great to see an analysis in source space. As the authors mention in their introduction, different brain areas, such as PPC, mPFC, and dlPFC have been implicated in serial biases. This begs the question of which brain areas contribute to the serial dependencies observed in the current study. For instance, it would be interesting to see whether attractive shifts in current representations and pre-stimulus reactivations of previous stimuli are evident in the same or different brain areas.

Reviewer #3 (Public Review):

Summary:

This study identifies the neural source of serial dependence in visual working memory, i.e., the phenomenon that recall from visual working memory is biased towards recently remembered but currently irrelevant stimuli. Whether this bias has a perceptual or post-perceptual origin has been debated for years - the distinction is important because of its implications for the neural mechanism and ecological purpose of serial dependence. However, this is the first study to provide solid evidence based on human neuroimaging that identifies a post-perceptual memory maintenance stage as the source of the bias. The authors used multivariate pattern analysis of magnetoencephalography (MEG) data while observers remembered the direction of two moving dot stimuli. After one of the two stimuli was cued for recall, decoding of the cued motion direction re-emerged, but with a bias towards the motion direction cued on the previous trial. By contrast, decoding of the stimuli during the perceptual stage was not biased.

Strengths:

The strengths of the paper are its design, which uses a retrospective cue to clearly distinguish the perceptual/encoding stage from the post-perceptual/maintenance stage, and the rigour of the careful and well-powered analysis. The study benefits from high within-participant power through the use of sensitive MEG recordings (compared to the more common EEG), and the decoding and neural bias analysis are done with care and sophistication, with appropriate controls to rule out confounds.

Weaknesses:

A minor weakness of the study is the remaining (but slight) possibility of an eye movement confound. A control analysis shows that participants make systematic eye movements that are aligned with the remembered motion direction during both the encoding and maintenance phases of the task. The authors go some way to show that this eye gaze bias seems unrelated to the decoding of MEG data, but in my opinion do not rule it out conclusively. They merely show that the strengths of the gaze bias and the strength of MEG-based decoding/neural bias are uncorrelated across the 10 participants. Therefore, this argument seems to rest on a null result from an underpowered analysis.

Impact:

This important study contributes to the debate on serial dependence with solid evidence that biased neural representations emerge only at a relatively late post-perceptual stage, in contrast to previous behavioural studies. This finding is of broad relevance to the study of working memory, perception, and decision-making by providing key experimental evidence favouring one class of computational models of how stimulus history affects the processing of the current environment.

Reviewer #1 (Public Review):

Summary:

In this study, the authors developed an organoid system that contains smooth muscle cells (SMCs) and interstitial cells of Cajal (ICCs; pacemaker) but few enteric neurons, and generates rhythmic contractions as seen in the developing gut. The stereotypical arrangements of SMCs and ICCs in the organoid allowed the authors to identify these cell types in the organoid without antibody staining. The authors took advantage of this and used calcium imaging and pharmacology to study how calcium transients develop in this system through the interaction between the two types of cells. The authors first show that calcium transients are synchronized between ICC-ICC, SMC-SMC, and SMC-ICC. They then used gap junction inhibitors to suggest that gap junctions are specifically involved in ICC-to-SMC signaling. Finally, the authors used an inhibitor of myosin II to suggest that feedback from SMC contraction is crucial for the generation of rhythmic activities in ICCs. The authors also show that two organoids become synchronized as they fuse and SMCs mediate this synchronization.

Strengths:

The organoid system offers a useful model in which one can study the specific roles of SMCs and ICCs in live samples.

Weaknesses:

Since only one blocker each for gap junction and myosin II was used, the specificities of the effects were unclear.

Reviewer #2 (Public Review):

Summary:

In this study, Yagasaki et al. describe an organoid system to study the interactions between smooth muscle cells (SMCs) and interstitial cells of Cajal (ICCs). While these interactions are essential for the control of rhythmic intestinal contractility (i.e., peristalsis), they are poorly understood, largely due to the complexity of and access to the in vivo environment and the inability to co-culture these cell types in vitro for long term under physiological conditions. The "gut contractile organoids" organoids described herein are reconstituted from stromal cells of the fetal chicken hindgut that rapidly reorganize into multilayered spheroids containing an outer layer of smooth muscle cells and an inner core of interstitial cells. The authors demonstrate that they contract cyclically and additionally use calcium imagining to show that these contractions occur concomitantly with calcium transients that initiate in the interstitial cell core and are synchronized within the organoid and between ICCs and SMCs. Furthermore, they use several pharmacological inhibitors to show that these contractions are dependent upon non-muscle myosin activity and, surprisingly, independent of gap junction activity. Finally, they develop a 3D hydrogel for the culturing of multiple organoids and found that they synchronize their contractile activities through interconnecting smooth muscle cells, suggesting that this model can be used to study the emergence of pacemaking activities. Overall, this study provides a relatively easy-to-establish organoid system that will be of use in studies examining the emergence of rhythmic peristaltic smooth muscle contractions and how these are regulated by interstitial cell interactions. However, further validation and quantification will be necessary to conclusively determine show the cellular composition of the organoids and how reproducible their behaviors are.

Strengths:

This work establishes a new self-organizing organoid system that can easily be generated from the muscle layers of the chick fetal hindgut to study the emergence of spontaneous smooth muscle cell contractility. A key strength of this approach is that the organoids seem to contain few cell types (though more validation is needed), namely smooth muscle cells (SMCs) and interstitial cells of Cajal (ICCs). These organoids are amenable to live imaging of calcium dynamics as well as pharmacological perturbations for functional assays, and since they are derived from developing tissues, the emergence of the interactions between cell types can be functionally studied. Thus, the gut contractile organoids represent a reductionist system to study the interactions between SMCs and ICCs in comparison to the more complex in vivo environment, which has made studying these interactions challenging.

Weaknesses:

The study falls short in the sense that it does not provide a rigorous amount of evidence to validate that the gut organoids are made of bona fide smooth muscle cells and ICCs. For example, only two "marker" proteins are used to support the claims of cell identity of SMCs and ICCs. At the same time, certain aspects of the data are not quantified sufficiently to appreciate the variance of organoid rhythmic contractility. For example, most contractility plots show the trace for a single organoid. This leads to a concern for how reproducible certain aspects of the organoid system (e.g. wavelength between contractions/rhythm) might be, or how these evolve uniquely over time in culture. Furthermore, while this study might be able to capture the emergence of ICC-SMC interactions as they related to muscle contraction and pacemaking, it is unclear how these interactions relate to adult gastrointestinal physiology given that the organoids are derived from fetal cells that might not be fully differentiated or might have distinct functions from the adult. Finally, despite the strength of this system, discoveries made in it will need to be validated in vivo.

Reviewer #3 (Public Review):

Summary:

The paper presents a novel contractile gut organoid system that allows for in vitro studying of rudimentary peristaltic motions in embryonic tissues by facilitating GCaMP-live imaging of Ca2+<br /> dynamics, while highlighting the importance and sufficiency of ICC and SMC interactions in generating consistent contractions reminiscent of peristalsis. It also argues that ENS at later embryonic stages might not be necessary for coordination of peristalsis.

Strengths:

The manuscript by Yagasaki, Takahashi, and colleagues represents an exciting new addition to the toolkit available for studying fundamental questions in the development and physiology of the hindgut. The authors carefully lay out the protocol for generating contractile gut organoids from chick embryonic hindgut, and perform a series of experiments that illustrate the broader utility of these organoids for studying the gut. This reviewer is highly supportive of the manuscript, with only minor requests to improve confidence in the findings and broader impact of the work. These are detailed below.

Weaknesses:

(1) Given that the literature is conflicting on the role GAP junctions in potentiating communication between intestinal cells of Cajal (ICCs) and smooth muscle cells (SMCs), the experiments involving CBX and 18Beta-GA are well-justified. However, because neither treatment altered contractile frequency or synchronization of Ca++ transients, it would be important to demonstrate that the treatments did indeed inhibit GAP junction function as administered. This would strengthen the conclusion that GAP junctions are not required, and eliminate the alternative explanation that the treatments themselves failed to block GAP junction activity.

(2) Given that 5uM blebbistatin increases the frequency of contractions but 10uM completely abolishes contractions, confirming that cell viability is not compromised at the higher concentration would build confidence that the phenotype results from inhibition of myosin activity. One could either assay for cell death, or perform washout experiments to test for recovery of cyclic contractions upon removal of blebbistatin. The latter may provide access to other interesting questions as well. For example, do organoids retain memory of their prior setpoint or arrive at a new firing frequency after washout?

(3) Regulation of contractile activity was attributed to ICCs, with authors reasoning that Tuj1+ enteric neurons were only present in organoids in very small numbers (~1%). However, neuronal function is not strictly dependent on abundance, and some experimental support for the relative importance of ICCs over Tuj1+ cells would strengthen a central assumption of the work that ICCs the predominant cell type regulating organoid contraction. For example, one could envision forming organoids from embryos in which neural crest cells have been ablated via microdissection or targeted electroporation. Another approach would be ablation of Tuj1+ cells from the formed organoids via tetrodotoxin treatment. The ability of organoids to maintain rhythmic contractile activity in the total absence of Tuj1+ cells would add confidence that the ICCs are indeed the driver of contractility in these organoids.

(4) Given the implications of a time lag between Ca++ peaks in ICCs and SMCs, it would be important to quantify this, including standard deviations, rather than showing representative plots from a single sample.

(5) To validate the organoid as a faithful recreation of in vivo conditions, it would be helpful for authors to test some of the more exciting findings on explanted hindgut tissue. One could explant hindguts and test whether blebbistatin treatment silences peristaltic contractions as it does in organoids, or following RCAS-GCAMP infection at earlier stages, one could test the effects of GAP junction inhibitors on Ca++ transients in explanted hindguts. These would potentially serve as useful validation for the gut contractile organoid, and further emphasize the utility of studying these simplified systems for understanding more complex phenomena in vivo.

(6) Organoid fusion experiments are very interesting. It appears that immediately after fusion, the contraction frequency is markedly reduced. Authors should comment on this, and how it changes over time following fusion. Further, is there a relationship between aggregate size and contractile frequency? There are many interesting points that could be discussed here, even if experimental investigation of these points is left to future work.

(7) Minor: As seen in Movie 6 and Figure 6A, 5uM blebbistatin causes a remarkable increase in the frequency of contractions. Given the regular periodicity of these contractions, it is a surprising and potentially interesting finding, but authors do not comment on it. It would be helpful to note this disparity between 5 and 10 uM treatments, if not to speculate on what it means, even if it is beyond the scope of the present study to understand this further.

(8) Minor: While ENS cells are limited in the organoid, it would be helpful to quantify the number of SMCs for comparison in Supplemental Figure S2. In several images, the number of SMCs appears quite limited as well, and the comparison would lend context and a point of reference for the data presented in Figure S2B.

(9) Minor: additional details in the Figure 8 legend would improve interpretation of these results. For example, what is indicated in orange signal present in panels C, G and H? Is this GCAMP?

Reviewer #1 (Public Review):

In the article by Dearlove et al., the authors present evidence in strong support of nucleotide ubiquitylation by DTX3L, suggesting it is a promiscuous E3 ligase with capacity to ubiquitylate ADP ribose and nucleotides. The authors include data to identify the likely site of attachment and the requirements for nucleotide modification.

While this discovery potentially reveals a whole new mechanism by which nucleotide function can be regulated in cells, there are some weaknesses that should be considered. Is there any evidence of nucleotide ubiquitylation occurring cells? It seems possible, but evidence in support of this would strengthen the manuscript. The NMR data could also be strengthened as the binding interface is not reported or mapped onto the structure/model, this seems of considerable interest given that highly related proteins do have the same activity.

The paper is for the most part well well-written and is potentially highly significant, but it could be strengthened as follows:

(1) The authors start out by showing DTX3L binding to nucleotides and ubiquitylation of ssRNA/DNA. While ubiquitylation is subsequently dissected and ascribed to the RD domains, the binding data is not followed up. Does the RD protein alone bind to the nucleotides? Further analysis of nucleotide binding is also relevant to the Discussion where the role of the KH domains is considered, but the binding properties of these alone have not been analysed.<br /> (2) With regard to the E3 ligase activity, can the authors account for the apparent decreased ubiquitylation activity of the 232-C protein in Figure 1/S1 compared to FL and RD?<br /> (3) Was it possible to positively identify the link between Ub and ssDNA/RNA using mass spectrometry? This would overcome issues associated with labels blocking binding rather than modification.<br /> (4) Furthermore, can a targeted MS approach be used to show that nucleotides are ubiquitylated in cells?<br /> (5) Do the authors have the assignments (even partial?) for DTX3L RD? In Figure 4 it would be helpful to identify the peaks that correspond to the residues at the proposed binding site. Also do the shifts map to a defined surface or do they suggest an extended site, particularly for the ssDNA.<br /> (6) Does sequence analysis help explain the specificity of activity for the family of proteins?<br /> (7) While including a summary mechanism (Figure 5I) is helpful, the schematic included does not necessarily make it easier for the reader to appreciate the key findings of the manuscript or to account for the specificity of activity observed. While this figure could be modified, it might also be helpful to highlight the range of substrates that DTX3L can modify - nucleotide, ADPr, ADPr on nucleotides etc.

Reviewer #2 (Public Review):

Summary:

The manuscript by Dearlove et al. entitled "DTX3L ubiquitin ligase ubiquitinates single-stranded nucleic acids" reports a novel activity of a DELTEX E3 ligase family member, DTX3L, which can conjugate ubiquitin to the 3' hydroxyl of single-stranded oligonucleotides via an ester linkage. The findings that unmodified oligonucleotides can act as substrates for direct ubiquitylation and the identification of DTX3 as the enzyme capable of performing such oligonucleotide modification are novel, intriguing, and impactful because they represent a significant expansion of our view of the ubiquitin biology. The authors perform a detailed and diligent biochemical characterization of this novel activity, and key claims made in the article are well supported by experimental data. However, the studies leave room for some healthy skepticism about the physiological significance of the unique activity of DTX3 and DTX3L described by the authors because DTX3/DTX3L can also robustly attach ubiquitin to the ADP ribose moiety of NAD or ADP-ribosylated substrates. The study could be strengthened by a more direct and quantitative comparison between ubiquitylation of unmodified oligonucleotides by DTX3/DTX3L with the ubiquitylation of ADP-ribose, the activity that DTX3 and DTX3L share with the other members of the DELTEX family.

Strengths:

The manuscript reports a novel and exciting observation that ubiquitin can be directly attached to the 3' hydroxyl of unmodified, single-stranded oligonucleotides by DTX3L. The study builds on the extensive expertise and the impactful previous studies by the Huang laboratory of the DELTEX family of E3 ubiquitin ligases. The authors perform a detailed and diligent biochemical characterization of this novel activity, and all claims made in the article are well supported by experimental data. The manuscript is clearly written and easy to read, which further elevates the overall quality of submitted work. The findings are impactful and will help illuminate multiple avenues for future follow-up investigations that may help establish how this novel biochemical activity observed in vitro may contribute to the biological function of DTX3L. The authors demonstrate that the activity is unique to the DTX3/DTX3L members of the DELTEX family and show that the enzyme requires at least two single-stranded nucleotides at the 3' end of the oligonucleotide substrate and that the adenine nucleotide is preferred in the 3' position. Most notably, the authors describe a chimeric construct containing RING domain of DTX3L fused to the DTC domain DTX2, which displays robust NAD ubiquitylation, but lacks the ability to ubiquitylate unmodified oligonucleotides. This construct will be invaluable in the future cell-based studies of DTX3L biology that may help establish the physiological relevance of 3' ubiquitylation of nucleic acids.

Weaknesses:

The main weakness of the study is in the lack of direct evidence that the ubiquitylation of unmodified oligonucleotides reported by the authors plays any role in the biological function of DTX3L. The study leaves plenty of room for natural skepticism regarding the physiological relevance of the reported activity, because, akin to other DELTEX family members, DTX3 and DTX3L can also catalyze attachment of ubiquitin to NAD, ADP ribose and ADP-ribosylated substrates. Unfortunately, the study does not offer any quantitative comparison of the two distinct activities of the enzyme, which leaves plenty of room for doubt. One is left wondering, whether ubiquitylation of unmodified oligonucleotides is just a minor and artifactual side activity owing to the high concentration of the oligonucleotide substrates and E2~Ub conjugates present in the in-vitro conditions and the somewhat lower specificity of the DTX3 and DTX3L DTC domains (compared to DTX2 and other DELTEX family members) for ADP ribose over other adenine-containing substrates such as unmodified oligonucleotides, ADP/ATP/dADP/dATP, etc. The intriguing coincidence that DTX3L, which is the only DTX protein capable of ubiquitylating unmodified oligonucleotides, is also the only family member that contains nucleic acid interacting domains in the N-terminus, is suggestive but not compelling. A recently published DTX3L study by a competing laboratory (PMID: 38000390), which is not cited in the manuscript, suggests that ADP-ribose-modified nucleic acids could be the physiologically relevant substrates of DTX3L. That competing hypothesis appears more convincing than ubiquitylation of unmodified oligonucleotides because experiments in that study demonstrate that ubiquitylation of ADP-ribosylated oligos is quite robust in comparison to ubiquitylation of unmodified oligos, which is undetectable. It is possible that the unmodified oligonucleotides in the competing study did not have adenine in the 3' position, which may explain the apparent discrepancy between the two studies. In summary, a quantitative comparison of ubiquitylation of ADP ribose vs. unmodified oligonucleotides could strengthen the study.

Reviewer #1 (Public Review):

The authors characterized a new non-coding RNA, which they named as PITAR. They first showed that the PITAR expression levels are higher in glioblastoma, and then demonstrated that knockdown of PITAR in glioblastoma cells decreased cell growth, induced G0/G1 arrest and apoptosis. They further identified the E3 ubiquitin ligase TRIM28 is the target of PITAR, and showed that PITAR bound to the TRIM28 mRNA and regulated the stability and expression of the latter. Since TRIM28 has been reported to be an E3 ubiquitin ligase for the tumor suppressor p53, the authors tried to link the PITAR function to p53 regulation. They showed that one PITAR siRNA increased the levels of p53 and p21, and the stability of p53, and these effects could be diminished by overexpression of TRIM28. They also showed that PITAR overexpression decreased the levels of adriamycin-induced p53/p21 expression and reversed DNA damage-induced G2/M arrest. Lastly, the authors showed that PITAR siRNA decreased the growth of glioblastoma, while PITAR overexpression increased glioblastoma growth and counteracted temozolomide for its anti-glioblastoma activity.

Overall, the manuscript has provided evidence supporting the important role of PITAR in the regulation of the growth of glioblastoma. The results supporting the regulation of PITAR on TRIM28 appear to be convincing. However, some weaknesses are also noted.

(1) More than one siRNA/shRNA should be used in critical experiments. For example, Fig 7A-E are important experiments demonstrating PITAR suppresses tumor growth. It is compelling that the siPITAR tumors disappeared at the end of the experiment. While this might be due to apoptosis, using another siRNA to confirm the results would be necessary. The authors may also need to use this model to test their hypothesis that PITAR regulates tumor growth through p53. They can check p53, p21, apoptosis levels in tumor sections.

(2) The data supporting that PITAR downregulates p53 stability and activity can be strengthened. The half-life of endogenous p53 protein is generally 20-30 min, and thus the cycloheximide chase experiments (Fig 5E) need to use shorter treatment time. The ubiquitinated p53 bands are not clear (Fig 5F), and the data suggesting that PITAR regulates p53 ubiquitination are not convincing. While the p53 protein level was largely altered by PITAR/TRIM28, the mRNA levels of its target genes, including p21 and MDM2 only marginally changed (Fig S6D). Other p53 targets, particularly proapoptotic genes, may need to be examined.

(3) The model depicting the role of PITAR in the cellular response to DNA-damaging agents is confusing. If DNA damaging agents like TMZ induce PITAR to inactivate p53, PITAR overexpression would confer TMZ resistance. However, Fig 7G did not support this. While the experimental design is quite problematic given that U87 cells already express a high level of PITAR, PITAR-overexpressing cells were still sensitive to TMZ treatment (this is apparent when checking the images in Fig 7F, although the large error bars shown in Fig 7G may lead to a "not significant" conclusion). The authors may need to test whether PITAR downregulation, which would increase p53 activity, has any effects on TMZ-insensitive tumors. Such results are more therapeutically relevant. It would also be helpful if the authors test whether PITAR is overexpressed in TMZ-resistant clinical samples.

Reviewer #2 (Public Review):

This study established an alternate way of p53 inactivation and proposed PITAR as a potential therapeutic target, so the impact is high. In addition, this manuscript has apparent strengths, including a logically designed research strategy, in vitro and in vivo study, and well-designed control.

This manuscript identified a long noncoding RNA, PITAR (p53 Inactivating TRIM28 associated RNA), as an inhibitor of p53. PITAR is highly expressed in glioblastoma (GBM) and glioma stem-like cells (GSC). The authors found that TRIM28 mRNA, which encodes a p53-specific E3 ubiquitin ligase, is a direct target of PITAR. PITAR interaction with TRIM28 RNA stabilized TRIM28 mRNA, which resulted in increased TRIM28 protein levels, enhanced p53 ubiquitination, and attenuated DNA damage response. While PITAR silencing inhibited the growth of WT p53 containing GSCs in vitro and reduced glioma tumor growth in vivo, its overexpression enhanced the tumor growth and promoted resistance to Temozolomide. DNA damage also activated PITAR, in addition to p53, thus creating an incoherent feedforward loop. Together, this study established an alternate way of p53 inactivation and proposed PITAR as a potential therapeutic target.

P53 is a well-established tumor suppressor gene contributing to cancer progression in many human cancers. It plays a vital role in preserving genome integrity and inhibiting malignant transformation. p53 is mutated in more than 50% of human cancers. In cancers that do not carry mutations in p53, the inactivation occurs through other genetic or epigenetic alterations. Therefore, further study of the mechanism of regulation of wt-p53 remains vital in cancer research. This study identified a novel LncRNA PITAR, which is highly expressed in glioblastoma (GBM) and glioma stem-like cells (GSCs) and interacts with and stabilizes TRIM28 mRNA, which encodes a p53-specific E3 ubiquitin ligase. TRIM28 can inhibit p53 through HDAC1-mediated deacetylation and direct ubiquitination in an MDM2-dependent manner. Thus, the overall impact of this study is high because of the identification of a novel mechanism in regulating wt-p53.

The other significant strengths of this manuscript included an apparent research strategy design and a clearly outlined and logically organized research approach. They provided both the in vitro and in vivo studies to evaluate the effect of PITAR. They offered reasonable control of the study by validating the results in cells with mutant p53. They also performed a rescue experiment to confirm the PITAR and TRIM28 relationship regulating p53. The conclusions were all supported by solid results. The overall data presentation is clear and convincing.

Reviewer #1 (Public Review):

The authors characterized a new non-coding RNA, which they named as PITAR. They first showed that the PITAR expression levels are higher in glioblastoma, and then demonstrated that knockdown of PITAR in glioblastoma cells decreased cell growth, induced G0/G1 arrest and apoptosis. They further identified the E3 ubiquitin ligase TRIM28 is the target of PITAR, and showed that PITAR bound to the TRIM28 mRNA and regulated the stability and expression of the latter. Since TRIM28 has been reported to be an E3 ubiquitin ligase for the tumor suppressor p53, the authors tried to link the PITAR function to p53 regulation. They showed that one PITAR siRNA increased the levels of p53 and p21, and the stability of p53, and these effects could be diminished by overexpression of TRIM28. They also showed that PITAR overexpression decreased the levels of adriamycin-induced p53/p21 expression and reversed DNA damage-induced G2/M arrest. Lastly, the authors showed that PITAR siRNA decreased the growth of glioblastoma, while PITAR overexpression increased glioblastoma growth and counteracted temozolomide for its anti-glioblastoma activity.

Overall, the manuscript has provided preliminary evidence supporting the important role of PITAR in the regulation of the growth and drug resistance of glioblastoma. The results supporting the regulation of PITAR on TRIM28 appear to be convincing. However, the study suffers significant weaknesses summarized as below.

(1) Only one PITAR siRNA was tested in majority of the experiments, which compromises the validity of the results. Some results are inconsistent. For example, Fig 2G indicates that PITAR siRNA caused G1 arrest. However, PITAR overexpression in the same cell line did not show any effect on cell cycle progression in Fig 5I.

(2) The conclusion that PITAR inactivates p53 through regulating TRIM28, which is highlighted in the title of the manuscript, is not supported by convincing results. Although the authors showed that a PITAR siRNA increased while PITAR overexpression decreased p53 level, the siRNA only marginally increased the stability of p53 (Fig 5E). The p53 ubiquitination level was barely affected by PITAR overexpression in Fig 5F. To convincingly demonstrate that PITAR regulates p53 through TRIM28, the authors need to show that this regulation is impaired/compromised in TRIM28-knockout conditions. The authors only showed that TRIM28 overexpression suppressed PITAR siRNA-induced increase of p53, which is not sufficient. Note that only one cell line was investigated in Fig 5.

(3) Another major weakness of this manuscript is that the authors did not provide any evidence indicating that the glioblastoma-promoting activities of PITAR were mediated by its regulation of p53 or TRIM28 (Fig 6 and Fig 7). Thus, the regulation of glioblastoma growth and the regulation of TRIM28/p53 appear to be disconnected.

(4) It is not clear what kind of message the authors tried to deliver in Fig 7F/G. Based on the authors' hypothesis, DNA damaging agents like TMZ would induce PITAR to inactivate p53, which would compromise TMZ's anti-cancer activity. However, the data show that TMZ was very effective in the inhibition of U87 growth. The authors may need to test whether PITAR downregulation, which would increase p53 activity, have any effects on TMZ-insensitive tumors. Such results are more therapeutically relevant.

(5) Lastly, the model presented in Fig 7H is confusing. It is not clear what the exact role of PITAR in the DNA damage response based on this model. If DNA damage would induce PITAR expression, this would lead to inactivation of p53 as revealed by this manuscript. However, DNA damage is known to activate p53. Do the authors want to imply that PITAR induction by DNA damage would help to bring down the p53 level at the end of DNA damage response? The presented data do not support this role unfortunately.

Reviewer #1 (Public Review):

SUMMARY:

The goal of Knudsen-Palmer et al. was to define a biological set of rules that dictate the differential RNAi-mediated silencing of distinct target genes, motivated by facilitating the long-term development of effective RNAi-based drugs/therapeutics. This work provides insights into how 1) cis-regulatory elements influence the RNAi-mediated regulation of genes; 2) determines that genes can "recover" from RNAi-silencing signals in an animal; and 3) pUGylation occurs exclusively downstream of the dsRNA trigger sequence, suggesting 3º siRNAs are not produced. In addition, the authors show that the speed at which RNAi-silencing is triggered does not correlate with the longevity of the silencing. Overall, the work presented supports the conclusions of the authors. The insights are significant because they suggest that if we understand the rules by which RNAi pathways effectively silence genes with different transcription/processing levels then we can design more effective synthetic RNAi-based therapeutics targeting endogenous genes.

MAJOR STRENGTH:

The authors use a combination of computational modeling, genetics, and RNAi function assays to reveal several criteria for effective RNAi-mediated silencing of two distinct targets.

WEAKNESS:

It may be beyond the scope of this study, but it would be interesting to know the typical expression levels and turnover rates of unc-22 and bli-1. Based on the results from the altered cis-regulatory regions of bli-1 and unc-22 in Fig 5, it seems like the transcription/turnover rates of each of these genes could also be used as a proof of principle for testing the model proposed in Figure 4. The strength of the model would be further increased if the RNAi sensitivity of unc-22 reflects differences in its transcription/turnover rates compared to bli-1.

Reviewer #2 (Public Review):

SUMMARY

This manuscript by Knudsen-Palmer et al. describes and models the contribution of MUT-16 and RDE-10 in the silencing through RNAi by the Argonaute protein NRDE-3 or others. The authors show that MUT-16 and RDE-10 constitute an intersecting network that can be redundant or not depending on the gene being targeted by RNAi. In addition, the authors provide evidence that increasing dsRNA processing can compensate for NRDE-3 mutants. Overall, the authors provide convincing evidence to understand the factors involved in RNAi in C. elegans by using a genetic approach.

MAJOR STRENGTHS

The author's work presents a compelling case for understanding the intricacies of RNA interference (RNAi) within the model organism Caenorhabditis elegans through a meticulous genetic approach. By harnessing genetic manipulation, they delve into the role of MUT-16 and RDE-10 in RNAi, offering a nuanced understanding of the molecular mechanisms at play in two independent case study targets (unc-22 and bli-1).

MAJOR WEAKNESSES

(1) It is unclear how the molecular mechanisms of amplification are different under the MUT-16 and RDE-10 branches of the regulatory pathway, since they are clearly distinct proteins structurally. It would be interesting to do some small-RNA-seq of products generated from unc-22 and bli-1, on wild-type conditions and some of the mutants studied (eg. mut-16, rde-10 and mut-16 + rde-10). That would provide some insights on whether the products of the 2 amplifications are the same in all conditions, just changing in abundance, or whether they are distinct in sequence patterns.

(2) In the same line, Figure 5 aims to provide insights to the sequence determinants that influence on the RNAi of bli-1. It is unclear whether the changes in transcript stability dictated by the 3'UTR are the sole factor governing the preference for the MUT-16 and RDE-10 branches of the regulatory pathway. In line with the mutant jam297, it might be interesting to test whether factors like codon optimality, splicing, ... of the ORF region upstream from bli-1-dsRNA can affect its sensitivity to the MUT-16 and RDE-10 branches of the regulatory pathway.

Reviewer #1 (Public Review):

Summary:

TMC7 knockout mice were generated by the authors and the phenotype was analyzed. They found that Tmc7 is localized to Golgi and is needed for acrosome biogenesis.

Strengths:

The phenotype of infertility is clear, and the results of TMC7 localization and the failed acrosome formation are highly reliable. In this respect, they made a significant discovery regarding spermatogenesis.

In the original version, I pointed out the gap between their pH/calcium imaging data and the hypothesis of ion channel function of TMC7 in the Golgi. Now the author agrees and has changed the description to be reasonable. Additional experiments were also performed, and I can say that they have answered my concern adequately.

I would say it is good to add any presumed mechanism for the observed changes in pH and calcium concentration in the cytoplasm this time.

Reviewer #2 (Public Review):

Summary:

This study presents a significant finding that enhances our understanding of spermatogenesis. TMC7 belongs to a family of transmembrane channel-like proteins (TMC1-8), primarily known for their role in the ear. Mutations to TMC1/2 are linked to deafness in humans and mice and were originally characterized as auditory mechanosensitive ion channels. However, the function of the other TMC family members remains poorly characterized. In this study, the authors begin to elucidate the function of TMC7 in acrosome biogenesis during spermatogenesis. Through analysis of transcriptomics datasets, they elevated levels of TMC7 in round spermatids in both mouse and human testis. They then generate Tmc7-/- mice and find that male mice exhibit smaller testes and complete infertility. Examination of different developmental stages reveals spermatogenesis defects, including with reduced sperm count, elongated spermatids and large vacuoles. Additionally, abnormal acrosome morphology are observed beginning at the early-stage Golgi phase, indicating TMC7's involvement in proacrosomal vesicle trafficking and fusion. They observed localization of TMC7 in the cis-Golgi and suggest that its presence is required for maintaining Golgi integrity, with Tmc7-/- leading to reduced intracellular Ca2+, elevated pH and increased ROS levels, likely resulting in spermatid apoptosis. Overall, the work delineates a new function of TMC7 in spermatogenesis and the authors propose that its ion channel and/or scramblase activity is likely important for Golgi homeostasis. This work is of significant interest to the community and is of high quality.

Strengths:

The biggest strength of the paper is the phenotypic characterization of the TMC7-/- mouse model, which has clear acrosome biogenesis/spermatogenesis defects. This is the main claim of the paper and it is supported with the data that are presented.

Weaknesses:

It isn't clear whether TMC7 functions as an ion channel from the current data presented in this paper, but the authors are careful in their interpretation and present this merely as a hypothesis supporting this idea.

Reviewer #3 (Public Review):

Summary:

In this study, Wang et al. have demonstrated that TMC7, a testis-enriched multipass transmembrane protein, is essential for male reproduction in mice. Tmc7 KO male mice are sterile due to reduced sperm count and abnormal sperm morphology. TMC7 co-localizes with GM130, a cis-Golgi marker, in round spermatids. The absence of TMC7 results in reduced levels of Golgi proteins, elevated abundance of ER stress markers, as well as changes of Ca2+ and pH levels in the KO testis. However, further confirmation is required because the analyses were performed with whole testis samples in spite of the differences in the germ cell composition in WT and KO testis. In addition, the causal relationships between the reported anomalies await thorough interrogation

Strengths:

By using PD21 testes, the revised assays have consolidated that depletion of TMC7 leads to a reduced level of Ca2+ and an elevated level of ROS in the male germ cells. The immunohistochemistry analyses have clearly indicated the reduced abundance of GM130, P115, and GRASP65 in the knockout testis.

Weaknesses:

Future studies are required to decipher how TMC7 stabilizes Golgi structure, coordinates vesicle transport, and maintains the germ cell homeostasis.

Reviewer #1 (Public Review):

Summary:

In this paper, the authors performed molecular dynamics (MD) simulations to investigate the molecular basis of association of alpha-synuclein chains under molecular crowding and salt conditions. Aggregation of alpha-synuclein is linked to the pathogenesis of Parkinson's disease, and the liquid-liquid phase separation (LLPS) is considered to play an important role in the nucleation step of the alpha-synuclein aggregation. This paper re-tuned the Martini3 coarse-grained force field parameters, which allows long-timescale MD simulations of intrinsically disordered proteins with explicit solvent under diverse environmental perturbation. Their MD simulations showed that alpha-synuclein does not have a high LLPS-forming propensity, but the molecular crowding and salt addition tend to enhance the tendency of droplet formation and therefore modulate the alpha-synuclein aggregation. The MD simulation results also revealed important intra and inter-molecule conformational features of the alpha-synuclein chains in the formed droplets and the key interactions responsible for the stability of the droplets. These MD simulation data add biophysical insights into the molecular mechanism underlying the association of alpha-synuclein chains, which may be useful for understanding the pathogenesis of Parkinson's disease.

Strengths:

(1) The re-parameterized Martini 3 coarse-grained force field enables the large-scale MD simulations of the intrinsically disordered proteins with explicit solvent, which will be useful for a more realistic description of the molecular basis of LLPS.

(2) This paper showed that the molecular crowding and salt contribute to the modulation of the LLPS through different means. The molecular crowding minimally affects surface tension, but adding salt increases surface tension. It is also interesting to show that the aggregation pathway involves the disruption of the intra-chain interactions arising from C-terminal regions, which potentially facilitates the formation of inter-chain interactions.

Weaknesses:

(1) Although the authors emphasized the advantage of the Martini3 force field for its explicit description of solvent, this paper did not analyze the water behavior contained in the simulation trajectories and discuss the water's role in the aggregation and LLPS.

(2) This paper discussed the effects of crowders and salt on the surface tension of the droplets. The calculation of the surface tension relies on the droplet shape. However, for the formed clusters in the MD simulations, the typical size is <10, which may be too small to rigorously define the droplet shape. As shown in previous work cited by this paper [Benayad et al., J. Chem. Theory Comput. 2021, 17, 525−537], the calculated surface tension becomes stable when the chain number is larger than 100.

(3) Both the sizes and volume fractions of the crowders can affect the protein association. It will be interesting to perform MD simulations by adding crowders with various sizes and volume fractions. In addition, in this work the crowders were modelled by fullerenes, which contribute to protein aggregation mainly by entropic means as discussed in the manuscript. It is not very clear how the crowder effect is sensitive to the chemical nature of the crowders (e.g., inert crowers with excluded volume effect or crowders with non-specific attractive interactions with proteins, etc).

Reviewer #2 (Public Review):

In the manuscript "Modulation of α-Synuclein Aggregation Amid Diverse Environmental Perturbation", Wasim et al describe coarse-grained molecular dynamics (cgMD) simulations of α-Synuclein (aSyn) at several concentrations and in the presence of molecular crowding agents or high salt. They begin by bench-marking their cgMD against all-atom simulations by Shaw. They then carry 2.4-4.3 µs cgMD simulations under the above-noted conditions and analyze the data in terms of protein structure, interaction network analysis, and extrapolated fluid mechanics properties. This is an interesting study because a molecular scale understanding of protein droplets is currently lacking.

Reviewer #1 (Public Review):

Summary:

In this manuscript, Fister et. al. investigate how amputational and burn wounds affect sensory axonal damage and regeneration in a zebrafish model system. The authors discovered that burn injury results in increased peripheral axon damage and impaired regeneration. Convincing experiments show altered axonal morphology and increased Ca2+ fluxes as a result of burn damage. Further experimental proof supports that early removal of the burnt tissue by amputation rescues axonal damage. Burn damage was also shown to markedly increase keratinocyte migration and increase localized ROS production as measured by the dye Pfbsf. These responses could be inhibited by Arp 2/3 inhibition and isotonic treatment.

Strengths:

The authors use state-of-the-art methods to study and compare transection and burn-induced tissue damage. Multiple experimental approaches (morphology, Ca2+ fluxing, cell membrane labeling) confirm axonal damage and the impaired regeneration time. Furthermore, the results are also accompanied by functional response tests of touch sensitivity. This is the first study to extend the role of tissue-damage related osmotic exposure beyond wound closure and leukocyte migration to a novel layer of pathology: axonal damage and regeneration.

The authors provide elegant experiments showing that early removal of the burnt tissue can rescue damage-induced axonal damage, which could also be interpreted in an osmotic manner. In the revised version of the paper the authors indeed show that tail fin transections close faster than burn wounds, allowing for lower hypotonic exposure time. However, their new experiments suggest that axonal damage and slow regeneration in tail fin burn wounds are not a direct consequence of the extended exposure time to hypotonic water.

Weaknesses:

The conclusions of the paper claiming a link between burn-induced epithelial cell migration, spatial redox signaling, and sensory axon regeneration are mainly based on correlative observations. Arp 2/3 inhibition impairs cell migration but has no significant effect on axon regeneration and restoration of touch sensitivity.

Genetic approaches have been tested during the revision process to directly prove the role of ROS production by targeting DUOX, however, the combination of DUOX morpholino and burn injury was lethal to the larvae and long-term pharmacological inhibition over 1 hour was also detrimental.

Reviewer #3 (Public Review):

Fister and colleagues use regeneration of the larval zebrafish caudal fin to compare the effects of two modes of tissue damage-transection and burn-on cutaneous sensory axon regeneration. The authors found that restoration of sensory axon density and function is delayed following burn injury compared to transection.

The authors hypothesized that thermal injury triggers signals within the wound microenvironment that impair sensory neuron regeneration. The authors identify differences in the responses of epithelial keratinocytes to the two modes of injury: keratinocytes migrate in response to burn but not transection. Inhibiting keratinocyte migration with a small-molecule inhibitor of Arp2/3 (CK666) resulted in decreased production of reactive oxygen species (ROS) at early, but not late, timepoints. Preventing keratinocyte migration by wounding in isotonic media resulted in increased sensory function 24 hours after burn.

Strengths of the study include the beautiful imaging and rigorous statistical approaches used by the authors. The ability to assess both axon density and axon function during regeneration is quite powerful. The touch assay adds a unique component to the paper and strengthens the argument that burns are more damaging to sensory structures and that different treatments help to ameliorate this.

A weakness of the study is the lack of genetic and cell autonomous manipulations. Additional comparisons between transection and burns, in particular with manipulations that specifically modulate ROS generation or cell migration without potentially confounding effects on other cell types or processes would help to strengthen the manuscript. In terms of framing their results, the authors refer to "sensory neurons" and "sensory axons" throughout the text - it should be made clear what type of neuron(s)/axon(s) are being visualized/assayed. Along these lines, a broader discussion of how burn injuries affect sensory function in other systems-and how the authors' results might inform our understanding of these injury responses-would be beneficial to the reader.

In summary, the authors have established a tractable vertebrate system to investigate different sensory axon wound healing outcomes in vivo that may ultimately allow for the identification of improved treatment strategies for human burn patients. Although the study implicates differences in keratinocyte migration and associated ROS production in sensory axon wound healing outcomes, the links between these processes could be more rigorously established.

Reviewer #1 (Public Review):

In the manuscript Chugh and co-workers utilize a suite of NMR relaxation methods to probe the dynamic landscape of the TAR RNA binding protein (TRBP) double-stranded RNA-binding domain 2 (dsRBD2) and compare these to their previously published results on TRBP dsRBD1. The authors show that, unlike dsRBD1, dsRBD2 is a rigid protein with minimal ps-ns or us-ms time scale dynamics in the absence of RNA. They then show that dsRBD2 binds to canonical A-form dsRNA with a higher affinity and with less changes in dynamics compared to dsRBD1.

Strengths:

The authors expertly use a variety of NMR techniques to probe protein motions over six-orders of magnitude in time. Other NMR titration experiments and ITC data support the RNA-binding model.

Weaknesses:

Generally, the data collection and analysis are sound. However, microsecond timescale dynamics for the RNA-bound form of dsRBD2 are inferred from a sample that is only 5% bound. Additionally, the manuscript lacks context with the much broader field of RNA-binding proteins. For example, many studies have shown that RNA recognition motif (RRM) domains have similar dynamic characteristics when binding diverse RNA substrates.

Reviewer #2 (Public Review):

Summary:

Proteins that bind to double-stranded RNA regulate various cellular processes, including gene expression and viral recognition. Such proteins often contain multiple double-stranded RNA-binding domains (dsRBDs) that play an important role in target search and recognition. In this work, Chug and colleagues have characterized the backbone dynamics of one of the dsRBDs of a protein called TRBP2, which carries two tandem dsRBDs. Using solution NMR spectroscopy, the authors characterize the backbone motions of dsRBD2 in the absence and presence of dsRNA and compare these with their previously published results on dsRBD1. The authors show that dsRBD2 is comparatively more rigid than dsRBD1 and claim that these differences in backbone motions are important for target recognition.

Strengths:

The strengths of this study are multiple solution NMR measurements to characterize the backbone motions of dsRBD2. These include 15N-R1, R2, and HetNOE experiments in the absence and presence of RNA and the analysis of these data using an extended-model-free approach; HARD-15N-experiments and their analysis to characterize the kex. The authors also report differences in binding affinities of dsRBD1 and dsRBD2 using ITC and have performed MD simulations to probe the differential flexibility of these two domains.

Weaknesses:

While it may be true that dsRBD2 is more rigid than dsRBD1, the manuscript lacks conclusive and decisive proof that such changes in backbone dynamics are responsible for target search and recognition and for the diffusion of TRBP2 along the RNA molecule.

Reviewer #1 (Public Review):

The study identifies the epigenetic reader SntB as a crucial transcriptional regulator of growth, development, and secondary metabolite synthesis in Aspergillus flavus, although the precise molecular mechanisms remain elusive. Using homologous recombination, researchers constructed sntB gene deletion (ΔsntB), complementary (Com-sntB), and HA tag-fused sntB (sntB-HA) strains. Results indicated that deletion of the sntB gene impaired mycelial growth, conidial production, sclerotia formation, aflatoxin synthesis, and host colonization compared to the wild type (WT). The defects in the ΔsntB strain were reversible in the Com-sntB strain.

Further experiments involving ChIP-seq and RNA-seq analyses of sntB-HA and WT, as well as ΔsntB and WT strains, highlighted SntB's significant role in the oxidative stress response. Analysis of the catalase-encoding catC gene, which was upregulated in the ΔsntB strain, and a secretory lipase gene, which was downregulated, underpinned the functional disruptions observed. Under oxidative stress induced by menadione sodium bisulfite (MSB), the deletion of sntB reduced catC expression significantly. Additionally, deleting the catC gene curtailed mycelial growth, conidial production, and sclerotia formation, but elevated reactive oxygen species (ROS) levels and aflatoxin production. The ΔcatC strain also showed reduced susceptibility to MSB and decreased aflatoxin production compared to the WT.

This study outlines a pathway by which SntB regulates fungal morphogenesis, mycotoxin synthesis, and virulence through a sequence of H3K36me3 modification to peroxisomes and lipid hydrolysis, impacting fungal virulence and mycotoxin biosynthesis.

The authors have achieved the majority of their aims at the beginning of the study, finding target genes, which led to catC mediated regulation of development, growth and aflatoxin metabolism. Overall most parts of the study are solid and clear.

Reviewer #2 (Public Review):

Summary:

Wu et al. explores the role of the histone reader protein SntB in Aspergillus flavus. They not only studied its function related to the growth, development, and secondary metabolite through gene knockout and complement, but also explored the underlying potential mechanisms by RNA-seq and ChIP-seq. The response of oxidative stress in ΔsntB strain and ΔcatC strain were further analyzed. Their study revealed a potential machinery that SntB regulated fungal morphogenesis, mycotoxin anabolism, and fungal virulence through the axle of from epigenetic modification to fungal virulence and mycotoxin bio-synthesis via SntB, i.e. H3K36me3 modification-SntB-Peroxisomes-Lipid hydrolysis-fungal virulence and mycotoxin bio-synthesis. This work is of great significance in revealing the regulatory mechanisms of pathogenic fungi in toxin production, pathogenicity, and in its prevention and pollution control.

Strengths:

One of the main advantages of this study is that the author constructed HA fused strains for ChIP seq analysis, rather than using antibodies related to epigenetic modifications. Nancy et al. reported the functions of sntB as a histone methylation regulator, but in addition to being an epigenetic regulator, there are also reports that it has transcriptional regulatory activity. Through integration analysis with RNA-seq data, it was found that SntB played key roles in oxidative stress response of A. flavus. This study can increase our understanding of more functions of the SntB in A. flavus.

Weaknesses:

The authors only studied the function of catC among the 7 genes related to oxidative response listed in Table S14.

Reviewer #3 (Public Review):

Summary:

The authors have devised an elegant stopped-flow fluorescence approach to probe the mechanism of action of the Hsp100 protein unfoldase ClpB on an unfolded substrate (RepA) coupled to 1-3 repeats of a folded titin domain. They provide useful new insight into the kinetics of ClpB action. The results support their conclusions for the model setup used.

Strengths:

The stopped-flow fluorescence method with a variable delay after mixing the reactants is informative, as is the use of variable numbers of folded domains to probe the unfolding steps.

Weaknesses:

The setup does not reflect the physiological setting for ClpB action. A mixture of ATP and ATPgammaS is used to activate ClpB without the need for its co-chaperones, Hsp70. Hsp40 and an Hsp70 nucleotide exchange factor. This nucleotide strategy was discovered by Doyle et al (2007) but the mechanism of action is not fully understood. Other authors have used different approaches. As mentioned by the authors, Weibezahn et al used a construct coupled to the ClpA protease to demonstrate translocation. Avellaneda et al used a mutant (Y503D) in the coiled-coil regulatory domain to bypass the Hsp70 system. These differences complicate comparisons of rates and step sizes with previous work. It is unclear which results, if any, reflect the in vivo action of ClpB on the disassembly of aggregates.

Reviewer #1 (Public Review):

In this study, the authors used a stopped-flow method to investigate the kinetics of substrate translocation through the channel in hexameric ClpB, an ATP-dependent bacterial protein disaggregase. They engineered a series of polypeptides with the N-terminal RepA ClpB-targeting sequence followed by a variable number of folded titin domains. The authors detected translocation of the substrate polypeptides by observing the enhancement of fluorescence from a probe located at the substrate's C-terminus. The total time of the substrates' translocation correlated with their lengths, which allowed the authors to determine the number of residues translocated by ClpB per unit time.

Strengths:

This study confirms a previously proposed model of processive translocation of polypeptides through the channel in ClpB. The novelty of this work is in the clever design of a series of kinetic experiments with an engineered substrate that includes stably folded domains. This approach produced a quantitative description of the reaction rates and kinetic step sizes. Another valuable aspect is that the method can be used for other translocases from the AAA+ family to characterize their mechanism of substrate processing.

Weaknesses:

The main limitation of the study is in using a single non-physiological substrate of ClpB, which does not replicate the physical properties of the aggregated cellular proteins and includes a non-physiological ClpB-targeting sequence. Another limitation is in the use of ATPgammaS to stimulate the substrate processing. It is not clear how relevant the results are to the ClpB function in living cells with ATP as the source of energy, a multitude of various aggregated substrates without targeting sequences that need ClpB's assistance, and in the presence of the co-chaperones.

The authors do not attempt to correlate the kinetic step sizes detected during substrate translocation and unfolding with the substrate's structure, which should be possible, given how extensively the stability and unfolding of the titin I27 domain were studied before. Also, since the substrate contains up to three I27 domains separated with unstructured linkers, it is not clear why all the translocation steps are assumed to occur with the same rate constant.

Some conclusions presented in the manuscript are speculative:

The notion that the emission from Alexa Fluor 555 is enhanced when ClpB approaches the substrate's C-terminus needs to be supported experimentally. Also, evidence that ATPgammaS without ATP can provide sufficient energy for substrate translocation and unfolding is missing in the paper.

Reviewer #2 (Public Review):

Summary:

The current work by Banwait et al. reports a fluorescence-based single turnover method based on protein-induced fluorescence enhancement (PIFE) to show that ClpB is a processive motor. The paper is a crucial finding as there has been ambiguity on whether ClpB is a processive or non-processive motor. Optical tweezers-based single-molecule studies have shown that ClpB is a processive motor, whereas previous studies from the same group hypothesized it to be a non-processive motor. As co-chaperones are needed for the motor activity of the ClpB, to isolate the activity of ClpB, they have used a 1:1 ratio ATP and ATPgS, where the enzyme is active even in the absence of its co-chaperones, as previously observed. A sequential mixing stop-flow protocol was developed, and the unfolding and translocation of RepA-TitinX, X = 1,2,3 repeats was monitored by measuring the fluorescence intensity with the time of Alexa F555 which was labelled at the C-terminal Cysteine. The observations were a lag time, followed by a gradual increase in fluorescence due to PIFE, and then a decrease in fluorescence plausibly due to the dissociation from the substrate allowing it to refold. The authors observed that the peak time depends on the substrate length, indicating the processive nature of ClpB. In addition, the lag and peak times depend on the pre-incubation time with ATPgS, indicating that the enzyme translocates on the substrates even with just ATPgS without the addition of ATP, which is plausible due to the slow hydrolysis of ATPgS. From the plot of substrate length vs peak time, the authors calculated the rate of unfolding and translocation to be ~0.1 aas-1 in the presence of ~1 mM ATPgS and increases to 1 aas-1 in the presence of 1:1 ATP and ATPgS. The authors have further performed experiments at 3:1 ATP and ATPgS concentrations and observed ~5 times increase in the translocation rates as expected due to faster hydrolysis of ATP by ClpB and reconfirming that processivity is majorly ATP driven. Further, the authors model their results to multiple sequential unfolding steps, determining the rate of unfolding and the number of amino acids unfolded during each step. Overall, the study uses a novel method to reconfirm the processive nature of ClpB.

Strengths:

(1) Previous studies on understanding the processivity of ClpB have primarily focused on unfolded or disordered proteins; this study paves new insights into our understanding of the processing of folded proteins by ClpB. They have cleverly used RepA as a recognition sequence to understand the unfolding of titin-I27 folded domains.

(2) The method developed can be applied to many disaggregating enzymes and has broader significance.

(3) The data from various experiments are consistent with each other, indicating the reproducibility of the data. For example, the rate of translocation in the presence of ATPgS, ~0.1 aas-1 from the single mixing experiment and double mixing experiment are very similar.

(4) The study convincingly shows that ClpB is a processive motor, which has long been debated, describing its activity in the presence of only ATPgS and a mixture of ATP and ATPgS.

(5) The discussion part has been written in a way that describes many previous experiments from various groups supporting the processive nature of the enzyme and supports their current study.

Weaknesses:

(1) The authors model that the enzyme unfolds the protein sequentially around 60 aa each time through multiple steps and translocates rapidly. This contradicts our knowledge of protein unfolding, which is generally cooperative, particularly for titinI27, which is reported to unfold cooperatively or utmost through one intermediate during enzymatic unfolding by ClpX and ClpA.

(2) It is also important to note that the unfolding of titinI27 from the N-terminus (as done in this study) has been reported to be very fast and cannot be the rate-limiting step as reported earlier(Olivares et al, PNAS, 2017). This contradicts the current model where unfolding is the rate-limiting step, and the translocation is assumed to be many orders faster than unfolding.

(3) The model assumes the same time constant for all the unfolding steps irrespective of the secondary structural interactions.

(4) Unlike other single-molecule optical tweezer-based assays, the study cannot distinguish the unfolding and translocation events and assumes that unfolding is the rate-limiting step.

Reviewer #1 (Public Review):

V.Mischley et al have applied several simple machine learning (ML)frameworks (which were widely used before the advent of deep learning methods) to distinguish (as the authors claimed) between interacting and non-interacting pairs. For this purpose, the authors have generated two sets of protein pairs, equal in their size (which is preferable for classification problems in ML). The first set comprises a non-redundant set of interacting proteins from the DOCKGROUND database, and the second set consists of presumably non-interacting protein pairs. Then, the authors trained and evaluated compared performance of the utilized ML frameworks using a set of well-described parameters. The authors also demonstrated the superior performance of their method in comparison to other metrics, such as ipTM and pdockQ. Finally, the authors applied their method to identify interacting pairs within the tumor necrosis factor superfamily. In general, the paper is well written, and the methodology applied is sound, however, I have a fundamental concern regarding the non-interacting set. As follows from the author's description, this set does not ensure that generated protein pairs do not interact as follows from the main paradigm of template-based docking (structurally similar proteins have similar binding modes). In my opinion, this set rather presents a set of non-cognate or weekly interacting protein pairs. That also explains the drop in performance for the pDockQ metric on the authors' set (AUC 0.71 in this paper opposite t0 0.87 in the original paper), as pDockQ was trained on the set of truly non-interacting proteins. In that respect, it would be interesting to see the performance of the authors' approach, but trained on the set described in the pDockQ paper (more or less the same set of interacting pairs but a different set of non-interacting proteins).

Reviewer #2 (Public Review):

Summary:

In this paper, the authors train a simple machine learning to improve the ability of AlphaFold-multimers ability to separate interacting from non-interacting pairs. The improvement is small compared with the default AlphaFold score (AUROC from 0.84 to 0.88).

Strengths:

The dataset seems to be carefully constructed.

Weaknesses:

The comparison with the state of the art is limited.<br /> - pDockQ comparison is (likely) incorrect (v2.1 should be used, not v1.0).<br /> - Comparison with ipTM should be complemented with RankingConfidence (the default AF2-score).<br /> - Several other scores than pDockQ have been developed for this task.<br /> - Other methods (by Jianlin Chen) to "improve" quality assessment of AF2-models have been presented - these should at least be cited.

Lack of ablation studies:

- Quite likely the most significant contributor is the ipTM (and other scores from AF2). This should be analyzed and discussed.

Lack of data:

- The GitHub repository does not contain the models - so the data can not be examined carefully. Nor can the model be retrained.

- No license is provided for the code in the Git repository.

Reviewer #3 (Public Review):

Due to AlphaFold's popularity, I see people taking the fact that AlphaFold predicted a decent protein complex structure between two proteins as strong support for protein-protein interaction (PPI) and even using such a hypothesis to guide their experimental studies. The scientific community needs to realize that just like the experimental methods to characterize PPIs, using AlphaFold to study PPIs has a considerate false positive and false negative rate.

Overall, I think it is solid work, but I have several concerns.

(1) In the benchmark set, the authors used about 1:1 ratio of positive (active) and negative controls. However, in real-life applications, the signal-to-noise ratio of PPI screening is very low. As they stated in their very nice introduction, there are expected to be "74,000 - 200,000" true PPIs in humans, whereas there are > 200,000,000 protein pairs. I am not suggesting that the authors need to make their tool able to handle such a high noise level, but at least some discussion along this line is helpful.

(2) The benchmark set from Dockground mostly consists of stable interactions that are actually relatively easily distinguished from non-interacting pairs. I suggest the authors test how well their tools will perform on weaker and transient interactions or discuss this limitation. For the more stable complexes, structural features at the interface are useful in predicting whether two proteins should interact, but I doubt this will be true for weaker and transient interactions.

(3) Given that the 1:1 benchmark set is a simplified task (see the first point) compared to real-life applications, the other task shown in this paper, i.e., the ligand/receptor pairings, seems to be more important. I think it is necessary to compare their tool against other simpler metrics for this more realistic task.

Reviewer #1 (Public Review):

Summary:

The authors' claims that CD9 and CD81 are key regulators of TNT formation and function are well-supported by the data. The use of KO and OE models provides strong evidence. The differential proteomic analysis between TNTs and EVPs and the functional assays justify the conclusion that these tetraspanins are critical for TNT biogenesis and functionality. Overall, the manuscript presents a nice study that advances our understanding of TNTs and their regulation by CD9 and CD81. Despite some limitations, the strengths of the experimental design and the robustness of the data justify the authors' conclusions. Future studies addressing the identified weaknesses would further solidify these findings and their implications in pathological contexts.

Strengths:

Novelty and Significance - this study addresses the composition and regulation of tunneling nanotubes (TNTs). By identifying the roles of CD9 and CD81 tetraspanins, the researchers offer insights into the molecular mechanisms underlying TNT formation. This could have implications for understanding cellular communication in pathological conditions such as cancer.

Methodological Accuracy - the authors employed a well-designed biochemical approach to isolate TNTs from U2OS cells, distinguishing them from extracellular vesicles and particles (EVPs). The use of multiple independent preparations and the application of LC-MS/MS for proteomic analysis ensure robustness and reproducibility of the data.

Complete Analysis - the study provides a detailed proteomic profile of TNTs, identifying 1177 proteins and highlighting key components. The comparative analysis between TNTs and EVPs further strengthens the findings by demonstrating distinct proteomic landscapes.

Functional Insights - using knockout (KO) and overexpression (OE) models, the authors convincingly demonstrate the distinct roles of CD9 and CD81 in TNT formation and function. CD9 is shown to stabilize TNTs, while CD81 facilitates vesicle transfer, likely by aiding membrane docking or fusion.

Experimental Design - the use of actin chromobody-GFP and various fluorescent markers enabled the authors to visualize TNTs and validate their isolation protocol. Additionally, the combination of electron microscopy, flow cytometry, and live-cell imaging provided convincing evidence for their claims.

Weaknesses:

Potential Contaminations - while the authors took steps to minimize contamination with other cellular structures, the presence of some nuclear proteins and the possible inclusion of small portions of cell bodies or ER in the TNT preparations cannot be entirely ruled out. This may affect the interpretation of some proteomic data.

Limited Cell Models - the experiments were conducted in U2OS and SH-SY5Y cells. While these are relevant models, in vivo validation of the findings would significantly enhance the impact and translational potential of the research.

Functional Mechanisms - although the study provides strong evidence for the roles of CD9 and CD81, the exact molecular mechanisms by which these tetraspanins regulate TNT formation and vesicle transfer remain partially speculative. Further biochemical and biophysical analyses would be necessary to elucidate these mechanisms in detail.

Reviewer #2 (Public Review):

Tunneling nanotubes (TNT) are common cellular protrusions that allow the transfer of multiple types of cargo between mammalian cells. TNTs are fragile, and lack any known unique marker, making it challenging to isolate and study them. Therefore, the content of TNTs is mostly unknown, and there are only a handful of proteins known to play a role in TNT formation or function.

In this paper, the authors developed a new protocol to isolate TNT fragments from a culture of adherent mammalian cells in a way that is distinctive of extracellular vesicle and identify the proteins within the TNT (referred to as TNTome) by mass spectrometry. The authors provide an analysis of the results in comparison to the extracellular vesicle (EV) proteome, and validate a few examples, thus providing valuable data for the TNT field. However, there is a big overlap between TNTome and EV proteome.

The authors further focus on two proteins, CD9 and CD81, that are enriched in TNTs. Using cells that are knocked out (KO) or over-expressing (OE) these proteins, the authors study their role in TNT formation and function. The authors focus on two major parameters, which are the percent of cells connected by TNT, and the percent of acceptor cells containing fluorescently labeled transferred vesicles. The authors use various assays, which are properly controlled, to measure these parameters. Their analysis provides convincing evidence that CD9 plays a partial role in TNT formation or stabilization and CD81 plays a partial role in forming fully elongated/connected TNT.

However, the authors overstate the importance of these proteins, since their absence only partially affects TNT formation and function, similar to what is seen when knocking out most any other protein implicated in TNT formation. Even their best results show just a 50% reduction for TNT formation and 70% vesicle transfer (in the double KO). Thus, these are not "key" regulators as the title suggests - no more than many other factors, some of them identified by the authors in previous publications. The model presented in Figure 7D is thus misleading, as it states that CD9 KO has "No TNT" which is incorrect (only a slight decrease according to Figure 3C), and states that CD81 KO has "Non-functional TNT" whereas there is still 50% vesicle transfer in this mutant.

In addition, the authors use vesicle transfer as a measure of function, but this is just one type of cargo amongst many others: ions, proteins, RNA, various organelles, and pathogens like viruses and bacteria. Since the authors clearly cannot test every type of cargo, the authors should at least be more accurate in their statements regarding functionality and mention the possibility that other types of cargo transfer could be less or more affected by the KO or OE of these proteins.

It is not completely clear from the text why the authors decided to focus on CD9 and CD81, which are also found in EV, instead of focusing on TNT-unique proteins, and in particular the cytoskeleton-related ones.

In summary, it is a good paper, that provides valuable data on the composition of TNT, and the role of additional players, bringing us closer to understanding the mechanism of TNT formation.

Reviewer #3 (Public Review):

Initially, the authors isolated TNTs from EVPs and cell bodies of cultured U2OS cells. Using transmission electronic microscopy and nanoflow cytometry, they demonstrated that these two structures are morphologically different. In engineered cells, they observed the presence of actin and CD9 in TNTs by immunofluorescence. Then they employed mass spectrometry techniques to analyze the EVPs and TNT fractions, discovering that their compositions significantly differ and that CD9 and CD81 are abundant in both structures.

Subsequently, they studied the role of CD9 and CD81 in the formation of TNTs by using SH-SY5Y cells, first confirming their presence in TNTs via immunofluorescence. CD9 knockout (KO) cells, but not CD81 KO, exhibited a reduced percentage of cells connected via TNTs. The percentage of TNT-connected double KO cells was even lower compared to CD9 KO cells. Additionally, CD9 overexpression (OE), but not CD81 OE increased the percentage of TNT-connected cells.

The authors then investigated the influence of CD9 and CD81 on the capacity of cells to transport material through TNTs by quantifying vesicle delivery between cells. The percentage of acceptor cells containing vesicles (I call it here the efficiency of vesicle transfer) was reduced in CD9 KO cells and CD81 KO cells, and even lower in double KO cells. CD9 OE or CD81 OE increased vesicle transfer efficiency.

Then, they studied possible redundant or complementary roles in the formation of TNTs through a combination of KO and OE of CD9 and CD81 and observed that CD81 does not play any role in TNT formation when CD9 is present, and vesicle transfer of CD81 KO cells can be efficient in CD9 OE conditions.

Incubation of WT cells and CD81 KO cells with an anti-CD9 monoclonal antibody caused CD9 and CD81 clustering, significantly increasing the percentage of TNT-connected cells and duration of TNTs. While the antibody enhanced vesicle transfer efficiency in WT cells, it did not affect vesicle transfer in CD81 KO cells.

The article is well-written and addresses an important biological question, providing some insightful results. However, I have concerns regarding the connection between the experimental data and some of the conclusions drawn by the authors. Below I summarize my points:

- The protocol used to separate TNTs from EVPs and the cell body to determine their protein composition appears problematic. The authors apply mechanical stress by vigorously shaking the samples to achieve this separation. I am skeptical that this method robustly isolates TNTs from other cellular structures/components. I am concerned that their proteomic analysis might not be analyzing the composition of TNTs exclusively, but rather a mixture that includes other structures. For example, the second and eighth most abundant proteins identified are histones (Table S1), and about 20% of the total TNT proteins identified are either mitochondrial or nuclear proteins. The authors should attempt to improve the proteomics section of their study. To differentiate structural TNT proteins from debris, the authors could use statistical analysis to compare multiple independent preparations. Structural TNT proteins will likely be consistently present across all preparations, while non-structural TNT proteins may not. If this approach proves ineffective, the authors might need to refine their TNT isolation procedure.

- Throughout the whole manuscript, the authors quantify the percentage of cells connected by TNTs but do not provide data on the total number of TNTs, which would offer additional valuable information not captured by the percentage of TNT-connected cells alone.

- To study TNT functionality, the authors quantified the efficiency of vesicle transfer by calculating the percentage of acceptor cells containing donor vesicles. How was this percentage computed? The actual number of vesicles delivered to acceptor cells would provide a more accurate metric of vesicle transfer efficiency.

- In Figure 7D, the authors provide a working model. They claim that CD9 KO cells are incapable of forming TNTs. However, this is not supported by their data. The percentage of TNT-connected cells in CD9 KO cells is only slightly lower than in WT cells (Figure 3C).

- In the abstract and discussion of Figure 7D, the authors also claim that CD81 is necessary for the functional transfer of vesicles through TNTs by regulating membrane docking/fusion with the opposing cell. Furthermore, they propose in the discussion section that CD81 is involved in the opening of the TNT. However, all these claims are purely speculative and not supported by their data. If CD81 played such a role, vesicles would accumulate at the tip of the TNTs, which does not appear to be the case. Vesicle transfer occurs in CD81 KO cells. Additionally, TNT formation and efficient vesicle transfer are observed in CD81 KO cells and CD9 OE conditions, suggesting that docking/fusion is not dependent on CD81. Can the authors justify their claims? It is possible that CD81 KO cells might form TNTs with smaller diameters, potentially hindering vesicle transfer. Quantifying the dependence of TNT diameter on CD81 and CD9 expression would address this hypothesis.

- The authors should explain the implications of their study. They need to elaborate on how their findings could impact our understanding of cellular communication and potential applications in therapeutic strategies.

- Tetraspanins are involved in cell migration. In the CRISPR knockout experiments, could the observed changes in the percentage of TNT-connected cells be attributed to variations in cell migration potential?

- The reason behind the clustering of CD9 and CD81 after CD9 antibody treatment should be discussed.