Author response:

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

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. 

We agree with the reviewer that access to both states requires a conformational change in the voltage sensor. This was stated in our revised article: “In contrast, to enter O2, all subunits must complete both voltage sensor transitions and the collective gating ring transition.” We interpret the two gating steps as sequential; the effective rotation of the intracellular ring would happen only once the sensor is in its fully activated position.

We also agree that the S4-S5 segment cannot be the only interaction mechanism, as we demonstrated in our earlier work (Lörinczi et al., 2015; Tomczak et al., 2017).  

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.

We observed the activity of a channel compatible with Kv10.1 ΔPAS-Cap (long openings at low-moderate potentials, very short flickery activity at strong depolarizations) in 12 patches from oocytes obtained from different frog operations over a period of two and a half months once the experimental conditions could be established. As stated in the text, we did not proceed to generate amplitude histograms because we could not resolve clear single-channel events at strong depolarizations. Astemizole abolished the activity and (remarkably) strongly reduced the noise in traces at strong depolarizations, which we interpret as partially caused by flicker openings.

Author response image 1.

We include two example recordings of Astemizole application (100µM) on two different patches. Both recordings are performed at -60 mV (to decrease the likelihood that the channel visits O2) with 100 mM internal and 60 mM external K+. In both cases, the traces in Astemizole are presented in red.

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 nonspecific 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.

In essence, we agree with the reviewer; we already have addressed these points in our revised article:

Regarding the voltage dependence we write “the κ/λ transition could reasonably be expected to be voltage independent because we related it to ring reconfiguration, a process that should occur as a consequence of a prior VSD transition. We have made some attempts to treat this transition as voltage independent but state-specific with upper-layer bias for states on the right and lower-layer bias for states on the left. This is in principle possible, as can already be gleaned from the similar voltage ranges of the left-right transition (α/β) and the κL/λ transition. However, this approach leads to a much larger number of free, less well constrained kinetic parameters and drastically complicated the parameter search. ” As you can see, we also formulated a strategy to free the model of the potentially spurious voltage dependence and (in bold here) explained why we did not follow this route in this study. 

Regarding the need for gating ring transitions after O1, we wrote, “Thus, the underlying gating events can be separated into two steps: The first gating step involves only the voltage sensor without engaging the ring and leads to a pre-open state, which is non-conducting in the WT but conducting in our mutants. The second gating event operates at higher depolarizations, involves a change in the ring, and leads to an open state both in WT and in the mutants. ” 

We interpret your statements such that you expect the conducting state to remain available once O1 is reached. However, the experimental evidence speaks against that the pore availability remains regardless of the further gating steps beyond O1. The description of model construction is informative here: “... we could exclude many possible [sites at which O1 connects to closed states] because the attachment site must be sufficiently far away from the conventional open state [O2]. Otherwise, the transition from "O1 preferred" to "O2 preferred" via a few closed intermediate states is very gradual and never produces the biphasic GV curves [that we observed]. ” 

In other words, voltage-dependent gating steps beyond the state that offers access to O1 appear to close the pore, after it was open. That might occur because only then (for states in which at least one voltage sensor exceeded the intermediate position) the ring is fixed in a particular state until all sensors completed activation. In the WT, closing the pore in deactivated states might rely on an interaction that is absent in the mutant because, at least in HERG: “the interaction between the PAS domain and the C-terminus is more stable in closed than in open KV11.1 (HERG) channels, and a single chain antibody binding to the interface between PAS domain and CNBHD can access its epitope in open but not in closed channels, strongly supporting a change in conformation of the ring during gating ”

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 CryoEM 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 sitedirected 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. 

We respectfully disagree. We think that your suggestions for alternative explanations are addressed in the current version of the article. We will rebut them once more below, but we feel the need to point out that our arguments are already laid out in the revised article.

I have some concerns related to the following points:

(1) Biphasic gating behavior

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. 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. 

We respectfully disagree. In contrast to other channels, like HERG, a common practice for Kv10 is not to use tail currents. It is long known that in this channel, tail currents and test-pulse steady-state currents can appear to be at odds because the channels deactivate extremely rapidly, at the border of temporal resolution of the measurements and with intricate waveforms. This complicates the estimation of the instantaneous tail current. Therefore, the outward current is commonly used to estimate conductance (Terlau et al., 1996; Schönherr et al., 1999; Schönherr et al., 2002; Whicher and MacKinnon, 2019), while the latter authors also use the extreme of the tail for some mutants.

Due to their activation at very negative voltage, the reversal potential in our mutants can be measured directly; we are, therefore, more confident with this approach. Nevertheless, we have determined the initial tail current in some experiments. The behavior of these is very similar to the average that we present in Figure 1. The biphasic behavior is unequivocally present.

Author response image 2.

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 referee’s observation is perfectly in line with the long-standing experience of several labs working with KV10: tail current amplitudes in KV10 appear to be out of proportion for the WT open state (O2). Importantly, this is due to the rapid closure, which is not present in O1. As a consequence, the initial amplitude of tail currents from O1 are easier to estimate correctly, and they are much more obvious in the graphs. Taken together, these differences between O1 and O2 explain the misconception the reviewer describes next.

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. 

As stated in the text and indicated in Author response image 3, the tail currents In Figure 1E increase in amplitude between +50 and +120 mV, as can be seen in the examples below from different experiments (+50 is presented in black, +120 in red). As stated above, the increase is not as evident as in traces from other mutants because the predominance of O2 also implies a much faster deactivation.

Author response image 3. 

We are aware that Ca2+-activated Cl- currents can represent a problem when interpreting electrophysiological data in oocytes. In fact, we show in Supplement 1 to Figure 8 that this can be the case during the Ca2+-CaM experiments, where the increase in Ca2+ would certainly augment Cl- contribution to the outward current. This is why we performed these experiments in Cl--free solutions. As we show in Figure 8, the biphasic behavior was also present in those experiments. 

Importantly, Cl- free bath solutions would not correct contamination during the tail, since this would correspond to Cl- exiting the oocyte. Yet, if there would be contamination of the outward currents by Cl-, one would expect it to increase with larger depolarizations as the typical Ca2+activated Cl- current in oocytes does. As the reviewer states, this does not seem to be the case.

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. 

We agree that this second component continues to increase; the reviewer brought this up in the first review, and we have already addressed this in our reply and in the discussion of the revised version: “This flicker block might also offer an explanation for a feature of the mutant channels, that is not explained in the current model version: the continued increase in current amplitude, hundreds of milliseconds into a strong depolarization (Supp. 4 to Fig. 9). If the relative stability of O2 and C2 continued to change throughout depolarization, such a current creep-up could be reproduced. However, this would require either the introduction of further layers of On ↔Cn states, or a non-Markovian modification of the model’s time evolution.” With non-Markovian, we mean a Langevin-type diffusive process. 

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.

For the readers of this conversation, we would like to clarify that the reviewer likely refers to experiments shown in Fig. 5 of the initial submission but shown in Fig. 6 of the revised version (“Hyperpolarization promotes access to a large conductance, slowly activating open state.” Fig. 5 deals with single channels). We agree that these data look different, but this is because the voltage protocols are completely different (compare Fig. 6A (fixed test pulse, varied prepulse) and Fig. 2A (varied test pulse, fixed pre-pulse). Therefore, no biphasic behavior is expected. 

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.

In summary, we respectfully disagree with all concerns raised in point (1). Our detailed arguments rebutting them are given above, but there is a more high-level concern about this entire exchange: the referee casts doubt on observations that are not new. Several labs have reported for a group of mutant KCNH channels: non-monotonic voltage dependence of activation (see, e.g., Fig. 6D in Zhao et al., 2017), multi-phasic tail currents (see e.g. Fig. 4A in Whicher and MacKinnon, 2019, in CHO cells where Cl- contamination is not a concern), and activation by high [Ca2+]i (Lörinczi et al., 2016). Our study replicates those observations and hypothesizes that the existence of an additional conducting state can alone explain all previously unexplained observations. We highlight the potency of this hypothesis with a Markov model that qualitatively reproduces all phenomena. We not only factually disagree with the individual points raised, but we also think that they don't touch on the core of our contribution

(2) Two step gating mechanism.

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.

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 voltagedependence of the second component that will suggest an implication of a voltage-sensing current.

We do not suggest that the gating ring motion is not voltage dependent. We would like to point out that voltage dependence can be conveyed by voltage sensor coupling to the ring; this is the widely accepted theory of how the ring can be involved. Should the reviewer mean it in a narrow sense, that the model should be constructed such that all voltage-dependent steps occur before and independently of ring reconfiguration and that only then an additional step that reflects the (voltage-independent) reconfiguration solely, we would like to point the reviewer to the article, where we write: “the κ/λ transition could reasonably be expected to be voltage independent because we related it to ring reconfiguration, a process that should occur as a consequence of a prior VSD transition. We have made some attempts to treat this transition as voltage independent but state-specific with upper-layer bias for states on the right and lower-layer bias for states on the left. This is in principle possible, as can already be gleaned from the similar voltage ranges of the left-right transition (α/β) and the κL/λ transition. However, this approach leads to a much larger number of free, less well constrained kinetic parameters and drastically complicated the parameter search. ” As you can see, we also formulated a strategy to free the model from the potentially spurious voltage dependence and (in bold here) explained why we did not follow this route in this study. 

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.

As stated in the panel heading, the figure legend, and the main text, we did not use 1-341 and 342-end as done in Tomczak et al. Instead, “we compared the behavior of ∆2-10 and ∆210.L341Split,”. Evidently, the additional deletion (2-10) causes a shift in activation that explains the difference you point out. However, as we do not compare L341Split and ∆210.L341Split but ∆2-10 and ∆2-10.L341Split, our conclusion remains that “As predicted, compared to ∆2-10, ∆2-10.L341Split showed a significant reduction in the first component of the biphasic GV (Fig. 2C, D).” Remarkably, the behavior of the ∆3-9 L341Split described in Whicher and MacKinnon, 2019 (Figure 5) matches that of our ∆2-10 L341Split, which we think reinforces our case.

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). 

There is a possibility that we are debating nomenclature here. In response to the suggestion that all our observations could be explained by inactivation, we attempted a disambiguation of terms in the reply and the article. As the argument is brought up again without reference to our clarification attempts, we will try to be more explicit here:

If, starting from deeply deactivated states, an open state is reached first, and then, following further activation steps, closed states are reached, this might be termed “inactivation”. In such a reading, our model features many inactivated states. The shortest version of such a model is C-O-I. It is for instance used by Raman and Bean (2001; DOI: 10.1016/S00063495(01)76052-3) to explain NaV gating in Purkinje neurons. If “inactivation” is meant in the sense that a gating transition exists, which is orthogonal to an activation/deactivation axis, and that after this orthogonal transition, an open state cannot be reached anymore, then all of the upper floor in our model is inactivated with respect to the open state O1. Finally, the state C2 is an inactivated state to O2. In this view, “inactivation” explains the observed phenomena. 

However, we must disagree if the referee means that a parsimonious explanation exists in which a single conducting state is the only source for all observed currents.   

There is a high-level reason: we found a single assumption that explains three different phenomena, while the inactivation hypothesis with one conducting state cannot explain one of them (the increase of the first component under raised CaM). But there is also a low-level reason: the tails in Tristani-Firouzi and Sanguinetti 1998 are fundamentally different from what we report herein in that they lack a third component. Thus, those tails are consistent with recovery from inactivation through a single open state, while a three-component tail is not. In the framework of a Markov model, the time constants of transitions from and to a given state (say O2), cannot change unless the voltage changes. During the tail current, the voltage does not change, yet we observe: 

i) a rapid decrease with a time constant of at most a few milliseconds (Fig 9 S2, 1-> 2),  ii) a slow increase in current, peaking after approximately 25 milliseconds and iii) a relaxation to zero current with a time constant of >50 ms. 

According to the reviewer’s suggestion, these processes on three timescales should all be explained by depopulating and repopulating the same open state while all rates are constant. There might well be a complicated multi-level state diagram with a single open state with different variants, like (open and open inactivated) that could produce triphasic tails with these properties if the system had not reached a steady state distribution at the end of the test pulse. It cannot, however, achieve it from an equilibrated system, and certainly, it cannot at the same time produce “biphasic activation” and “activation by CaM”. 

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. 

Again, we disagree. The model construction explains in detail that the transition from the first to the second phase is not gradual. Shifting equilibria cannot reproduce this. We have extensively tested that idea and can exclude this possibility.

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). 

We respectfully disagree. We explicitly model an open state that inactivates faster (O2->C2) than it activates. Once more, this is stated in the revised article, which we point to for details. Again, this alternative mechanism does not have the potential to explain all three effects. As discussed above about the chloride contamination concerns, this inactivation hypothesis was mentioned in the first review round and, therefore, addressed in our reply and the revised article. We also explained that “inactivation” has no specific meaning in Markov models. In the absence of O1, all transitions towards the lower layer are effectively “inactivation from closed states”, because they make access to the only remaining open state less likely”. But this is semantics. What is relevant is that no network of states around a single open state can reproduce the three effets in a more parsimonious way than the assumption of the second open state does.

(3) Single channel conductance.

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. 

We respectfully disagree. As previously shown, the conductance of the flickering wild-type open state is very difficult to resolve. Our recordings do not show that the two states have different single-channel conductances, and therefore the model assumes identical singlechannel conductance. 

The important point is that the single-channel recordings clearly show two different gating modes associated with the voltage ranges in which we predict the two open states. One has a smaller macroscopic current due to rapid flickering (aka “inactivation”). These recordings are another proof of the existence of two open states because the two gating modes occur.  Wild-type data can be found in Bauer and Schwarz, (2001, doi:10.1007/s00232-001-0031-3) or Pardo et al., (1998, doi:10.1083/jcb.143.3.767) for comparison.

We appreciate the effort editors and reviewers invested in assessing the revised manuscript. Yet, we think that the demanded revision of experimental conditions and quantification methods contradicts the commonly accepted practice for KV10 channels. Some of the reviewer comments are skeptical about the biphasic behavior, which is an established and replicated finding for many mutants and by many researchers. The alternative explanations for these disbelieved findings are either “semantics” or cannot quantitatively explain the measurements. Therefore, only the demand for more explanations and unprecedented resolution in singlechannel recordings remains. We share these sentiments.

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

(1) The authors must show that the second open state is not just an artifact of endogenous activity but represents the activity of the same EAG channels. I suggest that the authors repeat these experiments in Mes-based solutions. 

(2) Along the same lines, it is necessary to show that these currents can be blocked using known EAG channel blockers such as astemizole. Ultimately, it will be important to demonstrate using single-channel analysis that these do represent two distinct open states separated by a closed state. 

We have addressed these concerns using several approaches. The most substantial change is the addition of single-channel recordings on ΔPASCap. In those experiments, we could provide evidence of the two types of events in the same patch, and the presence of an outward current at -60 mV, 50 mV below the equilibrium potential for chloride. The channels were never detected in uninjected oocytes, and Astemizole silenced the activity in patches containing multiple channels. These observations, together with the maintenance of the biphasic behavior that we interpret as evidence of the presence of O1 in methanesulfonate-based solutions, strongly suggest that both O1 and O2 obey the expression of KV10.1 mutants.

(3) Currents should be measured by increasing the pulse lengths as needed in order to obtain the true steady-state G-V curves. 

We agree that the endpoint of activation is ill-defined in the cases where a steady-state is not reached. This does indeed hamper quantitative statements about the relative amplitude of the two components. However, while the overall shape does change, its position (voltage dependence) would not be affected by this shortcoming. The data, therefore, supports the claim of the “existence of mutant-specific O1 and its equal voltage dependence across mutants.”

(4) A more clear and thorough description should be provided for how the observations with the mutant channels apply to the behavior of WT channels. How exactly does state O1 relate to WT behavior, and how exactly do the parameters of the mathematical model differ between WT and mutants? How can this be interpreted at a structural level? What could be the structural mechanism through which ΔPASCap and E600R enable conduction through O1? It seems contradictory that O1 would be associated exclusively with voltage-sensor activation and not gating ring transitions, and yet the mutations that enable cation access through O1 localize at the gating ring - this needs to be better clarified. 

We have undertaken a thorough rewriting of all sections to clarify the structural correlates that may explain the behavior of the mutants. In brief, we propose that when all four voltage sensors move towards the extracellular side, the intracellular ring maintains the permeation path closed until it rotates. If the ring is altered, this “lock” is incompetent, and permeation can be detected (page 34). By fixing the position of the ring, calmodulin would preclude permeation in the WT and promote the population of O1 in the mutants.

(5) Rather than the t80% risetime, exponential fits should be performed to assess the kinetics of activation. 

We agree that the assessment of kinetics by a t80% is not ideal. We originally refrained from exponential fits because they introduce other issues when used for processes that are not truly exponential (as is the case here). We had planned to perform exponential fits in this revised version, but because the activation process is not exponential, the time constants we could provide would not be accurate, and the result would remain qualitative as it is now. In the experiments where we did perform the fits (Fig. 3), the values obtained support the statement made. 

(6) It is argued based on the G-V relations in Figure 2A that none of the mutations or deletions introduced have a major effect on state O1 properties, but rather affect state O2. However, the occupancy of state O2 is undetermined because activation curves do not reach saturation. It would be interesting to explore the fitting parameters on Fig.2B further to test whether the data on Fig 2A can indeed only be described by fits in which the parameters for O1 remain unchanged between constructs. 

We agree that the absolute occupancy of O2 cannot be properly determined if a steady state is not reached. This is, however, a feature of the channel. During very long depolarizations in WT, the current visually appears to reach a plateau, but a closer look reveals that the current keeps increasing after very long depolarizations (up to 10 seconds; see, e.g., Fig. 1B in Garg et al., 2013, Mol Pharmacol 83, 805-813. DOI: 10.1124/mol.112.084384). Interestingly, although the model presented here does not account for this behavior, we propose changes in the model that could. “If the relative stability of O2 and C2 continued to change throughout the depolarization such a current creep-up could be reproduced. However, this would require either the introduction of further layers of On↔Cn states or a non-Markovian modification of the model’s evolution.” Page 34.

(7) The authors interpret the results obtained with the mutants DPASCAP and E600R -tested before by Lorinczi et al. 2016, to disrupt the interactions between the PASCap and cNBHD domains- as a two-step gating mechanism with two open states. All the results obtained with the E600R mutant and DPASCap could also be explained by inactivation/recovery from inactivation behavior and a change in the equilibrium between the closed states closed/inactivated states and open states. Moreover, the small tails between +90 to +120 mV suggest channels accumulate in an inactive state (Fig 1E). It is not convincing that the two open-state model is the mechanism underlying the mutant's behavior.  

We respectfully disagree with the notion that a single open state can provide a plausible explanation for "All the results obtained with the E600R mutant and DPASCap". We think that our new single channel results settle the question, but even without this direct evidence, a quantitative assessment of the triphasic tail currents all but excludes the possibility of a single open state. We agree that it is, in principle, possible to obtain some form of a multiphasic tail with a single open state using the scheme suggested in this comment: at the end of the test pulse, a large fraction of the channels must be accumulated in inactive states, and a few are in the open state. The hyperpolarization to -100mV then induces a rapid depopulation of the open state, followed by slower replenishments from the inactive state. Exactly this process occurs in our model, when C2 empties through O2 (Supp. 5 to Fig 9, E600R model variant). However, this alone is highly unlikely to quantitatively explain the measured tail currents, because of the drastically different time scales of the initial current decay (submillisecond to at most a few milliseconds lifetime) and the much slower transient increase in current (several tens of milliseconds) and the final decay with time constants of >100 ms (see for instance data in Fig. 1 E for E600R +50 to +120mV test pulse). To sustain the substantial magnitude of slowly decaying current by slow replenishment of an open state with a lifetime of 1 ms requires vast amounts of inactivated channels. A rough estimation based on the current integral of the initial decay and the current integral of the slowly decaying current suggests that at the end of the test pulse, the ratio inactivated/open channels would have to be 500 to 1500 for this mechanism to quantitatively explain the observed tail currents. To put this in perspective: This would suggest that without inactivation all the expressed channels in an oocyte would provide 6 mA current during the +100 mV test pulse. While theoretically possible, we consider this a less likely explanation than a second open state.

(8) Different models should be evaluated to establish whether the results in Figure 4 can also be explained by a model in which states O1 and O2 have the same conductance. It would be desirable if the conductance of both states were experimentally determined - noise analysis could be applied to estimate the conductance of both states. 

In the modified model, O1 and O2 have the same single-channel conductance. The small conductance combined with the fast flickering did not allow an accurate determination, but we can state that there is no evidence that the single-channel conductance of the states is different.

(9) Although not included, it looks like the model predicts some "conventional inactivation" This can be appreciated in Fig 8, and in the traces at -60mV. Interestingly, the traces obtained in the absence of Cl- also undergo slow inactivation, or 'conventional inactivation' as referred to by the authors. Please revise the following statement "Conventional inactivation was never detected in any mutants after repeated or prolonged depolarization. In the absence of inactivation, the pre-pulse dependent current increase at +40 mV could be related to changes in the relative occupancy of the open states". 

We have carefully edited the manuscript to address this concern. The use of the term inactivation admittedly represents a challenge. We agree that the state that results from the flickering block (C2) could be defined as “inactivated” because it is preceded by an open state. Yet, in that case, the intermediate states that the channel travels between O1 and O2 would also be sensu stricto “inactivated”, but only in the mutants. We have made this clear in page 17.

Recommendations for improving the writing and presentation.

(1) Methods section: Please state the reversal potential calculated for the solution used. It looks like the authors used an Instantaneous I-V curve method to calculate the reversal potential; if that's correct, please show the I-V and the traces together with the protocol used. 

We have provided the calculated reversal potentials for excised patches. We cannot predict the reversal potential in whole oocytes because we have no control over the intracellular solution. The reversal potential was determined in the mutants through the current at the end of the stimulus because the mutants produced measurable inward currents. The differences in reversal potential were not significant among mutants.

Pulse protocols have been added to the figures.

(2) Figure 1 suggestion: Combine the two panels in panel D and move the F panel up so the figure gets aligned in the lower end.

Thank you, this has been done.

(3) Please clarify the rationale for using the E600R-specific mutant. I assume it is based on the Lorinzci et al. 2016 effect and how this is similar to the DPASCap phenotype, or is it due to the impact of this mutation in the interactions between the N-term and the cNBHD? 

We have explained the rationale for the use of E600R explicitly on page 6.

(4) Fig S1A is not present in the current version of the manuscript. Include a cartoon as well as a structural figure clearly depicting the perturbations introduced by E600R, ΔPASCap, and the other deletions that are tested. Additional structural information supporting the discussion would also be helpful to establish clearer mechanistic links between the experimental observations described here and the observed conformational changes between states in Kv10 channel structures. 

We have corrected this omission, thank you for pointing it out.

(5) It would be informative to see the traces corresponding to the I-V shown in Fig 7 A and B at the same indicated time points (0, 60, 150, and 300s). Did the authors monitor the Ca2+ signal rise after the I&T treatment to see if it coincides with the peak in the 60s? 

In Figure 7 (now Figure 8) we used voltage ramps instead of discrete I-V protocols because of the long time required for recording the latter. This is stated on page 19. Ca2+ was monitored through Cl- current after ionomycin/thapsigargin. The duration of the Ca2+ increase was reproducible among oocytes and in good agreement with the changes observed in the biphasic behavior of the mutants (Supplement 1 to Figure 8).

(6) Fig 4. Please state in the legend what the different color traces correspond to in E600R and DPASCap. Is there a reason to change the interpulse on DPASCap to -20mV and not allow this mutant to close? Please state. How do the authors decide the 10 ms interval for the experiments in Fig 2? 

Thank you for pointing this out, we have added the description. We have explained why we use a different protocol for ΔPASCap and the reason for using 10 ms interval (we believe the referee means Figure 4) on page 12.  

(7) Fig. 5. Since the pre-pulse is supposed to be 5s, but the time scale doesn't correspond with a pre-pulse of 5 s before the test pulse to +40mV. Has the pre-pulse been trimmed for representation purposes? If so, please state. 

The pre-pulse was 5s, but as the reviewer correctly supposed, the trace is trimmed to keep the +40 mV stimulus visible. This has now been clearly stated in the legend.

(8) The mutant L322H is located within the S4 helix according to the Kv10.1 structure (PDB 5K7L), not in the 'S3-S4 linker'; please correct. 

This has been done, thank you.

The introduction of this mutant should also shift the voltage dependence toward more hyperpolarizing potentials (around 30mV, according to Schoenherr et al. 1999). It looks like that shift is present within the first component of the G-V. Still, since the max amplitude from the second component could be contaminated by endogenous Cl- currents, this effect is minimized. Repeating these experiments in the no Cl- solutions will help clarify this point and see the effect of the DPASCap and E600R in the background of a mutation that accelerates the transitions between the closed states (see Major comment 1). Did the authors record L322H alone for control purposes? 

We have decided not to measure L322H alone or repeat the measurements in Cl--free solutions because we do not see a way to use the quantitative assessment of the voltage dependence of L322H and the L322H-variants of the eag domain mutants. Like in our answer to main point 3, we base our arguments not on the precise voltage dependence of the second component but on the shape of the G-V curves instead, specifically the consistent appearance of the first component and the local conductance minimum between the first and second components. After the introduction of L322H the first component is essentially absent.

We think that the measurements of the L322H mutants cannot be interpreted as a hyperpolarizing shift in the first component. The peak of the first conductance component occurs around -20 mV in ΔPASCap and E600R (Fig. 7 C, D). After a -30mV shift, in L322H+DPASCap and L322H+E600R, this first peak would still be detected within the voltage range in our experiments, but it is not. A contamination of the second component would have little impact on this observation, which is why we refrain from the suggested measurements.  

(9) The authors differentiate between an O1 vs. O2 state with different conductances, and maybe I missed it, but there's no quantitative distinction between the components; how are they different?

Please see the response to the main comments 1 and 2. This has been addressed in singlechannel recordings.

(10) Please state the voltage protocols, holding voltages, and the solutions (K+ concentration and Cl-presence/absence) used for the experiments presented in the legends on the figures. Hence, it's easier to interpret the experiments presented. 

Thank you, this has been done.

(11) The authors state on page 7 that "with further depolarizations, the conductance initially declined to rise again in response to strong depolarizations. This finding matches the changes in amplitude of the tail currents, which, therefore, probably reflect a true change in conductance" However, the tails in the strong voltage range (+50 to +120 mV) for the E600R mutant argue against this result. Please review.

The increase in the amplitude of the tail current is also present in E600R, but the relative increase is smaller. We have decided against rescaling these traces because the Figure is already rather complex. We indicated this fact with a smaller arrow and clarified it in the text (page 8).

(12) The authors mention that the threshold of activation for the WT is around -20mV; however, the foot of the G-V is more around -30 or -40mV. Please revise. 

Thank you. We have done this. 

(13) The authors state on page 9 that the 'second component occurs at progressively more depolarized potentials for increasingly larger N-terminal deletions" However E600R mutant that conserves the N-terminal intact has a shift as pronounced as the DPASCap and larger than the D2-10. How do the authors interpret this result? 

We have corrected this statement in page 10 : “…the second component occurs at progressively more depolarized potentials for increasingly larger N-terminal deletions and when the structure of the ring is altered through disruption of the interaction between N- and C-termini (E600R)”.

(14) The equation defined to fit the G-Vs, can also be used to describe the WT currents. If the O1 is conserved and present in the WT, this equation should also fit the WT data properly. The 1-W component shown could also be interpreted as an inactivating component that, in the WT, shifts the voltage-dependence of activation towards depolarizing potentials and is not visible. Still, the mutants do show it as if the transition from closed-inactivated states is controlled by interactions in the gating ring, and disturbing them does affect the transitions to the open state. 

Out of the two open states in the mutant, O2 is the one that shares properties with the WT (e.g. it is inaccessible during Ca2+-CaM binding) while O1 is the open state with the voltage dependence that is conserved across the mutants. We, therefore, believe that this question is based on a mix-up of the two open states. We appreciate the core of the question: does the pattern in the mutants’ G-V curves find a continuation in the WT channel? 

Firstly, the component that is conserved among mutants does not lead to current in the WT because the corresponding open state (O1) is not observed in WT. However, the gating event represented by this component should also occur in WT and –given its apparent insensitivity to eag domain mutations–  this gating step should occur in WT with the same voltage dependence as in all the mutants. This means that this first component sets a hard boundary for the most hyperpolarized G-V curve we can expect in the WT, based on our mutant measurements. Secondly, the second component shows a regular progression across mutants: The more intact the eag domain is, the more hyperpolarized the Vhalf values of transition term (1-W) and O2 activation. In Δ2-10, the transition term already almost coincides with O1 activation (estimated Vhalf values of -33.57 and -33.47 mV). A further shift of (1-W) in the WT is implausible because, if O1 activation is coupled to the earliest VSD displacement, the transition should not occur before O1 activation. Still, the second component might shift to more hyperpolarized values in the WT, depending on the impact of amino acids 2 to 10 on the second VSD transition.

In summary, in WT the G-V should not be more hyperpolarized than the first component of the mutants, and the (1-W)-component probably corresponds to the Δ2-10 (1-W)-component. In WT the second component should be no more depolarized than the second component of Δ2-10. The WT G-V (Fig.1B) meets all these predictions derived from the pattern in the mutant GVs: When we use Eq. 4 to fit the WT G-V with A1=0 (O1 is not present in WT) and the parameters of the transition term (1-W)  fixed to the values attained in Δ2-10, we obtain a fit for the O2 component with Vhalf\=+21mV. This value nicely falls into the succession of Vhalf values for Δeag, ΔPASCap, and Δ2-10 (+103mV,+80mV,+52mV) and, at the same time, it is not more hyperpolarized than the conserved first component (Vhalf -34mV). Our measurements therefore support that the O2 component in the mutants corresponds to the single open state in the WT. 

(15) Page 15, the authors state that 'The changes in amplitude and kinetics in response to rising intracellular Ca2+ support our hypothesis that Ca-CaM stabilized O1, possibly by driving the channels to deep closed states (Fig 5 and 6)' (pg 15). This statement seems contradictory; I can't quite follow the rationale since Ca2+ potentiates the current (Fig 7), and the addition of the L322H mutant in Fig 7 makes the shift of the first component to negative potentials visible.

Please check the rationale for this section. 

We have explained this more explicitly in the discussion (page 32). “Because access to O1 occurs from deep closed states, this could be explained by an increased occupancy of such deactivated states in response to CaM binding. This appears to be the case since CaM induces a biphasic behavior in the mutant channels that show reduced access to deep closed states; thus, L322H mutants behave like the parental variants in the presence of Ca2+-CaM. This implies a mechanistic explanation for the effect of Ca2+-CaM on WT since favoring entry into deep closed states would result in a decrease in current amplitude in the absence of (a permeable) O1”.

Also, Figs 5 and 6 seem miscited here. 

Thank you, we have corrected this.

(16) For Figure 5, it would be helpful if each of the current traces corresponding to a particular voltage had a different color. That way, it will be easier to see how the initial holding voltage modulates current. 

We have considered this suggestion, and we agree that it would make it easier to follow. Yet, since we have identified the mutants with different colors, it would be inconsistent if we used another color palette for this Figure. Supplement 3 to Figure 9 shows the differences in a clearer way.

(17) Add zero-current levels to all current traces.

We have done this.

(18) The mathematical model should be described better. Particularly, the states from which O1 can be accessed should be described more clearly, as well as whether the model considers any direct connectivity between states O1 and O2. The origin of the voltage-dependence for transitions that do not involve voltage-sensor movements should be discussed. Also, it separation of kappa into kappa-l and kappa-r should be described. 

We have extensively rewritten the description of the mathematical model to address these concerns.

(19) Page 4, "reveals a pre-open state in which the transmembrane regions of the channel are compatible with ion permeation, but is still a nonconducting state". Also, page 27, "renders a hydrophobic constriction wider than 8 Å, enough to allow K+ flow, but still corresponds to a non-conducting state". These sentences are confusing - how can the regions be compatible with ion permeation, and still not be conducting? Is cation conductance precluded by a change in the filter, or elsewhere? How is it established that it represents a non-conducting state? 

We have rephrased to clarify this apparent inconsistence. Page 4: “(…) in which the transmembrane regions of the channel are compatible with ion permeation (the permeation path is dilated, like in open states) but the intracellular gate is still in the same conformation as in closed states (Zhang et al., 2023).” Page 31: “The presence of an intact intracellular ring would preclude ionic flow in the WT, and its alteration would explain the permeability of this state in the mutants.”

Author response:

We thank the reviewers for their thorough comments on our manuscript. We appreciate their recognition of the strengths in our study, including addressing the significant problem of neonatal sepsis in preterm infants using a preterm piglet model, the robustness of our multi-omics dataset, and our multi-pronged approach to examining the physiological changes under different glucose management regimens.

This document addresses our initial responses to the main concerns of the 3 reviewers. We will provide more detailed responses to their comments and revise the manuscript at a later date.

In response to Reviewer #1, we acknowledge the concern about high blood glucose levels in the control group. This work is a follow-up from our previous work (Muk et al, JCI insight 2022) where we explored different PN glucose regimens. Taken together, our experiments suggest a linear relationship between glucose provision and infection severity, indicating increased glucose may heighten mortality risk, while radical reduction could reduce mortality due to sepsis, but cause hypoglycemia and brain damage. As for the discrepancy in survival rates between Figures 1B and 6B, this is due to a shortened follow-up time in the follow-up experiment. This was done to minimize animal suffering because relevant differences in immune-responses were detectable within 12 hours in the primary experiment. As for the relationship between bacterial burdens and glucose, we agree that lower bacterial density in piglets receiving the reduced glucose PN may result from slower bacterial growth. However, we analyzed the relationship between bacterial burdens and mortality and found that it did not correlate within each of the treatment groups. This finding inspired us to further explore the relationship between bacterial burdens and infection responses in our model which has resulted in our recent preprint: Wu et at. Regulation of host metabolism and defense strategies to survive neonatal infection. BioRxiv 2024.02.23.581534; doi: https://doi.org/10.1101/2024.02.23.581534

For Reviewer #2, The distinction between early (EOS) and late onset sepsis (LOS) in the time cut-off makes sense clinically because they are likely to be caused by different organisms and origins (EOS with maternal origin and LOS with postnatal origin) and therefore require different empirical antibiotics regimes. However, it is also important to acknowledge that the pathophysiology of “sepsis” may be similar despite timing and pathogen and depends on the degree of immune activation. Therefore, even though the infection in our model is initiated on the first day after birth the organism that we use, Staphylococcus epidermidis (most common bacteria detected in LOS), makes it a better model for LOS. As for neutrophil specific transcripts, we only collected the whole blood transcript during the experiments, which reflects the transcriptomic profile of all the leucocytes. Since we did not do single cell RNA sequencing during the experiment there is no possibility of isolating the neutrophil transcriptome at this time. As for the question of a “safe glucose infusion rate”, there likely is none as the immune responses to glucose intake do not seem binary but increase with glucose intake. Our reduced glucose PN was chosen as it corresponded with the low end of recommended guidelines for PN glucose intake. However, the reduced glucose intervention still resulted in significant morbidity and a 25% mortality within 22 hours. There is therefore still vast room for improvement, but even though further reduction in PN glucose intake would probably provide further protection, it would entail dangerous hypoglycemia. The findings in this paper have prompted us to explore several alternative strategies to both reduce infection-related mortality and maintain glucose homeostasis. Thus, the optimal PN for infected newborns would probably differ from standard PN in all macronutrients compartments and will require much more pre- and clinical research.

For Reviewer #3, we acknowledge the variability in data collected from animals at euthanasia. These endpoints represent snapshots of the animals' states at euthanasia, which is a clear limitation of our method. Therefore, we do not know what metabolic processes precede the development of lethal sepsis, although the increases in plasma lactate suggest a higher rate of glycolysis in animals on high glucose PN. However, we believe the data still heavily imply a causal relationship between energy metabolic processes, especially glycolytic breakdown of glucose, and the pro-inflammatory responses leading to sepsis. In our recent preprint mentioned above we further explored the metabolic responses in pigs that succumbed to sepsis, compared to those that survived and found that survival was strongly associated with increases in mitochondrial metabolism and reduction in glycolysis.

We hope these clarifications and our commitment to further research address your concerns satisfactorily. Thank you for your valuable feedback.

Author response:

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

Reviewer 1 (Public Review):

“but an obvious influencing factor that the authors could investigate in their own data set is the retinal input. In Fig1b, the authors even show these data in the form of gaze and pupil size. In these example data, by eye, it looks like the pupil size is positively correlated with the run speed. This would of course have large consequences on the activity in V1, but the authors do not do anything with these data. The study would improve substantially if the authors would correlate their run speed traces with other factors that they have recorded too, such as pupil size and gaze.”

Absolutely. We have added a first level of eye movement (and pupil size) analyses to the revised manuscript, resulting in an additional figure. In short, we found that eye movements are unlikely to play a significant role in our primary results, as the patterns of eye movements differed only slightly between running and stationary periods, and the measured impacts of such eye movements were also quantitatively much smaller than the primary effect sizes.

We also note that in analyzing the eye movements, we also found that pupil size was larger during running than stationary. This is suggestive evidence that running is correlated with increases in arousal. Although more work will be needed to calibrate and quantify how much this factor affects neural responses (and perhaps to dissociate it from running per se), the simple analysis we present suggest that the large differences we observe could be explained by a difference between how arousal and running are correlated in the monkey versus the mouse. Instead, it appears that both species have at least qualitatively similar relations between pupil size (a standard proxy for arousal) and running.

On this issue, we have added extensive discussion of the relevant recent work by Talluri et al. (2023) who attempted a similar cross-species analysis that considered spontaneous body movements and their effect on cortical activity (as well as the possibility that eye movements are a critical mediator in these modulations). Due to delays in revising our manuscript, we regret that our earlier submission had not cited this work, but we now do our best to highlight its importance and the synergy between these two papers. The full citation is listed below:

Talluri BC, Kang I, Lazere A, Quinn KR, Kaliss N, Yates JL, Butts DA, Nienborg H. Activity in primate visual cortex is minimally driven by spontaneous movements. Nat Neurosci. 2023 Nov;26(11):1953-1959. doi: 10.1038/s41593-023-01459-5.

There is a finer level of analysis that we hope to do in the future along these lines. It would rely on detailed characterization of each receptive field, building an image-computable model linking those receptive fields to the neural activity, and doing so at a finer time grain that links individual eye movements and changes in the spike train within a stimulus presentation (as opposed to working at the level of spike counts per stimulus presentation). Because these steps need to be accomplished together— and each requires substantial additional work and would go beyond the first-order findings we report in this work— we hope to report on such finer analyses in a standalone paper later. We are working on being able to do this in both marmoset and mouse.

More generally, we want to emphatically agree that what is missing from this paper is the “why?”! We have done our best to show that a fair comparison reveals quantitatively different phenomena in marmoset and mouse. In the revised discussion, we lay out many lines of work that we hope will gain traction on this deeper mechanistic point. There’s a lot to do, and several of the possibilities are already current topics of exploration in our ongoing work.

“Looking at the raster plot, however, shows that this strong positive correlation must be due entirely to the lower half of the neurons significantly increasing their firing rate as the mouse starts to run; in fact, the upper 25% or so of the neurons show exactly the opposite (strong suppression of the neurons as the mouse starts running). It would be more balanced if this heterogeneity in the response is at least mentioned somewhere in the text.”

We are also intrigued by the heterogeneity of effects at the single neuron level. That is why the next section of the paper is dedicated to analyzing effects on a cell-by-cell basis. The fractions of neurons showing either increases or decreases are described separately, to get at this very issue.

Reviewer 2 (Public Review)::

“For example, it is known that the locomotion gain modulation varies with layer in the 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, it was not reported from which cortical layer the neurons are from, leaving this point unanswered.”

Reviewer 2 called for more consideration of details that have been addressed in the mouse literature, such as the cortical layer of the cells, and related aspects of circuitry. We have greatly re-worked the Discussion to address several of these issues. In short, the manuscript’s set of data were collected without strong traction on layers or cell types, and it will be quite interesting to get a better handle on this using both refinements to our recording procedures as well as new techniques that are now possible in the marmoset for future studies.

“In this regard, it is worth noting that the authors report an interesting difference between the foveal and peripheral parts 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 the 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.”

Same page! We have expanded the discussion to better emphasize these points and are already deep in follow up experiments to explore the foveal and peripheral representations.

Reviewer 3 (Public Review)::

“However, the authors did not take full advantage of the quantity and diversity of the marmoset visual cortex recordings in their analyses. They mention recording and analyzing the activity of peripheral V1 neurons but mainly present results involving foveal V1 neurons. Foveal neurons, with their small receptive fields strongly affected by precise eye position, would seem to be less likely to be comparable to rodent data. If the authors have a reason for not doing so, they should provide an explanation.”

We agree, and hope the reviewer finds our overall reply, detailed response to Reviewer 1 (who raised a similar issue), and corresponding updates to the manuscript appropriate for this stage of understanding.

“Given that the marmosets are motivated to run with liquid rewards, the authors should provide more context as to how this may or may not affect marmoset V1 activity. Additionally, the lack of consideration of eye movements or position presents a major absence for the marmoset results, and fails to take advantage of one of the key differences between primate and rodent visual systems - the marmosets have a fovea, and make eye movements that fixate in various locations on the screen during the task.”

In addition to the response above, we have made edits to the manuscript to speak to issues of arousal and eye movements (also detailed in previous responses). Given the modest decrease in activity we see, the usual concerns about potential increases in neural activity related to eye movements (which we quantify in the revision) and other issues related to motivation are hard to specifically relate to existing literature. But in the revised Discussion we talk more about how future work can/should dissociate these factors, as has been done in the mouse literature.

“Finally, the model provides a strong basis for comparison at the level of neuronal populations, but some methodological choices are insufficiently described and may have an impact on interpreting the claims.”

We have also clarified the shared-gain model’s description, which we agree needed additional detail and clarification.

Author response:

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

eLife assessment

This study presents a useful comparison of the dynamic properties of two RNA-binding domains. The data collection and analysis are solid, making excellent use of a suite of NMR methods. However, evidence to support the proposed model linking dynamic behavior to RNA recognition and binding by the tandem domains remains incomplete. The work will be of interest to biophysicists working on RNA-binding proteins.

We thank eLife for taking the time and effort to review our manuscript. Evidence from the literature and our study shows a great deal of parity between the dynamic behavior of dsRBDs and its dsRNA-recognition and -binding that helped us culminate in proposing a fair model. As already mentioned in the manuscript, we have been working on the suggested experiments to support our proposed model further.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In the manuscript entitled "Differential conformational dynamics in two type-A RNA-binding domains drive the double-stranded RNA recognition and binding," 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 compared to dsRBD1 and does so without much alteration in protein dynamics. Using their previously published data, the authors propose a model whereby dsRBD2 recognizes dsRNA first and brings dsRBD1 into proximity to search for RNA bulge and internal loop structures.

We thank the Reviewer for sending us an encouraging review. We have combined the findings reported in the literature with new ones that led us to propose the dsRNA-binding model by tandem A-form dsRBDs.

We propose that dsRBD1 can first recognize a variety of sequential and structurally different dsRNAs. dsRBD2 assists the interaction with a higher affinity, thus fortifying the interaction between TRBP and a possible substrate. This may enable the other associated proteins like Dicer and Ago2 to perform critical biological functions.

However, we feel that a few statements in the comment above are factually incorrect.

Statement 1. “They then show that dsRBD2 binds to canonical A-form dsRNA with a higher affinity compared to dsRBD1 and does so without much alteration in protein dynamics.”

We have explicitly shown the perturbation in dsRBD2 dynamics upon RNA binding.

Statement 2. “Using their previously published data, the authors propose a model whereby dsRBD2 recognizes dsRNA first and brings dsRBD1 into proximity to search for RNA bulge and internal loop structures.”

Our previously published data suggests that dsRBD1, owing to its high conformational dynamics in solution, is able to recognize a variety of structurally and sequentially different dsRNAs ([Paithankar et al., 2022]). dsRBDs preferably bind to the double-stranded region (minor-major-minor-groove) of an A-form RNA ([Acevedo et al., 2016]; [Vuković et al., 2014]) and do not search for bulge and internal loop structures as a part of the binding event. Even though dsRBDs preferably bind to the double-stranded region, they can still accommodate perturbation in the A-form helix due to mismatch and bulges with decreased binding affinity ([Acevedo et al., 2015]). However, it is a matter of future research to identify how much of a deviation from the A-form structure can be accommodated by the dsRBDs. The diffusion event observed in the literature ([Koh et al., 2013]) also does not show any direct implication for searching for bulge and internal loop structures.

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:

The data collection and analysis are sound. The only weakness in the manuscript is the lack of 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. Furthermore, there was no discussion about the entropy of binding derived from ITC. It might be interesting to compare with dynamics from NMR.

We understand the reviewer’s point that this study is focused on a dsRNA-binding mechanism rather than addressing the much broader field of RNA-binding. There are multiple challenges in finding a single mechanism that works for all RNA-binding proteins. For instance, RRM is a single-stranded RNA binding domain that is able to read out the substrate base sequence. RRM behaves entirely differently than the dsRBD in terms of target specificity. Besides, several other RNA-binding domains, like the KH-domain, Puf domains, Zinc finger domains, etc., showcase a unique RNA-binding behavior. Thus, it would be really difficult to draw a single rule of thumb for RNA-recognition behavior for all these diverse domains.

Thank you for pointing out the entropy of binding from ITC. We have now included the entropy of binding discussion in the main text, page 7.

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 the diffusion of TRBP2 along the RNA molecule. To conclusively prove the central claim of this manuscript, the authors could have considered a larger construct that carries both RBDs. With such a construct, authors can probe the characteristics of these two tandem domains (e.g., semi-independent tumbling) and their interactions with the RNA. Additionally, mutational experiments may be carried out where specific residues are altered to change the conformational dynamics of these two domains. The corresponding changes in interactions with RNA will provide additional evidence for the model presented in Figure 8 of the manuscript. Finally, there are inconsistencies in the reported data between different figures and tables.

We thank the reviewer for the comprehensive and insightful review. A larger construct carrying both RBDs was not used because of the multiple challenges pertaining to dynamics study by NMR spectroscopy (intrinsic R2 rates of the dsRBD1-dsRBD2 construct would be high, resulting in broadened peaks) as per our previous experience ([Paithankar et al., 2022]). There would be additional dynamics in that construct coming from domain-domain relative motions, and it is difficult to deconvolute the dynamics information. Further, the dsRNA needed to bind to this construct will be longer, causing further line broadening in NMR.

Coming to mutational studies, careful designing of domain mutants remains as a challenge because the conformational dynamics in both the domains are distributed all through the backbone rather than only in the RNA-binding residues. The mutational studies would need an exhaustive number of mutations in protein as well as RNA to draw a parallel between the binding and dynamics. Having said that, we are working on making such mutations in the protein (at several locations to freeze the dynamics site-specifically) and the RNA (to change the shape of the dsRNA) to systematically study this mechanism, which will be out of scope of this manuscript.

The reviewer has rightly pointed out some subtle superficial differences in the reported data between different figures and tables. These superficial differences are present because of the context in which we are describing the data. For example, in Figure S4, we are talking about the average relaxation rates and nOe values for only the common residues we were able to analyze between two magnetic field strengths 600 and 800 MHz. Whereas in Figure 6, we are comparing the averages of the core (159-227) dsRBD residues at 600 MHz, in the presence and absence of D12RNA. The differences, however, are minute falls well within the error range.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

Suggestions for improved or additional experiments -

In regards to ITC data, dsRBD1 does not bind canonical A-form RNA with high affinity. What is dsRBD1 and dsRBD2 affinity to the miR-16 RNA?

We have not performed ITC-based studies with miR-16 RNA for the domains. The study by Acevedo et al. has shown the effect of lengths of Watson-Crick duplex RNAs upon TRBP2 dsRBD binding. In this study, they have compared the ds22 RNA to miRNA/miRNA* duplex. By using EMSA, they show that the Kd,app (μM) for dsRBD1 is 3.5±0.2 and for dsRBD2 is 1.7±0.1, indicating a higher affinity by the latter ([Acevedo et al., 2015]).

What was the amount of time used for the 1H saturation in the heteronuclear NOE experiment? Based on the average T1 (1/1.44 s-1) = 0.69 s, a recovery delay of >7 s should have been used for this experiment.

According to Cavanagh et al., a minimum recovery/recycle delay should be greater than 5*1/R1 to make sure that 99% of the 1HN and 15N magnetizations are restored ([“Protein NMR Spectroscopy, Principles and Practice, John Cavanagh, Wayne J. Fairbrother, Arthur G. Palmer III, and Nicholas J. Skelton. Academic Press, San Diego, 1995, 587 pages, $59.95. ISBN: 0-12-164490-1.,” 1996]). In our study, we have used a relaxation delay of 5 s, which is greater than 7*1/R1avg thus ensuring at least 99% of the 1HN and 15N recover their bulk magnetization.

Recommendations for improving writing and presentation -

Figure 3 - The legend in panel C is incomplete.

Figure 3 (Figure 4 in the revised manuscript) has been updated, and the legend now reads complete.

Figures 3 E and F - The three views can be combined into one as is done in Figures 4 C and D.

Thanks for the kind suggestion. We have depicted the kex in the three ranges to highlight the difference between the two domains at each range. Since there are three different exchange regimes with different populations, we believe this gives us an uncomplicated picture while classifying and comparing the dynamics between the two. Combining the three views into one becomes too overwhelming to visualize kex and population distribution in the protein.

Figure 3 - The residues indicated in the text (e.g., R200, L212, and R224) should be indicated in panels E and F.

We have marked the residues described in the text in Figure 4C (revised Figure 5C), and thus, they are not mentioned in Figures 3E and 3F (revised Figures 4E and 4F).

The results and discussion put these findings into minimal context. Most comparisons are made between dsRBD1 and dsRBD2. What about other RNA-binding proteins? There is a wealth of structure/dynamics/functional data about RNA recognition motifs, which do exactly the same thing as described here but are missing.

We understand the reviewer’s point that this study is focused on a dsRNA-binding mechanism rather than addressing the much broader field of RNA-binding. There are multiple challenges in finding a single mechanism that works for all RNA-binding proteins. For instance, RRM is a single-stranded RNA-recognition motif that can read out the substrate base sequence. RRM behaves entirely differently than the dsRBD in terms of sequence specificity. Besides, several other RNA-binding domains, like the KH-domain, Puf domains, Zinc-finger domains, etc., showcase a unique RNA-binding behaviour. Thus, with the current knowledge, it would not be possible to draw a single rule of thumb for RNA-recognition behaviour for all these diverse domains. Hence, the findings of this study are not comparable to those of other RNA-binding domains and are beyond the scope of this study.

Results, page 8 - I'm not sure that allosteric quenching is appropriately invoked here. The amount of residues showing dynamics in the apo state is small and the number only moderately increases upon RNA binding. The observation that some residues show an increase and a neighboring residue shows a decrease (or vice versa) upon RNA binding could just be random with the small number of observations. This observation would be more convincing if it were happening to larger regions within the protein.

We agree with the reviewer that the number of residues showing dynamics in the apo-state of the dsRBD2 is small when compared with that of dsRBD1, and the number only moderately increases upon RNA-binding. However, we believe it is quite important to invoke the allosteric quenching as all the new residues where dynamics is induced, do lie in the spatial proximity, as also observed in the dsRBD1 ([Paithankar et al., 2022]). It is a parameter to not only compare the differences and similarities in the two domains but also to highlight the presence of this phenomenon common in both the type-A dsRBDs of TRBP.

Minor corrections -

Introduction, page 2 - The order parameter should be defined for non-NMR experts.

Thank you for the suggestion. The definition of order parameter has now been included on page 2 of the revised manuscript.

Introduction, page 2 - TRBP should be defined in the main text the first time used.

We have now defined TRBP on page 2 of the revised manuscript, where it is used in the main text for the first time.

Results, page 5 - The reference for the HARD experiment should be given earlier in that paragraph.

Thank you for the suggestion. We have now referenced the HARD experiment earlier in the last paragraph on page 5 of the revised manuscript.

Results, page 7 - What is the limiting amount of RNA used for the D12-bound dsRBD2 spin relaxation measurements?

The limiting amount of RNA used for the D12-bound dsRBD2 spin relaxation measurements is 0.05 equivalent (RNA:Protein= 50 mM:1000 mM). It has now been included on page 7 of the revised manuscript.

Reviewer #2 (Recommendations For The Authors):

Throughout the manuscript, NMR datasets are not consistent with one another (a few examples are listed below).

Figures S4, 6, and Table S4: (a) It is unclear why relaxation data for certain residues are missing in Table S4 (e.g., S156, V168, E177, F192, etc.).

We thank the reviewer for pointing this out. We have now reanalyzed the data for all the above-mentioned residues and other missing residues. In the revised manuscript, we have added the data for the above-mentioned residues like E177, R189, and many more N- and C-terminal residues. Unfortunately, for some residues like V168, S184, F192, S209, and L222, we witnessed severe peak broadening while measuring the R2 rates and/or nOe. Hence, data for V168, S184, F192, S209, and L222 are missing in Table S4. We have explicitly mentioned this in the table legends about missing data for a few residues.

(b) The reported values are not consistent. For example, Figure S4 says that the average 15N-R2 rate is 10.85 +/- 0.36 s-1 whereas Figure 6 says the 15N-R2 rate is 11.02 +/- 0.39 s-1 for the same dataset.

The superficial differences are present because of the context in which we are describing the data (now mentioned in the methods section on page 13). In Figure S4, we are talking about the average relaxation rates and nOe values for only the common residues we could analyze between two magnetic field strengths, 600 and 800 MHz. Whereas in Figure 6 (revised figure 3), we compare the averages of all the analyzed core dsRBD residues at 600 MHz in the presence and absence of D12RNA. The differences, however, are insignificant, falling well within the error range.

(c) There is also a discrepancy in reported R2 values (at 600 MHz) in Table S4. It is unclear to me what the reported values are, as most of these are below 1 s-1.

Thank you very much for pointing out our mistake here. The Table S4 seems to have the wrong values for R2 at 600 MHz. However, the raw data submitted to the BMRB as entry 52077 holds the correct information. We have now updated the Table S4.

(d) It is also unclear as to why perfectly resolved residues (e.g., L230, A232, D234, etc.) have been omitted from these data (and other datasets such as 15N-CPMGs shown in Figure S6).

The residues L230, A232, D234, etc., are the C-terminal residues of TRBP-dsRBD2 beyond the core (159-227 aa) fold of dsRBD. They have now been included in the revised figures S6 and S11 for completeness.

(e) Figure 6 reports a 15N-R2 of 21 s-1 for one of the residues in the absence of RNA. This data point has been omitted from Figure S4.

In Figure S4, we are talking about relaxation rates and nOe values only for the common residues we could analyze between the two magnetic field strengths, 600 and 800 MHz. Thus, that 15N-R2 value has been omitted.

The S2 order parameters reported in Figures S5 and S10 are inconsistent with one another, as additional residues are shown in S10 (e.g., N159).

Thank you for pointing it out. We have now reanalyzed the data for S2 order parameter and Rex by including more residues (e.g., N159, R189, etc) in the core and have updated both Figures S5 and S10. Please see the revised supplementary information.

Tables S6 and S7 report values for residue R189. This residue has been omitted in every other dataset. Based on the 1H-15N HSQC spectrum shown in Figure S3, this residue gives a well-resolved crosspeak (which lies adjacent to V228). Can the authors explain why they omit data for this residue in Figures S4, 6, and Table S4?

The reviewer is correct in pointing out that data for R189 is missing in the fast dynamics data, such as Figure S4, Figure 6 (revised figure 3), and Table S4. We have now reanalyzed our raw data and included data for R189 and other missing residues in our updated manuscript. Please see the revised figures S4 and 6 (revised figure 3) and the revised table S4.  

Moreover, this residue lies in the loop2 region of this domain. Based on the MD simulations (Figure 2), this region is more flexible compared to the rest of the domain. Does the corresponding 15N-relaxation data support this claim?

Yes, the apo 15N-relaxation data do strongly support this claim. R189 showed a higher than core average R2 rate (R189 = 15.44 +/- 0.69 s-1; core = 10.92 +/- 0.37 s-1) and a lower than core average nOe (R189 = 0.49 +/- 0.05; core = 0.73 +/- 0.03) which indicate a higher flexibility than the rest of the core (updated Figure 3 and Table S4). Additionally, the S2 order parameter for R189 was found to be 0.52 +/- 0.03, slightly lower than the core average of 0.59 +/- 0.03, indicating a more flexible region than the core (updated Table S14). Moreover, the dynamics parameters extracted from HARD experimental data using the geoHARD method for apo TRBP2-dsRBD2 shown in Table S18 depict a high kex value of 31748.72 +/- 955.20 Hz for R189. This supports the claim that this residue is highly flexible with a high exchange rate.

Figure S9. I was not able to follow this dataset as the data points are not consistent between different residues.

In Figure S9, the residue-wise peak intensities plotted against the RNA concentration indicate that line broadening was witnessed for all the core residues (irrespective of the initial peak intensity). Another interesting observation is that the terminal residues do not undergo the same line broadening as seen in the core residues.

It is also unclear why residue G185 is highlighted.

It is taken as an example and magnified to show the extent of line broadening. This is now explicitly mentioned in the figure caption in the revised supplementary information.

It is also not clear exactly what the authors are trying to fit, as I see no chemical shift changes upon the addition of RNA (Fig. S8), and the equation used for data fitting (pg. 11) uses chemical shift changes (and not the changes in intensities).

The same equation can be used to fit the chemical shift perturbation and peak intensity perturbation as a function of ligand concentration. Here, we have tried to fit the intensity perturbation. We have now modified the statement on page 11 in the revised manuscript.

Table S2: The ITC analysis reports an n value of ~3. Can authors elaborate as to what this means?

The stoichiometry ~3 indicates the number of TBDP2-dsRBD2 that can interact with D12 RNA in a single binding event. The minimum binding register for dsRBDs is known to be >8 bp (12 bp for optimal binding) ([Ramos et al., 2000]), and one single domain only covers one-third of the face of the cylindrical RNA ([Masliah et al., 2018]). Hence, 3 dsRBD2 could interact with a 12-mer RNA in solution.

The reported Kd values between the main text (page 7) and Figure 5 are not consistent with one another (one lists 1.18 uM while the other says 1.11 uM). Table S2 does not list the parameters for interactions between dsRBD1 and D12.

Figure 5 (revised figure 6) depicts the information of a single isolated experiment out of a total of three, whereas in the main text, we say 1.18 μM as the average Kd value (table S2).

Figure S4: The red axis should read "211" instead of "111".

Thank you for your helpful insight. We have now changed it in the revised figure.

Table S3 lists the structural motifs of the two dsRBDs, which are nearly identical to one another, and yet the manuscript claims that these are different (page 4, paragraph 1).

We agree with the reviewer that the differences are minute but important, which we have tried to highlight in this paper. In particular, loop 2, critical for dsRNA-binding ([Masliah et al., 2012]), is 1 residue longer in dsRBD2 and has a possible effect in enhanced substrate binding.

Figure S8 shows severe signal attenuation for many residues upon the addition of 100 uM RNA. The most notable among these are residues M194, T195, and C196. Can the authors explain how they measure 15N-relaxation rates for these residues in the presence of 50 uM D12?

First, we have recorded the measured 15N-relaxation rates for these residues in the presence of 50 mM D12 (RNA:Protein= 50 mM:1000 mM)), corresponding to 0.05 equivalent RNA. The amount of RNA used is less than that used for the HSQC-based titration shown in Figure S8, 0.1 equivalent RNA (RNA:Protein = 5 mM:50 mM), where we witness line broadening for residues like M194, T195, and C196. Second, we increased the overall protein concentration from 50 mM (used in HSQC-based titration) to 1000 mM (used in relaxation measurements) to ensure a better signal-to-noise ratio in all the spectra.

Use the same coloring scheme for Figures S7 and S8.

Thank you for the suggestion. We have now edited Figure S8 accordingly.

Figures are often listed out-of-order, making it difficult to follow the manuscript.

Thank you for the suggestion. We have now amended the main text to refer to the figures sequentially. While doing so, we have renumbered Figure 6 as Figure 3, Figure 3 as Figure 4, Figure 4 as Figure 5, and Figure 5 as Figure 6.

Figure captions for the relaxation data should specify the temperature at which these datasets were collected.

Thanks for the valuable suggestion. We have now added the temperature wherever applicable.

References

Acevedo R, Evans D, Penrod KA, Showalter SA. 2016. Binding by TRBP-dsRBD2 Does Not Induce Bending of Double-Stranded RNA. Biophys J 110:2610–2617. doi:10.1016/j.bpj.2016.05.012

Acevedo R, Orench-Rivera N, Quarles KA, Showalter SA. 2015. Helical Defects in MicroRNA Influence Protein Binding by TAR RNA Binding Protein. PLoS ONE 10:e0116749. doi:10.1371/journal.pone.0116749

Koh HR, Kidwell MA, Ragunathan K, Doudna JA, Myong S. 2013. ATP-independent diffusion of double-stranded RNA binding proteins.

Masliah G, Barraud P, Allain FH-T. 2012. RNA recognition by double-stranded RNA binding domains: a matter of shape and sequence. Cell Mol Life Sci 70:1875–1895. doi:10.1007/s00018-012-1119-x

Masliah G, Maris C, König SL, Yulikov M, Aeschimann F, Malinowska AL, Mabille J, Weiler J, Holla A, Hunziker J, Meisner‐Kober N, Schuler B, Jeschke G, Allain FH. 2018. Structural basis of siRNA recognition by TRBP double‐stranded RNA binding domains. EMBO J 37:e97089. doi:10.15252/embj.201797089

Paithankar H, Tarang GS, Parvez F, Marathe A, Joshi M, Chugh J. 2022. Inherent conformational plasticity in dsRBDs enables interaction with topologically distinct RNAs. Biophys J 121:1038–1055. doi:10.1016/j.bpj.2022.02.005

Protein NMR Spectroscopy, Principles and Practice, John Cavanagh, Wayne J. Fairbrother, Arthur G. Palmer III, and Nicholas J. Skelton. Academic Press, San Diego, 1995, 587 pages, $59.95. ISBN: 0-12-164490-1. 1996. . J Magn Reson, Ser B 113:277. doi:10.1006/jmrb.1996.0189

Ramos A, Grünert S, Adams J, Micklem DR, Proctor MR, Freund S, Bycroft M, Johnston DS, Varani G. 2000. RNA recognition by a Staufen double‐stranded RNA‐binding domain. EMBO J 19:997–1009. doi:10.1093/emboj/19.5.997

Vuković L, Koh HR, Myong S, Schulten K. 2014. Substrate Recognition and Specificity of Double-Stranded RNA Binding Proteins. Biochemistry 53:3457–3466. doi:10.1021/bi500352s

Author response:

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

Reviewer #1 (Public Review): 

Summary: 

Campbell et al investigated the effects of light on the human brain, in particular the subcortical part of the 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 highresolution 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 a silent substitution approach to distinguish between different photoreceptors. This may be something to consider when designing the follow-up studies. 

Reviewer #1 (Recommendations For The Authors): 

(1) As suggested in the public review more information regarding the reasons behind the chosen light condition is needed. 

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 or cone opsins. In order to provide a more mechanistic insight into the light-driven effects on cognition ideally one would use a silent substitution approach to distinguish between different photoreceptors. 

(2) In support of this work, it was shown in mice that acute activation of ipRGCs using chemogenetics induces c-fos in some of the hypothalamic brain areas discussed here including LH (Milosavljevic et al, 2016 Curr Biol). Another study to consider including in the discussion is by Sonoda et al 2020 Science, in which the authors showed that a subset of ipRGCs release GABA. 

(3) Figure 1 looks squashed, especially the axes. Also, Figure 2 looks somewhat blurry. I would suggest that the authors edit the figures to correct this.

We thank the reviewer for their positive comments and agree with the weaknesses they pointed out. 

(1) The explanation regarding the choice of the illuminance is now included in the revised manuscript (PAGE 17): “Blue-enriched light illuminances were set according to the technical characteristics of the light source and to keep the overall photon flux similar to prior 3T MRI studies of our team (between ~1012 and 1014 ph/cm²/s) (Vandewalle et al., 2010, 2011). The orange light was introduced as a control visual stimulation for potential secondary whole-brain analyses. For the present region of interest analyses, we discarded colour differences between the light conditions and only considered illuminance as indexed by mel EDI lux. This constitutes a limitation of our study as it does not allow attributing the findings to a particular photoreceptor class.”

The revised discussion makes clear that these choices limit the interpretation about the photoreceptors involved (PAGES 12-13): “We based our rationale and part of our interpretations on ipRGC projections, which have been demonstrated in rodents to channel the NIF biological impact of light and incorporate the inputs from rods and cones with their intrinsic photosensitivity into a light signal that can impact the brain (Güler et al., 2008; Tri & Do, 2019). Given the polychromatic nature of the light we used, classical photoreceptors and their projections to visual brain areas are, however, very likely to have directly or indirectly contributed to the modulation by light of the regional activity of the hypothalamus.”

The discussion also points out the promises of silent substitution (PAGE 13): “Future human studies could isolate the contribution of each photoreceptor class to the impact of light on cognitive brain functions by manipulating prior light history (Chellappa et al., 2014) or through the use of silent substitutions between metameric light exposures (Viénot et al., 2012)”.

(2) We now refer to the studies by Milosavljevic et al. and Sonoda et al. 

PAGE 9: “Our data may therefore be compatible with an increase in orexin release by the LH with increasing illuminance. In line with this assumption, chemoactivation of ipRGCs lead to increase c-fos production, a marker of cellular activation, over several nuclei of the hypothalamus, including the lateral hypothalamus (Milosavljevic et al., 2016). If this initial effect of light we observe over the posterior part of the hypothalamus was maintained over a longer period of exposure, this would stimulate cognition and maintain or increase alertness (Campbell et al., 2023) and may also be part of the mechanisms through which daytime light increases the amplitude in circadian variations of several physiological features (BanoOtalora et al., 2021; Dijk et al., 2012).”

PAGE 10: “Chemoactivation of ipRGCs in rodents led to an increase activity of the SCN, over the inferior anterior hypothalamus, but had no impact on the activity of the VLPO, over the superior anterior hypothalamus (Milosavljevic et al., 2016). How our findings fit with these fine-grained observations and whether there are species-specific differences in the responses to light over the different part of the hypothalamus remains to be established.”

PAGE 10: “In terms of chemical communication, these changes in activity could be the results of an inhibitory signal from a subclass of ipRGCs, potentially through the release aminobutyric acid (GABA), as a rodent study found that a subset of ipRGCs release GABA at brain targets including the SCN (and intergeniculate leaflet and ventral lateral geniculate nucleus), leading to a reduction in the ability of light to affect pupil size and circadian photoentrainment (Sonoda et al., 2020). Whatever the signalling of ipRGC, our finding over the anterior hypothalamus could correspond to a modification of GABA signalling of the SCN which has been reported to have excitatory properties, such that the BOLD signal changes we report may correspond to a reduction in excitation arising in part from the SCN (Albers et al., 2017).”

(3) Figures 1 and 2 were modified. We hope their quality is now satisfactory. We are willing to provide separate figures prior to publication of the Version of Record.

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 illuminated 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 the 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. 

(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 rigor of the study. 

(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. 

(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. 

Weaknesses: 

(1) Foundation for Claims about Orexin and Histamine Systems: The manuscript needs to provide a clearer theoretical or empirical foundation for claims regarding the impact of light on the orexin and histamine systems in the abstract. 

(2) Inclusion of Cortical Correlates: While focused on the hypothalamus, the manuscript may benefit from discussing the role of cortical activation in cognitive performance, suggesting an opportunity to expand the scope of the manuscript. 

(3) Details of Light Exposure Control: More detailed information about how light exposure was controlled and standardized is needed to ensure the replicability and validity of the experimental conditions. 

(4) Rationale Behind Different Exposure Protocols: To clarify methodological choices, the manuscript should include more in-depth reasoning behind using different protocols of light exposure for executive and emotional tasks. 

Reviewer #2 (Recommendations For The Authors): 

Attention to English language precision and correction of typographical errors, such as "hypothalamic nuclei" instead of "hypothalamus nuclei," is necessary for enhancing the manuscript.

We thank the reviewer for recognising the interest and strength of our study.

(1) As detailed in the discussion, we do believe orexin and histamine are excellent candidates for mediating the results we report. As also pointing out, however, we are in no position to know which neurons, nuclei, neurotransmitter and neuromodulator underlie the results. The last sentence of the abstract (PAGE 2) was therefore removed as we agree the statement was too strong. We carefully reconsider the discussion and believe that no such overstatement was present.

(2) Hypothalamus nuclei are connected to multiple cortical (and subcortical) structures. The relevance of these projections will vary with the cognitive task considered. In addition, we have not yet considered the cortex in our analyses such that truly integrating cortical structures appears premature. 

We nevertheless added the following short statement (PAGE 11): “Subcortical structures, and particularly those receiving direct retinal projections, including those of the hypothalamus, are likely to receive light illuminance signal first before passing on the light modulation to the cortical regions involved in the ongoing cognitive process (Campbell et al., 2023).”

(3) We now include the following as part of the method section (PAGES 16-17): “Illuminance and spectra could not be directly measured within the MRI scanner due to the ferromagnetic nature of measurement systems. The coil of the MRI and the light stand, together with the lighting system were therefore placed outside of the MR room to reproduce the experimental conditions of the in a completely dark room. A sensor was placed 2 cm away from the mirror of the coil that is mounted at eye level, i.e. where the eye of the first author of the paper would be positioned, to measure illuminance and spectra. The procedure was repeated 4 times for illuminance and twice for spectra and measurements were averaged. This procedure does not take into account interindividual variation in head size and orbit shape such that the reported illuminance levels may have varied slightly across subjects. The relative differences between illuminance are, however, very unlikely to vary substantially across participants such that statistics consisting of tests for the impact of relative differences in illuminance were not affected. The detailed values reported in Supplementary Table 2 were computed combining spectra and illuminance using the excel calculator associated with a published work (Lucas et al., 2014).”

(4) The explanation regarding the choice of the illuminance is now included in the revised manuscript (PAGE 17): “Blue-enriched light illuminances were set according to the technical characteristics of the light source and to keep the overall photon flux similar to prior 3T MRI studies of our team (between ~1012 and 1014 ph/cm²/s) (Vandewalle et al., 2010, 2011). The orange light was introduced as a control visual stimulation for potential secondary whole-brain analyses. For the present region of interest analyses, we discarded colour differences between the light conditions and only considered illuminance as indexed by mel EDI lux. This constitutes a limitation of our study as it does not allow attributing the findings to a particular photoreceptor class.”

(5) The manuscript was thoroughly rechecked, and we hope to have spotted all typos and language errors.

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: 

(1) 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. 

(4) 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 #3 (Recommendations For The Authors): 

(1) The authors may wish to explain the importance of the orange light condition in the early section of the results -- i.e., when they first present the task structure. As it stands, I don't have a good appreciation of why the orange light was included -- was it a control condition? And if the differences between the light conditions (e.g., the narrow- vs. wide-band of light) were indeed ignored by focussing on the illuminance levels, are there any potential issues that the authors could then mitigate against with further experiments/analyses? 

(2) Are there other explanations for why illuminance levels might improve cognitive performance? For instance, the capacity to more easily perceive the stimuli in an experiment could plausibly make it easier to complete a given task. If this is the case, can the authors conceptualise a way to rule out this hypothesis? 

(3) Did the authors control for the differences in the number of voxels in each hypothalamic subregion? Or perhaps consider estimating the variance across voxels within the larger parcels, to determine whether the mean time series was comparable to the time series of the smaller parcels? 

(4) An alternative strategy that would mitigate against the differences in the size of hypothalamic parcels would be to conduct analyses on the hypothalamus without parcellation, but instead using dimensionality reduction techniques to observe the natural spread of responses across the hypothalamus. From the authors' results, my intuition is that these analyses will lead to similar conclusions, albeit without any of the potential issues with respect to differently-sized parcels. 

We thank the reviewer for acknowledging the originality and interest of our study. We agree that some methodological choices needed more explanation. We will address the weaknesses they pointed out as follows:

(1) The explanation regarding the choice of the illuminance is now included in the revised manuscript (PAGE 17): “Blue-enriched light illuminances were set according to the technical characteristics of the light source and to keep the overall photon flux similar to prior 3T MRI studies of our team (between ~1012 and 1014 ph/cm²/s) (Vandewalle et al., 2010, 2011). The orange light was introduced as a control visual stimulation for potential secondary whole-brain analyses. For the present region of interest analyses, we discarded colour differences between the light conditions and only considered illuminance as indexed by mel EDI lux. This constitutes a limitation of our study as it does not allow attributing the findings to a particular photoreceptor class.”

The revised discussion makes clear that these choices limit the interpretation about the photoreceptors involved (PAGE 12-13): “We based our rationale and part of our interpretations on ipRGC projections, which have been demonstrated in rodents to channel the NIF biological impact of light and incorporate the inputs from rods and cones with their intrinsic photosensitivity into a light signal that can impact the brain (Güler et al., 2008; Tri & Do, 2019). Given the polychromatic nature of the light we used, classical photoreceptors and their projections to visual brain areas are, however, very likely to have directly or indirectly contributed to the modulation by light of the regional activity of the hypothalamus.”

We further mention that (PAGE 13): “Furthermore, we cannot exclude that colour and/or spectral differences between the orange and 3 blue-enriched light conditions may have contributed to our findings. Research in rodent model demonstrated that variation in the spectral composition of light was perceived by the suprachiasmatic nucleus to set circadian timing (Walmsley et al., 2015). No such demonstration has, however, been reported yet for the acute impact of light on alertness, attention, cognition or affective state.”

Regarding the choice of tasks, we added the following the method section (PAGE 18): “Prior work of our team showed that the n-back task and emotional task included in the present protocol were successful probes to demonstrate that light illuminance modulates cognitive activity, including within subcortical structures (though resolution did not allow precise isolation of nuclei or subparts) (e.g. (Vandewalle et al., 2007, 2010)). When taking the step of ultra-high-field imaging, we therefore opted for these tasks as our goal was to show that illuminance affects brain activity across cognitive domains while not testing for task-specific aspects of these domains.”

We further added to the discussion (PAGE 8): “The pattern of light-induced changes was consistent across an executive and an emotional task which consisted of block and an event-related fMRI design, respectively. This suggests that a robust anterior-posterior gradient of activity modulation by illuminance is present in hypothalamus across cognitive domains.”

(2) We are unsure what the reviewer refers to when he states that the experiment could make it easier to perceive a stimulus. Aside from the fact that illuminance can increase alertness and attention such that a stimulus may be better or more easily perceived/processed, we do not see how blocks of ambient light, i.e. a long-lasting visual stimulus, may render auditory stimulation (letters or pseudo-words in the present) easier to perceive. To our knowledge multimodal or cross-modal integration has been robustly demonstrated for short visual/auditory cues that would precede or accompany auditory/visual stimulation. 

We are willing to clarify this issue in the text if we receive additional explanation from the reviewer.

(3) We added subpart size as covariate in the analyses (instead of subpart number) and it did not affect the output of the statistical analyses (Author response table 1). 

For completeness, we further computed standard deviation of the activity estimates of the voxels within each parcel for the main analysis of the n-back tasks and found a main effect of subpart (Author response table 2) indicating that the variability of the estimates varied across subparts. Post hoc contrast and the display included in Author response image1 show however that the difference were not related to subpart size per see. It is in fact the largest subpart (subpart 4) that shows the largest variability while one of the smallest subpart (subpart 2) shows the lowest variability. Though it may have contributed, it is therefore unlikely to explain our findings. We consider the analyses reported in (Author response table 1 and 2 and (Author response image 1 as very technical and did not include it in the supplementary material for conciseness. If the reviewer judges it essential, we can reconsider our decision.  

While computing these analyses, we realized that there were errors in the table 1 reporting the statistical outcomes of the main analyses of the emotional task. The main statistical outputs remain the same except for a nominal main effect of the task (emotional vs. neutral) and the fact that post hoc show a consistent difference between the posterior subpart (subpart 3) and all the other subparts, rather than all the other subparts except for the difference with superior tubular hypothalamus subpart: p-corrected = 0.09. We apologise for this slight error and were unable to isolate its origin. It does not modify the rest of the analyses (which were also rechecked) and the interpretations. 

Author response table 1.

Recomputations of the main GLMMs using subpart sizes rather than subpart numbers as covariate of interest.

Author response image 1.

Activity estimate variability per hypothalamus subpart and subpart size.  

Author response table 2.

Difference in activity estimate standard deviation between hypothalamus subparts during the n-back task.

Outputs of the generalized linear mixed model (GLMM) with subject as the random factor (intercept and slope), and task and subpart as repeated measures (ar(1) autocorrelation).

* The corrected p-value for multiple comparisons over 2 tests is p < 0.025.

# Refer to Fig.2A for correspondence of subpart numbers

The text referring to Table 1 was modified accordingly (PAGE 5): “A nominal main effect of the task was detected for the emotional task [p = 0.049; Table 1] but not for the n-back task. For both tasks, there was no significant main effect for any of the other covariates and post hoc analyses showed that the index of the illuminance impact was consistently different in the posterior hypothalamus subpart compared to the other subparts [pcorrected ≤ 0.05]”.

(4) We agree that a data driven approach could have constituted an alternative means to tests our hypothesis. We opted for an approach that we mastered best, while still allowing to conclusively test for regional differences in activity across the hypothalamus. Examination of time series of the very same data we used will mainly confirm the results of our analyses – an anterior-posterior gradient in the impact of illuminance - while it may yield slight differences in the boarders of the subparts of the hypothalamus undergoing decreased or increased activity with increasing illuminance. While the suggested approach may have been envisaged if we had been facing negative results (i.e. no differences between subparts, potentially because subparts would not reflect functional differences in response to illuminance change), it would constitute a circular confirmation of our main findings (i.e. using the same data). While we truly appreciate the suggestion, we do not consider that it would constitute a more parsimonious test of our hypothesis, now that we successfully applied GLM/parcellation and GLMM approaches.

We added the following statement to the discussion to take this comment into account (PAGE 12): “Future research may consider data-driven analyses of hypothalamus voxels time series as an alternative to the parcellation approach we adopted here. This may refine the delineation of the subparts of the hypothalamus undergoing decreased or increased activity with increasing illuminance.”

Response references

Albers, H. E., Walton, J. C., Gamble, K. L., McNeill, J. K., & Hummer, D. L. (2017). The dynamics of GABA signaling: Revelations from the circadian pacemaker in the suprachiasmatic nucleus. Frontiers in Neuroendocrinology, 44, 35–82. https://doi.org/10.1016/J.YFRNE.2016.11.003

Bano-Otalora, B., Martial, F., Harding, C., Bechtold, D. A., Allen, A. E., Brown, T. M., Belle, M. D. C., & Lucas, R. J. (2021). Bright daytime light enhances circadian amplitude in a diurnal

mammal. Proceedings of the National Academy of Sciences of the United States of America, 118(22), e2100094118. https://doi.org/10.1073/PNAS.2100094118/SUPPL_FILE/PNAS.2100094118.SAPP.PDF

Campbell, I., Sharifpour, R., & Vandewalle, G. (2023). Light as a Modulator of Non-Image-Forming Brain Functions Positive and Negative Impacts of Increasing Light Availability. Clocks & Sleep, 5(1), 116. https://doi.org/10.3390/CLOCKSSLEEP5010012

Chellappa, S. L., Ly, J. Q. M., Meyer, C., Balteau, E., Degueldre, C., Luxen, A., Phillips, C., Cooper, H. M., & Vandewalle, G. (2014). Photic memory for executive brain responses. Proceedings of the National Academy of Sciences of the United States of America, 111(16), 6087–6091. https://doi.org/10.1073/pnas.1320005111

Dijk, D. J., Duffy, J. F., Silva, E. J., Shanahan, T. L., Boivin, D. B., & Czeisler, C. A. (2012). Amplitude reduction and phase shifts of melatonin, cortisol and other circadian rhythms after a gradual advance of sleep and light exposure in humans. PloS One, 7(2). https://doi.org/10.1371/JOURNAL.PONE.0030037

Güler, A. D., Ecker, J. L., Lall, G. S., Haq, S., Altimus, C. M., Liao, H. W., Barnard, A. R., Cahill, H., Badea, T. C., Zhao, H., Hankins, M. W., Berson, D. M., Lucas, R. J., Yau, K. W., & Hattar, S. (2008). Melanopsin cells are the principal conduits for rod-cone input to non-image-forming vision. Nature, 453(7191), 102–105. https://doi.org/10.1038/nature06829

Lucas, R. J., Peirson, S. N., Berson, D. M., Brown, T. M., Cooper, H. M., Czeisler, C. A., Figueiro, M. G., Gamlin, P. D., Lockley, S. W., O’Hagan, J. B., Price, L. L. A., Provencio, I., Skene, D. J., & Brainard, G. C. (2014). Measuring and using light in the melanopsin age. Trends in Neurosciences, 37(1), 1–9. https://doi.org/10.1016/j.tins.2013.10.004

Milosavljevic, N., Cehajic-Kapetanovic, J., Procyk, C. A., & Lucas, R. J. (2016). Chemogenetic Activation of Melanopsin Retinal Ganglion Cells Induces Signatures of Arousal and/or Anxiety in Mice. Current Biology, 26(17), 2358–2363. https://doi.org/10.1016/j.cub.2016.06.057

Sonoda, T., Li, J. Y., Hayes, N. W., Chan, J. C., Okabe, Y., Belin, S., Nawabi, H., & Schmidt, T. M. (2020). A noncanonical inhibitory circuit dampens behavioral sensitivity to light. Science (New York, N.Y.), 368(6490), 527–531. https://doi.org/10.1126/SCIENCE.AAY3152

Tri, M., & Do, H. (2019). Melanopsin and the Intrinsically Photosensitive Retinal Ganglion Cells: Biophysics to Behavior. Neuron, 104, 205–226. https://doi.org/10.1016/j.neuron.2019.07.016

Vandewalle, G., Hébert, M., Beaulieu, C., Richard, L., Daneault, V., Garon, M. Lou, Leblanc, J., Grandjean, D., Maquet, P., Schwartz, S., Dumont, M., Doyon, J., & Carrier, J. (2011). Abnormal hypothalamic response to light in seasonal affective disorder. Biological Psychiatry, 70(10), 954–961. https://doi.org/10.1016/j.biopsych.2011.06.022

Vandewalle, G., Schmidt, C., Albouy, G., Sterpenich, V., Darsaud, A., Rauchs, G., Berken, P. Y., Balteau, E., Dagueldre, C., Luxen, A., Maquet, P., & Dijk, D. J. (2007). Brain responses to violet, blue, and green monochromatic light exposures in humans: Prominent role of blue light and the brainstem. PLoS ONE, 2(11), e1247. https://doi.org/10.1371/journal.pone.0001247

Vandewalle, G., Schwartz, S., Grandjean, D., Wuillaume, C., Balteau, E., Degueldre, C., Schabus, M., Phillips, C., Luxen, A., Dijk, D. J., & Maquet, P. (2010). Spectral quality of light modulates emotional brain responses in humans. Proceedings of the National Academy of Sciences of the United States of America, 107(45), 19549–19554. https://doi.org/10.1073/pnas.1010180107

Viénot, F., Brettel, H., Dang, T.-V., & Le Rohellec, J. (2012). Domain of metamers exciting intrinsically photosensitive retinal ganglion cells (ipRGCs) and rods. Journal of the Optical Society of America A, 29(2), A366. https://doi.org/10.1364/josaa.29.00a366

Walmsley, L., Hanna, L., Mouland, J., Martial, F., West, A., Smedley, A. R., Bechtold, D. A., Webb, A. R., Lucas, R. J., & Brown, T. M. (2015). Colour As a Signal for Entraining the Mammalian Circadian Clock. PLOS Biology, 13(4), e1002127. https://doi.org/10.1371/journal.pbio.1002127

Author response:

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

Reviewer #1 (Public Review):

Summary:

The current study aims to quantify associations between the 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 up to several years later in time. There are 6 respiratory outcomes included: risk of influenza, pneumonia, COVID-19, other respiratory tract infections, as well as COVID-19 severity and mortality).

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.

We are grateful for your pointing out the strengths in our article, particularly the assessment of e-values and the comparison with another medication to mitigate confounding by indication. We extend our sincere gratitude to the reviewer for identifying multiple concerns and offering constructive feedback to help improve our manuscript. We will incorporate these suggestions into our revisions.

Weaknesses:

(1) The main exposure of interest seems to be only measured at one time-point in time (at study enrollment) while patients are considered many years at risk afterwards without knowing their exposure status at the time of experiencing the outcome. As indicated by the authors, PPI are sometimes used for only short amounts of time. It seems biologically implausible that an infection was caused by using PPI for a few weeks many years ago.

We agree with the reviewer that PPIs are sometimes used for only short amounts of time, as indicated in our manuscript. We acknowledge that it is a limitation of the UK Biobank cohort, and we have discussed this in the discussion section as follows:

“Given that the PPI exposure was mainly assessed at the baseline recruitment, it was possible that a small proportion of PPI users was misclassified during the follow-up due to the medication discontinuation, which may result in an underestimation of potential risk.” (Page 14, Line 8-10)

In addition, to alleviate these concerns, we have conducted effect medication for the subgroup of potential long-term users, which were defined by participants with indications of PPI use. This information has been included in the discussion section:

“In addition, no effect moderation was observed in subgroup analyses for the main outcome among PPI users with indications (more likely to regularly use PPIs for a long period) compared to those without indications, indicating the risks remained increased among long-term PPI users.” (Page 14, Line 12-15)

We hope that in the future, the concerns highlighted by the reviewer can be resolved by utilizing datasets with close follow-up, especially regarding medication use:

“Since the follow-up prescription data was lacking in our study to precisely identifying the long-term users, further evaluation using cohorts with close follow-up is needed.” (Page 14, Line 15-17)

(2) Previous studies have shown that by focusing on prevalent users of drugs, one often induces several biases such as collider stratification bias, selection bias through depletion of susceptible, etc.

Because of the limitations of data from the UK Biobank, such as the absence of details on initiation of medications and regular monitoring, we were restricted to using a prevalent user design to assess the associations between PPI use and respiratory outcomes. We have discussed it in the limitation section:

“Given that the PPI exposure was mainly assessed at the baseline recruitment, it was possible that a small proportion of PPI users was misclassified during the follow-up due to the medication discontinuation, which may result in an underestimation of potential risk. However, the prevalent user design could underestimate the actual risks of PPI use for respiratory infections, which indicates the real effect might be stronger [38]……Since the follow-up prescription data was lacking in our study to precisely identifying the long-term users, further evaluation using cohorts with close follow-up is needed.” (Page 14, Line 8-17)

(3) It seems Kaplan Meier curves are not adjusted for confounding through e.g. inverse probability weighting. As such the KM curves are currently not informative (or the authors need to make clearer that curves are actually adjusted for measured confounding).

Your kind suggestions are greatly appreciated. We have plotted Kaplan Meier curves adjusted for confounding by inverse probability weighting with the measured confounders according to the reviewer’s advice. The methods and results are demonstrated as follows:

“The event-free probabilities were compared by Kaplan-Meier survival curves with inverse probability weights adjusting for the measured covariates.” (Page 8, Line 13-15)

“Regular PPI users had lower event-free probabilities for influenza and pneumonia compared to those of non-users (Supplementary Figure 2 A-B).” (Page 9, Line 21-23)

“PPI users had lower event-free probabilities for COVID-19 severity and mortality, but not COVID-19 positivity compared to those of non-users (Supplementary Figure 2 C-E).” (Page 10, Line 9-10)

(4) Throughout the manuscript the authors seem to misuse the term multivariate (using one model with e.g. correlated error terms to assess multiple outcomes at once) when they seem to mean multivariable.

We apologize for misusing the term “multivariate” and “multivariable” in our previous manuscript. We have corrected the misused terms throughout the manuscript:

“Univariate and multivariable Cox proportional hazards regression models were utilized to assess the association between regular use of PPIs and the selected outcomes.” (Page 7, Line 19-20)

“The remaining imbalanced covariates (standardized mean difference ≥ 0.1) after propensity score matching were further adjusted by multivariate multivariable Cox regression models to calculate HRs and 95% CIs.” (Page 8, Line 23-25)

(5) Given multiple outcomes are assessed there is a clear argument for accounting for multiple testing, which following the logic of the authors used in terms of claiming there is no association when results are not significant may change their conclusions. More high-level, the authors should avoid the pitfall of stating there is evidence of absence if there is only an absence of evidence in a better way (no statistically significant association doesn't mean no relationship exists).

We have revised our interpretation for the results, particularly for those without statically significant association based on the reviewer’s advice, and clearly recognize that the conclusions should be interpreted with cautions:

“In contrast, the risk of COVID-19 infection was not significant with regular PPI use…” (Page 2, Line 11-12)

“PPI users were associated with a higher risk of influenza (HR 1.74, 95%CI 1.19-2.54), but the risks with pneumonia or COVID-19-related outcomes were not evident.” (Page 2, Line 14-16)

“…while the effects on pneumonia or COVID-19-related outcomes under PPI use were attenuated when compared to the use of H2RAs.” (Page 2, Line 18-19, in the Abstract)

“…while their association with pneumonia and COVID-19-related outcomes is diminished after comparison with H2RA use and remains to be further explored.” (Page 15, Line 21-22, in the Conclusion)

(6) While the authors claim that the quantitative bias analysis does show results are robust to unmeasured confounding, I would disagree with this. The e-values are around 2 and it is clearly not implausible that there are one or more unmeasured risk factors that together or alone would have such an effect size. Furthermore, if one would use the same (significance) criteria as used by the authors for determining whether an association exists, the required effect size for an unmeasured confounder to render effects 'statistically non-significant' would be even smaller.

We agree with the reviewer that there might still exist one or more unmeasured risk factors that have effect sizes larger than 2. Hence, we cannot affirm that the findings are robust to unmeasured confounding in the current analysis, which is a limitation of our study. We have deleted the previous statement, and added more discussion in the limitation section:

“Moreover, patients with exacerbations of respiratory disorders (e.g., asthma, COPD) might suffer from a wide range of gastrointestinal symptoms that lead to the use of PPIs [38]. Due to the lack of data for respiratory severity and close follow-up for medication use, residual confounding might still exist due to the observational nature.” (Page 14, Line 23-27)

(7) Some patients are excluded due to the absence of follow-up, but it is unclear how that is determined. Is there potentially some selection bias underlying this where those who are less healthy stop participating in the UK biobank?

Thank you for your question. The reasons for the absence of follow-up are mainly classified into five categories, including: (1) Death reported to UK Biobank by a relative; (2) NHS records indicate they are lost to follow-up; (3) NHS records indicate they have left the UK; (4) UK Biobank sources report they have left the UK; (5) Participant has withdrawn consent for future linkage. According to the data from UK Biobank (https://biobank.ndph.ox.ac.uk/ showcase/field.cgi?id=190), the major reason for the loss of follow-up among participants is their departure from the UK (84.7% of participants who were lost to follow-up). In addition, not including those who were less healthy in the study might also underestimate the risk, leading to lower estimated effects of PPIs for respiratory infections. We have supplemented this in our revised manuscript:

“Among them, 1,297 participants without follow-up, which were mainly determined by reported death, departure from the UK, or withdrawn consent, had been removed after initial exclusion.” (Page 4, Line 25-27)

(8) Given that the exposure is based on self-report how certain can we be that patients e.g. do know that their branded over-the-counter drugs are PPI (e.g. guardium tablets)? Some discussion around this potential issue is lacking.

Thank you for your concerns. In the data collection by the UK Biobank, the participants can enter the generic or trade name of the treatment on the touchscreen to match the medications they used. We have added this important information to the method section:

“The exposure of interest was regular use of PPIs. The participants could enter the generic or trade name of the treatment on the touchscreen to match the medications they used (Supplementary Table S1).” (Page 5, Line 6-8)

We acknowledge that specific information on prescribed or over-the-counter use of medications is lacking in the UK Biobank. We have discussed it in the limitation section:

“Limitations exist in our study. Information on dose and duration of PPI use, discrimination between prescription and over-the-counter use of PPIs, health-seeking behavior, different types of pneumonia, and pneumococcus vaccination is currently not available from the UK Biobank.” (Page 14, Line 5-8)

(9) Details about the deprivation index are needed in the main text as this is a UK-specific variable that will be unfamiliar to most readers.

Thank you for your question on the definition of deprivation index. We have proved the details  about the deprivation index in the manuscript:

“…socioeconomic status (deprivation index, which was defined using national census information on car ownership, household overcrowding, owner occupation, and unemployment combined for postcode areas of residence)…” (Page 6, Line 14-17)

(10) It is unclear how variables were coded/incorporated from the main text. More details are required, e.g. was age included as a continuous variable and if so was non-linearity considered and how?

We apologize for not elucidating how variables were incorporated into the main text. Previously, the linearity between continuous variables and outcomes was assessed by Martingale residuals plots, while the variables detected with non-linearity were regarded as categorical variables for further analyses. For example, after evaluation with the Martingale residuals plot, age demonstrated non-linearity, and we incorporated it as a categorical variable for the analysis of COVID-related mortality.

We have supplemented the information in the method section:

“The linearity between continuous variables and outcomes was assessed by Martingale residuals plots, while the variables detected with non-linearity were regarded as categorical variables for further analyses.” (Page 6, Line 28 to Page 7, Line 1)

(11) The authors state that Schoenfeld residuals were tested, but don't report the test statistics. Could they please provide these, e.g. it would already be informative if they report that all p-values are above a certain value.

We are sorry for not providing the statistics about the Schoenfeld residual in our previous manuscript. We have supplemented the information in our revisions:

“Schoenfeld residuals tests were used to evaluate the proportional hazards assumptions, while no violation of the assumption was detected (Supplementary Table S3).” (Page 7, Line 27 to Page 8, Line 1)

(12) The authors would ideally extend their discussion around unmeasured confounding, e.g. using the DAGs provided in https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7832226/, in particular (but not limited to) around severity and not just presence/absence of comorbidities.

Thank you for your insightful suggestions that the discussion about unmeasured confounding should be extended. We agree with the reviewer that, in addition to the comorbidities themselves, their severity could also have an important impact on the use of PPIs. We have added the discussion in the limitation section with citing the article (PMC7832226):

“Moreover, patients with exacerbations of comorbid disorders (e.g., diabetes, asthma, COPD) might suffer from a wide range of gastrointestinal symptoms that lead to the use of PPIs [38] (Supplementary Figure S4). Due to the lack of data for respiratory severity and close follow-up for medication use, residual confounding might still exist due to the observational nature.” (Page 14, Line 23-27)

(13) The UK biobank is known to be highly selected for a range of genetic, behavioural, cardiovascular, demographic, and anthropometric traits. The potential problems this might create in terms of collider stratification bias - as highlighted here for example: https://www.nature.com/articles/s41467-020-19478-2 - should be discussed in greater detail and also appreciated more when providing conclusions.

We acknowledge the reviewer's point about the UK Biobank's highly selective nature potentially leading to collider stratification bias in the evaluation of COVID-19-related outcomes. We have discussed this in detail and are cautious when generating conclusions.

“Furthermore, the highly selective nature of the UK Biobank might create collider stratification bias for the evaluation of COVID-19-related outcomes, and thus the conclusions should be interpreted with cautions [39].” (Page 15, Line 2-4)

Reviewer #2 (Public Review):

Summary:

Zeng et al investigate in an observational population-based cohort study whether the use of proton pump inhibitors (PPIs) is associated with an increased risk of several respiratory infections among which are influenza, pneumonia, and COVID-19. They conclude that compared to non-users, people regularly taking PPIs have increased susceptibility to influenza, pneumonia, as well as COVID-19 severity and mortality. By performing several different statistical analyses, they try to reduce bias as much as possible, to end up with robust estimates of the association.

Strengths:

The study comprehensively adjusts for a variety of critical covariates and by using different statistical analyses, including propensity-score-matched analyses and quantitative bias analysis, the estimates of the associations can be considered robust.

We are grateful to the reviewer for pointing out the merits of our articles, which include adjusting for a wide range of covariates, employing diverse statistical analyses, and using robust data. We will revise our manuscript further based on the reviewer's suggestions.

Weaknesses:

As it is an observational cohort study there still might be bias. Information on the dose or duration of acid suppressant use was not available, but might be of influence on the results. The outcome of interest was obtained from primary care data, suggesting that only infections as diagnosed by a physician are taken into account. Due to the self-limiting nature of the outcome, differences in health-seeking behavior might affect the results.

Thank you for your questions for information on the dose/duration of acid suppressants, the source of diagnosis, and the health-seeking behavior of participants. For the data from the UK Biobank, the dose or duration of acid suppressant use was not available since the information was not collected as baseline or follow-up. In addition, the outcome of interest was also retrieved from the hospital ICD diagnosis. We apologize for not clarifying it in our previous manuscript. Moreover, we agree with the reviewer that the health-seeking behavior could have an impact on the analyses, whereas the correlated data are still not available from the UK Biobank. We have discussed them in the method and limitation section:

“Briefly, the first reported occurrences of respiratory system-related conditions within primary care data,  and hospital inpatient data defined by the International Classification of Diseases (ICD)- 10 codes were categorized by the UK Biobank.” (Page 5, Line 21-25)

“Limitations exist in our study. Information on dose and duration of PPI use, discrimination between prescription and over-the-counter use of PPIs, health-seeking behavior, different types of pneumonia, and pneumococcus vaccination is currently not available from the UK Biobank.” (Page 14, Line 5-8)

Reviewer #1 (Recommendations For The Authors):

Analysis code should be made available.

Thank you for your question. We have provide the sources of the analysis code we used for this study in our revised manuscript:

“The codes used in this study can be found at: https://epirhandbook.com/en/ and https://cran.r-project.org/doc/contrib/Epicalc_Book.pdf.” (Page 16, Line 21-22)

Reviewer #2 (Recommendations For The Authors):

It might be interesting to study whether including self-reported infections changes the results, as people using PPI may more easily consult their GP even for a self-limiting disease such as influenza and therefore are more likely diagnosed/confirmed with such a respiratory infection.

Thank you for your insightful suggestions on conducting analyses including self-reported infections. Therefore, we have included the self-reported cases as sensitivity analyses, and the results were not significantly altered, which confirms the robustness of our results:

“Self-reported infections, except for COVID-19-related outcomes due to the lack of data, were also included for the outcomes as sensitivity analyses. The self-reported cases were reported at the baseline or subsequent UK Biobank assessment center visit.” (Page 8, Line 17-19)

“Inclusion of the self-reported cases did not significantly alter the results (Supplementary Table S4).” (Page 9, Line 17-18)

Moreover, to address the above-mentioned, sub-analyses differentiating between over-the-counter and prescribed medication might be interesting.

Thank you for your questions on differentiating between over-the-counter and prescribed medication. We have thoroughly looked up the data provided by the UK Biobank, but it is a pity that they are not provided. We have discussed this in the limitation section:

“Information on dose and duration of PPI use, discrimination between prescription and over-the-counter use of PPIs, health-seeking behavior, different types of pneumonia, and pneumococcus vaccination is currently not available from the UK Biobank.” (Page 14, Line 5-8)

Author response:

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

Reviewer #1 (Public Review):

(1.1) I thought the manuscript was very clear. While I realize the authors included the reference to medulloblastoma in the introduction based on previous reviewer comments, I think this speculation is better left in the discussion.

Whilst we appreciate the reviewers feedback here, we felt it was important to include a reference to medulloblastoma and developmental disorders associated with the cerebellum to put this work into a broader context.

We removed the sentence “Medulloblastoma can be a consequence of uncontrolled proliferation of granule cell progenitors, with BMP overexpression being a potential therapeutic avenue to inhibit this proliferation” to limit the speculation in this statement.

(1.2) line 81: It would be better to cite the 2 original papers (Hendrikes et al 2022, Smith et al 2022) rather than the Phoenix commentary article. I'm not sure the Phoenix article needs to be cited at all within this paper.

We have cited the two suggested papers and removed the citation to Phoenix et al.

(1.3) line 102: confusing sentence with the unexpected separation of do and not: "the same conditional deletions of BMP pathway elements that fail to block early granule cell specification at the rhombic lip do result not in a larger cerebellum as might be expected, but either have no affect".

We thank the reviewer for pointing out this error and have corrected the text to “do not result in a larger cerebellum”.

(1.4) line 133: inconsistent acronyms (for example, W9 vs pcw9).

This has been corrected to PCW in all occurrences.

(1.5) line 139: coronal vs transverse? it seems like you show transverse sectioning but refer to it as coronal in the text.

We thank the reviewer for highlighting this and have corrected the text to “transverse”.

(1.6) fig 2C: would it be possible to provide a similar inset as 2D?

We thank the reviewer for this suggestion and have added the insets in 2C. We agree that this is now clearer and more consistent with the rest of the figure.

(1.7) line 368/369/435/436 missing arrows.

The arrows have been re-added- it appears that they did not show up on the uploaded PDF.

(1.8) line 517 missing word: rhombic-lip-derived.

This typo has been corrected.

Reviewer #2 (Public Review):

(2.1) Fig. 3 M Why are there asterisks both above and below the brackets?

This was a formatting error that has now been corrected.

(2.2) Fig. 8. The arrows (BMP up and BMP down) are touching the right ")" in the figure, which makes it hard to read.

This was also a formatting issue which has been corrected.

(2.3) Fig. 4 and 8 legends. There are spaces in the text which I believe are for arrows to be inserted "(BMP )", but the arrows have been omitted in the PDF that I read.

This is the same as reviewer 1’s comment- these have been re-added to the text and appears to have been an issue with the PDF upload.

(2.4) Fig. 3 legend gets very hard to read at the end, where it seems some punctuation is missing.

We have re-worded the legend for Fig. 3 to make it easier to read.

(2.5) Significant figures in some of the text are probably too much given the accuracy at which they can be measured with.

We appreciate the reviewer’s concerns here, however these were added in response to the original reviewer’s request to “provide some additional support to otherwise qualitative observations”.

Author response:

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

Reviewer #1 (Recommendations For The Authors):

In my opinion, the three most important controls (hopefully easy):

(1) Include no ATR controls for optogenetic activation experiments (not all, just one or two, e.g., Figure 4B, C, or D, for the highest activation condition). The concern is that it can be quite hard to use light to both monitor neural responses while also using light to activate the function of other neurons.

We thank the reviewer for the suggestions. We use a 2-photon 910-nm laser (which does not activate Chrimson) for imaging of GCaMP and a 624-nm LED (which does not activate GFP) for Chrimson activation. Calcium (GCaMP) signals are detected by PMT during Chrimson activation. With this setup, we are able to image GCaMP signals without crosstalk during activation of Chrimson.

We performed calcium imaging in animals that were not fed ATR and found that SS04185 showed no response to LED stimulation at the strongest intensity (µW/mm) (New Figure 4 – figure supplement 1B).

(2) Demonstrate that their RNAi constructs do indeed knock down the intended target gene. They showed nicely in Figure 5A that SeIN128 expresses GABA. Presumably, these neurons also express VGAT. Is it possible to check the expression of VGAT after RNAi knockdown? The concern is that using only a single RNAi introduces the possibility of off-target effects. Using multiple RNAi lines for VGAT or other parts of the pathway would also alleviate this (minor concern).

We thank the reviewer for raising this point. We agree that using only one RNAi line (HMS02355) for VGAT in Figure 5A is a weakness. 

Accordingly, we have performed additional experiments to quantify the effect of RNAi knockdown of VGAT using HMS02335 in all neurons, followed by subsequent immunostaining against GABA or VGAT. We found that both VGAT and GABA were significantly reduced in the neuropil (Figure 5 – figure supplement 1C and D). These data strongly suggest that HMS02355 knocks down VGAT and reduces GABA at axon terminals. We note that HMS02355 has been used previously for knocking down GABA signaling in the following studies.

(1) Kallman BR, Kim H, Scott K (2015). Excitation and inhibition onto central courtship neurons biases Drosophila mate choice. eLife 4:e11188. https://doi.org/10.7554/eLife.11188

(2) Zhao W, Zhou P, Gong C et al. (2019). A disinhibitory mechanism biases Drosophila innate light preference. Nat Commun 10, 124. https://doi.org/10.1038/s41467-018-07929-w

(3) Yamagata N, Ezaki T, Takahashi T, Wu H, Tanimoto H (2021). Presynaptic inhibition of dopamine neurons controls optimistic bias. eLife 10:e64907. https://doi.org/10.7554/eLife.6490

(3) Include genetic controls for their driver line.

In Figure 1, it would be nice to see one half or the other half of their split GAL4 line in their manipulations. The concern is that perhaps the phenotype is coming from something unexpected in the genetic background.

We thank the reviewer for the suggestion. We have added half of the GAL4 lines (AD or DBD) as controls (New Figure 1 – figure supplement 2). We found that SS04185 showed reduction of rolling, whereas AD only or DBD only (split control) did not (half of the split lines). 

In the discussion:

It seems that activation of SS014185 has additional effects beyond what the authors have quantified. Specifically, larvae do not appear to re-initiate rolling in the same manner as Basin activation alone. Also, there appears to be an off-response, turning.

We appreciate the reviewer’s comments. We have included a section in the discussion to consider the differences patterns of rolling observed during joint stimulation of Basins and SS04185 and during stimulation of Basins alone, as well as the increase in turning following the offset of joint stimulation of Basins and SS04185 compared with stimulation of Basins alone (lines 464 to 481). Although the reasons for these differences are beyond the scope of the paper, we have added Figure 2 – figure supplement 1K, which shows that co-activation of SS04185-MB and Basins is sufficient to evoke turning following the offset of stimulation, suggesting that the increased turning may be due to the activation of SS04185-MB neurons and independent of SS04185-DN neurons.  

The labeling of the Figure panels could be improved. In many places, it is not clear that Basins are being stimulated in the background, whereas in nearby panels, it is clearly labeled. This is confusing for the reader.

We thank the reviewer for the constructive suggestions. We have modified all relevant figures to read “Basins>Chrimson” above the pink line indicating the period of optogenetic activation.

Reviewer #2 (Recommendations For The Authors):

Claims, rigorousness, repeatability, and accuracy of terms.

(1) In line 254, the authors suggest that the slow response of SeIN128 neurons is due to the input they receive from SEZ, but in line 453, they suggest it is due to axo-axonal connections. However, their evidence does not support one factor over the other. Overall, only the axo-axonal connection was strongly suggested in the discussion. The authors could clarify that the delay of SeIN128 activity may also be caused by multisynaptic connections involving SEZ or other neurons in the last section of the Discussion.

Although SeIN128 primarily receives inputs from the SEZ, it also receives inputs within the VNC from Basin-2 (Figure 4 – figure supplement 2). Specifically, in the VNC, the axons of SeIN128 make inhibitory synaptic contacts onto the axon of Basin-2, which in turn makes reciprocal excitatory contacts onto the axon of SeIN128, thereby forming a feedback loop. However, by the time we wrote the original discussion, we had inadvertently focused on the potential of the negative feedback loop formed by these axo-axonal synapses in the VNC to mediate the slow response of SeIN128, overlooking the possibility that other as yet unidentified pathways could convey Basin or A00c activity indirectly to SeIN128 dendrites in the SEZ. Therefore, we have revised the original text, which read “These data suggest that the main synaptic inputs onto SeIN128 neurons in the SEZ mediate the slow responses upon activation of Basins or A00c neurons” to “These data suggest that the delay of SeIN128 activity may be caused by multi-synaptic connections involving the SEZ or a feedback loop involving axo-axonal connections between SeIN128 and Basin-2 or A00c” (revised, Lines 259 and 261). Accordingly, we have also adjusted the relevant discussion section to be consistent with this change (Lines 460 and 466).

(2) Please clarify the following: How does the algorithm define rolling and crawling? Healthy larvae complete 360{degree sign} rolls, in each roll they rotate from dorsal up to dorsal up. It is possible that a larva rolls for an incomplete cycle and straightens up. Does the algorithm simply label individual frames as “roll”, “non-roll”, or “unknown”, and defines rolling by the existence of “roll” frames? If so, then larvae that rolled for 90{degree sign} and straightened would be counted as “rolling” though they failed to complete a full rolling bout. Also, how were “hunch” “turn” and “back” identified? Lastly, is there any manual quality control involved? Address this and related issues in the methods:

a)  Expand the description of the classifier algorithm.

b)  How are rolling and non-rolling animals defined in the "rolling%" assay? Were all "rolling" animals able to do at least one 360{degree sign} roll?

c)  How are "rolling duration" and "end of 1st rolling" defined? Is the algorithm able to distinguish different rolling bouts? In these two assays, were the animals rolled for <1 second (in total or their "first roll") able to complete a 360{degree sign} roll?

The Multi-worm Tracker (MWT) records only the contours of animals (no real video image data). Thus, the data fed into the classifier algorithm only includes features based on contour time-series data. The algorism uses movement perpendicular to the body axis—the characteristic feature of larval rolling—to classify rollers and non-rollers. Although the algorithm cannot determine whether a rolling event involves a rotation of more than 360 degrees, we ensure that rolling events are at least 360 degrees by removing any events that are shorter than 0.2 s (the minimum time to complete a 360-degree roll).

We have accordingly revised the section of “Behavior detection” relating to the behavior classification algorithm in the methods section as follows (Lines 600 to 620).

“After extracting behavioral parameters from Choreography, we used an unsupervised machine learning behavior classification algorithm to detect and quantify the following behaviors: hunching (Hunch), headbending (Turn), stopping (Stop), and peristaltic crawling (Crawl) as previously reported (Masson et al., 2020). Escape rolling (Roll) was detected with a classifier developed using the Janelia Automatic Animal Behavior Annotator (JAABA) platform (Kabra et al., 2013; Ohyama et al., 2015). JAABA transforms the MWT tracking data into a collection of ‘per-frame’ behavioral parameters and regenerates 2D dorsal-view videos of the tracked larvae. Based on such videos, we defined rolling as a rotation around the body while the larva maintains a C-shape, which results in a movement perpendicular to larval body axis (Supplementary videos 1 and 2). Using this definition, we trained the algorithm in the JAABA platform by labeling ~10,000 randomly chosen frames as rolling or non-rolling to develop the rolling classifier. If a larva did not curl into a C-shape or move sideways, it was labeled as a “non-roller.” Every animal with at least one rolling event longer than 0.2 s in a given period was labeled as a “roller” (i.e., it was assumed to have rolled at least 360 degrees), based on the observation that when the start and end of rolling events were precisely measured, the algorithm could identify rolling events completed in 0.2 s.

The rejection of false positives, especially at the beginning and the end of each rolling bout, enhanced accuracy. The algorithm integrated these training labels and parameters generated with Choreography in a time series, such as speed, crabspeed, and body curvature, to generate a score for rolling detection. Above a certain threshold, the classifier labeled the frame as rolling. This classifier, which has false negative and false positive rates of 7.4% and 7.8%, respectively (n = 102), was utilized to detect rolling in this paper.”

Readability of text

(1) I suggest giving the SS04185 line and SeIN128 neuron common names that are easier to remember and follow (after mentioning their full name once).

We acknowledge the reviewer’s concerns. However, because SS04185 was initially named using the Janelia split-line pipeline, and SeIN128 was named independently in a more recent study (Ohyama et al., 2015), we have retained these designations in the present manuscript.

Figures and figure legends

(1) It would help if the authors could put visual representations of rolling and crawling, such as a cartoon larva performing the rolling-crawling switch, and still frames of rolling and crawling of real larvae, especially in Figure 1. Also, please consider including a video of rolling and crawling in real larvae (preferably comparing control and experimental groups).

We appreciate the reviewer’s suggestion. We have added a cartoon of the behavioral sequence in Figure 1A, as well as a Figure 1 supplement video based on MWT data, which shows rolling followed by crawling. 

(2) To give the reader a take-home message, it would help if the authors could make a simplified version of Figure 4A and put it at the end of the paper.

We thank the reviewer for the suggestion. To assist the reader, we have added schematics depicting how the circuit may function in panel I of Figure 8.

(3) In Figure 1A, add the text "activation " after the neuron names.

We have added “Chrimson” following “Basins>” to the new Figure 1B (old Figure 1A) and other figures (Figure 1C and D, Figure 5A, Figure 6A, and figure supplements).

(4) Figure 1G: a data point is misaligned (at the top of the graph). 

We have aligned the data point accordingly.

(5) Figure 1B can benefit from a better design. If possible, please separate the crawling speed into an independent graph (or at least use a different line shape to code for crawling speed and indicate it on the in-graph legend). Is the speed of Basin/SS04185 co-activation studied?

We appreciate the reviewer’s suggestion. We have separated the plots for rolling and crawling speed into different panels (Figure 1C and D). As shown in Figure 1D, the crawling speed observed during coactivation of Basins and SS04185 was similar to that during activation of Basins alone.

(6) Figure S1 uses a different color-coding scheme from Figure 1. I suggest making the color coding consistent between figures.

We are grateful for the reviewer’s suggestion. We have adjusted the color-coding scheme accordingly.

(7) Line 692 (Figure 2 legend), "Killer Zipper" is misspelled as "Kipper Zipper". Out of curiosity, is there a way to remove or reduce SS04185-DN expression in the same manner as SS04185-MB reduction?

We have corrected the text in the legend for Figure 2. As for the reviewer’s question, we did attempt to reduce or abolish SS04185-DN expression with tsh-LexA and LexAop-Kip+ but found no effect. Other identified LexA constructs with SeIN128 expression, however, all showed SS04185-MB expression. Consequently, we could not use these constructs because they inhibit both SeIN128 and SS04185-DN.

(8) The color coding of Figure 2 (especially in D) makes it hard to distinguish between the brown and red groups.

We thank the reviewer for the suggestion. Accordingly, we have changed the color for the brown group to orange.

(9) In line 926 (Figure S2 legends), the description of F and G seems inverted.

We appreciate the reviewer for pointing out the error. We have revised the text from “(F) has only SS04185-

MB expression, and (G) has both SS04185-DN and SS04185-MB expression” to “(F) has both SS04185DN and SS04185-MB expression, and (G) has only SS04185-MB expression.”

(10) Figure 7B: which line does the top group of asterisks belong to?

The top group of asterisks indicates that each experimental group differs significantly (p < 0.001) from the control group. We have revised the figure to clarify the comparisons indicated by the asterisks in Figure 7B, as well as the figure legend below (Line 890-894).

“(B) Cumulative plot of rolling duration. Statistics: Kruskal-Wallis test: H = 69.52, p < 0.001; Bonferronicorrected Mann-Whitney test, p < 0.001 between control and the GABA-B-R11, GABA-B-R12 and GABAB-R2 RNAi groups, p < 0.001 between GABA-A-R and all other experimental RNAi group. Sample size for the colored bars from top (control, black) to bottom (GABA-A-R, red); n = 520, 488, 387, 582, 306.”

(11) Figure S8 D and F: indicate Basin-2 or Basin-4 activation on graph.

We have revised Figure 8 – figure supplement D and F accordingly.

Reviewer #3 (Recommendations For The Authors):

(1) Lines 86-87: Text needs to be rewritten for clarity. Also, include the genotype in the corresponding figure legend (Figure 1B).

We thank the reviewer for pointing this out. We have clarified the text accordingly and included the genotype in the figure legend (lines 86 and 87). Specifically, we have revised Figure 1B (New Figure 1C and D) and adjusted the legend accordingly as follows. 

Lines 86 and 87: Crawling speed during the activation of all Basins following rolling was ~1.5 times that of the crawling speed at baseline (Figure 1D).

(2) Include the protocol for heat shock-FLP out experiments

We have added the following paragraph to the Methods section describing the heat shock-FlpOut experiments (lines 537 to 546).

“Heat shock FlpOut mosaic expression

First instar Drosophila larvae were exposed to heat shock in a water bath at 37°C for 12 min as previously described (Nern et al., 2015). With precise temporal and thermal control of heat shock, larvae with genotype

w+, hs(KDRT.stop)FLP/13xLexAop2-IVS-CsChrimson::tdTomato; R54B01-Gal4.AD/72F11LexA;20xUAS-(FRT.stop)-CsChrimson::mVenus/R46E07-Gal4.DBD showed sporadic

CsChrimson::mVenus expression driven by SS04185 split GAL4. As a result, the ratio of the larvae with SS04185-DN and SS04185-MB expression to those with only SS04185-MB expression was 1:1. Each larva was individually examined with optogenetic stimulation and behavior analysis. After behavioral experiments, mVenus expression in CNS was confirmed under the fluorescence microscope.”

(3) In the immunohistochemistry, the authors exclude the steps for washings. Recommend the authors to cite the previous literature. Similar to the other protocols detailed in the methods.

We have added a brief description of the steps involved in washing (lines 641 and 648). We have also provided a citation with similar immunohistology protocols (Patel, 1994).

(4) Keeping the same Y-axis scale for similar graphical representation would be helpful to compare across different experimental conditions and genotypes-for example, 2E and 2H for the start of the first crawl.

As suggested by the reviewer, we have adjusted the y-axis scales for Figure 2E and H to be identical.

(5) The color schematics used for the graph make it hard to visualize the data. The author might reconsider the better presentation of the data by avoiding darker colors.

We thank the reviewer for the constructive suggestion. We have lightened the shading of all violin plots. We have also modified the shading for the middle group in Figure 2C and E from dark brown to orange.

(6) Co-activation of the SS04185 and Basins in the figures represented as Basins+SS04185 (Figure 1A) and SS04185 (rest of the figures). Authors might reconsider this terminology to define and distinguish the coactivation of SS04185 and Basins neurons from the activation of SS04185 or Basins alone. It needs to be clarified in the figures.

We have adjusted the terminology by including “Basins>Chrimson” in all panels in which Basin neurons are optogenetically activated to trigger rolling in the background for all groups. Additionally, we have labeled the control group as “Control” and the experimental group as ”SS04185”. 

(7) Figure 4A, summarizes the synaptic connection and strength between different neurons - SeIN128, Basins, A00c and mdIV. However, the nature of these synaptic connections - excitatory and inhibitory- is not represented. Based on the previous and current studies, the authors consider providing the schematic for circuit mechanisms of escape behavior sequences in larvae. Also, discussing these findings in light of the downstream output circuit and motor regulation might be informative (See Cooney et al. 2023, PNAS).

As the reviewer correctly points out, the diagram of the connectome shown in Figure 4A does not indicate whether the connections are excitatory or inhibitory. Accordingly, we have added a new summary panel (Figure 8I) based on the results of examining GABAergic synapses (Figure 5A). The schematics in Figure 8I depict how the joint activity of inhibitory and excitatory synapses (indicated by arrowheads and blunt ends, respectively) may lead to rolling or fast crawling.

We have also added a section discussing the premotor circuits for crawling and rolling premotor circuit in discussion (Line 512 – 519).

(8) Percentage rolling present in figure 5B and 6A correspond to the control larvae 13xLexAop2-IVS-CsChrimson::mVenus; R72F11-lexA/+; HMS02355/+ and 13xLexAop2-IVS- Cs-Chrimson::mVenus; R72F11-lexA/+; UAS-TeTxLC.tnt/+. How does the author interpret the observed variability across the experiments? The author might consider discussing the genetic background effect on the observed behaviors, if any.

As pointed out by the reviewer, we noticed that rolling probability varied depending on genetic background. We have revised the text accordingly (Lines 277 to 280).

(9) Recheck the arrowheads in Figure 5A.

We have confirmed the positions of the arrowheads in Figure 5A and modified the figures by outlining the cells with dotted lines.

(10) Lines 295-298: Data presented in the supplementary figure and p-values in the text (p=0.11) suggest that the first crawl's onset is comparable to controls. Rewrite this text for clarity and include the statistical values in the supplemental figure 6.

We have revised the text as follows (Lines 302 to 305).

“Although the duration of each rolling bout, time to onset of the first rolling bout, and time to onset of the first crawling bout did not differ from those of controls (Figure 6–figure supplement 1D, E and G), the time to offset of the first rolling bout was delayed relative to controls (p = 0.013 for Figure 6–figure supplement 1F).”

(11) Lines 263-264: Data provide evidence for SS04185 receiving inputs Basin-2 and A00c neurons. SS04185, which provides inputs to other neurons, specifically A00c neurons, but still needs clarification.

We have revised the text as follows (Lines 264 to 266).

The results thus far indicate that, activation of SeIN128 neurons inhibits rolling (Figure 1A–C), SeIN128 neurons receive functional inputs from Basin-2 and A00c (Figure 4A-C); and SeIN128 neurons make anatomical connections onto Basin-2 and A00c (Figure 4A). 

(12) In the table that lists the genotypes, instead of '-' or the blank space in the label column, the author might consider using 'control,' consistent with the figures.

In accord with the reviewer’s suggestion, we have revised the notation of ‘-’ or the blank space, to ‘control’ for all figures.

(13) Check the typographical errors throughout the manuscript. Some below:

We have revised the text accordingly as suggested below.

a.  Lines 100, 142: SS4185 should be SS04185

b.  Line 230: A00C should be A00c

c.  Line 180: Expand VNC

d.  10xUAS-IVS-mry::GFP should be 10xUAS-IVS-myr::GFP

e.  Lines 444, 449: drosophila should be Drosophila

Author response:

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.

We thank the reviewer for their time in assessing our manuscript and are very grateful for the positive response. Below, we give a point-by-point response to the reviewer’s comments and indicate where we plan to adjust the manuscript.

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?

We thank the reviewer for their question. Indeed theoretically, if the production of Gremlin-1 exceeds the capacity of heparin to immobilise Gremlin-1, the protein may be released into solution and thus may enter the circulation. Whilst we have not addressed this possibility in our studies, we agree that it may be a mechanism worthwhile exploring in future studies.

(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?

We have thought of the possibility that Gremlin would bind other proteins such as BMPs, and thereby mask assay-antibody epitopes. To minimise this possibility, we used antibody pairs which bind different epitopes. We also used LC-MS for Gremlin-1 detection (data not shown in the manuscript), a method that is not affected by epitope masking. With the LC-MS analysis we did not pick up any gremlin-signal in plasma. We will mention the LC-MS data in the updated manuscript.

Also, we were able to detect circulating Gremlin-1 after treatment with anti-Gremlin-1 antibodies. As these were the same antibodies that were used in our assays, we should have not been able to detect Gremlin-1 if there had been a masking interaction with circulating high abundant plasma proteins such as albumin.

Finally, we believe that the assay antibodies would outcompete binding of any other proteins because of their high affinity and very high concentrations used in the assays.

In summary, we are very confident that Gremlin-1 is not present in circulation. We will though make some minor adjustments to the manuscript in order to stress this important point.

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

All reported results were derived from hard quantitative readouts obtained through assays that are not liable to subjective interpretation. This also applies to immunohistochemistry and RNAscope histologic quantification, using Visiopharm Integrator System software ver. 8.4 or HALO v3.5.3577 (Area Quantification v2.4.2 module), respectively. Therefore, no blinding was necessary 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.

We thank the reviewer for the suggestion and will add a reference to Figure 1F here.

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

As the y axis in Figure 1E is on the logarithmic scale, integer numbers would be very hard to read because of the large range of numbers. As we acknowledge that the notation used may be difficult to read, we will change it to superscript scientific notation.

(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).

Parallel to the work described in PMID 31384391 (Dutton et al., Oncotarget, 10: 4630-4639, 2019), we have tested a whole range of commercial and in-house gremlin-1 antibodies. We independently arrived at the same conclusion as Dutton et al namely that goat anti-gremlin antibody R&D Systems AF956 can stain the mouse or rat intestine in the muscularis layer and in the crypts/lower part of the villi, using FFPE sections. As per Dutton et al. we also corroborated this IHC staining by RNAscope - the mRNA was restricted to the muscularis and the connective tissue just below the crypts, suggesting that Gremlin-1 partially diffuses away from the cells that produce it. In contrast, none of the other commercial or in-house gremlin antibodies that we tested provided any useful staining on FFPE sections.

We also used the R&D Systems AF956 antibody on several rat MASH liver samples. We saw little or no staining in livers from chow-fed rats, with only occasional weak staining around portal areas. Depending on the rat model, we saw from little or no staining to at most weak staining in portal areas and fibrotic areas. Among the various models tested, we observed the strongest staining in the rat CDAA-HFD+cholesterol model, in line with the ISH data.

However, we were unable to establish IHC on human MASH liver samples using the R&D Systems AF956 antibody (or any other antibody) despite 98% sequence identity at the amino acid level between human and rat gremlin-1. Considering the results in Dutton et al. on rodent intestines, we tested the antibody on some human intestine samples, but the results on the available samples (inflamed appendices) were inconclusive.

We will include representative IHC staining images for Gremlin-1 protein on rat livers as a Supplementary Figure and mention in the manuscript that IHC for human Gremlin-1 did not work with the available antibodies.

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

The Atto-532 Gremlin-1 cell association assay was mainly intended to visualise the association of Gremlin-1 with cell surface proteoglycans and how this interaction is affected by heparin-displacing and non-displacing antibodies. We observed a possible, but inconclusive intracellular association of Atto-532 Gremlin-1. However, this assay was not specifically designed for this purpose, and we did not follow up on this. Therefore, we cannot draw any conclusions on whether cell surface bound Gremlin-1 can be internalised. However, we appreciate that internalisation of Gremlin-1 would be an interesting biological mechanism worth following up in future studies.

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

We have performed Grem1 ISH in the rat CDAA-HFD model and representative images of this are shown in Figure 1F. In chow-fed animals, Grem1 was expressed in a few cells in the portal tract, whereas after CDAA-HFD, Grem1 positive cells became more abundant in the portal tract and were also detectable in the fibrotic septa, as described in the respective results section. However, we performed no co-staining with other markers as we did for human liver samples.

(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.

In Supplementary Figure 4C we show expression of GFP mRNA and GREM1 mRNA in lysates of GFP-control and GREM1-GFP overexpressing LX-2 cells. The x-axis labels indicate the different lentiviruses. Therefore, the right panel in Supplementary Figure 4C shows that GREM1 overexpressing LX-2 cells expressed more GREM1 compared to GFP-control transduced LX-2, while GFP mRNA expression was comparable between the two.

The results in Supplementary Figure 4E look different because – as can also be seen from the % of GFP+ cells in Supplementary Figure 4D – the GREM1 lentivirus here was more effective in transducing the cells, which is why both GFP and GREM1 mRNA were increased with GREM1 lentivirus compared to the GFP-only control. Unlike LX-2, the lentivirally transduced HHSC were not sorted on GFP positive cells prior to qPCR, which may explain the differences in GFP mRNA expression pattern between the two cell types.

We acknowledge that the figure may be difficult to interpret and will adjust the figure annotation to improve on this.

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.

We thank the reviewer for their time in assessing our manuscript and are very grateful for the positive response. We acknowledge the fact that most of our results were derived from assays using in-house generated reagents which will therefore be hard to reproduce externally. Whilst for legal reasons we cannot share the sequences of the monoclonal antibodies, we will be able to share aliquots with fellow scientists upon request. We will include a sentence to this end to the data availability statement.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

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. To achieve this, the authors use a combination of computational modeling and RNAi function assays to reveal several criteria for effective RNAi-mediated silencing. This work provides insights into how (1) cis-regulatory elements influence the RNAi-mediated regulation of genes; (2) it is determined 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. These 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. The conclusions of this study are mostly supported by the data, but there are some aspects that need to be clarified.

We thank the reviewer for their kind words and for appreciating the practical utility of our approach and discoveries. 

(1) The methods do not describe the "aged RNAi plates feeding assay" in Figure 2E. The figure legend states that "aged RNAi plates" were used to trigger weaker RNAi, but the detail explaining the experiment is insufficient. How aged is aged? If the goal was to effectively reduce the dsRNA load available to the animals, why not quantitatively titrate the dsRNA provided? Were worms previously fed on the plates, or was simply a lawn of bacteria grown until presumably the IPTG on the plate was exhausted?

We have elaborated our methods section to describe that the plates were left at 4ºC for about 4 months before adding bacteria and performing the assay, with one possible reason for the weaker knockdown being that perhaps the IPTG in the RNAi plates is less effective. However, it is worth noting that the robustness of a feeding RNAi assay can vary from culture to culture and/or batch of plates. We therefore always perform RNAi assays with wild-type animals alongside test strains to gauge the strength of the RNAi assay for a given culture and batch of plates. We called the data in Figure 2E “weak” because of the response of wild-type animals was weak as evidenced by weak twitching in levamisole. Despite this reduced effect, we observed 100% penetrance in wild-type animals, enabling us to sensitively detect the reduced responses of the mutants. 

(2) Is the data presented in Figure 2F completed using the "aged RNAi plates" to achieve the partial silencing of dpy-7 observed? Clarification of this point would be helpful.

No. The only occasion when plates were older was as in response to comment 1 above.

(3) Throughout the manuscript the authors refer to "non-dividing cells" when discussing animals' ability to recover from RNA silencing. It is not clear what the authors specifically mean with the phrase "non-dividing cells", but as this is referred to in one of their major findings, it should be clarified. Do they mean the cells are somatic cells in aged animals, thus if they are "non-dividing" the siRNA pools within the cells cannot be diluted by cell division? Based on the methods, the animals of RNAi assays were L4/Young adults that were scored over 8 days after the initial pulse of dsRNA feeding. If this is the case, wouldn't these animals be growing into gravid adults after the feeding, and thus have dividing cells as they grew?

We thank the reviewer for highlighting the need to explain this point further. Our experiment test the silencing of the unc-22 gene, which is expressed and functions in body-wall muscle cells. Most of the body wall muscles in C. elegans are developed by the L1 stage (reviewed in Krause and Liu, 2012), and they do not divide between the L4 and adult stages. Therefore, during the duration of the experiment where we delivered a pulse of dsRNA and examined responses over days, none of these cells divide. We have added a statement in the main text to explicitly say that the recovery from silencing by dsRNA that we observed cannot be explained by dilution during cell divisions.

(4) What are the typical expression levels/turnover of unc-22 and bli-1? Based on the results from the altered cis-regulatory regions of bli-1 and unc-22 in Figure 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.

We can get a sense of the relative abundances of unc-22 and bli-1 across development from the RNA-seq experiments that have been performed by others in the field (see below). However, these data cannot be used to infer either the production or the turnover rates. Future experiments that measure production (the combined rate of transcriptional run-on, splicing, export from the nucleus, etc.) will be required to define the production rates. Similarly, assays that detect the rate of degradation of transcripts without confounding presence from continued production will be needed to establish turnover rates. Future efforts to obtain values for these in vivo rates for multiple genes will help further test the model.

Author response image 1.

Expression data for unc-22:

Author response image 2.

Expression data for 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).

We thank the reviewer for their kind words and for appreciating our genetic analysis.

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 mut16 + rde-10). That would provide some insights into 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.

As we highlight in the paper, MUT-16 and RDE-10 are indeed very different proteins. One possible hypothesis suggested by this difference is that different kinds of small RNAs are made when the underlying mechanism relies on MUT-16 versus on RDE-10. However, postulating such a difference is not necessary for explaining the data. Furthermore, since the amounts of 2º siRNAs do not have to be correlated with the strength of silencing (Figure 4E), this work raises caution against the over-reliance on small RNA sequencing for inferring gene silencing. Nevertheless, it is indeed an attractive possibility that the amounts of small RNA, their distributions along mRNA sequence, and/or the sequence biases of the accumulating small RNAs could be different when relying on MUT-16- or RDE-10-dependent mechanisms. Future work that directly examine the small RNAs that accumulate in different mutant strains after initiating RNAi can shed light on these possibilities.

(2) In the same line, Figure 5 aims to provide insights into the sequence determinants that influence 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.

In Figure 5, we eliminated the possibility that any gene that is transcribed using the bli-1 promoter would require NRDE-3, and showed using jam297 that modifications to the 3’ cis regulatory regions of a target can alter the dependence on NRDE-3 for knockdown. We agree that future experiments that control individual aspects of bli-1, potentially one feature at a time, can reveal the separate contributions of each characteristic of the gene to the observed dependence on NRDE-3 of the wild-type bli-1 gene. However, given the many ways that the same level of transcript knockdown can be achieved in our modeling (Figure 4 and its supplemental figures) we expect that multiple characteristics could contribute to NRDE-3 dependence. 

Recommendations For The Authors:

Reviewer #1 (Recommendations For The Authors):

(1) On page 5, the authors state that "MUT-16 and RDE-10 are redundantly or additively required for silencing unc-22"; however, based on their data in Figure 1D, it seems nearly 100% silencing of unc-22 is achieved in single mut-16 or rde-10 mutants. If this is the case, wouldn't it suggest that redundancy of MUT-16 and RDE-10, and not an "additive effect" of MUT-16 and RDE-10 function? Although, as the mutator complex nucleates around MUT-16, the data in Figure 1D suggests it is possible that the presence of MUT-16 or RDE-10 is sufficient for the recruitment of one or more factors that triggers the silencing of unc-22, and thus only one of these factors is necessary.

Because we are seeing 100% silencing in wild-type, mut-16(-), or rde-10(-) animals in Figure 1D, this assay (where the silencing response is strong) does not allow us to discriminate between differing levels of silencing. The “weak” RNAi assay in Figure 2E provides the opportunity to observe differences in the contributions made by MUT-16 or RDE-10, supporting the idea that the 2º siRNAs and relative contributions to silencing can indeed be additive, explaining the complete loss of silencing only in the double mutant. While MUT-16 has been shown to be required for the recruitment of other Mutators in the germline, Mutator foci are not detectable in the soma. Given that unc-22 and bli-1 are somatic targets, we are hesitant to assume a mechanism for the production of small RNAs that requires a similar MUT-16-dependent nucleation in somatic cells. MUT-16 is clearly required for full silencing. But, if it functions similarly in the soma and the germline remains an open question. Indeed the mechanism(s) for producing small RNAs in somatic cells could be different from that used for production of small RNAs in the germline because of known differences in the use of RNA-dependent RNA polymerases (e.g. Ravikumar et al., Nucleic Acids Res. 2019). Future studies that determine the subcellular localization(s) and potential biochemical function(s) of RDE-10 and MUT-16 in somatic cells are needed to further delineate mechanisms.

(2) On page 10, "rather than one that looks a frequency" - the "a" should be "at".

We thank the reviewer and have fixed this typo. 

(3) Figure 4 is very crowded, further dividing 4A (right) and 4B into subpanels would help the readability of the figure.

We thank the reviewer for identifying these figures as being particularly crowded. These panels are presented as single units because the left and right portions of each panel are intimately connected. In Fig. 4A, the outline of mechanism deduced on the left is based on experiments at various scales shown on the right. We have now clarified this in the figure legend. In Fig. 4B, the equations on the right define and use the constants depicted on the left and the definitions below apply to both parts. We have now adjusted both figure parts to make these connections clearer. 

(4) References to the subpanels of Figure 4 in the text on page 12 are off from the figure and figure legend.

For example:

"Overall, τkd and tkd were uncorrelated..." refers to 4C when it should refer to 4D. "However, the maximal amount of 2ºsiRNAs..." refers to 4D when it should refer to 4E. "Additionally, an increase in transcription..." refers to 4E when it should refer to 4F.

"When a fixed amount of dsRNA was exposed..." refers to 4F when it should refer to 4G.

We thank the reviewer for catching these errors and we have corrected these figure references.

Reviewer #2 (Recommendations For The Authors):

I would encourage the authors to follow up on some of the more mechanistic comments made above, that would strengthen and complement the genetic part of the work presented.

We agree that additional work is needed to elucidate differences in molecular mechanisms for amplifying small RNAs in an MUT-16-dependent vs. RDE-10-dependent manner. We hope to address these extensions of our work in future manuscripts that focus on the biochemistry of these proteins and the populations of small RNAs generated using them.

I appreciate the efforts to computationally model the dynamics of the system, but I am not sure that it helps that the mathematical modelling treats both branches of the pathway as functionally equals, since they could have some mechanistic specialisation that is not yet elucidated by the current work.

Our assumption that both branches are equivalent is the most parsimonious. If we allowed for differences, even more values for the parameters of the model will agree with experimental data. The strength of the model is that despite such conservative assumptions, it agrees with experimental data. Biochemical elaborations that make the MUT-16 and RDE-10 branches qualitatively different could exist in vivo as suggested by the reviewer. Even with such qualitative differences in detail, the overall impact on gene silencing is a quantitative and additive one as demonstrated by our experiments. Future experimental work focused on biochemistry could elucidate how a Maelstrom domain-containing protein (RDE-10) and an intrinsically disordered protein (MUT-16) act differently to ultimately promote small RNA production.

Author response:

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

Public Reviews:

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.

We appreciate your positive comments on our revised manuscript.

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 identify TMC7 as a transmembrane channel-like protein with elevated transcript levels 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 reduced sperm count, elongated spermatids, and large vacuoles. Additionally, abnormal acrosome morphology is 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 suggest that its ion channel 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 by the data that are presented.

Weaknesses:

The claim is that TMC7 functions as an ion channel. It is reasonable to assume this given what has been previously published on the more well-characterized TMCs (TMC1/2), but the data supporting this is preliminary here, and more needs to be done to solidify this hypothesis. The authors are careful in their interpretation and present this merely as a hypothesis supporting this idea.

We appreciate this constructive suggestion.

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:

The Discussion section contains sentences reiterating the Introduction and Results of this manuscript (e.g., Lines 79-85 and 231-236; Lines 175-179 and 259-263). Those read repetitive and can be removed.

We thank the reviewer for this import comment. We have modified the text according to your suggestion.

Future studies are required to decipher how TMC7 stabilizes Golgi structure, coordinates vesicle transport, and maintains the germ cell homeostasis.

Thanks. We appreciate this constructive suggestion. We totally agree the reviewer that future studies are required to decipher how TMC7 stabilizes Golgi structure, coordinates vesicle transport, and maintains the germ cell homeostasis.

Recommendations for the authors

Reviewer #1 (Recommendations For The Authors):

1. In Fig S6d, the bar of Tmc7-/- is broken in the middle for P-EIF2.

Thanks. We have remade Fig S6d according to your suggestion in the revised manuscript.

Reviewer #2 (Recommendations For The Authors):

None. The reviewers have adequately answered my points. Many thanks!

We thank the reviewer for accepting our revisions as sufficient.

Reviewer #3 (Recommendations For The Authors):

In the revised manuscript, the authors have addressed most of my concerns.

We are pleased that we were able to adequately address the reviewer’s concerns. We appreciate your suggestions to further improve our study.

Author response:

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

Reviewer #1

Summary:

In this paper, the authors performed molecular dynamics (MD) simulations to investigate the molecular basis of the 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 is important 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 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.

We thank the reviewer for pointing out the strengths of our study.

Weaknesses:

(1) Although the authors emphasized the advantage of the Martini3 force field for its explicit description of solvent, the whole paper did not discuss the water's role in the aggregation and LLPS.

We thank the reviewer for pointing this out. We agree that we have not explored or discussed the role of water in aS aggregation or LLPS. We would like to convey that we would like to explore that in detail in a separate study altogether. However we have updated the “Discussion” section with the following lines to convey to the readers the importance water plays in aggregation and LLPS of aS.

Page 24: “The significance of the solvent in alpha-synuclein (αS) aggregation remains underexplored. Recent studies [26, 55] underscore the pivotal role of water as a solvent in LLPS. It suggests that comprehending the solvent’s role, particularly water, is essential for attaining a deeper grasp of the thermodynamic and physical aspects of αS LLPS and aggregation. By delving into the solvent’s contribution, researchers can uncover additional factors influencing αS aggregation. Such insights hold the potential to advance our comprehension of protein aggregation phenomena, crucial for devising strategies to address diseases linked to protein misfolding and aggregation, notably Parkinson’s disease. Future investigations focusing on elucidating the interplay between αS, solvent (especially water), and other environmental elements could yield valuable insights into the mechanisms underlying LLPS and aggregation. Ultimately, this could aid in the development of therapeutic interventions or preventive measures for Parkinson’s and related diseases.”

(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.

We appreciate the insightful feedback from the reviewer. However, we would like to emphasize that the αS droplets exhibit a highly liquid-like behavior, characterized by frequent exchanges of chains between the dense and dilute phases, alongside a slow aggregation process. In the study by Benayad et al. (2020, JCTC) [ref. 30], FUS-LCD was the protein of choice at concentrations in the (mM) range. FUS-LCD is known to undergo very rapid LLPS at concentrations lower than 100 (μM) where for αS the critical concentration for LLPS is 500 (μM) and undergoes slower aggregation than FUS. Moreover, the diffusion constant of αS inside newly formed droplets (no liquid to solid phase transition has occurred) has been estimated to be 0.23-0.58 μm2/s (Ray et al, 2020, Nat. Comm.). The value of diffusion constant for FUS-LCD inside LLPS droplets has been estimated to be 0.17 μm2/s (Murthy et al. 2023, Nat. Struct. and Mol. Biol.). These prove that αS forms droplets that are less viscous than that formed by FUS-LCD. This dynamic nature impedes the formation of large droplets in the simulations, making it challenging to rigorously calculate surface tension from interfacial width, which, in turn, necessitates the computation of g(r) between water and the droplet.

Furthermore, it's essential to note that our primary aim in calculating surface tension was not to determine its absolute value. Rather, we aimed to compare surface tensions obtained for the three distinct environments explored in this study. Hence, our primary objective is to compare the distributions of surface tensions rather than focusing solely on the mean values obtained. The distributions shown in Figure 4a clearly show a trend which we have stated in the article.

(3) In this work, the Martini 3 force field was modified by rescaling the LJ parameters \epsilon and \sigma with a common factor \lambda. It has not been very clearly described in the manuscript why these two different parameters can be rescaled by a common factor and why it is necessary to separately tune these two parameters, instead of just tuning the coefficient \epsilon as did in a previous work [Larsen et al., PLoS Comput Biol 16: e1007870].

We thank the reviewer for the comment. We think that the distance of the first hydration layer also should have an impact on aggregation/LLPS. Here we are scaling both the epsilon and sigma. A higher epsilon of water-protein interactions mean higher the energy required for removal of water molecules (dehydration) when a chain goes from the dilute to the dense phase. A higher sigma on the other hand means that the hydration shell will also be at a larger distance making dehydration easier. Moreover, tuning both (either by same or different parameter) required a change of the overall protein-water interaction by only 1%, thereby requiring only considerably minimal change in forcefield parameters (compared to the case where only epsilon is being tuned which required 6-10% change in epsilon from its original values.) . Thus we think one of the ways of tuning water-protein interactions which requires minimal retuning of Martini 3 is by optimizing both epsilon and sigma. However whether a single scaling parameter is good enough requires further exploration and is outside the scope of the current study. More importantly it would introduce another free parameter into the system and the lesser the number of free parameters, the better. For this study, a single parameter sufficed as depicted in Figure 9. To inform the readers of why we chose to scale both sigma and epsilon, we have added the following in the main text:

Page 25-26: “Increasing the ϵ value of water-protein interactions results in a higher energy demand for removing water molecules (dehydration) as a chain transitions from the dilute to the dense phase. Conversely, a higher σ value implies that the hydration shell will be at a greater distance, facilitating dehydration if a chain moves into the dilute phase. Therefore, adjusting water-protein interactions based on the protein’s single-chain behavior may not significantly influence the protein’s phase behavior. Furthermore, fine-tuning both ϵ and σ parameters only requires a minimal change in the overall protein-water interaction (1%). As a result, this adjustment minimally alters the force field parameters.”

(4) 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 crowders with excluded volume effect or crowders with non-specific attractive interactions with proteins, etc) and therefore the force field parameters.

We thank the reviewer for a potential future direction. In this investigation our main focus was to simulate the inertness features of crowders only, to ensure that only entropic effect of the crowders are explored. Although this study focuses on the factors that enable aS to form an aggregates/LLPS under different environmental conditions, it would be interesting to explore in a systematic way the mechanism of action of crowders of varying shapes, sizes and interactions. Therefore we added the following lines in the “Discussion” section to let the readers know that this is also a future prospect of investigation.

Page 22: “Under physiological conditions, crowding effects emerge prominently. While crowders are commonly perceived to be inert, as has been considered in this investigation, the morphology, dimensions, and chemical interactions of crowding agents with αS in both dilute and dense phases may potentially exert considerable influence on its LLPS. Hence, a comprehensive understanding through systematic exploration is another avenue that warrants extensive investigation.”

Reviewer #1 (Recommendations For The Authors):

(1) Figure S1. The title of the figure and the description in the figure caption are inconsistent?

We thank the reviewer for the comment and we have updated the article with the correct caption.

(2) Page 14, line 3, the authors may want to provide more descriptions of the "ms1", "ms2", and "ms3" for better understanding.

We are grateful to the reviewer for pointing this out. We have added a line describing in brief what “ms1”, “ms2” and “ms3” represent. It reads “Subsequent to the investigation, we utilize three representative conformations, each corresponding to one of the macrostates. We designate these macrostates as 1 (ms1), 2 (ms2), and 3 (ms3) (Figure S7)” (Page 28)

(3) Page 20, the authors may want to briefly explain how the normalized Shannon entropy was calculated.

We thank the reviewer for pointing this out. This is plain Shannon Entropy and the word “normalized” should not have been there. To avoid confusion we have provided the equation we have used to calculate the Shannon entropy (Eq 8) (Page 21).

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 (αS) 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, but I have a number of concerns about how it is currently executed and presented.

We thank the reviewer for finding our study interesting.

(1) It is not clear whether the simulations have reached a steady state. If they have not, it invalidates many of their analysis methods and conclusions.

We have used the last 1 μs (1.5-2.5 1 μs) from each simulation for further analysis in this study. To understand whether the simulations have reached steady state or not, we plot the time profile of the concentration of the protein in the dilute phase for all three cases.

Author response image 1.

Except for the scenario of only αS (Figures a and b), the rest show very steady concentrations across various sections of the trajectory (Figures c-f). The larger sudden fluctuations observed inFigures a and b are due to the fact that only αS undergo very slow spontaneous aggregation and owing to the fact that the dense phase itself is very fluxional, addition/removal of a few chains to/from the dense to dilute phase register themselves as large fluctuations in the protein concentration in the dilute phase. For the other two scenarios (Figures c-f) aggregation has been accelerated due to the presence of crowders/salt. This causes larger aggregates to be formed. Therefore addition/removal of one or two chains does not significantly affect the concentration and we do not see such sudden large jumps. In summary, the large jumps seen in Figures a and b are due to slow, fluxional aggregation of pure αS and finite size effects. However as these still are only fluctuations, we posit that the systems have reached steady states. This claim is further supported by the following figure where the time profile of a few useful system wide macroscopic properties show no change between 1.5-2.5 µs.

We also have added a brief discussion in the Methods section (Page 29-30) with these figures in the Supplementary Information.

Author response image 2.

“In this study, we utilized the final 1 µs from each simulation for further analysis. To ascertain whether the simulations have achieved a steady state, we plotted the time profile of protein concentration in the dilute phase for all three cases. Except for minor intermittent fluctuation involving only αS in neat water (Figures S8a and S8b), the remaining cases exhibit notably stable concentrations throughout various segments of the trajectory (Figures S8 c-f). The relatively higher fluctuations observed in Figures S8a and b stem from the slow, spontaneous aggregation of αS alone, compounded by the inherently ambiguous nature of the dense phase.

Consequently, the addition or removal of a few chains from the dense to the dilute phase results in significant fluctuations in protein concentration within the dilute phase. Conversely, in the other two scenarios (Figures S8c-f), aggregation is expedited by the presence of crowders/salt, leading to the formation of larger aggregates. Consequently, the addition or removal of one or two chains has negligible impact on concentration, thereby mitigating sudden large jumps. In summary, the conspicuous jumps depicted in Figures S8a and b arise from the gradual, fluctuating aggregation of pure αS and finite size effects. However, since these remain within the realm of fluctuations, we assert that the systems have indeed reached steady states. This assertion is bolstered by the subsequent figure, where the time profile of several pertinent system-wide macroscopic properties reveals no discernible change between 1.5-2.5 µs (Figures S9).”

(2) The benchmarking used to validate their cgMD methods is very minimal and fails to utilize a large amount of available all-atom simulation and experimental data.

We disagree with the reviewer on this point. We have cited multiple previous studies [26, 27] that have chosen Rg as a metric of choice for benchmarking coarse-grained model and have used a reference (experimental or otherwise) to tune Martini force fields. Majority of the notable literature where Rg was used as a benchmark during generation of new coarse-grained force fields are works by Dignon et al. (PLoS Comp. Biol.) [ref. 25], Regy et al (Protein Science. 2021) [ref. 26], Joseph et al.(Nature Computational Science. 2021) [ref. 27] and Tesei et al (Open Research Europe, 2022) [ref. 28]. From a polymer physics perspective, tuning water-protein interactions is simply changing the solvent characteristics for the biopolymer and Rg has been generally considered a suitable metric in the case of coarse-grained model. Moreover we try to match the distribution of the Rg rather than only the mean value. This suggests that at a single molecule level, the cgMD simulations at the optimum water of water-protein interactions would allow the protein to sample the conformations present in the reference ensemble. We use the extensively sampled 70 μs all-atom data from DE Shaw Research to obtain the reference Rg distribution. Also we perform a cross validation by comparing the fraction of bound states in all-atom and cgMD dimer simulations which also seem to corroborate well with each other at optimum water-protein interactions. To let the readers understand the rationale behind choosing Rg we have added a section in the Methods section (Page 25) that explains why Rg is plausibly a good metric for tuning water-protein interactions in Martini 3, at least when dealing with IDPs.

Our optimized model is further supported by the FRET experiments by Ray et al. [6]. They found that interchain NAC-NAC interactions drive LLPS. Residue level contact maps obtained from our simulations also show decreased intrachain NAC-NAC interactions with an increased interchain NAC-NAC interactions inside the droplet. This corroborates well with the experimental observations and furthermore validates the metrics we have used for optimization of the water-protein interactions. However the comparison with the FRET data by Ray et al. was not present earlier and we have added the following lines in the updated draft.

Page17: “Thus we observed that increased inter-chain NAC-NAC regions facilitate the formation of αS droplets which also have previously been seen from FRET experiments on αS LLPS

droplets[6].”

(3) They also miss opportunities to compare their simulations to experimental data on aSyn protein droplets.

We thank the reviewer for pointing this out. We have tried to compare the results from our simulations to existing experimental FRET data on αS. Please see the previous response where we have described our comparison with FRET observations.

(4) Aspects such as network analysis are not contextualized by comparison to other protein condensed phases.

For a proper comparison between other protein condensed phases, we would require the position phase space of such condensates which is not readily available. Therefore we tried to explain it in a simpler manner to paint a picture of how αS forms an interconnecting network inside the droplet phase.

(5) Data are not made available, which is an emerging standard in the field.

We thank the reviewer for mentioning this. We have provided the trajectories between 1.5-2.5 μs, which we used for the analysis presented in the article, via a zenodo repository along with other relevant files related to the simulations (https://zenodo.org/records/10926368).

Firstly, it is not clear that these systems are equilibrated or at a steady state (since protein droplets are not really equilibrium systems). The authors do not present any data showing time courses that indicate the system to be reaching a steady state. This is problematic for several of their data analysis procedures, but particularly in determining free energy of transfer between the condensed and dilute phases based on partitioning.

We have addressed this concern as stated previously in the response. We have updated the article accordingly.

Secondly, the benchmarking that they perform against the 73 µs all-atom simulation of aSyn monomer by Shaw and coworkers provides only very crude validation of their cgMD models based on reproducing Rg for the monomer. The authors should make more extensive comparisons to the specific conformations observed in the DE Shaw work. Shaw makes the entire trajectory publicly available. There are also a wealth of experimental data that could be used for validation with more molecular detail. See for example, NMR and FRET data used to benchmark Monte Carlo simulations of aSyn monomer (as well as extensive comparisons to the Shaw MD trajectory) in Ferrie at al: A Unified De Novo Approach for Predicting the Structures of Ordered and Disordered Proteins, J. Phys. Chem. B 124 5538-5548 (2020)

DOI:10.1021/acs.jpcb.0c02924

I note that NMR measurements of aSyn in liquid droplets are available from Vendruscolo: Observation of an α-synuclein liquid droplet state and its maturation into Lewy body-like assemblies, Journal of Molecular Cell Biology, Volume 13, Issue 4, April 2021, Pages 282-294, https://doi.org/10.1093/jmcb/mjaa075.

In addition, there are FRET studies by Maji: Spectrally Resolved FRET Microscopy of α-Synuclein Phase-Separated Liquid Droplets, Methods Mol Biol 2023:2551:425-447. doi: 10.1007/978-1-0716-2597-2_27.

So the authors are missing opportunities to better validate the simulations and place their structural understanding in greater context. This is just based on my own quick search, so I am sure that additional and possibly better experimental comparisons can be found.

We have performed a comparison with existing FRET measurements by Ray et al. (2020) as discussed in a previous response and also updated the same in the article. The doi (10.1007/978-1-0716-2597-2_27) provided by the reviewer is however for a book on Methods to characterize protein aggregates and does not contain any information regarding the observations from FRET experiments. The other doi (https://doi.org/10.1093/jmcb/mjaa075) for the article from Vendrusculo group does not contain information directly relevant to this study. Moreover NMR measurements cannot be predicted from cgMD since full atomic resolution is lost upon coarse-graining of the protein . A past literature survey by the authors found very little scientific literature on molecular level characterization of αS LLPS droplets.

Thirdly, the small word network analysis is interesting, but hard to contextualize. For instance, the 8 Å cutoff used seems arbitrary. How does changing the cutoff affect the value of S determined? Also, how does the value of S compare to other condensed phases like crystal packing or amyloid forms of aSyn?

The 8 Å cutoff is actually arbitrary since a distance based clustering always requires a cutoff which is empirically decided. However 8 Å is quite large compared to other cutoffs used for distance based clustering. For example in ref 26, 5 Å was used as a cutoff for calculation of protein clusters. Larger cutoffs will lead to sparser network structures. However we used the same cutoff for all distance based clustering which makes the networks obtained comparable. We wanted to perform a comparison among the networks formed by αS under different environmental conditions.

Fourthly, I see no statement on data availability. The emerging standard in the computational field is to make all data publicly available through Github or some similar mechanism.

We thank the reviewer for pointing this out and we have provided the raw data between 1.5-2.5 μs for each scenario along with other relevant files via a zenodo repository (https://zenodo.org/records/10926368).

Finally, on page 16, they discuss the interactions of aSyn(95-110), but the sequence that they give is too long (seeming to contain repeated characters, but also not accurate). aSyn(95-110) = VKKDQLGKNEEGAPQE. Presumably this is just a typo, but potentially raises concerns about the simulations (since without available data, one cannot check that the sequence is accurate) and data analysis elsewhere.

This indeed is a typographical error. We have updated the article with the correct sequence. The validity of the simulations can be verified from the data we have shared via the zenodo repository (https://zenodo.org/records/10926368).

Author response:

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

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 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. 

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. 

We agree with the reviewer. We have tried many experiments to address this question. The data show that Arp 2/3 inhibition with CK666 is an effective way to inhibit initial keratinocyte migration. However, later migration still proceeds. What is interesting is that just inhibition of the early migration is sufficient to restore localized ROS production in the wound area in the first  hour post-burn, even if this is not sufficient to prevent ROS accumulation over time. There is also a trend toward improved sensory neuron function late after this early treatment. However, this is not statistically significant. We think it is likely that both migration and tissue scale ROS influence the regeneration defect of sensory neurons after burn. The data using isotonic solution supports this conclusion. We have tried many other ways to limit keratinocyte migration including depletion of talin and expression of a dominant negative Rac in basal epithelial cells, but these treatments were not compatible with survival of the fish after burn.

Pharmacological or genetic approaches should be used to prove the role of ROS production by directly targeting the known H2O2 source in the system: DUOX. 

We agree that pharmacologic or genetic approaches to directly manipulate ROS production would provide substantial support to the hypothesis that ROS, along with keratinocyte migration, is a main factor contributing to poor burn outcomes. To address this, we first tried using a morpholino to deplete DUOX. However, the combination of DUOX morpholino and burn injury was lethal to larvae. We also used pharmacologic inhibition of ROS production using DPI (Diphenyleneiodonium). With this treatment, ROS is inhibited for only the first hour post-burn as treatment is lethal for longer periods of time. Burned larvae have marginally improved axon density and touch sensitivity, suggesting the importance of ROS in burn outcomes, however it was not statistically significant. It is likely that an increased effect would be observed with longer treatment, but treatment for more than 1 hour was toxic. We have added a supplemental figure with this new DPI data.

While the authors provide clear and compelling proof that osmotic responses lie at the heart of the burn-induced axonal damage responses, they did not consider the option of further exploring any biology related to osmotic cell swelling. Could osmotic ATP release maybe play a role through excitotoxicity? Could cPLA2 activation-dependent eicosanoid production relate to the process? Pharmacological tests using purinergic receptor inhibition or blockage of eicosanoid production could answer these questions. 

We agree that the role of osmotic cell swelling in the burn response is an interesting avenue for future study. However, we make use of isotonic treatment in this study specifically for its effect on keratinocyte migration and broad-scale wound healing. As a result, we feel that pursuing the biology of this swelling phenomenon is outside the scope of this paper.

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: tail fin transections could close faster than burn wounds, allowing for lower hypotonic exposure time. Axonal damage and slow regeneration in tail fin burn wounds could be a direct consequence of extended exposure time to hypotonic water. 

We have done experiments using FM dye to test how long it takes burn and transection wounds to close (shown below). In these experiments, dye entry into wounded tissue is used as a readout of wound closure. Dye is only able to enter wounded tissue when the epithelial barrier is disrupted. Our data reveal that transections take approximately 10 minutes to fully close, while burns take approximately 20 minutes to close.

Author response image 1.

To test if this difference in wound closure time would have an effect on axon outcomes, we repeated, but slightly modified, the dual-wound experiment. We increased the amount of time the burn condition was exposed to hypotonic conditions by 10 additional minutes (by transecting burned tissue at 15 minutes post burn, shortly before closure) and compared axon outcomes to the 5 mpw control transection. These results show there was no difference in axon regeneration or function when secondary transection was performed at 5 or 15 minutes post burn, suggesting that increased exposure to hypotonic solution is not the reason for defects in axon outcomes after burn injury.

Author response image 2.

Reviewer #2 (Public Review): 

This is an interesting study in which the authors show that a thermal injury leads to extensive sensory axon damage and impaired regrowth compared to a mechanical transection injury. This correlates with increased keratinocyte migration. That migration is inhibited by CK666 drug treatment and isotonic medium. Both restrict ROS signalling to the wound edge. In addition, the isotonic medium also rescues the regrowth of sensory axons and recovery of sensory function. The findings may have implications for understanding non-optimal re-innervation of burn wounds in mammals. 

The interpretation of results is generally cautious and controls are robust. 

Here are some suggestions for additional discussion: 

The study compares burn injury which produces a diffuse injury to a mechanical cut injury which produces focal damage. It would help the reader to give a definition of wound edge in the burn situation. Is the thermally injured tissue completely dead and is resorbed or do axons have to grow into damaged tissue? The two-cut model suggests the latter. Also giving timescales would help, e.g. when do axons grow in relation to keratinocyte movement? An introductory cartoon might help. 

We thank the reviewer for these insightful comments and questions. The burn wound is defined as the area that is directly damaged as a result of increased heat (labeled by FM dye entry), and the burn wound edge as the first line of healthy cells adjacent to the burned cells. These definitions have been added to the text to clarify the areas referenced. Recent experiments lead us to believe the wound area is composed almost completely of dead cells, but we are currently working to discover the fate of these dead cells as well as the wound adjacent cells that migrate to the wound edge after burn. As a result, we do not know whether axons grow into damaged tissue or if the damaged tissue is extruded, but we do see growth cone formation within a few hours after wounding suggesting the axons are actively trying to regenerate after a burn.

Could treatment with CK666 or isotonic solution influence sensory axons directly, or through other non-keratinocyte cell types, such as immune cells? 

We have done experiments looking at the density of caudal fin innervation in CK666, isotonic, or DPI treated fins. The axon density is unchanged in all these treatments compared to control treated larvae, so we do not believe these treatments affect axon health homeostatically. These data have been added to supplemental figure 3. Additionally, one of the benefits of the larval zebrafish burn model is the simplicity of the system – the epidermis is primarily composed of sensory axons, mesenchymal cells and keratinocytes. The burn environment is proinflammatory so it does promote immune cell recruitment, but we do not believe the immune cells are interacting directly with sensory axons besides clearing axonal debris. Previous papers by our lab have shown that peak immune cell recruitment occurs at 6 hpw, but they localize to the damaged tissue in the burn area and not the wound edge.

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 the small-molecule inhibitor of Arp2/3 (CK666) resulted in decreased production of reactive oxygen species (ROS) at early, but not late, time points. 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.

The use of genetic and cell-autonomous approaches would strengthen our study, however, we were unable to do this due to the lethality of these genetic approaches (or cell autonomous approaches). Basal epithelial migration is necessary for embryonic development. We attempted to circumvent this by generation of larvae transiently expressing a dominant-negative form of Rac, a protein crucial to the migratory process. The chimeric expression of the dominant negative Rac was either damaging to the larvae or the mosaicism was too low to observe any effects on migration phenotype.

We also attempted a genetic approach to manipulate ROS production, as discussed above. We found that the DUOX morpholino was lethal to burned larvae. Finally, we attempted pharmacological inhibition of ROS production using the inhibitor DPI (Diphenyleneiodonium). With this treatment, burned larvae have marginally improved axon density and touch sensitivity, suggesting that dampening ROS may improve outcome. The DPI data have been added to 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. 

The inconsistency between “neuron” and “axon” has been noted and the text has been corrected accordingly. “Neuron” is used when referring to the cell as a whole, while “axon” is used when referring to the sensory processes in the caudal fin. We added information about burn in the introduction as suggested: “While epithelial tissue is well adapted to repair from mechanical damage, burn wounds heal poorly. Thermal injury results in chronic pain and lack of sensation in the affected tissue, suggesting that an abnormal sensory neuron response contributes to burn wound pathophysiology.”

We thank the reviewer’s for their comments.

Recommendations For The Authors:

Reviewer #1 (Recommendations For The Authors): 

Suggested experiments: 

(1) ROS measurements with the dye Pfbsf should be validated with more established ROS probes such as HyPer. 

Pfbsf has been used previously as a readout of ROS production, and its use is documented in zebrafish (Maeda et al., Angew Chem Int Ed Engl, 2004, and Niethammer et al, Nature, 2009). These sources have been added as references when introducing Pfbsf to provide context for its use. The probe was validated and compared to HyPer in Niethammer’s 2009 paper. In our hands, we have used both probes and have similar results with tail transection.

(2) To better support claims on ROS and H2O2 playing a central role in mediating axonal damage, the authors should consider pharmacological approaches such as rescue experiments with H2O2 and experiments using inhibitors such as DPI ar apocynin. 

While the above reagents and drugs have limitations and non-specific side effects, more convincing proof could result from genetic approaches including experiments on DOUX knockdown or knockout lines. 

To further dissect the role of ROS in the burn response, we conducted experiments using DPI, a potent ROS inhibitor that is well-documented in the literature. We found that 20 uM treatment of DPI (1 hour pretreatment, 1 hour post-burn) marginally improved axon density when quantified 24 hpw. Any higher dose, when in combination with a burn, proved to be lethal. Longer treatment with DPI was also not tolerated.

In addition to experiments with DPI, we attempted to burn larvae that were injected with DUOX morpholino. The combined use of burn and DUOX MO was lethal. We have dampened the conclusions and include the new data with the DPI in the revised manuscript.

Minor corrections: 

(1)A phrase/expression in the abstract is confusing: isotonic treatment does not "induce osmotic regulation". Cells exposed to hypo- or hypertonicity will respond by regulatory volume decrease or increase, respectively. Isotonic treatment maintains homeostasis. 

We appreciate this point and agree with the distinction. Revisions have been made in the text accordingly.

(2) Figures 4E and 5E would be better to show as an average of multiple experiments with statistical significance. 

The purpose of figures 4E and 5E are to demonstrate changes in fluorescence intensity and localization of ROS using the representative time series shown in 4D and 5D. The figure legend has been updated accordingly.

Reviewer #2 (Recommendations For The Authors): 

Figure 3D How can one distinguish between the two cellular elements that randomly meet or that there is actual coordination? Can the interactions be quantified? It is also unclear what the authors mean by "sensory neuron movement". The authors show that the neuronal cell bodies stay in their position, so only the axons change position. Do they do this by growth, i.e. the neuronal growth cones follow the keratinocytes or do keratinocytes displace the axon shafts? 

We have included supplemental movies that address this question in the new uploaded document. Figure 3D is comprised of still images taken from supplemental movie 2, which is a timelapse of keratinocytes/axons moving together after a burn injury.  This movie clearly shows keratinocytes and their ensheathed axons moving simultaneously, so keratinocytes are mechanically pulling sensory axon shafts with them. We have revised the text to say axon movement, not sensory neuron movement.

Over the time course of axonal movement (1 hour post-burn), it is not possible that neuronal growth cones contribute to movement, as this is too slow – previous work by other labs has shown that it takes several hours for axons to fully regenerate into amputated tissue, with movement not even noticeable until about 3 hours post-wound (Rieger and Sagasti, PLOS Biology, 2011).

Regarding the second point, “neuron” vs. “axon” is an inconsistency in the text that has been corrected. “Neuron” is used when referring to the cell as a whole, “axon” is used when referring to the processes that innervate the caudal fin. The axons are physically pulled along with keratinocytes as they migrate after burn application. From our observations, growth cones appear closer to the wound site after the movement has stopped.

Figure 4G It is surprising that the visual differences in the distribution of values are not statistically significant. 

The distribution of values in 4G was large and that is why there is no statistically-significant difference – we were also surprised at this result. We did all statistics with a statistician and this included rigorous criteria for significance.

Figure 4H The images seem to show a difference, whereas the quantification does not. I suggest choosing more representative images. 

Figure 4H has been updated to include a more representative image of axon patterning with CK666 treatment.

Figure 6A The text states that axon damage in the control and isotonic condition is comparable, yet in the image, it appears that the damage in the isotonic treatment at 0 hpw is more distal. 

This is a good observation that we consistently see in isotonic-treated fish after burn. Axon damage localizes more proximally in isotonic-treated samples because the keratinocytes distal to the notochord are likely dead, and the axons innervating those cells are likely immediately destroyed upon burn application. As a result, the distal axons are not present to express GCaMP. We believe isotonic treatment allows keratinocytes to live slightly longer, so axon damage is therefore prevented for longer. This is also the focus of continuing work to further understand the burn microenvironment.

Finally, the materials section could mention bias mitigation measures, e.g. withholding the treatment condition from the experimenter in the touch test. 

We minimized bias in experiments whenever possible, and the conservative statistical measures that were applied to our data further reduce the likelihood of false significance.

Reviewer #3 (Recommendations For The Authors): 

- Line numbers would have facilitated reviewer feedback. 

- Supplementary movies were missing in the submission. 

The lack of supplementary movies upon submission was a mistake and the movies have been uploaded along with the revised manuscript.

Introduction: 

- Pg. 3: "In response to tissue damage, sensory neurons undergo rapid and localized axonal degeneration 4,5." Not sure reference 4 (Reyes et al) is appropriate here as this study was not in the context of tissue damage. 

We have revised this section as suggested by the reviewer.

Results: 

- The expected expression pattern/localization of several transgenes was unclear. Please clearly state what cell type(s) each should label. For example, pg. 5 - "We next sought to further investigate sensory neuron function in burned tissue. For this, we assessed wound-induced axonal damage using zebrafish larvae that express the calcium probe GCaMP." Where is GCaMP expressed? 

The manuscript has been updated to include expression patterns for the included transgenes – in this mentioned case, GCaMP is expressed in neurons under the pan-neuronal Elavl3 promoter.

- Introducing the GCaMP labeling could use some clarification. Pg. 5 - "As shown previously by other groups, GCaMP labels degenerating neurons in real time35." This is confusing. Do the authors mean that GCaMP increases immediately prior to Wallerian degeneration as shown by Vargas et al. (PMID: 26558774)? 

Sustained elevated calcium levels are associated with axon damage. Previous work from other labs has shown that calcium influx follows axon injury (Ziv and Spira, EJN 1993, Adalbert et al., Neuroscience 2012). In these experiments, whenever there are CGaMP-positive punctae, this indicates axon damage. We have revised the manuscript to address this critique.

The Elavl3-GCaMP5 transgenic line will label when calcium levels increase in neurons. However, given the parameters used for imaging in our study (20x magnification, 100 ms exposure, and collection speed every 30 seconds for timelapses), we believe that only sufficiently large increases in calcium that are indicative of cell damage, and not physiological function, are being visualized.

- Figure 1E - Are these panels images of the same fish? Please specify in the legend. 

Figure 1E is comprised of one transected and one burned larva each, live-imaged over the course of six hours. The legend has been updated to include this information.

- Figure 1F - How was the damage area measured? Consider doing this measurement over time to match Figure 1E. 

Axon damage area measurements were performed similar to axon density measurements – maximum intensity z-projected confocal images of the caudal fin were generated using FIJI. For all experiments, the caudal fin area posterior to the notochord was outlined using the Polygon tool and measured to obtain a total surface area ROI. Axon fragments inside the outlined area were manually thresholded so all fragments posterior to the notochord were labeled and no saturated pixels were present, and an area measurement of these thresholded pixels was taken. We have added a section describing these measurements in the Methods section under “Axon damage quantification.”

- Pg. 5 - When introducing the ngn1 MO - please state the expected phenotype and cite the appropriate background literature_._ 

The ngn1 morpholino was cited in the Methods section with the appropriate literature (Cornell and Eisen, Development, 2002), from which we got the morpholino sequence. We thank the reviewer for pointing out the need for more introduction and clarification in the main text, so the ngn1 morpholino has been discussed in greater depth and cited in the main text as well using the same citation.

- The two-wound model is an elegant approach but could be more clearly described in the main text. 

An improved explanation of the two-wound experiment has been added to the text.

- For Figure 3, it would be helpful to have a schematic of the anatomy illustrating the relative positions of axons and epidermal cell types. 

- Figure 3C - should an additional control here be transected? Given that the krt4:lifeact transgene labels both layers of the epidermis, how were the superficial and basal keratinocytes separated? Interpretation of this section should be carefully worded. The authors state that "...suggesting that the superficial keratinocytes are being pulled by the motile basal keratinocytes" (pg.7 ) but isn't another possibility that the superficial cells are stationary? 

It is correct that the krt4:lifeact transgene labels both layers of keratinocytes, which together span 20-30 microns. These layers were separated from the same z-stack collected by confocal imaging. The first z-slice and last z-slice of the same stack were separated using FIJI and pseudocolored to appear as different colors. This clarification has been added to the Methods.

Prior observations with the krt4:lifeact and krt4:utrch (figure 3A) transgenic lines reveal that both keratinocyte layers will move distally after burn application.

- Pg. 7 - "The axons of sensory neurons are ensheathed within actin-rich channels running through basal keratinocytes 50,51." ref 51 is a C. elegans paper which does not have basal keratinocytes.

This was in error. The correct reference has replaced reference 51 (O’brien, J Comp. Neurol., 2012), in which electron microscopy is used to document the development of two layers of epithelial cells that also ensheath sensory neurons in a protective manner similar to glial cells in the central nervous system.

- Figures S1E and F - the authors state that RB and DRG soma don't move. However, it was unclear from the figure panels and legend whether the authors imaged neurons that actually innervate the caudal fin (rather than some other region of the animal). Please clarify. For comparison, Fig S1F needs a pre-injury image to be meaningful. 

The imaged cell bodies were those in the posterior trunk region, which are responsible for innervating the posterior sections of the fish including the caudal fin. From our observations, there was no movement of neuronal cell bodies after the burn.

- Figure 5 title - can the authors clarify what aspect of this figure relates to "sustained epidermal damage" 

The figure 5 title has been updated in response to the reviewer comments.

- Figure 6 - is touch sensitivity really "restored" as the authors suggest? Alternatively, sensitivity may never be lost in isotonic treatment. Or the loss may be delayed? 

We have modified the text accordingly by updating our phrasing – “restored” has been replaced with “improved” to indicate benefit over time.

- Can the authors further disentangle the effects of keratinocyte migration, ROS, and isotonic treatment on axon regeneration? For example, would the addition of CK666 to the Isotonic +1 hpw treatment improve axon regeneration? Can the authors directly manipulate ROS signaling (e.g., through exogenous addition of H2O2 or duox1 MO) to alter regeneration outcomes in their wounding assays? 

See the comments above.

- Figure 6 title - consider removing or clarifying the word "excessive" here 

The title has been revised according to the reviewer suggestion.

- hpw vs hpb were used inconsistently throughout the text 

The manuscript has been revised to use “hpw” when referring to the timeframe after injury application.

Methods: 

- Zebrafish transgenics are missing allele names 

References: 

- Many mistakes were noted in this section e.g., journal names missing, wrong authors, typos, DOIs misformatted 

The references section has been corrected to use formatting consistent with APA citation and eLife preferred guidelines.

Author response:

Generals:

We deeply appreciate the efforts by the Senior and Reviewing Editors, and also thank the three reviewers for their careful reading of the MS and their constructive comments, which are very helpful to improve our MS. We agree that we extend our efforts to elaborate the pharmacological analyses including clarification of the penetrance of GAP junction inhibitor(s), and effectiveness and specificity of the drugs. We plan to test at least L-type calcium channel blocker nifedipine. Concerning the reproducibility of the phenotypes, we indeed repeated experiments at multiple times for each of the analyses. While we demonstrated in the current version a series of representative data for simplicity along with explanation in the text that we conducted multiple times of experiments,  in a revised version we will improve the demonstration so that readers/reviewers can be convinced with the reproducibility of the data. We will also try to test other markers to look into cell types constituting the gut contractile organoid

Specifics:

Our provisional responses to “The weakness” raised by the reviewers are as follows:

Reviewer #1:

Please see the responses shown above (“Generals”).

Reviewer #2:

In addition to the responses in “Generals”, our response also includes the followings: We will look into wavelength between contractions/rhythm of the orgnaoid. We agree that our organoids derived from embryonic hind gut (E15) might not necessarily recapitulate the cell function in adult. However, it has well been accepted in the field of developmental biology that studies with embryonic tissue/cells make a huge contribution to unveil how complicated physiological cell function is underpinned. Nevertheless, we will carefully consider in the revised version so that the MS would not send misleading messages. Recent advances have also shown that 3D organoids can somehow “replace/substitute for” a complicated in vivo specimen when a particular cellular function is a focus of study.

Reviewer #3:

We appreciate a strong support of our findings.

(1) We plan to perform positive control experiments, for example, to test if the drugs we use would interfere cardiac muscle functions.

(2) We plan to do wach-out experiment to  confirm 10uM blebbistatin does not kill the cells. Thank you for this suggestion.

(3) We plan to conduct tetrodotoxin treatment. Since experiments with such toxic reagents are not enouraged by our institute, we will perform experiments with a necessary-minimum amount.

(4) We plant to address this point properly

5) It is well predictable that blebbistatin would stop the gut movement in an explanted hindgut, and it is also well established that gut contractions (movements) are concomitant with Ca2+ transients. It would indeed be interesting to see how GJ inhibitors affect such in vivo gut movement. However, since all the reviewers and the Reviewing Editor pointed out, sensitivity (concentration) and penetrance of the drug is an important point of concern, we think that the in vivo analyses will be a next step to go in near future.

(6) We have indeed noticed that contraction frequency is reduced after organoidal fusion. It seems as if cells communicate with each other to decide which rhythm they need to be adjusted to. Furthermore, contraction frequency tends to be slow down when the organoid becomes larger in size. It might be attributed to a delay in conductance between cells over growing distance. We plan to either quantify these potentially interesting phenomena or make a concise speculation in the revised version.

(7)-(10) Thank you for these comments. We will fix them.

Author response:

Public Reviews: 

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. 

We thank the reviewer for the suggestion. We have tested DTX3L RD for ssDNA binding using NMR (see Figure 4A and Figure S2), which showed that DTX3L RD binds ssDNA. We also tested the DTX3L KH domains for RNA/ssDNA binding using an FP experiment. However, the FP experiment did not show significant changes upon titrating RNA/ssDNA. It seems that the KH domains alone are not sufficient to bind RNA/ssDNA and both KH and RD domains are required for binding. Understanding how DTX3L binds RNA/ssDNA is an ongoing research in the lab. We will revise the Discussion on the KH domains.

(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? 

We will address this question in the revision.

(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. 

We have tried to use mass spectrometry to detect the linkage between Ub and ssDNA/RNA, but was unable to do so. We suspect that the oxyester linkage might be labile, posing a challenge for mass spectrometry techniques. Similarly, a recent preprint from Ahel lab, which utilises LC-MS, detects the Ub-NMP product rather than the linkage (https://www.biorxiv.org/content/10.1101/2024.04.19.590267v1.full.pdf).

(4) Furthermore, can a targeted MS approach be used to show that nucleotides are ubiquitylated in cells? 

This will require future development and improvement of the MS approach, specifically the isolation of labile oxyester-linked products from cells and the optimisation of the MS detection method.

(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.

We only collected HSQC spectra which was insufficient for assignments. We have performed a competition experiment using ADPr and labelled ssDNA, showing that ADPr competes against the ubiquitination of ssDNA (Figure 4D). We will provide an additional experiment showing that ssDNA with a blocked 3’-OH can compete against ubiquitination of ADPr. These data, together with our NMR analysis, will further strengthen the evidence that ssDNA and ADPr compete the same binding pocket in DTX3L RD. Understanding how DTX3L RD binds ssDNA/RNA is an ongoing research in the lab.

(6) Does sequence analysis help explain the specificity of activity for the family of proteins? 

We will performed sequence alignment of DTX proteins RD domains and discuss this point in the revision.

(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. 

We will modify this Figure as suggested.

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. 

We thank the reviewer for the constructive feedback. We agree that evidence for the biological function is lacking. While we have tried to detect Ub-ssDNA/RNA from cells, we found that Isolating and detecting labile oxyester-linked Ub-ssDNA/RNA products remain challenging due to (1) low levels of Ub-ssDNA/RNA products, (2) the presence of DUBs and nucleases that rapidly remove the products during the experiments, and (3) our lack of a suitable MS approach to detect the product. For these reasons, we feel that discovering the biological function will require future effort and expertise and is beyond the scope of our current manuscript.

In the manuscript (PMID: 38000390), the authors used PARP10 to catalyse ADP-ribosylation onto 5’-phosphorylated ssDNA/RNA. They used the following sequences which lacks 3’-adenosine, which could explain the lack of ubiquitination.

E15_5′P_RNA [Phos]GUGGCGCGGAGACUU

E15_5′P_DNA [Phos]GTGGCGCGGAGACTT

We will perform the experiment using this sequence to verify this. We have cited this manuscript but for some reasons, Pubmed has updated its published date from mid 2023 to Jan 2024. We will update the Endnote in the revised manuscript.

We agree that it is crucial to compare ubiquitination of oligonucleotides and ADPr by DTX3L to find its preferred substrate. We have challenged oligonucleotide ubiquitination by adding excess ADPr and found that ADPr efficiently competes with oligonucleotide (Figure 4D). We will perform more thorough competition experiments by titrating with increasing molar excess of either ADPr or ssDNA to examine the effect on the ubiquitination of ssDNA and ADPr, respectively.

Author response:

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

Reviewer #1 (Recommendations For The Authors):

Inclusion of other catalase, peroxidase or superoxide dismutase gene promoters (with ChiP-seq screen shots) and whether they contain sntB binding sites is important to provide other potential downstream pathways controlling oxidative stress mediated regulation of development and aflatoxin metabolism. This can be presented as supplementary material.

or

Some more examples of ChiP-seq peaks in the promoters of nsdC, nsdD, sclR, steA, wetA, veA, fluG, sod2, catA, catC would strengthen the paper for the reliability of the ChiP-seq data. Currently, visualisation of the ChIP-seq data is only limited to catC gene promoter, where background ChIP-seq signals are very high (Figure 5F).

The binding region and motif of SntB on the catA, catB, sod1, and sod2 genes were shown in Figure S7 and described in lane 531-536 and 881-884. The background of ChIP-seq signals is high, but the enrich level in the ip-sntB-HA samples is significant compared to IP-WT.

Figure 5F, letters are too small, and difficult to read. The same is true for Figure 4. Letters should be enlarged for the readers to read it without problem.

Thanks. We have revised the Figure 5F and Figure 4. Please see these Figures.

Reviewer #2 (Recommendations For The Authors):

The authors fully addressed my concerns and made appropriate changes in the manuscript. The quality of the manuscript is now improved.

Thanks. We would like to express our sincere gratitude for your affirmation and thoughtful feedback. Your positive comments have been extremely encouraging and have strengthened my confidence in my work. Your time and effort in reviewing my submission are greatly appreciated.

Author response:

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

Reviewer #1 (Public Review):

(1) Only one PITAR siRNA was tested in majority of the experiments, which compromises the validity of the results.

We thank the reviewer for this comment. We have now used two siRNAs to demonstrate PITAR functions in various assays. In the revised manuscript, we carried out additional experiments with two siRNAs, and the results are presented in Figures 2C, D, F, G, H, I, and J; Figures 5A, B, Supplementary Figure 2B, C, D, E, and F.

(2) 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.

The reason for the fact that PITAR silencing showed a robust G1 arrest, unlike PITAR overexpression, is as follows. Since glioma cells overexpress PITAR (which keeps the p53 suppressed), silencing PITAR (which will elevate p53 levels) in glioma cells shows a robust phenotype in cell cycle profile (in the form of increased G1 arrest). In contrast, the overexpression of PITAR in glioma cells fails to show robust changes in the cell cycle profile because glioma cells already have high levels of PITAR.

(3) 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.

We disagree with the fact that PITAR silencing only marginally increased the stability of p53. In the cycloheximide experiment in Figure 5E, the half-life of p53 is increased by 60 % (50 mins to 120 mins), which is quite significant in altering the DNA damage response by p53. Further, we also want to point out that the other arm of p53 degradation by Mdm2 remains intact under these conditions. We also provide an improved p53 ubiquitination western blot in the revised version (Figure 5F). 

(4) 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.

We thank the reviewer. In the revised manuscript, we demonstrate that PITAR overexpression fails to inhibit p53 in TRIM28 silenced cells (Supplementary Figure 5G; Figure 5K, L, M, N).

(5) Note that only one cell line was investigated in Fig 5.

In revised manuscript, the impact of PITAR silencing and PITAR overexpression on p53 functions are demontsrared for one more glioma cell line (Supplemenatry Figure 5B, C, D, and E).

(6) 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.

We would like to respectfully disagree with the reviewer on this particular point.  We have indeed provided the following evidence in the first version of the manuscript: glioblastoma-promoting activities of PITAR were mediated by its regulation of p53 or TRIM28.

(1) To show the importance of p53:

We show that PITAR silencing failed to inhibit the colony growth of p53-silenced U87 glioma cells (U87/shp53#1). We also show that while PITAR silencing decreased TRIM28 RNA levels in U87/shNT and U87/shp53#1 glioma cells, it failed to increase CDKN1A and MDM2 (p53 targets) at the RNA level in U87/shp53#1 cells unlike in U87/siNT cells (Supplementary Figure 6 Panels A, B, C, and D). 

(2) To show the importance of TRIM28 and p53:

The importance of p53 is also demonstrated in the context of patient-derived GSC lines. We demonstrate that PITAR silencing-induced reduction in the neurosphere growth (WT p53 containing patient-derived GSC line) is accompanied by a reduction in TRIM28 RNA and an increase in the CDKN1A RNA without a change in p53 RNA levels (Supplementary Figure 7 Panels A, B, C, D, and E). We also demonstrate that PITAR overexpression-induced neurosphere growth is accompanied by an increase in the TRIM28 RNA, and a decrease in CDKN1A RNA without a change in p53 RNA levels (Supplementary Figure 7 Panels F, G, H, and I). However, PITAR silencing failed to decrease neurosphere growth in mutant p53 containing GSC line (MGG8) (Supplementary Figure 7 Panels J, K, L, M, N, and F).

(3) We show that the TRIM28 protein level is drastically reduced in small tumors formed by U87/siPITAR cells (Supplementary Figure 7 Panel E).

(4) We show that glioma tumors formed by U87/PITAR OE cells express high levels of TRIM28 protein but reduced levels of p21 protein (Supplementary Figure 7 Panel B).

Further, we did additional experiments to prove the importance of TRIM28.

In the revised manuscript, we have carried out an additional experiment to prove the requirement of TRIM28 for tumor-promoting functions of PITAR overexpression. Earlier, we have shown that exogenous overexpression of PITAR promotes glioma tumor growth and imparts resistance to Temozolomide chemotherapy (Figure 7F and G; Supplementary Figure 9A and B). In the revised manuscript, we show that the tumor growth-promoting function of PITAR overexpression requires TRIM28. U87-Luc/PITAR OE cells formed a larger tumor compared to U87-Luc/VC cells (Figure 7H, and I; compare red line with blue line). U87-Luc/shTRIM28 cells formed very small-sized tumors (Figure 7H, and I; compare green line with blue line). Further, PITAR overexpression (U87-Luc/PITAR OE) was less efficient in promoting glioma tumor growth in TRIM28 silenced cells (Figure 7H, and I; compare pink line with red line). Thus, we prove that, as a whole, TRIM28 mediates the tumor growth-promoting functions of PITAR.

(7) 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.

Reviewer #1 rightly pointed out that TMZ induces PITAR expression, which should compromise TMZ's anti-cancer activity.

We demonstrate the same as below:

Figure 7F&G demonstrates the following two facts:1. PITAR overexpression increases the glioma-tumor growth (Figure 7G, compare red line with the blue line), 2. PITAR overexpressing glioma tumors are resistant to TMZ chemotherapy (Figure 7G, compare the pink line with the green line).

In addition, Figure 7 F and G also demonstrate that TMZ treatment of tumors formed by U87/VC glioma cells inhibited the growth but not eliminated the tumor growth completely (compare pink line with blue line). We believe that the inability of TMZ to eliminate the tumor growth completely is because of the chemoresistance imparted by the DNA damage induced PITAR.

Further, in Figure 2I, we indeed show that PITAR-silenced cells are more sensitive to TMZ and Adriamycin chemotherapy.

(8) 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.

We respect the views and questions raised by the reviewer.

We would like explain as below the importance of our model.

Yes, it is true that DNA damage induces p53. We show here that DNA damage also induces PITAR in a p53-independent manner, which, in turn, inhibits p53. Here is our explanation. Even though DNA damage activates p53, there exists an autoregulatory negative feedback loop that controls the extent and duration of p53 response to DNA damage (Wu et al., 1993; Haupt et al., 1997; Kubbutat, Jones and Vousden, 1997; Zhang et al., 2009).  It is proposed that the p53-Mdm2 feedback loop generates a “digital clock” that releases well-timed quanta of p53 until the damage is repaired or the cell dies (Lahave et al., 2004). In addition, it has also been shown that TRIM28, through its association with Mdm2, also contributes to p53 inactivation (Wang et al., 2005b; Czerwińska, Mazurek, and Wiznerowicz, 2017).

Based on the above reports and our current work, we propose that DNA damage-induced PITAR, through its ability to increase the TRIM28 levels, contributes to the control of the DNA damage response of p53 along with Mdm-2. The difference is as follows: Since Mdm-2 is also a transcriptional target of p53, the p53-Mdm-2 axis is an autoregulatory negative feedback loop to control the DNA damage response by p53. In contrast, PITAR is not a transcriptional target of p53, and DNA damage-induced activation of PITAR is p53-independent. Hence, the PITAR-TRIM28 axis in controlling the DNA damage response of p53 creates an Incoherent feedforward regulatory network.  The experimental evidence provided in the revised manuscript is as follows: 1) We have already (the first version of the manuscript) shown that exogenous overexpression of PITAR significantly inhibits DNA damage-induced p53 (Figures 6A, B, C, and D). 2) In the revised manuscript, we show that the DNA damage response of p53 (duration and extent of p53 activation after a pulse of ionizing radiation) in PITAR-silenced cells follows similar kinetics in terms of duration, but the extent of p53 activation was much stronger (Supplementary figures 8H, I, J, and K).  This is because the TRIM28 component in TRIM28/Mdm-2 axis is compromised as PITAR silencing reduces the TRIM28 levels. 3) We also demonstrate that DNA damage-induced TRIM28 is dependent on PITAR (Figure 6K; Supplementary Figure 5G)

Reviewer #1(Recommendations For The Authors):

(1) Fig 7A, what is the explanation for the observation that tumors disappeared in most of the mice in the siPITAR group? Did the authors check if apoptosis was induced here?

We agree to the point that the lack of tumor growth in the siPITAR group is likely due to the induction of apoptosis. We would like to point out that in vitro experiments indeed demonstrate that PITAR silencing induces apoptosis in Figure 2H and Supplementary Figure 2F.

(2) The authors need to explain why Fig 6 used a cell line different from other experiments. It would be better to check other cell lines.

The purpose of RG5 and MGG8 is as follows. 1) We wanted to establish the growth-promoting functions of PITAR in patient-derived GSC lines. 2) We also wanted to show the importance of WT p53 for the growth-promoting functions of PITAR.

However, in the revised manuscript we moved this portion under the subsection “PITAR inhibits p53 protein levels by its association with TRIM28 mRNA“.

Further,the experiments related to DNA damage induced activation of PITAR in p53-independent manner and its impact on DNA damage response by p53 is moved to a new section entitled “PITAR is induced by DNA damage in a p53-independent manner, which in turn diminishes the DNA damage response by p53”

(3) It would be more convincing if the authors could test more p53 target genes in addition to p21.

We thank the reviewer for this comment and the specific suggestions for checking additional p53 targets. In the revised manuscript, we have checked the MDM2 transcript levels in Supplementary Figure 6D. 

Reviewer #2 (Recommendations For The Authors):

(1) In the text, they mentioned " Figure 4J". There is no Figure 4J in Figure 4. It may be Figure 4K.

We thank reviewer #2. We corrected this information in the revised manuscript.

(2) The molecular weight markers in Western blots were missed in several Figure panels, including Figure 4J, Figure 5K, and Supple. Figure 3B, Supple. Figure 5G, H, Supple. Figures 6A and 7A.

We thank reviewer #2, and we have included the molecular weight markers in all the mentioned figures.

Author response:

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

eLife assessment:

This useful modeling study explores how the biophysical properties of interneuron subtypes in the basolateral amygdala enable them to produce nested oscillations whose interactions facilitate functions such as spike-timing-dependent plasticity. The strength of evidence is currently viewed as incomplete because of insufficient grounding in prior experimental results and insufficient consideration of alternative explanations. This work will be of interest to investigators studying circuit mechanisms of fear conditioning as well as rhythms in the basolateral amygdala.

We disagree with the overall assessment of our paper. The current reviews published below focus on two kinds of perceived inadequacies. Reviewer 1 (R1) was concerned that the fear conditioning paradigm used in the model is not compatible with some of the experiments we are modeling. The reviewer helpfully suggested in the Recommendations for the Authors some papers, which R1 believed exposed this incompatibility. In our reading, those data are indeed compatible with our hypotheses, as we will explain in our reply. Furthermore, the point raised by R1 is an issue for the entire field. We will suggest a solution to that issue based on published data.

Reviewer 2 (R2) said that there is no evidence that the BLA is capable of producing, by itself, the rhythms that have been observed during fear conditioning in BLA and, furthermore, that the paper we cited to support such evidence, in fact, refutes our argument. We believe that the reasoning used by reviewer 2 is wrong and that the framework of R2 for what counts as evidence is inadequate. We spell out our arguments below in the reply to the reviewers.

Finally, we believe this work is of interest far beyond investigators studying fear conditioning. The work shows how rhythms can create the timing necessary for spike-timing-dependent plasticity using multiple time scales that come from multiple different kinds of interneurons found both in BLA and, more broadly, in cortex. Thus, the work is relevant for all kinds of associative learning, not just fear conditioning. Furthermore, it is one of the first papers to show how rhythms can be central in mechanisms of higher-order cognition.

Reviewer #1

We thank Reviewer 1 for his kind remarks about our first set of responses and their understanding of the importance of the work. There was only one remaining point to be addressed:

Deficient in this study is the construction of the afferent drive to the network, which does elicit activities that are consistent with those observed to similar stimuli. It still remains to be demonstrated that their mechanism promotes plasticity for training protocols that emulate the kinds of activities observed in the BLA during fear conditioning.

It is true that some fear conditioning protocols involve non-overlapping US and CS, raising the question of how plasticity happens or whether behavioral effects may happen without plasticity. This is an issue for the entire field (Sun et al., F1000Research, 2020). Several papers (Quirk, Repa and LeDoux, 1995; Herry et al, 2007; Bordi and Ledoux 1992) show that the pips in auditory fear conditioning increase the activity of some BLA neurons: after an initial transient, the overall spike rate is still higher than baseline activity. The question remains as to whether the spiking is sustained long enough and at a high enough rate for STDP to take place when US is presented sometime after the stop of the CS.

Experimental recordings cannot speak to the rate of spiking of BLA neurons during US due to recording interference from the shock. However, evidence seems to suggest that ECS activity should increase during the US due to the release of acetylcholine (ACh) from neurons in the basal forebrain (BF) (Rajebhosale et al., 2024). Pyramidal cells of the BLA robustly express M1 muscarinic ACh receptors (Muller et al., 2013; McDonald and Mott, 2021) and M1 receptors target spines receiving glutamatergic input (McDonald et al., 2019). Thus, ACh from BF should elicit a long-lasting depolarization in pyramidal cells. Indeed, the pairing of ACh with even low levels of spiking of BLA neurons results in a membrane depolarization that can last 7 – 10 s (Unal et al., 2015). This implies that the release of ACh can affect the consequences of the CS in successive trials. This should include higher spiking rates and more sustained activity in the ECS neurons after the first presentation of US, thus ensuring a concomitant activation of ECS and fear (F) neurons necessary for STDP to take place. Hence, we suggest that a solution to the problem raised by R1 may be solved by considering the role of ACh release by BF. To the best of our knowledge, there is nothing in the literature that contradicts this potential solution. The model we have may be considered a “minimal” model that puts in by hand the higher frequency due to the cholinergic drive without explicitly modeling it. As R1 says, it is important for us to give the motivation of that higher frequency; in the next revision, we will be explicit about how the needed adequate firing rate can come about without an overlap of CS and US in any given trial.

Reviewer #2

The authors of this study have investigated how oscillations may promote fear learning using a network model. They distinguished three types of rhythmic activities and implemented an STDP rule to the network aiming to understand the mechanisms underlying fear learning in the BLA.

After the revision, the fundamental question, namely, whether the BLA networks can or cannot intrinsically generate any theta rhythms, is still unanswered. The author added this sentence to the revised version: "A recent experimental paper, (Antonoudiou et al., 2022), suggests that the BLA can intrinsically generate theta oscillations (3-12 Hz) detectable by LFP recordings under certain conditions, such as reduced inhibitory tone." In the cited paper, the authors studied gamma oscillations, and when they applied 10 uM Gabazine to the BLA slices observed rhythmic oscillations at theta frequencies. 10 uM Gabazine does not reduce the GABA-A receptor-mediated inhibition but eliminates it, resulting in rhythmic populations burst driven solely by excitatory cells. Thus, the results by Antonoudiou et al., 2022 contrast with, and do not support, the present study, which claims that rhythmic oscillations in the BLA depend on the function of interneurons. Thus, there is still no convincing evidence that BLA circuits can intrinsically generate theta oscillations in intact brain or acute slices. If one extrapolates from the hippocampal studies, then this is not surprising, as the hippocampal theta depends on extra-hippocampal inputs, including, but not limited to the entorhinal afferents and medial septal projections (see Buzsaki, 2002). Similarly, respiratory related 4 Hz oscillations are also driven by extrinsic inputs. Therefore, at present, it is unclear which kind of physiologically relevant theta rhythm in the BLA networks has been modelled.

Reviewer 2 (R2) says “the fundamental question, namely, whether the BLA networks can or cannot intrinsically generate any theta rhythms, is still unanswered.” In our revision, we cited (Antonoudiou et al., 2022), who showed that BLA can intrinsically generate theta oscillations (3-12 Hz) detectable by LFP recordings. R2 pointed out that this paper produces such theta under conditions in which the inhibition is totally removed. R2 then states that the resulting rhythmic populations burst at theta “are driven solely by excitatory cells. Thus, the results by (Antonoudiou et al., 2022) contrast with, and do not support, the present study, which claims that rhythmic oscillations in the BLA depend on the function of interneurons. Thus, there is still no convincing evidence that BLA circuits can intrinsically generate theta oscillations in intact brain or acute slices.”

This reasoning of R2 is faulty. With all GABAergic currents omitted, the LFP is composed of excitatory currents and intrinsic currents. Our model of the LFP includes all synaptic and membrane currents. In our model, the high theta comes from the spiking activity of the SOM cells, which increase their activity if the inhibition from VIP cells is removed. We are including a new simulation, which models the activity of the slice in the presence of kainate (as done in Antonoudiou et al., 2022), providing additional excitation to the network. If the BLA starts at high excitation, our model produces an ongoing gamma in the VIP cells that suppress SOM cells and allows a PING gamma to form between PV and F cells; with Gabazine (modeled as the removal of all the GABAergic synapses), this PING is no longer possible and so the gamma rhythm disappears. As expected, the simulation shows that the model produces theta with Gabazine; the model also shows that a PING rhythm is produced without Gabazine, and that this rhythm goes away with Gabazine because PING requires feedback inhibition (see Author response image 1). Thus, the theta increase with Gabazine in the (Antonoudiou et al., 2022) paper can be reproduced in our model, so that paper does support the model.

Author response image 1.

Spectral properties of the BLA network without (black) versus with Gabazine (magenta). Power spectra of the LFP proxy, which is the linear sum of AMPA, GABA (only present in the absence of Gabazine, D-, NaP-, and H-currents. Both power spectra are represented as mean and standard deviation across 10 network realizations. Bottom: inset between 35 and 50 Hz.

Nevertheless, we agree that this paper alone is not sufficient evidence that the BLA can produce a low theta. We have recently learned of a new paper (Bratsch-Prince et al., 2024) that is directly related to the issue of whether the BLA by itself can produce low theta, and in what circumstances. In this study, intrinsic BLA theta is produced in slices with ACh stimulation (without needing external glutamate input) which, in vivo, would be produced by the basal forebrain (Rajebhosale et al., eLife, 2024) in response to salient stimuli. The low-theta depends on muscarinic activation of CCK interneurons, a group of interneurons that overlaps with the VIP neurons in our model (Krabbe 2017; Mascagni and McDonald, 2003).

We suspect that the low theta produced in (Bratsch-Prince et al., 2024) is the same as the low theta in our model. We do not explicitly include ACh modulation of BLA in our paper, but in current work with experimentalists, we aim to show that ACh is essential to the theta by activating the BLA VIP cells. In our re-revised version, we will discuss Bratsch-Prince et al., 2024 and its connection to our hypothesis that the theta oscillations can be produced within the BLA.

Note that we have already included a paragraph stating explicitly that our hypothesis in no way contradicts the idea that inputs to the BLA may include theta oscillations. Indeed, the following paragraphs in the revised paper describe the complexity of trying to understand the origin of brain rhythms in vivo. R2 did not appear to take this complexity, and the possible involvement of neuromodulation, into account in their current position that the theta rhythms cannot be produced intrinsically in the BLA.

From revised paper: “Where the rhythms originate, and by what mechanisms. A recent experimental paper, (Antonoudiou et al. 2022), suggests that the BLA can intrinsically generate theta oscillations (3-12 Hz) detectable by LFP recordings under certain conditions, such as reduced inhibitory tone. They draw this conclusion in mice by removing the hippocampus, which can volume conduct to BLA, and noticing that other nearby brain structures did not display any oscillatory activity. Our model also supports the idea that intrinsic mechanisms in the BLA can support the generation of the low theta, high theta, and gamma rhythms.

Although the BLA can produce these rhythms, this does not rule out that other brain structures also produce the same rhythms through different mechanisms, and these can be transmitted to the BLA. Specifically, it is known that the olfactory bulb produces and transmits the respiratory-related low theta (4 Hz) oscillations to the dorsomedial prefrontal cortex, where it organizes neural activity (Bagur et al., 2021). Thus, the respiratory-related low theta may be captured by BLA LFP because of volume conduction or through BLA extensive communications with the prefrontal cortex. Furthermore, high theta oscillations are known to be produced by the hippocampus during various brain functions and behavioral states, including during spatial exploration (Vanderwolf, 1969) and memory formation/retrieval (Raghavachari et al., 2001), which are both involved in fear conditioning. Similarly to the low theta rhythm, the hippocampal high theta can manifest in the BLA. It remains to understand how these other rhythms may interact with the ones described in our paper.”

We believe our current paper is important to show how detailed biophysical modeling can unearth the functional implications of physiological details (such as the biophysical bases of rhythms), which are often (indeed, usually) ignored in models, and why rhythms may be essential to some cognitive processes (including STDP). Indeed, for evaluating our paper it is necessary to go back to the purpose of a model, especially one such as ours, which is “hypothesis/data driven”. The hypotheses of the model serve to illuminate the functional roles of the physiological details, giving meaning to the data. Of course, the hypotheses must be plausible, and we think that the discussion above easily clears that bar. Hypotheses should also be checked experimentally, and a model that explains the implications of a hypothesis, such as ours, provides motivation for doing the hard work of experimental testing. We think that R1 understands this and has been very helpful.

—————

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

eLife assessment

This useful modeling study explores how the biophysical properties of interneuron subtypes in the basolateral amygdala enable them to produce nested oscillations whose interactions facilitate functions such as spike-timing-dependent plasticity. The strength of evidence is currently viewed as incomplete because the relevance to plasticity induced by fear conditioning is viewed as insufficiently grounded in existing training protocols and prior experimental results, and alternative explanations are not sufficiently considered. This work will be of interest to investigators studying circuit mechanisms of fear conditioning as well as rhythms in the basolateral amygdala. 

Most of our comments below are intended to rebut the sentence: “The strength of evidence is currently viewed as incomplete because the relevance to plasticity induced by fear conditioning is viewed as insufficiently grounded in existing training protocols and prior experimental results, and alternative explanations are not sufficiently considered”. 

We believe this work will be interesting to investigators interested in dynamics associated with plasticity, which goes beyond fear learning. It will also be of interest because of its emphasis on the interactions of multiple kinds of interneurons that produce dynamics used in plasticity, in the cortex (which has similar interneurons) as well as BLA. We note that the model has sufficiently detailed physiology to make many predictions that can be tested experimentally. Details are below in the answer to reviewers.

Reviewer #1 (Public Comments):  

(1) … the weakness is that their attempt to align with the experimental literature (specifically Krabbe et al. 2019) is performed inconsistently. Some connections between cell types were excluded without adequate justification (e.g. SOM+ to PV+). 

In order to constrain our model, we focused on what is reported in (Krabbe et al., 2019) in terms of functional connectivity instead of structural connectivity. Thus, we included only those connections for which there was strong functional connectivity. For example, the SOM to PV connection is shown to be small (Krabbe et al., 2019, Supp. Fig. 4, panel t). We also omitted PV to SOM, PV to VIP, SOM to VIP, VIP to excitatory projection neurons; all of these are shown in (Krabbe et al. 2019, Fig. 3 (panel l), and Supp. Fig. 4 (panels m,t)) to have weak functional connectivity, at least in the context of fear conditioning. 

We reply with more details below to the Recommendations for the Authors, including new text.

(2) The construction of the afferent drive to the network does not reflect the stimulus presentations that are given in fear conditioning tasks. For instance, the authors only used a single training trial, the conditioning stimulus was tonic instead of pulsed, the unconditioned stimulus duration was artificially extended in time, and its delivery overlapped with the neutral stimulus, instead of following its offset. These deviations undercut the applicability of their findings.  

Regarding the use of a single long presentation of US rather than multiple presentations (i.e., multiple trials): in early versions of this paper, we did indeed use multiple presentations. We were told by experimental colleagues that the learning could be achieved in a single trial. We note that, if there are multiple presentations in our modeling, nothing changes; once the association between CS and US is learned, the conductance of the synapse is stable. Also, our model does not need a long period of US if there are multiple presentations.  

We agree that, in order to implement the fear conditioning paradigm in our in-silico network, we made several assumptions about the nature of the CS and US inputs affecting the neurons in the BLA and the duration of these inputs. A Poisson spike train to the BLA is a signal that contains no structure that could influence the timing of the BLA output; hence, we used this as our CS input signal. We also note that the CS input can be of many forms in general fear conditioning (e.g., tone, light, odor), and we wished to de-emphasize the specific nature of the CS. The reference mentioned in the Recommendations for authors, (Quirk, Armony, and LeDoux 1997), uses pulses 2 seconds long. At the end of fear conditioning, the response to those pulses is brief. However, in the early stages of conditioning, the response goes on for as long as the figure shows. The authors do show the number of cells responding decreases from early to late training, which perhaps reflects increasing specificity over training. This feature is not currently in our model, but we look forward to thinking about how it might be incorporated. Regarding the CS pulsed protocol used in (Krabbe et al., 2019), it has been shown that intense inputs (6kHz and 12 kHz inputs) can lead to metabotropic effects that last much longer than the actual input (200 ms duration) (Whittington et al., Nature, 1995). Thus, the effective input to the BLA may indeed be more like Poisson.

Our model requires the effect of the CS and US inputs on the BLA neuron activity to overlap in time in order to instantiate fear learning. Despite paradigms involving both overlapping (delay conditioning, where US coterminates with CS (Lindquist et al., 2004), or immediately follows CS (e.g., Krabbe et al., 2019)) and non-overlapping (trace conditioning) CS/US inputs existing in the literature, we hypothesized that concomitant activity in CS- and US-encoding neuron activity should be crucial in both cases. This may be mediated by the memory effect, as suggested in the Discussion of our paper, or by metabotropic effects as suggested above, or by the contribution from other brain regions. We will emphasize in our revision that the overlap in time, however instantiated, is a hypothesis of our model. It is hard to see how plasticity can occur without some memory trace of US. This is a consequence of our larger hypothesis that fear learning uses spiketiming-dependent plasticity; such a hypothesis about plasticity is common in the modeling literature. 

We reply with more details below to the Recommendations for the Authors, including new text.

Reviewer #1 (Recommendations For The Authors): 

Major points: 

(1) This paper draws extensively from Krabbe et al. 2019, but it does not do so consistently. The paper would be strengthened if it tried to better match the circuit properties and activations.

Specifically: 

a. Krabbe found that PV interneurons were comparably activated by the US (see Supp Fig 1). Your model does not include that. The basis for the Krabbe 2019 claim that PV US responses are weaker is that they have a slightly larger proportion of cells inhibited by the US, but this is not especially compelling. In addition, their Fig 2 showed that VIP and SOM cells receive afferents from the same set of upstream regions. 

b. The model excluded PV-SOM connections, but this does not agree with Krabbe et al. 2019, Table 2. PV cells % connectivity and IPSC amplitudes were comparable to those from VIP interneurons. 

c. ECS to PV synapses are not included. This seems unlikely given the dense connectivity between PV interneurons and principal neurons in cortical circuits and the BLA (Woodruff and Sah 2007 give 38% connection probability in BLA). 

We thank the Reviewer for raising these points, which allow us to clarify how we constrained our model and to do more simulations. Specifically: 

a. (Wolff et al., Nature, 2014), cited by (Krabbe et al. 2018), reported that PV and SOM interneurons are on average inhibited by the US during the fear conditioning. However, we agree that (Krabbe et al., 2019) added to this by specifying that PV interneurons respond to both CS+ and US, although the fraction of US-inhibited PV interneurons is larger. As noted by the Reviewer, in the model we initially considered the PV interneurons responding only to CS+ (identified as “CS” in our manuscript). For the current revision, we ran new simulations in which the PV interneuron receives the US input, instead of CS+. It turned out that this did not affect the results, as shown in the figure below: all the network realizations learn the association between CS and fear. In the model, the PING rhythm between PV and F is the crucial component for establishing fine timing between ECS and F, which is necessary for learning. Having PV responding to the same input as F, i.e., US, facilitates their entrainment in PING and, thus, successful learning. 

As for afferents of VIP and SOM from upstream regions, in (Krabbe et al., 2019) is reported that “[…] BLA SOM interneurons receive a different array of afferent innervation compared to that of VIP and PV interneurons, which might contribute to the differential activity patterns observed during fear learning.” Thus, in the model, we are agnostic about inputs to SOM interneurons; we modeled them to fire spontaneously at high theta.

To address these points in the manuscript, we added some new text in what follows:

(1) New Section “An alternative network configuration characterized by US input to PV, instead of CS, also learns the association between CS and fear” in the Supplementary information:

“We constrained the BLA network in Fig. 2 with CS input to the PV interneuron, as reported in (Krabbe et al., 2018). However, (Krabbe et al., 2019) notes that a class of PV interneurons may be responding to US rather than CS. Fig. S3 presents the results obtained with this variation in the model (see Fig. 3 A,B for comparison) and shows that all the network realizations learn the association between CS and fear. In the model, the PING rhythm between PV and F is the crucial component for establishing fine timing between ECS and F, which is necessary for learning. Having PV responding to the same input as F, i.e., US, facilitates their entrainment in PING and, thus, successful fear learning.

We model the VIP interneuron as affected by US; in addition, (Krabbe et al. 2019) reports that a substantial proportion of them is mildly activated by CS. Replacing the US by CS does not change the input to VIP cells, which is modeled by the same constant applied current. Thus, the VIP CS-induced activity is a bursting activity at low theta, similar to the one elicited by US in Fig. 2.”

(2) Section “With the depression-dominated plasticity rule, all interneuron types are needed to provide potentiation during fear learning” in Results: “Finally, since (Krabbe et al., 2019) reported that a fraction of PV interneurons are affected by US, we have also run the simulations for single neuron network with the PV interneuron affected by US instead of CS. In this case as well, all the network realizations are learners (see Fig. S3). ”

(3) Section “Conditioned and unconditioned stimuli” in Materials and Methods: “To make Fig. S3, we also considered a variation of the model with PV interneurons affected by US, instead of CS, as reported in (Krabbe et al. 2019).”

b. Re the SOM to PV connection: As reported in the reply to the public reviews, we considered the prominent functional connections reported in (Krabbe et al., 2019), instead of structural connections. That is, we included only those connections for which there was strong functional connectivity. For example, the SOM to PV connection is shown to be small (Supp. Fig. 4, panel t, in (Krabbe et al., 2019)). We also omitted PV to SOM, PV to VIP, SOM to VIP, and VIP to excitatory projection neurons; all of these are shown in (Krabbe et al. 2019, Fig. 3 (panel l), and Supp. Fig. 4 (panels m,t)) to have weak functional connectivity, at least in the context of fear conditioning.

In order to clarify this point, in Section “Network connectivity and synaptic currents” in Materials and Methods, we now say:

“We modeled the network connectivity as presented in Fig. 2B, derived from the prominent functional, instead of structural, connections reported in (Krabbe et al., 2019).”

c. Re the ECS to PV synapses: We thank the Reviewer for the reference provided; as the Reviewer says, the ECS to PV synapses are not included. Upon adding this connection in our network, we found that, unlike the connection suggested in part a above, introducing these synapses would, in fact, change the outcome. Thus, the omission of this connection must be considered an implied hypothesis. Including those synapses with a significant strength would alter the PING rhythm created by the interactions between F and PV, which is crucial for ECS and F fine timing. Thanks very much for showing us that this needs to be said. Our hypothesis does not contradict the dense connections mentioned by the Reviewer; such dense connectivity does not mean that all pyramidal cells connect to all interneurons. This hypothesis may be taken as a prediction of the model.

The absence of this connection is now discussed at the end of a new Section of the Discussion entitled “Assumptions and predictions of the model”, which reads as follows:

“Finally, the model assumes the absence of significantly strong connections from the excitatory projection cells ECS to PV interneurons, unlike the ones from F to PV. Including those synapses would alter the PING rhythm created by the interactions between F and PV, which is crucial for ECS and F fine timing. We note that in (Woodruff and Sah, 2007) only 38% of the pyramidal cells are connected to PV cells. The functional identity of the connected pyramidal cells is unknown. Our model suggests that successful fear conditioning requires F to PV connections and that ECS to PV must be weak or absent.”

(2) Krabbe et al. 2019 and Davis et al. 2017 were referenced for the construction of the conditioned and unconditioned stimulus pairing protocol. The Davis citation is not applicable here because that study was a contextual, not cued, fear conditioning paradigm. Regarding Krabbe, the pairing protocol was radically different from what the authors used. Their conditioned stimulus was a train of tone pips presented at 0.9 Hz, which lasted 30 s, after which the unconditioned stimulus was presented after tone offset. The authors should determine how their network behaves when this protocol is used. Also, note that basolateral amygdala responses to tone stimuli are primarily brief onset responses (e.g. Quirk, Armony, and LeDoux 1997), and not the tonic activation used in the model.  

We replied to this point in our responses to the Reviewer’s Public Comments as follows:

“We agree that, in order to implement the fear conditioning paradigm in our in-silico network, we made several assumptions about the nature of the CS and US inputs affecting the neurons in the BLA and the duration of these inputs. A Poisson spike train to the BLA is a signal that contains no structure that could influence the timing of the BLA output; hence, we used this as our CS input signal. We also note that the CS input can be of many forms in general fear conditioning (e.g., tone, light, odor), and we wished to de-emphasize the specific nature of the CS. The reference mentioned in the Recommendations for authors, (Quirk, Armony, and LeDoux 1997), uses pulses 2 seconds long. At the end of fear conditioning, the response to those pulses is brief. However, in the early stages of conditioning, the response goes on for as long as the figure shows. The authors do show the number of cells responding decreases from early to late training, which perhaps reflects increasing specificity over training. This feature is not currently in our model, but we look forward to thinking about how it might be incorporated. Regarding the CS pulsed protocol used in (Krabbe et al., 2019), it has been shown that intense inputs (6kHz and 12 kHz inputs) can lead to metabotropic effects that last much longer than the actual input (200 ms duration) (Whittington et al., Nature, 1995). Thus, the effective input to the BLA may indeed be more like

Poisson.”

Current answer to the Reviewer:

There are several distinct issues raised by the Reviewer in the more detailed critique. We respectfully disagree that the model is not applicable to context-dependent fear learning where the context acts as a CS, though we should have been more explicit. Specifically, our CS input can describe both the cue and the context. We included the following text in the Results section “Interneuron rhythms provide the fine timing needed for depression-dominated STDP to make the association between CS and fear”:

“In our simulations, the CS input describes either the context or the cue in contextual and cued fear conditioning, respectively. For the context, the input may come from the hippocampus or other non-sensory regions, but this does not affect its role as input in the model.”

The second major issue is whether the specific training protocols used in the cited papers need to be exactly reproduced in the signals received by the elements of our model; we note that there are many transformations that can occur between the sensory input and the signals received by the BLA. In the case of auditory fear conditioning, a series of pips, rather than individual pips, are considered the CS (e.g., (Stujenske et al., 2014; Krabbe et al. 2019)). Our understanding is that a single pip does not elicit a fear response; a series of pips is required for fear learning. This indicates that it is not the neural code of a single pip that matters, but rather the signal entering the amygdala that incorporates any history-dependent signaling that could lead to spiking throughout the sequence of pips.  Also, as mentioned above, intense inputs at frequencies about 6kHz and 12kHz can lead to metabotropic effects that last much longer than each brief pip (~200 ms), thus possibly producing continuous activity in neurons encoding the input. Thus, we believe that our use of the Poisson spike train is reasonable. 

However, we are aware that the activity of neurons encoding CS can be modulated by the pips: neurons encoding auditory CS display a higher firing rate when each pip is presented and a Poisson-like spike train between pips (Herry et al., Journal of Neuroscience, 2007). Here we confirm that potentiation is present even in the presence of the fast transient response elicited by the pips. We said in the original manuscript that there is learning for a Poisson spike train CS input at ~50 Hz; this describes the neuronal activity in between pips. For the revision, we asked whether learning is preserved when CS is characterized by higher frequencies, which would describe the CS during and right after each pip. We show in the new Fig. S4 that potentiation is ensured for a range of CS frequencies. The figure shows the learning speed as a function of CS and US frequencies. For all the CS frequencies considered, i) there is learning, ii) learning speed increases with CS frequency. Thus, potentiation is present even when pips elicit a faster transient response.

To better specify this in the manuscript, 

We added the following sentences in the Results section “With the depressiondominated plasticity rule, all interneuron types are needed to provide potentiation during fear learning”: 

“We note that the CS and US inputs modeled as independent Poisson spike trains represent stimuli with no structure. Although we have not explicitly modeled pulsating pips, as common in auditory fear conditioning (e.g., (Stujenske 2014; Krabbe 2019)), we show in Fig. S4 that potentiation can be achieved over a relatively wide range of gamma frequencies. This indicates that overall potentiation is ensured if the gamma frequency transiently increases after the pip.”

We added the section “The full network potentiates for a range of CS frequencies“ and figure S4 in the Supplementary Information:

We included in Materials and Methods “Conditioned and unconditioned stimuli” the following sentences:

“Finally, for Fig.S4, we considered a range of frequencies for the CS stimulus. To generate the three Poisson spike trains with average frequencies from 48 to 64 Hz in Fig. S4, we set 𝜆 = 800, 1000, 1200.”

Finally, to address the comment about the need for CS and US overlapping in time to instantiate fear association, we added the following text in the Results section “Assumptions and predictions of the model”:

“Finally, our model requires the effect of the CS and US inputs on the BLA neuron activity to overlap in time in order to instantiate fear learning. Despite paradigms involving both overlapping (delay conditioning, where US co-terminates with CS (e.g., (Lindquist et al., 2004)), or immediately follows CS (e.g., Krabbe et al., 2019)) and non-overlapping (trace conditioning) CS/US inputs exist, we hypothesized that concomitant activity in CS- and US-encoding neuron activity should be crucial in both cases. This may be mediated by the memory effect due to metabotropic effects (Whittington et al., Nature, 1995) as suggested above, or by the contribution from other brain regions (see section “Involvement of other brain structures” in the Discussion). The fact that plasticity occurs with US memory trace is a consequence of our larger hypothesis that fear learning uses spike-timing-dependent plasticity; such a hypothesis about plasticity is common in the modeling literature.”

(3) As best as I could tell, only a single training trial was used in this study. Fair enough, especially given that fear learning can occur with a single trial. However, most studies of amygdala fear conditioning have multiple trials (~5 or more). How does the model perform when multiple trials are given?  

The association between CS and fear acquired after one trial, i.e., through a potentiated ECS to F connection, is preserved in the presence of multiple trials.  Indeed, the association would be weakened or erased (through depression of the ECS to F connection) only if ECS and F did not display good fine timing, i.e., F does not fire right after ECS most of the time. However, the implemented circuit supports the role of interneurons in providing the correct fine timing, thus preventing the association acquired from being erased.  

In the second paragraph of the Results section “With the depression-dominated plasticity rule, all interneuron types are needed to provide potentiation during fear learning”, we made the above point by adding the following text:

“We note that once the association between CS and fear is acquired, subsequent presentations of CS and US do not weaken or erase it: the interneurons ensure the correct timing and pauses in ECS and F activity, which are conducive for potentiation.”

(4) The LFP calculations are problematic. First, it is unclear how they were done. Did the authors just take the transmembrane currents they included and sum them, or were they scaled by distance from the 'electrode' and extracellular conductivity (as one would derive from the Laplace equation)? Presumably, the spatial arrangement of model neurons was neglected so distance was not a factor. 

Second, if this is the case, then the argument for excluding GABAergic conductances seems flawed. If the spatial arrangement of neurons is relevant to whether to include or exclude GABAergic conductances, then wouldn't a simulation without any spatial structure not be subject to the concern of laminar vs. nuclear arrangement? 

Moreover, to the best I can tell, the literature the authors use to justify the exclusion of

GABAergic currents does not make the case for a lack of GABAergic contribution in non-laminar structures. Instead, those studies only argue that in a non-laminar structure, AMPA currents are detectable, not that GABA cannot be detected. Thus, the authors should either include the GABAergic currents when calculating their simulated LFP, or provide a substantially better argument or citation for their exclusion. 

We thank the Reviewer for pointing this out; this comment helped us rethink how to model the LFP. The origin of the LFP signal in BLA has not been fully determined, but factors thought to be important include differences in the spatial extension of the arborization in excitatory and inhibitory neurons, in the number of synaptic boutons, and spatial distributions of somata and synapses (Lindén et al 2011; Łęski 2013; Mazzoni et al. 2015). In the first version of the manuscript, we excluded the GABAergic currents because it is typically assumed that they add very little to the extracellular field as the inhibitory reversal potential is close to the resting membrane potential. For the revision, we re-ran the simulations during pre and post fear conditioning and we modeled the LFP as the sum of the AMPA, GABA and NaP-/H-/D- currents. With this new version of the LFP, we added a new Fig. 6 showing that there is a significant increase in the low theta power, but not in the high theta power, with fear learning (Fig. 6 C, D, E). This increase in the low theta power was mainly due to the AMPA currents created by the newly established connection from ECS to F, which allowed F to be active after fear conditioning in response to CS. 

However, as the Reviewer mentioned, our network has no spatial extent: neurons are modeled as point cells. Thus, our current model does not include the features necessary to model some central aspects of the LFP. Despite that, our model does clearly demonstrate how rhythmic activity in the spike timing of neurons within the network changes due to fear learning (Fig. 6B). The spiking outputs of the network are key components of the inputs to the LFP, and thus we expect the rhythms in the spiking to be reflected in more complex descriptions of the LFP. But we also discovered that different LFP proxies provide different changes in rhythmic activity comparing pre- and post-fear learning; although we have no principled way to choose a LFP proxy, we believe that the rhythmic firing is the essential finding of the model.

We have added the following to the manuscript:

(1) In the new version of Fig. 6, we present the power spectra of the network spiking activity (panel B), along with the power spectra of the LFP proxy that includes the GABA, AMPA, and NaP-/H-/D- currents (panels C, D, E). 

(2) We modified the conclusion of the Results section entitled “Increased low-theta frequency is a biomarker of fear learning” by saying:

“In this section, we explore how plasticity in the fear circuit affects the network dynamics, comparing after fear conditioning to before. We first show that fear conditioning leads to an increase in low theta frequency power of the network spiking activity compared to the pre-conditioned level (Fig. 6 A,B); there is no change in the high theta power. We also show that the LFP, modeled as the linear sum of all the AMPA, GABA, NaP-, D-, and H- currents in the network, similarly reveals a low theta power increase and no significant variation in the high theta power (Fig. 6 C,D,E). These results reproduce the experimental findings in (Davis et al., 2017), and (Davis et al., 2017), and Fig 6 F,G show that the low theta increase is due to added excitation provided by the new learned pathway. The additional unresponsive ECS and F cells in the network were included to ensure we had not biased the LFP towards excitation. Nevertheless, although both the AMPA and GABA currents contribute to the power increase in the low theta frequency range (Fig. 6F), the AMPA currents show a dramatic power increase relative to the baseline (the average power ratio of AMPA and GABA post- vs pre-conditioning across 20 network realizations is 3*103 and 4.6, respectively). This points to the AMPA currents as the major contributor to the low theta power increase. Specifically, the newly potentiated AMPA synapse from ECS to F ensures F is active after fear conditioning, thus generating strong currents in the PV cells to which it has strong connections (Fig. 6G). Finally, the increase in power is in the low theta range because ECS and F are allowed to spike only during the active phase of the low theta spiking VIP neurons. We have also explored another proxy for the LFP (see Supplementary Information and Fig. S6).”

In the Supplementary Information, we included a figure and some text in the new section entitled “A higher low theta power increase emerges in LFP approximated with the sum of the absolute values of the currents compared to their linear sum”:

“Given that our BLA network comprises a few neurons described as single-compartment cells with no spatial extension and location, the LFP cannot be computed directly from our model’s read-outs. In the main text, we choose as an LFP proxy the linear sum of the AMPA, GABA, and P-/H-/D-currents. We note that if the LFP is modeled as the sum of the absolute value of the currents, as suggested by (Mazzoni et al. 2008; Mazzoni et al. 2015), an even higher low theta power increase arises after fear conditioning compared to the linear sum. Differences in the power spectra also arise if other LFP proxies (e.g., only AMPA currents, only GABA currents) are considered. A principled description of an LFP proxy would require modeling the three-dimensional BLA anatomy, including that of the interneurons VIP and SOM; this is outside the scope of the current paper. (See (Feng et al. 2019) for a related project in the BLA.)”

(3) We updated the Materials and Methods section “Local field potentials and spectral analysis” to explain how we compute the LFP in the revised manuscript: 

“We considered as an LFP proxy as the linear sum of all the AMPA, GABA, NaP, D, and H currents in the network. The D-current is in the VIP interneurons, and NaP-current and H-current are in SOM interneurons.”

Although it is beyond the scope of the current work, an exploration of the most accurate proxy of the LFP in the amygdala is warranted. Such a study could be accomplished by adopting a similar approach as in (Mazzoni et al., 2015), where several LFP proxies based on point-neuron leaky-integrate and fire neuronal network were compared with a “groundtruth” LFP obtained in an analogous realistic three-dimensional network model. 

To explicitly mention this issue in the paper, we add a paragraph in the “Limitations and caveats” section in the Discussion, which reads as follows:

“LFPs recorded in the experiments are thought to be mainly created by transmembrane currents in neurons located around the electrode and depend on several factors, including the morphology of the arborization of contributing neurons and the location of AMPA and GABA boutons (Katzner et al. 2009; Lindén et al 2011; Łęski 2013; Mazzoni et al. 2015). Since our model has no spatial extension, we used an LFP proxy; this proxy was shown to reflect the rhythmic output of the network, which we believe to be the essential result (for more details see Results “Increased low-theta frequency is a biomarker of fear learning”, and Supplementary Information “A higher low theta power increase emerges in LFP approximated with the sum of the absolute values of the currents compared to their linear sum”).”

(4)     We have removed the section “Plasticity between fear neuron and VIP slows down overall potentiation” in Results and sections “Plasticity between the fear neuron (F) and VIP slows down overall potentiation” and “Plastic F to VIP connections further increase lowtheta frequency power after fear conditioning” in the Supplementary Information. This material is extraneous since we are using a new proxy for LFP.

Minor points: 

(1) In Figure 3C, the y-axis tick label for 0.037 is written as "0.37."

We thank the reviewer for finding this typo; we fixed it.

(2) Figure 5B is unclear. It seems to suggest that the added ECS and F neurons did not respond to either the CS or UCS. Is this true? If so, why include them in the model? How would their inclusion change the model behavior? 

It is correct that the added ECS and F neurons did not respond to the CS or US (UCS); they are constructed to be firing at 11 Hz in the absence of any connections from other cells.  These cells were included to be part of our computation of the LFP.  Specifically, adding in those cells would make the LFP take inhibition into account more, and we wanted to make sure that were not biasing our computation away from the effects of inhibition.  As shown in the paper (Fig. 6B), even with inhibition onto these non-responsive cells, the LFP has the properties claimed in the paper concerning the changes in the low theta and high-theta power, because the LFP is dominated by new excitation rather than the inhibition. 

First, in the Results section “Network with multiple heterogeneous neurons can establish the association between CS and fear”, we commented on the added ECS and F neurons that do not respond to either CS or US by saying the following:

“The ECS cells not receiving CS are inhibited by ongoing PV activity during the disinhibition window (Fig. 5B); they are constructed to be firing at 11 Hz in the absence of any connections from other cells. The lack of activity in those cells during fear conditioning implies that there is no plasticity from those ECS cells to the active F. Those cells are included for the calculation of the LFP (see below in “Increased low-theta frequency is a biomarker of fear learning”.)”

Furthermore, we add the following sentence in the Results section “Increased low-theta frequency is a biomarker of fear learning”: 

“The additional unresponsive ECS and F cells in the network were included to ensure we had not biased the LFP towards excitation.”

(3) Applied currents are given as current densities, but these are difficult to compare with current levels observed from whole-cell patch clamp recordings. Can the currents be given as absolute levels, in pA/nA. 

In principle, it is possible to connect current densities with absolute levels, as requested. However, we note that the number of cells in models is orders of magnitude smaller than the number being modeled. It is common in modeling to adjust physiological parameters to achieve the qualitative properties that are important to the model, rather than trying to exactly match particular recordings.

We added to the Methods description why we choose units per unit area, rather than absolute units. 

“All the currents are expressed in units per area, rather than absolute units, to avoid making assumptions about the size of the neuron surface.”

(4) Regarding: "We note that the presence of SOM cells is crucial for plasticity in our model since they help to produce the necessary pauses in the excitatory projection cell activity. However, the high theta rhythm they produce is not crucial to the plasticity: in our model, high theta or higher frequency rhythms in SOM cells are all conducive to associative fear learning. This opens the possibility that the high theta rhythm in the BLA mostly originates in the prefrontal cortex and/or the hippocampus (Stujenske et al., 2014, 2022)." The chain of reasoning in the above statement is unclear. The second sentence seems to be saying contradictory things. 

We agree that the sentence was confusing; thank you for pointing it out. We have revised the paragraph to make our point clearer. The central points are: 1) having the SOM cells in the BLA is critical to the plasticity in the model, and 2) these cells may or may not be the source of the high theta observed in the BLA during fear learning.

We deleted from the discussion the text reported by the Reviewer, and we added the following one to make this point clearer:

“We note that the presence of SOM cells is crucial for plasticity in our model since they help to produce the necessary pauses in the excitatory projection cell activity. The BLA SOM cells do not necessarily have to be the only source of the high theta observed in the BLA during fear learning; the high theta detected in the LFP of the BLA also originates from the prefrontal cortex and/or the hippocampus (Stujenske et al., 2014, 2022).”

(5) Regarding: "This suggests low theta power change is not just an epiphenomenon but rather a biomarker of successful fear conditioning." Not sure this is the right framing for the above statement. The power of the theta signal in the LFP reflects the strengthening of connections, but it itself does not have an impact on network activity. Moreover, whether something is epiphenomenal is not relevant to the question of whether it can serve as a successful biomarker. A biomarker just needs to be indicative, not causal. 

We intended to say why the low theta power change is a biomarker in the sense of the Reviewer. That is: experiments have shown that, with learning, the low theta power increases. The modeling shows in addition that, when learning does not take place, the low power does not increase. That means that the low theta power increases if and only if there is learning, i.e., the change in low theta power is a biomarker. To make our meaning clearer, we have changed the quoted sentences to read: 

“This suggests that the low theta power change is a biomarker of successful fear conditioning: it occurs when there is learning and does not occur when there is no learning.”

Reviewer #2 (Public Comments): 

We thank the Reviewer for raising these interesting points. Below are our public replies and the changes we made to the manuscript to address the Reviewer’s objections.

(1) Gamma oscillations are generated locally; thus, it is appropriate to model in any cortical structure. However, the generation of theta rhythms is based on the interplay of many brain areas therefore local circuits may not be sufficient to model these oscillations.

Moreover, to generate the classical theta, a laminal structure arrangement is needed (where neurons form layers like in the hippocampus and cortex)(Buzsaki, 2002), which is clearly not present in the BLA. To date, I am not aware of any study which has demonstrated that theta is generated in the BLA. All studies that recorded theta in the BLA performed the recordings referenced to a ground electrode far away from the BLA, an approach that can easily pick up volume conducted theta rhythm generated e.g., in the hippocampus or other layered cortical structure. To clarify whether theta rhythm can be generated locally, one should have conducted recordings referenced to a local channel (see Lalla et al., 2017 eNeuro). In summary, at present, there is no evidence that theta can be generated locally within the BLA. Though, there can be BLA neurons, firing of which shows theta rhythmicity, e.g., driven by hippocampal afferents at theta rhythm, this does not mean that theta rhythm per se can be generated within the BLA as the structure of the BLA does not support generation of rhythmic current dipoles. This questions the rationale of using theta as a proxy for BLA network function which does not necessarily reflect the population activity of local principal neurons in contrast to that seen in the hippocampus.

In both modeling and experiments, a laminar structure does not seem to be needed to produce a theta rhythm. A recent experimental paper, (Antonoudiou et al. 2022), suggests that the BLA can intrinsically generate theta oscillations (3-12 Hz) detectable by LFP recordings under certain conditions, such as reduced inhibitory tone. The authors draw this conclusion by looking at mice ex vivo slices. The currents that generate these rhythms are in the BLA, since the hippocampus was removed to eliminate hippocampal volume conduction and other nearby brain structures did not display any oscillatory activity. Also, in the modeling literature, there are multiple examples of the production of theta rhythms in small networks not involving layers; these papers explain the mechanisms producing theta from non-laminated structures (Dudman et al., 2009, Kispersky et al., 2010, Chartove et al. 2020).  We are not aware of any model description of the mechanisms of theta that do require layers.

We added the following text in the introduction of the manuscript to make this point clearer:  “A recent rodent experimental study (Antonoudiou et al. 2022) suggests that BLA can intrinsically generate theta oscillations (3-12 Hz).”

(2) The authors distinguished low and high theta. This may be misleading, as the low theta they refer to is basically a respiratory-driven rhythm typically present during an attentive state (Karalis and Sirota, 2022; Bagur et al., 2021, etc.). Thus, it would be more appropriate to use breathing-driven oscillations instead of low theta. Again, this rhythm is not generated by the BLA circuits, but by volume conducted into this region. Yet, the firing of BLA neurons can still be entrained by this oscillation. I think it is important to emphasize the difference.

Many rhythms of the nervous system can be generated in multiple parts of the brain by multiple mechanisms. We do not dispute that low theta appears in the context of respiration; however, this does not mean that other rhythms with the same frequencies are driven by respiration. Indeed, in the response to question 1 above, we showed that theta can appear in the BLA without inputs from other regions. In our paper, the low theta is generated in the BLA by VIP neurons. Using intrinsic currents known to exist in VIP neurons (Porter et al., 1998), modeling has shown that such neurons can intrinsically produce a low theta rhythm. This is also shown in the current paper. This example is part of a substantial literature showing that there are multiple mechanisms for any given frequency band. 

To elaborate more on this in the manuscript, we added the following new section in the discussion:

“Where the rhythms originate, and by what mechanisms. A recent experimental paper, (Antonoudiou et al. 2022), suggests that the BLA can intrinsically generate theta oscillations (3-12 Hz) detectable by LFP recordings under certain conditions, such as reduced inhibitory tone. They draw this conclusion in mice by removing the hippocampus, which can volume conduct to BLA, and noticing that other nearby brain structures did not display any oscillatory activity. Our model also supports the idea that intrinsic mechanisms in the BLA can support the generation of the low theta, high theta, and gamma rhythms. 

Although the BLA can produce these rhythms, this does not rule out that other brain structures also produce the same rhythms through different mechanisms, and these can be transmitted to the BLA. Specifically, it is known that the olfactory bulb produces and transmits the respiratory-related low theta (4 Hz) oscillations to the dorsomedial prefrontal cortex, where it organizes neural activity (Bagur et al., 2021). Thus, the respiratory-related low theta may be captured by BLA LFP because of volume conduction or through BLA extensive communications with the prefrontal cortex. Furthermore, high theta oscillations are known to be produced by the hippocampus during various brain functions and behavioral states, including during spatial exploration (Vanderwolf, 1969) and memory formation/retrieval (Raghavachari et al., 2001), which are both involved in fear conditioning. Similarly to the low theta rhythm, the hippocampal high theta can manifest in the BLA. It remains to understand how these other rhythms may interact with the ones described in our paper.”

We also note that the presence of D-currents in the BLA VIP interneurons should be confirmed experimentally, and that the ability of VIP interneurons to generate the BLA low theta rhythm constitutes a prediction of our computational model. These points are specified in the first paragraph in the Discussion entitled “Assumptions and predictions of the model”:

“The interneuron descriptions in the model were constrained by the electrophysiological properties reported in response to hyperpolarizing currents (Sosulina et al., 2010). Specifically, we modeled the three subtypes of VIP, SOM, and PV interneurons displaying bursting behavior, regular spiking with early spike-frequency adaptation, and regular spiking without spike-frequency adaptation, respectively. Focusing on VIP interneurons, we were able to model the bursting behavior by including the D-type potassium current. This current is thought to exist in the VIP interneurons in the cortex (Porter et al., 1998), but whether this current is also found in the VIP interneurons the BLA is still unknown. Similarly, we endowed the SOM interneurons with NaP- and H-currents, as the OLM cells in the hippocampus. Due to these currents, the VIP and SOM cells are able to show  low- and high-theta oscillations, respectively. The presence of these currents and the neurons’ ability to exhibit oscillations in the theta range during fear conditioning and at baseline in BLA, which are assumptions of our model, should be tested experimentally.”

(3) The authors implemented three interneuron types in their model, ignoring a large fraction of GABAergic cells present in the BLA (Vereczki et al., 2021). Recently, the microcircuit organization of the BLA has been more thoroughly uncovered, including connectivity details for PV+ interneurons, firing features of neurochemically identified interneurons (instead of mRNA expression-based identification, Sosulina et al., 2010), synaptic properties between distinct interneuron types as well as principal cells and interneurons using paired recordings. These recent findings would be vital to incorporate into the model instead of using results obtained in the hippocampus and neocortex. I am not sure that a realistic model can be achieved by excluding many interneuron types.

The interneurons and connectivity that we used were inspired by the functional connectivity reported in (Krabbe et al., 2019) (see above answer to Reviewer #1). As reported in (Vereczki et al., 2021), there are multiple categories and subcategories of interneurons; that paper does not report on which ones are essential for fear conditioning. We did use all the highly represented categories of the interneurons, except NPYcontaining neurogliaform cells.

The Reviewer says “I am not sure that a realistic model can be achieved by excluding many interneuron types”. We agree with the Reviewer that discarding the introduction of other interneurons subtypes and the description of more specific connectivity (soma-, dendrite-, and axon-targeting connections) may limit the ability of our model to describe all the details in the BLA. However, this work represents a first effort towards a biophysically detailed description of the BLA rhythms and their function. As in any modeling approach, assumptions about what to describe and test are determined by the scientific question; details postulated to be less relevant are omitted to obtain clarity. The interneuron subtypes we modeled, especially VIP+ and PV+, have been reported to have a crucial role in fear conditioning (Krabbe et al., 2019). Other interneurons, e.g. cholecystokinin and SOM+, have been suggested as essential in fear extinction. Thus, in the follow-up of this work to explain fear extinction, we will introduce other cell types and connectivity. In the current work, we have achieved our goals of explaining the origin of the experimentally found rhythms and their roles in the production of plasticity underlying fear learning. Of course, a more detailed model may reveal flaws in this explanation, but this is science that has not yet been done.

We elaborate more on this in a new section in the Discussion entitled “Assumptions and predictions of the model”. The paragraph related to this point reads as follows:

“Our model, which is a first effort towards a biophysically detailed description of the BLA rhythms and their functions, does not include the neuron morphology, many other cell types, conductances, and connections that are known to exist in the BLA; models such as ours are often called “minimal models” and constitute the majority of biologically detailed models. Such minimal models are used to maximize the insight that can be gained by omitting details whose influence on the answers to the questions addressed in the model are believed not to be qualitatively important. We note that the absence of these omitted features constitutes hypotheses of the model: we hypothesize that the absence of these features does not materially affect the conclusions of the model about the questions we are investigating. Of course, such hypotheses can be refuted by further work showing the importance of some omitted features for these questions and may be critical for other questions. Our results hold when there is some degree of heterogeneity of cells of the same type, showing that homogeneity is not a necessary condition.”

(4) The authors set the reversal potential of GABA-A receptor-mediated currents to -80 mV. What was the rationale for choosing this value? The reversal potential of IPSCs has been found to be -54 mV in fast-spiking (i.e., parvalbumin) interneurons and around -72 mV in principal cells (Martina et al., 2001, Veres et al., 2017).

A GABA-A reversal potential around -80 mV is common in the modeling literature (Jensen et al., 2005; Traub et al., 2005; Kumar et al., 2011; Chartove et al., 2020). Other computational works of the amygdala, e.g. (Kim et al., 2016), consider GABA-A reversal potential at -75 mV based on the cortex (Durstewitz et al., 2000). The papers cited by the reviewer have a GABA-A reversal potential of -72 mV for synapses onto pyramidal cells; this is sufficiently close to our model that it is not likely to make a difference. For synapses onto PV+ cells, the papers cited by the reviewer suggest that the GABA-A reversal potential is -54 mV; such a reversal potential would lead these synapses to be excitatory instead of inhibitory. However, it is known (Krabbe et al., 2019; Supp. Fig. 4b) that such synapses are in fact inhibitory. Thus, we wonder if the measurements of Martina and Veres were made in a condition very different from that of Krabbe. For all these reasons, we consider a GABA-A reversal potential around -80 mV in amygdala to be a reasonable assumption.

In section “Network connectivity and synaptic currents” in “Materials and Methods” we provided references to motivate our choice of considering a GABA-A reversal potential around -80 mV:

“The GABAa current reversal potential (𝐸!) is set to −80        𝑚𝑉, as common in the modeling literature (Jensen et al., 2005; Traub et al., 2005; Kumar et al., 2011; Chartove et al., 2020).”

(5) Proposing neuropeptide VIP as a key factor for learning is interesting. Though, it is not clear why this peptide is more important in fear learning in comparison to SST and CCK, which are also abundant in the BLA and can effectively regulate the circuit operation in cortical areas.

Other peptides seem to be important in overall modulation of fear, but VIP is especially important in the first part of fear learning, the subject of our paper. Re SST: we hypothesize that SST interneurons are critical in fear extinction and preventing fear generalization, but not to initial fear learning. The peptide of the CCK neurons, which overlap with VIP cells, has been proposed to promote the switch between fear and safety states after fear extinction (Krabbe al. 2018). Thus, these other peptides are likely more important for other aspects of fear learning.  

In the Discussion, we have added:

“We hypothesize that SST peptide is critical in fear extinction and preventing fear generalization, but not to initial fear learning. Also, the CCK peptide has been proposed to promote the switch between fear and safety states after fear extinction (Krabbe al. 2018).”

Reviewer #2 (Recommendations For The Authors): 

We note that Reviewer #2’s Recommendations For The Authors have the same content as the Public Comments. Thus, the changes to the manuscript we implemented above address also the private critiques listed below.

(1) As the breathing-driven rhythm is a global phenomenon accompanying fear state, one might restrict the analysis to this oscillation. The rationale beyond this restriction is that the 'high' theta in the BLA has an unknown origin (since it can originate from the ventral hippocampus, piriform cortex etc.). 

In response to point 4 made by Reviewer 1 (Recommendations for the Authors) (p. 13), referring to high theta in the BLA, we previously wrote: 1) having the SOM cells in the BLA is critical to the plasticity in the model, and 2) these cells may or may not be the source of the high theta observed in the BLA during fear learning.

In the Public Critiques, Reviewer 2 relates the respiratory rhythm to the low theta. We answered this point in point 2 of the Reviewer’s Public Comments (at p. 15).

(2) I would include more interneurons in the network model incorporating recent findings. 

This point was answered in our response to point 3 of the Reviewer’s Public Comments.

(3) The reversal potential for GABA-A receptor-mediated currents would be good to set to measured values. In addition, I would use AMPA conductance values that have been measured in the BLA. 

We addressed this objection in our response to point 4 of the Reviewer’s Public Comments.

Reviewer #3 (Public comments):

Weaknesses: 

(1) The main weakness of the approach is the lack of experimental data from the BLA to constrain the biophysical models. This forces the authors to use models based on other brain regions and leaves open the question of whether the model really faithfully represents the basolateral amygdala circuitry. 

(2) Furthermore, the authors chose to use model neurons without a representation of the morphology. However, given that PV+ and SOM+ cells are known to preferentially target different parts of pyramidal cells and given that the model relies on a strong inhibition form SOM to silence pyramidal cells, the question arises whether SOM inhibition at the apical dendrite in a model representing pyramidal cell morphology would still be sufficient to provide enough inhibition to silence pyramidal firing.

3) Lastly, the fear learning relies on the presentation of the unconditioned stimulus over a long period of time (40 seconds). The authors justify this long-lasting input as reflecting not only the stimulus itself but as a memory of the US that is present over this extended time period. However, the experimental evidence for this presented in the paper is only very weak.

We are repeating here the answers we gave in response to the public comments, adding further relevant points.

(1) Our neurons were constrained by electrophysiology properties in response to hyperpolarizing currents in the BLA (Sosulina et al., 2010). We can reproduce these electrophysiological properties by using specific membrane currents known to be present in similar neurons in other brain regions (D-current in VIP interneurons in the cortex, and NaP- and H-currents in OLM/SOM cells in the hippocampus). Also, though a much more detailed description of BLA interneurons was given in (Vereczki et al., 2021), it is not clear that this level of detail is relevant to the questions that we were asking, especially since the experiments described were not done in the context of fear learning.

(2) It is true that we did not include the morphology, which undoubtedly makes a difference to some aspects of the circuit dynamics. Furthermore, it is correct that the model relies on a strong inhibition from SOM and PV to silence the excitatory projection neurons. We agree that the placement of the SOM inhibition on the pyramidal neurons can make a difference on some aspects of the circuit behavior. We are assuming that the inhibition from the SOM cells can inhibit the pyramidal cells firing, which can be seen as a hypothesis of our model. It is well known that VIP cells disinhibit pyramidal cells through inhibition of SOM and PV cells (Krabbe et al. 2019); hence, this hypothesis is generally believed. This choice of parameters comes from using simplified models: it is standard in modeling to adjust parameters to compensate for simplifications.

Re points 1) and 2), in a new paragraph (“Assumptions and predictions of the model”) in the Discussion reported in response to Reviewer #2 (public comments)’s point 3, we stated that modeling requires the omission of many details to bring out the significance of other details.

(3) 40 seconds is the temporal interval we decided to use to present the results. In the Results, we also showed that there is learning over a shorter interval of time (15 seconds) where CS and US/memory of US should both be present. Thus, our model requires 15 seconds over a single or multiple trials for associative learning to be established. We included references to additional experimental papers to support our reasoning in the last paragraph of section “Assumptions and predictions of the model” in the Discussion, also reported in response to Reviewer #1 point 2 (Recommendations for the Authors). We said there that some form of memory or overlap in the activity of the excitatory projection neurons is necessary for spike-timing-dependent plasticity.

The authors achieved the aim of constructing a biophysically detailed model of the BLA not only capable of fear learning but also showing spectral signatures seen in vivo. The presented results support the conclusions with the exception of a potential alternative circuit mechanism demonstrating fear learning based on a classical Hebbian (i.e. non-depression-dominated) plasticity rule, which would not require the intricate interplay between the inhibitory interneurons. This alternative circuit is mentioned but a more detailed comparison between it and the proposed circuitry is warranted.

Our model accounts for the multiple rhythms observed in the context of fear learning, as well as the known involvement of multiple kinds of interneurons. We did not say explicitly enough why our complicated model may be functionally important in ways that cannot be fulfilled with a simpler model with the non depression-dominated Hebbian rule. To explain this, we have added the following in the manuscript discussion: 

“Although fear learning can occur without the depression-dominated rule, we hypothesize that it is necessary for other aspects of fear learning and regulation. That is, in pathological cases, there can be overgeneralization of learning. We hypothesize that the modulation created by the involvement of these interneurons is normally used to prevent such overgeneralization. However, this is beyond the scope of the present paper.”

We have also written an extra paragraph about generalization in the Discussion “Synaptic plasticity in our model”:

“With the classical Hebbian plasticity rule, we show that learning can occur without the involvement of the VIP and SOM cells. Although fear learning can occur without the depressiondominated rule, we hypothesize that the latter is necessary for other aspects of fear learning and regulation. Generalization of learning can be pathological, and we hypothesize that the modulation created by the involvement of VIP and SOM interneurons is normally used to prevent such overgeneralization. However, in some circumstances, it may be desirable to account for many possible threats, and then a classical Hebbian plasticity rule could be useful. We note that the involvement or not of the VIP-SOM circuit has been implicated when there are multiple strategies for solving a task (Piet et al., 2024). In our situation, the nature of the task (including reward structure) may determine whether the learning rule is depression-dominated and therefore whether the VIP-SOM circuit plays an important role.”

Reviewer #3 (Recommendations For The Authors): 

We thank the Reviewer for all the recommendations. We replied to each of them below.

In general, there are some inconsistencies in the naming (e.g. sometimes you write PV sometimes PV+,...), please use consistent abbreviations throughout the manuscript. You also introduce some of the abbreviations multiple times. 

We modified the manuscript to remove all the inconsistencies in the naming. 

Introduction: 

- In the last section you speak about one recent study but actually cite two articles. 

We removed the reference to (Perrenoud and Cardin, 2023), which is a commentary on the Veit et al. article.

Results: 

- 'Brain rhythms are thought to be encoded and propagated largely by interneurons' What do you mean by encoded here? 

We agree with the Reviewer that the verb “to encode” is not accurate. We modified the sentence as follows:

“Brain rhythms are thought to be generated and propagated largely by interneurons”.

- The section 'Interneurons interact to modulate fear neuron output' could be clearer. Start with describing the elements of the circuit, then the rhythms in the baseline. 

We reorganized the section as follows:

“Interneurons interact to modulate fear neuron output. Our BLA network consists of interneurons, detailed in the previous section, and excitatory projection neurons (Fig. 2A). Both the fear-encoding neuron (F), an excitatory projection neuron, and the VIP interneuron are activated by the noxious stimulus US (Krabbe et al., 2019). As shown in Fig. 2A (top, right), VIP disinhibits F by inhibiting both SOM and PV, as suggested in (Krabbe et al., 2019). We do not include connections from PV to SOM and VIP, nor connections from SOM to PV and VIP, since those connections have been shown to be significantly weaker than the ones included (Krabbe et al., 2019). The simplest network we consider is made of one neuron for each cell type. We introduce a larger network with some heterogeneity in the last two sections of the Results.

Fig. 2A (bottom) shows a typical dynamic of the network before and after the US input onset, with US modeled as a Poisson spike train at ~50 Hz; the network produces all the rhythms originating from the interneurons alone or through their interactions with the excitatory projection neurons (shown in Fig. 1). Specifically, since VIP is active at low theta during both rest and upon the injection of US, it then modulates F at low theta cycles via SOM and PV. In the baseline condition, the VIP interneuron has short gamma bursts nested in low theta rhythm. With US onset, VIP increases its burst duration and the frequency of low theta rhythm. These longer bursts make the SOM cell silent for long periods of each low theta cycle, providing F with windows of disinhibition and contributing to the abrupt increase in activity right after the US onset. Finally, in Fig. 2A, PV lacks any external input and fires only when excited by F. Thanks to their reciprocal interactions, PV forms a PING rhythm with F, as depicted in Fig.1C.”

- Figure 3C: The lower dashed line has the tick label '0.37' which should read '0.037'. 

We fixed it.

- The section describing the network with multiple neurons could be clearer, especially, it is not really clear how these different ECS and F neurons receive their input. 

We answered the same objection in the reply to Reviewer #1 in point 2 under “minor issues.”

Discussion: 

- The paragraph 'It has also been suggested that ventral tegmental area has a role in fear expression (Lesas et al.,2023). Furthermore, it has been reported that the prelimbic cortex (PL) modulates the BLA SOM cells during fear retrieval, and the latter cells are crucial to discriminate non-threatening cues when desynchronized by the PL inputs (Stujenske et al., 2022).' is merely stating facts but I don't see how they relate to the presented work. 

We thank the Reviewer for pointing out that this was confusing. What we meant to emphasize was that later stages of fear conditioning and extinction appear to require more than the BLA. We specifically mention the discrimination of non-threatening cues at the end of the paragraph, which now reads as follows:

“Other brain structures may be involved in later stages of fear responsiveness, such as fear extinction and prevention of generalization. It has been reported that the prelimbic cortex (PL) modulates the BLA SOM cells during fear retrieval, and the latter cells are crucial to discriminate non-threatening cues when desynchronized by the PL inputs (Stujenske et al., 2022). Brain structures such as the prefrontal cortex and hippocampus have been documented to play a crucial role also in fear extinction, the paradigm following fear conditioning aimed at decrementing the conditioned fearful response through repeated presentations of the CS alone. As reported by several studies, fear extinction suppresses the fear memory through the acquisition of a distinct memory, instead of through the erasure of the fear memory itself (Harris et al., 2000; Bouton, 2002; Trouche et al., 2013; Thompson et al., 2018). Davis et al., 2017 found a high theta rhythm following fear extinction that was associated with the suppression of threat in rodents. Our model can be extended to include structures in the prefrontal cortex and the hippocampus to further investigate the role of rhythms in the context of discrimination of non-threatening cues and extinction. We hypothesize that a different population of PV interneurons plays a crucial role in mediating competition between fearful memories, associated with a low theta rhythm, and safety memories, associated with a high theta rhythm; supporting experimental evidence is in (Lucas et al., 2016; Davis et al., 2017; Chen et al., 2022).”

- The comparison to other models BLA is quite short and seems a bit superficial. A more indepth comparison seems warranted. 

We thank the reviewer for suggesting that a more in-depth comparison between our and other models in the literature would improve the manuscript. We rewrote entirely the first paragraph of that section. The new content reads as follows:

“Comparison with other models. Many computational models that study fear conditioning have been proposed in the last years; the list includes biophysically detailed models (e.g., (Li 2009; Kim et al., 2013a)), firing rate models (e.g., Krasne 2011; Ball 2012; Vlachos 2011), and connectionist models (e.g., Moustafa 2013; Armony 1997; Edeline 1992) (for a review see (Nair et al., 2016)). Both firing rate models and connectionist models use an abstract description of the interacting neurons or regions. The omission of biophysical details prevents such models from addressing questions concerning the roles of dynamics and biophysical details in fear conditioning, which is the aim of our model.  There are also biophysically detailed models (Li 2009; Kim 2013; Kim 2016; Feng 2019), which differ from ours in both the physiology included in the model and the description of how plastic changes take place.  One main difference in the physiology is that we differentiated among types of interneurons, since the fine timing produced for the latter was key to our use of rhythms to produce spike-time dependent plasticity. The origin of the gamma rhythm (but not the other rhythms) was investigated in Feng et al 2019, but none of these papers connected the rhythms to plasticity.

The most interesting difference between our work and that in (Li 2009; Kim 2013; Kim 2016) is the modeling of plasticity.  We use spike-time dependent plasticity rules.  The models in (Li 2009; Kim 2013; Kim 2016) were more mechanistic about how the plasticity takes place, starting with the known involvement of calcium with plasticity.  Using a hypothesis about back propagation of spikes, the set of papers together come up with a theory that is consistent with STDP and other instantiations of plasticity (Shouval 2002a; Shouval 2002b).  For the purposes of our paper, this level of detail, though very interesting, was not necessary for our conclusions.  By contrast, in order for the rhythms and the interneurons to have the dynamic roles they play in the model, we needed to restrict our STDP rule to ones that are depression-dominated.  Our reading of (Shouval 2002) suggests to us that such subrules are possible outcomes of the general theory.  Thus, there is no contradiction between the models, just a difference in focus; our focus was on the importance of the much-documented rhythms (Seidenbecher et al., 2003; Courtin et al., 2014b; Stujenske et al., 2014; Davis et al., 2017) in providing the correct spike timing.  We showed in the Supplementary Information (“Classical Hebbian plasticity rule, unlike the depression-dominated one, shows potentiation even with no strict pre and postsynaptic spike timing”) that if the STDP rule was not depression dominated, the rhythms need not be necessary.  We hypothesize that the necessity of strict timing enforced by the depression-dominated rule may foster the most appropriate association with fear at the expense of less relevant associations.”

- The paragraph 'This could happen among some cells responding to weaker sensory inputs that do not lead to pre-post timing with fear neurons. This timing could be modified by the "triconditional rule", as suggested in (Grewe et al., 2017).' is not very clear. What exactly is 'this' in the first sentence referring to? If you mention the 'tri-conditional rule' here, please briefly explain it and how it would solve the issue at hand here.  

We apologize that the sentence reported was not sufficiently clear. “This” refers to “depression”. We meant that, in our model, depression during fear conditioning happens every time there is no pre-post timing between neurons encoding the neutral stimuli and fear cells; poor pre-post timing can characterize the activity of neurons responding to weaker sensory inputs and does not lead to associative learning. We modified that paragraph as follows:

“The study in (Grewe et al., 2017) suggests that associative learning resulting from fear conditioning induces both potentiation and depression among coactive excitatory neurons; coactivity was determined by calcium signaling and thus did not allow measurements of fine timing between spikes. In our model, we show how potentiation between coactive cells occurs when strict pre-post spike timing and appropriate pauses in the spiking activity arise. Depression happens when one or both of these components are not present. Thus, in our model, depression represents the absence of successful fear association and does not take part in the reshaping of the ensemble encoding the association, as instead suggested in (Grewe et al., 2017). A possible follow-up of our work involves investigating how fear ensembles form and modify through fear conditioning and later stages. This follow-up work may involve using a tri-conditional rule, as suggested in (Grewe et al. 2017), in which the potential role of neuromodulators is taken into account in addition to the pre- and postsynaptic neuron activity; this may lead to both potentiation and depression in establishing an associative memory.”

- In the limitations and caveats section you mention that the small size of the network implies that they represent a synchronous population. What are the potential implications for the proposed rhythm-dependent mechanism? What are your expectations for larger networks? 

We apologize if we were not adequately clear. We are guessing that the Reviewer thought we meant the entire population was synchronous, which it is not. We meant that, when we use a single cell to represent a subpopulation of cells of that type, that subpopulation is effectively synchronous. For larger networks in which each subtype is represented by many cells, there can be heterogeneity within each subtype. We have shown in the paper that the basic results still hold under some heterogeneity; however, they may fail if the heterogeneity is too large.

We mentioned in a new section named “Assumptions and predictions of the model” in response to point 3 made by Reviewer #2.

- The discussion is also missing a section on predictions/new experiments that can be derived from the model. How can the model be confirmed, what experiments/results would break the model? 

To answer this question, we put in a new section in the Discussion entitled “Assumptions and predictions of the model”. The first paragraph of this section is in the reply to Reviewer #2 point 2; the second paragraph is in the reply to Reviewer #2 point 3; the last paragraph is in the Reply to Reviewer #1 point c; the rest of the section reads as follows:

“Our study suggests that all the interneurons are necessary for associative learning provided that the STDP rule is depression-dominated. This prediction could be tested experimentally by selectively silencing each interneuron subtype in the BLA: if the associative learning is hampered by silencing any of the interneuron subtypes, this validates our study. Finally, the model prediction could be tested indirectly by acquiring more information about the plasticity rule involved in the BLA during associative learning. We found that all the interneurons are necessary to establish fear learning only in the case of a depression-dominated rule. This rule ensures that fine timing and pauses are always required for potentiation: interneurons provide both fine timing and pauses to pyramidal cells, making them crucial components of the fear circuit. 

The modeling of the interneurons assumes the involvement of various intrinsic currents; the inclusion of those currents can be considered hypotheses of the model. Our model predicts that blockade of D-current in VIP interneurons (or silencing VIP interneurons) will both diminish low theta and prevent fear learning. Finally, the model assumes the absence of significantly strong connections from the excitatory projection cells ECS to PV interneurons, unlike the ones from F to PV. Including those synapses would alter the PING rhythm created by the interactions between F and PV, which is crucial for fine timing between ECS and F needed for LTP.”

We would like to thank you and the reviewers for your thoughtful comments that assisted us to improve the manuscript. We carefully followed the reviewers’ recommendations and provide a detailed point-by-point account of our responses to the comments. 

Please find below the important changes in the updated manuscript.

(1) We changed the title according to the comments provided by reviewer #1.

(2) We edited the introduction, results, and discussion to improve the link between the objectives of the study, the findings, and their discussion, as reviewer #2 recommended.

(3) We clarified the link between camouflage and fitness, which is now presented as a hypothesis, as reviewer #1 suggested.

(4) We added new analyses and figures in the main text and in the supplementary materials to better emphasize sex differences in landing force, foraging strategies and hunting success, following reviewer #1 suggestion.

(5) According to reviewer #2 comments, we edited the results adding key information about methods to help the reader understand the findings without reading the Methods section.

(6) We added important details about the model selection approach along with a discussion of the low R-square values reported in our analyses on hunting success, as reviewer #2 suggested.

eLife assessment 

This fundamental work substantially advances our understanding of animals' foraging behaviour, by monitoring the movement and body posture of barn owls in high resolution, in addition to assessing their foraging success. With a large dataset, the evidence supporting the main conclusions is convincing. This work provides new evidence for motion-induced sound camouflage and has broad implications for understanding predator-prey interactions. 

Public Reviews: 

Reviewer #1 (Public Review): 

In this paper, Schalcher et al. examined how barn owls' landing force affects their hunting success during two hunting strategies: strike hunting and sit-and-wait hunting. They tracked tens of barn owls that raised their nestlings in nest boxes and utilized high-resolution GPS and acceleration loggers to monitor their movements. In addition, camcorders were placed near their nest boxes and used to record the prey they brought to the nest, thus measuring their foraging success. 

This study generated a unique dataset and provided new insights into the foraging behavior of barn owls. The researchers discovered that the landing force during hunting strikes was significantly higher compared to the sit-and-wait strategy. Additionally, they found a positive relationship between landing force and foraging success during hunting strikes, whereas, during the sit-and-wait strategy, there was a negative relationship between the two. This suggests that barn owls avoid detection by generating a lower landing force and producing less noise. Furthermore, the researchers observed that environmental characteristics affect barn owls' landing force during sit-and-wait hunting. They found a greater landing force when landing on buildings, a lower landing force when landing on trees, and the lowest landing force when landing on poles. The landing force also decreased as the time to the next hunting attempt decreased. These findings collectively suggest that barn owls reduce their landing force as an acoustic camouflage to avoid detection by their prey. 

The main strength of this work is the researchers' comprehensive approach, examining different aspects of foraging behavior, including high-resolution movement, foraging success, and the influence of the environment on this behavior, supported by impressive data collection. The weakness of this study is that the results only present a partial biological story contained within the data. The focus is on acoustic camouflage without addressing other aspects of barn owls' foraging strategy, leaving the reader with many unanswered questions. These include individual differences, direct measurements of owls' fitness, a detailed analysis of the foraging strategy of males and females, and the collective effort per nest box. However, it is possible that these data will be published in a separate paper. 

We greatly appreciate your recognition of the comprehensive approach and extensive data collection. Our primary objective was to study the role of acoustic camouflage. Nonetheless, the manuscript now includes a detailed analysis of the foraging strategy and hunting success of males and females (lines 164-225).

The results presented support the authors' conclusion that lower landing force during sit-andwait hunting increases hunting success, likely due to a decreased probability of detection by their prey, resulting in acoustic camouflage. The authors also argue that hunting success is crucial for survival, and thus, acoustic camouflage has a direct link to fitness. While this statement is reasonable, it should be presented as a hypothesis, as no direct evidence has been provided here.

Thank you for the comment. We agree and thus have edited the language accordingly.  

However, since information about nestling survival is typically monitored when studying behavior during the breeding period, the authors' knowledge of the effect of acoustic camouflage on owls' fitness can probably be provided. Furthermore, it will be interesting to further examine the foraging strategies used by different individuals during foraging, the joint foraging success of both males and females within each nest box, and the link between landing force and foraging success if the data are available.

We are currently writing a manuscript on these topics. We are aware that several scientific questions regarding the foraging ecology of the barn owl still need our attention. Regarding the link between landing force and foraging success, we believe that our revised manuscript addresses this specific topic, please see specific responses below.

However, even without this additional analysis on survival, this paper provides an unprecedented dataset and the first measurement of landing force during hunting in the wild. It is likely to inspire many other researchers currently studying animal foraging behavior to explore how animals' movements affect foraging success.

Reviewer #2 (Public Review): 

Summary: 

The authors provide new evidence for motion-induced sound camouflage and can link the hunting approach to hunting success (detailing the adaptation and inferring a fitness consequence). 

Strengths: 

Strong evidence by combining high-resolution accelerometer data with a ground-truthed data set on prey provisioning at nest boxes. A good set of co-variates to control for some of the noise in the data provides some additional insights into owl hunting attempts. 

Weaknesses: 

There is a disconnect between the hypotheses tested and the results presented, and insufficient detail is provided on the statistical approach. R2 values of the presented models are very small compared to the significance of the effect presented. Without more detail, it is impossible to assess the strength of the evidence.

In the revised manuscript, we changed the way results are presented and we improved the link between the hypotheses and the results. The R2 values are indeed small. It is however important to keep in mind that we are assessing the outcome of one specific behavior (i.e. landing force during sit-and-wait hunts) on hunting success in a wild environment, where many complex ecological interactions likely influence hunting success. Nonetheless, the coefficients (as reported in the results) show that for every 1 N increase in landing force, there is a 15% reduction in hunting success, which is substantial. In the discussion we also note that 50 Hz is a relatively low sampling frequency for estimating the peak ground reaction force. We have gone back over the presentation of our results and made our discussion more nuanced to acknowledge this aspect. 

We have also added a detailed description about our model selection process in the methods section and provide a model selection table for each analysis in the supplementary materials.

The authors seem to overcome persisting challenges associated with the validation and calibration of accelerometer data by ground-truthing on-board measures with direct observations in captivity, but here the methods are not described any further and sample sizes (2 owls - how many different loggers were deployed?) might be too small to achieve robust behavioural classifications.

Thank you for the comment. Details of our methods of behavioural identification are provided in lines 385 – 429. There are two reasons why our results should not be limited by the sample size. First, we used the temporal sequence of changes in acceleration, and rates of change in acceleration data, which make the methods robust to individual differences in acceleration values. Furthermore, our methods for behavioural identification were not based on machine learning. Instead, we use a Boolean based approach (as described in Wilson et al. 2018. MEE), which is more robust to small differences in absolute values that might occur e.g. in relation to slight changes in device position. 

Recommendation for the authors:

Reviewer #1 (Recommendations For The Authors): 

Comment 1. This study provides new insights into animals' foraging behavior and will probably inspire other researchers to examine foraging behavior in such high resolution.

We hope so, thank you.

Comment 2. However, it is necessary to describe better the measured landing force and the hunting strike and perching behavior so the readers can understand these methods when reading the results (and without reading the Methods).

We have now changed the text in the “Results” to help the reader understand the key methods while reading the results.

Comment 3. In addition, make sure you use the same terminology for hunting strategies during the entire paper and especially in all figures and corresponding result descriptions.

We now use consistent terminology throughout the text and figures. We hope that this is now clear in the revised manuscript.

Comment 4. In addition, although I find your statement about the link between acoustic camouflage and fitness reasonable, it should be described as a hypothesis or examined if you want to keep the direct link statement. I believe showing a direct link can add an additional outstanding aspect to this paper, but I also understand that it can be addressed in a separate paper.

We agree that the relationship between hunting success and barn owl fitness is an important topic, but it necessitates a consideration of both hunting strategies, including hunting on the wing, which extends beyond the limits of our current study. Indeed, our primary objective was to conduct a detailed examination of the interplay between acoustic camouflage and the success of the sit-and-wait technique.

However, we have edited the manuscript to explicitly describe the link between acoustic camouflage and fitness as a hypothesis. We believe this adjustment provides a more accurate representation of our approach. We hope this clarifies the specific emphasis of our work and its contribution to the understanding of barn owl hunting behavior.

Here are my detailed comments about the paper: 

Comment 5. Title: Consider changing the title to "Acoustic camouflage predicts hunting success in a wild predator." 

We would like to thank you for your nice proposition. However, we opted for a different title, which is now “Landing force reveals new form of motion-induced sound camouflage in a wild predator”.

Comment 6. Line 91-93: Please provide additional information about the collected dataset, including: 

Description of the total period of observations, an average and standard deviation of perching and hunting attempt events per individual per night, number of foraging trips per individual per night, details about the geographic location and characteristics of the habitat, season, and reproductive state. 

The revised manuscript now includes detailed information about the collected dataset (i.e. study area, reproductive state, etc…). “We used GPS loggers and accelerometers to record high resolution movement data during two consecutive breeding seasons (May to August in 2019 and 2020) from 163 wild barn owls (79 males and 84 females) breeding in nest boxes across a 1,000 km² intensive agricultural landscape in the western Swiss plateau.” Results section, lines 79 – 82

Details about the number of foraging trips per individuals and per night are now presented in the results: “Sexual dimorphism in body mass was marked among our sampled individuals. Males were lighter than females (84 females, average body mass: 322 ± 22.6 g; 79 males, average body mass 281 ± 16.5 g, Fig S6) and provided almost three times more prey per night than females (males: 8 ± 5 prey per night; females: 3 ± 3 prey per night; Fig.S7). Males also displayed higher nightly hunting effort than females (Males: 46 ± 16 hunting attempts per night, n= 79; Females: 25 ± 11 hunting attempts per nights, n=84; Fig. 3A, Fig S8). However, females were more likely to use a sit and wait strategy than males (females: 24% ± 15%, males: 13% ± 10%, Fig.S9). As a result, the number of perching events per night was similar between males and females (Females: 76 ± 23 perching events per nights; Males: 69 ± 20 perching events per night; Fig S8).” (lines 165 – 174) 

Comment 7. In addition, state if the information describes breeding pairs of males and females and provides statistics on the number of tracked pairs and the number of nest boxes.

The revised manuscript now includes a description of the number of tracked breeding pairs and the number of nest boxes. “Of these individuals, 142 belonged to pairs for which data were recovered from both partners (71 pairs in total, 40 in 2019, 31 in 2020). The remaining 21 individuals belonged to pairs with data from one partner (11 females and 1 male in 2019; 4 females and 5 males in 2020).” (lines 82 – 85.)

Comment 8. Line 93: Briefly define the term "landing force" and explain how it was measured (and let the reader know that there is a detailed description in the Methods).

We now include a brief definition of the “landing force” along with a brief explanation of how it was measured in the results section. “We extracted the peak vectoral sum of the raw acceleration during each landing and converted this to ground reaction force (hereafter “landing force”, in Newtons) using measurements of individual body mass (see methods for detailed description).” (lines 92 – 95).

Comment 9. Line 94: All definitions, including "pre-hunting force," need to be better described in the Results section.

Thank you for this suggestion. We now provided a better description of those key definitions directly in the results section: 

Measurement of landing force: “Barn owls employing a sit-and-wait strategy land on multiple perches before initiating an attack, with successive landings reducing the distance to the target prey (Fig. 2C). 

We used the acceleration data to identify 84,855 landings. These were further categorized into perching events (n = 56,874) and hunting strikes (n = 27,981), depending whether barn owls were landing on a perch or attempting to strike prey on the ground (Fig. 1A and B, see methods for specific details on behavioral classification).” (lines 88 – 95)

Pre-hunt perching force predicts hunting success: “Finally, we analyzed whether the landing force in the last perching event before each hunting attempt (i.e. pre-hunt perching force) predicted variation in hunting success” (lines 229 – 230)

Comment 10. Line 102: Remove "Our analysis of 27,981 hunting strikes showed that" and add "n = 27,981" after the statistics. You have already stated your sample size earlier. There is no need to emphasize it again, although your sample size is impressive.

We modified the text in the results section as suggested.

Comment 11. Line 104: The results so far suggest that the difference in landing force between males and females is an outcome of their different body masses. However, it is not clear what is the reason for the difference in the number of hunting strike attempts between males and females (Lines 104-106). Can you compare the difference in landing force between males and females with similar body mass (females from the lower part of the distribution and males from the upper part)? Is there still a difference?

Thank you, following your comment we made some new analyses that clarified the situation around landing force involved in perching and hunting strike events between sexes. But firstly, we wanted to clarify why there is a difference in number of hunting attempts between males and females. During the breeding season, females typically perform most of the incubation, brooding, and feeding of nestlings in the nest, while the male primarily hunts food for the female and chicks. The female supports the male providing food in a very irregular way, and this changes from pair to pair (paper in prep.). The differences in number of hunting attempts between males and females reflects this asymmetry in food provisioning between sexes during this specific period. We specified this in the revised version of the manuscript (lines 164 – 174). 

We also provide a new analysis to investigate sex differences in mass-specific landing force (force/body mass). We found that males and females produce similar force per unit of body mass during perching events. This demonstrates that the overall higher perching force in females (see Fig. 4C in the manuscript) is therefore driven by their higher body mass. (lines 194 – 199)

Comment 12. Line 154: I believe Boonman et al. (2018) is relevant to this part of the discussion. Boonman, Arjan, et al. found that barn owl noise during landing and taking off is worth considering. ["The sounds of silence: barn owl noise in landing and taking off."

Behavioral Processes 157 (2018): 484-488.]

We now cited this paper in the discussion.

Comment 13. Line 164: Your results do not directly demonstrate a link to fitness, although they potentially serve as a proxy for fitness (add a reference). However, you might have information regarding nestlings' survival - that will provide a direct link for fitness. Change your statement or add the relevant data.

We appreciated your feedback, and we adjusted the language accordingly.

Comment 14. Line 213: If the poles are closer to the ground - is it possible that the higher trees and buildings serve for resting and gathering environmental information over greater distances? For example, identifying prey at farther distances or navigating to the next pole?

Yes, this is indeed the most likely explanation for the fact that owls land more on buildings and trees than on poles until the last period (about 6 minutes) before hunting. In these last minutes, barn owls preferentially use poles, as we showed in figure 2B. The revised manuscript now includes this explanation in the discussion (lines 269 – 284).

Comment 15. Line 250: The product "AXY-Trek loggers" does not appear on the Technosmart website (there are similar names, but not an exact match). Are you sure this is the correct name of the tracking device you used? 

Thank you for pointing out this detail that we missed. The device we used is now called "AXY-Trek Mini" (https://www.technosmart.eu/axy-trek-mini/). We have corrected this error directly in the revised manuscript.

Comment 16. Line 256: Please explain how the devices were recovered. Did you recapture the animals? If so, how? Additionally, replace "after approximately 15 days" with the exact average and standard deviation. Furthermore, since you have these data, please state the difference in body mass between the two measurements before and after tagging.

The birds were recaptured to recover the devices. Adults barn owls were recaptured at their nest sites, again using automatic sliding traps that are activated when birds enter the nest box. The statement "after approximately 15 days" was replaced by the exact mean and standard deviation, which were 10.47 ± 2.27 days. Those numbers exclude five individuals from the total of 163 individuals included in this study. They could not be recaptured in the appropriate time window but were re-encountered when they initiated a second clutch later in the season (4 individuals) or a new clutch the year after (1 individual).

We integrated this previously missing information in the revised manuscript (lines 370 – 372).

Comment 17. Line 259: What was the resolution of the camera? What were the recording methods and schedule? How did you analyze these data? 

The resolution was set to 3.1 megapixel. Motion sensitive camera traps were installed at the entrance to each nest box throughout the period when the barn owls were wearing data loggers, and each movement detected triggered the capture of three photos in bursts. The photos recorded were not analyzed as such for this study, but were used to confirm each supply of prey, which had previously been detected from the accelerometer data. We added these details in the revised manuscript (lines 377 – 380)

Comment 18_1. Figure 1: 

Panel A) Include the sex of the described individual. 

The sex of the described individual is now included in the figure caption.

Comment 18_2. It would be interesting to show these data for both males and females from the same nest box (choose another example if you don't have the data for this specific nest box). 

Although we agree that showing tracks of males and females from the same nest is very interesting, the purpose of this figure was to illustrate our data annotation process and we believe that adding too many details on this figure will make it appear messy. However, the revised manuscript now includes a new figure (Fig. 3A) which shows simultaneous GPS tracks of a male and a female during a complete night, with detailed information about perching and hunting behaviors.

Comment 18_3. Add the symbol of the nest box to the legend. 

Done

Comment 18_4. Provide information about the total time of the foraging trip in the text below. 

The duration of the illustrated foraging trip has been included in the figure caption.

Comment 18_5. To enhance the figure’s information on foraging behavior, consider color coding the trajectory based on time and adding a background representing the landscape. Since this paper may be of interest to researchers unfamiliar with barn owl foraging behavior, it could answer some common questions. 

For similar reasons explained in our answer above (Comment 18_2), we would rather keep this figure as clean as possible. However, we followed your recommendations and included these details in the new Figure 3 described above. In this new figure, GPS tracks are color coded according to the foraging trip number and includes a background representing the landscape. To provide even more detail about the landscape, we added another figure in the supplementary materials (Fig. S2) which provides illustration of barn owls foraging ground and nest site that we think might be of interest for people unfamiliar with barn owls.

Comment 18_6. Inset panels) provide a detailed description of the acceleration insert panels. 

Done

Comment 18_7. Color code the acceleration data with different colors for each axis, add x and y axes with labels, and ensure the time frame on the x-axis is clear. How was the self-feeding behavior verified (should be described in the methods section)? 

We kept both inset panels as simple as possible since they serve here as examples, but a complete representation of these behaviors (with time frame, different colors and labels) is provided in the supplementary materials (figure S3). We included this statement in the figure caption and added a reference to the full representations from the supplementary materials: 

In the Figure caption: “Inset panels show an example of the pattern of the tri-axial acceleration corresponding to both nest-box return and self-feeding behaviors (but see Fig S3for a detailed representation of the acceleration pattern corresponding to each behavior).” 

In the Method section: “Self-feeding was evident from multiple and regular acceleration peaks in the surge and heave axes (resulting in peaks in VeDBA values > 0.2 g and < 0.9 g, Fig.S3D), with each peak corresponding to the movement of the head as the prey was swallowed whole.”.

Comment 18_8. Panel B) Note in the caption that you refer to the acceleration z-axis.

We believe that keeping the statement “the heave acceleration…” in the figure caption is more informative than referring to the “z-axis” as it describes the real dimension to which we are referring. The use of the x, y and z axes can be misleading as they can be interchanged depending on the type and setting of recorders used.

Comment 18_9. Present the same time scale for both hunting strategies to facilitate comparison. You can achieve this by showing only part of the flight phase before perching. 

Done

Comment 18_10. Panel C) Presenting the data for both hunting strategy and sex would provide more comprehensive information about the results and would be relatively easy to implement. 

We agree with your comment. We present the differences in landing force for both landing contexts and sexes in the new Figure 3 as well as in the supplementary materials (Figure S10) of this revised manuscript.

Comment 19. Figure 2: Please provide an explanation of the meaning of the circles in the figure caption.  

Done

Comment 20. Figure 3: 

Panel A) It is unclear how the owl illustration is relevant to this specific figure, unlike the previous figures where it is clear. Also, suggest removing the upper black line from the edge of the figure or add a line on the right side. 

Done (now in Figure 2).

Panel B) "Density" should be capitalized. 

Done

Panel C) Add a scale in meters, and it would be helpful to include an indication of time before hunting for each data point. 

Done

Comment 21. Figure S1: Mark the locations of the nest boxes and ensure that trajectories of different individuals and sexes can be identified. 

The purpose of this figure was to show the spatial distribution of the data. We think that adding nest locations and coloring the paths according to individuals and/or sex will make the figure less clear. However, the new Figure 3 highlights those details.

Comment 22. Figure S2: Show the pitch angle similarly to how you showed the acceleration axes, and explain what "VeDBA" stands for. Provide a description of the perching behavior, clearly indicating it on the figure. Add axes (x, y, z) to the illustration of the acceleration explanation. 

We edited this figure (now figure S3) to show the pitch angle and provide an explanation of what “VeDBA” stands for in the figure caption. The figure caption now also provides a better description of the perching behavior. For the axes (i.e. X, Y, Z), we prefer to refer to the heave, surge, and sway as this is more informative and refers to what is usually reported in studies working with tri-axial accelerometers.

Comment 23. Table S1: Improve the explanation in the caption and titles of the table. 

Done

Reviewer #2 (Recommendations For The Authors): 

Comment 1. From the public review and my assessment there, the authors can be assured that I thoroughly enjoyed the read and am looking forward to seeing a revised and improved version of this paper. 

We thank the reviewer for this comment. We revised the manuscript according to their comments.

Comment 2. In addition to my major points stated above, I would like to add the following recommendations: 

The manuscript is overall well written, but it uses a very pictorial language (a little as if we were in a David Attenborough documentary) that I find inappropriate for a research paper (especially in the abstract and introduction, "remarkable" (2x), "sophisticated" (are there any unsophisticated adaptations? We are referring to something under selection after all) etc.

We appreciated that you found the paper overall well written, and we understand the comment about pictorial language. We therefore slightly changed the text to make sure that the adjective used to describe adaptive strategies are not over-emphasized.

Comment 3. Abstract 

"While the theoretical benefits of predator camouflage are well established, no study has yet been able to quantify its consequences for hunting success." - This claim is actually not fully true: 

Nebel Carina, Sumasgutner Petra, Pajot Adrien and Amar Arjun 2019: Response time of an avian prey to a simulated hawk attack is slower in darker conditions, but is independent of hawk colour morph. Soc. open sci.6:190677 

We edited our claim to specify that the consequences of predator camouflage on hunting success has never been quantified in natural conditions and cited the reference in the introduction.

Comment 4. Line 23. Rephrase to: "We used high-resolution movement data to quantify how barn owls (Tyto alba) conceal their approach when using a sit-and-wait strategy, as well as the power exerted during strikes." 

We edited this sentence in the abstract, as suggested.

Comment 5. Results 

There is a disconnect between the objectives outlined at the end of the introduction and the following results that should be improved. 

The authors state: "Using high-frequency GPS and accelerometer data from wild barn owls (Tyto alba), we quantify the landing dynamics of this sit-and-wait strategy to (i) examine how birds adjust their landing force with the behavioral and environmental context and (ii) test the extent to which the magnitude of the predator cue affects hunting success." But one of the first results presented are sex differences. 

This is a fair point. We have now changed our statement in the end of the introduction as well as the order of the results to improve the link between the objectives outlined in the introduction and the way result are presented. 

Comment 6. At this stage, the reader does not even know yet that we are presented with a size-dimorphic species that also has very different parental roles during the breeding season. This should be better streamlined, with an extra paragraph in the introduction. And these sex differences are then not even discussed, so why bring them up in the first place (and not just state "sex has been fitted as additional co-variate to account for the size-dimorphism in the species" without further details). 

We edited the way the objectives are outlined in the introduction to cover the size dimorphism (lines 70 – 76). We also completely changed the way the sex differences are presented in the results, including a new analysis that we believe provides a better comprehensive understanding of barn owl foraging behavior (lines 164 – 206). Finally, we added a new paragraph in the discussion to consider those results (lines 319 – 339).

Comment 7. It is not clear to me where and how high-resolution GPS data were used? The results seem to concentrate on ACC – why GPS was used and how it features should be foreshadowed in a few lines in the introduction. I definitively prefer having the methods at the end of a manuscript, but with this structure, it is crucial to give the reader some help to understand the storyline. 

GPS data were used to validate some behavioral classifications (prey provisioning for example), but most importantly they were used to link each landing event with perch types. We edited the text in the result section to clarify where GPS and/or ACC data were used.

Comment 8. Discussion 

Move the orca example further down, where more detail can be provided to understand the evidence. 

After our extensive edits in the discussion, we felt this example was interrupting the flow. We now cite this study in the introduction. 

Comment 9. Size dimorphism and evident sex differences are not discussed. 

The revised manuscript now includes a new paragraph in the discussion in which sex differences are discussed (lines 319 – 339).

Comment 10. Be more precise in the terminology used (for example, land use seems to be interchangeable with habitat characteristics?). 

We modified “land use” with “habitat data” in the revised manuscript.

Comment 11. Methods 

Please provide a justification for the very high weight limit (5%; line 256). This limit is outdated and does not fulfill the international standard of 3% body weight. I assume the ethics clearance went through because of the short nature of the study (i.e., the birds were not burdened for life with the excess weight? But a line is needed here or under the ethics considerations to clarify this). 

The 5% weight limit was considered acceptable due to the short deployment period, and we now edited the ethics statement to emphasize this point. However, it is important to note that there is no real international standard, with both 3% and 5% weight limits being commonly used. Both limits are arbitrary and the impact of a fixed mass on a bird varies with species and flight style. All owls survived and bred similarly to the non-tagged individuals in the population (lines 373 – 376 & lines 558 – 561)

EDITORIAL COMMENT: We strongly encourage you to provide further context and clarification on this issue, as suggested by the Reviewer. On a related point, the ethics statement refers to GPS loggers, rather than GPS and ACC devices; we encourage you to clarify wording here.

Thank you for highlighting this point that indeed needed some clarifications.

Although we have used the terminology "GPS recorders", the authorization granted by the Swiss authorities for this study effectively covers the entire tracking system, which combines both GPS and ACC recorders in the same device. We have therefore changed the wording used in the ethics statement to avoid any misunderstanding (lines 373 – 376 & lines 558 – 561)

Comment 12. Please provide more information on the model selection approach, what does "Non-significant terms were dropped via model simplification by comparing model AIC with and without terms." mean? Did the authors use a stepwise backward elimination procedure (drop1 function)? Or did they apply a complete comparison of several candidate models? I think a model comparison approach rather than stepwise selection would be more informative, as several rather than only one model could be equally probable. This might also improve model weights or might require a model averaging procedure - current reported R2values are very small and do not seem to support the results well. 

We apologize for the lack of details about this important aspect of the statistical analysis. We applied an automated stepwise selection using the dredge function from the R package “MuMin”, therefore applying a complete comparison of several candidate models. The final models were chosen as the best models since the number of candidate models within ∆AIC<2 was relatively low in each analysis and thus a model averaging was not appropriate here. We edited the methods section to ensure clarity, and added model selection tables for each analysis, ranked according to AICc scores, in the supplementary materials (lines 532 – 552)

In addition, we agree that the reported R-squared values in our analyses are quite low, specifically regarding the influence of pre-hunt perching force on hunting success (cond R2 = 0.04). Nonetheless, landing impact still has a notable effect size (an increase of 1N reduces hunting success by 15%). The reported values are indicative of the inherent complexity in studying hunting behavior in a wild setting where numerous variables come into play. We specifically investigated the hypothesis that the force involved during pre-hunt landings, and consequently the emitted noise, influences the success of the next hunting attempt in wild barn owls. Factors such as prey behavior and micro-habitat characteristics surrounding prey (such as substrate type and vegetation height) are most likely to be influential but hard, or nearly impossible, to model. We now cover this in a more nuanced way in the discussion (lines 266 – 268)

Comment 13. Please explain why BirdID was nested in NightID - this is not clear to me.

Probably here there is a misunderstanding because we wrote that we nested NightID in BirdID (and not BirdID in NightID). 

Comment 14. I hope the final graphs and legends will be larger, they are almost impossible to read. 

We enlarged the graphs and legends as much as possible to improve readability. However, looking at the graphs in the published version they seem clear and readable.

Comment 15. Figure S1: Does "representation" mean the tracks don't show all of the 163 owls? If so, be precise and tell us how many are illustrated in the figure. 

Figure S1 represent the tracks for each of the 163 barn owls used in the study. We changed the terminology used in the figure caption to avoid any misunderstanding.

Comment 16. Figure S4: Please adjust the y-axis to a readable format. 

Done

Author response:

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

Response to Reviewer #1 comments:

(1) SY1 aggregation enhances (in terms of number of aggregates) when Sphingolipid biosynthesis is blocked.

a. Line no 132-133: I agree that there is circumstantial evidence that the maturation pathway of SY1 IB is perturbed by knocking down sphingolipid biosynthesis. However, to prove this formally, a time course of IB maturation needs to be reported in the knock-down strains.

Please see Figure 2-figure supplement 1 for the time course of SY1 IB maturation in the knock-down strains. We have added the result to the manuscript, please see lines 129-131on page 5 in the revised version.

b. It will be good to have formal evidence that sphingolipids are indeed downregulated when these genes are downregulated (knocked down).

This issue has been clearly evidenced in previous reports, and we have added the appropriate references in the main text. For example, down-regulation of LCB1 or SPT in yeast decreased sphingolipid levels by Huang et al (https://doi.org/10.1371/journal.pgen.1002493). According to the report from Tafesse FG, et al (https://doi.org/10.1371/journal.ppat.1005188), in mammalian cells in which Sptlc2 was knocked down by CRISPR/Cas9, sphingolipid and glucosylceramide production is almost completely blocked. In addition, the levels of sphingosine, sphingomyelin, and ceramide were significantly lower compared to control cells. Please see lines 143-144 on pages 6 and lines 232-233 on pages 9 in the revised version.

(2) In a normal cell (where sphingolipid biosynthesis is not hampered), the aggregate of SY1 (primarily the Class I aggregate) is localized only on the mitochondrial endomembrane system. These results have been published for other aggregation-prone proteins and are partly explained in the literature. However, their role in the context of maturation is relatively unclear. The authors however provide no strong evidence to show if mitochondria are preferentially involved in any of the stages of IB maturation. Specifically:

a. Line 166-167: It is not clear from Figure 4B that this is indeed the case. Only the large IB seems to colocalize in all three panels (Class I, 2, 3) with Mitotracker. The smaller IBs in 2 and 3 do not show any obvious co-localization. It is also possible that they do co-localize, but it is not clear from the images. I would appreciate it if the authors either provide stronger evidence (better image) or revise this statement. This point is crucial in some claims made later in the manuscript. (pls see comment #5A).

Based on the reviewer's suggestion, we replaced the images in Figure 4B. In addition, we added the 3D reconstruction results of the interrelationship between Class 3 and Mitotracker in Figure 4-figure supplement 1B, to further show their relationship.

(3) The localization is due to the association of SY1 (aggregates) with mitochondrial proteins like Tom70, Tim44 etc. There are some critical points (that can strengthen the manuscript) that are not addressed here. Primarily, the important role of mitochondria in the context of toxicity is neglected. Although the authors have mentioned in the discussion that it was not their main focus, I believe that this is the novel part of the manuscript and this part is potentially a beautiful addition to literature. The questions I found unanswered are:

a. Is the localization completely lost upon deleting these genes? I see only a partial loss in shape/localization. This is not properly explained in the manuscript. The shape of the IB seems to remain intact while the localization is slightly altered. This indicates that even when sphingolipid is present, SY1 localization is dictated by the (lipid-raft embedded) proteins. Interestingly, it shows that even in the absence of mitochondrial localization the shape of the aggregates is not altered in these deletion strains! How do the authors explain this if mitochondrial surface sphingolipids are important for IB maturation? (the primary screen found that sphingolipid biosynthesis promotes the formation of Class I IBs).

We agree that mutation in one mitochondrial binding protein only a partial loss in shape/localization, and we have replaced “association” with “surrounding” in the manuscript. Please see lines 163-166 on page 6 in the revised version. In mutants that interact with SY1, we counted the proportion of Class 3 aggregates formed by SY1 and found an increase in the proportion of SY1 Class 3 aggregates in the deletion mutants compared to controls, partially lost interaction of SY1 with mitochondria has effect on shape of aggregates, as detailed in line 184 on page 7 and Figure 4-figure supplement 1D. We think that SY1 interactions with mitochondrial proteins are important for the localization of SY1 IB in mitochondria, whereas sphingolipids play an important role in facilitating the formation of Class 1 IBs from Class 3 aggregates.

b. What happens to the toxicity when the aggregates are not localized on mitochondria?

We thank the reviewer for the comments, however to investigate this issue, since a single mutant can only partially affect the phenotype, it may be necessary to construct groups of mutants of different genes to observe the effect, which we will further elucidate in our future studies. What we want to show in this work is that SY1 achieves binding to mitochondria by interacting with these mitochondrial proteins.

c. It is important to note that sphingolipids may affect the whole process indirectly by altering pathways involved in protein quality control or UPR. UPR may regulate the maturation of IBs. It is therefore important to test if any of the effects seen could be of direct consequence.

We agree with the reviewer's comments, but there was no significant enrichment for protein quality control or UPR-related pathways in our genome-wide screen, so it is unlikely that sphingolipids indirectly cause maturation of IBs by affecting these two pathways. We addressed this issue in our discussion. Please see lines 325-328 on page 12 in the revised version.

d. In Figure 4D, the authors find SY1 when they pull down Tom70, Tom37 or Tim44. Tim44 is a protein found in the mitochondrial matrix, how do the authors explain that this protein is interacting with a protein outside the mitochondrial outer membrane?

This interaction could be potentially due to that some of the soluble SY1 enter the mitochondrial matrix and interact with Tim44.

e. Is it possible that the authors are immunoprecipitating SY1 since IBs have some amount of unimported mitochondrial proteins in aggregates formed during proteotoxic stress (https://doi.org/10.1073/pnas.2300475120) (Liu et al. 2023).

Our Co-IP experiments were performed in the soluble state supernatant, so mitochondrial proteins in aggregates were not detected.

f. Line 261 (Discussion): Does deletion of Tom70 or one of the anchors increase Class III aggregation and increase toxicity? Without this, it is hard to say if mitochondria are involved in detoxification.

We thank the reviewer for the comments, please see our response to comment 3b.

(4) This fuels the loss of mitochondrial function.

a. Line 218-219: Although the change is significant, the percentage change is very slight. Is this difference enough to be of physiological relevance in mitochondrial function? In our hands, the DCF fluorescence is much more variable.

We agree with the reviewer that there is a small difference (but significant). To which extend such a difference be of physiological relevance in mitochondrial function need to be further investigated.

b. Is SY1-induced loss of mitochondrial function less in knockouts of Tom70 or the other ones found to be important for localizing the SY1 aggregate to mitochondria?

We examined mitochondrial membrane potential (indicated by Rho 123 fluor intensity) in tom70Δ, tom37Δ and control his3Δ strains and found that the knocking out of Tom70 or Tom37 reduced the mitochondrial toxicity caused by SY1 expression. Please see lines 212-214 on page 8 in the revised version, and Figure 5-figure supplement 2.

(5) Mitochondrial function is further abrogated when there is a block in sphingolipid biosynthesis.

a. Myriosin acted like the deletion strains that showed less structured aggregates. There were more aggregates (Class 3) but visually they seemed to be spread apart. The first comment (#2A) on aggregate classes and their interaction with mitochondria may become relevant here.

According to a recent review article (https://doi.org/10.3389/fcell.2023.1302472), sphingolipids are present in the mitochondrial membrane, bind to many mitochondrial proteins and have emerged as key regulators of mitochondrial morphology, distribution and function. Dysregulation of sphingolipid metabolism in mitochondria disrupts many mitochondrial processes, leading to mitochondrial fragmentation, impaired bioenergetics and impaired cellular function. Myriocin treatment, which affects sphingolipid metabolism, causes mitochondria to become more fragmented, which may explain why the aggregates appear visually spread apart. Regarding the interaction with mitochondria, we counted the proportion of SY1 aggregates surrounded by mitochondria after treatment with myriocin, and the results were not significantly different compared to the control. Please see lines 168-169 on page 6 in the revised version, and Figure 4-figure supplement 1C.

(6) A similar phenomenon is conserved in mammalian cell lines.

a. Line 225-226: Did the authors confirm that this was the only alteration in the genome? Or did they complement the phenotype, genetically?

We performed SPTLC2 gene complementation experiments in knockout cell lines and found that SPTLC2 gene complementation was able to reduce the number of cells forming IBs and the percentage of dispersed irregular IBs compared to controls. Please see lines 240-242 on page 9 in the revised version, and Figure 6-figure supplement 2B.

b. Line 241-245: One of the significant phenotypes observed by downregulating sphingolipid biosynthesis in yeast and mammalian cells, was the increase in the number of aggregates. This is not shown in myriocin treatment in mammalian cells. This needs to be shown to the main concordance with the original screen and the data presented with the KO mammalian cell line.

Please see Figure 7-figure supplement 1A for the data on the proportion of cells forming SY1 IBs after myriocin treatment in mammalian cells, and myriocin treatment in mammalian cells was the same as in the KO mammalian cell line.

Minor Comments:

Line 273-275: How is this statement connected to the previous statement? Was it observed that aggregate fusion was advantageous to the cells?

Yes, aggregate/oligomer fusion is advantageous to the cells, and we have modified the previous statement. Please see line 280 on page 10 in the revised version.

Line 293-294: I am not sure I understand this statement.

We have modified this statement. Please see lines 302-303 on page 11 in the revised version.

Line 295-296: But the authors have commented at multiple places that mitochondria detoxify the cell from SY1 aggregates. I find this link fascinating and worth investigating. Most of the current work has some known links in literature (not everything). The mitochondrial connection being the most fascinating one.

We have removed this sentence. We have added a validation experiment for the role of mitochondrial activity in SY1 IB maturation in the revised version.

Line 318: Do the authors mean: The open question is...

Thanks to the reviewer, we have corrected it.

Response to Reviewer #2 comments:

I recommend considering live cell microscopy to analyze whether sphingolipid-dependent formation of SY1 IB takes place at the mitochondrial outer membrane. The IBs could also be produced at other membranes and then transported to the mitochondrial outer membrane for storage.

As shown in Figure 4A, SY1 IB primarily interacts with mitochondria.

I recommend analyzing whether mitochondrial activity is needed for sphingolipid-dependent SY1 IB formation. Are these IBs localized to mitochondrial membrane solely as scaffold or are these organelles needed to provide the energy for driving IB formation in concert with sphingolipids? This point could be addressed with rho0 strains lacking mitochondrial DNA.

We thank the reviewer for this recommendation. We expressed SY1 protein in BY4741 rho0 strain as suggested and found that the maturation and mitochondrial surrounding state of SY1 IB was not affected by mitochondrial activity. Please see lines 185-187 on page 7 in the revised version, and Figure 4-figure supplement 1E and 1F.

The authors should be more precise in the statistical methods used in their study (method, pre-/post-tests, number of replicates...).

We thank the reviewer for the comment and we have provided a more precise description of the statistical methods. Please see lines 531-534 on page 19 and figure legends in the revised version.

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This is an interesting study that utilizes a novel epigenome profiling technology (single molecule imaging) in order to demonstrate its utility as a readout of therapeutic response in multiple DIPG cell lines. Two different drugs were evaluated, singly and in combination. Sulfopin, an inhibitor of a component upstream of the MYC pathway, and Vorinostat, an HDAC inhibitor. Both drugs sensitized DIPG cells, but high (>10 micromolar) concentrations were needed to achieve half-maximal effects. The combination seemed to have some efficacy in vivo, but also produced debilitating side-effects that precluded the measurement of any survival benefit.

We thank the reviewer for deeply evaluating our work and acknowledging the use of multiple experimental strategies to explore the effect of combination therapy on DMG cells. Of note, all mice in our experiment experienced deterioration (including the control mice and those treated with single agents). Thus, it is not the combination of drugs that led to the debilitating side-effects; the mice deteriorated due to the extremely aggressive tumor cells, forming relatively large tumors prior to the treatment onset, calling for further optimization of the therapeutic regime.

We modified the text in the results section to clarify this point (lines 238-241): “This rapid deterioration is likely a result of the aggressiveness of the transplanted tumors and does not represent side effects of the treatment, as mice from all groups, including the non-treated mice, showed similar signs of deterioration”.  

We also elaborate on this in the discussion (lines 272-276): “Notably, despite a significant reduction in tumor size in-vivo, the combined treatment did not increase mice survival. This is perhaps due to the relatively large tumors already formed at the onset of treatment, leading to rapid deterioration of mice in all experimental groups. Thus, further optimization of the modeling system and therapeutic regime is needed.” We truly hope that further studies will allow better assessment of this drug combination in various models.

Strengths:

Interesting use of a novel epigenome profiling technology (single molecule imaging).

Weaknesses:

The use of this novel imaging technology ultimately makes up only a minor part of the study. The rest of the results, i.e. DIPG sensitivity to HDAC and MYC pathway inhibition, have already been demonstrated by others (Grasso Monje 2015; Pajovic Hawkins 2020, among others). The drugs have some interesting opposing effects at the level of the epigenome, demonstrated through CUT&RUN, but this is not unexpected in any way. The drugs evaluated here also didn't have higher efficacy, or efficacy at especially low concentrations, than inhibitors used in previous reports. The combination therapy attempted here also caused severe side effects in mice (dehydration/deterioration), such that an effect on survival could not be determined. I'm not sure this study advances knowledge of targeted therapy approaches in DIPGs, or if it iterates on previous findings to deliver new, or more efficient, mechanistic or therapeutic/pharmaclogic insights. It is a translational report evaluating two drugs singly and in combination, finding that although they sensitise cells in vitro, efficacy in vivo is limited at best, as this particular combination cannot progress to human translation.

We thank the reviewer for pointing out the strengths and weaknesses of our work. As far as we know, while many studies demonstrated upregulation of the MYC pathway in DIPG, this is the first study that shows inhibition of this pathway (via PIN1) as a therapeutic strategy. While it is clear from the literature that MYC inhibition may pose therapeutic benefit, the development of potent MYC inhibitors is highly challenging due to its structure and cellular localization. Of note, in the 2020 paper, Pajovic and colleagues inhibited MYC by transfecting the cells with a plasmid expressing a specific inhibitory MYC peptide (Omomyc); while this strategy works well for cell cultures, the clinical translation requires different delivery strategies. Sulfopin is a small molecule inhibitor that can be used in-vivo and potentially in clinical studies. Thus, we believe that our study offers a novel strategy, as well as mechanistic insights, regarding the potential use of Sulfopin and Vorinostat to treat DIPG.

As noted above, the combination therapy did not cause side effects, but rather the aggressiveness of the tumors. We did not notice specific toxicity in the mice treated with Sulfopin alone, or the combined treatment. Furthermore, Dubiella et al. extensively examined toxicity issues and did not observe adverse effects or weight loss when administrating Sulfopin in a dose of 40 mg kg–1.

Optimization of the model and treatment regime (# of cells injected, treatment starting point, etc.) may have allowed us to reveal survival benefits. Yet, these are highly complicated and expensive experiments; unfortunately, we did not have the resources to perform them within the scope of this revision. Importantly, within the current manuscript, we show the effect of this drug combination in reducing the growth of DMG cells in-vitro and in-vivo, laying the framework for follow-up exploration in future studies. Furthermore, the epigenetic and transcriptomic profiling shed light on the molecular mechanisms that drive these aggressive tumors.

Reviewer #2 (Public Review):

Summary:

The study by Algranati et al. introduces an exciting and promising therapeutic approach for the treatment of H3-K27M pediatric gliomas, a particularly aggressive brain cancer predominantly affecting children. By exploring the dual targeting of histone deacetylases (HDACs) and MYC activation, the research presents a novel strategy that significantly reduces cell viability and tumor growth in patient-derived glioma cells and xenograft mouse models. This approach, supported by transcriptomic and epigenomic profiling, unveils the potential of combining Sulfopin and Vorinostat to downregulate oncogenic pathways, including the mTOR signaling pathway. While the study offers valuable insights, it would benefit from additional clarification on several points, such as the rationale behind the dosing decisions for the compounds tested, the specific contributions of MYC amplification and H3K27me3 alterations to the observed therapeutic effects, and the details of the treatment protocols employed in both in-vitro and in-vivo experiments.

We thank the reviewer for evaluating our work and recognizing its potential for the DMG research field. We address in detail below the important comments regarding the treatment protocols and dosing decisions.

Clarification is needed on how doses were selected for the compounds in Figure S2A and throughout the study. Understanding the basis for these choices is crucial for interpreting the results and their potential clinical relevance. IC50s are calculated for specific patient derived lines, but it is not clear how these are used for selecting the dose.

We thank the reviewer for these important comments. For the epigenetic drugs shown in Figure S2A, we followed published experimental setups; for EPZ6438, GSKJ4, Vorinostat and MM-102 we chose the treating concentrations according to Mohammad et al. 2017, Grasso et al. 2015 and Furth et al. 2022, accordingly. For Sulfopin, we conducted a dedicated dose curve analysis (shown in Figure 1E), indicating only a mild effect on viability and relatively high IC-50 values as a single agent. Since we aimed to test the ability of a combined treatment to additively reduce viability, we used a sub-IC50 concentration for Sulfopin in these experiments. We added this information in lines 123 and 131-132.

Finally, following the results obtained in the experiment shown in Figure S2A, we conducted a full dose-curve analysis of the combined treatment in multiple DMG patient-derived cells (figure 2B and S2C), to identify a combination of concentrations that provides an additive effect (as indicated by BLISS index in figure 2C and S2E). Of note, for downstream analysis of the molecular mechanisms underlying the treatment response (RNAseq and Cut&Run), we intentionally used concentrations that provide an additive BLISS index, but do not completely abolish the culture, to allow for cellular analysis (i.e. 10uM Sulfopin and 1uM Vorinostat).

The introduction mentions MYC amplification in high-grade gliomas. It would be beneficial if the authors could delineate whether the models used exhibit varying degrees of MYC amplification and how this factor, alongside differences in H3K27me3, contributes to the observed effects of the treatment.

The reviewer highlights an important part of our study relating to the MYC-dependent sensitivity of the proposed treatment combination. Since high expression of MYC can be mediated by different molecular mechanisms and not only genomic amplification, we directly quantified mRNA levels of MYC by qPCR (shown in figure S2G) in order to explore its relationship with cellular response to Sulfopin and Vorinostat. Indeed, cultures that express high levels of MYC mRNA were more sensitive to Sulfopin treatment alone (figure S1P) and to the combined treatment (figure 2D-E). We also relate to these findings in lines 103-106 and 142-147 of the results section. Importantly, in cultures that express high levels of MYC (SU-DIPG13 as an example), we see downregulation of MYC targets upon the combined treatment, supporting the notion that this treatment affects viability by attenuation of MYC signaling.

In Figure 2A, the authors outline an optimal treatment timing for their in vitro models, which appears to be used throughout the figure. It would be helpful to know how this treatment timing was selected and also why Sulfopin is dosed first (and twice) before the vorinostat. Was this optimized?

As PIN1 regulates the G2/M transition, its inhibition by Sulfopin delays cell cycle progression (Yeh et al. 2007). Thus, in order to observe a strong viability difference in culture, a prolonged treatment period of 8-9 days is required (Dubiella et al., 2021). To maintain an active concentration of the drug during this long time period, we added a Sulfopin pulse (2nd dose) to achieve a stronger effect on cell viability. We and others noticed that, unlike Sulfopin, the effect of Vorinostat on viability is rapid and can be clearly seen after 2-3 days of treatment. Thus, we added this drug only after the 2nd dose of Sulfopin. We now relate to the mode of action of Sulfopin in lines 79-81.

It should be clarified whether the dosing timeline for the combination drug experiments in Figure 3 aligns with that of Figure 2. This information is also important for interpreting the epigenetic and transcriptional profiling and the timing should be discussed if they are administered sequentially (also shown in Figure 2A).I have the same question for the mouse experiments in Figure 4.

The reviewer is correct that this information is critical for evaluating the results. In order to link the expression changes to the epigenetic changes, we kept the same experimental conditions in both the Cut&Run and RNA-seq experiments (shown in figures 2-3). We added this information to the text in line 184.

For the in-vivo studies of HDAC inhibition (Figure 4), we followed published protocols (Ehteda et al. 2021). In these experiments both drugs were administrated simultaneously every day. We added this information to the text in line 231-232.  It may be that changing the admission regime may improve the efficacy of the drug combination, which remains to be tested in future studies.

The authors mention that the mice all had severe dehydration and deterioration after 18 days. It would be helpful to know if there were differences in the side effects for different treatment groups? I would expect the combination to be the most severe. This is important in considering the combination treatment.

As noted in our response to Reviewer #1, all mice in our experiment experienced deterioration (including the control mice and those treated with single agents- we could not observe any differences between the groups). This is due to the extremely aggressive tumor cells, forming relatively large tumors prior to the treatment onset, calling for further optimization of the system and therapeutic regime (# of cell injected, treatment starting point, etc.). Unfortunately, this model is very challenging (especially the injection of cells to the pons of the mice brains, which requires unique expertise and is associated with mortality of some of the mice). Thus, these are highly complicated and expensive experiments; unfortunately, we did not have the resources to repeat and optimize the treatment protocol within the scope of this revision. Of note, Dubiella et al. extensively examined toxicity issues and did not observe adverse effects or weight loss when administrating Sulfopin in a dose of 40 mg kg–1. In our model, the side effects were caused by the tumors rather than the drugs.

Minor Points:

(1) For Figure 1F, reorganizing the bars to directly compare the K27M and KO cell lines at each dose would improve readability of this figure.

We have changed figure 1F as the reviewer suggested.

(2) In Figure 4D, it would be helpful to know how many cells were included (or a minimum included) to calculate the percentages.

We added the number of H3-K27M positive cells detected per FOV to the figure legend and method section (n=13-198 cells per FOV). Of note, while we analyzed similar-sized FOVs, the number of tumor cells varied between the groups, with the treated group presenting a lower number of H3-K27M cells (due to the effect of the treatment on tumor growth). To account for this difference, we calculated the portion of mTOR-positive cells out of the tumor cells.   

Reviewer #3 (Public Review):

Summary:

The authors use in vitro grown cells and mouse xenografts to show that a combination of drugs, Sulfopin and Vorinostat, can impact the growth of cells derived from Diffuse midline gliomas, in particular the ones carrying the H3 K27M-mutations (clinically classified as DMG, H3 K27M-mutant). The authors use gene expression studies, and chromatin profiling to attempt to better understand how these drugs exert an effect on genome regulation. Their main findings are that the drugs reduce cell growth in vitro and in mouse xenografts of patient tumours, that DMG, H3 K27M-mutant tumours are particularly sensitive, identify potential markers of gene expression underlying this sensitivity, and broadly characterize the correlations between chromatin modification changes and gene expression upon treatment, identifying putative pathways that may be affected and underlie the sensitive (and thus how the drugs may affect the tumour cell biology).

Strengths:<br /> It is a neat, mostly to-the-point work without exploring too many options and possibilities. The authors do a good job not overinterpreting data and speculating too much about the mechanisms, which is a very good thing since the causes and consequences of perturbing such broad epigenetic landscapes of chromatin may be very hard to disentangle. Instead, the authors go straight after testing the performance of the drugs, identifying potential markers and characterizing consequences.

Weaknesses:

If anything, the experiments done on Figure 3 could benefit from an additional replicate.<br />

We thank the reviewer for evaluating our work, and for the positive and insightful comments.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

Perhaps a more substantial drug screen, or CRISPR screen, that utilises single molecule imaging as a readout would identify pharmacologic candidates that are either more effective, or novel.

While out of scope for the current study, this is a highly interesting suggestion, which will be considered in future studies. Here, we focused on the potential use of the novel MYC inhibitor, Sulfopin. While the dependency of DMG cells on MYC signaling has been documented, to the best of our knowledge, pharmacological inhibition of MYC has not been tested for this disease due to the severe lack of potent MYC inhibitors. We show preliminary evidence for the use of this inhibitor, in combination with HDAC inhibition, to attenuate DMG growth in-vitro and in-vivo.  

Reviewer #2 (Recommendations For The Authors):

In Figure 1B, it is hard to tell if there are error bars for HSP90 and E2F2. Is there a potential error here? Seems unlikely to not have an error with a RT-qPCR?

We thank the reviewer for the careful evaluation of the figures. We included error bars for all genes shown in Figure 1B. We have now increased the line width with the hope of making this information more accessible. As stated in the figure legend, these error bars represent the standard deviation of two technical repeats.

I noticed that many experiments only had technical replicates. Incorporating biological (independent) replicates, where feasible, would strengthen the study's findings.

We agree with the reviewer regarding the importance of biological replicates. While some of the panels present error estimates based on technical repeats, the main results were repeated independently with complementary approaches or various biological systems for validation.

The RNAseq analysis presented in figure 1 was conducted in triplicates and then independently validated by qPCR (Figure 1A-B). Similarly, the transcriptomic analysis presented in figures 2G-I was verified by both western blot (figure 2J) and qPCR (figure S2O). Of note, this later validation was conducted for two different DMG-patient derived cultures.

To verify the robust effects on cellular viability, we analyzed the response to each drug and the combination on eight different DMG-patient-derived cultures, each representing a completely independent experiment. We show very similar trends in response to treatment between cultures that share the same H3-K27M variant. Thus, while for each culture technical repeats are shown, we provide multiple, independent repeats by examining the different cultures. Similarly, in figure 1F we examined the dependency of Sulfopin treatment on the expression of the H3-K27M oncohistone in two independent isogenic systems.

Reviewer #3 (Recommendations For The Authors):

A few questions and suggestions:

(1) To avoid confusion is important to state if the cells used in each experiment are or not K27M mutants (e.g. SU-DIPG13 on line 63).

We thank the reviewer for pointing this out and have now added this information when appropriate across the manuscript.

2) Line 72 - confirming epigenetic silencing of these genes upon PIN1 inhibition (Fig. 1C, S1D)

Considering that the mechanism of down regulation of MYC targets is likely H3K27me3-independent if it is also happening in DMG H3 K27M-mutants (high H3K27me3 here may rather be a consequence of less MYC binding?), I would strike this sentence out and just point out the correlation between lower expression and higher H3K27me3.

We agree with the reviewer that the exact molecular mechanism mediating the silencing is yet to be characterized. We have modified the text in line 72 accordingly.

3) (line 78) Are MYC targets also down regulated in Sulfopin treated DMG, H3 K27M-mutant lines? Any qPCR or previously done RNA-seq data to use?

In addition to the extensive analysis done on SU-DIPG13 cells (Figure 1 and S1), in light of the reviewer`s comment we examined specific MYC targets in an additional H3-K27M mutant DMG culture (SU-DIPG6) treated with Sulfopin, followed by qPCR. We observed a mild reduction in two prominent targets, E2F2 and mTOR (new figure S1D). Unfortunately, within this study, we only conducted full RNA-sequencing analysis on SU-DIPG13 cells treated with Sulfopin, and thus, we could not examine the global effect of Sulfopin on the transcriptome of other DMG cultures. This will, of course, be of high interest for future studies.

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This manuscript aims at a quantitative model of how visual stimuli, given as time-dependent light intensity signals, are transduced into electrical currents in photoreceptors of macaque and mouse retina. Based on prior knowledge of the fundamental biophysical steps of the transduction cascade and a relatively small number of free parameters, the resulting model is found to fairly accurately capture measured photoreceptor currents under a range of diverse visual stimuli and with parameters that are (mostly) identical for photoreceptors of the same type.

Furthermore, as the model is invertible, the authors show that it can be used to derive visual stimuli that result in a desired, predetermined photoreceptor response. As demonstrated with several examples, this can be used to probe how the dynamics of phototransduction affect downstream signals in retinal ganglion cells, for example, by manipulating the visual stimuli in such a way that photoreceptor signals are linear or have reduced or altered adaptation. This innovative approach had already previously been used by the same lab to probe the contribution of photoreceptor adaptation to differences between On and Off parasol cells (Yu et al, eLife 2022), but the present paper extends this by describing and testing the photoreceptor model more generally and in both macaque and mouse as well as for both rods and cones.

Strengths:

The presentation of the model is thorough and convincing, and the ability to capture responses to stimuli as different as white noise with varying mean intensity and flashes with a common set of model parameters across cells is impressive. Also, the suggested approach of applying the model to modify visual stimuli that effectively alter photoreceptor signal processing is thought-provoking and should be a powerful tool for future investigations of retinal circuit function. The examples of how this approach can be applied are convincing and corroborate, for example, previous findings that adaptation to ambient light in the primate retina, as measured by responses to light flashes, mostly originates in photoreceptors.

Weaknesses:

In the current form of the presentation, it doesn't become fully clear how easily the approach is applicable at different mean light levels and where exactly the limits for the model inversion are at high frequency. Also, accessibility and applicability by others could be strengthened by including more details about how parameters are fixed and what consensus values are selected.

Thank you - indeed a central goal of writing this paper was to provide a tool that could be easily used by other laboratories. We have clarified and expanded four points in this regard: (1) we have stated more clearly that mean light levels are naturally part of inversion process, and hence the approach can be applied across a broad range of light levels (lines 292-297); (2) we have expanded our analysis of the high frequency limits to the inversion and added that expanded figure to the main text (new Fig 5); (3) we have included additional detail about our calibration procedures, including our calibration code, to facilitate transfer to other labs; and, (4) we have detailed the procedure for identification of consensus parameters (line 172-182, 191-199 and Methods section starting on line 831).

Reviewer #2 (Public Review):

Summary:

This manuscript proposes a modeling approach to capture nonlinear processes of photocurrents in mammalian (mouse, primate) rod and cone photoreceptors. The ultimate goal is to separate these nonlinearities at the level of photocurrent from subsequent nonlinear processing that occurs in retinal circuitry. The authors devised a strategy to generate stimuli that cancel the major nonlinearities in photocurrents. For example, modified stimuli would generate genuine sinusoidal modulation of the photocurrent, whereas a sinusoidal stimulus would not (i.e., because of asymmetries in the photocurrent to light vs. dark changes); and modified stimuli that could cancel the effects of light adaptation at the photocurrent level. Using these modified stimuli, one could record downstream neurons, knowing that any nonlinearities that emerge must happen post-photocurrent. This could be a useful method for separating nonlinear mechanisms across different stages of retinal processing, although there are some apparent limitations to the overall strategy.

Strengths:

(1) This is a very quantitative and thoughtful approach and addresses a long-standing problem in the field: determining the location of nonlinearities within a complex circuit, including asymmetric responses to different polarities of contrast, adaptation, etc.

(2) The study presents data for two primary models of mammalian retina, mouse, and primate, and shows that the basic strategy works in each case.

(3) Ideally, the present results would generalize to the work in other labs and possibly other sensory systems. How easy would this be? Would one lab have to be able to record both receptor and post-receptor neurons? Would in vitro recordings be useful for interpreting in vivo studies? It would be useful to comment on how well the current strategy could be generalized.

We agree that generalization to work in other laboratories is important, and indeed that was a motivation for writing this as a methods paper. The key issue in such generalization is calibration. We have expanded our discussion of our calibration procedures and included that code as part of the github repository associated with the paper. Figure 10 (previously Figure 9) was added to illustrate generalization. We believe that the approach we introduce here should generalize to in vivo conditions. We have expanded the text on these issues in the Discussion (sections starting on line 689 and 757).

Weaknesses:

(1) The model is limited to describing photoreceptor responses at the level of photocurrents, as opposed to the output of the cell, which takes into account voltage-dependent mechanisms, horizontal cell feedback, etc., as the authors acknowledge. How would one distinguish nonlinearities that emerge at the level of post-photocurrent processing within the photoreceptor as opposed to downstream mechanisms? It would seem as if one is back to the earlier approach, recording at multiple levels of the circuit (e.g., Dunn et al., 2006, 2007).

Indeed the current model is limited to a description of rod and cone photocurrents. Nonetheless, the transformation of light inputs to photocurrents can be strongly nonlinear, and such nonlinearities can be difficult to untangle from those occurring late in visual processing. Hence, we feel that the ability to capture and manipulate nonlinearities in the photocurrents is an important step. We have expanded Figure 10 to show an additional example of how manipulation of nonlinearities in phototransduction can give insight into downstream responses. We have also noted in text that an important next step would be to include inner segment mechanisms (section starting on line 661); doing so will require not only characterization of the current-to-voltage transformation, but also horizontal cell feedback and properties of the cone output synapse.

(2) It would have been nice to see additional confirmations of the approach beyond what is presented in Figure 9. This is limited by the sample (n = 1 horizontal cell) and the number of conditions (1). It would have been interesting to at least see the same test at a dimmer light level, where the major adaptation mechanisms are supposed to occur beyond the photoreceptors (Dunn et al., 2007).

We have added an additional experiment to this figure (now Figure 10) which we feel nicely exemplifies the approach. The approach that we introduce here really only makes sense at light levels where the photoreceptors are adapting; at lower light levels the photoreceptors respond near-linearly, so our “modified” and “original” stimuli as in Figure 10 (previously Figure 9) would be very similar (and post-phototransduction nonlinearities are naturally isolated at these light levels).

Reviewer #3 (Public Review):

Summary:

The authors propose to invert a mechanistic model of phototransduction in mouse and rod photoreceptors to derive stimuli that compensate for nonlinearities in these cells. They fit the model to a large set of photoreceptor recordings and show in additional data that the compensation works. This can allow the exclusion of photoreceptors as a source of nonlinear computation in the retina, as desired to pinpoint nonlinearities in retinal computation. Overall, the recordings made by the authors are impressive and I appreciate the simplicity and elegance of the idea. The data support the authors' conclusions but the presentation can be improved.

Strengths:

-  The authors collected an impressive set of recordings from mouse and primate photoreceptors, which is very challenging to obtain.

-  The authors propose to exploit mechanistic mathematical models of well-understood phototransduction to design light stimuli that compensate for nonlinearities.

-  The authors demonstrate through additional experiments that their proposed approach works.

Weaknesses:

-  The authors use numerical optimization for fitting the parameters of the photoreceptor model to the data. Recently, the field of simulation-based inference has developed methods to do so, including quantification of the uncertainty of the resulting estimates. Since the authors state that two different procedures were used due to the different amounts of data collected from different cells, it may be worthwhile to rather test these methods, as implemented e.g. in the SBI toolbox (https://joss.theoj.org/papers/10.21105/joss.02505). This would also allow them to directly identify dependencies between parameters, and obtain associated uncertainty estimates. This would also make the discussion of how well constrained the parameters are by the data or how much they vary more principled because the SBI uncertainty estimates could be used.

Thank you - we have improved how we describe and report parameter values in several ways. First, the previous text erroneously stated that we used different fitting procedures for different cell types - but the real difference was in the amount of data and range of stimuli we had available between rods and cones. The fitting procedure itself was the same for all cell types. We have clarified this along with other details of the model fitting both in the main text (lines 121-130) and in the Methods (section starting on line 832). We also collected parameter values and estimates of allowed ranges in two tables. Finally, we used sloppy modeling to identify parameters that could covary with relatively small impact on model performance; we added a description of this analysis to the Methods (section starting on line 903).

-  In several places, the authors refer the reader to look up specific values e.g. of parameters in the associated MATLAB code. I don't think this is appropriate, important values/findings/facts should be in the paper (lines 142, 114, 168). I would even find the precise values that the authors measure interesting, so I think the authors should show them in a figure/table. In general, I would like to see also the average variance explained by different models summarized in a table and precise mean/median values for all important quantities (like the response amplitude ratios in Figures 6/9).

We have added two tables with these parameters values and estimates of allowable ranges. We also added points to show the mean (and SD) across cells to the population figures and added those numerical values to the figure legends throughout.

-  If the proposed model is supposed to model photoreceptor adaptation on a longer time scale, I fail to see why this can be an invertible model. Could the authors explain this better? I suspect that the model is mainly about nonlinearities as the authors also discuss in lines 360ff.

For the stimuli that we use we see little or no contribution of slow adaptation in phototransduction. We have expanded the description of this point in the text and referred to Angueyra et al (2022) which looks at this issue in more detail for primate cones (paragraph starting on line 280).

-  The important Figures 6-8 are very hard to read, as it is not easy to see what the stimulus is, the modified stimulus, the response with and without modification, what the desired output looks like, and what is measured for part B. Reworking these figures would be highly recommended.

We have reworked all of the figures to make the traces clearer.

-  If I understand Figure 6 correctly, part B is about quantifying the relative size of the response to the little first flash to the little second flash. While clearly, the response amplitude of the second flash is only 50% for the second flash compared to the first flash in primate rod and cones in the original condition, the modified stimulus seems to overcompensate and result in 130% response for the second flash. How do the authors explain this? A similar effect occurs in Figure 9, which the authors should also discuss.

Indeed, in those instances the modified stimulus does appear to overcompensate. We suspect this is due to differences in sensitivity of the specific cells probed for these experiments and those used in the model construction. We now describe this limitation in more detail (lines 524-526). A similar point comes up for those experiments in which we speed the photoreceptor responses (new FIgure 9B), and we similarly note that the cells used to test those manipulations differed systematically from those used to fit the model (lines 558-560).

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

I only have a few minor questions and suggestions for clarification.

It hasn't become fully clear to me how general the model is when different mean light levels (on long-time scales) are considered. Are there slow adaptation processes not captured in the model that affect model performance? And how should one go about setting the mean light level when, for example, probing ganglion cells with a stimulus obtained through model inversion? Should it work to add an appropriate DC component to the current that is provided as input to the inverted model? (Presumably, deriving a stimulus and then just adding background illumination should not work, or could this be a good approximation, given a steady state that is adapted to the background?)

We have clarified in the main text that slow adaptation does not contribute substantially to responses to the range of stimuli we explored (lines 281-289). We have also clarified that the stimulus in the model inversion is specified in isomerizations per second - so the mean value of the stimulus is automatically included in the model inversion (lines 293-298).

Furthermore, a caveat for the model inversion seems to be the potential amplification of high-frequency noise. The suggested application of a cutoff temporal frequency seems appropriate, but data are shown only for a few example cells. Is this consistent across cells? (Given that performance between, e.g., mouse cones can vary considerably according to Fig. 4B?) I would also like to suggest moving the corresponding Supplemental Figure (4.1) into the main part of the manuscript, as it seems quite important.

We have added population analysis to the new Figure 5 (which was Figure 4 - Figure Supplement 1). We have also clarified that the amplification of high frequency noise is an issue only when we try to apply model inversion to measured stimuli. When we use model inversion to identify stimuli that elicit desired responses, the target responses are computed from a linear model that has no noise, so this is not a concern in applications like those in Figures 6-10.

Also, could the authors explain more clearly what the effect of the normalization of the estimated stimulus by the power of the true stimulus is? Does this simply reduce power at high frequency or also affect frequencies below the suggested cutoff (where the stimulus reconstruction should presumably be accurate even without normalization)?

Indeed this normalization reduces high frequency power and has little impact on low frequencies where the inversion is accurate; this is now noted in the text (line 363). As for amplification of high frequency noise (previous comment), the normalization by the stimulus power is only needed when inverting measured responses (i.e. responses with noise) and is omitted when we are identifying stimuli that elicit desired responses (e.g. in Figures 6-10).

While the overall performance of the model to predict photoreceptor currents is impressive, it seems that particular misses occur for flashes right after a step in background illumination and for the white-noise responses at low background illumination (e.g. Figure 1B). Is that systematic, and if so what might be missing in the model?

Indeed the model (at least with fixed parameters across stimuli) appears to systematically miss a few aspects of the photoreceptor responses. These include the latency of the response to a bright flash and the early flashes in the step + flash protocol in Figure 1B. Model errors for the variable mean noise stimulus (Figure 2) showed little dependence on time even when responses were sorted by mean light level and by previous mean level. Model errors did not show a clear systematic dependence on light level; this likely reflects, at least in part, the use of mean-square-error to identify model parameters. We have expanded our discussion of these systematic errors in the text (lines 164-166).

I was also wondering whether this is related to the fact that in Figure 9B, the gain in the modified condition is actually systematically higher when there is more background light. Do the authors think that this could be a real effect or rather an overcompensation from the model? (By the way, is it specified what "Delta-gain" really is, i.e., ratio or normalized difference?)

We suspect this is an issue with the sensitivity of the specific cells for which we did these experiments (i.e. variability in the gamma parameter between cells). This sensitivity varies between cells, and such variations are likely to place the strongest limitation on our ability to use this approach to manipulate responses in different retinas. We now note those issues in the Results (lines 523-526, 557-559 and 591-593) with reference to Figures 9 (previously Figure 8) and 10 (previously Figure 9), and describe this limitation more generally in the Discussion (section starting on line 649). We have also changed delta-gain to response ratio, which seemed more intuitive.

Maybe I missed this, but it seems that the parameter gamma is fitted in a cell-type-specific fashion (e.g. line 163), but then needs to be fixed for held-out cells. How was this done? Is there much variability of gamma between cells?

There is variability in gamma between cells, and this likely explains some of systematic differences between data and model (see above and Methods, lines 902-903). For the consensus models in Figure 2B, gamma was allowed to vary for each cell while the remaining consensus model parameters were fixed. Gamma was set equal to the mean value across cells for model inversion (i.e. for all of the analyses in Figures 4-10). We have described the fitting procedure in considerably more detail in the revised Methods (starting on line 832).

For completeness, it would be nice to have the applied consensus model parameters in the manuscript rather than just in the Matlab code (especially since the code has not been part of the submission). Also, some notes on how the numerical integration of the differential equations was done would be nice (time step size?).

We have added tables with consensus parameters and estimates of the sensitivity of model predictions to each parameter. We have also added additional details about the numerical approaches (including the time step) to Methods.

Similarly, it would be nice to explicitly see the relationships that are used to fix certain model parameters (lines 705ff). And can the constants k and n (lines 709-710) be assumed identical for different species and receptor types?

We have added more details to the model fitting to the methods, including the use of steady-state conditions to hold certain parameters fixed (lines 862 and 866). We are not aware of any direct comparisons of k and n across species and receptor types. We have noted that model performance was not improved by modest changes in these parameters (due to compensation by other model parameters). More generally, we have explained how some parameters trade for others and hence the logic of fixing some even when exact values were not available.

For the previous measurements of m and beta (lines 712-713), is there a reference or source?

We have added references for these values.

Did the authors check for differences in the model parameters between cone types (e.g., S vs. M)?

We did not include S cones here. They are harder to record from and collecting a fairly large data set across a range of stimuli would be challenging. Our previous work shows that S cones have slower responses than L and M cones, and this would certainly be reflected in differences in model parameters. We have noted this in the text (Methods, line 808-810).

For the stated flash responses time-to-peak (lines 183-184), is this for a particular light intensity with no background illumination?

Those are flashes from darkness - now noted in the text.

Figure 2 - Supplement 1 doesn't have panel labels A and B, unlike the legend.

Fixed - thank you.

Reviewer #2 (Recommendations For The Authors):

(1) Fig. 2B - for some cells, the consensus model seems to fit better than the individual model. How is this possible?

This was mostly an error on our part (we inadvertently included responses to more stimuli in fitting the individual models, which slightly hampered their performance). Even with this correction, however, a few cells remain for which the consensus model outperforms and individual model. We believe this is because there is more data to constrain model parameters for the consensus models (since they are fit to all cells at the same time), and that can compensate for improvements associated with customizing parameters to specific cells.

(2) Fig. 2 Supplement 1, it would be useful to see a blow-up of the data in an inset, as in Fig. 2B.

Thanks - added.

(3) Line 400 - this paragraph could include additional quantification and statistics to back up claims re 'substantially reduced', 'considerably lower'.

We quantify that in the next sentence by computing the mean-square-error between responses and sinusoidal fits (also in Figure 7B, which now includes statistics as well). We have made that connection more direct in the text.

(4) Maybe a supplement to Fig. 8 could show the changes to the stimulus required to alter the kinetics in both directions - to give more insight into part B., especially.

Good suggestion - we have added the stimuli to all of the panels of the figure (now Figure 9).

(5) Fig. 8B - in 'Speed response up' condition - there seems to be error in the model for the decay time of the response - especially for the 'original' condition, which is not quantified in 8C. Was it generally difficult to predict responses to flashes?

That seems largely to reflect that the cells used for those experiments had faster initial kinetics than the average cells (responses to the control traces are also faster than model predictions in these cells - black traces in Figure 9B). We have added this to the text.

(6) Line 678, possibly notes that 405 nm equally activates S and M photopigments in mice, since most of the cones co-express the two photopigments (Rohlich et al., 1994; Applebury et al., 2000; Wang et al., 2011).

Thanks - we have added this (lines 827-829).

(7) The discussion could include a broader description of the various approaches to identifying nonlinearities within retinal circuitry, which include (incomplete list): recording at multiple levels of the circuit (e.g., Kim and Rieke 2001; Rieke, 2001; Baccus and Meister, 2002; Dunn et al., 2006; 2007; Beaudoin et al., 2007; Baccus et al., 2008); recording currents vs. spiking responses in a ganglion cell (e.g., Kim and Rieke, 2001; Zaghloul et al., 2005; Cui et al., 2016); neural network modeling approaches (e.g., Maheswaranathan et al., 2023); optogenetic approaches to studying filtering/nonlinear behavior at synapses (e.g., Pottackal et al., 2020; 2021).

Good suggestion - we have added this to the final paragraph of the Discussion.

Reviewer #3 (Recommendations For The Authors):

-  I am personally not a fan of the style: "... as Figure 4A shows..." or comparable and much prefer a direct "We observe that X is the case (Figure 4A)". If the authors agree, they may want to revise their paper in this way.

We have revised the text to avoid the “... as Figure xx shows” construction. We have retained multiple instances which follow a “Figure xx shows that …” construction (which is both active rather than passive and does not use a personal pronoun).

-  I am not a fan of the title. Light-adaption clamp caters only to a very specialized audience.

We have changed the title to “Predictably manipulating photoreceptor light responses to reveal their role in downstream visual responses.”

-  The parameter fitting procedure should not only be described in Matlab code, but in the paper.

Thanks - we have expanded this in the Methods considerably (section starting on line 832).

-  The authors should elaborate on why different fitting procedures were used.

We did not describe that issue clearly. The fitting procedures used across cells were identical, but we had different data available for different cell types due to experimental limitations. We have substantially revised that part of the main text to clarify this issue (paragraph starting on line 121).

-  The authors state in line 126 that the input stimulus is supposed to mimic eye movements mouse, monkey, or human? Please clarify.

Thanks - we have changed this sentence to “abrupt and frequent changes in intensity that characterize natural vision.”

-  Please improve the figure style. For example, labels should be in consistent capitalization and ideally use complete words (e.g. Figure 2B, 4B, and others).

We have made numerous small changes in the figures to make them more consistent.

-  Is the fraction of variance calculated on held-out-data? Linear models should be added to Figure 2B.

The fraction of variance explained was not calculated on held out data because of limitations in the duration of our recordings. Given the small number of free parameters, and the ability of the model to capture held out cells, we believe that the model generalizes well. We have added a supplemental figure with linear model performance (Figure 2 - Figure Supplement 2).

-  Fig. 9A is lacking bipolar cell and amacrine cell labels. Currently, it looks like HC is next to the BC in the schematic.

Thanks - we have updated that figure (now Figure 10A)

-  Maybe I am misunderstanding something, but it seems like the linear model prediction shown in Figure 2A for the rod could be easily improved by scaling it appropriately. Is this impression correct or why not?

We have clarified how the linear model is constructed (by fitting the linear model to low contrast responses of the full model at the mean stimulus intensity). We also added a supplemental figure, following the suggestion above, showing the linear model performance when a free scaling factor is included for each cell.

-  The verification experiment in Fig. 5 is only anecdotal and is elaborated only in Figure 6. If I am not mistaken, this does not necessitate its own figure/section but could rather be merged.

We have kept this figure separate (now Figure 6) as we felt that it was important to highlight the approach in general in a figure before getting into quantification of how well it works.

-  Figure 5 right is lacking labels. What is red and grey?

Thanks for catching that - labels are added now.

-  The end of the Discussion is slightly unusual. Did some text go missing?

Thanks - we have rearranged the Discussion so as not to end on Limitations.

-  There is a bonus figure at the end which seems also not to belong in the manuscript.

Thanks - the bonus figure is removed now.

-  The methods should also describe briefly what kind of routines were used in the Matlab code, e.g. gradient descent with what optimizer?

We’ve added that information as well.

Author response:

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

Reviewer 1:

(1) Peptides were synthesized with fluorescein isothiocyanate (FITC) and Tat tag, and then PEGylated with methoxy PEG Succinimidyl Succinate.

I have two concerns about the peptide design. First, FTIC was intended "for monitoring" (line 129), but was never used in the manuscript. Second, PEGylation targets the two lysine sidechains on the Tat, which would alter its penetration property.

We conducted an analysis of the cellular trafficking of FITC-tagged peptides following their permeabilization into cells.

Author response image 1.

However, we did not include it in the main text because it is a basic result.

(2) As can be seen in the figure above, after pegylation and permeabilization, the cells were stained with FITC. It appears that this does not affect the ability to penetrate into the cells.

(2) "Superdex 200 increase 10/300 GL column" (line 437) was used to isolate mono/di PEGylated PDZ and separate them from the residual PEG and PDZ peptide. "m-PEG-succinimidyl succinate with an average molecular weight of 5000 Da" (lines 133 and 134).

To my knowledge, the Superdex 200 increase 10/300 GL column is not suitable and is unlikely to produce traces shown in Figure 1B.

As Superdex 200 increase 10/300 GL featrues a fractionation range of 10,000 to 600,000 Da, we used it to fractionate PEGylated products including DiPEGylated PDZ (approx. 15 kDa) and MonoPEGylated PDZ (approx. 10 kDa) from residuals (PDZ and PEG), demonstrating successful isolation of PEGylated products (Figure 1C). Considering the molecular weights of PDZ and PEG are approximately 4.1 kDa and and 5.0 kDa, respectively, the late eluting peaks from SEC were likely to represent a mixed absorbance of PDZ and PEG at 215 nm.

However, as the reviewer pointed out, it could be unreasonable to annotate peaks representing PDZ and PEG, respectively, from mixed absorbance detected in a region (11-12 min) beyond the fractionation range.

In our revised manuscript, therefore, multiple peaks in the late eluting volume (11-12 min) were labeled as 'Residuals' all together. As a reference, the revised figure 1B includes a chromatogram of pure PDZ-WT under the same analytic condition.

Therefore, we changed Fig.1B to new results as followed:

(3) "the in vivo survival effect of LPS and PDZ co-administration was examined in mice. The pretreatment with WT PDZ peptide significantly increased survival and rescued compared to LPS only; these effects were not observed with the mut PDZ peptide (Figure 2a)." (lines 159-160).

Fig 2a is the weight curve only. The data is missing in the manuscript.

We added the survived curve into Fig. 2A as followed:

(4) Table 1, peptide treatment on ALT and AST appears minor.

In mice treated with LPS, levels of ALT and AGT in the blood are elevated, but these levels decrease upon treatment with WT PDZ. However, the use of mut PDZ does not result in significant changes. Figure 3A shows inflammatory cells within the central vein, yet no substantial hepatotoxicity is observed during the 5-day treatment with LPS. Normally, the ranges of ALT and AGT in C57BL6 mice are 16 ~ 200 U/L and 46 ~ 221 U/L, respectively, according to UCLA Diagnostic Labs. Therefore, the values in all experiments fall within these normal ranges. In summary, a 5-day treatment with LPS induces inflammation in the liver but is too short a duration to induce hepatotoxicity, resulting in lower values.

(5) MitoTraker Green FM shouldn't produce red images in Figure 6.

We changed new results (GREEN one) into Figs 6A and B as followed:

(6) Figure 5. Comparison of mRNA expression in PDZ-treated BEAS-2B cells. Needs a clearer and more detailed description both in the main text and figure legend. The current version is very hard to read.

We changed Fig. 5A to new one to understand much easier and added more detailed results and figure legend as followed:

Results Section in Figure 5:

“…we performed RNA sequencing analysis. The results of RNA-seq analysis showed the expression pattern of 24,424 genes according to each comparison combination, of which the results showed the similarity of 51 genes overlapping in 4 gene categories and the similarity between each comparison combination (Figure 5a). As a result, compared to the control group, it was confirmed that LPS alone, WT PDZ+LPS, and mut PDZ+LPS were all upregulated above the average value in each gene, and when LPS treatment alone was compared with WT PDZ+LPS, it was confirmed that they were averaged or downregulated. When comparing LPS treatment alone and mut PDZ+LPS, it was confirmed that about half of the genes were upregulated. Regarding the similarity between comparison combinations, the comparison combination with LPS…”

Figure 5 Legend Section:

“Figure 5. Comparison of mRNA expression in PDZ-treated BEAS-2B cells.

BEAS-2B cells were treated with wild-type PDZ or mutant PDZ peptide for 24 h and then incubated with LPS for 2 h, after which RNA sequencing analysis was performed. (a) The heat map shows the general regulation pattern of about 51 inflammation-related genes that are differentially expressed when WT PDZ and mut PDZ are treated with LPS, an inflammatory substance. All samples are RED = upregulated and BLUE = downregulated relative to the gene average. Each row represents a gene, and the columns represent the values of the control group treated only with LPS and the WT PDZ and mut PDZ groups with LPS. This was used by converting each log value into a fold change value. All genes were adjusted to have the same mean and standard deviation, the unit of change is the standard deviation from the mean, and the color value range of each row is the same. (b) Significant genes were selected using Gene category chat (Fold change value of 2.00 and normalized data (log2) value of 4.00). The above pie chart shows the distribution of four gene categories when comparing LPS versus control, WT PDZ+LPS/LPS, and mut PDZ+LPS/LPS. The bar graph below shows RED=upregulated, GREEN=downregulated for each gene category, and shows the number of upregulated and downregulated genes in each gene category. (c) The protein-protein interaction network constructed by the STRING database differentially displays commonly occurring genes by comparing WT PDZ+LPS/LPS, mut PDZ+LPS/LPS, and LPS. These nodes represent proteins associated with inflammation, and these connecting lines denote interactions between two proteins. Different line thicknesses indicate types of evidence used in predicting the associations.”

Reviewer 2:

(1) In this paper, the authors demonstrated the anti-inflammatory effect of PDZ peptide by inhibition of NF-kB signaling. Are there any results on the PDZ peptide-binding proteins (directly or indirectly) that can regulate LPS-induced inflammatory signaling pathway? Elucidation of the PDZ peptide-its binding partner protein and regulatory mechanisms will strengthen the author's hypothesis about the anti-inflammatory effects of PDZ peptide

As mentioned in the Discussion section, we believe it is crucial to identify proteins that directly interact with PDZ and regulate it. This direct interaction can modulate intracellular signaling pathways, so we plan to express GST-PDZ and induce binding with cellular lysates, then characterize it using the LC-Mass/Mass method. We intend to further research these findings and submit them for publication.

(2) The authors presented interesting insights into the therapeutic role of the PDZ motif peptide of ZO-1. PDZ domains are protein-protein interaction modules found in a variety of species. It has been thought that many cellular and biological functions, especially those involving signal transduction complexes, are affected by PDZ-mediated interactions. What is the rationale for selecting the core sequence that regulates inflammation among the PDZ motifs of ZO-1 shown in Figure 1A?

The rationale for selecting the core sequence that regulates inflammation among the PDZ motifs of ZO-1, as shown in Figure 1A, is grounded in the specific roles these motifs play in signal transduction pathways that are crucial for inflammatory processes. PDZ domains are recognized for their ability to function as scaffolding proteins that organize signal transduction complexes, crucial for modulating cellular and biological functions. The chosen core sequence is particularly important because it is conserved across ZO-1, ZO-2, and ZO-3, indicating a fundamental role in maintaining cellular integrity and signaling pathways. This conservation suggests that the sequence’s involvement in inflammatory regulation is not only significant in ZO-1 but also reflects a broader biological function across the ZO family.

(3) In Figure 3, the authors showed the representative images of IHC, please add the quantification analysis of Iba1 expression and PAS-positive cells using Image J or other software. To help understand the figure, an indication is needed to distinguish specifically stained cells (for example, a dotted line or an arrow).

We added the semi-quantitative results into Figs. 4d,e,f as followed:

Result section: “The specific physiological mechanism by which WT PDZ peptide decreases LPS-induced systemic inflammation in mice and the signal molecules involved remain unclear. These were confirmed by a semi-quantitative analysis of Iba-1 immunoreactivity and PAS staining in liver, kidney, and lung,respectively (Figures 4d, e, and f). To examine whether WT PDZ peptide can alter LPS-induced tissue damage in the kidney, cell toxicity assay was performed (Figure 3g). LPS induced cell damage in the kidney, however, WT PDZ peptide could significantly alleviate the toxicity, but mut PDZ peptide could not. Because cytotoxicity caused by LPS is frequently due to ROS production in the kidney (Su et al., 2023; Qiongyue et al., 2022), ROS production in the mitochondria was investigated in renal mitochondria cells harvested from kidney tissue (Figure 3h)....”

Figure legend section: “Indicated scale bars were 20 μm. (d,e,f) Semi-quantitative analysis of each are positive for Iba-1 in liver and kidney, and positive cells of PAS in lung, respectively. (g) After the kidneys were harvested, tissue lysates were used for MTT assay. (h) After...”

(4) In Figure 6G, H, the authors confirmed the change in expression of the M2 markers by PDZ peptide using the mouse monocyte cell line Raw264.7. It would be good to add an experiment on changes in M1 and M2 markers caused by PDZ peptides in human monocyte cells (for example, THP-1).

We thank you for your comments. To determine whether PDZ peptide regulates M1/M2 polarization in human monocytes, we examined changes in M1 and M2 gene expression in THP-1 cells. As a result, wild-type PDZ significantly suppressed the expression of M1 marker genes (hlL-1β, hIL-6, hIL-8, hTNF-ɑ), while increasing the expression of M2 marker genes (hlL-4, hIL-10, hMRC-1). However, mutant PDZ did not affect M1/M2 polarization. These results suggest that PDZ peptide can suppress inflammation by regulating M1/M2 polarization of human monocyte cells. These results are for the reviewer's reference only and will not be included in the main content.

Author response image 2.

Author response image 3.

Minor point:

The use of language is appropriate, with good writing skills. Nevertheless, a thorough proofread would eliminate small mistakes such as:

- line 254, " mut PDZ+LPS/LPS (45.75%) " → " mut PDZ+LPS/LPS (47.75%) "

- line 296, " Figure 6f " → " Figure 6h "

We changed these points into the manuscript.

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):  

In this study, Hunt et al investigated the role of the ubiquitin-conjugating enzyme UBE2D/effete (eff) in maintaining proteostasis during aging. Utilizing Drosophila as a model, the researchers observed diverse roles of E2 ubiquitinconjugating enzymes in handling the aggregation-prone protein huntingtin-polyQ in the retina. While some E2s facilitated aggregate assembly, UBE2D/eff and other E2s were crucial for degradation of hL-polyQ. The study also highlights the significance of UBE2D/eff in skeletal muscle, showing that declining levels of eff during aging correlate with proteostasis disruptions. Knockdown of eff in muscle led to accelerated accumulation of poly-ubiquitinated proteins, shortened lifespan, and mirrored proteomic changes observed in aged muscles. The introduction of human UBE2D2, analogous to eff, partially rescued the deficits in lifespan and proteostasis caused by eff-RNAi expression in muscles. 

The conclusions of this paper are mostly well supported by data, although a more precise mechanistic explanation of phenotypes associated with UBE2D/eff deficiency would have strengthened the study. Additionally, some aspects of image quantification and data analysis need to be clarified and/or extended.  

We thank reviewer #1 for the thoughtful assessment of our work. We have amended the discussion to better explain the phenotypes associated with UBE2D/eff deficiency. We have also improved the methods describing the procedures for image quantification and data analysis.

Reviewer #2 (Public Review):  

Important findings: 

- Knockdown of UBE2D increases HTT aggregation. 

- Knockdown of UBE2D leads to an accumulation of ubiquitinated proteins and reduces the lifespan of Drosophila, which is rescued by an ectopic expression of the human homolog. 

- UBE2D protein levels decline with aging. 

- UBE2D knockdown is associated with an up- and downregulation of several different cellular pathways, including proteostasis components. 

Thank you for reviewing our manuscript.

Caveats: 

- The readout of HTT aggregation (with methods that are not suitable) as a proxy for the role of UBE2D in proteostasis is not convincing. It would probably improve the manuscript to start with the proteomic analysis of UBE2D to demonstrate that its protein levels decrease with aging. The authors could then induce UBE2D in aged animals to assess the role of UBE2D in the proteome with aging.  

While presenting the data in a different order would be possible, we prefer to keep the current order in which from a general screen with a proteostasis readout (HTT aggregates; see the answer below for a discussion on the methods) we proceed to identify a candidate (UBE2D) which is then studied in more detail with additional focused analyses in the retina and skeletal muscle during aging. Concerning the induction of UBE2D in aged animals, our analyses in Figure 4E demonstrate that muscle-specific induction of UBE2D2 throughout life does not increase lifespan alone: this could be explained by UBE2D2 only partially recapitulating the function and substrate diversity of Drosophila eff/UBE2D due to divergence from a single Drosophila UBE2D enzyme (eff) to multiple UBE2D enzymes in humans (UBE2D1/2/3/4).

- UBE2D knockdown increases the number of HTT foci (Figure 1A), but the quantification is less convincing as depicted in Figure 1B, and other E2 enzymes show a stronger effect (e.g. Ubc6 that is only studied in Figures 1 and 2 without an explanation and Ubc84D). The graph is hard to interpret. What is the sample size and which genetic conditions show a significant change? P values and statistical analyses are missing.  

The full data underlying this genetic screen is reported in Supplementary Table 1. The role of UBC6/UBE2A/B is thoroughly examined in Hunt et al 2021 (PMID: 33658508). We agree that Ubc84D has an important effect and that it should be considered for future studies. We have amended the legend of Figure 1 to indicate that each data point in the graph represents a single RNAi line targeting the corresponding gene. The mean of 5 biological replicates is shown for each RNAi, with each biological replicate representing a single eye imaged from a distinct fly. Therefore, the data points that do not show large magnitude changes may indicate RNAi lines that were not effective at knocking down the target protein (or that did not affect HTT aggregates). The E2s worth pursuing were identified because of multiple RNAi lines scoring consistently: this is the case of UBC6 (studied previously in PMID: 33658508) and eff/UBE2D (pursued in this study). This screen was therefore utilized to identify and select candidate genes (i.e. eff/UBE2D) for more in-depth studies on proteostasis.

- The quantification of the HTT fluorescence cannot be used as a proxy for HTT aggregation. The authors should assess HTT aggregation by e.g. SDD-AGE, FRAP, filter retardation, etc. The quantification of the higher MW species of HTT in the SDS-PAGE is not ideal either as this simply reflects material that is stuck in the wells that could not enter the gel. Aggregation and hence high MW size could be one reason, but it can also be HTT trapped in cell debris, etc.  

We agree that the use of multiple methods is a good way to assess the impact of E2 enzymes on HTT protein aggregation. In this regard, we estimated HTT aggregates by fluorescence microscopy and by western blot. Microscopy-based analyses demonstrate both the accumulation of the HTT-GFP pathogenic protein into aggregates (HTT polyQ polypeptides aggregating into one spatial region; Fig. 1 and Fig. 2B) as well as their potential cytotoxicity, resulting in the disruption of the ommatidial ultrastructure and cellular degeneration (Fig. 2A). Similar to native gels and filter retardation, we have utilized SDS-PAGE and western blotting of cellular samples isolated with strong chaotropic and denaturing reagents (8M urea plus detergents and reducing reagents used in the lysis). These experimental conditions maintain the higher-order organization of HTT into high-molecular-weight aggregates that are not broken down into individual polypeptides and that therefore do not readily travel through a gel or filter. Therefore, the biochemical methods we have used are equivalent to those proposed by the reviewer. In addition to combining microscopy-based and biochemical approaches to examine the impact of eff/UBE2D on the HTT aggregates, we have analyzed eff/UBE2D during skeletal muscle aging and found consistent phenotypes as those observed in the HTT model: RNAi for eff/UBE2D leads to the accumulation of detergent-insoluble ubiquitinated proteins that associate with protein aggregates.

- Does UBE2D ubiquitinate HTT? And thus, is HTT accumulation a suitable readout for the functional assessment of the E2 enzyme UBE2D? 

We propose that the accumulation of HTT in response to eff/UBE2D RNAi may be due to a generalized loss of protein quality control rather than to a direct decline in the ubiquitination of HTT by eff/UBE2D. In a previous study that examined the UBE2D interactome (Hunt et al. 2023; PMID: 37963875), we did not find an interaction between UBE2D and HTT, suggesting that HTT may not be directly modulated by eff/UBE2D via ubiquitination.

- The proteomic analyses could help to identify potential substrates for UBE2D.

The proteomic analyses in Figure 5 identify several proteins that are modulated by RNAi for eff and by its human homolog, UBE2D2. Such eff/UBE2D2-modulated proteins may indeed be potential substrates for UBE2D-mediated ubiquitination. For example, this is the case for Pex11 and Pex13, which were found to be upregulated upon UBE2D RNAi also in human cells, where they are ubiquitinated in a UBE2D-dependent manner (Hunt et al. 2023; PMID: 37963875).

- Are there mutants available for UBE2D or conditional mutants? One caveat of RNAi is: first not complete knockdown and second, variable knockdown efficiencies that increase variability.

There are potential hypomorphic alleles of eff/UBE2D that may be available, but they would present the same caveats of incomplete loss of eff/UBE2D function as RNAi. Given the strong phenotype that we find with partial eff knockdown, a caveat of full eff/UBE2D knockout is that this could be lethal.

- The analysis of the E3 enzymes does not add anything to this manuscript. 

The analysis of E3 enzymes relates to our recent publication (Hunt et al. 2023; PMID: 37963875) that reports the physical interactions between E2 and E3 enzymes. Analysis of these E2-E3 pairs in the genetic screen in Fig.1 therefore follows this IP-MS study to provide insight into the functional interaction between these E2-E3 pairs in proteostasis.

- Figure 2B: the fluorescence intensities in images 2 and 4 are rather similar, yet the quantification shows significant differences. 

Please note that some of the GFP fluorescence in image 4 is not punctate, but rather diffuse fluorescence that is not related to HTT-GFP aggregates. Our image quantitation methods utilized thresholding to identify GFP-positive puncta while eliminating background fluorescence not corresponding to HTT-GFP puncta.

- The proteomic analyses could provide insights into the functional spectrum of UBE2D or even the identification of substrates. Yet apart from a DAVID analysis, none of the hits were followed up. In addition, only a few hits were labelled in the volcano plots (Figure 5). On what basis did the authors select those?

Please see the previous answer above regarding the identification of eff/UBE2D protein substrates from our proteomic analysis in Fig. 5. Only some of the top-regulated hits could be labeled in Fig.5 to avoid overcrowding.

- The manuscript remains at this stage rather descriptive. 

Our study has demonstrated a key role for the eff/UBE2D ubiquitin-conjugating enzyme in regulating protein quality control during aging in the Drosophila retina and skeletal muscle. Our study has identified key proteins that are modulated by eff/UBE2D RNAi in Drosophila muscle, that are rescued by expression of human UBE2D2, and that may underlie the accelerated decline in proteostasis that occurs upon eff/UBE2D RNAi. While more could be known about the regulation of these eff/UBE2D-modulated proteins in Drosophila, we have previously demonstrated that some of the proteins that are upregulated by UBE2DRNAi in human cells (e.g. some peroxins) are indeed direct ubiquitination targets of UBE2D via associated E3 ubiquitin ligases (Hunt et al. 2023; PMID: 37963875).

Reviewer #3 (Public Review):  

This is a potentially quite interesting paper that defines E2 and E3 genes in Drosophila that can impact the accumulation of the Q72-GFP protein in the fly eye. The authors then focus on the eff gene, showing which human homolog can rescue fly knockdown. They extend to skeletal muscle, from the hL protein, to show that eff by TMT mass spec decreases with age normally in the fly muscle and that there is a significant overlap of proteins that are disrupted with eff knockdown in young animals in muscle vs aged animals normally in muscle. 

Overall these data suggest eff decrease with age may contribute to the increase in ubiquitinated proteins in muscle with age, and that upregulation of eff activity might be of interest to extending lifespan. Because eff function can be performed by a human homologue, the findings may also apply to human situations of aging. 

These data are overall interesting and are of relevance for those interested in neurodegenerative disease and aging, although a number of points from the figures seem confusing and need more explanation or clarity. 

Thank you for reviewing our manuscript, we have improved the explanations and clarity of the manuscript.

Recommendations for the authors:

We would like to keep the manuscript title as it is currently to report the partial overlap in the proteomic changes induced by aging and effRNAi (Fig. 6).

Reviewer #1 (Recommendations For The Authors): 

(1) A significant concern arises from the unexpected outcome observed in the UBE2D/eff loss-of-function experiments. Despite its role as a ubiquitin-conjugating enzyme (E2), the reduction in UBE2D/eff levels paradoxically increased polyubiquitinated proteins and p62 accumulation, presenting a more intricate and seemingly unrelated phenotype to its anticipated function. 

eff/UBE2D represents one out of 21 different Drosophila E2 ubiquitin-conjugating enzymes and therefore eff RNAi alone is unlikely to reduce the total pool of ubiquitinated proteins. The generalized increase in insoluble polyubiquitinated proteins results from an overall derangement of protein quality control caused by effRNAi. In agreement with this scenario, the protein categories that were found to be modulated by effRNAi (Fig. 5) include proteins associated with protein quality control such as proteasome components and chaperones. Therefore, derangement in the levels of a wide range of regulators of proteostasis may lead to a generalized loss of protein quality control upon effRNAi.

I believe elucidating the mechanisms underlying the impact of UBE2D/eff deficiency on the observed phenotypes would contribute to a more comprehensive understanding of the study's implications. For instance, investigating whether the loss of UBE2D/eff influences muscle proteostasis by impeding proteasome assembly or function, modulating autophagy, etc. 

We have previously utilized luciferase assays to measure the proteolytic activity of the proteasome in human cells treated with siRNAs targeting UBE2D1/2/3/4 but found no effect of UBE2D knockdown compared to control nontargeting siRNAs (Hunt et al. 2023; PMID: 37963875). In Drosophila muscles, we have examined the levels of GFP-CL1 (a GFP fused with a proteasomal degron) and found that effRNAi does not impact GFP-CL1 levels (data shown in author response image 1). Overall, these results suggest that effRNAi reduces protein quality control without affecting proteasome activity.

Author response image 1.

(2) Related to Figures 1B-C: It is not clear to this reviewer the quantification methodology used in the experiment. Does each point represent the Average +/- SD for each replicate? If so, it appears that not all cases align with the n=5 as indicated in the figure legend. Additionally, how many animals per replicate were quantified? 

We have amended the legend of Figure 1 to indicate that each data point in the graph represents a single RNAi line targeting the corresponding gene. The mean of 5 biological replicates is shown for each RNAi line, with each biological replicate representing a single eye imaged from a distinct fly. Therefore, the data points that do not show large magnitude changes may indicate RNAi that were not effective at knocking down the target protein (or with no effect on HTT aggregates).  

(3) Related to the previous point: The analysis of pathogenic Huntingtin aggregation in the Materials and Methods section lacks information regarding the number of individuals, replicates, etc. 

Please see the response above.

(4) Related to Figure 1 B: In the case of eff/UBE2D, it appears that 3 out of 9 replicates demonstrate a significant increase in HL-polyQ aggregates. Considering the strength of this result, it raises questions about whether it justifies using eff for future analyses. 

Please see the response to point (2) above. These results indicate that 3 distinct UAS-RNAi lines targeting eff/UBE2D produced the same effect whereas 6 other effRNAi lines did not, possibly because they are less efficacious in knocking down eff/UBE2D. We have now amended the legend of Fig. 1B to better explain these results.

(5) Related to Figure 1 D-E: Could the authors provide clarification regarding the tissue type and animal age utilized in these experiments? 

Whole flies were utilized at 1 week of age.

(6) Related to Figure 3: Incorporating the normal accumulation of poly-ubiquitinated proteins during aging could provide context to better interpret the effect of eff/UBE2D KD at 3 weeks of age. 

Several papers from us and others have previously demonstrated a progressive increase in the insoluble levels of poly-ubiquitinated proteins during aging in Drosophila skeletal muscle (PMID: 36640359; PMID: 31249065; PMID: 33773104; PMID: 33658508; PMID: 24092876; PMID: 21111239; PMID: 24244197; PMID: 25199830; PMID: 28878259; PMID: 36213625). Our analyses now indicate that such age-related loss of protein quality control is accelerated by eff/UBE2D knockdown.

(7) Related to Figure 3: Would it be possible for the authors to include a list or table detailing the specific E2, deubiquitinating enzymes, and E3s identified in the comparative analysis of the old vs young proteome? This would provide a clear reference for the identified regulatory proteins involved in the age-related proteomic changes. 

We have added a tab to Supplementary Table 2 to report the list of age-regulated deubiquitinating enzymes (DUBs) and E1, E2, and E3 enzymes.

(8) Related to Figures 3 and 4: Given that the comparative analysis of the old versus young proteome identified 10 out of 21 E2 ubiquitin-conjugating enzymes, exploring the impact of eff/UBE2D overexpression becomes pivotal to understanding its role in age-related changes in proteostasis and lifespan. Conducting an experiment involving eff overexpression could provide valuable insights into whether restoring eff levels mitigates aging-related phenotypes. 

Although we have not done this experiment with eff overexpression, Fig. 4E reports that the overexpression of human UBE2D2 in skeletal muscle does not appear to influence lifespan by itself (green line in Fig. 4E), although it can partially rescue the short lifespan of flies with muscle-specific effRNAi (purple line in Fig. 4E).

(9) Providing a more detailed description of the Supplementary Tables would significantly enhance the reader's comprehension of their content. 

A description has been added at the end of the methods.

Reviewer #2 (Recommendations For The Authors): 

In addition, to the points listed above: 

- The title does not reflect the content of the manuscript and should be changed. There is no evidence that UBE2D maintains a "youthful" (needs to be changed as well) proteome. Rather, its expression declines with aging and its depletion leads to an increase of ubiquitinated proteins. This is true for essentially the entire proteostasis network. 

While proteostasis generally declines with aging, it is incompletely understood what specific components of the proteostasis network are dysregulated with aging. Our study now identifies the E2 ubiquitin-conjugating enzyme eff/UBE2D as a key regulator of proteostasis that is transcriptionally downregulated with aging. Comparison of the proteomic changes induced by aging versus those induced by effRNAi in young age indicates a partial overlap (Fig. 6), indicating that eff/UBE2D is, at least in part, necessary to maintain the proteome composition that is found in young age (“youthful”). On this basis, we would like to keep the current title but have amended the manuscript to indicate that such regulation of the proteome composition is only in part dependent on eff/UBE2D.

- Molecular weight markers are missing for the gels/western blot depicted in Fig 1E, 2C, 3E, and 4A. 

Thank you for pointing this out, these have been added.

- Fig. 4A, the Ponceau staining for the detergent insoluble samples shows almost no signal for lane 7 and the data should hence not be analyzed. 

The western blot membrane in Fig. 4A shows a reliable signal in all lanes (including lane 7) when probed with antibodies for ubiquitin, Ref(2)P, and tubulin. Therefore, there is no reason for excluding lane 7 from the analysis. Ponceau S staining is provided as an additional loading control but was not used to normalize the data.

Reviewer #3 (Recommendations For The Authors): 

There are a number of confusing or not sufficiently explained points in the figures that require clarity. 

In Figure 1, panels B and C, one assumes the gray broad line across means no difference from control. For the genes, many have points that are scattered both above and below that control line. What do the dots and range represent for each gene, and why are the data so scattered. How do the authors explain data ranging from no effect, to a negative effect to a positive effect, all for the same gene? Akt1 and Hsp83 are controls but are not quantitated to appreciate how variable the assay is. Can they explain the figure better, and also why the data for any one gene are so variable?

We have amended the legend of Figure 1 to indicate that each data point in the graph represents a single RNAi line targeting the corresponding gene. The mean of 5 biological replicates is shown for each RNAi line, with each biological replicate representing a single eye imaged from a distinct fly. Therefore, the data points that do not show large magnitude changes may indicate RNAi lines that were not effective at knocking down the target protein (or that did not affect HTT aggregates). Therefore, the variability in the analysis of a single gene arises because different RNAi lines targeting that gene may have different efficacy. RNAi lines for Akt1 and Hsp83 are merely used as controls (these have been quantified in Jiao et al. 2023; PMID: 36640359).

In Figure 2A, it is not clear which animals have the hL-Q72-GFP (which eyes are "rough eyes"?). Also, do ubc6-RNAi and eff-RNAi have an impact on the normal eye? That is, can they explain the images and genotypes more clearly. 

UBC6 and eff RNAi produce these rough eye phenotypes in the absence of HTT-polyQ and these are rescued by the expression of their human homologs. The panel images indicated in bold here below are those that have “rough eye” phenotypes: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 (a green R has been added to these panels in Fig. 2A).

In Figure 2B, panel 3 looks very different from 1 and 4 and yet is not different from them by quantitation. Can they replace it with a more representative panel or is 3 lower (but not significantly so)? 

Please note that some of the GFP fluorescence in image 4 is not punctate, but rather diffuse fluorescence that is not related to HTT-GFP aggregates. Our image quantitation methods utilized thresholding to identify GFP-positive puncta while eliminating background fluorescence not corresponding to HTT-GFP puncta.

In Figures 3E and F, it would be helpful in F to put the detergent soluble bar graphs all on the left so that those data are on the left in both E and F, and then detergent-insoluble in E and F to the right. This would make the figure and quantitation easier to follow. 

Done.

The same point as above for Figures 4 A and B. 

Done.

In Figure 3A, CG7656 is nearly as reduced with age as eff. One wonders if that gene would give a different or similarly overlapping proteome with age as eff. Was CG7656 not focused on because not conserved? 

As indicated in Figure 1B, CG7656 is orthologous to UBE2R1 (also called CDC34) and UBE2R2 in humans. In this screen, however, RNAi targeting CG7656 did not appear to influence HTT aggregates and therefore was not selected for further analyses. However, it may play a role in skeletal muscle proteostasis during aging.

In Figure 6, the R2 value correlating age with eff-RNAi is weak. Although they discuss this in the text, it might also be helpful to include Venn diagrams for gene overlaps and the significance to make the argument more clear that there is a significant correlation in proteins up and down to indicate that eff largely recapitulates the changes of aging. Correlating this with proteins that are restored with UBE2D in muscle in a more clear manner may also be helpful for readers interested in aging. 

We have amended the text to indicate that this relatively low correlation (R2\=~0.2, but corresponding to a consistent regulation of 70% of proteins by aging and effRNAi) could indicate that eff/UBE2D is only in part responsible for maintaining a youthful composition of the muscle proteome during aging. Other changes that occur with aging likely account for non-correlated alterations in protein levels. We have also added Venn diagrams (Fig. 6E) to further display the overlap in protein regulation by aging vs. effRNAi.

In Figure 7, they might indicate that the accumulated insoluble protein is ubiquitinated. That is left out of the figure, although indicated in the legend. 

Done.

Author response:

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

Our revised version of the manuscript addresses all the comments and suggestions raised, as clarified in our point-by-point answer to the reviewers. We have performed additional experiments regarding the effects on proliferation and differentiation of additional cell types in the muscle, such as myogenic and mesenchymal progenitors as well as chondrogenesis in parental hMSCs that did not express exogenous ACVR1. Moreover, as suggested by reviewer #2, we performed all the chondrogenic experiments with addition of TGFβ in the differentiation media and analyzed chondrogenesis by both Alcian blue staining and qPCR analysis of gene markers (Sox9, Acan, Col2a1 and Mmp3). We also extended our RNA-seq analysis and included new data using both hMSCs expression wild type or R206H ACVR1 receptor, with or without different ACVR1 ligands (BMP6 and Activin A) and treated or not with the inhibitor BYL719. The new data suggests that BYL719 is able to inhibit the expression of genes involved in ossification and osteoblast differentiation irrespective of the presence of the mutation. We also discuss the effect of BYL719 in mTOR signaling and addressed all the minor comments suggested by both reviewers.

We addressed the specific comments of the reviewers as follows:

Reviewer # 1:

Specific points:

Point #1 and #2. The authors showed that BYL719 inhibited HO in FOP model mice. Did they have HO not only in the muscle but also in the bone marrow? The progenitor cells of chondrocytes and osteoblasts may differ between the muscle and bone marrow. The authors should examine the effects of BYL719 on some other types of cells in the muscle, such as myoblasts and fibro-adipogenic cells, in addition to the bone marrow-derived MSCs. Furthermore, it was unclear whether they were human or murine MSCs in the text.

The inhibitory effect of BYL719 on HO in FOP mice was clear, but the molecular mechanisms or target cells were still unclear because BYL719 affected multiple types of cells and molecules. The authors are encouraged to show clearer mechanisms and target cells' critical inhibition of HO. Again, this reviewer believes that in vivo and in vitro experiments using muscle and bone marrow and cells prepared from them should provide additional critical information.

As detailed in the introduction, it is known that Heterotopic Ossification develops in the skeletal muscle and connective tissues. Consistent with the current knowledge of the field, none of the mice showed HO in the bone marrow. Additionally, since activation of the mutant allele is achieved by injection of CRE-expressing adenovirus and cardiotoxin in the muscle hindlimb, it is unlikely that mesenchymal progenitors in the bone marrow would be strongly affected. Interestingly, single-cell RNA sequencing from multiple mouse tissues identified a very strong transcriptional similarity between FAPs and non-muscle mesenchymal progenitors (PMID: 37599828). As suggested, we examined the effects of BYL719 in proliferation and differentiation in additional cell types such as muscle progenitors. In this new version of the manuscript, we show that BYL719 reduces the proliferation of muscle and mesenchymal progenitors while it blocks myoblast differentiation in vitro (Figure 7, Figure Supplement 1). MSCs were murine on those experiments shown in Figure 3; whereas assays shown in Figures 5 and 6 were of human origin. We have further clarified this in the respective Figure legends.

All the data generated strongly suggests that there is not a single mechanism supporting all the effects of BYL719 in HO. Instead, BYL719 affects multiple cell types involved in efficient HO (e.g. reduction in proliferation and osteochondrogenic specification of mesenchymal precursors (MPs), reduction on proliferation, migration, and inflammatory gene expression on monocytes, etc.). Interestingly, our data suggests that BYL719 is able to inhibit these effects on MPs and monocytes irrespective of the presence of the ACVR1-R206H mutation (Figures 5, 6 and 7). Additionally, there are several signaling mechanisms affected. BYL719 reduces SMAD1/5, p38, AKT and mTOR signaling in parental MPs or with mutations in ACVR1 (Figure 3 and our previous publication PMID: 31373426), being all these pathways required for efficient osteochondrogenic specification of MPs. We consider that the different detailed mechanisms by which BYL719 inhibits osteochondrogenic specification enhances the robustness of the findings in this study.

Point #3. In FOP model mice, ACVR1 was mutated as Q207D. However, R206H was used in in vitro experiments. Do they have the same characteristics? This reviewer would like to recommend examining the effect of BYL719 on wild-type ACVR1, R206H, and Q207D simultaneously in each experiment.

We already performed these experiments, assaying in parallel ACVR1-WT, ACVR1-Q207D and ACVR1-R206H, in the transcriptional responses of MPs in our previous work (PMID: 31373426). Both mutations had similar responses, being ACVR1-Q207D stronger than ACVR1-R206H, as it has been shown in vivo in mouse models of HO (PMID: 34633114). In any case, BYL719 inhibits these transcriptional responses induced by both mutant alleles.

Point #4. Figure 5: What was the effect of BYL719 on the differentiation of parental cells that did not express exogenous ACVR1?

We performed new assays of chondrogenic differentiation of hMSCs that are shown in the new Figure 5. BYL719 inhibits chondrogenic differentiation of parental hMSCs and also inhibits chondrogenic specification irrespective of the expression of either wild type or mutant ACVR1.

Point #5. Figure 6: In this experiment, gene expression was examined in pretreated MSCs-ALK2 (ACVR1?) R206H with and without BYL719. It was clear whether suppression of gene expression by BYL719 was specifically caused in cells expressing R206H. What were the effects of BYL719 on parental cells that did not express exogenous ACVR1?

To be consistent, we relabeled ALK2 to ACVR1 in the figure. We expanded the conditions analyzed in the RNA-sequencing. We included conditions where we activate ACVR1 (either WT or R206H) with their known physiological ligand BMP6. In both, human MSCs expressing ACVR1-R206H and human MSCs expressing Wild Type ACVR1, we observed a downregulation of differentially expressed genes upon addition of BYL719, irrespective of ligand (BMP6 or Activin A) or receptor (RH or WT) (added new Figure 6: B and C).

Point #6. Figure 7: BYL719 suppressed cell proliferation of all cells examined partially at 2 uM and almost completely at 10 uM, respectively. There is a possibility that BYL719 inhibits HO by inhibiting osteochondroprogenitor proliferation. The authors are encouraged to show data on the effect of BYL719 on the proliferation of other types of cells, such as myoblasts, fibro-adipogenic cells, or bone marrow cells.

We examined the effects of BYL719 in proliferation in additional cell types such as muscle and mesenchymal progenitors. BYL719 slightly reduced the proliferation of myoblasts and mesenchymal cells in vitro (Figure 7, Figure Supplement 1). However, the reduction in the proliferation in myoblasts or MPs did not reach the extent to that observed in monocytes or macrophages (Figure 7).

Point #7. Figure 8: How was the effect of BYL719 on muscle regeneration in wild-type? It was reported that mTOR signaling is important in HO in FOP. The authors are encouraged to show the effect of BYL719 on mTOR signaling.

Muscle regeneration in wild-type mice has also been shown in our previous results PMID: 31373426. In addition, we included images of the muscle regeneration after 23 days of treatment with BYL719 in mice ACVR1Q207D with or without PI3Kα deletion after induction of HO in the new Figure 2, Figure Supplement 2. These mice showed full muscle regeneration or small calcifications surrounded by muscle at most. The effects of PI3Kα inhibitors, either BYL719 or A66, on mTOR signaling had been previously shown by our group (PMID: 31373426). Both inhibitors strongly reduced signaling of mTOR, visualized by activation of p70 S6-kinase, a surrogate marker of mTOR activity.

Minor points:

(9) SMAD 1/5 should be SMAD1/5.

(10) The source of human MSCs should be indicated in the text.

(11) ALK2 should be ACVR1 in Figure 6A.

(12) The protein levels of each receptor should be examined in Fig. 4.

We introduced the suggested changes in the manuscript and Figure 6 and indicated the source of human MSCs in Materials and Methods. We also examined the levels of each receptor that are shown in the new Figure 4, Figure Supplement 1.

Reviewer # 2:

Specific points:

Point #1. Because the involvement of PI3K in HO of FOP, was already reported by authors' group and also others (Hino et al, Clin Invest, 2017), the main purpose of this study was to disclose the mechanism of how PI3K was activated in FOP cells. In the published study (Hino et al, Clin Invest, 2017), PI3K was activated by the ENPP2-LPA-LPR cascade. Unfortunately, there were no new data for this important issue.

The main purpose of this study is to demonstrate that the pharmacological and genetic inhibition of PI3Kα in HO progenitors at injury sites reduces HO in vivo, to extend the insights into the molecular and cellular mechanisms responsible for the therapeutic effect of PI3K inhibition, and to optimize the timing of the administration of BYL719. Class I PI3Ks are heterodimers of a p110 catalytic subunit in complex with a regulatory subunit. They engage in signaling downstream of tyrosine kinases, G protein-coupled receptors and monomeric small GTPases. Therefore, a plethora of growth factors, cytokines, inflammatory agents, hormones and additional external and internal stimuli are able to activate PI3Kα (PMID: 31110302). In fact, TGF-β family members, including activin A, are able to activate PI3K and mediate some of their non-canonical responses (PMID: 19114990). Multiple factors with known increased expression in the ossifying niche in HO and FOP (e.g. activin A, TGF-β, inflammatory agents such as TNFα, IL6, IL3, etc.) are known activators of PI3K (PMID: 30429363). Interestingly, in our RNA-seq analysis in hMSCs we did not observe increased expression levels of Enpp2 when comparing wild type and R206H mutated cells treated with activin A.

Point #2. The HO formation of ACVR1/Q207D model mice in this study is extremely unstable (Figure 1B, DMSO). Even the bone volume of some red symbols, which indicate the presence of HO, is located on the base (0.00) line. I would examine carefully the credibility of the data. Also, it is well known that the molecular behavior of mice Acvr1/Q207D and human ACVR1/R206H was different.

We agree with the reviewer that induction of HO is variable between mice showing variations in penetrance and intensity of the ossifying lesions. This variability is a known common trend that appears in all the models of HO published so far (e.g. PMID: 28758906, PMID: 26333933). Accordingly, we did not exclude any animal that has been injected with CRE-expressing adenovirus plus cardiotoxin in the μCT analysis. Regarding the behavior of mice Acvr1/Q207D and human ACVR1/R206H, it is well known that Q207D produces more robust and stronger responses in terms of signaling and formation of heterotopic ossification (PMID: 34633114). Therefore, reduction of HO by BYL719 would be more stringent in the Acvr1/Q207D model.

Point #3. The experimental design of Figure 5 experiments is confusing. Although the authors mentioned that the data in Figure 5A were taken seven days after chondrogenic induction, I am skeptical whether the chondrogenic induction was successful. Based on the description of Material and Methods, the authors did not include TGFβ in their "Differentiation Medium", which is an essential growth factor to induce chondrogenic differentiation of human MSC. Why did the ALP activity increase after chondrogenic induction? The authors should demonstrate the evidence of successful chondrogenic induction by showing the expression of key chondrogenic genes such as SOX9, ACAN, or COL2A1. The data in Figure 5B-E are also confusing. The addition of Activin A showed no difference between ACVR1/WT and ACVR1/R206H cells, suggesting that these cells did not reproduce the situation of FOP.

We performed new assays of chondrogenic differentiation of hMSCs that are shown in the new Figure 5. We included TGFβ1 in the differentiation medium and also included the parental cell line in the analysis. In addition of being a marker of osteoblast differentiation, alkaline phosphatase (ALPL) has also been shown to be induced during chondroblast differentiation in vitro (PMID: 19855136; PMID: 9457080; PMID: 18377198; PMID: 23388029). Moreover, expression data of SOX9, COL2A1, ACAN and MMP13 of cells after chondrogenic differentiation is included in the new Figure 5. Expression of some markers (e.g. ACAN) are increased by the expression of ACVR1R206H, however, we did not observe significant differences in chondroblast differentiation gene expression between ACVR1wt and ACVR1R206H expressing cells. In any case, BYL719 could inhibit chondrogenic differentiation of parental hMSCs and also the chondrogenic specification irrespective of the expression of either wild type or mutant ACVR1.

Point #4. The experimental design and data analyses of RNA-seq were inappropriate and insufficient, which is disappointing for the reviewer because this will be a key experiment in this study. Because the most important point is to identify the signal for PI3Kα induced by Activin A via ACVR1/R206H, they should also use hMSC-ACVR1/WT for this experiment. Because the authors clearly demonstrated that TGFBR were not targets of BYL719, they should compare the expression profiles between MSC-ACVR1/WT and MSC-ACVR1/WT with BYL719 to identify the targets of BYL719 unrelated to Activin A signal. Then the expression profiles of ACVR1/R206H cells treated with Activin A and Activin A plus BYL719 were compared. Among down-regulated signals by BYL719, those found also in MSC-ACVR1/WT should be discarded. It is important to investigate whether the GO term of ossification or osteoblast differentiation is found also in MSC-ACVR1/WT. If it is so, the effect of BYL719 is not specific for FOP cells.

We extended our RNA sequencing analysis with additional experimental conditions and comparisons. In new Figure 6, we now compare hMSCs expressing wild type or R206H receptors, with or without BYL719 inhibition, and with or without different ligand activations (BMP6 or Activin A) (New Figure 6A). New Figure 6B shows the Gene ontology analysis of the differentially expressed genes between cells expressing WT and RH receptors under control conditions. We can observe that ossification (GO:0001503) and osteoblast differentiation (GO:0001649) were detected within the top 10 significantly differentially regulated biological processes between these conditions. Therefore, we analyzed these relevant identified GO terms in 5 different comparisons upon GO enrichment analysis (Figure 6C). In addition to the comparison between cells expressing WT and RH receptors under control conditions explained above, we also compared cells expressing WT or RH receptor, with different ACVR1 ligands (BMP6 and Activin A), and with or without BYL719 inhibitor. The addition of BYL719 resulted in a downregulation of the GO terms “ossification” and “osteoblast differentiation” (new Figure 6C). These results confirm the inhibitory effect of BYL719 on ossification and osteoblast differentiation biological processes, and inform that this inhibitory effect remains consistent upon BMP6 or Activin A ligand activation, and with ACVR1 WT and RH expression.

Point #5. The data in Figure 7 were not related to the aim of this study because cell lines used in these experiments did not have ACVR1/R206H mutations. It is not appropriate to extrapolate these data in the FOP situation.

We utilized immune cell lines where we could activate ACVR1 with their known physiological ligand BMP6. Mutated ACVR1 gains response to activin A in addition to maintaining the physiological response to BMP6 as the wild type form. Therefore, in these assays we interrogated in vitro, with addition of BMP6, the effects of BYL719 in the growth, migration and inflammatory gene expression upon conditions of activated ACVR1 receptor downstream signaling. We consider that understanding the effects of PI3Kα inhibition in the regulation of proliferation, migration and inflammatory cytokine expression in monocytes, macrophages and mast cells is essential to better define the potential outcome of BYL719 treatment for heterotopic ossifications.

Minor comments:

(1) The legends for Figure 1C were those for Figure 1D, and there were no descriptions for Figure 1C in the legends and methods section. The reviewer was unable to understand the meaning of BV/TV. What is TV?

(2) “However, in PI3Kα deficient mice ACVR1Q207D expression only led to minor ectopic calcifications that were already surrounded by fully regenerated muscle tissue on the 23rd day after injury (Figure 2D, Figure 2-Figure Supplement 1B)": There were no histological data either Figure 2D, Figure 2-Figure Supplement 1B), which showed muscle tissues.

(3) "The overexpression of Acvr1R206H increased basal and activin dependent expression of canonical (Id1 and Sp7) and non-canonical (Ptgs2) BMP target genes (Figure 3C),": There was no increase of Ptgs2 gene in basal level.

(4) Materials and Methods. Production of human fetal mesenchymal stem cells expressing ACVR1.: Is it derived from a fetus?

(5) Figure 6C: There was no description of the meaning of each column. What does AA mean and what is the number?

We introduced the missing information in the manuscript, Figure legends and material and methods section for points #1, 4 and 5. AA was Activin A, the number was the number of replicates. This has been detailed in the figure legend. We included images of the muscle regeneration after 23 days of treatment with BYL719 in mice after induction of HO in the new Figure 2, Figure Supplement 2 (point #2). We corrected the mistake in the manuscript refraining for suggesting increase of Ptgs2 gene expression by ACVR1-R206 at the basal level (Point #3).

Author response:

Reviewer #1 (Public Review):

Weaknesses:

There are some minor weaknesses.

Notably, there are not a lot of new insights coming from this paper. The structural comparisons between MCC and PCC have already been described in the literature and there were not a lot of significant changes (outside of the exo- to endo- transition) in the presence vs. absence of substrate analogues.

We agree that the structures of the human MCC and PCC holoenzymes are similar to their bacterial homologs. That is due to the conserved sequences and functions of MCC and PCC across different species.

There is not a great deal of depth of analysis in the discussion. For example, no new insights were gained with respect to the factors contributing to substrate selectivity (the factors contributing to selectivity for propionyl-CoA vs. acetyl-CoA in PCC). The authors state that the longer acyl group in propionyl-CoA may mediate stronger hydrophobic interactions that stabilize the alpha carbon of the acyl group at the proper position. This is not a particularly deep analysis and doesn't really require a cryo-EM structure to invoke. The authors did not take the opportunity to describe the specific interactions that may be responsible for the stronger hydrophobic interaction nor do they offer any plausible explanation for how these might account for an astounding difference in the selectivity for propionyl-CoA vs. acetyl-CoA. This suggests, perhaps, that these structures do not yet fully capture the proper conformational states.

We appreciate this comment. Unfortunately, in the cryo-EM maps of the PCC holoenzymes, the acyl groups were not resolved (fig. S6), so we were unable to analyze the specific interactions between the acyl-CoAs and PCC. We will discuss this limitation in our revised manuscript.

The authors also need to be careful with their over-interpretation of structure to invoke mechanisms of conformational change. A snapshot of the starting state (apo) and final state (ligand-bound) is insufficient to conclude *how* the enzyme transitioned between conformational states. I am constantly frustrated by structural reports in the biotin-dependent enzymes that invoke "induced conformational changes" with absolutely no experimental evidence to support such statements. Conformational changes that accompany ligand binding may occur through an induced conformational change or through conformational selection and structural snapshots of the starting point and the end point cannot offer any valid insight into which of these mechanisms is at play.

Point accepted. We will revise our manuscript to use "conformational differences" instead of "conformational changes" to describe the differences between the apo and ligand-bound states.

Reviewer #2 (Public Review):

Comments and questions to the manuscripts:

I'm quite impressed with the protein purification and structure determination, but I think some functional characterization of the purified proteins should be included in the manuscript. The activity of enzymes should be the foundation of all structures and other speculations based on structures.

We appreciate this comment. However, since we purified the endogenous BDCs and the sample we obtained was a mixture of four BDCs, the enzymatic activity of this mixture cannot accurately reflect the catalytic activity of PCC or MCC holoenzyme. We will acknowledge this limitation in the discussion section of our revised manuscript.

In Figure 1B, the structure of MCC is shown as two layers of beta units and two layers of alpha units, while there is only one layer of alpha units resolved in the density maps. I suggest the authors show the structures resolved based on the density maps and show the complete structure with the docked layer in the supplementary figure.

We appreciate this comment. We have shown the cryo-EM maps of the PCC and MCC holoenzymes in fig. S8 to indicate the unresolved regions in these structures. The BC domains in one layer of MCCα in the MCC-apo structure were not resolved. However, we think it would be better to show a complete structure in Fig. 1 to provide an overall view of the MCC holoenzyme. We will revise Fig. 1B and the figure legend to clearly point out which domains were not resolved in the cryo-EM map and were built in the structure through docking.

In the introduction, I suggest the author provide more information about the previous studies about the structure and reaction mechanisms of BDCs, what is the knowledge gap, and what problem you will resolve with a higher resolution structure. For example, you mentioned in line 52 that G437 and A438 are catalytic residues, are these residues reported as catalytic residues or this is based on your structures? Has the catalytic mechanism been reported before? Has the role of biotin in catalytic reactions revealed in previous studies?

Point accepted. It was reported that G419 and A420 in S. coelicolor PCC, corresponding to G437 and A438 in human PCC, were the catalytic residues (PMID: 15518551). The same study also reported the catalytic mechanism of the carboxyl transfer reaction. The role of biotin in the BDC-catalyzed carboxylation reactions has been extensively studied (PMIDs: 22869039, 28683917). We will include these information in the introduction section of our revised manuscript.

In the discussion, the authors indicate that the movement of biotin could be related to the recognition of acyl-CoA in BDCs, however, they didn't observe a change in the propionyl-CoA bound MCC structure, which is contradictory to their speculation. What could be the explanation for the exception in the MCC structure?

We appreciate this comment. We do not have a good explanation for why we did not observe a change in the propionyl-CoA bound MCC structure. It is noteworthy that neither acetyl-CoA nor propionyl-CoA is the natural substrate of MCC. Recently, a cryo-EM structure of the human MCC holoenzyme in complex with its natural substrate, 3-methylcrotonyl-CoA, has been resolved (PDB code: 8J4Z). In this structure, the binding site of biotin and the conformation of the CT domain closely resemble that in our acetyl-CoA-bound MCC structure. Therefore, the movement of biotin induced by acetyl-CoA binding mimics that induced by the binding of MCC's natural substrate, 3-methylcrotonyl-CoA, indicating that in comparison with propionylCoA, acetyl-CoA is closer to 3-methylcrotonyl-CoA regarding its ability to bind to MCC. We will discuss this possibility in our revised manuscript.

In the discussion, the authors indicate that the selectivity of PCC to different acyl-CoA is determined by the recognition of the acyl chain. However, there are no figures or descriptions about the recognition of the acyl chain by PCC and MCC. It will be more informative if they can show more details about substrate recognition in Figures 3 and 4.

We appreciate this comment. Unfortunately, in the cryo-EM maps of the PCC holoenzymes, the acyl groups were not resolved (fig. S6), so we were unable to analyze the specific interactions between the acyl-CoAs and PCC. We will discuss this limitation in our revised manuscript.

How are the solved structures compared with the latest Alphafold3 prediction?

Since AlphaFold3 was not released when our manuscript was submitted, we did not compare the solved structures with the AlphaFold3 predictions. We have now carried out the predictions using Alphafold3. Due to the token limitation of the AlphaFold3 server, we can only include two α and six β subunits of human PCC or MCC in the prediction. The overall assembly patterns of the Alphafold3-predicted structures are similar to that of the cryo-EM structures. The RMSDs between PCCα, PCCβ, MCCα, and MCCβ in the apo cryo-EM structures and those in the AlphaFold3-predicted structures are 7.490 Å, 0.857 Å, 7.869 Å, and 1.845 Å, respectively. The PCCα and MCCα subunits adopt an open conformation in the cryo-EM structures but adopt a closed conformation in the AlphaFold-3 predicted structures, resulting in large RMSDs.

Author response:

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

eLife assessment

This study presents an important contribution to cardiac arrhythmia research by demonstrating long noncoding RNA Dachshund homolog 1 (lncDACH1) tunes sodium channel functional expression and affects cardiac action potential conduction and rhythms. The evidence supporting the major claims are solid. The work will be of broad interest to cell biologists and cardiac electrophysiologists.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In this study, the authors show that a long-non coding RNA lncDACH1 inhibits sodium currents in cardiomyocytes by binding to and altering the localization of dystrophin. The authors use a number of methodologies to demonstrate that lncDACH1 binds to dystrophin and disrupt its localization to the membrane, which in turn downregulates NaV1.5 currents. Knockdown of lncDACH1 upregulates NaV1.5 currents. Furthermore, in heart failure, lncDACH1 is shown to be upregulated which suggests that this mechanism may have pathophysiological relevance.

Strengths:

(1) This study presents a novel mechanism of Na channel regulation which may be pathophysiologically important.

(2) The experiments are comprehensive and systematically evaluate the physiological importance of lncDACH1.

Reviewer #2 (Public Review):

This manuscript by Xue et al. describes the effects of a long noncoding RNA, lncDACH1, on the localization of Nav channel expression, the magnitude of INa, and arrhythmia susceptibility in the mouse heart. Because lncDACH1 was previously reported to bind and disrupt membrane expression of dystrophin, which in turn is required for proper Nav1.5 localization, much of the findings are inferred through the lens of dystrophin alterations.

The results report that cardiomyocyte-specific transgenic overexpression of lncDACH1 reduces INa in isolated cardiomyocytes; measurements in whole heart show a corresponding reduction in conduction velocity and enhanced susceptibility to arrhythmia. The effect on INa was confirmed in isolated WT mouse cardiomyocytes infected with a lncDACH1 adenoviral construct. Importantly, reducing lncDACH1 expression via either a cardiomyocyte-specific knockout or using shRNA had the opposite effect: INa was increased in isolated cells, as was conduction velocity in heart. Experiments were also conducted with a fragment of lnDACH1 identified by its conservation with other mammalian species. Overexpression of this fragment resulted in reduced INa and greater proarrhythmic behavior. Alteration of expression was confirmed by qPCR.

The mechanism by which lnDACH1 exerts its effects on INa was explored by measuring protein levels from cell fractions and immunofluorescence localization in cells. In general, overexpression was reported to reduce Nav1.5 and dystrophin levels and knockout or knockdown increased them.

The strengths of this manuscript include convincing evidence of a link between lncDACH1 and Na channel function. The identification of a lncDACH1 segment conserved among mammalian species is compelling. The observation that lncDACH1 is increased in a heart failure model and provides a plausible hypothesis for disease mechanism.

One limitation of the fractionation approach is the uncertain disposition of Na channel protein deemed "cytoplasmic." It seems likely that the membrane fraction includes ER membrane. The signal may reasonably be attributed to Na channel protein in stalled transport vesicles, or alternatively in stress granules, but this was not directly addressed.

Reviewer #3 (Public Review):

Summary:

In this manuscript, the authors report the first evidence of Nav1.5 regulation by a long noncoding RNA, LncRNA-DACH1, and suggest its implication in the reduction in sodium current observed in heart failure. Since no direct interaction is observed between Nav1.5 and the LncRNA, they propose that the regulation is via dystrophin and targeting of Nav1.5 to the plasma membrane.

Strengths:

(1) First evidence of Nav1.5 regulation by a long noncoding RNA.

(2) Implication of LncRNA-DACH1 in heart failure and mechanisms of arrhythmias.

(3) Demonstration of LncRNA-DACH1 binding to dystrophin.

(4) Potential rescuing of dystrophin and Nav1.5 strategy.

Weaknesses:

(1) The fact that the total Nav1.5 protein is reduced by 50% which is similar to the reduction in the membrane reduction questions the main conclusion of the authors implicating dystrophin in the reduced Nav1.5 targeting. The reduction in membrane Nav1.5 could simply be due to the reduction in total protein.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

Weaknesses:

(1) What is indicated by the cytoplasmic level of NaV1.5, a transmembrane protein?

This is still confusing. Since Nav1.5 is an integral membrane protein, I am not sure what is really meant here by cytosolic fraction. From the workflow, it seems a separate organelle fraction is also collected. Is the amount of Nav1.5 in this fraction (which I assume includes for e.g. lysosome) also increased with lncDACH1? I recommend the authors to refer to the Nav channels not at the plasma membrane as 'intracellular' rather than 'cytoplasmic'.

Thanks for the insightful comment. We completely agree. Accordingly, we have changed “cytoplasmic” to “ intracellular“.

Line 226. "In consistent with the results" Perhaps unnecessary to have "in"

Thank you for the insightful comment. We have corrected it.

Line 228. Is it optimal or optical?

Sorry for the mistake, it should be optical. We have corrected it.

Reviewer #3 (Recommendations For The Authors):

I still have an issue with the total reduction in Nav1.5 which is about the same as the reduction in membrane and currents. The authors argue that there is an increase in cytoplasmic Nav1.5. However the controls that they provide for membrane and cytoplasmic fractions are not convincing.

Thank you for the insightful comment. We can not rule out the possibility that the reduction in membrane Nav1.5 maybe be due to the reduction in total protein. Our data indicates that the membrane and total protein levels of Nav1.5 were reduced by 50%. However, the intracellular Nav1.5 was not decreased, but increased in the hearts of lncDACH1-TG mice than WT controls, which indicates that the intracellular Nav1.5 failed to traffic to the membrane.

Author response:

First we thank the reviewers for a thorough reading of our paper and some useful comments. A recurrent remark of the reviewers concerns the appearance of kRas-expressing cells (labelled by a nuclear blue fluorescent marker) which we attribute to the progeny of the initially induced cell. The reviewers suggest that these cells may have been obtained through activation of the Cre-recombinase in other cells by cyclofen released from light scattering, via diffusion, leakiness, etc. These remarks are perfectly reasonable from people not familiar with the cyclofen uncaging approach that we are using but are unwarranted as we shall show below.

We have been using cyclofen uncaging with subsequent activation of a Cre-recombinase (or some other proteins) since 2010 (see ref.34, Sinha et al., Zebrafish 7, 199-204 (2010) and our 2018 review (ref.35, Zhang et al., ChemBioChem 19,1-8 (2018)). In our experiments, the embryos are incubated in the dark in 6M caged cyclofen (cCyc) and washed in E3 medium (or transferred to a new medium with no cCyc). In these conditions, over many years we never observed activation of the recombinase, i.e. the appearance of the associated fluorescent label in cells of embryos grown in E3 medium. Hence leakiness can be ruled out (in presence of cCyc or in its absence).

Following transfer of the embryos to new E3 medium we illuminate the embryos locally with light at 405nm. In these conditions, cCyc is only partially uncaged and results in activation of Cre-recombinase in only a few cells (1,2, 3, …) within the illuminated region only, namely in the appearance of the kRas-associated nuclear blue fluorescent label in usually one cell (and sometimes in a few more; data and statistics will be incorporated in a revised manuscript). In absence of any further treatment (e.g. activation of a reprogramming factor) these fluorescently labelled cells disappear within a few days (either via shut-down of their promotor, apoptosis or some other mechanism). The crucial point here is that we see less and not more kRas expressing cells (i.e. with nuclear blue fluorescence). This observation rules out activation of Cre-recombinase in other cells days after illumination due to leakiness, cyclofen released by light or diffusing from the illumination spot.

To observe many more fluorescent cells days after activation of the initial cell, one needs to transiently activate VentX-GR by overnight incubation in dexamethasone (DEX) (Injecting the embryos at 1-cell stage with VentX-GR or incubating them in DEX does not result in the appearance of more blue fluorescent cells). Following activation of VentX-GR, the fluorescent cells observed a couple of days after initiation are visualized in E3 medium (i.e. in absence of cyclofen) and are localized to the vicinity of the otic vesicle (the region where the initial cell was activated). In a revised manuscript we will present images of these fluorescent cells taken a few days apart from the same embryo in which a single cell was initially activated. Hence, we attribute these cells to the progeny of the activated cell. Obviously, single cell tracking via time-lapse microscopy would nail down this issue and provide fascinating insight into the initial stages of tumor growth. Unfortunately, immobilization of embryos in the usual medium (e.g. MS222, tricaine) over 5-6 days to track the division and motion of single cells is not possible. We are considering some other possibilities (immobilization in bungarotoxin or via photo-activation of anionic channels), but these challenging experiments are for a future paper.

Reviewer #1 (Public Review):

The authors then performed allotransplantations of allegedly single fluorescent TICs in recipient larvae and found a large number of fluorescent cells in distant locations, claiming that these cells have all originated from the single transplanted TIC and migrated away. The number of fluorescent cells showed in the recipient larve just after two days is not compatible with a normal cell cycle length and more likely represents the progeny of more than one transplanted cell.

As mentioned in the manuscript, we measure the density of cells/nl and inject in the yolk of 2dpf Nacre embryos a volume containing about 1 cell, following published protocols (S.Nicoli and M.Presta, Nat.Prot. 2,2918 (2007)). We further image the injected cell(s) by fluorescence microscopy immediately following injection, as shown in Fig.4A and Fig.S8B. We might miss a few cells but not many. With a typical cell cycle of ~10h the images of tumors in larvae at 3dpt (and not 2dpt as misunderstood by this reviewer) correspond to ~100 cells. In any case the purpose of this experiment was not to study tumorigenesis upon transplantation but to show that the progeny of the initially induced cells is capable of developing into a tumor in a naïve fish, which is the operational definition of cancer that we adopted here.

The ability to migrate from the injection site should be documented by time-lapse microscopy.

As stated above our purpose here is not to study tumor formation from transplanted cell(s) but to use that assay as an operational test of cancer. Besides as mentioned earlier single cell tracking in larvae over 3-4dpt is not a trivial task.

Then, the authors conclude that "By allowing for specific and reproducible single cell malignant transformation in vivo, their optogenetic approach opens the way for a quantitative study of the initial stages of cancer at the single cell level". However, the evidence for these claims are weak and further characterization should be performed to:

(1) show that they are actually activating the oncogene in a single cell (the magnification is too low and it is difficult to distinguish a single nucleus, labelling of the cell membrane may help to demonstrate that they are effectively activating the oncogene in, or transplanting, a single cell)

In a revised manuscript we will provide larger magnification of the initial induced cell and show examples of oncogene activation in more than one cell.

(2) the expression of the genes used as markers of tumorigenesis is performed in whole larvae, with only a few transformed cells in them. Changes should be confirmed in FACS sorted fluorescent cells

When the oncogene is activated in a whole larvae all cells are fluorescent and thus FACS is of no use for cell sorting. Sorting could be done in larvae where single cells are activated, but then the efficiency of FACS is not good enough to isolate the few fluorescent cells among the many more non-fluorescent ones. We agree that the change in expression of the genes used as markers of tumorigenesis is an underestimate of their true change, but our goal at this time is not to precisely measure the change in expression level, but to show that the pattern of change is different from the controls and corresponds to what is expected in tumorigenesis.

(3) the histology of the so called "tumor masses" is not showing malignant transformation, but at the most just hyperplasia.

The histology of the hyperplasic tissues displays cellular proliferation with a higher density of nuclear material which is characteristic of tumors, Fig.S4C. Besides the increased expression of pERK in these tissues, Fig.S4A,B is also a hallmark of cancer.

In the brain, the sections are not perfectly symmetrical and the increase of cellularity on one side of the optic tectum is compatible with this asymmetry.

The expected T-shape formed by the sections of the tegmentum and hypothalamus are compatible with the symmetric sections shown in Fg.2D. The asymmetry in the optic tectum is a result of the hyperplasic growth.

(4) The number of fluorescent cells found dispersed in the larvae transplanted with one single TIC after 48 hours will require a very fast cell cycle to generate over 50 cells. Do we have an idea of the cell cycle features of the transplanted TICs?

As answered above, the transplanted larvae are shown at 3dpt (and not 2dpt as misunderstood by this reviewer). With a cell cycle of about 10h, a single cell can give rise to about 100 cells in that time lapse.

Reviewer #2 (Public Review):

Summary:

This paper describes a genetically tractable and modifiable system …which could be used to study an array of combinations and temporal relationships of these cancer drivers/modifiers.

We thank this referee for its positive comments. We would also like to point out that our approach provides for the first quantitative means to estimate the probability of tumorigenesis from a single cell, an estimate which is crucial in any assessment of cancer malignancy and the effectiveness of prophylactics.

Weaknesses:

There is minimal quantitation of … the efficiency of activation of the Ras-TFP fusion (Fig 1) in, purportedly, a single cell. …, such information seems essential.

In a revised manuscript we will add more images of induction of a single (or a few cells) and a table where the efficiency of RAS activation is detailed.

The authors indicate that a single cell is "initiated" (Fig 2) using the laser optogenetic technique, but without definitive genetic lineage tracing, it is not possible to conclude that cells expressing TFP distant from the target site near the ear are daughter cells of the claimed single "initiated" cell. A plausible alternative explanation is 1) that the optogenetic targeting is more diffuse (i.e. some of the light of the appropriate wavelength hits other cells nearby due to reflection/diffraction), so these adjacent cells are additional independent "initiated" cells or 2) that the uncaged tamoxifen analogue can diffuse to nearby cells and allow for CreER activation and recombination.

We have addressed this point in our general comments to the reviewers’ remarks. The possibilities mentioned by this reviewer would result in cells expressing TFP in absence of VentX activation, which is not the case. Cells expressing TFP away from the initial site are observed days after activation of the oncogene (and TFP) in a single cell and only upon activation of VentX.

In Fig 2B, the claim is made that "the activated cell has divided, giving rise to two cells" - unless continuously imaged or genetically traced, this is unproven.

We have addressed this remark previously. Tracking of larvae over many days is not possible with the usual protocol using tricaine to immobilize the larvae. Nonetheless, in a revised version we will present images of an embryo imaged at various times post activation where proliferation of the cells can be observed. We are pursuing other alternatives for time-lapse microscopy over many days since, besides convincing the sceptics, a single cell tracking experiment (possibly coupled with in-situ spatial transcriptomics) will shed a new and fascinating light on the initial stages of tumor growth.

In addition, it appears that Figures S3 and S4 are showing that hyperplasia can arise in many different tissues (including intestine, pancreas, and liver, S4C) with broad Ras + Ventx activation …. This should be clarified in the manuscript).

This is true and will be clarified in the new version.

In Fig S7 where single cell activation and potential metastasis is discussed, similar gut tissues have TFP+ cells that are called metastatic, but this seems consistent with the possibility that multiple independent sites of initiation are occurring even when focal activation is attempted.

As mentioned previously this is ruled out by the fact that these cells are observed days after cyclofen uncaging (and TFP activation) and if and only if VentX is activated.

Although the hyperplastic cells are transplantable (Fig 4), the use of the term "cells of origin of cancer" or metastatic cells should be viewed with care in the experiments showing TFP+ cells (Fig 1, 2, 3) in embryos with targeted activation for the reasons noted above.

The purpose of this transplantation experiment was to show that cell in which both kRas and VentX have been activated possess the capacity to metastasize and develop a tumor mass when transplanted in a naïve zebrafish. This - to the best of our knowledge - is the operational definition of a malignant tumor.

Reviewer #3 (Public Review):

Summary:

This study employs an optogenetics approach … to examine tumourigenesis probabilities under altered tissue environments.

We thank this reviewer for this remark, since we believe that the opportunity to assess the probability of tumorigenesis from a single cell is possibly the most significant contribution of this work. To the best of our knowledge this has never been done before.

Weaknesses:

Lack of Methodological Clarity: The manuscript lacks detailed descriptions of methodologies,

In a revised manuscript we will include additional detail of our methodology.

Sub-optimal Data Presentation and Quality:

Lack of quantitative data and control condition data obtained from images of higher magnification limits the ability to robustly support the conclusions.

In a revised version we will include more images at higher magnification and quantitative data to support the main report of targeted single cell induction.

Here are some details:

Authors might want to provide more evidence to support their claim on the single cell KRAS activation.

More images and a data on activation of single or few cells in the illumination field will be provided in a revised version.

· Stability of cCYC: The manuscript does not provide information on the half-life and stability of cCYC. Understanding these properties is crucial for evaluating the system's reliability and the likelihood of leakiness, which could significantly influence the study's outcomes.

We have been using the cCyc system for about 14 years. We refer the reader to our previous papers and reviews on this methodology (e.g. ref. 34,35). Briefly, cCyc is stable when not illuminated with light around 375nm. Typically, we incubate our embryos in the dark for about 1h before transferring them into E3 medium and illuminating them. Assessing the leakiness of the system is easy as expression of the fluorescent marker is permanently turned on. We have observed none in the conditions of our experiment.

· Metastatic Dissemination claim: However, the absence of a supportive cellular compartment within the fin-fold tissue makes the presence of mTFP-positive metastatic cells there particularly puzzling. This distribution raises concerns about the spatial specificity of the optogenetic activation protocol … The unexpected locations of these signals suggest potential ectopic activation of the KRAS oncogene,

We have addressed this remark in the introduction and above. Specifically, metastatic and proliferative mTFP-positive cells are observed if and only if VentX is also activated concomitant with activation of kRAS in a single cell. No proliferative cells are observed in absence of VentX activation, or in presence of VentX or Dex alone, or if kRAS has not been activated by cyclofen uncaging.

· Image Resolution Concerns: The cells depicted in Figure 3C β, which appear to be near the surface of the yolk sac and not within the digestive system as suggested in the MS, underscore the necessity for higher-resolution imaging. Without clearer images, it is challenging to ascertain the exact locations and states of these cells, thus complicating the assessment of experimental results.

Better images will be provided in the revised version.

· The cell transplantation experiment is lacking protocol details:

Details will be provided in the revised version. We have followed regular protocols for transplantation: S.Nicoli and M.Presta, Nat.Prot. 2,2918 (2007).

• If the cells are obtained from whole larvae with induced RAS + VX expression, it is notable and somewhat surprising that the larvae survived up to six days post-induction (6dpi) before cells were harvested for transplantation. This survival rate and the subsequent ability to obtain single cell suspensions raise questions about the heterogeneity of the RAS + VX expressing cells that transplanted.

From Fig.S4D, about 50% of the embryos survive at 6dpi. Though an interesting question by itself we have not (yet) addressed the important issue of the heterogeneity of the outgrowth obtained from a single cell. Our purpose here was just to show that cells in which both kRAS and VentX have been activated possess the capacity to metastasize and develop a tumor mass when transplanted in a naïve zebrafish. This - to the best of our knowledge - is the operational definition of a malignant tumor.

· Unclear Experimental Conditions in Figure S3B: …It is not specified whether the activation of KRAS was targeted to specific cells or involved whole-body exposure.

This was whole body (global) illumination and will be specified in the revised version.

· Contrasting Data in Figure S3C compared to literature: The graph in Figure S3C indicates that KRAS or KRAS + DEX induction did not result in any form of hyperplastic growth. The authors should provide detailed descriptions of the conditions under which the experiments were conducted in Figure S3B and clarifying the reasons for the discrepancies observed in Figure S3C are crucial. The authors should discuss potential reasons for the deviation from previous reports.

This discrepancy will be discussed in the revised version. First the previous reports consider the development of tumors over a longer time-span (4-5 weeks) which we have not studied here. Second, the expression of the oncogene in these reports might be stronger than in ours. Third, the stochastic appearance of tumors in these reports suggest that some other mechanism (transient stress-induced reprogramming?) might have activated the oncogene in the initial cell.

Further comments:

Throughout the study, KRAS-activated cell expansion and metastasis are two key phenotypes discussed that Ventx is promoting. However, the authors did not perform any experiments to directly show that KRAS+ cells proliferate only in Ventx-activated conditions.

Yes, we did. See Fig. S1 and compare with Fig.S3B, or Fig.S8A in comparison with Fig.2A,B.

The authors also did not show any morphological features or time-lapse videos demonstrating that KRAS+ cells are motile, even though zebrafish is an excellent model for in vivo live imaging. This seems to be a missed opportunity for providing convincing evidence to support the authors' conclusions.

Performing single cell time-lapse microscopy on larvae over many (4-5) days is not possible with the regular tricaine protocol for immobilization. We are definitely planning such experiments, but they will require some other protocol, perhaps using bungarotoxin or some optogenetic inhibitory channels. Nonetheless, in the revised version we will show images of the same embryos at various times post single cell induction displaying proliferation of cells.

There were minimal experimental details provided for the qPCR data presented in the supplementary figures S5 and S6, therefore, it is hard to evaluate result obtained.

More details will be given in the revised version.

Author response:

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In this manuscript, Tutak et al use a combination of pulldowns, analyzed by mass spectrometry, reporter assays, and fluorescence experiments to decipher the mechanism of protein translation in fragile X-related diseases. The topic is interesting and important.

Although a role for Rps26-deficient ribosomes in toxic protein translation is plausible based on already available data, the authors' data are not carefully controlled and thus do not support the conclusions of the paper.

Strengths:

The topic is interesting and important.

Weaknesses:

In particular, there is very little data to support the notion that Rps26-deficient ribosomes are even produced under the circumstances. And no data that indicate that they are involved in the RAN translation. Essential controls (for ribosome numbers) are lacking, no information is presented on the viability of the cells (Rps26 is an essential protein), and the differences in protein levels could well arise from block in protein synthesis, and cell division coupled to differential stability of the proteins.

We agree that presented data could benefit from addition of suggested experiments. We will  address the ribosome content, global translation rate and cell viability upon RPS26 depletion. We are also planning to apply polysome profiling to determine if RPS26-depleted ribosomes are translationally active.

Specific points:

(1) Analysis of the mass spec data in Supplemental Table S3 indicates that for many of the proteins that are differentially enriched in one sample, a single peptide is identified. So the difference is between 1 peptide and 0. I don't understand how one can do a statistical analysis on that, or how it would give out anything of significance. I certainly do not think it is significant. This is exacerbated by the fact that the contaminants in the assay (keratins) are many, many-fold more abundant, and so are proteins that are known to be mitochondrial or nuclear, and therefore likely not actual targets (e.g. MCCC1, PC, NPM1; this includes many proteins "of significance" in Table S1, including Rrp1B, NAF1, Top1, TCEPB, DHX16, etc...).

The data in Table S6/Figure 3A suffer from the same problem.

Tables S3 and S6 show the mass spectrometry output data from MaxQuant analysis  without any flittering.  Certain identifications, i.e. those denoted as contaminants (such as keratins) were removed during statistical analysis in Perseus software. Regarding the data presented in Table S6 (SILAC data), we argue that these data are of very good quality. More than 2000 proteins were identified in a 125min gradient, with over 80% of proteins that were identified with at least 2 unique peptides. However, we acknowledge that the description of Tables S3 and S6 may lead to misunderstanding, thus we will clarify their explanation.

I am not convinced that the mass spec data is reliable.

(2) The mass-spec data however claims to identify Rps26 as a factor binding the toxic RNA specifically. The rest of the paper seeks to develop a story of how Rps26-deficient ribosomes play a role in the translation of this RNA. I do not consider that this makes sense.

Indeed, we identified RPS26 as a protein co-precipitated with FMR1 RNA containing expanded CGG repeats. However, we do not claim that they interact directly. Downregulation of FMRpolyG biosynthesis could be an outcome of the alteration of ribosomal assembly, changes in efficiency and fidelity of PIC scanning or impeded elongation or more likely combination of some of these processes. We will  provide better explanation regarding those issues in the revised version of the manuscript.

(3) Rps26 is an essential gene, I am sure the same is true for DHX15. What happens to cell viability? Protein synthesis? The yeast experiments were carefully carried out under experiments where Rps26 was reduced, not fully depleted to give small growth defects.

We agree with the Reviewer 1 that RPS26 is an essential protein. Previously, it was shown that cell viability in cells with mutated C-terminal deletion of RPS26 is decreased (Havkin-Solomon T, Nucleic Acids Res 2023). We will address the question regarding the suppression of FMRpolyG in models with partial RPS26 knock-down.

(4) Knockdown efficiency for all tested genes must be shown to evaluate knockdown efficiency.

Missing experiments showing efficiency of knock-down will be included in the revised version of the manuscript.

(5) The data in Figure 1E have just one mock control, but two cell types (control si and Rps26 depletion).

We will clarify this ambiguity in the revised version of the manuscripts.

(6) The authors' data indicate that the effects are not specific to Rps26 but indeed also observed upon Rps25 knockdown. This suggests strongly that the effects are from reduced ribosome content or blocked protein synthesis. Additional controls should deplete a core RP to ascertain this conclusion.

We agree that observed effect may stem partially from reduced ribosome content, however, we argue that this is not the only explanation. In the publication concerning RPS25 regulation of G4C2-related RAN translation (Yamada SB, 2019, Nat Neurosci), it was shown that RPS25 KO does not affect global translation. Our experiments (SUnSET assay, unpublished) indicated that RPS26 KD also did not reduce global translation rate significantly. We will present that data in the revised version of the manuscript.

(7) Supplemental Figure S3 demonstrates that the depletion of S26 does not affect the selection of the start codon context. Any other claim must be deleted. All the 5'-UTR logos are essentially identical, indicating that "picking" happens by abundance (background).

Results shown in Fig.S3 does not imply that RPS26 does not affect the selection of start codon context entirely. We just tested a few hypotheses. We decided to test -4 position, because this position was indicated as the most sensitive to RPS26 regulation in yeast (Ferretti M, 2017, Nat Struct Mol Biol). Regarding WebLOGO analysis; we wrote in the manuscript that we did not identify any specific motif or enrichment within analysed transcripts in comparison to background. We will clarify this ambiguity in revised version of the manuscript.

(8) Mechanism is lacking entirely. There are many ways in which ribosomes could have mRNA-specific effects. The authors tried to find an effect from the Kozak sequence, unsuccessfully (however, they also did not do the experiment correctly, as they failed to recognize that the Kozak sequence differs between yeast, where it is A-rich, and mammalian cells, where it is GGCGCC). Collisions could be another mechanism.

As in (7).

Reviewer #2 (Public Review):

Summary:

Translation of CGG repeats leads to the accumulation of poly G, which is associated with neurological disorders. This is a valuable paper in which the authors sought out proteins that modulate RAN translation. They determined which proteins in Hela cells bound to CGG repeats and affected levels of polyG encoded in the 5'UTR of the FMR1 mRNA. They then showed that siRNA depletion of ribosomal protein RPS26 results in less production of FMR1polyG than in control. There are data supporting the claim that RPS26 depletion modulates RAN translation in this RNA, although for some results, the Western results are not strong. The data to support increased aggregation by polyG expression upon S26 KD are incomplete.

Strengths:

The authors have proteomics data that show the enrichment of a set of proteins on FMR1 RNA but not a related RNA.

Weaknesses:

- It is insinuated that RPS26 binds the RNA to enhance CGG-containing protein expression. However, RPS26 reduction was also shown previously to affect ribosome levels, and reduced ribosome levels can result in ribosomes translating very different RNA pools.

We agree that presented data could benefit from addition of some experiments. Therefore we will address questions regarding the ribosome content, global translation rate and cell viability upon RPS26 depletion. We are also planning to apply polysome profiling to determine if RPS26-depleted ribosomes are translationally active. However, we did not state that RPS26 binds directly to RNA with expanded CGG repeats and that this interaction is crucial for translation regulation of studied RNA. We just tested such hypotheses. We will improve the text narration in revised version of the manuscript to make major conclusions clearer.

- A significant claim is that RPS26 KD alleviates the effects of FMRpolyG expression, but those data aren't presented well.

We thank the Reviewer 2 for this comment. We will show the data derived from a few different cell models that we already have obtained. Moreover, we will include results of experiments with luminescence readout for FMRpolyG fused with luciferase upon RPS26 KD.

Reviewer #3 (Public Review):

Tutak et al provide interesting data showing that RPS26 and relevant proteins such as TSR2 and RPS25 affect RAN translation from CGG repeat RNA in fragile X-associated conditions. They identified RPS26 as a potential regulator of RAN translation by RNA-tagging system and mass spectrometry-based screening for proteins binding to CGG repeat RNA and confirmed its regulatory effects on RAN translation by siRNA-based knockdown experiments in multiple cellular disease models and patient-derived fibroblasts. Quantitative mass spectrometry analysis found that the expressions of some ribosomal proteins are sensitive to RPS26 depletion while approximately 80% of proteins including FMRP were not influenced. Since the roles of ribosomal proteins in RAN translation regulation have not been fully examined, this study provides novel insights into this research field. However, some data presented in this manuscript are limited and preliminary, and their conclusions are not fully supported.

(1) While the authors emphasized the importance of ribosomal composition for RAN translation regulation in the title and the article body, the association between RAN translation and ribosomal composition is apparently not evaluated in this work. They found that specific ribosomal proteins (RPS26 and RPS25) can have regulatory effects on RAN translation（Figures 1C, 2B, 2C, 2E, 4A, 5A, and 5B), and that the expression levels of some ribosomal proteins can be changed by RPS26 knockdown (Figure 3B, however, the change of the ribosome compositions involved in the actual translation has not been elucidated). Therefore, their conclusive statement, that is, "ribosome composition affects RAN translation" is not fully supported by the presented data and is misleading.

We thank Reviewer 3 for critical comments and suggestions. We agree that the proposed title may be misleading and the presented data does not fully support the aforementioned statement regarding ribosomal composition affecting FMRpolyG synthesis. Hence, we will change the title together with a narrative regarding these unfortunate statements that go beyond the presented results.

(2) The study provides insufficient data on the mechanisms of how RPS26 regulates RAN translation. Although authors speculate that RPS26 may affect initiation codon fidelity and regulate RAN translation in a CGG repeat sequence-independent manner (Page 9 and Page 11), what they really have shown is just identification of this protein by the screening for proteins binding to CGG repeat RNA (Figure 1A, 1B), and effects of this protein on CGG repeat-RAN translation. It is essential to clarify whether the regulatory effect of RPS26 on RAN translation is dependent on CGG repeat sequence or near-cognate initiation codons like ACG and GUG in the 5' upstream sequence of the repeat. It would be better to validate the effects of RPS26 on translation from control constructs, such as one composed of the 5' upstream sequence of FMR1 with no CGG repeat, and one with an ATG substitution in the 5' upstream sequence of FMR1 instead of near-cognate initiation codons.

We will address the question regarding the influence of the content of CGG repeats and START codon selection (including different near-cognate start codons) on RPS26-sensitive translation, and present these data in revised version of the manuscript.

(3) The regulatory effects of RPS26 and other molecules on RAN translation have all been investigated as effects on the expression levels of FMRpolyG-GFP proteins in cellular models expressing CGG repeat sequences Figures 1C, 2B, 2C, 2E, 4A, 5A, and 5B). In these cellular experiments, there are multiple confounding factors affecting the expression levels of FMRpolyG-GFP proteins other than RAN translation, including template RNA expression, template RNA distribution, and FMRpolyG-GFP protein degradation. Although authors evaluated the effect on the expression levels of template CGG repeat RNA, it would be better to confirm the effect of these regulators on RAN translation by other experiments such as in vitro translation assay that can directly evaluate RAN translation.

We agree that there are multiple factors affecting final translation of investigated mRNA including aforementioned processes. We evaluated the level of FMR1 mRNA, which turned out not to be affected upon RPS26 depletion (Figure 2B&C), however, we will address other possibilities as well.

(4) While the authors state that RPS26 modulated the FMRpolyG-mediated toxicity, they presented limited data on apoptotic markers, not cellular viability (Figure 1E), not fully supporting this conclusion. Since previous work showed that FMRpolyG protein reduces cellular viability (Hoem G et al., Front Genet 2019), additional evaluations for cellular viability would strengthen this conclusion.

We thank Reviewer 3 for this suggestion. We addressed the effect of RPS26 KD on apoptotic process induced by FMRpolyG. We will perform other experiments regarding different aspects of FMRpolyG-mediated cell toxicity as well.

Author response:

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

Public Review:

This article is a direct follow-up to the paper published last year in eLife by the same group. In the previous article, the authors discovered a zinc finger protein, Kipferl, capable of guiding the HP1 protein Rhino towards certain genomic regions enriched in GRGGN motifs and packaged in heterochromatin marked by H3K9me3. Unlike other HP1 proteins, Rhino recruitment activates the transcription of heterochromatic regions, which are then converted into piRNA source loci. The molecular mechanism by which Kipferl interacts specifically with Rhino (via its chromodomain) and not with other HP1 proteins remained enigmatic. 

In this latest article, the authors go a step further by elucidating the molecular mechanisms important for the specific interaction of Rhino and not other HP1 proteins with Kipferl. A phylogenetic study carried out between the HP1 proteins of 5 Drosophila species led them to study the importance of an AA Glycine at position 31 located in the Rhino chromodomain, an AA different from the AA (aspartic acid) found at the same position in the other HP1 proteins. The authors then demonstrate, through a series of structure predictions, biochemical, and genetic experiments, that this specific AA in the Rhino-specific chromodomain explains the difference in the chromatin binding pattern between Rhino and the other Drosophila HP1 proteins. Importantly, the G31D conversion of the Rhino protein prevents interaction between Rhino and Kipferl, phenocopying a Kipferl mutant. 

Strengths: 

The authors' effective use of phylogenetic analyses and protein structure predictions to identify a substitution in the chromodomain that allows Rhino's specific interaction with Kipferl is very elegant. Both genetic and biochemical approaches are applied to rigorously probe the proposed explanation. They used a point mutation in the endogenous locus that replaces the Rhino-specific residue with the aspartic acid residue present in all other HP1 family members. This novel allele largely phenocopies the defects in hatch rate, chromatin organization, and piRNA production associated with kipferl mutants, and does not support Kipferl localization to clusters. The data are of high quality, the presentation is clear and concise, and the conclusions are generally well-supported.

Weaknesses: 

The reviewers identified potential ways to further strengthen the manuscript.

(1) The one significant omission is RNAseq on the rhino point mutant, which would allow direct comparison to cluster, transposon, and repeat expression in kipferl mutants. 

In this eLife Advances submission, we aim to elucidate the molecular interaction between Rhino and the zinc finger protein Kipferl and how it evolved. Using various assays, of which piRNA sequencing is the most relevant and comprehensive, we show that the rhino[G31D] mutation phenocopies a rhino loss-of-function situation for Kipferl and a kipferl loss-of-function situation for Rhino. Further confirmation of this statement by additional RNA-seq experiments to probe the extent of selective TE de-repression would indeed be a possibility. We decided to test for TE de-repression phenotypes using sensitive RNA-FISH experiments of a handful of TEs that are deregulated in kipferl loss of function flies (Baumgartner at al. 2022). This showed that the same TEs are also deregulated in rhino[G31D] flies, further confirming the similarity of the two genotypes. We have added these data to the text and to Figure 5-figure supplement 2, which shows representative RNA FISH images.

(2) The manuscript would benefit from adding more evolutionary comparisons. The following or similar analyses would help put the finding into a broader evolutionary perspective:

i) Is Kipferl's surface interacting with Rhino also conserved in Kipferl orthologs? In other words, are the Rhino-interacting amino acids of Kipferl under any pressure to be conserved?

We performed an analysis of the Kipferl interface that interacts with the Rhino chromodomain in those species where Kipferl could be unambiguously identified. This showed that the residues involved in the Rhino interaction are generally conserved. We have added this analysis to Figure 1-figure supplement 4.

ii) The remarkable conservation of Rhino's G31 is at odds with the arms race that is proposed to be happening between the fly's piRNA pathway proteins and transposons. Does this mean that Rhino's chromodomain is "untouchable" for such positive selection? 

We agree that the conservation of the G31 residue argues against this binding interface being under positive selection in Rhino. Without understanding the pressures acting on Rhino that underlie the previously published positive selection, we find it difficult to draw firm conclusions. Mutating G31 in fly species that lack Kipferl would be an interesting experiment.

Recommendations for the authors:

(1) RNAseq is important to the full characterization of the phenotype and should be included. It's now clear that the major piRNA clusters are not required for fertility, so I would also include an analysis of piRNA production and Rhino binding to regions flanking isolated insertions. 

See our response to raised weakness #1 above. Briefly, we have now added an analysis of TE de-repression based on RNA-FISH experiments (Figure 5-figure supplement 2). Regarding the proposed analysis of piRNA production and Rhino binding to regions flanking isolated TE insertions: this is an important issue that we carefully analysed in our previous work characterising the kipferl mutant (Baumgartner et al. 2022). In the present work, we focused on generating a rhino mutant that uncouples Rhino from Kipferl.

(2) The authors do not provide direct biochemical evidence that the chromodomain substitution blocks Rhino binding to Kipferl. However, Rhino protein is very low abundance, making analysis of the endogenous protein very difficult.

Based on our previous work (Baumgartner et al 2022), the Rhino chromodomain interacts directly with the fourth zinc finger of Kipferl. Mutation of a single residue in the predicted interface (Rhino[G31D]) phenocopies a kipferl mutant, strongly suggesting that this mutation disrupts the Rhino-Kipferl interaction. Definitive evidence will have to await the reconstitution of this interaction using recombinant proteins. Our attempts to purify recombinant Kipferl (expressed in bacteria or in insect cells) or the protein fragments relevant to the interaction were unsuccessful so far. While we obtained soluble fractions of the first ZnF array, there was always a high level of co-purifying nucleic acids that we were not able to remove.

(3) Even if the Rhino G31D mutant retains its ability to interact with H3K9me3 it does not localize correctly on the chromatin preventing certain regions such as locus 80F from being converted into piRNA source loci. However other regions such as satellite regions attract the Rhino mutant protein converting them into super piRNA source loci, phenocopying the effects observed in a Kipferl mutant. Why Rhino when not bound to Kipferl concentrates in satellite regions is a question that remains unanswered.

This is a very interesting question indeed. We have not been able to elucidate the molecular basis of how Rhino is recruited to satellite repeats in Kipferl mutants. For example, we performed a proximity biotinylation experiment with GFP-Rhino in Kipferl mutant ovaries, but this experiment did not reveal any protein that would explain the observed accumulation of Rhino at the complex satellite repeats.

(4) In the phylogenetic analysis the authors identified two residues as Rhino-specific and conserved sequence alterations, the D31G mutation and the G62 insertion. However, the authors limit their study to D31G mutation, and nothing is performed on the G32 insertion. It would have been interesting to know the impact of this insertion on Rhino's biology. 

The role, if any, of the Rhino-specific G62 insertion and its effect on Rhino localisation or function is an interesting topic for further study. We have not investigated the G62 residue experimentally. In the current manuscript, we limited our efforts to the analysis of the G31D mutation, as the goal was to identify the mode of interaction with Kipferl, and the G62 residue is not predicted to contact Kipferl according to AlphaFold.

(5) The authors report that the G31D mutation of Rhino phenocopies the Kipferl mutant. Rhino is wrongly localized in the nucleus, and Rhino G31D recruitment in certain Kipferl-enriched regions is affected, as at the 80F locus, which correlates with a strong drop in piRNA production from this locus. To go a step further in demonstrating that G31D phenocopies the Kipferl mutant, it would have been informative to analyse how much TE piRNAs are affected and whether TEs are deregulated.

See our response to similar comments above. We have added RNA-FISH experiments to illustrate that the TE de-repression phenotypes are comparable between rhino[G31D] and kipferl loss of function ovaries (Figure 5-figure supplement 2). Analyses of TE-mapping piRNAs also show well correlated phenotypes (Figure 5-figure supplement 1).

(6) Figure 3A: To homogenize with the immunostaining presented in Figure 3B, can the authors add on the bar graph depicting female fertility the results obtained with kipferl-/- and rhino-/- genotype? 

rhino mutants are completely (100%) sterile and the fertility of kipferl mutants was previously measured to range between 15% and 40% (Baumgartner et al. 2022).

(7) Figure 4A: It would have been interesting to show Venn diagrams showing the overlap of genomic regions enriched for Kipferl versus regions enriched for Rhi in a WT and in a Rhi G31D mutant. 

We consider the analysis presented in Figure 4 to be more meaningful, as a Venn diagram would require binary cut-offs.

(8) Figure 1B: In the phylogenic analysis for Rhino/HP1d two D. simulans lines are presented. Can the authors clarify this point?

There are two Rhino paralogs in D. simulans: one paralog (NCBI: AAY34025.1) is more similar to D. melanogaster Rhino, contains one intron and is located at chromosome chr2R (assembly Apr. 2005, WUGSC mosaic 1.0/droSim1: 12256895-12258668). The second paralog (XP_002106478.1) is located on chromosome X (6734493-6735248) and does not contain an intron. We have added a clarifying statement to the corresponding figure legend.

(9) To determine whether Rhino G31D point mutation affects the overall function of Rhino, the authors analysed Kipferl-independent piRNA source loci by looking at Responder and 1,688 family satellites. I'm not sure that these loci can be classified as Kipferl-independent piRNA source loci since a strong increase of piRNA production from these loci in Kipferl mutant is observed. In my point of view, the 42AB and 38C are real Kipferl-independent piRNA source loci as piRNA production from these loci is not affected by Kipferl KD. 

Indeed, the Rsp and 1,688 family satellites are not completely independent of Kipferl, as their expression and Rhino occupancy differ between wild-type and kipferl loss-of-function phenotypes (including rhino[G31D]). However, we believe that this increase is due to a strong dependence on different sequestration mechanisms and is not mediated by a direct function of Kipferl at these sites. Similarly, we observe slight differences in piRNA production for the peripheral parts of cluster 42AB, as well as differences in Rhino occupancy despite an unaltered piRNA profile at cluster 38C (Baumgartner et al. 2022). Thus, different flavours of Kipferl-independence exist, with the only truly Kipferl-independent piRNA sources likely to be the piRNA clusters in the testis. A clear classification is further complicated by previously observed compensatory effects in the piRNA pathway, leading us to adopt the current definition of "requiring Kipferl for Rhino recruitment" to distinguish Kipferl-dependent from Kipferl-independent sites.

(10) The authors report that the G31D mutation of Rhino phenocopies the Kipferl mutant. Rhino is wrongly localized in the nucleus, and Rhino G31D recruitment in certain Kipferl-enriched regions is affected, as at 80F locus, which correlates with a strong drop in piRNA production from this locus. To go a step further in demonstrating that G31D phenocopies the Kipferl mutant, it would have been interesting to look at how much TE piRNAs are affected and whether TEs (and which class of TE) are deregulated by RNAseq and/or in situ hybridization. 

See our response to similar comments above. Our new RNA-FISH experiments and TE-mapping piRNA analysis extend the comparison of phenotypes between kipferl mutants and rhino[G31D] mutants and are consistent with our previous conclusions (Figure 5-figure supplements 1 and 2).

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In this study, Jellinger et al. performed engram-specific sequencing and identified genes that were selectively regulated in positive/negative engram populations. In addition, they performed chronic activation of the negative engram population over 3 months and observed several effects on fear/anxiety behavior and cellular events such as upregulation of glial cells and decreased GABA levels.

Strengths:

They provide useful engram-specific GSEA data and the main concept of the study, linking negative valence/memory encoding to cellular level outcomes including upregulation of glial cells, is interesting and valuable.

Weaknesses:

A number of experimental shortcomings make the conclusion of the study largely unsupported. In addition, the observed differences in behavioral experiments are rather small, inconsistent, and the interpretation of the differences is not compelling.

Major points for improvement:

(1) Lack of essential control experiments

With the current set of experiments, it is not certain that the DREADD system they used was potent and stable throughout the 3 months of manipulations. Basic confirmatory experiments (e.g., slice physiology at 1m vs. 3m) to show that the DREADD effects on these vHP are stable would be an essential bottom line to make these manipulation experiments convincing.

In previous work from our lab performing long-term activation of Gq DREADD receptors in the vHPC, we quantify the presence of Gq receptor expression over 3-, 6- and 9-month timepoints and show that there is no decrease in receptor expression, as measured via fluorescence intensity (Suthard et al., 2023). In this study, we also address that even if our manipulation is only working for 1 month, rather than 3 months, we are observing the long-term effects of this shorter-term stimulation. This is still relevant, and only changes how we interpret these findings, as shorter-term stimulation or disruption of neuronal activity can still have detrimental effects on behavior.

Furthermore, although the authors use the mCherry vector as a control, they did not have a vehicle/saline control for the hM3Dq AAV. Thus, the long-term effects such as the increase in glial cells could simply be due to the toxicity of DREADD expression, rather than an induced activity of these cells.

For chemogenetic studies, our experimental rationale utilized a standard approach in the field, which includes one of two control options: 1) active receptor vs. control vector + ligand or 2) active receptor + ligand or saline control. We chose the first option, as this more properly controls for the potential off-target effects of the ligand itself, as shown in other previous work (Xia et al., 2017). This is particularly important for studies using CNO, as many off-target effects have been noted as a limitation (Manvich et al., 2018). We chose to use DCZ as it is closely related to CNO and newer ligands, but comes with added benefits of high specificity, low off-target effects, high potency and brain penetrance (Nagai et al., 2020), but any potential off-target effects of DCZ are yet to be completely investigated as this ligand is very new.

Evidence of DREADD toxicity has been shown at high titer levels of AAV2/7- CamKIIα-hM4D(Gi)-mCherry in the hippocampus at 5 weeks, as the reviewer pointed out in their above comment (Goossens et al., 2021). Our viral strategy is targeted to a much smaller number of cells using AAV9-DIO-Flex-hM3Dq-mCherry at a lower titer, unlike expression within a much larger population of CaMKII+ excitatory neurons in this study. Additionally, visual comparison of their viral load and expression with ours shows much more intense expression that spans a larger area of the hippocampus (Goossens et al, 2021; Figure 1D), whereas ours is isolated to a smaller region of vHPC (see Figure 1B).

Further, we attempted to quantify a decrease in neuronal health (Yousef et al., 2017) resulting from DREADD expression via NeuN counts within multiple hippocampal subregions for the 6- and 14-month groups across active Gq receptor and mCherry conditions and did not observe significant decreases in NeuN as a result (Supplemental Figure 1). However, immunohistochemistry of an individual marker may not be sufficient to capture the entire health profile of an individual neuron and future work should consider other markers of cell death or inflammation, which we have added to the Limitations & Future Work section of our Discussion.

(2) Figure 1 and the rest of the study are disconnected

The authors used the cFos-tTA system to label positive/negative engram populations, while the TRAP2 system was used for the chronic activation experiments. Although both genetic tools are based on the same IEG Fos, the sensitivity of the tools needs to be validated. In particular, the sensitivity of the TRAP2 system can be arbitrarily altered by the amount of tamoxifen (or 4OHT) and the administration protocols. The authors should at least compare and show the percentage of labeled cells in both methods and discuss that the two experiments target (at least slightly) different populations. In addition, the use of TRAP2 for vHP is relatively new; the authors should confirm that this method actually captures negative engram populations by checking for reactivation of these cells during recall by overlap analysis of Fos staining or by artificial activation.

We thank the reviewer for their comments and opportunity to discuss the marked differences between TRAP2 and DOX systems. In particular, we agree that while both systems rely on the the Fos promoter to drive an effector of interest, their efficacy and temporal resolution vary substantially depending on genetic cell-type, brain region, temporal parameters of Dox or 4-OHT delivery, subject-by-subject metabolic variability, and threshold to Fos induction given the promoter sequences inherent to each system. For example, recent studies have reported the following:

- The TRAP2 line labels a subset of endogenously activeCA1 pyramidal cells (e.g. 5-18%) while the DOX system labels 20-40% of CA1 pyramidal cells (DeNardo et al, 2019; Monasterio et al, BioRxiv 2024 ).

- The temporal windows for each range from hours in TRAP2 to 24-48 hours for DOX (DeNardo et al, 2019; Denny et al, 2014; Liu & Ramirez et al, 2012).

- The efficacy of “tagging” a population of cells with TRAP2 vs with DOX will constrain the number of possible cells that may overlap with cFos upon re-exposure to a given experience (e.g. see the observed overlaps in vCA1 - BLA circuits (Kim & Cho, 2020), compared to vCA1 in general (Ortega-de San Luis et al, 2023) and valence-specific vCA1 populations (Shpokayte et al, 2022).

- Tagging vCA1 cells with both the TRAP2 and DOX systems are nonetheless sufficient to drive corresponding behaviors (e.g. vCA1 terminal stimulation drives behavioral changes with the DOX and TRAP2 system (Shpokayte et al, 2022) and vCA1 stimulation of an updated fear-linked ensemble drives light-induced freezing in a neutral context utilizing the TRAP2 and DOX systems (Ortega-de San Luis et al, 2023)).

Finally, and promisingly, as more studies continue to link the in vivo physiological dynamics of these cell populations tagged using each system (e.g. compare Pettit et al, 2022 with Tanaka et al, 2018) and correlating their activity to behavioral phenotypes, our field is in the prime position to uncover deeper principles governing hippocampus-mediated engrams in the brain. Together, we believe a more comprehensive understanding of these systems is fully warranted, especially in the service of further cataloging cellular similarities and differences within such tagged populations.

(3)  Interpretation of the behavior data

In Figures 3a and b, the authors show that the experimental group showed higher anxiety based on time spent in the center/open area. However, there were no differences in distance traveled and center entries, which are often reduced in highly anxious mice. Thus, it is not clear what the exact effect of the manipulation is. The authors may want to visualize the trajectories of the mice's locomotion instead of just showing bar graphs.

Our findings show that our experimental group displays higher levels of anxiety-like behaviors as measured via time spent in center/open area, while there are no differences in distance traveled or center entries. For distance traveled, our interpretation is in line with complementary research (Jimenez et al, 2018; Kheirbek et al, 2013) that shows no changes in distance traveled/distance traveled in the center coupled with changes in anxiety levels as a result of manipulation within anxiety-related circuits. More broadly, any locomotion-related deficit could cause a change in distance traveled that is unrelated to anxiety-like behaviors alone. For example, a reduction in distance traveled could be coupled with a decrease in time spent in the center, but could also result only from motor or exploratory deficits. We hope that this explanation clarifies our interpretation of the open field and elevated plus maze findings in light of other literature.

In addition, the data shown in Figure 4b is somewhat surprising - the 14MO control showed more freezing than the 6MO control, which can be interpreted as "better memory in old". As this is highly counterintuitive, the authors may want to discuss this point. The authors stated that "Mice typically display increased freezing behavior as they age, so these effects during remote recall are expected" without any reference. This is nonsense, as just above in Figure 4a, older mice actually show less freezing than young mice. Overall, the behavioral effects are rather small and random. I would suggest that these data be interpreted more carefully.

In Figure 4B, we present our findings from remote recall and observe increased freezing levels in control mice with age, as mentioned by the reviewer, indicating increased memory. This is in line with previous work from Shoji & Miyakawa, 2019 which has been added as a reference for the quotation described above; we thank the reviewer for pointing this error out. As the reviewer has pointed out, above in Figure 4A, we measured freezing levels across all groups during contextual fear conditioning before the start of chronic stimulation, as this was the session we ‘tagged’ a negative memory in. Although it appears that there may be slightly lower levels of freezing in older (14-month old) mice, our findings do not determine statistical significance for difference between age group, only effects of time and subject which are expected as freezing increases within the session and animals display high levels of variability in freezing levels across many experiments (Figure 4A i-iii). We also find in previous work that control mice receiving 3-, 6- and 9-months of chronic DCZ stimulation in the vHPC with empty vector (mCherry) receptor show an increase in freezing with age (Suthard et al, 2023; Figure 2A ii).

(4) Lack of citation and discussion of relevant study

Khalaf et al. 2018 from Gräff lab showed that experimental activation of recall-induced populations leads to fear attenuation. Despite the differences in experimental details, the conceptual discrepancy should be discussed.

As mentioned by the reviewer, Khalaf et al. 2018 showed that experimental activation of recall-induced populations in the dentate gyrus leads to fear attenuation. Specifically, they pose that this fear attenuation occurs in these ensembles through updating or unlearning of the original memory trace via the engagement, rather than suppression, of an original traumatic experience. Despite the differences in experimental details with our current study and this work, we agree that the conceptual discrepancy should be discussed. First, one major difference is that we are reactivating an ensemble that was tagged during fear memory encoding, while Khalaf et al. are activating a remote recall-induced ensemble that was tagged one month after encoding. Although there is high overlap between the encoding and recall ensembles when mice are exposed to the conditioning context, these ensembles are not identical and may result in different behavioral phenotypes when chronically reactivated. Further, Khalaf et al rely on reactivation of the recall-induced ensemble during extinction to facilitate rapid fear attenuation. This differs from our current work, as their reactivation is occurring during the extinction process in the previously conditioned context, while we are reactivating chronically in the animal’s home cage over the course of a longer time period. It may be necessary that the memory is first reactivated, and thus, more liable to re-contextualization, in the original context compared to an unrelated homecage environment where there are presumably no related cues present. Importantly, this previous work tests the attenuation of fear shortly after an extinction process, while we are not traditionally extinguishing the context with aid of the memory reactivation. Finally, we are testing remote recall (3 months post-conditioning), while they are testing at a shorter time interval (28 days). In line with these ideas, future work may seek to tease out the mechanistic differences between recent and remote memory extinction both in terms of natural memory recall and chronically manipulated memory-bearing cells.

Reviewer #2 (Public Review):

Summary:

Jellinger, Suthard, et al. investigated the transcriptome of positive and negative valence engram cells in the ventral hippocampus, revealing anti- and pro-inflammatory signatures of these respective valences. The authors further reactivated the negative valence engram ensembles to assay the effects of chronic negative memory reactivation in young and old mice. This chronic re-activation resulted in differences in aspects of working memory, and fear memory, and caused morphological changes in glia. Such reactivation-associated changes are putatively linked to GABA changes and behavioral rumination.

Strengths:

Much of the content of this manuscript is of benefit to the community, such as the discovery of differential engram transcriptomes dependent on memory valence. The chronic activation of neurons, and the resultant effects on glial cells and behavior, also provide the community with important data. Laudable points of this manuscript include the comprehensiveness of behavioral experiments, as well as the cross-disciplinary approach.

Weaknesses:

There are several key claims made that are unsubstantiated by the data, particularly regarding the anthropomorphic framing of "rumination" on a mouse model and the role of GABA. The conclusions and inferences in these areas need to be carefully considered.

(1) There are many issues regarding the arguments for the behavioural data's human translation as "rumination." There is no definition of rumination provided in the manuscript, nor how rumination is similar/different to intrusive thoughts (which are psychologically distinct but used relatively interchangeably in the manuscript), nor how rumination could be modelled in the rodent. The authors mention that they are attempting to model rumination behaviours by chronically reactivating the negative engram ("To understand if our experimental model of negative rumination..."), but this occurs almost at the very end of the results section, and no concrete evidence from the literature is provided to attempt to link the behavioural results (decreased working memory, increased fear extinction times) to rumination-like behaviours. The arguments in the final paragraph of the Discussion section about human rumination appear to be unrelated to the data presented in the manuscript and contain some uncited statements. Finally, the rumination claims seem to be based largely upon a single data figure that needs to be further developed (Figure 6, see also point 2 below).

(2) The staining and analysis in Figure 6 are challenging to interpret, and require more evidence to substantiate the conclusions of these results. The histological images are zoomed out, and at this resolution, it appears that only the pyramidal cell layer is being stained. A GABA stain should also label the many sparsely spaced inhibitory interneurons existing across all hippocampal layers, yet this is not apparent here. Moreover, both example images in the treatment group appear to have lower overall fluorescence intensity in both DAPI and GABA. The analysis is also unclear: the authors mention "ROIs" used to measure normalized fluorescence intensity but do not specify what the ROI encapsulates. Presumably, the authors have segmented each DAPI-positive cell body and assessed fluorescence however, this is not explicated nor demonstrated, making the results difficult to interpret.

Based on the collective discussion from all reviewers on the completeness of our GABA quantification and its implications, we have decided to remove this figure and perform more substantive analysis of this E/I imbalance in future work.

(3) A smaller point, but more specific detail is needed for how genes were selected for GSEA analysis. As GSEA relies on genes to be specified a priori, to avoid a circular analysis, these genes need to be selected in a blind/unbiased manner to avoid biasing downstream results and conclusions. It's likely the authors have done this, but explicitly noting how genes were selected is an important context for this analysis.

As mentioned in our Methods section, gene sets were selected based on pre-existing biology and understanding of genes canonically involved in “neurodegeneration” such as those related to apoptotic pathways and neuroinflammation or “neuroprotection” such as brain-derived neurotrophic factor, to name a few. A limitation of this method is that we must avoid making strong claims about the actual function of these up- or down-regulated genes without performing proper knock-in or knock-out studies, but we hope that this provides an unbiased inventory for future experiments to perform causal manipulations.

Reviewer #3 (Public Review):

Summary:

The authors note that negative ruminations can lead to pathological brain states and mood/anxiety dysregulation. They test this idea by using mouse engram-tagging technology to label dentate gyrus ensembles activated during a negative experience (fear conditioning). They show that chronic chemogenetic activation of these ensembles leads to behavioral (increased anxiety, increased fear generalization, reduced fear extinction) and neural (increases in neuroinflammation, microglia, and astrocytes).

Strengths:

The question the authors ask here is an intriguing one, and the engram activation approach is a powerful way to address the question. Examination of a wide range of neural and behavioral dependent measures is also a strength.

Weaknesses:

The major weakness is that the authors have found a range of changes that are correlates of chronic negative engram reactivation. However, they do not manipulate these outcomes to test whether microglia, astrocytes, or neuroinflammation are causally linked to the dysregulated behaviors.

Recommendations For The Authors:

Reviewer #1 (Recommendations For The Authors):

- Figure 2c should include Month0, the BW before the start of the manipulation.

Regrettably, we do not have access to the Month 0 body weights at this time as this project changed hands over the course of the past year or so. This is an inherent limitation that we missed during analysis and we pose this as a limitation in the Results section after describing this finding. Therefore, it is possible that over the first month of stimulation (Month 0-1), there may have been a drop in body weight that rebounded by the first measurement at Month 1 that continued to increase normally through Months 2-3, as shown in our Figure 1. Thank you for this note.

- Figure 6a looks confusing - the background signal in the green channel is very different between control and experimental groups. Were representative images taken with different microscope settings?

The representative images were taken with the same microscope power settings, but were adjusted in brightness/contrast within FIJI for clarity in the Figure – we apologize that this was misleading in any way and thank the reviewer for their feedback. Further, based on the collective discussion from all reviewers on the completeness of our GABA quantification and its implications, we have decided to remove this figure and perform more substantive analysis of this E/I imbalance in future work.

- Typo mChe;try

This typo was fixed

- "During this contextual... mice in the 6- and 14- month groups..." Isn't it 3- and 11- month respectively at the time of fear conditioning? Throughout the manuscript, this point was written very confusingly.

Yes, we thank the reviewer for pointing this out. It has been corrected to 3- and 11-month old mice at the timing of fear conditioning and clarified throughout the manuscript where applicable.

- "GABAergic eYFP fluorescence" Where does the eYFP come from? The methods state that GABA quantification is based on IHC staining.

Based on the collective discussion from all reviewers on the completeness of our GABA quantification and its implications, we have decided to remove this figure and perform more substantive analysis of this

E/I imbalance in future work. We discuss this E/I balance not being directly assessed in the Limitations & Future Directions section of our Discussion, noting the importance of detailed quantification of both excitatory and inhibitory markers within the hippocampus.

Reviewer #2 (Recommendations For The Authors):

(1) There is a full methods section ("Analysis of RNA-seq data") that mostly describes RNA-seq analysis that seemingly does not appear in the paper. This section should be reviewed.

We have included this portion of the methods that explain the previous workflow from Shpokayte et al., 2022 where this dataset was generated and this has been noted in the “Analysis of RNA-seq data” section of the methods.

(2) Figure 6: GABA staining should be more critically analyzed, as discussed above, and validated with another GABA antibody for rigor. From the representative images provided in Figure 6, it looks possibly as though the hM3Dq images were simply not fully in the focal plane when being imaged or were over-washed, as DAPI staining also appears to be lower in these images.

Based on the collective discussion from all reviewers on the completeness of our GABA quantification and its implications, we have decided to remove this figure and perform more substantive analysis of this E/I imbalance in future work. Specifically, it will be necessary to rigorously investigate both excitatory and inhibitory markers within this region to ensure these claims are substantiated. Thank you for this suggestion.

(3) The first claim that human GABAergic interneurons cause rumination is uncited. (Page 19, first sentence beginning with: "Evidence from human studies suggests...").

Based on the collective discussion from all reviewers on the completeness of our GABA quantification and its implications, we have decided to remove this figure and perform more substantive analysis of this E/I imbalance in future work. Apologies for the lack of citation in-text, the proper citation for this finding is Schmitz et al, 2017.

(4) Gene names throughout the manuscript and figure are written in the wrong format for mice (eg: Page 13, second line: SPP1, TTR, and C1QB1 instead of Spp1, Ttr, C1qb1).

This was corrected throughout the manuscript.

(5) Tense on Page 15 third sentence of the second paragraph: "...spatial working memory was assessed...".

This was corrected throughout the manuscript.

(6) Supplemental Figure 1 would benefit from normalization of the NeuN+ cell counts. The inclusion of an excitatory and inhibitory neuron marker in this figure might benefit the argument that there is a change in the excitation/inhibition of the hippocampus - as the numbers of excitatory neurons outweigh the numbers of inhibitory neurons that would be assayed here.

In an effort to normalize the NeuN+ cell counts, for each of our ROIs (6-8 single tiles for each brain region (DG, vCA1, vSub) x 3-5 coronal slices = ~18 single tiles per mouse x 3-4 mice) we captured a 300 x 300 micrometer, single-tile z-stack at 20x magnification. These ROIs were matched for dimensions and brain regions across all groups for each hippocampal subregion quantified. We initially proposed to normalize these NeuN counts over DAPI, but because DAPI includes all nuclei (microglia, oligodendrocytes, astrocytes and neurons), we weren’t sure this was the most optimal tool. We do agree that further quantification of excitatory and inhibitory cell markers would be vital to more concrete interpretation of our findings and we have added this to our Limitations & Future Work section of the Discussion.

Reviewer #3 (Recommendations For The Authors):

(1) The DOX tagging window lacks temporal precision. I suggest the authors note this as a limitation.

We thank the reviewer for noting this, and we have added this limitation to the Methods section with the context of the 24-48 hour DOX window being longer than other methods like TRAP.

(2) Is there a homeostatic response to chronic engram stimulation? That is, is DCZ as effective in increasing neuronal excitability on day 90 as it is on day 1. This could be addressed with electrophysiology, or with IEG induction. Alternatively, the authors could refer to previous literature-- for example, Xia et al (2017) eLife-- that examined whether there was any blunting of the effects of DREADD ligands after sustained delivery via drinking water. There, of course, may be other papers as well.

As noted by the reviewer, it is important to determine if DCZ maintains its effects on neuronal excitability throughout the 3 month administration period. To address this, previous work has shown that CNO administration in drinking water over one month consistently inhibited hM4Di+ neurons without altering baseline neuronal excitability as measured by firing rate and potassium currents (Xia et al, 2017). Although this is only for one month, it is administered via the same oral route as our DCZ protocol and suggests that at least for that amount of time we are likely producing consistent effects. In our reply above to Reviewer #1’s comment, we also note that even if DCZ is only having an effect for one month, rather than 3 months, we are still observing enduring changes that resulted from this short-term disturbance.

(3) Please double check there is no group effect on weight in 6-month-old mice in Figure 2C.

Two-way RM ANOVA showed no main effect of Group within the 6-month-old control and hM3Dq groups.

Group: F(1,17) = 1.361, p=0.2594.

(4) The shock intensity is much higher than is typical for fear conditioning studies in mice. Why was this the case?

Yes, we do agree that this shock intensity is on the higher side of typical paradigms in mice, however, our lab has utilized 0.75mA to 1.5mA intensity foot shocks for contextual fear conditioning in the past (Suthard & Senne et al, 2023; 2024; Dorst & Senne et al, 2023; Grella et al., 2022; Finkelstein et al., 2022) and we maintained this protocol for internal consistency. However, it would be interesting to systematically investigate how differing intensities of foot shock, subsequent tagging of this ensemble and reactivation would uniquely impact behavioral state acutely and chronically in mice.

(5) Remote freezing is very low. The authors should comment on this-- perhaps repeated testing has led to some extinction?

A reviewer above suggested a similar phenomenon may be occuring, specifically fear attenuation as a result of chronic stimulation. They referenced previous work from Khalaf et al. 2018, where they reactivated a recall-induced ensemble, while we reactivated an ensemble tagged during encoding. We expand upon this work in light of our findings within the Limitations & Future Work section of our Discussion. However, we do appreciate the lower levels of freezing observed in remote recall and sought out other literature to understand the typical range of remote freezing levels. One thing that we note is that our remote recall is occurring 3 months after conditioning, which is much longer than typical 14-28 day protocols. However, we find that freezing levels at remote timepoints from 21-45 days results in contextual freezing levels of between 20-50% approximately (Kol et al., 2020), as well as 40-75% approximately in a variety of 28 day remote recall experiments (Lee et al., 2023). This information, together with our current experimental protocol demonstrates a wide range of remote freezing levels that may depend heavily on the foot shock intensity, duration of days after conditioning, and animal variability.

(6) "mice display increased freezing with age": please add a reference.

Apologies, we missed the citation for that claim and it has been added in-text and in the references list (Shoji & Miyakawa, 2019).

(7) Related to the low freezing levels for remote memory, why is generalization minimal? Many studies have shown that there is a time-dependent emergence of generalized fear, yet here this is not seen. Is it linked to extinction (as above)? Or genetic background?

Previous work has shown that rats receiving multiple foot shocks during conditioning displayed a time-dependent generalization of context memory, while those receiving less shocks did not (Poulos et al., 2016), as the reviewer noted in their comment. In our current study, we observe low levels of generalization in all of our groups compared to freezing levels displayed in the conditioned context at the remote timepoint, in opposition to this time-dependent enhancement of generalization. It is possible that the genetic background of our C57BL/6J mice compared to the Long-Evans rat strain in this previous work accounts for some of this difference. In addition, it is possible that the longer duration of time (3 months) compared to their remote timepoint (28 days) resulted in time-dependent decrease in generalization that decreases with greater durations of time from original conditioning. As noted above, it is indeed plausible that the reactivation of a contextual fear ensemble over time is attenuating freezing levels for both the original and similar contexts (Khalaf et al, 2018). We discuss the differences in our study and this 2018 work more comprehensively above.

(8) Morphological phenotypes of astrocytes/microglia. Would be great to do some transcriptomic profiling of microglia/astrocytes to couple with the morphological characterization (but appreciate this is beyond the scope of current work).

We thank the reviewer this suggestion, we agree that would be an incredibly informative future experiment and have added this to our Limitations & Future Experiments section of the Discussion.

(9) The authors could consider including a limitations section in their discussion which discusses potential future directions for this work:

- causal experiments.

- E/I balance is not assessed directly (interestingly, in this regard, expanded engrams are linked to increased generalization [e.g., Ramsaran et al 2023]).

Thank you for this suggestion, we have added a Limitations & Future Directions section to our Discussion and have expanded upon these suggested points.

(10) For Figure 10, consider adding an experimental design/timeline.

We are making the assumption that the reviewer meant Figure 1 instead of Figure 10 here, but note that there is a description of the viral expression duration (D0-D10), followed by an off Dox period of 48 hours (D10-D12), with subsequent engram tagging of a negative (foot shock) or positive (male-to-female exposure) on D12. In our experiments (Shpokayte et al., 2022), Dox was administered for 24 hours (D12-D13), which was followed by sacrificing the animal for cell suspension and sequencing of the positive and negative engram populations. This figure also shows the viral strategy for the Tet-tag system (Figure 1A), as well as representative viral expression in vHPC (Figure 1B). We are happy to add additional experimental design/timeline information to this figure that would be helpful to the reviewer.

Author response:

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

In light of some reviewer comments requesting more clarity on the relationship between our model and prior theoretical studies of systems consolidation, we propose a modification to the title of our manuscript: “Selective consolidation of learning and memory via recall-gated plasticity.” We believe this title better reflects the key distinguishing feature of our model, that it selectively consolidates only a subset of memories, and also highlights the model’s applicability to task learning as well as memory storage.

Major comments:

Reviewer #3’s primary concern with the paper is the following: “The main weakness of the paper is the equation of recall strength with the synaptic changes brought about by the presentation of a stimulus. In most models of learning, synaptic changes are driven by an error signal and hence cease once the task has been learned. The suggested consolidation mechanism would stop at that point, although recall is still fine. The authors should discuss other notions of recall strength that would allow memory consolidation to continue after the initial learning phase.”

We thank the reviewer for drawing attention to this issue, which primarily results from a poor that memories should be interpreted as actual synaptic weight updates,∆𝑤and thus in the context choice of notation on our part. Our decision to denote memories as gives the impression of supervised learning would go to zero when the task is learned. However, in the formalism of our model, memories are in fact better interpreted as target values of synaptic weights, and the synaptic model/plasticity rule is responsible for converting these target values into synaptic weight updates. We were unclear on this point in our initial submission, because our paper primarily considers binary synaptic weights, where target synaptic weights have a one-to-one correspondence with candidate synaptic weight updates. We have updated the paper to use w* to refer to memories, which we hope resolves this confusion, and have updated our introduction to the term “memory” to reflect their interpretation as target synaptic weight values. We have also updated the paper’s language to more clearly disambiguate between the “learning rule,” which determines how the memory vector (target synaptic weight vectors) are derived from task variables, and the “plasticity rule,” which governs how these are translated into actual synaptic weight updates. We acknowledge that our manuscript still does not explicitly consider a plasticity rule that is sensitive to continuous error error signals, as our analysis is restricted to binary weights. However, we believe that the updated notation and exposition makes it more clear that our model could be applied in such a case.

Reviewer #1 brought up that our framework cannot capture “single-shot learning, for example, under fear conditioning or if a presented stimulus is astonishing.” Reviewer #2 raised a related question of how our model “relates to the opposite more intuitive idea, that novel surprising experiences should be stored in memory, as the familiar ones are presumably already stored.”

We agree that the built-in inability to consolidate memories after a single experience is a limitation of our model, and that extreme novelty is one factor (among others, such as salience or reward) that might incentivize one-shot consolidation. We have added a comment to the discussion to acknowledge these points (added text in bold): “ Moreover, in real neural circuits, additional factors besides recall, such as reward or salience, are likely to influence consolidation as well. For instance, a sufficiently salient event should be stored in long-term memory even if encountered only once. Furthermore, while in our model familiarity drives consolidation, certain forms of novelty may also incentivize consolidation, raising the prospect of a non-monotonic relationship between consolidation probability and familiarity.” We agree that future work should address the combined influence of recall (as in our model) and other factors on the propensity to consolidate a memory.

Reviewer #1 requested, “a comparison/discussion of the wide range of models on synaptic tagging for consolidation by various types of signals. Notably, studies from Wulfram Gerstner's group (e.g., Brea, J., Clayton, N. S., & Gerstner, W. (2023). Computational models of episodic-like memory in food-caching birds. Nature Communications, 14(1); and studies on surprise).”

We thank the reviewer for the reference, which we have added to the manuscript. The model of Brea et al.(2023) is similar to that of Roxin & Fusi (2013), in that consolidation consists of “copying” synaptic weights from one population to another. As a result, just like the model of Roxin & Fusi (2013), this model does not provide the benefit that our model offers in the context of consolidating repeatedly recurring memories. However, the model of Brea et al. does have other interesting properties – for instance, it affords the ability to decode the age of a memory, which our model does not. We have added a comment on this point in the subsection of the Discussion tilted “Other models of systems consolidation.”

Reviewer #2 noted, “While the article extensively discusses the strengths and advantages of the recall-gated consolidation model, it provides a limited discussion of potential limitations or shortcomings of the model, such as the missing feature of generalization, which is part of previous consolidation models. The model is not compared to other consolidation models in terms of performance and how much it increases the signal-to-noise ratio.”

We agree that our work does not consider the notion of generalization and associated changes to representational geometry that accompany consolidation, which is the focus of many other studies on consolidation. We have further highlighted this limitation in the discussion. Regarding the comparison to other models, this is a tricky point as the desiderata we emphasize in this study (the ability to recall memories that are intermittently reinforced) is not the focus of other studies. Indeed, our focus is primarily on the ability of systems consolidation to be selective in which memories are consolidated, which is somewhat orthogonal to the focus of many other theoretical studies of consolidation. We have updated some wording in the introduction to emphasize this focus.

Additional comments made by reviewer #1

Reviewer #1 pointed out issues in the clarity of Fig. 2A. We have added substantial clarifying text to the figure caption.

Reviewer #1 pointed out lack of clarity in our introduction to the terms “reliability” and “reinforcement.” We have now made it more clear what we mean by these terms the first time they are used.

We have updated our definition of “recall” to use the term “recall factor,” which is how we refer to it subsequently in the paper.

We have made explicit in the main text our simplifying assumption that memories are mean-centered.

We have made consistent our use of “forgetting curve” and “memory trace”.

Additional comments made by reviewer #2

We have added a comment in the discussion acknowledging alternative interpretations of the result of Terada et al. (2021)

We have significantly expanded the discussion of findings about the mushroom body to make it accessible to readers who do not specialize in this area. We hope this clarifies the nature of the experimental finding, which uncovered a circuit that performs a strikingly clean implementation of our model.

The reviewer expresses concern that the songbird study (Tachibana et al., 2022) does not provide direct evidence for consolidation being gated by familiarity of patterns of activity. Indeed, the experimental finding is one-step removed from the direct predictions of our model. That said, the finding – that the rate of consolidation increases with performance – is highly nontrivial, and is predicted by our model when applied to reinforcement learning tasks. We have added a comment to the discussion acknowledging that this experimental support for our model is behavioral and not mechanistic.

We do not regard it as completely trivial that the parallel LTM model performs roughly the same as the STM model, since a slower learning rate can achieve a higher SNR (as in Fig. 2C). Nevertheless we have added wording to the main text around Fig. 4B to note that the result is not too surprising.

We have added a sentence that clarifies the goal / question of our paper earlier on in the introduction.

We have updated Figure 3 by labeling the key components of the schematics and adding more detail to the legend, as suggested by the reviewer. We also reordered the figure panels as suggested.

Additional comments made by reviewer #3:

We have clarified in the main text that Fig. 2C and all results from Fig. 4 onward are derived from an ideal observer model (which we also more clearly define).

We have now emphasized in the main text that the derivations of the recall factors for specific learning rules are derived in the Supplementary Information.

We have highlighted more clearly in the main text that the recall factors associated with specific learning rules may correspond to other notions that do not intuitively correspond to “recall,” and have added a pointer to Fig. 3A where these interpretations are spelled out.

We have added references corresponding to the types of learning rules we consider.

The cutoffs / piecewise-looking behavior of plots in Fig. 4 are primarily the result of finite N, which limits the maximum SNR of the system, rather than coarse sampling of parameter values.

Thank you for pointing out the error in the legend in Fig. 5D (also affected Supp Fig. S7/S8), which is now fixed.

The reference to the nonexistence panel Fig. 5G has been removed.

As the reviewer points out, the use of a binary action output in our reinforcement learning task renders it quite similar to the supervised learning task, making the example less compelling. In the revised manuscript we have updated the RL simulation to use three actions. Note also that in our original submission the network outputs represented action probabilities directly (which is straightforward to do for binary actions, but not for more than two available actions). In order to parameterize a policy when more than two actions are available, we sample actions using a softmax policy, as is more standard in the field and as the reviewer suggested. The associated recall factor is still a product of reward and a “confidence factor,” and the confidence factor is still the value of the network output in the unit corresponding to the chosen action, but in the updated implementation this factor is equal to , similar (though with a sign difference) to the reviewer’s suggestion. We believe these updates make our RL implementation and simulation more compelling, as it allows them to be applied to tasks with arbitrary numbers of actions.

Additional minor comments

The reviewers made a number of other specific line-by-line wording suggestions, typo corrections,

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

The mechanisms of how axonal projections find their correct target requires the interplay of signalling pathways, and cell adhesion that act over short and long distances. The current study aims to use the small ventral lateral clock neurons (s-LNvs) of the Drosophila clock circuit as a model to study axon projections. These neurons are born during embryonic stages and are part of the core of the clock circuit in the larval brain. Moreover, these neurons are maintained through metamorphosis and become part of the adult clock circuit. The authors use the axon length by means of anti-Pdf antibody or Pdf>GFP as a read-out for the axonal length. Using ablation of the MB- the overall target region of the s-LNvs, the authors find defects in the projections. Next, by using Dscam mutants or knock-down they observe defects in the projections. Manipulations by the DNs - another group of clock neurons- can induce defects in the s-LNvs axonal form, suggesting an active role of these neurons in the morphology of the s-LNvs.

Strengths:

The use of Drosophila genetics and a specific neural type allows targeted manipulations with high precision.

Proposing a new model for a small group of neurons for axonal projections allows us to explore the mechanism with high precision.

Weaknesses:

It is unclear how far the proposed model can be seen as developmental.

The study of changes in fully differentiated and functioning neurons may affect the interpretation of the findings.

We appreciate the reviewer's feedback on the strengths and weaknesses of our study.

We acknowledge the strengths of our research, particularly the precision afforded by using Drosophila genetics and a specific neural type for targeted manipulations, as well as the proposal of a new model for studying axonal projections in a small group of neurons.

We understand the concerns about the developmental aspects of our proposed model and the use of Pdf-GAL4 >GFP as a read-out for the axonal length (revised manuscript Figure 1--figure supplement 1). However, even with the use of Clk856-GAL4 that began to be expressed at the embryonic stage (revised manuscript Figure 3--figure supplement 1) to suppress Dscam expression, the initial segment of the dorsal projection of s-LNvs (the vertical part) remained unaffected. Instead, the projection distance is severely shortened towards the midline, and this defect persists until the adult stage. It is for this reason that we delineate the dorsal projections of s-LNvs into two distinct phases: the vertical and horizontal parts, rather than a mere expansion in correspondence with the development of the larval brain.

Thank you for your valuable feedback, and we have incorporated these considerations into our revised manuscript to enhance the clarity and depth of our research.

Reviewer #2 (Public Review):

Summary:

The paper from Li et al shows a mechanism by which axons can change direction during development. They use the sLNv neurons as a model. They find that the appearance of a new group of neurons (DNs) during post-embryonic proliferation secretes netrins and repels horizontally towards the midline, the axonal tip of the LNvs.

Strengths:

The experiments are well done and the results are conclusive.

Weaknesses:

The novelty of the study is overstated, and the background is understated. Both things need to be revised.

We appreciate your acknowledgment that the experiments were well-executed and the results conclusive. This validation reinforces the robustness of our findings.

We take note of your feedback regarding the novelty of the study being overstated and the background being understated. While axonal projections navigate without distinct landmarks, like the midline or the layers, columns, and segments, they pose more challenges and uncertainties. As highlighted, our key contribution lies in elucidating how axonal projections without clear landmarks are guided, with our research demonstrating how a newly formed cluster of cells at a specific time and location provides the necessary guidance cues for axons.

We value your insights, and we have carefully addressed these points in our manuscript revision to improve the overall quality and presentation of our research.

Recommendations For The Authors:

Reviewer #1 (Recommendations For The Authors):

The overall idea of using the s-LNvs as a model is indeed intriguing. There are genetic tools available to tackle these cells with great precision.

However, based on the stage at which these cells are investigated raises some issues, that I feel are critical to be addressed.

These neurons develop their axonal projections during embryogenesis and are fully functioning when the larvae hatch, thus to investigate axonal pathfinding one would have to address embryonic development.

The larval brain indeed continues to grow during larval life, however extensive work from the Hartenstein lab, Truman lab, and others have shown that the secondary (larval born) neurons do not yet wire into the brain, but stall their axonal projections.

It is thus quite unclear, what the authors are actually studying.

One interpretation could be that the authors observe changes in axon length due to morphological changes in the brain. Indeed, the fact that the MB expands the anatomy of the surrounding neuropil changes too.

Moreover, it is unclear when exactly the Pdf-Gal4 (and other drivers) are active, thus how far (embryonic) development of s-LNvs is affected, or if it's all happening in the differentiated, functioning neuron. (Gal4 temporal delay and dynamics during embryonic development may further complicate the issue). As far as I am aware the MB drivers might already be active during embryonic stages.

Since the raised issue is quite fundamental, I am not sure what might be the best and most productive fashion to address this.

Eg. either to completely re-focus the topic on "neural morphology maintenance" or to study the actual development of these cells.

We thank the reviewer for the detailed and insightful feedback on our study. We have tested whether Pdf-Gal4 could effectively label s-LNv, and tracked the s-LNv projection in the early stage after larvae hatching. We did not observe the PDF antibody staining signal and the GFP signal driven by Pdf-GAL4 when the larvae were newly hatched. At 2-4 hours ALH, PDF signals were primarily concentrated at the end of axons, while GFP signals were mainly concentrated at the cell body. Helfrich-Förster initially detected immunoreactivity for PDF in the brains approximately 4-5 hours ALH. The GFP signal expressed by Pdf-GAL4 driver does have signal delay. However, at 8 hours ALH, the GFP signal strongly co-localized with the PDF signal within the axons (see revised manuscript lines 98-101) (Figure 1—figure supplement 1).

Based on previous research findings and our staining of Clk856-GAL4 >GFP, it is indeed confirmed that the dorsal projection of s-LNvs in Drosophila is formed during the embryonic stage (Figure 3—figure supplement 1). The s-LNvs in first-instar larval Drosophila are capable of detecting signal output and may play a role in regulating certain behaviors. Our selection of tools for characterizing the projection pattern of s-LNv was not optimal, leading us to overlook the crucial detail that the projection had already formed during its embryonic stage.

However, even when employing Clk856-GAL4 to suppress Dscam expression from the embryonic stage, the initial segment of the dorsal projection of s-LNvs (the vertical part) remains unaffected. Instead, the projection distance is severely shortened towards the midline, and this defect persists until the adult stage. It is for this reason that we delineate the dorsal projections of s-LNvs into two distinct phases: the vertical and horizontal parts, rather than a mere expansion in correspondence with the development of the larval brain.

From the results searched in the Virtual Fly Brain (VFB) database (https://www.virtualflybrain.org/), it is clear that the neurons that form synaptic connections with s-LNvs at the adult stage are essentially completely different from the neurons that are associated with them at the L1 larval stage. Thus, most neurons that form synapses with s-LNvs in the early larvae either cease to exist after metamorphosis or assume other roles in the adult stage. Similar to the scenario where Cajal-Retzius cells and GABAergic interneurons establish transient synaptic connections with entorhinal axons and commissural axons, respectively, these cells form a transient circuit with presynaptic targets and subsequently undergo cell death during development. In our model, the neurons that synapse with s-LNvs in early development serve as "placeholders," offering positive or negative cues to guide the axonal targeting of s-LNvs towards their ultimate destination.

Thank you again for your valuable feedback, and we have incorporated these considerations into our revised manuscript to enhance the clarity and depth of our research.

Reviewer #2 (Recommendations For The Authors):

Major:

In the introduction too many revisions are cited and very few actual research papers. This should be corrected and the most significant papers in the field should be cited. For example, there is no reference to the pioneering work from the Christine Holt lab or the first paper looking at axon guidance and guideposts by Klose and Bentley, Isbister et al 1999.

The introduction should encapsulate the actual knowledge based on actual research papers.

We acknowledge your concern regarding the citation of review papers rather than primary research papers in the introduction. Following your suggestion, we have revised the introduction section to incorporate references to relevant research papers.

In the introduction and discussion: The authors cite revisions where the signals that guide axons across different regions including turning are shown and they end up saying: "However, how the axons change their projection direction without well-defined landmarks is still unclear." I think the sentence should be changed. Many things are still not clear but this is not a good phrasing. Maybe they could focus on their temporal finding?

We appreciate the reviewer's feedback and insightful suggestions. We agree that emphasizing the temporal aspect is crucial in our study. However, we also recognize the significance of understanding the origin of signals that guide axonal reorientation at specific locations. While axonal projections navigating without distinct landmarks pose more challenges and uncertainties compared to those guided by prominent landmarks like the midline, our research demonstrates the crucial role of a specific cell population near turning points in providing accurate guidance cues to ensure precise axonal reorientation. We have revised our phrasing in the introduction and discussion to better reflect these key points (see revised manuscript lines 69-71 and 350-354). Thank you for highlighting the significance of focusing on our temporal findings and the complexities involved in studying axonal projection.

Many rather old papers have looked into the effect of repulsive guideposts to guide axon projections. In particular, I can think of the paper from Isbister et al. 1999 (DOI: 10.1242/dev.126.9.2007) that not only shows how semaphoring guides Ti axon projection but also shows how the pattern of expression of sema 2a changes during development to guide the correct projection. I really think that the novelty of the paper should be revised in light of the actual knowledge in the field.

We appreciate the reviewer's reference to the seminal work by Isbister et al. (1999) and the importance of guidepost cells in axon projection guidance, which we have already cited in our revised manuscript. It is crucial to recognize that segmented patterns such as the limb segment traversed by Ti1 neuron projections or neural circuits formed in a layer- or column-specific manner also serve as intrinsic "guideposts," offering valuable insights into axonal pathfinding processes. In our model, explicit guidance cues are lacking. As highlighted, our key contribution lies in elucidating how axonal projections without clear landmarks are guided, with our research demonstrating how a newly formed cluster of cells at a specific time and location provides the necessary guidance cues for axons (see revised manuscript lines 350-354). We have ensured that our revised manuscript reflects these insights and emphasizes the significance of studying axonal guidance in the absence of distinct guideposts. Thank you for underscoring these essential points, which enhance our understanding of axonal projection dynamics.

Minors:

Line 54, the authors start talking about floorplate at the end of a section on Drosophila. Please use “In vertebrates”, or “in invertebrates” or “in Drosophila” etc.. when needed to put things in context.

We thank the reviewer for this suggestion and have modified this sentence. Please refer to lines 62-63 of the revised manuscript.

Line 69: many factors change the axonal outgrowth. The authors are missing the paper from Fernandez et al. 2020, who have shown that unc5 the receptor of netrin induces the stalling for sLNvs projections before the turn. https://doi.org/10.1016/j.cub.2020.04.025

We thank the reviewer for this suggestion and have added this research article. Please refer to line 79 of the revised manuscript.

Line 99: "precisely at the pivotal juncture". It I hard to see how it was done in the figures shown. Can the authors add a small panel with neuronal staining showing this (please no HRP)?

For all figures, tee magenta is too strong and it is really hard to see the sLNvs projections. Can this be sorted, please?

We have depicted the pivotal juncture in the schematic diagram on the left side of Figure 1C. Additionally, we have included a separate column of images without HRP in Figure 1A. Moreover, we have modified the pseudo-color of HRP from magenta to blue to enhance the visualization of the s-LNv projection. The figure legends have also been correspondingly modified.

Line 407: Spatial position relationship between calyx and s-LNvs. OK107-GAL4 labels ... calyx and s-LNvs labeled by, which which.

We have modified it according to your suggestion. Please refer to lines 430-432 of the revised manuscript.

Line 137 typo RPRC

We thank the reviewer for noticing this mistake, which has now been corrected. Please refer to line 148-149 of the revised manuscript.

Section 158-164. the paper from Zhang et al 2019 needs to be cited since they have found the same effect of decreasing Dscam even if they didn't think about horizontal projection.

Thanks to the suggestion, we have included in the manuscript the phenotype observed by Zhang et al. (2019) upon knocking down Dscam1-L in adults. Please refer to lines 170-172 of the revised manuscript.

Line 176: typo senses (instead of sensor).

Thank you for pointing out our mistake. We have modified it according to your suggestion. Please refer to line 189 of the revised manuscript.

Line 193: more than Interesting it is Notable. Add "ubiquitus" knockdown.

Thank you for the suggestion. We have included the word "ubiquitus" to enhance the precision of the narrative. Please refer to line 206 of the revised manuscript.

Line 224: the pattern of expression of the crz cells is not visible where the projections of sLNvs are located. Are they in that region? Or further away?

We've changed the pseudo-color of HRP, and in the updated Figure 5- figure supplement 1, you can see the projection pattern of crz+ cells, positioned close to the end of the s-LNv axon terminal.

Line 243: applied? Do you mean "used"

Thank you for the suggestion. We have revised it at line 256.

Figure 5 Sup1: the schematic shows DNs proliferation that is not visible on the GFP image. Please comment.

We have modified the Figure 5 figure supplementary 1 for 120 h per-GAL4, Pdf-GAL80 >GFP expression pattern. Due to the strong GFP intensity in some DN neurons, there was a loss of GFP signal. Additionally, in Figure 6 figure supplementary 1, we have added co-localization images of DN and s-LNv at 72 h and 96 h. To better illustrate the co-localization information, we have shown only a portion of the layers in the right panel. We hope these additions clarify your concerns.

Line 251: cite Fernandez et al. 2020 with Purohit et al 2012.

We have modified it according to your suggestion. Please refer to line 264 of the revised manuscript.

Line 272: you have not shown synergistic effects because you have not modulated both pathways at the same time. You should talk about complementary.

We have modified it according to your suggestion at lines 25, 285, 439.

Author response:

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

Reviewer #1:

(1) Point for more elaborate discussion: Apparently the timescale of negative feedback signals is conserved between endothelial cell migration in vitro (with human cells) and endothelial migration during the formation of ISVs in zebrafish. What do you think might be an explanation for such conserved timescales? Are there certain processes within cytoskeletal tension build up that require this quantity of time to establish? Or does it relate to the time that is needed to begin to express the YAP/TAZ target genes that mediate feedback?

The underlying mechanisms responsible for the conserved timescale is a major direction that we continue to explore. Localization of YAP/TAZ to the nucleus is likely not rate-limiting. We showed previously that acute RhoA activation produced significant YAP/TAZ nuclear localization within minutes, while subsequent co-transcriptional activity aligned with the gene expression dynamics observed here (Berlew et al., 2021). We hypothesize that the dynamics of YAP/TAZdependent transcription and the translation of those target genes are rate-limiting for initial feedback loop completion (tic = 4 hours). This is supported by work from us and others in a variety of cell lines showing YAP/TAZ transcriptional responses take place during the first few hours after activation. (Franklin et al., 2020; Mason et al., 2019; Plouffe et al., 2018) While our data identify mediators of initial feedback loop completion, the molecular effectors that determine the timescale of new cytoskeletal equilibrium establishment (teq = 8 hours) remain unclear.

(2) Do you expect different timescales for slower endothelial migratory processes (e.g. for instance during fin vascular regeneration which takes days)?

We selected the ISV development model because it exhibits similar migratory kinetics to our previously-explored human ECFC migration in vitro. The comparable kinetics allowed us to study dynamics of the feedback loop in vivo on similar time scales, but we have not explored models featuring either slower or faster dynamics. 

It would be interesting to test how feedback dynamics are impacted in distinct endothelial migratory processes. Our data suggest that the feedback loop is necessary for persistent migration; however, YAP and TAZ respond to a diversity of upstream regulators in addition to mechanical signals, which might depend on the process of vascular morphogenesis. For example, after fin amputation, inflammation and tissue regeneration may impact the biochemical and mechanical environment experienced by the endothelium. Additionally, cells display different migratory behaviors in ISV morphogenesis compared to fin regeneration. During ISV formation, sprouting tip cells migrate dorsally through avascular tissue, followed by stalk cells. (Ellertsdóttir et al., 2010) In contrast, the fin vasculature regenerates by forming an intermediate vascular plexus, where some venous-derived endothelial cells migrate towards the sprouting front, while others migrate against it. (Xu et al., 2014) We are excited to study the role of this feedback loop in these different modes of neovessel formation in future studies.

(3) Is the ~4hrs and 8hrs feedback time window a general property or does it differ between specific endothelial cell types? In the veins the endothelial cells generate less stress fibers and adhesions compared to in the arteries. Does this mean that there might be a difference in the feedback time window, or does that mean that certain endothelial cell types may not have such YAP/TAZcontrolled feedback system?

Recent studies suggest that venous endothelial cells are the primary endothelial subtype responsible for blood vessel morphogenesis. (Lee et al., 2022, 2021; Xu et al., 2014) They are highly motile and mechanosensitive, migrating against blood flow. (Lee et al., 2022) The Huveneers group has shown that the actin cytoskeleton is differently organized in adult arteries and veins in response to biomechanical properties of its extracellular matrix, rather than intrinsic differences between arterial and venous cells. (van Geemen et al., 2014) This suggests that arterial and venous cells have distinct cytoskeletal setpoints due to mechanical cues in their environment (Price et al., 2021). We expect this to impact the degree of cytoskeletal remodeling and cell migration at equilibrium, rather than the kinetics of the feedback loop per se, though we have not yet tested this hypothesis. Testing these predictions on cytoskeletal setpoint stability and adaptation is a major direction that we continue to explore. 

(4) The experiments are based on perturbations to prove that transcriptional feedback is needed for endothelial migration. What would happen if the feedback systems is always switched on? An experiment to add might be to analyse the responsiveness of endothelial cells expressing constitutively active YAP/TAZ.

This is a problem that we are actively pursuing. Though the feedback system forms a coherent loop, we anticipate that the identity of the node of the loop selected for constitutive activation will influence the outcome, depending on whether that node is rate-limiting for feedback kinetics and the extent of intersection of that node with other signaling events in the cell. For example, we have observed that constitutive YAP activation drives profound changes to the transcriptional landscape including, but not limited to, RhoA signaling (Jones et al., 2023). We further anticipate that constitutive activation of feedback loop nodes may alter feedback dynamics, while dynamic or acute perturbation will be required to dissect these contributions in real time. For these reasons, ongoing work in the lab is pursuing these questions using optogenetic tools that enable precise spatial and temporal control (Berlew et al., 2021).   

(5) To investigate the role of YAP-mediated transcription in an accurate time-dependent manner the authors may consider using the recently developed optogenetic YAP translocation tool: https://doi.org/10.15252/embr.202154401

We are enthusiastic about the power of optogenetics to interrogate the nodes and timescales of this feedback system, and we are now funded to pursue this line of research. 

Reviewer #2:

The idea is intriguing, but it is not clear how the feedback actually works, so it is difficult to determine if the events needed could occur within 4 hrs. Specifically, it is not clear what gene changes initiated by YAP/TAZ translocation eventually lead to changes in Rho signaling and contractility. Much of the evidence to support the model is preliminary. Some of the data is consistent with the model, but alternative explanations of the data are not excluded. The fish washout data is quite interesting and does support the model. It is unclear how some of the in vitro data supports the model and excludes alternatives.

Major strengths:

The combination of in vitro and in vivo assessment provides evidence for timing in physiologically relevant contexts, and a rigorous quantification of outputs is provided. The idea of defining temporal aspects of the system is quite interesting.

Major weaknesses:

The evidence for a "loop" is not strong; rather, most of the data can also be interpreted as a linear increase in effect with time once a threshold is reached. Washout experiments are key to setting up a time window, yet these experiments are presented only for the fish model. A major technical challenge is that siRNA experiments take time to achieve depletion status, making precise timing of events on short time scales problematic. Also, Actinomycin D blocks most transcription so exposure for hours likely leads to secondary and tertiary effects and perhaps effects on viability. No RNA profiling is presented to validate proposed transcriptional changes.

We thank the reviewer for these helpful suggestions. We have expanded our explanation of the history and known mediators of the feedback loop in the introduction. We and, independently, the Huveneers group recently reported that human endothelial cells maintain cytoskeletal equilibrium for persistent motility through a YAP/TAZ-mediated feedback loop that modulates cytoskeletal tension. (Mason et al., 2019; van der Stoel et al., 2020) Because YAP and TAZ are activated by tension of the cytoskeleton (Dupont et al., 2011), suppression of cytoskeletal tension by YAP/TAZ transcriptional target genes constitutes a negative feedback loop (Fig. 1A). We described key components of this cell-intrinsic feedback loop, which acts as a control system to maintain cytoskeletal homeostasis for persistent motility via modulation of Rho-ROCK-myosin II activity. (Mason et al., 2019) Both we and the Huveneers group found that perturbation of genes and pathways regulated by YAP/TAZ mechanoactivation can functionally rescue motility in YAP/TAZ-depleted cells (e.g., RhoA/ROCK/myosin II, NUAK2, DLC1). (Mason et al., 2019; van der Stoel et al., 2020) We further showed previously that both YAP/TAZ depletion and acute YAP/TAZ-TEAD inhibition consistently increased stress fiber and FA maturation and arrested cell motility, accounting for these limitations of siRNA. (Mason et al., 2019)

Enduring limitations to the temporal, spatial, and cell-specific control of the genetic and pharmacologic methods have inspired us to initiate alternative approaches, which are the subject of ongoing efforts. Further research will be necessary in the zebrafish to determine the extent to which the observed migratory dynamics are driven by cytoskeletal arrest. 

To identify early YAP/TAZ-regulated transcriptional changes, we have added RNA profiling of control and YAP/TAZ depleted cells cultured on stiff matrices for four hours. Genes upregulated by YAP/TAZ depletion were enriched for Gene Ontology (GO) terms associated with Rho protein signal transduction, vascular development, cellular response to vascular endothelial growth factor (VEGF) stimulus, and endothelial cell migration (Fig. 9B). These data support a role for YAP and TAZ as negative feedback mediators that maintain cytoskeletal homeostasis for endothelial cell migration and vascular morphogenesis.  

Reviewer #3:

The authors used ECFC - endothelial colony forming cells (circulating endothelial cells that activate in response to vascular injury).

Q: Did the authors characterize these cells and made sure that they are truly endothelial cells - for example examine specific endothelial markers, arterial-venous identity markers & Notch signalling status, overall morphology etc prior to the start of the experiment. How were ECFC isolated from human individuals, are these "healthy" volunteers - any underlying CVD risk factors, cells from one patient or from pooled samples, what injury where these humans exposed to trigger the release of the ECPFs into the circulation, etc. The materials & methods on ECFC should be expanded.

Human umbilical cord blood-derived ECFCs were isolated at Indiana University School of Medicine and kindly provided by Dr Mervin Yoder. Cells were cultured as described by the Yoder group (Rapp et al., 2011) and our prior paper (Mason et al., 2019). We have expanded the materials and methods section to describe the source and characterization of these cells.

The authors suggest that loss of YAP/TAZ phenocopies actinomycin-D inhibition - "both transcription inhibition and YAP/TAZ depletion impaired polarization, and induced robust ventral stress fiber formation and peripheral focal adhesion maturation". However, the cell size of actinomycin-D treated cells (Fig. 1B, top right panel), differs from the endothelial cell size upon siYAP/TAZ (Fig. 1E, top right panel) - and vinculin staining seems more pronounced in actinomycin-D treated cells (B, bottom right) when compared to siYAP/TAZ group. Cell shape is defined by acto-myosin tension.

Q: Besides Fraction of focal adhesion >1um; focal adhesion number did the authors measure additional parameters related to cytoskeleton remodelling / focal adhesions that can substantiate their statement on similarity between loss of YAP/TAZ and actinomycin-D treatment. Would it be possible to make a more specific genetic intervention (besides YAP/TAZ) interfering with the focal adhesion pathway as opposed to the broad spectrum inhibitor actinomyocin-D.

Our previous paper (Mason et al., 2019) delineated the mechanistic relationships between YAP/TAZ signaling, focal adhesion turnover, actomyosin polymerization, and the intervening mechanisms of myosin regulation. Specifically, we demonstrated that YAP/TAZ regulate the myosin phosphatase kinase, NUAK2, and ARHGAP genes to mediate this feedback. Expanding on this work, the current study aimed to define the temporal kinetics of the cytoskeletal mechanotransductive feedback in vitro and in vivo. We used actinomycin-D and YAP/TAZ depletion to interrogate the role of transcriptional regulation and YAP/TAZ signaling, respectively. In this revision, we have added RNA profiling that identifies early YAP/TAZ-regulated transcriptional changes and further points to other molecular mediators of focal adhesions (e.g. TRIO, RHOB, THBS1) that will be the subjects of future studies.    

Q: Does the actinomycin-D treatment affect responsiveness to Vegf? induce apoptosis or reduce survival of the ECFC?

We have not looked specifically at the effect of actinomycin-D treatment on responsiveness to VEGF. However, actinomycin-D has been reported to reduce transcription of VEGF receptors (E et al., 2012). In contrast, we found that YAP/TAZ depletion upregulated GO terms associated with endothelial cell migration and response to VEGF stimulus (Fig. 9B), as well as receptors to angiogenic growth factors, including KDR and FLT4 (Fig. 9E). These results suggest YAP/TAZ depleted cells may be more sensitive to VEGF stimulation but remain nonmotile due to cytoskeletal arrest.

We showed previously that long-term treatment with actinomycin-D reduces ECFC survival (Mason et al., 2019).

Q: Which mechanism links ECM stiffness with endothelial surface area in the authors scenario. In zebrafish, activity of endothelial guanine exchange factor Trio specifically at endothelial cell junctions (Klems, Nat Comms, 2020) and endoglin in response to hemodynamic factors (Siekmann, Nat Cell Biol 2017) have been show to control EC shape/surface area - do these factors play a role in the scenario proposed by the authors.

Our new transcriptional profiling indicates both Trio and endoglin are regulated through YAP and TAZ in human ECFCs. We plan to follow up on these findings.

Q: The authors report that EC migrate faster on stiff substrate, and concomitantly these cells have a larger surface area. What is the physiological rationale behind these observations. Did the authors observe such behaviors in their zebrafish ISV model? How do these observations integrate with the tip - stalk cell shuffling model (Jakobsson & Gerhardt, Nat Cell Biol, 2011) and Notch activity in developing ISVs.

This question raises important distinctions between the mode of migration in ISV morphogenesis and endothelial cells adherent to substrates. Cells behave and respond to mechanical cues differently in 2D vs. 3D matrices. (LaValley and Reinhart-King, 2014) Additionally, the microenvironment in vivo is much more complex, combining numerous biochemical signals and changing mechanical properties. (Whisler et al., 2023) We are actively investigating the downstream targets of YAP/TAZ mechanotransduction and how that integrates with other pathways known to regulate vascular morphogenesis, such as Notch signaling. 

The authors examined the formation of arterial intersegmental vessels in the trunk of developing zebrafish embryos in vivo. They used a variety of pharmacological inhibitors of transcription and acto-myosin remodelling and linked the observed morphological changes in ISV morphogenesis with changes in endothelial cell motility.

Q: Reduced formation and dorsal extension of ISVs may have several reasons, including reduced EC migration and proliferation. The Tg(fl i1a:EGFP) reporter however is not the most suitable line to monitor migration of individual endothelial cells. Can the authors repeat the experiments in Tg(fl i1a:nEGFP); Tg(kdrl:HRAS-mCherry) double transgenics to visualize movement-migration of the individual endothelial cells and EC proliferation events, in the different treatment regimes.

So far, we have not tracked individual endothelial cells during ISV morphogenesis. We agree this is the best approach and are pursuing a similar technique for these experiments.

ISV formation is furthermore affected by Notch signalling status and a series of (repulsive) guidance cues.

Q: Does de novo blockade of gene expression with Actinomycin D affect Notch signalling status, expression of PlexinD - sFlt1, netrin1 or arterial-venous identify genes.

While we have not performed gene expression analysis under the Actinomycin D condition, Actinomycin D functions as a broad transcription inhibitor. We are currently pursuing the downstream targets of YAP/TAZ mechanotransduction in both ECFCs and zebrafish.

Remark: The authors may want to consider using the Tg(fl i1:LIFEACT-GFP) reporter for in vivo imaging of actin remodelling events.

We thank the reviewer for their helpful suggestion.

Remark: the authors report "As with broad transcription inhibition, in situ depletion of YAP and TAZ by RNAi arrested cell motility, illustrated here by live-migration sparklines over 10 hours: siControl: , siYAP/TAZ: (25 μm scale-bar: -)". Can the authors make a separate figure panel for this, how many cells were measured?

Please refer to our previous publication for the complete details on this data (Mason et al., 2019). We have added the citation in the text.

Remark: in the wash-out experiments, exposure to the inhibitors is not the same in the different scenarios - could it be that the longer exposure time induces "toxic" side effect that cannot be "washed out" when compared to the short treatment regimes?

This is a possible limitation of the pharmacological approach and have included it in the discussion section. We are currently exploring alternative approaches to interrogate the timescale of the feedback loop more precisely.  

References

Berlew EE, Kuznetsov IA, Yamada K, Bugaj LJ, Boerckel JD, Chow BY. 2021. Single-Component Optogenetic Tools for Inducible RhoA GTPase Signaling. Advanced Biology 5:2100810. doi:10.1002/adbi.202100810

Dupont S, Morsut L, Aragona M, Enzo E, Giulitti S, Cordenonsi M, Zanconato F, Le Digabel J,Forcato M, Bicciato S, Elvassore N, Piccolo S. 2011. Role of YAP/TAZ in mechanotransduction. Nature 474:179–183. doi:10.1038/nature10137

E G, Cao Y, Bhattacharya S, Dutta S, Wang E, Mukhopadhyay D. 2012. Endogenous Vascular Endothelial Growth Factor-A (VEGF-A) Maintains Endothelial Cell Homeostasis by Regulating VEGF Receptor-2 Transcription. J Biol Chem 287:3029–3041. doi:10.1074/jbc.M111.293985

Ellertsdóttir E, Lenard A, Blum Y, Krudewig A, Herwig L, Affolter M, Belting H-G. 2010. Vascular morphogenesis in the zebrafish embryo. Developmental Biology, Special Section: Morphogenesis 341:56–65. doi:10.1016/j.ydbio.2009.10.035

Franklin JM, Ghosh RP, Shi Q, Reddick MP, Liphardt JT. 2020. Concerted localization-resets precede YAP-dependent transcription. Nat Commun 11:4581. doi:10.1038/s41467-02018368-x

Jones DL, Hallström GF, Jiang X, Locke RC, Evans MK, Bonnevie ED, Srikumar A, Leahy TP, Nijsure MP, Boerckel JD, Mauck RL, Dyment NA. 2023. Mechanoepigenetic regulation of extracellular matrix homeostasis via Yap and Taz. Proceedings of the National Academy of Sciences 120:e2211947120. doi:10.1073/pnas.2211947120

LaValley DJ, Reinhart-King CA. 2014. Matrix stiffening in the formation of blood vessels. Advances in Regenerative Biology 1:25247. doi:10.3402/arb.v1.25247

Lee H-W, Shin JH, Simons M. 2022. Flow goes forward and cells step backward: endothelial migration. Exp Mol Med 54:711–719. doi:10.1038/s12276-022-00785-1

Lee H-W, Xu Y, He L, Choi W, Gonzalez D, Jin S-W, Simons M. 2021. Role of Venous Endothelial Cells in Developmental and Pathologic Angiogenesis. Circulation 144:1308–1322. doi:10.1161/CIRCULATIONAHA.121.054071

Mason DE, Collins JM, Dawahare JH, Nguyen TD, Lin Y, Voytik-Harbin SL, Zorlutuna P, Yoder MC, Boerckel JD. 2019. YAP and TAZ limit cytoskeletal and focal adhesion maturation to enable persistent cell motility. Journal of Cell Biology 218:1369–1389. doi:10.1083/jcb.201806065

Plouffe SW, Lin KC, Moore JL, Tan FE, Ma S, Ye Z, Qiu Y, Ren B, Guan K-L. 2018. The Hippo pathway effector proteins YAP and TAZ have both distinct and overlapping functions in the cell. J Biol Chem 293:11230–11240. doi:10.1074/jbc.RA118.002715

Price CC, Mathur J, Boerckel JD, Pathak A, Shenoy VB. 2021. Dynamic self-reinforcement of gene expression determines acquisition of cellular mechanical memory. Biophysical Journal 120:5074–5089. doi:10.1016/j.bpj.2021.10.006

Rapp BM, Saadatzedeh MR, Ofstein RH, Bhavsar JR, Tempel ZS, Moreno O, Morone P, Booth DA, Traktuev DO, Dalsing MC, Ingram DA, Yoder MC, March KL, Murphy MP. 2011. Resident Endothelial Progenitor Cells From Human Placenta Have Greater Vasculogenic Potential Than Circulating Endothelial Progenitor Cells From Umbilical Cord Blood. Cell Med 2:85–96. doi:10.3727/215517911X617888

Tammela T, Zarkada G, Nurmi H, Jakobsson L, Heinolainen K, Tvorogov D, Zheng W, Franco CA, Murtomäki A, Aranda E, Miura N, Ylä-Herttuala S, Fruttiger M, Mäkinen T, Eichmann A, Pollard JW, Gerhardt H, Alitalo K. 2011. VEGFR-3 controls tip to stalk conversion at vessel fusion sites by reinforcing Notch signalling. Nat Cell Biol 13:1202–1213. doi:10.1038/ncb2331

van der Stoel M, Schimmel L, Nawaz K, van Stalborch A-M, de Haan A, Klaus-Bergmann A, Valent ET, Koenis DS, van Nieuw Amerongen GP, de Vries CJ, de Waard V, Gloerich M, van Buul JD, Huveneers S. 2020. DLC1 is a direct target of activated YAP/TAZ that drives collective migration and sprouting angiogenesis. Journal of Cell Science 133:jcs239947. doi:10.1242/jcs.239947

van Geemen D, Smeets MWJ, van Stalborch A-MD, Woerdeman LAE, Daemen MJAP, Hordijk PL, Huveneers S. 2014. F-Actin–Anchored Focal Adhesions Distinguish Endothelial Phenotypes of Human Arteries and Veins. Arteriosclerosis, Thrombosis, and Vascular Biology 34:2059–2067. doi:10.1161/ATVBAHA.114.304180

Whisler J, Shahreza S, Schlegelmilch K, Ege N, Javanmardi Y, Malandrino A, Agrawal A, Fantin A, Serwinski B, Azizgolshani H, Park C, Shone V, Demuren OO, Del Rosario A, Butty VL, Holroyd N, Domart M-C, Hooper S, Szita N, Boyer LA, Walker-Samuel S, Djordjevic B, Sheridan GK, Collinson L, Calvo F, Ruhrberg C, Sahai E, Kamm R, Moeendarbary E. 2023. Emergent mechanical control of vascular morphogenesis. Science Advances 9:eadg9781. doi:10.1126/sciadv.adg9781

Xu C, Hasan SS, Schmidt I, Rocha SF, Pitulescu ME, Bussmann J, Meyen D, Raz E, Adams RH, Siekmann AF. 2014. Arteries are formed by vein-derived endothelial tip cells. Nat Commun 5:5758. doi:10.1038/ncomms6758

Author response:

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

Reviewer #1 (Recommendations For The Authors): 

Major points about revised manuscript 

(1) While I acknowledge that the Laccase2 vector is probably the best available in terms of its clean circRNA-expression potential, the authors still lack an estimation of the circRNA overexpression efficiency, specifically the circular-to-linear expression ratio. In their second rebuttal letter, the authors argue that they do not have the option to use another probe and that they are limited by the Backsplicing junction (BSJ)-specific one. I assume they mean that such a probe might only partially hybridize with the linear form and therefore give a poor or no signal in the Northern blot. However, in this referee's opinion, it is precisely because of this limitation that the authors should have used another probe against both the linear and circular RNAs to simultaneously and quantitatively detect both isoforms. This would have allowed them to reliably estimate a circular-to-linear ratio. Perhaps the linear isoform is indeed not expressed or is very low for this circRNA overexpression vector, but the probe used by the authors does not prove it. I think that this addition to the manuscript is not strictly necessary at this stage, but it would certainly improve the results.  

We fully agree with this recommendation. Our efforts to show this using northern blotting was unfortunately unsuccesful due to background signal. To accommodate the question about circ-to-linear ratio, we instead used an RT-qPCR strategy to measure the linear vs circRNA expression derived from the LaccasecircHIPK3 expression constructs/cell lines. To be able to compare obtained results from different amplicons, we measured primer efficiencies (using amplification standard curves – not shown) of two linear Laccase version amplicons and our divergent primers targeting circHIPK3, which were found to be directly comparable. Using these primer sets in RT-qPCR on the same RNA preparation (total cellular RNA) from the northern blot (Supplementary figure S5H) revealed a ~4 fold higher expression of circHIPK3 compared to linear precursor RNA (Supplementary Figure S5I). 

This demonstrates that the Laccase vector system efficiently produces circHIPK3 RNA as expected. 

The few changes to the manuscript (results section text and reference to Supplementary Figure S5I) has been highlighted in yellow. The materials and methods section and Table S1 have been modified to include description of RTqPCR and specific primers.

Author response:

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

Joint Public Review:

Xie et al. propose that the asymmetric segregation of the NuRD complex is regulated in a V-ATPase-dependent manner, and plays a crucial role in determining the differential expression of the apoptosis activator egl-1 and thus critical for the life/death fate decision.

Remaining concerns are the following:

The authors should provide the point-by-point response to the following issues. In particular, authors should provide clear reasoning as to why they did not address some of the following comments in the previous revisions. The next response should be directly answering to the following concerns.

(1) Discussion should be added regarding the criticism that NuRD asymmetric segregation is simply a result of daughter cell size asymmetry. It is perfectly fine that the NuRD asymmetry is due to the daughter cell size difference (still the nucleus within the bigger daughter would have more NuRD, which can determine the fate of daughter cells). Once the authors add this clarification, some criticisms about 'control' may become irrelevant.

We thank the reviewer for this suggestion. We will add the following text in the revised discussion on page 14, line 26:

“…We cannot rule out the possibility that NuRD asymmetric segregation results from daughter cell size asymmetry. According to this perspective, the nucleus in the larger daughter cell could possess more NuRD, potentially influencing the fate of the daughter cells. However, it is important to note that the nuclear protein histone or the MYST family histone acetyltransferase is equally segregated in daughter cells of different sizes.….”

(2) ZEN-4 is a kinesin that predominantly associates with the midzone microtubules and a midbody during mitosis. Given that midbodies can be asymmetrically inherited during cell division, ZEN-4 is not a good control for monitoring the inheritance of cytoplasmic proteins during asymmetric cell division. Other control proteins, such as a transcriptional factor that predominantly localizes in the cytoplasm during mitosis and enters into nucleus during interphase, are needed to clarify the concern.

We clarified the issue of ZEN-4 below:

The critique assumes that "midbodies can be asymmetrically inherited during cell division." However, this assumption does not apply to our study of Q cell asymmetric divisions. In our earlier research, we demonstrated that midbodies in Q cells are released post-division and subsequently engulfed by surrounding epithelial cells (Chai et al., Journal of Cell Biology, 2012). Moreover, we have shown that midbodies from the first cell division in C. elegans embryos are also released and engulfed by the P1 cell (Ou et al., Cell Research, 2013). Therefore, the notion of midbody asymmetric inheritance is irrelevant to this manuscript. Additionally, our manuscript already presents the example of the MYST family histone acetyltransferase, illustrating a nuclear protein that predominantly localizes in the cytoplasm during mitosis and symmetrically enters the nucleus during interphase.

As for pHluorin experiments, symmetric inheritance of GFP and mCherry is not an appropriate evidence to estimate the level of pHluorin during asymmmetric Q cell division. This issue remains unsolved.

We acknowledge the limitation of pHluorin in measuring the pH level in a living cell. Future studies could be performed to measure the dynamics of pH levels when advanced tools are available.

(3) Q-Q plot (quantile-quantile plot) in Figure S10 can be used for visually checking normality of the data, but it does not guarantee that the distribution of each sample is normal and has the standard deviation compared with the other samples. I recommend the authors to show the actual statistical comparison P-values for each case. The authors also need to show the number of replicate experiments for each figure panel.

We thank the reviewer for pointing this out. We will provide P-values for each case and the number of replicate experiments in the revised Figure 5-figure supplement 1 ( corresponding to Figure S10) and the figure legend.

The authors left inappropriate graphs in the revised manuscript. In Figure 3E, some error bars are disconnected and the other are stuck in the bars. In Figure S4C, LIN-53 in QR.a/p graph shows lines disconnected from error bars.

We thank the reviewer for pointing this out. We will correct these error bars.

I am bit confused with the error bars in Figure 2B. Each dot represents a fluorescent intensity ratio of either HDA-1 or LIN-53 between the two daughter cells in a single animal. Plots are shown with mean and SEM, but several samples (for example, the left end) exhibit the SEM error bar very close to a range of min and max. I might misunderstand this graph but am concerned that Figure 2B may contain some errors in representing these data sets. I would like to ask the authors to provide all values in a table format so that the reviewers could verify the statistical tests and graph representation.

We thank the reviewer for pointing this out. We apologize for the typo in Figure 2B figure legend. We will correct SEM to SD.

(4) The authors still do not provide evidence that the increase in sAnxV::GFP and Pegl-1gfp or the increase in H3K27ac at the egl-1 gene in hda-1(RNAi) and lin-53(RNAi) animals is not a consequence of global effects on development. Indeed, the images provided in Figure S7B demonstrate that there are global effects in these animals. no causal interactions have been demonstrated.

We cannot exclude the global effects and have discussed this issue in our previous manuscript on page 9, line 26:

“...Considering the pleiotropic phenotypes caused by loss of HDA-1, we cannot exclude the possibility that ectopic cell death might result from global changes in development, even though HDA-1 may directly contribute to the life-versus-death fate determination.”

(5) Figure 4: Due to the lack of appropriate controls for the co-IP experiment (Fig. 4), I remain unconvinced of the claim that the NuRD complex and V-ATPase specifically interact. Concerning the co-IP, the authors now mention that the co-IP was performed three times: "Assay was performed using three biological replicates. Three independent biological replicates of the experiment were conducted with similar results." However, the authors did not use ACT-4::GFP or GFP alone as controls for their co-IP as previously suggested. This is critical considering that the evidence for a specific HDA-1::GFP - V-ATPase interaction is rather weak (compare interactions between HDA-1::GFP and V-ATPase subunits in Fig 4B with those of HDA-1::GFP and subunits of NuRD in Fig S8B).

We conducted GFP pull-down experiments and MS spectrometric analysis for HDA-::GFP and ACT-4::GFP using identical protocols, yielding consistent results. We agree with the reviewer that in our Western blot, inclusion of ACT-4::GFP is a more effective negative control compared to empty beads.

(6) Based on Fig 5E, it appears that Bafilomycin treatment causes pleiotropic effects on animals (see differences in HDA-1::GFP signal in the three rows). The authors now state: "Although BafA1-mediated disruption of lysosomal pH homeostasis is recognized to elicit a wide array of intracellular abnormalities, we found no evidence of such pleiotropic effects at the organismal level with the dosage and duration of treatment employed in this study". However, the 'evidence' mentioned is not shown. It is critical that the authors provide this evidence.

We thank the Reviewer for pointing out this issue. We only checked the viability of the L1 larvae and morphology of animals at the organismal level with the BafA1 dosage and duration of treatment and did not notice any death of the animals and apparent abnormality in morphology (N > 20 for each treatment). However, as the reviewer pointed out, there can be some abnormalities at the cellular level. We thus revised this above description as the following, on page 11, line 27:

“…Although BafA1-mediated disruption of lysosomal pH homeostasis is recognized to elicit a wide array of intracellular abnormalities, we did not observe any larval deaths and apparent abnormality in morphology at the organismal level (N > 20 for each treatment) at the dose and duration of treatment employed in this study...”

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

eLife assessment

The authors propose that the asymmetric segregation of the NuRD complex in C. elegans is regulated in a V-ATPase-dependent manner, that this plays a crucial role in determining the differential expression of the apoptosis activator egl-1, and that it is therefore critical for the life/death fate decision in this species. If proven, the proposed model of the V-ATPase-NuRD-EGL-1-Apoptosis cascade would shed light onto the mechanisms underlying the regulation of apoptosis fate during asymmetric cell division, and stimulate further investigation into the intricate interplay between V-ATPase, NuRD, and epigenetic modifications. However, the strength of evidence for this is currently incomplete.

Public Review:

Xie et al. propose that the asymmetric segregation of the NuRD complex is regulated in a V-ATPase-dependent manner, and plays a crucial role in determining the differential expression of the apoptosis activator egl-1 and thus critical for the life/death fate decision.

While the model is very intriguing, the reviewers raised concerns regarding the rigor of the method. One issue is with statistics (either insufficient information or inadequate use of statistics), and second is the concern that the asymmetry observed may be caused by one cell dying (resulting in protein degradation, RNA degradation etc). We recommend that the authors address these issues.

We extend our sincere thanks to the Editors and Reviewers for their insightful comments on this study.

Major #1:

There are still many misleading statements/conclusions that are not rigorously tested or that are logically flawed. These issues must be thoroughly addressed for this manuscript to be solid.

(1) Asymmetry detected by scRNA seq vs. imaging may not represent the same phenomenon, thus should not be discussed as two supporting pieces of evidence for the authors' model, and importantly each method has its own flaw. First, for scRNA seq, when cells become already egl-1 positive, those cells may be already dying, and thus NuRD complex's transcripts' asymmetry may not have any significance. The data presented in FigS1D, E show that there are lots of genes (6487 out of 8624) that are decreased in dying cells. Thus, it is not convincing to claim that NuRD asymmetry is regulated by differential RNA amount.

We agree with the reviewer's comment. Indeed, scRNA-seq reveals phenomena different from those observed in protein imaging, and NuRD asymmetry may not be regulated by differential RNA levels. Seven years ago, when we started this project, NuRD asymmetry during asymmetric neuroblast division was unknown. We first found NuRD mRNA asymmetry using scRNA-seq and then NuRD protein asymmetry using fluorescence imaging. We have documented the whole process of discovering NuRD asymmetry, although the asymmetry of NuRD complex transcripts does not necessarily imply protein asymmetry. We have revised statements related to "NuRD asymmetry being regulated by differential RNA amounts" and discussed this issue in the revised manuscript on page 14, line 2:

" The transcript asymmetry detected by scRNA-seq may not correspond to the protein asymmetry detected by microscopic imaging. Our scRNA-seq data shows that 6487 out of 8624 genes were not detected in egl-1-positive cells, the putative apoptotic cells. Cells that are egl-1 positive may be undergoing apoptosis, rendering the asymmetry of NuRD complex transcripts insignificant in inferring protein asymmetry. Thus, the observed transcript asymmetry of the NuRD subunits between live and dead cells may be coincidental with NuRD protein asymmetry during asymmetric neuroblast division, rather than serving as a regulatory mechanism."

(2) Regarding NuRD protein's asymmetry, there are still multiple issues. Most likely explanation of their asymmetry is purely daughter size asymmetry. Because one cell is much bigger than the other (3 times larger), NuRD components, which are not chromatin associated, would be inherited to the bigger cell 3 times more than the smaller daughter. Then, upon nuclear envelope reformation, NuRD components will enter the nucleus, and there will be 3 times more NuRD components in the bigger daughter cell. It is possible that this is actually the underling mechanism to regulate gene expression differentially, but this possibility is not properly acknowledged. Currently, the authors use chromatin associated protein (Mys-1) as 'symmetric control', but this is not necessarily a fair comparison. For NuRD asymmetry to be meaningful, an example of protein is needed that is non-chromatin associated in mitosis, distributed to daughter cells proportional to daughter cell size, and re-enter nucleus after nuclear envelope formation to show symmetric distribution. And if daughter size asymmetry is the cause of NuRD asymmetry, other lineages that do not undergo apoptosis but exhibit daughter size asymmetry would also show NuRD asymmetry. The authors should comment on this (if such examples exist, it is fine in that in those cell types, NuRD asymmetry may be used for differential gene expression, not necessarily to induce cell death, but such comparison provides the explanation for NuRD asymmetry, and puts the authors finding in a better context).

For more than one decade, we have meticulously explored the relationship between protein asymmetry and cell size asymmetry during ACDs of Q cells. A notable example of even protein distribution is the cytokinetic kinesin ZEN-4, as documented in our 2012 publication in the Journal of Cell Biology (Chai et al., JCB, 2012). This study, primarily focusing on the fate of the midbody post-cell division, also showcased the dynamics of GFP-tagged ZEN-4 during ACDs of QR.a cells in movie S1. Intriguingly, beyond its role in the cytokinetic ring, we observed a uniform dispersal of ZEN-4 throughout the cytoplasm. Remarkably, following cell division, ZEN-4 transitions evenly into the nuclei of the daughter cells, a phenomenon with implications yet to be fully understood. One hypothesis is that ZEN-4's nuclear localization may prevent the formation of ectopic microtubule bundles in the cytosol during interphase. Below, we present a snapshot from our original movie, clearly showing the symmetrical distribution of ZEN-4 into the nuclei of the two daughter cells.

(3) For the analysis of protein asymmetry between two daughters in Fig S4C, the method of calibration is unclear, making it difficult to interpret the results.

In Figure S4C, we quantified the relative total fluorescence of the Q cell, with the quantification method illustrated in Figure S4A. To further clarify our quantification approach, we have updated Figure S4A and the "Live-Cell Imaging and Quantification" section in the Materials and Methods:

“…To determine the ratios of fluorescence intensities in the posterior to anterior half (P/A) of Q.a lineages or A/P of Q.p lineages, the cell in the mean intensity projection was divided into posterior and anterior halves. ImageJ software was used to measure the mean fluorescence intensities of two halves with background subtraction. The slide background's mean fluorescence intensity was measured in a region devoid of worm bodies. The background-subtracted mean fluorescence intensities of the two halves were divided to calculate the ratio. The same procedure was used to determine the fluorescence intensity ratios between two daughter cells. Total fluorescence intensity was the sum of the posterior and anterior fluorescence intensities or the sum of fluorescence intensities from two daughter cells (Figure S4A). …”

(4) As for pHluorin experiments, the authors were asked to test the changes in fluorescence observed are due to changes in pH or changes in the amount of pHluorin protein. They need to add a ratio-metric method in this manuscript. A brief mention to Page 12 line 12 is insufficient to clarify this issue.

We appreciate the concerns about potential changes in pH or pHluorin protein levels. While we cannot completely dismiss the impact of changes in the amount of pHluorin protein, it appears improbable that the asymmetry of pHluorin fluorescence is attributed to an asymmetric amount of pHluorin protein. This inference is supported by the observation that other fluorescent proteins, such as GFP or mCherry, did not exhibit any asymmetry during ACDs of Q cells. An example of GFP alone during the ACD of QL.p is illustrated in figure 5A from Ou and Vale, JCB, 2009. The fluorescence intensities in the large QL.pa cell and the small QL.aa are indistinguishable.

Major #2:

Some issues surrounding statistics must be resolved.

(1) Fig. 1FG, 2D, 3BDEG, 5BD and 6B used either one-sample t-test or unpaired two-tailed parametric t-test for statistical comparison. These t-tests require a verification of each sample fitting to a normal distribution. The authors need to describe a statistical test used to verify a normal distribution of each sample.

(2) Fig. 2D, 3D, and 3G have very small sample size (N=3-4, N=6, N=3, respectively), it is possible that a normal distribution cannot be verified. How can the authors justify the use of one-sample t-test and unpaired parametric t-test ?

(3) Statistical comparison in Fig. 2D and Fig. 6B should be re-assessed. For Fig. 2D, the authors need to compare the intensity ratio of HDA-1/LIN53 between sister cells dying within 35 min and those over 400 min. For Fig. 6B, they need to compare the intensity ratio of VHA-17 between DMSO- and BafA1- treated cells at the same time point after anaphase.

We appreciate the reviewer's advice on the statistical analysis of our data. In response, we performed normality tests on the datasets presented in Figures 1F, 1G, 3B, 5B, 5D, and 6B, all of which passed the tests (as demonstrated in Figure S10). We also acknowledge the reviewer's comment on the inadequate sample sizes in Figures 2D, 3D, 3E, and 3G for fitting a normal distribution. Therefore, we have revised our statistical analysis methods for these figures and updated both the figures and their legends. The revised statistical results support the primary conclusions of this study.

In response to the reviewer's observation regarding the small sample size in Figure 2D , which precluded normality verification, and the suggestion to compare sister cells that die within 35 minutes to those surviving over 400 minutes, we adapted our approach. We implemented the Kruskal-Wallis test to evaluate the differences among the groups. To assess the specific differences between each group and the 400 min MSpppaap group, we conducted the Dunn’s multiple comparisons test. The revised Figure 2D illustrates the updated statistical significance.

For Figure 3D, due to the small sample size precluding normality verification, we applied the Wilcoxon test with 1 as the theoretical median. The revised Figure 3D illustrates the updated statistical significance.

For Figure 3E, where the sample size also hindered normality verification, we conducted the Kruskal-Wallis test to evaluate the overall effect. Additionally, Dunn’s multiple comparisons test was utilized to examine the differences between groups. The revised Figure 3E illustrates the updated statistical significance.

For Figure 3G, the reviewer pointed out the small sample size and the limited statistical power due to having only three data points per group. To address this, we revised the figure to visually present each data point, aiming to more clearly illustrate the variation trends.

For Figure 6B, following the reviewer's suggestion, we compared the DMSO group directly with the Baf A1 group, updating Figure 6B to reflect this comparison as advised.

These adjustments have been made to ensure the statistical analyses are robust and appropriate given the sample sizes and to align with the reviewer's recommendations, enhancing the clarity and accuracy of our findings.

Recommendations for the authors:

We recommend using grey scale (instead of 'heatmap' representation) to show the protein distribution of interest. Heatmap does not help at all, because 'total protein amount per cell' (instead of signal intensity on each pixel) is what matters in the context of this paper. Heatmap presentation does not allow readers to integrate signal intensity with their eyes.

We thank the editor for pointing this out. We have changed heatmaps to inverted fluorescence images in grey scale.

Author response:

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

eLife assessment

The study presents a valuable tool for searching molecular dynamics simulation data, making such data sets accessible for open science. The authors provide convincing evidence that it is possible to identify useful molecular dynamics simulation data sets and their analysis can produce valuable information.

Public Reviews

Reviewer #1 (Public Review):

Summary:

Tiemann et al. have undertaken an original study on the availability of molecular dynamics (MD) simulation datasets across the Internet. There is a widespread belief that extensive, well-curated MD datasets would enable the development of novel classes of AI models for structural biology. However, currently, there is no standard for sharing MD datasets. As generating MD datasets is energy-intensive, it is also important to facilitate the reuse of MD datasets to minimize energy consumption. Developing a universally accepted standard for depositing and curating MD datasets is a huge undertaking. The study by Tiemann et al. will be very valuable in informing policy developments toward this goal.

Strengths:

The study presents an original approach to addressing a growing concern in the field. It is clear that adopting a more collaborative approach could significantly enhance the impact of MD simulations in modern molecular sciences.

The timing of the work is appropriate, given the current interest in developing AI models for describing biomolecular dynamics.

Weaknesses:

The study primarily focuses on one major MD engine (GROMACS), although this limitation is not significant considering the proof-of-concept nature of the study.

We thank the reviewer for his/her comments. Moving forward, our plan includes expanding this research to encompass other MD engines used in biomolecular simulations and materials sciences, such as NAMD, Charmm, Amber, LAMMPS, etc. However, this requires parsing associated files to supplement the sparse metadata generally available for the related datasets

Reviewer #2 (Public Review):

Summary:

Molecular dynamics (MD) data is deposited in public, non-specialist repositories. This work starts from the premise that these data are a valuable resource as they could be used by other researchers to extract additional insights from these simulations; it could also potentially be used as training data for ML/AI approaches. The problem is that mining these data is difficult because they are not easy to find and work with. The primary goal of the authors was to discover and index these difficult-to-find MD datasets, which they call the "dark matter of the MD universe" (in contrast to data sets held in specialist databases).

The authors developed a search strategy that avoided the use of ill-defined metadata but instead relied on the knowledge of the restricted set of file formats used in MD simulations as a true marker for the data they were looking for. Detection of MD data marked a data set as relevant with a follow-up indexing strategy of all associated content. This "explore-and-expand" strategy allowed the authors for the first time to provide a realistic census of the MD data in non-specialist repositories.

As a proof of principle, they analyzed a subset of the data (primarily related to simulations with the popular Gromacs MD package) to summarize the types of simulated systems (primarily biomolecular systems) and commonly used simulation settings.

Based on their experience they propose best practices for metadata provision to make MD data FAIR (findable, accessible, interoperable, reusable).

A prototype search engine that works on the indexed datasets is made publicly available. All data and code are made freely available as open source/open data.

Strengths:

The novel search strategy is based on relevant data to identify full datasets instead of relying on metadata and thus is likely to have many true positives and few false positives.

The paper provides a first glimpse at the potential hidden treasures of MD simulations and force field parametrizations of molecules.

Analysis of parameter settings of MD simulations from how researchers *actually* run simulations can provide valuable feedback to MD code developers for how to document/educate users. This approach is much better than analyzing what authors write in the Methods sections.

The authors make a prototype search engine available.

The guidelines for FAIR MD data are based on experience gained from trying to make sense of the data.

Weaknesses:

So far the work is a proof-of-concept that focuses on MD data produced by Gromacs (which was prevalent under all indexed and identified packages).

As discussed in the manuscript, some types of biomolecules are likely underrepresented because different communities have different preferences for force fields/MD codes (for example: carbohydrates with AMBER/GLYCAM using AMBER MD instead of Gromacs).

Materials sciences seem to be severely under-represented --- commonly used codes in this area such as LAMMPS are not even detected, and only very few examples could be identified. As it is, the paper primarily provides an insight into the *biomolecular* MD simulation world.

The authors succeed in providing a first realistic view on what MD data is available in public repositories. In particular, their explore-expand approach has the potential to be customized for all kinds of specialist simulation data, whereby specific artifacts are used as fiducial markers instead of metadata. The more detailed analysis is limited to Gromacs simulations and primarily biomolecular simulations (even though MD is also widely used in other fields such as the materials sciences). This restricted view may simply be correlated with the user community of Gromacs and hopefully, follow-up studies from this work will shed more light on this shortcoming.

The study quantified the number of trajectories currently held in structured databases as ~10k vs ~30k in generalist repositories. To go beyond the proof-of-principle analysis it would be interesting to analyze the data in specialist repositories in the same way as the one in the generalist ones, especially as there are now efforts underway to create a database for MD simulations (Grant 'Molecular dynamics simulation for biology and chemistry research' to establish MDDB' DOI 10.3030/101094651). One should note that structured databases do not invalidate the approach pioneered in this work; if anything they are orthogonal to each other and both will likely play an important role in growing the usefulness of MD simulations in the future.

We thank the reviewer for his/her comments. As mentioned to Reviewer 1, we intend to extend this work to other MD engines in the near future to go beyond Gromacs and even biomolecular simulations. Furthermore, as the value of accessing and indexing specialized MD databases such as MDDB, MemprotMD, GPCRmd, NMRLipids, ATLAS, and others has been mentioned by the reviewer, it is indeed one of our next steps to continue to expand the MDverse catalog of MD data. This indexing may also extend the visibility and widespreaded adoptability of these specific databases.

Reviewer #3 (Public Review):

Molecular dynamics (MD) simulations nowadays are an essential element of structural biology investigations, complementing experiments and aiding their interpretation by revealing transient processes or details (such as the effects of glycosylation on the SARS-CoV-2 spike protein, for example (Casalino et al. ACS Cent. Sci. 2020; 6, 10, 1722-1734 https://doi.org/10.1021/acscentsci.0c01056) that cannot be observed directly. MD simulations can allow for the calculation of thermodynamic, kinetic, and other properties and the prediction of biological or chemical activity. MD simulations can now serve as "computational assays" (Huggins et al. WIREs Comput Mol Sci. 2019; 9:e1393.

https://doi.org/10.1002/wcms.1393). Conceptually, MD simulations have played a crucial role in developing the understanding that the dynamics and conformational behaviour of biological macromolecules are essential to their function, and are shaped by evolution. Atomistic simulations range up to the billion atom scale with exascale resources (e.g. simulations of SARS-CoV-2 in a respiratory aerosol. Dommer et al. The International Journal of High Performance Computing Applications. 2023; 37:28-44. doi:10.1177/10943420221128233), while coarse-grained models allow simulations on even larger length- and timescales. Simulations with combined quantum mechanics/molecular mechanics (QM/MM) methods can investigate biochemical reactivity, and overcome limitations of empirical forcefields (Cui et al. J. Phys. Chem. B 2021; 125, 689 https://doi.org/10.1021/acs.jpcb.0c09898).

MD simulations generate large amounts of data (e.g. structures along the MD trajectory) and increasingly, e.g. because of funder mandates for open science, these data are deposited in publicly accessible repositories. There is real potential to learn from these data en masse, not only to understand biomolecular dynamics but also to explore methodological issues. Deposition of data is haphazard and lags far behind experimental structural biology, however, and it is also hard to answer the apparently simple question of "what is out there?". This is the question that Tiemann et al explore in this nice and important work, focusing on simulations run with the widely used GROMACS package. They develop a search strategy and identify almost 2,000 datasets from Zenodo, Figshare and Open Science Framework. This provides a very useful resource. For these datasets, they analyse features of the simulations (e.g. atomistic or coarse-grained), which provides a useful snapshot of current simulation approaches. The analysis is presented clearly and discussed insightfully. They also present a search engine to explore MD data, the MDverse data explorer, which promises to be a very useful tool.

As the authors state: "Eventually, front-end solutions such as the MDverse data explorer tool can evolve being more user-friendly by interfacing the structures and dynamics with interactive 3D molecular viewers". This will make MD simulations accessible to non-specialists and researchers in other areas. I would envisage that this will also include approaches using interactive virtual reality for an immersive exploration of structure and dynamics, and virtual collaboration (e.g. O'Connor et al., Sci. Adv.4, eaat2731 (2018). DOI:10.1126/sciadv.aat2731)

The need to share data effectively, and to compare simulations and test models, was illustrated clearly in the COVID-19 pandemic, which also demonstrated a willingness and commitment to data sharing across the international community (e.g. Amaro and Mulholland, J. Chem. Inf. Model. 2020, 60, 6, 2653-2656 https://doi.org/10.1021/acs.jcim.0c00319; Computing in Science & Engineering 2020, 22, 30-36 doi: 10.1109/MCSE.2020.3024155). There are important lessons to learn here, for simulations to be reproducible and reliable, for rapid testing, for exploiting data with machine learning, and for linking to data from other approaches. Tiemann et al. discuss how to develop these links, providing good perspectives and suggestions.

I agree completely with the statement of the authors that "Even if MD data represents only 1 % of the total volume of data stored in Zenodo, we believe it is our responsibility, as a community, to develop a better sharing and reuse of MD simulation files - and it will neither have to be particularly cumbersome nor expensive. To this end, we are proposing two solutions. First, improve practices for sharing and depositing MD data in data repositories. Second, improve the FAIRness of already available MD data notably by improving the quality of the current metadata."

This nicely states the challenge to the biomolecular simulation community. There is a clear need for standards for MD data and associated metadata. This will also help with the development of standards of best practice in simulations. The authors provide useful and detailed recommendations for MD metadata. These recommendations should contribute to discussions on the development of standards by researchers, funders, and publishers. Community organizations (such as CCP-BioSim and HECBioSim in the UK, BioExcel, CECAM, MolSSI, learned societies etc) have an important part to play in these developments, which are vital for the future of biomolecular simulation.

We thank the reviewer for his/her comments. Beyond the points mentioned to Reviewers 1 and 2, as the reviewer suggested, it would be of great interest to combine innovative and immersive approaches to visualize and possibly interact with the data collected. This is indeed more and more amenable thanks to technologies such as WebGL and programs such as Mol*, or even - as also pointed out by the reviewer - through virtual reality, for example with the mentioned Narupa framework or with the UnityMol software. For a comprehensive review on MD trajectory visualization and associated challenges, we refer to our recent review article https://doi.org/10.3389/fbinf.2024.1356659.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

Some minor text editing would improve the readability of the manuscript.

It would be very useful if the authors could share their perspectives on the best and most efficient approach to sharing datasets and code associated with a publication. My concern lies in the fact that Github, which is currently the dominant platform for sharing code, is not well-suited for hosting large MD datasets. As a result, researchers often need to adopt a workflow where code is shared on Github and datasets are stored elsewhere (e.g., Zenodo). While this is feasible, it adds extra work. Ideally, a transparent process could be developed to seamlessly share code and datasets linked to a study through a unified interface.

We thank the reviewer for this excellent suggestion. To our knowledge, there is yet no easy framework to jointly store and share code and data, linked to their scientific publication. Of course, code can be submitted to “generic” databases along with the data, but at the current state, those do not provide such useful features like collaborative work & track recording as done to the extent of GitHub.

Although GitHub is indeed a suitable platform to deposit code, we strongly advise researchers to archive their code in Software Heritage. In addition to preserving source code, Software Heritage provides a unique identifier called SWHID that unambiguously makes reference to a specific version of the source code.

So far, it is the responsibility of the scientific publication authors to link datasets and source codes (whether in GitHub or Software Heritage) in their paper, but also to make the reverse link from the data and code sharing platforms to the paper after publication.

As mentioned by the reviewer, a unified interface that could ease this process would significantly contribute to FAIR-ness in MD.

Reviewer #2 (Recommendations For The Authors):

L180: I am not aware that TRR files contain energy terms as stated here, my understanding was that EDR files primarily served that purpose.

“…available in one dataset. Interestingly, we found 1,406 .trr files, Which contain trajectory but also additional information such as velocities, energy of the system, etc’ While the file is especially useful in terms of reusability, the large size (can go up to several 100GB) limits its deposition in most…”

Indeed, our formulation was ambiguous. The EDR files contain the detailed information on energies, whereas TRR files contain numerous values from the trajectory such as coordinates, velocities, forces and to some extent also energies

(https://manual.gromacs.org/current/reference-manual/file-formats.html#trr)

L207: The text states that the total time was not available from XTC files, only the number of frames. However, XTC files record time stamps in addition to frame numbers. As long as these times are in the Gromacs standard of picoseconds, the simulation time ought to be available from XTCs.

“…systems and the number of frames available in the files (Fig. 3-B). Of note, the frames do not directly translate to the simulation runtime - more information deposited in other files (e.g. .mdp files) is needed to determine the complete runtime of the simulation. The system was up…”.

Thank you for the useful comment, we removed this sentence. We now mention that studying the simulation time would be of interest in the future, especially when we will perform an exhaustive analysis of XTC files.

“Of note, as .xtc files also contain time stamps, it would be interesting to study the relationship between the time and the number of frames to get useful information about the sampling. Nevertheless, this analysis would be possible only for unbiased MD simulations. So, we would need to decipher if the .xtc file is coming from biased or unbiased simulations, which may not be trivial.”

Analysis of MDP files: Were these standard equilibrium MD or can you distinguish biased MD or free energy calculations?

Currently we do not distinguish between biased and unbiased MD, but in the future we may attempt to do so, e.g. by correlating it with standard equilibration force-fields/parameters, timesteps or similar. Nevertheless, a true distinction will remain challenging.

L336: typo: pikes -> spikes (or peaks?)

“…simulations of Lennard-Jones models (Jeon et al., 2016). Interestingly, we noticed the appearance of several pikes at 400K, 600K and 800K, which were not present before the end of the year 2022. These peaks correspond to the same study related to the stability of hydrated crystals (Dybeck et al., 2023)’ Overall, thhis analysis revealed that a wide range of temperatures have been explored,…”

Thank you. We have corrected this typo.

Make clear how multiple versions of data sets are handled, e.g., if v1, v2, and v3 of a dataset are provided in Zenodo then which one is counted or are all counted?

We collected the latest version only of datasets, as exposed by default by the Zenodo API. To reflect this, we added the following sentence to the Methods and Materials section, Initial data collection sub-section:

“By default, the last version of the datasets was collected.”

L248 Analysis of GRO files seems fairly narrow because PDB files are very often used for exactly the same purpose, even in the context of Gromacs simulations, not the least because it is familiar to structural biologists that may be interested in representative MD snapshots. Despite all the shortcomings of abusing the PDB format for MD, it is an attempt at increased interoperability. Perhaps the authors can make sure that readers understand that choosing GRO for analysis may give a somewhat skewed picture, even within Gromacs simulations.

Thanks for this comment. We collected about 12,000 PDB files that could indeed be output from Gromacs simulations and easily be shared due to the universality of this format, but that could as well come from different sources (like other MD packages or the PDB database itself). We purposely decided to limit our study to files strictly associated with the Gromacs package, like MDP and XTC file types. However, we will extend our survey to all other structure-like formats and especially the PDB file type. We reflected this purpose in the following sentence (after line 281)

“Beyond .gro files, we would like to analyze the ensemble of the ~12,000 .pdb files extracted in this study (see Figure 2-B) to better characterize the types of molecular structures deposited.”

A simple template metadata file would be welcome (e.g., served from a GitHub/GitLab repository so that it can be improved with community input).

Thank you for this suggestion that we fundamentally agree with. However, the generation of such a file is a major task, and we believe that the creation of a metadata file template requires far-reaching considerations, therefore is beyond the scope of this paper and should not be decided by a small group of researchers. Indeed, this topic requires a large consensus of different stakeholders, from users, to MD program developers, and journal editors. It would be especially useful to organize dedicated workshops with representatives of all these communities to tackle this specific issue, as mentioned by Reviewer3 in his/her public review. As a basis for this discussion, we humbly proposed at the end of this manuscript a few non-constraining guidelines based on our experience retrieving the data.

To emphasize this statement, we added the following sentence at the end of the “Guidelines for better sharing of MD simulation data” section (line 420):

“Converging on a set of metadata and format requires a large consensus of different stakeholders from users, to MD program developers, and journal editors. It would be especially useful to organize specific workshops with representatives of all these communities to collectively tackle this specific issue.”

In "Data and code availability" it would be good to specify licenses in addition to stating "open source". Thank you for pointing out that GitLab/GitHub are not archives and that everyone should be strongly encouraged to submit data to stable archival repositories.

We added the corresponding licenses for code and data in the “Data and code availability” section.

Reviewer #3 (Recommendations For The Authors)

The paper is well written, with very few typographical or other minor errors.

Minor points:

Line 468-9 "can evolve being more user-friendly" should be "can evolve to being more user-friendly", I think.

Thank you, we have changed the wording accordingly.

Author response:

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

eLife assessment

This valuable study reports on the packing of molecules in cellular compartments, such as actin-based protrusions. The study provides solid evidence for parameters that enable the building of a biophysical model of filopodia, which is required to gain a complete understanding of these important actin-based structures. Some areas of the manuscript require further clarification.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

The manuscript proposes an alternative method by SDS-PAGE calibration of Halo-Myo10 signals to quantify myosin molecules at specific subcellular locations, in this specific case filopodia, in epifluorescence datasets compared to the more laborious and troublesome single molecule approaches. Based on these preliminary estimates, the authors developed further their analysis and discussed different scenarios regarding myosin 10 working models to explain intracellular diffusion and targeting to filopodia.

Strengths:

Overall, the paper is elegantly written and the data analysis is appropriately presented.

Weaknesses:

While the methodology is intriguing in its descriptive potential and could be the beginning of an interesting story, a good portion of the paper is dedicated to the discussion of hypothetical working mechanisms to explain myosin diffusion, localization, and decoration of filopodial actin that is not accompanied by the mandatory gain/loss of function studies required to sustain these claims.

To be fair, the detailed mechanisms that we raise related to diffusion, localization, and decoration are based on extensive work by others. Many prior papers use domain deletions of Myo10 and fall in the category of gain/loss-of-function studies. It is true that we have not repeated those extensive studies, but it seems appropriate to connect with and cite their work where appropriate.

Reviewer #2 (Public Review):

Summary:

The paper sought to determine the number of myosin 10 molecules per cell and localized to filopodia, where they are known to be involved in formation, transport within, and dynamics of these important actin-based protrusions. The authors used a novel method to determine the number of molecules per cell. First, they expressed HALO tagged Myo10 in U20S cells and generated cell lysates of a certain number of cells and detected Myo10 after SDS-PAGE, with fluorescence and a stained free method. They used a purified HALO tagged standard protein to generate a standard curve which allowed for determining Myo10 concentration in cell lysates and thus an estimate of the number of Myo10 molecules per cell. They also examined the fluorescence intensity in fixed cell images to determine the average fluorescence intensity per Myo10 molecule, which allowed the number of Myo10 molecules per region of the cell to be determined. They found a relatively small fraction of Myo10 (6%) localizes to filopodia. There are hundreds of Myo10 in each filopodia, which suggests some filopodia have more Myo10 than actin binding sites. Thus, there may be crowding of Myo10 at the tips, which could impact transport, the morphology at the tips, and dynamics of the protrusions themselves. Overall, the study forms the basis for a novel technique to estimate the number of molecules per cell and their localization to actin-based structures. The implications are broad also for being able to understand the role of myosins in actin protrusions, which is important for cancer metastasis and wound healing.

Strengths:

The paper addresses an important fundamental biological question about how many molecular motors are localized to a specific cellular compartment and how that may relate to other aspects of the compartment such as the actin cytoskeleton and the membrane. The paper demonstrates a method of estimating the number of myosin molecules per cell using the fluorescently labeled HALO tag and SDS-PAGE analysis. There are several important conclusions from this work in that it estimates the number of Myo10 molecules localized to different regions of the filopodia and the minimum number required for filopodia formation. The authors also establish a correlation between number of Myo10 molecules filopodia localized and the number of filopodia in the cell. There is only a small % of Myo10 that tip localized relative to the total amount in the cell, suggesting Myo10 have to be activated to enter the filopodia compartment. The localization of Myo10 is log-normal, which suggest a clustering of Myo10 is a feature of this motor.

Weaknesses:

One main critique of this work is that the Myo10 was overexpressed. Thus, the amount in the cell body compared to the filopodia is difficult to compare to physiological conditions. The amount in the filopodia was relatively small - 100s of molecules per filopodia so this result is still interesting regardless of the overexpression. However, the overexpression should be addressed in the limitations.

This is a reasonable perspective and we now note this caveat in the Limitations section so that readers will take note. Our goal here was to understand a system in which Myo10 is the limiting reagent for filopodia, rather than a native system that expresses high Myo10 on its own. Because U2OS cells do not express detectable levels of Myo10 (see below), the natural perturbation here is overexpressing Myo10 to stimulate filopodial growth.

The authors have not addressed the potential for variability in transfection efficiency. The authors could examine the average fluorescence intensity per cell and if similar this may address this concern.

Indeed, cells are heterogenous and will naturally express different levels of Myo10 not only due to transfection efficiency, but also due to their state (cell cycle stage, motile behavior, and more). In fact, we measure the transfection efficiency of each bioreplicate and account for it in our calibration procedure. We also measure the fluorescence intensity per cell, which lets us calculate the total Myo10s per cell and the cell-to-cell variability. These Myo10 distributions across cells are shown in Fig. 1D-E.

We note here an error that we made in applying this transfection efficiency correction in the first submission. When we obtain the total Myo10 molecules by SDS-PAGE, we should divide by the total number of transfected cells. However, due to an operator precedence error, the transfection efficiency appeared in the numerator rather than the denominator. We have now corrected this error, which has the effect of increasing the number of molecules in all of our measurements. The effect of this correction has strengthened one of the paper’s main conclusions, that Myo10 is frequently overloaded at filopodial tips.

The SDS PAGE method of estimating the number of molecules is quite interesting. I really like this idea. However, I feel there are a few more things to consider. The fraction of HALO tag standard and Myo10 labeled with the HALO tagged ligand is not determined directly. It is suggested that since excess HALO tagged ligand was added we can assume nearly 100% labeling. If the HALO tag standard protein is purified it should be feasible to determine the fraction of HALO tagged standard that is labeled by examining the absorbance of the protein at 280 and fluorophore at its appropriate wavelength.

This is a fair point raised by the reviewer, and we have now measured a labeling efficiency of 90% in Supplementary Figure 2A-C. We have adjusted all values according to this labeling efficiency.

The fraction of HALO tagged Myo10 labeled may be more challenging to determine, since it is in a cell lysate, but there may be some potential approaches (e.g. mass spec, HPLC).

As noted, this value is considerably more challenging. Instead, we determined conditions under which labeling in cells is saturated. We have now stained with a concentration range for both fixed and live cell samples. Saturation occurs with ~0.5 μM HaloTag ligand-TMR in fixed/permeabilized cells and in live cells (Supplementary Figure 2D-E). This comparison of live cells vs. permeabilized cells allows us to say that the intact plasma membrane is not limiting labeling under these conditions.

In Figure 1B, the stain free gel bands look relatively clean. The Myo10 is from cell lysates so it is surprising that there are not more bands. I am not surprised that the bands in the TMR fluorescence gel are clean, and I agree the fluorescence is the best way to quantitate.

Figure 1B shows the focused view at high MW, and there is not much above Myo10. The full gel lanes shown in Supp. Fig. 1C show the expected number of bands from a cell lysate.

In Figure 3C, the number of Myo10 molecules needed to initiate a filopodium was estimated. I wonder if the authors could have looked at live cell movies to determine that these events started with a puncta of Myo10 at the edge of the cell, and then went on to form a filopodia that elongated from the cell. How was the number of Myo10 molecules that were involved in the initiation determined? Please clarify the assumptions in making this conclusion.

We thank the reviewer (and the other reviewers) for this excellent suggestion. We have now carried out these live cell experiments. These experiments were quite challenging, because we needed to collect snapshots of ~50 cells to measure the mean fluorescence intensity of transfected cells and then acquire movies of several cells for analysis. The U2OS cells were also highly temperature-sensitive and would retract their filopodia without objective heating.

We have now analyzed filopodial initiation events and measured considerably more Myo10 at the first signs of accumulation– in the 100s of molecules. The dimmer spots that we measured in the first draft were likely unrelated to filopodial initiation, and we have corrected the discussion on this point.

We now also track further growth from a stable filopodial tip (the phased-elongation mechanism from Ikebe and coworkers) and find approximately 500 molecules bud off in those events. We also track filopodial elongation rates as a function of Myo10 numbers. We have added additional live cell imaging sections that include these results.

It is stated in the discussion that the amount of Myo10 in the filopodia exceeds the number of actin binding sites. However, since Myo10 contains membrane binding motifs and has been shown to interact with the membrane it should be pointed that the excess Myo10 at the tips may be interacting with the membrane and not actin, which may prevent traffic jams.

This is also an excellent point to consider, and we have expanded the relevant discussion along these lines. We agree that the Myo10 at the filopodial tip is likely membrane-bound. We now estimate the 2D membrane area occupied by Myo10, and find that it reaches nearly full packing in many cases (under a number of assumptions that we spell out more fully in the manuscript).

Reviewer #3 (Public Review):

Summary:

The unconventional myosin Myo10 (aka myosin X) is essential for filopodia formation in a number of mammalian cells. There is a good deal of interest in its role in filopodia formation and function. The manuscript describes a careful, quantitative analysis of Myo10 molecules in U2OS cells, a widely used model for studying filopodia, how many are present in the cytosol versus filopodia and the distribution of filopodia and molecules along the cell edge. Rigorous quantification of Myo10 protein amounts in a cell and cellular compartment are critical for ultimately deciphering the cellular mechanism of Myo10 action as well as understand the molecular composition of a Myo10-generated filopodium.

Consistent with what is seen in images of Myo10 localization in many papers, the vast majority of Myo10 is in the cell body with only a small percentage (appr 5%) present in filopodia puncta. Interestingly, Myo10 is not uniformly distributed along the cell edge, but rather it is unevenly localized along the cell edge with one region preferentially extending filopodia, presumably via localized activation of Myo10 motors. Calculation of total molecules present in puncta based on measurement of puncta size and measured Halo-Myo10 signal intensity shows that the concentration of motor present can vary from 3 - 225 uM. Based on an estimation of available actin binding sites, it is possible that Myo10 can be present in excess over these binding sites.

Strengths:

The work represents an important first step towards defining the molecular stoichiometry of filopodial tip proteins. The observed range of Myo10 molecules at the tip suggests that it can accommodate a fairly wide range of Myo10 motors. There is great value in studies such as this and the approach taken by the authors gives one good confidence that the numbers obtained are in the right range.

Weaknesses:

One caveat (see below) is that these numbers are obtained for overexpressing cells and the relevance to native levels of Myo10 in a cell is unclear.

A similar concern was raised by Reviewer 2; please see above.

An interesting aspect of the work is quantification of the fraction of Myo10 molecules in the cytosol versus in filopodia tips showing that the vast majority of motors are inactive in the cytosol, as is seen in images of cells. This has implications for thinking about how cells maintain this large population in the off-state and what is the mechanism of motor activation. One question raised by this work is the distinction between cytosolic Myo10 and the population found at the ‘cell edge’ and the filopodia tip. The cortical population of Myo10 is partially activated, so to speak, as it is targeted to the cortex/membrane and presumably ready to go. Providing quantification of this population of motors, that one might think of as being in a waiting room, could provide additional insight into a potential step-by-step pathway where recruitment or binding to the cortical region/plasma membrane is not by itself sufficient for activation.

As mentioned in our response to Reviewer 2, we have now carried out quantitation in live cells to capture Myo10 transitions from cell body into filopodial movement. We attempted to identify this membrane-bound population of motors in our new live cell experiments but were unable to make convincing measurements. Notably, we see no noticeable enrichment of Myo10 at the cortex relative to the cytosol. Although we believe there is a membrane-bound waiting room (akin to the 3D-2D-1D mechanism of Molloy and Peckham), we suspect that the 2D population is diffusing too rapidly to be detected under our imaging conditions.

Specific comments:

(1) It is not obvious whether the analysis of numbers of Myo10 molecules in a cell that is ectopically overexpressing Myo10 is relevant for wild type cells. It would appear to be a significant excess based on the total protein stained blot shown in Fig S1E where a prominent band the size of tagged Myo10 seen in the transfected sample is almost absent in the WT control lane.

Even “wildtype” cells vary considerably in their Myo10 expression levels. For example, melanoma cells often heavily upregulate Myo10, while these U2OS cells produce nearly none (Supplementary Figure 1E). Thus, there is no single, widely acceptable target for Myo10 expression in wildtype cells.

Please note that the new Supplementary Figure 1E is a Myo10 Western blot, not total protein staining as before.

Ideally, and ultimately an important approach, would be to work with a cell line expressing endogenously tagged Myo10 via genome engineering. This can be complicated in transformed cells that often have chromosomal duplications.

Indeed, we chose U2OS cells for this work because they do not express detectable levels of Myo10, and thus we can avoid all of these complications. Here we can examine how Myo10 levels control filopodial production through ectopic expression.

However, even though there is an excess of Myo10 it would appear that activation is still under some type of control as the cytosolic pool is quite large and its localization to the cell edge is not uniform. But it is difficult to gauge whether the number of molecules in the filopodium is the same as would be seen in untransfected cells. Myo10 can readily walk up a filopodium and if excess numbers of this motor are activated they would accumulate in the tip in large numbers, possibly creating a bulge as and indeed it does appear that some tips are unusually large. Then how would that relate to the normal condition?

As noted above, the normal condition depends on the cellular system. However, endogenous Myo10 also accumulates in bulges at filopodial tips, so this is not a phenotype unique to Myo10 overexpression. For example, the images from Figure 1 of the Berg and Cheney (2002) citation show bulges from endogenous Myo10 in endothelial cells.

(2) Measurements of the localization of Myo10 focuses in large part on ‘Myo10 punctae’. While it seems reasonable to presume that these are filopodia tips, the authors should provide readers with a clear definition of a puncta. Is it only filopodia tips, which seems to be the case? Does it include initiation sites at the cell membrane that often appear as punctae?

We define puncta as any clusters/spots of Myo10 signal detected by segmentation, not limited to any location within the surface-attached filopodia. We exclude puncta that appear in the cell interior (~5 of which appear in Fig. 1A). These are likely dorsal filopodia, but there are few of these compared to the surface attached filopodia of U2OS cells. In Figure 2, “puncta” includes all Myo10 clusters along the filopodia shaft, though a majority happen to be tip-localized (please see Supplementary Figure 4B). We have edited the main text for clarification.

Along those lines, the position of dim punctae along the length of a filopodium is measured (Fig 3D). The findings suggest that a given filopodium can have more than one puncta which seems at odds if a puncta is a filopodia tip. How frequently is a filopodium with two puncta seen? It would be helpful if the authors provided an example image showing the dim puncta that are not present at the tip.

We have now provided an example image of dim puncta along filopodia in Supplementary Figure 4C.

(3) The concentration of actin available to Myo10 is calculated based on the deduction from Nagy et al (2010) that only 4/13 of the actin monomers in a helical turn are accessible to the Myo10 motor (discussion on pg 9; Fig S4). Subsequent work (Ropars et al, 2016) has shown that the heads of the antiparallel Myo10 dimer are flattened, but the neck is rather flexible, meaning that the motor can a variable reach (36 - 52 nm). Wouldn’t this mean that more actin could be accessible to the Myo10 motor than is calculated here?

Although we see why the reviewer might believe otherwise, the 4/13 fraction of accessible actin holds. This fraction is obtained from consideration of the fascin-actin bundle structure alone, independent of the reach of any particular myosin motor. Every repeating layer of 13 actin subunits (or 36 nm) has 4 accessible myosin binding-sites. The remaining 9 sites are rejected because a single myosin motor domain will have a steric clash with a neighboring actin filament in the bundle. A myosin with an exceptionally long reach might reach the next 13 subunit layer, but that layer also has only 4 binding sites. Thus, we can calculate the number of binding sites per unit length along the filopodium. This number would hold for a dimeric myosin with any reach, including myosin-5 or myosin-2.

(4) Quantification of numbers of Myo10 molecules in filopodial puncta (Fig 3C) leads the authors to conclude that ‘only ten or fewer Myo10 molecules are necessary for filopodia initiation’ (pg 7, top). While this is a reasonable based on the assumption that the formation of a puncta ultimately results from an initiation event, little is known about initiation events and without direct observation of coalescence of Myo10 at the cell edge that leads to formation of a filopodium, this seems rather speculative.

As noted above, we have now performed the necessary live cell imaging of filopodial nucleation events and have updated our conclusions accordingly.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

I have made a series of comments that might help the authors improve their manuscript:

- A full calibration of the methodology would require testing a wider range of protein amounts, to exhaustively detect the dynamic range of the technique. The authors acknowledge in the discussion that “Furthermore, our estimates of molecules are predicated on the calibration curve of the Halo Standard Protein on the SDS-PAGE gels, which is likely the highest source of error on our molecule counts”. A good way of convincing a nasty reviewer is to provide a calibration with more than 3 reference points. At least this will help exclude from the analysis cells where Myo10 estimates are not in the linear regime of detection.

We completely agree with the reviewer’s suggestion to build a robust calibration curve. The SDS gel shown in Figure 1C originally contained 4 reference points, but the highest HaloTag standard protein point oversaturated the detector at the set exposure in the TMR channel and was omitted. We have now re-run the SDS gel to include a HaloTag standard protein curve comprising 5 points, alongside all three bioreplicates from the fixed cell experiments and all three bioreplicates from the live cell experiments (updated in Figure 1B-C). We had saved frozen lysates from the original fixed cell work, so we were able to reanalyze our data with the new set of standards. The Myo10 quantities are consistent, but with much tighter CIs from the standard curve.

- As already said this methodology is intriguing, however, a correlative validation with a conventional SMLM approach to address the bona-fide of the method would be ideal.

Unfortunately, single molecule approaches for validation are impractical for us. Due to the relatively high magnification of our TIRF microscope and the large spread area of the U2OS cells, single cells typically extend beyond the field of view. We acknowledge the benefits of SMLM quantitative techniques and other approaches cited in the introduction section. To avoid use of special tools/instruments, we offer our methodology, based off Pollard group’s quantitative Western blotting of GFP, as a simpler alternative accessible to anyone.

- TMR is a small ligand likely interacting also with Halo in its denatured state. However, to clear any doubts a parallel Native-PAGE investigation should be included, or if existing a specific reference should be provided.

Perhaps there is a misunderstanding here. One of the key advantages of the HaloTag labeling system is that the engineered dehalogenase is covalently modified by the ligand (the TMR-ligand is a suicide substrate). This means that the TMR remains bound even under denaturing conditions, which allows its detection in SDS-PAGE. Native gels are unnecessary here.

- Moreover, SDS-PAGE is run at alkaline pH, have the authors considered these points when designing the methodology? Fluorescence images were taken in PBS, which has a different pH. Could the authors, or the literature, exclude these aspects as potential pitfalls in the methodology? Also temperature is affecting fluorescence emission, but it is easier to control with certain tolerance in the room-temperature regime.

Our method does not compare fluorescence values that cross the experimental systems (SDS-PAGE vs. microscopy). Cellular proteins and HaloTag protein standards are compared in a single setting of SDS-PAGE to obtain the average number of Myo10s per transfected cell. Likewise, all measurements on intact (live or fixed) cells are conducted in that single setting to obtain average fluorescence per cell. Thus, there is no issue with the different buffers or temperatures affecting fluorescence emission.

- The authors should test their approach also with truncation variants of Myosin10 (for instance lacking the PH or motor domain). This is a classical approach that might prove the potential of the technique when altering the capacity of the protein to interact with a main binding partner. Also, treatments that induced filopodia formation might prove useful (i.e., hypotonic media induce filopodia formation in some fibroblast cell lines in our hands).

The reviewer raises interesting suggestions that we aim to address in future experiments, but truncation variants and environmental perturbations are beyond the focus of the current manuscript. Here, we report on the otherwise unperturbed state when we add exogenous full-length Myo10 to the U2OS cells. But indeed, experiments with Myo10 domain truncations, PI3K and PTEN inhibition, and cargo protein / activating cofactor knock-downs (among others) are on our drawing board.

- Most of the mechanisms hypothesized in the discussion are sound and plausible. However, the authors have chosen an experimental model where transient transfection of exogenous Myo10 in U2OS is performed. This approach poses two main and fundamental questions that are not resolved by the data provided:

A) how do different expression levels affect the Myo10 counting?

Our counting procedure does not assume uniform expression across a population of cells– quite the opposite, in fact. We directly measure Myo10 expression levels on a cell-by-cell basis with microscopy, once we know the number of molecules in our total pool (see the Methods for details). As an example of the final output, Figs. 1D and 1E show the total number of Myo10 molecules per cell for fixed and live cells, respectively.

B) how does endogenous and unlabeled Myo10 hamper the bonafide of counts? The authors claimed “U2OS cells express low levels of Myo10, so there is a small population of unlabeled endogenous Myo10 unaddressed by this paper”. As presented, the low levels of endogenous Myo10 sound an arbitrary parameter, and there are no data presented that can limit if not exclude this bias in the analysis. To produce data in a genetically modified cell line with Halo-tag on the endogenous protein will represent a much cleaner system. Alternatively, the authors should look for Myo10 KO cell lines where they can back-transfect their Halo-Tagged Myo10 construct in a more consistent framework, focusing on cells with low-to-mid levels of expression.

We agree, this is an important point to nail down (and is often neglected in the literature). We have now measured the endogenous Myo10 levels in U2OS cells by Western blotting and found that it is undetectable compared to our HaloTagged construct expression. Please see Supp. Fig 1E. Thus, for all intents and purposes, every Myo10 molecule in these experiments came from our expression plasmid. Accordingly, we have removed this caveat from the paper.

Minor points

- Figure 1B. To help the reader SDS-PAGE gels annotations should be clearer already from the figure.

We have updated the annotations for clarity.

- Methods should be organized in sessions. As it stands, it is hard for the reader to look for technical details.

We have expanded and added subsections to the Methods as requested.

- The good practice of indicating the gene and transcript entry numbers and the primer used to amplify and clone into the backbone vectors is getting lost in many papers. I would strongly encourage the authors to add this information to the methods.

We have included the gene entries to the methods and will include a full FASTA file of the coding sequence as supplementary information to avoid any ambiguity here.

The authors write “It is unclear how myosins navigate to the right place at the right time, but our results support an important interplay between Myo10 and the actin network.” It is a bit scholastic to say that Myo10 and actin have an important interplay, they are major binding partners. What is the new knowledge contained in this sentence?

Agreed– we have deleted the sentence in question.

Reviewer #2 (Recommendations For The Authors):

The authors should address all the weaknesses indicated in the public review.

There were a few other places that require clarification.

On page 4, the last paragraph. It is stated that the targeting of Myo10 was reported/proposed based on previous work (ref 31). The next few sentences are not referenced and thus likely refer to ref 31. The authors did not measure the parameters discussed in these sentences, so it is important to clarify that they are referring to previous work and not the current study.

Indeed, the next few sentences still refer to old reference 31, so we have now edited the paragraph for clarity.

On page 7, the reference to Figure 3A indicates that the trend of higher Myo10 correlating with more filopodia. However, the reference to Figure 3B indicates total intracellular Myo10 weakly correlates with more filopodia. However, the x-axis on Figure 3B is filopodia molecules not the intracellular Myo10. Please clarify.

We appreciate the reviewer for catching our mistake. Those plots are now in Fig. 2 and have been edited accordingly.

Reviewer #3 (Recommendations For The Authors):

The Discussion of results at the end of each section is rather brief and could be expanded on a bit more.

Before we were operating under the constraints of an eLife Short Report. We have now expanded the discussion for a full article.

The authors mention that actin filaments at the tips of filopodia could be frayed, citing Medalia et al, 2007 (ref 40). That paper describes an early cryoEM analysis of filopodia from the amoeba Dictyostelium. EM images of mammalian filopodia tips, e.g. Svitkina et al, 2003, JCB, do not show quite the same organization of actin as seen in the Dictyostelium filopodia tips. However, recent work from the Bershadsky lab, Li et al, 2023, presents a few cryoEM images of tips of left-bent filopodia that are tightly adhered to a substrate and there it looks like actin filaments become disorganized in tips, along with membrane bulging. The authors should consider expanding discussion of the filopodia tips to take into account what is known for mammalian filopodia.

We thank the reviewer for bringing these enlightening papers to our attention. We have now included these citations in the discussion.

Fig 1D - The x-axis is a bit odd, it goes from 0 then to 2.5e+06 with no indication of the bin size. Can this be re-labelled or the scale displayed a bit differently?

We have double-checked the axis breaks, which are large because the underlying values are large. We have also provided the bin size as requested for all histograms.

Fig 4A - What is the bin size for the histogram?

As above, we have now updated the figure legends (now in Fig. 3) to include the bin size.

Methods -

- Please provide an accession number for the Myo10 nucleotide sequence used for this work as there are at least two known isoforms.

Thank you for noting this. We are using the full-length, not the headless isoform. We have now updated the Methods accordingly.

- No mention is made of the SDS sample buffer used, was that also added to the sample?

We have now updated the Methods accordingly.

- How are samples boiled at 70 deg C? Do the authors actually mean ‘heated’?

Indeed. We have now corrected “boiled” to “heated.”

- Could the authors please briefly explain the connected component analysis used to identify filopodia?

We have now updated the Methods accordingly.

- The intensity of filopodia was determined by dividing tip intensity by the total bioreplicate sum of intensities then multiplying it by the total pool, if this reviewer understands correctly. It sounds like intensities are being averaged across a whole cell population instead of cell-by-cell. Is that correct? If so, can the authors please provide the underlying rationale for this? If not, then please better describe what was actually done.

We apologize for the confusion. Intensities are being averaged (summed) across a whole cell population, but importantly that step is only used to obtain a scale factor that converts the fluorescence signal at the microscope to the number of molecules. We then use that scale factor for all cells imaged in the bioreplicate, to both 1) find the total Myo10 in that cell, and 2) find the total amount of that Myo10 in any given location within that cell.

To further clarify, each bioreplicate has a known total number of Myo10 molecules associated with the number of cells loaded onto the SDS gel. From the SDS gel, we have an average number of Myo10 molecules per positively transfected cell. If 50 cell images are analyzed, then there is a Myo10 ‘total pool’ of (50 cells) * (average Myo10 molecules/cell). The fluorescence signal intensities in microscopy were summed for all cells within the bioreplicate (50 cells in this example). However, due to variation in expression, not every cell has the same signal intensity when imaged under the same conditions. It would be inaccurate to assume each cell contains the average Myo10 molecules/cell. Therefore, to get the number of molecules within a given Myo10 cell (or punctum), the summed cell (punctum) intensity was divided by the bioreplicate fluorescence signal intensity sum and multiplied by ‘total pool.’

- The authors quantify Myo10 protein amounts by western blotting using Halo tag fluorescence, a method that should provide good accuracy. The results depend on the transfection efficiency and it is rarely the case that it is 100%. The authors state that they use a ‘value correction for positively transfected cells’ (pg 11). It is likely that there was a range of expression levels in the cells, how was a cut-off for classifying a cell as non-expressing determined or set?

As described in the Methods, “microscopy was used to count the percentage of transfected cells from ~105-190 randomly surveyed cells per bioreplicate.” Cells were labeled and located with DAPI. If no TMR signal could be visually detected by microscopy, then the cell was deemed to be non-Myo10 expressing. We did not set a cutoff fluorescence value, as untransfected cells have no detectable signal. Please see Supplementary Figure 1F for examples.

- “In-house Python scripts” are used for image analysis. Will these be made publicly available?

Yes, we will package these up on GitHub.

Author response:

Public Reviews: 

Reviewer #1 (Public Review): 

This study is convincing because they performed time-resolved X-ray crystallography under different pH conditions using active/inactive metal ions and PpoI mutants, as with the activity measurements in solution in conventional enzymatic studies. Although the reaction mechanism is simple and may be a little predictable, the strength of this study is that they were able to validate that PpoI catalyzes DNA hydrolysis through "a single divalent cation" because time-resolved X-ray study often observes transient metal ions which are important for catalysis but are not predictable in previous studies with static structures such as enzyme-substrate analog-metal ion complexes. The discussion of this study is well supported by their data. This study visualized the catalytic process and mutational effects on catalysis, providing new insight into the catalytic mechanism of I-PpoI through a single divalent cation. The authors found that His98, a candidate of proton acceptor in the previous experiments, also affects the Mg2+ binding for catalysis without the direct interaction between His98 and the Mg2+ ion, suggesting that "Without a proper proton acceptor, the metal ion may be prone for dissociation without the reaction proceeding, and thus stable Mg2+ binding was not observed in crystallo without His98". In future, this interesting feature observed in I-PpoI should be investigated by biochemical, structural, and computational analyses using other metal-ion dependent nucleases. 

We appreciate the reviewer for the positive assessment as well as all the comments and suggestions.

Reviewer #2 (Public Review): 

Summary: 

Most polymerases and nucleases use two or three divalent metal ions in their catalytic functions. The family of His-Me nucleases, however, use only one divalent metal ion, along with a conserved histidine, to catalyze DNA hydrolysis. The mechanism has been studied previously but, according to the authors, it remained unclear. By use of a time resolved X-ray crystallography, this work convincingly demonstrated that only one M2+ ion is involved in the catalysis of the His-Me I-PpoI 19 nuclease, and proposed concerted functions of the metal and the histidine. 

Strengths: 

This work performs mechanistic studies, including the number and roles of metal ion, pH dependence, and activation mechanism, all by structural analyses, coupled with some kinetics and mutagenesis. Overall, it is a highly rigorous work. This approach was first developed in Science (2016) for a DNA polymerase, in which Yang Cao was the first author. It has subsequently been applied to just 5 to 10 enzymes by different labs, mainly to clarify two versus three metal ion mechanisms. The present study is the first one to demonstrate a single metal ion mechanism by this approach. 

Furthermore, on the basis of the quantitative correlation between the fraction of metal ion binding and the formation of product, as well as the pH dependence, and the data from site-specific mutants, the authors concluded that the functions of Mg2+ and His are a concerted process. A detailed mechanism is proposed in Figure 6. 

Even though there are no major surprises in the results and conclusions, the time-resolved structural approach and the overall quality of the results represent a significant step forward for the Me-His family of nucleases. In addition, since the mechanism is unique among different classes of nucleases and polymerases, the work should be of interest to readers in DNA enzymology, or even mechanistic enzymology in general. 

Thank you very much for your comments and suggestions.

Weaknesses: 

Two relatively minor issues are raised here for consideration: 

p. 4, last para, lines 1-2: "we next visualized the entire reaction process by soaking I-PpoI crystals in buffer....". This is a little over-stated. The structures being observed are not reaction intermediates. They are mixtures of substrates and products in the enzyme-bound state. The progress of the reaction is limited by the progress of the soaking of the metal ion. Crystallography has just been used as a tool to monitor the reaction (and provide structural information about the product). It would be more accurate to say that "we next monitored the reaction progress by soaking....". 

We appreciate the clarification regarding the description of our experimental approach. We agree that our structures do not represent reaction intermediates but rather mixtures of substrate and product states within the enzyme-bound environment. We will revise the text accordingly to more accurately reflect our methodology.

p. 5, the beginning of the section. The authors on one hand emphasized the quantitative correlation between Mg ion density and the product density. On the other hand, they raised the uncertainty in the quantitation of Mg2+ density versus Na+ density, thus they repeated the study with Mn2+ which has distinct anomalous signals. This is a very good approach. However, there is still no metal ion density shown in the key Figure 2A. It will be clearer to show the progress of metal ion density in a figure (in addition to just plots), whether it is Mg or Mn. 

Thank you for your insightful comments. We recognize the importance of visualizing metal ion density alongside product density data. As you commented, distinguishing between Mg2+ and Na+ is challenging, and in Fig 2A, no distinguishable density was observed at 20s. Mn2+, with its higher electron density, is detectable even at low occupancy. To address this, we will include figure panels in Figure 3 or supplementary figures to present Mn2+ and product densities concurrently.

Author response:

a) that the investigation is very interesting and inventive, and has the potential to reveal some novel insights.

We thank the reviewers and are excited to improve upon the manuscript through their suggestions.

b) that the problem of temporal autocorrelation in the fMRI and behavioral data has not been dealt with clearly and convincingly

We agree that convincingly accounting for fMRI temporal autocorrelation is important to our claims. To reduce its effects, we used field standard methods: prewhitening and autocorrelation modeling with SPM’s FAST algorithm (shown by Olszowy et al. 2019 to be superior to SPM’s default setting), as well as a high-pass filter of 128 Hz. There is still some first-order autocorrelation structure present across voxels in the left hippocampal beta series: across participants there is slightly positive autocorrelation between the betas of decision trials on successive trials, that decays to ~0 at subsequent lags. We note that our task is a narrative, and some patterns over time are expected; instead of attempting to fully eliminate all temporal structure in the data, we aim to show that the temporal distance between trials is unlikely to explain our effects.

In the within versus between social dimension representational similarity analysis, the average temporal distance between trials is the same within and between dimensions. The clustering analysis is a between subject analysis about individual differences–and the same overall temporal structure is experienced by all participants.

The trajectory analysis does not focus on consecutive trials across characters, but rather on consecutive trials within characters, where the time gap between successive trials is relatively large and highly variable. An average of over a minute of time elapses between successive decision trials for a given character (versus ~20 seconds across characters), which is on average almost 11 narrative slides and 3 decision trials. Across characters, the temporal gap between decision trials ranges between 12 seconds to more than 10 minutes, reducing the likelihood that temporal autocorrelation drives character-related estimates. We also highlight the shuffled choices control model, which shares the same temporal autocorrelation structure as the model of interest but had significantly poorer social location decoding–a strong indication that temporal autocorrelation alone can’t explain these results. For each participant, we shuffled their choices and re-computed trajectories that preserved the origin and end locations but produced different locations along the way. Our model decoded location significantly better than this null model, and this difference in performance can't be explained by differences in temporal autocorrelation in the neural or behavioral data.

In the revision, we will further address this concern. For example, we will report more details on the task structure to aid in interpretation and will more precisely characterize the temporal autocorrelation profile. Where appropriate, we will also improve on and/or add more control analyses that preserve the autocorrelation structure.

c) that a number of important interesting questions have not been addressed: Are the differences between social partners encoded in the hippocampus? Are the social dimensions encoded in a consistent manner across social partners?

We believe that we should be able to decode other interesting task- and relationship-related features from the hippocampal patterns, as suggested by the reviewers. In the revision, we will attempt several such analyses, while taking care to control for temporal autocorrelation.

d) that the cluster analysis in the brain-behavior correlation analysis is not well motivated or validated and should be clarified.

We agree with the reviewers that this clustering analysis should be better described and validated. We aimed to ask whether less diverse and distinctive cognitive representations of the relationship trajectories relate to smaller real-world social networks. This question of impoverished cognitive maps was first raised by Edward Tolman; we think it is relevant here, as well. In the revision, we will clarify its motivations and implications, and better evaluate it for its robustness. Here, we address a few comments made by the reviewers.

Reviewer 2 noted that other analyses could be used to ask whether social cognitive map complexity relates to real-world social network complexity. While the proposed alternatives are interesting (e.g., correlating decoding accuracy with social network size), we believe these analyses ask different questions. The current co-clustering analysis was intended to estimate map complexity jointly from the behavioral and neural signatures of the social map across characters. In contrast, the spline location decoding is within character; the accuracy of this decoding does not say much about representations across characters. And although we think character decoding is an interesting possible addition to this manuscript, its accuracy may reflect other aspects of the relationships, beyond just spatial representation. Thus, we will provide a clearer and better validated version of the current analysis to address this question.

We would also like to clarify that we did not collect the Social Network Index questionnaire in the Initial sample; as such these results are more tentative than the other analyses, due to the inability to confirm them in a separate sample. Reviewer 2 also suggests that a single outlier could drive this effect; but estimating the effect with robust regression also returns a right-tailed p < 0.05, showing that the relationship is robust to outliers.

References

Olszowy, W., Aston, J., Rua, C. & Williams, W.B. Accurate autocorrelation modeling substantially improves fMRI reliability. Nature Communications. (2019).

Author response:

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

Reviewer 1:

Comment 0: In this paper, the authors develop a comprehensive program to investigate the organization of chromosome structures at 100 kb resolution. It is extremely well executed. The authors have thought through all aspects of the problem. The resulting software will be most useful to the community. Interestingly they capture many experimental observations accurately.

I have very few complaints.

We appreciate the reviewer’s strong assessment of the paper’s significance, novelty, and broad interest, and we thank them for the detailed suggestions and comments.

Comment 1: The number of parameters in the energy function is very large. Is there any justification for this? Could they simplify the functions?

We extend our gratitude to the reviewer for their insightful remarks. The parameters within our model can be categorized into two groups: those governing chromosome-chromosome interactions and those governing chromosome-nuclear landmark interactions.

In terms of chromosome-chromosome interactions, the parameter count is relatively modest compared to the vast amount of Hi-C data available. For instance, while the whole-genome Hi-C matrix at the 100KB resolution encompasses approximately 303212 contacts, our model comprises merely six parameters for interactions among different compartments, along with 1000 parameters for the ideal potential. As outlined in the supporting information, the ideal potential is contingent upon sequence separation, with 1000 chosen to encompass bead separations of up to 100MB. While it is theoretically plausible to reduce the number of parameters by assuming interactions cease beyond a certain sequence separation, determining this scale a priori presents a challenge.

During the parameterization process, we observed that interchromosomal contacts predicted solely based on compartmental interactions inadequately mirrored Hi-C data. Consequently, we introduced 231 additional parameters to more accurately capture interactions between distinct pairs of autosomes. These interactions may stem from factors such as non-coding RNA or proteins not explicable by simple, non-specific compartmental interactions.

Regarding parameters concerning chromosome-nuclear landmark interactions, we have 30321 parameters for speckles and 30321 for the nuclear lamina. To streamline the model, we opted to assign a unique parameter to each chromatin bead. However, it is conceivable that many chromatin beads share a similar mechanism for interacting with nuclear lamina or speckles, potentially allowing for a common parameter assignment. Nonetheless, implementing such simplification necessitates a deeper mechanistic understanding of chromosome-nuclear landmark interactions, an aspect currently lacking.

As our comprehension of nuclear organization progresses, the interpretability of parameter counts may improve, facilitating their reduction.

Comment 2: What would the modification be if the resolution is increased?

To increase the resolution of chromatin, we can in principle keep the same energy function as defined in Eq. S6. In this case, we only need to carry out further parameter optimization.

However, transitioning to higher resolutions may unveil additional features not readily apparent at 100kb. Notably, chromatin loops with an average size of 200kb or smaller have been identified in high-resolution Hi-C data [1]. To effectively capture these loops, new terms in the energy function must be incorporated. For instance, Qi and Zhang [2] employed additional contact potentials between CTCF sites to account for loop formation. Alternatively, an explicit loop-extrusion process could be introduced to model loop formation more accurately.

Comment 3: They should state that the extracted physical values are scale-dependent. For example, viscosity.

We thank the reviewer for the comment and would like to clarify that our model does not predict the viscosity. The nucleoplasmic viscosity was set as 1Pa · s to produce a diffusion coefficient that reproduces experimental value. The exact value for the nucleoplasmic viscosity is still rather controversial, and our selected value falls in the range of reported experimental values from 10−1Pa·s to 102Pa · s.

We have modified the main text to clarify the calculation of the diffusion coefficient.

“The exponent and the diffusion coefficient Dα = (27±11)×10−4μm2 · s−α both match well with the experimental values [cite], upon setting the nucleoplasmic viscosity as 1Pa · s (see Supporting Information Section: Mapping the reduced time unit to real time for more details).”

Reviewer 2:

Comment 0: In this work, Lao et al. develop an open-source software (OpenNucleome) for GPU-accelerated molecular dynamics simulation of the human nucleus accounting for chromatin, nucleoli, nuclear speckles, etc. Using this, the authors investigate the steady-state organization and dynamics of many of the nuclear components.

We thank the reviewer for summary of our work.

Comment 1: The authors could introduce a table having every parameter and the optimal parameter value used. This would greatly help the reader.

We would like to point out that model parameters are indeed provided in Table S1, S2, S3, S4, and Fig. S7. In these tables, we further provided details on how the parameters were determined.

Given the large number of parameters for the ideal potential (1000), we opted to plot it rather than listing out all the numbers. We added three new figures to plot the interaction parameters between chromosomes, between chromosomes and speckles, and between chromosomes and the nuclear lamina. Numerical values can be found online in the GitHub repository (parameters).

Comment 2: How many total beads are simulated? Do all beads have the same size?

The total number of the coarse-grained beads is 70542, including 60642 chromatin beads, 300 nucleolus beads, 1600 speckle beads, and 8000 nuclear lamina beads. The radius of the chromatin, nucleolus, and speckle beads is 0.25, while that of the lamina bead is 0.5. More information of the size and number of the beads are discussed in the Section: Components of the whole nucleus model.

Comment 3: In Equation S17, what is the 3rd and 4th powers mean? What necessitates it?

The potential defined in Equation S17 follows the definition of class2 bond in the LAMMPS package (LAMMPS docs). Compared to a typical harmonic potential, the presence of higher order terms produces sharper increase in the energy at large distances (Author response image 1). This essentially reduces the flucatuation of bond length in simulations.

Author response image 1.

Comparison between the Class2 potential (defined in Eq. S17) and the Harmonic potential (K(r − r0)2, with K = 20 and r0 = 0.5).

Comment 4: What do the X-axis and Y-axis numbers in Figure 5A and 5B mean? What are their units?

We apologize for the lack of clarify in our original figure. In Fig. 5A, the X and Y axis depicts the simulated and experimental radius of gyration (Rg) for individual chromosomes, as indicated in the title of the figure. Similarly, in Fig. 5B, the X and Y axis depicts the simulated and experimental radial position of individual chromosomes.

We have converted the chromosome Rg values into reduced units and labeled the corresponding axes in the updated figure (Fig. 5). The normalized radial position is unitless and its detailed definition is included in the supporting information Section: Computing simulated normalized chromosome radial positions. We updated the figure caption to provide an explicit reference to the SI text.

Reviewer 3:

Comment 0: In this work, the authors present the development of OpenNucleome, a software for simulating the structure and dynamics of the human nucleus. It provides a detailed model of nuclear components such as chromosomes and nuclear bodies, and uses GPU acceleration for better performance based on the OpenMM package. The work also shows the model’s accuracy in comparisons with experimental data and highlights the utility in the understanding of nuclear organization. While I consider this work a good tool for the genome architecture scientific community, I have some comments and questions that could further clarify the usage of this tool and help potential users. I also have a few questions that would help to clarify the technique and results and some suggestions for references.

We appreciate the reviewer’s strong assessment of the paper’s significance, novelty, and broad interest, and we thank them for the detailed suggestions and comments.

Comment 1: Could the authors elaborate on what they consider to be ’well-established and easily adoptable modeling tools’?

By well established, we meant that models that have been extensively validated and verified, and are highly regarded by the community.

By easily adoptable, we meant that tools that are well documented and can be relatively easily learned by new groups without help from the developers.

We have revised the text to clarify our meaning.

“Despite the progress made in computational modeling, the absence of well-documented software with easy-to-follow tutorials pose a challenge.”

Comment 2: Recognizing the value of a diverse range of tools in the community, the Open-MiChroM tool is also an open-source platform built on top of OpenMM. The documentation shows various modeling approaches and many tutorials that contain different approaches besides the MiChroM energy function. How does OpenNucleome compare in terms of facilitating crossvalidation and user accessibility? The two tools seem to be complementary, which is a gain to the field. I recommend adding one or two sentences in the matter. Also, while navigating the OpenNucleome GitHub, I have not found the tutorials mentioned in the text. I also consider a barrier in the process of generating necessary input files. I would suggest expanding the tutorials and documentation to help potential users.

We thank the reviewer for the excellent comments. We agree that while many of the tutorials were included in the original package, they were not as clearly documented. We have revised them extensively to to now present:

• A tutorial for optimizing chromosome chromosome interactions.

• A tutorial for optimizing chromosome nuclear landmark interactions.

• A tutorial for building initial configurations.

• A tutorial for relaxing the initial configurations.

• A tutorial for selecting the initial configurations.

• A tutorial for setting up performing Langevin dynamics simulations.

• A tutorial for setting up performing Brownian dynamics simulations.

• A tutorial for setting up performing simulations with deformed nucleus.

• A tutorial for analyzing simulation trajectories.

• A tutorial for introducing new features to the model.

These tutorials and our well-documented and open source code (https://zhanggroup-mitchemistry.github.io/OpenNucleome) should significantly promote user accessibility. Our inclusion of python scripts for analyzing simulation trajectorials shall allow users to compute various quantities for evaluating and comparing model quality.

We added a new paragraph in the Section: Conclusions and Dicussion of the main text to compare OpenNucleosome with existing software for genome modeling.

“Our software enhances the capabilities of existing genome simulation tools [cite]. Specifically, OpenNucleome aligns with the design principles of Open-MiChroM [cite], prioritizing open-source accessibility while expanding simulation capabilities to the entire nucleus. Similar to software from the Alber lab [cite], OpenNucleome offers highresolution genome organization that faithfully reproduces a diverse range of experimental data. Furthermore, beyond static structures, OpenNucleome facilitates dynamic simulations with explicit representations of various nuclear condensates, akin to the model developed by [citet].”

Comment 3: Lastly, I would appreciate it if the authors could expand their definition of ’standardized practices’.

We apologize for any confusion caused. By ”standardized practices,” we refer to the fact that different groups often employ unique procedures for structural modeling. These procedures differ in the representation of chromosomes, the nucleus environment, and the algorithms for parameter optimization. This absence of a consensus on the optimal practices for genome modeling can be daunting for newcomers to the field.

We have revised the text to the following to avoid confusion:

“Many research groups develop their own independent software, which complicates crossvalidation and hinders the establishment of best practices for genome modeling [3–5].”

Comment 4: On page 7, the authors refer to the SI Section: Components of the whole nucleus model for further details. Could the authors provide more information on the simulated density of nuclear bodies? Is there experimental data available that details the ratio of chromatin to other nuclear components, which was used as a reference in the simulation?

We thank the reviewer for the comment. Imaging studies have provided quantitative measures about the size and number of various nuclear bodies. For example, there are 2 ∼ 5 nucleoli per nucleus, with the typical size RNo ≈ 0.5μm [6–10]. In the review by Spector and Lamond [11], the authors showed that there are 20 ∼ 50 speckles, with the typical size RSp ≈ 0.3μm. We used these numbers to guide our simulation of nuclear bodies. These information was mentioned in the Section: Chromosomes as beads on the string polymers of the supporting information.

The chromatin density is fixed by the average size of chromatin bead and the nucleus size. We chose the size of chromatin based on imaging studies as detailed in the Subsection: Mapping chromatin bead size to real unit of the supporting information. Upon fixing the bead size, the chromatin volume is determined.

Comment 5: In the statement, ’the ideal potential is only applied for beads from the same chromosome to approximate the effect of loop extrusion by Cohesin molecules for chromosome compaction and territory formation,’ it would be helpful if the authors could clarify the scope of this potential. Specifically, the code indicates that the variable ’dend ideal’ is set at 1000, suggesting an interaction along a 100Mb polymer chain at a resolution of 100Kb per bead. Could the authors elaborate on their motivation for the Cohesin complex’s activity having a significant effect over such long distances within the polymer chain?

We thank the reviewer for the insight comment. They are correct that the ideal potential was introduced to capture chromosome folding beyond the interactions between compartments, including loop extrusion. Practically, we parameterized the ideal potential such that the simulated average contact probabilities as a function of sequence separation match the experimental values. The reviewer is correct that beyond a specific value of sequence separation, one would expect the impact of loop extrusion on chromosome folding should be negligible, due to Cohesin dissociation. Correspondingly, the interaction potential should be zero at large sequence separations.

However, it is important to note that the precise separation scale cannot be known a priori. We chose 100Mb as a conservative estimation. However, as we can see from Fig. S7, our parameterization scheme indeed produced interaction parameters are mainly zero at large sequence separations. Interesting, the scale at which the potential approaches 0 (∼ 500KB), indeed agree with the estimated length traveled by Cohesin molecules before dissociation [12].

Comment 6: On pages 8 and 9, the authors discuss the optimization process. However, in reviewing the code and documentation available on the GitHub page, I could not find specific sections related to the optimization procedure described in the paper. In this context, I have a few questions: Could the authors provide more details or direct me to the parts of the documentation and the text/SI that address the optimization procedure used in their study? Additional clarification on the cost/objective function employed during the optimization process would be highly beneficial, as this was not readily apparent in the text.

We thank the reviewer for the comment. We revised the SI to include the definition of the cost function for the Adam optimizer.

“During the optimization process, our aim was to minimize the disparity between experimental findings and simulated data. To achieve this, we defined the cost function as follows:

where the index i iterates over all the constraints defined in Eq. S28.”

The detailed optimization procedure was included in the SI as quoted below

“The details of the algorithm for parameter optimization are as follows

(1) Starting with a set of values for and we performed 50 independent 3-million-step long MD simulations to obtain an ensemble of nuclear configurations. The 500K steps of each trajectory are discarded

as equilibration. We collected the configurations at every 2000 simulation steps from the rest of the simulation trajectories to compute the ensemble averages defined on the left-hand side of Eq. S13.

(2) Check the convergence of the optimization by calculating the percentage of error

defined as . The summation over i includes all the average contact probabilities defined in Eq. S28.

(3) If the error is less than a tolerance value etol, the optimization has converged, and we stop the simulations. Otherwise, we update the parameters, α, using the Adam optimizer [13]. With the new parameter values, we return to step one and restart the iteration.”

Previously, the optimization code was included as part of the analysis folder. To avoid confusion and improve readability, a separate folder named optimization has been created. This folder provides the Adam optimization of chromosome-chromosome interactions (chr-chr optimization) and chromosome-nuclear landmarks interactions (chr-NL optimization).

Comment 7: What was the motivation for choosing the Adam algorithm for optimization? Adam is designed for training on stochastic objective functions. Could the authors elucidate on the ’stochastic’ aspect of their function to be optimized? Why the Adam algorithm was considered the most appropriate choice for this application?

We thank the reviewer for the comment. As defined in Eq. R1, the cost function measures the difference between the simulated constraints with corresponding experimental values. The estimation of simulation values, by averaging over an ensemble of chromosome configurations, is inherently noisy and stochastic. Exact ensemble averages can only be achieved with unlimited samples obtained from infinite long simulations.

In the past, we have used the Newton’s method for parameterization, and the detailed algorithm can be found in the SI of Ref. 14. However, we found that Adam is more efficient as it is a first-order approximation method. The Newton’s method, on the other hand, is second-order approximation method and requires estimation of the Hessian matrix. When the number of constraints is large, as is in our case, the computational cost for estimating the Hessian matrix can be significant. Another advantage of the Adam algorithm lies in its adjustment of the learning rate along the optimization to further speedup convergence.

Comment 8: The authors mention that examples of setting up simulations, parameter optimization, and introducing new features are provided in the GitHub repository. However, I was unable to locate these examples. Could the authors guide me to these specific resources or consider adding them if they are not currently available?

We thank the reviewer for the comment. We have improved the GitHub repository and all the tutorials can be found using the links provided in Response to Comment 2.

Comment 9: Furthermore, the paper states that ’a configuration file that provides the position of individual particles in the PDB file format is needed to initialize the simulations.’ It would be beneficial for new users if the authors could elaborate on how this file is generated. And all other input files in general. Detailing the procedures for a new user to run their system using OpenNucleome would be helpful.

We thank the reviewer for the comment. The procedure for generating initial configurations was explained in the SI Section: Initial configurations for simulations and quoted below.

“We first created a total of 1000 configurations for the genome by sequentially generating the conformation of each one of the 46 chromosomes as follows. For a given chromosome, we start by placing the first bead at the center (origin) of the nucleus. The positions of the following beads, i, were determined from the (i − 1)-th bead as . v is a normalized random vector, and 0.5 was selected as the bond length between neighboring beads. To produce globular chromosome conformations, we rejected vectors, v, that led to bead positions with distance from the center larger than 4σ. Upon creating the conformation of a chromosome i, we shift its center of mass to a value ri com determined as follows. We first compute a mean radial distance, with the following equation

where Di is the average value of Lamin B DamID profile for chromosome i. Dhi and Dlo represent the highest and lowest average DamID values of all chromosomes, and 6σ and 2σ represent the upper and lower bound in radial positions for chromosomes. As shown in Fig. S6, the average Lamin B DamID profiles are highly correlated with normalized chromosome radial positions as reported by DNA MERFISH [cite], supporting their use as a proxy for estimating normalized chromosome radial positions. We then select as a uniformly distributed random variable within the range . Without loss of generality, we randomly chose the directions for shifting all 46 chromosomes.

We further relaxed the 1000 configurations to build more realistic genome structures. Following an energy minimization process, one-million-step molecular dynamics (MD) simulations were performed starting from each configuration. Simulations were performed with the following energy function

where UGenome is defined as in Eq. S7. UG-La is the excluded volume potential between chromosomes and lamina, i.e, only the second term in Eq. S24. Parameters in UGenome were from a preliminary optimization. The end configurations of the MD simulations were collected to build the final configuration ensemble (FCE).”

The tutorial for preparing initial configurations can be found at this link.

Comment 10: In the section discussing the correlation between simulated and experimental contact maps, as referenced in Figure 4A and Figure S2, the authors mention a high degree of correlation. Could the authors specify the exact value of this correlation and explain the method used for its computation? Considering that comparing two Hi-C matrices involves a large number of data points, it would be helpful to know if all data points were included in this analysis.

We have updated Fig 4A and S2 to include Pearson correlation coefficients next to the contact maps. The reviewer is correct in that all the non-redundant data points of the contact maps are included in computing the correlation coefficients.

For improved clarity, we added a new section in the supporting information to detail the calculations. The section is titled Computing Pearson correlation coefficients between experimental and simulated contact maps, and the relevant text is quoted below.

“We computed the Pearson correlation coefficients (PCC) between experimental and simulated contact maps in Fig. 4A and Fig. S2 as

xi and yi represent the experimental and simulated contact probabilities, and n is the total number of data points. Only non-redundant data points, i.e., half of the pairwise contacts, are used in the PCC calculation.”

Comment 11: In addition, the author said: ”Moreover, the simulated and experimental average contact probabilities between pairs of chromosomes agree well, and the Pearson correlation coefficient between the two datasets reaches 0.89.” How does this correlation behave when not accounting for polymer compaction or scaling? An analysis presenting the correlation as a function of genomic distance would be interesting.

Author response image 2.

Pearson correlation coefficient between experimental and simulated contact probabilities as a function of the sequence separation within specific chromosomes. For each chromosome, we first gathered a set of experimental contacts alongside a matching set of simulated ones for genomic pairs within a particular separation range. The Pearson correlation coefficient at the corresponding sequence separation was then determined using Equation R4. We limited the calculations to half of the chromosome length to ensure the availability of sufficient data.

We thank the reviewer for the comment. The analysis presenting the correlation as a function of genomic distance (sequence separation) for each chromosome is shown in Figure S12 and also included in the SI. While the correlation coefficients decreases at larger separation, the values around 0.5 is quite reasonable and comparable to results obtained using Open-Michrom.

We also computed the correlation of whole genome contact maps after excluding intra-chromosomal contacts. The PCC decreased from 0.89 to 0.4. Again, the correlation coefficient is quite reasonable considering that these contacts are purely predicted by the compartmental interactions and were not directly optimized.

Comment 12: I recommend using the web-server that is familiar to the authors to benchmark the OpenNucleome tool/model: ”3DGenBench: A Web-Server to Benchmark Computational Models for 3D Genomics.” Nucleic Acids Research, vol. 50, no. W1, July 2022, pp. W4-12.

We appreciate the reviewer’s suggestion. Unfortunately, the website is no longer active during the time of the revision. However, as detailed in Response to comment 11, we used the one of the popular metrics to exclude polymer compact effect and evaluate the agreement between simulation and experiments.

Comment 13: Regarding the comparison of simulation results with microscopy data from reference 34. Given their different resolutions and data point/space groupings, how do the authors align these datasets? Could the authors describe how they performed this comparison? How were the radial positions calculated in both the simulations and experiments? Since the data from reference 34 indicates a non-globular shape of the nucleus; how did this factor into the calculation of radial distributions?

We thank the reviewer for the comment and apologize for the confusion. First, the average properties we examined, including radial positions and interchromosomal contacts, were averaged over all genomic loci. Therefore, they are independent of data resolution.

Secondly, instead of calculating the absolute radial positions, which are subject to variations in nucleus shape and size, we defined the normalized radial positions. They measure the ratio between the distance from the nucleus center to the chromosome center and the distance from the nucleus center to the lamina. This definition was frequently used in prior imaging studies to measure chromosome radial positions.

The calculation of the simulated normalized radial positions and the experimental normalized radial positions are discussed in the Section: Computing simulated normalized chromosome radial positions

“For a given chromosome i, we first determined its center of mass position denoted as Ci. Starting from the center of the nucleus, O, we extend the the vector vOC to identify the intersection point with the nuclear lamina as Pi. The normalized chromosome radial position i is then defined as , where ||·|| represents the L2 norm.

and Section: Computing experimental normalized chromosome radial positions.

“We followed the same procedure outlined in Section: Computing simulated normalized chromosome radial positions to compute the experimental values. To determine the center of the nucleus using DNA MERFISH data, we used the algorithm, minimum volume enclosing ellipsoid (MVEE)[15], to fit an ellipsoid for each genome structure. The optimal ellipsoid defined as is obtained by optimizing subjecting to the constraint that . xi correspond to the list of chromatin positions determined experimentally.”

Comment 14: In the sentence: ”It is evident that telomeres exhibit anomalous subdiffusive motion.” I recommend mentioning the work ”Di Pierro, Michele, et al., ”Anomalous Diffusion, Spatial Coherence, and Viscoelasticity from the Energy Landscape of Human Chromosomes.” Proceedings of the National Academy of Sciences, vol. 115, no. 30, July 2018, pp. 7753-58.”.

We have revised the sentence to include the citation as follows.

“In line with previous research [cite], telomeres display anomalous subdiffusive motion. When fitted with the equation , these trajectories yield a spectrum of α values, with a peak around 0.59.”

Comment 15: Regarding the observation that ’chromosomes appear arrested and no significant changes in their radial positions are observed over timescales comparable to the cell cycle,’ could the authors provide more details on the calculations or analyses that led to this conclusion? Specifically, information on the equilibration/relaxation time of chromosome territories relative to rearrangements within a cell cycle would be interesting.

Our conclusion here was mostly based on the time trace of normalized radial positions shown in Figure 6A of the main text. Over the timescale of an entire cell cycle (24 hours), the relatively little to no changes in the radial positions supports glassy dynamics of chromosomes. We further determined the mean squared displacement (MSD) for chromosome center of masses. As shown in the left panel of Fig. S12, the MSDs are much smaller than the average size of chromosomes (see Rg values in Fig. 5A), supporting arrested dynamics.

We further computed the auto-correlation function of the normalized chromosome radial position as

where t indexes over the trajectory frames and ¯r is the mean position. As shown in Fig. S12, the positions are not completely decorrelated over 10 hours, again supporting slow dynamics. It would be interesting to examine the relaxation timescale more closely in future studies.

Comment 16: The authors also comment on the SI ”Section: Initial configurations for simulations provides more details on preparing the 1000 initial configurations.” and related to reference 34 mentioning that ”the average Lamin B DamID profiles are highly correlated with chromosome radial positions as reported by DNA MERFISH”. How do the authors account for situations where homologous chromosomes are neighbors or have an interacting interface? Ref. 34 indicates that distinguishing between these scenarios can be challenging, potentially leading to ’invalid distributions’ that are filtered out. Clarification on how such cases were handled in the simulations would be helpful.

We would like to first clarify that when comparing with experimental data, we averaged over the homologous chromosomes to obtain haploid data. We added the following text in the manuscript to emphasize this point

“Given that the majority of experimental data were analyzed for the haploid genome, we adopted a similar approach by averaging over paternal and maternal chromosomes to facilitate direct comparison. More details on data analysis can be found in the Supporting Information Section: Details of simulation data analysis.”

Furthermore, we used the processed DNA MERFISH data from the Zhuang lab, which unambiguously assigns a chromosome ID to each data point. Therefore, the issue mentioned by the reviewer is not present in the procssed data. In our simulations, since we keep track of the explicit connection between genomic segments, the trace of individual chromosomes can be determined for any configuration. Therefore, there is no ambiguity in terms of simulation data.

Comment 17: When discussing the interaction with nuclear lamina and nuclear envelop deformation, I suggest mentioning the following studies: The already cited ref 52 and ”Contessoto, Vin´ıcius G., et al. ”Interphase Chromosomes of the Aedes Aegypti Mosquito Are Liquid Crystalline and Can Sense Mechanical Cues.” Nature Communications, vol. 14, no. 1, Jan. 2023, p. 326.”

We updated the text to include the suggested reference.

“Numerous studies have highlighted the remarkable influence of nuclear shape on the positioning of chromosomes and the regulation of gene expression [16, 17].”

Comment 18: The authors state that ’Tutorials in the format of Python Scripts with extensive documentation are provided to facilitate the adoption of the model by the community.’ However, as I mentioned, the documentation appears to be limited, and the available tutorials could benefit from further expansion. I suggest that the authors consider enhancing these resources to better assist users in adopting and understanding the model.

As detailed in the Response to Comment 2, we have updated the GitHub repository to better document the included Jupyter notebooks and tutorials.

Comment 19: In the Methods section, the authors discuss using Langevin dynamics for certain simulations and Brownian dynamics for others. Could the authors provide more detailed reasoning behind the choice of these different dynamics for different aspects of the simulation? Furthermore, it would be insightful to know how the results might vary if only one of these dynamics was utilized throughout the study. Such clarification would help in understanding the implications of these methodological choices on the outcomes of the simulations.

We thank the reviewer for the comment. As detailed in the supporting information Section: Mapping the Reduced Time Unit to Real Time, the Brownian dynamics simulations provide a rigorous mapping to the biological timescale. By choosing a specific value for the nucleoplasmic viscosity, we determined the time unit in simulations as τ = 0.65s. With this time conversion, the simulated diffusion coefficients of telomeres match well with experimental values. Therefore, Brownian dynamics simulations are recommended for computing time dependent quantities and the large damping coefficients mimics the complex nuclear environment well.

On the other hand, the large damping coefficient slows down the configuration relaxation of the system significantly. For computing equilibrium statistical properties, it is useful to use a small coefficient and the Langevin integrator with large time steps to facilitate conformational relaxation.

References

[1] Rao, S. S.; Huntley, M. H.; Durand, N. C.; Stamenova, E. K.; Bochkov, I. D.; Robinson, J. T.; Sanborn, A. L.; Machol, I.; Omer, A. D.; Lander, E. S.; others A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 2014, 159, 1665–1680.

[2] Qi, Y.; Zhang, B. Predicting three-dimensional genome organization with chromatin states. PLoS computational biology 2019, 15, e1007024.

[3] Yildirim, A.; Hua, N.; Boninsegna, L.; Zhan, Y.; Polles, G.; Gong, K.; Hao, S.; Li, W.; Zhou, X. J.; Alber, F. Evaluating the role of the nuclear microenvironment in gene function by population-based modeling. Nature Structural & Molecular Biology 2023, 1–14.

[4] Junior, A. B. O.; Contessoto, V. G.; Mello, M. F.; Onuchic, J. N. A scalable computational approach for simulating complexes of multiple chromosomes. Journal of molecular biology 2021, 433, 166700.

[5] Fujishiro, S.; Sasai, M. Generation of dynamic three-dimensional genome structure through phase separation of chromatin. Proceedings of the National Academy of Sciences 2022, 119, e2109838119.

[6] Caragine, C. M.; Haley, S. C.; Zidovska, A. Nucleolar dynamics and interactions with nucleoplasm in living cells. Elife 2019, 8, e47533.

[7] Brangwynne, C. P.; Mitchison, T. J.; Hyman, A. A. Active liquid-like behavior of nucleoli determines their size and shape in Xenopus laevis oocytes. Proceedings of the National Academy of Sciences 2011, 108, 4334–4339.

[8] Farley, K. I.; Surovtseva, Y.; Merkel, J.; Baserga, S. J. Determinants of mammalian nucleolar architecture. Chromosoma 2015, 124, 323–331.

[9] Qi, Y.; Zhang, B. Chromatin network retards nucleoli coalescence. Nature Communications 2021, 12, 6824.

[10] Caragine, C. M.; Haley, S. C.; Zidovska, A. Surface fluctuations and coalescence of nucleolar droplets in the human cell nucleus. Physical review letters 2018, 121, 148101.

[11] Spector, D. L.; Lamond, A. I. Nuclear speckles. Cold Spring Harbor perspectives in biology 2011, 3, a000646.

[12] Banigan, E. J.; Mirny, L. A. Loop extrusion: theory meets single-molecule experiments. Current opinion in cell biology 2020, 64, 124–138.

[13] Kingma, D. P.; Ba, J. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 2014,

[14] Zhang, B.; Wolynes, P. G. Topology, structures, and energy landscapes of human chromosomes. Proceedings of the National Academy of Sciences 2015, 112, 6062–6067.

[15] Moshtagh, N.; others Minimum volume enclosing ellipsoid. Convex optimization 2005, 111, 1–9.

[16] Brahmachari, S.; Contessoto, V. G.; Di Pierro, M.; Onuchic, J. N. Shaping the genome via lengthwise compaction, phase separation, and lamina adhesion. Nucleic Acids Res. 2022, 50, 1–14.

[17] Contessoto, V. G.; Dudchenko, O.; Aiden, E. L.; Wolynes, P. G.; Onuchic, J. N.; Di Pierro, M. Interphase chromosomes of the Aedes aegypti mosquito are liquid crystalline and can sense mechanical cues. Nature Communications 2023, 14, 326.

Author response:

eLife assessment:

This important work provides another layer of regulatory mechanism for TGF-beta signaling activity. The evidence supports the involvement of microtubules as a reservoir of Smad2/3, however, additional evidence to convincingly demonstrate the functional involvement of Rudhira in this process is highly appreciated. The work will be of broad interest to developmental biologists in general and molecular biologists in the field of growth factor signaling.

Public Reviews:

Reviewer #1 (Public Review):

Summary

This manuscript aimed to study the role of Rudhira (also known as Breast Carcinoma Amplified Sequence 3), an endothelium-restricted microtubules-associated protein, in regulating of TGFβ signaling. The authors demonstrate that Rudhira is a critical signaling modulator for TGFβ signaling by releasing Smad2/3 from cytoskeletal microtubules and how Rudhira is a Smad2/3 target gene. Taken together, the authors provide a model of how Rudhira contributes to TGFβ signaling activity to stabilize the microtubules, which is essential for vascular development.

Strengths

The study used different methods and techniques to achieve aims and support conclusions, such as Gene Ontology analysis, functional analysis in culture, immunostaining analysis, and proximity ligation assay. This study provides an unappreciated additional layer of TGFβ signaling activity regulation after ligand-receptor interaction.

We thank the reviewer for acknowledging the importance of our study and providing a clear summary of our findings.

Weaknesses

(1) It is unclear how current findings provide a better understanding of Rudhira KO mice, which the authors published some years ago.

Our previous study demonstrated that Rudhira KO mice have a predominantly developmental cardiovascular phenotype that phenocopies TGFβ loss of function (Shetty, Joshi et al., 2018). Additionally, we found that at the molecular level, Rudhira regulates cytoskeletal organization (Jain et al., 2012; Joshi and Inamdar, 2019). Our current study builds upon these previous findings, showing an essential role of Rudhira in maintaining TGFβ signaling and controlling the microtubule cytoskeleton during vascular development. On one hand Rudhira regulates TGFβ signaling by promoting the release of Smads from microtubules, while on the other, Rudhira is a TGFβ target essential for stabilizing microtubules. Thus, our current study provides a molecular basis for Rudhira function in cardiovascular development.

(2) Why do they use HEK cells instead of SVEC cells in Figure 2 and 4 experiments?

Our earlier studies have characterized the role of Rudhira in detail using both loss and gain of function methods in multiple cell types (Jain et al., 2012; Shetty, Joshi et al., 2018; Joshi and Inamdar, 2019). As endothelial cells are particularly difficult to transfect, and because the function of Rudhira in promoting cell migration is conserved in HEK cells, it was practical and relevant to perform these experiments in HEK cells (Figures 2 and 4E).

(3) A model shown in Figure 5E needs improvement to grasp their findings easily.

We have modified Figure 5E for clarity.

Reviewer #2 (Public Review):

Summary

It was first reported in 2000 that Smad2/3/4 are sequestered to microtubules in resting cells and TGF-β stimulation releases Smad2/3/4 from microtubules, allowing activation of the Smad signaling pathway. Although the finding was subsequently confirmed in a few papers, the underlying mechanism has not been explored. In the present study, the authors found that Rudhira/breast carcinoma amplified sequence 3 is involved in the release of Smad2/3 from microtubules in response to TGF-β stimulation. Rudhira is also induced by TGF-β and is probably involved in the stabilization of microtubules in the delayed phase after TGF-β stimulation. Therefore, Rudhira has two important functions downstream of TGF-β in the early as well as delayed phase.

Strengths:

This work aimed to address an unsolved question on one of the earliest events after TGF-β stimulation. Based on loss-of-function experiments, the authors identified a novel and potentially important player, Rudhira, in the signal transmission of TGF-β.

We thank the reviewer for the critical evaluation and appreciation of our findings.

Weaknesses:

The authors have identified a key player that triggers Smad2/3 released from microtubules after TGF-β stimulation probably via its association with microtubules. This is an important first step for understanding the regulation of Smad signaling, but underlying mechanisms as well as upstream and downstream events largely remain to be elucidated.

We acknowledge that the mechanisms regulating cytoskeletal control of Smad signaling are far from clear, but these are out of scope of this manuscript. This manuscript rather focuses on Rudhira/Bcas3 as a pivot to understand vascular TGFβ signaling and microtubule connections.

(1) The process of how Rudhira causes the release of Smad proteins from microtubules remains unclear. The statement that "Rudhira-MT association is essential for the activation and release of Smad2/3 from MTs" (lines 33-34) is not directly supported by experimental data.

We agree with the reviewer’s comment. Although we provide evidence that the loss of Rudhira (and thereby deduced loss of Rudhira-MT association) prevents release of Smad2/3 from MTs (Fig 3C), it does not confirm the requirement of Rudhira-MT association for this. In light of this, we have modified the statement to ‘Rudhira associates with MTs and is essential for the activation and release of Smad2/3 from MTs”.

(2) The process of how Rudhira is mobilized to microtubules in response to TGF-β remains unclear.

Our previous study showed that Rudhira associates with microtubules, and preferentially binds to stable microtubules (Jain et al., 2012; Joshi and Inamdar, 2019). Since TGFβ stimulation is known to stabilize microtubules, we hypothesize that TGFβ stimulation increases Rudhira binding to stable microtubules. We have mentioned this in our revised manuscript.

(3) After Rudhira releases Smad proteins from microtubules, Rudhira stabilizes microtubules. The process of how cells return to a resting state and recover their responsiveness to TGF-β remains unclear.

We show that dissociation of Smads from microtubules is an early response and stabilization of microtubules is a late TGFβ response. However, we agree that the sequence of these molecular events has not been characterized in-depth in this or any other study, making it difficult to assign causal roles (eg. whether release of Smads from MTs is a pre-requisite for MT stabilization by Rudhira) or reversibility. However, the TGFβ pathway is auto regulatory, leading to increased turnover of receptors and Smads and increased expression of inhibitory Smads, which may recover responsiveness to TGFβ. Additionally, the still short turnover time of stable microtubules (several minutes to hours) may also promote quick return to resting state.

We have discussed this in our revised manuscript.

Author response:

eLife assessment

This important study provides new insight into the dynamics that underlie the development of therapy resistance in prostate cancer by revealing that divergent tumor evolutionary paths occur in response to different treatment timing and that these converge on common resistance mechanisms. The use of barcoded lineage tracing and characterization of isolated tumor clonal populations provides compelling evidence supporting the importance of clonal dynamics in a tumor ecosystem for treatment resistance. Several open questions remain, however, raising the possibility of alternative interpretations of the data set in its current form. Overall, the findings deepen our understanding of prostate cancer evolution and hold promising implications for how drug resistance can be addressed or prevented.

We are pleased the reviewers found our work reporting distinct evolutionary paths to resistance based on timing of treatment to be important and supported by compelling evidence.  We also acknowledge the need for additional work to clarify some details, particularly regarding the mechanism of clonal cooperativity as a catalyst of resistance.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Lee, Eugine et al. use in vivo barcoded lineage tracing to investigate the evolutionary paths to androgen receptor signaling inhibition (ARSI) resistance in two different prostate cancer clinical scenario models: measurable disease and minimal residual disease. Using two prostate cancer cell lines, LNCaP/AR and CWR22PC, the authors find that in their minimal residual disease models, the outgrowth of pre-existing resistant clones gives rise to ARSI-resistant tumors. Interestingly, in their measurable disease model or post-engraftment ARSI setting, these pre-existing resistant clones are depleted and rather a subset of clones that give rise to the treatment of naïve tumors adapt to ARSI treatment and are enriched in resistant tumors. For the LNCaP/AR cell line, characterization of pre-existing resistant clones in treatment naïve and ARSI treatment settings reveal increased baseline androgen receptor transcriptional output as well as baseline upregulation of glucocorticoid receptor (GR) as the primary driver of pre-existing resistance. Similarly, the authors found induction of high GR expression over long-term ARSI treatment in ARSI-sensitive clones for adaptive resistance to ARSI. For CWR22Pc cells, HER3/NRG1 signaling was the primary driver for ARSI resistance in both measurable disease and minimal residual disease models. Not only were these findings consistent with the authors' previous reports of GR and NRG1/Her3 as the molecular drivers of ARSI resistance in LNCaP/AR and CWR22Pc, respectively, but also demonstrate conserved resistance mechanisms despite pre-existing or adaptive evolutionary paths to resistance. Lastly, the authors show adaptive ARSI resistance is dependent on interclonal cooperation, where the presence of pre-existing resistant clones or "helper" clones is required to promote adaptive resistance in ARSI-sensitive clones.

Strengths:

The authors employ DNA barcoding, powerful a tool already demonstrated by others to track the clonal evolution of tumor populations during resistance development, to study the effects of the timing of therapy as a variable on resistance evolution. The authors use barcoding in two cell line models of prostate cancer in two clinical disease scenarios to demonstrate divergent evolutionary paths converging on common resistant mechanisms. By painstakingly isolating clones with barcodes of interest to generate clonal cell lines from the treatment of naïve cell populations, the authors are able to not only characterize pre-existing resistance but also show cooperativity between resistant and drug-sensitive populations for adaptive resistance.

Weaknesses:

While the finding that different evolutionary paths result in common molecular drivers of ARSI resistance is novel and unexpected, this work primarily confirms the authors' previous published work identifying the resistance mechanisms in these cell lines. The impact of the work would be greater with additional studies understanding the specific molecular/genetic mechanisms by which cells become resistant or cooperate within a population to give rise to resistant population subclones.

We agree that additional insights into the mechanism of adaptive resistant and the role of cell-cell cooperativity are clear next steps for this work. We propose to do so through single cell characterization (RNA-seq, ATAC-seq) of tumor evolution in a time course experiment where we can track each clone using expressed barcodes. This will allow us to explore the dynamics of interaction between the "adaptable" and "helper" clones. Unfortunately, the barcode methodology used in this initial report is DNA-based; therefore, a follow-up study using a transcribable barcode library is needed to address these fascinating questions.

This study would also benefit from additional explanation or exploration of why the two resistance driver pathways described (GR and NRG1/Her3) are cell line specific and if there are genetic or molecular backgrounds in which specific resistance signaling is more likely to be the predominant driver of resistance.

In the case of NRG1/HER3 pathway mediated resistance, we know that this mechanism requires that the PTEN/PIK3CA pathway be wildtype.  This is the case for the CWR22Pc model described in the manuscript. Furthermore, we have data showing that PTEN deletion in these cells rescues the phenotype, meaning that CWR22Pc cells with PTEN deletion are no longer dependent on NRG1/HER3 signaling for ARSI resistance.

In contrast, LNCaP/AR cells are PTEN null at baseline and therefore must evolve alternative mechanisms of ARSI resistance. Since our initial identification of the GR mechanism, we and others have extended the finding to additional models (VCaP, LAPC4) (PMID: 24315100; PMID: 28191869). Another recent insight is the importance of RB1 and TP53 status in maintenance of luminal lineage identity during ARSI therapy, and the recognition of lineage plasticity as a resistance mechanism in cell lines/tumor models that lack these two tumor suppressors. In summary, baseline genetics clearly plays a role in which ARSI resistance pathway is  likely to emerge. We will clarify this point in the revision with additional discussion.

Reviewer #2 (Public Review):

Summary

The authors aimed to characterise the evolutionary dynamics that occur during the resistance to androgen receptor signalling inhibition, and how this differs in established tumours vs. residual disease, in prostate cancer. By using a barcoding method, they aimed to both characterise the distribution of clones that support therapy resistance in these settings, while also then being able to isolate said clones from the pre-graft population via single-cell cloning to characterise the mechanisms of resistance and dependency on cooperativity.

While, interestingly, the timing of combination therapies has been shown to be critical to avoid cross-resistance, the timing of therapy has not been specifically considered as a factor dictating resistance pathways. Additionally, the role of residual disease and dormant populations in driving relapse is of increasing interest, yet a lot remains to be understood of these populations. The question of whether different clinical manifestations of therapy resistance follow similar evolutionary pathways to resistance is therefore interesting and relevant for the field.

The methods applied are elegant and the body of work is substantial. The proposed divergent evolutionary pathways pose interesting questions, and the findings on cooperativity provide insight. However, whether the model truly reflects minimal residual disease to the extent that the authors suggest may limit the relevance of the findings at this stage. Certain patterns in the DNA barcoding results also call into question whether the results fully support the strong claims of the authors, or whether alternative explanations could exist. While the potential to isolate individual clones in the pre-graft setting is a great strength of the method applied and the isolation of these clones is a huge body of work in itself, the limited number of clones that could be isolated also somewhat limits the validation of the findings.

Strengths

Very relevant and interesting question, clear clinical relevance, applying elegant methods that hold the potential to provide a novel understanding of multiple aspects of therapy resistance, through from evolutionary patterns to intracellular and cooperative mechanisms of resistance.

The text is clearly written, logical, and the structure is easy to follow.

Weaknesses

(1) The extent to which the model used truly mimics residual disease

The main conclusions of the paper are built upon results using a model for minimal residual disease. However, the extent to which this truly recapitulates minimal residual disease, particularly with regard to their focus on the timings of therapy, could be discussed further. If in the clinical setting residual disease occurs following the existence of a tumour and its microenvironment, there might be many aspects of the process that are missed when coinciding treatment with engraftment of a xenograft tumour with pre-castration. If any characterisation of the minimal residual disease was possible (such as histologically or through RNA sequencing), this may help demonstrate in what ways this model recapitulates minimal residual disease.

We appreciate the reviewer's feedback on this point and acknowledge that the pre-ARSI setting used in our studies is not precisely identical to minimal residual disease (MRD) seen clinically, where a patient typically undergoes primary treatment (radical prostatectomy surgery or local radiotherapy) then relapses with distant disease from micrometastases that were not initially detectable.  Having uncovered a key difference in the path to resistance using our pre-ARSI model, we believe our data provide a strong rationale to invest additional effort in designing newer MRD models that more closely mimic the clinical scenario, perhaps through surgical resection of a primary tumor that could “seed” micrometatases prior to therapy. We will highlight this aspect in our revised manuscript and provide clarity on the limitations and scope of our study.

(2) Whether the observed enrichment of pre-resistant clones is truly that

The authors strongly make the case that their barcoding experiments provide evidence for pre-existing resistance in the context of minimal residual disease. However, it seems that the clones enriched in the ARSIR tumours are consistently the most enriched clones in the pregraft. Is it possible that the high selective pressure in the pre-engraftment ARSI condition simply leads to an enrichment of the most populous clones from the pregraft? Whereas in the control setting, the reduced selective pressure at the point of engraftment allows for a wider variety of clones to establish in the tumour?

The reviewer raises an important point about enrichment of ARSI resistance clones in the pregraft but we do not believe that explains the subsequent in vivo data for the following reasons:

(1) The two most enriched clones in the Pre-ARSIR tumors are the second and third the most enriched clones in pre-graft, not first (Supplementary figure 1E). If the clones were enriched in resistant tumors based on their abundance in starting population, we expect to find the most enriched clone in the tumor.

(2) By varying the androgen concentration in the pregraft culture media, we could selectively deplete or enrich the same clones enriched in the Pre-ARSIR tumors in vivo, indicating the enrichment of these clones in the resistant tumors is unlikely to be solely based on their relative frequency in the pregraft (Supplementary figure 2).

We will clarify these points in the revised manuscript.

Additionally, is there the possibility that the clones highly enriched in the pregraft are in fact a heterogeneous group of cells bearing the same barcode due to stochastic events in the process of viral transduction? Addressing these questions would greatly improve the study.

The barcode library was deep sequenced to confirm even distribution of the barcodes before it was transferred from Novartis (PMID: 258491301) and we intentionally used a low multiplicity of infection (MOI) to generate barcode lines to ensure single copy insertion. That said, we cannot entirely rule out the possibility that the second and third most enriched clones in the pregraft originated from the same ancestral clone and subsequently acquired two different barcodes.  We will clarify this point in the revised manuscript.

(3) The robustness of the subsequent work based on 1-2 pre-resistant clones

While appreciating the volume of work involved in isolating and culturing individual pre-resistant clones, given the previous point, the conclusions would benefit from very robust validations with these single-cell clones. There are only two clones, and the results seem to focus more on one than the other, for which the data is less convincing. For instance, the Enz IC50 data, which in the case for pre-ARSI R2 is restricted to the supplementary, compares the clones A-D. In Figure S8 B, pre-ARSI R2 is compared to clone B, which is, of the four clones shown in the main figure when compared to R1, the one with the lowest Enz IC50. Therefore, while the resistant clones seem to have a significantly higher Enz IC50, comparing both clones to clones A-D may not have achieved this significance. It would also be useful to know how abundant the resistant clones were in the original barcode experiments.

We acknowledge that studies relying on 1-2 biological samples indeed have limitations. Given our extensive prior work into the role of GR in the development of ARSI resistance (and that of other labs), we focused on demonstrating that both pre-ARSIR1 and pre-ARSIR2 clones exhibit pre-existing GR expression and are primed to further upregulate GR levels under ARSI conditions, thereby relying on GR function to sustain resistance. Given the redundancy of resistant mechanisms of the two clones, we made efforts to isolate additional clones enriched in Pre-ARSIR tumors. However, despite our attempts, we were unable to identify further clones. Pre-ARSIR1 and pre-ARSIR2 are second and third most enriched clones in pre-graft (2.1% and 1.7% respectively).

(4) The logic used in the final section requires further explanation

In the final section, the authors suggest that a pre-ARSIR clone is able to cooperate with a pre-Intact clone to aid adaptive ARSI resistance. If this is true, then could it not be that rare, pre-resistant clones support adaptive resistance in established tumours? And, therefore, the mechanism underlying resistance could be through pre-existing resistant clones in both settings. The work would benefit from a discussion to clarify this discrepancy in the interpretation of the findings. This is particularly necessary given the strong wording the authors use regarding their findings, such as that they have provided 'conclusive evidence' for acquired resistance.

We agree that rare, pre-resistant clones could support adaptive resistance (and therefore resistance in this adaptive setting could, technically be called “pre-existing”) but it is critical to recognize that these rare, pre-resistant “helper” clones are vastly outnumbered by pre-Intact clones that “acquire” resistance through their “help.” We find this to be fascinating biology and we will clarify this logic in the resubmission, as well as future experimental approaches to unravel the mechanism.

Author response:

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

Reviewer #1 (Public Review):

Summary:

In this manuscript, Chowdhury and co-workers provide interesting data to support the role of G4-structures in promoting chromatin looping and long-range DNA interactions. The authors achieve this by artificially inserting a G4-containing sequence in an isolated region of the genome using CRISPR-Cas9 and comparing it to a control sequence that does not contain G4 structures. Based on the data provided, the authors can conclude that G4-insertion promotes long-range interactions (measured by Hi-C) and affects gene expression (measured by qPCR) as well as chromatin remodelling (measured by ChIP of specific histone markers).

Whilst the data presented is promising and partially supports the authors' conclusion, this reviewer feels that some key controls are missing to fully support the narrative. Specifically, validation of actual G4-formation in chromatin by ChIP-qPCR (at least) is essential to support the association between G4-formation and looping. Moreover, this study is limited to a genomic location and an individual G4-sequence used, so the findings reported cannot yet be considered to reflect a general mechanism/effect of G4-formation in chromatin looping.

Strengths:

This is the first attempt to connect genomics datasets of G4s and HiC with gene expression. The use of Cas9 to artificially insert a G4 is also very elegant.

Weaknesses:

Lack of controls, especially to validate G4-formation after insertion with Cas9. The work is limited to a single G4-sequence and a single G4-site, which limits the generalisation of the findings.

In the revised version we validated G4 formation inside cells at the insertion site using the reported G4-selective antibody BG4. Significant BG4 binding (by ChIP-qPCR) was clear in the G4-array insert, and not in the G4-mutated insert, supporting formation of G4s by the inserted G4-array (included as Figure S4).

To directly address the second point, we inserted the G4-sequence, or the mutated control, at a second relatively isolated locus (at the 10 millionth position on Chr12, denoted as 10M site in text). First, BG4 ChIP was done to confirm intracellular G4 formation by the inserted array. BG4 ChIP-qPCR binding was significant within the inserted region, and not in the negative control region (Figure S8), consistent with the 79M locus. Together these demonstrate intracellular G4 formation by inserted sequences at two different loci.

We next checked the state of chromatin of the G4-array inserted at the 10M locus, or its negative control. Histone marks H3K4Me1, H3K27Ac, H3K27Me3, H3K9me3 and H3K4Me3 were tested at the G4-array, or the negative control locus. Relative increase in the enhancer histone marks was evident, relative to the control sequence. This was largely similar to the 79M locus, supporting an enhancer-like state. Interestingly, here we further noted presence of the H3K27me3 histone mark. The presence of the H3K27Me3 repressor histone mark, along with H3K4Me1/H3K27Ac enhancer histone marks, support a poised enhancer-like status of the inserted G4 region, as has been observed earlier in other studies. Together, although data from the two distinct G4 insertion sites support the enhancer-like state, there are contextual differences likely due to the sequence/chromatin of the sites adjacent to the inserted sequence.

Effect of the 10M G4-insertion on activation of surrounding genes (10 Mb window), and not the G4-mutant insert, was evident for most genes. Consistent with the enhancer-like state of the G4-array insert; in line with the 79M G4-array insert.

These results have been added as the final section in the revised version, data is shown in Figure 7.

Reviewer #2 (Public Review):

Summary:

Roy et al. investigated the role of non-canonical DNA structures called G-quadruplexes (G4s) in long-range chromatin interactions and gene regulation. Introducing a G4 array into chromatin significantly increased the number of long-range interactions, both within the same chromosome (cis) and between different chromosomes (trans). G4s functioned as enhancer elements, recruiting p300 and boosting gene expression even 5 megabases away. The study proposes a mechanism where G4s directly influence 3D chromatin organization, facilitating communication between regulatory elements and genes.

Strength:

The findings are valuable for understanding the role of G4-DNA in 3D genome organization and gene transcription.

Weaknesses:

The study would benefit from more robust and comprehensive data, which would add depth and clarity.

(1) Lack of G4 Structure Confirmation: The absence of direct evidence for G4 formation within cells undermines the study's foundation. Relying solely on in vitro data and successful gene insertion is insufficient.

Using the reported G4-specific antibody, BG4, we performed BG4 ChIP-qPCR at the 79M locus. In addition, a second G4-insertion site was created and BG4 ChIP-qPCR was used to validate intracellular G4 formation. Briefed below, more details in the response above.

In the revised version we validated G4 formation inside cells at the insertion site using the reported G4-selective antibody BG4. Significant BG4 binding (by ChIP-qPCR) was clear in the G4-array insert, and not in the G4-mutated insert, supporting formation of G4s by the inserted G4-array (included as Figure S4).

Further, we inserted the G4-sequence, or the mutated control, at a second relatively isolated locus (at the 10 millionth position on Chr12, denoted as 10M site in text). First, BG4 ChIP was done to confirm intracellular G4 formation by the inserted array. BG4-ChIP-qPCR was significant within the G4-array inserted region, and not in the negative control region (Figure S8), consistent with the 79M locus. Together these demonstrate intracellular G4 formation by inserted sequences at two different loci. Added in revised text in the second and the final sections of results, data shown in Figures 7, S4 and S8.

(2) Alternative Explanations: The study does not sufficiently address alternative explanations for the observed results. The inserted sequences may not form G4s or other factors like G4-RNA hybrids may be involved.

As mentioned in response to the previous comment, we confirmed that the inserted sequence indeed forms G4s inside the cells. RNA-DNA hybrid G4s can form within R-loops with two or more tandem G-tracks (G-rich sequences) on the nascent RNA transcript as well as the non-template DNA strand (Fay et al., 2017, 28554731). A recent study has observed that R-loop-associated G4 formation can enhance chromatin looping by strengthening CTCF binding (Wulfridge et al., 2023, 37552993). As pointed out by the reviewer, the possibility of G4-RNA hybrids remains, we have mentioned this possibility for readers in the second last paragraph of the Discussion.

(3) Limited Data Depth and Clarity: ChIP-qPCR offers limited scope and considerable variation in some data makes conclusions difficult.

We noted variation with one of the primers in a few ChIP-qPCR experiments (in Figures 2 and 3D). The changes however were statistically significant across replicates, and consistent with the overall trend of the experiments (Figures 2, 3 and 4). Enhancer function, in addition to ChIP, was also confirmed using complementary assays like 3C and RNA expression.

(4) Statistical Significance and Interpretation: The study could be more careful in evaluating the statistical significance and magnitude of the effects to avoid overinterpreting the results.

We reconfirmed our statistical calculations from biological replicate experiments. We carefully looked at potential overinterpretations, and made appropriate changes in the manuscript (details of the changes given below in response to comment to authors).

Reviewer #3 (Public Review):

Summary:

This paper aims to demonstrate the role of G-quadruplex DNA structures in the establishment of chromosome loops. The authors introduced an array of G4s spanning 275 bp, naturally found within a very well-characterized promoter region of the hTERT promoter, in an ectopic region devoid of G-quadruplex and annotated gene. As a negative control, they used a mutant version of the same sequence in which G4 folding is impaired. Due to the complexity of the region, 3 G4s on the same strand and one on the opposite strand, 12 point mutations were made simultaneously (G to T and C to A). Analysis of the 3D genome organization shows that the WT array establishes more contact within the TAD and throughout the genome than the control array. Additionally, a slight enrichment of H3K4me1 and p300, both enhancer markers, was observed locally near the insertion site. The authors tested whether the expression of genes located either nearby or up to 5 Mb away was up-regulated based on this observation. They found that four genes were up-regulated from 1.5 to 3-fold. An increased interaction between the G4 array compared to the mutant was confirmed by the 3C assay. For in-depth analysis of the long-range changes, they also performed Hi-C experiments and showed a genome-wide increase in interactions of the WT array versus the mutated form.

Strengths:

The experiments were well-executed and the results indicate a statistical difference between the G4 array inserted cell line and the mutated modified cell line.

Weaknesses:

The control non-G4 sequence contains 12 point mutations, making it difficult to draw clear conclusions. These mutations not only alter the formation of G4, but also affect at least three Sp1 binding sites that have been shown to be essential for the function of the hTERT promoter, from which the sequence is derived. The strong intermingling of G4 and Sp1 binding sites makes it impossible to determine whether all the observations made are dependent on G4 or Sp1 binding. As a control, the authors used Locked Nucleic Acid probes to prevent the formation of G4. As for mutations, these probes also interfere with two Sp1 binding sites. Therefore, using this alternative method has the same drawback as point mutations. This major issue should be discussed in the paper. It is also possible that other unidentified transcription factor binding sites are affected in the presented point mutants.

Since the sequence we used to test the effects of G4 structure formation is highly G-rich, we had to introduce at least 12 mutations to be sure that a stable G4 structure would not form in the mutated control sequence. Sp1 has been reported to bind to G4 structures (Raiber et al., 2012). Therefore, Sp1 binding is likely to be associated with the G4-dependent enhancer functions observed here. We also appreciate that apart from Sp1, other unidentified transcription factor binding sites might be affected by the mutations we introduced. We have discussed these possibilities in the fourth paragraph of the Discussion section in the revised manuscript.

Reviewer #1 (Recommendations For The Authors):

Whilst the data presented is promising and partially supports the authors' conclusion, this reviewer feels that some key controls are missing to fully support the narrative used. Below are my main concerns:

(1) The main thing missing in the current manuscript is to validate the actual formation of G4 in chromatin context for the repeat inserted by CRISPR-Cas. Whilst I appreciate this will form promptly a G4 in vitro, to fully support the conclusions proposed the authors would need to demonstrate actual G4-formation in cells after insertion. This could be done by ChIP-qPCR using the G4-selective antibody BG4 for example. This is an essential piece of evidence to be added to link with confidence G4-formation to chromatin looping.

To address the concern regarding whether the inserted G4 sequence forms G4s in cells, as suggested, we used the G4-selective antibody BG4. PCR primers in the study were designed keeping multiple points in mind: Primers should not bind to any site of G/C alteration in the mutated control insert; either the forward/reverse primer is from the adjacent region for specificity; covers adjacent regions for studying any effects on chromatin; and, PCRs optimized keeping in mind the repeats within the inserted sequence. Given these, primer pairs R1-R4 were chosen for further work following optimizations (Figure 2, top panel). For BG4 ChIP-qPCR we used primer pairs R2, which covered >100 bases of the inserted G4-array, or the G4-mutated control. Significant BG4 binding was clear in the G4-array insert, and not in the G4-mutated insert, demonstrating formation of G4s by the inserted G4-array (Figure S4).

In response to comment #3 below, we inserted the G4-forming sequence (or its mutated control) at a second locus. This insertion was near the 10 millionth position of chromosome 12 (10M insertion locus in text). Here also, BG4 binding was significant within the G4-array inserted region, and not in the negative control region (Figure S8). Together these demonstrate G4 formation by the inserted sequence at two different loci.

(2) I found the LNA experiment very elegant. However, what would be the effect of LNA treatment on the control sequence that does not form G4s? This control is essential to disentangle the effect of LNA pairing to the sequence itself vs disrupting the G4-structure.

As per the reviewer’s suggestion, we performed a control experiment where we treated the G4-mutated insert (control) cells with the G4-disrupting LNA probes. The changes in the expression of the surrounding genes in this case were not significant, indicating that the effects observed in the G4-array insert cells were possibly due to disruption of the inserted G4 structures. This data is presented in Figure S5.

(3) The authors describe their work and present its conclusion as if this were a genome-wide study, whilst the work is focused on a specific genomic location, and the looping, along with the effect on histone acetylation and gene expression, is limited to this. The authors cannot conclude, therefore, that this is a generic effect and the discussion should be more focused on the specific G4s used and the genomic location investigated. Ideally, insertion of a different G4-forming sequence or of the same in a different genomic location is recommended to really claim a generic effect.

To address this we inserted the G4-array sequence, or the G4-mutated control sequence, at another relatively isolated locus – at the 10 millionth position of chromosome 12 – denoted as 10M. Using BG4 ChIP-qPCR intracellular G4 formation was confirmed. We observed that the enhancer-like features in terms of enhancer histone marks and increase in the expression of surrounding genes were largely reproduced at the 10M locus on G4 insertion (Figure 7). These results are added as the final section under Results.

Reviewer #2 (Recommendations For The Authors):

The study proposes a mechanism where G4s directly influence 3D chromatin organization, facilitating communication between regulatory elements and genes.

While the present manuscript presents an interesting hypothesis, it would benefit from enhanced novelty and more robust data. The study complements existing G4 research (e.g., PMID: 31177910). While the conclusions hold biological relevance, they largely reiterate established knowledge. Furthermore, the presented data appear preliminary and still lack depth and clarity.

Hou et al., 2019 (PMID: 31177910) showed presence of potential G4-forming sequences correlated with TAD boundaries, along with enrichment of architectural proteins and transcription factor binding sites. Also, other studies noted enrichment of potential G4-forming sequences at enhancers along with nucleosome depletion and higher transcription factor binding (Hou et al., 2021; Williams et al., 2020). These studies proposed the role of G4s in chromatin/TAD states based on analysis of potential G4-forming sequences using correlative bioinformatics analyses. Here we sought to directly test causality. Insertion of G4 sequence, and formation of intracellular G4s in an isolated, G4-depleted region resulted in altered characteristics of chromatin, and not in the negative control insertion that does not form G4s. These, in contrast to earlier studies, directly demonstrates the causal role of G4s as functional elements that impact local and distant chromatin.

Major concerns:

(1) Lack of G4 Structure Confirmation: Implement G4-specific antibodies or fluorescent probes to verify G4 structures inside the cells.

Detailed response given above. Briefly, in the revised version we validated G4 formation inside cells at the insertion site using the reported G4-selective antibody BG4. Significant BG4 binding (by ChIP-qPCR) was clear in the G4-array insert, and not in the G4-mutated insert, supporting formation of G4s by the inserted G4-array (included as Figure S4).

Further, we inserted the G4-sequence, or the mutated control, at a second relatively isolated locus (at the 10 millionth position on Chr12, denoted as 10M site in text). First, BG4 ChIP was done to confirm intracellular G4 formation by the inserted array. BG4 ChIP-qPCR binding was significant within the G4-array inserted region, and not in the negative control region (Figure S8), consistent with the 79M locus. Together these demonstrate intracellular G4 formation by inserted sequences at two different loci. Added in revised text in the second and the final sections of results, data shown in Figures 7, S4 and S8.

(2) Alternative Explanations: Explore the possibility that the sequences may not form G4s or that other factors like G4-RNA hybrids are involved.

Response provided in the public reviews section.

(3) Limited Data Depth and Clarity: ChIP-qPCR offers limited scope. Consider employing G4 ChIP-seq for genome-wide analysis of G4 association with histone modifications. Address inconsistencies in data like H3K27me3 variation and incomplete H3K9me3 data sets.

A recent study performed G4 CUT&Tag (Lyu et al., 2022, 34792172) and observed G4 formation at both active promoters and active and poised enhancers. We have discussed this in the sixth paragraph of the Discussion. The H3K27Me3 occupancy at the 79M locus insertions did not have any significant G4-dependent changes, however, at the second insertion site at the 10M locus (introduced in the revised manuscript, Figure 7) there was significant G4-dependent increase in H3K27Me3 occupancy along with the H3K4Me1 and H3K27Ac enhancer histone marks, indicating formation of a poised enhancer-like element.

We completed the H3K9me3 data sets for both insertion sites.

(4) Statistical Significance and Interpretation: Re-evaluate the statistical significance of results and interpret them in the context of relevant biological knowledge. Avoid overstating the impact of minor changes.     

We revised several lines to avoid overstating results. Some of the changes are as below (changes underline/strikethrough)

- There was an a relatively modest increase in the recruitment of both p300 and a substantial increase in the recruitment of the more functionally active acetylated p300/CBP to the G4-array when compared against the mutated control.

- As expected, although modest, a decrease in the H3K4Me1 and H3K27Ac enhancer histone modifications was evident within the insert upon the LNAs treatment.

- Moreover, the enhancer marks were relatively reduced, although not markedly, when the inserted G4s were specifically disrupted.

(5) Unexplored Aspects: Investigate the relationship between G4 DNA and R-loops, and consider the role of CTCF and cohesin proteins in mediating long-range interactions. Integrate existing research to build a more comprehensive framework and draw more robust conclusions.

As mentioned in response to one of the earlier comments, a recent publication extensively studied the association between G4s, R-loops, and CTCF binding (Wulfridge et al., 2023). While, here we focused on the primary features of a potential enhancer, further work will be necessary to establish how G4s influence the coordinated action between cohesin and CTCF and consequent chromatin looping. We have described this for readers in the second last paragraph of the Discussion in the revised version.

Minor Concern:

(1) Enhancer Definition: The term "enhancer" requires specific criteria. Modify the section heading or provide evidence demonstrating the G4 sequence fulfills all conditions for being an enhancer, such as position independence and long-range effects.

Although we checked some of the primary features of a potential enhancer: Like expression of surrounding genes, enhancer histone marks, chromosomal looping interactions, and recruitment of transcriptional coactivators, further aspects may need to be validated. As suggested, in the revised manuscript the section heading has been modified to ‘Enhancer-like features emerged upon insertion of G4s.’

Reviewer #3 (Recommendations For The Authors):

In addition to the points in my public review, I would like to mention some less significant points.

The authors mention that "the array of G4-forming sequences used for insertion was previously reported to form stable G4s in human cells" (Lim et al., 2010; Monsen et al., 2020; Palumbo et al., 2009). However, upon reading the publications, I found that these observations were made in vitro. I may have missed something, but there are now several mappings of folded-G4 in human cells based on different approaches. It would be beneficial to investigate whether the hTERT promoter is a site of G-quadruplex formation in vivo. If confirmed, a similar analysis should be conducted on the 275 bp region inserted into the ectopic region to determine if it also has the ability to form a structured G4.

We performed BG4 ChIP to confirm in vivo G4 formation by the inserted G4-array as suggested (Figures S4, S8). Detail response given above. Briefly, in the revised version we validated G4 formation inside cells at the insertion site using the reported G4-selective antibody BG4. Significant BG4 binding (by ChIP-qPCR) was clear in the G4-array insert, and not in the G4-mutated insert, supporting formation of G4s by the inserted G4-array (included as Figure S4).

Further, we inserted the G4-sequence, or the mutated control, at a second relatively isolated locus (at the 10 millionth position on Chr12, denoted as 10M site in text). First, BG4 ChIP was done to confirm intracellular G4 formation by the inserted array. BG4-ChIP-qPCR was significant within the inserted region, and not in the negative control region (Figure S8). Consistent with the 79M locus. Together these demonstrate intracellular G4 formation by inserted sequences at two different loci. Added in revised text in the second and the final sections of results, data shown in Figures 7, S4 and S8.

The inserted sequence originates from a well-characterized promoter. The authors suggest that placing it in an ectopic position creates an enhancer-like region, based on the observation of increased levels of H3K27Ac and H3K4me1 on the WT array. To provide a control that it is not a promoter, it would be useful to also analyze a specific mark of promoter activity, such as H3K4me3.

As suggested by reviewer, we also analysed the H3K4Me3 promoter activation mark at both the 79M and 10M (introduced in the revised manuscript, Figure 7) insertion loci. We did not observe any significant G4-dependent changes in the recruitment of H3K4Me3 (Figures 2, 7).

In the discussion, the authors mention "it was proposed that inter-molecular G4 formation between distant stretches of Gs may lead to DNA looping". To investigate this further, it would be worthwhile to examine whether the promoter regions of activated genes (PAWR, PPP1R12A, NAV3, and SLC6A15) contain potentially forming G-quadruplexes (pG4). Additionally, sites that establish more contact with the G4 array described in Figure 6F could be analyzed for enrichment in pG4.

Thank you for pointing this out. We found promoters of the four genes (PAWR, PPP1R12A, NAV3, and SLC6A15) harbour potential G4-forming sequences (pG4s). Also as suggested, we analysed the contact regions in Fig 6F, along with the whole locus, for pG4s. Relative enrichment in pG4 was seen, particularly within the significantly enhanced interacting regions, which at times spreads beyond the interacting regions also. This is shown in the lower panel of Figure 6F in the revised version. We have described this in Discussion for readers.

Author response:

eLife assessment

This important study addresses the idea that defective lysosomal clearance might be causal to renal dysfunction in cystinosis. They observe that restoring expression of vATPase subunits and treatment with Astaxanthin ameliorate mitochondrial function in a model of renal epithelial cells, opening opportunities for translational application to humans. The data are convincing, but the description of methodologies is incomplete.

Public Reviews:

Reviewer #1 (Public Review):

Cystinosis is a rare hereditary disease caused by biallelic loss of the CTNS gene, encoding two cystinosin protein isoforms; the main isoform is expressed in lysosomal membranes where it mediates cystine efflux whereas the minor isoform is expressed at the plasma membrane and in other subcellular organelles. Sur et al proceed from the assumption that the pathways driving the cystinosis phenotype in the kidney might be identified by comparing the transcriptome profiles of normal vs CTNS-mutant proximal tubular cell lines. They argue that key transcriptional disturbances in mutant kidney cells might not be present in non-renal cells such as CTNS-mutant fibroblasts.

Using cluster analysis of the transcriptomes, the authors selected a single vacuolar H+ATPase (ATP6VOA1) for further study, asserting that it was the "most significantly downregulated" vacuolar H+ATPase (about 58% of control) among a group of similarly downregulated H+ATPases. They then showed that exogenous ATP6VOA1 improved CTNS(-/-) RPTEC mitochondrial respiratory chain function and decreased autophagosome LC3-II accumulation, characteristic of cystinosis. The authors then treated mutant RPTECs with 3 "antioxidant" drugs, cysteamine, vitamin E, and astaxanthin (ATX). ATX (but not the other two antioxidant drugs) appeared to improve ATP6VOA1 expression, LC3-II accumulation, and mitochondrial membrane potential. Respiratory chain function was not studied. RTPC cystine accumulation was not studied.

In this manuscript, as an initial step, we have studied the first step in respiratory chain function by performing the Seahorse Mito Stress Test to demonstrate that the genetic manipulation (knocking out the CTNS gene and plasmid-mediated expression correction of ATP6V0A1) impacts mitochondrial energetics. We did not investigate the respirometry-based assays that can identify locations of electron transport deficiency, which we plan to address in a follow-up paper.

We would like to draw attention to Figure 3D, where cystine accumulation has been studied. This figure demonstrates an increased intracellular accumulation of cystine.

The major strengths of this manuscript reside in its two primary findings.

(1) Plasmid expression of exogenous ATP6VOA1 improves mitochondrial integrity and reduces aberrant autophagosome accumulation.

(2) Astaxanthin partially restores suboptimal endogenous ATP6VOA1 expression.

Taken together, these observations suggest that astaxanthin might constitute a novel therapeutic strategy to ameliorate defective mitochondrial function and lysosomal clearance of autophagosomes in the cystinotic kidney. This might act synergistically with the current therapy (oral cysteamine) which facilitates defective cystine efflux from the lysosome.

There are, however, several weaknesses in the manuscript.

(1) The reductive approach that led from transcriptional profiling to focus on ATP6VOA1 is not transparent and weakens the argument that potential therapies should focus on correction of this one molecule vs the other H+ ATPase transcripts that were equally reduced - or transcripts among the 1925 belonging to at least 11 pathways disturbed in mutant RPTECs.

The transcriptional profiling studies on ATP6V0A1 have been fully discussed and publicly shared. Table 2 lists the v-ATPase transcripts that are significantly downregulated in cystinosis RPTECs. We have also clarified and justified the choice of further studies on ATP6V0A1, where we state the following: "The most significantly perturbed member of the V-ATPase gene family found to be downregulated in cystinosis RPTECs is ATP6V0A1 (Table 2). Therefore, further attention was focused on characterizing the role of this particular gene in a human in vitro model of cystinosis."

(2) A precise description of primary results is missing -- the Results section is preceded by or mixed with extensive speculation. This makes it difficult to dissect valid conclusions from those derived from less informative experiments (eg data on CDME loading, data on whole-cell pH instead of lysosomal pH, etc).

We appreciate the reviewer highlighting areas for further improving the manuscript's readership. In our resubmission, we have revised the results section to provide a more precise description of the primary findings and restrict the inferences to the discussion section only.

(3) Data on experimental approaches that turned out to be uninformative (eg CDME loading, or data on whole=cell pH assessment with BCECF).

We have provided data whether it was informative or uninformative. Though lysosome-specific pH measurement would be important to measure, it was not possible to do it in our cells as they were very sick and the assay did not work. Hence we provide data on pH assessment with BCECF, which measures overall cytoplasmic and organelle pH, which is also informative for whole cell pH that is an overall pH of organelle pH and cytoplasmic pH.

(4) The rationale for the study of ATX is unclear and the mechanism by which it improves mitochondrial integrity and autophagosome accumulation is not explored (but does not appear to depend on its anti-oxidant properties).

We have provided rationale for the study of ATX; provided in the introduction and result section, where we mentioned the following: “correction of ATP6V0A1 in CTNS-/- RPTECs and treatment with antioxidants specifically, astaxanthin (ATX) increased the production of cellular ATP6V0A1, identified from a custom FDA-drug database generated by our group, partially rescued the nephropathic RPTEC phenotype. ATX is a xanthophyll carotenoid occurring in a wide variety of organisms. ATX is reported to have the highest known antioxidant activity and has proven to have various anti-inflammatory, anti-tumoral, immunomodulatory, anti-cancer, and cytoprotective activities both in vivo and in vitro”.

We are still investigating the mechanism by which ATX improves mitochondrial integrity and this will be the focus of a follow-on manuscript.

(5) Thoughtful discussion on the lack of effect of ATP6VOA1 correction on cystine efflux from the lysosome is warranted, since this is presumably sensitive to intralysosomal pH.

We have provided a thoughtful discussion in the revised manuscript on some possible mechanisms that may result in an effect of ATP6V0A1 correction on cysteine efflux from the lysosome.

(6) Comparisons between RPTECs and fibroblasts cannot take into account the effects of immortalization on cell phenotype (not performed in fibroblasts).

The purpose of examining different tissue sources of primary cells in nephropathic cystinosis was to assess if any of the changes in these cells were tissue source specific. We used primary cells isolated from patients with nephropathic cystinosis—RPTECs from patients' urine and fibroblasts from patients' skin—these cells are not immortalized and can therefore be compared. This is noted in the results section - “Specific transcriptional signatures are observed in cystinotic skin-fibroblasts and RPTECs obtained from the same individual with cystinosis versus their healthy counterparts”.

We next utilized the immortalized RPTEC cell line to create CRISPR-mediated CTNS knockout RPTECs as a resource for studying the pathophysiology of cystinosis. These cells were not compared to the primary fibroblasts.

(7) This work will be of interest to the research community but is self-described as a pilot study. It remains to be clarified whether transient transfection of RPTECs with other H+ATPases could achieve results comparable to ATP6VOA1. Some insight into the mechanism by which ATX exerts its effects on RPTECs is needed to understand its potential for the treatment of cystinosis.

In future studies we will further investigate the effect of ATX on RPTECs for treatment of cystinosis- this will require the conduct of Phase 1 and Phase 2 clinical studies which are beyond the scope of this current manuscript.

Reviewer #2 (Public Review):

Sur and colleagues investigate the role of ATP6V0A1 in mitochondrial function in cystinotic proximal tubule cells. They propose that loss of cystinosin downregulates ATP6V0A1 resulting in acidic lysosomal pH loss, and adversely modulates mitochondrial function and lifespan in cystinotic RPTECs. They further investigate the use of a novel therapeutic Astaxanthin (ATX) to upregulate ATP6V0A1 that may improve mitochondrial function in cystinotic proximal tubules.

The new information regarding the specific proximal tubular injuries in cystinosis identifies potential molecular targets for treatment. As such, the authors are advancing the field in an experimental model for potential translational application to humans.

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Weaknesses:

The authors fail to truly define codon optimality, rare codons, and stalling sequences in their work, all of which are distinct terminologies. They use reporters with rare codon usage but do not mention what metrics they use to determine this, such as cAI, codon usage bias, or tAI. The distinction between the type of codon sequences that DDX6 affects is very important to differentiate and should be done here as certain stretches of codons are known to lead to different quality control RNA decay pathways that are not reliant on canonical mRNA decay factors.

Thank you for the reviewer’s feedback on our work. Clearly defining codon optimality, rare codons, and stalling sequences is indeed crucial. We will emphasize this distinction more in our revisions to help readers better understand our analysis and findings.

Likewise, the authors sort their Ribo-seq data to determine genes that might exhibit a DDX6 specific mRNA decay effect but fail to go into great depth about common features shared among these genes other than GO term analysis, GC content, and coding sequence (CDS) length. The authors then sort out 35 genes that are both upregulated at the mRNA level and have increased local ribosome footprint along the ORF. They are then able to show that 6 out of 9 of those genes had a DDX6-dependent mRNA decay effect. There was no comment or effort as to why 2 out of those 6 genes tested did not show as strong of a DDX6-dependent decay effect relative to the other targets tested. Thus, the efforts to identify mRNA features at a global level that exhibited DDX6-dependent mRNA decay effects are lacking in this analysis.

We appreciate the reviewer's insightful comments regarding the need to further characterize the genes influenced by DDX6-mediated mRNA decay. To address this, we carried out additional analyses to identify potential traits of these genes. Our findings revealed that DDX6-regulated coding sequences tend to be longer and exhibit lower predicted mRNA stability scores compared to the average across the transcriptome. This observation indicates a possible connection to codon optimality. It suggests that DDX6 could play a role in regulating a specific subset of mRNAs with inherently lower stability, potentially shedding light on why some genes may exhibit varied decay patterns when DDX6 is depleted.

Overall, the work done by Weber et al. is sound, with the proper controls. The authors expand significantly on the knowledge of what we know about DDX6 in the process of mRNA decay in humans, confirming the evolutionary conservation of the role of this factor across eukaryotes. The analysis of the RNA-seq and Ribo-seq data could be more in-depth, however, the authors were able to show with certainty that some transcripts containing known repetitive sequences or polybasic sequences exhibited a DDX6-mRNA decay effect.

We appreciate the reviewer’s acknowledgment of the soundness of our work and the inclusion of proper controls. We are committed to refining our manuscript to meet your expectations and ensure the accuracy and depth of our findings.

Reviewer #2 (Public Review):

The experiments were well-performed, and the results clearly demonstrated the requirement of DDX6 in mRNA degradation induced by slowed ribosomes. However, in some cases, the authors interpreted their data in a biased way, possibly influenced by the yeast study, and drew too strong conclusions. In addition, the authors should have cited important studies about codon optimality in mammalian cells. This lack of information hinders placing their important discoveries in a correct context.

(1) Although the authors concluded that DDX6 acts as a sensor of the slowed ribosome, it is not clear if DDX6 indeed senses the ribosome speed. What the authors showed is a requirement of DDX6 for mRNA decay induced by rare codons, and DDX6 binds to the ribosome to exert this role. For example, DDX6 may bridge the sensor and decay machinery on the ribosome. Without structural or biochemical data on the recognition of the slowed ribosome by DDX6, the role of DDX6 as a sensor remains one of the possible models. It should be described in the discussion section.

We greatly appreciate the reviewer’s comments and suggestions. We agree that our study does not directly establish that DDX6 senses ribosome speed. We also agree that without structural or biochemical data demonstrating recognition of the slowed ribosome by DDX6, the role of DDX6 as a sensor remains one of the possible models. We will incorporate this point into the discussion section and acknowledge it as an important direction for future research.

(2) It is not clear if DDX6 directly binds the ribosome. The authors used ribosomes purified by sucrose cushion, but ribosome-associating and FDF motif-interacting factors might remain on ribosomes, even after RNaseI treatment. Without structural or biochemical data of the direct interaction between the ribosome and DDX6, the authors should avoid description as if DDX6 directly binds to the ribosome.

We agree with the reviewer’s perspective that, even after RNase I treatment, factors associated with the ribosome and interacting with the FDF motif might still remain on the ribosomes that were purified via a sucrose cushion. In the revised manuscript, we will describe the relationship between DDX6 and the ribosome more cautiously, avoiding the depiction of DDX6 directly binding to the ribosome.

(3) Although the authors performed rigorous reporter assays recapitulating the effect of ribosome-retardation sequences on mRNA stability, this is not the first report showing that codon optimality determines mRNA stability in human cells. The authors did not cite important previous studies, such as Wu et al., 2019 (PMID: 31012849), Hia et al., 2019 (PMID: 31482640), Narula et al., 2019 (PMID: 31527111), and Forrest et al., 2020 (PMID: 32053646). These milestone papers should be cited in the Introduction, Results, and Discussion.

Thank you for the reviewer’s correction. We apologize for the oversight in our references. In the revised manuscript, we will ensure these key studies are appropriately cited.

(4) While both DDX6 and deadenylation by the CCR4-NOT were required for mRNA decay by the slowed ribosome, whether DDX6 is required for deadenylation was not investigated. Given that the CCR4-NOT deadenylate complex directly interacts with the empty ribosome E-site in yeast and humans (Buschauer et al., 2020 PMID: 32299921 and Absmeier et al., 2023 PMID: 37653243), whether the loss of DDX6 also affected the action of the CCR4-NOT complex is an important point to investigate, or at least should be discussed in this paper.

We sincerely appreciate the reviewer's valuable suggestions. This point is indeed crucial, and we have addressed it in the revised version of our manuscript. We have included experimental results confirming that the knockout of DDX6 does not impact the CCR4-NOT complex’s deadenylation function. This addition will contribute to a more comprehensive discussion of the relevant issues and refine our manuscript.

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

The authors should explain what they use to determine rare codons in their system and distinguish this feature from codon optimality. Codon optimality is a distinct feature from rare codon usage, and both should be defined better in the context of the paper. The authors interchange between the use of codon optimality, rare codon usage, and translation stalling sequences frequently and should explain and clarify these terms or consider only referring to translation stalling sequences for their discussion.

We appreciate the reviewer's valuable feedback, we have been able to improve the clarity and rigor of the relevant statements in the manuscript. In the revised manuscript, we have provided more explicit and detailed explanations regarding the definition and use of rare codons, and differentiated this from codon optimality, in order to help readers better understand the basis of our analysis and research findings. Furthermore, in the revised manuscript, we are now referring exclusively to 'translation stalling sequences' in our discussion, in order to provide greater clarity.

Reviewer #2 (Recommendations For The Authors):

Interestingly, the translation efficiency of zinc-finger domain mRNAs was increased in DDX6 KO cells. This finding is consistent with the previous study reporting that mRNAs encoding zinc-finger domains are enriched with non-optimal codons and unstable. (Diez et al., 2022 PMID: 35840631). The authors might want to cite this paper and mention the consistency of the two studies.

Thank you for noting the relevance of the increased translation efficiency of zinc-finger domain mRNAs in DDX6 KO cells. We will reference the study by Diez et al. (2022) and emphasize the consistency between their findings and ours, which supports the idea that DDX6 is involved in regulating the translation of mRNAs with these characteristics.

A mutagenesis analysis of the poly-basic residues of BMP2 would further strengthen the authors' claim that this sequence is a primal cause of ribosome slowdown and mRNA decay.

We greatly appreciate the reviewer’s suggestion to conduct a mutagenesis analysis of the poly-basic residues of BMP2. We agree that such an analysis could potentially strengthen our claim. However, considering the constraints we are currently encountering, and our study has already provided substantial evidence to support our findings, we believe that at this stage of our research, conducting this analysis may not be the most immediate priority. We will consider undertaking a mutagenesis analysis in future studies to further validate our conclusions.

In the Introduction, RQC is not commonly referred to as "ribosome-based quality control." Please consider the use of "ribosome-associated quality control."

We appreciate the reviewer providing this suggestion. During the revision process, we corrected the relevant terminology to ensure more precise and appropriate usage.

In the Introduction, the authors should avoid introducing NMD as a part of RQC. NMD was discovered and defined independently of RQC.

Thank you for pointing out this important distinction. We recognize that NMD was discovered and defined independently from RQC, and should not be presented as an integral part of the RQC process. In the revised manuscript, we have made sure to avoid introducing nonsense-mediated decay (NMD) as a component of ribosome-associated quality control (RQC).

Author response:

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

Reviewer #1:

Detection of early-stage colorectal cancer is of great importance. Laboratory scientists and clinicians have reported different exosomal biomarkers to identify colorectal cancer patients. This is a proof-of-principle study of whether exosomal RNAs, and particularly predicted lncRNAs, potential biomarkers of early-stage colorectal cancer and its precancerous lesions.

Strengths:

The study provides a valuable dataset of the whole-transcriptomic profile of circulating sEVs, including miRNA, mRNA, and lncRNA. This approach adds to the understanding of sEV-RNAs' role in CRC carcinogenesis and facilitates the discovery of potential biomarkers.

The developed 60-gene t-SNE model successfully differentiated T1a stage CRC/AA from normal controls with high specificity and sensitivity, indicating the potential of sEV-RNAs as diagnostic markers for early-stage colorectal lesions.

The study combines RNA-seq, RT-qPCR, and modelling algorithms to select and validate candidate sEV-RNAs, maximising the performance of the developed RNA signature. The comparison of different algorithms and consideration of other factors enhance the robustness of the findings.

Weaknesses:

Validation in larger cohorts would be required to establish as biomarkers, and to demonstrate whether the predicted lncRNAs implicated in these biomarkers are indeed present, and whether they are robustly predictive/prognostic.

Thank you for your careful evaluation and valuable suggestions, which have provided valuable guidance for the improvement of our paper. In response to your feedback, we have implemented the following improvements.

(1) More detail about how lncRNA and miRNA candidates were defined, and how this compares to previously published miRNA and lncRNA predictions. The Suppl Methods section for lncRNAs does not describe in detail how the "CPC/CNCI/Pfam" "methods" were combined to define lncRNAs here.

Author response and action taken: Thanks for your comments. In the Supplementary Methods section titled " Selection of Predictive Biomarkers", we have provided a more detailed illustration regarding the screening process for candidate RNA biomarkers. The revised section is as follows: To ensure the predictive performance of the sEV-RNA signature, candidate sEV-RNAs were ultimately selected based on their fold change in colorectal cancer/ precancerous advanced adenoma, absolute abundance, and module attribution. In detail, we initially selected the top 10 RNAs from each category (mRNA, miRNA, and lncRNA) with a fold change greater than 4. In cases where fewer than 10 RNAs were meeting this criterion, all RNAs with a fold change greater than 4 were included. Subsequently, we filtered out RNAs with low abundance, and we selected the top-ranked RNAs from each module based on the fold change ranking for inclusion in the final model.

Compared to most previous studies on EV biomarkers, the overall discriminative performance of the biomarker model we constructed is considerable, holding clinical value for practical application. In contrast, the supplementary merit of this study lies in uncovering the heterogeneity at the whole transcriptome level among samples of different categories, providing a more comprehensive insight into the dynamic changes of biological states. For instance, we inferred the cell subtypes of EV origins through ssGSEA and correlated them with the tumor microenvironment status. The regulatory relationships among different RNA categories were delineated, and their impacts on biological signaling pathways were analyzed, a feat challenging to accomplish solely through sequencing of a single RNA category.

In the Supplementary Methods section titled " Identification of mRNAs and lncRNAs", we have provided a more detailed explanation regarding how the "CPC/CNCI/Pfam" methods were combined to define lncRNAs. The revised section is as follows: Three computational approaches including CPC (Coding Potential Calculator)/CNCI (Coding-Non-Coding Index)/Pfam were combined to sort non-protein coding RNA candidates from putative protein-coding RNAs in the unknown transcripts. CPC is a sequence alignment-based tool used to assess protein-coding capacity. By aligning transcripts with known protein databases, CPC evaluates the biological sequence characteristics of each coding frame of the transcript to determine its coding potential and identify non-coding RNAs.1 CNCI analysis is a method used to distinguish between coding and non-coding transcripts based on adjacent nucleotide triplets. This tool does not rely on known annotation files and can effectively predict incomplete transcripts and antisense transcript pairs.2 Pfam divides protein domains into different protein families and establishes statistical models for the amino acid sequences of each family through protein sequence alignment.3 Transcripts that can be aligned are considered to have a certain protein domain, indicating coding potential, while transcripts without alignment results are potential lncRNAs. Putative protein-coding RNAs were filtered out using a minimum length and exon number threshold. Transcripts above 200 nt with more than two exons were selected as lncRNA candidates and further screened by CPC/CNCI/Pfam. We distinguished lncRNAs from protein-coding genes by intersecting the results of the three determination methods mentioned above.

(2) The role and function of many lncRNAs are unknown, and some lncRNA species may simply be the product of pervasive transcription. Although this is an exploratory and descriptive study of potential biomarkers, it would benefit from some discussion of potential mechanisms because the proposed prediction models include lncRNAs. Do the authors have a hypothesis as to why lncRNAs were informative and predictive in this study? Are these lncRNAs well-studied and/or known to be functional? Or are they markers for pervasive transcription, for example?

Author response and action taken: Thanks for your comments. Whole transcriptome sequencing results facilitate the discussion of regulatory mechanisms between different biomarkers, supplying evidence for future investigations. Among the three lncRNAs involved in this study, lnc-MKRN2-42:1 is involved in the occurrence and development of Parkinson's disease4. The other two lncRNAs, however, lack relevant reports. Therefore, we cannot confirm that these lncRNAs have specific biological functions. In the Supplementary Methods section titled " Identification of mRNAs and lncRNAs", we acknowledge the limited understanding of sEV-lncRNAs in current research. In contrast, many miRNAs in the model have been proven to participate in the occurrence and development of colorectal cancer, such as miR-36155, miR-425-5p6, and miR-106b-3p7. These data provide biological support for the performance of the model, which is particularly valuable for model prediction.

(3) In the Results section "Cell-specific features of the sEV-RNA profile indicated the different proportion of cells of sEV origin among different groups", the sEV-RNA profiles were correlated with existing transcriptome profiles from specific cell types (ssGSEA) and used to estimate "tumour microenvironment-associated scores". This transcriptomic correlation is a valuable observation, but there is no further evidence provided that the sEV-RNAs profiles truly reflect differential cell types of sEV origin between the sample subgroups.

Could the authors clarify the strength of evidence for the cells-of-origin estimates, which are based only on sEV-RNA transcriptome profiles? Would sEV-RNA-derived cells-of-origin be expected to correlate with histopath-derived scores (tumour microenvironment; immune infiltrate) for example? Or is this section intended as an exploratory description of sEV-RNAs, perhaps a check on the plausibility of the sEV-RNA profiles, rather than an accurate estimation of cells-of-origin in each subgroup?

Author response: Thanks for your comments. This section explores the proportional distribution of EVs from different cellular subgroups solely based on transcriptome profiles and algorithms, rather than providing precise estimates of cellular origins within each subgroup.

(4) Software and R package version numbers should be provided.

Author response and action taken: Thanks for your comments. We have added version information for relevant R packages at the first mention in the original text (e.g., WGCNA (version 1.61), Rtsne (version 0.15), GSVA (version 1.42.0), ESTIMATE (version 1.0.13), DOSE (version 3.8.0)).

References

(1) Kong L, et al. CPC: assess the protein-coding potential of transcripts using sequence features and support vector machine. Nucleic Acids Res. 35, W345-349 (2007).

(2) Sun L, et al. Utilizing sequence intrinsic composition to classify protein-coding and long non-coding transcripts. Nucleic Acids Res. 41, e166 (2013).

(3) Finn RD, et al. Pfam: the protein families database. Nucleic Acids Res. 42, D222-230 (2014).

(4) Wang Q, et al. Integrated analysis of exosomal lncRNA and mRNA expression profiles reveals the involvement of lnc-MKRN2-42:1 in the pathogenesis of Parkinson's disease. CNS Neurosci Ther. 26, 527-537 (2020).

(5) Zheng G, et al. Identification and validation of reference genes for qPCR detection of serum microRNAs in colorectal adenocarcinoma patients. PLoS One. 8, e83025 (2013).

(6) Liu D, Zhang H, Cui M, Chen C, Feng Y. Hsa-miR-425-5p promotes tumor growth and metastasis by activating the CTNND1-mediated β-catenin pathway and EMT in colorectal cancer. Cell Cycle. 19, 1917-1927 (2020).

(7) Liu H, et al. Colorectal cancer-derived exosomal miR-106b-3p promotes metastasis by down-regulating DLC-1 expression. Clin Sci (Lond). 134, 419-434 (2020).

Author response:

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

eLife assessment

In this study, Ger and colleagues present a valuable new technique that uses recurrent neural networks to distinguish between model misspecification and behavioral stochasticity when interpreting cognitivebehavioral model fits. Evidence for the usefulness of this technique, which is currently based primarily on a relatively simple toy problem, is considered incomplete but could be improved via comparisons to existing approaches and/or applications to other problems. This technique addresses a long-standing problem that is likely to be of interest to researchers pushing the limits of cognitive computational modeling.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Ger and colleagues address an issue that often impedes computational modeling: the inherent ambiguity between stochasticity in behavior and structural mismatch between the assumed and true model. They propose a solution to use RNNs to estimate the ceiling on explainable variation within a behavioral dataset. With this information in hand, it is possible to determine the extent to which "worse fits" result from behavioral stochasticity versus failures of the cognitive model to capture nuances in behavior (model misspecification). The authors demonstrate the efficacy of the approach in a synthetic toy problem and then use the method to show that poorer model fits to 2-step data in participants with low IQ are actually due to an increase in inherent stochasticity, rather than systemic mismatch between model and behavior.

Strengths:

Overall I found the ideas conveyed in the paper interesting and the paper to be extremely clear and wellwritten. The method itself is clever and intuitive and I believe it could be useful in certain circumstances, particularly ones where the sources of structure in behavioral data are unknown. In general, the support for the method is clear and compelling. The flexibility of the method also means that it can be applied to different types of behavioral data - without any hypotheses about the exact behavioral features that might be present in a given task.

Thank you for taking the time to review our work and for the positive remarks regarding the manuscript. Below is a point-by-point response to the concerns raised.

Weaknesses:

That said, I have some concerns with the manuscript in its current form, largely related to the applicability of the proposed methods for problems of importance in computational cognitive neuroscience. This concern stems from the fact that the toy problem explored in the manuscript is somewhat simple, and the theoretical problem addressed in it could have been identified through other means (for example through the use of posterior predictive checking for model validation), and the actual behavioral data analyzed were interpreted as a null result (failure to reject that the behavioral stochasticity hypothesis), rather than actual identification of model-misspecification. I expand on these primary concerns and raise several smaller points below.

A primary question I have about this work is whether the method described would actually provide any advantage for real cognitive modeling problems beyond what is typically done to minimize the chance of model misspecification (in particular, post-predictive checking). The toy problem examined in the manuscript is pretty extreme (two of the three synthetic agents are very far from what a human would do on the task, and the models deviate from one another to a degree that detecting the difference should not be difficult for any method). The issue posed in the toy data would easily be identified by following good modeling practices, which include using posterior predictive checking over summary measures to identify model insufficiencies, which in turn would call for the need for a broader set of models (See Wilson & Collins 2019). Thus, I am left wondering whether this method could actually identify model misspecification in real world data, particularly in situations where standard posterior predictive checking would fall short. The conclusions from the main empirical data set rest largely on a null result, and the utility of a method for detecting model misspecification seems like it should depend on its ability to detect its presence, not just its absence, in real data.

Beyond the question of its advantage above and beyond data- and hypothesis-informed methods for identifying model misspecification, I am also concerned that if the method does identify a modelinsufficiency, then you still would need to use these other methods in order to understand what aspect of behavior deviated from model predictions in order to design a better model. In general, it seems that the authors should be clear that this is a tool that might be helpful in some situations, but that it will need to be used in combination with other well-described modeling techniques (posterior predictive checking for model validation and guiding cognitive model extensions to capture unexplained features of the data). A general stylistic concern I have with this manuscript is that it presents and characterizes a new tool to help with cognitive computational modeling, but it does not really adhere to best modeling practices (see Collins & Wilson, eLife), which involve looking at data to identify core behavioral features and simulating data from best-fitting models to confirm that these features are reproduced. One could take away from this paper that you would be better off fitting a neural network to your behavioral data rather than carefully comparing the predictions of your cognitive model to your actual data, but I think that would be a highly misleading takeaway since summary measures of behavior would just as easily have diagnosed the model misspecification in the toy problem, and have the added advantage that they provide information about which cognitive processes are missing in such cases.

As a more minor point, it is also worth noting that this method could not distinguish behavioral stochasticity from the deterministic structure that is not repeated across training/test sets (for example, because a specific sequence is present in the training set but not the test set). This should be included in the discussion of method limitations. It was also not entirely clear to me whether the method could be applied to real behavioral data without extensive pretraining (on >500 participants) which would certainly limit its applicability for standard cases.

The authors focus on model misspecification, but in reality, all of our models are misspecified to some degree since the true process-generating behavior almost certainly deviates from our simple models (ie. as George Box is frequently quoted, "all models are wrong, but some of them are useful"). It would be useful to have some more nuanced discussion of situations in which misspecification is and is not problematic.

We thank the reviewer for these comments and have made changes to the manuscript to better describe these limitations. We agree with the reviewer and accept that fitting a neural network is by no means a substitute for careful and dedicated cognitive modeling. Cognitive modeling is aimed at describing the latent processes that are assumed to generate the observed data, and we agree that careful description of the data-generating mechanisms, including posterior predictive checks, is always required. However, even a well-defined cognitive model might still have little predictive accuracy, and it is difficult to know how much resources should be put into trying to test and develop new cognitive models to describe the data. We argue that RNN can lead to some insights regarding this question, and highlight the following limitations that were mentioned by the review: 

First, we accept that it is important to provide positive evidence for the existence of model misspecification. In that sense, a result where the network shows dramatic improvement over the best-fitting theoretical model is easier to interpret compared to when the network shows no (or very little) improvement in predictive accuracy. This is because there is always an option that the network, for some reason, was not flexible enough to learn the data-generating model, or because the data-generating mechanism has changed from training to test. We have now added this more clearly in the limitation section. However, when it comes to our empirical results, we would like to emphasize that the network did in fact improve the predictive accuracy for all participants. The result shows support in favor of a "null" hypothesis in the sense that we seem to find evidence that the change in predictive accuracy between the theoretical model and RNN is not systematic across levels of IQ. This allows us to quantify evidence (use Bayesian statistics) for no systematic model misspecification as a function of IQ. While it is always possible that a different model might systematically improve the predictive accuracy of low vs high IQ individuals' data, this seems less likely given the flexibility of the current results.  

Second, we agree that our current study only applies to the RL models that we tested. In the context of RL, we have used a well-established and frequently applied paradigm and models. We emphasize in the discussion that simulations are required to further validate other uses for this method with other paradigms.  

Third, we also accept that posterior predictive checks should always be capitalized when possible, which is now emphasized in the discussion. However, we note that these are not always easy to interpret in a meaningful way and may not always provide details regarding model insufficiencies as described by the reviewer. It is very hard to determine what should be considered as a good prediction and since the generative model is always unknown, sometimes very low predictive accuracy can still be at the peak of possible model performance. This is because the data might be generated from a very noisy process, capping the possible predictive accuracy at a very low point. However, when strictly using theoretical modeling, it is very hard to determine what predictive accuracy to expect. Also, predictive checks are not always easy to interpret visually or otherwise. For example, in two-armed bandit tasks where there are only two actions, the prediction of choices is easier to understand in our opinion when described using a confusion matrix that summarizes the model's ability to predict the empirical behavior (which becomes similar to the predictive estimation we describe in eq 22).  

Finally, this approach indeed requires a large dataset, with at least three sessions for each participant (training, validation, and test). Further studies might shed more light on the use of optimal epochs as a proxy for noise/complexity that can be used with less data (i.e., training and validation, without a test set).

Please see our changes at the end of this document.  

Reviewer #2 (Public Review):

SUMMARY:

In this manuscript, Ger and colleagues propose two complementary analytical methods aimed at quantifying the model misspecification and irreducible stochasticity in human choice behavior. The first method involves fitting recurrent neural networks (RNNs) and theoretical models to human choices and interpreting the better performance of RNNs as providing evidence of the misspecifications of theoretical models. The second method involves estimating the number of training iterations for which the fitted RNN achieves the best prediction of human choice behavior in a separate, validation data set, following an approach known as "early stopping". This number is then interpreted as a proxy for the amount of explainable variability in behavior, such that fewer iterations (earlier stopping) correspond to a higher amount of irreducible stochasticity in the data. The authors validate the two methods using simulations of choice behavior in a two-stage task, where the simulated behavior is generated by different known models. Finally, the authors use their approach in a real data set of human choices in the two-stage task, concluding that low-IQ subjects exhibit greater levels of stochasticity than high-IQ subjects.

STRENGTHS:

The manuscript explores an extremely important topic to scientists interested in characterizing human decision-making. While it is generally acknowledged that any computational model of behavior will be limited in its ability to describe a particular data set, one should hope to understand whether these limitations arise due to model misspecification or due to irreducible stochasticity in the data. Evidence for the former suggests that better models ought to exist; evidence for the latter suggests they might not.

To address this important topic, the authors elaborate carefully on the rationale of their proposed approach. They describe a variety of simulations - for which the ground truth models and the amount of behavioral stochasticity are known - to validate their approaches. This enables the reader to understand the benefits (and limitations) of these approaches when applied to the two-stage task, a task paradigm commonly used in the field. Through a set of convincing analyses, the authors demonstrate that their approach is capable of identifying situations where an alternative, untested computational model can outperform the set of tested models, before applying these techniques to a realistic data set.

Thank you for reviewing our work and for the positive tone. Please find below a point-by-point response to the concerns you have raised.

WEAKNESSES:

The most significant weakness is that the paper rests on the implicit assumption that the fitted RNNs explain as much variance as possible, an assumption that is likely incorrect and which can result in incorrect conclusions. While in low-dimensional tasks RNNs can predict behavior as well as the data-generating models, this is not *always* the case, and the paper itself illustrates (in Figure 3) several cases where the fitted RNNs fall short of the ground-truth model. In such cases, we cannot conclude that a subject exhibiting a relatively poor RNN fit necessarily has a relatively high degree of behavioral stochasticity. Instead, it is at least conceivable that this subject's behavior is generated precisely (i.e., with low noise) by an alternative model that is poorly fit by an RNN - e.g., a model with long-term sequential dependencies, which RNNs are known to have difficulties in capturing.

These situations could lead to incorrect conclusions for both of the proposed methods. First, the model misspecification analysis might show equal predictive performance for a particular theoretical model and for the RNN. While a scientist might be inclined to conclude that the theoretical model explains the maximum amount of explainable variance and therefore that no better model should exist, the scenario in the previous paragraph suggests that a superior model might nonetheless exist. Second, in the earlystopping analysis, a particular subject may achieve optimal validation performance with fewer epochs than another, leading the scientist to conclude that this subject exhibits higher behavioral noise. However, as before, this could again result from the fact that this subject's behavior is produced with little noise by a different model. Admittedly, the existence of such scenarios *in principle* does not mean that such scenarios are common, and the conclusions drawn in the paper are likely appropriate for the particular examples analyzed. However, it is much less obvious that the RNNs will provide optimal fits in other types of tasks, particularly those with more complex rules and long-term sequential dependencies, and in such scenarios, an ill-advised scientist might end up drawing incorrect conclusions from the application of the proposed approaches.

Yes, we understand and agree. A negative result where RNN is unable to overcome the best fitting theoretical model would always leave room for doubt regarding the fact that a different approach might yield better results. In contrast, a dramatic improvement in predictive accuracy for RNN is easier to interpret since it implies that the theoretical model can be improved. We have made an effort to make this issue clear and more articulated in the discussion. We specifically and directly mention in the discussion that “Equating RNN performance with the generative model should be avoided”.   

However, we would like to note that our empirical results provided a somewhat more nuanced scenario where we found that the RNN generally improved the predictive accuracy of most participants. Importantly, this improvement was found to be equal across participants with no systematic benefits for low vs high IQ participants. We understand that there is always the possibility that another model would show a systematic benefit for low vs. high IQ participants, however, we suggest that this is less likely given the current evidence. We have made an effort to clearly note these issues in the discussion.  

In addition to this general limitation, the paper also makes a few additional claims that are not fully supported by the provided evidence. For example, Figure 4 highlights the relationship between the optimal epochs and agent noise. Yet, it is nonetheless possible that the optimal epoch is influenced by model parameters other than inverse temperature (e.g., learning rate). This could again lead to invalid conclusions, such as concluding that low-IQ is associated with optimal epoch when an alternative account might be that low-IQ is associated with low learning rate, which in turn is associated with optimal epoch. Yet additional factors such as the deep double-descent (Nakkiran et al., ICLR 2020) can also influence the optimal epoch value as computed by the authors.

An additional issue is that Figure 4 reports an association between optimal epoch and noise, but noise is normalized by the true minimal/maximal inverse-temperature of hybrid agents (Eq. 23). It is thus possible that the relationship does not hold for more extreme values of inverse-temperature such as beta=0 (extremely noisy behavior) or beta=inf (deterministic behavior), two important special cases that should be incorporated in the current study. Finally, even taking the association in Figure 4 at face value, there are potential issues with inferring noise from the optimal epoch when their correlation is only r~=0.7. As shown in the figures, upon finding a very low optimal epoch for a particular subject, one might be compelled to infer high amounts of noise, even though several agents may exhibit a low optimal epoch despite having very little noise.

Thank you for these comments. Indeed, there is much we do not yet fully understand about the factors that influence optimal epochs. Currently, it is clear to us that the number of optimal epochs is influenced by a variety of factors, including network size, the data size, and other cognitive parameters, such as the learning rate. We hope that our work serves as a proof-of-concept, suggesting that, in certain scenarios, the number of epochs can be utilized as an empirical estimate. Moreover, we maintain that, at least within the context of the current paradigm, the number of optimal epochs is primarily sensitive to the amount of true underlying noise, assuming the number of trials and network size are constant. We are therefore hopeful that this proofof-concept will encourage research that will further examine the factors that influence the optimal epochs in different behavioral paradigms.  

To address the reviewer's justified concerns, we have made several amendments to the manuscript. First, we added an additional version of Figure 4 in the Supplementary Information material, where the noise parameter values are not scaled. We hope this adjustment clarifies that the parameters were tested across a broad spectrum of values (e.g., 0 to 10 for the hybrid model), spanning the two extremes of complete randomness and high determinism. Second, we included a linear regression analysis showing the association of all model parameters (including noise) with the optimal number of epochs. As anticipated by the reviewer, the learning rate was also found to be associated with the number of optimal epochs. Nonetheless, the noise parameter appears to maintain the most substantial association with the number of optimal epochs. We have also added a specific mentioning of these associations in the discussion, to inform readers that the association between the number of optimal epochs and model parameters should be examined using simulation for other paradigms/models. Lastly, we acknowledge in the discussion that the findings regarding the association between the number of optimal epochs and noise warrant further investigation, considering other factors that might influence the determination of the optimal epoch point and the fact that the correlation with noise is strong, but not perfect (in the range of 0.7).

The discussion now includes the following:

“Several limitations should be considered in our proposed approach. First, fitting a data-driven neural network is evidently not enough to produce a comprehensive theoretical description of the data generation mechanisms. Currently, best practices for cognitive modeling \citep{wilson2019ten} require identifying under what conditions the model struggles to predict the data (e.g., using posterior predictive checks), and describing a different theoretical model that could account for these disadvantages in prediction. However, identifying conditions where the model shortcomings in predictive accuracy are due to model misspecifications rather than noisier behavior is a challenging task. We propose leveraging data-driven RNNs as a supplementary tool, particularly when they significantly outperform existing theoretical models, followed by refined theoretical modeling to provide insights into what processes were mis-specified in the initial modeling effort.

Second, although we observed a robust association between the optimal number of epochs and true noise across varying network sizes and dataset sizes (see Fig.~\ref{figS2}), additional factors such as network architecture and other model parameters (e.g., learning rate, see .~\ref{figS7}) might influence this estimation. Further research is required to allow us to better understand how and why different factors change the number of optimal epochs for a given dataset before it can be applied with confidence to empirical investigations. 

Third, the empirical dataset used in our study consisted of data collected from human participants at a single time point, serving as the training set for our RNN. The test set data, collected with a time interval of approximately $\sim6$ and $\sim18$ months, introduced the possibility of changes in participants' decision-making strategies over time. In our analysis, we neglected any possible changes in participants' decision-making strategies during that time, changes that may lead to poorer generalization performance of our approach. Thus, further studies are needed to eliminate such possible explanations.

Fourth, our simulations, albeit illustrative, were confined to known models, necessitating in-silico validation before extrapolating the efficacy of our approach to other model classes and tasks. Our aim was to showcase the potential benefits of using a data-driven approach, particularly when faced with unknown models. However, whether RNNs will provide optimal fits for tasks with more complex rules and long-term sequential dependencies remains uncertain.

Finally, while positive outcomes where RNNs surpass theoretical models can prompt insightful model refinement, caution is warranted in directly equating RNN performance with that of the generative model, as seen in our simulations (e.g., Figure 3). We highlight that our empirical findings depict a more complex scenario, wherein the RNN enhanced the predictive accuracy for all participants uniformly. Notably, we also provide evidence supporting a null effect among individuals, with no consistent difference in RNN improvement over the theoretical model based on IQ. Although it remains conceivable that a different datadriven model could systematically heighten the predictive accuracy for individuals with lower IQs in this task, such a possibility seems less probable in light of the current findings.”

Reviewer #1 (Recommendations For The Authors):

Minor comments:

Is the t that gets fed as input to RNN just timestep?

t = last transition type (rare/common). not timestep

Line 378: what does "optimal epochs" mean here?

The number of optimal training epochs that minimize both underfitting and overfitting (define in the line ~300)

Line 443: I don't think "identical" is the right word here - surely the authors just mean that there is not an obvious systematic difference in the distributions.

Fixed

I was expecting to see ~500 points in Figure 7a, but there seem to be only 50... why weren't all datasets with at least 2 sessions used for this analysis?

We used the ~500 subjects (only 2 datasets) to pre-train the RNN, and then fine-tuned the pre-trained RNN on the other 54 subjects that have 3 datasets. The correlation of IQ and optimal epoch also hold for the 500 subjects as shown below. 

Author response image 1.

Reviewer #2 (Recommendations For The Authors):

Figure 3b: despite spending a long time trying to understand the meaning of each cell of the confusion matrix, I'm still unsure what they represent. Would be great if you could spell out the meaning of each cell individually, at least for the first matrix in the paper.

We added a clarification to the Figure caption. 

Figure 5: Why didn't the authors show this exact scenario using simulated data? It would be much easier to understand the predictions of this figure if they had been demonstrated in simulated data, such as individuals with different amounts of behavioral noise or different levels of model misspecifications.

In Figure 5 the x-axis represents IQ. Replacing the x-axis with true noise would make what we present now as Figure 4. We have made an effort to emphasize the meaning of the axes in the caption. 

Line 195 ("...in the action selection. Where"). Typo? No period is needed before "where".

Fixed

Line 213 ("K dominated-hand model"). I was intrigued by this model, but wasn't sure whether it has been used previously in the literature, or whether this is the first time it has been proposed.

This is the first time that we know of that this model is used.  

Line 345 ("This suggests that RNN is flexible enough to approximate a wide range of different behavioral models"): Worth explaining why (i.e., because the GRUs are able to capture dependencies across longer delays than a k-order Logistic Regression model).

Line 356 ("We were interested to test"): Suggestion: "We were interested in testing".

Fixed

Line 389 ("However, as long as the number of observations and the size of the network is the same between two datasets, the number of optimal epochs can be used to estimate whether the dataset of one participant is noisier compared with a second dataset."): This is an important claim that should ideally be demonstrated directly. The paper only illustrates this effect through a correlation and a scatter plot, where higher noise tends to predict a lower optimal epoch. However, is the claim here that, in some circumstances, optimal epoch can be used to *deterministically* estimate noise? If so, this would be a strong result and should ideally be included in the paper.

We have now omitted this sentenced and toned down our claims, suggesting that while we did find a strong association between noise and optimal epochs, future research is required to established to what extent this could be differentiated from other factors (i.e., network size, amount of observations).

Author response:

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

eLife assessment

This study provides important new insights into how multisensory information is processed in the lateral cortex of the inferior colliculus, a poorly understood part of the auditory midbrain. By developing new imaging techniques that provide the first optical access to the lateral cortex in a living animal, the authors provide convincing in vivo evidence that this region contains separate subregions that can be distinguished by their sensory inputs and neurochemical profiles, as suggested by previous anatomical and in vitro studies. Additional information and analyses are needed, however, to allow readers to fully appreciate what was done, and the comparison of multisensory interactions between awake and anesthetized mice would benefit from being explored in more detail.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In this paper, the authors provide a characterisation of auditory responses (tones, noise, and amplitude-modulated sounds) and bimodal (somatosensory-auditory) responses and interactions in the higher-order lateral cortex (LC) of the inferior colliculus (IC) and compare these characteristics with the higher order dorsal cortex (DC) of the IC - in awake and anaesthetised mice. Dan Llano's group has previously identified gaba'ergic patches (modules) in the LC distinctly receiving inputs from somatosensory structures, surrounded by matrix regions receiving inputs from the auditory cortex. They here use 2P calcium imaging combined with an implanted prism to - for the first time - get functional optical access to these subregions (modules and matrix) in the lateral cortex of IC in vivo, in order to also characterise the functional difference in these subparts of LC. They find that both DC and LC of both awake and anaesthetised mice appear to be more responsive to more complex sounds (amplitude-modulated noise) compared to pure tones and that under anesthesia the matrix of LC is more modulated by specific frequency and temporal content compared to the gabaergic modules in LC. However, while both LC and DC appear to have low-frequency preferences, this preference for low frequencies is more pronounced in DC. Furthermore, in both awake and anesthetized mice, somatosensory inputs are capable of driving responses on their own in the modules of LC, but very little (possibly not at all) in the matrix. However, bimodal interactions may be different under awake and anesthesia in LC, which warrants deeper investigation by the authors: They find, under anesthesia, more bimodal enhancement in modules of LC compared to the matrix of LC and bimodal suppression dominating the matrix of LC. In contrast, under awake conditions bimodal enhancement is almost exclusively found in the matrix of LC, and bimodal suppression dominates both matrix and modules of LC.

The paper provides new information about how subregions with different inputs and neurochemical profiles in the higher-order auditory midbrain process auditory and multisensory information, and is useful for the auditory and multisensory circuits neuroscience community.

Strengths:

The major strength of this study is undoubtedly the fact that the authors for the first time provide optical access to a subcortical region (the lateral cortex of the inferior colliculus (i.e. higher order auditory midbrain)) which we know (from previous work by the same group) have optically identifiable subdivisions with unique inputs and neurotransmitter release, and plays a central role in auditory and multisensory processing. A description of basic auditory and multisensory properties of this structure is therefore very useful for understanding auditory processing and multisensory interactions in subcortical circuits.

Weaknesses:

I have divided my comments about weaknesses and improvements into major and minor comments. All of which I believe are addressable by the reviewers to provide a more clear picture of their characterisation of the higher-order auditory midbrain.

Major comment:

(1) The differences between multisensory interactions in LC in anaesthetised and awake preparations appear to be qualitatively different, though the authors claim they are similar (see also minor comment related to figure 10H for further explanation of what I mean). However, the findings in awake and anaesthetised conditions are summarised differently, and plotting of similar findings in the awake figures and anaesthetised figures are different - and different statistics are used for the same comparisons. This makes it very difficult to assess how multisensory integration in LC is different under awake and anaesthetised conditions. I suggest that the authors plot (and test with similar statistics) the summary plots in Figure 8 (i.e. Figure 8H-K) for awake data in Figure 10, and also make similar plots to Figures 10G-H for anaesthetised data. This will help the readers understand the differences between bimodal stimulation effects on awake and anaesthetised preparations - which in its current form, looks very distinct. In general, it is unclear to me why the awake data related to Figures 9 and 10 is presented in a different way for similar comparisons. Please streamline the presentation of results for anaesthetised and awake results to aid the comparison of results in different states, and explicitly state and discuss differences under awake and anaesthetised conditions.

We thank the reviewer for the valuable suggestion. We only highlighted the similarities between the data obtained from anesthetized and awake preparations to indicate the ability to reproduce the technique in awake animals for future assessment. Identifying those similarities between the two experimental setups was based on the comparison between modules vs matrix or LC vs DC within each experimental setup (awake vs anesthetized). Therefore, the statistics were chosen differently for each setup based on the size of the subjects (n) within each experimental preparation. However, we agree with the reviewer’s comment that there are differences between the anesthetized and awake data. To examine these differences, we ran the same statistics for Figure 5 (tonotopy of LC vs. DC-anesthetic animals) and Figure 9 (tonotopy of LC vs DC-awake animals). In addition, we added a new figure after Figure 9 to separate the statistical analysis from the maps. Accordingly, Figures 4 and 5 (maps and analysis, respectively -anesthetized animals) now match Figures 9 and 10 (maps and analysis, respectively – awake animals). We also did the same thing for Figures 7 (microprism imaging of the LC - anesthetized animals), 8 (imaging of the LC from the dorsal surface - anesthetized animals) as well as Figure 11 or old Figure 10 (microprism imaging of the LC - awake animals) to address the similarities and differences of the multisensory data between awake and anesthetized animals. We edited the text accordingly in the result and discussion sections.

(2) The claim about the degree of tonotopy in LC and DC should be aided by summary statistics to understand the degree to which tonotopy is actually present. For example, the authors could demonstrate that it is not possible/or is possible to predict above chance a cell's BF based on the group of other cells in the area. This will help understand to what degree the tonotopy is topographic vs salt and pepper. Also, it would be good to know if the gaba'ergic modules have a higher propensity of particular BFs or tonotopic structure compared to matrix regions in LC, and also if general tuning properties (e.g. tuning width) are different from the matrix cells and the ones in DC.

Thank you for the reviewer’s suggestion. We have examined the tonotopy of LC and DC using two regression models (linear and quadratic polynomial) between the BFs of the cells and their location on the anatomical axis. Therefore, the tonotopy is indicated by a significant regression fit with a high R2 between the BFs the cells, and their location within each structure. For the DC, there was a significant regression fit between the BFs of the cells and their locations over the rostromedial to the caudolateral axis. Additionally, the R2 of the quadratic polynomial fit was higher than that of the linear fit, which indicates a nonlinear distribution of cells based on their BFs, which is consistent with the presence of high-low-high tuning over the DC surface. Given that the microprism cannot image the whole area of the LC, and it images a slightly different area in each animal, it was very difficult to get a consistent map for the LC as well as a solid conclusion about the LC tonotopy. However, we have examined the regression fit between the BFs of cells and their location along the main four anatomical axes of the field of view obtained from each animal (dorsal to ventral), (rostral to caudal), (dorsocaudal to ventrorostral) (dorsorostral to ventrocoudal). Unlike the DC, the LC imaged via microprism showed a lower R2 for both linear and quadratic regression mostly in the dorsoventral axis. We show the fitting curves of these regressions in Figure 4-figure supplement 1 (anesthetized data) and Figure 9-figure supplement 1 (awake data). Despite the inconsistent tonotopy of the LC imaged via microprism, the modules were found to have a higher BFs median at 10 kHz compared to matrix that had a lower BFs median at 7.1 kHz, which was consistent across the anesthetized and awake animals. We have added these results in the corresponding spot in the results section (lines 193-197 and 361-364). We have examined the tuning width using the binarized receptive field sum (RFS) method in which each neuron was given a value of 1 if it responds to a single frequency (Narrow RF), but this value increases if the neuron responds to more neighbor frequencies (wide RF). We did this calculation across all the sound levels. Both DC and LC of the anesthetized animals had higher RFS mean and median than those of awake animals given that ketamine was known to broaden the RF. However, in both preparations (anesthetized and awake), the DC had a higher RFS mean than that of the LC, which could be consistent with the finding that the DC had a relatively lower SMI than the LC. To show these new data, we made a new Figure 10-figure supplement 1, and we edited the text accordingly [lines 372-379 & 527-531].

(3) Throughout the paper more information needs to be given about the number of cells, sessions, and animals used in each panel, and what level was used as n in the statistical tests. For example, in Figure 4 I can not tell if the 4 mice shown for LC imaging are the only 4 mice imaged, and used in the Figure 4E summary or if these are just examples. In general, throughout the paper, it is currently not possible to assess how many cells, sessions, and animals the data shown comes from.

Thank you for the reviewer’s comment. We do apologize for not adding this information. We added all the information regarding the size of the statistical subjects (number of cells or number of animals used) for every test outcome. To keep the flow of the text, we added the details of the statistical tests in the legends of the figures.

(4) Throughout the paper, to better understand the summary maps and plots, it would be helpful to see example responses of the different components investigated. For example, given that module cells appear to have more auditory offset responses, it would be helpful to see what the bimodal, sound-only, and somatosensory responses look like in example cells in LC modules. This also goes for just general examples of what the responses to auditory and somatosensory inputs look like in DC vs LC. In general example plots of what the responses actually look like are needed to better understand what is being summarised.

Thank you for the reviewer’s comment and suggestion. We modified Figure 6 and the text accordingly to include all the significant examples of cells discussed throughout the work.

Reviewer #2 (Public Review):

Summary:

The study describes differences in responses to sounds and whisker deflections as well as combinations of these stimuli in different neurochemically defined subsections of the lateral and dorsal cortex of the inferior colliculus in anesthetised and awake mice.

Strengths:

The main achievement of the work lies in obtaining the data in the first place as this required establishing and refining a challenging surgical procedure to insert a prism that enabled the authors to visualise the lateral surface of the inferior colliculus. Using this approach, the authors were then able to provide the first functional comparison of neural responses inside and outside of the GABA-rich modules of the lateral cortex. The strongest and most interesting aspects of the results, in my opinion, concern the interactions of auditory and somatosensory stimulation. For instance, the authors find that a) somatosensory-responses are strongest inside the modules and b) somatosensory-auditory suppression is stronger in the matrix than in the modules. This suggests that, while somatosensory inputs preferentially target the GABA-rich modules, they do not exclusively target GABAergic neurons within the modules (given that the authors record exclusively from excitatory neurons we wouldn't expect to see somatosensory responses if they targeted exclusively GABAergic neurons), and that the GABAergic neurons of the modules (consistent with previous work) preferentially impact neurons outside the modules, i.e. via long-range connections.

Weaknesses:

While the findings are of interest to the subfield they have only rather limited implications beyond it. The writing is not as precise as it could be. Consequently, the manuscript is unclear in some places. For instance, the text is somewhat confusing as to whether there is a difference in the pattern (modules vs matrix) of somatosensory-auditory suppression between anesthetized and awake animals. Furthermore, there are aspects of the results which are potentially very interesting but have not been explored. For example, there is a remarkable degree of clustering of response properties evident in many of the maps included in the paper. Taking Figure 7 for instance, rather than a salt and pepper organization we can see auditory responsive neurons clumped together and non-responsive neurons clumped together and in the panels below we can see off-responsive neurons forming clusters (although it is not easy to make out the magenta dots against the black background). This degree of clustering seems much stronger than expected and deserves further attention.

Thank you for the reviewer’s comment. We do apologize if some areas in the manuscript were imprecisely written. For anesthetized and awake data, we have only emphasized the similarities between the two setups to show the ability to use microprism in awake animals for future assessment. To highlight the differences between anesthetized and awake animals, we have now run uniform statistics for all the data collected from both setups. Accordingly, we have edited Figures 4 and 5 (tonotopy-anesthetized) to match Figures 9 and new Figure 10 (tonotopy-awake). Also, we edited Figures 7 and 8 (multisensory- anesthetized) to match Figure 11 or old Figure 10 (multisensory- awake). We edited the text accordingly in the results section and discussed the possible differences between anesthetized and awake data in the discussion section [lines 521-553].

We agree with the reviewer’s comment that the cells were topographically clustered based on their responses. Some of these clusters include the somatosensory responsive cells, which were located mostly in the modules (Figures 7D and 8E). Also, the auditory responsive cells with offset responses were clustered mostly in the modules (Figures 7C and 8F). Accordingly, we have edited the text to emphasize this finding.

We noticed also that some responsive cells to the tested stimulations were surrounded by nonresponsive cells. By comparing the response of the cells to different stimuli we found that while Figures 7 and 11 (old Figure 10) showed only the response of the cells to auditory stimulation (unmodulated broadband noise at 80 dB) and somatosensory stimulation (whisker deflection), some nonresponsive cells to these specific stimulations were found to be responsive to pure tones of different frequencies and amplitudes. As an indicator of the cells' viability, we additionally examined the spontaneous activity of the nonresponsive cells across different data sets. We note that spontaneous activity was rare for all cells even among the responsive cells to sound or somatosensory stimulations. This finding could be related to the possibility that the 2P imaging of calcium signals may not be sensitive enough to track spontaneous activity that may originate from single spikes. However, in some data sets, we have found that the cells that did not respond to any tested stimuli showed spontaneous activity when no stimulation was given indicating the viability of those cells. We have addressed the activity of the non-responsive cells in the text along with a new Figure 11-figure supplement 1.

We changed the magenta into a green color to be suitable for the dark background. Also, we have completely changed the color palette of all of our images to be suitable for color-blind readers as suggested by reviewer 1.

Reviewer #3 (Public Review):

The lateral cortex of the inferior colliculus (LC) is a region of the auditory midbrain noted for receiving both auditory and somatosensory input. Anatomical studies have established that somatosensory input primarily impinges on "modular" regions of the LC, which are characterized by high densities of GABAergic neurons, while auditory input is more prominent in the "matrix" regions that surround the modules. However, how auditory and somatosensory stimuli shape activity, both individually and when combined, in the modular and matrix regions of the LC has remained unknown.

The major obstacle to progress has been the location of the LC on the lateral edge of the inferior colliculus where it cannot be accessed in vivo using conventional imaging approaches. The authors overcame this obstacle by developing methods to implant a microprism adjacent to the LC. By redirecting light from the lateral surface of the LC to the dorsal surface of the microprism, the microprism enabled two-photon imaging of the LC via a dorsal approach in anesthetized and awake mice. Then, by crossing GAD-67-GFP mice with Thy1-jRGECO1a mice, the authors showed that they could identify LC modules in vivo using GFP fluorescence while assessing neural responses to auditory, somatosensory, and multimodal stimuli using Ca2+ imaging. Critically, the authors also validated the accuracy of the microprism technique by directly comparing results obtained with a microprism to data collected using conventional imaging of the dorsal-most LC modules, which are directly visible on the dorsal IC surface, finding good correlations between the approaches.

Through this innovative combination of techniques, the authors found that matrix neurons were more sensitive to auditory stimuli than modular neurons, modular neurons were more sensitive to somatosensory stimuli than matrix neurons, and bimodal, auditory-somatosensory stimuli were more likely to suppress activity in matrix neurons and enhance activity in modular neurons. Interestingly, despite their higher sensitivity to somatosensory stimuli than matrix neurons, modular neurons in the anesthetized prep were far more responsive to auditory stimuli than somatosensory stimuli (albeit with a tendency to have offset responses to sounds). This suggests that modular neurons should not be thought of as primarily representing somatosensory input, but rather as being more prone to having their auditory responses modified by somatosensory input. However, this trend was reversed in the awake prep, where modular neurons became more responsive to somatosensory stimuli than auditory stimuli. Thus, to this reviewer, the most intriguing result of the present study is the dramatic extent to which neural responses in the LC changed in the awake preparation. While this is not entirely unexpected, the magnitude and stimulus specificity of the changes caused by anesthesia highlight the extent to which higher-level sensory processing is affected by anesthesia and strongly suggest that future studies of LC function should be conducted in awake animals.

Together, the results of this study expand our understanding of the functional roles of matrix and module neurons by showing that responses in LC subregions are more complicated than might have been expected based on anatomy alone. The development of the microprism technique for imaging the LC will be a boon to the field, finally enabling much-needed studies of LC function in vivo. The experiments were well-designed and well-controlled, and the limitations of two-photon imaging for tracking neural activity are acknowledged. Appropriate statistical tests were used. There are three main issues the authors should address, but otherwise, this study represents an important advance in the field.

(1) Please address whether the Thy1 mouse evenly expresses jRGECO1a in all LC neurons. It is known that these mice express jRGECO1a in subsets of neurons in the cerebral cortex, and similar biases in the LC could have biased the results here.

Thank you for the reviewer’s comment. In the work published by Dana, et al, the expression of jRGECO1a in all Thy1 mouse lines was determined by the brightness of the jRGECO1a in the soma. Given that some cells do not show a detected level of jRGECO1a fluorescence until activated, the difference in expression shown in different brain regions could be related to the level of neuronal activity at the time of sample processing and not the expression levels of the indicator itself. To the best of our knowledge, there is no antibody for jRGECO1a, which can be used for detecting the expression levels of the indicator regardless of the neuronal activity. To test the hypothesis that DC and LC have different levels of jRGECO1a, we examined the expression levels of jRGECO1a after we perfused the mice with high potassium saline to elicit a general neuronal depolarization in the whole brain. Then we immunostained against NeuN (the neuronal marker) to quantify the percentage of the neurons expressing jRGECO1a to the total number of neurons (indicated by NeuN). To have a fair comparison, we restricted our analysis to include the areas imaged only by 2P as some regions were not accessible by microprism such as the deep ventral regions of the LC. There is a similar % of cells expressing jRGECO1a in DC and LC. As expected, the neurons expressing jRGECO1a were only nonGABAergic cells. We addressed these findings in the new Figure 3-figure Supplement 1 as well as the corresponding text in the results [lines 178-184] and methods sections [lines 878-892].

(2) I suggest adding a paragraph or two to the discussion to address the large differences observed between the anesthetized and awake preparations. For example, somatosensory responses in the modules increased dramatically from 14.4% in the anesthetized prep to 63.6% in the awake prep. At the same time, auditory responses decreased from 52.1% to 22%. (Numbers for anesthetized prep include auditory responses and somatosensory + auditory responses.). In addition, the tonotopy of the DC shifted in the awake condition. These are intriguing changes that are not entirely expected from the switch to an awake prep and therefore warrant discussion.

Thank you for the reviewer’s comment. To determine if differences exist between anesthetized and awake data, we have now used the same statistics and edited Figures 4,5,7,8,9, and 10 as well as added a new Figure 11. Accordingly, we have edited the result section and added a paragraph addressing the possible differences between the two preparations in the Discussion section [lines 521-553]..

(3) For somatosensory stimuli, the authors used whisker deflection, but based on the anatomy, this is presumably not the only somatosensory stimulus that affects LC. The authors could help readers place the present results in a broader context by discussing how other somatosensory stimuli might come into play. For example, might a larger percentage of modular neurons be activated by somatosensory stimuli if more diverse stimuli were used?

We agree with the reviewer’s point. Indeed, the modules are receiving different inputs from different somatosensory sources such as somatosensory cortex and dorsal column nuclei, which could indicate that the activity of the cells in the modular areas could be evoked by different types of somatosensory stimulations, which is an open area for future studies. We have discussed this point in the revised Discussion section [lines 516-520].

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors):

Minor comments:

(1) Figure 3H: The lateral surface seems quite damaged by the prism. An example slice of the imaging area of each mouse would help the reader better understand the extent of damage the prism leaves in the area of interest.

Thank you for the reviewer’s comment. We already have included such images in Figures 4A, 7A, and 9A to present the field of view of all prism experiments. However, we need to clarify the point of tissue damage. The insertion of microprism may be associated with some tissue damage as a result of making the pocket for the microprism to be inserted, but it is not possible to get neuronal signals from a damaged field of view. Therefore, we do not believe that there is tissue damage to the parts of the LC imaged by microprism. However, there may be some areas where the microprism is not in direct contact with the LC surface. These areas are located mostly in the periphery of the field of view, and they are completely black as they are out of focus (i.e., the left side of Figure 3B). The right side of Figure 3b as well as Figure 3A have some black areas, which present the vasculatures, where there are no red signals because of the lack of jRGECO1a expression in those areas.

(2) In relation to the data shown in Figure 4E it is claimed that LC is tuned to higher frequencies (lines 195-196). However, the majority of cells appear to be tuned to frequencies below 14kHz (with a median of 7.5 kHz), which is quite low for the mouse. I assume that the authors mean frequencies that are relatively higher than the DC, but it is worth mentioning in the text that the BFs found in the LC are quite low-frequency responses for the mouse.

Thank you for the reviewer’s comment, which we agree with. We edited this part by acknowledging that around 50% of the LC cells had a low-frequency bias to 5 and 7.1 kHz. Then we mentioned that most of the LC cells are tuned to relatively higher frequencies than those of the DC [lines 215-218].

(3) Figure 5A-C: Is it the tone-responsive cells plus an additional ~22% of cells that respond to AM, or are there also cells that respond to tones that do not respond to AM. Please break down to which degree the tone and AM responsive cells are overlapping.

Thank you for the reviewer’s comment and suggestion. We broke down the responsive cells into cells responsive only to pure tone (tone selective cells or Tone-sel) or to only AM-noise (noise selective cells or Noise-sel) as well as cells responding to both sounds (nonselective cells or Non-sel). We examined the fractions of these categories of cells in both LC and DC within all responsive neurons. Accordingly, we have edited Figure 5A-C as well as the text [lines 229-243].

(4) Figure 5D. It is unclear to me how a cell is classified as SMI or TMI responsive after computing the SMI or TMI for each cell. What statistic was used to determine if the cell was responsive or not?

Thank you for the reviewer’s comment. We do apologize for the confusion caused by Figures 5D and E. These figures do not show the values of SMI or TMI, respectively. Rather, the figures show the percentage of the spectrally or temporally modulated cells, respectively. At each sound level, the cells were categorized into two main types. The spectrally modulated cells are those responsive to pure tones or unmodulated noise, so they can detect the spectral features of the sound (old Figure 5D or new Figure 5E). The temporally modulated cells are those responsive to AM-noise, so they can detect the temporal features of the sound of complex spectra like the broadband noise (old Figure 5E or new Figure 5F). To clear this confusion, we removed the words SMI and TMI from the figures, and then we renamed the x-axis label into “% of spectrally modulated cells” and “% of temporally modulated cells” for Figures 5D (new 5E) and E (new 5F), respectively.

(5) Figure 5 D, E: Is the decrease in SMI and TMI modulated cells in the modules a result of simply lower sensitivity to sounds (i.e. higher response thresholds)? If a cell responds to neither tone, AM, or noise it will have a low SMI and TMI index. If this is the case that affects the interpretation, as it is then not a decrease in sensitivity to spectral or temporal modulation, but instead a difference in overall sound sensitivity.

Thank you for the reviewer’s comment. We apologize for the confusion about Figures 5E and D, which did not show the SMI and TMI values. Rather, they show the percentage of spectrally or temporally modulated cells, respectively, as explained in our previous response. Therefore, Figure 5D shows the percentage of cells that can detect the spectral features of sound, while Figure 5E shows the percentage of cells that can detect the temporal features of sounds of complex spectra like broadband noise. Accordingly, Figures 5D and E show the sensitivity to different features of sound and not the overall sound sensitivity.

(6) Figure 7 and 8: What is the false positive rate expected of the responsive cells using the correlation cell flagging criteria? Especially given that the fraction of cells responsive to somatosensory stimulation in LC (matrix) is 0.88% and 1.3% in DC, it is important to know what the expected false positive rate is in order to be able to state that there are actually somatosensory responses there or if this is what you would expect from false positives given the inclusion test used. Please provide an estimate of the false positive rate given your inclusion test and show that the rate found is statistically significantly above that level - and show this rate with a line in Figure 7 H, I.

Thank you for the reviewer’s comment. To test the efficiency of the correlation method to determine the responsive cells, we initially ran an ROC curve comparing the automated method to a blinded human interpretation. The AUC of the ROC curve was 0.88. This high AUC value indicates that the correlation method can rank the random responsive cells than the random nonresponsive cells. At the correlation coefficient (0.4), which was the cutoff value to determine the responsive cells for somatosensory stimulation, the specificity was 87% and the sensitivity 72%, the positive predictive value was 73%, and the negative predictive value was 86%. Although the above percentages indicate the efficiency of the correlation method, we excluded all the false responsive cells from the analysis. Therefore, the fractions of cells in the graphs are the true responsive cells with no contamination of the non-responsive cells. We also modified Figures 7H and I to match the other data sets obtained from awake animals. Therefore, Figures 7H and I no longer show the average of the responsive cells. Instead, they show the % of different fractions of responsive cells within each cellular motif (modules and matrix). Accordingly, we believe that there is no need to include a rate line on the graph. We added the section describing the validation part to the methods section [lines 808-815].

(7) Figure 7: Please clarify what is meant by a cell responding to 'both responding to somatosensory and auditory stimulation'. Does it mean that the cell has responses to both auditory and somatosensory stimulation when presented individually or if it responds to both presented together? If it is the former, I don't understand how the number to both can be higher than the number of somatosensory alone (as both requires it also to respond to somatosensory alone). If it is the latter (combined auditory and somatosensory) then it seems that somatosensory inputs remove the responsiveness of most cells that were otherwise responsive to auditory alone (e.g. in the module while 42% respond to sound alone, combined stimulation would leave only 10% of cells responsive). Please clarify what exactly the authors are plotting and stating here.

Thank you for the reviewer’s comment. The responsive cells in Figure 7 are divided into three categories. Each category has a completely different group of cells. The first category is for the cells responding only to auditory stimulation (auditory-selective cells or Aud-sel). The second category is for the cells that respond only to somatosensory stimulation (somatosensory selective cells or Som-sel). The third category is for the cells that respond to both auditory and somatosensory stimulations when both stimulations are presented individually (auditory/somatosensory nonselective cells or Aud/Som-nonsel). Accordingly, the number of cells may be different across all these categories. We have clarified this part in the text [lines 299-303]. We have modified Figures 7, 8, and 11 (old Figure 10) to match the data from anesthetized and awake animals, so Figures 7H and I now show the collective % of the cells from all animals within modules vs matrix.

(8) Why are the inferential statistics used in Figure 9F (chi-square test) and Figure 5A-C (t-test) when it tests the same thing (the only difference is one is anaesthetised data and the other awake)? Indeed, all Figure 9 and 10 (awake data figures) plots use chi-square tests to test differences in percentages instead of t-tests used in earlier (anaesthetised data figures) plots to test differences in percentages between groups. Please clarify the reason for this change in statistics used for similar comparisons.

Thank you for the reviewer’s comment. Imaging the LC via microprism from awake animals confirmed the ability to run this technique with no interference to the ambulatory functions of the animals. Therefore, the main goal was to highlight the similarities between the data obtained from awake and anesthetized setups by highlighting the comparison between the LC and DC or between modules and matrix within each preparation (anesthetized vs awake). Accordingly, the statistics used to run these comparisons were chosen based on the number of the tested animals at each setup (7 anesthetized animals and 3 awake animals for prism insertion). The low number of animals used for awake data made us use the number of cells collectively from all animals instead of the number of animals, so we used the Chi-square test to examine the differences in percentages.

(9) Figure 10H: The main text describes the results shown here as similar to what was seen in anaesthetised animals. But it looks to me like the results in awake animals are qualitatively different from the multisensory interaction seen in anaesthetised animals. In anaesthetised animals the authors find that there is a higher chance of auditory responses being enhanced by somatosensory inputs when cells are in the modules compared to in the matrix. However, in awake data, this relationship is flipped, with more bimodal enhancement found in the matrix compared to the modules. Furthermore, almost all cells in the modules are suppressed by combined somatosensory input which looks like it is different from what is found in anaesthestised mice and what is described in the discussion: 'we observed that combined auditory-somatosensory stimulation generally suppressed neural responses to auditory stimuli and that this suppression was most prominent in the LC matrix'.

Thank you for the reviewer’s comment. Our statement was meant to show how the data obtained from awake and anesthetized animals were generally similar. However, we agree that the statement may not be suitable due to the possible differences between awake and anesthetized animals. To address a fair comparison between the anesthetized and awake preparations, we ran similar statistics and graphs for Figures 7, 8, and 11 (old Figure 10). Given that the areas occupied by modules and matrix are different across animals due to the irregular shape of the modules, we chose to run a chi-square test for all the data to quantify the collective % of responding cells within modules vs matrix from all tested animals for each experimental setup (anesthetized vs awake). The anesthetized and awake animals similarly showed that modules and matrix had higher fractions of auditory responsive cells. However, matrix had more cells responding to auditory stimulations than modules, while modules had more cells responding to somatosensory stimulation than matrix. In contrast, while the anesthetized animals showed higher fractions of offset auditory-responsive cells, which were mostly clustered in the modules, the offset auditory-responsive cells were very rare in awake animals (6 cells/one animal).

Based on the fractions of cells with suppressed or enhanced auditory response induced by bimodal stimulation, the data obtained from anesthetized and awake animals showed that the auditory response in the matrix was suppressed more than enhanced by bimodal stimulation. In contrast, modules had different profiles across the experimental setups and locations. For instance, the modules imaged via microprism in the anesthetized and awake animals showed suppressed more than enhanced auditory responses, but modules imaged from the dorsal surface in anesthetized animals showed enhanced more than suppressed auditory responses. Additionally, modules had less suppressed and more enhanced auditory responses compared to matrix in the anesthetized animals regardless of the location of the modules (microprism or dorsal surface). Yet, modules from awake animals had more suppressed and less enhanced auditory responses compared to matrix. We have addressed these differences in the results and discussion section.

Additional minor comments that I think the authors could use to aid their manuscript clarity:

(1) The figure colour selection - especially in Figures 7 and 8 - is really hard to tell apart. Please choose more distinct colours, and a colour scheme that is appropriate for colour blind readers.

Thank you for the reviewer’s suggestion. We have noticed that the magenta color assigned for the cells with offset responses was very difficult to distinguish from the black background. We have changed the magenta color to green to be different from the color of other cells. Using Photoshop, we chose a color scheme that is suitable for color-blind readers in all our maps.

(2) The sentence in lines 331-334 should be rephrased for clarity.

Thank you for the reviewer’s suggestion. We have rephrased the statement for clarity [lines 364-371].

Reviewer #2 (Recommendations For The Authors):

As mentioned in the public review the strong clustering evident in some of the maps (some of which may be related to module/matrix differences but certainly not all of it) seems worth scrutinizing further. Would we expect such a strong spatial segregation of auditory responsive and non-responsive neurons? Would we expect response properties (e.g. off-responsiveness) other than frequency tuning to show evidence of a topographic arrangement in the IC? In addressing this it would, of course, be important to rule out that this clustering is not down to some trivial experimental variables and truly reflects functional organization. For instance, are the patches of non-responsive neurons found in parts of the field of view with poor visibility, poor labelling, etc which may explain why it is difficult to pick up responses there? Are the neurons in non-responsive areas otherwise active (i.e. do they show spontaneous activity) or could they be 'dead'? Could the way neuropil signals are dealt with play a role here (it is weighted by 0.4 which strikes me as quite low)? In relation to this, I am also wondering to what extent the extreme overrepresentation (Figure 4) of neurons with a BF of 5kHz (some of this is, of course, down to the fact that the lower end of the frequency range was 5kHz and that the step size was 0.5 octaves), especially in the DC, is to be interpreted.

Thank you for the reviewer’s comment. Before analysis, the ROIs of all cells were set around the cell bodies using the jRGECO1a signals as a reference, so all cells (responsive and nonresponsive) were collected from areas of good visibility of jRGECO1a signals. In other words, no cells were collected from regions having poor jRGECO1a signals. In Figures 7, 8, and 11 (old Figure 10), the cells showed response either only to unmodulated broadband noise at 80 dB as an auditory stimulus or to whisker deflection with specific speed and power as a somatosensory stimulus. Given that the two stimuli above had specific parameters, the remaining non-responsive cells may respond to auditory or somatosensory stimulations with other features. For instance, some nonresponsive cells to the unmodulated broadband noise were responding to pure tones with different amplitudes and frequencies or to different AM-noise with different amplitudes and modulation frequencies.  Also, these nonresponsive cells may not respond to any of our tested stimuli and may respond to other sensory stimulations. Some of the non-responsive cells showed spontaneous activity when no stimulations were presented. However, we can not rule out the possibility that some of these nonresponsive cells may not be viable. We have addressed the clustering properties in the revised version of the manuscript in the corresponding spots of the results and discussion sections. We have added a new supplementary figure (Figure 11- Figure Supplement 1) to show how the nonresponsive cells to the unmodulated noise may respond to other types of sound and to show the spontaneous activity of some non-responsive cells.

For the neuropil, previous reports used the contamination factor (r) in a range of 0.3-0.7 (we referenced these studies in the method section [line 776) based on the tissue or cells imaged, vasculatures, and the objective used for imaging. Therefore, we optimized the contamination factor (r) to be 0.4 through a preliminary analysis based on the tissue we image (LC), and the objective used (16x with NA = 0.8 and 3 mm as a working distance).

We agree that there is an overrepresentation of 5 kHz as the best tuning frequency for DC cells. The previous report (A. B. Wong & Borst, 2019) showed a large zone of the DC where cells were tuned to (2-8 kHz). Given that 5kHz was the lowest tested frequency in our experiment, we think that the low-frequency bias of the DC surface is consistent between studies. This finding also could be supported by the electrophysiology data obtained by spanning the recording electrodes through the IC tissue along the dorsoventral axis. In those experiments, the cells were tuned to lower frequencies at the dorsal surface of the IC.

We have changed the magenta-colored cells to green ones, so it will be easier to identify the cells. As required by another reviewer, we changed the color pallets of some images and cellular maps to be suitable for color-blind readers. 

The manuscript would benefit from more precise language in a number of places, especially in the results section.

Line 220/221, for instance: "... a significant fraction of cells that did not respond to pure tones did respond to AM-noise" Strictly speaking, this sentence suggests that you considered here only the subset of neurons that did not respond to pure tones and then ran a test on that subset. The test that was done seems to suggest though that the authors tested whether the percentage of responsive cells was greater for pure tones or for AM noise.

Thank you for the reviewer’s comment. We do apologize for the confusion. In the revised manuscript, we categorized the cells according to their response into cells responding to pure tone only (tone-selective cells or Tone-sel), Am-noise only (noise-selective cells or Nose-sel), and to both pure tone and am-noise (nonselective cells or Non-sel). We have modified Figure 5 accordingly. We did the same thing for the data obtained from awake animals and showed that in a new figure to easily match the analysis done for the anesthetized animals.

Please refer to the figure panels in the text in consecutive order. 2B, for instance, is mentioned after 2H.

Thank you for the reviewer’s comment. Throughout the paper, we kept the consecutive order of the figure panels within each figure to be in a smooth flow with the text. Yet, figure 2 was just the only exception for a good reason. Figure 2 is a complex one that includes many panels to show a parallel comparison between LC imaged via microprism and DC through single photon images, two-photon images, validating laser lesioning, and histology. Accordingly, we navigated many panels of the figure to efficiently highlight the aspects of this comparison. We prefer to keep Figure 2 as one figure with its current format to show this parallel comparison between LC and DC.

The legend for Figure 2 could be clearer. For instance, there are two descriptions for panel D. Line 1009: "(C-E)" [i.e. C, D, E] and line 1010: "(D and F)".

Thank you for the reviewer’s comment. It should be C and E, not C-E. We have fixed the mistake [line 1224]

Line 275: What does 'with no preference' mean?

Thank you for the reviewer’s comment. We do apologize for the confusion. There are three categories of cells. Some cells respond only to auditory stimulation, while others respond to only somatosensory stimulation. However, there is another group of cells that respond nonselectively to auditory and somatosensory stimulations or Aud/Som-nonsel cells. We edited the sentence to be clearer [lines 303-304].

Line 281 (and other places): What does 'normalized against modules' mean?

Thank you for the reviewer’s comment. This normalization was done by dividing the number of responsive cells of the same response type in the matrix by that in the modules. Therefore, the value taken by modules was always 1 and the value taken by the matrix is something around 1. Accordingly, the value for matrix could be > 1 if matrix had more cells than modules. In contrast, the value of matrix would be < 1 if matrix had fewer cells than modules. In the revised version, we used this normalization method to make the revised Figures 5C and 10C to describe the cell fractions responding to pure tone only, AM-noise only, or to both stimuli in the matrix vs modules. 

Sentence starting on line 288. I don't find that point to be as obvious from the figures as the sentences seem to suggest. Are we to compare magenta points (auditory off cells) from 7C with green points in 7F?

Thank you for the reviewer’s comment. We came to this conclusion based on our visual comparison of magenta points (now green in the revised version to increase the visibility) representing the auditory offset cells in Figure 7C and the green points in Figure 7F representing the cells responding to both somatosensory and auditory stimulations. In the revised manuscript, we statistically examined if the percentage of onset auditory response and offset auditory responses are different within the responsive cells to both somatosensory and auditory stimulations in the modules vs matrix. We have found that most of the cells responding to both somatosensory and auditory stimulations inside the modules had offset auditory responses, which could indicate a level of multisensory integration between somatosensory input and the offset auditory responses in these cells. We have added the statistical results to the revised manuscript to address this effect [lines 312-317]

Lines 300-302: "These data suggest that the module/matrix system permits preservation of distinct multimodal response properties in the face of massive integration of inputs in the LC". First, I'm not quite sure what that sentence means. Second, it would be more appropriate for the discussion. Third, the fact that we are more likely to find response enhancement in the modules than in the matrix is nicely consistent with the idea (supported by work from the senior author's lab and others) that excitatory somatosensory input predominantly targets neurons in the modules (which is why we see mostly response enhancement in the modules) and that this input targets GABAergic neurons which then project to and inhibit neurons both outside and inside of their module. Therefore, I would recommend that the authors replace the aforementioned sentence with one that interprets these results in light of what we know about this somatosensory-auditory circuitry.

Thank you for the reviewer’s comment. Despite the massive multimodal inputs, the LC receives from auditory vs nonauditory regions, the module/matrix system is a platform for distinct multimodal responses indicated by more somatosensory responsive cells in modules versus more auditory responsive cells in matrix, which matches the anatomical differences that were reported before. We edited the sentence in the light of the comparison between the data obtained from awake and anesthetized animals and moved it to the discussion section [lines 503-506].

The term 'LC imaged via microprism' is used dozens of times throughout the manuscript. Replacing it with a suitable acronym or initialism could improve the flow of the text and would make some of the sentences less cumbersome.

Thank you for the reviewer’s suggestion. We changed the term “LC imaged via microprism” into LC (microprism) throughout the revised manuscript.

5A-C: It is unclear what is being compared here. What are the Ns? Different animals?

Thank you for the reviewer’s comment. We do apologize for this missing information. We have added the number of subjects used in every statistical test in each corresponding figure legend.

5G: minus symbol missing on the y-axis.

Thank you for the reviewer’s comment. We gladly have fixed that.

Figure 6: Are these examples or population averages?

Thank you for the reviewer’s question. Every figure panel of the old Figure 6 represents a single trace of an example cell. However, we modified Figure 6 to include more examples of cells showing different responses complying with another reviewer’s suggestion. Each panel of the new Figure 6 represents the average response of 5 stimulations of the corresponding stimulus type. We preferred to show the average signal because it was the one used for the subsequent analysis.

How are module borders defined?

Thank you for the reviewer’s question. The modules were defined based on the intensity of the green channel that shows the expression of the GFP signals. The boundaries of modules were determined according to the distinction between high and low GFP signal boundaries of the modules. This step was done before data analysis to avoid any bias.

7JKL: How are these to be interpreted? Does panel 7K, for instance, indicate that the fraction of neurons showing 'on' responses was roughly twice as large in the matrix than in the modules and that the fraction of neurons showing 'off' responses was roughly 10 times larger in the modules than in the matrix (the mean seems to be at about 1/10).

Thank you for the reviewer’s comment. The data represented by Figures 7J-L defined the normalization of the number of cells of the same response type in the matrix against the modules. This normalization was done per animal, and then the data of the matrix were plotted against the normalization line at 1 representing the modules. The matrix will be claimed to have more cells than modules if the median of the matrix values > 1. In contrast, the matrix will be claimed to have fewer cells than the modules if the median of the matrix values < 1. Finally, if the median of matrix values = 1, this means there is no difference between matrix and modules. However, to match the data obtained from anesthetized animals (Figures 7 and 8) with those obtained from awake animals (Figure 11 or old Figure 10), we ran all data through the Chi-square test in the revised manuscript. Therefore, the format of Figures 7K-L was changed in the revised manuscript, so they became new Figures 7I-K.

10A suggests that significantly more than half the neurons shown here are not auditory responsive. Perhaps I am misinterpreting something here but isn't that in contrast to what is shown in panel 9F?

Thank you for the reviewer’s comment. The data shown in Figure 10A (or revised Figure 11A) represents the cellular response to only one stimulus (broadband noise at 80 dB with no modulation frequency), while Figure 9F (revised 10B) represents the cells responding to varieties of auditory stimulations of different combinations of frequencies and amplitudes (pure tones) as well as to AM-noise of different amplitudes and modulation frequencies. Accordingly, the old Figure 9F or revised Figure 10B shows different cell types based on their responses. For instance, some cells respond only to pure tone. Others respond only to AM-noise or to both pure tones and AM-noise. This may also support that the nonresponsive cells in Figure 10A (revised 11A) can respond to other types of sound features.

The way I understood panels 7L and 8K there were more suppressed neurons in the matrix than in the modules (line 296: "cells in the modules had a higher odds of having an enhancement response to bimodal stimulation than matrix, while cells in the matrix had a higher odds of having a suppressive response to bimodal stimulation"). Now, panel 10F indicates that in awake mice there is a greater proportion of suppressed neurons in the modules than in the matrix. I may very well have overlooked or misread something but I may not be the only reader confused by this so please clarify.

Thank you for the reviewer’s comment. We do apologize for this confusion. The ambiguity between Figures 7 and 8 (anesthetized animals) as well as Figure 10 (awake animals) comes from the fact that different statistics have been used for each preparation. In the revised version, we have fixed that by running the same statistics for all the data, and we accordingly revised Figures 7, 8, and 10 (new Figure 11). In brief, the matrix preserves a higher percentage of cells with suppressed auditory responses than those with enhanced auditory responses induced by bimodal stimulation in all conditions (anesthetized vs awake). In contrast, modules act differently across all tested conditions. While modules had more cells with enhanced auditory responses induced by bimodal interaction in anesthetized animals, they had more cells with suppressed response in awake animals indicating that modules could be sensitive to the effect of anesthesia compared to matrix. We addressed this effect in the discussion of the revised manuscript [lines 521-553].

Line 438: ...as early AS...

Thank you for the reviewer’s comment. We gladly fixed that [line 512].  

Reviewer #3 (Recommendations For The Authors):

My minor recommendations for the authors are as follows:

(1) The text can be a bit difficult to follow in places. This is partly due to the convoluted nature of the results, but I suggest a careful read-through to look for opportunities to improve the prose. In particular, there is a tendency to use long sentences and long paragraphs. For example, the third paragraph of the introduction runs for almost fifty lines.

Thank you for the reviewer’s comment and suggestion. We have fixed that.

(2) This might be due to journal compression, but some of the bar graphs in the figures are difficult to read. For example, the individual data points, especially when filled with striped background colors get lost. Axes can become invisible, like the y-axis in 7L, and portions of bars, like in 7F, are not always rendered correctly. Error bars are sometimes hidden behind data points, as in 5C. Increasing line thickness and shifting individual data points away from error bars might help with this.

Thank you for the reviewer’s comment and suggestion. We made all the data points with black color and filled circles to make the data points visible. We put all the data points behind the main columns, so they don’t block the error bars. We have fixed figures 7 and 5.

(3) Throughout the manuscript, the authors use a higher SMI to indicate a preference of cells for auditory stimuli with "greater spectral... complexity" (e.g., lines 219 and 372). I find this interpretation a bit challenging since SMI compares a neuron's preference for tones over noise, and to me, tones seem like the least spectrally complex of all auditory stimuli. Perhaps some clarification of what the authors mean by this would help. For example, is the assumption that a neuron that prefers tones over noise is, either directly or indirectly, receiving input sculpted by inhibitory processes?

Thank you for the reviewer’s comment. In general, higher SMI values indicate an increase in the preference of the cells to respond to pure tones than noise with no modulation (less spectral complexity). We will clarify this statement throughout the manuscript. However, the SMI value was not mentioned in lines 219 and 372. The statement mentioned in line 219 describes the revised figure 5C (old 5B), where more cells in matrix specifically respond to AM-noise compared to modules, which indicates the preference of the matrix to respond to sounds of greater spectral and temporal complexity. The statement in 372 in the discussion section refers to the finding in revised figures 5E and F (old 5D and E). In the revised figure 5E or old 5D, the data show that matrix has more cells responding to pure tones or noise with no modulation than modules, so matrix has a lower threshold to detect the spectral features of sound (revised figure 5E or old 5D). In the revised figure 5F or old 5E, the data show that matrix has more cells responding to AM-noise than modules, which indicates that matrix functions more to process the temporal features of sound. As explained above, all findings were related to the percentage of cells responding to specific sound stimuli and not the exact SMI values. We have revised the figures accordingly by removing the terms SMI and TMI from the figures, and we have clarified that in the text.

(4) Lines 250-253: How does a decrease in SMI correspond to "an increase in pure tone responsiveness?" Doesn't a decrease suggest the opposite?

Thank you for the reviewer’s comment, which we agree with. We do apologize for that. We have fixed this statement [lines 275-277] and any related findings accordingly.

(5) Line 304: Add "imaged via microprism" or similar after "response profiles with the LC.".

Thank you for the reviewer’s suggestion. We have fixed that. However, we changed the term “LC imaged via microprism” into “LC(microprism)” for simplicity as suggested by another reviewer [line 330].

(6) Figure 5A and C: Both plots show that more neurons responded to AM-noise than tones, but it would be interesting to know how much the tone-responsive and AM-noise responsive populations overlapped. Were all tone-responsive neurons also responsive to AM-noise?

Thank you for the reviewer’s comment. We have categorized the cells based on their response to pure tone only, AM-only, and both pure tone and AM-noise when each stimulus is presented individually. We have modified Figures 5A and C, and they are now Figures 5B and D.

(7) Figure 5G: Missing negative sign before "0.5.".

Thank you for the reviewer’s suggestion. We gladly have fixed that. However, old Figure 5G became a revised Figure 5H.  

(8) Figure 7 legend, Line 1102: Missing period after "(C and E)".

Thank you for the reviewer’s suggestion. We think that the period should be placed before (C and E) at the end of “respectively”. The parentheses refer to the statements after them. We gladly fixed that. [line 1394]

Author response:

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

Public Reviews:

Reviewer #1 (Public Review):

Summary:

This study reports that IT neurons have biased representations toward low spatial frequency

(SF) and faster decoding of low SFs than high SFs. High SF-preferred neurons, and low SF-preferred neurons to a lesser degree, perform better category decoding than neurons with other profiles (U and inverted U shaped). SF coding also shows more sparseness than category coding in the earlier phase of the response and less sparseness in the later phase. The results are also contrasted with predictions of various DNN models.

Strengths:

The study addressed an important issue on the representations of SF information in a high-level visual area. Data are analyzed with LDA which can effectively reduce the dimensionality of neuronal responses and retain category information.

We would like to express our sincere gratitude for your insightful and constructive comments which greatly contributed to the refinement of the manuscript. We appreciate the time and effort you dedicated to reviewing our work and providing suggestions. We have carefully considered each of your comments and addressed the suggested revisions accordingly.

Weaknesses:

The results are likely compromised by improper stimulus timing and unmatched spatial frequency spectrums of stimuli in different categories.

The authors used a very brief stimulus duration (35ms), which would degrade the visual system's contrast sensitivity to medium and high SF information disproportionately (see Nachmias, JOSAA, 1967). Therefore, IT neurons in the study could have received more degraded medium and high SF inputs compared to low SF inputs, which may be at least partially responsible for higher firing rates to low SF R1 stimuli (Figure 1c) and poorer recall performance with median and high SF R3-R5 stimuli in LDA decoding. The issue may also to some degree explain the delayed onset of recall to higher SF stimuli (Figure 2a), preferred low SF with an earlier T1 onset (Figure 2b), lower firing rate to high SF during T1 (Figure 2c), somewhat increased firing rate to high SF during T2 (because weaker high SF inputs would lead to later onset, Figure 2d).

We appreciate your concern regarding the course-to-fine nature of SF processing in the vision hierarchy and the short exposure time of our paradigm. According to your comment, we repeated the analysis of SF representation with 200ms exposure time as illustrated in Appendix 1 - Figure 4. Our recorded data contains the 200ms version of exposure time for all neurons in the main phase. As can be seen, the results are similar to what we found with 33 ms experiments.

Next, we bring your attention to the following observations:

(1) According to Figure 2d, the average firing rate of IT neurons for HSF could be higher than LSF in the late response phase. Therefore, the amount of HSF input received by the IT neurons is as much as LSF, however, its impact on the IT response is observable in the later phase of the response. Thus, the LSF preference is because of the temporal advantage of the LSF processing rather than contrast sensitivity.

(2) According to Figure 3a, 6% of the neurons are HSF-preferred and their firing rate in HSF is comparable to the LSF firing rate in the LSF-preferred group. This analysis is carried out in the early phase of the response (70-170 ms). While most of the neurons prefer LSF, this observation shows that there is an HSF input that excites a small group of neurons. Furthermore, the highest separability index also belongs to the HSF-preferred profile in the early phase of the response which supports the impact of the HSF part of the input.

(3) Similar LSF-preferred responses are also reported by Chen et al. (2018) (50ms for SC) and Zhang et al. (2023) (3.5 - 4 secs for V2 and V4) for longer duration times.

Our results suggest that the LSF-preferred nature of the IT responses in terms of firing rate and recall, is not due to the weakness or lack of input source (or information) for HSF but rather to the processing nature of the SF in the vision hierarchy.

To address this issue in the manuscript:

Figure Appendix 1 - Figure 4 is added to the manuscript and shows the recall value and onset for R1-R5 with 200ms of exposure time.

We added the following description to the discussion:

“To rule out the degraded contrast sensitivity of the visual system to medium and high SF information because of the brief exposure time, we repeated the analysis with 200ms exposure time as illustrated in Appendix 1 - Figure 4 which indicates the same LSF-preferred results. Furthermore, according to Figure 2, the average firing rate of IT neurons for HSF could be higher than LSF in the late response phase. It indicates that the amount of HSF input received by the IT neurons in the later phase is as much as LSF, however, its impact on the IT response is observable in the later phase of the response. Thus, the LSF preference is because of the temporal advantage of the LSF processing rather than contrast sensitivity. Next, according to Figure 3(a), 6\% of the neurons are HSF-preferred and their firing rate in HSF is comparable to the LSF firing rate in the LSF-preferred group. This analysis is carried out in the early phase of the response (70-170ms). While most of the neurons prefer LSF, this observation shows that there is an HSF input that excites a small group of neurons. Additionally, the highest SI belongs to the HSF-preferred profile in the early phase of the response which supports the impact of the HSF part of the input. Similar LSF-preferred responses are also reported by Chen et. al. (2018) (50ms for SC) and Zhang et. al. (2023) (3.5 - 4 secs for V2 and V4). Therefore, our results show that the LSF-preferred nature of the IT responses in terms of firing rate and recall, is not due to the weakness or lack of input source (or information) for HSF but rather to the processing nature of the SF in the IT cortex.”

Figure 3b shows greater face coding than object coding by high SF and to a lesser degree by low SF neurons. Only the inverted-U-shaped neurons displayed slightly better object coding than face coding. Overall the results give an impression that IT neurons are significantly more capable of coding faces than coding objects, which is inconsistent with the general understanding of the functions of IT neurons. The problem may lie with the selection of stimulus images (Figure 1b). To study SF-related category coding, the images in two categories need to have similar SF spectrums in the Fourier domain. Such efforts are not mentioned in the manuscript, and a look at the images in Figure 1b suggests that such efforts are likely not properly made. The ResNet18 decoding results in Figure 6C, in that IT neurons of different profiles show similar face and object coding, might be closer to reality.

Because of the limited number of stimuli in our experiments, it is hard to discuss the category selectivity, which needs a higher number of stimuli. To overcome the limited number of stimuli in our experiment, we fixed 60% (nine out of 15 stimuli) while varying the remaining stimuli to reduce the selective bias. To check the coding capability of the IT neurons for face and non-face objects, we evaluated the recall of face vs. non-face classification in intact stimuli (similar to classifiers stated in the manuscript). Results show that at the population level, the recall value for objects is 90.45%, and for faces is 92.45%. However, the difference is not significant (p-value=0.44). On the other hand, we note that a large difference in the SI value does not translate directly to the classification accuracy, rather it illustrates the strength of representation.

Regarding the SF spectrums, after matching the luminance and contrast of the images we matched the power of the images concerning SF and category. Powers are calculated using the sum of the absolute value of the Fourier transform of the image. Considering all stimuli, the ANOVA analysis shows that various SF bands have similar power (one-way ANOVA, p-value=0.24). Furthermore, comparing the power of faces and images in all SF bands (including intact) and both unscrambled and scrambled images indicates no significant difference between face and object (p-vale > 0.1). Therefore, the result of Figure 3b suggests that IT employs various SF bands for the recognition of various objects.

Comparing the results of CNNs and IT shows that the CNNs do not capture the complexities of the IT cortex in terms of SF. One of the sources of this difference is because of the behavioral saliency of the face stimulus in the training of the primate visual system.

To address this issue in the manuscript:

The following description is added to the discussion:

“… the decoding performance of category classification (face vs. non-face) in intact stimuli is 94.2%. The recall value for objects vs. scrambled is 90.45%, and for faces vs. scrambled is 92.45% (p-value=0.44), which indicates the high level of generalizability and validity characterizing our results.”

The following description is added to the method section, SF filtering.

“Finally, we equalized the stimulus power in all SF bands (intact, R-R5). The SF power among all conditions (all SF bands, face vs. non-face and unscrambled vs. scrambled) does not vary significantly (p-value > 0.1). SF power is calculated as the sum of the square value of the image coefficients in the Fourier domain.”

Reviewer #2 (Public Review):

Summary:

This paper aimed to examine the spatial frequency selectivity of macaque inferotemporal (IT) neurons and its relation to category selectivity. The authors suggest in the present study that some IT neurons show a sensitivity for the spatial frequency of scrambled images. Their report suggests a shift in preferred spatial frequency during the response, from low to high spatial frequencies. This agrees with a coarse-to-fine processing strategy, which is in line with multiple studies in the early visual cortex. In addition, they report that the selectivity for faces and objects, relative to scrambled stimuli, depends on the spatial frequency tuning of the neurons.

Strengths:

Previous studies using human fMRI and psychophysics studied the contribution of different spatial frequency bands to object recognition, but as pointed out by the authors little is known about the spatial frequency selectivity of single IT neurons. This study addresses this gap and they show that at least some IT neurons show a sensitivity for spatial frequency and

interestingly show a tendency for coarse-to-fine processing.

We extend our sincere appreciation for your thoughtful and constructive feedback on our paper. We are grateful for the time and expertise you invested in reviewing our work. Your detailed suggestions have been instrumental in addressing several key aspects of the paper, contributing to its clarity and scholarly merit. We have carefully considered each of your comments and have made revisions accordingly.

Weaknesses and requested clarifications:

(1) It is unclear whether the effects described in this paper reflect a sensitivity to spatial frequency, i.e. in cycles/ deg (depends on the distance from the observer and changes when rescaling the image), or is a sensitivity to cycles /image, largely independent of image scale. How is it related to the well-documented size tolerance of IT neuron selectivity?

Our stimuli are filtered using cycles/images and knowing the distance of the subject to the monitor, we can calculate the cycles/degrees. To the best of our knowledge, this is also the case for all other SF-related studies. To find the relation of observations to the cycles/image and degree/image, one should keep one of them fixed while changing the other, for example changing the subject's distance to the monitor will change the SF content in terms of cycle/degree. With our current data, we cannot discriminate this effect. To address this issue, we added the following description to the discussion. To address this issue, we added the following description to the discussion:

“Finally, since our experiment maintains a fixed SF content in terms of both cycles per degree and cycles per image, further experiments are needed to discern whether our observations reflect sensitivity to cycles per degree or cycles per image.”

(2) The authors' band-pass filtered phase scrambled images of faces and objects. The original images likely differed in their spatial frequency amplitude spectrum and thus it is unclear whether the differing bands contained the same power for the different scrambled images. If not, this could have contributed to the frequency sensitivity of the neurons.

After equalizing the luminance and contrast of the images, we equilized their power concerning SF and category. The powers were calculated using the sum of the absolute values of the Fourier transform of the images. The results of the ANOVA analysis across all stimuli indicate that various SF bands exhibit similar power (one-way ANOVA, p-value = 0.24). Additionally, a comparison of power between faces and objects in all SF bands (including intact), for both unscrambled and scrambled images, reveals no significant differences (p-value > 0.1). To clarify this point, we have incorporated the following information into the Methods section.

“Finally, we equalized the stimulus power in all SF bands (intact, R-R5). The SF power among all conditions (all SF bands, face vs. non-face and unscrambled vs. scrambled) does not vary significantly (ANOVA, p-value > 0.1).”

(3) How strong were the responses to the phase-scrambled images? Phase-scrambled images are expected to be rather ineffective stimuli for IT neurons. How can one extrapolate the effect of the spatial frequency band observed for ineffective stimuli to that for more effective stimuli, like objects or (for some neurons) faces? A distribution should be provided, of the net responses (in spikes/s) to the scrambled stimuli, and this for the early and late windows.

The sample neuron in Figure 1c is chosen to be a good indicator of the recorded neurons. In the early response phase, the average firing rate to scrambled stimuli is 26.3 spikes/s which is significantly higher than the response in -50 to 50ms which is 23.4. In comparison, the mean response to intact face stimuli is 30.5 spikes/s, while object stimuli elicit an average response of 28.8 spikes/s. Moving to the late phase, T2, the responses to scrambled, face, and object stimuli are 19.5, 19.4, and 22.4 spikes/s, respectively. Moreover, when the classification accuracy for SF exceeds chance levels, it indicates a significant impact of SF bands on the IT response. This raises a direct question about the explicit coding for SF bands in the IT cortex observed for ineffective stimuli and how it relates to complex and effective stimuli, such as faces. To show the strength of neuron responses to the SF bands in scrambled images, We added Appendix 1 - Figure 2 and also added Appendix 1 - Figure 1, according to comment 4, which shows the average and std of the responses to all SF bands. The following description is added to the results section.

“Considering the strength of responses to scrambled stimuli, the average firing rate in response to scrambled stimuli is 26.3 Hz, which is significantly higher than the response observed between -50 and 50 ms, where it is 23.4 Hz (p-value=3x10-5). In comparison, the mean response to intact face stimuli is 30.5 Hz, while non-face stimuli elicit an average response of 28.8 Hz. The distribution of neuron responses for scrambled, face, and non-face in T1 is illustrated in Appendix 1 - Figure 2.

[…]

Moreover, the average firing rates of scrambled, face, and non-face stimuli are 19.5 Hz, 19.4 Hz, and 22.4 Hz, respectively. The distribution of neuron responses is illustrated in Appendix 1 Figure 2.”

(4) The strength of the spatial frequency selectivity is unclear from the presented data. The authors provide the result of a classification analysis, but this is in normalized units so that the reader does not know the classification score in percent correct. Unnormalized data should be provided. Also, it would be informative to provide a summary plot of the spatial frequency selectivity in spikes/s, e.g. by ranking the spatial frequency bands for each neuron based on half of the trials and then plotting the average responses for the obtained ranks for the other half of the trials. Thus, the reader can appreciate the strength of the spatial frequency selectivity, considering trial-to-trial variability. Also, a plot should be provided of the mean response to the stimuli for the two analysis windows of Figure 2c and 2d in spikes/s so one can appreciate the mean response strengths and effect size (see above).

The normalization of the classification result is just obtained by subtracting the chance level, which is 0.2, from the whole values. Therefore the values could still be interpreted in percent as we did in the results section. To make this clear, we removed the “a.u.” from the figure and we added the following description to the results section.

“The accuracy value is normalized by subtracting the chance level (0.2).”

Regarding the selectivity of the neuron, as suggested by your comment, we added a new figure in the appendix section, Appendix 1 - figure 2. This figure shows the strength of SF selectivity, considering trial-to-trial variability. The following description is added to the results section:

“The strength of SF selectivity, considering the trial-to-trial variability is provided in Appendix 1 Figure 2, by ranking the SF bands for each neuron based on half of the trials and then plotting the average responses for the obtained ranks for the other half of the trials.”

The firing rates of Figures 2c and 2d are normalized for better illustration since the variation in firing rates is high across neurons, as can be observed in Figure Appendix 1 - Figure 1. Since we seek trends in the response, the absolute values are not important (since the baseline firing rates of neurons are different), but the values relative to the baseline firing rate determine the trend. To address the mean response and the strength of the SF response, the following description is added to the results section.

“Considering the strength of responses to scrambled stimuli, the average firing rate in response to scrambled stimuli is 26.3 Hz, which is significantly higher than the response observed between -50 and 50 ms, where it is 23.4 Hz (p-value=3x10-5). In comparison, the mean response to intact face stimuli is 30.5 Hz, while non-face stimuli elicit an average response of 28.8 Hz. The distribution of neuron responses for scrambled, face, and non-face in T1 is illustrated in Appendix 1 - Figure 2.

[…]

Moreover, the average firing rates of scrambled, face, and non-face stimuli are 19.5 Hz, 19.4

Hz, and 22.4 Hz, respectively. The distribution of neuron responses is illustrated in Appendix 1 Figure 2.”

Furthermore, we added a figure, Appendix 1 - Figure 3, to illustrate the strength of SF selectivity in our profiles. The following is added to the results section:

“To check the robustness of the profiles, considering the trial-to-trial variability, the strength of SF selectivity in each profile is provided in Appendix 1 - Figure 3, by forming the profile of each neuron based on half of the trials and then plotting the average SF responses with the other

half of the trials.”

(5) It is unclear why such brief stimulus durations were employed. Will the results be similar, in particular the preference for low spatial frequencies, for longer stimulus durations that are more similar to those encountered during natural vision?

Please refer to the first comment of Reviewer 1.

(6) The authors report that the spatial frequency band classification accuracy for the population of neurons is not much higher than that of the best neuron (line 151). How does this relate to the SNC analysis, which appears to suggest that many neurons contribute to the spatial frequency selectivity of the population in a non-redundant fashion? Also, the outcome of the analyses should be provided (such as SNC and decoding (e.g. Figure 1D)) in the original units instead of undefined arbitrary units.

The population accuracy is approximately 5% higher than the best neuron. However, we have no reference to compare the effect size (the value is roughly similar for face vs object while the chance levels are different). However, as stated in Methods, SNC is calculated for two label modes (LSF and HSF) and it can not be directly compared to the best neuron accuracy. Regarding the unit of SNC, it can be interpreted directly to percent by multiplying by a factor of 100. We removed the “a.u.” to prevent misunderstanding and modified the results section for clearance.

“… SNC score for SF (two labels, LSF (R1 and R2) vs. HSF (R4 and R5)) and category … (average SNC for SF=0.51\%±0.02 and category=0.1\%±0.04 …”

(7) To me, the results of the analyses of Figure 3c,d, and Figure 4 appear to disagree. The latter figure shows no correlation between category and spatial frequency classification accuracies while Figure 3c,d shows the opposite.

In Figure 3c,d, following what we observed in Figure 3a,b about the category coding capabilities in the population of neurons based on the profile of the single neurons, we tested a similar idea if the coding capability of single neurons in SF/category could predict the coding capability of population neurons in terms of category/SF. Therefore, both analyses investigate a relation between a characteristic of single neurons and the coding capability of a population of similar neurons. On the other hand, in Figure 4, the idea is to check the characteristics of the coding mechanisms behind SF and category coding. In Figure 4a, we check if there exists any relation between category and SF coding capability within a single neuron activity without the impact of other neurons, to investigate the idea that SF coding may be a byproduct of an object recognition mechanism. In Figure 4b, we investigated the contribution of all neurons in population decision, again to check whether the mechanisms behind the SF and category coding are the same or not. This analysis shows how individual neurons contribute to SF or category coding at the population level. Therefore, the experiments in Figures 3 and 4 are different in the analysis method and what they were designed to investigate and we cannot directly compare the results.

(8) If I understand correctly, the "main" test included scrambled versions of each of the "responsive" images selected based on the preceding test. Each stimulus was presented 15 times (once in each of the 15 blocks). The LDA classifier was trained to predict the 5 spatial frequency band labels and they used 70% of the trials to train the classifier. Were the trained and tested trials stratified with respect to the different scrambled images? Also, LDA assumes a normal distribution. Was this the case, especially because of the mixture of repetitions of the same scrambled stimulus and different scrambled stimuli?

In response to your inquiry regarding the stratification of trials, both the training and testing data were representative of the entire spectrum of scrambled images used in our experiment. To address your concern about the assumption of a normal distribution, especially given the mixture of repetitions of the same scrambled stimulus and different stimuli, our analysis of firing rates reveals a slightly left-skewed normal distribution. While there is a deviation from a perfectly normal distribution, we are confident that this skewness does not compromise the robustness of the LDA classifier.

(9) The LDA classifiers for spatial frequency band (5 labels) and category (2 labels) have different chance and performance levels. Was this taken into account when comparing the SNC between these two classifiers? Details and SNC values should be provided in the original (percent difference) instead of arbitrary units in Figure 5a. Without such details, the results are impossible to evaluate.

For both SNC and CMI calculations in SF, we considered two labels of HSF (R4 and R5) and LSF (R1 and R2). This was mentioned in the Methods section, after equation (5). According to your comment, to make it clear in the results section, we also added this description to the results section.

“… illustrates the SNC score for SF (two labels, LSF (R1 and R2) vs. HSF (R4 and R5)) and category (face vs. non-face) … conditioned on the label, SF (LSF (R1 and R2) vs. HSF (R4 and R5)) or category, to assess the information.”

The value of SNC can also be directly converted to the percent by a factor of 100. To make it clear, we removed “a.u.” from the y-axis.

(10) Recording locations should be described in IT, since the latter is a large region. Did their recordings include the STS? A/P and M/L coordinate ranges of recorded neurons?

We appreciate your suggestion for the recording location. Nevertheless, given the complexities associated with neurophysiological recordings and the limitations imposed by our methodologies, we face challenges in precisely localizing every unit if they are located in STS or not. To address your comment, We added Appendix 1 - Figure 5 which shows the SF and category coding capability of neurons along their recorded locations.

(11) The authors should show in Supplementary Figures the main data for each of the two animals, to ensure the reader that both monkeys showed similar trends.

We added Appendix 2 which shows the consistency of the main results in the two monkeys.

(12) The authors found that the deep nets encoded better the spatial frequency bands than the IT units. However, IT units have trial-to-trial response variability and CNN units do not. Did they consider this when comparing IT and CNN classification performance? Also, the number of features differs between IT and CNN units. To me, comparing IT and CNN classification performances is like comparing apples and oranges.

Deep convolutional neural networks are currently considered the state-of-the-art models of the primate visual pathway. However, as you mentioned and based on our results, they do not yet capture various complexities of the visual ventral stream. Yet studying the similarities and differences between CNN and brain regions, such as the IT cortex, is an active area of research, such as:

a. Kubilius, Jonas, et al. "Brain-like object recognition with high-performing shallow recurrent ANNs." Advances in neural information processing systems 32 (2019).

b. Xu, Yaoda, and Maryam Vaziri-Pashkam. "Limits to visual representational correspondence between convolutional neural networks and the human brain." Nature Communications, 12.1 (2021).

c. Jacob, Georgin, et al. "Qualitative similarities and differences in visual object representations between brains and deep networks." Nature Communications, 12.1 (2021).

Therefore, we believe comparing IT and CNN, despite all of the differences in terms of their characteristics, can help both fields grow faster, especially in introducing brain-inspired networks.

(13) The authors should define the separability index in their paper. Since it is the main index to show a relationship between category and spatial frequency tuning, it should be described in detail. Also, results should be provided in the original units instead of undefined arbitrary units. The tuning profiles in Figure 3A should be in spikes/s. Also, it was unclear to me whether the classification of the neurons into the different tuning profiles was based on an ANOVA assessing per neuron whether the effect of the spatial frequency band was significant (as should be done).

Based on your comment, we added the description of the separability index to the methods section. However, since the separability index is defined as the division of two dispersion matrices, it has no units by nature. The tuning profiles in Figure 3a are normalized for better illustration since the variation in firing rates is high. Since we seek trends in the response, the absolute values are not important. Regarding the SF profile formation, to better present the SF profile assignment, we updated the method section. Furthermore, The strength of responses for scrambled stimuli can be observed in Appendix 1 - Figures 1 and 2.

(14) As mentioned above, the separability analysis is the main one suggesting an association between category and spatial frequency tuning. However, they compute the separability of each category with respect to the scrambled images. Since faces are a rather homogeneous category I expect that IT neurons have on average a higher separability index for faces than for the more heterogeneous category of objects, at least for neurons responsive to faces and/or objects. The higher separability for faces of the two low- and high-pass spatial frequency neurons could reflect stronger overall responses for these two classes of neurons. Was this the case? This is a critical analysis since it is essential to assess whether it is category versus responsiveness that is associated with the spatial frequency tuning. Also, I do not believe that one can make a strong claim about category selectivity when only 6 faces and 3 objects (and 6 other, variable stimuli; 15 stimuli in total) are employed to assess the responses for these categories (see next main comment). This and the above control analysis can affect the main conclusion and title of the paper.

We appreciate your concern regarding category selectivity or responsiveness of the SF profiles. First, we note that we used SI since it overcomes the limitations of the accuracy and recall metrics as they are discrete and can be saturated. Using SI, we cannot directly calculate face vs object with SI, since this index only reports one value for the whole discrimination task. Therefore, we have to calculate the SI for face/object vs scrambled to obtain a value per category. However, as you suggested, it raises the question of whether we assess how well the neural responses distinguish between actual images (faces or objects) and their scrambled versions or if we just assess the responsiveness. Based on Figure 3b, since we have face-selective (LSF and HSF preferred profiles), object-selective (inverse U), and the U profile, where SI is the same for both face and object, we believe the SF profile is associated with the category selectivity, otherwise we would have the same face/object recall in all profiles, as we have in the U shape profile.

To analyze this issue further, we calculated the number of face/object selective neurons in 70-170ms. We found 43 face-selective neurons and 36 object-selective neurons (FDR corrected p-value < 0.05). Therefore, the number of face-selective and object-selective neurons is similar. Next, we check the selectivity of the neurons within each profile. Number of face/object selective neurons is LP=13/3, HP=6/2, IU=3/9, U=14/13, and the remaining belong to the NP group. Results show higher face-selective neurons in LP and HP and a higher number of object-selective neurons in the IU class. The U class contains roughly the same number of face and object-selective neurons. This observation supports the relationship between category selectivity and profiles.

Next, we examined the average neuron response to the face and object in each profile. The difference between the firing rate of the face and object in none of the profiles was significant (Ranksum with a significance level of 0.05). However, the rates are as follows. The average firing rate (spikes/s) of face/object is LP=36.72/28.77, HP=28.55/25.52, IU=21.55/27.25, U=38.48/36.28. While the differences are not significant, they support the relationship between profiles and categories instead of responsiveness.

The following description is added to the results section to cover this point of view.

“To assess whether the SF profiles distinguish category selectivity or merely evaluate the neuron's responsiveness, we quantified the number of face/non-face selective neurons in the 70-170ms time window. Our analysis shows a total of 43 face-selective neurons and 36 non-face-selective neurons (FDR-corrected p-value < 0.05). The results indicate a higher proportion of face-selective neurons in LP and HP, while a greater number of non-face-selective neurons is observed in the IU category (number of face/non-face selective neurons: LP=13/3, HP=6/2, IU=3/9). The U category exhibits a roughly equal distribution of face and non-face-selective neurons (U=14/13). This finding reinforces the connection between category selectivity and the identified profiles. We then analyzed the average neuron response to faces and non-faces within each profile. The difference between the firing rates for faces and non-faces in none of the profiles is significant (face/non-face average firing rate (Hz): LP=36.72/28.77, HP=28.55/25.52, IU=21.55/27.25, U=38.48/36.28, Ranksum with significance level of 0.05). Although the observed differences are not statistically significant, they provide support for the association between profiles and categories rather than mere responsiveness.”

About the low number of stimuli, please check the next comment.

(15) For the category decoding, the authors employed intact, unscrambled stimuli. Were these from the main test? If yes, then I am concerned that this represents a too small number of stimuli to assess category selectivity. Only 9 fixed + 6 variable stimuli = 15 were in the main test. How many faces/ objects on average? Was the number of stimuli per category equated for the classification? When possible use the data of the preceding selectivity test which has many more stimuli to compute the category selectivity.

We used only the main phase recorded data, which contains 15 images in each session. Each image results in 12 stimuli (intact, R1-R5, and phase-scrambled version). Thus, there exists a total of 180 unique stimuli in each session. Increasing the number of images would have increased the recording time. We compensated for this limitation by increasing the diversity of images in each session by picking the most responsive ones from the selectivity phase. On average, 7.54 of the stimuli were face in each session. We added this information to the Methods section. Furthermore, as mentioned in the discussion, for each classification run, the number of samples per category is equalized. We note that we cannot use the selectivity data for analysis, since the SF-related stimuli are filtered in different bands.

Recommendations For The Authors:

Reviewer #1 (Recommendations For The Authors):

I suggest that the authors double-check their results by performing control experiments with longer stimulus duration and SF-spectrum-matched face and object stimuli.

Thanks for your suggestion, according to your comment, we added Appendix 1 - Figure 3.

In addition, I had a very difficult time understanding the differences between Figure 3c and Figure 4a. Please rewrite the descriptions to clarify.

Thanks for your suggestion, we tried to revise the description of these two figures. The following description is added to the results section for Figure 3c.

“Next, to examine the relation between the SF (category) coding capacity of the single neurons and the category (SF) coding capability of the population level, we calculated the correlation between coding performance at the population level and the coding performance of single neurons within that population (Figure 3 c and d). In other words, we investigated the relation between single and population levels of coding capabilities between SF and category. The SF (or category) coding performance of a sub-population of 20 neurons that have roughly the same single-level coding capability of the category (or SF) is examined.”

Lines 147-148: The text states that 'The maximum accuracy of a single neuron was 19.08% higher than the chance level'. However, in Figure 4, the decoding accuracies of individual neurons for category and SF range were between 49%-90% and 20%-40%, respectively.

Please explain the discrepancies.

The first number is reported according to chance level which is 20%, thus the unnormalized number is 39% which is consistent with the SF accuracy in Figure 4. We added the following description to prevent any misunderstanding.

“… was 19.08\% higher than the chance level (unnormalized accuracy is 49.08\%, neuron \#193, M2).”

Lines 264-265: Should 'the alternative for R3 and R4' be 'the alternative for R4 and R5'?

Thanks for your attention, it's “R4 and R5”. We corrected that mistake.

Lines 551-562: The labels for SF classification are R1-R5. Is it a binary or a multi-classification task?

It’s a multi-label classification. We made it clear in the text.

“… labels were SF bands (R1, R2, ..., R5, multi-label classifier).”

Figure 4b: Neurons in SF/category decoding exhibit both positive and negative weights. However, in the analysis of sparse neuron weights in Equation 6, only the magnitude of the weights is considered. Is the sign of weight considered too?

We used the absolute value of the neuron weight to calculate sparseness. We also corrected Equation 6.

Reviewer #2 (Recommendations For The Authors):

(1) Line 52: what do the authors mean by coordinate processing in object recognition?

To avoid any potential misunderstanding, we used the exact phrase in Saneyoshi and Michimata (2015). It is in fact, coordinate relations processing. Coordinate relations specify the metric information of the relative locations of objects.

(2) About half of the Introduction is a summary of the Results. This can be shortened.

Thanks for your suggestion.

(3) Line 134: Peristimulus time histogram instead of Prestimulus time histogram.

Thanks for your attention. We corrected that.

(4) Line 162: the authors state that R1 is decoded faster than R5, but the reported statistic is only for R1 versus R2.

It was a typo, the p-value is only reported for R1 and R5.

(5) Line 576: which test was used for the asses the statistical significance?

The test is Wilcoxon signed-rank. We added it to the text.

(6) How can one present a 35 ms long stimulus with a 60 Hz frame rate (the stimuli were presented on a 60Hz monitor (line 470))? Please correct.