Endogenous hydrogen peroxide positively regulates secretion of a gut-derived peptide in neuroendocrine potentiation of the oxidative stress response in C. elegans

Abstract The gut-brain axis mediates bidirectional signaling between the intestine and the nervous system and is critical for organism-wide homeostasis. Here we report the identification of a peptidergic endocrine circuit in which bidirectional signaling between neurons and the intestine potentiates the activation of the antioxidant response in C. elegans in the intestine. We identify a FMRF-amide-like peptide, FLP-2, whose release from the intestine is necessary and sufficient to activate the intestinal oxidative stress response by promoting the release of the antioxidant FLP-1 neuropeptide from neurons. FLP-2 secretion from the intestine is positively regulated by endogenous hydrogen peroxide (H 2 O 2 ) produced in the mitochondrial matrix by sod-3 /superoxide dismutase, and is negatively regulated by prdx-2 /peroxiredoxin, which depletes H 2 O 2 in both the mitochondria and cytosol. H 2 O 2 promotes FLP-2 secretion through the DAG and calcium-dependent protein kinase C family member pkc-2 and by the SNAP25 family member aex-4 in the intestine. Together, our data demonstrate a role for intestinal H 2 O 2 in promoting inter-tissue antioxidant signaling through regulated neuropeptide-like protein exocytosis in a gut-brain axis to activate the oxidative stress response.


Introduction
The gut-brain axis is critical for communication between the intestine and the nervous system to regulate behavior and maintain homeostasis, and altered gut-brain signaling is associated with neurodegeneration, obesity and tumor proliferation (Carabotti et al. 2015;Grenham et al. 2011;Mayer, Nance, and Chen 2022;Mehrian-Shai et al. 2019;Vitali et al. 2022).Over the last decade the importance of peptides that function as signals in gut-brain signaling has gained recognition.Numerous gut peptides are distributed throughout the gastrointestinal (GI) tract with regional specificity (Haber et al. 2017), and gut-secreted peptides can modulate neurocircuits in regulation of feeding behavior and glucose metabolism (Batterham and Bloom 2003;Han et al. 2018;Song et al. 2019), inflammatory responses against pathogenic bacteria (Campos-Salinas et al. 2014;Yu et al. 2021), and satiety (Batterham et al. 2002;Chelikani, Haver, and Reidelberger 2005;Gibbs, Young, and Smith 1973;Lutz et al. 1995;Lutz, Del Prete, and Scharrer 1994;West, Fey, and Woods 1984).A gut-released peptide suppresses arousal through dopaminergic neurons during sleep in Drosophila (Titos et al. 2023).In C. elegans, gut-derived peptides regulate rhythmic behavior and behavioral responses to pathogenic bacteria (Lee and Mylonakis 2017;Singh and Aballay 2019;Wang et al. 2013).Conversely, peptides released from the nervous system regulate many aspects of intestinal function including gut mobility, inflammation and immune defense (Browning and Travagli 2014;Furness et al. 2014;Lai, Mills, and Chiu 2017).In C. elegans, the secretion of peptides from various neurons regulates the mitochondrial unfolded protein response (UPR mt ), the heat shock response, and the antioxidant response in the intestine (Jia and Sieburth 2021;Maman et al. 2013;Prahlad, Cornelius, and Morimoto 2008;Shao, Niu, and Liu 2016).In spite of the many roles of peptides in the gut-brain axis, the mechanisms underlying the regulation of intestinal peptide secretion and signaling are not well defined.
Here we demonstrate a role of endogenous H2O2 signaling in the intestine in regulating the release of the intestinal FMRFamide-like peptide, FLP-2, to modulate a neurocircuit that activates the anti-oxidant response in the intestine in C. elegans.Intestinal FLP-2 signaling functions by potentiating the release of the antioxidant neuropeptide-like protein FLP-1 from AIY interneurons, which in turn activates the antioxidant response in the intestine.FLP-2 secretion from the intestine is rapidly and positively regulated by H2O2.Intestinal H2O2 levels and FLP-2 secretion are positively regulated by superoxide dismutases in the mitochondrial matrix and cytosol, and are negatively regulated by the peroxiredoxin-thioredoxin system in the cytosol.Intestinal FLP-2 release is mediated by aex-4/SNAP25-dependent exocytosis of dense core vesicles and H2O2-induced FLP-2 secretion is dependent upon the production of intestinal diacylglycerol and on pkc-2/PKCa/b kinase activity.

Results
Neuronal FLP-1 secretion is regulated by neuropeptide signaling from the intestine.
We previously showed that 10 minute treatment with the mitochondrial toxin juglone leads to a rapid, reversible and specific increase in FLP-1 secretion from AIY, as measured by a two-fold increase in coelomocyte fluorescence in animals expressing FLP-1::Venus fusion proteins in AIY (Fig. 1A and B, (Jia and Sieburth 2021)).Coelomocytes take up secreted neuropeptides by bulk endocytosis (Fares and Greenwald 2001) and the fluorescence intensity of Venus in their endocytic vacuoles is used as a measure of regulated neuropeptide secretion efficacy (Ailion et al. 2014;Ch'ng, Sieburth, and Kaplan 2008;Sieburth, Madison, and Kaplan 2006).To determine the role of the intestine in regulating FLP-1 secretion, we first examined aex-5 mutants.aex-5 encodes an intestinal subtilisin/kexin type 5 prohormone convertase (PCSK5), that functions to proteolytically process peptide precursors into mature peptide fragments in dense core vesicles (DCVs) (Edwards et al. 2019;Thacker and Rose 2000), and aex-5 mutants are defective in peptide signaling from the intestine (Mahoney et al. 2008).We found that aex-5 mutants expressing FLP-1::Venus in AIY exhibited no significant difference in coelomocyte fluoresce compared to wild type controls in the absence of juglone (Fig. 1B).However, coelomocyte fluorescence did not significantly increase in aex-5 mutants treated with juglone.Expression of aex-5 cDNA selectively in the intestine (under the ges-1 promoter) fully restored normal responses to juglone to aex-5 mutants, whereas aex-5 cDNA expression in the nervous system (under the rab-3 promoter) failed to rescue (Fig. 1B).Thus, peptide processing in intestinal DCVs is necessary for juglone-induced FLP-1 secretion from AIY.
flp-1 protects animals form the toxic effects of juglone (Jia and Sieburth 2021).We reasoned that the intestinal signal that regulates FLP-1 secretion should also protect animals from juglone-induced toxicity.We identified the FMRF-amide neuropeptide-like protein, flp-2, in an RNA interference (RNAi) screen for neuropeptides that confer hypersensitivity to juglone toxicity upon knockdown (Jia and Sieburth 2021).flp-2 signaling has been implicated in regulating lifespan, reproductive development, locomotion during lethargus, and the mitochondrial unfolded protein response (UPR mt ) (Chai et al. 2022;Chen et al. 2016;Kageyama et al. 2022;Shao et al. 2016).Putative flp-2(ok3351) null mutants, which eliminate most of the flp-2 coding region, are superficially as healthy as wild type animals, but they exhibited significantly reduced survival in the presence of juglone compared to wild type controls (Fig. 1C).The reduced survival rate of flp-2 mutants was similar to that of flp-1 mutants, and flp-1; flp-2 double mutants exhibited survival rates that were not more severe than those of single mutants (Fig. 1C), suggesting that flp-1 and flp-2 may function in a common genetic pathway.
To determine whether flp-2 signaling regulates FLP-1 secretion from AIY, we examined FLP-1::Venus secretion.flp-2 mutants exhibited normal levels of FLP-1 secretion in the absence of stress, but FLP-1 secretion failed to significantly increase following juglone treatment of flp-2 mutants (Fig. 1D).flp-2 is expressed in a subset of neurons as well as the intestine (Chai et al. 2022), and flp-2 functions from the nervous system for its roles in development and the UPR mt (Chai et al. 2022;Chen et al. 2016;Kageyama et al. 2022;Shao et al. 2016).Expressing flp-2 genomic DNA (gDNA) in the nervous system failed to rescue the FLP-1::Venus defects of flp-2 mutants, whereas expressing flp-2 selectively in the intestine fully restored juglone-induced FLP-1::Venus secretion to flp-2 mutants (Fig. 1D).Intestinal overexpression of flp-2 had no effect on FLP-1 secretion in the absence of stress, but significantly enhanced the ability of juglone to increase FLP-1 secretion (Fig. 1D).These results indicate that although flp-2 signaling has minimal impacts on FLP-1 secretion under normal conditions, flp-2 originating from the intestine is necessary and sufficient to positively regulate FLP-1 release from AIY in the presence of juglone to promote oxidative stress resistance.

FLP-2 secretion from the intestine is H2O2-regulated.
To directly investigate the mechanisms underlying the regulation of FLP-2 secretion, we examined FLP-2::Venus fusion proteins expressed in the intestine under various conditions (Fig. 2A).FLP-2::Venus fusion proteins adopted a punctate pattern of fluorescence throughout the cytoplasm of intestinal cells and at the plasma membrane (Fig. 2A), and FLP-2::Venus puncta co-localized with the DCV cargo protein AEX-5/PCSK5 tagged to mTurquoise2 (AEX-5::mTur2, Fig. 2B).FLP-2::Venus fluorescence was also observed in the coelomocytes (marked by mCherry) (Fig. 2A, C), indicating that FLP-2 is released from the intestine.SNAP25 forms a component of the core SNARE complex, which drives vesicular membrane fusion and transmitter release (Chen and Scheller 2001;Goda 1997;Jahn and Scheller 2006).aex-4 encodes the C. elegans homolog of SNAP25, and mutations in aex-4 disrupt the secretion of neuropeptides from the intestine (Lin-Moore et al. 2021;Mahoney et al. 2008;Wang et al. 2013).We found that aex-4 null mutations significantly reduced coelomocyte fluorescence in FLP-2::Venus expressing animals, and expression of aex-4 cDNA selectively in the intestine fully restored FLP-2 secretion to aex-4 mutants (Fig. 2C).Together these results suggest that intestinal FLP-2 can be packaged into DCVs that undergo release via SNARE-dependent exocytosis.
Next, to determine if H2O2 can regulate FLP-2 secretion, we examined the effects of acute exogenous H2O2 exposure on coelomocyte fluorescence of FLP-2::Venus-expressing animals.We found that 10 minute treatment with H2O2 increased FLP-2::Venus secretion to a similar extent as juglone treatment.aex-4/SNAP25, or aex-6/Rab27 mutants exhibited no increase in FLP-2 secretion in response to H2O2 treatment compared to untreated controls (Fig. 3G).In contrast, sod-1 or sod-3 mutants (or sod-1; sod-3 double mutants) exhibited an increase in FLP-2 secretion in response to H2O2 that was similar to that of wild type controls (Fig. 3H), suggesting that exogenous H2O2 can bypass the requirement of SODs but not SNAREs to promote FLP-2 secretion.Together these results suggest that H2O2 generated by SODs can positively regulate intestinal FLP-2 exocytosis form DCVs (Fig. 3I).

SOD-1 and SOD-3 regulate intestinal mitochondrial H2O2 levels.
To directly monitor H2O2 levels in the intestine, we generated transgenic animals expressing HyPer7, which is a pH stable genetically encoded H2O2 sensor in which the ratio of GFP/CFP fluorescence intensity increases in the presence of H2O2 (Pak et al. 2020).We targeted HyPer7 to either the mitochondrial matrix (matrix-HyPer7) by generating fusion proteins with the cytochrome c MLS, or to the cytosolic face of the outer mitochondrial membrane (OMM-HyPer7) by generating fusion proteins with TOMM-20.When co-expressed in the intestine with the OMM marker TOMM-20::mCherry, matrix-HyPer7 formed round structures throughout the cytoplasm that were surrounded by the OMM, and OMM-HyPer7 formed ring-like structures throughout the cytoplasm that colocalized with the OMM marker (Fig. 3J).Ten minutes treatment with H2O2 significantly increased the fluorescence intensity by about two-fold of both matrix-HyPer7 and OMM-HyPer7 without altering mitochondrial morphology or abundance (Fig. 3J, S3B and C), validating the utility of HyPer7 as a sensor for acute changes in H2O2 levels in and around intestinal mitochondria.
To determine whether juglone treatment impacts H2O2 levels in the intestine, we first treated matrix-HyPer7 expressing animals with juglone for 10 minutes.Juglone treatment led to a similar two-fold increase in matrix-HyPer7 fluorescence as H2O2 treatment (Fig. 3J).sod-3 mutations did not alter baseline H2O2 levels in the matrix, but they completely blocked juglone-induced increases H2O2 levels, whereas sod-1 mutations had no effect on either baseline or juglone-induced increases H2O2 levels (Fig. 3J).These results indicate that superoxide produced by juglone treatment is likely to be converted into H2O2 by SOD-3 in the matrix (Fig. 3I).
Next, we examined H2O2 levels on the outer surface of mitochondria using OMM-HyPer7 and we found that juglone treatment led to a two-fold increase in OMM-HyPer7 fluorescence, similar to H2O2 treatment (Fig. 3J).sod-3 or sod-1 mutations did not alter baseline H2O2 levels on the OMM, but sod-1 single mutations attenuated, juglone-induced increases in OMM H2O2 levels, while sod-3 mutations had no effect (Fig. 3J).In sod-1; sod-3 double mutants, the juglone-induced increase in OMM H2O2 levels was completely blocked, whereas baseline H2O2 levels in the absence of stress were unchanged (Fig. 3J).These results suggest that sod-3 and sod-1 are exclusively required for H2O2 production by juglone and that both mitochondrial SOD-3 and cytosolic SOD-1 contribute to H2O2 levels in the cytosol.One model that could explain these results is that juglonegenerated superoxide is converted into H2O2 both by SOD-3 in the matrix, and by SOD-1 in the cytosol, and that the H2O2 generated in the matrix can exit the mitochondria to contribute to cytosolic H2O2 levels needed to drive FLP-2 secretion (Fig. 3I).The peroxiredoxin-thioredoxin system regulates endogenous H2O2 levels and FLP-2 secretion.
To determine whether endogenous H2O2 regulates FLP-2 secretion, we examined mutations in the peroxiredoxin-thioredoxin system.Peroxiredoxins and thioredoxins detoxify excessive H2O2 by converting it into water and they play a critical role in maintaining cellular redox homeostasis (Netto and Antunes 2016) (Fig. 4A).C. elegans encodes two peroxiredoxin family members, prdx-2 and prdx-3, that are expressed at high levels in the intestine (Taylor et al. 2021).Null mutations in prdx-2 significantly increased FLP-2::Venus secretion compared to wild type animals in the absence of stress (Fig. 4B), whereas null mutations in prdx-3 had no effect on FLP-2 secretion (Fig. S4A).We observed a corresponding increase in both matrix-HyPer7 and OMM-HyPer7 fluorescence intensity in prdx-2 mutants (Fig. 4C and D), demonstrating that endogenous H2O2 is neutralized by peroxiredoxin and establishing a correlation between endogenous H2O2 levels and FLP-2 secretion.The increase in FLP-2 secretion in prdx-2 mutants was not further increased by juglone treatment (Fig. 4B).These results suggest that H2O2 generated under either normal conditions or by juglone-treatment can positively regulate FLP-2 secretion.
Once oxidized, peroxiredoxins are reduced by thioredoxins (TRXs) for reuse [refs].TRX-3 is an intestine-specific thioredoxin promoting protection against specific pathogen infections (Jiménez-Hidalgo et al. 2014; Miranda-Vizuete, Damdimopoulos, and Spyrou 2000; Netto and Antunes 2016).Mutations in trx-3 elevated FLP-2::Venus release in the absence of juglone and expressing trx-3 transgenes in the intestine restored wild type FLP-2 release to trx-3 mutants (Fig. 4B).Juglone treatment failed to further enhance FLP-2::Venus release in trx-3 mutants (Fig. 4B).Mutations in cytoplasmic sod-1 but not in mitochondrial sod-3 reduced the elevated FLP-2::Venus release in trx-3 mutants to wild type levels (Fig. 4B).Mutations in trx-3 increased H2O2 levels in the OMM but had no effect on matrix H2O2 levels (Fig. 4C and D).Thus TRX-3 likely functions in the cytosol but not in the matrix to neutralize H2O2, and elevated H2O2 levels in the cytosol are sufficient to drive FLP-2 secretion without SOD-3-mediated H2O2 generation in the matrix.
H2O2 functions as a cellular signaling molecule by oxidizing reactive cysteines to sulfenic acid, and this modification on target proteins can regulate intracellular signaling pathways (García Santamarina, Boronat i Llop, and Hidalgo Hernando 2014).One of the validated targets of H2O2 signaling is the protein kinase C (PKC) family of serine threonine kinases (Jia and Sieburth 2021;Konishi et al. 1997Konishi et al. , 2001;;Min, Kim, and Exton 1998).C. elegans encodes four PKC family members including pkc-1 and pkc-2, which are expressed at highest levels in the intestine (Islas-Trejo et al. 1997;Taylor et al. 2021).pkc-1 null mutants had no effect on baseline or juglone-induced FLP-2 secretion (Fig. S5A).pkc-2 null mutations did not alter baseline intestinal FLP-2 secretion, but they eliminated juglone-induced FLP-2 secretion (Fig. 5A).pkc-2 encodes a calcium and diacylglycerol (DAG) stimulated PKCα/β protein kinase C that regulates thermosensory behavior by promoting transmitter secretion (Edwards et al. 2012;Land and Rubin 2017).Expressing pkc-2 cDNA selectively in the intestine fully restored juglone-induced FLP-2 secretion to pkc-2 mutants (Fig. 5A), whereas expressing a catalytically inactive pkc-2(K375R) variant ( Van et al. 2021) failed to rescue (Fig. 5A).The intestinal site of action of pkc-2 is in line with prior studies showing that pkc-2 can function in the intestine to regulate thermosensory behavior (Land and Rubin 2017).pkc-2 mutants exhibited wild type H2O2 levels in the mitochondrial matrix and OMM of the intestine in both the presence and absence of juglone (Fig. 5B and C).Increasing H2O2 levels by either acute H2O2 treatment or by prdx-2 mutation failed to increase FLP-2 secretion in pkc-2 mutants (Fig. 5D and E).To determine whether pkc-2 can regulate the secretion of other peptides from the intestine, we examined expulsion frequency, which is a measure of NLP-40 secretion (Mahoney et al. 2008;Wang et al. 2013).pkc-2 mutants showed wild type expulsion frequency (Fig. S5B), indicating that intestinal NLP-40 release is largely unaffected.Together, these results show that pkc-2 is not a general regulator of intestinal peptide secretion, and that PKC-2 functions in the intestine downstream or in parallel to H2O2 to promote FLP-2 secretion by a mechanism that involves phosphorylation of target proteins.
DAG kinase converts DAG into phosphatidic acid (PA), and is therefore a negative regulator or DAG levels (Fig. 6A) (van Blitterswijk and Houssa 2000;Topham 2006).In C. elegans, dgk-2/DGKε is the highest expressing DAG kinase in the intestine.Mutations in dgk-2 elevated FLP-2::Venus secretion (Fig. 6E) without altering H2O2 levels in the intestinal mitochondrial matrix or OMM (Fig. 6F and G).Expressing dgk-2 transgenes selectively in the intestine restored normal FLP-2::Venus secretion to dgk-2 mutants (Fig. 6E).Finally, the increase in FLP-2 secretion in dgk-2 mutants was not further increased by juglone treatment, but it was completely blocked by pkc-2 mutations or aex-4/SNAP25 mutations (Fig. 6E and H).These results show that FLP-2 secretion can be regulated bidirectionally by DAG, and they suggest that DAG and H2O2 function in a common genetic pathway upstream of pkc-2 to promote FLP-2 secretion (Fig. 6A).

Discussion
By screening for intercellular regulators of FLP-1 signaling from the nervous system in promoting the anti-oxidant response, we have uncovered a function for peptidergic signaling in mediating gut-to-neuron regulation of the anti-oxidant response in C. elegans.We identified the neuropeptide-like protein FLP-2 as an inter-tissue signal originating in the intestine to potentiate stress-induced FLP-1 release from AIY neurons and the subsequent activation of SKN-1 in the intestine.We found that H2O2 generated endogenously in the intestine or exogenously by acute oxidant exposure increases FLP-2 secretion from intestinal DCVs.H2O2 promotes FLP-2 exocytosis through PKC-2, and AEX-4/SNAP25.The use of oxidant-regulated peptide secretion exemplifies a mechanism that can allow the gut and the nervous system to efficiently and rapidly communicate through endocrine signaling to promote organism-wide protection in the face of intestinal stress (Fig. 6I).A new function for flp-2 signaling in the antioxidant response.
Previous studies have identified roles for flp-2 signaling in development and in stress responses.flp-2 promotes locomotion during molting (Chen et al. 2016) promotes entry into reproductive growth (Chai et al. 2022), regulates longevity (Kageyama et al. 2022), and activates the UPR mt cell-non autonomously during mitochondrial stress (Shao et al. 2016).The function we identified for flp-2 in the antioxidant response has some notable similarities with flp-2's other functions.First, flp-2 mediates its effects at least in part by regulating signaling by other peptides.flp-2 signaling increases the secretion of the neuropeptide like protein PDF-1 during lethargus (Chen et al. 2016) and INS-35/insulin-like peptide for its roles in reproductive growth choice and longevity (Kageyama et al. 2022), in addition to regulating AIY FLP-1 secretion (Fig. 5G).In mammals, release of the RF-amide neuropeptide kisspeptin from the anteroventral periventricular nucleus (AVPV) regulates reproduction by inducing the release of gonadotropins via its stimulatory action on GnRH neurons (Han et al. 2005).Second, the secretion of FLP-2 is dynamic.FLP-2 secretion decreases during lethargus (Chen et al. 2016) and increases under conditions that do not favor reproductive growth (Kageyama et al. 2022), as well increasing in response to oxidants (Fig. 2C and D).However, in some instances, the regulation of FLP-2 secretion may occur at the level of flp-2 expression (Kageyama et al. 2022), rather than at the level of exocytosis (Fig. 2C).Finally, genetic analysis of flp-2 has revealed that under normal conditions, flp-2 signaling may be relatively low, since flp-2 mutants show no defects in reproductive growth choice when animals are well fed (Chai et al. 2022), show only mild defects in locomotion during molting in non-sensitized genetic backgrounds (Chen et al. 2016), and do not have altered baseline FLP-1 secretion or antioxidant gene expression in the absence of exogenous oxidants (Fig. 1D and E).It is notable that increased ROS levels are associated with molting (Back et al. 2012;Knoefler et al. 2012), ageing (Back et al. 2012;Van Raamsdonk and Hekimi 2010), starvation (Tao et al. 2017), and mitochondrial dysfunction (Dingley et al. 2010), raising the possibility that flp-2 may be used in specific contexts associated with high ROS levels to affect global changes in physiology, behavior and development.
One major difference we found for flp-2 signaling in our study is that intestinal, but not neuronal flp-2 activates the oxidative stress response, whereas flp-2 originates from neurons for its reported roles in development and the UPR mt .The intestine is uniquely poised to relay information about diet to the rest of the animal, and secretion of a number of neuropeptide-like proteins from the intestine (e.g, INS-11, PDF-2 and INS-7) is proposed to regulate responses to different bacterial food sources (Lee and Mylonakis 2017;Murphy, Lee, and Kenyon 2007;O'Donnell et al. 2018).Since bacterial diet can impact ROS levels in the intestine (Pang and Curran 2014), secretion of FLP-2 from the intestine could function to relay information about bacterial diet to distal tissues to regulate redox homeostasis.In addition, the regulation of intestinal FLP-2 release by oxidants may meet a unique spatial, temporal or concentration requirement for activating the antioxidant response that cannot be met by its release from the nervous system.AIY as a target for flp-2 signaling.
AIY interneurons receive sensory information from several neurons primarily as glutamatergic inputs to regulate behavior (Bargmann et al. 2007;Clark et al. 2006;Satoh et al. 2014).Our study reveals a previously undescribed mechanism by which AIY is activated through endocrine signaling originating from FLP-2 secretion from the intestine.FLP-2 could act directly on AIY, or it may function indirectly through upstream neurons that relay FLP-2 signals to AIY. frpr-18 encodes an orexin-like GPCR that can be activated by FLP-2-derived peptides in transfected mammalian cells (Larsen et al. 2013;Mertens et al. 2005), and frpr-18 functions downstream of flp-2 in the locomotion arousal circuit (Chen et al. 2016).frpr-18 is expressed broadly in the nervous system including in AIY (Chen et al. 2016), and loss-of-function frpr-18 mutations lead to hypersensitivity to certain oxidants (Ouaakki et al. 2023).FRPR-18 is coupled to the heterotrimeric G protein Gaq (Larsen et al. 2013;Mertens et al. 2005), raising the possibility that FLP-2 may promote FLP-1 secretion from AIY by directly activating FRPR-18 in AIY.However, flp-2 functions independently of frpr-18 in the reproductive growth circuit, and instead functions in a genetic pathway with the GPCR npr-30 (Chai et al. 2022).In addition, FLP-2-derived peptides can bind to the GPCRs DMSR-1, or DMSR-18 in transfected cells (Beets et al. 2023).Identifying the FLP-2 receptor and its site of action will help to define the circuit used by intestinal flp-2 to promote FLP-1 release from AIY.
FLP-1 release from AIY is positively regulated by H2O2 levels generated from mitochondria (Jia and Sieburth 2021).Here we showed that FLP-1 release is also regulated by intestinal FLP-2 signaling.Interestingly, H2O2 treatment is not sufficient to promote FLP-1 secretion in the absence of flp-2, and intestinal-derived FLP-2 is not sufficient to promote FLP-1 secretion in the absence of H2O2 (Fig. 1D).These results point to a model whereby AIY must integrate at least two stress signals: increased mt H2O2 levels and increased flp-2 signaling input, and only when both conditions are met will FLP-1 secretion increase.AIY shows a sporadic Ca 2+ response regardless of the presence of explicit stimulation (Ashida, Hotta, and Oka 2019;Bargmann et al. 2007;Clark et al. 2006), and FLP-1 secretion from AIY is calcium dependent (Jia and Sieburth 2021).How mitochondrial H2O2 levels are established in AIY by intrinsic or extracellular inputs, and how AIY integrates H2O2 and flp-2 signaling to control FLP-1 secretion remain to be defined.A role for endogenous H2O2 in regulated neuropeptide secretion.
Using HyPer7, we showed that acute juglone exposure results in a rapid elevation of endogenous H2O2 levels inside and outside intestinal mitochondria and a corresponding increase of FLP-2 release from the intestine that depends on the cytoplasmic superoxide dismutase sod-1, and mitochondrial sod-3.We favor a model whereby superoxide generated by juglone in the mitochondria is converted to H2O2 by SOD-3 in the matrix and by SOD-1 in the cytosol.In this case, both the superoxide generated by juglone and the H2O2 generated by SOD-3 would have to be able to exit the mitochondria and enter the cytosol.Superoxide and H2O2 can be transported across mitochondrial membranes through anion channels and aquaporin channels, respectively (Bienert and Chaumont 2014;Ferri et al. 2003;Han et al. 2003;Kontos et al. 1985).The observation that both SOD-1 and SOD-3 activity are necessary to drive FLP-2 release suggests that H2O2 levels much reach a certain threshold in the cytoplasm to promote FLP-2 release, and this threshold requires the generation of H2O2 by both SOD-1 and SOD-3.
We identified a role for the antioxidant peroxiredoxin-thioredoxin system, encoded by prdx-2 and trx-3, in maintaining low endogenous H2O2 levels in the intestine and in negatively regulating FLP-2 secretion.We showed that the prdx-2b isoform functions to inhibit FLP-2 secretion and to lower H2O2 levels in both the mitochondrial matrix and on the cytosolic side of mitochondria.These observations are consistent with a subcellular site of action for PRDX-2B in either the matrix only or in both the matrix and cytosol.In contrast, trx-3 mutations do not alter mitochondrial H2O2 levels, suggesting that TRX-3 functions exclusively in the cytosol.Thus, the PRDX-2B-TRX-3 combination may function in the cytosol, and PRDX-2B may function with a different TRX family member in the matrix.There are several thioredoxin-domain containing proteins in addition to trx-3 in the C. elegans genome that could be candidates for this role.Alternatively, prdx-2 may function alone or with other redox proteins.PRDX-2 may function without thioredoxins in its roles in light sensing and stress response in worms (Li et al. 2016;Oláhová et al. 2008;Oláhová and Veal 2015).PRDX-2B contains a unique N terminal domain that is distinct from the catalytic domain and is not found on the other PRDX-2 isoforms.This domain may be important for targeting PRDX-2B to specific subcellular location(s) where it can regulate FLP-2 secretion.Regulation of FLP-2 exocytosis by PKC-2/PKCa/b and AEX-4/SNAP25.
We demonstrated that pkc-2 mediates the effects of H2O2 on intestinal FLP-2 secretion, and H2O2-and DAG-mediated PKC-2 activation are likely to function in a common genetic pathway to promote FLP-2 secretion.Our observations that DAG is required for the effects of juglone (Fig. 6B), are consistent with a two-step activation model for PKC-2, in which H2O2 could first modify PKC-2 in the cytosol, facilitating subsequent PKC-2 recruitment to the membrane by DAG.Alternatively, DAG could first recruit PKC-2 to membranes, where it is then modified by H2O2.We favor a model whereby H2O2 modification occurs in the cytosol, since H2O2 produced locally by mitochondria would have access to cytosolic pools of PKC-2 prior to its membrane translocation.
We defined a role for aex-4/SNAP25 in the fusion step of FLP-2 containing DCVs from the intestine under normal conditions as well as during oxidative stress.In neuroendocrine cells, phosphorylation of SNAP25 on Ser187 potentiates DCV recruitment into releasable pools (Nagy et al. 2002;Shu et al. 2008;Yang et al. 2007), and exocytosis stimulated by the DAG analog phorbol ester (Gao et al. 2016;Shu et al. 2008), without altering baseline SNAP25 function.Interestingly, the residue corresponding to Ser187 is conserved in AEX-4, raising the possibility that PKC-2 potentiates FLP-2 secretion by phosphorylating AEX-4.Since SNAP25 phosphorylation on Ser187 has been shown to increase its interaction with syntaxin and promote SNARE complex assembly in vitro (Gao et al. 2016;Yang et al. 2007), it is possible that elevated H2O2 levels could promote FLP-2 secretion by positively regulating SNARE-mediated DCV fusion at intestinal release sites on the basolateral membrane through AEX-4/SNAP25 phosphorylation by PKC-2.Prior studies have shown that PKC-2 phosphorylates the SNARE-associated protein UNC-18 in neurons to regulate thermosensory behavior (Edwards et al. 2012;Land and Rubin 2017).Thus, PKC-2 may have multiple targets in vivo and target selection may be dictated by cell type and/or the redox status of the cell.

Similar molecular mechanisms regulating FLP-1 and FLP-2 release.
The molecular mechanisms we identified that regulate FLP-2 secretion from the intestine are similar in several respects to those regulating FLP-1 secretion from AIY.First, the secretion of both peptides is positively regulated by H2O2 originating from mitochondria.Second, in both cases, H2O2 promotes exocytosis of neuropeptidecontaining DCVs by a mechanism that depend upon the kinase activity of protein kinase C. Finally, the secretion of both peptides is controlled through the regulation of H2O2 levels by superoxide dismutases and by the peroxiredoxin-thioredoxin system.H2O2-regulated FLP-1 and FLP-2 secretion differ in the identity of the family members of some of the genes involved.prdx-3-trx-2 and sod-2 family members regulate H2O2 levels in AIY, whereas prdx-2-trx-3 and sod-1/sod-3 family members regulate H2O2 levels in the intestine.In addition, pkc-1 promotes H2O2 induced FLP-1 secretion from AIY whereas pkc-2 promotes H2O2 -induced FLP-2 secretion from the intestine.Nonetheless, it is noteworthy that two different cell types utilize largely similar pathways for the H2O2mediated regulation of neuropeptide release, raising the possibility that similar mechanisms may be utilized in other cell types and/or organisms to regulate DCV secretion.

Strains and transgenic lines
C. elegans strains were maintained at 20°C in the dark on standard nematode growth medium (NGM) plates seeded with OP50 Escherichia coli as food source, unless otherwise indicated.All strains were synchronized by picking mid L4 stage animal either immediately before treatment (for coelomocyte imaging and intestine imaging) or 24h before treatment (for Pgst-4::gfp imaging).The wild type strain was Bristol N2.Mutants used in this study were out-crossed at least four times.

Toxicity Assay
Stock solution of 50mM juglone in DMSO was freshly made on the same day of liquid toxicity assay, working solution of juglone in M9 buffer was prepared using stock solution before treatment.Around 60-80 synchronized adult animals were transferred into a 1.5mL Eppendorf tube with fresh M9 buffer and washed three times.Working solution of juglone was added to the animals at indicated concentration.Animals were incubated in dark for 4h on rotating mixer before being transferred onto fresh NGM plates seeded with OP50 to recover in dark at 20°C.Percentage of survival was assayed by counting the number of alive and dead animals.Toxicity assays were performed in triplicates.

RNAi Interference
Plates for feeding RNAi interference were prepare as described (Kamath 2003).Around 20-25 gravid adult animals with indicated genotype were transferred onto the RNAi plates that were seeded with HT115(DE3) bacteria transformed with L4440 vectors with targeted gene inserts or empty L4440 vectors.Eggs were collected for 4h to obtain synchronized populations.L4 stage animals were collected for further assays.RNAi clones were from Ahringer or Vidal RNAi library, or made from genomic DNA.Details were listed in the Supplementary Table.

Behavioral assays
The defecation motor program was assayed as previously described (Liu and Thomas 1994).Twenty to thirty L4 animals were transferred onto a fresh NGM plate seeded with OP50 E. coli and were stored in a 20°C incubator for 24 hours.After 24 hours, ten consecutive defecation cycles were observed from three independent animals and the mean and the standard error was calculated for each genotype.The pBoc and aBoc steps were recorded using custom Etho software (James Thomas Lab website:http://depts.washington.edu/jtlab/software/otherSoftware.html)

Microscopy and Fluorescence Imaging
Approximately 30-40 age matched animals were paralyzed with 30mg/mL 2,3butanedione monoxime (BDM) in M9 buffer and mounted on 2% agarose pads.Images were captured using the Nikon eclipse 90i microscope equipped with Nikon Plan Apo 20x, 40x, 60x, and 100x oil objective (N.A.=1.40), and a Photometrics Coolsnap ES2 camera or a Hamamatsu Orca Flash LT+CMOS camera.Metamorph 7.0 software (Universal Imaging/Molecular Devices) was used to capture serial image stacks and to obtain the maximum intensity projection image for analysis.
For transcriptional reporter imaging, young adult animals with indicated genotype were transferred into a 1.5mL Eppendorf tube with M9 buffer, washed three times and incubated in working solution of juglone with indicated concentration for 4h in dark on rotating mixer before recovering on fresh NGM plates with OP50 for 1h in dark at 20°C.The posterior end of the intestine was imaged with the 60x objective and quantification for average fluorescence intensity of a 16-pixel diameter circle in the posterior intestine was calculated using Metamorph.
For coelomocyte imaging, L4 stage animals were transferred in fresh M9 buffer on a cover slide, washed six times with M9 before being exposed to juglone or H2O2 in M9 buffer.Animals were then paralyzed in BDM and images of coelomocytes next to the posterior end of intestine were taken using the 100x oil objective.Average fluorescence intensity of Venus from the endocytic compartments in the posterior coelomocytes was measured in ImageJ.
For fusion protein fluorescence imaging, L4 stage animals were exposed to M9 buffer or indicated oxidants for 10min before being paralyzed in BDM and taken images of the posterior end of the intestine using 100x oil objective.For HyPer7 imaging, Z stacks were obtained using GFP (excitation/emission: 500nm/520nm) and CFP (excitation/emission: 400nm/520nm) filter sets sequentially, HyPer7 fluorescence signal was quantified as the ratio of GFP to CFP fluorescence intensity changes with respect to the baseline [(Ft − F0)/F0].CRISPR/Cas9 Editing prdx-2b(vj380) knock-out mutants were generated using a co-CRISPR protocol (Arribere et al. 2014).A sgRNA and a repair single stranded oligodeoxynucleotides (ssODN) targeting dpy-10 were co-injected with a sgRNA for genes of interest and a ssODN that induces homology-directed repaire (HDR) to introduce Cas9 mediated mutagenesis.Fifteen young adult animals were injected to produce around thirty singled F1 animals carrying Dpy or Rol phenotype.F2 animals were genotyped for mutations based on PCR and enzyme digest.Homozygous mutants were outcrossed with wild type animals at least four times before being used for assays.

Statistics
Statistical analysis was performed on GraphPad Prism 9. Unpaired t-test with two tails was used for two groups and one-way ANOVA with multiple comparison corrections was used to three or more groups to determine the statistical significance.Statistical details and n are specified in the figure legends.All comparisons are conducted based on wild-type controls unless indicated by lines between genotypes.Bar graphs with plots were generated using GraphPad Prism 9.

Acknowledgement
C. elegans strains used in this study were provided by the Caenorhabditis Genetics Centre (CGC), which is funded by the NIH National Center for Research Resources (NCRR).We thank members of the Sieburth lab for critical reading and discussion of the manuscript.This work was supported by grants from National Institute of Health NINDS R01NS071085 and R01NS110730 to D.S.