Avoidance of hydrogen sulfide is modulated by external and internal states in C. elegans

Reviewer #3 (Public review):

Summary:

The manuscript explores behavioral responses of C. elegans to hydrogen sulfide, which is known to exert remarkable effects on animal physiology in a range of contexts. The possibility of genetic and precise neuronal dissection of responses to H2S motivates the study of responses in C. elegans. The revised manuscript does not seem to have significantly addressed what was lacking in the initial version.

The authors have added further characterization of possible ASJ sensing of H2S by calcium imaging but ASJ does not appear to be directly involved. Genetic and parallel analysis of O2 and CO2 responsive pathways do not reveal further insights regarding potential mechanisms underlying H2S sensing. Gene expression analysis extends prior work. Finally, the authors have examined how H2S-evoked locomotory behavioral responses are affected in mutants with altered stress and detoxification response to H2S, most notably hif-1 and egl-9. These data, while examining locomotion, are more suggestive that observed effects on animal locomotion are secondary to altered organismal toxicity as opposed to specific behavioral responedse

Overall, the manuscript provides a wide range of intriguing observations, but mechanistic insight or a synthesis of disparate data is lacking.

Reviewer #4 (Public review):

Summary:

The authors establish a behavioral paradigm for avoidance of H2S and conduct a large candidate screen to identify genetic requirements. They follow up by genetically dissecting a large number of implicated pathways - insulin, TGF-beta, oxygen/HIF-1, and mitochondrial ROS, which have varied effects on H2S avoidance. They additionally assay whole-animal gene expression changes induced by varying concentrations and durations of H2S exposure.

Strengths:

The implicated pathways are tested extensively through mutants of multiple pathway molecules. The authors address previous reviewer concerns by directly testing the ability of ASJ to respond to H2S via calcium imaging. This allows the authors to revise their previous conclusion and determine that ASJ does not directly respond to H2S and likely does not initiate the behavioral response.

Weaknesses:

Despite the authors focus on acute perception of H2S, I don't think the experiments tell us much about perception. I think they indicate pathways that modulate the behavior when disrupted, especially because most manipulations used broadly affect physiology on long timescales. For instance, genetic manipulation of ASJ signaling, oxygen sensing, HIF-1 signaling, mitochondrial function, as well as starvation are all expected to constitutively alter animal physiology, which could indirectly modulate responses to H2S. The authors rule out effects on general locomotion in some cases, but other physiological changes could relatively specifically modulate the H2S response without being involved in its perception.

I am actually not convinced that H2S is directly perceived by the C. elegans nervous system at all. As far as I can tell, the avoidance behavior could be a response to H2S-induced tissue damage rather than the gas itself.

Author response:

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

Reviewer #1 (Public Review):

Summary:

This paper sets out to achieve a deeper understanding of the effects of hydrogen sulfide on C. elegans behavior and physiology, with a focus on behavior, detection mechanism(s), physiological responses, and detoxification mechanisms.

Strengths:

The paper takes full advantage of the experimental tractability of C. elegans, with thorough, welldesigned genetic analyses.

Some evidence suggests that H₂S may be directly detected by the ASJ sensory neurons. The paper provides interesting and convincing evidence for complex interactions between responses to different gaseous stimuli, particularly an antagonistic role between H₂S and O2 detection/response. Intriguing roles for mitochondria and iron homeostasis are identified, opening the door to future studies to better understand the roles of these components and processes.

We thank the reviewer for the supportive comments.

Weaknesses:

The claim that worms' behavioral responses to H₂S are mediated by direct detection is incompletely supported. While a role for the chemosensory neuron ASJ is implicated, it remains unclear whether this reflects direct detection. Other possibilities, including indirect effects of ASJ and the guanylyl cyclase daf-11 on O2 responses, are also consistent with the authors' data.

We thank the reviewer for the insightful comment and agree that the role of ASJ neurons in H₂S detection was not clear. We included new experiments and revised our text to make it clearer.

Since our initial analyses suggest a role of ASJ neurons in H₂S-evoked locomotory responses (Figure 2F and G), We thought that this would offer us a starting point to dissect the neuronal circuit involved in H₂S responses. Expression of the tetanus toxin catalytic domain in ASJ, which blocks neurosecretion, inhibited H₂S-evoked locomotory speed responses (Figure 2H), suggesting that neurosecretion from ASJ promotes H₂S-evocked response (Lines 162–165). We then performed calcium imaging of ASJ neurons in response to H₂S exposure. However, while we observed CO₂-evoked calcium transients in ASJ using GCaMP6s, we did not detect any calcium response to H₂S, under several conditions, including animals on food, off food, and with different H₂S concentrations and exposure times (Figure2—Figure supplement 2E and F) (Lines 166–170). Since signaling from ASJ neurons regulates developmental programs that modify sensory functions in C. elegans (Murakami et al., 2001), the involvement of ASJ neurons is not specific to H₂S and ASJ neurons are unlikely to serve as the primary H₂S sensor (Discussed in Line 449–458). Therefore, the exact sensory neuron, circuit and molecular triggers mediating acute H₂S avoidance remain to be elucidated.

Our subsequent investigation on mitochondrial components suggests that a burst of mitochondrial ROS production may be the trigger for H₂S avoidance, as transient exposure to rotenone substantially increases baseline locomotory speed (Figure 7E) (Line 391–396). However, to initiate avoidance behavior to H₂S, mitochondrial ROS could potentially target multiple neurons and cellular machineries, making it challenging to pinpoint specific sites of action. Nevertheless, we agree that further dissection of the neural circuits and mitochondrial signaling in H₂S avoidance will be important and should be explored in future studies.

The role of H₂S-mediated damage in behavioral responses, particularly when detoxification pathways are disrupted, remains unclear.

We thank the reviewer for the insightful comment and fully agree with the concern raised. The same issue was also noted by the other reviewers. We agree that decreased locomotory responses in H₂S-sensitized animals can arise from distinct causes, either systemic toxicity or behavioral adaptation, and distinguishing between these is critical. We have included new experiments and revised the text to clarify this issue.

Our data suggest that increased initial omega turns and a rapid loss of locomotion in hif-1 and detoxification-defective mutants including sqrd-1 and ethe-1 likely reflect an enhanced sensitivity to H₂S toxicity due to their failure to induce appropriate adaptative responses (Figure 5D–F, Figure 5J–L, Figure 5—Figure supplement 1F–P). Supporting this, hif-1 mutants become less responsive to unrelated stimuli (near-UV light) after 30 minutes of H₂S exposure (Figure 5I).

In contrast, egl-9 and SOD-deficient animals show reduced initial omega-turn and reduced speed responses (Figure 5B, Figure 7G, Figure 5—Figure supplement 1A and B, and Figure 7—Figure supplement 1F and G), although both egl-9 and sod mutants respond normally to the other stimuli prior or after H₂S exposure (Figure 5I, Figure 5—Figure supplement 1C, and Figure 7—Figure supplement 1H). Since disrupting egl-9 stabilizes HIF-1 and upregulates the expression of numerous genes involved in cellular defense against H₂S toxicity, the enhanced detoxification capacity in egl-9 mutants likely increases animals’ tolerance to H₂S, thereby reducing avoidance to otherwise toxic H₂S levels. Similarly, persistently high ROS in SOD deficient animals activates a variety of stress-responsive signaling pathways, including HIF-1, NRF2/SKN-1 and DAF-16/ FOXO signaling (Lennicke & Cocheme, 2021; Patten et al., 2010), facilitating cellular adaptation to redox stress and reducing animals’ responsiveness to toxic H₂S levels. Taken together, these findings support the view that reduced locomotory speed during H₂S exposure can arise from distinct mechanisms: early systemic toxicity in hif-1 and detoxificationdefective mutants, versus enhanced cellular adaptation in egl-9 and SOD mutants. We have integrated the relevant information across the result section and discussed this in Lines 485-536.

The findings of the paper are somewhat disjointed, such that a clear picture of the relationships between H₂S detection, detoxification mechanisms, mitochondria, and iron does not emerge from these studies. Most importantly, the relative roles of H₂S detection and integration, vs. general and acute mitochondrial crisis, in generating behavioral responses are not convincingly resolved.

We thank the reviewer for this comment and agree that our presentation did not fully connect different findings into a cohesive picture. To address this, we have acquired new data, and revised the abstract, results and discussion sections to clarify two phases of H₂S-evoked responses: an initial avoidance behavior upon H₂S exposure, followed by a later phase of adaption and detoxification when the escape is not successful.

In brief, we began with the basic characterization of H₂S-induced locomotory speed response, followed by a candidate gene screen to identify key molecules and pathways involved in initial speed response to H₂S. Subsequently, we focused on three major intersecting pathways that contributed to the acute behavioral response to H₂S. These include cGMP signaling, which led to the identification of ASJ neurons; nutrient-sensitive pathways that modulate behavioral responses to both H₂S and CO2; and O2sensing signaling, whose activation inhibits responses to H₂S. However, the molecules and neurons in these pathways, including ASJ, likely play modulatory roles and are unlikely to serve as the primary H₂S sensors. Our subsequent analysis, however, suggests that mitochondria play a critical role in triggering avoidance behavior upon H₂S exposure. Brief treatment with rotenone, a potent inducer of ROS, led to marked increase in locomotory speed (Figure 7E). This suggests the possibility that a burst of ROS production triggered toxic levels of H₂S (Jia et al., 2020) may initiate the avoidance behavior.

When the initial avoidance fails, H₂S detoxification programs are induced as a long-term survival strategy. The induction of detoxification programs appears to enhance tolerance to H₂S exposure and contributes to the gradual decrease of locomotory speed in H₂S. We now provide a clearer image of how different pathways modulate H₂S detoxification and adaptation (see our responses to other comments). Briefly, mutants defective in detoxification, such as hif-1 and other detoxification-defective mutants, showed stronger initial omega-turn response and a rapid loss of locomotion. This loss of locomotion is likely caused by early cellular toxicity as the mutants failed to respond to other unrelated stimuli (nearUV light) after 30 minutes of H₂S exposure (Figure 5I). Likewise, smf-3 mutants and BP-treated animals were hypersensitive to H₂S (Figure 6D and E, and Figure 6—Figure supplement 1G and I), likely due to impaired H₂S detoxification under low iron conditions, as iron is a co-factor required for the activity of the H₂S detoxification enzyme ETHE-1 (Figure 5K and Figure 5—Figure supplement 1E).

In contrast, reduced locomotion and response in other contexts such as egl-9 mutants and SODdeficient animals reflect H₂S-induced adaptive mechanism rather than toxicity as they remain responsive to the other stimuli after H₂S exposure. Since disrupting egl-9 stabilizes HIF-1 and upregulates the expression of numerous genes involved in cellular defense against H₂S toxicity, the enhanced detoxification capacity in egl-9 mutants likely increases animals’ tolerance to H₂S, thereby reducing avoidance to otherwise toxic H₂S levels. Similarly, persistently high ROS in SOD deficient animals activates a variety of stress-responsive signaling pathways, including HIF-1, NRF2/SKN-1 and DAF-16/ FOXO signaling (Lennicke & Cocheme, 2021; Patten et al., 2010), facilitating cellular adaptation to redox stress and reducing animals’ responsiveness to toxic H₂S levels. Therefore, different animals decline their locomotory speed to the effects of H₂S through distinct mechanisms. We have integrated the relevant information across the result section and discussed this in Lines 485-536.

Reviewer #2 (Public Review):

Summary:

H₂S is a gas that is toxic to many animals and causes avoidance in animals such as C. elegans. The authors show that H₂S increases the frequency of turning and the speed of locomotion. The response was shown to be modulated by a number of neurons and signaling pathways as well as by ambient oxygen concentrations. The long-term adaptation involved gene expression changes that may be related to iron homeostasis as well as the homeostasis of mitochondria.

Strengths:

Overall, the authors provide many pieces that will be important for solving how H₂S signals through neuronal circuits to change gene expression and physiological programs. The experiments rely mostly on a behavioral assay that measures the increase of locomotion speed upon exposure to H₂S. This assay is then combined with manipulations of environmental factors, different wild-type strains, and mutants. The mutants analyzed were obtained as candidates from the literature and from transcriptional profiling that the authors carried out in worms that were exposed to H₂S. These studies imply several genetic signaling pathways, some neurons, and metabolism-related factors in the response to H₂S. Hence the data provided should be useful for the field.

We thank the reviewer for the supportive comments.

Weaknesses:

On the other hand, many important aspects of the underlying mechanisms remain unsolved and the reader is left with many loose ends. For example, it is not clear how H₂S is actually sensed, how sensory neurons are activated and signal to downstream circuits, and what the role of ciliated and RMG neurons is in this circuit. It remains unclear how signals lead to gene expression and physiological changes such as metabolic rewiring. Solving all this would clearly be beyond the scope of a single manuscript. Yet, the manuscript also does not focus on understanding one of these central aspects and rather is all over the place, which makes it harder to understand for readouts that are not in this core field. Multiple additional methods and approaches exist to dig deeper into these mechanisms in the future, such as neuronal calcium imaging, optogenetics, and metabolic analysis. To generate a story that will be interesting to a broad readership substantial additional experimentation would be required. Further, in the current manuscript, it is often difficult to understand the rationales of the experiments, why they were carried out, and how to place them into a context. This could be improved in terms of documentation, narration/explanation, and visualization.

We thank the reviewer for the comment, which has also been raised by the other reviewers. We agree that our initial submission was poorly presented. We also acknowledge the fact that some aspects, such as detailed neural circuit and sensory transduction, still remain unresolved. We have now included additional experiments and revised the manuscript to clarify the logic of our experiments, provided better context for our findings, and improved both the narrative flow and data visualization to make the manuscript more accessible to readers. We now provide a clearer image of how different pathways interact to modulate the initial avoidance response, and the H₂S detoxification and behavioral habituation during prolonged H₂S exposure. The following response is similar to the one for reviewer #1.

In brief, we began with the basic characterization of H₂S-induced locomotory speed response, followed by a candidate gene screen to identify key molecules and pathways involved in initial speed response to H₂S. Subsequently, we focused on three major intersecting pathways that contributed to the acute behavioral response to H₂S. These include cGMP signaling, which led to the identification of ASJ neurons; nutrient-sensitive pathways that modulate behavioral responses to both H₂S and CO2; and O2sensing signaling, whose activation inhibits responses to H₂S. However, the molecules and neurons in these pathways, including ASJ, likely play modulatory roles and are unlikely to serve as the primary H₂S sensors. Our subsequent analysis, however, suggests that mitochondria play a critical role in triggering avoidance behavior upon H₂S exposure. Brief treatment with rotenone, a potent inducer of ROS, led to marked increase in locomotory speed (Figure 7E). This suggests the possibility that a burst of ROS production triggered toxic levels of H₂S (Jia et al., 2020) may initiate the avoidance behavior.

When the initial avoidance fails, H₂S detoxification programs are induced as a long-term survival strategy. The induction of detoxification programs appears to enhance tolerance to H₂S exposure and contributes to the gradual decrease of locomotory speed in H₂S. We now provide a clearer image of how different pathways modulate H₂S detoxification and adaptation (see our responses to other comments). Briefly, mutants defective in detoxification, such as hif-1 and other detoxification-defective mutants, showed stronger initial omega-turn response and a rapid loss of locomotion. This loss of locomotion is likely caused by early cellular toxicity as the mutants failed to respond to other unrelated stimuli (nearUV light) after 30 minutes of H₂S exposure (Figure 5I). Likewise, smf-3 mutants and BP-treated animals were hypersensitive to H₂S (Figure 6D and E, and Figure 6—Figure supplement 1G and I), likely due to impaired H₂S detoxification under low iron conditions, as iron is a co-factor required for the activity of the H₂S detoxification enzyme ETHE-1 (Figure 5K and Figure 5—Figure supplement 1E).

In contrast, reduced locomotion and response in other contexts such as egl-9 mutants and SODdeficient animals reflect H₂S-induced adaptive mechanism rather than toxicity as they remain responsive to the other stimuli after H₂S exposure. Since disrupting egl-9 stabilizes HIF-1 and upregulates the expression of numerous genes involved in cellular defense against H₂S toxicity, the enhanced detoxification capacity in egl-9 mutants likely increases animals’ tolerance to H₂S, thereby reducing avoidance to otherwise toxic H₂S levels. Similarly, persistently high ROS in SOD deficient animals activates a variety of stress-responsive signaling pathways, including HIF-1, NRF2/SKN-1 and DAF-16/ FOXO signaling (Lennicke & Cocheme, 2021; Patten et al., 2010), facilitating cellular adaptation to redox stress and reducing animals’ responsiveness to toxic H₂S levels. Therefore, different animals decline their locomotory speed to the effects of H₂S through distinct mechanisms. We have integrated the relevant information across the result section and discussed this in Lines 485-536.

Reviewer #3 (Public Review):

Summary:

The manuscript explores the behavioral responses of C. elegans to hydrogen sulfide, which is known to exert remarkable effects on animal physiology in a range of contexts. The possibility of genetic and precise neuronal dissection of responses to H₂S motivates the study of responses in C. elegans. The manuscript is well-written in communicating the complex physiology around C. elegans behavioral responses to H₂S and in appropriately citing prior and related relevant work.

There are three parts to the manuscript.

In the first, an immediate behavioral response-increased locomotory rate-upon exposure to H₂S is characterized. The experimental conditions are critical, and data are obtained from exposure of animals to 150ppm H₂S at 7% O2. The authors provide evidence that this is a chemosensory response to H₂S, showing a requirement for genes encoding components of the cilia apparatus and implicating a role for tax-4 and daf-11. Neuron-specific rescue in the ASJ neurons suggests the ASJ neurons contribute to the response to H₂S. One caveat is that previous work has shown that the dauer-constitutive phenotype of daf-11 mutants can be suppressed by ASJ ablation, suggesting that there may be pervasive changes in animal nervous system signaling that are ASJ-dependent in daf-11 mutants, which may indirectly alter chemosensory responses to H₂S. More direct methods to assess whether ASJ senses H₂S, e.g. using calcium imaging, would better assess a direct role for the ASJ neurons in a behavioral response to H₂S. The authors also point out interesting parallels between the response to H₂S and CO2 though provide some genetic data separating the two responses. Importantly, the authors note that when aerotaxis (O2sensing and movement) in the presence of bacterial food is intact, as in npr-1 215F animals, the response to H₂S is abrogated. Mutation in gcy-35 in the npr-1 215F background restores the H₂S chemosensory response.

There is a second part of the paper that conducts transcriptional profiling of the response to H₂S that corroborates and extends prior work in this area.

The final part of the paper is the most intriguing, but for me, also the most problematic. The authors examine how H₂S-evoked locomotory behavioral responses are affected in mutants defective in the stress and detoxification response to H₂S, most notably hif-1. Prior genetic studies have established the pathways leading to HIF-1 activation/stabilization, as well as potential downstream mechanisms. The authors conduct logical genetic analysis to complement studies of the hif-1 mutant and in part motivated by their transcriptional profiling studies, examine the role of iron sequestration/free iron in the locomotory response to H₂S, and further speculate on how the behavior of mutants defective in mitochondrial function might be affected by exposure to H₂S.

In some regard, this part of the manuscript is interesting because the analysis begins to connect how the behavior of an animal to a toxic compound is affected by mutations that affect sensitivity to the toxic compound. However, what is unclear is what is being studied at this point. In the context, of noting that H₂S at 150ppm is known to be lethal, its addition to mutants clearly sensitized to its effects would be anticipated to have pervasive effects on animal physiology and nervous system function. The authors note that the continued increased locomotion of wild-type animals upon H₂S exposure might be due to the byproducts of detoxification or the detrimental effects of H₂S. The latter explanation seems much more likely, in which case what one may be observing is the effects of general animal sickness, or even a bit more specifically, neuronal dysfunction in the presence of a toxic compound, on locomotion. As such, what is unclear is what conclusions can be taken away from this part of the work.

Strengths:

(1) Characterization of a motor behavior response to H₂S

(2) Transcriptional profiling of the response to H₂S corroborating prior work.

We thank the reviewer for the supportive comments.

Weaknesses:

Unclear significance and experimental challenges regarding the study of locomotory responses to animals sensitized to the toxic effects of H₂S under exposure to H₂S.

We thank the reviewer for the comment, which has also been raised by the other reviewers. We agree that our initial submission left several important questions open, and we acknowledge the fact that some aspects, such as detailed neural circuit and sensory transduction, still remain unresolved. Nevertheless, we acquired new data and revised our text, aiming to clarify the distinct mechanisms underlying the reduced locomotion in different mutants during prolonged H₂S exposure.

Our data suggest that increased initial omega turns and a rapid loss of locomotion in hif-1 and detoxification-defective mutants including sqrd-1 and ethe-1 likely reflect an enhanced sensitivity to H₂S toxicity due to their failure to induce appropriate adaptative responses (Figure 5D–F, Figure 5J–L, Figure 5—Figure supplement 1F–P). Supporting this, hif-1 mutants become less responsive to unrelated stimuli (near-UV light) after 30 minutes of H₂S exposure (Figure 5I).

In contrast, egl-9 and SOD-deficient animals show reduced initial reorientation and reduced speed responses (Figure 5B, Figure 7G, Figure 5—Figure supplement 1A and B, and Figure 7—Figure supplement 1F and G), although both egl-9 and sod mutants respond normally to the other stimuli prior or after H₂S exposure (Figure 5I, Figure 5—Figure supplement 1C, and Figure 7—Figure supplement 1H). Since disrupting egl-9 stabilizes HIF-1 and upregulates the expression of numerous genes involved in cellular defense against H₂S toxicity, the enhanced detoxification capacity in egl-9 mutants likely increases animals’ tolerance to H₂S, thereby reducing avoidance to otherwise toxic H₂S levels. Similarly, constant high ROS in SOD deficient animals activates a variety of stress-responsive signaling pathways, including HIF-1, NRF2/SKN-1 and DAF-16/ FOXO signaling (Lennicke & Cocheme, 2021; Patten et al., 2010), facilitating cellular adaptation to redox stress and reducing animals’ responsiveness to toxic H₂S levels. Taken together, these findings support the view that reduced locomotory speed during H₂S exposure can arise from distinct mechanisms: early systemic toxicity in hif-1 and detoxification-defective mutants, versus enhanced cellular adaptation in egl-9 and SOD mutants. We have integrated the relevant information across the result section and discussed this in Lines 485-536.

Reviewer #1 (Recommendations For The Authors):

To better substantiate a role for H₂S detection, it would be useful for the authors to image Ca responses to H₂S in ASJ in WT and unc-13, and to rule out the possibility that the requirement for daf-11 in ASJ reflects a role in O2 rather than H₂S detection.

We thank the reviewer for this comment. As suggested, we performed calcium imaging of ASJ neurons using GCaMP6s. As previously described, 3% CO₂ evoked a calcium transient in ASJ (Figure 2—figure supplement 2F). To investigate whether H₂S evoked a calcium transient in ASJ neurons, we tested several conditions, including animals on food or off food, with different H₂S concentrations (~75 or ~150ppm), and different exposure time (4 or 8 mins). However, we did not detect a calcium response to H₂S in ASJ under any of the conditions tested (Figure2—figure supplement 2E) (Lines 166–168). Given that neuronspecific rescue of daf-11 or tax-4 mutants pointed to a role of ASJ neurons in promoting H₂S responses, we sought to determine how ASJ neurons were involved. Expression of the tetanus toxin catalytic domain in ASJ neurons, which blocks neurosecretion, inhibited H₂S-evoked locomotory speed responses (Figure 2H), similar to the phenotypes observed in daf-11 and daf-7 mutants (Figure 2C and D) (Lines 162–165). These results confirm that ASJ activity and neurosecretion contribute to the H₂S responses, although ASJ is unlikely to serve as the primary H₂S sensor. One potential explanation is that DAF-7 released by ASJ controls the starvation program, which in turn modulates the animal’s response to H₂S. We also discussed this in Lines 449–458.

The paper would be significantly strengthened by testing the possibility (as the authors acknowledge in lines 348-52) that disruption of detoxification mechanisms reduces sustained behavioral responses to H₂S because of physiological damage. Authors use acute exposure to high O2 for this purpose earlier in the paper, but not to probe the consequences of loss of hif-1 and detoxification factors.

We thank the reviewer for the valuable suggestion. As the reviewer highlighted, we attributed the brief locomotory speed responses to H₂S observed in hif-1 mutants to the lack of detoxification response, leading to the rapid intoxication of the animals. Several lines of evidence support this conclusion. First, we observed that hif-1 and the detoxification mutants displayed a stronger initial reorientation response (omega turns) and a more rapid decline in speed and reversals compared to wild type (Figure 5 D–F). Second, to test if hif-1 mutants were indeed more susceptible to H₂S toxicity, we exposed WT and hif-1 animals to H₂S for 30 mins and subsequently tested their ability to respond to near-UV light. Unlike WT animals, the speed response to near-UV light was inhibited in hif-1 mutants (Figure 5I), suggesting that exposure to H₂S for 30 min causes a stronger toxicity in animals deficient of HIF-1 signaling. Third, hif-1 and detoxification mutants displayed a sustained high speed in response to 1% O₂ , suggesting the specific impairment of H₂S response. The data were presented in Lines 318–347, and were further discussed this in Lines 485–508.

To better understand whether mitochondrial damage has a role in H₂S-evoked behavior, it might be useful for the authors to determine whether general ROS response pathways are important for H₂S behavioral responses.

We thank the reviewer for this insightful comment. As suggested, we investigated whether ROS detoxification pathways contribute to H₂S-evoked locomotory speed responses by analyzing mutants in the superoxide dismutase (SOD) family. These experiments, together with other observations, suggest that mitochondrial ROS play a dual role in H₂S-evoked locomotion. The relevant results were presented in Lines 401–425, and were further discussed in Lines 509–536.

First, we found that increased mitochondrial ROS formation, either induced pharmacologically by rotenone or genetically in mitochondrial electron transport chain (ETC) mutants (Ishii et al., 2013; Ochi et al., 2016; Ramsay & Singer, 1992; Yang & Hekimi, 2010; Zorov, Juhaszova, & Sollott, 2014), suppressed the behavioral response to toxic H₂S (Figure 7A–E). This indicates that mitochondrial ROS plays a significant role in H₂S-evoked responses. One likely explanation is that high ROS formation may dampen the H₂S-triggered ROS spike, or may impair other H₂S signaling processes required to initiate avoidance. Second, consistent with previous reports (Onukwufor et al., 2022), we observed that shortterm rotenone exposure (<1 hour) significantly increased baseline locomotory speed. Given that toxic H₂S levels promote ROS formation (Jia et al., 2020), our findings suggest that acute mitochondrial ROS production by toxic levels of H₂S exposure may serve as a trigger for the avoidance response.

In contrast, animals with sustained mitochondrial ROS production do not have an increased baseline locomotory speed. This effect was observed after 2 hours of rotenone exposure, in mitochondrial ETC mutants, and in animals lacking all SOD enzymes (Figures 7A–K). A likely explanation for the reduced basal locomotory speed during sustained mitochondrial ROS production is the activation of ROSresponsive signaling pathways including HIF-1, NRF2/SKN-1, and DAF-16/FOXO (Lennicke & Cocheme, 2021; Patten, Germain, Kelly, & Slack, 2010), which may promote adaptation to prolonged oxidative stress (Figure 7H). Notably, unlike hif-1 mutants, SOD-deficient animals remained as responsive as WT to other stimuli after 30 minutes of H₂S exposure (Figure 7—figure supplement 1H), indicating that elevated ROS levels do not compromise overall viability or the ability to detoxify H₂S.

Taken together, these results support a model in which mitochondrial ROS exerts a biphasic effect on H₂S-induced avoidance. It enhances detection and avoidance under acute stress but contributes to locomotory suppression when ROS levels remain elevated chronically.

Reviewer #2 (Recommendations For The Authors):

The way the manuscript is presented could be improved without much effort by rewriting/editing. For the reader, it is hard at present to understand the rationales of the experiments, why they were carried out, and how to place them into a context. This could be improved on three levels:

(1) Documentation

(2) Narration/Explanation

(3) Visualization

(1) Documentation

Not all of the results in the text are well documented. The results should be described with more details in the written text and improved documentation and quantification of the results. Example:

Turning behavior is mentioned as an important aspect of the response to H₂S. There is no citation given but this effect is not well documented. The authors image the animals and could provide video footage of the effect, could quantify eg turning/pirouettes, and provide the data. At the moment the manuscript largely relies on measuring the increase in speed, but the reader is left wondering what other behavioral effects occur and how this is altered in all of the mutant and other conditions tested. Just quantifying speed reduces the readout and seems like an oversimplification to characterize the behavioral response.

We are grateful for this comment. We now provide a video footage of the H₂S effects (Figure 1—Video 1). As suggested, we analyzed the recordings to extract reorientation (omega-turns) and reversals. These analyses are now included in the Supplemental file 1 with representative panels displayed in Figure 5 and supplements to Figures 2, 3, 5, 6 and 7. Even though the mutant effects on omega-turns were often subtle, and reversal responses showed considerable variability (likely due to differences in population density, food availability, or animals’ physiological state prior to the assay), this analysis has proven valuable for distinguishing mutants that exhibit adaptation from those that display hypersensitivity to H₂S toxicity. For instance, although both SOD-deficient and BP-treated animals failed to increase their locomotory speed in H₂S (Figure 6E and Figure 7G), they exhibited distinct omega-turn responses (Figure 6—figure supplement 1I and Figure 7—figure supplement 1F), suggesting that different mechanisms likely underlie the locomotory defects of these two animals. We have integrated the omega-turn and reversal data into the text and discussed under relevant contexts.

(2) Narration/Description.

Generally, the description of the results part is very brief and it is often not clear why a certain experiment was carried out and how. Surely it is possible to check the methods but this interrupts the flow of reading and it would be easier for the reader to be guided through the results with more information what the initial motivation for an experiment is, what the general experimental outline is, and what specific experiments are carried out.

We apologize for the lack of clarity and logical structure in the initial submission. In the revised manuscript, we have thoroughly revised the text to improve its organization and readability.

Examples:

Line 97ff: The authors performed a candidate screen yet it is not described why which genes were chosen. Are there also pathways that were tested that turned out to not be involved?

We thank the reviewer for the suggestion. To address this, we have added a new section, explaining the rationale for selecting genes and pathways in our candidate screen. Briefly, we focused on genes known or predicted to be involved in sensory responses to gaseous stimuli in C. elegans and mammals, including globins and guanylate cyclases (21% O₂ sensing), potassium channels (acute hypoxia), and nutrientsensitive pathways (CO₂ responses). We also included mutants defective in sensory signal transduction and neurotransmission. In addition, mitochondrial mutants were analyzed because mitochondria play a central role in H₂S detoxification. The pathways that contributed to the acute H₂S response included cGMP, insulin, and TGF-β signaling, as well as mitochondrial components. In contrast, globins, potassium channels, and biogenic amine signaling did not appear to play significant roles under our assay conditions. The results of the candidate screen are described in Lines 106–138 and summarized in Supplementary File 1.

line 262ff: the paragraph starts with explaining ferritin genes that are important for iron control but the reader does not yet know why. Then it is explained that a ferritin gene is DE in the H₂S transcriptomes. then a motivation to look into the labile iron pool is described. Why not first explain what genes are strongly regulated and why they are selected based on their DE? Then explain what is known about these genes and pathways, and then motivate a set of experiments.

We agree with the reviewer that our initial description could have been more logically organized. We reframed this section to first present the RNA-seq data, followed by an explanation of their known biological functions and the motivation for the subsequent experiments (Lines 350–357).

nhr-49 appears suddenly in the results part and it is not clear why it was tested and how the result links. Is nhr-49 a key transcription factor that is activated by H₂S sensory or physiological response, and does it control the signaling or protective changes induced by H₂S?

We thank the reviewer for the comment. As suggested, we revised the text to present the information more clearly. In our candidate gene screen, a set of mutants exhibiting reduced speed responses to H₂S has previously been shown to be defective in response to CO₂ stimulation (Hallem & Sternberg, 2008). These included animals deficient in nutrient-sensitive pathways, including insulin, TGF-beta, and NHR49, which were reported by Sternberg’s lab to exhibit dampened responses to CO₂ (Hallem & Sternberg, 2008) (Lines 173–179). We also included a simply cartoon to further illustrate this (Figure 3C).

The nuclear hormone receptor NHR-49 has been implicated in a variety of stress responses, including starvation (Van Gilst, Hadjivassiliou, & Yamamoto, 2005), bacterial pathogen (Naim et al., 2021; Wani et al., 2021), and hypoxia (Doering et al., 2022). The nhr-49 mutants exhibited a rapid decline in locomotory speed during H₂S exposure, implicating a role in sustaining high speed in the presence of H₂S. Furthermore, we observed that fmo-2, a well-characterized target gene of NHR-49, was significantly upregulated after 1 hour of exposure to 50 and 150 ppm H₂S (Supplementary file 2), suggesting that NHR-49 signaling is rapidly activated by H₂S exposure. Exactly how NHR-49 contributes to H₂S response requires further investigation.

(3) Visualization

Adding a model/cartoon summary that describes the pathways tested and their interaction would be helpful in some of the figures for the reader to keep an overview of the pathways that were tested. Also, a final summary cartoon that integrates all the puzzle pieces into one larger picture would be helpful. Such a final cartoon overview could also point to the key open questions of the underlying mechanisms.

We thank the reviewer for this suggestion. We have added a series of models/cartoons to illustrate the different pathways and their interactions. These include starvation regulatory mechanisms (Figure 3C), 21% O₂ sensing mechanisms (Figure 3G), HIF-1 signaling and detoxification (Figure 5—figure supplement 1E), HIF-1 signaling and the regulation of labile iron (Figure 6H), as well as ROS signaling and regulation (Figure 7L). To help interpretation and to elaborate on these models, we have also included explanatory sentences in the corresponding figure legends.

Other comments:

Introduction and line 93: The authors mention that 50 ppm H₂S has beneficial effects on lifespan yet does not have a detectable phenotype." Are there any concentrations of H₂S that cause attraction of C. elegans and what is the preferred range if it exists? Could this be measured in an H₂S gradient?

We thank the reviewer for the insightful comment. We performed an H₂S gradient assay, which suggests that wild type animals are attracted toward low concentrations of H₂S around 40 ppm (Figure 1G and H) (Lines 95–104). These results suggest that H₂S acts as a strong repellent for C. elegans at high concentrations but as an attractant at low levels. This dual role may be ecologically relevant, as wild C. elegans lives in complex and dynamic environments where H₂S levels likely fluctuate over short distances (Adams, Farwell, Pack, & Bamesberger, 1979; Budde & Roth, 2011; Morra & Dick, 1991; Patange, Breen, Arsuffi, & Ruvkun, 2025; Rodriguez-Kabana, Jordan, & Hollis, 1965; Romanelli-Cedrez, Vairoletti, & Salinas, 2024).

Line 146: "Local H₂S concentrations could also be significantly higher in decomposing substances where wild C. elegans thrives" please provide a citation.

As suggested, we included a set of references that have described the H₂S enrichment in the natural environment in early field studies (Adams et al., 1979; Morra & Dick, 1991; Rodriguez-Kabana et al., 1965), as well as references that have discussed and implied this in C. elegans studies (Budde & Roth, 2011; Patange et al., 2025; Romanelli-Cedrez et al., 2024). They can be found in the introduction (Lines 59–62) and in the result (Lines 197–199).

Line 311 "Wild C. elegans isolates thrive in the decomposing matters, where the local concentrations of O2 are low while the levels of CO2 and H₂S could be high. These animals have adapted their behavior in such an environment, displaying increased sensitivity to high O2 exposure but dampened responses to CO2." Please provide citations for these statements.

As suggested, we cited the relevant articles or books describing the variation of O₂ and CO₂ levels in the decomposing matters including several C. elegans papers that mentioned this in Lines 197–199 (Bretscher, Busch, & de Bono, 2008; Gea, Barrena, Artola, & Sanchez, 2004; Hallem & Sternberg, 2008; Oshins, Michel, Louis, Richard, & Rynk, 2022), and the above-mentioned articles for H₂S (Adams et al., 1979; Budde & Roth, 2011; Morra & Dick, 1991; Patange et al., 2025; Rodriguez-Kabana et al., 1965; Romanelli-Cedrez et al., 2024).

For C. elegans’ sensitivity to O2 and CO2, these articles were cited in Lines 201–203 (Beets et al., 2020; Bretscher et al., 2008; Carrillo, Guillermin, Rengarajan, Okubo, & Hallem, 2013; Hallem & Sternberg, 2008; Kodama-Namba et al., 2013; McGrath et al., 2009).

Reviewer #3 (Recommendations For The Authors):

More work could be conducted establishing the neuronal circuitry involved in the initial, tractable response to H₂S.

We thank the reviewer for the insightful comment. Since our initial analyses suggest a role of ASJ neurons in H₂S-evoked locomotory responses (Figure 2F and G), We thought that this would offer us an entry point to dissect the neuronal circuit involved in H₂S responses. Expression of the tetanus toxin catalytic domain in ASJ, which blocks neurosecretion, inhibited H₂S evoked locomotory responses (Figure 2H), suggesting that neurosecretion from ASJ promotes the speed response to H₂S (Lines 162– 165). We then performed calcium imaging of ASJ neurons in response to H₂S exposure. However, while we observed CO₂ -evoked calcium transients in ASJ using GCaMP6s, we did not detect any calcium response to H₂S, under several conditions, including animals on food, off food, and with different H₂S concentrations and exposure times (Figure2—Figure supplement 2E and 2F) (Lines 166–168). Since signaling from ASJ neurons regulates developmental programs that modify sensory functions in C. elegans, including CO₂ and O₂ responses (Murakami, Koga, & Ohshima, 2001), the involvement of ASJ neurons is not specific to H₂S responses and ASJ neurons are unlikely to serve as a primary H₂S sensor (Discussed in Line 449–458). Therefore, the exact sensory neuron, circuit and molecular triggers mediating acute H₂S avoidance behavior remain to be elucidated.

Our subsequent investigation on mitochondrial components suggests that a burst of mitochondrial ROS production may be the trigger for H₂S avoidance, as transient exposure to rotenone substantially increases baseline locomotory activity (Figure 7E) (Line 391–396). However, mitochondrial ROS could potentially target multiple neurons and cellular machineries to initiate avoidance behavior to H₂S, making it challenging to pinpoint specific sites of action. Nevertheless, we agree that further dissection of the neural circuits and mitochondrial signaling in H₂S avoidance will be important and should be explored in future studies. We discussed this in Lines 509–536.

I am not sure how to overcome the challenges involved in reaching conclusions from the decreased locomotory responses of animals that are sensitized to the effects of H₂S. Perhaps this conundrum could be discussed in more detail in the text.

We thank the reviewer for this important comment. We agree that decreased locomotory speed during H₂S exposure can arise from distinct causes, either systemic toxicity or adaptation, and distinguishing between these is critical. We have included new experiments and revised the text to clarify this issue.

Our data suggest that increased initial omega turns and a rapid loss of locomotion in hif-1 and detoxification-defective mutants including sqrd-1 and ethe-1 likely reflect an enhanced sensitivity to H₂S toxicity due to their failure to induce appropriate adaptative responses (Figure 5D–F, Figure 5J–L, Figure 5—Figure supplement 1F–P). Supporting this, hif-1 mutants become less responsive to unrelated stimuli (near-UV light) after 30 minutes of H₂S exposure (Figure 5I).

In contrast, egl-9 and SOD-deficient animals show reduced initial reorientation and reduced speed responses (Figure 5B, Figure 7G, Figure 5—Figure supplement 1A and B, and Figure 7—Figure supplement 1F and G), although both egl-9 and sod mutants respond normally to the other stimuli prior or after H₂S exposure (Figure 5I, Figure 5—Figure supplement 1C, and Figure 7—Figure supplement 1H). Since disrupting egl-9 stabilizes HIF-1 and upregulates the expression of numerous genes involved in cellular defense against H₂S toxicity, the enhanced detoxification capacity in egl-9 mutants likely increases animals’ tolerance to H₂S, thereby reducing avoidance to otherwise toxic H₂S levels. Similarly, persistently high ROS in SOD deficient animals activates a variety of stress-responsive signaling pathways, including HIF-1, NRF2/SKN-1 and DAF-16/ FOXO signaling (Lennicke & Cocheme, 2021; Patten et al., 2010), facilitating cellular adaptation to redox stress and reducing animals’ responsiveness to toxic H₂S levels. Taken together, these findings support the view that reduced locomotory speed during H₂S exposure can arise from distinct mechanisms: early systemic toxicity in hif-1 and detoxificationdefective mutants, versus enhanced cellular adaptation in egl-9 and SOD mutants. We have integrated the relevant information across the result section and discussed this in Lines 485–536.

References

Adams, D. F., Farwell, S. O., Pack, M. R., & Bamesberger, W. L. (1979). Preliminary Measurements of Biogenic Sulfur-Containing Gas Emissions from Soils. Journal of the Air Pollution Control Association, 29(4), 380-383. doi:Doi 10.1080/00022470.1979.10470805

Beets, I., Zhang, G., Fenk, L. A., Chen, C., Nelson, G. M., Felix, M. A., & de Bono, M. (2020). NaturaL Variation in a Dendritic Scaffold Protein Remodels Experience-Dependent Plasticity by Altering Neuropeptide Expression. Neuron, 105(1), 106-121 e110. doi:10.1016/j.neuron.2019.10.001

Bretscher, A. J., Busch, K. E., & de Bono, M. (2008). A carbon dioxide avoidance behavior is integrated with responses to ambient oxygen and food in Caenorhabditis elegans. Proc Natl Acad Sci U S A, 105(23), 8044-8049. doi:10.1073/pnas.0707607105

Budde, M. W., & Roth, M. B. (2011). The response of Caenorhabditis elegans to hydrogen sulfide and hydrogen cyanide. Genetics, 189(2), 521-532. doi:10.1534/genetics.111.129841

Carrillo, M. A., Guillermin, M. L., Rengarajan, S., Okubo, R. P., & Hallem, E. A. (2013). O-2-Sensing Neurons Control CO2 Response in C. elegans. Journal of Neuroscience, 33(23), 9675-9683. doi:10.1523/Jneurosci.4541-12.2013

Doering, K. R. S., Cheng, X., Milburn, L., Ratnappan, R., Ghazi, A., Miller, D. L., & Taubert, S. (2022). Nuclear hormone receptor NHR-49 acts in parallel with HIF-1 to promote hypoxia adaptation in Caenorhabditis elegans. Elife, 11. doi:10.7554/eLife.67911

Gea, T., Barrena, R., Artola, A., & Sanchez, A. (2004). Monitoring the biological activity of the composting process: Oxygen uptake rate (OUR), respirometric index (RI), and respiratory quotient (RQ). Biotechnol Bioeng, 88(4), 520-527. doi:10.1002/bit.20281

Hallem, E. A., & Sternberg, P. W. (2008). Acute carbon dioxide avoidance in Caenorhabditis elegans. Proc Natl Acad Sci U S A, 105(23), 8038-8043. doi:10.1073/pnas.0707469105

Ishii, T., Miyazawa, M., Onouchi, H., Yasuda, K., Hartman, P. S., & Ishii, N. (2013). Model animals for the study of oxidative stress from complex II. Biochim Biophys Acta, 1827(5), 588-597. doi:10.1016/j.bbabio.2012.10.016

Jia, J., Wang, Z., Zhang, M., Huang, C., Song, Y., Xu, F., . . . Cheng, J. (2020). SQR mediates therapeutic effects of H(2)S by targeting mitochondrial electron transport to induce mitochondrial uncoupling. Sci Adv, 6(35), eaaz5752. doi:10.1126/sciadv.aaz5752

Kodama-Namba, E., Fenk, L. A., Bretscher, A. J., Gross, E., Busch, K. E., & de Bono, M. (2013). Crossmodulation of homeostatic responses to temperature, oxygen and carbon dioxide in C. elegans. PLoS Genet, 9(12), e1004011. doi:10.1371/journal.pgen.1004011

Lennicke, C., & Cocheme, H. M. (2021). Redox metabolism: ROS as specific molecular regulators of cell signaling and function. Mol Cell, 81(18), 3691-3707. doi:10.1016/j.molcel.2021.08.018

McGrath, P. T., Rockman, M. V., Zimmer, M., Jang, H., Macosko, E. Z., Kruglyak, L., & Bargmann, C. I. (2009). Quantitative mapping of a digenic behavioral trait implicates globin variation in C. elegans sensory behaviors. Neuron, 61(5), 692-699. doi:10.1016/j.neuron.2009.02.012

Morra, M. J., & Dick, W. A. (1991). Mechanisms of h(2)s production from cysteine and cystine by microorganisms isolated from soil by selective enrichment. Appl Environ Microbiol, 57(5), 14131417. doi:10.1128/aem.57.5.1413-1417.1991

Murakami, M., Koga, M., & Ohshima, Y. (2001). DAF-7/TGF-beta expression required for the normal larval development in C-elegans is controlled by a presumed guanylyl cyclase DAF-11. Mechanisms of Development, 109(1), 27-35. doi:Doi 10.1016/S0925-4773(01)00507-X

Naim, N., Amrit, F. R. G., Ratnappan, R., DelBuono, N., Loose, J. A., & Ghazi, A. (2021). Cell nonautonomous roles of NHR-49 in promoting longevity and innate immunity. Aging Cell, 20(7). doi:ARTN e13413 10.1111/acel.13413

Ochi, R., Dhagia, V., Lakhkar, A., Patel, D., Wolin, M. S., & Gupte, S. A. (2016). Rotenone-stimulated superoxide release from mitochondrial complex I acutely augments L-type Ca2+ current in A7r5 aortic smooth muscle cells. Am J Physiol Heart Circ Physiol, 310(9), H1118-1128. doi:10.1152/ajpheart.00889.2015

Onukwufor, J. O., Farooqi, M. A., Vodickova, A., Koren, S. A., Baldzizhar, A., Berry, B. J., . . . Wojtovich, A. P. (2022). A reversible mitochondrial complex I thiol switch mediates hypoxic avoidance behavior in C. elegans. Nat Commun, 13(1), 2403. doi:10.1038/s41467-022-30169-y

Oshins, C., Michel, F., Louis, P., Richard, T. L., & Rynk, R. (2022). Chapter 3 - The composting process. In R. Rynk (Ed.), The Composting Handbook (pp. 51-101): Academic Press.

Patange, O., Breen, P., Arsuffi, G., & Ruvkun, G. (2025). Hydrogen sulfide mediates the interaction between C. elegans and Actinobacteria from its natural microbial environment. Cell Reports, 44(1), 115170. doi:10.1016/j.celrep.2024.115170

Patten, D. A., Germain, M., Kelly, M. A., & Slack, R. S. (2010). Reactive oxygen species: stuck in the middle of neurodegeneration. J Alzheimers Dis, 20 Suppl 2, S357-367. doi:10.3233/JAD-2010100498

Ramsay, R. R., & Singer, T. P. (1992). Relation of superoxide generation and lipid peroxidation to the inhibition of NADH-Q oxidoreductase by rotenone, piericidin A, and MPP+. Biochem Biophys Res Commun, 189(1), 47-52. doi:10.1016/0006-291x(92)91523-s

Rodriguez-Kabana, R., Jordan, J. W., & Hollis, J. P. (1965). Nematodes: Biological Control in Rice Fields: Role of Hydrogen Sulfide. Science, 148(3669), 524-526. doi:10.1126/science.148.3669.524

Romanelli-Cedrez, L., Vairoletti, F., & Salinas, G. (2024). Rhodoquinone-dependent electron transport chain is essential for Caenorhabditis elegans survival in hydrogen sulfide environments. J Biol Chem, 300(9), 107708. doi:10.1016/j.jbc.2024.107708

Van Gilst, M. R., Hadjivassiliou, H., & Yamamoto, K. R. (2005). A Caenorhabditis elegans nutrient response system partially dependent on nuclear receptor NHR-49. Proc Natl Acad Sci U S A, 102(38), 13496-13501. doi:10.1073/pnas.0506234102

Wani, K. A., Goswamy, D., Taubert, S., Ratnappan, R., Ghazi, A., & Irazoqui, J. E. (2021). NHR- 49/PPAR-α and HLH-30/TFEB cooperate for host defense via a flavin-containing monooxygenase. Elife, 10. doi:ARTN e62775 10.7554/eLife.62775

Yang, W., & Hekimi, S. (2010). A mitochondrial superoxide signal triggers increased longevity in Caenorhabditis elegans. PLoS Biol, 8(12), e1000556. doi:10.1371/journal.pbio.1000556

Zorov, D. B., Juhaszova, M., & Sollott, S. J. (2014). Mitochondrial reactive oxygen species (ROS) and ROS-induced ROS release. Physiol Rev, 94(3), 909-950. doi:10.1152/physrev.00026.2013