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

Learn more at Review Commons

---

Referee #4

Evidence, reproducibility and clarity

This is a short report featuring an innovative proximity labeling approach to the identification of proteins enriched in distinct types of chromatin domains. The domains compared are centromeric heterochromatin and X-linked hyperactive chromatin in Drosophila cells. These are relatively well-described domains, thus serving as an excellent test for the targeting of biotinylation in the permeabilized nucleus via interaction of specific antibodies with ProteinA-Apex2 provided exogenously. In parallel with the signature chromatin proteins CID or MSL2 as baits, the authors also target proteins in proximity to specific histone tail PTMs. Taking the work one step further, they compare the recovery of proteins +/- pretreatment of nuclei with RNase. They conclude that in each case selective interactions are specifically lost with pre-treatment of RNase.

Major comment:

As mentioned above, the approach is innovative and raises the possibility of a simpler MS method to identify protein-protein interactions. The RNase result is also provocative. However, in each case the specificity of potentially novel results are not explored further. Thus, the work is of interest but clearly still preliminary.

Significance

Did the authors dig deeper into novel interactions without obtaining convincing validation? Did they conclude that the MS approach is worth pursuing further or not? Admittedly the RNase result is difficult to follow up, but additional discussion of prior related work as well as consideration of future experiments would help improve the manuscript.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

The manuscript entitled 'The role of RNA in the maintenance of chromatin domains as revealed by antibody mediated proximity labelling coupled to mass spectrometry' by Choudhury et al. describe a new method, which they termed AMPL-MS (Antibody mediated proximity labelling mass spectrometry). The technique is based on proximity labelling but uses antibodies instead of fusion proteins. They use this method to characterize chromatin domains containing specific signature proteins or histone modifications and focus on the composition of chromocenter as well as the chromosome territory containing the hyperactive X-chromosome in Drosophila. Last but not least they include data that show that RNA is involved in maintaining the integrity of chromatin domains by RNAse treatment and mass spec analysis. The technique works well and the results are very clear. I therefore expect that, in the right hands, it is very reproducible.

There are a few points that the authors may want to address:

1. Title 'The' role of RNA in the maintenance of chromatin domains as..., seems too much of a statement. The title is therefore an overstatement that needs to be fixed.
2. Figure 1 In Figure 1 the authors show very convincingly that the methods works well in their hands. They report on 172 proteins that localized in proximity to CID containing centromeric chromatin but do not provide the list of proteins as far as I can tell. Especially the RNA binders should be named.
3. Figure 2 Using the hyperactive X is very clever when addressing RNA function but it should be stated in the discussion that there may be certain aspects that are specific to the male x and that is impossible to discriminate general and specific effects uncovered by this method.
4. Figure 3 The authors should state more clearly the new findings of this figure since it is not fully obvious from its current representation.
5. Figure 4 These are certainly interesting data but the authors remain in the very descriptive state. This is fine for a methods paper but then, the authors should hypothesize more on what the results mean. Are certain RNA dependent factors specific or general and they then recruit a specific set of factors that fall off upon RNAse treatment as a secondary effect or because they bind RNA directly. I feel like there may be more information that they authors got get out of there data than what they currently provide.
6. Discussion The authors state: 'While we have not identified the RNAs responsible for the formation of theses domains, we clearly observe that they do confer specificity for the domains as we observe very little overlap in the factors lost from the corresponding domains (Fig 4h). the 'specificity' is hard to determine since factors bound to these regions are different, and therefore different factors will fall off, regardless of whether the RBP are specific unless the RNA is involved in recruiting the factors specifically, which the authors have not shown. Therefore, this result is suggestive and interesting but the statement is too strong and not backed by their results.

Significance

Overall, this is an interesting method that has been used in the past to identify protein modifications with high quality antibodies available. The authors show here that the method can also be used to different nuclear proteins and detect changes in protein complex composition. As it is it is primarily a methods paper, and for that the results are very clear. Gain of new info is not large but it is a useful technique to continue research on this subjectand is a nice start of many new avenues into how RNA effects chromatin,

My expertises are in epigenetics, chromatin biology, RNA and Drosophila genetics

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

This manuscript by Choudhury et al. describes a new method for antibody-mediated proximity labeling and applies it in the cell nucleus. In short, nuclei are isolated, fix and permeabilized, proteins are labeled with primary antibodies, a bacterially expressed/purified protein-A-APEX2 fusion protein is added, conventional H2O2/biotin phenol labeling of proximate protein is performed, proteins are un-crosslinked and biotin-affinity captured for MS analysis. The application to nuclear proteins and results seems appropriate. The method is highly similar to and more complicated than prior methods as described in more detail below. I would focus the impact of this paper towards its biological results and not the novelty of the methods used.

Prior methods that effectively accomplish the same outcome (fixed cells/tissues, antibodies and proximity labeling for AP-MS) have been published before. Perhaps most recently it was reinvented as the so called BAR method in PMID 29256494. That paper was cited here but incorrectly as BirA-related, which it is not. Of course that prior manuscript itself ignored prior methods from years back (2008, 2012, 2014, 2015, PMID 18495923, 22936677, 24706754, 25829300) using the same approaches of antibody targeted peroxidase for the same purposes of proximity labeling.

This method seems a somewhat Rube Golderbergian approach to antibody-mediated proximity labeling, which has been performed previously in multiple reports. APEX/2was developed to function inside of living cells since HRP does not. The value of doing the proximity labeling in living cells was either to capture protein associations over time, as with BioID/TurboID, or to get snapshots of protein associations in living cells with APEX/2. HRP does however function quite well for proximity labeling outside of cells, or in fixed/permeabilized cells, as has been demonstrated in the prior methods/papers that are referenced above. Replacing commercially available secondary antibodies fused to HRP with homemade protein-A-fused to APEX2 seems counterintuitive and/or unnecessary.

Could the authors explain the mechanisms that underly the reported enhanced sensitivity of AMPL-MS compared to conventional APEX2 in living cells. Is there something about the nuclear isolation that reduces interfering background, the loss of small soluble molecules in the nucleus after isolation and/or permeabilization that enhance the proximity labeling, penetration issues with the biotin-phenol in living cells, and/or something else?

There seems to be the use of various controls based on the figures and legends, but they are not clearly described in the results or methods.

All MS results should be provided, preferably in an Excel file format.

Significance

This manuscript by Choudhury et al. describes a new method for antibody-mediated proximity labeling and applies it in the cell nucleus. In short, nuclei are isolated, fix and permeabilized, proteins are labeled with primary antibodies, a bacterially expressed/purified protein-A-APEX2 fusion protein is added, conventional H2O2/biotin phenol labeling of proximate protein is performed, proteins are un-crosslinked and biotin-affinity captured for MS analysis. The application to nuclear proteins and results seems appropriate. The method is highly similar to and more complicated than prior methods as described in more detail below. I would focus the impact of this paper towards its biological results and not the novelty of the methods used.

Prior methods that effectively accomplish the same outcome (fixed cells/tissues, antibodies and proximity labeling for AP-MS) have been published before. Perhaps most recently it was reinvented as the so called BAR method in PMID 29256494. That paper was cited here but incorrectly as BirA-related, which it is not. Of course that prior manuscript itself ignored prior methods from years back (2008, 2012, 2014, 2015, PMID 18495923, 22936677, 24706754, 25829300) using the same approaches of antibody targeted peroxidase for the same purposes of proximity labeling.

This work may be of interest to investigators studying the nuclear proteins/structures to which the APML-MS was applied.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

Summary:

This manuscript describes the application of a new variation of proximity (biotin) labelling (antibody mediated proximity labelling coupled to mass spectrometry, AMPL-MS). Combining protein- or histone variant-specific antibodies with a APEX2-proteinA fusion protein, they characterise the proteome of nuclear subdomains and demonstrate that RNA is important for the integrity of two tested domains, the Drosophila chromocenter and the chromosomal territory of the hyperactive X chromosome.

Major comments:

The vast majority the experimental results, statistical analysis, and conclusions drawn by the authors appear sound and are described in way that should allow reproduction (however, see my comments below for some suggestions for minor improvements). The authors rigorously test their method, using the Drosophila chromodomain as 'playground', before applying it to other chromosomal areas and histone variants/modifications. Besides providing proteomes of the targeted nuclear subcompartments, they show that RNase treatment of the cells radically changes the proteome(s) and conclude a role for RNA in the integrity of the corresponding compartments. This is shown by immunofluorescence staining as well as proteomic analysis of the biotinylated proteins. The images in figure 4b (and to lesser extent 4c) show an increased intensity and more diffuse labelling. Can the authors exclude that RNase treatment simply leads to an increase in accessibility for the biotin-phenol, hence a visibly higher biotinylation? Along these lines, have the authors maybe observed an increase in overall labelling/pulldown efficiency or for biotinylated proteins in their proteomic data?

Minor comments:

1. In figures 1a and 4a (as well as in the Methods section), the authors use the term 'biotin-tyramide' as labelling agent, but in the main text and figure legends 'biotin-phenol' is used. For clarity, only one term should be used.
2. Figure 2a shows a magnified cell/nucleus in the last column. To what cells do the magnifications in this last column refer to? Maybe these cells could be boxed in the second last column?
3. In figures 4b + c:, the figure legend mentions the individual rows as '(I)' and '(II)' but no such label seen in the corresponding panel(s).
4. The Quantification method for co-localization (e.g. 1c and 2b) is insufficiently described to the reader (reference simply relates to Fiji package). What module/script within the Fiji package has been used?
5. The RNase treatment is not described at all in the methods section or the supplementary information and should be added.
6. The sentence on page 6 ('As expected, neither the targeted signature factor or proteins that mainly interact with them protein-protein interactions such as MSL1,3 and MOF for MSL2 or Cenp-C for Cid are not affected by RNAase treatment') should be rephrased as it is not comprehensible in the current form.

Significance

The findings in this manuscript advance our portfolio of proximity labelling techniques although this advancement is not a major step forward. As the authors state themselves, antibody-based proximity labelling has already been introduced, even in the context of chromosomal proteomes (e.g. Gan et al., 2022; https://doi.org/10.1016/j.gpb.2021.09.003). One major technical advance is the finding that modifications or protein variants can now reliably be targeted for proximity labelling, using their method. Furthermore, the number of cells that are required for proximity labelling and detection of biotinylated proteins could be significantly reduced compared to previous approaches (although this might simply be due to the use of a more advanced proximity labelling enzyme). I should state her that as I am not an expert in the field of chromatin domains, I cannot be certain if the proteomes and changes of proteomes the authors report are providing a significant increase in our knowledge on these domains, especially related to their individual functions.

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

Learn more at Review Commons

---

Reply to the reviewers

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

In the paper entitled GOT1 primes the cellular response to hypoxia by supporting glycolysis and HIF1α stabilization, Grimm and co-authors investigate the metabolic adaptations of cancer cells upon acute hypoxia. By measuring metabolite levels at early time points upon hypoxia, they observe the accumulation of lactate and depletion of aspartate, along with other TCA cycle metabolites. Importantly, they demonstrate that these metabolic changes are independent of the HIF alpha-dependent transcriptional response. The authors investigate the role of aspartate during these initial phases of hypoxia. To this aim, they characterize cells devoid of glutamate oxaloacetate transaminase (GOT1), in which aspartate accumulates and can no longer be used for replenishing NAD+ via the downstream conversion of oxaloacetate to malate, via malate dehydrogenase. These cells have lower cytosolic NAD+ which affects glycolytic flux through the rate-limiting, NAD+-dependent enzyme GAPDH. GOT1 KO cells have a decrease in glucose consumption, lactate secretion and metabolite levels downstream of GAPDH upon early hypoxia, however ATP levels and viability are only affected with additional lactate dehydrogenase (LDH) impairment. Finally, the authors demonstrate that GOT1 KO cells have higher alpha-ketoglutarate (aKG) levels during early hypoxia, which could contribute to higher prolyl-hydroxylation and subsequent degradation of HIF, regulating the transcriptional response mediated by transcription factor.

* Major comments *

1. The authors claim that they were unable to supplement cells with aspartate (Figure S3), (even though an increase of aspartate is instead observed in cells treated with sodium aspartate) and had to resort to the GOT1 knock-out model to "prevent aspartate from decreasing in hypoxia". This approach implicitly assumes that Got1 is the main driver of aspartate depletion upon hypoxia. However, although steady-state levels of aspartate are indeed higher in these cells, there is still a strong decrease upon hypoxia, which the authors acknowledge but merely ascribe to "attenuated production from glutamine". This seems an insufficient explanation, considering the very fast depletion upon hypoxia originally observed. The authors should provide further information regarding why aspartate is depleted in these conditions and consider other aspartate-consuming enzymes such as GOT2, ASNS, or even nucleotide biosynthesis and urea cycle enzymes. These observations could be made using the labeling experiments already acquired. In addition, to corroborate their hypothesis, the authors could supplement 13 C-aspartate at a supraphysiological concentration (i.e. 5-10 mM) to determine to what extent it is consumed by GOT1 or other pathways. > We thank the reviewer for this comment that helped us to recognise, in retrospect, that by focusing on GOT1ko as a means to rescue aspartate levels detracted from our main finding and extensive mechanistic insights into the role of GOT1 in sustaining the increase in glycolysis in early hypoxia. As we detail in our response to the Reviewer’s point 2, we have now re-written our results section to better clarify why we focused on GOT1 (lines 175-223 of the revised manuscript – please note that line numbering corresponds to the word document with the track changes off). However, we also agree that, because the motivation that led us to GOT1 was the counter-correlation between aspartate and lactate, expanding on the pathways that determine aspartate levels in hypoxia would be useful to the reader.

2. To address the reviewer’s point, in revised Fig. S3E, we present new data where we incubated cells in normoxia or hypoxia for 3h in the presence of 1.5 mM 13C-aspartate. We opted for an intermediate aspartate concentration which was enough to observe intracellular labelling while minimising significant perturbation to cells. We found that the amount of labelled aspartate that accumulates intracellularly is not significantly different between normoxia and hypoxia. At the same time, we observe a vast depletion of unlabelled aspartate. We accept that aspartate labelling may not have reached isotopic steady state within the 3h time point we are confined to for our experiments. However, if increased consumption contributed significantly to aspartate depletion within this timeframe, the amount of labelled aspartate that accumulated would be lower in hypoxia compared to normoxia. Therefore, the data in Fig. S3E indicate that, at least within the timeframe of our experiments, the magnitude of aspartate consumption is not likely to increase to such an extent that could significantly contribute to the depletion in aspartate.

We had, indeed considered other aspartate-consuming pathways, however, in light of the above results and our subsequent finding that GOT1 is needed for increased glycolysis, we did not pursue these investigations any further and focused on the role of GOT1 instead.

• In revised Figure S3, and also in response to one of the Reviewer’s other comments below, we have now replotted the data from the experiment in the original manuscript to show both absolute and fractional isotopologue abundances of TCA intermediates from cells labelled with 13C-glucose or 13C-glutamine. Based on these re-plotted data, we find that the amounts of labelled intermediates from both labels decreases; the apparent decrease from glutamine appears greater than that from glucose, likely because glutamine labels more rapidly a greater fraction of TCA intermediates. Moreover, glutamate fractional labelling from glutamine decreases, but modestly increases from glucose over time in hypoxia compared to normoxia. These data raise the possibility that TCA intermediates are diverted to glutamate synthesis. However, as we point out in the revised text, the fact that only glutamine has reached an isotopic steady state by the end of the time course precludes us from making a more accurate quantitative statement and therefore we have refrained from further elaborating on these observations.

Taking the above observations together, in the revised text we do not dismiss increased consumption as a factor in decreased aspartate levels and rather state that “within the timeframe tested, decreased production is a significant contributor to the low aspartate levels in early hypoxia.” (lines 187-188).

In line with the previous comment, the conclusion that "GOT1 activity, rather than a decrease in aspartate concentration itself, is required to sustain the increase in glycolysis in early hypoxia." seems questionable, especially considering the failed aspartate supplementation. The authors suspect low expression of plasma membrane aspartate transporters as the reason and quote Garcia-Bermudez et al.2018 (PMID: 29941933). This paper contains ranked SLC1A2 mRNA expression data from the Cancer Cell Line Encyclopedia (CCLE). The authors may apply aspartate supplementation and "early hypoxia" to a cancer cell line expressing SLC1A2 or other aspartate transporters. Alternatively, they could try introducing the transporter by overexpression.

> We concede that the way we phrased this statement was not ideal and has rightly led to the reviewer’s criticism. In particular, referring to a “decrease in aspartate concentration”, could mislead the reader into thinking that we were referring to the process of aspartate consumption, rather than the low aspartate levels themselves, which is what we aimed to explore. In the revised text, we now carefully make this distinction; we show new data (Figure S3G) supporting the idea that low aspartate levels are not necessary for increased lactate; we explain that, given the known role of the malate-aspartate shuttle in coordinating redox balance and potentially affecting glycolytic flux, the fact that aspartate didn’t appear to be limiting was surprising and we therefore asked whether GOT1, which depends on aspartate, had a role in the increased glycolysis in early hypoxia. Given that GOT1ko attenuated the increase in glycolysis we subsequently focused on the mechanism underlying this observation. In more detail:

As Reviewer 2 noted in point 1 of their review, the increase in lactate became more apparent after 2 h, when aspartate levels had almost reached their minimum. This successive timing of abundance changes raised the possibility that low aspartate levels precede, and possibly drive, the increased lactate. Therefore, we sought to test whether this was the case by preventing depletion of aspartate in hypoxia with exogenous aspartate. We agree that, to address the comment of Reviewer 1 here, overexpression of an aspartate transporter would have been a good way to overcome poor aspartate uptake by MCF7 cells, however, at the time we initiated this study, SLC1A2 was not known as an aspartate transporter. We, therefore, cultured MCF7 cells for several weeks in media containing 0.5 mM aspartate (which is normally absent in our standard media formulation) because we expected that cells would adapt to take up more aspartate. We, thereby, obtained a derivative cell line that we called MCF7Asp. In new Figure S3G, we show that addition of 0.5 mM aspartate in the media of MCF7Asp cells largely prevented the decrease in intracellular aspartate seen in parental MCF7 cells after 3h in 1% O2; however, the increase in lactate was similar between MCF7 and MCF7Asp cells. These data are consistent with the idea that the low aspartate levels in hypoxia are not the likely cause for the increase in lactate.

As the Reviewer notes in point 3 below, production of malate m+1 from 2H-glucose does not decrease below the levels found in normoxia (Fig. 4H), even though aspartate levels are depleted (Fig. 1C). Together with the fact that maintaining aspartate levels to near-normoxic levels does not further boost lactate levels (Figure S3G), these findings speak against the notion that the lack of increased GOT1-MDH1 flux is due to insufficient aspartate and are aligned with the idea that the malate-aspartate shuttle is saturated (PMID: 35973426, 21982705).

• The observation that labelled m+1 malate produced from [4-2H]-glucose is similar in normoxia and hypoxia (Figure 4G), does not support the notion that GOT1-MDH axis is increased at low oxygen and seems to suggest that the depletion of aspartate observed in early hypoxia is unrelated to this axis. The authors should resolve this discrepancy.*

> In our manuscript, we do not claim that the flux through the GOT1-MDH1 axis is increased but, instead, we emphasise the fact that, as the reviewer observed, malate labelling from 2H-glucose is unchanged (e.g. see text in our original manuscript - lines 519-522 of the revised version: “Importantly, a model where increased upper glycolysis due to the Pasteur effect overwhelms GAPDH capacity also elucidates the apparent increase in the reliance of glycolysis on GOT1-MDH1 in hypoxia, even though flux through this pathway is not elevated.”). As we also detail in our responses to comments 1 and 2, above, in the revised manuscript, we have re-written the discussion to better explain that the reliance on GOT1 in hypoxia is not driven by increased flux through this pathway (which is likely saturated as outline in our response to point 2, above), but rather from the increased demand imposed by the elevation in incoming glucose carbons due to the Pasteur effect (lines 504-531). This is akin to a situation where increased demand for a product drives its price up if the manufacturer does not boost production to increase supply. We hope that the reviewed discussion makes this clearer and addresses the reviewer’s comment.

• The alpha-KG level regulation by Got1 and the subsequent HIF1alpha "priming" seem quite promising and likely the most novel part of the manuscript. However, further proof should be added to support this strong claim. First, aKG to succinate ratio, rather than aKG alone, is a better indicator of aKG-dependent dioxygenases activity. So. the authors should provide this measurement. *

In line with the reviewer’s excellent suggestion, in the revised manuscript, we added new panel in Figure 6F (discussed in lines 457-458) that shows αKG levels alongside the corresponding αKG/succinate ratios. These data agree with our original interpretation that cofactor levels in GOT1ko cells favour increased dioxygenase activity.

*Second, the authors should rule out the possibility that the differential hydroxylation of HIF is due to the redistribution of intracellular oxygen due to alterations in mitochondrial function. To do this, they could determine whether cytosolic oxygen levels differ in the two conditions. *

The reviewer raises the interesting hypothesis that, given the decreased respiration in hypoxic GOT1ko cells, one could expect increased availability of oxygen that could contribute to the destabilisation of HIF1α. To the best of our knowledge, measuring absolute cytosolic O2 concentration, particularly in hypoxia, would require specialised equipment [e.g. phosphorescence lifetime imaging (PMID: 26065366), or phosphorescence quenching oxymetry (PMID: 21912692); unfortunately, we do not have access to such equipment. In the revised manuscript, we acknowledge the reviewer’s point with added new text in the discussion (lines 576-577).

Finally, the authors could test whether α-ketoglutarate or 2-hydroxyglutarate supplementation affects HIF stability in their experimental conditions.

> We thank the reviewer for this suggestion. In the revised manuscript (new Figure S6H and lines 453-455) we show that addition of DM-αKG, a cell-permeable form of αKG, to the media of MCF7 cells incubated at 1% O2, decreases HIF1α protein levels in a dose-dependent manner and, at the highest dose, to a degree comparable to that of GOT1ko cells.

Minor comments:

- The glycerol-3-phosphate shuttle is another means of re-oxidizing NADH and α-GP is indeed higher in GOT1 KO. According to this, in Fig 5C a clear increase in a-GP is observed in LDH KO cells. Would the phenotype be stronger upon additional GPD1 knockout or inhibition?

> The main phenotype of combined LDHA/GOT1 inhibition is a deficit in ATP and decreased cell survival. While increased flux through GPD1 could, indeed, provide more NAD+, this would come at the expense of glucose carbons that would otherwise need to flow into lower glycolysis to produce ATP. Consistent with this idea, our data show that, even if GPD1 or other dehydrogenases reoxidise NADH, as would be the case in both the LDHAko and GOT1ko cells where α-GP is elevated, they are not sufficient to compensate for the decrease in LDH and GOT1 activity. Therefore, we did not pursue this hypothesis further.

* - Aspartate and lactate levels appear unchanged in MDA-MB231 upon hypoxia. Can these changes be ascribed to a pseudohypoxic state? The authors should comment on this observation.*

> In Figure S2A, we show that MDA-MB-231 cells have increased basal levels of HIF1α compared to the almost undetectable HIF1α seen in BT474 (same figure, adjacent panel) or MCF7 cells (Figure 2A). We, therefore, agree with the reviewer’s hypothesis that the attenuated changes in aspartate or lactate levels in MDA-MB-231 cells are likely due to a pseudohypoxic state. As this is speculative, we have refrained from elaborating on this point further in the manuscript.

* - Figure S3B: The authors do not provide information on the length of hypoxia for these experiments. *> The data shown in original Figure S3B (new Fig. S3A-B) are a time course. Cells were incubated at 21% or 1% O2 with the respective isotope label for increasing lengths of time, with the longest time point shown (6h) being the longest time we incubated cells in hypoxia. If the reviewer meant another panel, the length of hypoxia would be 3h unless otherwise stated.

- Glucose and glutamine isotopic labelling should be accompanied by graphs showing the total pool levels of these metabolites, and also the uptake of glucose and glutamine (and their specific isotopologue distribution). It would be important to show the isotopologue distribution of aKG in all the conditions tested, in particular, because of its proposed regulation by Got1.

> In the revised manuscript, new Fig. S3 panels A-D, we now show absolute and fractional isotopologue distributions for TCA intermediates for both glucose and glutamine labelling. We have omitted showing αKG in this figure as we could not reliably quantify it in the glutamine-labelling experiment. Also, unfortunately, quantification of glutamine in our GC-MS datasets is not reliable due to conversion to 5-oxoproline.

- Malate generated by MDH1 can be converted by ME1 into Pyruvate, which could be further processed by LDH. Have the authors measured this conversion in their dataset.

> In the figure below we labelled cells with [U-13C]-glutamine for 3 h at 21% or 1% O2 and plotted the fractional labelling for all observable isotopologues in malate, pyruvate and lactate. These data show that there is minimal labelling in pyruvate and lactate (- Aspartate absolute levels across cell lines appear different. Is this due to differences in cell volume? Can the authors comment on this observation?

> To address the reviewer’s hypothesis, we focused on MCF7 and MDA-MB-231, the two cell lines with the highest and lowest aspartate levels, respectively. The volume of MCF7 is approx. 19% higher than that of MDA-MB-231 (calculated based on cell size data from PMID: 31015463). Based on this calculation, and bearing in mind that cell volume is a good predictor of biomass content (PMID: 18595067), cell volume differences may contribute to, but cannot fully account for the one order of magnitude difference in aspartate abundance we see between these cell lines (Figures 1C and S1A).

The cell lines we used in this manuscript (MCF7, BT474, MDA-MB-231, MCF10A) represent different breast cancer (or untransformed, in the case of MCF10A) cell types, with different oncogenic mutation content (PMID: 17157791, 22460905) and proliferation rates (PMID: 22628656); all these factors can be related to steady-state cellular metabolite levels (PMID: 31015463). In the figure below, we have plotted aspartate abundance data (from PMID: 31068703) in 928 cell lines of various origins. These data show that aspartate levels can differ as much as 2 orders of magnitude between cancer cell lines and about half an order of magnitude between MCF7 and MDA-MB-231 or BT474 (MCF10A was not present in this dataset); they also show that aspartate levels in the three cell lines rank in the same order as in our manuscript (MCF7>BT474>MDA-MB-231), although, it is unclear if cells in this dataset were also cultured in dialysed serum as in ours, so we cannot confidently compare the absolute aspartate measurements between our studies.

In conclusion, we suspect that cell volume differences together with other factors, such as proliferation rates and metabolic network differences may account for the differences in intracellular aspartate levels.

- Under hypoxia the contribution of glutamine (labelled fraction, Fig. S3) to TCA cycle intermediates decreases. However, this is not paralleled by an increase in the contribution of glucose, as also supported by an increase in the m+0 in the glutamine labeling but not in the glucose one. How do the authors explain this apparent inconsistency? Are there sources of unlabelled TCA cycle during the hypoxic experiment?

> While glucose and glutamine are the major carbon sources in many cultured cancer cell lines, incl. MCF7 as indicated by the data in Figure S3A-D, other nutrients (such as amino acids, other than glutamine, and fatty acids) can also provide carbons at various points of the TCA cycle. The fact that fractional labelling of glutamate from glutamine is decreased in hypoxia would suggest that the source of decreased contribution of glutamine into the TCA is unlabelled glutamate. We can exclude uptake of exogenous glutamate, because all our metabolic measurements are performed with cells incubated in media without glutamate and supplemented with dialysed serum. However, we observe a modest increase in the fractional labelling from glucose into glutamate (Figure S3A). As glucose labelling into the TCA cycle is not at steady-state even after 5h, it is hard to assess whether, increased labelling from glucose suffices to explain the dilution of glutamine-derived labelling into glutamate a quantitative conclusion but it points to efflux of intermediates out of the TCA cycle (discussed in lines 181-183 of the revised manuscript).

We thank the reviewer for their time and thoughtful comments that helped us improve the presentation of our work.

**Referees cross-commenting**

Referee 2 raises important questions that are in part aligned with referee 1 and are reasonable and doable is the time frame proposed. These are all important questions and comments to consolidate the central hypothesis of the work and I believe are required for publication.

*

Reviewer #1 (Significance (Required)):*

Overall, this is an exciting and well-executed piece of work focusing on the early biochemical consequences of hypoxia that the wide metabolism/biochemistry audience will appreciate. While most of these observations are not entirely unexpected, the work brings a sufficiently novel perspective and insights to the field and deserves publication. However, some conclusions are not fully supported by the data and some additional experiments are suggested to bring clarification and strengthen the authors' conclusions.

We are a lab expert in cancer metabolism.

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

Summary ** This manuscript represents an interesting and novel description of the role of a cytosolic transaminase, glutamic-oxaloacetate transaminase 1 (GOT1) on both cytosolic redox (and therefore glycolysis through its functional linkage with malate dehydrogenase 1) and the availability of alpha-ketoglutarate for stabilisation of HIF1a in hypoxia. Some of the most interesting data are the evidence for increased cytosolic NAD+ regeneration through the combined action of LDHA (known) and GPD1 (less well-described increase in activity in hypoxia). The manuscript as a whole describes the multiple systems required for the early response to hypoxia, but the focus of the title and way the article is written do not entirely reflect this. For example, the title focuses on GOT1 as the enzymes whose activity is responsible for the early response to hypoxia. However, this is not reflected in some of the data - the deuteron labelling in particular - which shows that LDH and GPD1 are responsible for the biggest redox activity (i.e. support of glycolysis). A degree of reframing of the article may therefore be of benefit.

We thank the reviewer for their constructive suggestions. In the revised manuscript, we have re-written the title and the relevant parts of the results section, and we have significantly re-structured the discussion section to reflect the fact that multiple enzyme systems, one of which is GOT1, converge to support the glycolytic increase and cell survival in early hypoxia. Furthermore, in our point-by-point responses, below, we highlight in detail how we have streamlined the way we present our results.

*Major comments. *

In Figure 1 C and D, the data suggest significant changes in the decrease in cellular aspartate between 1-2 hours, which then slow. This is followed by a change in lactate concentrations from 2 hours onwards, which is observed in the cells (D) and media (F). The rapid decrease in aspartate concentration suggests a relatively large change, which does not correspond to the later lack of alteration in deuteron labelling from d4-glucose (Figure 4H-J) in m+1 malate. This therefore suggests that the biggest determinant of decreased aspartate is not coupled to MDH1 activity directly. If the manuscript is focused on the relevance of GOT1 activity to the early hypoxic response, this should be better resolved. Given that this could undermine the strength of the case being made for GOT1 activity playing a significant role (through MDH1), could the authors perform the same experiments but in the GOT1KO cells to show how NADH is handled under these conditions by LDHA and GPD1? If the focus of the manuscript is shifted, these experiments would likely not be necessary.

> We thank the reviewer for these comments, which, together with those by Reviewer 1, highlighted that the way we presented our results warranted improvement. First, we would like to clarify that by referring to a “decrease in aspartate concentration”, we may have misled the reader into thinking that we were referring to the process of aspartate consumption; rather we wanted to explore whether the low aspartate level itself could be causing the increase in lactate. This is because, as the Reviewer points out, the rate of lactate accumulation picked up after aspartate had almost reached its minimum. Furthermore, by not elaborating on the cause of decreased aspartate and by focusing on GOT1ko as a means to rescue aspartate levels implied a hypothesis whereby GOT1 was the main aspartate consumer, thereby detracting from our main finding and extensive mechanistic insights into the role of GOT1 in sustaining the increase in glycolysis in early hypoxia (regardless the contribution of GOT1 activity in the observed depletion of aspartate).

In the revised text, we have re-written parts of the results section to better clarify these points (e.g. lines 175-223 - please note that line numbering corresponds to the word document with the track changes off). In summary, and as detailed below, we explore the glucose and glutamine data further and present new data with 13C-Asp, which, together support the idea that decreased aspartate in early hypoxia is largely attributable to decreased synthesis and, to a lesser extent, if at all, to increased degradation. We then explain that, given the known role of the malate-aspartate shuttle in coordinating redox balance and potentially affecting glycolytic flux, we asked whether GOT1, which depends on aspartate, still had a role in the increased glycolysis vis-à-vis the low aspartate levels in early hypoxia. Given that GOT1ko did attenuate the increase in glycolysis we subsequently focused on the mechanism underlying this observation. We have re-structured the discussion, to highlight that GOT1 is one the multiple systems required for survival in early hypoxia. We also explain that the reliance on GOT1 in hypoxia is not driven by increased flux through the GOT1-MDH1 axis (which is likely saturated), but rather from the increased demand imposed by the elevation in incoming glucose carbons due to the Pasteur effect (lines 504-531). A relatable situation is when increased demand for a product drives its price up if the manufacturer does not boost production to increase supply. We hope that the revised text better clarifies these points.

Below, we detail the new experimental evidence/analyses we referred to above:

• In revised Figure S3A-D, we have now replotted the data from the experiments in the original manuscript to show both absolute and fractional isotopologue abundances of TCA intermediates from cells labelled with 13C-glucose or 13C-glutamine. Based on these re-plotted data, we find that the amounts of labelled intermediates from both labels decreases; the apparent decrease from glutamine appears greater than that from glucose, likely because glutamine labels more rapidly a greater fraction of TCA intermediates. Moreover, glutamate fractional labelling from glutamine decreases, but modestly increases from glucose over time in hypoxia compared to normoxia. These data raise the possibility that TCA intermediates are diverted to glutamate synthesis. However, as we point out in the revised text, the fact that only glutamine has reached an isotopic steady state by 5h precludes us from making a more accurate quantitative statement and therefore we have refrained from further elaborating on these observations.

• In revised Fig. S3E, we present new data where we incubated cells in normoxia or hypoxia for 3h in the presence of 1.5 mM 13C-aspartate. We found that the amount of labelled aspartate that accumulates intracellularly is not significantly different between normoxia and hypoxia. At the same time, we observe a vast depletion of unlabelled aspartate. We accept that aspartate labelling may not have reached isotopic steady state within the 3h time point we are confined to for our experiments. However, if increased consumption contributed significantly to aspartate depletion within this timeframe, the amount of labelled aspartate that accumulated would be lower in hypoxia compared to normoxia. Therefore, the data in Fig. S3E indicate that, at least within the timeframe of our experiments, the magnitude of aspartate consumption is not likely to increase to such an extent that could significantly contribute to the depletion in aspartate.

Together with the data in Fig. S3A-D, these findings suggest that decreased aspartate in early hypoxia is to a great degree driven by decreased production.

• The authors present data in Figure 1 and 3 using 2DG as a surrogate for glucose uptake. 2DG has been previously shown not to always be a surrogate for glucose uptake (Sinclair et al. Immunometabolism 2020). Given that this paper highlighted warns in particular about assuming SLC2A1 and SLC2A3 activities based on 2DG uptake, and that these two transporters are the major glucose transporters regulated by hypoxia, a cautious approach to these data is recommended. Assuming that 2DG uptake is a surrogate for glucose in this system (panel C), the effect of GOT1 appears to be at the level of glucose uptake even at 3 hours - it has been marked as being significant by the authors. This suggests that loss of GOT1 has an effect on glucose uptake prior to any transcriptional response is observed. Is the plasma membrane occupancy by the SLC2A1 or SLC2A3 been reduced after GOT1 KO? The same is true for Figure 1 - as intracellular aspartate and lactate and extracellular lactate is shown, could change in extracellular glucose not be presented as a direct measure?*

The reviewer raises two points: (a) that using 2DG may not faithfully report transporter-mediated glucose uptake and (b) that, if our observations with 2DG are valid, they could point to the possibility that attenuation of glycolysis in GOT1ko cells may be attributable to effects in glucose uptake. In brief, we cannot use glucose measurements in media as an indicator of glucose uptake rates because we do not observe measurable glucose depletion from media within the relevant timeframe (3h) of our experiments.

(a) Given that we did not have access to a set up for using radionuclides, we explored both 2DG-based and glucose depletion from media as potential means to assess glucose uptake. We found that, over 24h, MCF7 cells deplete glucose faster than cells incubated in normoxia for the same amount of time (figure below, A). The magnitude of this increase is similar to that we report using 2-DG (~3-fold, Fig. 1E and 3C). However, we observed only minimal depletion of glucose in the first 3-5 h of culturing cells with fresh media (figure below, B). This is perhaps not surprising given that studies that look at metabolite exchange rates (incl. glucose) typically sample over a period of one to several days rather than hours (e.g. PMID: 31015463, 22628656). In conclusion, we reasoned that detecting a positive change in signal (intracellular 2DG) would provide a more sensitive means than a decrease in extracellular glucose to enable assessment of glucose use within the early time-points that our manuscript is mainly concerned with.

(b) Indeed, we were initially intrigued by the decrease in glucose uptake by GOT1ko cells as it could explain decreased lactate production. However, the upregulation of upstream glycolytic intermediates in GOT1ko cells in both normoxia and hypoxia (Figure 4A) together with the evidence of increased α-GP production from glucose (Figure 4K-L) suggested that, even if less glucose is taken up by GOT1ko cells, there is still a bottleneck at the GAPDH step that prevents maximal flow of glycolytic intermediates to lower glycolysis. We therefore did not pursue further the cause of decreased glucose uptake by GOT1ko cells at this stage.

• The data shown in Figure 2D suggests that there is little change in overall contribution to citrate from glucose in hypoxia compared to normoxia, and that HIF1 is does not play a role in the hypoxic response at this point. However, the data presented are overall fractional labelling, and therefore do not focus on the main hypoxia-dependent point of control highlighted before this by the authors - pyruvate oxidation through PDH. Could the authors consider plotting m+2 isotopomer of citrate either alongside or instead of the total fractional label (which includes hypoxia-independent PC activity and cycling carbons). *

We agree with the reviewer’s suggestion. In the revised manuscript, we added a new panel in Fig. 2D that shows the m+2 citrate isotopologue alongside the original fractional labelling data. This new panel is shown as a bar graph to enable the presentation of individual datapoints and statistical test results.

Additionally, the experimental set-up means that average incorporation over the time shown is represented - i.e. the 3h timepoint is incorporation over the first two hours, while the 24 hour timepoint is averaged over the whole period. It is therefore likely under-representing the decrease in glucose contribution to citrate at 24 hours - the authors could point this out, or OPTIONALLY perform a more time-resolved experiment where flux over shorter periods is assessed for each of the timepoints (i.e. 0-1, 2-3, 5-6, 23-24).

Indeed, we did consider a more time-resolved labelling experiment as the reviewer suggests, however, we decided against this approach as we were concerned that even if we pre-equilibrated the labelling media in hypoxia, it would be challenging to avoid perturbations associated with handling of the cells during addition of the isotopically labelled compound. The new panel in Fig. 2D that shows absolute citrate m+2 abundances should address this point, however, in the revised text (lines 162-164) we added new text that points out this issue.

• Figure 3 data are key for the GOT1 theme of the manuscript, as the authors show that loss of GOT1 increases cellular aspartate in both normoxia and hypoxia - suggesting that GOT1 is an aspartate-consuming enzyme in both conditions. Indeed the magnitude of the change in aspartate after GOT1 knockdown appears similar in both conditions (Panel B). These are interesting data, as they contrast with a recently published study (Altea-Manzano et al. Molecular Cell 2022) suggesting that in respiration-deficient cells (a condition with parallels with hypoxia), GOT1 activity may be aspartate producing to supply aspartate to the mitochondria for GOT2. It would be important for the authors to discuss the differences between studies.*

Following the reviewer’s suggestion, in the revised manuscript (lines 547-556), we have now expanded our previous discussion on the functions of GOT1 in cells with respiration defects.

• Panel E shows data at 5 hours, while the rest of the panels here are a mix of 1 and 3h timepoints. Equally panel E also presents concentration, while D presents relative abundance of lactate - could a consistent approach to presenting the results be taken?*

We agree. Taking into consideration that the data in this panel show one time point of the full time-course in Figure S3F, and to streamline the presentation of these data, in the revised manuscript, we have moved the time-course graph to the main figure.

• In Figure S3, the authors show the lack of direct aspartate uptake, or supplementation through the use of an esterified form. OPTIONAL: they could consider using the expression of SLC1A3 (Tajan et al. Cell Metabolism 2018; Hart et al eLife 2023) to increase aspartate uptake in order to test their hypothesis. *

We agree that, to address this point, overexpression of an aspartate transporter would have been a good way to overcome poor aspartate uptake by MCF7 cells, however, at the time we initiated this study, SLC1A2 was not known as an aspartate transporter. We, therefore, cultured MCF7 cells for several weeks in media containing 0.5 mM aspartate (which is normally absent in our standard media formulation) because we expected that cells would adapt to take up more aspartate. We, thereby, obtained a derivative cell line that we called MCF7Asp. In new Figure S3G, we show that addition of 0.5 mM aspartate in the media of MCF7Asp cells largely prevented the decrease in intracellular aspartate seen in parental MCF7 cells after 3h in 1% O2. However, the increase in lactate was similar between MCF7 and MCF7Asp cells. These data are consistent with the idea that the low aspartate levels in hypoxia are not the likely cause for the increase in lactate.

*Figure S3B-E - the authors suggest based on these data that aspartate decrease in hypoxia is through decreased glutamine contribution. Indeed they could also interrogate the data further, as the defect is observed in glutamate, perhaps suggesting that glutamine metabolism through glutaminase is altered. *

To address the Reviewer’s point, in revised Figure S3, we have now replotted the data from the experiment in the original manuscript to show both absolute and fractional isotopologue abundances of TCA intermediates from cells labelled with 13C-glucose or 13C-glutamine. We have elaborated on these results in our response to point 1, and we re-iterate our conclusions here for the Reviewer’s convenience: Based on these re-plotted data, we find that the amounts of labelled intermediates from both labels decreases; the apparent decrease from glutamine appears greater than that from glucose, likely because glutamine labels more rapidly a greater fraction of TCA intermediates. Moreover, glutamate fractional labelling from glutamine decreases but modestly increases from glucose over time in hypoxia compared to normoxia. These data raise the possibility that TCA intermediates are diverted to glutamate synthesis. However, as we point out in the revised text, the fact that only glutamine has reached an isotopic steady state by 5h precludes us from making a more accurate quantitative statement and therefore we have refrained from further elaborating on these observations.

*Figure S3D and E - the authors show data from 3 hours of labelling, which is not at steady-state (observable from the timecourse also shown in B and C). To be able to compare the glucose and glutamine labelling, a timepoint in which (pseudo)steady-state is achieved would be better chose. *

In the revised manuscript, this concern is now addressed by showing both absolute and relative isotopologue abundances for all available time points. We agree that quantitative comparison of labelling must be done at steady-state conditions, however, as we also point out in the revised text (lines 180-181), only glutamine reaches isotopic steady state by 5h whereas glucose hasn’t.

Additionally, within the aspartate isotopomers arising from glutamine, there is an odd m+1 for aspartate not observed in the other proximal metabolites. Is this a technical defect or is there a biological reason for the significant fractional amount in normoxia?

We thank the reviewer for pointing this irregularity, which we should have clearly identified as such during proofreading of the manuscript. Probed by the reviewer’s comment, we reviewed the corresponding data tables used to plot these data and found that M+1 had exactly the same values as M+0. We then inspected the original data and confirmed that this resulted from an error during the copying of the data from the R-script output data table to GraphPad Prism for plotting (the line containing the replicates for the m+0 isotopologue was pasted again in the line of the M+1 isotopologues). This issue is now obsolete, as, in the revised manuscript Fig S3 new panels A-D, we have replaced the fractional data with detailed absolute and fractional labelling showing all isotopologues. We apologise for this error.

• Figure S6F - all samples from GOT1 KO cells have less actin - could an appropriately loaded western blot be presented?*

In the revised manuscript, we added a new panel with the Ponceau (27/02/2018) staining of the same membrane used for immunoblotting. This staining shows equal loading between all lanes. It is unclear why despite equal loading, the actin signal differs between the two lines.

• In Figure S5B, the authors present ATP data in wild-type control cells, and LDHA-KO with LDHA re-expression. These should be phenotypically similar, but clearly are not. It suggests that there is something not correct with the system being used.*

As shown in the western blot of this figure, expression of exogenous LDH only reaches a fraction of endogenous levels, which likely explains the partial, albeit significant, rescue of the ATP depletion observed in the LDHAko cells. We have not been able to achieve higher LDH expression in our cell preparations that would enable us to address this point further.

*

*

Minor comments

1. • PHDs need iron, alpha-ketoglutarate, oxygen and critically ascorbate (Introduction page 2)*

     We thank the reviewer for highlighting this critical omission. In the revised manuscript, we have now added this information (line 58).

* PDK1 phosphorylation of PDH leads to a reduction in pyruvate oxidation, rather than entry of glucose carbons to the TCA cycle (Introduction page 3)*

We agree with the reviewer that our wording was not accurate, and, in the revised text, we have re-written this part (lines 72-74): “…[PDK1] catalyses the inhibitory phosphorylation of pyruvate dehydrogenase (PDH), leading to attenuated pyruvate oxidation and, consequently, decreased contribution of glucose-derived carbons into the tricarboxylic acid (TCA) cycle.”

* SLC25A51 has been identified as being required for NAD transport into the mitochondria (Kori et al. Science Advances 2020), so it is incorrect to say that the inner mitochondrial membrane is impermeable to this metabolite (page 7)*

We agree that, in light of the Kori et al. study, the phrasing in our text presented an outdated view of pyridine nucleotide compartmentalisation. The data in Kory et al. support SLC25A51 as a mitochondrial NAD+ transporter, however, it is not clear if NADH is also a substrate. Furthermore, as the authors also point out, SLC25A51 has a relatively low affinity for NAD+ and therefore unlikely to interfere with the functions of the malate-aspartate shuttle. Taking all this into consideration, in the revised text (line 249), we acknowledge the existence of a low-affinity mitochondrial NAD+ transporter and retained the statement about impermeability specifically for NADH.

* Figure S6D - authors shows a highly significant increase in the mRNA for EGLN3, which is a HIF1 target gene, as well as encoding PHD3, which acts to hydroxylate HIF1a alongside PHD2. This should be commented on in the text.*

In the revised discussion (lines 577-578), we acknowledge that increased PHD3 (together with increased oxygen availability, related to Reviewer 1’s comment), may additionally contribute to HIF1α destabilisation. Please note that we have also added new data (Figure S6H) in response to Reviewer 1, where we show that exogenous αKG causes HIF1α destabilisation in hypoxia, further supporting the notion that boosting intracellular αKG, alone, can destabilise HIF1α.

* Figure S5G - could it be made clear on the graph whether this is at 21% or 1% O2?*

We thank the reviewer for pointing out this omission. We now state clearly both in the revised corresponding legend (line 937) and revised figure that these data are at 1% O2.

• Figure 5I shows ATP level against % labelling of alpha-GP. It isn't clear whether this is abundance or fractional label, but if the latter this it potentially misleading, as if the concentration of alpha-GP increases as fractional label decreases, there is effectively no change. Could the authors extract the steady-state data from the analysis and use this to calculate amount of m+3 label instead of fraction? Similarly for Figure S1H showing fractional labelling of lactate from glucose. It is likely that the title of this graph is a typo, and that m+3 instead was meant. Additionally, measurement of fractional labelling does not demonstrate increased concentrations of the metabolite, but the glucose carbons making up this isotopomer in the pool.*

For Figure 5I, we confirm that what we show is based on abundance of α-GP m+3 labelling from glucose and, in the revised manuscript (line 895), we amended the legend to clarify this important point.

We concede that the way we had originally written this sentence, suggested that we derived our conclusion that increased lactate in media was due to increased glycolysis based solely on the fractional data in Fig. S1H. In the revised manuscript, we have re-phrased the relevant sentence (lines 136-137) to indicate that our conclusion is based on the fractional data, together with the total lactate data that we show in Fig. 1F.

For all our GC-MS experiments we used ions that we detected reliably in all our experiments – in the case of lactate this is m/z 117. This is a 2-carbon fragment as indicated in the original legend; the molecular formula of the derivatised fragment is shown in Table S2. In the revised manuscript (line 671) we clarify that this fragment contains carbons 2 and 3 of lactate (which we concluded from experiments where labelling with 3,4-13C-glucose (which labels lactate at C1) led to partial decrease in this isotopologue); therefore changes in 117 m+2 indicate changes in glycolysis rather glycolysis and the PPP.

* Figure S2G - the purpose of the measurement of cysteine is unclear; measurement of NAC directly within cells would be a clearer demonstration of its uptake, and to demonstrate direct contribution to antioxidant response would instead require measurement of cellular antioxidants rather than cysteine itself.*

We agree with the reviewer’s comment that, ideally, we would have measured antioxidants, however, unfortunately our GC-MS experiments do not detect glutathione; we, therefore, opted to show cysteine as the best available proof that NAC was added to these cells from the same experiments where we measured aspartate and lactate.

* There is no Figure S3F (page 6 of text)*

In the original version of our manuscript we had awkwardly placed Figure S3F at the top right side of the figure due to space limitations, so, understandably, the reviewer may have missed it. In the revised manuscript, we have now moved this panel to the main Figure 3E, to also address the reviewer’s point 5, above (presentation of lactate data).

* Figure 2E, lactate excretion into the media is presenting an odd profile, suggesting that between 3 and 6 hour there is uptake by cells. Equally, the 24 hour timepoint is being presented as p

The overlap of the error bars arises from error propagation as we report the values at each time point relative to t=0h. The statistical difference we reported was calculated on the original values at 24 h alone, so to avoid this discrepancy we have opted for removing the results of this statistical test altogether.

*

Reviewer #2 (Significance (Required)):*

* The data throughout this paper provide some strong evidence for an early and likely HIF-independent metabolic response - while this is understood, detailed studies have not been performed into the various redox balancing cytosolic pathways, which are presented here. The focus on GOT1 is also interesting and novel, but represents part of a larger overall picture presented, which is not reflected in the title.*

* This is suitable for a relatively broad audience, as the phenotype is likely not cancer specific.

*__Reviewer #3 (Evidence, reproducibility and clarity (Required)): __

* Here, Grimm and colleagues investigate the immediate cellular response to hypoxia, prior to onset of HIF1a stabilization/activity. Consistent with established findings they describe that glycolysis is rapidly upregulated under hypoxia, in a HIF1 alpha independent manner, this correlates with an decreased aspartate levels. From this basis, they describe a key role for GOT1 activity in regulating the early hypoxic response, demonstrating its requirement for glycolysis, maintaining the NAD/NADH balance and - in combination with LDHA - maintaining ATP homeostasis in hypoxia. Finally they describe a role for GOT1 (though alpha KG depletion) in contributing to HIF1 alpha stabilization.*

* In sum, the authors present a compelling study investigating the mechanistic basis of early response to hypoxia, placing GOT1 as a key metabolic regulator of this response. The question of how cell metabolically adapt in the short term to hypoxia is, in my view, an often overlooked area of investigation but clearly has importance across biology, not least in cancer biology - thus the area of investigation is topical. The authors conclusions are supported by their data, often in multiple cell lines and/or through orthologous methods. I would support publication of this study as is.*

* Reviewer #3 (Significance (Required)):*

* Significance is stated in my review above, an understudied area of investigation (early hypoxic responses) but clearly important since without a transient response, the long-term impact of HIF1 stress responses would not be possible*

We thank the reviewer for their time assessing our manuscript and for their positive feedback.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

Here, Grimm and colleagues investigate the immediate cellular response to hypoxia, prior to onset of HIF1 stabilization/activity. Consistent with established findings they describe that glycolysis is rapidly upregulated under hypoxia, in a HIF1 alpha independent manner, this correlates with an decreased aspartate levels. From this basis, they describe a key role for GOT1 activity in regulating the early hypoxic response, demonstrating its requirement for glycolysis, maintaining the NAD/NADH balance and - in combination with LDHA - maintaining ATP homeostasis in hypoxia. Finally they describe a role for GOT1 (though alpha KG depletion) in contributing to HIF1 alpha stabilization.

In sum, the authors present a compelling study investigating the mechanistic basis of early response to hypoxia, placing GOT1 as a key metabolic regulator of this response. The question of how cell metabolically adapt in the short term to hypoxia is, in my view, an often overlooked area of investigation but clearly has importance across biology, not least in cancer biology - thus the area of investigation is topical. The authors conclusions are supported by their data, often in multiple cell lines and/or through orthologous methods. I would support publication of this study as is.

Significance

Significance is stated in my review above, an understudied area of investigation (early hypoxic responses) but clearly important since without a transient response, the long-term impact of HIF1 stress responses would not be possible

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Summary

This manuscript represents an interesting and novel description of the role of a cytosolic transaminase, glutamic-oxaloacetate transaminase 1 (GOT1) on both cytosolic redox (and therefore glycolysis through its functional linkage with malate dehydrogenase 1) and the availability of alpha-ketoglutarate for stabilisation of HIF1a in hypoxia. Some of the most interesting data are the evidence for increased cytosolic NAD+ regeneration through the combined action of LDHA (known) and GPD1 (less well-described increase in activity in hypoxia). The manuscript as a whole describes the multiple systems required for the early response to hypoxia, but the focus of the title and way the article is written do not entirely reflect this. For example, the title focuses on GOT1 as the enzymes whose activity is responsible for the early response to hypoxia. However, this is not reflected in some of the data - the deuteron labelling in particular - which shows that LDH and GPD1 are responsible for the biggest redox activity (i.e. support of glycolysis). A degree of reframing of the article may therefore be of benefit.

Major comments

1. In Figure 1 C and D, the data suggest significant changes in the decrease in cellular aspartate between 1-2 hours, which then slow. This is followed by a change in lactate concentrations from 2 hours onwards, which is observed in the cells (D) and media (F). The rapid decrease in aspartate concentration suggests a relatively large change, which does not correspond to the later lack of alteration in deuteron labelling from d4-glucose (Figure 4H-J) in m+1 malate. This therefore suggests that the biggest determinant of decreased aspartate is not coupled to MDH1 activity directly. If the manuscript is focused on the relevance of GOT1 activity to the early hypoxic response, this should be better resolved. Given that this could undermine the strength of the case being made for GOT1 activity playing a significant role (through MDH1), could the authors perform the same experiments but in the GOT1KO cells to show how NADH is handled under these conditions by LDHA and GPD1? If the focus of the manuscript is shifted, these experiments would likely not be necessary.
2. The authors present data in Figure 1 and 3 using 2DG as a surrogate for glucose uptake. 2DG has been previously shown not to always be a surrogate for glucose uptake (Sinclair et al. Immunometabolism 2020). Given that this paper highlighted warns in particular about assuming SLC2A1 and SLC2A3 activities based on 2DG uptake, and that these two transporters are the major glucose transporters regulated by hypoxia, a cautious approach to these data is recommended. Assuming that 2DG uptake is a surrogate for glucose in this system (panel C), the effect of GOT1 appears to be at the level of glucose uptake even at 3 hours - it has been marked as being significant by the authors. This suggests that loss of GOT1 has an effect on glucose uptake prior to any transcriptional response is observed. Is the plasma membrane occupancy by the SLC2A1 or SLC2A3 been reduced after GOT1 KO? The same is true for Figure 1 - as intracellular aspartate and lactate and extracellular lactate is shown, could change in extracellular glucose not be presented as a direct measure?
3. The data shown in Figure 2D suggests that there is little change in overall contribution to citrate from glucose in hypoxia compared to normoxia, and that HIF1 is does not play a role in the hypoxic response at this point. However, the data presented are overall fractional labelling, and therefore do not focus on the main hypoxia-dependent point of control highlighted before this by the authors - pyruvate oxidation through PDH. Could the authors consider plotting m+2 isotopomer of citrate either alongside or instead of the total fractional label (which includes hypoxia-independent PC activity and cycling carbons). Additionally, the experimental set-up means that average incorporation over the time shown is represented - i.e. the 3h timepoint is incorporation over the first two hours, while the 24 hour timepoint is averaged over the whole period. It is therefore likely under-representing the decrease in glucose contribution to citrate at 24 hours - the authors could point this out, or OPTIONALLY perform a more time-resolved experiment where flux over shorter periods is assessed for each of the timepoints (i.e. 0-1, 2-3, 5-6, 23-24).
4. Figure 3 data are key for the GOT1 theme of the manuscript, as the authors show that loss of GOT1 increases cellular aspartate in both normoxia and hypoxia - suggesting that GOT1 is an aspartate-consuming enzyme in both conditions. Indeed the magnitude of the change in aspartate after GOT1 knockdown appears similar in both conditions (Panel B). These are interesting data, as they contrast with a recently published study (Altea-Manzano et al. Molecular Cell 2022) suggesting that in respiration-deficient cells (a condition with parallels with hypoxia), GOT1 activity may be aspartate producing to supply aspartate to the mitochondria for GOT2. It would be important for the authors to discuss the differences between studies.
5. Panel E shows data at 5 hours, while the rest of the panels here are a mix of 1 and 3h timepoints. Equally panel E also presents concentration, while D presents relative abundance of lactate - could a consistent approach to presenting the results be taken?
6. In Figure S3, the authors show the lack of direct aspartate uptake, or supplementation through the use of an esterified form. OPTIONAL: they could consider using the expression of SLC1A3 (Tajan et al. Cell Metabolism 2018; Hart et al eLife 2023) to increase aspartate uptake in order to test their hypothesis. Figure S3B-E - the authors suggest based on these data that aspartate decrease in hypoxia is through decreased glutamine contribution. Indeed they could also interrogate the data further, as the defect is observed in glutamate, perhaps suggesting that glutamine metabolism through glutaminase is altered. Figure S3D and E - the authors show data from 3 hours of labelling, which is not at steady-state (observable from the timecourse also shown in B and C). To be able to compare the glucose and glutamine labelling, a timepoint in which (pseudo)steady-state is achieved would be better chose. Additionally, within the aspartate isotopomers arising from glutamine, there is an odd m+1 for aspartate not observed in the other proximal metabolites. Is this a technical defect or is there a biological reason for the significant fractional amount in normoxia?
7. Figure S6F - all samples from GOT1 KO cells have less actin - could an appropriately loaded western blot be presented?
8. In Figure S5B, the authors present ATP data in wild-type control cells, and LDHA-KO with LDHA re-expression. These should be phenotypically similar, but clearly are not. It suggests that there is something not correct with the system being used.

Minor comments

1. PHDs need iron, alpha-ketoglutarate, oxygen and critically ascorbate (Introduction page 2)
2. PDK1 phosphorylation of PDH leads to a reduction in pyruvate oxidation, rather than entry of glucose carbons to the TCA cycle (Introduction page 3)
3. SLC25A51 has been identified as being required for NAD transport into the mitochondria (Kori et al. Science Advances 2020), so it is incorrect to say that the inner mitochondrial membrane is impermeable to this metabolite (page 7)
4. Figure S6D - authors shows a highly significant increase in the mRNA for EGLN3, which is a HIF1 target gene, as well as encoding PHD3, which acts to hydroxylate HIF1a alongside PHD2. This should be commented on in the text.
5. Figure S5G - could it be made clear on the graph whether this is at 21% or 1% O2?
6. Figure 5I shows ATP level against % labelling of alpha-GP. It isn't clear whether this is abundance or fractional label, but if the latter this it potentially misleading, as if the concentration of alpha-GP increases as fractional label decreases, there is effectively no change. Could the authors extract the steady-state data from the analysis and use this to calculate amount of m+3 label instead of fraction? Similarly for Figure S1H showing fractional labelling of lactate from glucose. It is likely that the title of this graph is a typo, and that m+3 instead was meant. Additionally, measurement of fractional labelling does not demonstrate increased concentrations of the metabolite, but the glucose carbons making up this isotopomer in the pool.
7. Figure S2G - the purpose of the measurement of cysteine is unclear; measurement of NAC directly within cells would be a clearer demonstration of its uptake, and to demonstrate direct contribution to antioxidant response would instead require measurement of cellular antioxidants rather than cysteine itself.
8. There is no Figure S3F (page 6 of text)
9. Figure 2E, lactate excretion into the media is presenting an odd profile, suggesting that between 3 and 6 hour there is uptake by cells. Equally, the 24 hour timepoint is being presented as p<0.01 for 4 replicates with error bars that cross the mean of one of the values. Could the authors possibly check that this is indeed the case?

Significance

The data throughout this paper provide some strong evidence for an early and likely HIF-independent metabolic response - while this is understood, detailed studies have not been performed into the various redox balancing cytosolic pathways, which are presented here. The focus on GOT1 is also interesting and novel, but represents part of a larger overall picture presented, which is not reflected in the title.

This is suitable for a relatively broad audience, as the phenotype is likely not cancer specific.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

In the paper entitled GOT1 primes the cellular response to hypoxia by supporting glycolysis and HIF1α stabilization, Grimm and co-authors investigate the metabolic adaptations of cancer cells upon acute hypoxia. By measuring metabolite levels at early time points upon hypoxia, they observe the accumulation of lactate and depletion of aspartate, along with other TCA cycle metabolites. Importantly, they demonstrate that these metabolic changes are independent of the HIF alpha-dependent transcriptional response. The authors investigate the role of aspartate during these initial phases of hypoxia. To this aim, they characterize cells devoid of glutamate oxaloacetate transaminase (GOT1), in which aspartate accumulates and can no longer be used for replenishing NAD+ via the downstream conversion of oxaloacetate to malate, via malate dehydrogenase. These cells have lower cytosolic NAD+ which affects glycolytic flux through the rate-limiting, NAD+-dependent enzyme GAPDH. GOT1 KO cells have a decrease in glucose consumption, lactate secretion and metabolite levels downstream of GAPDH upon early hypoxia, however ATP levels and viability are only affected with additional lactate dehydrogenase (LDH) impairment. Finally, the authors demonstrate that GOT1 KO cells have higher alpha-ketoglutarate (aKG) levels during early hypoxia, which could contribute to higher prolyl-hydroxylation and subsequent degradation of HIF, regulating the transcriptional response mediated by transcription factor.

Major comments

1. The authors claim that they were unable to supplement cells with aspartate (Figure S3), (even though an increase of aspartate is instead observed in cells treated with sodium aspartate) and had to resort to the GOT1 knock-out model to "prevent aspartate from decreasing in hypoxia". This approach implicitly assumes that Got1 is the main driver of aspartate depletion upon hypoxia. However, although steady-state levels of aspartate are indeed higher in these cells, there is still a strong decrease upon hypoxia, which the authors acknowledge but merely ascribe to "attenuated production from glutamine". This seems an insufficient explanation, considering the very fast depletion upon hypoxia originally observed. The authors should provide further information regarding why aspartate is depleted in these conditions and consider other aspartate-consuming enzymes such as GOT2, ASNS, or even nucleotide biosynthesis and urea cycle enzymes. These observations could be made using the labeling experiments already acquired. In addition, to corroborate their hypothesis, the authors could supplement 13 C-aspartate at a supraphysiological concentration (i.e. 5-10 mM) to determine to what extent it is consumed by GOT1 or other pathways.
2. In line with the previous comment, the conclusion that "GOT1 activity, rather than a decrease in aspartate concentration itself, is required to sustain the increase in glycolysis in early hypoxia." seems questionable, especially considering the failed aspartate supplementation. The authors suspect low expression of plasma membrane aspartate transporters as the reason and quote Garcia-Bermudez et al.2018 (PMID: 29941933). This paper contains ranked SLC1A2 mRNA expression data from the Cancer Cell Line Encyclopedia (CCLE). The authors may apply aspartate supplementation and "early hypoxia" to a cancer cell line expressing SLC1A2 or other aspartate transporters. Alternatively, they could try introducing the transporter by overexpression.
3. The observation that labelled m+1 malate produced from [4-2H]-glucose is similar in normoxia and hypoxia (Figure 4G), does not support the notion that GOT1-MDH axis is increased at low oxygen and seems to suggest that the depletion of aspartate observed in early hypoxia is unrelated to this axis. The authors should resolve this discrepancy.
4. The alpha-KG level regulation by Got1 and the subsequent HIF1alpha "priming" seem quite promising and likely the most novel part of the manuscript. However, further proof should be added to support this strong claim. First, aKG to succinate ratio, rather than aKG alone, is a better indicator of aKG-dependent dioxygenases activity. So. the authors should provide this measurement. Second, the authors should rule out the possibility that the differential hydroxylation of HIF is due to the redistribution of intracellular oxygen due to alterations in mitochondrial function. To do this, they could determine whether cytosolic oxygen levels differ in the two conditions. Finally, the authors could test whether α-ketoglutarate or 2-hydroxyglutarate supplementation affects HIF stability in their experimental conditions.

Minor comments:

• The glycerol-3-phosphate shuttle is another means of re-oxidizing NADH and α-GP is indeed higher in GOT1 KO. According to this, in Fig 5C a clear increase in a-GP is observed in LDH KO cells. Would the phenotype be stronger upon additional GPD1 knockout or inhibition?
• Aspartate and lactate levels appear unchanged in MDA-MB231 upon hypoxia. Can these changes be ascribed to a pseudohypoxic state? The authors should comment on this observation.
• Figure S3B: The authors do not provide information on the length of hypoxia for these experiments.
• Glucose and glutamine isotopic labelling should be accompanied by graphs showing the total pool levels of these metabolites, and also the uptake of glucose and glutamine (and their specific isotopologue distribution). It would be important to show the isotopologue distribution of aKG in all the conditions tested, in particular, because of its proposed regulation by Got1.
• Malate generated by MDH1 can be converted by ME1 into Pyruvate, which could be further processed by LDH. Have the authors measured this conversion in their dataset.
• Aspartate absolute levels across cell lines appear different. Is this due to differences in cell volume? Can the authors comment on this observation?
• Under hypoxia the contribution of glutamine (labelled fraction, Fig. S3) to TCA cycle intermediates decreases. However, this is not paralleled by an increase in the contribution of glucose, as also supported by an increase in the m+0 in the glutamine labeling but not in the glucose one. How do the authors explain this apparent inconsistency? Are there sources of unlabelled TCA cycle during the hypoxic experiment?

Referees cross-commenting

Referee 2 raises important questions that are in part aligned with referee 1 and are reasonable and doable is the time frame proposed. These are all important questions and comments to consolidate the central hypothesis of the work and I believe are required for publication.

Significance

Overall, this is an exciting and well-executed piece of work focusing on the early biochemical consequences of hypoxia that the wide metabolism/biochemistry audience will appreciate. While most of these observations are not entirely unexpected, the work brings a sufficiently novel perspective and insights to the field and deserves publication. However, some conclusions are not fully supported by the data and some additional experiments are suggested to bring clarification and strengthen the authors' conclusions.

We are a lab expert in cancer metabolism.

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

Learn more at Review Commons

---

Reply to the reviewers

Manuscript number: RC-2023-01935

Corresponding author(s): Vincent Mirouse

1. General Statements [optional]

First of all, we would like to thank the three reviewers for the interest they expressed in our work. Moreover, we believe that, aided by their suggestions, we managed to significantly improve our manuscript.

2. Point-by-point description of the revisions

From here, Reviewers’ comments are in black and our reply in italic dark blue

__Reviewer #1 (Evidence, reproducibility and clarity): ______

This manuscript by Dennis et al. reports a study of the polarized secretion of basement membrane Collagen IV in the Drosophila (fruit fly) follicular epithelium. Using genetic manipulations and confocal imaging, the authors show that Rab-GTPases Rab8 and Rab10, both known to be required for proper basal secretion of Collagen IV (work by the labs of Sally Horne-Badovinac and Trudi Schupbach, respectively), mediate two alternative secretion routes: Rab8 mediates basal-most secretion of soluble Collagen IV that is incorporated homogenously into the basement membrane, whereas Rab10 mediates basal-lateral secretion of Collagen IV that produces insoluble fibers. The authors additionally study the relation between Rab10 and Dystroglycan/Dystrophin (Dystrophin-associated protein complex, DAPC), which they previously showed to be essential for fibril formation (Cerqueira-Campos et al., 2020). They show here that Dystrophin and Rab10 colocalize at the basal trailing side of follicle cells and that overexpressed Dystroglycan can recruit Rab10 to the plasma membrane; however, they also show that Dystrophin mutants fail to display an effect on Rab10 localization, leaving the significance of the proposed Rab10-DAPC interaction unresolved. Finally, the authors present convincing evidence that the exocyst complex opposes fibril formation, and suggestive but comparatively weaker results pointing that this opposition is due to two independent separate exocyst roles: an inhibitory interaction exocyst-Dystrophin (Dystrophin being required for fibril formation), and a positive role in the alternative Rab8 non-fibril route.

Major comment:

• There are several instances throughout the study in which the authors seem to have problems quantifying results. This affects some assertions central to the message of the paper that are not supported by the quantifications presented. It also casts doubts on accessory points deduced from quantitative differences (or lack of difference) that do not seem fully reliable. I would urge the authors to reevaluate their quantification methods.

a) Rab8 KD does not significantly increase apical fraction of Collagen IV with respect to control (Fig. 1H). The image in 1C clearly shows that Col IV is present apically, something that has been shown by others and that never occurs in the wild type. Failure of the quantifying method to detect a difference can only mean the quantifying method is not adequate. A 10% average in the control when it's clear that no Col IV at all is found apically in the wild type suggests that the authors are quantifying background signal that they should not be acquiring, and, if acquired, they should be subtracting. Rab8/Rab10 double knock down is said to show a synergistic effect, when an additive effect would be more consistent with alternative routes. Other problematic deductions drawn from apical fraction quantifications are found in Fig. 5J (Dys- enhancing Rab8 KD but not Rab10 KD) and Fig. 7D (Exo70- enhancing Rab10 KD but not Rab8 KD).

We agree that this quantification was not optimal. We improved it by quantifying a narrower and more precise region for each domain. The new results are shown in Figure 1H. This improvement reduces the apical signal in the control from 10% to 6% and allows us to detect a significant increase between the control and Rab8 KD, thus resolving the problem raised. After verification, we did not subtract the background because there was no electronic background in our images (i.e. black is really black and equal to zero). Thus, the remaining signal is the true cytoplasmic GFP signal and it may not be appropriate to subtract it. Other data (fig 5J and 7D, now named fig 5G and 7H) were also re-analyzed with no major change.

1. b) Similar to apical fraction, measurements of planar polarization (trailing/lateral ratio) show average ratios near 1 for Dg, Rab10 and Dys, which is striking given that the localization of these proteins is so clearly polarized. Ratios lower than 1, which are reported for many individual cells in these graphs, should mean reversed polarity. In light of this, I would not be too confident on the effects reported in 5O-Q. In fact, on two occasions, the authors obtain significant differences in these planar polarization measurements that they themselves disregard: Fig. 6J (Rab10 in Exo70-) and Fig. 7I (Dys in Rab8 KD). *We agree that this quantification could be improved. Our initial quantification of the planar polarized proteins, Rab10 and Dys, found at the trailing edge, was confounded by their lateral spread. We have now reported with only the front half of the lateral side. By doing this for instance on Figure 5, we increased the ratio in the control conditions, with almost no points below the value of 1, while the conditions in which polarity is visually affected are unchanged and still close to 1. Thus, this new quantitative approach reinforced our conclusions on this figure. *

*For the figure 6, this new analysis confirm our previous observations : we observed a significant effect of Exo70 mutant, but not of Exo70 overexpression, on Rab10 localization (Figure 6J) while both impact Dys localization (Figure 6F). Main text mentions these two results. *

• Regarding the effect of Rab8 on Dys localization, we indeed observed a slight decrease of its polarization that we currently cannot explain (Figure 7). The important point here is that this effect is opposite to the one observed in Exo70 mutants. Thus, Exo70 effect on Dys cannot be explained by the fact that Rab8 route is blocked in this context. Text has been modified: “Conversely, Rab8 KD slightly affected Dys localization, but, importantly, this effect is opposite to the one observed in the Exo70 null mutant (Figure 7I).”*

• c) Quantifications of lateral fraction Col IV in mosaic experiments do not support decreased lateral secretion in Rab8 OE (3G) or Dys- (S5C), which are central tenets of the study. *We have endeavoured to detect such differences in Dys mutants and Rab8 OE and do not see any possible improvement in the quantification method and we therefore attempted, instead, additional experiments. *

*With respect to Rab8 OE, we suspect that this gain of function is not sufficiently effective under the specific conditions of the experimental setup described in Figure 3, as its effect appears to be more subtle than that of Rab10 OE in Figure 2. We therefore tried to repeat this experiment in a sensitized background in which Rab10 function was partially affected. Unfortunately, we did not see an improvement. However, since downregulation of Rab10 is not sufficient on its own to induce significant differences in this experimental setup, such an experiment is unconclusive and was not added the article. Nonetheless, we modified the results and the discussion to underly the data we got that strongly support that Rab8 route is targeted towards the basal domain with for instance the fact that Exo70 is required for Rab8 route and for basal secretion of collagen. *

*Regarding Dystrophin, we attempted to see whether its effect could be specific on its canonical ECM ligand that is Laminin A. Though we did not have the proper construct (UAS:LanA-GFP) to reproduce the same experiment set-up as with collagen, we tried to see whether Dg overexpressing clones, in presence or absence of Dys, were able to target LanA-GFP( under its own promoter) to the lateral domain of the cells. However, the result was negative and the experiment has not been included in the article. Thus, potential explanations of our results involving Dystrophin and Dystroglycan are detailed in the discussion. *

Minor comments:

• It is stated that Rab10 and Dys associate with tubular endosomes, but no data here support identification as endosomes of these tubular structures, to my understanding.

*We agree with this comment and we modified the text accordingly, mentioning a “tubular compartment” or” a subcellular compartment, with structures reminiscent of tubular endosomes.” *

• The authors call sup-basal the cell region immediately apical to the most basal. Is there sufficient reason to not call this lateral? If a new term is needed, shouldn't it be supra-basal?

*It was changed everywhere for supra-basal. *

• In Fig. S1A and B, Col IV is labeled as green but represented in cyan.

*Sorry for this mistake that has been corrected. *

• Fig. S1A should present a wild type control.

A control has been added.

• It is not clear where Y2H results in Fig 6A come from.

*The Legend has been modified to make it clearer : “ scheme of Dys domains and the fragments identified in a yeast two-hybrid screen with Exo70 as prey (Formstecher et al, 2005).” *

• Fig. 3C'-E' label suggests a gradient made from multiple images, but it looks like just two images and two colors.

*It is actually a true color gradient depending on z axis but the signal is indeed mainly found at the extremities of the z-stack. *

• Graphs in Fig. 3H-J, S5D and 7B are not legible.

• Fig. S1B does not seem to make a significant point in the context of this study.

*Although we understand this comment, we followed suggestion of R#2 who asked in its major comments for more details with other cell polarity markers. *

• I suggest drawing a summary scheme to aid readers better assess interpretations alternative to the ones given in the text.

*Such a summary scheme is now shown in the last figure. *

__Reviewer #1 (Significance (Required)): __

This study reports important new information on the secretion of Collagen IV by polarized cells of the Drosophila follicular epithelium. It complements previous studies on the roles of Rab8, Rab10 and Dystroglycan/Dystrophin, additionally uncovering a role for the exocyst complex. Addressing some issues with quantitative imaging should increase confidence in its most critical conclusions.

__Reviewer #2 (Evidence, reproducibility and clarity (Required)): __

Summary: In this manuscript, Dennis et al. identify different secretory routes and cell exit sites involved in basement membrane secretion and diversification in epithelial cells. Using the follicular epithelium of the Drosophila ovary as their model system coupled with genetics, imaging, and image analysis approaches, they show that two previously identified RabGTPases, Rab8 and Rab10, work in parallel routes for basement membrane secretion. These two small GTPases work in a partially redundant manner, where Rab8 promotes basal secretion leading to a homogenous basement membrane, while Rab10 promotes lateral and planer-polarized secretion, leading to the formation of fibrils. The authors also show that Rab10 and the dystrophin-associated protein act together to regulate lateral secretion, and dystrophin (Dys) is necessary for dystroglycan (Dg) to recruit Rab10. Furthermore, DAPC is shown to be essential for fibril formation and is sufficient to reorient Collagen IV to the Rab10-dependent secretory route. Dys was also shown to interact directly with exocyst subunit Exo70. Using overexpression and loss of function approaches the authors claim that Exo70 limits the planer polarization of Dys, and as a result, Rab10, hence limiting basement membrane fibril formation. Finally, the authors state that the Exocyst (Exo70) is also required for the Rab8-dependent basement membrane route. Overall, the data described in this manuscript are convincing and the authors' claims are supported by the presented data. We have mainly minor comments and only a few major comments that need to be addressed.

Major Comments:

• In the text for Figure 1G-H (page 4), the authors stated that the basal secretion was not restored in Rab8, 10, and 11 triple KD, in our opinion, it is unclear how the authors came to this strong conclusion from the presented data. It would be good if the authors explicitly explain how they come to this conclusion. Is it only based on the weak Coll-IV-GFP signal in the Rab8, 10, and 11 triple KD data compare to the control? If so, the authors should statistically quantify the difference with the control. In Figure 1H, no statistical analysis is provided between the control and triple KD conditions.

We agree that it was not entirely appropriate to give such conclusions on the basis of the quantifications available. A new graph showing basal fluorescence intensity (new Figure 1H) (and not just the ratio of apical to apical plus basal as in Figure 1I) has been added to better support the text. A relevant statistical comparison has been added to Figure 1H (old Figure 1I). We apologize for this oversight.

• From the data presented in Figure S1B, the authors state that the basement membrane mislocalization observed in Rab8/10KD has no major impact on polarity maintenance. They based this statement only on the localization of the apical marker aPKC. Although the aPKC data are convincing, it would be more compelling if the authors observe the distribution of other polarity proteins such as Dlg, E-Cadherin, and armadillo to better assess if the overall epithelial polarity is maintained in this condition. *Dlg and Ecad staining in these different genotypes were added to figure S1 with no major impact on the conclusions. *

Minor Comments: General comments: • In the text describing their data, we recommend that the authors clearly indicate which panel(s) they are referring to.

*We paid attention to this point in the revised manuscript. *

• The authors should also be consistent with the diction throughout the manuscript when referring to the cortical domain or region of the cell (back/rear/trailing edge/leading edge). *We tried to be more consistent. We now only speak about the “front” for one side and for the other of “trailing edge” or “rear half of the cell”, the latter corresponding to a more extended part of the cell than the previous. A scheme on figure 2A illustrates these terms. *

The following specific comments are in order of appearance in the manuscript. Introduction Section: The following statements in the introduction should be supported by specific references: • "BM is critical for tissue development, homeostasis and regeneration, as exemplified in humans by its implication in many congenital and chronic disorders."

We added the following reference: (Sekiguchi and Yamada, 2018)

• "BM is assembled from core components conserved throughout evolution: type IV collagen (Col IV), the heparan sulfate proteoglycan perlecan, and the glycoproteins laminin and nidogen." We added the following reference: (Mouw et al, 2014)

• "During development, the dynamic interplay between cells and BM participates in sculpting organs and maintaining their shape." We added the following references: (Sherwood, 2021; Jayadev and Sherwood, 2017; Walma and Yamada, 2020; Pastor-Pareja, 2020).

• "BM protein secretion shows some specificities, mainly because of the large size of the protein complexes (e.g., procollagen) that must transit from the endoplasmic reticulum to the cell surface". This statement could be supported with references including specific Drosophila references. Additionally, the authors need to clarify what they mean by "some specifies". We added the following references: (Ke et al ,2018; Feng et al, 2021).

Results section: • In the text describing Fig. 2 (page 5), the authors describe two different basement membrane types: fibrils and homogenous. Moreover, the manuscript focuses on the role of Rab8 and Rab10 in the formation of these two structures. Thus, the authors must better describe the two different types of basement membrane structures and their known roles. This will be helpful for the readers to analyze the presented data, especially for those that are not familiar with the system.

We rewrite the beginning of this paragraph : “ *Follicle BM is composed of an homogenous matrix from the very first stages while BM fibrils are added during the collective cell migration (Figure 2A, top) (Haigo and Bilder, 2011; Isabella 2016). Although the exact contribution of each of these BM types is not yet fully understood, genetic manipulation indicated that they are both required for the proper morphogenesis of the future egg (Haigo and Bilder, 2011; Isabella et al, 2016; Cerqueira Campos et al, 2020). Findings mainly based on gain of function experiments suggest that Rab10 participates in the follicle cell BM diversification by contributing to the formation of BM fibrils that are deposited as the cells migrate (Isabella et al, 2016). On the other hand, the route to generate homogenous BM remains unknown.” *

In Figure 2A, the authors describe stage 3 basement membrane as uniform BM, do they mean homogenous?

*Figure 2A has been corrected. *

• In the text describing the data for Fig. 3 (page 6), the authors should clearly explain the reason to use anti-GFP antibodies in a non-permeabilized condition (i.e., to detect specifically the extracellular secretion of BM proteins). This will help the readers to interpret the data presented. It is now explained as following “ Thus, detection of Col IV with an anti-GFP antibody and a Cy3- or Cy5-conjugated secondary antibody without permeabilization allowed discriminating secreted collagen from the total protein.*“ *

• On page 9, the authors stated that the precise localization of Dg in follicle cells is unknown. This statement is incorrect. It has been shown, using a Dg antibody, that Dg localizes at a high level at the basal side of the follicle cells and at a lower level at the apical side (Deng et al, 2003 and Denef et al. 2008). It has been corrected : “Endogenous Dg was described by immunostaining to be mainly enriched on the basal side of the cells (Denef et al, 2008; Deng et al, 2003).*“ *

Discussion Section: • The following statement is not clear: "Thus, three different Rab proteins are targeted towards the three distinct domains of epithelial cells defined by apical basal polarity, and at least of them is also planar polarized". The authors should rephrase and describe specifically which Rabs they are talking about.

*Text has been changed as following “Thus, these three different Rab proteins, Rab11, Rab10 and Rab8, are targeted towards the three distinct domains of epithelial cells defined by apical basal polarity, apical, lateral and basal, respectively. “ *

• This statement is vague: "These three Rab GTPases have been jointly involved in different processes (Knödler et al, 2010; Sato et al, 2014; Vogel et al, 2015; Eguchi et al, 2018; Häsler et al, 2020)". The authors could also mention the processes in which Rab8, 10, and 11 are involved. We tried to be more precise : “The same three Rab GTPases have been jointly involved in different processes such as ciliogenesis, targeted exocytosis or lysosome homeostasis where they have been proposed to act in a redundant manner”

• The following statements need to be supported by references. "Therefore, more investigations are required to define exactly how the DAPC allows the formation of BM fibrils. Nonetheless, given the importance of the DAPC and BM proteins in muscular dystrophies, our results will pave the way to determine whether a similar function is present also in muscle cells. Interestingly, the extracellular matrix is different between the myotendinous junction and the interjunctional sarcolemmal basement membrane and may provide another developmental context where several routes targeted to different subcellular domains may be implicated". *The following reference has been added : (Jacobson et al, 2020). *

Experimental Procedure Section: • In the dissection and immunostaining section (p14), there is a typo: it should be for "20 min" instead of "2for 0 min"

*It has been corrected. * • For the GST pulldown experiments, the authors mention that they use a standard protocol to produce S35 Exo 70 and the GST pulldown experiments. The authors should provide references.

A reference has been added.

Figure and Figure Legend: • General comment: The orientation of the images showing the rotation and leading and trailing edges need to be consistent in the different figures (e.g., In Figures 3 and 7, the leading edge is oriented to the top while in Figures 4, S4, 5, 6, the leading edge is oriented to the bottom). This will help the readers to analyze the data.

*We apologize for this, and we carefully checked image orientation throughout the figures. *

• In Figure 1 C-G the scale bars are missing and should be added as Fig. 1B.• In Figure 4, some scale bars are missing.• In Figure 6, some scale bars are missing. *Scale bars have been added. *

• Figure S1A: The data presented in Figure S1A is convincing. However, a control panel should be added showing the absence of apical Coll IV for comparison. This information will help with the interpretation of the data. A control has been added.

• In Figure 3 legend: it should be "immunostained" for GFP instead of stain for f-actin and GFP.• In Figure 4 legend: it should be "(A, E)" after (i.e 0.8 µm above the basal surface) instead of "(C, G)". In Figure 5 legend (p23), it should be "plane" and not "plan". *Legends of figure 3, 4, 5 have been corrected. *

• In Figure 5A-E, the authors show quantification of the fibril fraction for Dys-, Rab10 OE, and Rab10OE+Dys, Rab8KD, and Rab8KD+Dys-, and images of the collagen fibril for all the conditions except Dys-, it will be informative that the authors present a representative image of the Coll IV fibril in Dys- condition for comparison. The above comment also applies to Figure 5F-J, and it will be also informative to have a representative image of Dys- condition. The requested panels have been added.

• Overall, the legend for Fig. S5 is not clear and we recommend the authors to clearly described the different panels. (e.g., it should be "(D)" instead of "(H-J)") *Legend is now detailed as requested. *

__Reviewer #2 (Significance): __

Despite the important roles of the basement membrane for mechanical support, tissue and organ development, and function, the mechanisms that control the polarized deposition of basement membrane proteins are largely unknown. The contribution of Rab 8 and Rab 10 in the polarized deposition of the basement membrane was previously shown. However, by identifying two competitive secretory routes for the basal secretion of the basement membrane proteins that required these two different RabGTPases, controlled by the DAPC and the exocyst complexes, the authors make a novel contribution to our understanding of the mechanism that leads to the polarized secretion of basement membrane proteins (in that case Collagen IV). Since the basement membrane has critical roles in tissue and organ morphogenesis and functions, and its misregulation has been associated with developmental defects and pathological conditions, this research sheds light on the mechanisms important in these morphogenetic processes and will give insights into their deregulations in pathological conditions.

__Reviewer #3 (Evidence, reproducibility and clarity ): __

In the present work the authors have elucidated a novel mechanistic model of basement membrane morphogenesis using Drosophila ovarian follicle cells as a model. The authors have employed extensive quantification approaches to justify the spatio-temporal expression of the molecules under study such as Collagen IV, Rab8, Rab10, DAPC, etc. The authors suggest distinct exit domains of BM protein Collagen IV facilitated by Rab8 and Rab10 via distinct routes as they interact with each other. The authors further show that DAPC plays an essential role in Rab10 mediated baso-lateral fibrillar BM synthesis whereas Rab8 functions are more Exocyst (Exo-70 dependent).

Major comments: Result 1: The authors use RNAi lines to arrive at their conclusions, however, the extent of inhibition of gene expression achieved by the RNAi, has not been justified. Also observations from only one RNAi stock may not be completely conclusive:

i) Efficiency of RNAi has not been tested or shown. No supporting data. Rab10-RNAi stock is 26289 BDSC which is in Valium10, which is a weak RNAi line and needs a Dicer. ii) Can same observations be made using classic alleles or generate somatic clones on follicular epithelial cells?

*R#3 raised several questions regarding the efficiency of RNAi, the use of different lines and/or the use of classical mutants as an alternative method. *

For Rab10, we tested three different lines with similar results as shown now in Figure S1A-B. These data are also consistent with those obtained by overexpression of a dominant-negative form of Rab10 (Lerner et al, 2013). Unfortunately, Rab10 is located extremely close to the X chromosome centromere and is even more proximal than the FRT transgenes. It is therefore impossible to generate somatic mutant clones.

*Regarding Rab8, it is already published that Rab8 RNAi, expression of a dominant-negative form of Rab8 and Rab8 mutant cells obtained by somatic clones give similar defects (Devergne et al, 2017). The text has been modified to better illustrate the available data validating our approach. *

In addition, mutant clones would not allow analysis of genetic interactions in complex genetic contexts such as double and triple KDs. Similarly, the choice of the Rab10 line was motivated by the ease of obtaining the appropriate genetic combination according to their genomic location.

iii) Intensity of Collagen IV in the basement membrane in Rab11 knock-down mutants seems to be significantly low as compared to the Rab8 and Rab10 knock downs in supplementary Fig 1B. Are the authors very sure that Rab11 has no functions in basement membrane basal organization?

Good catch! Indeed, Rab11 RNAi significantly reduces basal secretion as now shown on fig 1H. Rab11 has pleiotropic functions in epithelial cells notably for their polarity (Choubey and Roy, 2017, Fletcher et al, 2012… and Fig S1). Thus, the reason for such a decrease is unclear and could be an indirect consequence of an overall abnormal epithelial structure. Thus, we now report this observation but have not taken its interpretation too far.

1. iv) Authors need to show where and how fluorescence intensities have been measured. *Magenta rectangles with dashed lines on Figure 1A illustrate the ROIs used for this analysis and more details have been added in the ‘experimental procedures’ section. *

Result 2: Confusing diagram. The authors should clarify whether the BM fibrils indicate lateral or planar BM components which they show to be more prominently expressed in Rab10 over-expression mutants.

A short note or an accompanying explanatory diagram on the source of the BM fibrils in the cellular context should make things less confusing.

*Schemes on figure 2A have been improved to make it clearer. *

FF calculation is an ingenious way of trying to look into functions.

The term Opposite effects/functions may be reconsidered as Rab8 and Rab10 compete with each other to deposit Collagen at spatially distinct domains. Opposite functions may give an impression that Rab8 actually represses Rab10 activity or vice versa, which may not be the case here.

*Text has been modified as suggested, speaking about “contrary effects” and “distinct functions”. *

Result 3: Why was anti-GFP Ab detected with Cy3-Cy5 secondary Ab. GFP itself is green so why detect it with a Red secondary? Logic? How clone Collagen GFP and ECM collagen GFP was differentiated? Please justify

It is now explained as following “ Thus, detection of Col IV with an anti-GFP antibody and a Cy3- or Cy5-conjugated secondary antibody ithout permeabilization allowed discriminating secreted collagen from the total protein.*“ *

A panel with a dotted line joining the peripheral or lateral Collagen as shown in panels D' E' of Fig 3 would support the cartoon provided and link the cartoon to the actual microscopic images.

*Figure has been modified as suggested. *

Result4:

The authors suggest a UAS-Rab10-RFP transgene show same results as endogenous Rab10-YFP as compared to spatial expression pattern. This is worrisome as expression of full length functional gene tagged with a fluorophore may be an overexpression. A control experiment would be helpful in suggesting/comparing with the Rab10 OE phenotype and that will be more convincing.

We are not sure that we fully understand the reviewer's comment. However, we initially compared endogenous Rab10 and UAS-RAB10 at 25°C, a temperature at which the latter has no visible impact on BM structure (Cerqueira-Campos et al, 2020). Furthermore, even when higher expression was induced (by increasing the temperature and therefore Gal4 activity) and this had an impact on BM structure, this did not change the subcellular localization of Rab10, i.e. it was still planarly polarized, as shown in Fig 5S. The text has been modified to emphasize this point.

When the authors mention back of cells, where do the authors exactly mean? A cartoon of "the back of follicle cells", wrt the entire ovarian follicle would be helpful.

*As asked by R#2, we are now more consistent throughout the paper, and a scheme illustrates these terms on Fig 2A. *

The authors suggest basal Rab10 expression domain near the Golgi exit point. Can the authors use a Trans-Golgi marker in order to confirm this statement other than the references stated?

*Trans golgi marker and Rab10 are now shown on figure S4. *

Result 5

The authors may provide a Rab10 expression profile in DAPC null or KD mutants which would make their claims more comprehensive.

*Data showing Rab10 localization in Dys mutant cells are shown on Figure S5A-B. *

Result 6:

Exo 70 is a versatile molecule and Rho kinases such as Cdc42 can direct Polarised exocytosis through interaction of Rab effectors with Exo 70. Have the authors considered this?

*We agree that it is an interesting prospect, but we consider it as beyond the scope of this article. *

In general some immunostainings should be carried out if not in all at least in some experiments with some cell domain specific markers, more specifically PCP markers such as Flamingo/Vangl and basolateral markers such as Lgl/Dlg. This makes the positions specific claims of the authors more valid in the eyes of the reader.

We agree that this may help the reader but the mentioned pcp markers are not expressed in this tissue. However, the tissue planar orientation is now systematically indicated and consistent in all figures. We did not generally perform immunostaining for lateral markers but routinely included F-actin staining to detect cellular cortex. Our quantifications or cortical segmentations were based on the cell outline provided by this stain. On the basis of this staining, the outline of the cells was added on certain figures to facilitate understanding of the images.

Reviewer #3 (Significance):

The findings impinge on a critical cellular process of Rab protein interactions in the genesis of the basement membrane which is of potential interest. This falls under basic research. Since Rab molecules have emerged as molecules governing membrane morphogenesis, Cell and Molecular Biologists as well as a wide audience including clinicians will be interested on this.

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

Learn more at Review Commons

---

Reply to the reviewers

Reviewer #1 Evidence, reproducibility and clarity: Bompierre et al have presented a set of interesting findings that demonstrate the interconnected roles of multiple PDEs in striatal cholinergic interneurons. They show that the regulation of neurons by PDEs differs between ChiNs and MSNs. Disentangling the complex and interconnected biology of PDE subtypes and calcium signalling is notoriously difficult and I commend the authors for not shying away from it. The data are interesting and compelling however I have a number of readily addressable concerns regarding their statistical analysis.

We are very grateful to the reviewer for the careful reading of our manuscript, positive comments and useful suggestions to enhance its readability. We also thank the reviewer for highlighting the difficulties resulting from the scarcity of these neurons. As described in the manuscript, during the course of several other projects, we noticed this sparse neuronal population that clearly departed from the vast majority of striatal neurons, the medium-sized spiny neurons. This particular neuronal population was identified as ChINs only later with immunohistochemistry. We were quite surprised that our visual identification was confirmed by immunohistochemistry in all cases, as described, now with more details, in the manuscript. We then performed a number of new experiments focused on ChINs while we could also re-analyze older experiments performed for other projects and in which ChINs were visually identified - all our experiments are terminated with the acquisition of a Z stack which allows a precise observation of each neuron in the imaging field. This explains some changes in the drugs presented in this manuscript.

Major concerns: • Statistics, general comments: o When performing multiple comparisons (as in figures 2-6) the authors should be using a one-way ANOVA or Friedman's test (for non-parametric).

Since data normality of a few samples was rejected by the Shapiro-Wilk test, and in line with our previous publication, we used non-parametric statistics for all of this study. The Friedman’s test is now applied to all situations comparing more than two conditions.

o When comparing an effect size between experimental conditions e.g. cAMP following NMDA with or without Lu AF64193, the conditions need to be compared directly, not just inferred from being 0.05. [ see Makin and Orban de Xivry 2019 PMID: 31596231 and Nieuwnhuis et al 2011 PMID: 31596231 for more details].

We agree with this comment and the recommended tests have been performed.

o Please include P values and degrees of freedom for each analysis, rather than just * indicating PP values are now indicated in the text - although a smaller P value does not imply that the difference is “more” significant. Degrees of freedom do not apply with Friedman and Wilcoxon rank tests.

• Introduction: o Page 3 para, describing the various PDEs reported to operate in ChiNs vs MSNs, is very convoluted and hard to follow. I recommend replacing this paragraph with a table: column 1: PDE subtype, 2: neuron type, 3: effect reported 4: ref.

We agree that this paragraph was difficult to understand. We tried to build a table as suggested, but such table could not convey all the aspects that we think should be made clear (mRNA or protein data, for example). Instead, we propose a revised paragraph that we hope will be easier to read.

• Methods: o Please include explanation for estimated cAMP concentration (right axis on figs 2c, 3b,5c-f

This was described in “Methods / Estimates of biosensor activation level”. We also added a similar estimate for cGMP concentration based on our previous work with the cGMP biosensor cyGNAL in Figures 4 and 6.

o Please specify slice thickness and age of mice

Slice thickness was missing and now added in methods (300 µm). The age was indicated in the first paragraph of Methods: “7 to 12 days”.

• Figure 1: o describe if 13 uM fsk is supposed to be a maximal concentration, what is the justification for this concentration. Is there a previously published dose-response curve?

A biochemical study of adenylyl cyclase intracellular domains (VC1 and IIC2 heteromer) reports a Kd for forskolin of 0.1 µM {Dessauer et al., 1997, J Biol Chem, 272, 22272-7}, well below the dose we used. To our knowledge, there is no published dose-response analysis of forskolin effect in ChINs and we did not perform this measurement. We routinely use 13 µM for practical reasons, our stock being 25 mM diluted 2,000-fold. ChINs display a much lower cAMP response than MSNs at this dose: we think that this interesting observation deserves to be reported (but we would like to point out that this is only a marginal aspect of our study). We totally agree that this point deserves more explanation and a paragraph has been added in the discussion: “The low responsiveness of ChINs to cAMP-activating signals such as forskolin is striking but the underlying cellular mechanisms remain to be determined. All adenylyl cyclases except AC9 are activated by forskolin {Defer et al., 2000, Am J Physiol Renal Physiol, 279, F400-16}. AC1, AC2 and AC5 are widely expressed in the brain, including the striatum {Matsuoka et al., 1997, J Neurochem, 68, 498-506}. Cluster analysis of mRNA transcript suggest that AC1 and AC2 predominate in ChINs whereas AC5 is mainly expressed in MSNs {Saunders et al., 2018, Cell, 174, 1015-1030.e16}. This indicates that the reduced responsiveness to forskolin in ChINs compared to MSNs does not result from ChINs lacking adenylyl cyclases sensitive to forskolin.”.

• Figure 2c: o Clarify which replicates are shown on the graph (I assume it's the n=11 i.e. number of slices)

Yes, the plot shows the 11 replicates of the same protocol, with each set of 3 points of a same color linked with a line showing the measurements for one experiment. This is now stated more clearly in the figure legend.

o You say "PDE3 thus contributes importantly to the regulation of cAMP level, and when this phosphodiesterase is inactivated pharmacologically, cAMP level becomes controlled by PDE4." But in fig 2c panel 2 inhibition of PDE4 with Picla increases cAMP without PDE3 being already inhibited. This doesn't agree with your statement, which implies that PDE4 only controls cAMP once PDE3 has been inhibited.

This is unfortunate since we did not want to suggest that PDE4 has no effect unless PDE3 is inhibited. Indeed, this experiment together with the next actually demonstrate that PDE3 and 4 are simultaneously engaged in cAMP regulation. This ambiguity was also noted by reviewer 3. We thus rephrased the link between between these two paragraphs.

o Data described but not shown: "In the absence of forskolin, application of PDE3 inhibitor (cilostamide, 1 µM, N=3; n=3; A=3) or PDE4 inhibitor (rolipram, 1 µM, N=2; n=2; A=2) or their combination (N=6; n=8; A=3) induced no significant ratio change in ChINs (as well as in MSNs)".

We removed the mention to the experiments with cilostamide and rolipram alone since N was less than 5. The effect of rolipram and cilostamide together are displayed in Figure 2, with proper statistics indicated in the text.

Also please explain why you have changed PDE inhibitor?

This project developed over several years and the phosphodiesterase inhibitors used in the team changed depending on their availability. In addition, some data such as the lack of effects on baseline cAMP level were extracted from experiments performed for other purposes, hence some differences in the inhibitors we used. Please note that given the scarcity of this neuronal population, many more trials and mice would be required to reproduce these experiments with a single inhibitor. We further consider that our approach contributes to the desired reduction in the use of animals in research (the 3-R).

• Figure 4: o Please clarify in text that data in 4b is from ChiNs and not pooled with data from MSNs? i.e. is the summary data from MSNs not shown?

Figure 4B and C only show average calcium responses in ChINs. This is now indicated in the results section “Figure 4B shows the calcium response to NMDA in ChINs with the average trace (left), and baseline and peak amplitude (right) for individual ChINs.” The calcium response to NMDA uncaging in MSNs has already been published (Betolngar 2019) and is not shown in this study.

o Response to DHPG is not formally compared between MSNs and ChiNs

In MSNs, the calcium response to DHPG showed variability, with a lack of response in most experiments but a clear calcium signal in a few MSNs. The cause of this variability was not the focus of this study but, nevertheless, we wanted to mention this qualitative observation to stimulate future studies on this subject. However, if a quantitative measurement of this variability is required, the description of the DHPG effect on MSNs will be removed.

• Figure 5: o Explain reasoning for switching PDE4 inhibitor (roflumilast).

At the time these experiments were performed, we were using roflumilast to inhibit PDE4.

Explain change in fsk concentration (0.5 uM in figure 5 vs 13 uM in earlier figs)

We wanted to study PDE1 in isolation, i.e. in conditions in which both PDE3 and PDE4 were inhibited. Simultaneous inhibition of PDE3 and 4 leads to biosensor saturation (Figure 2), a situation in which a PDE1-mediated decrease in cAMP level might be difficult to resolve. The forskolin concentration was therefore reduced to decrease adenylyl cyclase activation. This is now explained in the manuscript: ”In order to stimulate a moderate cAMP production and thus maintain the visibility of PDE1 action, a lower concentration of forskolin (0.5 µM) was employed in these experiments.“

o Text states: "These effects were blocked with Lu AF64196 (1-10 µM), a potent and selective PDE1 inhibitor" but this was not formally tested and in the figure a response is still visible (smaller than before Lu, but not blocked)

Indeed, a small change in the average ratio trace is still visible, so we changed “blocked” by “largely reduced”. The effect of Lu AF64196 in the NMDA condition was tested as follows: the change in ratio level induced by NMDA uncaging was calculated in control and Lu AF64196 conditions. This ratio change was compared between control and Lu condition with a Wilcoxon rank test. The same test was applied to compare control and DHPG condition. This is now indicated in the manuscript.

o Please explain the reason for the different time courses in figure 5ab vs 5c-f)

Figure 5A,B are illustrative experiments. Figure 5C-F are average traces from several experiments. This is indicated in the figure legend.

o Data not shown: "Of note, the addition of the PDE1 inhibitor did not increase the steady-state cAMP level, demonstrating that PDE1 did not exhibit a tonic activity before its activation by the calcium signal"

In our experimental conditions, the cAMP level is too high to faithfully report an increase in cAMP upon Lu AF64196 application, so we removed this sentence. However, the cGMP level in the presence of DEANO is farther from saturation, allowing to perform this control: the application of Lu AF64196 produced no significant increase in cGMP level, indicating that PDE1 is not active in our experimental conditions. This has been added in the results.

• Figure 6: o Comment on why using quisqualate (mixed AMPA and mGLuR) rather than DHPG as used previously?

These early experiments were performed with these drugs until we made sure that group 1 mGlu was responsible for this effect and the more specific agonist DHPG was used. The drug combination, however, is specific of group I mGlu activation.

o Again, effect reported as being blocked, but not formally tested and a response is still evident on the graph. If the authors believe that response is an artefact from the stim then please show data with NMDAR antagonists.

We agree that a small change in the average ratio trace is still visible, so we changed “blocked” by “largely reduced”. The effect of Lu AF64196 was tested as described above for cAMP, which is now indicated in the manuscript. We agree that NMDA as well as mGlu stimulation, by increasing calcium, can affect cyclic nucleotides by other mechanisms than PDE1 activation. This was already reported in our previous work, in particular in the hippocampus and cortex (Betolngar 2019). We are not sure that NMDAR antagonists would clarify the situation since NMDA receptor blockade would probably suppress the cAMP change that is still visible in the presence of the PDE1 inhibitor. Nonetheless our manuscript reports experimental conditions in which the vast majority of the effect that we focus on is blocked by the PDE1 selective inhibitor.

• Discussion: o Add references in para 2 describing the PDEs expressed by ChiNs

Done

o Figure 7: cartoon indicates that calcium will exclusively activate PDE1A, is this for simplicity or is their evidence to support this?

Single-cell RT-PCR data {Saunders et al., 2018, #42891} indicate that ChINs express PDE1A but not PDE1B. This is now indicated in the introduction. The cartoon has been changed accordingly.

o Please comment on the specificity of the pharmacological tools used throughout the study.

Phosphodiesterases bear a cleft-shaped catalytic site that is particularly amenable to chemical inhibition, and phosphodiesterase thus constitute a therapeutic target of great interest: a large number of highly specific phosphodiesterase inhibitors have been developed and tested by pharmaceutical companies. It nonetheless remains that our demonstration relies on the specificity of the phosphodiesterase inhibitors. We added a sentence in the discussion “We used highly specific phosphodiesterase inhibitors to acutely test the functional contribution of these phosphodiesterases.” to acknowledge this.

o In the context of regulation of striatal cholinergic and dopaminergic signalling by NO (via cGMP), I believe the authors should cite Hartung et al PMID: 21508928

This very valuable citation has been added in the discussion.

Minor comments: • Page 3 paragraph 1 and 3 the authors have used ellipsis instead of finishing their sentences.

Done.

• In figure 1a please indicate you are recording in dorsomedial region of the striatum (you state in methods this is your recording location, but it would be helpful to show it here)

We thought of improving the cartoon in Figure 1 by drawing a rectangle over the dorso-medial striatum on the image of the brain slice, but that would suggest that the slices had been cut at this stage of the preparation, which was not the case. Adding another image of a brain slice would clutter the figure. We leave it to the Editor to decide how this should be handled.

• The authors use a lot of different drugs, I think a table listing the drugs, concentrations and their targets would help limit confusion.

This will certainly make it easier for the reader. This table has been added in Methods / Chemicals and drugs.

• I found the graphs with each replicate in a different colour quite difficult to read. My personal preference would be for graphs to show each replicate as a transparent line/small symbol, with the mean and SEM shown larger and in bold.

We tried to enhance the visibility as suggested. SEM should not be shown since our statistics are non-parametric.

Significance: General strengths: The question of how PDEs interact to regulate striatal output is extremely interesting and notoriously difficult to tackle. The authors have relied upon pharmacological manipulation of PDEs. A pharmacological approach has both strengths (intact system with little compensation occurring between PDEs, which would occur with a genetic strategy) and weaknesses (relying on each of the drugs to act selectively and specifically). By investigating multiple PDEs in the same system in two neuron types, I believe the authors are illustrating interesting findings. For these findings to be more concrete I believe they need a more robust statistical approach, but the experiments and questions are valid. PDEs are of interest to most neuroscientists and understanding cholinergic function is of interest to anyone studying the striatum of basal ganglia more broadly. My expertise is investigating the interactions between striatal cholinergic and dopaminergic signalling.

Reviewer #2 Evidence, reproducibility and clarity : The work characterizes in depth the dynamics of the cyclic nucleotide signaling in cholinergic interneurons (ChINs) in the striatum and the interconnection with calcium signaling. The study is ambitious and risky since it targets a minority of neurons representing only 1% of the total population of striatal neurons. For that they used genetically encoded biosensors, at a very low infection rate, and highly specific phosphodiesterase inhibitors. With these tools they defined PDE1, PDE3 and PDE4 as the key regulators of cAMP levels in ChINs and the interplay with incoming signals raising cGMP and free-calcium levels after nitric oxide or glutamate activation of NMDA or mGlu1/5 receptors, respectively. The conclusions of the study are solid and well supported by the experimental results.

We would like to thank the reviewer for his/her positive comments about our work.

The team has the necessary technical and conceptual background in the field. This is very important to trust the criteria they used to identify ChINs, a fundamental hallmark in this study.

Again, we are very grateful to the reviewer for this very positive comment.

Still, confirmation by immunohistochemical labeling with ChAT antibodies sounds important and perhaps it should had been performed in more experiments.

We agree that our qualitative immunostaining validation of ChAT expression was too terse. We re-analyzed our archived data to provide a more precise account of our observations. We first identified ChINs from their morphology in the biosensor image stack, then checked whether these neurons were positive for ChAT. This is now explained in detail in “Identification of Cholinergic Interneurons in a brain slice”: “11 brain slices were fixed after the biosensor experiment and later processed for ChAT immunoreactivity. In these slices, 15 neurons were visually identified as ChINs during the biosensor recording session. All of these neurons showed a positive ChAT labelling.”

Significance: The results represent an important conceptual advance in the field. To understand better the signaling that regulates firing of cholinergic neurons in the striatum might be relevant to explain pathological responses and they could be useful to define better strategies for the treatment of Parkinson's patients, for instance. In this regard, this study fills an existing gap since this elusive neuronal population was not functionally characterized before. The basic aspects of the study could be of interest to a broad audience.

Reviewer #3 Evidence, reproducibility and clarity: Summary: The author's present very elegant findings regarding how NO regulates cAMP in striatal cholinergic interneurons (ChINs). The major strength of the manuscript lies in the approach, which enables single-cell imaging of neuronal signaling in acute brain slices. A clever combination of pharmacological tools were then utilized to dissect PDE contribution toward cAMP alterations in ChINs. While the manuscript is high-quality, there are a few controls and points of discussion that need to be considered.

We would like to thank the reviewer for his/her careful reading of our manuscript and very positive comments about our study.

Major: There are numerous references to results seemingly missing from the figures. - Figure 2 TP-10 - Figure 2 PF-05 traces - Figure 2 data in absence of FSK (rolipram, cilo) - Figure 3 ODQ

We thought that these experiments showing a lack of effect were of little interest to the reader and therefore omitted raw traces and statistics from the manuscript. However, we fully agree to display more of our data, as long as it does not clutter the main points of our manuscript. We now illustrate the lack of effect of PDE2A inhibition with a typical experiment in Figure 2C. The ratio level is now shown in Figure 2D for roli-cilo and TP-10, with matching statistics in the text. Figure 3 now shows the lack of effect of ODQ.

Have the author's considered an alternative perspective that slow cAMP detection in ChIN, relative to MSN, could be due to the size of neuron? The significantly greater volume of ChIN soma could conceivably require more cAMP to reach the detection threshold of the biosensor. Therefore, how do the author's reconcile such technical caveats?

This is a very interesting hypothesis that is supported by many theoretical and experimental data: it takes longer for membrane adenylyl cyclases to fill up a void volume in which both phosphodiesterases and the biosensor reside. We could rule-out the buffering effect of the biosensor by the experiment described in Figure 1B, but a lower surface to volume ratio such as that observed for ChINs vs MSNs could indeed explain a biologically slower onset in cAMP level. However, a lower surface to volume ratio should not affect the steady-state level that will be eventually reached upon continuous forskolin application: it takes longer to fill up the volume but, if waiting long enough, the final level will be only determined by the equilibrium between cyclase and phosphodiesterase. Forskolin applications of more than 10 min (Figure 2) led to steady-state levels that were far below biosensor saturation, while it did reach saturation in MSNs. Therefore, while we certainly acknowledge the importance of the peculiar neuronal morphology of ChINs, there must be additional specific differences between ChINs and MSNs. In any case, we agree that this important point was missing in our manuscript and we added a paragraph in the discussion to discuss differences in adenylyl cyclases and cell geometry.

In the traces from Figure 2, it is unclear why PDE3 and PDE4 have differential contributions toward cAMP elevation depending on the order of inhibition.

Thank you for pointing out our poor wording, which has also been noted by the first reviewer. Indeed, the data in Figure 2 shows that PDE3 and PDE4 are simultaneously engaged in cAMP regulation. This part has been rewritten.

Moreover, we cannot conclude that PDE3 and PDE4 the major PDEs in ChIN from such experiment based on the result that IBMX did not further raise cAMP. Likely, the biosensor has reached the detection ceiling. This should be discussed as a possibility.

It is an interesting possibility that cAMP levels higher than biosensor saturation level ([cAMP] above 100 µM) could be modulated after PDE3 and PDE4 inhibition, in concentration ranges that go beyond biosensor saturation level. However, our experiments clearly demonstrate that the combined action of PDE3 and PDE4 constitutes the first line of cAMP control since the concomitant inhibition of PDE3 and PDE4 raises [cAMP] beyond physiological relevance. When both PDE3 and PDE4 were inhibited, cAMP indeed reached the level of biosensor saturation which led us to state “The ratio was not further raised by the non-specific phosphodiesterase inhibitor IBMX (200 µM), indicating the saturation level of the biosensor by cAMP (Rmax)”, a conclusion that remains valid.

The author's intepretation ignores the influence of calcium on the activity of various types of ACs, which seems to be a critical feature given the experimental design that first broadly stimulates ACs with forskolin.

We agree with the Reviewer that calcium will certainly activate AC1 (possibly present in ChINs) and inhibit AC5 (expressed in MSNs and possibly also in ChINs). Our experimental design relies on the highly specific PDE1 inhibitor to isolate the selective contribution of PDE1, but minor changes in cAMP and cGMP remained even after PDE1 inhibition, as also pointed out by the first Reviewer. We found experimental conditions in which the contribution of PDE1 could be largely visible, which was the point of this study. However, we certainly agree with the Reviewer that a calcium signal will affect cyclic nucleotide levels through a number of other mechanisms. Therefore, we added the sentence “It should also be noted that, in the presence of the PDE1 inhibitor Lu AF64196, some changes in cAMP level still remained, which can result either from incomplete PDE1A inhibition and/or from NMDA effects on other targets, such as calcium-modulated adenylyl cyclases.”

A missing control in Figure 5 is the effect of Lu AF64196 by itself on cAMP (and in the presence of FSK pre stimulation).

We agree that this is an important control. However, this could not be tested on this protocol with cAMP since the ratio was too close to Rmax, and an increase in cAMP level following PDE1 inhibition would be undetectable. However, with cGMP imaging, the ratio reached with DEANO was farther from saturation such that an increase in cGMP resulting from the inhibition of PDE1 should be detectable. However, Lu AF64196 showed no significant effect. This measurement was added in the manuscript: “As a control, we verified that PDE1inhibition had no effect on the steady-state cGMP level elicited by 10 µM DEANO (Figure 7G: in DEANO: 0.78 of Rmax; in DEANO and Lu AF64196: 0.81; N=5, n=6, a=5; Wilcoxon P=0.094).”

The mechanism would be signicantly bolstered by measuring cAMP from PDE4 inhibition following forskolin and DEANO (Figure 3).

It is true that, in the presence of forskolin and DEANO, the only PDE that remains is PDE4: its inhibition should increase cAMP to the maximal level. Unfortunately, this experimental scheme was not tested. However, Figure 3 is focusing on the regulation of PDE3 by cGMP and at this stage of the reasoning, it might be confusing to get back to the question of which PDE remains after PDE3 has been blocked. In this manuscript, we want to highlight that PDE3 is blocked via the NO-cGMP signaling pathway, and we think that the data in Figure 3 demonstrates this point clearly.

Minor: The introduction should discuss the critical role of ChIN in striatum rather than simply stating "critical role in striatal functions"

A paragraph has been added in the Introduction to highlight the importance of ChINs in striatal function.

PDE should be included as abbreviation in introduction after first mention of phosphodiesterases.

We agree that this was inducing confusion. We keep PDE1-11 as it is the official names of proteins, but we replaced all occurrences of “PDE” by “phosphodiesterases” when alluding to the general concept of this class of enzymes.

Light sources (e.g. laser) for excitation during imaging are missing from the methods.

This was indicated in Methods / Biosensor imaging: “LED light sources (420 nm with a 436 nm excitation filter and 360 nm) were purchased from Mightex (Toronto, Canada).”

The study certainly provides implication for diseases associated with striatal dysfunction such as Parkinson's disease. However, it may be important to note that experiments were performed on slice preparations from very young animals, which could have inherent differences in functionality relative to an aged or diseased context.

We agree that our preparation of brain slices from mice pups is not representative of what could be found in adults, and even less in pathological conditions. Nonetheless, we believe that we identified a crosstalk mechanism that has never been reported in neurons, and that is not taken into account in theories of striatal functions. We hope that this novel understanding will lead to the development of experiments on adult mice that might confirm the functional importance of this effect, in particular pharmacological studies in adult animals with PDE3 inhibitors.

Significance: General assessment: The study utilizes pharmacological tools to selectively target enzymes and receptors in the cAMP cascade to mechanistically dissect how cAMP is handled in ChINs. The major strength is ability to perform such experiments in acute brain slices, i.e. a "native" neuronal context. An exciting aspect is stimulation of NMDAR by agonist-uncaging, thereby revealing an endogenous signaling route that modulates cAMP. A major physiological limitation is reliance on fsk to induce AC activity, however the approach is suitable to obtain mechanistic information. Moreover, conducting experiments in a Parkinsonian disease model would provide tremendous value, although such pursuits are beyond the scope of the work here. Advance: The study builds off robust studies previously published by the author's. The work is also similar to AC-cAMP investigations on acute brain slices performed by the Sabatini (24765076, 29154125) and Martemyanov (29298426, 31644915) labs, which unfortunately were not discussed/cited here. This is perhaps a missed opportunity to highlight the significance of the author's study. For instance, the Sabatini lab investigated calcium influence on cAMP but in the hippocampus. The Martemyanov lab investigated striatal cAMP, but not in ChINs or through calcium or cGMP mechanisms. Therefore there is a gap toward understanding striatal ChINs, which is clearly demonstrated in the author's work here.

These very important references have been left out of our manuscript intentionally since not pertaining directly to the topic of the manuscript. We would like to point out that we also omitted citations to our own work which had also described dopamine, adenosine and acetylcholine responses measured with various biosensors in striatal neurons (Castro 2013 23551948; Yapo 2017 28782235; Nair 2019 24560149). Indeed, this research field is flourishing and we hope that people interested in the general question of cAMP signal integration in neurons will easily find many such relevant publications.

Audience: I find this basic research to have a relatively broad appeal. For example, my lab is working on behavioral aspects of motor dysfunction. Therefore, the pharmacological insight here is very intriguingly. The mechanistic nature of the work may also appeal to those working on the signaling aspect of such diseases.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

Summary:

The author's present very elegant findings regarding how NO regulates cAMP in striatal cholinergic interneurons (ChINs). The major strength of the manuscript lies in the approach, which enables single-cell imaging of neuronal signaling in acute brain slices. A clever combination of pharmacological tools were then utilized to dissect PDE contribution toward cAMP alterations in ChINs. While the manuscript is high-quality, there are a few controls and points of discussion that need to be considered.

Major:

There are numerous references to results seemingly missing from the figures. - Figure 2 TP-10 - Figure 2 PF-05 traces - Figure 2 data in absence of FSK (rolipram, cilo) - Figure 3 ODQ

Have the author's considered an alternative perspective that slow cAMP detection in ChIN, relative to MSN, could be due to the size of neuron? The significantly greater volume of ChIN soma could conceivably require more cAMP to reach the detection threshold of the biosensor. Therefore, how do the author's reconcile such technical caveats?

In the traces from Figure 2, it is unclear why PDE3 and PDE4 have differential contributions toward cAMP elevation depending on the order of inhibition. Moreover, we cannot conclude that PDE3 and PDE4 the major PDEs in ChIN from such experiment based on the result that IBMX did not further raise cAMP. Likely, the biosensor has reached the detection ceiling. This should be discussed as a possibility.

The author's intepretation ignores the influence of calcium on the activity of various types of ACs, which seems to be a critical feature given the experimental design that first broadly stimulates ACs with forskolin.

A missing control in Figure 5 is the effect of Lu AF64196 by itself on cAMP (and in the presence of FSK pre stimulation).

The mechanism would be signicantly bolstered by measuring cAMP from PDE4 inhibition following forskolin and DEANO (Figure 3).

Minor:

The introduction should discuss the critical role of ChIN in striatum rather than simply stating "critical role in striatal functions"

PDE should be included as abbreviation in introduction after first mention of phosphodiesterases.

Light sources (e.g. laser) for excitation during imaging are missing from the methods.

The study certainly provides implication for diseases associated with striatal dysfunction such as Parkinson's disease. However, it may be important to note that experiments were performed on slice preparations from very young animals, which could have inherent differences in functionality relative to an aged or diseased context.

Significance

General assessment:

The study utilizes pharmacological tools to selectively target enzymes and receptors in the cAMP cascade to mechanistically dissect how cAMP is handled in ChINs. The major strength is ability to perform such experiments in acute brain slices, i.e. a "native" neuronal context. An exciting aspect is stimulation of NMDAR by agonist-uncaging, thereby revealing an endogenous signaling route that modulates cAMP. A major physiological limitation is reliance on fsk to induce AC activity, however the approach is suitable to obtain mechanistic information. Moreover, conducting experiments in a Parkinsonian disease model would provide tremendous value, although such pursuits are beyond the scope of the work here.

Advance:

The study builds off robust studies previously published by the author's. The work is also similar to AC-cAMP investigations on acute brain slices performed by the Sabatini (24765076, 29154125) and Martemyanov (29298426, 31644915) labs, which unfortunately were not discussed/cited here. This is perhaps a missed opportunity to highlight the significance of the author's study. For instance, the Sabatini lab investigated calcium influence on cAMP but in the hippocampus. The Martemyanov lab investigated striatal cAMP, but not in ChINs or through calcium or cGMP mechanisms. Therefore there is a gap toward understanding striatal ChINs, which is clearly demonstrated in the author's work here.

Audience:

I find this basic research to have a relatively broad appeal. For example, my lab is working on behavioral aspects of motor dysfunction. Therefore, the pharmacological insight here is very intriguingly. The mechanistic nature of the work may also appeal to those working on the signaling aspect of such diseases. Please define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

The work characterizes in depth the dynamics of the cyclic nucleotide signaling in cholinergic interneurons (ChINs) in the striatum and the interconnection with calcium signaling. The study is ambitious and risky since it targets a minority of neurons representing only 1% of the total population of striatal neurons. For that they used genetically encoded biosensors, at a very low infection rate, and highly specific phosphodiesterase inhibitors. With these tools they defined PDE1, PDE3 and PDE4 as the key regulators of cAMP levels in ChINs and the interplay with incoming signals raising cGMP and free-calcium levels after nitric oxide or glutamate activation of NMDA or mGlu1/5 receptors, respectively. The conclusions of the study are solid and well supported by the experimental results. The team has the necessary technical and conceptual background in the field. This is very important to trust the criteria they used to identify ChINs, a fundamental hallmark in this study. Still, confirmation by immunohistochemical labeling with ChAT antibodies sounds important and perhaps it should had been performed in more experiments.

Significance

The results represent an important conceptual advance in the field. To understand better the signaling that regulates firing of cholinergic neurons in the striatum might be relevant to explain pathological responses and they could be useful to define better strategies for the treatment of Parkinson's patients, for instance. In this regard, this study fills an existing gap since this elusive neuronal population was not functionally characterized before. The basic aspects of the study could be of interest to a broad audience.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

Bompierre et al have presented a set of interesting findings that demonstrate the interconnected roles of multiple PDEs in striatal cholinergic interneurons. They show that the regulation of neurons by PDEs differs between ChiNs and MSNs. Disentangling the complex and interconnected biology of PDE subtypes and calcium signalling is notoriously difficult and I commend the authors for not shying away from it. The data are interesting and compelling however I have a number of readily addressable concerns regarding their statistical analysis.

Major concerns:

• Statistics, general comments:
  • When performing multiple comparisons (as in figures 2-6) the authors should be using a one-way ANOVA or Friedman's test (for non-parametric).
  • When comparing an effect size between experimental conditions e.g. cAMP following NMDA with or without Lu AF64193, the conditions need to be compared directly, not just inferred from being < or > 0.05. [ see Makin and Orban de Xivry 2019 PMID: 31596231 and Nieuwnhuis et al 2011 PMID: 31596231 for more details].
  • Please include P values and degrees of freedom for each analysis, rather than just * indicating P<0.05
• Introduction:
  • Page 3 para, describing the various PDEs reported to operate in ChiNs vs MSNs, is very convoluted and hard to follow. I recommend replacing this paragraph with a table: column 1: PDE subtype, 2: neuron type, 3: effect reported 4: ref.
• Methods:
  • Please include explanation for estimated cAMP concentration (right axis on figs 2c, 3b,5c-f
  • Please specify slice thickness and age of mice
• Figure 1:
  • describe if 13 uM fsk is supposed to be a maximal concentration, what is the justification for this concentration. Is there a previously published dose-response curve?
• Figure 2c:
  • Clarify which replicates are shown on the graph (I assume it's the n=11 i.e. number of slices)
  • You say "PDE3 thus contributes importantly to the regulation of cAMP level, and when this phosphodiesterase is inactivated pharmacologically, cAMP level becomes controlled by PDE4." But in fig 2c panel 2 inhibition of PDE4 with Picla increases cAMP without PDE3 being already inhibited. This doesn't agree with your statement, which implies that PDE4 only controls cAMP once PDE3 has been inhibited.
  • Data described but not shown: "In the absence of forskolin, application of PDE3 inhibitor (cilostamide, 1 µM, N=3; n=3; A=3) or PDE4 inhibitor (rolipram, 1 µM, N=2; n=2; A=2) or their combination (N=6; n=8; A=3) induced no significant ratio change in ChINs (as well as in MSNs)". Also please explain why you have changed PDE inhibitor?
• Figure 4:
  • Please clarify in text that data in 4b is from ChiNs and not pooled with data from MSNs? i.e. is the summary data from MSNs not shown?
  • Response to DHPG is not formally compared between MSNs and ChiNs
• Figure 5:
  • Explain reasoning for switching PDE4 inhibitor (roflumilast). Explain change in fsk concentration (0.5 uM in figure 5 vs 13 uM in earlier figs)
  • Text states: "These effects were blocked with Lu AF64196 (1-10 µM), a potent and selective PDE1 inhibitor" but this was not formally tested and in the figure a response is still visible (smaller than before Lu, but not blocked)
  • Please explain the reason for the different time courses in figure 5ab vs 5c-f)
  • Data not shown: "Of note, the addition of the PDE1 inhibitor did not increase the steady-state cAMP level, demonstrating that PDE1 did not exhibit a tonic activity before its activation by the calcium signal"
• Figure 6:
  • Comment on why using quisqualate (mixed AMPA and mGLuR) rather than DHPG as used previously?
  • Again, effect reported as being blocked, but not formally tested and a response is still evident on the graph. If the authors believe that response is an artefact from the stim then please show data with NMDAR antagonists.
• Discussion:
  • Add references in para 2 describing the PDEs expressed by ChiNs
  • Figure 7: cartoon indicates that calcium will exclusively activate PDE1A, is this for simplicity or is their evidence to support this?
  • Please comment on the specificity of the pharmacological tools used throughout the study.
  • In the context of regulation of striatal cholinergic and dopaminergic signalling by NO (via cGMP), I believe the authors should cite Hartung et al PMID: 21508928

Minor comments:

• Page 3 paragraph 1 and 3 the authors have used ellipsis instead of finishing their sentences.
• In figure 1a please indicate you are recording in dorsomedial region of the striatum (you state in methods this is your recording location, but it would be helpful to show it here)
• The authors use a lot of different drugs, I think a table listing the drugs, concentrations and their targets would help limit confusion.
• I found the graphs with each replicate in a different colour quite difficult to read. My personal preference would be for graphs to show each replicate as a transparent line/small symbol, with the mean and SEM shown larger and in bold.

Significance

General strengths: The question of how PDEs interact to regulate striatal output is extremely interesting and notoriously difficult to tackle. The authors have relied upon pharmacological manipulation of PDEs. A pharmacological approach has both strengths (intact system with little compensation occurring between PDEs, which would occur with a genetic strategy) and weaknesses (relying on each of the drugs to act selectively and specifically).

By investigating multiple PDEs in the same system in two neuron types, I believe the authors are illustrating interesting findings. For these findings to be more concrete I believe they need a more robust statistical approach, but the experiments and questions are valid. PDEs are of interest to most neuroscientists and understanding cholinergic function is of interest to anyone studying the striatum of basal ganglia more broadly. My expertise is investigating the interactions between striatal cholinergic and dopaminergic signalling.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

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

Learn more at Review Commons

---

Reply to the reviewers

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

Summary

In this manuscript, the authors characterized the molecular function of the brain-enriched kinase KIS by combining transcriptome-wide approaches with molecular and functional studies. They uncover that KIS regulates isoform selection of genes involved in neuronal differentiation and inhibits through phosphorylation the capacity of the splicing regulator PTB2 to interact with both target RNAs and protein partners.

Major comments

- This is a very clear and well-written manuscript presenting high-quality and carefully controlled experimental results. The authors used an impressive range of approaches (transcriptome-wide exon usage, phospho-proteomic, imaging, biochemical assays..) to profile exon usage alterations upon KIS knock down and provide a mechanistic understanding of how KIS regulate the splicing activity of PTBP2. Specifically, they convincingly demonstrate that the phosphorylation of PTBP2 by KIS leads to both dismantling of PTBP2 protein complexes and impaired RNA binding.

My only main concerns relate to the understanding of the biological context in which the mechanism studied may be at play. That KIS can counteract PTB2 activity through direct phosphorylation has been very clearly shown by the authors using overexpression of KIS and /or PTB constructs in different contexts (HEK293T cells, N2A cell line, hippocampal neurons). Whether this occurs endogenously in the context of neuronal differentiation, and how much this contributes to the overall phenotypes induced by KIS inactivation, is less clear. While fully investigating the interplay between KIS and PTB2 in the context of neuronal differentiation is beyond the scope of this study, the three following points could be addressed to provide some evidence in this direction.

1- Building on the experiments they perform in a KIS knock-down context (e.g. Fig. 3B, or previously described spine phenotype), the authors should investigate whether inhibiting PTBP2 in this context (through shRNA or expression of a phospho-mimetic construct) might suppress the phenotypes observed when inactivating KIS.

As suggested by the reviewer, we have added a new results section showing the effects on spine maturation in hippocampal neurons expressing PTBP2 phosphomutants and in a PTBP2-KIS double knockdown scenario (Fig4 and S4 Fig; Results section: P6 L12-P7 L21). First, PTBP2-overexpression effects on post-synaptic protrusion density are exacerbated by the phosphoablated mutant. Intriguingly, the phosphomimetic mutant still has a negative impact in spine formation, suggesting either a residual ability of this protein to interact with its normal partners or the existence of additional roles of PTBP2 in spine development that are Matrin3 and hnRNPM independent. Second, KIS knockdown partially suppresses the defects in mature spine formation produced by the loss of PTBP2. In all, these data support the notion of KIS being a phosphorylation-mediated inhibitor of PTBP2 activity during neuronal differentiation.

2- Based on Figures 1E and 3A, it seems that KIS downregulation affects both exon inclusion and exon skipping, and that its function in exon usage is only partly explained by modulation of PTBP2-dependent exons. Have the authors analyzed the populations of PTBP2-dependent exons that are regulated by KIS in an opposite manner? This may point to specific classes of transcripts (in terms of expression pattern, function, molecular signature) important in the context of endogenous neuronal differentiation.

We have analyzed the GO terms of genes with KIS-upregulated exons by that are either downregulated or upregulated by PTBP2. In both groups we found an enrichment of genes in terms associated with calcium ion activity but with different specific functions. Genes with exons downregulated by PTBP2 are more involved in transmembrane transfer of calcium ions from intracellular stores whereas genes with exons upregulated by PTBP2 facilitate the diffusion of calcium ions through transmembrane postsynaptic. Interestingly, with respect to the cytoskeleton, the two groups show a clearly different term enrichment. Genes with exons downregulated by PTBP2 are significantly associated with the tubulin cytoskeleton, whereas genes with exons upregulated by PTBP2 are associated with the actin cytoskeleton. We have added a paragraph in the results section (P8 L6-15) and a new panel in S4C Fig.

3- The authors should better discuss when and where they think PTBP2 phosphorylation by KIS might be relevant. Is there evidence that this process (or PTBP2 complex assembly) might be regulated upon differentiation or plasticity?

We have modified the Discussion section (P11 L35-P12 L14) as follows:

Regarding neuronal differentiation, it is worth noting that KIS expression increases during brain development (Bièche et al, 2003) and in vitro differentiation of hippocampal cultures (Fig. 1B), coinciding with postnatal decrease of PTBP2 levels (Zheng et al, 2012). Therefore, the phosphorylation-dependent inhibition of PTPB2 by KIS and the concerted relative inversion in their levels would generate a molecular switch linking transcriptional and alternative exon usage programs in neuronal development (Fig 6D). In mature neurons, alternative splicing has a well-established role in expanding proteome diversity (Mauger & Scheiffele, 2017). Although the connection between synaptic activity and the control of KIS expression and/or kinase activity is not yet established, the contraposition of PTBP2 and KIS in splicing may constitute a fine-tuning mechanism to modulate proteome diversity as a function of plasticity-inducing signals. In this regard, single-cell transcriptomic data from hippocampal neurons show that expression variability of KIS and PTBP2 is much higher compared to actin (Perez et al, 2021) (S6B Fig). Thus, differences in the expression of these two splicing regulators at a single neuron level would increase protein isoform variability and expand diversity in neural circuits, a crucial property in information processing (Miller et al, 2019).

Minor comments

1- Figures and associated legends are overall very clear and well-organized. Addressing the following points would however help improving the clarity of some Figures:

• In Figure 2EV2C legend, the characteristics of the 3SA constructs are not described

We have modified the legend of Fig 2EV2C (S2C Fig in revised version) to clarify this point.

- the difference between Figure EV1A and Figure 1H classifications is unclear, nor the interpretation regarding the different GO classes identified

The gene lists used for the two GO term analyses are different. In Fig EV1A (now S1A Fig) the gene list is more restrictive than in Fig 1H as we choose genes with more than one exon upregulated by KIS. In contrast, the analysis in Fig 1H includes all genes with one exon upregulated by KIS.

2- Whether PTBP2 is endogenously the major target of KIS explaining transcriptome-wide changes in exon selection is a possibility that remains to be demonstrated. Thus, the authors should correct and tune down the following sentences:

"KIS phosphorylation counteracts PTBP2 activity and thus alters isoform expression patterns ..." (end of introduction)

"PTBP2 being one of the most relevant phosphotargets" (results, end of the second section)

We agree with the reviewer and we have modified the two sentences (P4 L6-8) and (P6 L10) in the revised version of the paper.

Reviewer #1 (Significance (Required)):

• The splicing regulator PTBP2 is a known master regulator of neuronal fate whose tightly controlled expression drives the progenitor-to-neuron transition as well as the establishment of neuronal differentiation programs. How this protein is regulated at the post-translational level has so far remains poorly investigated. In this manuscript, the authors provide a thorough mechanistic understanding of how KIS-mediated phosphorylation of PTB2 impacts on its regulation of exon usage. They also provide a transcriptome-wide view on the function of the brain-enriched KIS kinase in exon usage, uncovering its broad functions in alternative splicing.

If the physiological context in which KIS-mediated phosphorylation of PTB2 is induced remains to be precisely defined, this work opens interesting new perspectives on regulatory mechanisms at play during neuronal differentiation. Providing extra lines of evidence indicating that KIS acts on neuronal functions through PTBP2 phosphorylation will help further strengthen this aspect.

• This manuscript will be of interest to different large communities interested on one hand on the regulation of gene expression programs underlying neuronal differentiation and on the other hand on the molecular regulation of major complexes involved in alternative splicing and isoform selection. It opens new perspectives related to the spatiotemporal regulation of neuronal isoform selection.

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

Summary

The manuscript by Moreno-Aguilera et al. shows that the brain enriched protein kinase KIS targets the well known neuronal splicing regulator PTBP2 and several of its interaction partners. As a consequence, PTBP2 activity is down-regulated. Using cultured primary immature neurons they show that KIS expression increases during differentiation and that shRNA knockdown of KIS alters the splicing of many alternative exons. Phosphoproteomic analysis of HEK293 cells transfected with KIS or a kinase dead mutant (K545A) show that it phosphorylates both PTBP2 as well as a cluster of proteins that are known to interact with PTBP2 or its paralog PTBP1. By comparing the new data on KIS-dependent splicing with previous data-sets on PTBP2-dependent splicing targets they show that KIS appears to act antagonistically with PTBP2 when it acts as a repressive regulator, but not when it is an activator. Using combinations of wt and kinase-dead KIS with PTBP2 mutants in the 3 main phosphorylation sites (3SA - non-phosphorylatable, S3D - phosphomimetic) to look at the effects on a known PTBP2 functional target, PSD95, they show that the likely effect of KIS is to antagonise PTBP2 function by phosphorylation at one or more of three residues (S178, S308, S434). Finally, they show that transfected KIS (but not K54A) reduces known protein-protein interactions of PTBP2 and that the triple phosphomimetic PTBP2 mutant shows reduced binding to RNA. Alphafold2 predictions show that the S178 phosphomimetic mutant might alter the conformation of the RRM2 domain, in particular altering the environment of Y244, which has been shown in PTBP1 to be critical for interaction with MATR3 and other coregulators.

Major points

In general, the conclusions drawn are consistent with the data. I have a few suggestions where the authors could either extend their findings with a few straightforward additional experiments, or clarify some of the existing data.

FigEV4 (also introductory text on p3): RRMs 3 and 4 of PTBP1/2/3 fold as a single back to back packed didomain - with the so-called linker contributing to the didomain fold (e.g. PMID: 24688880, PMID: 16179478) and also extending the RNA binding surface by creating a positive patch (e.g. PMID:20160105 PMID: 24957602). AlphaFold successfully predicts the didomain in full length PTBP2 (https://alphafold.ebi.ac.uk/entry/A0A7I2RVZ4). The authors should therefore use AlphaFold2 to predict the RRM3-4 di-domain structure of wt and phosphomimetic mutant PTBP2s. Phosphorylation of S434 or S434D, which is on the C-terminal end of RRM3 may have no predicted effect on RRM3 alone (FigEV4), but it could conceivably disrupt didomain packing, which could itself have important knock-on consequences for RNA binding. In addition, the inrtoduction of negative charges at S434 might affect the ability of R438, K440 & K441 to interact with RNA. An image of the didomain charge density of WT and mutant PTBP2 would be useful to address this.

As suggested by the reviewer, we have considered the di-domain structure of RRM3 and RRM4, and AlphaFold2 predicted no effects by the phosphomimetic residues. We have added the di-domain predictions to S6B Fig.

S434 lies at the very end of RRM3 and limits with a basic region that would not bind RNA in a canonical RRM-dependent manner. In addition, as predicted by AlphaFold2, this basic region is not structured and the effects of a nearby negative charge may be difficult to predict.

Figure 4 could also easily go further in experimentally testing the effects of individual phosphomimetic mutations upon protein-protein interactions (Alphafold predicts that S178D, but not S308 or S434D, should affect Y244 mediated interactions, such as MATR3). The co-IP approach in Fig 4A could readily be used with FLAG-PTBP2 mutants. Likewise, consequences of individual mutations upon RNA binding (Fig 4D) could be tested. The use of a Y244N mutant here would test whether the loss of RNA binding is a consequence of the loss of protein-protein interactions. Such experiments are not essential, but they are readily carried out and have the potential to unravel the consequences of the individual phosphorylation events (more correctly of phosphomimetic mutants).

After building the Y244N mutant we tested PTBP2 interactions with protein partners and observed no significant changes in the levels of Matrin3 and hnRNPM proteins in FLAG-PTBP2 immunoprecipitates nor in the RNA binding ability of PTBP2. These data suggest that, although Y244 is involved in the interaction between PTBP1 and PRI-containing proteins such as Raver1, the interaction between Matrin3 and PTBP2 would involve structural determinants other than the Matrin3 PRI and the PTPB2 Y244 residue. Compared to PTBP1, the nearby flexible loop between RRM2 and RRM3 in PTBP2 is very different and could accommodate specific interaction determinants with Matrin3.

Minor points

Do KIS regulated exons show enrichment of motifs associated with PTBP2, consistent with the proposed model - particularly CU-rich motifs upstream of exons that are more repressed upon KIS shRNA treatment.

We have not observed a significant enrichment of CU-rich sequences upstream of the top 100 exons upregulated by KIS. Indeed, our data suggest that only a fraction of exons upregulated by KIS are inhibited by PTBP2.

For the splicing analysis pipeline, how were exon-exon junction reads treated? If "only exons with more than 5 reads in all samples" were considered, will this not exclude highly regulated exons that are completely skipped under one condition?

This sentence has been corrected as "only exons with more than 5 reads in all samples of one condition..." (P17 L9)

The Introduction mentions U2AF homology (UHM) domains, but neglects to discuss their known binding partners - ULMs (UHM ligand motifs), which contain an essential tryptophan. It would be useful for the discussion to highlight whether any direct KIS interactors possess ULMs and how this relates to the phospho-targets identified here. The authors may wish to draw the parallel with the structurally analogous way that PTBP1 (and presumably PTBP2) interact with their short peptide ligand motifs.

As suggested by the reviewer, we have searched for ULM sequences in the identified KIS phosphotargets, but we only found clear ULMs in SUGP1, which contains KRKRKSR__W__385 and KMG__W__573K. The absence of ULM motifs in most of the proteins identified in the KIS phosphoproteome would indicate that phosphorylation does not require stable protein-protein interactions. We have added these lines to the Discussion section (P10 L34-P11 L2). We completely agree with the reviewer that, in future work, it would be very interesting to test the possibility that KIS binding modulates the composition and functional properties of splicing complexes through ULM domains.

Figure EV2C. The S3A and S308A mutations clearly reduce phosphorylation. However, the effects of S178A and S434A are far less clear. Presumably the quantitation shown in the lower panel of EV2C relies on normalization to PTBP2 protein input, which appears quite variable in the Coomassie gel. It might be better to repeat the experiment with uniform protein inputs. Minimally, details of the quantitation approach should be added to Materials and Methods.

The different levels of reduction in 32P incorporation displayed by the single phosphonull mutants suggests that phosphorylation follows a hierarchical pattern, S308-P facilitating phosphorylation of the other two phosphosites. We have added this comment to the revised version of the paper (P5 L25-28). As mentioned by the reviewer, 32P incorporation was made relative to the total amount of PTBP2 present in the assay, which was deduced from cold Coomassie-stained gels run in parallel to the radioactive gels with same amount of proteins. We have added details of the quantification in the Methods section from 3 independent experiments

Fig 3D shows PTBP2 overexpression, but the main text (p7) states KIS overexpression.

The text and panel order in Fig 3D were misleading and have been corrected (P7 L9-14).

Fig 4B should have a scale bar for the FRET signal

          Done (now Fig 5B)

Fig 4E should indicate the location of S178

          Done (now Fig 6C)

Reviewer #2 (Significance (Required)):

Significance

This interesting, clear and concise manuscript provides important new insights into the way that a neuron specific kinase can regulate neuronal splicing networks by phosphorylating and thereby downregulating the known neuronal splicing regulator PTBP2. Alternative splicing is known to play a particularly important role in neurons, so this demonstration of an additional layer of regulation by post-translational modification should make the manuscript of wide interest to investigators of splicing regulation, neuronal differentiation and maturation.

Issues that are not addressed in the manuscript include; i) how does KIS specifically target PTBP2 and related proteins? The UHM domain can mediate interaction with ULM containing splicing factors (such as U2AF2, SF3B1), but none of the identified targets have known ULMs. ii) the consequences of individual phoshomimetic mutants upon protein-protein interactions and RNA binding could readily be explored further using computational and experimental methods already used in the manuscript.

For context, this reviewer has a direct interest in the mechanisms of regulation of alternative splicing, but not in the context of neurons (though I am familiar with a lot of the relevant literature), and I do not have expertise in neuronal cell biology.

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

In this manuscript, the authors explored the function of the protein kinase KIS in splicing regulation associated with neuronal differentiation in vitro. KIS is a serine threonine kinase known to phosphorylate splicing factors such as SF1 and SUGP1, and to be preferentially expressed in adult brain in mammals. Using an shRNA based approach, the authors characterize cassette exon usage upon partial KIS depletion in cultured mouse cortical neurons.

In parallel using mass spectrometry of proteins in KIS overexpressing HEK293 cells, they identify potential KIS substrates including the splicing regulator PTBP2. They confirm that recombinant KIS can phosphorylates PTBP2 in vitro. They show a correlation between KIS-activated and PTBP2-inhibited exons using published data for this factor. They report opposite effects of KIS and PTBP2 on CamKIIB splicing and Finally, coimmunoprecipitation and FRET experiments suggest that KIS inhibits the interactions of PTBP2 with known protein binders, hnRNPM and Matrin3 as well as with RNA. Altogether these data suggest that KIS downregulates PTBP2 during neuronal differentiation.

Majors comments:

Overall the manuscript is well written and the data are interesting.

However several points could have been more extensively studied or discussed to achieve a stronger demonstration of the role of KIS in PTBP2 phosphorylation and neuronal differentiation.

1) To minimize possible off target problems, the RNAseq analysis would be more convincing if replicated with a second shRNA to knockdown KIS.

The efficiency of the selected shRNA had been validated both by the supplier (Merck-Sigma) and in our previous work, which also included a complementation assay (see Fig. 4A-C in Pedraza et al 2014).

2) Part of the reported splicing changes might reflect an indirect consequence of an altered differentiation contributing to the correlation observed in figure 1F. It would be interesting to confirm splicing changes using shorter incubation times with the shRNA compared to the 11 days used in this study.

The levels of splicing regulators such as PTBP1 and PTBP2 change quite markedly during the initial phases of neuronal differentiation (Zheng et al 2012). However, we observed no change in their levels when comparing KIS knockdown to control conditions, suggesting no major upstream effects on the differentiation program per se. But, more important, whereas the splicing pattern of CamKIIβ transcript was clearly affected by KIS knockdown at 18 DIV (Fig 3B), we observed no changes at 14 DIV.

3) Standard deviation is more relevant to describe data dispersion in all figures.

For non-parametric statistics we prefer the interquartile range as a measure of dispersion. For the parametric statistics of 3 independent experiments we show the standard error of the mean as a measure of dispersion.

4) Previous papers of the group described a function of KIS in translation (Cambray et al 2009, Pedraza et al 2014). This is not discussed here. For example, the possibility that RBPs are regulated by KIS at the translation level is not excluded by the analysis in Fig EV2a.

It is an interesting comment by the reviewer that we have considered during the course of this work. Nevertheless, in our experiments coexpressing KIS and PTBP2 in HEK293 cells we did not observe any reduction in splicing factor levels. We have included a representative immunoblot (S5C Fig) of input samples from the experiments shown in the corresponding main figure (Fig 5A in the revised version).

Minor comments:

Figure 1: The authors state that "KIS...accumulates in nuclear sub-structures adjacent to those formed by splicing factors". As the figure presents in fact GFP-KIS, it should be mentioned, and how this localisation is relevant for endogenous KIS should be addressed.

We have corrected the text to mention that GFP-KIS was used to analyse its nuclear localization pattern as shown in Fig 1A (P4 L12-13). We had previously validated the nuclear localization (Boehm et al, 2002) of an N-terminal fusion of KIS to a fluorescent protein (Cambray et al, 2009).

Fig EV1: SI range in pannel D is very different from that in pannel C and Fig1E.

In this figure we plot the average SI obtained from bins with 2500 exons, which would necessarily narrow the SI range obtained from individual exons. Our data indicate that protein disorder would only constitute a minor, but significant, factor in exon usage.

On page 4 "KIS expression reached maximal levels in hippocampal cultures (Fig 1B)." However the figure legend indicate that this analysis was performed with cortical neurons. The use of cortical or hippocampal neurons along the manuscript should be clarified.

We apologize for the typing mistake, and it has been corrected in the revised version (P4 L17-18)

page 4 " KISK54A, a point mutant without kinase activity" The authors should indicate the reference.

The reference to Maucuer et al (1997) has been added (P5 L15)

Figure EV2C: It is not clear whether the Coomassie staining and autoradiography do correspond to the same gel.

32P incorporation was made relative to the total amount of PTBP2 present in the assay, which was deduced from cold Coomassie-stained gels run in parallel to the radioactive gels with same amount of proteins.

Figure 3C. The authors use a dual fluorescence reporter to analyse PSD95 exon 18 splicing. However the well to well variability in such experiments might be elevated. Not only the cells number in a single well but also the number of replicates should be indicated and well to well variability reported.

As stated in the figure legend, the dual fluorescence reporter experiment has been analyzed at a single-cell level. Using ImageJ software, we analyzed the fluorescence of 1054, 970 and 672 cells expressing KIS, KISK54A or none, respectively, from 3 independent experiments.

Figure 3D. The precise timing for the transfection and culture of cells before staining is unclear

Hippocampal neurons were transfected at 5DIV and fixed at 12 DIV. This description has been added to the legend in Fig 3D.

Figure 4A. The input should be loaded to evaluate the coIP efficiencies and ascertain that KIS does not downregulate Matrin3 and hnRNPM levels.

We agree with the reviewer. We have included a representative immunoblot (S5C Fig) of input samples from the experiments shown in the corresponding main figure (Fig 5A in the revised version).

Figure EV4A. No difference of Matrin3 binding is to be seen on the gel. In addition, the authors should confirm that PTBP2 or binders are phosphorylated by recombinant KIS. The preparation of GST-KIS is not described.

We agree with the reviewer that in the in vitro assays the differences in phosphorylation are not as clear as in the in vivo experiments. Fig S2C shows an in vitro kinase assay of PTBP2 by recombinant KIS. Finally, we include a reference (Pedraza et al, 2009) for the preparation of recombinant GST-KIS (P14 L7)

Page 6: "We found that PTBP2-inhibited exons are significantly (FDR=0.001) enriched in KIS knockdown neurons, supporting the notion that KIS acts on AS, at least in part, by inhibiting PTBP2 activity." This should be rephrased as in fact PTBP2-inhibited exons are enriched among KIS activated exons.

The sentence has been rephrased as “We found that PTBP2-inhibited exons are significantly (FDR=0.001) among KIS activated exons…” (P6 L17-18)

Page 10: "SUGP1 is one of the most enriched proteins in our KIS phosphoproteome (see Fig 2A)". Phosphorylation and interaction with KIS was already reported by Arfelli and coll. 2023 supplementary figure 2.

We have modified the Discussion section to add this information (P11 L14-15)

Page 10: " It forms part of the spliceosome complex, interacts with the general splicing factor U2AF2 and has been reported to play an important role in branch recognition by its association with SF3B1." A reference is needed there.

A reference to Zhang et al (2019) has been added.

Page 10: The authors previously reported a differentiation defect in cultured neurons 'Cambray et al, 2008' that was not observed by another group (Manceau et al., PLOS One 2012). This should be discussed in view of these more recent results. Is there any differentiation defect in the experiments reported there?

We have added a new results section showing the effects on spine maturation in hippocampal neurons expressing PTBP2 phosphomutants and in a PTBP2-KIS double knockdown scenario (Fig4 and S4 Fig; Results section: P6 L12-P7 L21). First, PTBP2-overexpression effects on post-synaptic protrusion density are exacerbated by the phosphoablated mutant. Related to the point raised by the reviewer, KIS knockdown also decreased spine emergence and maturation, but partially suppressed the defects produced by the loss of PTBP2. In all, these data support the notion of KIS being a phosphorylation-mediated inhibitor of PTBP2 activity during neuronal differentiation.

Statistical values are difficult to read in the figures. Please use larger fonts.

Done

Reviewer #3 (Significance (Required)):

This manuscript brings new elements supporting the function of the protein kinase KIS in splicing regulation in neurons. In particular it identifies for the first time the splicing regulator PTBP2 as a substrate for KIS.

It will be of interest to a specialized audience of researchers interested in splicing regulators in neuronal differentiation.

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

Learn more at Review Commons

---

Reply to the reviewers

The authors do not wish to provide a response at this time.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

Mouse models have contributed significantly to the description of the anti-tumor and the tumor-promoting activity of gd-T cells, which (in mouse) can be divided into an IFNg and an IL-17-producing subgroup based on their CD27 expression. The study describes Ly6C+ and Ly6C cells as "new" subsets of IFNg-producing murine CD27+ gd T cells and proposes Ly6C+ cells as mouse counterparts to human Vd2 T cells. The use of Ly6C expression as a potential marker for functionally distinct CD27+gd T cell subpopulations is based on single-cell transcriptomics of murine lung gd T cells and was validated by flow cytometry of surface antigens and assay of cells for proliferation/cell death after CD3/CD28 stimulation, killing of tumor cells in vitro and in vivo and analysis of the conversion of Ly6C- to Ly6C+ cells and vice versa. Ly6C cells were defined as "immature" and Ly6C cells as "mature". Ly6C+ cells show increased tumor cell kill in vitro and better IFNg production in vitro. In a tumor model, adoptively transferred ex vivo expanded Ly6C+ but not Ly6C- cells prolonged survival. Based on the data shown in Fig. 6, it is postulated that IL-27 is important for the maintenance of LyC6+ cells in vitro and in vivo and for the increased IFN-γ production of Ly6C+-derived cells. It also modulates killing of tumor cells in vitro. Finally, the effects of IL-27 on human Vd1 and Vd2 T cells are analyzed and Vg9Vd2 T cells are postulated as a human counterpart to Ly6C+ gd-T cells. At times, I found the work hard to read, and hope that as someone who doesn't work with mouse gd-T cells, I fully understood the work.

I agree with the characterization of LyC6+ cells as a "more mature/differentiated" subset of IFNg-producing CD27 gd T cells. My doubts mainly concern the interpretation of the transcriptomic data (Fig. 1) and the data shown in Fig.6 and 7 (erroneously referred to as Fig. 8) which described the effects of IL-27 and the postulated similarity of LyC6+ cells and human Vg9Vd2 T cells.

I find it confusing to talk about Ly6 expression of Ly6+ and Ly6- cells. Perhaps it would be better to talk about these expanded cells Ly6+ derived and Ly6- derived cells (or something similar).

Fig. 1 While the identification of clusters 0 and 1 helps to identify LyC6 expression as a marker for CD27+ gd T cell differentiation, I do not understand how the gene signatures and the expression patterns in human lymphocyte subgroups "highlight the high similarity between gd T cells between species", since Fig. 1 E and 1 F also show high similarities between both clusters and non-gd T cells. In fact, the greatest similarities exist between cluster-0 cells and NK cells, followed by mature Vd1 and finally by total Vg9Vd2 T cells and CD8 cells. Naïve Vd1 and naïve (ab T) lymphocytes cells show mainly similarities to cluster 1 cells.

Fig. 2 Please name the organ from which the cells were taken. Please also present in the extended data figures a representative staining with a complete gating strategy, starting with and live gating, to give a wider audience an idea of how the subpopulations of CD27+gd T cells were identified. I also wonder if the expression levels (not the frequency) of CD27 and Ly6C are the same for different organs.

Fig. 3. It appears that Ly6C+ -derived cells show an effect against the E0771 tumor. In the Kaplan Meyer plot these were compared only with the PBS control. Please show the P values for Ly6C+ and Ly6C - derived cells.

Fig. 4 Please indicate the absolute number of gd-T cells and CD27+gd-T cells in organs and tumors.

Fig. 5 Please indicate the absolute number of gd-T cells and CD27+gd-T cells in organs and tumors.

Fig. 6A-C What is meant by control? No cytokines or IL-2 and IL-15 as in the other experiments plus the indicated cytokines. Please specify.

Fig. 6D-G Please indicate the absolute cell counts. Please give also cell counts and frequencies of CD27-Ly6C+ and Ly6C- cells directly taken ex vivo for WT and IL-27R-/-.

Figs. 6E and G and extended data Fig. 4. Was the significance of the differences between WT and IL-27-/- also tested for Ly6C-derived cells, as it was done for L6C+-derived cells. At least in some cases, I would be surprised if such differences were not observed (e.g. Extended data Fig. 4A for NKG2A expression in LN and lung). In addition, in some cases, the lack of differences of for Ly6C- cells between WT and IL-27R-/- cells may be (in part) reflect to their low abundance, which makes it difficult to find statistically significant differences in Il-27 dependence.

Fig. 6F is aimed to show that there is no statistically significant difference in the Ly6C expression of Ly6C- derived cells between IL-27R/- and WT animals, while there is such a difference for the Ly6C+ derived cells. To me, it looks like the lack of significance of differences between the Ly6C-derived WT and IL-27R-/- cells is due to a single outlier. Especially, since only three animals were analyzed compared to five or more in most other experiments I am not convinced of the biological (non-) significance of these differences.

Fig. 6E is indeed "incongruent" since it contradicts the claim that IL-27 supports the anti-tumor response of CD27+ gdT cells, as stated in the ONE-SENTENCE-SUMMARY. To claim a positive effect of IL-27 on tumor control by CD27+ gdT cells, an experiment as shown in Fig. 3 with gd T cells from IL-27R/- and WT mice or with cells expanded with or without IL-27 is required. Another question concerns the purity of cells after CD3/CD28 stimulation. Could it be that contaminating cells (e.g. NK cells) contribute to killing and that IL-27 effects are due to such cells?

Fig. 7 (incorrectly referred to as 8) and extended data Fig. 5. Does the MFI indicate the MFI of the positively gated cells or the alll cells? Please specify. Please also indicate the proportion of Vd1+ and Vd2+ cells before and after expansion for all donors. Fig. 5 A shows an approximately 30% inhibitory effect of IL-27 on Vd1 expansion and a 70-75% reduction in expansion of Vd2 cells. This contradicts the postulated (functional) Vg9Vd2 T cell homology with Ly6C+ cells, since in mice ( Fig. 6) IL-27 reduces the expansion of Ly6C cells and accelerates the expansion of Ly6C+ cells. Another problem could be the comparison of the mouse subgroups with the total Vd1 and Vg9Vd2 T cells. In view of the data in Fig. 1, it would be better to compare mature and naïve cells from both populations.

Mat. and Meth. been mentions FTOC experiments but I did not find them?

Mat. and Meth. also mentions that WT and Tcrb/- mice were used as a source of CD27+Ly6C- and CD27+Ly6C+ cells in the in vivo expansion experiment. I didn't find any mention of the use of Tcrb/- mice in the text of the illustrations. Please explain.

Ly6C is encoded by the genes Ly6c1 and Ly6c2. Ly6c2 is the dominant expressed gene, but in some populations (CD4+ T cells), a significant proportion of Ly6C-positive cells express Ly6C1 https://doi.org/10.4049/immunohorizons.2100114. This should be mentioned in the text. An optional experiment would be to test the expression of mRNA for both genes in naïve and expanded gd-T cells.

With respect to possible revisions. For the core of the paper description of the Ly6C (derived populations (Fig. 1-5) 1-2 months should be sufficient. More complicated is to address the issue of Il-27. Fig. 6. A good part of my concerns are on the evidence of the lack of an Il-27 response of Ly6C- derived cells. Putting less emphasis on this point and might be largely sufficient, since in my eyes it is not so central anyway. A problem is the statement on the importance of IL-27 in control of tumor by Ly6C derived cells. This would need in vivo experiments which may take much longer (-6 months?).

Significance

The work is certainly of interest to most immunologists working with mouse gd-T cells. The characterization of the Ly6C-defined subgroups of CD27+gd-T cells is plausible except for details and supported by the results. The data on IL-27 in Fig. 6 and extended Fig. 4A are less clear, in particular the lack of effect of IL-27 on Ly6C cells may have been overestimated sometimes. Also the importance of IL-27 for tumor control by gd-T cells in vivo has yet to proven. Both would broaden the audience, nd may increase translational/clinical relevance.

The weakest part is the postulated similarity between human gd T cell groups and the two mouse gd T cell subgroups. The cell populations of the mouse could indeed have human counterparts, but this has yet to be shown. Given the strong species-specific differences between both species especially for unconventional T cells, it is also conceivable that non-gd T cells take over the functions of certain gd T cell populations.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

The manuscript submitted by Wiesheu et al focuses on the intrinsic nature of murine gd T cells in tumor immune response. Authors used their previously published scRNA-seq data and using a refined strategy, they dissected the role of Ly6C subset-specific CD27+ gd T cells. The authors represented experimental evidence for the possible differentiation of CD27+Ly6C gd T cells and revealed the essential role of IL-27 for the maintenance of CD27+Ly6C+ gd T cells. Finally, the authors found parallels in cellular features of gd T-cell subsets between mice and humans. Overall, this study is well-designed and expands knowledge of gd T-cell biology. However, the authors are suggested to consider the following comments:

Major:

1. One of the major concerns is that the submitted manuscript is the extension of their previous work (Edwards et al 2022) published in J Exp Med. This previous work clearly delineated the differential role of PD-1 and TIM-3 using the scRNA-seq approach. Using the same scRNA-seq data, authors are now defining gd T-cell subsets within the CD27+ population. While providing the functional basis for Ly6C expression, the author shall also define the role of PD-1 or TIM3 in this context. Additionally, this suggestion is based on the literature reports Zhu et al, 2015 Nature Communications (https://doi.org/10.1038/ncomms7072) and Huang et al, 2019 J Exp Med (https://doi.org/10.1084/jem.20190173). It would be interesting to find out if IRF1 also plays a mechanistic role in gd T cells like CD8 T cells. So further experiments are needed in this context, which will complete the functional features of CD27+ gd T cells.
2. As stated in the manuscript, Ly6C is a marker for the myeloid compartment in mice. Also, the differentiation (as authors described "conversion") phenomenon of monocytes based on the Ly6C marker is already reported by Mildner et al 2017 Immunity (https://doi.org/10.1016/j.immuni.2017.04.018). In the case of monocytes, Mildner et al showed that Ly6C+ monocytes differentiate into Ly6C- cells. But, in the current manuscripts, authors are proposing the opposite differentiation trajectory for gd T cells. Also, in the corresponding analysis with human scRNA-seq data (fig. 1E-F), are there no analogous expression of genes in the monocyte cluster. Could you please explain? Sorry, this is a bit unclear to me!
3. I think, it is a bit strong statement that CD27+Ly6C- gd T-cell converts into CD27+Ly6C+ gd T cells. This statement needs to be supported and validated by more experiments:
   • (a) authors shall perform trajectory analysis in scRNA-seq data, which will, hopefully, support the conversion hypothesis;
   • (b) authors shall represent the purity and sorting strategy of flow cytometry experiments. For such experiments, it is very important to have the highest purity without any 'false negative' sorting!
   • (c) Usually, a conversion of one phenotype into another phenotype requires an intermediate stage. For example, Mildner et al 2017 Immunity reported three stages of monocytes (Ly6C+, Ly6Cint, Ly6C-). However, none of the flow cytometry plots in the current manuscript show any 'intermediate' population.
   • (d) check for protein expression of homing receptors like Ccr7, SELL, and S1pr1 to completely exclude that there is no 'homing'
   • (e) Can authors consider a third possibility that there is a loss/gain of Ly6C protein expression through molecule signal orchestrated by epigenetics mechanism? If yes, this possibility is worth testing!
   • (f) Last, but not least: One way to prove the conversion would be tagging Ly6C protein with a marker (e.g. GFP) to track its fate during the differentiation. Would it be possible?
4. The statement in line 137 is not entirely convincing as Fig. 1C-D shows some overlap of Ly6C and CCR7 gene expression in opposing clusters. At least, there are no clear 'black-and-white' opposing expressing levels. Sorry, this is my observation. Is there any other way of representation to make it clearer?
5. Authors have nicely shown the cytotoxic potential of CD27+Ly6C+ gd T cells in the in vivo setting. Would they observe the similar cytotoxic potential of expanded CD27+Ly6C+ gd T cells? Also, it would be great if memory CD8 T cells were added to the experimental set-up as a positive control, in addition to naïve CD8 as a negative control.

Minor:

1. Authors have used t-SNE representation. However, UMAP representation is preferred as UMAP preserves local and global structure in the data. Can authors show the similar segregation of clusters in UMAP representation?
2. Authors are suggested to add a supplementary figure on Ly6C2 and CCR7 expression in CD27- gd T cells. This is just a supporting evidence.
3. If I have understood the fig. 1E-F correctly, has the gene signatures mapped to the B-cell cluster as well, not only to naïve Vd1 and ab T cells?
4. Line 281-283: I think, it is redundant and can be moved to supplement.
5. Line 309: I am not sure if the authors are referring to Fig. 6F. It should be Fig. 6E, right?
6. In all figures, labels need to be precise. For example, figure labels say "% Ly6C+ cells". I think, the authors want to represent "% Ly6C+ CD27+ gd cells", right? Similarly, for other protein markers.
7. Fig. 3I needs to include a figure legend for color bars.

Significance

This study is a good example of single-cell genomic data can be used to find novel immune cell subsets and their functions. Such approaches can be combined with human immune cells. Thus, it brings direct significance and relevance to the basic understanding and translational approaches. However, such study needs in-depth examination of claims and observations.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

Wiesheu et al. studied the maturation and functional maintenance of IFNg-producing γδ T cells in a mouse tumor model and by using NGS methods. The study provides insights into underlying mechanisms in the differentiation from Ly6C- to Ly6C+ γδ cells and their anti-tumor function in tissues. The study identifies IL-27 as one key cytokine that maintains these cytotoxic Ly6C+ γδ T cells. They further compare transcriptomes of murine and human γδ T cells. The study is well performed, and provides new insights into the functional differentiation of gd T cells in the tumor response. I have few remarks as outlined below.

Major comments:

Comparison of murine and human γδ T cells: This is an important point and should be addressed in more detail. In Figure 1e-f, the transcriptomes of mouse cells are nicely mapped to human transcriptomes. I agree that there are similarities between naive Vδ1 T cells and αβ T cells, but in Figure 1f it appears that Vγ9Vδ2 T cells lie between the two clusters. Could the authors analyse this in more detail? It is well known that Vγ9Vδ2 T cells are more innate than other γδ T cell subsets (e.g. Vd1 cells). I suggest that this should also be taken into account in the comparison. How do these results relate to the in vitro assays on human γδ T cells in Figure 7? Second, do the authors see common pathways to IL-27 responsiveness of murine and human γδ T cells? Are the molecules described in Fig. 6H also upregulated in human γδ T cells? Similarities and differences could be better described/discussed.

Survival of Ly6C+ γδ T cells and implications for functionality: In the adoptive transfer experiments, tumour-bearing mice receive repeated injections of Ly6C+ γδ T cells. Would the system work with one transfer to further test survival or survival supported by IL-27? Do the transferred cells remain functional or would they be exhausted?

Minor comments:

In Figure 2A the CCR7 staining looks very weak despite the use of FMO. In my opinion, an isotype control or a positive control for CCR7 (e.g. expression on ab T cells) should be included. The Figure 1f is poor in resolution.

Significance

The study provides new insights into the functional differentiation and maintenance of murine γδ T cells in the tumor response, and compares this to human γδ T cell subsets at steady-state. A special focus was set on Ly6C, which has been rarely addressed in the past (please see references within the manuscript). A more in depth comparison of murine and human γδ T cells would be beneficial to address a broad readership and an ongoing debate on the translation of findings in murine systems to human immunology.

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

Learn more at Review Commons

---

Reply to the reviewers

1. General Statements [optional]

We are happy to receive the comments from the reviewers and grateful for their suggestions on how to improve our manuscript. We note that both reviewers find the work extensive and meaningful.

Based on the comments from the reviewers, we have performed a comprehensive set of additional experiments, which will result in one additional figure and a substantial restructuring of two figures with new data, considerably expanding both the preclinical as well as the mechanistic findings of our manuscript.

In short, reviewer 1 finds that we have done extensive work to understand the role of CDK12/CDK13 in glioblastoma and would like to see additional mechanistic details. Reviewer 2 recognizes the value of our work in exploring the potential usefulness of CDK12/13 inhibition in treatment of aggressive brain tumors and would like to see additional experiments, which demonstrate the efficacy of CDK12/13 inhibition in complex environments to reinforce our proof-of-concept.

To address this feedback, our response plan includes two lines of experiments, which will strengthen both the preclinical and mechanistic parts of our work:

1. A) We have established a migration assay using GSC G7 in organotypic mice brain slices and tested the effect of CDK12/CDK13 inhibition on glioma migration and we will include these data in the revised manuscript.
2. B) To further understand the mechanisms involved in the transcriptional inhibition following CDK12/CDK13 inhibition on DNA replication in glioma cells, we have performed the following additional experiments:
3. Comparative mass-spectrometry to identify changes in the total and phospho-proteome. This revealed that major regulators of DNA replication and repair are impaired following CDK12/CDK13 inhibition.
4. iPond (Identification of proteins on nascent DNA) assays that demonstrate that CDK12/CDK13 inhibition changes the composition of replication forks, with a strong reduction PCNA abundance early after treatment. PCNA tethers the DNA polymerase catalytic unit to the DNA template ensuring rapid and processive DNA synthesis. This reduction of PCNA occurs before EdU incorporation/DNA replication is reduced, suggesting that loss of DNA polymerase clamping and processivity explains the subsequent arrest of DNA replication.
5. DNA fiber assays showing that the origin firing is heavily downregulated in GSCs following CDK12/CDK13 inhibition. Further analyses using immunofluorescence microscopy reveal that the markers of DNA damage response and cell cycle progression are not affected following CDK12/CDK13 inhibition at early time-points, thereby ruling out activation of cell-cycle checkpoints and/or DNA damage response as potential explanation for replication block in GSCs following CDK12/CDK13 inhibition. The results from these experiments strengthen our main findings that inhibiting CDK12/CDK13 has a potential therapeutic value in glioblastoma treatment. Our work also offers mechanistic insights into how the glioblastoma stem cells have acquired transcriptional addiction to CDK12/CDK13 involving phosphorylation of RNAPII CTD, nascent RNA synthesis and DNA replication dependent on CDK12/CDK13 activity.

2. Description of the planned revisions

A point-by-point plan in blue is described below.

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

*The authors in this manuscript studied the role of a transcriptional cyclin-dependent kinase CDK12/CDK13 in glioblastoma. These cyclin-dependent kinases phosphorylate at ser2 residue in the C-terminal of RNA Pol II. Pharmacological inhibition of CDK12/CDK13 kinase with inhibitor decreases cell proliferation in multiple glioma cell lines and in patient-derived organoids. The CDK12/CDK13 inhibitor also reduces tumor growth in a mouse xenograft model. Mechanistically, the authors showed that genome-wide inhibition of CDK12/CDK13 attenuates RNA Pol II phosphorylation, disrupting transcriptional elongation and decreasing cell cycle progression. So, the authors proposed that targeting CDK12/CDK13 kinases can be used as a therapeutic strategy in glioblastoma. The authors have done extensive work in this manuscript to understand the role of CDK12/CDK13 in glioblastoma, but it is still a descriptive paper lacking mechanistic details.

*

RESPONSE: We appreciate the reviewer’s recognition of the extensive efforts behind this manuscript, and we are thankful for being pointed towards strengthening the mechanistic insights. In brief, we would like to corroborate our key findings that inhibition of CDK12/CDK13 abrogates RNAPII phosphorylation, nascent RNA synthesis and DNA replication. We have expanded the mechanistic characterization using the following experiments:

• Using DNA fiber assay, we find that origin firing is heavily downregulated in GSCs following CDK12/CDK13 inhibition. Furthermore, we have done in-depth characterization of the effect of THZ531 treatment on cell cycle regulators and DNA damage response in GSCs, and found that these were not affected by CDK12/CDK13 inhibition within six hours. This indicates that activation of a cell cycle checkpoint or DDR machinery was not the reason for replication block.
• To further characterize the rapid effect of CDK12/CDK13 inhibition, we have done comparative mass spectrometry following CDK12/CDK13 inhibition in GSCs to identify changes in total and phosphorylated proteins and identified major regulators of DNA replication and repair machinery that are strongly affected.
• We have implemented iPOND (identification of proteins on nascent DNA) to study the effect of CDK12/CDK13 inhibition on protein composition at the replication fork. On this basis, we find that the abundance of the DNA clamp PCNA is substantially reduced after two hours of THZ531 treatment. PCNA tethers the DNA polymerases together on the fork and adds processivity to the speed of DNA replication. EdU incorporation was not affected by two hours of THZ531 treatment, and loss of PCNA from the replication fork is a likely explanation for the DNA replication block observed after six hours of THZ531 treatment.

*Comments: 1. Figure 1 shows that CDK12/CDK13 inhibitor decreases cell viability, colony-forming ability, cell competition assay, and cell migration. The rationale behind choosing CDK12/CDK13 inhibitor in glioma is unclear from the manuscript. What is the CDK12/CDK13 expression in multiple glioma cells vs non-glioma cells? The authors should include normal astrocytes as a control for cell viability assay. The p value is missing in numerous Figure panels. *

RESPONSE: We have investigated the possibility of targeting transcriptional regulation in glioma cells by using inhibitors targeting transcriptional cyclin-dependent kinases which included CDK7, CDK9 and CDK12/CDK13.

• We found that glioma cell proliferation was most sensitive to CDK12/CDK13 inhibitors compared to other cancer cells (Figure 1A), whereas there was no specificity for CDK7 and CDK9 inhibitors on glioma cell proliferation compared to other cancer cells (Supplementary figure 1D). The selective inhibition of glioma cells by CDK12/CDK13 inhibitors was the rationale for choosing CDK12/CDK13 inhibitors for further studies. This is mentioned in the introduction, and the result section has been updated to reflect this.
• We have performed expression analyses of CDK12/CDK13 at the mRNA levels using RT-qPCR in the cell lines that are used in the study, and we did not find any correlation of CDK12/CDK13 expression in glioma versus non-glioma cells (Supplementary figure 1B). Thus, the propensity of cells to become addicted to CDK12/13 signaling for their survival seems not related to total transcript levels, but must rely on the function of CDK12/CDK13 as a selective regulator of transcriptional program required for glioblastoma proliferation.
• We will perform the cell viability assays on normal astrocytes.

• p-values will be added in the figure panels.

• Figure 2A shows the expression of CDK12 by immunohistochemistry in glioblastoma tissues. Including the non-glioma tissue samples as another control and including a quantification graph with the statistics is essential. In Figure 2B-D, the authors discussed the treatment of glioma patient-derived organoids with CDK12/CDK13 inhibitors. From the Figure, the organoids are resistant to THZ531 and SR-4835 inhibitors. To rule out this possibility, the immunoblot assay with cleave PARP will be essential to execute. Again, statistics need to be included in Figure 2C-D. *

RESPONSE: We want to point out that the immunohistochemistry for non-glioma tissue and additional controls are shown in Figure 2A, top right panel and supplementary Figure 2A.

Regarding the next statement, we do not think that there is any indication that the organoids (GBOs) are resistant to THZ531 and SR-4835. We would like to stress that data presented on Fig 2B-D shows the efficacy of THZ531, abemaciclib and SR-4835 inhibitors in GBOs. GBOs showed high resistance only to lomustine. We apologize for any part of the figure which may lack clarity and lead to potential misconceptions. We would very much like to improve on this, if we are able to identify which figure component that may give the impression that the organoids are resistant to THZ531 and SR-4835. One option would be to remove the 0 hr time point in Figure 2B, if that is the cause for misinterpretation. To emphasize the drug efficacy better, we plan to perform the following amendments to the revised manuscript:

• We will provide statistical analysis of the IC50 and AUC analysis in the supplementary table xxx. These analyses will further highlight the robustness of the evaluation of drug responses in comparison to lomustine.
• We will provide one-way Annova comparison of the efficacy of the four assessed drugs in Fig 3D.
• The cell viability assay applied in GBOs is based on the CellTiterGlow technology, which is applicable to small organoid cultures of
• The mouse subcutaneous xenograft experiment was carried out in U87 cells with CDK12/CDK13 inhibitors. However, the glioma stem cells are a more appropriate model for glioma biology, and it is not clear why authors suddenly chose U87 cells. Again, statistics are absent in multiple sub-panels. *

* *RESPONSE: We note reviewer’s acknowledgement of using GSCs as a more appropriate model for glioma biology and we want to emphasize that in this work, we have used 15 different glioma patient derived glioma cells (11 GSCs in Figure-1 and 4 GBOs in Figure-2) from two different research environments to show that CDK12/CDK13 inhibition compromises glioma proliferation in vitro. GSCs/GBOs used in our study are xenografted orthotopically in the brain to model glioma in vivo and since our drugs do not sufficiently cross the BBB, the GSCs/GBOs were not considered for the in vivo validation and instead, a subcutaneous xenograft model was best to assess the efficacy of the drug(s). Considering that these models require a high number of cells (eight million cells per xenograft were used in our experiment), we had to base our decision on feasibility and chose a type of cells that could be propagated to the required extent. Considering the reviewer’s criticism, we are open to moving the xenograft data are presented to the supplementary section. Appropriate statistics will be done and shown.

• The authors have performed CUT & RUN experiments in G7 cells with CDK12/CDK13 inhibitors and decided to use 1hr and 6hr time points for the assay. Although the inhibitor THZ531 is supposed to inhibit RNA Pol II phosphorylation at the Ser2 residue, it decreases the Pol II phosphorylation at the ser5 residue quite a bit. Therefore, it is crucial to determine the effect coming from ser2 vs ser5 phosphorylation and gene expression regulation. **

*

RESPONSE: This is a good point. To address the relationship further, we will perform quantitation of Ser2 and Ser5 signals as well as the changes in these over time. We will then correlate this to the transcriptional changes to assess which of the relationships that are most strongly correlated. In addition, we will perform non-parametric statistical testing of significance of ranked data.

• There are a lot of supplementary Figures where axes are not labeled correctly or missing. **

*

RESPONSE: This will be addressed.

• The statistical section needs to be included in the manuscript. **

*

RESPONSE: This will be included.

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

In this manuscript, the authors studied the role of CDK12/CDK13 in glioblastoma and performed extensive studies to uncover the importance of these kinases in glioblastoma. Understanding more mechanistic details of how these kinases are involved in glioma progression will uncover more therapeutic opportunities in glioblastoma.

*

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

*

*Summary: ** Lier et al. present a set of results showing that pharmacological inhibition of CDK12/13, cyclin-dependent kinases that phosphorylate RNA polymerase II (RNAPII), alters the proliferative behavior and transcriptional program of glioblastoma cells. A set of 2D and 3D cultures of patient-derived cell lines with stem-like properties (GSC), as well as subcutaneous xenografts of the U87 cell line, were used as in vitro and in vivo models, respectively. Among the CDKs tested, only CDK13 expression was found to be associated with worse patient survival, while CDK12-immunoreactive cells were detected in patient glioblastoma tissues. The response of GSCs to the CDK12 and CDK13 inhibitor TZH541 included cell cycle blockade and decreased migration. Reduction in RNAPII phosphorylation in TZH541-treated cells was verified using one of the GSC lines. Genome-wide exploration of the transcriptional consequences of TZH541 treatment of 2 GSCs using CUT&RUN and SLAM-seq technologies revealed major transcriptional repression, particularly of genes associated with cell proliferation. *

*Main comments: ** Although I found this study very interesting, I noted points requiring clarification, particularly in order to fully support the authors conclusions. My recommendations focus on the glioblastoma cell biology experiments, my area of expertise.

*

RESPONSE: We are grateful for the reviewer's keen interest in our manuscript and appreciate various insightful observations on the challenges within glioblastoma biology. Recognizing the necessity of validating CDK12/CDK13 requirements in complex environments, we have undertaken a migration assay using GSC, G7 cells in organotypic mice brain slices. The ongoing assessment of CDK12/CDK13 inhibition on glioma migration will be included in the revised manuscript. We have also more carefully explained how the organoid models used in this study address the requested need to recapitulate the complexity seen in the patient tissue and tumor environment. Moreover, we have related immunohistochemistry assessments of CDK12 levels to the proliferation marker Ki-67. Finally, we have strengthened the mechanistic insights provided in the manuscript by the inclusion of new proteomics data, iPond data on nascent chromatin, and chromatin fiber assays, altogether showing that replication origins firing as well as PCNA function is heavily reduced and identifying key proteins in DNA replication that are affected. These points are thoroughly discussed and explained in the comments below.

1. • The rationale for studying only CDK12 expression in patient glioblastoma tissues needs clarification. In contrast with CDK13, the authors found no association between CDK12 expression levels and patient survival (Sup Fig. 1A). Do the authors obtain similar results using independent datasets of glioblastoma tissue transcriptomes (e.g. CGGA)? With regard to the major effect of CDK12/13 inhibition on glioblastoma cell proliferation, determining whether CDK12/13 expression is observed in proliferating areas of the patients' tumor tissues (Ki67 IHC) would help support the authors' conclusion that their "results provide proof-of-concept for the potential of CDK12 and CDK13 as therapeutic targets for glioblastoma". The main data regarding CDK expression the status in patients' tumors and their possible association with patient survival should be rearranged in the same figure and described in the same paragraph of the results. * RESPONSE: We have performed our analyses on CGGA dataset, which matches with the TCGA data. We will show analyses from both TCGA and CGGA in Sup Fig. 1.

CDK12 and CDK13 are functionally redundant, which is one of the reasons that they do not score in genome-wide CRISPR/Cas9 dropout screens. As a result, GSC proliferation is only partially dependent on the individual expression of CDK12 and CDK13, as we observe in Figure 1E. However, GSCs are dependent on the combined CDK12/CDK13 activity and therefore are sensitive to inhibitors targeting both. Possibly, this functional redundancy makes the interpretation of the relationship between the individual expression of CDK12/CDK13 and glioma patient survival less straightforward.

With regards to the immunohistochemistry (IHC) staining evaluating the expression of CDK12 and CDK13 in glioma patient samples, we tested several antibodies for both CDK12 and CDK13. However, we were only able to identify an antibody for CDK12 which worked reliably in IHC.

We will perform Ki-67 IHC to test whether CDK12 expression matches with proliferative areas of the tumor tissues.

• Fig.1 caption "Inhibition of CDK12/13 specifically affects proliferation of glioma cells" is not entirely consistent with the results. This inhibition also appears to induce cell death, at least in some of the GSC tested, as indicated with cell counts (Fig. 1C., sup Fig.1 G) and an 8-fold increase in the % of apoptotic cells after a 24h-TZH treatment shown in Fig. 5E. All data concerning the effects of TZH on proliferation and survival (including detailed effects on the cell cycle) should be brought together rather than split between the 1st and last figure. *

RESPONSE: We appreciate these comments and will be addressed it in the manuscript.

*3. The reason for which serum-treated GSC were used should be explicated (sup Fig. 1C). Serum being usually used to trigger GSC "differentiation", did the authors want to verify whether CDK12/13 inhibitors affected GSC in a specific manner? If yes, it is necessary to demonstrate that serum-treated GSC have lost their stem-like properties. *

RESPONSE: This is a good point that we appreciate being able to expound on. GSCs are grown in serum-free media with N2 and B27 supplements together with EGF/FGFb whereas the control cells, including breast cancer and Hela/U2OS cells are grown in media containing serum. Serum-containing media was used to assess whether the diverse set of macromolecules present in serum would affect the bioavailability and/or response to the drug, and our data clearly demonstrated that this was not the case and that glioma stem cells are susceptible to the drug regardless of serum presence. In order to minimize the effect of serum on GSC differentiation, serum was added in the media immediately before the drug treatment.

• The viability of patient-derived 3D organoids (GBO) was assessed by measuring ATP production. It is therefore not possible to distinguish between decreased cell proliferation and increased cell death as responsible for the signal decrease. This limitation in the interpretation of the results needs to be made explicit. I was also misled by the use of GBO. This abbreviation is currently used to designate fragments of patient tumor tissue amplified in culture, which retain the cellular heterogeneity and the extracellular matrix of the original tumor and therefore provide an actual ex vivo model of the tumor. To avoid any misunderstanding, I recommend referring to experimental models obtained from dissociated patient-derived cell lines as "3D organoids" or "cellular spheroids", and avoiding to designate them as ex vivo models since they do not recapitulate the complexity of the tumor. *

RESPONSE: We apologize for providing insufficient details concerning our GBO modelling, and we have now updated the description in the methods to avoid misconceptions and unclarity. Our GBOs are not derived from cell lines. We derive GBOs from patient tumors by short-term culture of tissue fragments in 3D conditions. Such organoids are of a very primary nature and contain extracellular matrix and tumor microenvironment components. To avoid propagation in vitro, we perform implantation of GBOs to immunodeficient animals to create patient-derived orthotopic xenografts (PDOXs). We have established that serial propagation of patient material via series of short-term GBO cultures and PDOXs allow for multiplication of GBM patient tumors without major clonal selection and genetic/phenotypic adaptation (Golebiewska, 2020, DOI: 10.1007/s00401-020-02226-7). To perform robust drug screening ex vivo in GBOs, we further developed a specific protocol based on the material isolated directly from well-established and characterized PDOXs (Oudin, 2021, DOI: 10.1016/j.xpro.2021.100534). The protocol includes reconstitution of 3D GBOs of uniform size, which allows for reliable ex vivo readouts. Importantly, GBM primary cells are able to reassemble into 3D structures of heterogeneous nature, including reconstitution of extracellular matrix. In the revised manuscript, we will provide a clear description of the GBO modelling in the material and methods as well as in the associated results.

• Although the abstract contains a statement indicating that CDK12/13 genetic ablation inhibits cell migration, I did not find the corresponding results in the article. The demonstration that CDK12/13 inhibition decreases cell migration is weaker than the demonstration of its effect on proliferation. Contrary to the experiments evaluating cell proliferation, cell migration was assessed using a single technical approach. Moreover, the method used to assay TZH effects on cell migration rather measures cell motility than cell migration over long distances in a 3D and complex environment as observed in diffuse glioma. Since these data add nothing significant to the article, I would delete them. *

RESPONSE: We thank the reviewer for pointing out the comment in first sentence, which is addressed in the abstract now.

It is correct that strictly speaking our assay measured the effect of CDK12/CDK13 inhibition on glioma motility rather than migration, we have corrected this sentence in the abstract. We have however also now strengthened the methodology in the manuscript by establishing and using migration assays of GSC G7 cells on organotypic mouse brain slices. Organotypic mouse brain slices have a preserved cytoarchitecture that allows analysis of migration over longer distances in a physiological environment. We are currently analyzing the data. These results will be included in the revised manuscript.

• In my opinion, the information from the in vivo experiments is limited and should be presented in a supplementary rather than a main figure. The data were obtained with a single cell model, U87 cells of uncertain origin, and using subcutaneous xenografts that provide an environment totally different from the patient's actual tumor. In this context, the data obtained provide little information on the response of cancer cells in a complex and specific environment well known to promote tumor growth and resistance to therapies. I understand that the use of intracerebral xenografts is not feasible, since the inhibitor does not appear to reach the brain. With this technical limitation, an alternative would be to deliver the compound directly inside the brain tumor. A cannula can be implanted into the tumor after it has formed, and connected to an Alzet minipump filled with the drug. These experiments are technically difficult, however, and success is not guaranteed. Another alternative would be to use GBO, as described by Jacob et al (2019) as a surrogate for tumor tissue, provided the authors can obtain tissue fragments from patient surgical resections or intracerebral xenografts of patient-derived cell lines. These alternatives are optional. *

RESPONSE: We thank the reviewer for pointing out the difficulties in testing currently available compounds in vivo. Following the reviewers’ comments, we are open to placing the in vivo experiments in U87 xenografts in the supplementary material. We would like to reemphasize the clinical significance of our data in GBOs (please see the response above), which relies on models of equal complexity compared to the Jacob’s protocol and represent 3D compact and complex structures ex vivo derived from the GBM patient tumors propagated as orthotopic patient-derived xenografts.

Minor comments: ** - Fig. 4A and Fig. 5E-F: Results from a single experiment? If yes, they must be repeated at least once.

RESPONSE: They are representative of a minimum of three independent biological experiments, which will be mentioned in the manuscript.

*- For the sake of clarity, all y-axes in graphs presenting MTT or CellTiter-Glo assay results should be labeled "cell viability index", as they only provide a measure of overall cell or organoid metabolic activity, and thus an indirect assessment of cell viability. *

RESPONSE: We thank the reviewer for this suggestion and will incorporate it in the revision.

*- Statistical analyses are missing for 3 of the 4 cell lines presented in Figure 1F. *

RESPONSE: This will be addressed.

*- Some GO terms are truncated in sup Fig. 3. *

* *RESPONSE: This will be fixed in the revised ms.

- The legend to Fig. 5B-D shows the mean and SD of 2 replicates. Please show individual points.

RESPONSE: This suggestion will be addressed in the revision.

- Sup Fig1 D-F: unit of concentration is missing (M?) ** RESPONSE: This is addressed.

*Reviewer #2 (Significance (Required)): **

Significance: Despite growing interest in the roles of CDK12/13 roles in cancers and their targeting for cancer therapy, their involvement in glioblastoma growth remains unexplored. The results presented in this study outline the potential of CDK12/13 inhibition in controlling the growth of glioblastoma, at least in vitro, and thus provide meaningful information on its potential usefulness for this aggressive brain tumor with a high proliferation rate. Obtaining the full proof-of-concept that CDK12/13 constitute relevant targets for glioblastoma therapies will however require additional experiments demonstrating efficacy of CDK12/13 inhibition in complex environments, as encountered in the patients' tumor. *

3. Description of the revisions that have already been incorporated in the transferred manuscript

We have addressed following of the reviewers’ comments.

Reviewer-1:

• Major comment-1 is partially incorporated in the text.

• Major comments-5 and 6 are incorporated. Reviewer-2:

• Major comment 1 is partially addressed.

• Major comment 2, 3 and 4 are addressed in writing.
• Major comment 5 is partially addressed in writing.
• Major comment 6 is addressed.
• All minor comments are incorporated in writing.

4. Description of analyses that authors prefer not to carry out

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Summary:

Lier et al. present a set of results showing that pharmacological inhibition of CDK12/13, cyclin-dependent kinases that phosphorylate RNA polymerase II (RNAPII), alters the proliferative behavior and transcriptional program of glioblastoma cells. A set of 2D and 3D cultures of patient-derived cell lines with stem-like properties (GSC), as well as subcutaneous xenografts of the U87 cell line, were used as in vitro and in vivo models, respectively. Among the CDKs tested, only CDK13 expression was found to be associated with worse patient survival, while CDK12-immunoreactive cells were detected in patient glioblastoma tissues. The response of GSCs to the CDK12 and CDK13 inhibitor TZH541 included cell cycle blockade and decreased migration. Reduction in RNAPII phosphorylation in TZH541-treated cells was verified using one of the GSC lines. Genome-wide exploration of the transcriptional consequences of TZH541 treatment of 2 GSCs using CUT&RUN and SLAM-seq technologies revealed major transcriptional repression, particularly of genes associated with cell proliferation.

Main comments:

Although I found this study very interesting, I noted points requiring clarification, particularly in order to fully support the authors conclusions. My recommendations focus on the glioblastoma cell biology experiments, my area of expertise.

• The rationale for studying only CDK12 expression in patient glioblastoma tissues needs clarification. In contrast with CDK13, the authors found no association between CDK12 expression levels and patient survival (Sup Fig. 1A). Do the authors obtain similar results using independent datasets of glioblastoma tissue transcriptomes (e.g. CGGA)? With regard to the major effect of CDK12/13 inhibition on glioblastoma cell proliferation, determining whether CDK12/13 expression is observed in proliferating areas of the patients' tumor tissues (Ki67 IHC) would help support the authors' conclusion that their "results provide proof-of-concept for the potential of CDK12 and CDK13 as therapeutic targets for glioblastoma". The main data regarding CDK expression the status in patients' tumors and their possible association with patient survival should be rearranged in the same figure and described in the same paragraph of the results.
• Fig.1 caption "Inhibition of CDK12/13 specifically affects proliferation of glioma cells" is not entirely consistent with the results. This inhibition also appears to induce cell death, at least in some of the GSC tested, as indicated with cell counts (Fig. 1C., sup Fig.1 G) and an 8-fold increase in the % of apoptotic cells after a 24h-TZH treatment shown in Fig. 5E. All data concerning the effects of TZH on proliferation and survival (including detailed effects on the cell cycle) should be brought together rather than split between the 1st and last figure.
• The reason for which serum-treated GSC were used should be explicated (sup Fig. 1C). Serum being usually used to trigger GSC "differentiation", did the authors want to verify whether CDK12/13 inhibitors affected GSC in a specific manner? If yes, it is necessary to demonstrate that serum-treated GSC have lost their stem-like properties.
• The viability of patient-derived 3D organoids (GBO) was assessed by measuring ATP production. It is therefore not possible to distinguish between decreased cell proliferation and increased cell death as responsible for the signal decrease. This limitation in the interpretation of the results needs to be made explicit. I was also misled by the use of GBO. This abbreviation is currently used to designate fragments of patient tumor tissue amplified in culture, which retain the cellular heterogeneity and the extracellular matrix of the original tumor and therefore provide an actual ex vivo model of the tumor. To avoid any misunderstanding, I recommend referring to experimental models obtained from dissociated patient-derived cell lines as "3D organoids" or "cellular spheroids", and avoiding to designate them as ex vivo models since they do not recapitulate the complexity of the tumor.
• Although the abstract contains a statement indicating that CDK12/13 genetic ablation inhibits cell migration, I did not find the corresponding results in the article. The demonstration that CDK12/13 inhibition decreases cell migration is weaker than the demonstration of its effect on proliferation. Contrary to the experiments evaluating cell proliferation, cell migration was assessed using a single technical approach. Moreover, the method used to assay TZH effects on cell migration rather measures cell motility than cell migration over long distances in a 3D and complex environment as observed in diffuse glioma. Since these data add nothing significant to the article, I would delete them.
• In my opinion, the information from the in vivo experiments is limited and should be presented in a supplementary rather than a main figure. The data were obtained with a single cell model, U87 cells of uncertain origin, and using subcutaneous xenografts that provide an environment totally different from the patient's actual tumor. In this context, the data obtained provide little information on the response of cancer cells in a complex and specific environment well known to promote tumor growth and resistance to therapies. I understand that the use of intracerebral xenografts is not feasible, since the inhibitor does not appear to reach the brain. With this technical limitation, an alternative would be to deliver the compound directly inside the brain tumor. A cannula can be implanted into the tumor after it has formed, and connected to an Alzet minipump filled with the drug. These experiments are technically difficult, however, and success is not guaranteed. Another alternative would be to use GBO, as described by Jacob et al (2019) as a surrogate for tumor tissue, provided the authors can obtain tissue fragments from patient surgical resections or intracerebral xenografts of patient-derived cell lines. These alternatives are optional.

Minor comments:

• Fig. 4A and Fig. 5E-F: Results from a single experiment? If yes, they must be repeated at least once.
• For the sake of clarity, all y-axes in graphs presenting MTT or CellTiter-Glo assay results should be labeled "cell viability index", as they only provide a measure of overall cell or organoid metabolic activity, and thus an indirect assessment of cell viability.
• Statistical analyses are missing for 3 of the 4 cell lines presented in Figure 1F.
• Some GO terms are truncated in sup Fig. 3.
• The legend to Fig. 5B-D shows the mean and SD of 2 replicates. Please show individual points.
• Sup Fig1 D-F: unit of concentration is missing (M?)

Significance

Despite growing interest in the roles of CDK12/13 roles in cancers and their targeting for cancer therapy, their involvement in glioblastoma growth remains unexplored. The results presented in this study outline the potential of CDK12/13 inhibition in controlling the growth of glioblastoma, at least in vitro, and thus provide meaningful information on its potential usefulness for this aggressive brain tumor with a high proliferation rate. Obtaining the full proof-of-concept that CDK12/13 constitute relevant targets for glioblastoma therapies will however require additional experiments demonstrating efficacy of CDK12/13 inhibition in complex environments, as encountered in the patients' tumor.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

The authors in this manuscript studied the role of a transcriptional cyclin-dependent kinase CDK12/CDK13 in glioblastoma. These cyclin-dependent kinases phosphorylate at ser2 residue in the C-terminal of RNA Pol II. Pharmacological inhibition of CDK12/CDK13 kinase with inhibitor decreases cell proliferation in multiple glioma cell lines and in patient-derived organoids. The CDK12/CDK13 inhibitor also reduces tumor growth in a mouse xenograft model. Mechanistically, the authors showed that genome-wide inhibition of CDK12/CDK13 attenuates RNA Pol II phosphorylation, disrupting transcriptional elongation and decreasing cell cycle progression. So, the authors proposed that targeting CDK12/CDK13 kinases can be used as a therapeutic strategy in glioblastoma. The authors have done extensive work in this manuscript to understand the role of CDK12/CDK13 in glioblastoma, but it is still a descriptive paper lacking mechanistic details.

Comments:

1. Figure 1 shows that CDK12/CDK13 inhibitor decreases cell viability, colony-forming ability, cell competition assay, and cell migration. The rationale behind choosing CDK12/CDK13 inhibitor in glioma is unclear from the manuscript. What is the CDK12/CDK13 expression in multiple glioma cells vs non-glioma cells? The authors should include normal astrocytes as a control for cell viability assay. The p value is missing in numerous Figure panels.
2. Figure 2A shows the expression of CDK12 by immunohistochemistry in glioblastoma tissues. Including the non-glioma tissue samples as another control and including a quantification graph with the statistics is essential. In Figure 2B-D, the authors discussed the treatment of glioma patient-derived organoids with CDK12/CDK13 inhibitors. From the Figure, the organoids are resistant to THZ531 and SR-4835 inhibitors. To rule out this possibility, the immunoblot assay with cleave PARP will be essential to execute. Again, statistics need to be included in Figure 2C-D.
3. The mouse subcutaneous xenograft experiment was carried out in U87 cells with CDK12/CDK13 inhibitors. However, the glioma stem cells are a more appropriate model for glioma biology, and it is not clear why authors suddenly chose U87 cells. Again, statistics are absent in multiple sub-panels.
4. The authors have performed CUT & RUN experiments in G7 cells with CDK12/CDK13 inhibitors and decided to use 1hr and 6hr time points for the assay. Although the inhibitor THZ531 is supposed to inhibit RNA Pol II phosphorylation at the Ser2 residue, it decreases the Pol II phosphorylation at the ser5 residue quite a bit. Therefore, it is crucial to determine the effect coming from ser2 vs ser5 phosphorylation and gene expression regulation.
5. There are a lot of supplementary Figures where axes are not labeled correctly or missing.
6. The statistical section needs to be included in the manuscript.

Significance

In this manuscript, the authors studied the role of CDK12/CDK13 in glioblastoma and performed extensive studies to uncover the importance of these kinases in glioblastoma. Understanding more mechanistic details of how these kinases are involved in glioma progression will uncover more therapeutic opportunities in glioblastoma.

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

Learn more at Review Commons

---

Reply to the reviewers

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

---

The authors Martiěnez-Balsalobre and colleagues found that the regenerative capacity of the zebrafish caudal fin is not limited by the lack of telomerase and showed that the length of telomeres does not decrease substantially after repeated amputations in telomerase-deficient zebrafish. These findings prompt the authors to explore an alternative mechanism that would explain the maintenance of telomere length in this regeneration setting. They produced suggestive evidence for the role of the ALT (Alternative Lengthening of Telomeres) mechanism in the maintenance of telomere length in the absence of telomerase in a regeneration setting.

In my view, several points need to be addressed and clarified.

**There are three major points:**

1.When working with tert mutants, the age at which these fish show a telomere phenotype (namely, loss of body mass and reduced fertility) varies. Therefore, it would be important to state if the fish used in this study were already showing these phenotypic characteristics at each time point studied, namely 4, 8 and 11 months of age

The premature aging phenotype of tert mutant fish has been previously characterized in the paper by Anchelin et al 2013 referenced in the manuscript. We used young fish with no phenotype (4 months old), and aged fish (8 and 11 months old) presenting the already described premature aging phenotypes, such as spinal curvature, loss of fertility, loss of body mass and loss of pigmentation.

The following sentence regarding this has been included in the revised version of the manuscript.

“The fish used showed non-detectable aging phenotype at 4 months old, whereas at 8- and 11-months fish presented the typical tert mutant premature aging phenotypes, i.e. backbone curvature, loss of body mass and hypopigmentation”

2.The knockdown experiments were performed using morpholinos. To confidently use morpholinos it is fundamental to demonstrate first their knockdown efficiency and their specificity. This is lacking in the manuscript.

In this work we have used 3 different morpholinos; tert morpholino has been already used and characterized in the work by Imamura and collaborators in 2008. atr morpholino has been already used and characterized in the paper by Stern at al., 2005.

However, nbs1 morpholino has been designed for this work. A Supplemental Figure (Figure S2) and the following paragraph have been added in the revised version of the manuscript to show the knock-down efficiency of the nbs1 morpholino:

“The knock-down efficiency of the atr morpholino was characterized by Stern and colleagues (Stern et al, 2005). The injection of the nbs1 morpholino in zebrafish eggs resulted in the reduction of the expression of nbs1 mRNA at 3dpf (Fig. S2A). Furthermore, PCR using cDNA as a template detected nbs1 mRNA species that retained the intron one of the gene as a result of the morpholino effect in blocking the splicing (Fig. S2B).

“The tert morpholino knock-down efficiency has been already showed (Imamura et al, 2008)”

3.The involvement of ALT mechanism in the regeneration process in the absence of telomerase is only suggestive, as the authors show an increase of C-circles and heterogenous telomerase length in telomerase-deficient zebrafish but when trying to establish a functional link the authors resort to the knowndown of genes that may be associated with ATL. Looking at the levels of TERRA and the number of C-circles in the knowndown caudal fins would be essential for their claim.

We have now performed caudal fin regeneration experiments in tert mutant fish microinjected with mo-atr and mo-nbs1 and analyzed the levels of TERRA RNAs and C-circles amount. The results are shown in Supplemental Figure S4. As expected, regeneration capacity decreased in fish microionjected with both morpholinos compared to control fish (FigS4 F). Consistently, TERRA RNAs levels, as well as C-circles amount, increased in the regenerating tissue and this induction was lower when atr and nbs1 gene expression was downregulated by mo-injection (Fig S4 G-J). Taking altogether, these results indicate that ALT mechanism is induced upon amputation and operates in the regenerating tissue of tert deficient fish.

**And several other points:**

4.The regeneration experiments were performed at 32 degrees and this option was never explained nor discussed.

The regeneration experiments in zebrafish typically are performed at 32 °C to accelerate regeneration process. Otherwise, the amount of regenerated blastema at 48 hpa or 72hpa would not be enough to perform any kind of analysis. Furthermore, it could happen that some experimental modifications, for instance the effects of the morpholino injection, do not last if the regeneration process is kept more than 84-96hpa at 28 °C.

This procedure have been used previously by other laboratories (PMID: 8601496, Johnson and Weston,1995; PMID: 12015289 Nechiporuk et al.,2003 and PMID: 16273523 Thumnel et al 2006) to increase the rate of regeneration approximately two fold, a temperature of 33°C was used for the regeneration experiments. In addition, It has been demonstrated normal regeneration at 33°C in wild-type fish

5.When referring to the ALT mechanism, the authors state that "... in about 10% of tumors cells, telomere length is maintained by the Alternative Lengthening of Telomeres (ALT) mechanism ..." and I think it would be more accurate to talk about cancer cells instead of tumor cells.

This has been corrected in the revised version

6.The sentence about C-circles is incorrect. C-circles are mostly single-stranded and not double-stranded as stated.

This has been corrected in the revised version

7.After Figure 2, the authors never mention the age of the fish used.

All the fish used in the amputation experiments after Fig2 are 4 -6 months of age

8.In Figure 1A. The site of amputation does not fit the one described in Mat & Met that states 2 cm from the base of the caudal peduncle. The same stands for Figure 2A.

This is corrected in the new version with a new Figure 1A and 2A

9.In Figure 1B

The Y axis should be named regeneration area instead of rate as the values are a percentage of the area reached after a certain time point after amputation. The same stands for Figure 2B, C. It would be nice to see the real caudal fin images for the relevant time points: before amputation, 0 dpa, when the fins reach 50% of regeneration area and then the last time point.

This has been changed in the new version

The authors should discuss why are the caudal fins reaching more than 100% of regeneration are

This is an intriguing question for which we currently lack an answer. Nonetheless, it does not impact the focus of our ongoing study

10.In Figure 2B. The meaning of ". .. ." on the right side of the graph is not clear. The same stands for Figure 2D.

This has been a mistake when handling the figure folder and has been corrected in the revised version

11.In Figure 2C .Why is the clip 10, 11 and 12 missing from the tert+/- and tert-/- ?

This has been changed in the new version and recalculated the statistical significance. We appreciate the feedback

12.In Figure 2E The proximity of all points at the 12 Clip is indicative of lack of statistical significance, therefore the **** related to which comparisons?

We have modified the data of fig 2E and recalculated the statistical significance

13.In Figure 2D, E

For the measurement of telomere length, the authors state that "Data are average of at least 2 independent experiments." What does this mean exactly? How many animals were used in each experiment?

In the experiments in Fig2, 6 fish total were used per group sampled in at least 2 independent experiments. This has been included in the figure legend and in the Mat&Met section

14.In Figure 3

The authors state that "Data are average of at least 2 independent experiments." What does this mean exactly? How many animals were used in each experiment?

The experiment in Fig 3A was done 3 times with 2 fish per group pooled in each experiment. The telomere length experiment has been done 2 times. This has been added to the figure legend and to the Mat&Met section.

Why were the c-circles evaluated at hpa while the telomere length evaluated at dpa? This should be discussed.

We expect to observe an effect on telomere length after several days of continuous cell proliferation in order to completely regenerate the caudal fin. However, the presence of C-circles in the regenerating tissue is expected to be found as early as 24hpa as a consequence of the action of the ALT mechanism of telomere maintenance, which has to be active from the very beginning. The following sentence has been included in the Discussion section: “ALT activation is expected to happen, and in fact detected, very early in the regeneration process, and eventually results in telomere length heterogenicity several days after amputation, when a lot of cell divisions and telomere recombination have occurred”.

15.In Figure 3A

The meaning of ". .. ." on the top side of the graph is not clear.

t0 should be removed and replaced by 0 hpa and 24hpa and 48hpa for coherence.

This has been a mistake when handling the figure folder and has been corrected in the revised version.

16.In Figure 3B,C 0 hpa replace by 0 dpa

This has been replaced in the new version

17.In Figure 3B

The blue and red stainings in the panels are labelling exactly what? This should be stated in the image and in the legend.

Red staining represents the telomeres and the blue staining are the nuclei. It is shown in the Figure and stated in the figure legend.

18.In Figure 3D

There is a mistake in the legend the should be corrected as follows "Very long telomeres have a higher fluorescence of 200,000 AUF and very short telomeres have a lower fluorescence of 30,000 AUF."

This has been corrected

19.In Figure 4

t0 should be removed and replaced by 0 hpa.

This has been corrected

The meaning of ". .. ." on the top side of the graph is not clear.

This has been a mistake when handling the figure folder and has been corrected in the revised version

The title is an overstatement, as the genes studied are DNA damage genes that may associate with ALT.

The title has been corrected to “The expression of ALT-associated genes is modulated in regenerative tissue of”

20.In Figure 4A, B

The expression of nbs1 and atr in tert-/- increases at 48hpa but the same seems to be true for the tert+/+ and this is never discussed by the authors.

This result would support the idea that both telomerase-dependent and ALT mechanisms operate in the regeneration process in a wild type animal. A sentence in the results and discussion sections has been added to mention and discuss this point:

“These genes were quantified in the regenerated tissue at 24 and 48 hpa. nbs1 and atr mRNA levels increased in telomerase deficient fish at 48 hpa compared to time 0 (0hpa) (Fig. 4A, 4B). The same effect in the expression of these genes was found in wild type fish regenerating fins. Interestingly, atrx and daxx expression decreased (Fig. 4C, 4D) at 24 and 48 hpa, in agreement with published data on ALT in cells (Amorim et al., 2016; Ren et al., 2018; Yost et al., 2019).”

“Curiously we observed an increased expression of ALT activator proteins in both wild type and telomerase deficient zebrafish, and a decrease in ALT inhibitor proteins suggesting that the main players of ALT and their mechanisms are conserved during evolution, and that both mechanisms of telomere maintenance could co-exist in the regeneration process in wild type fish”..

21.In Figure 4C, D

The differences in the expression of atrx and daxx decreases over time in a in tert-/- and this is never discussed by the authors.

As mentioned, and referenced in the manuscript, the proteins are ALT inhibitors, and mutations in these proteins are described to be promoting the activation of ALT mechanisms. Thus, it is expected that in the regenerating fins where ALT is activated, their expression decreases.

22.In Figure 5

An ideal control would be the direct comparison between microinjected+electroporated mo-std in the ventral part of the fin while the dorsal part would be microinjected+electroporated with the mo-gene of interest. This would discard any effect of microinjection+electroporation in the regeneration efficiency.

These experiments are not convincing to show that there is an ALT mechanism is operating here. What this experiment shows if the relevance of these genes for the regenerative capacity of the caudal fin. To show that this is related to the ALT mechanism the authors should investigate the C-circles in these regenerating fins.

We have performed regeneration experiments using WT fish to address this issue. We analyzed the regenerated area of control and morpholino injected fish and then obtained regenerating blastema and analyzed the expression of tert and atr. The results are shown in Supplemental Figure S4 (A-E). The regeneration capacity is inhibited in tissues injected with a mix of mo-std+mo-ter, a mix of mo-std+mo-atr, or a mix of mo-tert+mo-atr compared with a control injected with a double dosis of mo-std (std 2x, Fig S4B). In addition, the expression of tert and atr is decreased in the regenerated blastema upon morpholino injection (Fig S4 C and D) indicating that the genetic inhibition of the expression of these genes was efficient. Finally, the levels of TERRA RNAs are increased upon amputation and this induction is reduced when we mo-atr or a combination of mo-atr+tert were microinjected (Fig S4E).

We have also performed caudal fin regeneration experiments in tert mutant fish microinjected with mo-atr and mo-nbs1 and analyzed the levels of TERRA RNAs and C-circles amount. The results are shown in Supplemental Figure S4. As expected, regeneration capacity of the caudal fins of fish microionjected with both morpholinos decreased compared to control fish (Fig S4 F-H). Consistently, TERRA RNAs levels, as well as C-circles amount, increased in the regenerating tissue and this induction was lower when atr and nbs1 gene expression was decreased by mo-injection (Fig S4 I and J). Taking altogether, these results indicate that ALT mechanism is induced upon an injury and operating in the regenerating tissue of both wild type and tert deficient fish.

The amputation red lines are not placed in the exact amputation position in some of the panels.

Regeneration rate should be regeneration area.

This has been corrected

23.In Figure 5C, E

Why is the mo-tert more inhibitory of regeneration (Figure 5E - around 30%) than the tert-/- mutant (Figure 5C - around 60%)? This should be discussed.

This point is now discussed: “

24.In Figure 6A

The 2 adult zebrafish shown in the tank with the ATR inhibitor IV should have an amputated caudal fin.

This has been modified

Control is exactly what? Untreated? Treated with vehicle?

The control is fish treated with the same amount of DMSO (vehicle). This is now shown in the panel

Why was the ATR inhibitor IY added immediately after fin amputation while the mo-atr was injected at 48 hpa?

The ATR inhibitor was added immediately after amputation because ALT is then inhibited from the starting of the regeneration process. However, in the case of the atr morpholino we need some regenerated tissue to perform the microinjection within and inhibit atr expression specifically in this tissue.

25.Figure 6D, E, F

These panels are a bit out of the focus of this paper. If presented should go to a supplementary figure.

These panels are now moved to the Supplemental Figure S5

26.In Figure S2

The relevant bands should be identified.

We have performed new regeneration experiments in wild type adult fish using ATR inhibitor. The results show that treating fish with ATR inhibitor provokes a clear decrease in the overall phosphorylation status of ATR/ATR substrates within the regenerated tissue (Figure 5B and C). In this case, the intensity of the whole lane was used for quantification.

The gel identifies DMSO, 10uM and 50 uM but the quantification graph identifies Control, 50uM and 100uM.

This has been corrected in the new version

There are no error bars

In the new experiments are now shown.

The authors say that the quantification of various western blot bands was done but how many exactly?

In the new experiments, 3 western blots are quantified

27.In Figure S3

The primers for rps11 are repeated twice.Were these primers design de novo by the authors or did they used previous reported primers, in this case the references should be given.

Tert F2 and R1 should be replaced for F and R for consistency.

This has been corrected and references for the primers used are added in the new Supplemental Figure S6

28.In Figure S4

The sequence of tert mo is missing.

This has been corrected

29.In the methods the genotyping protocol of tert mutants is not described.

A protocol for genotyping the tert deficient zebrafisn has been added in the Mat&Met section.

30.The method to calculate the area of the fin pre- and post-amputation is not described.

The method is already described in the Mat&Met section: “In order to calculate the percentage area of growth between the injected and non-injected part, the values were inserted in the following formula: (Dorsal 48 hpi - Dorsal 0 hpi)/(Ventral 48 hpi -Ventral 0 hpi)*100, where Dorsal is the regenerative area of the MO-treated tissue and Ventral is the regenerative area of the corresponding uninjected half”

Reviewer #1 (Significance (Required)):

---

The manuscript by Martiěnez-Balsalobre and colleagues deals with a very interesting question on the importance of telomere lengthening during regenerative processes and its relation to ageing. To this end the authors made use of the tert mutant, a telomerase-deficient zebrafish. The authors show a surprising phenotype that telomerase-deficient zebrafish can still regenerate their caudal fins and are able to maintain telomere length during consecutive amputations and I say surprising because it has been shown that telomerase-deficient zebrafish are unable to regenerate their hearts efficiently.

Taking these novel findings, the authors propose that in the zebrafish caudal fin and in the absence of telomerase, telomere length is maintained through the activation of an alternative mechanism called ALT. To my knowledge, the role of ALT as a mechanism of telomere lengthening has never been described in the context of regenerating organs in zebrafish.

We fully appreciate the reviewer´s comments on the significance of the manuscript!

**Referees cross-commenting**

I agree with the comments made by the other reviewers. I would stress the need to tone down the role of ALT during fin regeneration in zebrafish as all the experiments are only indicative of the possible of the involvement of ALT.

We have conducted additional experiments that further support the involvement of ALT. Please read the responses to the other reviewers for more details.

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

Using zebrafish as a model for regeneration, the authors find that telomere maintenance by recombination can occur in the absence of telomerase.

Title to Figure 4 perhaps may be too strong, 'ALT mechanism is activated', since only a few features of ALT are assessed. Perhaps, 'ALT features are activated'?

The title to Figure 4 has been changed to “The expression of ALT-related genes is modulated…”

mRNA levels of NBS, ATR are also increased in WT animals (Figure 4A and 4B), but ATRX and DAXX mRNA levels are not decreased in WT animals. Is the increase why the authors in part suggest that ALT is being used in WT animals. If so, what would be the trigger for the use of ALT, as opposed to the trigger to use ALT in tert-/- animals?

Our results indicate the utilization of both telomere maintenance mechanisms to support cell division in regenerative fins among wild-type animals. Consequently, we propose that the signals instigating regeneration are shared between both mechanisms and are present in both wild-type and tert-deficient animals, albeit with varying degrees of contribution.

In Figure 5C, if tert-/- animals are downregulated for nbs1 and atr, would it be expected that the effect on regeneration be more pronounced compared to tert+/+ downregulated for nbs1 and atr than what is observed?

We agree with the reviewer comment, and that is what actually happens. The inhibition of the regeneration in wild type fish is about 40% in mo-nbs1 injection and around 70% in mo-atr injected animals. However, in tert mutants, the decrease in regeneration observed in mo-nbs1 injection is about 56%, whereas is 82% in mo-atr injection.

What are the telomere lengths in tert-/- animals treated with mo-atr or mo-nbs1 or in tert+/+ animals treated with mo-tert and mo-atr compared to singly treated?

The telomere length does not change in mo-atr or mo-nbs1 injected tert mutants compared to mo-std animals.

The telomere length in mo-tert and mo-atr injected wild type animals does not change compared to mo-std injected animals.

This results are now shown in Supplemental Figure S3

Reviewer #2 (Significance (Required)):

Reported findings are novel, timely and model of possible therapeutic value for screening compounds for ALT and/or telomerase inhibitors. Mechanisms of co-existence of ALT and telomerase can also be explored using this model.

We fully appreciate the reviewer´s comments on the significance of the manuscript!

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

---

**Summary:**

Martinez-Balsalobre have examined caudal fin regeneration following surgical transection in WT and telomerase-deficient (tert+/- and tert-/-) zebrafish adults of several ages, and in one experiment, in embryos. They conclude: (1) regeneration efficiency decrease with aging in all genotypes (2) telomere length is maintained, even in a tert-mutant background (3) ALT (alternative lengthening of telomeres) is involved in supporting cell proliferation in tailfin regeneration. The experimental system employs a quantitative area-based measurement as a measure of the degree of regeneration. Functional studies used antisense morpholino gene knockdown and chemical inhibition to implicate ALT involvement.

**Major comments:**

The experimental logic is appropriate, and in general, the data support the conclusions. Strengths of the work include: (1) The quantitative measure of % regeneration appears to be quite objective; (2) the internally controlled experimental design of the morpholino knockdown experiments of Fig 5.

We thank the reviewer for the comment

The Western blot in Fig S2 has some issues. The image is a montage. The experiment appears to have been done only once. The band's identifications by kDa are imprecise (where is the 82 kDa band on the gel? - there are bands smaller and larger than 82 kDa, but none of 82kDa; the 50 kDa band is close to background; the DMSO lane is underloaded relative to the two test lanes (but as the observation is a reduction in the test samples, this does not result in a misinterpretation). What concentration of ATRinhIV were used? The blot has 10 and 50 microM, Fig S1B has 50 and 100 microM, and the text says 1-50 microM).

We have performed new regeneration experiments in wild type adult fish using ATR inhibitor. The results show that treating fish with ATR inhibitor results in a clear decrease in the overall phosphorylation status of ATR/ATR substrates within the regenerated tissue (Figure 5B and C). In this case, the intensity of the whole lane was used for quantification. As mentioned in the text, we used concentrations of 1, 10 and 50 microM, but we do not observe any difference with the 1 microM concentration, thus do not show it. Then we measured the regeneration capacity in both wild type and tert mutant fish using 10microM concentration

The MO-knockdown studies are interpreted as showing synergy of atr and tert knockdown.

There are two problems with them interpretation of synergy: (1) the single result of a greater effect with both MOs does not distinguish between an additive or synergistic effect (and synergistic action is by definition a greater than additive action;

We agree with the reviewer´s comment, and have removed the sentence “Interestingly a synergistic effect was observed when both mechanisms are inhibited” from the Results section.

(2) MO dose is not controlled by a group with an equal total MO doses (mo-std+mo-atr and mo-std+mo-tert). While acknowledging that the issues of using local MO delivery in an adult model are very different from global delivery in an embryonic model, the "synergy" interpretation still requires these experiments/controls be done. These experiments were not accompanied by any molecular evidence that either of the morpholinos targeted expression of the intended gene (which would likely have to be derived from their assessment in another system) - a control that can be challenging, but one that is regarded as essential in the field (https://doi.org/10.1242/dev .001115 ). While this will be difficult to do in the adult setting, it is still appropriate to validate the activity/molecular efficacy of the MO sequence in an experimentally tractable scenario. The specificity of this experiment and interpretation would also be enhanced and corroborated independently by undertaking the atr knockdown in the tert -/- mutant background. Overall, these experiments were preliminary and require further work that could be done withiin 3 months.

We have performed regeneration experiments using WT fish to address this issue. We analyzed the regenerated area of control and morpholino injected fish and then obtained regenerating blastema and analyzed the expression of tert and atr .The results are shown in Supplemental Figure S4 (A-E). The regeneration capacity is inhibited in tissues injected with a mix of mo-std+mo-ter, a mix of mo-std+mo-atr, or a mix of mo-tert+mo-atr compared with a control injected with a double dosis of mo-std (std 2x, Fig S4B). In addition, the expression of tert and atr is decreased in the regenerated blastema upon morpholino injection (Fig S4 C and D) indicating that the genetic inhibition of the expression of these genes was efficient. Finally, the levels of TERRA RNAs are increased upon amputation and this induction is reduced when we mo-atr or a combination of mo-atr+tert were microinjected (Fig S4E).

We have also performed caudal fin regeneration experiments in tert mutant fish microinjected with mo-atr and mo-nbs1, and analyzed the levels of TERRA RNAs and C-circles amount. The results are shown in Supplemental Figure S4. As expected, regeneration capacity of the caudal fins of fish microionjected with both morpholinos decreased compared to control fish (Fig S4 F-H). Consistently, TERRA RNAs levels, as well as C-circles amount, increased in the regenerating tissue and this induction was lower when atr and nbs1 gene expression was decreased by mo-injection (Fig S4 I and J). Taking altogether, these results indicate that ALT mechanism is induced upon an injury and operating in the regenerating tissue of both wild type and tert deficient fish.

Note - the tert MO sequence is missing from the table in Fig S4.

The sequence has been added

The adult experiments have used n=6-10 animals/group. There is no consideration of statistical power (is the analysis of Fig 1C adequately powered?).

The type of statistical test applied in Fig 1C (2-way ANOVA, plus Dunnett´s post-test) compares means of every clip among the 3 genotypes. This is the test that is recommended for this kind of data and experiment.

The degree and nature of replication is not clear in all cases. For example, in Fig 1, were the 6 fish run as one cohort of 6 animals in parallel (which would be just one experiment with 6 animals, each animal being a biological replicate), or were there 6 animals injured at different times (representing multiple independent experiments and represented a greater degree of reproducibility), or something in between. A similar question applies to the other figures.

In the experiments, 6 fish total were used per group sampled in at least 2 independent experiments. This has been included in the figure legend and in the Mat&Met section

For the experiment of Fig 6F, although there are >=100 larvae per group, it is not clear that this experiment has been done more than once.

In the conducted experiments, three independent trials were conducted. The total number of larvae per group utilized in each of the three distinct experiments surpassed 100 larvae per group (approximately 40 larvae in each independent experiment). This data has been incorporated into both the figure legend and the Materials and Methods section."

A few comments about data presentation. "Regeneration rate" and its derivatives are presented as mean +/- SEM. The parameter measured is correctly defined in methods as "Percent fin regeneration", however the graphs where it is plotted have the y-axis labelled as "regeneration rate (%)" (for example. Fig 1B), which is incorrect. The plotted parameter is not a rate - although there is a time dimension (x-axis), what is plotted at each time point is "% regeneration".

This has been corrected and y-axis is now labeled as Regeneration area (% of initial fin area.

Also, in most figures, such as Fig 1B and 1C, mean +/- SD would be more appropriate, as here each of the n=6 data points represents a single observation from one individual in the population, not the mean of 6 small samples of groups of individuals from the population. Furthermore, at these small n-values (6-10 through the report), scatter plots are considered a more appropriate way of displaying the data (some succinct references: DOI: 10.4103/2229-3485.100662 ; from a Nature group journal DOI: 10.1038/s41551-017-0079 ; from a PLOS journal https://doi.org/10.1371/journal.pbio.1002128 ).

This was a mistake in the figure legend, since Fig 1B was already showing mean +/- SD. Fig 1C is now showing mean+/- SD and has been represented with scatter plots.

The use of mean +/- SEM in Fig 4 could be appropriate, but as n is "at least two independent experiments" scatter plots would again be appropriate. Readers would then know which data sets had only two values.

In two instances, the same data are presented in two different ways (Fig 1B, 1C; the column graphs and arrows of Fig 3D).

Fig4 is presented now as scatter plot graphs

How does "data are average of at least 2 independent experiments" apply to Fig 3C?

In the experiments in Fig3C, “Data are average of 2 independent experiments of 3 fish per group pooled”. This has been included in the figure legend.

**Minor comments:**

The paper is written clearly overall. There are multiple minor grammatical/typographical errors, but these did not detract from understanding the manuscript. These were most abundant in the discussion.

A few points:

Discussion p1 - what is meant by "prematurely aged 11-month fish"

This refers to 11 months old tert-/- fish, which has been shown to present accelerated aging features at this age compared to wild type

Discussion p2 - you mean "doubled" rather than duplicated?

Yes; this has been corrected in the new version

tert +/+, tert +/- and tert -/- genotypes for experiments - how were these obtained and genotypically verified? (heterozygous incrosses? WT x homozygous mutant outcrosses?)

All the fish adult fish of the 3 genotypes were obtained from heterozygous incrosses. Then fish were genotyped by PCR. A protocol for genotyping has been added in the Mat&Met section. The wild type larvae used in the tail fin regeneration experiments inhibiting ATR were obtained by wild type cross, whereas the tert-/- were obtained by tert-/- incross

The last paragraph of the discussion makes some valid points, but it seemed out of place and I wondered if it was misplaced.

This paragraph is added to highlight that our work describes new in vivo model to perform drug screening to inhibit ALT mechanism of telomere maintenance, which is of particular importance for the survival of ALT positive tumor cells.

The rps11 primers appear in the Table of Fig S3 twice.

This has been corrected

Reviewer #3 (Significance (Required)):

---

The authors claim that this is the first in vivo model examining ALT in regeneration.

The paper contributes to the relatively small body of literature using adult zebrafish models (rather than embryonic larval models) in biomedical research. I cannot comment on the telomere/telomerase literature.

This report will be of interest to those working in regenerative medicine, telomere biology, cancer research, and those interested in zebrafish models of disease and physiological processes.

My expertise encompasses zebrafish disease models and functional studies; I do not have special expertise in telomerase or ALT pathways.

We fully appreciate the reviewer´s comments!

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

Summary:

Martinez-Balsalobre have examined caudal fin regeneration following surgical transection in WT and telomerase-deficient (tert+/- and tert-/-) zebrafish adults of several ages, and in one experiment, in embryos. They conclude: (1) regeneration efficiency decrease with aging in all genotypes (2) telomere length is maintained, even in a tert-mutant background (3) ALT (alternative lengthening of telomeres) is involved in supporting cell proliferation in tailfin regeneration. The experimental system employs a quantitative area-based measurement as a measure of the degree of regeneration. Functional studies used antisense morpholino gene knockdown and chemical inhibition to implicate ALT involvement.

Major comments:

The experimental logic is appropriate, and in general, the data support the conclusions. Strengths of the work include: (1) The quantitative measure of % regeneration appears to be quite objective; (2) the internally controlled experimental design of the morpholino knockdown experiments of Fig 5.

The Western blot in Fig S2 has some issues. The image is a montage. The experiment appears to have been done only once. The band's identifications by kDa are imprecise (where is the 82 kDa band on the gel? - there are bands smaller and larger than 82 kDa, but none of 82kDa; the 50 kDa band is close to background; the DMSO lane is underloaded relative to the two test lanes (but as the observation is a reduction in the test samples, this does not result in a misinterpretation). What concentration of ATRinhIV were used? The blot has 10 and 50 microM, Fig S1B has 50 and 100 microM, and the text says 1-50 microM).

The MO-knockdown studies are interpreted as showing synergy of atr and tert knockdown. There are two problems with them interpretation of synergy: (1) the single result of a greater effect with both MOs does not distinguish between an additive or synergistic effect (and synergistic action is by definition a greater than additive action; (2) MO dose is not controlled by a group with an equal total MO doses (mo-std+mo-atr and mo-std+mo-tert). While acknowledging that the issues of using local MO delivery in an adult model are very different from global delivery in an embryonic model, the "synergy" interpretation still requires these experiments/controls be done. These experiments were not accompanied by any molecular evidence that either of the morpholinos targeted expression of the intended gene (which would likely have to be derived from their assessment in another system) - a control that can be challenging, but one that is regarded as essential in the field (https://doi.org/10.1242/dev.001115). While this will be difficult to do in the adult setting, it is still appropriate to validate the activity/molecular efficacy of the MO sequence in an experimentally tractable scenario. The specificity of this experiment and interpretation would also be enhanced and corroborated independently by undertaking the atr knockdown in the tert -/- mutant background. Overall, these experiments were preliminary and require further work that could be done withiin 3 months. Note - the tert MO sequence is missing from the table in Fig S4.

The adult experiments have used n=6-10 animals/group. There is no consideration of statistical power (is the analysis of Fig 1C adequately powered?).

The degree and nature of replication is not clear in all cases. For example, in Fig 1, were the 6 fish run as one cohort of 6 animals in parallel (which would be just one experiment with 6 animals, each animal being a biological replicate), or were there 6 animals injured at different times (representing multiple independent experiments and represented a greater degree of reproducibility), or something in between. A similar question applies to the other figures. For the experiment of Fig 6F, although there are >=100 larvae per group, it is not clear that this experiment has been done more than once.

A few comments about data presentation. "Regeneration rate" and its derivatives are presented as mean +/- SEM. The parameter measured is correctly defined in methods as "Percent fin regeneration", however the graphs where it is plotted have the y-axis labelled as "regeneration rate (%)" (for example. Fig 1B), which is incorrect. The plotted parameter is not a rate - although there is a time dimension (x-axis), what is plotted at each time point is "% regeneration". Also, in most figures, such as Fig 1B and 1C, mean +/- SD would be more appropriate, as here each of the n=6 data points represents a single observation from one individual in the population, not the mean of 6 small samples of groups of individuals from the population. Furthermore, at these small n-values (6-10 through the report), scatter plots are considered a more appropriate way of displaying the data (some succinct references: DOI: 10.4103/2229-3485.100662 ; from a Nature group journal DOI: 10.1038/s41551-017-0079 ; from a PLOS journal https://doi.org/10.1371/journal.pbio.1002128). The use of mean +/- SEM in Fig 4 could be appropriate, but as n is "at least two independent experiments" scatter plots would again be appropriate. Readers would then know which data sets had only two values. In two instances, the same data are presented in two different ways (Fig 1B, 1C; the column graphs and arrows of Fig 3D). How does "data are average of at least 2 independent experiments" apply to Fig 3C?

Minor comments:

The paper is written clearly overall. There are multiple minor grammatical/typographical errors, but these did not detract from understanding the manuscript. These were most abundant in the discussion. A few points: Discussion p1 - what is meant by "prematurely aged 11-month fish" Discussion p2 - you mean "doubled" rather than duplicated? tert +/+, tert +/- and tert -/- genotypes for experiments - how were these obtained and genotypically verified? (heterozygous incrosses? WT x homozygous mutant outcrosses?) The last paragraph of the discussion makes some valid points, but it seemed out of place and I wondered if it was misplaced. The rps11 primers appear in the Table of Fig S3 twice.

Significance

The authors claim that this is the first in vivo model examining ALT in regeneration.

The paper contributes to the relatively small body of literature using adult zebrafish models (rather than embryonic larval models) in biomedical research. I cannot comment on the telomere/telomerase literature.

This report will be of interest to those working in regenerative medicine, telomere biology, cancer research, and those interested in zebrafish models of disease and physiological processes.

My expertise encompasses zebrafish disease models and functional studies; I do not have special expertise in telomerase or ALT pathways.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Using zebrafish as a model for regeneration, the authors find that telomere maintenance by recombination can occur in the absence of telomerase.

Title to Figure 4 perhaps may be too strong, 'ALT mechanism is activated', since only a few features of ALT are assessed. Perhaps, 'ALT features are activated'?

mRNA levels of NBS, ATR are also increased in WT animals (Figure 4A and 4B), but ATRX and DAXX mRNA levels are not decreased in WT animals. Is the increase why the authors in part suggest that ALT is being used in WT animals. If so, what would be the trigger for the use of ALT, as opposed to the trigger to use ALT in tert-/- animals?

In Figure 5C, if tert-/- animals are downregulated for nbs1 and atr, would it be expected that the effect on regeneration be more pronounced compared to tert+/+ downregulated for nbs1 and atr than what is observed?

What are the telomere lengths in tert-/- animals treated with mo-atr or mo-nbs1 or in tert+/+ animals treated with mo-tert and mo-atr compared to singly treated?

Significance

Reported findings are novel, timely and model of possible therapeutic value for screening compounds for ALT and/or telomerase inhibitors. Mechanisms of co-existence of ALT and telomerase can also be explored using this model.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

The authors Martínez-Balsalobre and colleagues found that the regenerative capacity of the zebrafish caudal fin is not limited by the lack of telomerase and showed that the length of telomeres does not decrease substantially after repeated amputations in telomerase-deficient zebrafish. These findings prompt the authors to explore an alternative mechanism that would explain the maintenance of telomere length in this regeneration setting. They produced suggestive evidence for the role of the ALT (Alternative Lengthening of Telomeres) mechanism in the maintenance of telomere length in the absence of telomerase in a regeneration setting. In my view, several points need to be addressed and clarified.

There are three major points:

1.When working with tert mutants, the age at which these fish show a telomere phenotype (namely, loss of body mass and reduced fertility) varies. Therefore, it would be important to state if the fish used in this study were already showing these phenotypic characteristics at each time point studied, namely 4, 8 and 11 months of age.

2.The knockdown experiments were performed using morpholinos. To confidently use morpholinos it is fundamental to demonstrate first their knockdown efficiency and their specificity. This is lacking in the manuscript.

3.The involvement of ALT mechanism in the regeneration process in the absence of telomerase is only suggestive, as the authors show an increase of C-circles and heterogenous telomerase length in telomerase-deficient zebrafish but when trying to establish a functional link the authors resort to the knowndown of genes that may be associated with ATL. Looking at the levels of TERRA and the number of C-circles in the knowndown caudal fins would be essential for their claim.

And several other points:

4.The regeneration experiments were performed at 32 degrees and this option was never explained nor discussed.

5.When referring to the ALT mechanism, the authors state that "... in about 10% of tumors cells, telomere length is maintained by the Alternative Lengthening of Telomeres (ALT) mechanism ..." and I think it would be more accurate to talk about cancer cells instead of tumor cells.

6.The sentence about C-circles is incorrect. C-circles are mostly single-stranded and not double-stranded as stated.

7.After Figure 2, the authors never mention the age of the fish used.

8.In Figure 1A The site of amputation does not fit the one described in Mat & Met that states 2 cm from the base of the caudal peduncle. The same stands for Figure 2A. The experimental procedure refers 1 dpa but this time point is not plotted in the graph in Figure 1B.

9.In Figure 1B The Y axis should be named regeneration area instead of rate as the values are a percentage of the area reached after a certain time point after amputation. The same stands for Figure 2B, C. It would be nice to see the real caudal fin images for the relevant time points: before amputation, 0 dpa, when the fins reach 50% of regeneration area and then the last time point. The authors should discuss why are the caudal fins reaching more than 100% of regeneration area.

10.In Figure 2B The meaning of ". .. ." on the right side of the graph is not clear. The same stands for Figure 2D.

11.In Figure 2C Why is the clip 10, 11 and 12 missing from the tert+/- and tert-/- ?

12.In Figure 2E The proximity of all points at the 12 Clip is indicative of lack of statistical significance, therefore the **** related to which comparisons?

13.In Figure 2D, E For the measurement of telomere length, the authors state that "Data are average of at least 2 independent experiments." What does this mean exactly? How many animals were used in each experiment?

14.In Figure 3 The authors state that "Data are average of at least 2 independent experiments." What does this mean exactly? How many animals were used in each experiment? Why were the c-circles evaluated at hpa while the telomere length evaluated at dpa? This should be discussed.

15.In Figure 3A The meaning of ". .. ." on the top side of the graph is not clear. t0 should be removed and replaced by 0 hpa and 24hpa and 48hpa for coherence.

16.In Figure 3B,C 0 hpa replace by 0 dpa

17.In Figure 3B The blue and red stainings in the panels are labelling exactly what? This should be stated in the image and in the legend.

18.In Figure 3D There is a mistake in the legend the should be corrected as follows "Very long telomeres have a higher fluorescence of 200,000 AUF and very short telomeres have a lower fluorescence of 30,000 AUF."

19.In Figure 4 t0 should be removed and replaced by 0 hpa. The meaning of ". .. ." on the top side of the graph is not clear. The title is an overstatement, as the genes studied are DNA damage genes that may associate with ALT.

20.In Figure 4A, B The expression of nbs1 and atr in tert-/- increases at 48hpa but the same seems to be true for the tert+/+ and this is never discussed by the authors.

21.In Figure 4C, D The differences in the expression of atrx and daxx decreases over time in a in tert-/- and this is never discussed by the authors.

22.In Figure 5 An ideal control would be the direct comparison between microinjected+electroporated mo-std in the ventral part of the fin while the dorsal part would be microinjected+electroporated with the mo-gene of interest. This would discard any effect of microinjection+electroporation in the regeneration efficiency. These experiments are not convincing to show that there is an ALT mechanism is operating here. What this experiment shows if the relevance of these genes for the regenerative capacity of the caudal fin. To show that this is related to the ALT mechanism the authors should investigate the C-circles in these regenerating fins. The amputation red lines are not placed in the exact amputation position in some of the panels. Regeneration rate should be regeneration area.

23.In Figure 5C, E Why is the mo-tert more inhibitory of regeneration (Figure 5E - around 30%) than the tert-/- mutant (Figure 5C - around 60%)? This should be discussed.

24.In Figure 6A The 2 adult zebrafish shown in the tank with the ATR inhibitor IV should have an amputated caudal fin. Control is exactly what? Untreated? Treated with vehicle? Why was the ATR inhibitor IY added immediately after fin amputation while the mo-atr was injected at 48 hpa?

25.Figure 6D, E, F These panels are a bit out of the focus of this paper. If presented should go to a supplementary figure.

26.In Figure S2 The relevant bands should be identified. The gel identifies DMSO, 10uM and 50 uM but the quantification graph identifies Control, 50uM and 100uM. There are no error bars. The authors say that the quantification of various western blot bands was done but how many exactly?

27.In Figure S3 The primers for rps11 are repeated twice. Were these primers design de novo by the authors or did they used previous reported primers, in this case the references should be given. Tert F2 and R1 should be replaced for F and R for consistency.

28.In Figure S4 The sequence of tert mo is missing.

29.In the methods the genotyping protocol of tert mutants is not described.

30.The method to calculate the area of the fin pre- and post-amputation is not described.

Significance

The manuscript by Martínez-Balsalobre and colleagues deals with a very interesting question on the importance of telomere lengthening during regenerative processes and its relation to ageing. To this end the authors made use of the tert mutant, a telomerase-deficient zebrafish. The authors show a surprising phenotype that telomerase-deficient zebrafish can still regenerate their caudal fins and are able to maintain telomere length during consecutive amputations and I say surprising because it has been shown that telomerase-deficient zebrafish are unable to regenerate their hearts efficiently.

Taking these novel findings, the authors propose that in the zebrafish caudal fin and in the absence of telomerase, telomere length is maintained through the activation of an alternative mechanism called ALT. To my knowledge, the role of ALT as a mechanism of telomere lengthening has never been described in the context of regenerating organs in zebrafish.

Referees cross-commenting

I agree with the comments made by the other reviewers. I would stress the need to tone down the role of ALT during fin regeneration in zebrafish as all the experiments are only indicative of the possible of the involvement of ALT.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

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

Learn more at Review Commons

---

Reply to the reviewers

'The authors do not wish to provide a response at this time.'

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

Learn more at Review Commons

---

Referee #4

Evidence, reproducibility and clarity

Abdullah Alieh and colleagues generate comprehensive transcriptome annotations in FACS-sorted murine cortical neural stem cells, precursor cells and neurons by combining (existing) short-read RNA-seq (SRS) data with long-read sequencing (LRS) data. They identify around 50,000 novel transcripts and show that they are enriched in neural functions and have a strong tendency of increasing inclusion of differential splicing events during differentiation. Several examples are validated by PCR. They show, by means of AlphaFold2 prediction of protein structure, that many splice isoforms likely cause either overall structural differences or switches in secondary structure.

Major points:

1. The authors generate data using a previously characterized mouse model. However, they need to reconfirm expression of markers for the three cell types they analyse, particularly since one is identified by lack of expression of fluorescent tags.
2. Validation is only performed by RT-PCR on 11 novel splicing events and not at all on novel TSS and termination sites. It would greatly benefit the reliability of novel isoforms if the authors could compare them with those detected previously by LRS in neural cells, or overlay novel TSS with data such as CAGE or 3'-end sequencing.
3. Are divergent structural regions between isoforms often within regions of low model confidence? This would impact the relevance of the discovered changes.
4. In the Discussion, the authors assert that '...AS alone was revealed to have a much greater impact in remodeling the transcriptome [...] than previously thought and independently from changes in gene expression.' However, this latter aspect is not demonstrated. To what extent does apparent change in AS derive from differential expression of isoforms from alternative TSS?
5. The statement in the Discussion that 'Our study supports this notion [that differential inclusion of disordered segments can affect protein-protein interaction] with a significant increase in disordered isoforms arising concomitantly with neurogenic commitment' is not supported by the results presented. The authors only show that alternatively spliced proteins in their dataset have a higher propensity for disordered regions than the proteome at large, which is not a new observation.
6. The statement in the Discussion that structural changes ostensibly caused by alternative splicing were 'similarly the case both when the structural change occurred within the AS event as well, more remarkably, when the event was far away' is not supported by the results as presented.
7. Supplementary material is mentioned but not included with the manuscript.

Minor points:

1. Fig. 1A: Why are there two numbers for transcripts (70,658, 71,760) in the overlap of pipelines 1 and two?
2. Fig. 2F: Statements that events either low in NSC and rising, or high in NSC and declining, represent the 'least represented' isoform in NSC or N, respectively, do not seem to take into account that there may be other transcript isoforms for which inclusion of the event in question stays constant (e.g., skipped). The authors could make use of their LRS to confirm that at least for selected events.
3. p8: How many unique new transcription start and end sites were identified?
4. Fig. 2C: were categories selected for display (and if so, how), or are these all the categories identified?
5. Fig. 2F-H: How many of the detected AS events, including neural microexons, are novel?
6. Was the propensity to elicit nonsense-mediated decay taken into account when AS events were mapped to transcripts that did not contain them?
7. How did 212 genes selected for modeling in Fig. 3 correspond to 987 isoforms? When genes comprised more than two isoforms, how were the changes in quantified properties attributed to the splicing events for which they were selected vs other isoforms or alternative translation start and stop sites?
8. Fig. 3D: Coloring the structures by chain would make this figure easier to interpret.
9. Details of Alphafold modeling are not provided.
10. The authors should acknowledge that integrating SRS and LRS is a standard approach to generating annotations in organisms for which no reliable annotation exists, as well as approaches aimed at doing so to improve annotations in mammals, such as PMID: 37779246, 35468141, 32461551 etc.

Significance

While a combination of SRS and LRS sequencing along stages of neuronal differentiation has not been used in the same way to identify novel transcript isoforms, substantial work has been done employing LRS in neural contexts, including in single cells (e.g., work from the Tilgner, Waldmann lab).

Although it is not entirely clear from the results presented how many of the detected AS events are novel, as opposed to transcript isoforms, their characteristics are similar to previously known neural-differential events, thus supporting their veracity. The main advance in this manuscript lies in the insights derived from structural modeling of splice isoforms, which supports the potential relevance of many splicing events. This is a question relevant for both fundamental research and clinical audiences. However, several of the author's claims are not well supported, or else are not novel (see major points).

This reviewers' expertise lies in the field of molecular biology of alternative splicing; they have experience with RNA-seq and structural modeling of splice variants.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

Summary:

Haj Abdullah Alieh at al., describe re-analysis of an existing short read RNA-Seq dataset consisting of 3 replicates of 3 FAC sorted cell populations of the E14.5 Btg2::RFP/Tubb3::GFP mouse cortex: neural stem cells (NSC; RFP-/GFP-), neural precursors (NP; RFP+/GFP-) and neurons (N; GFP+), for the purpose of investigating alternative splicing isoform switching during neuronal cell-type specification. They generate a one replicate PacBio dataset of these same sorted cells, with the aim of identifying full-length transcript isoforms, which are difficult to discern with short-read data alone. The key conclusions are the discovery of ~50,000 novel transcript isoforms containing ~2,500 novel splice junctions; the discovery of isoform switches between NSC -> neuron that contain a high proportion of microexon inclusion events and the finding that many of these switches are predicted by Alphafold2 to have a structural impact.

The data is interesting and the bioinformatics approach of investigating potential impacts of splice variants on protein structure using Alphafold2 is also interesting, however at present the paper would be better presented as a resource, unless effort is undertaken to experimentally validate some potential biological findings. However, for the paper to be useful as a resource, links to newly generated data and analysis code need to be provided. The capacity for exploration of these newly identified splice isoforms, or further analysis using the new GTF, could then be one of the attractions of this work.

Major comments:

• Are the key conclusions convincing?
• Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?
• Would additional experiments be essential to support the claims of the paper? Request additional experiments only where necessary for the paper as it is, and do not ask authors to open new lines of experimentation. Figure 1 The discovery of ~50,000 novel transcript isoforms containing ~2,500 novel splice junctions As far as I can see the description of novelty is based on them being not present in either Ensembl (GRCm38.p6), NCBI_RefSeq, or Gencode (vM10) - note here the numbers are genome assembly versions and do not refer to the GTF annotation versions compared against - these should be provided as they are frequently updated. The claim is that they are not present in these references because the unique cell samples have not been analysed before. For transcript isoforms to be included in these references they must have a good level of support. I have a couple of concerns about the support for these isoforms: The numbers in figure 1A do not add up. For long read sequencing two pipelines are used resulting in 76,077 and 80,782 isoforms - in the venn diagram 1A the overlapping circle has two numbers of isoforms in it: 70,658 and 71,760 so it is unclear, are 70,658 isoforms found by both pipelines or 71,760? Then we are told the union of these transcripts is taken forward to the next venn diagram. However this diagram is labelled with 82,046 transcript isoforms. Pipeline 1 has labelled 5419 unique isoforms, pipeline 2 has 9,022 unique isoforms so 5419 + 9022 + 70658(71760) = 85,099(86201) not 82,046 - perhaps some extra filtering has occurred that should be labelled/described? Again the final number of transcripts at the end of everything is off - if the 82,046 transcripts from long read are combined with the 16,070 unique to the short read this equals 98,116, not 97,240. The authors decide to use long read sequencing to assemble the isoforms as short-read sequencing is unreliable for assembling full length isoforms - however for their final list they merge isoforms assembled by StringTie from short read data with the isoforms assembled from the PacBio long read data, it seems likely that the isoforms detected only by short-read Stringtie assembly would be unreliable and shouldn't be included in the final total. The authors perform only one biological replicate of PacBio long read sequencing of three different samples, so it is not possible to easily determine the reproducibility of the findings. I appreciate PacBio is expensive, the authors could consider other ways to evaluate the reproducibility - perhaps by looking at the detection of transcripts expected to be uniformly expressed between the different conditions? The authors provide no quality information for their PacBio sequencing run - eg. length distribution of reads, how many reads are left after quality filtering, quality across the length of reads, ie. I do not know if most isoforms reported are supported by 5 full length isoform reads, or if it is rare in the dataset to get full length isoform reads .etc is the quality comparable across the three PacBio samples? How many of the novel isoforms are supported by both short read and long read data? How many of the novel isoforms are supported only by short reads? How many isoforms are found in all three PacBio samples? Does gene expression measured with the PacBio data match the previous results of measuring gene expression in the short read data? Adding these kinds of analyses would give more confidence in the results. This section of methods is confusing, I don't really understand what has been done or what part of the manuscript this refers to: "​​Events were assigned to an inclusion isoform if their coordinates overlapped, at least partially, with an exon or to an exclusion isoform if they were located within an intron. AS events without a corresponding inclusion or exclusion isoform were assigned to an Ensembl or NCBI_RefSeq isoform using the criteria above. Only AS events assigned to at least one inclusion and one exclusion isoform were considered for further analysis." VastDB is a splicing database created by Manuel Irimia/Ben Blencowe containing a lot of neural samples across development - how many of the 'novel' splice sites are present in VastDB? Similarly, how many of the 'novel' splice isoforms were previously detected by Zhang et al., 2016, Cell.

Figure 2: over neuronal maturation the major splicing change is for cassette exons to become more included, 50% of those measured being microexons Overall this section is strongest, the conclusions are well supported. Figure 2D - there are no genome coordinates given to allow the reader to check the highlighted events out for themselves. Figure 2F is very confusing, consider an alternative way to present this. Figure 2G, the premise of this analysis is interesting! But confused on the numbers - in 2F its shown that 226 exons become more included between both NSC->NP->N, so why are 441 exons plotted in 2G? Whilst I appreciate genes must be expressed in both NSCs and neurons to be able to calculate differential splicing, one thing not addressed is whether expression of a lot of these genes also goes up in neurons, i.e. could it be that when these genes are lowly expressed in NSCs their splicing is not particularly well regulated but it doesn't really matter because they are not really required in NSCs? This becomes relevant later where you start to address the functionality of isoform switches - if the gene is expressed to the same degree in NSC vs. N this would suggest that both isoforms are functional, if a gene is very lowly expressed in NSC but highly expressed in N, then maybe only the N isoform needs to be functional. Gene ontology methodology is not described in the methods. What were the spliced genes compared against? Given these are neural samples, lots of expressed genes will have neural functions, so is this really informing us about the alternatively spliced genes? The manuscript would benefit by integration of its data with other published datasets - especially with the microexons - how do these behave in other datasets of neuronal maturation (such as those from vastdb or zhang 2016)? The authors could consider looking at motifs around regulated microexons to try and establish if any specific RBPs might be involved in this regulation, although this would benefit from follow up experiments.

Figure 3: exon inclusion in neuronal specific transcripts confers different structures to translated proteins, suggesting these events are important functionally Here, Alphafold2 is used to predict the structures of switching isoforms, whilst an interesting approach to inform further experiments, presented alone, it remains hypothetical. Hook2 is highlighted as one example, where inclusion of a microexon introducing two amino acids to the translated protein is predicted to cause a structural change that will impact its binding to microtubules. It's hard to determine if this really will have a functional impact without doing experiments in the lab. For this manuscript to serve as a research (rather than resource) article, it would benefit from an example experiment expressing neuronal vs. NSC Hook2 isoform in a cell line and measuring co-localisation with microtubules via IF microscopy, or something similar to address the proposed function. In the second half of this figure, more subtle local structural changes are investigated and the example of an alpha-helix to beta-strand switch predicted in Kctd13 is presented. The figure would benefit from showing the splicing change at the RNA level and relating that to the change seen at the protein sequence level as it is a bit confusing - the region of deletion is labelled as 'AS REGION' however, two amino acids preceding this box are different between the two isoforms (KVEF vs. KVRG) - so presumably the splicing change starts earlier than denoted? In the discussion the authors state: "While these regions are long known to exist, their structural switch was assumed to be dependent on substantial changes in their structural and sequence contexts (Gendoo and Harrison, 2011; W. Li et al., 2015) as opposed to, as observed in our study, being triggered by small perturbations within nearly identical sequence contexts." It's not clear whether these small local predictions are accurate and would require some additional structural data to validate. - Are the suggested experiments realistic in terms of time and resources? It would help if you could add an estimated cost and time investment for substantial experiments. Suggestions of additional computational analysis are very realistic and shouldn't take longer than a month or two. The addition of experimental data to support Figure 3 would take considerable time and resources, potentially collaboration with other labs. Perhaps focusing on making this dataset an accessible resource would be a better route to publication. - Are the data and the methods presented in such a way that they can be reproduced? No, no source code, software versions or supplementary data/materials is provided. - Are the experiments adequately replicated and statistical analysis adequate? Having one replicate of the PacBio experiment is a bit concerning, but I am aware that it is expensive. Given they have three samples of different conditions with PacBio data perhaps showing the quality control of the libraries, reproducibility of transcripts that don't change in the three conditions, etc. would give more confidence in the data.

Minor comments:

• Specific experimental issues that are easily addressable. Made above.
• Are prior studies referenced appropriately? Yes. Except for this section of introduction: "While great effort is being made to overcome these limitations, capturing cell type-specific AS dynamics that is both quantitative and comprehensive of full-length transcript information currently requires combination of both SRS and LRS performed in parallel on the same cell pool. This was seldom attempted (Gupta et al., 2018; Joglekar et al., 2021) and, to the best of our knowledge, never for specific cell types of the developing mammalian brain. Even more limiting, systematic assessment of the consequences of AS on protein structure and putative function in cell fate commitment is entirely lacking. "

LRS has allowed for whole transcriptome determination and quantification in a number of cases, especially in non-model organisms, below I mention some examples from human and mouse: Nanopore use in GTEX + short reads: Glinos et al., 2022 Nature https://www.nature.com/articles/s41586-022-05035-y PacBio SMRT-Seq + short reads human and mouse cortex: Leung et al. Cell Reports 2021 https://www.cell.com/cell-reports/pdf/S2211-1247(21)01504-7.pdf PacBio IsoSeq + short reads in human and mouse sperm: Sun et al., 2021 Nature Communications https://www.nature.com/articles/s41467-021-21524-6 Single cell long read RNA-Seq has also been described in several scenarios and is worth referencing in the introduction: Samples from various mouse and human sources: Tian et al., 2021 Genome Biology https://link.springer.com/article/10.1186/s13059-021-02525-6 differential isoform usage in myeloma cell lines: Phillpott et al., 2021, Nature Biotech https://www.nature.com/articles/s41587-021-00965-w Single cell long read isoform analysis in human immune cells: Volden and Vollmers, 2022, Genome Biology https://genomebiology.biomedcentral.com/articles/10.1186/s13059-022-02615-z - Are the text and figures clear and accurate? Mostly, I've highlighted where numbers in figures don't make sense to me. Generally the text could use some going over and tightening up (eg. sentence on page 12 needs revising for clarity and typo "The fact that within this helical packing resides the protein domain essential for Hook2 function to bind microtubules, implies that such a negligible AS switch by two ammino acids may result in a completely altered function. ") - Do you have suggestions that would help the authors improve the presentation of their data and conclusions? I have made suggestions above about figures that are unclear to me.

Referees cross-commenting

After reading the reviews of other reviewers, it seems we are much in agreement over the main concerns relating to this manuscript. Namely: concerns over the PacBio being single replicate, concerns over indiscriminately merging PacBio and SRS transcripts, concerns about lack of validation of the structural changes predicted by AlphaFold2. On the question of novelty and significance we also seem to be aligned.

Significance

• Describe the nature and significance of the advance (e.g. conceptual, technical, clinical) for the field.

The main general findings of the work have been described elsewhere: that microexon inclusion increases in many transcripts during neuronal cell fate commitment has previously been described, the suggestions of important isoform structural changes in Hook2 and Kctd13 are not backed up by any experimental data and so are not reliable. The description of a huge number of novel isoforms is not particularly useful because it's not clear if these have been found by other similar studies, because the data is not compared, furthermore we have no information about these isoforms to be able to pursue further research about them. The main output of the work would be the data and transcript annotations for other people to follow up on, but this is not provided in any accessible way. The paper might be better reframed as a resource, if it is not possible to follow up on the biological conclusions. - Place the work in the context of the existing literature (provide references, where appropriate).

Previously, alternative splicing has been studied in purified cell types of the developing mouse cortex using short read sequencing eg. in Zhang 2016, Cell. In this previous study, VZ NPCs (EGFP−) and non-VZ cells (EGFP+) were isolated from E14.5 Tbr2-EGFP mouse cerebral cortex. The double reporter mouse model used in the present study allows for better cell sorting into NSC, NPC and neurons, and the long read sequencing allows for whole transcript identification, however the present study has made no effort to compare the data, so it's not clear how much new biology this leads to. In Zhang 2016, the authors also predict disruption to protein domains caused by AS, but go further to perform experiments to validate the impact of some of these predictions. - State what audience might be interested in and influenced by the reported findings.

Researchers of this cell fate transition might want to look at their favourite genes to see if there are novel isoforms reported (however this is currently not possible because this information is not provided). Researchers of Hook2 or Kctd13 may want to further explore the described predicted structural changes. Researchers generally studying alternative splicing may want to include the novel isoforms in their analyses (again currently not possible because they are not provided). Generally this paper would probably be best seen as a resource. - Define your field of expertise with a few keywords to help the authors contextualize your point of view.

Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate. Bioinformatics, Splicing, RBP biology

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

In this manuscript the authors attempt to characterize alternative splicing in neurogenic progenitors during corticogenesis and the consequence of such alternative splicing on protein conformation. To do this the authors used previously published short-read sequencing data from neural stem cells, neural progenitors, and neurons at E14.5 and expanded on this dataset by adding long-read sequencing data.

Major comments:

1. According to the methods section, new long-read sequencing data was generated for each of the NSC, NP, and N cell types. It is unclear to me how these were processed in terms of replicates. From figure 1 is seems that the samples were sequenced individually but then pooled for transcriptome assembly. It would really be helpful to understand the quality of the samples better. Are there replicates for each of the cell types included? What did the read count and transcript detection look like for each of the individual samples? Are the 3 samples really equal enough to be pooled together or will 1 sample dominate when assembling the transcriptome?
2. On page 9, end of 2nd paragraph the authors state: "... these findings highlight the extent of AS within the neurogenic lineage underscoring its potential to regulate corticogenesis to a much greater degree than previously appreciated." Would it be possible to do a direct comparison between the number of AS detected or the type of AS detected between published data and the current paper? The authors provide a very coarse description of AS events during corticogenesis based on GO terms. The GO terms to surface are not surprising and seem not very meaningful in distinguishing the three cell types. Are there lower level GO terms that are specific to a subset of the cell populations?
3. The authors show that cell types moving from NSC to NP to N gain exons. This raises the questions whether there is a specific set of genes that gains exons during development and/or there are different RNA binding proteins present in the three cell populations that could contribute to the differential splicing patterns seen in the three cell populations?

Minor comments:

1. What was the background chosen for gene ontology analysis?
2. For this paper the focus was on development of neurons. Certain non-neuronal populations arise from NSC and it would be interesting to compare the non-neuronal lineage as well. To what extent is the splicing pattern a differentiation/maturation hallmark and to what extent is it specific to the neuronal lineage.

Significance

• General assessment:
  • Strengths: This manuscript describes a potential strategy to investigate the effect of alternative splicing events on the protein output. By combining short- and long-read sequencing the authors are able to capture a wide variety of splicing events in the neuronal lineage at one timepoint during development. The modeling of potential protein structures that arise from the alternatively spliced transcripts is critical to start to understand the biological effects of alternative splicing in ever changing systems like the brain during development.
  • Limitations: Main limitations are the wet-lab experimental setup. The analysis was performed on a limited number of samples (n=1?) per cell type for just 1 time point. It is not known what the variability in AS events between individuals is and will limit statistical testing.
• This manuscript is mostly a proof-of-concept but does not provide enough solid proof to claim new discoveries.
• This manuscript serves a specialized audience interested in alternative splicing and biological effects of splicing events.
• Filed of expertise: single cell transcriptomics, long-read, alternative splicing, mouse brain development.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

The authors FACS-sorted neuronal cells and conducted both short- and long-read sequencing to delineate the process of neurogenic differentiation. They went on to verify certain new splice junctions via RT-PCR and employed AlphaFold2 to forecast the outcomes. There are several issues the authors need to address.

1. It's unclear why the author decided to superimpose the GTF file created by StringTie (intended for SRS) onto those generated by two distinct LRS pipelines. Given that long-read sequencing doesn't match the accuracy of NGS, which could result in discrepancies in splice junction coordinates, this approach seems questionable. Additionally, the presence of alternative start sites or polyadenylation sites could further reduce the concordance rate, as evidenced by the mere 15% transcript overlap between the methods depicted in Figure 1A. The updated version of StringTie, StringTie2, offers an improved protocol for assembling short-reads using long-read data as a guide. The author should contemplate the use of these more advanced tools rather than combining them in a potentially incompatible manner.
2. The main text and figure legends of Figure 1 do not specify the number of replicates used.
3. The author needs to depict alternative splicing events with gene annotations, such as those seen in a sashimi plot in panel 1C. The existing panel does not adequately differentiate whether the splice junctions presented are novel. Furthermore, the author should provide the PSI for each splicing event and contrasts these with the PSI derived from RT-PCR data.
4. In the discussion section, the author asserts that their methodology, combining Short Read Sequencing (SRS) and Long Read Sequencing (LRS), is novel. However, similar approaches have been reported in previous studies, for instance in references 10.1371/journal.pcbi.1009730 and 10.1098/rsob.220206.

Significance

While the sequencing data and the integration of AlphaFold2 are new, the authors fall short of experimentally demonstrating the biological significance of their findings.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

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

Learn more at Review Commons

---

Reply to the reviewers

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

Summary:

The current study investigates the metabolic regulation of hematopoietic cell differentiation through chromatin modification and gene expression. Using the primary CD34+ human cord blood cells, the authors show that transient pharmacological inhibition of glycolysis, PPP, and glutamine/glutamate metabolism alters the dynamics of chromatin structures and gene expression, leading to the impacts on cell proliferation, morphology, and the long-term differentiation capacity. Following are specific comments:

Major:

1. The rationale behind the selection of the metabolic targets and the working hypothesis regarding specific effects on cellular consequence is not explicitly conveyed, which makes it difficult to judge if the experiment design is appropriate and if the results address the questions:
2. The operational definition of "Metabolic perturbation" or "Metabolic stress" needs to be provided and the validation of inhibitory effects needs to be clarified. Fig. 3D and S1 Fig are supposed to indicate the inhibition of targeted metabolic pathways but it is not clear if the authors believe the inhibitors exert expected metabolic effects based on the presented data. The author should explain why they target the selected pathways (i.e. glycolysis, PPP and glutamine/glutamate metabolism) and precisely point out which up or down regulation (in Fig. 3D and S1 Fig, for example) indicate sufficient and specific inhibitory effects for each inhibitor to operationally define "metabolic perturbation". Thank you for bringing this point to our attention. We extended the Introduction section (page 3) with a paragraph better explaining the notion of metabolic perturbation or stress. Indeed, a clear definition of the metabolic targets is also required. Consequently, the update includes a more detailed presentation of the metabolic steps and the rationale as to why we selected them as targets (pages 3 to 4). Additionally, we have also incorporated an extra figure (S1 Fig) to illustrate the major metabolic pathways affected by the various inhibitors.

In this study, we have used single time-point detections of steady-state metabolite levels. The single time-point detection of individual metabolite levels alone does not allow clear understanding of the precise metabolic alterations. The network of metabolic reactions is highly interconnected with complex regulatory loops that makes precise predictions difficult. More detailed metabolic flux studies will be required to characterize the perturbations. There are considerable challenges in carrying out such flux experiments with the limited amount of cells (which cannot all be from a single patient source), making such experiments well beyond the scope of this study. However, even with single time-point steady inhibitor studies, we observe significant and inhibitor-specific cellular reactions involving cell division rate, morphology, cell surface marker distribution and changes in bulk metabolite levels. Therefore, we interpret these changes as collectively reflecting the metabolic impact of the inhibitors, which can be qualified as metabolic perturbation or stress. The manuscript has been modified (page 5) to clarify this point.

1. Given that the major goal of the study is to characterize the long-term effects of transient metabolic perturbation, it is particularly important to address how soon after the treatment (and how soon after removal) of the inhibitor, the authors observed the expected changes of the targeted metabolic pathways. *The cells were cultured in the presence of inhibitors for 4 days, with day 0 being the beginning of the experiment. The effect on chromatin was detectable by ATAC-seq as early as 12 hours. Given the dramatic changes observed at 24h and early changes (detected at the chromatin level and observed in Time-Lapse), it is reasonable to infer that changes occur almost immediately after the addition of the inhibitors. The first time point that was analyzed after the removal of inhibitors was on day 7 (i.e. 3 days culture without inhibitors), then on day 10 and 14. The cells of the four conditions exhibited distinct evolution even after the inhibitors were removed. *

The chromatin-independent and transcriptional-independent mechanisms are not considered. Intermediate metabolites are known to directly modify protein activity, alter cell signaling resulting changes in differentiation potentials. The authors should acknowledge this possibility and examining their data to speculate which specific gene expression and related cell-fate changes are likely (or not likely) the direct result of epigenetic modulation.

We completely agree with the reviewer that cellular memory mechanisms other than chromatin modifications were not investigated. Fluctuations of the energy metabolism can also impact the post-translational modifications of cellular proteins. However almost nothing is known so far on the role of these modifications in cellular memory processes, and in the consolidation of phenotypic characteristics of a cell lineage. This idea is of course very exciting, but studying this aspect would necessitate an entirely separate investigation, using alternative methods. At this stage we believe that this is well out of the scope of the present study. We have added the idea in the Discussion section (page 16).

The samples of primary cells have heterogenic cell populations. The cellular characterization in bulk may confound the results regarding cell-fate programming versus the cell selection effect.

In Fig 3 and Fig6, how would the authors determine whether the inhibitor or rescue treatments alter cell differentiation program or selectively allow proliferation or survival of non-differentiated cells?

The question of the first selective hit followed by the amplification of the surviving cells is highly relevant. The CD34+cell population is inherently very heterogenous, and we used inhibitor concentrations close to the IC50 values. Collectively, we observe that the surviving cells exhibited greater resistance, which is likely due to their more resistant metabolic state. Our metabolic MS analysis was conducted on a bulk population, precluding conclusions at the single-cell level. However, time-lapse, cytometry, single-cell ATAC and RNA-seq analyses all provide information at the single-cell level. ATAC-seq revealed initial differences between control and treated cells approximately 12 hours after stimulation. By 24 hours, 16 different subsets of cells were identified using single-cell ATAC-seq chromatin accessibility profiling. All four conditions were represented in all subsets in variable proportions. Previous studies [1,9] indicated that at 24 hours, these cells couldn't be clustered into distinct groups based on their gene expression patterns, suggesting that chromatin changes precede gene expression changes by several hours. Notably, at the time of analysis, these cells had not undergone division yet. Time-lapse microscopy revealed that the first division occurred in control and 2-DG cells 24 hours later, while in DON and AOA cells, it occurred only around 72 hours later. At this point, single-cell RNA-seq data clustering identified 17 different subsets of cells. Particularly, AOA cells exhibited a distinctly different gene expression pattern, forming separate clusters. Based on these observations, we think that although some selection occurs during the initial hours, the differences observed between the inhibitors cannot be solely explained by it. Instead, chromatin differences between cells appear before the first division of the cells surviving the initial shock. These differences then gradually develop over the initial 96 hours. The inhibitors were removed at this point, and the cells primed by the different inhibitors were subsequently cultured under identical conditions. It is likely that cells exhibiting differential gene expression patterns possessed varying proliferation capacities, contributing to the observed evolution of cell populations as detected on days 7, 10, and 14. We have added this paragraph to the manuscript in the Discussion section for better clarity (pages 14 and 15).

1. Trajectory analysis may further elucidate that the effects of metabolic perturbation on cell differentiation program are permissive or more instructive (towards/against specific lineage commitment). Although we were able to identify 17 subsets of cells based on their transcriptome profiles, any of them could be assigned to a specific hematopoietic lineage. It is presumably too early. As it was shown (Moussy et al 2017), at this stage, just 96 hours after stimulation most of the cells are still “hesitant” with fluctuating gene expression profiles and morphology. Their commitment to a specific lineage is not robust making the definition of trajectories impossible.

Minor:

1. Fig. 1A is missing figure legends. We clarified the legend (see page 40).

The cell clusters in fig 3 needs to be at least deconvoluted based on the differentiation or cell-identity markers and annotated accordingly in the main figure.

Indeed, we conducted this analysis, but the results weren't conclusive enough to be included in the manuscript. We extracted the list of differentially expressed genes for each cluster (for a more detailed description, refer to the answer to Reviewer 2's Question 2 regarding the analysis of cluster 8). The list of extracted biomarkers was studied, and the top 20 for each cluster are shown on the heat-map in S6 Fig. However, for many clusters, canonical markers couldn't be identified to easily match the clusters to known cell types. For others, a few markers were detected, but with inconsistent mixes, such as in cluster 7 (LYZ and CD14 associated with CD14+ Mono, CST3 associated with DC, NKG7 associated with NK, IL7R and S100A4 associated with Memory CD4+, and MS4A7 associated with B cells) or in cluster 12 (PPBP associated with platelets, S100A4 associated with memory CD4+ cells and FCER1A associated with DC). At this very early stage, the cells are just exiting the multi-lineage primed stage, and it's likely that their identity is not yet fully determined, explaining the mix of markers from different lineages. We also attempted a Gene Ontology analysis on the lists of biomarkers, but most terms were general cellular functioning terms, making it impossible to assign the cells in the various clusters to specific cell types.

The statements in abstract and introduction broadly mention the environmental changes and metabolic adaptation in the process of differentiation. The study, however, address only the setting in vitro. As the mobilization of the hematopoiesis process is not possible to be address with the data presented in the current study. The author should revise the manuscript to better introduce relevant questions of the study.

With all due respect, we do not agree with this comment. The question we are seeking the answer to is defined in the Introduction section (page 3): “Does the change of the metabolic setup of the cells precede and trigger the non-specific chromatin opening?”. For better clarity, now we extended this question by a second one (page 3). It is true that in vitro studies cannot reproduce faithfully all the in vivo conditions such as the mobilization of the hematopoiesis process. However, the objective of our study was only to ask if the external restriction of the energy metabolism modifies the cellular differentiation process. From this perspective, utilizing metabolic inhibitors is a possible way to model restricted access to some substrates in a stressful environment. Indeed, this is the entire philosophy and value of in vitro experiments. The time resolution used in this study is impossible to achieve currently in any in vivo setting. The use of human CD34+ cells was motivated by the fact that this is a very well-studied in vitro model that retains many characteristics of cell differentiation in general. We only hope that our hypothesis and the observations done here are robust enough to be generalizable to other models and to cell differentiation in general. Obviously, confirmation by complementary studies on various other cellular models will be required.

Reviewer #1 (Significance (Required)):

Overall, we appreciate the author using untrivial experiments with purified/primary human cells and highly parallel omics analyses to test an interesting hypothesis. However, we think the specific question(s) and objective(s) of the study need to be specified/clarified and to be better addressed by more conclusive results.

This study will be of fundamental interest to the field of stem cell biology, cell metabolism and developmental biology. Our expertise is adult stem cell biology and dietary research.

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

Summary:

The authors evaluate the impact of metabolic perturbations on chromatin structure and the transcriptional landscape of undifferentiated hematopoietic progenitor cells following stimulation with early acting cytokines. Of note, the authors find very early changes in chromatin structure, associated with more long-term changes in transcriptional profiles, modulating the differentiation potential of these progenitors.

Major Comments:

-The authors show a significantly larger impact of AOA than DON on the chromatin and transcription responses of CD34+ progenitors even though they are both impacting glutamine metabolism. Alpha-ketoglutarate rescued CD34+ progenitors from the effect of AOA but did not rescue DON-treated cells which should also have an attenuated generation of alpha-ketoglutarate. How do the authors interpret this apparent discrepancy? In this regard, the MS data are confusing to this reviewer; alpha ketoglutarate levels were much higher in AOA-treated cells than in DON (or even 2-DG-treated) cells, potentially suggesting that DON had more of an impact on glutamine metabolism than AOA. Additionally, glutamine levels are low in DON-treated cells (where GLS is inhibited) but not in AOA-treated cells (this reviewer would have expected higher levels in both) and lactate is high in 2-DG treated cells (low levels would have been expected).

We were surprised by the metabolite levels found by mass spectrometry in the cells at 24 hours. In many cases these levels were different than what one would intuitively expect. This is why we have repeated the experiments many times. One possible explanation is to consider that these metabolites are produced and consumed simultaneously by many different alternative biochemical reactions. Inhibiting one of them induces immediate compensations by others. The metabolic network is complex and its state at a given moment is difficult to predict. Our measurement provides only a snapshot (which are steady-state measurements at that time). The significant change in the abundance of many metabolic intermediates indicates the fact that the network function is perturbed. To understand in detail the exact nature of these perturbations a single time-point measurement is not sufficient, detailed metabolic flux studies will be able to identify the modified metabolic fluxes. This is at present challenging, because the sources of cells are from different patients, at different times, and will require overcoming substantial experimental challenges. More specifically, the reason why AOA had a greater impact on the chromatin than DON and could be rescued by alpha-ketoglutarate may reside in the structure of the glutamine metabolizing pathway. The effect of DON inhibition on alpha-ketoglutarate can be relatively easily compensated by other amino acids, given that glutamine is a non-essential amino acid. This aligns with the observed recovery of surviving cells after an initial setback, where they subsequently resume their proliferation and differentiation following a brief lag period. Conversely, compensating for the inhibition caused by AOA is more challenging due to the direct involvement of transaminases in αKG production.

*The manuscript has been completed in the Results section (page 5) and in the Discussion section (pages 15 and 16). *

-The authors' finding of a single cluster of cells following AOA treatment (cluster 8) is extremely impressive. Can the authors better define this cluster?

Indeed, scRNA-seq analysis at 96hrs revealed very specific transcriptomic profiles for the AOA condition (Fig.3BC). Although some cells appeared in small numbers in clusters common to other conditions (clusters 4, 7, 10 and 13), most were grouped in completely distinct clusters (clusters 8, 11, 14 and 15). In particular, cluster 8 contained 70.2% of the cells from the AOA condition, i.e. 3598 cells out of 5126 analyzed for this condition before normalization. Given the small size of clusters 11, 14 and 15, attention was focused on cluster 8 for further characterization.

*First, we were able to confirm that this cluster was real and significant because even at a lower resolution than that initially used for the study (resolution 0.6 in Fig.3B), the cluster persists, so it is not an artefact of the clustering algorithm (cluster 1 on the figure on the left corresponds to cluster 8 on Fig.3B). *

Overall, the analysis of gene expression profile revealed that the cluster 8 was better defined by the genes that were down regulated rather than those overexpressed compared to the other clusters. However, the Gene Ontology analysis conducted on these gene lists was inconclusive. The extracted biomarkers do not allow for associating the cells with a specific mature cell type, 96hours is too early in the differentiation process. We think that this observation is not sufficiently conclusive at this stage to be included in the manuscript. Deeper analyses would be necessary to better understand their specificity, but it was out of the scope of the present study.

*Here is the detailed description of the analysis: *

*We searched for specific markers to characterize this cluster using the FindAllMarkers function in the Seurat package. This analysis compares each cluster against all others, identifying genes with differential expression. In the generated output, pct.1 represents the proportion of cells within the cluster where a specific gene is detected, while pct.2 signifies the average proportion of cells across all other clusters where the gene is detected. To refine our results, we filter the positive markers, retaining those with a difference > 0.25 between pct.1 and pct.2, alongside a p_val_adj

ID

Ont.

Description

Gene Ratio

geneID

Count

GO:0071392

BP

cellular response to estradiol stimulus

45171

CRHBP/NRIP1

2

GO:0017046

MF

peptide hormone binding

45232

CRHBP/NPR3

2

GO:0042562

MF

hormone binding

45232

CRHBP/NPR3

2

*The study of genes overexpressed in this cluster 8 is therefore inconclusive. When we look at the heatmap with the top 20 markers for each cluster, it seems that cluster 8 is characterized by the under-expression of certain genes, genes that are also under-expressed in clusters 14 and 15 and over-expressed in clusters 11 and 16: GPNMB, LGALS3, MMP9, CTSD, CXCL8, CTSB, SOD2, IFI30, PSAP, CHI3L1, CYP1B1, CSTB, ACP5, MARCKS, S100A11, FCER1G, LIPA. We conducted a Gene Ontology analysis on this new list, and this time, 53 terms were identified. The figure below shows the top 25 terms. Several terms related to immune cells and neutrophils are observed. The standard analysis doesn't provide us with additional insights into the cells within cluster 8. *

-The authors find an increase in cells expressing the CD36 marker, especially following 2-DG treatment. However, they never discuss the functional significance of CD36 as a fatty acid translocase (FAT), serving as a receptor for long chain fatty acids, and potentially as a compensatory mechanism under conditions where glucose metabolism is inhibited. We thank the reviewer for drawing our attention to this omission. It is indeed highly relevant and important to mention it in the paper. It fits perfectly with the basic idea of metabolic adaptation as a driving force. We introduced this point with references in the manuscript in the Results section (page 11).

__Minor Comments: __

-A schematic showing the different inhibitors and metabolic pathways would be helpful. A schematic representation of the main metabolic pathways and the steps affected by inhibitors has been added as S1 Fig (see page 32 and 40). Consequently, the other supplementary figures have been renumbered.

Reviewer #2 (Significance (Required)):

General comments:

The impact of metabolic perturbations on a progenitor cell with the potential to differentiate to multiple lineages is of much interest to the field. The authors have performed extensive single cell analyses, incorporating both scATACseq and scRNAseq together with cell morphology analyses and cell surface protein evaluations, to monitor short- and long-term impacts. They find very rapid changes in chromatin structure with long-lasting effects, despite the cessation of the metabolic perturbation. This has important implications for our understanding of the crosstalk between metabolic alterations, chromatin structure, and gene expression, coming together to regulate progenitor cell survival, expansion, and differentiation.

Assessments: strengths and limitations

Strengths and Advances:

The authors should be commended for their use of primary hematopoietic progenitors and a close evaluation of the impact of metabolic perturbations during the first 24h of stimulation. Their studies have added significantly to our understanding of cell differentiation, showing that changes in metabolic circuits rapidly modulate cytokine-induced epigenetic chromatin states.

Limitations:

Because CD34+ progenitors represent a heterogeneous population, metabolic perturbations are likely impacting the different subsets in distinct manners. The single cell data presented here can be exploited to assess how these subsets (clusters) change at very early time points following perturbation. It will also be important to confirm the effects of different inhibitors on specific metabolites in a cell line(s) since the changes reported here do not appear to be specific. It is possible that these differences are due to an overall decrease in the activation state of a cytokine-stimulated progenitor leading to a global decrease in metabolites.

Audience: This study will be of much interest to scientists/clinicians studying stem cells, hematopoietic stem cells, metabolism, and epigenomic/transcriptomic landscapes. As such, it will be of interest to a large community.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Summary:

The authors evaluate the impact of metabolic perturbations on chromatin structure and the transcriptional landscape of undifferentiated hematopoietic progenitor cells following stimulation with early acting cytokines. Of note, the authors find very early changes in chromatin structure, associated with more long-term changes in transcriptional profiles, modulating the differentiation potential of these progenitors.

Major Comments:

• The authors show a significantly larger impact of AOA than DON on the chromatin and transcription responses of CD34+ progenitors even though they are both impacting glutamine metabolism. Alpha-ketoglutarate rescued CD34+ progenitors from the effect of AOA but did not rescue DON-treated cells which should also have an attenuated generation of alpha-ketoglutarate. How do the authors interpret this apparent discrepancy? In this regard, the MS data are confusing to this reviewer; alpha ketoglutarate levels were much higher in AOA-treated cells than in DON (or even 2-DG-treated) cells, potentially suggesting that DON had more of an impact on glutamine metabolism than AOA. Additionally, glutamine levels are low in DON-treated cells (where GLS is inhibited) but not in AOA-treated cells (this reviewer would have expected higher levels in both) and lactate is high in 2-DG treated cells (low levels would have been expected).
• The authors' finding of a single cluster of cells following AOA treatment (cluster 8) is extremely impressive. Can the authors better define this cluster?
• The authors find an increase in cells expressing the CD36 marker, especially following 2-DG treatment. However, they never discuss the functional significance of CD36 as a fatty acid translocase (FAT), serving as a receptor for long chain fatty acids, and potentially as a compensatory mechanism under conditions where glucose metabolism is inhibited.

Minor Comments:

• A schematic showing the different inhibitors and metabolic pathways would be helpful.

Significance

General comments:

The impact of metabolic perturbations on a progenitor cell with the potential to differentiate to multiple lineages is of much interest to the field. The authors have performed extensive single cell analyses, incorporating both scATACseq and scRNAseq together with cell morphology analyses and cell surface protein evaluations, to monitor short and long term impacts. They find very rapid changes in chromatin structure with long-lasting effects, despite the cessation of the metabolic perturbation. This has important implications for our understanding of the crosstalk between metabolic alterations, chromatin structure, and gene expression, coming together to regulate progenitor cell survival, expansion, and differentiation.

Assessments: strengths and limitations

Strengths and Advances: The authors should be commended for their use of primary hematopoietic progenitors and a close evaluation of the impact of metabolic perturbations during the first 24h of stimulation. Their studies have added significantly to our understanding of cell differentiation, showing that changes in metabolic circuits rapidly modulate cytokine-induced epigenetic chromatin states. Limitations: Because CD34+ progenitors represent a heterogeneous population, metabolic perturbations are likely impacting the different subsets in distinct manners. The single cell data presented here can be exploited to assess how these subsets (clusters) change at very early time points following perturbation. It will also be important to confirm the effects of different inhibitors on specific metabolites in a cell line(s) since the changes reported here do not appear to be specific. It is possible that these differences are due to an overall decrease in the activation state of a cytokine-stimulated progenitor leading to a global decrease in metabolites.

Audience:

This study will be of much interest to scientists/clinicians studying stem cells, hematopoietic stem cells, metabolism, and epigenomic/transcriptomic landscapes. As such, it will be of interest to a large community.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

The current study investigates the metabolic regulation of hematopoietic cell differentiation through chromatin modification and gene expression. Using the primary CD34+ human cord blood cells, the authors show that transient pharmacological inhibition of glycolysis, PPP, and glutamine/glutamate metabolism alters the dynamics of chromatin structures and gene expression, leading to the impacts on cell proliferation, morphology, and the long-term differentiation capacity. Following are specific comments:

Major:

1. The rationale behind the selection of the metabolic targets and the working hypothesis regarding specific effects on cellular consequence is not explicitly conveyed, which makes it difficult to judge if the experiment design is appropriate and if the results address the questions:
   • i. The operational definition of "Metabolic perturbation" or "Metabolic stress" needs to be provided and the validation of inhibitory effects needs to be clarified. Fig. 3D and S1 Fig are supposed to indicate the inhibition of targeted metabolic pathways but it is not clear if the authors believe the inhibitors exert expected metabolic effects based on the presented data. The author should explain why they target the selected pathways (i.e. glycolysis, PPP and glutamine/glutamate metabolism) and precisely point out which up or down regulation (in Fig. 3D and S1 Fig, for example) indicate sufficient and specific inhibitory effects for each inhibitor to operationally define "metabolic perturbation".
   • ii. Given that the major goal of the study is to characterize the long-term effects of transient metabolic perturbation, it is particular important to address how soon after the treatment (and how soon after removal) of the inhibitor, the authors observed the expected changes of the targeted metabolic pathways.
2. The chromatin-independent and transcriptional-independent mechanisms are not considered. Intermediate metabolites are known to directly modify protein activity, alter cell signaling resulting changes in differentiation potentials. The authors should acknowledge this possibility and examining their data to speculate which specific gene expression and related cell-fate changes are likely (or not likely) the direct result of epigenetic modulation.
3. The samples of primary cells have heterogenic cell populations. The cellular characterization in bulk may confound the results regarding cell-fate programming versus the cell selection effect.
   • i. In Fig 3 and Fig6, how would the authors determine whether the inhibitor or rescue treatments alter cell differentiation program or selectively allow proliferation or survival of non-differentiated cells?
   • ii. Trajectory analysis may further elucidate that the effects of metabolic perturbation on cell differentiation program are permissive or more instructive (towards/against specific lineage commitment).

Minor:

1. Fig. 1A is missing figure legends.
2. The cell clusters in fig 3 needs to be at least deconvoluted based on the differentiation or cell-identity markers and annotated accordingly in the main figure.
3. The statements in abstract and introduction broadly mention the environmental changes and metabolic adaptation in the process of differentiation. The study, however, address only the setting in vitro. As the mobilization of the hematopoiesis process is not possible to be address with the data presented in the current study. The author should revise the manuscript to better introduce relevant questions of the study.

Significance

Overall, we appreciate the author using untrivial experiments with purified/primary human cells and highly parallel omics analyses to test an interesting hypothesis. However, we think the specific question(s) and objective(s) of the study need to be specified/clarified and to be better addressed by more conclusive results.

This study will be of fundamental interest to the field of stem cell biology, cell metabolism and developmental biology. Our expertise is adult stem cell biology and dietary research.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

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

Learn more at Review Commons

---

Reply to the reviewers

The two reviewers are very positive and emphasize the relevance of this study. Reviewer 1 notes that “the humoral immune responses but also parasite transcriptomics data is examined for the first time”. Reviewer 2 notes that our study “tries to mimic the infection in nature by reinfecting the Aotus monkeys with different stains of the parasite and then assesses the immune response with main emphasis on antibody response to the infection. This model is important to facilitate vaccine development and understanding the immune response against particular vaccines.”

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

To elucidate whether humoral immunity and/or genetic polymorphisms contribute to the protection against P. vivax blood-stage infection, the Authors assessed whether P. vivax strain-transcendent immunity can be achieved by repeated infection in Aotus monkeys. They infected six Aotus monkeys with blood stages of the P. vivax Salvador 1 (SAL-1) strain until obtaining sterile immunity, and then challenged with the heterologous AMRU-1 strain. Sterile immunity was achieved after two homologous infections, and partial protection against a heterologous AMRU-1 challenge was also achieved. IgG levels against parasite lysate by ELISA and protein microarray increased with repeated infections and correlated with the level of homologous protection. Although there were transcriptional differences in the P. vivax gene repertoire between SAL-1 and AMRU-1, there was no evidence of major antigenic switching upon homologous or heterologous challenge. These findings suggest that the partial protection observed during heterologous challenge is caused by genetic polymorphism between strains, rather than immune evasion by antigenic switching.

Major Comments: 1) Title There are several non-human primate models, therefore, please specify "Aotus monkey model" in the title.

Concur: We have added “Aotus monkey model” to the title.

2) Protein array Lines 373-374 "against major immunogenic blood stage antigens (Ags)" Please add selection criteria for how they select these 244 antigens.

We have added the following paragraph in the methods to address this comment:

“All P. vivax sequences for the array used in this study were derived from the SAL-1 strain, which allowed the evaluation of greater breadth of antigens (but limited evaluation of antigenic variation). Antigens on this array were down selected from larger arrays probed with reactive sera derived from various endemic regions. Only antigens demonstrating seroreactivity across all tested sera were included [105].”

They prepared protein array (n=244) based on the SAL-1 sequence. Please add a discussion of how the data was affected by the sequence difference between SAL-1 and AMRU-1 strains. They described this point only on the top 7 targets (Lines 283-287). Any further difference in antibody reactivity between polymorphic and conserved antigens (SAL-1 and AMRU-1).

We agree with this concern and have added two sentences to the discussion.

*“In comparing the SAL-1 and AMRU-1 strains to the PvP01 reference strain, the sequence data demonstrated clear differences between the isolates in the whole genome analysis. Therefore, this suggests that the current iteration of the microarray (n=244) used in the study did not capture the sequence target(s) responsible for the partial protection observed.” *

Please also add a discussion on how they can interpret their protein microarray data because the E. coli-based IVTT proteins array detects antibody responses against linear epitopes of the printed antigens.

*The IVTT cell-free E.coli express system used to generate the protein microarrays represents an unbiased systems biology approach to antigen identification (Davies DH et al PMID: 26428458). The focus is intentionally on linear epitopes as attempting to capture correctly folded whole proteins is a notoriously difficult venture (Vedadi M et al Mol Biochem Parasitol. PMID: 17125854; Mehlin C et al. Mol Biochem Parasitol. PMID: 16644028). The system has shown proven utility across several disease in identifying important antigenic targets which can then be explored in greater detail using other methods (Wager LE et al. Nat Med. PMID: 33432170; Nakajima R et al. mSphere PMID: 30541779; Virgil A et al. Future Microbiol. PMID: 20143947; Vankatesh A et al Sci Rep. PMID: 35654904). *

The following text and references have been included into the discussion:

“This approach was supported by previous studies which demonstrated the utility of the IVTT platform in high throughput antigen discovery across several disease areas (Jan S et al. Front Immunol PMID: 37533862; Nakajima R et al. mSphere PMID: 30541779; King CL et al. Am J Trop Med Hyg PMID: 26259938; Vankatesh A et al. Methods Mol Biol. PMID: 34115357; Vankatesh A et al. Malar J PMID: 30995911).”

3) Weakness Please summarize the weak points of this study (i.e. small number of animals used) in the Discussion section.

We have added and combined a few phrases with limitations in the discussion section:

“The partial protection observed in the heterologous AMRU-1 challenges may therefore be due to major genetic differences and hence antibody epitope variation between the two strains [50]. In comparing the SAL-1 and AMRU-1 strains to the PvP01 reference strain, the sequence data demonstrated clear differences between the isolates in the whole genome analysis. Therefore, this suggests that the current iteration of the microarray (n=244) used in the study did not capture the sequence target(s) responsible for the partial protection observed. To overcome this limitation and induce high levels of protective antibodies, we propose use of an immunization regime with whole parasite antigen pools from a mixture of genetically diverse strains. Another limitation of this study is the small number of subjects. The study can be considered as exploratory (i.e. looking for patterns of response rather than hypothesis testing [95]), hence the number of subjects used in the only group studied is typical of such exploratory research with humans [35, 96] and NHP [38].”

Minor Comments: 4) Line 129 "inoculation level II" Please reword this to "2nd inoculation" throughout the manuscript because "inoculation level" is a bit confusing for the readers.

Do not concur: It is easier to understand, in the figures in particular. Unless the editor insists, we would rather keep as is.

5) Line 320 "pir genes" Please spell out because this is the first appearance in this manuscript.

Done. Plasmodium interspersed repeat (PIR) genes.

6) Line 373 "IVTT" Please spell out because this is the first appearance in this manuscript.

Done. in vitro transcription/translation reaction (IVTT)

7) Line 404 "VIR antigens" Please spell out because this is the first appearance in this manuscript.

Done. Plasmodium vivax interspersed repeat (VIR) antigens.

8) Line 498 "Goat anti-monkey Rhesus macaque)" This may be HRP-labelled? Please correct.

Concur: We have added HRP labelled to: "Goat anti-monkey Rhesus macaque HRP-labelled"

9) Line 512 "temperature Plates" should be "temperature. Plates". 10) Line 514 "sulphuric acid 2.5M" should be "2.5M sulphuric acid".

Concur: Changed to "2.5M sulphuric acid".

11) Line 516 "Plasmodium falciparum" should be "Plasmodium vivax".

Concur: Changed to "Plasmodium vivax".

12) Line 524 "Escherichia.coli" should be "Escherichia coli".

Concur: Changed to "Escherichia coli".

13) Line 604 "is spleen-dependent (ref)" Please add a reference.

This paragraph has been removed as the data are not included in this study.

14) Line 1099 "core genes" Please add a description of what core genes mean.

Has now been added in the text line 319.

15) Figure S2 Please label each panel in Figure S2 A&B. Maybe I, II, III, IV from the left. Please also revise the label of the X-axis in Figure S2C because "Inoculation level" is misleading.

We have added the labeling to S2A and B.

**Referees cross-commenting**

I agree with Reviewer#2 comments.

Reviewer #1 (Significance (Required)):

1) General assessment: This is a valuable and important study conducted by qualified experts in this research field. All the works were carefully designed, and clearly presented, and the manuscript is well written.

(1) Strongest and most important aspects? Aotus monkey study with intensive data acquisition including humoral immune response and detailed parasite transcriptomic investigation.

(2) Weakness The number of animals used is rather small.

2) Advance: Does the study extend the knowledge in the field and in which way? Not only the humoral immune responses but also parasite transcriptomics data is examined for the first time.

3) Audience: Malariologists will be interested in or influenced by this research The data in this study will be the basis of future whole-parasite-based vaccine development.

My field of expertise is malariology and malaria vaccine research.

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

This study focuses on the development of the model which can be further used as the model for developing a vaccine for the malaria parasite Plasmodium vivax. The researchers infected Aotus monkeys with one strain, achieved immunity, and then exposed them to a different strain. Four monkeys became immune to the initial strain, and three showed partial protection against the second strain. The researchers found that differences in genetic diversity and gene expression between strains are responsible for the varying levels of protection. This study provides insights for testing candidate vaccines against P. vivax. This model is unique and important for facilitating vaccine developments.

• The researchers provide a clear methodology and suitable for the proposed research questions.
• Did researchers observed any gametocytes after inoculations especially in the asymptomatic one or the prolong parasitemia. If they found, whether those gametocyte are infectious?

*We did not focus on gametocytes in this study, hence no mosquito infection experiments were performed. *

Reviewer #2 (Significance (Required)):

The asymptomatic infections are common in malaria endemic areas but it is hard to identify the underlying immune mechanism in response to the disease. This model tries to mimic the infection in nature by reinfecting the Aotus monkeys with different stains of the parasite and then assesses the immune response with main emphasis on antibody response to the infection. This model is important to facilitate vaccine development and understanding the immune response against particular vaccines.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

This study focuses on the development of the model which can be further used as the model for developing a vaccine for the malaria parasite Plasmodium vivax. The researchers infected Aotus monkeys with one strain, achieved immunity, and then exposed them to a different strain. Four monkeys became immune to the initial strain, and three showed partial protection against the second strain. The researchers found that differences in genetic diversity and gene expression between strains are responsible for the varying levels of protection. This study provides insights for testing candidate vaccines against P. vivax. This model is unique and important for facilitating vaccine developments.

• The researchers provide a clear methodology and suitable for the proposed research questions.
• Did researchers observed any gametocytes after inoculations especially in the asymptomatic one or the prolong parasitemia. If they found, whether those gametocyte are infectious?

Significance

The asymptomatic infections are common in malaria endemic areas but it is hard to identify the underlying immune mechanism in response to the disease. This model tries to mimic the infection in nature by reinfecting the Aotus monkeys with different stains of the parasite and then assesses the immune response with main emphasis on antibody response to the infection. This model is important to facilitate vaccine development and understanding the immune response against particular vaccines.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

To elucidate whether humoral immunity and/or genetic polymorphisms contribute to the protection against P. vivax blood-stage infection, the Authors assessed whether P. vivax strain-transcendent immunity can be achieved by repeated infection in Aotus monkeys. They infected six Aotus monkeys with blood stages of the P. vivax Salvador 1 (SAL-1) strain until obtaining sterile immunity, and then challenged with the heterologous AMRU-1 strain. Sterile immunity was achieved after two homologous infections, and partial protection against a heterologous AMRU-1 challenge was also achieved. IgG levels against parasite lysate by ELISA and protein microarray increased with repeated infections and correlated with the level of homologous protection. Although there were transcriptional differences in the P. vivax gene repertoire between SAL-1 and AMRU-1, there was no evidence of major antigenic switching upon homologous or heterologous challenge. These findings suggest that the partial protection observed during heterologous challenge is caused by genetic polymorphism between strains, rather than immune evasion by antigenic switching.

Major Comments:

1. Title There are several non-human primate models, therefore, please specify "Aotus monkey model" in the title.
2. Protein array Lines 373-374 "against major immunogenic blood stage antigens (Ags)" Please add selection criteria for how they select these 244 antigens. They prepared protein array (n=244) based on the SAL-1 sequence. Please add a discussion of how the data was affected by the sequence difference between SAL-1 and AMRU-1 strains. They described this point only on the top 7 targets (Lines 283-287). Any further difference in antibody reactivity between polymorphic and conserved antigens (SAL-1 and AMRU-1). Please also add a discussion on how they can interpret their protein microarray data because the E. coli-based IVTT proteins array detects antibody responses against linear epitopes of the printed antigens.
3. Weakness Please summarize the weak points of this study (i.e. small number of animals used) in the Discussion section.

Minor Comments:

1. Line 129 "inoculation level II" Please reword this to "2nd inoculation" throughout the manuscript because "inoculation level" is a bit confusing for the readers.
2. Line 320 "pir genes" Please spell out because this is the first appearance in this manuscript.
3. Line 373 "IVTT" Please spell out because this is the first appearance in this manuscript.
4. Line 404 "VIR antigens" Please spell out because this is the first appearance in this manuscript.
5. Line 498 "Goat anti-monkey Rhesus macaque)" This may be HRP-labelled? Please correct.
6. Line 512 "temperature Plates" should be "temperature. Plates".
7. Line 514 "sulphuric acid 2.5M" should be "2.5M sulphuric acid".
8. Line 516 "Plasmodium falciparum" should be "Plasmodium vivax".
9. Line 524 "Escherichia.coli" should be "Escherichia coli".
10. Line 604 "is spleen-dependent (ref)" Please add a reference.
11. Line 1099 "core genes" Please add a description of what core genes mean.
12. Figure S2 Please label each panel in Figure S2 A&B. Maybe I, II, III, IV from the left. Please also revise the label of the X-axis in Figure S2C because "Inoculation level" is misleading.

Referees cross-commenting

I agree with Reviewer#2 comments.

Significance

1. General assessment: This is a valuable and important study conducted by qualified experts in this research field. All the works were carefully designed, and clearly presented, and the manuscript is well written.
2. (1) Strongest and most important aspects? Aotus monkey study with intensive data acquisition including humoral immune response and detailed parasite transcriptomic investigation
3. (2) Weakness The number of animals used is rather small.
4. Advance: Does the study extend the knowledge in the field and in which way? Not only the humoral immune responses but also parasite transcriptomics data is examined for the first time.
5. Audience: Malariologists will be interested in or influenced by this research The data in this study will be the basis of future whole-parasite-based vaccine development.

My field of expertise is malariology and malaria vaccine research.

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

Learn more at Review Commons

---

Reply to the reviewers

1. General Statements

We express our gratitude to the reviewers for their time and insightful comments, which have significantly contributed to the enhancement of our manuscript. We believe that the thoughtful critiques and suggestions have substantially improved the overall quality of our work. The changes made in the revised manuscript were highlighted in red. Below, we provide a point-by-point response to each comment, addressing the concerns raised by the reviewers.

2. Point-by-point description of the revisions

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

*Summary: *

*In the current study, Li et al investigated how TGF-beta signaling is controlled by protein abundances. Computational modeling and experiments indicated that the abundance of TGFBR1 and TGFBR2 affects the signaling, and those with lower abundance affect the signaling more, resembling Liebig's law of the minimum. Specifically, they showed that by using multiple cell lines with a different abundance of receptors, modulation of expression of the less abundant receptor impacts the signaling, which is measured by SMAD2 nuclear-to-cytosol ratio and/or relative phospho-SMAD2 level. Also, by using a light-induced interaction system, they showed that the signaling is dependent on the concentration of receptor complex when both receptors are expressed at similar amounts. *

*Major comments: *

*Computational predictions support the authors' idea. The computation and the experiments are well-documented. And it would gain substantially if the authors fill the gap between the predictions and the experiments as follows. *

*In Figure 4, the authors showed that perturbation on receptors with lower expression levels in each cell line changes the phospho-SMAD2 level. Although the data looks consistent with their claim, the result is only qualitative. The authors established a computational model in the former sections, thus it would be of great interest to assess if the experimental results quantitatively match the computational prediction. *

Response: The reviewer suggests that our work could benefit from a quantitative comparison between computational predictions and experimental data shown in Figure 4. We appreciate this suggestion. Given the challenges in obtaining precise quantification of TGFBR1 protein due to antibody issues (see the response to comment #2 from reviewer 2), a direct quantitative comparison between model predictions and experimental results is difficult. Our model predictions about the control principle with Liebig's law of the minimum should be interpreted qualitatively, rather than a strict quantitative law. We have explicitly indicated in the revised manuscript that our siRNA knockdown experiments are to qualitatively test our model predictions.

*In Figure 5, the authors computationally predicted that the expression level of receptors is correlated with SMAD2 N2C levels 1 hour after stimulation, and the strength of negative feedback with SMAD2 N2C levels 8 hours after stimulation. Because the authors employed iRFP-SMAD2 system, the prediction could be verified experimentally, at least the prediction on SMAD2 N2C 1 hour after stimulation could be checked. (In a sense, this is partially verified by the data in Figure 7, where both receptors are expressed at similar levels). It would gain substantially if the authors could verify the computational prediction in Figure 6. Since the authors stated in the introduction that "The same TGF-beta ligand can initiate different signaling responses depending on the cellular context, but the underlying control principle remains unclear...Together, these results revealed an effect of the minimum control in the TGF-beta pathway, which may be an important principle of control in signaling pathways with context-dependent outputs.", experimental verification of the prediction done in Figures 4-6 will be very important. Or the authors should stress that these points are only predicted by computational models. *

__Response: __The reviewer recommends verifying the model predictions in Figure 6 experimentally, particularly regarding SMAD2 N2C levels 1 hour after stimulation. We appreciate this valuable suggestion, which was also raised by reviewer 2. In response, we conducted experiments as recommended by reviewer #2, in which imbalanced expression of TGFBR1 and TGFBR2 was achieved by transfecting optoTGFBR1 or optoTGFBR2 plasmids into optoTGFBRs-HeLa cells, which initially expressed similar levels of both receptors. Western blot analysis confirmed the desired imbalance (Figure S13).

Consistent with the model predictions (Figure 6), the strong correlation between SMAD2 N2C fold change response at 1h and optoTGFBR2-tdTomato expression levels persisted in single cells when optoTGFBR1 was overexpressed (Figure 8A). Conversely, the high correlation between nuclear SMAD2 signaling and optoTGFBR2-tdTomato expression levels vanished at single cell level when optoTGFBR2 was overexpressed (Figure 8B). These experimental results validate our model predictions, confirming that the SMAD2 signaling is determined by the low abundance TGF-beta receptor in single cells. Incorporating these experimental validations enhances the quantitative support for our model predictions and clarifies the relationship between TGF-beta receptor abundance and signaling outcomes in single cells.

*As written in the below "Significance" section, the result is, in a sense, obvious. It should be stated that because the study utilized a slightly high concentration of TGF-beta in the experiments, it might be natural that the low-abundance receptor becomes a bottleneck of the signaling. It would gain to assess how receptor abundance affects signaling with the stimulation of lower concentrations of TGF-beta, or to examine the computational model if the low abundance of a receptor becomes a bottleneck of signaling because of saturation. Also, it is highly recommended to discuss the physiological implication of the current study, taking into account the experimental conditions used. *

Response: We appreciate the reviewer's insightful comments regarding the concentration of TGF-beta used in our experiments and the potential influence on the model predictions. In our experiments and model simulations, we utilized 100 pM TGF-beta, equivalent to 2.5 ng/mL (not 4.4 ng/mL as calculated by the reviewer). This concentration is a widely used dose in TGF-beta signaling studies. The reviewer's suggestion to explore how varying TGF-beta concentrations might influence the minimum control concept prompted us to extend our computational simulations. We used the extended model to perform simulations with lower TGF-beta concentrations (25 pM, equivalent to 0.625 ng/mL, and 10 pM, equivalent to 0.25 ng/mL). The results, depicted in Figure S7 of the revised manuscript, reaffirm that even at lower TGF-beta stimulations, a low abundance of a TGF-beta receptor acts as a bottleneck for SMAD2 signaling.

Following the reviewer’s suggestion, we have incorporated additional paragraphs to discuss the physiological implications and potential limitations of our study (Page 16-17 in the Main text).

It is pertinent to note that while the concept of TGF-beta signaling response being dictated by the minimum abundance of TGF-beta receptors may seem intuitive or even obvious, theoretical and experimental validations are crucial. As demonstrated in Figure S1B, our new simulation results from the minimal model illustrate similar response profiles when a high binding affinity (K1) is set for ligand-receptor interactions (Figure S1A). However, with a small binding affinity (K1), the minimal model indicates that TGF-beta signal response remains proportional to the product of TGFBR1 and TGFBR2 abundance and can be sensitive to the change of high abundance receptor in some region (Figure S1B). This highlights that the observed response patterns aligning with Liebig's law of the minimum depend on the binding affinity of ligand-receptor interactions in our minimal model. Consequently, the intuitive idea about Liebig's law of the minimum is not necessarily true theoretically. Moreover, given the non-linearity of the TGF-beta network, this complexity introduces an additional layer of uncertainty regarding the applicability of the minimum control principle to TGF-beta responses. This uncertainty led us to develop an extended model, with parameter values either experimentally measured or estimated from time course experimental data. The extended model predicted a similar minimum control principle at the TGF-beta receptor level, inspiring us to validate this prediction through diverse experiments. While we acknowledge the intuitive nature of our findings, we believe it is important for the field to prove this expectation, as emphasized by reviewer 4.

Reviewer #1 (Significance (Required)):

*TGF-beta signaling is one of the most rigorously studied pathways both computationally and experimentally. As written in the introduction of the manuscript, it is still unknown how the variability of responses arises not only between cell types but also differences among cells of single cell type. Studies showed that protein abundance accounts at least partly for a source of cell variability in TGF-beta signaling. While former studies examined the variability in SMAD protein abundance, the uniqueness of this study is that it focused on the abundance of TGF-beta receptors. *

*Given that both TGFBR1 and TGFBR2 are involved in the signaling, however, it's not difficult to imagine that a less abundant receptor affects the signaling more than the other, and serves as a bottleneck for the signaling. Specifically, because a slightly high concentration (100pM = 4.4 ng/mL of TGF-beta; other studies used much lower conc., e.g. 0, 0.03, 0.04, 0.07, and 2.4 ng/mL in Frick et al, PNAS, 2017, and 0, 1, 2.5, 5, 25, and 100 pM in Strasen et al, Mol Syst Biol, 2017) is used throughout the experiments to check cell-cell variability and the effect of receptor abundance in the current study, the formation of the receptor-ligand complex may be quite fast and be saturated at the level where the receptor with lower abundance is exhausted. In the reviewer's humble opinion, the authors' statement that this is Liebig's law of the minimum sounds a bit exaggerated. *

Nevertheless, the study is of some value because it utilized both computational and experimental analysis to show it is indeed the case. Of note, the current study showed that the variability in the different proteins leads to the variability in different time points, namely, the variability in the receptor abundance leads to the variability 1 hour after stimulation, while that in negative feedback strength leads to the variability 8 hours after stimulation. If the authors fill a small gap between their computational analysis and experimental verification, the study will be of interest to the specialist in the field.

__Response: __We are grateful for the valuable feedback provided by the reviewer. The concerns related to the TGF-beta dose have been thoroughly addressed in our responses to previous comments. Regarding the observation that the term "Liebig's law of the minimum" may sound a bit exaggerated, we acknowledge this consideration. We have refined the title to "Liebig’s Law of the Minimum in the TGF-β/SMAD Pathway," specifying its relevance to SMAD signaling exclusively, as non-SMAD signaling was not within the scope of this study. We appreciate the reviewer's constructive feedback and hope these adjustments enhance the specificity and accuracy of our manuscript.

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

Li et al. present an interesting and intuitive concept for the sensitivity and heterogeneity of biological networks: When two or more proteins form a functional complex, it is the limiting component with the lowest concentration that is most sensitive to perturbations and whose fluctuations dictate cell-to-cell variability of complex function. The authors apply this concept to the TGFb pathway and discuss sensitivity of SMAD signaling towards TGFb receptor I and II fluctuations. The paper is clearly written and convincing, but some improvements in the experimental validation would be beneficial as detailed further below.

1) The authors claim that the ratio of TGFb receptor I and II is very different across cell lines (Fig. 1) and use this observation for the validation of their model in Fig. 4. However, the relative expression TGFb receptor levels are purely based on RNAseq data which does not necessarily imply similar behavior at the protein level, especially on the cell surface. To address this issue, the authors should ideally provide absolute Western blot measurements of TGFbRI at the protein level to complement their absolute quantification of TGFbRII (Fig. S2). At the very least they should show that the observed relative expression levels of TGFbRI and II at the protein level (Figure S7) are correlated to differences in RNA levels (Fig. 1) using protein quantification. They should also confirm that similar receptor ratios for these receptors at the RNA level are observed in other published RNAseq datasets of the same cell lines (e.g., ENCODE for HepG2 and published RNAseq studies in HaCaT). Furthermore, they might take into account published mass spec datasets for quantifications of TGFbR protein levels.

Response: We appreciate the reviewer's thorough evaluation and constructive suggestions.

(A) Absolute quantification of TGFBR1: We acknowledge the importance of obtaining absolute quantification of TGFBR1 protein similar as what we have done for TGFBR2 protein (Figure S2). Despite significant efforts, our attempts to achieve this were hindered by challenges with available TGFBR1 antibodies and recombinant TGFBR1 proteins. Many commercial antibodies failed negative controls with TGFBR1 knockdown samples, while others validated TGFBR1 antibodies could not recognize the available recombinant TGFBR1 protein standards.

Although many mass spectrometry proteomics data available for different cell lines, it is difficult to convert these MS quantitative values to absolute protein abundance as mentioned in a recent publication (Nusinow et al.,bioRxiv 2020.02.03.932384): “Importantly, these values are all relative values to the other values for that same protein and not absolute values. This means that comparing the levels of different proteins to each other without using something like a correlation to standardize values won’t produce meaningful results.”

We share the reviewer's concern and fully agree that obtaining this absolute quantification is crucial. However, at the present stage, technical limitations prevent us from providing this information for TGFBR1. We commit to pursuing this aspect when feasible in the future.

(B) Validation of relative TGF-beta receptor expression ratios: Following the reviewer's suggestion, we conducted additional analyses to validate the relative expression ratios of TGFBR1 and TGFBR2 using different RNA-Seq databases. The results, presented in Table S1, demonstrate consistent imbalances in TGFBR1-to-TGFBR2 ratios across HepG2 and RH30 cell lines from various data sources, reinforcing the reliability of our observations.

(C) Correlation between RNA and protein expression: We appreciate the reviewer highlighting the challenges associated with correlating RNA and protein expression. Indeed, the correlations between RNA and protein levels vary widely, and direct comparisons can be challenging. To address this, we referenced a recent study (Nusinow et al., Cell 2020, 180:387), which reported that the protein data of TGFBR1 and TGFBR2 were highly correlated with the corresponding RNA data from the same cell line (Spearman’s correlation: 0.672 for TGFBR1, 0.771 for TGFBR2) based on quantitative proteomics and RNA expression data from 375 cancer cell lines.

2) Figure 4: To better judge the reproducibility of the knockdown titration, it would be good to show the different siRNA concentrations as a color code- Alternatively, TGFBR expression could be plotted as a function of the siRNA concentration in a Supplemental Figure, showing the effects of individual replicates.

Response: We thank the reviewer for the suggestion to enhance the clarity of the knockdown titration data. In response, we have now presented the quantified experimental data from three replicates with different colors in Figure 4. Additionally, we have created Figure S9 that plots the expression levels of relative TGFBR1 and TGFBR2 as a function of siRNA concentration, providing a more detailed view of the effects across individual replicates.

3) The simulations in Figs. 5 and 6 show that SMAD signaling fluctuations are mainly determined by cell-to-cell variability of receptor levels when using the SMAD nucleocytoplasmic ratio as a readout, and this is especially true for early time points. For downstream cellular responses, the absolute concentration of phosphorylated SMAD (complexes) in the nucleus is likely more relevant. Based on the authors work and evidence from the literature, I expect that this quantity will likely be heavily be influenced by receptor levels as well, but fluctuations in SMAD expression will play an important role as well. The authors should discuss this issue, and clarify that normalized quantities like SMAD N2C and pSMAD/SMAD mostly characterize receptor-level fluctuations while filtering SMAD fluctuations.

__Response: __We acknowledge the importance of discussing the relevance of different readouts in our study. In the revised manuscript, we have incorporated a discussion addressing this issue. Specifically, we highlight that while the SMAD nucleocytoplasmic ratio is sensitive to cell-to-cell variability in low abundance receptor levels, the absolute concentration of phosphorylated SMAD in the nucleus may be more relevant for downstream cellular responses (e.g.: gene expression). We have cited the work by Lucarelli et al, which demonstrated that variations in SMAD abundance could modulate the balance of different SMAD complexes, thereby regulating heterogeneous gene expression in diverse cell types (Lucarelli et al., Cell Systems 2018).

4) The single-cell measurements in Fig. 7 are interesting, but can only partially be seen as a direct validation of the model predictions, as it seems expected that varying the total input by introducing co-fluctuations in both receptors heavily influence the SMAD level. Wouldn't it be possible to design more specific validation experiments, in which the receptor co-expression construct (Fig. 7C) is used for baseline optoTGFBR expression and combined with an individual expression construct for one of the opto-receptors? This way, the authors could establish different regimes, in which one of the two receptors becomes dominant, and the impact fluctuations could be analyzed in a larger receptor expression space. Of course, a full validation of all possible scenarios is not necessary, but it would, for instance, be valuable to see whether the strong dependency of SMAD signaling of TGFBR2 levels vanishes when TGFBR2 is expressed at a higher level than TGFBR1.

Response: We appreciate the insightful comments and suggestions provided by the reviewer. Based on these recommendations, we have conducted additional experiments to further validate our model predictions. Reviewer 1 also raised this point, we quote our aforementioned response here: “consistent with the model predictions (Figure 6), the strong correlation between SMAD2 N2C fold change response at 1h and optoTGFBR2-tdTomato expression levels persisted in single cells when optoTGFBR1 was overexpressed (Figure 8A). Conversely, the high correlation between nuclear SMAD2 signaling and optoTGFBR2 expression levels vanished at single cell level when optoTGFBR2 was overexpressed (Figure 8B). These experimental results validate our model predictions, confirming that the SMAD2 signaling is determined by the low abundance TGF-beta receptor in single cells. Incorporating these experimental validations enhances the quantitative support for our model predictions and clarifies the relationship between TGF-beta receptor abundance and signaling outcomes in single cells.”

**Referees cross-commenting**

Comments from R2: I agree with most comments of the other reviewers, and highlight the most important overlaps with my comments below.

I agree with R1 that the model validation in Fig. 7 is incomplete and think that this will be a key point to improve the quality of the manuscript (see also my reviewer comment 4)

In line with R3 and R4, I think that the SMAD N/C simulations do not necessarily imply effects on TGFb target gene expression, cell fate decisions or human pathologies. The significance of the results for cellular behavior should be discussed (see also my comment 3)

__Response: __We are grateful for the reviewer's thoughtful comments. These comments have been now addressed (see our responses to the corresponding comments).

Reviewer #2 (Significance (Required)):

The manuscript presents an interesting and intuitive concept for the sensitivity and heterogeneity of biological networks. The authors apply this concept to the TGFb pathway and discuss sensitivity of SMAD signaling towards TGFb receptor I and II fluctuations.

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

*Summary: *

*This is an interesting study that examines the output of the TGF-Beta pathway and how abundance/dosage can determine the signaling response in single cells across multiple cell types. The study is primarily mathematical. The focus is on the Type 1 and 2 TGF-Beta receptors driving nuclear SMAD2 expression. The authors observe that SMAD2 phosphorylation is sensitive to variations in the lower levels of either receptor but robust at variations of high abundance of the receptor reflected through SiRNA experiments shown in Figure 4. Their conclusion is that the feature is consistent with Liebig's law of the minimum- where in this case- a low abundance of the receptor serves as the rate-limiting step in signaling for this pathway. *

*Major comments: *

*- While the data as presented are interesting, it is unclear as to whether the abundance regulates biological function. SMAD2 phosphorylation is shown with some nuclear translocation. However, TGF-Beta target gene activation is not shown, and this needs to be completed. *

Response: We appreciate the reviewer's constructive comment. We have conducted new experiments and included quantitative real-time PCR data in the revised manuscript to evaluate the impact of TGFBR1 and TGFBR2 knockdown on the expression of TGF-beta target genes, such as SMAD7, PAI1, and JUNB. The results, presented in Figure S11, demonstrate differential sensitivity of these genes to the downregulation of TGFBR1 and TGFBR2 in various cell lines (HaCaT, HepG2, and RH30). Specifically, the expression of SMAD7, PAI1, and JUNB is sensitive to TGFBR2 knockdown in RH30 cells, while it is sensitive to TGFBR1 knockdown in HepG2 cells. HaCaT cells, expressing similar levels of both receptors, show comparable sensitivities to reductions in both TGFBR1 and TGFBR2. These findings provide additional insights into the regulatory role of TGF-beta receptor abundance on downstream target gene activation, complementing our study's focus on SMAD2 phosphorylation and nuclear translocation.

*- In addition, it is unclear as to what happens to SMAD3 and SMAD4 which are expressed endogenously in this setting. How are these other TGF-Beta signaling molecules addressed by these observations? *

__Response: __Thank you for bringing up this important point. In our study, the expression levels of endogenous SMAD2 and SMAD4 were found to be similar across HaCaT, RH30, and HepG2 cells. However, SMAD3 expression was notably lower in RH30 and HepG2 compared to HaCaT cells. The central conclusion of our study is based on the observed common control principle, which hinges on the relative expression levels of TGFBR1 and TGFBR2. Consequently, the applicability of this principle is more pertinent when comparing signal responses within the same cell type.

We acknowledge the relevance of endogenous SMAD proteins, and in the revised manuscript, we have expanded our discussion on how differences in SMAD protein expression levels and potential mutations (page 16 in main text), as observed in certain cancers, could influence the formation of homo- and hetero-oligomeric SMAD complexes. These considerations contribute to a more comprehensive understanding of downstream gene expression responses, as discussed in the work of Lucarelli et al. (Cell Systems 2018).

*-Specific biological readouts- cell differentiation etc. are not examined and would need to be provided and discussed. Therefore, the claims put forward while interesting require additional experiments examining SMAD2 target gene activation and biological readouts. *

__Response: __We appreciate this valuable suggestion. While we acknowledge the importance of exploring long-term biological responses, including cell differentiation, it is crucial to note that specific biological readouts are not solely dependent on SMAD signaling; they also involve other non-SMAD signaling pathways. Additionally, these responses are highly cell type-specific. Undertaking extensive investigations into these responses would extend beyond the current scope of our work. Nevertheless, we have discussed this topic in the revised manuscript (page 16 in main text).

Following the reviewers’ suggestion on examining TGF-beta target genes, we have performed experiments examining the expression of SMAD7, PAI1, and JUNB with respect to the changes of TGFBR1 and TGFBR2, respectively (see our response to the first major comment of this reviewer).

*- Lastly, statistical analyses are not provided and would need to be provided. For instance, in Figure 4, how many experiments were replicated and statistical analysis performed for this Figure? *

__Response: __In addressing this concern, we conducted three siRNA knockdown titration experiments for each cell line, as detailed in the figure legend. Due to batch effects, different percentages of TGF-beta receptors were knocked down in different experiments using the same concentration of siRNA. To transparently present the data, we utilized a scatter plot. Following the suggestion from reviewer 2, we have further enhanced the clarity of our data presentation by labeling the results of different experiments with a color code. In addition, we have performed statistical analysis of TGF-β receptor fold-change effects leading to a 50% reduction in the P-Smad2 response compared to that in the non-targeting siRNA control group (EC50) during siRNA knockdown experiments (Figure S10). The results of this analysis unveil significant differences in the sensitivities of pSMAD2 responses to variations in TGFBR1 and TGFBR2 within RH30 and HepG2 cells.

Reviewer #3 (Significance (Required)):

*- Conceptually this is an important study because dosage is a prominent issue in TGF-Beta signaling. For instance, in my field of expertise- mouse models of TGF-beta signaling e.g. SMAD2 knockouts- the cancer phenotypes are evident in haploid animals. Yet how and why dosage plays such a large role in tumorigenesis remains unclear. *

__Response: __We sincerely appreciate your recognition of the conceptual importance of our study in addressing the dosage-related complexities of TGF-beta signaling. Your insights into dosage effects in mouse models, particularly in haploid animals, highlight the relevance of our work underlying tumorigenesis. We have incorporated relevant citations and expanded our discussion in the revised manuscript, providing additional context to the importance of dosage in tumorigenesis (page 18 in main text).

Reviewer #4 (Evidence, reproducibility and clarity (Required)):

Summary: In this study, Li and co-workers combined computational modeling and experimental analysis to study the dependence of the output of the TGF-beta pathway on the abundance of signaling molecules in the pathway, mainly the most upstream regulators of SMAD2, TGFbeta type I and type II receptors. They showed by a combination of biochemical studies (mainly pSmad2 WB and type I/II receptor expression profiling) in HaCaT and HeLa cells as well as stable optogenetical receptor variants expressed by those cell lines, that TGF-beta receptor abundance influences signaling outputs using the concept of Liebigs law of the minimum, meaning that the output-modifying factor is the signaling protein that is most limited, to determine signaling responses across cell types and in single cells.

*Major comments: *

The study is very interesting, the combination of biochemistry and computational modeling to better understand the compexity of the TGFbeta pathway is very much required in the field and should stimulate others to further expand this approach.

__Response: __Thank you for the positive evaluation of this work.

*However, the authors must further explain that the model depicted here to explain pathway kinetics and dynamics lacks multiple crossroads and feedbacks and is until now oversimplified in the manuscript. They have mentioned receptor internalization and recycling, nuclear import and export of SMAD protein, and the feedback regulations e.g. by SMADs regulating receptor expression. Beyond, there is non- SMAD signaling (Derynck et al.; SMAD Linker regulation, deRobertis et al.), different receptor oligomerization modes (Ehrlich/Henis et al.) and heteromeric receptor complexes of TGFbeta receptors known (Hill et al.), that further diversify beyond these mentioned mechanisms. It is understandable that the mathematical model cannot include those considerations to date, however, they must be further explained and commented on to allow that this model can be expanded in the future. *

Response: We acknowledge that there are multiple crossroads and feedbacks that exist in the TGF-beta signaling pathway that have not been explicitly incorporated into our model. We appreciate the reviewer's understanding that current model cannot include these considerations and his/her suggestions for potential future extensions. In the revised manuscript, we have mentioned one of the limitations of our model: non-Smad signaling and crosstalk with other signaling pathways were not considered for simplicity. We have also discussed how to expand this model by including these regulations when more quantitative data are available in the future (page 16-17 in main text).

*A myriad of research labs focus on these intricate fine tuning ot the TGFbeta pathway by those mechanisms which makes the difference between "good" TGFbeta signaling and "bad" TGFbeta signaling in different context and this complexity must be acknowledged by more introduction and discussion. *

Response: In the revised manuscript, we have added an introduction and discussion about the dual role of TGF-beta signaling (page 4 and page 18 in main text).

*The model here will be important to explain *

*A: the mode of heterooligomeric TGFbeta/BMP receptor assemblies as e.g. found in pathological conditions and *

B: Can maybe explain the formation of mixed SMAD complexes as activated by lateral signaling comprising TGFbeta *and BMP receptors once one receptor is of lower abundance to form a high affinity complex. *

*It is therefore required to comment on these aspects at multiple points in the manuscript. *

*It is very important that the visual model used in this manuscript depicts on the possibility, that a TGFbeta type I receptor can team up with e.g. another TGFbeta type I receptor together with two TGFbeta type II receptors but also with an activin type II receptor or that a BMP type I receptor (e.g. ALK1) can form heterooligomeric complexes with ALK5 (TGFbeta type I). *

__Response: __Thank you for this comment. We cited the relevant work (Ramachandran et al, eLife 2018; Szilagyi et al, BMC Biology 2022) and added a discussion about the complexity of the mode of heterooligomeric TGFbeta/BMP receptor assemblies and its effect on the induction of mixed SMAD complexes (page 17 in the main text).

*While the use of optogenetical TGFbeta receptor biosensors is highly interesting, their mode of oligomerization is not yet fully described. It is not known if those biosensors behave like wt receptors in terms of oligomerization and ligand binding. This should be mentioned somewehere. For this reason, the authors should also consider to draw the TGFbeta receptor complex in the cartoons with more detail towards the heterooligomeric assembly that is standard to the field. *

__Response: __The reviewer is correct that the optogenetic TGF-beta receptors might behave differently from the natural TGF-beta receptor system in terms of ligand binding. We have added this point in the Discussion part to highlight the potential difference between the optogenetic TGF-beta systems and the wild-type system (page 16 in the main text).

*While the general finding is not surprising (manipulating the receptor with the lowest abundancy has the biggest impact on signaling output) the methods and models used here are very important to the field to proof that this expectation is actually true and can be experimentally addressed by a combination of bioinformatics and biochemistry. The model developed will be valuable to expand to much more complex and interesting questions in TGFbeta signaling and possibly also BMP signaling e.g. in pathological context (see below). *

*Minor comments: *

*The authors should discuss their findings in the context of: *

• non-Smad signaling outputs (similar or different to the observations on pSMAD2)*
• What do these findings mean for e.g. human pathologies, where type I or type II receptor expression is altered? *
• Can those findings integrate into the "switch" in TGFbeta signaling? *
• How do these findings translate towards BMP SMAD 1/5/9 signaling? * Response: First, we sincerely appreciate the reviewer’s recognition that our work is very important to the field in proving that manipulating the receptor with the lowest abundance has the biggest impact on signaling output. The reviewer’s suggestions about discussing our work in the context of non-Smad signaling, BMP SMAD1/5/9 branch, and the relevance to the dual role of TGF-beta signaling are all constructive. We have incorporated these suggestions and discussed them in the revised manuscript (page 17 in the main text).

Reviewer #4 (Significance (Required)):

*The manuscript is novel and interesting, partiular the combination of bioinformatical and biochemical approaches. The use of optogenetics is state-of-art while some more care should be given to interpretation of results with optogenetical TGfbeta receptor biosensors, is is not known if they really behave similar in terms of receptor oligomerization and signaling. Also it is not shown how their interactome in terms of effector proteins looks like that can potentially influence SMAD signaling output (e.g. Phosphathases to SMADs known to interact with wt receptors). *

*The models drawn need to depict more accurately on the nature of type I and type II receptor complexes (heterotetrameric) and high affinity towards the ligand. The current versions are too oversimplified at this stage. The pathway crosstalks and feedbacks need to be more visible, in order for non experts to not draw too simple conclusions from the visual representations presented in this MS. Particularly the work by Hill and co-workers on receptor oligomerization and SMAD shuttling and feedback need to be included. *

Overall, the manuscript is very significant to the field.

__Response: __We would like to thank the reviewer again for his/her positive evaluation of the novelty and significance of our work. We have taken the reviewer's comments into consideration and made revisions to the manuscript. We now provide more information on the limitations of our current model and the optogenetic TGF-beta receptor biosensors in the Discussion section. We have also included more details about the receptor complex nature and the high affinity towards the ligand. The ligand receptor complex in the model is now drawn as heterotetrametric complex (1 ligand dimer with two TGFBR1s and two TGFBR2s). Additionally, we have incorporated information about pathway crosstalks and feedbacks, giving a more comprehensive view for non-experts. The work by Hill and co-workers on receptor oligomerization, SMAD shuttling, and feedback has been included in the revised manuscript to provide a more complete and accurate representation of the current knowledge in the field.

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

Learn more at Review Commons

---

Referee #4

Evidence, reproducibility and clarity

Summary:

In this study, Li and co-workers combined computational modeling and experimental analysis to study the dependence of the output of the TGF-β pathway on the abundance of signaling molecules in the pathway, mainly the most upstream regulators of SMAD2, TGFbeta type I and type II receptors.

They showed by a combination of biochemical studies (mainly pSmad2 WB and type I/II receptor expression profiling) in HaCaT and HeLa cells as well as stable optogenetical receptor variants expressed by those cell lines, that TGF-β receptor abundance influences signaling outputs using the concept of Liebigs law of the minimum, meaning that the output-modifying factor is the signaling protein that is most limited, to determine signaling responses across cell types and in single cells.

Major comments:

The study is very interesting, the combination of biochemistry and computational modeling to better understand the compexity of the TGFbeta pathay is very much required in the field and should stimulate others to further expand this approach.

However, the authors must further explain that the model depicted here to explain pathway kinetics and dynamics lacks multiple crossroads and feedbacks and is until now oversimplified in the manuscript. They have mentioned receptor internalization and recycling, nuclear import and export of SMAD protein, and the feedback regulations e.g. by SMADs regulating receptor expression. Beyond, there is non- SMAD signaling (Derynck et al.; SMAD Linker regulation, deRobertis et al.), different receptor oligomerization modes (Ehrlich/Henis et al.) and heteromeric receptor complexes of TGFbeta receptors known (Hill et al.), that further diversify beyond these mentioned mechanisms. It is understandable that the mathematical model can not include those considerations to date, however, they must be further explained and commented on to allow that this model can be expanded in the future. A myriad of research labs focus on these intricate fine tuning ot the TGFbeta pathway by those mechanisms which makes the difference between "good" TGFbeta signaling and "bad" TGFbeta signaling in different context and this complexity must be acknowledged by more introduction and discussion.

The model here will be important to explain

A: the mode of heterooligomeric TGFbeta/BMP receptor assemblies as e.g. found in pathological conditions and

B: Can maybe explain the formation of mixed SMAD complexes as activated by lateral signaling comprising TGFbeta nd BMP receptors once one receptor is of lower abundance to form a high affinity complex.

It is therefore required to comment on these aspects at multiple points in the manuscript.

While the use of optogenetical TGFbeta receptor biosensors is highly interesting, their mode of oligomerization is not yet fully described. It is not known if those biosensors behave like wt receptors in terms of oligomerization and ligand binding. This should be mentioned somewehere.

For this reason, the authors should also consider to draw the TGFbeta receptor complex in the cartoons with more detail towards the heterooligomeric assembly that is standard to the field.

It is very important that the visual model used in this manuscript depicts on the possibility, that a TGFbeta type I receptor can team up with e.g. another TGFbeta type I receptor together with two TGFbeta type II receptors but also with an activin type II receptor or that a BMP type I receptor (e.g. ALK1) can form heterooligomeric complexes with ALK5 (TGFbeta type I).

While the general finding is not surprising (manipulationg the receptor with the lowest abundancy has the biggest impact on signaling output) the methods and models used here are verxy important to the field to proof that this expactation is actually true and can be experimentally adressed by a combination of bioinformatics and biochemistry. The model developed will be valuable to expand to much more complex and interesting questions in TGFbeta signaling and possibly also BMP signaling e.g. in pathological context (see below).

Minor comments:

The authors should discuss their findings in the context of: 1. non- Smad signaling outputs (similar or different to the observations on pSMAD2) 2. What do these findings mean for e.g. human pathologies, where type I or type II receptor expression is altered? 3. Can those findings intergate into the "switch" in TGFbeta signaling? 4. How do these findings translate towards BMP SMAD 1/5/9 signaling?

Significance

The manuscript is novel and interesting, partiular the combination of bioinformatical and biochemical approaches. The use of optogenetics is state-of-art while some more care should be given to interpretation of results with optogenetical TGfbeta receptor biosensors, is is not known if they really behave similar in terms of receptor oligomerization and signaling. Also it is not shown how their interactome in terms of effector proteins looks like that can potentially influence SMAD signaling output (e.g. Phosphathases to SMADs known to interact with wt receptors).

The models drawn need to depict more accurately on the nature of type I and type II receptor complexes (heterotetrameric) and high affinity towards the ligand. The current versions are too oversimplified at this stage. The pathway crosstalks and feedbacks need to be more visible, in order for non experts to not draw too simple conclusions from the visual representations presented in this MS. Particularly the work by Hill and co-workers on receptor oligomerization and SMAD shuttling and feedback need to be included.

Overall, the manuscript is very significant to the field.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

Summary:

This is an interesting study that examines the output of the TGF-Beta pathway and how abundance/dosage can determine the signaling response in single cells across multiple cell types. The study is primarily mathematical. The focus is on the Type 1 and 2 TGF-Beta receptors driving nuclear SMAD2 expression. The authors observe that SMAD2 phosphorylation is sensitive to variations in the lower levels of either receptor but robust at variations of high abundance of the receptor reflected through SiRNA experiments shown in Figure 4. Their conclusion is that the feature is consistent with Liebig's law of the minimum- where in this case- a low abundance of the receptor serves as the rate-limiting step in signaling for this pathway.

Major comments:

• While the data as presented are interesting, it is unclear as to whether the abundance regulates biological function. SMAD2 phosphorylation is shown with some nuclear translocation. However, TGF-Beta target gene activation is not shown, and this needs to be completed.
• In addition, it is unclear as to what happens to SMAD3 and SMAD4 which are expressed endogenously in this setting. How are these other TGF-Beta signaling molecules addressed by these observations?
• Specific biological readouts- cell differentiation etc. are not examined and would need to be provided and discussed.
• Therefore, the claims put forward while interesting require additional experiments examining SMAD2 target gene activation and biological readouts.
• Lastly, statistical analyses are not provided and would need to be provided. For instance in Figure 4, how many experiments were replicated and statistical analysis performed for this Figure?

Significance

• Conceptually this is an important study because dosage is a prominent issue in TGF-Beta signaling.
• For instance, in my field of expertise- mouse models of TGF-beta signaling e.g. SMAD2 knockouts- the cancer phenotypes are evident in haploid animals. Yet how and why dosage plays such a large role in tumorigenesis remains unclear.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Li et al. present an interesting and intuitive concept for the sensitivity and heterogeneity of biological networks: When two or more proteins form a functional complex, it is the limiting component with the lowest concentration that is most sensitive to perturbations and whose fluctuations dictate cell-to-cell variability of complex function. The authors apply this concept to the TGFb pathway and discuss sensitivity of SMAD signaling towards TGFb receptor I and II fluctuations. The paper is clearly written and convincing, but some improvements in the experimental validation would be beneficial as detailed further below.

1. The authors claim that the ratio of TGFb receptor I and II is very different across cell lines (Fig. 1) and use this observation for the validation of their model in Fig. 4. However, the relative expression TGFb receptor levels are purely based on RNAseq data which does not necessarily imply similar behavior at the protein level, especially on the cell surface. To address this issue, the authors should ideally provide absolute Western blot measurements of TGFbRI at the protein level to complement their absolute quantification of TGFbRII (Fig. S2). At the very least they should show that the observed relative expression levels of TGFbRI and II at the protein level (Figure S7) are correlated to differences in RNA levels (Fig. 1) using protein quantification. They should also confirm that similar receptor ratios for these receptors at the RNA level are observed in other published RNAseq datasets of the same cell lines(e.g., ENCODE for HepG2 and published RNAseq studies in HaCaT). Furthermore, they might take into account published mass spec datasets for quantifications of TGFbR protein levels.
2. Figure 4: To better judge the reproducibility of the knockdown titration, it would be good to show the different siRNA concentrations as a color code- Alternatively, TGFBR expression could be plotted as a function of the siRNA concentration in a Supplemental Figure, showing the effects of individual replicates.
3. The simulations in Figs. 5 and 6 show that SMAD signaling fluctuations are mainly determined by cell-to-cell variability of receptor levels when using the SMAD nucleocytoplasmic ratio as a readout, and this is especially true for early time points. For downstream cellular responses, the absolute concentration of phosphorylated SMAD (complexes) in the nucleus is likely more relevant. Based on the authors work and evidence from the literature, I expect that this quantity will likely be heavily be influenced by receptor levels as well, but fluctuations in SMAD expression will play an important role as well. The authors should discuss this issue, and clarify that normalized quantitites like SMAD N2C and pSMAD/SMAD mostly characterize receptor-level fluctuations while filtering SMAD fluctuations.
4. The single-cell measurements in Fig. 7 are interesting, but can only partially be seen as a direct validation of the model predictions, as it seems expected that varying the total input by introducing co-fluctuations in both receptors heavily influence the SMAD level. Wouldn't it be possible to design more specific validation experiments, in which the receptor co-expression construct (Fig. 7C) is used for baseline optoTGFBR expression and combined with an individual expression construct for one of the opto-receptors? This way, the authors could establish different regimes, in which one of the two receptors becomes dominant, and the impact fluctuations could be analyzed in a larger receptor expression space. Of course, a full validation of all possible scenarios is not necessary, but it would, for instance, be valuable to see whether the strong dependency of SMAD signaling of TGFBR2 levels vanishes when TGFBR2 is expressed at a higher level than TGFBR1.

Referees cross-commenting

Comments from R2: I agree with most comments of the other reviewers, and highlight the most important overlaps with my comments below.

I agree with R1 that the model validation in Fig. 7 is incomplete and think that this will be a key point to improve the quality of the manuscript (see also my reviewer comment 4)

In line with R3 and R4, I think that the SMAD N/C simulations do not necessarily imply effects on TGFb target gene expression, cell fate decisions or human pathologies. The significance of the results for cellular behavior should be discussed (see also my comment 3)

Significance

The manuscript presents an interesting and intuitive concept for the sensitivity and heterogeneity of biological networks. The authors apply this concept to the TGFb pathway and discuss sensitivity of SMAD signaling towards TGFb receptor I and II fluctuations.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

Summary:

In the current study, Li et al investigated how TGF-beta signaling is controlled by protein abundances. Computational modeling and experiments indicated that the abundance of TGFBR1 and TGFBR2 affects the signaling, and those with lower abundance affect the signaling more, resembling Liebig's law of the minimum. Specifically, they showed that by using multiple cell lines with a different abundance of receptors, modulation of expression of the less abundant receptor impacts the signaling, which is measured by SMAD2 nuclear-to-cytosol ratio and/or relative phospho-SMAD2 level. Also, by using a light-induced interaction system, they showed that the signaling is dependent on the concentration of receptor complex when both receptors are expressed at similar amounts.

Major comments:

Computational predictions support the authors' idea. The computation and the experiments are well-documented. And it would gain substantially if the authors fill the gap between the predictions and the experiments as follows.

In Figure 4, the authors showed that perturbation on receptors with lower expression levels in each cell line changes the phospho-SMAD2 level. Although the data looks consistent with their claim, the result is only qualitative. The authors established a computational model in the former sections, thus it would be of great interest to assess if the experimental results quantitatively match the computational prediction.

In Figure 5, the authors computationally predicted that the expression level of receptors is correlated with SMAD2 N2C levels 1 hour after stimulation, and the strength of negative feedback with SMAD2 N2C levels 8 hours after stimulation. Because the authors employed iRFP-SMAD2 system, the prediction could be verified experimentally, at least the prediction on SMAD2 N2C 1 hour after stimulation could be checked. (In a sense, this is partially verified by the data in Figure 7, where both receptors are expressed at similar levels). It would gain substantially if the authors could verify the computational prediction in Figure 6.

Since the authors stated in the introduction that "The same TGF-β ligand can initiate different signaling responses depending on the cellular context, but the underlying control principle remains unclear...Together, these results revealed an effect of the minimum control in the TGF-β pathway, which may be an important principle of control in signaling pathways with context-dependent outputs.", experimental verification of the prediction done in Figures 4-6 will be very important. Or the authors should stress that these points are only predicted by computational models.

As written in the below "Significance" section, the result is, in a sense, obvious. It should be stated that because the study utilized a slightly high concentration of TGF-beta in the experiments, it might be natural that the low-abundance receptor becomes a bottleneck of the signaling. It would gain to assess how receptor abundance affects signaling with the stimulation of lower concentrations of TGF-beta, or to examine the computational model if the low abundance of a receptor becomes a bottleneck of signaling because of saturation. Also, it is highly recommended to discuss the physiological implication of the current study, taking into account the experimental conditions used.

Significance

TGF-beta signaling is one of the most rigorously studied pathways both computationally and experimentally. As written in the introduction of the manuscript, it is still unknown how the variability of responses arises not only between cell types but also differences among cells of single cell type. Studies showed that protein abundance accounts at least partly for a source of cell variability in TGF-beta signaling.

While former studies examined the variability in SMAD protein abundance, the uniqueness of this study is that it focused on the abundance of TGF-beta receptors.

Given that both TGFBR1 and TGFBR2 are involved in the signaling, however, it's not difficult to imagine that a less abundant receptor affects the signaling more than the other, and serves as a bottleneck for the signaling. Specifically, because a slightly high concentration (100pM = 4.4 ng/mL of TGF-beta; other studies used much lower conc., e.g. 0, 0.03, 0.04, 0.07, and 2.4 ng/mL in Frick et al, PNAS, 2017, and 0, 1, 2.5, 5, 25, and 100 pM in Strasen et al, Mol Syst Biol, 2017) is used throughout the experiments to check cell-cell variability and the effect of receptor abundance in the current study, the formation of the receptor-ligand complex may be quite fast and be saturated at the level where the receptor with lower abundance is exhausted. In the reviewer's humble opinion, the authors' statement that this is Liebig's law of the minimum sounds a bit exaggerated.

Nevertheless, the study is of some value because it utilized both computational and experimental analysis to show it is indeed the case. Of note, the current study showed that the variability in the different proteins leads to the variability in different time points, namely, the variability in the receptor abundance leads to the variability 1 hour after stimulation, while that in negative feedback strength leads to the variability 8 hours after stimulation. If the authors fill a small gap between their computational analysis and experimental verification, the study will be of interest to the specialist in the field.

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

Learn more at Review Commons

---

Reply to the reviewers

Reviewer #1

We thank this Reviewer for the time spent assessing our manuscript, and for suggesting approaches to strengthen the robustness of the differences (e.g., TL vs FL) reported in our results. We have carefully addressed each point raised by this and other reviewers, providing new analyses and data - see list below. Indeed, these analyses combined helped us to make our main results reproducible, corroborating the main findings and refining the message of the manuscript.

New analyses/data added:

1. *Effect of batch due to different lanes - comparison of DEGs (TL/FL) obtained when samples in different lanes are tested individually (new Figure S15). *
2. Effect of batch correction on our results - comparison of the DEGs (TL/FL) obtained with and without batch removal (new Figure S15).
3. Sensitivity of our enrichment results for GWAS significance – we performed the enrichment of GWAS genes using different GWAS thresholds, 10-6, 10-7, 5x10-8 (new Figure S14).
4. Expression analysis of GRIN2A and SLC12A5 in Allen Brain Atlas data and qPCR results of GRIN2A and SLC12A5 in patients with frontal and temporal lobe traumatic injury (new Figure S12, Table S3).
5. Comparison of the DEGs (TL/FL) with DEGs (autism/Ctrl) obtained from single cell RNA seq (new Figure S16, Table S7).
6. Comparison of the results using the GWAS genes derived from Trubetskoy et al. with our gene lists (new Figure S17).

7. Description of the data quality (Figure S2) Major points:

8. The main limitation of the work is the small starting sample size. The authors studied 1 frontal lobe sample and 2 temporal lobe samples. Although this information was in Table S3 it would be good to include upfront in the Methods. snRNA-seq was generated on the 10x platform. It would be helpful to know if the 10x step and sequencing was performed as one batch, or as individual batches. Similarly, were the sample libraries all sequenced on the same lane, or different lanes. The authors do not state in the Methods how many nuclei they were targeting and this should be included. Sample pre-processing was well described and standard. We now provide additional details about the sequencing step (nuclei, sample pre-processing, etc.) in the revised manuscript (see Methods, and text below). The potential batch effect of the lane is discussed and addressed in the next point.

‘10X Genomics uses a microfluidic system for cell sorting. Cells and enzymes, combined with Gel Beads, enter the oil phase to form GEMs. The resulting sample libraries were sequenced on separate lanes. To enhance sequencing depth, the primary target number of nuclei for the two samples from TL is set at 10,000, considering an RNA integrity number (RIN) of 6.5. In contrast, for the sample from FL, the target is set at 20,000 nuclei due to a higher RIN of 8.1.’

In relation to Batch correction - as with any batch correction method, it is unclear whether the correction is adjusting for biological differences or technical. Since this is a study of the differences between FL and TL, would it not be more appropriate not to correct for batch, particularly as the samples were analysed individually - particularly if batch effects were carefully controlled for in the initial study design. The authors should test whether the results are robust to batch correction or not.

Since the samples are sequenced by different lanes of 10X platform we can’t exclude potential batch effects. To address this, we corrected the batch by CCA (canonical correlation analysis) which enhanced the clustering and the UMAP visualization, which is now less affected by batch-specific variations.

Moreover, in an attempt to account for the sample size limitation, we employed 3 approaches to confirm the main transcriptional differences between the 2 regions, and that these are “robust” to batch correction, as is shown in new Figure S15: (1) Comparison of the gene expression differences (2 TL vs 1FL) with and without removing batches (new Figure S15. a, c); (2) The results obtained by comparing the differences between each individual TL sample (processed in different lanes) and FL sample are contrasted with the results after batch removal (new__ Figure S15. b, d__); (3) To confirm a limited effect of lane, we provide analysis of the expression similarity of three samples which demonstrates, consistently for each major cell type and neuronal sub-types, a strong correlation between the two TL samples (form different lanes) as compared with FL (new Figure S15. e).

As shown in panel a, c, below, the majority of DEGs (2TL vs FL) identified with batch effects largely overlap with the DEGs (2TL vs FL) without considering batch effects for both major cell types and neuronal sub-types. In panel b, d, we show that the majority of DEGs with batch correction (2TL vs FL) overlap with the individual DEGs found in each TL vs FL comparison. In panel e, we show that the transcriptomic profiles of 2 TL exhibit higher similarity compared with the sample from FL. Overall, based on these analyses we concluded that our results are robust to batch correction.

In addition, we highlight that, differently from other tissues, it is very difficult to obtain the “fresh” human samples of brain cortex, which most likely provides different transcriptome information than the more commonly used post-mortem brain samples. These analyses offered another evidence supporting the differences between TL and FL, which complement (and align with) the comparative analyses using the data from Allen Brain Atlas (see Figure S9, original results).

Figure S15. Comparison of biological (gene expression) differences in each major cell type and neuron-subtype between the 2 regions with and without batch effect removal. ____a, c. Comparison of the DEGs (2 TL vs 1FL) with and without removing batches (a, up-regulated in TL; c, up-regulated in FL). b, d. Comparison of the DEGs (2TL vs FL) following the removal of batch effects with the DEGs calculated by individual TL vs FL samples. __e. __Expression correlation between each sample (without batch correction for lane), showing higher transcriptional similarity within the same tissue type than across tissues, consistently in major cell-types and neuronal subtypes.

3.Differential gene expression analyses between the FL and TL was undertaken using edgeR. It is unclear if this was performed on aggregated counts or not - i.e., sum of counts per gene per cell type. If it was, then with such a small sample size (1 frontal lobe and 2 temporal lobe samples), it is unclear how well edgeR will perform. Similarly, if the DE analysis was performed using individual gene per cell counts, then there is a type 2 error risk due to pseudoreplication. It is reassuring that the primary results were replicated in a second dataset. Moreover, the downstream analyses (functional enrichment analysis, heritability enrichment analysis etc) are designed to cope with noisy data so I'm happy with the broad conclusions.

We acknowledge the reviewer’s point, and here we specify that edgeR performs differential expression analysis at the level of individual genes across individual cells, and we performed DE analysis for each cell type. We and others consider edgeR a robust tool for analyzing RNA-Seq data; edgeR has been extensively benchmarked alongside other widely used statistical methods, e.g., edgeR-LRT and edgeR-QLF which showed high performance1. Another study about different tools for differential expression in single cell data demonstrated that edgeR (and others) has usually higher precision, larger than 0.9, yielding lower false positive2. Therefore, based on previous formal assessments showing the robustness of edgeR, we select this approach for DE analysis.

Moreover, it has been previously documented that edgeR can be used also to analyses small samples due to several inherent features. First, edgeR uses an empirical Bayes framework to estimate dispersion, which is a measure of the biological variability in gene expression. This approach uses information across genes, helping to stabilize the variance estimates even when sample sizes are small. This makes edgeR more robust in cases with a limited number of replicates. Second, edgeR accounts for overdispersion, which can effectively handle small sample sizes and provide more accurate statistical tests. In the revised manuscript, we now discuss the advantages of edgeR in Methods, in particular for edgeR performance on small sample size in single cell RNA seq.

It is unclear if this was performed on aggregated counts or not - i.e., sum of counts per gene per cell type

We specify that edgeR performs differential expression analysis at the level of individual genes across individual cells, and we performed DE analysis for each cell type. This is now indicated in Methods.

*4.To calculate the enrichment of "genetic risk" associated with psychiatric disorders, the authors used a hypergeometric test for the overlap between cell type specific genes and the GWAS variant-mapped genes for each disease, which is widely used to evaluate the enrichment of genetic risk genes. To identified GWAS variant mapped genes the authors used a GWAS SNP threshold of To test the sensitivity of the enrichment analysis, we selected the GWAS genes with each threshold respectively: 10-6, 10-7, 5x10-8. The new results are largely consistent with those obtained using a P-value of 10-5. Susceptibility genes for neuropsychiatric disorders are enriched for expression in neuronal cell types for each P-value. With respect to neuronal subtypes, we found stronger enrichment in INH than in EX sub-clusters, with INH PVALB, SST and EX L5 being the neuronal sub-clusters mostly enriched for expression of GWAS genes (new __Figure S14).

Figure S14 Cell type for expression of neuropsychiatric disorder associated GWAS genes with each threshold respectively: 10-6, 10-7, 5x10-8. a-c. adjusted P-value of enrichment in each 7 major cell type. d-f. adjusted P-value of enrichment in each neuron subtype.

Moreover, the Reviewer suggests using an alternative tool, FUMA, which requires the whole set of SNP GWAS associations. While these can be available for single diseases and GWAS data (assuming the authors made all data available, and assuming one obtains approval by the consortia managing the GWAS data), unfortunately these SNPs data are not available for several diseases in the NHGRI-EBI GWAS catalog, which provides only SNPs with a max P=10-5. Since in our study we wanted to consider GWAS data from 7 neuropsychiatric diseases, we pragmatically opted for obtaining data from NHGRI-EBI GWAS catalog rather than seeking GWAS SNP data from individual studies.

We also acknowledge the limitations for the variant to gene mapping (revised Discussion, page 17, line 17), and we also highlight that several other studies rely on the variant to gene mapping from NHGRI-EBI GWAS catalog for enrichment analyses3-5. There are also studies that investigate the enrichment of mapped genes (from NHGRI-EBI GWAS catalog) in different cell types using the hypergeometric test 6-7, as we do in our study. Therefore, the methods used in our manuscript are consistent with approaches adopted in previously published studies. Perhaps more importantly, in the revised manuscript, we replicated the main GWAS enticement results (e.g., in INH neurons and in PVLAB from the temporal lobe) in the Brain Allen Atlas datasets, which shows that, despite these limitations of variant to gene mapping, our main enrichment results are replicable. We discussed these limitations in our paper (see Discussion, page 17, line 6).

However, where individual genes are mentioned then the authors may wish to confirm the results from edgeR for a few selected genes with a second technique such as qPCR. For example, GRIN2A and SLC12A5.

To address this point, first, we check the expression of the 2 genes using the data from Allen Brain Atlas data, which show significantly high expression in TL (new Figure S12. b, and below). In addition, we carried out new qPCR analysis, and found the mRNA expression levels of GRIN2A and SLC2A5 in patients with traumatic brain injury in the temporal lobe region were higher than those in patients with frontal lobe injury (new Figure S12. c).

Figure S12. b. Expression level of GRIN2A and SLC12A5 in 2 regions using Brain Allen Atlas. ***P-value-ΔΔCt method. Significance was determined through T-test (two-tailed). qPCR for each TL or FL sample was repeated 3 times.

Reviewer #2

We thank this Reviewer for the time spent evaluating our manuscript. In the revised manuscript we have now included several new analyses and data that allowed us to replicate and strengthen our main findings, and especially we considered the psychoactive drug target genes using the whole psychoactive drugs DB. We believe these new data helped us to refine the message and overall improve reproducibility of the main findings presented. We have carefully addressed each point raised by this and other reviewers, by providing revisions and explanations, and adding new data to our manuscript, as follows:

New analyses/data added:

1. *Effect of batch due to different lanes - comparison of DEGs (TL/FL) obtained when samples in different lanes are tested individually (new Figure S15). *
2. Effect of batch correction on our results - comparison of the DEGs (TL/FL) obtained with and without batch removal (new Figure S15).
3. Sensitivity of our enrichment results for GWAS significance – we performed the enrichment of GWAS genes using different GWAS thresholds, 10-6, 10-7, 5x10-8 (new Figure S14).
4. Expression analysis of GRIN2A and SLC12A5 in Allen Brain Atlas data and qPCR results of GRIN2A and SLC12A5 in patients with frontal and temporal lobe traumatic injury (new Figure S12, Table S3).
5. Comparison of the DEGs (TL/FL) with DEGs (autism/Ctrl) obtained from single cell RNA seq (new Figure S16, Table S7).
6. Comparison of the results using the GWAS genes derived from Trubetskoy et al. with our gene lists (new Figure S17).
7. Description of the data quality (Figure S2) 1.The manuscript is unfortunately lacking (supplemental) figures showing the preprocessing, batch effect correction, and cell type annotation of single nucleus RNAseq data. Although this part is described in the methods in detail, it is hard to judge if these parts were done properly if data is not shown in any of the figures. Regarding the batch effect correction, it reads as if the batch effects have been removed for both brain regions separately. This potentially introduces a bias between brain regions that hugely questions the later performed analysis of differential expression analysis in FL vs TL. In any case, this analysis is not convincing since it has been performed on n=3 vs. n=3 samples and is thus tremendously underpowered.

We thank the reviewer for the suggestions. First, we added the cell type annotation process for the major cell type by showing the expression of known markers in Figure S2. f. To show the validity of our cell classification, we calculated the significance of overlap with major cell type markers derived from known study in Figure S2. e. __We also provide the distribution of nUMI, nGenes, percentage of mitochondrial genes after quality control in Figure S2. b __to show the large number of cells contributing to the overall quality and depth of the scRNA-seq dataset despite the small number of individual samples.

Figure S2. Description of snRNA-seq data. b. Distribution of nUMI, nGenes, percentage of mitochondrial genes after QC. e. Significance of overlap with major cell type markers derived from known study. f. Expression of known markers for each cell type.

Since the samples are sequenced by different lanes of 10X platform, therefore, we can’t exclude potential batch effects. To account for this potential batch effect, we corrected the batch by doing CCA (canonical correlation analysis) which enhanced the clustering and the UMAP visualization more biologically meaningful and less driven by batch-specific variations.

Moreover, in an attempt to account for the sample size limitation, we employed 3 approaches to confirm the main transcriptional differences between the 2 regions, and that these are “robust” to batch correction, as is shown in new Figure S15 (see next page): (1) Comparison of the gene expression differences (2 TL vs 1FL) with and without removing batches (new Figure S15. a, c); (2) The results obtained by comparing the differences between each individual TL sample (processed in different lanes) and FL sample are contrasted with the results after batch removal (new Figure S15. b, d); (3) To confirm a limited effect of lane, analysis of the expression similarity of three samples demonstrates, consistently for each major cell type and neuronal sub-types, a strong correlation between the two TL samples (form different lanes) as compared with FL (new__ Figure S15. e__).

As shown in panel a, c, below, the majority of DEGs (2TL vs FL) identified with batch effects largely overlap with the DEGs (2TL vs FL) without considering batch effects for both major cell types and neuronal sub-types. In panel b, d, we show that the majority of DEGs with batch correction (2TL vs FL) overlap with the individual DEGs found in each TL vs FL comparison. In panel e, we identified that the transcriptomic of 2 TL exhibit higher similarity compared with the sample from FL.

Overall, based on these analyses we concluded that the results are robust to batch correction.

Figure S15. Comparison of biological (gene expression) differences in each major cell type and neuron-subtype between the 2 regions with and without batch effect removal. a, c. Comparison of the DEGs (2 TL vs 1FL) with and without removing batches (a, up-regulated in TL; c, up-regulated in FL). b, d. Comparison of the DEGs (2TL vs FL) following the removal of batch effects with the DEGs calculated by individual TL vs FL samples. __e. __Expression correlation between each sample (without batch correction for lane), showing higher transcriptional similarity within the same tissue type than across tissues, consistently in major cell-types and neuronal subtypes.

In addition, we highlight that, differently from other tissues, it is very difficult to obtain the “fresh” human samples of brain cortex, which most likely provides different transcriptome information than the more commonly used post-mortem brain samples. These analyses offered another evidence supporting the differences between TL and FL, which complement (and align with) the comparative analyses using the data from Allen Brain Atlas (Figure S9, original results).

2.Furthermore, the way that the authors treat GWAS data for disease does not seem to follow best practices. For schizophrenia, last year the largest GWAS so far was published (Trubetskoy et al, Nature, 2022) with very careful prioritization of genes. The authors should re-analyze their data using the gene list from this paper (and similar from other disorders) rather than the gene list that they came up with using their approach. The approach to select genes from different GWAS introduced seems highly arbitrary and leaves the reader unsure about statistical rigor.

We have carefully considered the suggestion regarding the treatment of GWAS data, particularly with respect to the gene list derived from the recent schizophrenia GWAS by Trubetskoy et al. (Nature, 2022). In this paper, the author mainly identified 120 genes (106 protein-coding) that are likely to underpin associations with schizophrenia which implicate fundamental processes related to neuronal function including synaptic organization, differentiation and transmission.

With respect to our study, first, we found there is significant overlap between prioritized genes in Trubetskoy et al’ study and GWAS genes included in our study. We showed the P value for overlap significance below, and listed the 27 genes. Among the prioritized genes, GRIN2A is also identified to be important in neuropsychiatric disorder, which is also confirmed to differ between the 2 regions and dysregulated in disease brain.

Enrichment of genes obtained from the prioritized schizophrenia-associated genes in Trubetskoy et al. Significant overlap (P=0.013, hypergeometric test) between schizophrenia-associated genes (120 prioritized genes from Trubetskoy et al.) and our GWAS genes (from GWAS catalogue).

Second, we conducted a supplementary analysis focused on the 120 genes prioritized by Trubetskoy et al, as shown below. We found the 120 prioritized genes in this paper are significantly enriched in excitatory and inhibitory neurons (panel b, below), aligning with our main findings conducted by schizophrenia related genes in our previous GWAS gene lists. Within the neuronal subcluster, we found a significant enrichment in L4, LAMP5 and PVALB cells (panel c); L4 and PVALB are largely consistent with our previous results (shown in Figure 3. c). Furthermore, we also found the 120 schizophrenia-associated genes are highly significantly enriched in DEGs (TL/FL) in VIP and PVALB subtypes (panel d).

b-c. Enrichment of 120 prioritized schizophrenia-associated genes in major cell types and neuronal subtypes. d. For each cell type, the enrichment of 120 genes is calculated with respect to the set of DEGs (TL/FL). Approach used for enrichment analysis is hypergeometric test (significance level, P-valueThese results suggest that while new gene lists from larger GWAS studies (e.g., Trubetskoy et al) come up regularly, the lists of GWAS genes prioritized in our enrichment analysis has some overlap with the newest GWAS. We agree that including more (larger) GWAS studies will strengthen the manuscript, but based on the analyses above, we believe our GWAS enrichment results are robust. In the revised manuscript, the new analysis including the detailed comparison with schizophrenia GWAS by Trubetskoy et al. (Nature, 2022) are reported in new Figure S17.

To improve on the GWAS enrichment analysis, we carried out additional sensitivity analyses to support our GWAS enticement results. We selected additional thresholds to evaluate the robustness of our results to the choice of gene lists to test the sensitivity of the enrichment analysis, we selected the thresholds: 10-6, 10-7, 5x10-8. The new results are largely consistent with those obtained using P-value of 10-5. Susceptibility genes for neuropsychiatric disorders are enriched for expression in neuronal cell types for each P-value. With respect to neuronal subtypes, we found stronger enrichment in INH than in EX sub-clusters, with INH PVALB, SST and EX L5 being the neuronal sub-clusters mostly enriched for expression of GWAS genes. These results are reported in new Figure S14.

Figure S14. Enrichment of cell type expression of neuropsychiatric disorder-associated GWAS genes for different GWAS-thresholds. a-c. Adjusted P-value of enrichment in each 7 major cell type. d-f. Adjusted P-value of enrichment in each neuron subtype.

3.Similarly, the choice of data set for disease-related differentially expressed genes is unclear as much larger (two orders of magnitude) published data sets exist for many of the disorders. For three of those DEG analyses performed on bulk RNAseq data, for the remaining two the DEG list of papers is used directly -making a comparison complicated. One would have to run DEG analysis in a standardized way for all 5 datasets/ disorders. It would be good to also indicate the respective sample size in Fig. 5a. (On a different note, the OCD publication is Piantadosi et al. 2021, not Sean C.et.al..) In addition, the authors matched brain regions to their regions of interest (frontal and temporal lobe) as shown in Fig. 5a. Still, they vary across disorders, which makes it hard to compare their findings across disorders and does not allow for a general statement about frontal vs. temporal lobe. ____To generalize for any of those psychiatric disorders I would recommend including more RNA-seq studies of the same disorder. Nowadays there are getting more and more case-control single nuclei studies on such disorders published. The authors could also include those by transforming them to pseudo bulk datasets and running their DEG analysis with edgeR as documented.

We acknowledge there might be a bias introduced by using the DEGs from the original paper directly. In addition, there is a general limitation affecting all bulk-RNA studies in complex tissues with different anatomical structures (e.g., kidney, brain, etc.), which form a great part of the publicly available data sources. In brain research, it is also more difficult to collect fresh human brain samples from patients with psychiatric disorders, which poses additional tissue availability constraints. Despite these limitations, we argue that bulk-RNA studies in anatomically complex tissues, and the DEGs reported therein, can be useful for GWAS enrichment analysis and not all DEGs are due to spurious or artificial signals. Furthermore, due to the lower sequencing depth inherent in single-cell RNA sequencing compared to bulk RNA sequencing, we set up to contrast our findings with results found by bulk-RNA seq.

We agree with the Reviewer that “One would have to run DEG analysis in a standardized way for all 5 datasets/ disorders”, however this approach assumes that the raw data are directly available and/or that the authors are keen to share the raw data. Both these assumptions are – unfortunately – not valid in many cases. (In several instances, we did contact authors to have access to raw data, with no success). Furthermore, when a commonly shared gene set in the DE genes is identified when using “heterogenous DE gene lists”, this might suggest a strongest convergence, or a convergence that is “robust” despite the differences between the heterogeneous DE lists (from authors or newly generated by us). Therefore, despite the limitations, our approach was motivated by practical considerations.

In addition, the brain region differences can be more prevalent and have a larger impact for specific psychiatric disorders. In our manuscript, for MDD we specially looked at only the BA8/9 which come from dorsolateral prefrontal cortex. Regarding OCD, BP, and MDD, several studies showed that there are no significant functional differences clinically observed between the orbitofrontal cortex and dorsolateral prefrontal cortex (Schoenbaum G, Setlow B. Integrating orbitofrontal cortex into prefrontal theory: common processing themes across species and subdivisions. Learning & Memory, 20018. Golkar A, Lonsdorf T B, Olsson A, et al. Distinct contributions of the dorsolateral prefrontal and orbitofrontal cortex during emotion regulation. PloS one, 20129). In the case of ASD, Brodmann area 41, 42, 22 refers to a subdivision of the cytoarchitecturally defined temporal region of cerebral cortex, exhibiting similar functionality to the temporal gyrus. Therefore, ASD and SCHI may arise from specific regions within the temporal lobe, while OCD, MDD, and BP may be associated with regions within the frontal lobe.

To address the Reviewer’s point more directly - we carried out additional analyses to investigate the effect of this factor on our main results. One of our aims was to understand how regional gene expression differences (TL/FL) in PVALB neurons are associated with gene dysregulation in the brain of neuropsychiatric disease patients. We have now extended these analyses to a separate dataset, and tested whether the dysregulated genes in neuropsychiatric disease are expressed mainly in TL and FL using single cell data from Brain Allen Atlas (4 patients, each with 6 brain regions profiled). The new results are shown in new Figure S11 b-f (and reported in the next page).

Briefly, we found that the percentage of dysregulated genes in SCHI, BP, OCD, and MDD that are expressed in MTG (SCHI: 75%, BP: 81%, OCD: 68%, MDD: 71%) and CgG (SCHI: 77%, BP: 80%, OCD: 60%, MDD: 77%) is higher compared with those in all other regions included in Brain Allen Atlas dataset. The percentage of ASD dysregulated genes expressed in the 6 regions from Brain Allen Atlas are quite similar. This analysis suggests that, despite the potential impact of heterogeneity of regions, the DEGs in psychiatric conditions are typically expressed at higher level in MTG (TL) and CgG (FL) compared with other regions, therefore highlighting the potential role of these two regions in psychiatric conditions. Therefore, we believe that despite the heterogeneity of regions included in the published RNA-seq studies, the strongest signal of enrichment for DEGs is detected consistently in TL and FL, i.e., in the 2 brain regions where the DEGs are also most highly expressed compared with other regions. These new data, reported in a new Figure S11 of the revised manuscript, provide additional evidence to support our main conclusions.

Due to the difficulties obtaining the human sample of psychiatric disorders causing limited public data resource, we found one study about molecular changes of ASD revealed by single cell RNA seq coming from Velmeshev et al. Science. 2019; 364(6441):685-689 (PMID: 31097668), including 22 ASD samples and 19 control samples. We compared the DEGs (TL/FL) with the DEGs (ASD/Ctrl), and report the results in new Figure S16. Briefly, the results show that except LAMP5, Endo, and L4, ASD-associated dysregulated genes significantly overlap with DEGs between FL and TL in several cell types, especially in VIP and astrocytes. While PVALB is not the most apparent cluster reflecting regional differences contributing to ASD, we found a moderate association (R2 =0.11, P=0.04) between changes in TL/FL and those in ASD/Ctrl brain. These findings suggest that gene expression differences between the 2 regions may contribute to ASD disorder, providing additional evidence to support our main conclusions.

Figure S16. Overlap of genes dysregulated in ASD and genes differentially expressed between TL and FL in each major cell type and neural subtype. Venn diagram plots in a-m showing the number of overlapped genes. Dot plot in each panel shows the relationship between the log2FC(TL/FL) [our study] and log2FC(ASD/Ctrl) [Velmeshev et al. Science. 2019 study]. Significance of the overlap: *0.001-0.01, **0.0001-0.001, ***0.00001-0.0001, ****4.For cell type enrichment of disease signal based on GWAS signal several carefully controlled studies exist using more sophisticated statistical methods (Skene et al., Nature Genetics, 2018, Bryois et al., et al. Nature Genetics 2020, MJ Zhang et al Nature Genetics 2022 to mention a few). I applaud that the authors aim to go beyond this basic characterization but I think it is worrisome that by using less sophisticated (and importantly less controlled) statistical and genetic approaches they reach a different signal -and then they go on and analyze this signal. It is potentially interesting they reach a different conclusion, but they need to provide a careful statistical analysis to explain how the chosen method is superior or at least different to previous efforts.

The Reviewer suggests the use of alternative approaches to link GWAS variants to genes, like MAGMA, LDSC, FUMA to improve the gene mapping from GWAS signals, and are better than the gene mapping based on proximity alone. While these approaches can provide some advantages, most of these methods do require the whole set of SNP GWAS associations, including non-significant associations. While these can be available for single diseases and specific GWAS data (assuming the authors made all data available, and assuming one obtains approval by the consortia managing the GWAS data) these SNPs are not available for several diseases in the NHGRI-EBI GWAS catalog, which provides only SNPs with a max P=10-5. Since in our study we considered GWAS data from 7 neuropsychiatric diseases, we (pragmatically) opted for obtaining data from NHGRI-EBI GWAS catalog rather than seeking GWAS SNP data from individual studies.

We now acknowledge the limitations for the variant to gene mapping (revised Discussion, page 17, line 17), and we also report that several other studies rely on the variant to gene mapping from NHGRI-EBI GWAS catalog for enrichment analyses4-6. There are also studies that investigate the enrichment of mapped genes (from NHGRI-EBI GWAS catalog) in different cell types using the hypergeometric test 7-8, as we do in our study. Perhaps more importantly, in the revised manuscript, we replicated the main GWAS enticement results (e.g., in INH neurons and in PVLAB from the temporal lobe) in the Brain Allen Atlas datasets, which shows that, despite these limitations of variant to gene mapping, our main enrichment results are replicable.

(Other comments)

- only n=3, ~45 000 cells making it hard to generalize

- no supplementary figures for the methods (i.e. preprocessing, cell type annotations), thus hard to judge if done properly if they do not show any data - much higher level of transparency needed

- The methods part is not clear, in general, it is only descriptive, with no equations

- Unconvincing determination of DEGs for each disorder

- DEGs and pathways based on n1=1 vs n2=2 feals handwavy

- DEG analysis and cell type annotation are mixed up and it is unclear how DEGs were determined

While we acknowledge the limitation of sample size in our study, we also emphasize again the challenges in of availability of fresh human sample, which provide more transcriptomic information than postern sample. Despite the small number of individual samples, the large number of cells (~45,000) contributes to the overall quality and depth of the scRNA-seq dataset. Hence, our study provides a foundational perspective on the gene expression between the frontal lobe (FL) and temporal lobe (TL), and valuable data source for further investigations.

With respect to the additional description of the data processing and cell annotation process, in the revised manuscript we now elucidate the cell type annotation process by showing the expression of some known markers in new Figure S2. f, the significance of overlap with major cell type markers derived from known study in new Figure S2. e, the distribution of nUMI, nGenes, percentage of mitochondrial genes after quality control in new __Figure S2. b. __

To strengthen the differential gene expression analysis, we replicated our main findings through SMART RNA-seq from Brain Allen Atlas including the DEGs identified in our study (Figure S9).

More technical details are provided in the revised manuscript, as detailed below:

In the revised Methods section – (1) Differential expression analysis in FL vs TL and pathway enrichment analysis, we added more details about how the DEGs are identified and how this is robust to batch correction. (2) Replication analyses in human Brain Allen Atlas, we provide more details about how we replicated the DEGs using Allen Brain Atlas dataset. (3) Enrichment of neuropsychiatric disease GWAS genes in brain cell clusters, we now added more methodological details about the enrichment analysis.

__ __

Reviewer #3

We thank the Reviewer for his/her overall positive comments. In the revised manuscript we have now included several new analyses requested by this and other reviewers (see list below), which allowed us to replicate and strengthen our main findings. We also add details of the method used in this paper. We believe these new analyses and data helped us to improve reproducibility and strengthen the main findings presented in our manuscript.

New analyses/data added:

1. *Effect of batch due to different lanes - comparison of DEGs (TL/FL) obtained when samples in different lanes are tested individually (new Figure S15). *
2. Effect of batch correction on our results - comparison of the DEGs (TL/FL) obtained with and without batch removal (new Figure S15).
3. Sensitivity of our enrichment results for GWAS significance – we performed the enrichment of GWAS genes using different GWAS thresholds, 10-6, 10-7, 5x10-8 (new Figure S14).
4. Expression analysis of GRIN2A and SLC12A5 in Allen Brain Atlas data and qPCR results of GRIN2A and SLC12A5 in patients with frontal and temporal lobe traumatic injury (new Figure S12, Table S3).
5. Comparison of the DEGs (TL/FL) with DEGs (autism/Ctrl) obtained from single cell RNA seq (new Figure S16, Table S7).
6. Comparison of the results using the GWAS genes derived from Trubetskoy et al. with our gene lists (new Figure S17).
7. Description of the data quality (Figure S2) 1.The authors integrated the brain snRNA-seq data with GWAS data to annotate the cell type specific expression, which is one of the key points for this analysis, however a more detailed description of the method is lacking.

We have made changes to the text to improve and clarify this aspect. In the revised Methods section, we now specify: “To calculate the enrichment of genetic risk associated with psychiatric disorders, we used a hypergeometric test for the overlap between cell type specific genes (DEGs between one cell with other cell types, log2FC>0.5, adjusted.P __2.The authors found a set of genes which is associated with psychiatric disorders and specific cell types, for example inhibitory neurons are the most vulnerable cell type to genetic susceptibility through their analysis. The correlation of each cell type and each psychiatric disorders can be discussed.*__

We thank the Reviewer for this suggestion; we have now added more details discussing the relationship between other cell types with psychiatric disorders other than PVALB-neuron in this part – see Discussion in the revised manuscript, where we added: “Astrocyte, OPC are also associated with psychiatric disorders, and play essential roles in maintaining brain homeostasis, regulating synaptic transmission, and supporting neuronal function. Astrocytes also contribute to maintaining the integrity of the blood-brain barrier (BBB) and interact closely with neurons. Disruptions in this communication impact neural circuitry, which is relevant to many psychiatric disorders. OPCs generate oligodendrocytes, producing myelin crucial for signal conduction and brain structural integrity, which potentially impacts brain connectivity and communication between brain regions. Among neuronal subtypes, our data suggest that disruption of specific biological process in PVALB, SST and L5 neurons may contribute to neuropsychiatric disorders. PVALB cells are believed to activate pyramidal neurons only if the signal from excitatory neurons is sufficient and optimize the signaling in both EX and INH72. SST neurons gate excitatory input onto pyramidal neurons within cortical microcircuits, mainly coming from L5 layer of excitatory neuron which is involve in motor control, decision-making, and information transfer between the cortex and subcortical structures73. These signaling processes, when dysregulated, have been implicated in psychiatric diseases74. The relationship between psychiatric disorders and other layers of the cerebral cortex is still under investigation. *L2-3 neurons handle local processing, relevant to conditions like schizophrenia and autism. L6 neurons in thalamocortical circuits are crucial for sensory processing and information relay, involving sensory perception abnormalities.” *

3.The authors have found a group of interesting genes, such as GRIN2A, DGKI, and SHISA9 and confirmed them with the Allen Brain Atlases. Experimental validation would be helpful to confirm such findings.

In our manuscript, we emphasized that GRIN2A and SLC12A5 (both implicated in schizophrenia and bipolar disorder) were significantly upregulated in TL PVALB neurons and in psychiatric disease patients’ brain. To address this point, first, we check the expression of the 2 genes using the data from Allen Brain Atlas data, which showed significantly high expression in TL (new Figure S12. b). By means of new qPCR analysis in primary TL/FL samples, we found the mRNA expression levels of GRIN2A and SLC2A5 in patients with traumatic brain injury in the temporal lobe region were higher than those in patients with frontal lobe injury (new Figure S12. c).

Figure S12. b. Expression level of GRIN2A and SLC12A5 in 2 regions using Brain Allen Atlas. ***P-value-ΔΔCt method. Significance was determined through T-test (two-tailed). qPCR for each TL or FL sample was repeated 3 times.

Lastly, we want to highlight that since we believe in “Data Democratization” and sharing our data resources, upon publication, we will make all our data (including the single cell in “fresh” (surgically resected) brain tissue samples) and corresponding detailed results available to the scientific community.

We believe our study (which is focused on psychiatric diseases) will prompt other groups to use our single cell data and to dig deep into the role of temporal and frontal lobes in other neurogenerative diseases.

__ __

References

1. Squair, J.W., Gautier, M., Kathe, C. et al. Confronting false discoveries in single-cell differential expression. Nat Commun 12, 5692 (2021).
2. Wang, T., Li, B., Nelson, C.E. et al. Comparative analysis of differential gene expression analysis tools for single-cell RNA sequencing data. BMC Bioinformatics 20, 40 (2019).
3. Bhattacherjee A, Djekidel MN, Chen R, Chen W, Tuesta LM, Zhang Y. Cell type-specific transcriptional programs in mouse prefrontal cortex during adolescence and addiction. Nat Commun. 2019 Sep 13;10(1):4169.
4. Grubman A, Chew G, Ouyang JF, Sun G, Choo XY, McLean C, Simmons RK, Buckberry S, Vargas-Landin DB, Poppe D, Pflueger J, Lister R, Rackham OJL, Petretto E, Polo JM. A single-cell atlas of entorhinal cortex from individuals with Alzheimer's disease reveals cell-type-specific gene expression regulation. Nat Neurosci. 2019 Dec;22(12):2087-2097
5. Przytycki, P.F., Pollard, K.S. CellWalker integrates single-cell and bulk data to resolve regulatory elements across cell types in complex tissues. Genome Biol 22, 61 (2021).
6. Swindell, William R., et al. "RNA-Seq analysis of IL-1B and IL-36 responses in epidermal keratinocytes identifies a shared MyD88-dependent gene signature." Frontiers in immunology 9 (2018): 80.
7. Geirsdottir, Laufey, Eyal David, Hadas Keren-Shaul, Assaf Weiner, Stefan Cornelius Bohlen, Jana Neuber, Adam Balic et al. "Cross-species single-cell analysis reveals divergence of the primate microglia program." Cell 179, no. 7 (2019): 1609-1622.
8. Schoenbaum G, Setlow B. Integrating orbitofrontal cortex into prefrontal theory: common processing themes across species and subdivisions[J]. Learning & Memory, 2001, 8(3): 134-147.
9. Golkar A, Lonsdorf T B, Olsson A, et al. Distinct contributions of the dorsolateral prefrontal and orbitofrontal cortex during emotion regulation[J]. PloS one, 2012, 7(11): e48107

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

In the manuscript "Decoding frontotemporal and cell type-specific vulnerabilities to neuropsychiatric disorders and psychoactive drugs", the authors integrated brain with no history of neuropsychiatric disorder snRNA-seq data with public GWAS data among 7 psychiatric disorders to explore the heterogeneity between temporal lobe (TL) and frontal lobe (FL), the genetic risk factors and potential drug responsible genes. Multiple bioinformatics technics have been used in the manuscript. The authors found critical pathways and key genes that are related to the psychiatric disorders and GWAS genes enriched cells such as PVALB cells, which can help the understanding in the field. Overall, the manuscript is well written and organized, but there are some issues need to be addressed.

1. The authors integrated the brain snRNA-seq data with GWAS data to annotate the cell type specific expression, which is one of the key points for this analysis, however a more detailed description of the method is lacking.
2. The authors found a set of genes which is associated with psychiatric disorders and specific cell types, for example inhibitory neurons are the most vulnerable cell type to genetic susceptibility through their analysis. The correlation of each cell type and each psychiatric disorders can be discussed.
3. The authors have found a group of interesting genes, such as GRIN2A, DGKI, and SHISA9 and confirmed them with the Allen Brain Atlases. Experimental validation would be helpful to confirm such findings.

Significance

Strength: this manuscript is strong in bioinformatics analysis. Limitation: wet-lab validation of some of the findings would be helpful.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Paper review: Decoding frontotemporal and cell type-specific vulnerabilities to neuropsychiatric disorders and psychoactive drugs

In their manuscript with the title Decoding frontotemporal and cell type-specific vulnerabilities to neuropsychiatric disorders and psychoactive drugs, the authors describe their work on integrating snRNAseq data from "fresh" human frontal and temporal lobe of three healthy donors with genetic risk factors of 7 psychiatric disorders, bulk RNAseq data from healthy and disease human cortex/ DEG lists from previous studies for 5of the psychiatric disorders, and gene targets for commonly prescribed psychoactive drugs. The authors claim that PVALB neurons in the temporal lobe are most vulnerable to genetic risk factors and even more to psychoactive drugs for psychiatric diseases and suggest GRIN2A and SLC12A5 as the genes that most contribute to their vulnerability.

According to my overall impression, the paper has major problems in terms of quality, clarity, and statistical power. I do not recommend publishing this manuscript in its current form.

The manuscript is unfortunately lacking (supplemental) figures showing the preprocessing, batch effect correction, and cell type annotation of single nucleus RNAseq data. Although this part is described in the methods in detail, it is hard to judge if these parts were done properly if data is not shown in any of the figures. Regarding the batch effect correction, it reads as if the batch effects have been removed for both brain regions separately. This potentially introduces a bias between brain regions that hugely questions the later performed analysis of differential expression analysis in FL vs TL. In any case, this analysis is not convincing since it has been performed on n=3 vs. n=3 samples and is thus tremendously underpowered.

Furthermore, the way that the authors treat GWAS data for disease does not seem to follow best practices. For schizophrenia, last year the largest GWAS so far was published (Trubetskoy et al, Nature, 2022) with very careful prioritization of genes. The authors should re-analyze their data using the gene list from this paper (and similar from other disorders) rather than the gene list that they came up with using their approach. The approach to select genes from different GWAS introduced seems highly arbitrary and leaves the reader unsure about statistical rigor. Similarly, the choice of data set for disease-related differentially expressed genes is unclear as much larger (two orders of magnitude) published data sets exist for many of the disorders. For three of those DEG analyses performed on bulk RNAseq data, for the remaining two the DEG list of papers is used directly -making a comparison complicated. One would have to run DEG analysis in a standardized way for all 5 datasets/ disorders. It would be good to also indicate the respective sample size in Fig. 5a. (On a different note, the OCD publication is Piantadosi et al. 2021, not Sean C.et.al..) In addition, the authors matched brain regions to their regions of interest (frontal and temporal lobe) as shown in Fig. 5a. Still, they vary across disorders, which makes it hard to compare their findings across disorders and does not allow for a general statement about frontal vs. temporal lobe. To generalize for any of those psychiatric disorders I would recommend including more RNAseq studies of the same disorder. Nowadays there are getting more and more case-control single nuclei studies on such disorders published. The authors could also include those by transforming them to pseudo bulk datasets and running their DEG analysis with edgeR as documented. For cell type enrichment of disease signal based on GWAS signal several carefully controlled studies exist using more sophisticated statistical methods (Skene et al., Nature Genetics, 2018, Bryois et al., et al. Nature Genetics 2020, MJ Zhang et al Nature Genetics 2022 to mention a few). I applaud that the authors aim to go beyond this basic characterization but I think it is worrisome that by using less sophisticated (and importantly less controlled) statistical and genetic approaches they reach a different signal -and then they go on and analyze this signal. It is potentially interesting that they reach a different conclusion, but they need to provide a careful statistical analysis to explain how the chosen method is superior or at least different to previous efforts.

Plus:

• Flash-frozen human tissue with little post-mortem delay
• TL and FL comparison: interesting
• Multiple comparison corrections
• Replication analysis included The drug target genes angle is interesting

Minus:

• only n=3, ~45 000 cells making it hard to generalize
• no supplementary figures for the methods (i.e. preprocessing, cell type annotations), thus hard to judge if done properly if they do not show any data - much higher level of transparency needed
• The methods part is not clear, in general, it is only descriptive, with no equations
• Unconvincing determination of DEGs for each disorder
• DEGs and pathways based on n1=1 vs n2=2 feals handwavy
• DEG analysis and cell type annotation are mixed up and it is unclear how DEGs were determined

Unclear:

• Is the background dataset used for enrichment of genetic risk calculation different for each region and cell type? If so? How is this a fair comparison?
• Which subset of GWAS genes is used for Gene co-expression networks

Significance

Mainly weaknesses

Advancement provided by the study remains modest due to low confidence in the findings. Potentially interesting approach but needs to utilize state-of-the-art methodology and data sets. A typical audience would be journals targeting molecular/biological psychiatry

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

The authors integrated snRNA-seq analysis with genetic risk and drug-specific signatures to investigate brain regional differences in risk for neuropsychiatric disease and drug response. To replicate the main findings, the authors also analyzed single nuclei data from the Brain Allen Atlas. Overall, the manuscript is very well written. The methods are comprehensive, clear and well-wrtten which is very welcome. The authors have undertaken a large number of bioinformatic investigations using appropriate methodology and careful design.<br /> The main limitation of the work is the small starting sample size. The authors studied 1 frontal lobe sample and 2 temporal lobe samples. Atlhough this information was in Table S3 it would be good to include upfront in the Methods. snRNA-seq was generated on the 10x platform. It would be helpful to know if the 10x step and sequencing was performed as one batch, or as individual batches. Similarly, were the sample libraries all sequenced on the same lane, or different lanes. The authors do not state in the Methods how many nuclei they were targeting and this should be included. Sample pre-processing was well described and standard. In relation to Batch correction - as with any batch correction method, it is unclear whether the correction is adjusting for biological differences or technical. Since this is a study of the differences between FL and TL, would it not be more appropriate not to correct for batch, particularly as the samples were analysed individually - particularly if batch effects were carefully controlled for in the initial study design. The authors should test whether the results are robust to batch correction or not. Differential gene expression analyses between the FL and TL was undertaken using edgeR. It is unclear if this was performed on aggregated counts or not - i.e., sum of counts per gene per cell type. If it was, then with such a small sample size (1 frontal lobe and 2 temporal lobe samples), it is unclear how well edgeR will perform. Similarly, if the DE analysis was performed using individual gene per cell counts, then there is a type 2 error risk due to pseudoreplication. It is reassuring that the primary results were replicated in a second dataset. Moreover, the downstream analyses (functional enrichment analysis, heritability enrichment analysis etc) are designed to cope with noisy data so I'm happy with the broad conclusions. However, where individual genes are mentioned then the authors may wish to confirm the results from edgeR for a few selected genes with a second technique such as qPCR. For example, GRIN2A and SLC12A5. To calculate the enrichment of "genetic risk" associated with psychiatric disorders, the authors used a hypergeometric test for the overlap between cell type specific genes and the GWAS variant-mapped genes for each disease, which is widely used to evaluate the enrichment of genetic risk genes. To identified GWAS variant mapped genes the authors used a GWAS SNP threshold of <10-5, and mapped SNPs to genes using the GWAS DB. The background set of genes is appropriate as is the statistical method. Given the small sample size however, I think it would be helpful to see a sensitivity analysis of the results that (a) uses different GWAS thresholds e.g., 10-6, 10-7, 5x10-8 and (b) uses an alternative SNP to gene mapping tool such as FUMA. Overall, whilst well written, the manuscript as a whole feels overly long. I think it could be improved by a more stringent focus on the most important biological and translational findings.

Significance

Overall, the manuscript is very well written. The methods are comprehensive, clear and well-wrtten which is very welcome. The authors have undertaken a large number of bioinformatic investigations using appropriate methodology and careful design.

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

Learn more at Review Commons

---

Reply to the reviewers

The authors do not wish to provide a response at this time +

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

Summary:

Oxytocin (OXT) is a neuro-hypophysial hormone and exerts its effects through binding to the oxytocin receptor (OXTR). OXTR is expressed by various types of cells, including leukocytes and gastrointestinal cells. Previous studies demonstrated that OXT alleviates experimental colitis and regulates anti-inflammatory response. The authors' group reported that conditional deletion of OXTR in macrophages and dendritic cells exacerbated dextran sulfate sodium (DSS)-induced colitis. In the present study, they aimed to uncover the essential function of OXT signaling in colonic carcinogenesis and colitis by using intestinal epithelium cell (IEC)-specific OXTR knockout (KO) mice. IEC-specific KO mice exhibited markedly increased susceptibility to DSS-induced colitis and Azoxymethane (AOM)/DSS-induced colitis-associated colorectal cancer (CAC) compared to wild-type mice. Mechanistically, OXTR depletion in IECs impaired the inner mucus of the colon epithelium. Furthermore, oxytocin was found to regulate MUC2 maturation through B3GNT7-mediated fucosylation. In human colitis and CAC colon samples, there was a positive correlation between B3GNT7 expression and OXTR expression. Moreover, the administration of oxytocin significantly alleviated tumor burden. These results suggested oxytocin's promising potential as an effective therapeutic intervention for individuals affected by colitis and CAC.

Major comments:

1. Figure 1: The expression levels of many genes are altered in cancer cells. It is unclear whether decreased OXTR expression is the cause or the consequence of CAC in both human cases and mouse experiments. In Figure 1B, the background staining on the control tissue is very high. On the CAC tissue section, the staining appears uneven and OXTR staining appears high where the background is high. Thus, the result is not convincing.
2. OXTR is expressed by many types of cells, including leukocytes, and OXTR expressed by leukocytes is reported to have an anti-inflammatory activity (Mehdi et al., Front Immunol, 2022, 13:864007. doi: 10.3389/fimmu.2022.864007). In this study, the importance of OXTR expressed by leukocytes is not considered.
3. Oxytocin is usually administered by injection. It is unclear how it was administered. Oral administration is probably not effective. There is also no description about the source of oxytocin.
4. The effects of oxytocin could be different between males and females. It might be interesting to present data of males and females separately and comment on the finding. It was previously shown that OT plasma levels (pg / ml, mean {plus minus} SD) were significantly higher in women than in men (4.53 {plus minus} 1.18 vs 1.53 {plus minus} 1.19, p ˂ 0.001), and such differences might be related to behaviors, attitudes, as well as susceptibility to stress response, resilience and social emotions specific of women and men (Marazziti et al., Clin. Pract. Epidemiol. Ment. Health. 2019; 15: 58-63). Male mice are more susceptible to DSS-induced colitis and this could be due to different oxytocin levels.
5. Figure S3: Is the antibody directed to sugar? In the absence of OXTR, MUC2 is almost absent. It is hard to believe that the expression/production of MUC2 is almost completely dependent on oxytocin.
6. Figure 5: The authors indicated that reduced fucosylation was due to the decreased B3GNT7 expression. Addition of L-fucose may not result in increased fucosylation.
7. The effects of fucose supplementation was studied using a colitis model, whereas the effects of oxytocin supplementation was studied using a colon cancer model. Thus, the effects by these two agents cannot be compared.
8. Figure 6: Oxytocin treatment could activate OXTR expressed on both leukocytes and epithelial cells. There is no comment on this subject.
9. Figure 6I: A few mice died after 20 days. Is it correct? What does "day 0" mean in this figure?
10. Figure 7B: Again, the result is not convincing. Leukocytes are reported to express OXTR and many leukocytes are in the colon tissues, especially in colitis tissue. But, they are not positive.
11. Figure 7M: It is good to have a summary figure; however, it appears not accurate. There is no data showing the floxed mice have "Tolerant immune response" and KO mice have "dysregulated immune response".
12. The authors' group previously reported that OT activated IECs to release prostaglandin E2 that was required for the repair of intestinal epithelium after injury (Ref. 11 in this manuscript). What happened to this mechanism?

Minor comments:

1. Mice: There is no reference for the floxed mice. It would be also helpful to add the strain number.
2. Figure 1D: What was the expression level "1"? There is only a small difference between 6.4 (Control) and 6.1 (AOM/DSS).
3. Figure 1J: What caused the increase in spleen weight? Is this a marker for increased inflammatory responses or cancer cell growth?
4. Figure 1K: What was the meaning of the increased expression of each cytokine?
5. Figure 1L: The photos are too small to see the detail.
6. Figure 1M: What was the end point?
7. Page 3: "OXTR Deficiency in IEC Facilitates CAC Depends on Inflammation" This is not a sentence.
8. Figure 2A and I: "% weight loss" were less then 1%. Is it correct?
9. Figure 2N: Hard to see any differences. Too small.
10. Fig 5K: Labels are not complete.
11. Figure 5: Silver color is used for lines and columns but difficult to see.
12. It is unclear at which time point samples were prepared.

Significance

General assessment:

The role of OXT and its receptor OXTR in DSS-induced colitis was previously reported by using systemic or myeloid cell-specific OXTR KO mice. Here, the authors used IEC-specific OXTR KO mice and found that OXTR expressed by IEC cells plays an important role in inflammation-associated colitis and CAC by promoting the post-transcriptional modification of MUC2 via B3GNT7. This is a strength. The major imitation is that they used only IEC-specific KO mice and the relative importance of OXTR on IECs is unclear. In addition, important information necessary to understand the results is missing throughout the manuscript. There are other questions as listed in the comments.

Advance:

Interaction of OXT and OXTR has been demonstrated to protect mice from inflammation-associated colitis. This study went one step further by demonstrating that OXTR expressed by IEC cells plays a protective role in inflammation-associated colitis and CAC by promoting the post-transcriptional modification of MUC2 via B3GNT7.

Audience:

Basic research and translational/clinical.

Field of expertise:

The reviewer is expertized in the field of "cancer and inflammation" but not in "MUC2" or "glycosylation". Keywords: inflammation, chemokines, leukocyte trafficking, tumor microenvironment.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Summary: In their paper "Oxytocin alleviated colitis and colitis-associated colorectal tumorigenesis by targeting fucosylated MUC2" the authors describe the contribution of oxytocin (OXT) to colonic mucus formation, and its protective contribution to colitis and colon cancer. The authors also describe the mechanism by which OXT execute its alleviating effects on the colon mucus; enhanced B3GNT7-mediated fucosylation. The findings are demonstrated in cell cultures, various mice models, and samples from human patients.

Major comments:

1. A conceptually confusing issue in this work is whether a decrease in OXTR expression is a predisposition, or a result of colonic illness. On one hand, the experiments with OXTR KO mice and cultured cells suggest that pre-existing lower levels of the receptor sensitize the tissue. However, in the AOM/DSS model the control mice present normal OXTR levels whereas mice that received AOM/DSS had lower expression, suggesting that changes in OXTR levels are not a predisposition but a result of the treatment/illness. Additionally, when tissue from CAC patients were analyzed, decreased levels of OXTR were found in sites of wounds but not in adjacent healthy tissue, implying that this decrease is not a genetic treat but a result of external cue. This inconsistency must be sorted out and clearly demonstrated.
2. The study describes a new regulatory pathway for colonic mucin 2, and colon related conditions. Why did the author choose to generate mice lacking OXTR in the entire intestine (small+large) and not a large-intestine specific deficiency? And is there any way to demonstrate that the absence of OXTR in the small intestine does not interfere with the results presented here?
3. The commonly used fixative for mucus and secreted mucins is Carnoy fixative (can be found in many of Hannson G.C and in Johansson M.E.V papers, and many other papers describing colon staining), while the use of formaldehyde and glutaraldehyde is less preservative for mucus layer. This raises a concern regarding the data obtained from aldehyde-fixed mucus samples.
4. The authors found that mice lacking OXTR have lowered levels of B3GNT7, which leads to a decrease in mucin 2 fucosylation and to further damage in the colon. What is the mechanism by which supplementation with L-fucose alleviates these outcomes given that the enzyme that regulates the addition of the fucose to mucin 2 is downregulated?

Minor comments:

1. Some IHC images don't show comparable or similar areas. Specifically, Figure 1 B, Figure 7 F, I.
2. There is a discrepancy between the dosage of DSS used to induce chronic colitis in the text (2%) and in the methods (2.2%). In addition, the difference between concentrations of DSS used to induce chronic and acute colitis (2.2% vs. 2.5%, respectively) is significantly smaller than what is reported in many other papers using these models.
3. In Figure 2 A, I, Y-axis labeling doesn't seem right (compare with Figure 5 G). It looks like the decimal point is a mistake.
4. All Western blots presented in this study lack the molecular weight of the proteins. In many cases it would have been more convincing to see a larger portion of the membrane.
5. Mucin 2 in a large protein (more than 5000 amino acids in human mucin 2), and many disulfide bonds. The authors do not mention if any reducing or denaturing agents were added to the lysis buffers, and whether any other special conditions were employed to separate this huge globular protein on SDS-PAGE gel.
6. The following sentence should be revised: " To examine the effects of fucosylation regulated by OXT on LS174T cells and colonic organoids, we found that..."

Significance

Though the concept of OXT-mediated suppression of colon cancer has been reported (For example: PMID: 34528509, and 31920487), the regulatory pathway by which it exerts its alleviating effect, and all the mechanistic components described in this paper, were not known before. This pathway may be a potential target for therapeutic intervention in various colonic diseases. Moreover, additional mucins may be regulated by OXT in a similar manner, which can extend the importance of these findings to other organs and disease-conditions. This type of findings is of interest to the broad audience of general cell biology as well as to GI clinicians. However, as stated in my comments there are some major issues with the hypothesis, the way data is presented, and in key methods that fundamentally limits my ability to evaluate this paper.

Key words for my field of expertise: Disulfide catalysis, Golgi, mucin

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

Summary

In this study, Wang and colleagues present data linking oxytocin signaling to protection against colitis and colitis associated cancer development. To this end, they utilize a Villin-Cre line to specifically remove OXTR only from intestinal epithelia and make use of AOM and/or DSS treatment to induce colitis, colorectal cancer or CAC. OXTRdeltaIEC mice consistently develop worse symptoms and more severe colitis/CAC compared to non-Cre expressing littermates, which appears to be associated with defects in the mucus layer. Using RNA-Seq, they identify the glycosyltransferase B3GNT7 as a differentially expressed gene. Due to its role in O-linked glycosylation, they investigate whether B3GNT7 is involved in mucin production and are able to show that OXT-induced upregulation of MUC2 protein is abolished when B3GNT7 is knocked down. In vivo, co-treatment with oxytocin reduces experimental CAC, which is an interesting that OXT may present a potential treatment option in CAC. The study is quite interesting in that it provides a potential treatment option where the mucus layer, which is often disturbed in IBD, can be impacted in a positive way. Still, there are some things missing to really be able to interpret the full picture and these should be experimentally addressed.

Major comments

• The staining for OXTR in Fig 1B is very strong (especially as others have reported that they were not able to demonstrate OXTR in human samples, Ohlsson et al. PMID: 16678285) - it would be beneficial to confirm OXTR distribution in steady state in mice, especially as you have a great negative control were OXTR staining should be absent from the IECs. Given the observations further in the study, I would also be very curious to see whether OXTR expression is specific to goblet cells, so co-staining with Muc2 would be interesting to include in later figures. The information for the OXTR antibody is also missing from the supplementary methods.
• Fig 1D, Fig S2 - is this whole colon RNA or specifically epithelial cells? This can have major impact on observed expression levels as the relative amounts of epithelial vs. other cell types can drastically change, thereby falsely giving the impression that expression levels in the epithelial cells change, so this really needs to be analyzed in purified epithelial cells. In Fig S2 there is significant OXTR expression remaining in the deltaIEC mice, so this suggests to me that non-IEC cell types are also included.
• The interpretation of the study would benefit from including some steady state/untreated data for the OXTRdIEC mice. For example in Fig 2 the researchers report increased spleen size, increased cytokines etc upon CRC in these mice, but it is important to also show steady state data as these parameters may already be significantly increased in basal conditions in these OXTR deficient mice (especially seeing as Fig 3 claims that under basal conditions the mucus layer is extensively damaged you would expect some phenotype in these mice).
• Special care needs to be taken to preserve the mucus layer during fixation, and from the methods it is not clear whether the authors have taken these technical difficulties into account. Only PFA fixation is mentioned, but it is well-established that the golden standard for imaging the mucus layer is to fix tissues in water-free Carnoy's fixative, as the mucus layer tends to collapse using formaldehyde (see also Johansson & Hansson, PMID: 22259139).
• In Fig 5, the message and conclusions become a bit more fuzzy. Overall fucosylation is measured, but it is unclear whether MUC2 itself is increasingly fucosylated due to OXTR signaling, or that this represents more global changes in the secretory pathway that eventually lead to more efficient MUC2 production. Perhaps an IP using fucose-specific lectin combined with western blotting for MUC2 may be an option to demonstrate whether MUC2 itself becomes increasingly fucosylated due to OXTR signaling?
• The title broadly claims that OXT "alleviates" colitis and CAC through MUC2 fucosylation and Fig 6 indeed shows that OXT treatment affects the outcome of mice in a CAC model, which is very promising, but it also loses the link with the mechanistic insights surrounding MUC2 fucosylation in previous figures. To really definitively make the claim in the title, it's important to investigate whether these OXT treated mice indeed have restored B3GNT7 levels and a thicker mucus layer after AOS/DSS regimen compared to non-OXT treated mice (as one would expect based on the in vitro data using LS174T cells and organoids). Studying the effect of OXT treatment in the regular DSS colitis model would also provide additional support for this claim.
• Optional as I am not a specialist in OXT signaling: I would assume that there are quite some differences between males and females when it comes to OXT and OXTR. Have the authors ever observed differences in staining pattern or expression levels between males and females? The methods state that all groups are sex matched, but I wonder if it may be necessary to include gender as a variable in the analysis?

Minor comments

• I would suggest to include another reference in the introduction and/or discussion, as MUC2 deficient mice are also known to develop colorectal cancer (Velcich et al. PMID: 11872843) and this serves as additional support for why it is important to discover how we can positively impact the mucin layer in IBD patients.
• In Fig 1A GEO data is reanalyzed, but it's not immediately clear what the original samples were (i.e. colon biopsies). At first glance, the figure itself adds to this confusion with the titles 'hypothalamus' and 'hypophysis' - it's not very clear that these labels indicate synthesis location of the respective hormone and not the tissue where expression was measured.
• Fig 1C - I could not find in the methods what software was used for these quantifications.
• Fig 1C - N=5 is mentioned in the figure legend and there are 10 datapoints in each group. Were 2 biopsies quantified per patient then? Please state this more clearly.
• Fig 3A, B and all other western blots - please include molecular weight indications in the figure
• Several figures use light grey bars and datapoints, but this color was very hard to see after printing the manuscript.
• The conclusion statement for Fig 3 should be revisited, as the expression of Muc2 mRNA is not affected at all by OXTR genotype (Fig S2F). Conclusion should make it clear that specifically (mature) protein levels are affected.
• Fig 4A-D - would be nice to include the full list of DE genes in supplement, it's an important resource. For example, there are other factors known that influence the mucus layer (such as AGR2), so I would be interested to see how these are behaving in the knockout mice.
• In Fig 4 H-J it would be informative to show the MUC2 mRNA expression level in these cultures as this could provide support for the mice data - i.e. do the cultures also display normal MUC2 mRNA levels, with a specific defect in the mucin maturation (as appears to be the case in mice)?
• Fig S3H-K - this figure and the validation of the siRNA is not mentioned in the main text
• It is interesting that L-fucose seems to partly reverse the effect of DSS, but I wonder whether mechanistically this is explained by restoration of B3GNT7 expression?
• Please check the accession codes for the reanalyzed datasets, figure legends mention two accessions, while the Data availability statement mentions three accessions.
• The number of repeats for each experiment is a bit unclear. It is now buried in the statistics statement in the methods, but it may be more clear if it is included in each figure legend.

Significance

This study shows a -for me- quite unexpected link between oxytocin and protection against colitis and colitis-associated cancer development. Disturbances in the mucin layer are a very common phenomenon in IBD and colitis and there has been a great interest in this in the scientific community for quite some years (Johansson, PMID: 25025717, Yao et al. PMID: 34902790, and many others). Current IBD treatment options are generally aimed at reducing inflammation, but this does not necessarily restore the mucin layer quality. It is therefore quite interesting to see that this is apparently heavily influenced by oxytocin (which already has applications in human medicine), and this provides significant advance to our current fundamental understanding of mucin barrier regulation.

As mentioned in the comments, the study can be further improved. To me, a more detailed investigation into the steady state phenotype of these mice, and a more detailed confirmation of where the oxytocin receptor is expressed is necessary to fully put the results into a broader framework. Also Fig 6, where the actual interventional effect of oxytocin is evaluated, no longer demonstrates whether this indeed happens through the same mechanism as outlined in the previous figures and this should be developed more.

I expect this study to be of interest primarily to a basic research audience, though I assume that a more clinical audience would be intrigued by the findings as well.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

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

Learn more at Review Commons

---

Reply to the reviewers

__Points raised by both reviewers in their cross-comments __

1. “emphasizing the acute nature of the study is important as well as the use of only male rats” RESPONSE: Thank you for pointing this out. It has been clarified throughout the manuscript, including the abstract, limitations section, and conclusions.

“The need for improvement of the presentation cannot be stressed enough”.

RESPONSE: The manuscript has undergone extensive revisions to enhance the clarity of data presentation and discussion, and to highlight its novelty in comparison to our prior studies. We have reduced the use of technical terminology and abbreviations, and when they do appear, they are explained with their first use and in the new Glossary section. The manuscript has been better organized, ensuring a logical flow of data and conclusions.

Reviewer #1

Major comments

“the extensive statistical analysis done for the gene expression would require assistance unless the in-house expertise already existed. If these are in place the work could be reproduced with the details provided.”

RESPONSE: Terms and abbreviations used in statistical and correlation analyses are thoroughly explained in the text and in the newly added Glossary section in the revised manuscript to the extent acceptable in a biological paper.

All statistical codes are accessible in the public GitHub repository at https://github.com/YaromirKo/biostatistics-nms. These codes may be utilized for the purpose of replicating studies.

Minor comments

“It is not clear how the genes that were studied here were picked. It is clearly stated what groups the genes fall into and their relevance to the study but it isn't clear how these were decided upon. Clarifying this would be helpful.”

RESPONSE: There is currently no consensus regarding the classification of genes as related to neuroplasticity. In particular, there is no agreement on lists of genes consistently associated with neuroplasticity across studies, and providers of mRNA analysis platforms do not offer panels of neuroplasticity-related genes. Most companies, such as Thermofisher, Illumina, and Nanostring, provide "Neurological" or "Neuropathology Research" panels that contain genes related to neuroplasticity. However, these panels are not specifically designed for targeted analysis of neuroplasticity-related genes.

The gene selection is arbitrary, and the chosen genes may vary across different studies depending on their objectives. In the present study, genes were selected based on several significant works having determined that these genes were likely related to neuroplasticity. Each gene's selection is justified by citing these works in the "Materials and Methods" section and we made every effort to avoid any bias. We do not assert that the gene set is all-encompassing. This matter is addressed in the Limitations section of the revised manuscript.

“It is not always clear what had been done in the previous work and what is completely new in this work, that could be addressed better.”

RESPONSE: Thank you for emphasizing that. It has been thoroughly addressed in the revised manuscript. While our previous study has discovered a left-sided neuroendocrine system, the current work delves into its organizational principles, which are equally crucial. We have shown that this system is bipartite and mirror asymmetric, and that its left and right counterparts can be targeted differently by pharmacological means. Additionally, we have revealed the left-right side-specific gene regulatory networks that operate in the neuroendocrine system and which activities are laterally coordinated by this system along the neuraxis.

“The text and figures are quite complex and require thorough reading the knowledge of the background to understand, therefore not making this work for a general audience.”

“Given the complexity of the work the reading of the results is quite dense and difficult to maneuver unless you have some prior understanding. My suggestion would be to try to simplify this but I wouldn't know exactly how to go about this.”

RESPONSE: We appreciate the Reviewer’s comments here, and agree that this is a complex work. We have endeavored to find a balance between a comprehensive presentation of the methods and results while also providing a level of simplification that will allow the reader who is not versed in this field to still appreciate this work. However, because of the nature of the experimental designs and of the findings that we report, we believe it to be important to provide a comprehensive explanation of the work and results. We believe that we have struck a balance between simplification and comprehensiveness with this revision. We have simplified the presentation of the results, their statistical analysis, and the analysis of gene regulatory networks for easier understanding. We also provide detailed explanations of technical terms in the newly added Glossary section. Please also refer to our response to point 2.

We believe that the revised manuscript has a level of complexity in data presentation and density similar to that of most combined physiological and molecular studies, complemented with advanced statistical and bioinformatics analysis. See please, for example papers published in Plos Biology (doi.org/10.1371/journal.pbio.3002328; doi.org/10.1371/journal.pbio.3002282; doi.org/10.1371/journal.pbio.3001465) and eLIFE (doi.org/10.7554/eLife.85756; https://doi.org/10.7554/eLife.90511.1).

General assessment

“The limitation would be understanding exactly what was done before and how this work expands on that, often it required the reader to look up references and prior work.”

RESPONSE: The introduction and discussion have been modified accordingly in order to comply with this comment. We have clarified how this study expands upon our previous work. In addition, please see the response to Comment 5 that also addresses this issue.

“The audience would be rather specialized, although it does gear towards clinical translation, this aspect could be highlighted better in the introduction and discussion.”

RESPONSE: Clinical aspects of the findings have been further highlighted in the revised manuscript. In the introduction, we note that the discovered phenomenon could contribute to asymmetrical neurological deficits following stroke and TBI. In the discussion section, we examine mechanical similarities between hindlimb asymmetry in rats and spastic dystonia in patients and hypothesize that the rat asymmetries may model this human neuropathology. In the concluding remarks, we state that it is crucial to examine the balance between neural and endocrine pathways in their contribution to neurological impairments, and to establish pharmacological approaches targeting the neuroendocrine system to restore the disturbed neurohormonal equilibrium.

“Those interested in brain injury/neurodegeneration as well as how signaling of motor control could be affected by not just damage to electrical descending motor tracts but to neuroendocrine signaling would be the specific audience.”

RESPONSE: We agree that the experts in neurotrauma, stroke and motor control may be interested in this study. However, the left-right side-specific neuroendocrine signaling may be a general biological phenomenon essential for regulation of lateralized brain functions, and, in a broader biological perspective, regulation of the body plan along the left-right axis.

Furthermore, the study presents what, to the best of our knowledge, is the first evidence for the existence of the left and right side-specific gene regulatory networks in the CNS. They operate in the neuroendocrine system and its peripheral target, and are coordinated across them via the humoral pathway. This is a novel molecular dimension in asymmetric organization of the generally mirror-symmetric CNS.

We are confident that experts in the establishment of the body plan and functional and molecular brain asymmetries will be interested in the concept formulated in this study.

Reviewer #2

Major comments:

“It should be made clear in the introduction that an acute complete cervical SCI is used and the discussion should be extended to include advantages and disadvantages of the used model and the alternatives.”

RESPONSE: Thank you for your suggestions. The introduction and discussion have been supplemented with the requested information. Specifically, we have noted that hindlimb postural asymmetry, a proxy model for neurological deficits, has enabled the discovery and characterization of the left-right side-specific neuroendocrine system. It is a binary model with two qualitatively different responses generated on either the left or right side. On the other hand, it cannot be used to analyze awake animals, and knowledge of its mechanisms is limited. A role for the neuroendocrine phenomenon in the persistent left-right specific biological and pathophysiological processes requires further investigation. This can be addressed by analyzing the effects of unilateral TBI in subchronic experiments with awake animals whose spinal cords are completely transected to disable neural pathways. The methodology could involve an integrated evaluation of hindlimb function during body weight-supported stepping, utilizing behavioral, electrophysiological, and biomechanical measures.

“A similar concern poses the use of pentobarbital and the interpretation of the results of the deafferentation. Were timing of the application and dosage strictly controlled between the different groups? It's effects on somatosensory afferent transmission through presynaptic inhibition are a concern.”

RESPONSE: Thank you for the remark. We have paid special attention to this issue. The rats were deeply anesthetized with the same dose and timing of anesthesia. These parameters were thoroughly controlled in all of the experiments. The depth of pentobarbital anesthesia was characterized by a barely perceptible corneal reflex and a lack of overall muscle tone. Of note, the side and magnitude of postural asymmetry do not apparently depend on anesthesia and its type; the asymmetry was virtually the same after brain injury in rats under deep pentobarbital or isoflurane anesthesia (this study and Lukoyanov et al., 2021; Watanabe et al., 2020; Watanabe et al., 2021; Zhang et al., 2020) and also in decerebrate unanesthetized rats (Zhang et al., 2020). Similar left-right differences were observed in the rats with left and right brain injury which were deafferentated 3 days later, and then analyzed under isoflurane anesthesia (Zhang et al., 2020). This is discussed in the revised manuscript.

Furthermore, no nociceptive stimulation was applied and tactile stimulation was negligible in the course of the asymmetry analysis; the legs were stretched by pulling the threads glued to nails of the toes. The application of lidocaine to the toes, which were pulled during stretching, had no impact on the formation of asymmetry. After all, the stretch and postural limb reflexes are immediately abolished and remain so for several days, and markedly decreased under anesthesia as it was firmly established in many studies. As these reflexes likely do not play a role in the formation of the asymmetric hindlimb posture, their afferent mechanisms could not be a cause of variations in our experiments.

In summary, three main arguments speak against an interference of pentobarbital with asymmetry formation in rats after rhizotomy. First, a similar asymmetry phenomenon developed in pentobarbital anesthetized rats, isoflurane anesthetized rats, and decerebrate un-anesthetized rats. Second, in rats that underwent rhizotomy, the primary sensory nerve fibers were entirely severed. Thus, the hypothetical link between pentobarbital's impact on asymmetry through its effect on presynaptic inhibition could be eliminated. Third, although there may be some variability in the depth of anesthesia among animals, the probability of such strong and statistically significant differences in the effects of brain injury and deafferentation arising from bias in the depth of anesthesia among groups of animals likely to be negligible.

*“Only two test for the asymmetry of spinal processing were used and the two tests are likely measuring very similar phenomena (tonic flexor over activation). Additional reflex tests could shed light onto underlying mechanisms.” *

RESPONSE: We agree. In previous studies, we also analyzed asymmetry in withdrawal reflexes between the left and right hindlimbs as an indicator of the effects of brain injury (Lukoyanov et al., 2021; Watanabe et al., 2021; Zhang et al., 2020). In the present study, we do not focus on the neurophysiological mechanisms of postural asymmetry. We instead prioritize characterizing the phenomenology and organizational principle of the left-right side-specific neuroendocrine system using the postural asymmetry model as a "black box" and as a robust and reliable readout.

Of note, there are several other equally important issues that remain to be addressed, including the identification of signaling pathways from the injured cortex to the hypothalamic-pituitary system, the identification of signaling molecules in the blood that convey information about the side of the brain injury, and the dissection of encoding and decoding mechanisms in the hypothalamus and spinal cord, respectively. No single study could investigate all of these mechanisms.

Minor comments:

“Figure 3 shows only the magnitude of the postural asymmetry in response to the different opioid receptor antagonists, yet the directionality is of interest, especially in case of the control animals. Pre2 values are missing too.”

RESPONSE: We appreciate the reviewer's comment and apologize for any errors in our previous version. The legend for Figure 3 has been revised and simplified. It is unnecessary to include PAS (Postural Asymmetry Size) in addition to MPA as the direction of PAS in all animals in each group was the same. This is stated in the revised manuscript's Legend for Figure 3. MPA was used to compare the left and right UBI groups, which had positive and negative PAS values, respectively. This comparison could not be carried out with PAS.

“Too many abbreviations are used which makes the text and figures very difficult to read at times.” “Terminology is sometimes inconsistent (e.g., delta vs contrast).”

RESPONSE: The manuscript now features a reduced amount of abbreviations. Technical terms and abbreviations are defined upon their first use and are also included in the newly added Glossary section. Corrections have been made to the use of the term "contrast" and its abbreviation "delta" in Figures. Additionally, the term "deltaW" as the left-right difference is no longer utilized within the manuscript.

“The section "correlation patterns in the hypothalamus and spinal cord" was almost impossible for me to understand and could use rephrasing.”

RESPONSE: We apologize for the previous version, and have simplified the presentation of molecular data. We believe that the level of complexity in the revised manuscript's statistics and data presentation is now comparable to that of many other molecular studies featuring system-level analyses; please see also response to Comment # 6 of the first reviewer.

“Only male rats are used.”

RESPONSE: This limitation has been addressed in the Limitation section. It is important to investigate whether identical or distinct neurohormones are responsible for the outcomes of left and right brain injury in male and female rats. However, this requires prior identification of most hypothalamic neurohormones and neuropeptides that regulate the asymmetric processes. Their number may be considerable, given the constellation of left and right gene regulatory networks in the hypothalamus.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Watanabe et al. build on their previous work to show that the left-right side specific effect of unilateral brain injury after acute complete spinal transection is indeed mediated by side-specific endocrine signaling. This is done by looking at a cervical spinal transection as opposed to a thoracic as in previous work. They further characterize the side-specific humoral hypothalamus-lumbar spinal cord pathways using gene expression patterns, application of opioid receptor antagonists, and dorsal root rhizotomy. Overall the evidence is very convincing and excludes mediation through the sympathetic system in addition to central descending tracts. Curiously the deafferentation, while having an effect on both sides only reversed the postural asymmetry caused by left-sided brain injury, and gene-gene co-expression revealed ipsilateral coordination.

Major comments:

• It is possible that many of the observations in the paper are dependent on the acute state of the spinal cord injury. This is mentioned in the limitations section and it is clear that the presented experiments are important and advance our understanding of this curious phenomenon. Yet, it should be made clear in the introduction that an acute complete cervical SCI is used and the discussion should be extended to include advantages and disadvantages of the used model and the alternatives.
• A similar concern poses the use of pentobarbital and the interpretation of the results of the deafferentation. Were timing of the application and dosage strictly controlled between the different groups? It's effects on somatosensory afferent transmission through presynaptic inhibition are a concern.
• Only two test for the asymmetry of spinal processing were used and the two tests are likely measuring very similar phenomena (tonic flexor over activation). Additional reflex tests could shed light onto underlying mechanisms.
• All major comments shouldn't be seen as a request for additional data but only require discussion.

Minor comments:

• Figure 3 shows only the magnitude of the postural asymmetry in response to the different opioid receptor antagonists, yet the directionality is of interest, especially in case of the control animals. Pre2 values are missing too.
• Too many abbreviations are used which makes the text and figures very difficult to read at times.
• Terminology is sometimes inconsistent (e.g., delta vs contrast).
• The section "correlation patterns in the hypothalamus and spinal cord" was almost impossible for me to understand and could use rephrasing.
• Only male rats are used.

Referees cross-commenting

I agree with reviewer #1's comments; most of them are in line with mine. The need for improvement of the presentation cannot be stressed enough. This is excellent and important work, which makes it even more important to convey it in an accessible way (be clear about prior work and what the novel results add, reduce number of abbreviations, guide the reader in how to interpret the figures, etc.). Otherwise, the audience will be limited.

Significance

General assessment: The manuscript provides clear evidence that there is a side-specific effect of UBI that is mediated by humoral signaling. Specifically, the present work excludes the sympathetic system. This is a very important finding that was missing in previous work. Further characterization of this recently discovered non-neuronal component of UBI is of very high importance as the potential for clinical implications are high.

Advance: The study provides a clear advance of our understanding of side-specific endocrine signaling to the spinal cord.

Audience: This study should be of interest to a wide audience, particularly for neuroscientists and neurologists who deal with the motor system.

My field of expertise: Neural control of locomotion, spinal cord injury, motor control, sensorimotor integration.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

Summary

Watanabe et al. show in "Bipartite left-right sided endocrine system: processing contralateral effects of brain injury" a continuation of previously published work, that hindlimb postural asymmetry (HL-PA) is due to the neuroendocrine signaling and not the cervical parasympathetic pathways in anesthetized spinal C6-C7 fully transected unilateral brain injured (UBI) rats. Further, this differential neuroendocrine control of the left-right side-specific hormonal signaling is affected differently by either right or left unilateral hindlimb sensorimotor cortex brain injuries. However, bilateral deafferentation (L1-S1) showed that only left-side brain injury was altered, indicating differing inputs. Adding to the previous finding that blocking opioid signaling in UBI non-injured spinal rats leads HL-PA, here they demonstrated this finding holds with right-left differences following a spinal transection. Furthering the previous findings of left-right lumbar spinal gene expression differences, this time they found hypothalmus and lumbar spinal cord gene expression differences that were ipsilaterally coordinated and affected by brain injury.

Major comments

• Are the claims and the conclusions supported by the data or do they require additional experiments or analyses to support them? Yes, the claims are supported by the data presented in this manuscript.
• Are the data and the methods presented in such a way that they can be reproduced? The data and methods have been presented in a way that could be reproduced, however given the expertise of this laboratory in developing new systems not for purchase it is likely it would take a given expertise to replicate the data. Additionally, the extensive statistical analysis done for the gene expression would require assistance unless the in-house expertise already existed. If these are in place the work could be reproduced with the details provided.
• Are the experiments adequately replicated and statistical analysis adequate? Yes, the experiments have been adequately replicated and statistical analysis to my understanding is adequate.

Minor comments

• Specific experimental issues that are easily addressable. It is not clear how the genes that were studied here were picked. It is clearly stated what groups the genes fall into and their relevance to the study but it isn't clear how these were decided upon. Clarifying this would be helpful.
• Are prior studies referenced appropriately? It is not always clear what had been done in the previous work and what is completely new in this work, that could be addressed better. The references themselves are extensive and well-used throughout the work.
• Are the text and figures clear and accurate? The text and figures are quite complex and require thorough reading the knowledge of the background to understand, therefore not making this work for a general audience.
• Do you have suggestions that would help the authors improve the presentation of their data and conclusions? Given the complexity of the work the reading of the results is quite dense and difficult to maneuver unless you have some prior understanding. My suggestion would be to try to simplify this but I wouldn't know exactly how to go about this.

Referees cross-commenting

Reviewer #2 makes a crucial point that emphasizing the acute nature of the study is important as well as the use of only male rats. Otherwise, reviewer #2's comments overlap partially with my own in increasing the accessibility of the work. Neither recommended changes would require new experimental data.

Significance

General assessment:

I would this topic quite intriguing and a novel understanding of motor control. The multiple experiments that were performed that addressed various contingencies of HL-PA may occur after UBI were addressed here (ie. parasympathetic and sensory input). Further experiments expanded on previous findings of the involvement of opioids, the pituitary gland, and spinal gene networks. The limitation would be understanding exactly what was done before and how this work expands on that, often it required the reader to look up references and prior work.

Advance:

Although this is my first encounter with the work, it is a follow-up study on work that was published previously in eLife in 2021. Therefore, given some of the overlap it wouldn't be entirely conceptually new but it would be addressing open questions which arose from that work and further add to our understanding of the mechanism involved in this phenomenon.

Audience:

The audience would be rather specialized, although it does gear towards clinical translation, this aspect could be highlighted better in the introduction and discussion. Those interested in brain injury/ neurodegeneration as well as how signaling of motor control could be affected by not just damage to electrical descending motor tracts but to neuroendocrine signaling would be the specific audience. My expertise is in spinal cord injury, sensorimotor coordination of hindlimbs and gene expression. Although not an expert in brain injury or neuroendocrine signaling, my background allows me to understand the experiments performed here and the relevance of the work.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

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

Learn more at Review Commons

---

Reply to the reviewers

We thank the reviewers for their constructive and detailed reviews. We have been able to resolve all issues raised by the reviewers with additional experiments and changes in the text:

• In response to two of the reviewers we've changed the nomenclature of the residues. As we would like to avoid assigning roles in the naming, we now use 'critical residue 3' and 'critical residue 4', with Cys and His forming critical residue number 1 and 2 respectively.
• We analyzed the role of the negative charge in the fourth critical residue of USP1, by mutating this Asp to Asn to assess the importance of a charged residue in these positions (Supplementary figure 2), resulting in complete loss of activity just like the alanine mutant. We also tested the effect of mutating the third critical residue to Asn in USP1, which causes a minor decrease in activity. This highlights the importance of the highly conserved aspartate (fourth critical residue), and shows that precise residue found in the position is important for catalysis. Additionally, these mutants address potential effects of the ‘holes’ left by the original Ala mutations.
• Importantly, we were able to perform single-turnover assays to expand on our analysis of the precise roles of the critical residues and give more fundamental insight in the defects of the mutants. These assays further elaborate on the variability observed between these USPs. In USP15, these experiments explain the defect in catalysis for the third critical residue mutant and provides insight how a successful nucleophilic attack is combined with defective catalysis (updated Figure 4), which is not observed in the other USPs we tested. In these other USPs, the single turnover experiments reveal that the nucleophilic attack performed by the third and fourth critical residue mutant of USP7 and USP40 happens with low efficiency, even lower efficiency for USP48 and that this ability is lost entirely in USP1.
• We included a number of important textual changes to better explain the choices and variation in USPs tested, highlight prior USP2 data and the implications for drug discovery.
• We updated Ub-PA conjugation assays (updated Figure 4) for better contrast, and repeated the Ub-PA assay for USP1 and USP48 with longer incubation (Supplementary Figure 6). More details are given in the point-by-point response below. All in all, we are convinced that this much improved manuscript is now ready for publication and hope that all reviewers will agree.

---

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

Summary: The authors study the functional role of two adjacent active site residues as candidates for polarising the catalytic histidine in the "Asn/Asp" box from five phylogenetically unrelated ubiquitin specific proteases (USP1, USP7, USP15, USP40 and USP48). One of these residues is more variable across USPs (Asn, Asp, Ser), whereas the second one is absolutely conserved (Asp). To this end they use alanine mutants in kinetic experiments and test their ability to crosslink to ubiquitin propargyl as a proxy for testing the nucleophilicity of the catalytic cysteine. They then further evaluate the activity of the USP1 mutants in processing PCNA-Ub in RPE1 cells. They find that the role of these two residues differs between the different USPs studied, which is in line with previous work that has shown that in USP7, the amongst USPs less conserved residue takes on the major role of polarising the histidine, whereas in the more distantly related USP2, the absolutely conserved Asp is more important (Zhang W, et al. Contribution of active site residues to substrate hydrolysis by USP2: insights into catalysis by ubiquitin specific proteases. Biochemistry. 2011 50(21):4775-85. doi: 10.1021/bi101958h). This study expands on these findings to evaluate the role of these residues in four other USPs.

Major comments: 1. The authors compare highly diverse USPs; USP1 requires UAF1 for full activity and the complex is used in the study, USP7 requires a C-terminal tail peptide for full activity, USP40 and USP48 belong to the CHN class, whereas USP7, USP15 and USP1 belong to the CHD class of USPs. The rationale for selecting this diverse set of USPs is therefore not clear and makes direct comparisons of the findings more difficult. It is certainly interesting that the previously published differences between USP2 and USP7 with respect to these residues are also found in four other divergent USPs, but for this reason it isn't as "surprising" as the title suggests. The title, omission of background knowledge on USP2 in the abstract and presentation of the findings in a graph that makes direct comparisons (Figure 5) are therefore a bit misleading, which needs addressing.

• We apologize that it seemed as if we had overlooked USP2, for which both critical residues are important, and we agree that our abstract previously focused too much on the perception of the field and its focus on USP7. We have changed the abstract and introduction to highlight the USP2 data for a more balanced perspective.
• The reviewer is correct that the set of USPs is diverse, but we see this as a strength, given that this is the first manuscript in which these residues are analyzed in a comparative side-by-side manner for multiple DUBs. We find that our results are not directly related to the CHN/CHD diversity (i.e. changes in the third catalytic residue), nor apparently to activation by a C-terminal tail (as both USP7 and USP40 have this mechanism). Since these are structurally conserved enzymes with a common fold, we do find the comparison is informative. Furthermore, we felt that it was important to clearly signal the variation in different steps of the mechanism, something which appears to largely remain unnoticed by the field. Figure 5 is helpful in understanding that these changes have multiple dimensions. We agree that it is important to signal the diversity as possible source for these differences and we have added the following sentences to paragraph 3 of the results: “These USPs vary in domain architecture and allosteric regulation, and therefore represent different aspects of the USP family. USP1, USP7 and USP15 both harbor two aspartates as third and fourth critical residue and USP40 and USP48 harbor an asparagine and aspartate as third and fourth critical residue respectively, allowing us to examine the importance of a negative charge in position of the third critical residue.”
• We used the word surprising in the title to indicate the variability we observed in the two dimensions of the mechanism, as indicated in Fig. 5.

The study relies on single alanine mutations, which will inevitably change the hydrogen bonding patterns and the local environment which could impact the conclusion. The authors should verify in kinetic assays at least for USP1, which is the main focus, that Asp to Asn mutants still display the same effects.

• We are thankful for this suggestion. We have made these additional USP1 mutants through insect cell expression and tested these in different assays. As expected, both Asn mutants follow the alanine mutations. The results are reported in Supplementary fig 2BC.

While neither mutant unfolds below 40 degrees, there are clear differences in thermal stability between some of the proteins used in the study (Supp. Fig. 1B). A full table of measured Tms by NanoDSF for all Wt and mutant proteins should be provided so that the reader can evaluate how the results may be impacted by local effects that impact the thermal stability. It is noticeable that USP40 and USP15 mutants in particular display large differences in thermal stability, which could directly affect the results. The authors should clearly discuss these limitations of the study.

• We have added supplemental table S2 to report the melting temperatures. The effect observed for USP15 is addressed in the results: “While both mutants of USP15 have a decreased thermal stability compared to USP15wt, these variants retain stability until 50 °C, indicating that they are still well-folded and suitable for kinetic assays at room temperature.”. For USP40, it is not the actual measured Tm that deviates a lot, but the measured 350/330 ratio, which is addressed in the legend of supplementary figure 1B “Ratios measured (350 mm/330 mm) varied between some of the mutants (Eg. USP40wt), but this did not affect the measured inflection points (Supplementary table 1)”.

  Minor comments: 1. For USP48 and USP40 no published structures are available at present, so it isn't clear whether there are any differences in orientation of the studied residues. An unpublished USP40 structure is referred to but not shown. The general conclusion that structures do not reveal any differences in these residues may therefore not be valid for all the studied USPs. Please revise.

• We apologize if this was not clear. We did however not refer to a USP40 structure, but a USP40 manuscript in preparation that studies biochemistry USP40 activation through activation by its C-terminal tail.

• The existing structures do not show observable differences in the active site residues, nor in the immediate surrounding, and therefore do not give insight which residue is critical for catalysis. We now mention this more explicitly. “It was previously shown that there are no structural differences in the positioning of the catalytic triad and the fourth critical residue between USP2 and USP7, despite their third and fourth critical residues behaving differently (Zhang et al., 2011). We superimposed the currently available crystal structures of USP catalytic domains (Table 1, Figure 1E) and also found only minor differences in the positioning of these two adjacent residues.”
• As the AlphaFold predictions for USP40 and USP48 closely resemble the known structures in Figure 1E, we have added this information as follows : “While the structures of USP40 and USP48 have not been solved, they contain the conserved USP catalytic domain and AlphaFold predictions for USP40 (Uniprot: Q9NVE5) and USP48 (Uniprot: Q86UV5) do not suggest major changes in their catalytic domains."

The introduction of the new terms "critical residue 1 and 2" are confusing and partially disproved by the study itself (replace with e.g. less conserved versus absolutely conserved 3rd triad residue or similar), please revise.

• Thank you, this issue is also mentioned by reviewer 2. We aimed at a solution that would not make inferences on mechanisms. We settled on "critical residue 3" and "critical residue 4", with the active site Cys and His being the first two.

p. 3/4: please add pH information to buffers used in the stability studies. "Previous publication" and "manuscript in preparation" are contradictions.

• pH information has been added.
• Thank you for the comments, we've adjusted the text.

p. 4. Assay buffer for USP1, USP7 and USP48 pH information is missing

• We have corrected the omission.

p. 6: last heading: typo is dispensable

• Typo was corrected.

p. 8: please explain choice of USP1 C90R mutation

• Other mutations tend to increase affinity for free ubiquitin, and in cells this can change ubiquitin homeostasis. The Cys to Arg mutation was shown to avoid this problem in some DUBs. (Morrow et al, Embo Rep 2018 Oct;19(10):e45680. doi: 10.15252/embr.201745680). We have added the reference in both the methods and results sections.

Explain choice of pH range 7-9 studied with regards to anticipated pKas

• We primarily aimed to look at the catalytic cysteine, which needs to be deprotonated in order to allow for catalysis. The sentence on pKas has been removed to avoid confusion. Since the catalytic cysteine in USPs typically has a high pKa, we decided to look at an increased pH to favor partial Cys deprotonation. To that, we have added a reference on USP7, in which it was previously shown USP7 is activated by a higher pH, which holds true for both full-length and its catalytic domain (Faesen et al., (2011). Molecular Cell, 44(1), 147–159. https://doi.org/10.1016/j.molcel.2011.06.034).

Importance of mutagenesis for studying enzymatic mechanisms is clear but limitations also need to be discussed; introduction of local changes etc.. this should be added to the discussion

• We have extended the discussion of limitations as requested. Importantly, the new USP1 asparagine mutants relieve some of the limitations of using alanine substitutions, which we also addressed in this section of the discussion: “While alanine mutations leave open an empty space, or take away the negative charge whenever an aspartate is mutated, mutating both critical residues to asparagine in USP1 did not alleviate the decrease in catalytic competence. Additionally, all single critical residue mutants remained stable and some mutants retaining most of their catalytic competence suggests that these enzymes still function properly.”

  Table 1: linear not lineair

• Thank you. We have made the change.

Table 2: add information for mutant names (exact residue numbers) these data correspond to to improve clarity

• Thank you. We have made the change.

Fig. 1D which structure is shown?

• USP7 (1NBF), we have adjusted the legend.

Fig. 4 bands for USP1/UAF1 D752A and USP15 WT/mutants very faint so difficult to see whether there is crosslinking or not, please comment

• We performed the experiment again and made new figures with better contrast.

Fig.5: please see above for comment about graph and remove or revise.

• We have adjusted the legend to make the diversity more clear: “These five USPs share the conserved USP catalytic domain but vary considerably in domain architecture and allosteric regulation, and therefore represent a part of the diversity found in the USP family.”

Suppl. Table 2: global fit analysis not appropriate for when a poor fit was obtained or where the mutants were barely active (Figs S2, S3). These constants should be removed from the table or more information on the fitting provided. There seems to be some correlation between barely active mutants and the thermal stability, please comment.

• We prefer to do the global fit analysis, as it enables us to share rate constants and get meaningful comparisons. All USP variants were fit simultaneously using the global fit approach where k1 and k-3 rate constants were fixed, k-1 and k3 were shared for all the data sets of the same USP and only k2 was fitted for each data set separately. The quality of the global fit correlates with standard errors of k-1 and k3 rate constants. So, the model we use fits reasonably well with all the data sets all together. Even though a few fitted curves are not aligned well with some of the data for mutants with low activity the value of k2 is still important to report since it gives an approximation of magnitude for the catalytic activity and high standard error reflects the quality of the fit for those specific data sets. In addition, kcat/Km values for all the proteins, including low activity mutants, calculated from global fit approach correlate well with the values calculated from Michaelis-Menten analysis. We clarified this in the legend of supplementary figure 3: “Our kinetic model fits the data well. No fit could be obtained for USP15D880A since no activity was detected. We got relatively poorer fits to USPs with low activity, USP1D752A, USP7D481A). Still, for these low activity USPs the reported Kcat/Km gives an approximation of the magnitude for the catalytic activity and the poorer fit is reflected by their relatively higher standard errors reported in supplementary table 3.”

Suppl. Fig. 1B: See above.

• See comment on 3.

  **Referees cross-commenting**

  reviewers' comments are balanced

  Reviewer #1 (Significance (Required)):

  The study builds on previous work on USP7 and USP2 and while not a conceptual advance, adds to our understanding and knowledge of USP mechanisms. The in cellulo work of probing critical residues in USP1 for processing PCNA-Ub adds a new dimension. However, the limitations of some of the experimental design, stability of mutants and choice of USPs (as outlined above) somewhat hamper the direct comparisons the study makes and previous work needs to be adequately represented (USP2). The work will be of interest to basic researchers and medicinal chemists in particular.

• We very much appreciate the enthusiasm of the reviewer for our cellular validation.

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

  Dr. Sixma is a leading expert in DUB enzymology, especially the enzymology of USP family members. This manuscript is a welcome addition to the field and her body of work to date. Exploring the possibility of redundant or entirely new catalytic residues in USPs is indeed an important venture for differentiating these highly homologous enzymes. The paper is well-written, and the experiments are simplistic and understandable. However, as a whole, the work is not ground-breaking, and the mechanistic explanation of the experimental observation lacks substantiating evidence. The manuscript should be recommended for publication in an appropriate journal after some revision.

  Major comments: - A major concern of the article is about the mechanistic explanation of the role of the second critical residue Asp. The authors proposed two different possible mechanisms, including 1. the residue is flexible to position itself to replace the role of the canonical general base "first" critical residue; 2. Cys/His forms a dyad as seen in other cysteine proteases, and the "second critical residue" Asp participates in the oxyanion hole to stabilize the activated substrate. However, as the authors argue in their discussion, both mechanisms are speculative and have major issues: mechanism #1 requires the catalytic His to flip, and the conformation of the His and "second" critical residue is not optimal for them to form a hydrogen bond directly. The author suggested it may be mediated by a water molecule. However, no such structure has been reported. Mechanism #2 also has the trouble of lacking experimental evidence, and since the tetrahedral oxyanion intermediate is negatively charged, the same negatively charged Asp would be unfavourable. Without mechanistic evidence, the observation of the second (more) critical residue Asp is a very interesting one but beyond that, most of the discussions are speculative. The activity-based labelling experiment using Ub-PA, and the cellular experiments using the mutants only confirmed the observation but can not approve any of these mechanisms.

• Indeed, we do not come with a full mechanistic explanation which explains catalysis in all USPs. Instead, we show that individual USPs have greatly different dependence on their catalytic residue, and thus display important mechanistic distinctions, both for nucleophilic attack and for completion of the reaction. The new Asn mutations do show that negative charge in the 4th critical residue is critical for USP1 function, while the new stopped-flow analysis reveals that USP15 is trapped after the first turnover when the 4th critical residue is lost, and that this is not the case for the other USPs tested.

  • The possibility of substrate trapping in some mutants is of interest. Paragraph 5 of the discussion even mentions this. I think this should be investigated by single-turnover assay techniques.

• We are very thankful for this great suggestion. We performed fast kinetics assays (stopped flow) for all USP wildtype and alanine variants. Together with the Ub-PA labelling experiments these assays shed new light on the ability of these USPs to perform a nucleophilic attack. In terms of substrate trapping, it does indeed turn out that USP15 is inactivated after the first turnover (Figure 4B).

  Minor general concern: - The naming of the Asp/Asn/Ser in the canonical triad is a bit confusing. It is called "the third catalytic residue" and then the "first critical residue" (Intro, last paragraph). This is confusing because, in the catalytic triad, Cys/His are also critical residues. Given the importance of the fourth Asp residue, maybe the authors should come up with a different naming system. One suggestion could be calling the Asp/Asn/Ser the **general base residue** (in the canonical triad terms, Cys is the nucleophile, His is the general acid-base residue, Asp is the general base residue), and the 4th Asp as the "alternative general base residue"?

• Reviewer 1 also did not like the naming. To address the issue we have settled on: "critical residue 3" and "critical residue 4", with the active site Cys and His being the first two. This avoids assigning mechanistic roles to particular residues, but still stresses their importance.

  • The augment at the end of the discussion that this alternative Asp residue could lead to new inhibitors for this difficult class of cysteine proteases is a stretch. The majority, if not all, structurally defined inhibitors of USPs (USP7, USP1, USP14) are allosteric inhibitors that do not target the catalytic triad directly. I doubt the discovery of Asp will change that. The most variability of activity regulation of USPs comes from auxiliary domains of the FL USPs, or cofactor proteins, as the authors' lab has previously demonstrated for many of the USPs, including USP7, USP4, USP1, etc., and there lie more opportunities for new inhibitor discovery.

• We agree that current inhibitors would not make use of these variations, but we feel that our findings could spark an interest in developing new classes that would benefit from the variability. We have adjusted the discussion to make that point more explicitly: “The variety in catalytic mechanisms might allow for development of new types of inhibitors with improved specificities.”

  • Similarly, it is a fancy term to cite of DUBTACs, but I don't see much relevance of this alternative residue applied to DUBTACs. The authors could explore the idea a bit if they decide to cite this.

• Indeed, only if the such new inhibitors can be made. We’ve removed the sentence on DUBTACS.

  Minor comments and grammar: editing is difficult without the inclusion of line numbers. I have attempted to address errors the best I can, considering this.

  • Synopsis: "..., the majority of USPs **does** not..." should be "**do**"

• Correction was made

  • Synopsis: "..., either critical **residues** can..." should be "**residue**"

• Correction was made

  • Intro: "Subsequently a tetrahedral..." should have a comma after subsequently

• Correction was made

  • Intro: 2nd paragraph, line 6, be more specific to be "peptide bond."

• Correction was made

  • Intro: in the 3rd paragraph, the residue numbers of the catalytic residues should be stated.

• The numbers were added

  • Intro: the first line of paragraph 4. The statement is confusing and should be made clearer by simply stating, "The third catalytic residue in USPs is either Asp, Asn, or Ser."

• Correction was made

  • Intro: second last paragraph, be a bit more specific on what "resembles USP15 and USP7" could be "... USP8, another USP whose catalytic triad resembles those of USP15 and USP7" because the domain structure of these FL USPs is very different, only the triad is similar.

• We agree and we apologize for this oversight, we have deleted the sentence on USP8 as it is not relevant in this context.

  • Intro: the last paragraph mentions the loss of function USP15 mutation behaves like wild type and USP1. The term "loss of function" is misleading. If mutation to the canonical 3rd catalytic residue has no effect on activity, then it is not a loss of function mutant. Please specify the alanine mutation.

• We've made this change

  • Intro: last paragraph, "Michaelis Menten," should have a hyphen in between.

• Correction was made

  • Methods: please add a space between values and units; this comes up multiple times throughout the manuscript

• Corrections have been made

  • Methods: all taxonomic names should be italicized, i.e., E. coli

• Correction was made

  • Methods: protein stability section, "**build**-in" should be "**built**-in" (build-in is repeated elsewhere and needs to be fixed)

• Correction was made

  • Methods: structure superposition section, "... bound to ubiquitin were **use** whenever..." should be "...bound to ubiquitin were **used** whenever..."

• Correction was made

  • Methods: pH analysis section, "duplo" should be duplicate

• Correction was made

  • Methods: Expression of USP1 in RPE1 cells section, please briefly state how you determined the expression level of USP1 in transduced RPE1 USP1KO cells when selecting clones with comparable levels to RPE1 wt cells

• We have added an extended description on how we selected these single clones. “To select clones with similar USP1 levels compared to endogenous, single clones were incubated with 1 µg/ml doxycycline for 44 hours and were lysed using RIPA buffer (1% NP40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 0.01 M sodium phosphate pH 7.5, 2 mM EDTA), containing cOmplete™, EDTA-free Protease Inhibitor Cocktail (Roche, 11873580001), 1 mM 2-chloroacetamide and 0.25 U/µl benzonase (SC-202391, Santa Cruz Biotechnology). Total protein concentration in the lysate was determined using a BCA assay (23227, Thermo Scientific) so that equal amounts could be loaded on gel. Samples were loaded on 4-12% Bolt gels (NW04127, Thermo Scientific), and run for 40 minutes at 180 V in MOPS running buffer (B0001, Thermo Scientific). Proteins were transferred to nitrocellulose membrane (10600002, Amersham Protran 0.45 NC nitrocellulose). Membranes were stained with a USP1 antibody (14346-1-AP, Proteintech). After incubation with HRP coupled secondary antibody the blots were imaged using a Bio-Rad Chemidoc XRS+. Using Bio-Rad ImageLab 5.1 software, USP1 levels were quantified by measuring the volume intensities of each USP1 band for each clone and compared this to endogenous USP1 levels in RPE1 cells. Clones with comparable expression levels were selected and used for further experiments.”

  • Methods: tCoffee webserver should be "T-Coffee"

• We realized that multiple sequence alignment was performed using Clustal Omega, not T-Coffee, which has now been corrected. We apologize for this oversight.

  • Methods: MSA. Can the authors provide more details on when doing BLAST, what were the criteria of selecting sequences from the result?

• Details have been added: “Catalytic domains as defined by Uniprot of the resulting human USPs were used for multiple sequence alignment. For USPs with multiple isoforms, the canonical isoform (isoform 1) was selected. In case of the USP17 gene family, USP17L2/DUB3 was selected (Komander et al., 2009). In order to properly align USP1, its inserts were removed from the catalytic domain following (Dharadhar et al., 2021). In order to properly align USP40, a shorter sequence was used (residues 250-480).”

  • Methods: please provide the details for determining the concentration of the enzymes used.

• Details on how we determined the concentrations of enzymes have been added.

  • Methods: Please provide the manufacturers of the Pherastar plate reader and the 384-well plate (please correct from "384 well-plate").

• Info on the manufacturers has been added.

  • Results: In paragraph 1, "lies a **much better** conserved..." you should use "more highly."

• Correction was made

  • Results: paragraph 1, "USP50 does not harbor either of" should be "USP50 harbors neither of"

• We corrected this: “This aspartate is present in all USPs except CYLD and USP50. The latter misses the third critical residue as well and therefore may be inactive.”

  • Supp Fig 2: USP39 does not have glutamate in position of the first critical residue, it is glutamine (Q)

• Correction was made

  • Results: second subsection title **"The first critical residue is dispenUSP1..."** needs to be fixed

• Correction was made as follows: The third critical residue is dispensable in USP1/UAF1, USP15, USP40 and USP48

  • Results: pg. 8 last line "to crosslink", the word crosslink is not proper for the reaction between Ub-PA with USPs. It usually refers to a reactive linker that links two molecules. Words like "conjugate", "conjugation," or "covalent react with", and "activity-based labelling" are probably better choices depending on the context.

• We have corrected this throughout the manuscript.

  • Figure 1: figure legend describing B, C, and D are mixed up.

• Correction was made

  • Results: In paragraph 9, the statement that your data on 5 USPs is representative of most of the 57 members in that the third catalytic is dispensable is not a sound statement for the small sample size. I think more emphasis on the diversity of USP1, USP7, USP15, USP40, and USP48 needs to be stated to help bolster such a claim. The statement to follow, which mentions sequence analysis alone is not able to predict the catalytic residue, is also somewhat contradictory to the opening statement and insinuates that all active USPs should be tested, while you only examined 5.

• We have changed this to ”Our findings demonstrate that for the majority of tested USPs…”. The diversity of tested USPs is clarified earlier in the manuscript: “These USPs vary in domain architecture and allosteric regulation, and therefore represent different aspects of the USP family, known for its structural variety and modular architecture”. The statement about sequence analysis has been removed from the results section and is now only mentioned in the discussion. However, we do think that precise active site assignment for other USPs will require mutagenesis support.

  • Figure 4: legend title, the critical residues are not responsible for **performing** nucleophilic attack per se; that is the job of Cys. The title of the figure should be altered to clear this up.

• Correction was made as follows: " Variation in the ability of USP critical residue mutants to successfully and efficiently facilitate a nucleophilic attack.”

  • Discussion: paragraph 3, since the Hu 2002 USP7 mechanism is not valid for other USPs tested, the "consensus USP catalytic mechanism" should be referred to as the "canonical."

• Indeed! Correction was made.

  • Discussion: paragraph 4, "USP7, USP15 and USP40 all **three** have misaligned..." should be "USP7, USP15 and USP40 all have misaligned..."

• Correction was made.

  • Discussion: paragraph 8, "negative charge itself could **contributes**..." should be "negative charge itself could **contribute**..."

• Correction was made

  • Discussion: pg. 10, 3rd paragraph. Is the first sentence a statement of fact or a hypothesis? The writing is not clear to differentiate the two possibilities.

• Parts of the discussion have been rewritten, but the corresponding sentence has been rewritten as follows: “Canonically, it is thought that the fourth critical residue is involved in oxyanion hole formation.”

  • Discussion: pg. 10, 3rd paragraph, line 3, which "critical residue" does it refer to, the general base residue or the alternative residue?

• We've changed the text as follows: ". A dual role, with the third or fourth critical residue stabilizing catalytic histidine and oxyanion hole formation simultaneously is unlikely”.

  • Discussion: pg. 10, second last paragraph. Can the statement that "inaccurate assumptions about the catalytic triad ... be substantiated with an example?

• We apologize for the possible confusion, but our point here was to point out that it could be misdirecting conclusions if you strictly follow the canonical assignment of the catalytic triad. We have rewritten the sentence to make that more clear: “Additionally, assumptions about the catalytic triad solely based on the canonical catalytic triad assignment in USP could affect conclusions made regarding loss of function mutations in genetic screens. For example, we find that some USPs retain full or most of their activity once their canonical third catalytic residue is mutated.”

  • Table 1, "ubiquitin variant" is mostly often used in the literature to refer to the ubiquitin mutants generated by phage display pioneered by the Sidhu lab or designed mutants. "ubiquitin and homolog derivatives" is a better term for "ubiquitin variant" in this article.

• We have changed this to ubiquitin-like proteins

  • Table 1, the USP21 line "Lineair" is a typo, it should be "linear."

• Correction was made

  • References: citations for Cadzow, 2020. and Tsefou, 2021 do not appear in the bibliography.

• Correction was made

  • Add a hyphen to "Ubiquitin-specific proteases."

• Correction was made

  Reviewer #2 (Significance (Required)):

  General assessment:

  Based on the studies of prototypical ubiquitin-specific protease USP7, the field generally accepts that USPs are a class of cysteine proteases that contain a catalytic triad with a cysteine, a histidine and a general base residue (asparagine, aspartate, or serine). This manuscript described the importance of an alternative, highly conserved aspartate that plays a critical role in catalysis using an enzyme kinetics study on five out of 57 USPs. The work is a very interesting observation that could change the perception in the field. However, the atomic details of how this fourth, or alternative residue, plays its role in catalysis are not clear without the structure evidence of an intermediate/transition state-bound complex.

  Advance:

  The study provided the first systematic enzymology study of the role of a fourth conserved residue critical for the catalysis of USPs. It is a conceptual advance and a first step to elucidate possibly a new catalytic mechanism of USPs.

  Audience: The manuscript will be of interest to biochemists in the field of ubiquitination and drug discovery.

  Reviewers' expertise

  The reviewers are structural biologists with expertise in the structure, function and enzymology of ubiquitin enzymes in general, with practical experience in drug discovery targeting the DUB and kinase families.

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

  The article by Keijzer and colleagues describes an interesting study comparing the active site of multiple USPs (the largest subfamily of deubiquitinases) and elucidating the importance of specific residues lining the active site for catalysis. The authors carried out a careful analysis of the kinetic properties of 5 representative USPs and mutants thereof revealing a remarkable variety in their function that highlights that the majority of USPs studied do not require the canonical third residue of the catalytic triad of USPs for activity but instead rely on a highly conserved second critical residue. Furthermore, the authors apply complementary experimental approaches (mutagenesis, pH dependence of activity, crosslinking with Ub-PA) to allow distinguishing between residues important for the nucleophilic attack versus oxyanion hole stabilisation.

  This is a well-written, thorough enzymatic study of high technical quality. The experiments are described in sufficient detail to allow others to reproduce the experimental set up. The data presented fully support the claims of the paper and no additional experiments are required to further support the conclusions. It is great to see that the authors have carried out thermal stability assays on all WT and mutant proteins under investigation to ensure that any effects observed are not due to protein misfolding.

  Minor comments:

  • There are a few typos in the manuscript the authors should correct.

• Thank you, we have removed the typos from the manuscript.

  • The panels/paper legends to Figure 1B/C/D are mixed up. Please correct.

• Correction was made

  -It would be helpful to use different colours in the alignment shown in Supplementary Figure1 to indicate the position of the first and second critical residue.

• Thank you, we have highlighted these residues

  • I wonder if the authors could comment on how representative the 5 USPs characterised in this work are of the entire family.

• We address the variation of these USPs in more detail, both in the results as in the legend of figure 5: “These USPs vary in domain architecture and allosteric regulation, and therefore represent different aspects of the USP family, known for its structural variety and modular architecture”

Reviewer #3 (Significance (Required)): Deubiquitinating enzymes (DUBs) play essential roles in many cellular processes and their activity is associated with a variety of diseases. There is a lot of interest in targeting DUBs for therapeutic purposes and a number of small molecule inhibitors are undergoing clinical studies. While the structure and mechanism of multiple DUBs have been studied over the years, many open questions about their detailed catalytic mechanism remain and the importance of specific residues might often have been inferred based on sequence conservation alone without accompanying experimental support. This work makes an important contribution to the field by systematically examining 5 members of the USP family and defining the precise role of the first and second critical residue for the catalytic cycle. This work will be of interest to those studying the mechanism of DUBs in general and those trying to target specific DUBs with small molecules. In addition, this study will also be interesting more generally for those studying enzyme kinetics as it highlights the importance of experimental validation of a catalytic mechanism that has been predicted based on sequence conservation or structural studies.