Neuronal SKN-1B modulates nutritional signalling pathways and mitochondrial networks to control satiety

The feeling of hunger or satiety results from integration of the sensory nervous system with other physiological and metabolic cues. This regulates food intake, maintains homeostasis and prevents disease. In C. elegans, chemosensory neurons sense food and relay information to the rest of the animal via hormones to control food-related behaviour and physiology. Here we identify a new component of this system, SKN-1B which acts as a central food-responsive node, ultimately controlling satiety and metabolic homeostasis. SKN-1B, an ortholog of mammalian NF-E2 related transcription factors (Nrfs), has previously been implicated with metabolism, respiration and the increased lifespan incurred by dietary restriction. Here we show that SKN-1B acts in two hypothalamus-like ASI neurons to sense food, communicate nutritional status to the organism, and control satiety and exploratory behaviours. This is achieved by SKN-1B modulating endocrine signalling pathways (IIS and TGF-β), and by promoting a robust mitochondrial network. Our data suggest a food-sensing and satiety role for mammalian Nrf proteins.

I found this paper very interesting and that the experimental evidence sufficiently supported their claims. However, I do have some minor comments to improve the clarity of some experiments and results. We are extremely pleased that you share our enthusiasm for this study and that they find the experimental evidence compelling. We have addressed your comments by carrying out a small number of simple, defined experiments.

1.
Because there are so many different ways to control the amount of food in the environment and is related to such a crucial result, I would have found a brief description of the dietary restriction protocols used and the main differences helpful, whether in the methods or the body of the results, rather than just citing other papers. We agree that this would be helpful. We have added a table (Table S3) which details each DR protocol and their relationship to skn-1. This is referred to in the text where DR protocols are discussed, the legend of Figure 1, and the methods.

I think it would be useful to measure quiescence in WT and skn-1b mutants under fed conditions.
The authors show that skn-1b mutants explore the plate much less than WT (Fig. 2A) and are larger than WT animals (Fig. 2E), leading one to think this must be due to skn-1b mutants eating more while on the lawn. However, after fasting (Fig. 2C), skn-1b mutants actually spend much of that time in quiescence and not eating. The phenotypes seen in the animal size (Fig. 2E) and food ingestion (Fig. 2F) experiments seem quite large given the modest effects skn-1b has on pumping. Perhaps an extra figure panel showing pumping period would also be helpful to add more context to the pumping data in Fig. 2D. Thank you for suggesting this experiment. We report that an absence of skn-1b caused strikingly reduced exploratory behaviour (in fed conditions), which could be explained by increases in quiescence, particularly as we observe skn-1b mutants to exhibit increased satiety induced quiescence in response to fasting and refeeding. To test this we measured the worm's movement on a continuous layer of food using a worm tracker (collaboration with Magnitude Biosciences) and quantified the periods of time when individual worms were not moving i.e. in quiescence. We found that skn-1b mutants actually moved more than WT in fed conditions, meaning that they are highly unlikely to be quiescing more. This contrasts to the role for skn-1b in suppressing satiety induced quiescence. To address your original point, it is therefore unlikely that the slight increase in size and pumping rate is the main driving factor for entry into quiescence. Our new data has been added to Figure 2H and discussed in the m/s. "To test this we measured quiescence in fed conditions using an automated tracking system that detects the time C. elegans spends stationary vs moving. We assumed that stationary worms are quiescing as both roaming and dwelling involve movement. We found that in fed conditions, skn-1b mutants spent less time stationary than WTs, indicating that they quiesce less in fed conditions ( Figure 2H). This contrasts with our data in fasted and re-fed conditions, and indicates that skn-1b is specifically required for satiety induced quiescence." Methods for these automated measurements are also now included in the appropriate section. From this data we could extrapolate that if skn-1b mutants roam less but quiesce the same, that they therefore dwell for longer -we comment on this in the discussion. "We found that SKN-1B acts specifically to suppress satiety induced quiescence ( Figure 2H). In fact, our movement data in fed conditions suggests that skn-1b mutants may move slightly more than WT ( Figure 2H). As fed skn-1b mutants explore less than WT, we could extrapolate that in fed conditions, they spend more time dwelling. Therefore SKN-1B acts to control different behaviours depending on food status." Fig. 5D could use some clarification. The skn-1b;daf-2;daf-16 triple mutant was more quiescent than the skn-1b;daf-2 double mutant, despite the fact that daf-16 is required for the increased quiescence of daf-2 mutants. This could indicate that there is an unknown factor that becomes important only in the absence of skn-1b and daf-16. However, given that TGF is known to regulate insulin signaling (Greer et al. 2008, Shaw et al. 2007) and quiescence, perhaps this data is bringing attention to the interaction of these two pathways. In other words, skn-1b could be directly regulating the expression of DAF-7/TGF, which then is what is actually responsible for regulating insulin signaling. This would be fairly easy to test by looking at various double mutants between these pathways. This is an incredibly interesting point, and we thank the reviewer for raising it. Our data so far suggests that SKN-1B acts upstream of both IIS and TGFβ to control behaviour, placing it as a new central node in the ASI behavioural response pathways. The suggested relationship between TGFβ and IIS is certainly possible. Increased nuclear localisation of DAF-16 is observed in daf-7 mutants (Shaw et al 2007) however, this does not marry with the idea that daf-7 mutants do not quiesce (Gallagher et al., 2013;You et al., 2008;Figure 4A), and that daf-7 suppresses the quiescence of skn-1b mutants ( Figure 4A). However as SKN-1B is acting so centrally, it was possible that in its absence this relationship alters. To test whether or not IIS and TGFβ interact to control quiescence downstream of skn-1b we measured satiety induced quiescence in daf-7 and daf-7; skn-1b mutants in the presence and absence of daf-16 RNAi. However, daf-16 RNAi did not cause any change in quiescence duration or numbers of worms entering quiescence in either daf-7 or daf-7; skn-1b mutants. These data have been included in Figure 4E and Figure S6F and commented on in the text (see below). NB: daf-16 RNAi is effective at suppressing the increase quiescence of daf-2 mutants ( Figure 5E) confirming that technically, this experiment is sound. We conclude that IIS and TGFβ do not interact to control quiescence downstream of skn-1b.

3.The data in
"TGF-β and Insulin signalling also interact and increased nuclear localisation of DAF-16 is observed in daf-7 mutants (Shaw et al 2007). Although daf-7 mutants do not quiesce (Gallagher et al., 2013;You et al., 2008;Figure 4A), and daf-7 fully suppresses skn-1b mutant quiescence ( Figure 4A), we wondered if removal of DAF-16 had the potential to alter this relationship. To test whether IIS and TGF-β interact to control quiescence downstream of skn-1b we measured satiety induced quiescence in daf-7 and daf-7; skn-1b mutants with and without daf-16 RNAi. However, daf-16 RNAi did not cause any changes in quiescence for either daf-7 or daf-7; skn-1b mutants ( Figure 4E and Figure S6F). We conclude that IIS and TGF-β do not interact to control quiescence downstream of  Reviewer #2 (Significance (Required)): This manuscript represents a conceptual advance in the understanding of SKN-1/Nrf function, satiety quiescence regulation, and fills in a gap between sensory perception and the resulting endocrine signaling. The authors find a new function of SKN-1/Nrf as a central node to regulate endocrine signaling responses to environmental food cues. Further, this function is specific to a single isoform, skn-1b, and is independent of dietary restriction-mediated longevity, which had previously been the main link between SKN-1 activity and metabolism (Bishop and Guarente, 2007). This paper would be of interest broadly to anyone studying how sensory perception affects behavior, endocrine signaling, and mitochondrial networks. My expertise lies in C. elegans, aging, andinsulin signaling. Bishop N a, Guarente L. 2007. Two neurons mediate diet-restriction-induced longevity in C. elegans. Nature 447:545-549. doi:10.1038/nature05904 We fully appreciate the recognition that our data provides a conceptual advance in how we view sensory perception and endocrine signalling, as well as identifying a novel function of SKN-1/Nrfs. As the reviewer points out, all past work has focused on SKN-1B and dietary restriction but this study demonstrates that SKN-1B has much broader metabolic roles. We agree that this paper would be of interest to a broad audience i.e. those from metabolic, behavioural, endocrine, mitochondrial, and ageing biology backgrounds, and that it suggests novel metabolic functions of mammalian Nrf proteins. From this perspective, PLoS Genetics is an ideal journal.

Reviewer #3 (Evidence, reproducibility and clarity (Required)):
Tataridas-Pallas and Thompson et al. describe a role for the ASI-expressed SKN-1 isoform SKN-1B in mediating food sensing and food-related behaviour in C. elegans. The authors show that SKN-1B modulates satiety quiescence by supressing DAF-7 expression within the ASI neurons, which alters downstream TGF-beta signalling. SKN-1B affects the morphology of mitochondria in distal tissues, which modulates the behavioural feeding state of the animal in response to altered food conditions. **Major comments:** *Are the key conclusions convincing?* 1. Although it has been previously shown that skn-1 plays a role in DR-longevity, the statement that "skn-1b contributes to, but is not essential for DR longevity" is not convincing based on the presented data. The eat-2 genetic model is a complex model for dietary restriction (for example PMID: 30965033), and skn-1b was required for longevity enhancement only for the genetic (eat-2) DR model, and not in the bacterial-dilution DR model used. Given that, among other food-sensing/eating phenotypes, eat-2 mutants have reduced pharyngeal pumping, and the evidence that follows in the manuscript show that skn-1b mutants have increased pharyngeal pumping/food intake, these data suggest that eat-2; skn-1b double mutants are likely to have wild-type food intake, i.e. are not dietary restricted at all. Thus, it cannot be concluded that skn-1b has any effect on the longevity by dietary restriction but affect the efficacy of using the eat-2 mutant as a genetic model for DRlongevity itself.
• This is an interesting point. To be clear, in our statement " skn-1b contributes to DR mediated longevity under some conditions, it is not necessarily essential" we intended to refer not only to the eat-2 and bacterial dilution experiments in Figures 1A-B, but also to the bacterial dilution experiment previously published by the Guarente lab (Bishop and Guarente 2007). In this 2007 work, they showed that a skn-1(zu135) mutant (which lacks all skn-1 isoforms) did not exhibit an increased lifespan in response to DR. However, the DR longevity response was rescued by expressing skn-1b specifically in the ASI neurons. Thus, in that scenario skn-1b did indeed contribute to DR lifespan. This paper was cited in this section of the manuscript, but we will also add it following this statement to increase clarity.
"We conclude that although skn-1b contributes to DR mediated longevity under some conditions, it is not necessarily essential ( Figures 1A-B) (Bishop and Guarente, 2007).
• We agree that the eat-2 DR model is complex. To address the issue of pumping rate in eat-2; skn-1b double mutants confounding its DR potential, we measured and compared pumping rate in eat-2 vs eat-2; skn-1b animals. We found no difference in pumping rate between these two strains suggesting that the requirement of skn-1b for eat-2 lifespan is not due to a change in eat-2 food intake. We have included this information in the methods section.. " "Note JMT7 was genotyped used a PCR for skn-1b(tm4241) and a pumping rate assay for eat-2. eat-2 pumping was ~90pumps/min (compared to ~250pumps/min for WT) but no difference in pumping rate (p=0.66) was detected between eat-2 and eat-2; skn-1b." This is an appropriate place for this control data as it was generated at the same time as the eat-2; skn-1b double mutant was constructed, as part of the genotyping protocol.
• Whilst discussing the methods, it seems a good place to note that in the revised version we removed a sentence from the methods that included information on an additional, transcriptional reporter strain that we had made. We had referenced it as (data not shown). For clarity we have now removed it as it is not relevant to this m/s. It is marked on the m/s.

2.
That skn-1b is regulating exploration behavior, and acts in the same genetic pathway as the gyanylate-cyclase daf-11 in exploration behaviour is convincing. As is the data showing the skn-1b and daf-11 regulate the appropriate exploration behavioural response to fasting and refeeding. DAF-11 enables ASIs to sense the external environment, and is required for the skn-1b expression response to fasting, however how daf-11 alters skn-1b levels, given their different subcellular locations is unclear. The data clearly show that DAF-7 was required for the increased satiety quiescence in the skn-1b mutant, and loss of skn-1b leads to increased daf-7 expression in the ASI neurons. However, the opposing roles of daf-7 and skn-1b in quiescence are not applicable to the exploration behaviour, where daf-7 and skn-1b appear to both promote exploration. Given that daf-11 is required for daf-7 expression in the ASI (PMID: 11677050), placing this ASI gene regulatory program into more context may make these findings more convincing and clear.
• Thank you for this insightful comment and we are glad you find our data relating to skn-1b and daf-11 exploratory behaviour convincing. We believe that this nicely links environmental food status with the main body of the ASI neuron. Understanding how daf-11 ultimately impacts on skn-1b activity is fascinating but beyond the scope of this study. The behavioural aspects of these molecular interactions are highly testable though and boil down to differences in behaviour in response to different food conditions. In fasted: refed conditions daf-11 and daf-7 are required for satiety induced quiescence (Gallagher et al., 2013) and here we show that skn-1b suppresses satiety induced quiescence.
We tested the contribution of skn-1b to daf-11 satiety induced quiescence and found that similarly to daf-7, daf-11 was also required for skn-1b mutants to increase satiety induced quiescence. This data has been added to Figure 2D and commented on in the text. "One factor controlling satiety induced quiescence is daf-11, and daf-11 mutants cannot quiesce (You et al., 2008). To test the relationship between skn-1b and daf-11 in this regard we measured quiescence in daf-11; skn-1b double mutants. Although daf-11 mutation did slightly reduce the % of skn-1b mutants entering quiescence by ~20%, we found that daf-11 completely suppressed the long quiescence duration of skn-1b mutants ( Figure 2D and Figure S6B). This suggests that SKN-1B and DAF-11 have opposing roles in controlling satiety induced quiescence, but that SKN-1B requires functional DAF-11 to act as a molecular switch." In addition, we measured skn-1b movement compared to WT animals in fully fed conditions using an automated system. We found that skn-1b mutants actually move more than WTs, and combined with our exploration data we can extrapolate from this that they may well dwell more (similarly to daf-7 mutants). We present this data in Figure 2H and comment on it in the results and in the discussion.
RESULTS "To test this we measured quiescence in fed conditions using an automated tracking system that detects the time C. elegans spends stationary vs moving. We assumed that stationary worms are quiescing as both roaming and dwelling involve movement. We found that in fed conditions, skn-1b mutants spent less time stationary than WTs, indicating that they quiesce less in fed conditions ( Figure  2H). This contrasts with our data in fasted and re-fed conditions, and indicates that skn-1b is specifically required for satiety induced quiescence. " DISCUSSION " We found that SKN-1B acts specifically to suppress satiety induced quiescence ( Figure  2H). In fact, our movement data in fed conditions suggests that skn-1b mutants may move slightly more than WT ( Figure 2H). As fed skn-1b mutants explore less than WT, we could extrapolate that in fed conditions, they spend more time dwelling. Therefore SKN-1B acts to control different behaviours depending on food status. Thank you for this comment. However, we believe that our data does show that skn-1b interacts with the Insulin/daf-2 pathway: To control DAF-16 nuclear localisation; exploration behaviour; and satiety induced quiescence. This interaction may not be linear in every case, and may vary depending on food status but it is definitely there. Examining each food condition separately.

The data and conclusions about SKN-1B's interaction with insulin
• Exploratory behaviour in fed conditions: We show that skn-1b cannot further suppress the exploration deficiency of two different daf-2 alleles, daf-2(e1370) (Figure 5B WT vs daf-2(e1370) p=0.0001); daf-2 vs daf-2; skn-1b p=NS)) and daf-2(e1368) (Figure S11A (WT vs daf-2(e1368) p=0.0001); daf-2 vs daf-2; skn-1b p=NS) mutants. The reviewer suggests that it would be challenging to further suppress (the low) exploration of the daf-2(e1370) mutant, but we also showing the same effect using a milder daf-2(e1368) allele (Gems et al., 1998), which exhibits an intermediate reduction in exploration. Overall, we believe that SKN-1B and IIS do act in the same pathway to control exploration in fed conditions. • Exploratory behaviour in fasted, re-fed conditions: Similarly to the fed scenario, we found that skn-1b could not further suppress the daf-2(e1370) fasted, re-fed exploration phenotype. As above, the reviewer suggests that it would be challenging to further suppress (the low) exploration of daf-2(e1370) mutant in these conditions. Therefore, we repeated this experiment using the milder e1368 allele. We found that fasting and re-feeding caused a further reduction in daf-2(e1368) exploration, but that the effect was more intermediate compared to e1370. We did not observe any further suppression of this effect with skn-1b mutation. This strengths our argument that IIS and skn-1b interact epistatically in relation to exploratory behaviour in fasted: re-fed conditions. This data has been added to Figure S11B and reference to the figure added at each point where e1370 data is discussed.
The e1368 data is discussed in the text as follows.
"As the class 1 allele daf-2(e1370) already exhibits very low exploratory behaviour it may be difficult to suppress further, so we also tested a class 2 allele daf-2(e1368) which exhibits a milder exploratory defect. Similarly to our e1370 results however, daf-2(e1368) and skn-1b exploratory defects were nonadditive in both fed, and fasted and re-fed conditions ( Figure S11A and S11B). " • Satiety induced quiescence behaviour: The reviewer notes that an additive quiescence effect is seen between daf-2 and skn-1b mutations following fasting and refeeding. That could imply that the two act independently. However, while the key transcriptional mediator of IIS (daf-16) is required for the high quiescence of daf-2 mutants, in the absence of skn-1b this effect is lost. Taken together with the information that skn-1b is required for daf-16 to be maintained in intestinal nuclei we interpret this as suggesting that IIS and skn-1b interact but that additional factors are required. We have now carried out additional controls, and found that daf-16 RNAi cannot contribute to quiescence in either a WT or skn-1b mutants (although similar percentages of each population enter quiescence). Therefore, we conclude that IIS needs to be reduced for daf-16 to be required for satiety induced quiescence. This data is now presented in Figure 5D and Figure S6D and the following added to the text.
"We found that whilst daf-16 RNAi had no effect on either WT or skn-1b mutant quiescence, daf-2 mutation enhanced quiescence compared to WT and effect suppressed by daf-16 RNAi ( Figures 5D-E and Figures S6D-E)."

4.
The data on mitochondrial networks are unclear. The fluorescence microscopy analysis demonstrates that skn-1b mutant has a high proportion of fragmented and very fragmented mitochondria compared to wild-type, which is exacerbated with fasting. However, the electron microscopy data shows that skn-1b mutants exhibit a more fused-like state than wild type. It is unclear what the explanation is for these opposing results.
Thank you for raising this. We believe that the concept of the neuroendocrine system and skn-1b/Nrf impacting on mitochondrial dynamics and ultimately behaviour is a unique and fascinating aspect of our story. Conceptually however, you have hit on a technical point that should be made to a wide audience; that direct comparison of fluorescent and TEM images do not always match. We do not believe that our results are opposing, but agree that we could add further information to support this. We have tackled this in 3 ways: • 1: We have expanded our genetic experiments to examine both mitochondrial fission and fusion. We treated animals with eat-3, fzo-1 and drp-1 RNAi (causing increased mitochondrial fission, fission and fusion respectively) and compared fluorescent and TEM images in each case. This demonstrates the mitochondrial network pattern observed when known mitochondrial network proteins are disrupted. One thing that is noticeable from this analysis is that a "spotty" fluorescent image indicates that the mitochondrial network is disrupted, but can indicate either a fused or fissioned network. This fluorescent vs TE microscopy comparative data is highly informative and presented in Figures 6A-D and Figures 14A and 14B and we make this point in the text.
"In our hands, although the fluorescent images provided evidence of mitochondrial disruption in each case, it was the electron microscopy that showed the precise nature of the disruption (Figures 6A-E and Figures S14A-B). " "Mitochondrial membrane proteins are required for mitochondrial fusion and fission: eat-3/Opa1 and Fzo-1/Mfn1 promote fusion and drp-1/Drp1 promotes fission (Spurlock et al., 2020). We examined muscle mitochondrial networks, in C. elegans fed either eat-3, fzo-1 or drp-1 RNAi using both fluorescent (myo-3::GFP(mit) and electron microscopy. Mitochondria in animals fed eat-3 or fzo-1 RNAi are smaller and more disjointed (as the mitochondria are unable to fuse), whereas those in drp-1 RNAi fed animals are more elongated (as they cannot fission) ( Figure S14A-B)." • 2: Expanded our data set examining the link between skn-1b, mitochondrial dynamics and behaviour. We had originally shown that eat-3 RNAi (causing mitochondrial fission) is sufficient to reverse the skn-1b reduced exploration phenotype. We have now tested the effect of both mitochondrial fission and fusion on skn-1b behaviour. We fed WT or skn-1b mutants with eat-3, fzo-1 or drp-1 RNAi and measured their exploration. These new data support our claim that breaking up the skn-1b fused mitochondrial network with either eat-3 or fzo-1 RNAi is sufficient to rescue skn-1b exploratory behaviour. However, fusing the mitochondrial network using drp-1 RNAi had no effect on behaviour in either a WT or skn-1b mutant background. The fluorescent and TEM data demonstrating the efficacy of each RNAi treatment is shown in Figures 14A and 14B and the behavioural data is in Figures 6F and 6G. demonstrate the effectiveness of each genetic intervention on the mitochondrial network. The text has been altered to include. Figures 6G and 6H). drp-1 RNAi however, had no effect on either WT or skn-1b behavioural patterns ( Figures 6G and 6H). "

"Mitochondrial dynamics have previously been implicated in behavioural responses (Byrne et al., 2019). So, given the behavioural role of skn-1b and its importance for maintaining mitochondrial networks, we tested whether the two were linked. Strikingly, we found that whilst neither eat-3 or fzo-1 RNAi had any effect on WT exploratory behaviour, both completely rescued skn-1b mutant exploration to normal levels (
• 3: Altered our descriptive language of mitochondrial changes to improve clarity. Given our data in Figures 6DA-D and Figures 14A-B showing differences between fluorescent and TEM, we have altered the language in Figure 6C (which shows a more qualitative analysis of this mitochondrial disruption, determined by examining fluorescent images and grading them).
The key now reads, "Very disorganised, Disorganised, Some disorganisation and Very organised". This does not alter the data presented but removes assumptions about the precise nature of the mitochondrial networks which could be construed as misleading. We have also changed our language in the text to match e.g. " We found the networks in skn-1b mutants to have a disorganised appearance, covering significantly less surface area than that of the WT (Figures 6A-C and Figure S12C). " In addition, we have added in data from a neat experiment we did rescuing the skn-1b mitochondrial phenotype using our SKN-1B::GFP transgene. These data have been added to Figure  6D and are commented on in the text.
"This phenotype could be rescued by re-introducing SKN-1B::GFP into skn-1b mutants ( Figure 6D)." *Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?*

The leaving assay on different bacteria (fig1I) is insufficient to show that skn-1b acts to sense food types, this would require testing whether skn-1b chooses one bacteria over another in a food choice assay (see PMC1352325). Instead the data show that skn-1b mutants behave in a satiated manner, regardless of the "quality" of the bacteria.
To address this we have altered the text from "As almost all skn-1b mutants are present on an OP50 lawn, it implies that while WT worms adapt to preferentially consume some foods, skn-1b mutants do not" to; "As almost all skn-1b mutants are present on an OP50 lawn, it implies that they are behaving in a satiated manner".

2.
The data forming the conclusion that SKN-1B::GFP expression responds to different bacterial types and dilutions, and responds selectively to the amount of food available requires more explanation and interpretation. As the two bacteria (PY79 and PA14) have the opposite effect on the behaviour, and that skn-1B was not required for sensing and avoiding PA14, please clarify how the skn-1b expression level is reconciled with opposite outputs in behaviour. This is an interesting point. We have added the following to the discussion.
"It is intriguing that the response of SKN-1B::GFP expression to diet and associated behavioural responses are not always consistent i.e. similar SKN-1B::GFP increases in expression with B. Subtilis (PY79) and Pseudomonas (PA14) leads to opposite behavioural responses ( Figures 3A-B and Figure  S5C). The transcriptional outputs of other SKN-1 isoforms are known to differ depending on the stimulus (Oliveira et al., 2009b) and it is possible that this is also the case for SKN-1B. Alternatively, in the case of the pathogen response, perhaps additional immune signals are sufficient to override any satiety behaviour in the skn-1b mutants."

Loss of SKN-1B does not affect DAF-16 localisation in either fasted or fed conditions, but was required for DAF-16::GFP to be maintained in the nucleus after fasting/refeeding. Could this be a downstream effect of the increased feeding behaviour of the skn-1b mutant? Given that skn-1b mutants have a much higher food intake and pumping rate, does this lead to a "satiated/fed" metabolic state much faster than in wild-type worms? If the skn-1b mutant recovers from fasting far more quickly than wild-type worms, does this then lead to a fast recovery of daf-16 localisation?
We know that skn-1b mutants spend longer in quiescence following fasting and refeeding but we don't know if this equates to skn-1b mutants recovering faster than WTs from the fasting. The skn-1b mutants only have a very moderate increase in body size, gut E. coli and pumping rate ~10% under these conditions, whereas daf-16 mutants demonstrate ~50% increase in food consumption e.g. Wu et al., 2019 PMID: 30905669. If the assumption is that DAF-16 acts to limit feeding then its seems counterintuitive that it would be "lost" in the skn-1b mutants where it could potentially be considered more important to limit food intake. There are also examples of mutants where food intake is increased e.g. daf-7 mutants (Gallagher et al., 2013)  • We agree that the eat-2 DR model is complex. To address the issue of pumping rate in eat-2; skn-1b double mutants confounding its DR potential, we measured and compared pumping rate in eat-2 vs eat-2; skn-1b animals. We found no difference in pumping rate between these two strains suggesting that the requirement of skn-1b for eat-2 lifespan is not due to a change in eat-2 food intake. We have included this information in the methods section..

"Note JMT7 was genotyped used a PCR for skn-1b(tm4241) and a pumping rate assay for eat-2. eat-2 pumping was ~90pumps/min (compared to ~250pumps/min for WT) but no difference in pumping rate (p=0.66) was detected between eat-2 and eat-2; skn-1b."
This is an appropriate place for this control data as it was generated at the same time as the eat-2; skn-1b double mutant was constructed, as part of the genotyping protocol. • Our data show that SKN-1B is not generally essential for DR (when comparing two protocols in our m/s with one published from the Guarente lab) so we are doubtful that testing multiple other protocols would be informative. In C. elegans, DR protocols give vastly different lifespan extensions and there is even more variability between labs. It is possible that the % extension in lifespan incurred by a DR protocol affects the ability of SKN-1B to contribute to it i.e. both the Guarente DR and eat-2 are fairly modest extensions whilst the Moroz DR extension is huge. Defining this further would be very challenging, and to paraphrase the reviewers own words below, "very time consuming and not essential for publication". The main focus of the manuscript is our work examining the novel role of SKN-1B in food related behaviour.

SKN-1B and DAF-11 act in the same genetic pathway to regulate exploration behavior. SKN-1B acts
to inhibit satiety quiescence and pharyngeal pumping after fasting, while previous data (in discussion, You et al 2008) show that daf-11 promotes satiety quiescence after fasting/refeeding. How does this tie in with exploration, as increased quiescence is the reason why skn-1B mutants do not explore? Performing the quiescence-duration assay in the daf-11, daf-11;skn-1b strains would aid in clarifying whether daf-11 or skn-1b has the dominant role in regulating pharyngeal pumping.
We agree that these are interesting questions and have already discussed (above) the interesting interplay between skn-1b, daf-11 and daf-7 in a single pair of neurons (the ASIs). To address this we tested the contribution of skn-1b to daf-11 satiety induced quiescence and found that similarly to daf-7, daf-11 was also required for skn-1b mutants to increase satiety induced quiescence. This data has been added to Figure 2D and commented on in the text.
"One factor controlling satiety induced quiescence is daf-11, and daf-11 mutants cannot quiesce (You et al., 2008). To test the relationship between skn-1b and daf-11 in this regard we measured quiescence in daf-11; skn-1b double mutants. Although daf-11 mutation did slightly reduce the % of skn-1b mutants entering quiescence by ~20%, we found that daf-11 completely suppressed the long quiescence duration of skn-1b mutants ( Figure 2D and Figure S6B). This suggests that SKN-1B and DAF-11 have opposing roles in controlling satiety induced quiescence, but that SKN-1B requires functional DAF-11 to act as a molecular switch." In addition, we measured skn-1b movement compared to WT animals in fully fed conditions using an automated system. We found that skn-1b mutants actually move more than WTs, and combined with our exploration data we can extrapolate from this that they may well dwell more (similarly to daf-7 mutants). We present this data in Figure 2H and comment on it in the results and in the discussion.
RESULTS "To test this we measured quiescence in fed conditions using an automated tracking system that detects the time C. elegans spends stationary vs moving. We assumed that stationary worms are quiescing as both roaming and dwelling involve movement. We found that in fed conditions, skn-1b mutants spent less time stationary than WTs, indicating that they quiesce less in fed conditions ( Figure  2H). This contrasts with our data in fasted and re-fed conditions, and indicates that skn-1b is specifically required for satiety induced quiescence. " DISCUSSION " We found that SKN-1B acts specifically to suppress satiety induced quiescence ( Figure  2H). In fact, our movement data in fed conditions suggests that skn-1b mutants may move slightly more than WT ( Figure 2H). As fed skn-1b mutants explore less than WT, we could extrapolate that in fed conditions, they spend more time dwelling. Therefore SKN-1B acts to control different behaviours depending on food status." Together with published data (You et al., 2008;Gallagher et al., 2013), this provides a complete picture of how each ASI gene impacts on satiety induced quiescence, and provides further information relating to the behavioural role of SKN-1B in fed conditions. Figure 5D would benefit by having the wild-type control for comparison, but requires skn-1b alone and with daf-16 RNAi to be able to interpret these data appropriately.

The IIS quiescence measurements presented in
We have added data measuring quiescence duration and % of population in quiescence for the following: WT, skn-1b, WT daf-16 RNAi and skn-1b daf-16 RNAi. These data are presented in Figure  5D and Figure S6D and the following added to the text.
"We found that whilst daf-16 RNAi had no effect on either WT or skn-1b mutant quiescence, daf-2 mutation enhanced quiescence compared to WT and effect suppressed by   Figures S6D-E)."

4.
The eat-3 RNAi rescue of skn-1b exploration shows that inhibiting mitochondrial fusion rescues the skn-1b phenotype, given that the fluorescence microscopy assay showed the opposite effect, then performing a couple of control experiments may help unravel these opposing results: Does the eat-3 RNAi lead to increased fragmentation in the mito::GFP assay? Also, does drp-1 RNAi (preventing fission) have any affect on skn-1b behaviour and mitochondrial morphology? As discussed in detail above, have added data from a set of genetic experiments that fully defines the roles of mitochondrial fission and fusion in regulating behaviour (Figures 6F and 6G). This is complemented by fluorescent and TEM images in each condition to clearly demonstrate what each genetic manipulation (eat-3, fzo-1 and drp-1 RNAi) is doing to the mitochondrial networks ( Figures  14A and 14B). These new data sets fully support our idea that mitochondrial fission promotes normal food-responsive behaviour but suggests that over-fusing mitochondria is not sufficient to alter behaviour (as also shown by Bryne et al., 2019 PMID: 30840087 Figure Figure 4A, this is mentioned in the text but the cites are not present. The text has been altered to read: "In agreement with published work, WT animals showed increased *Do you have suggestions that would help the authors improve the presentation of their data and conclusions?* 1. I would suggest bringing the focus and developing the logical arguments and presentation of the experiments around food sensing, behaviour and metabolism rather than DR-longevity, as it appears in the first figure that mediating DR-longevity is not the biological function of skn-1b, but rather a consequence of its role in feeding/food-sensing.

Please cite which published work is in agreement with
We have tried writing this manuscript in a number of ways and understand this comment. However, we have decided that placing the lifespan data at the start (to get it out the way) is the best way forward. However, We have softened the angle relating to DR lifespan in the abstract to move the focus to food sensing and behaviour. We do not feel that there is an over-emphasis on this data for the rest of the study (and it is not in the title) particularly with all the new data presented. We also note that there are a considerable number of people in the SKN-1 and lifespan fields who would like the SKN-1B DR story clarified, and these experiments will be informative for them.

I found the explanation of daf-11 regulating skn-1b levels a little confusingly written. The exploration data implies an epistatic relationship, the expression pattern data (and what is known of daf-11 function) implies daf-11 drives the link between the external environment and SKN-1B expression levels.
The new data and explanations in the text will hopefully have clarified this section. *Describe the nature and significance of the advance (e.g. conceptual, technical, clinical) for the field.* The ASI neurons are key drivers of food-related behaviour in the worm, and understanding the complexities (and subtleties) of the transcriptional and signalling networks that govern the functional output of these neurons is very interesting. This work adds an additional component to the significant body of work the exists about ASI function. Given that many of the observations and pathways described have been previously published, the nature of this work is less a significant conceptual advance, but rather an extension of our knowledge and understanding of how the ASI neurons can integrate signals from the external environment and propagate these signals, in part through skn-1b activity to the internal metabolic state of the animal.

The inclusion of
Our work identifies SKN-1B as a novel component of the neuroendrocrine network that underlies ASI control of behaviour and mitochondrial physiology. Identifying SKN-1B as a metabolic regulator in this context significantly expands and consolidates what we know of other reported behavioural regulators (daf-11, daf-7, daf-2 and daf-16), forming a coherent picture of how food is sensed by the animal, information computed by the ASI neurons, and the appropriate behavioural and physiological responses instigated. In addition, SKN-1/Nrfs have not been described in these roles before, so this work opens up the novel idea that mammalian Nrfs could also be involved in food sensing and satiety related behaviour. Taken together, this is an important, conceptual, extension of our knowledge.
*Place the work in the context of the existing literature (provide references, where appropriate).* There is a large body of work that demonstrates the importance of the ASI neurons, TGF-beta and insulin-like signalling in feeding, metabolic regulation and food-related behaviour. This work places skn-1b as an environmentally-responsive transcription factor acting in the ASI neurons that modulates outputs of the ASI neurons.
*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.* Feeding behaviour, neuronal transcriptional programs, metabolism. Not a longevity expert or an electron microscopy expert.

REFEREES CROSS-COMMENTING
The review of Reviewer 1 was inappropriate in my opinion. Very few manuscripts are that perfect and having no concerns or suggestions for improvement is worrying. The review system is obviously not there to hinder publication but to provide an unbiased view of the experimentation, data and conclusions drawn.
Reviewer 2 had similar concerns as myself. The manuscript is of interest however the suggested experiments and clarifications by both of us would improve the manuscript.