Neuropeptides encoded by nlp-49 modulate locomotion, arousal and egg-laying behaviours in Caenorhabditis elegans via the receptor SEB-3

Neuropeptide signalling has been implicated in a wide variety of biological processes in diverse organisms, from invertebrates to humans. The Caenorhabditis elegans genome has at least 154 neuropeptide precursor genes, encoding over 300 bioactive peptides. These neuromodulators are thought to largely signal beyond ‘wired’ chemical/electrical synapse connections, therefore creating a ‘wireless’ network for neuronal communication. Here, we investigated how behavioural states are affected by neuropeptide signalling through the G protein-coupled receptor SEB-3, which belongs to a bilaterian family of orphan secretin receptors. Using reverse pharmacology, we identified the neuropeptide NLP-49 as a ligand of this evolutionarily conserved neuropeptide receptor. Our findings demonstrate novel roles for NLP-49 and SEB-3 in locomotion, arousal and egg-laying. Specifically, high-content analysis of locomotor behaviour indicates that seb-3 and nlp-49 deletion mutants cause remarkably similar abnormalities in movement dynamics, which are reversed by overexpression of wild-type transgenes. Overexpression of NLP-49 in AVK interneurons leads to heightened locomotor arousal, an effect that is dependent on seb-3. Finally, seb-3 and nlp-49 mutants also show constitutive egg-laying in liquid medium and alter the temporal pattern of egg-laying in similar ways. Together, these results provide in vivo evidence that NLP-49 peptides act through SEB-3 to modulate behaviour, and highlight the importance of neuropeptide signalling in the control of behavioural states. This article is part of a discussion meeting issue ‘Connectome to behaviour: modelling C. elegans at cellular resolution’.

IB, 0000-0002-0968-4360; WRS, 0000-0002-6676-8034 Neuropeptide signalling has been implicated in a wide variety of biological processes in diverse organisms, from invertebrates to humans. The Caenorhabditis elegans genome has at least 154 neuropeptide precursor genes, encoding over 300 bioactive peptides. These neuromodulators are thought to largely signal beyond 'wired' chemical/electrical synapse connections, therefore creating a 'wireless' network for neuronal communication. Here, we investigated how behavioural states are affected by neuropeptide signalling through the G protein-coupled receptor SEB-3, which belongs to a bilaterian family of orphan secretin receptors. Using reverse pharmacology, we identified the neuropeptide NLP-49 as a ligand of this evolutionarily conserved neuropeptide receptor. Our findings demonstrate novel roles for NLP-49 and SEB-3 in locomotion, arousal and egg-laying. Specifically, high-content analysis of locomotor behaviour indicates that seb-3 and nlp-49 deletion mutants cause remarkably similar abnormalities in movement dynamics, which are reversed by overexpression of wild-type transgenes. Overexpression of NLP-49 in AVK interneurons leads to heightened locomotor arousal, an effect that is dependent on seb-3. Finally, seb-3 and nlp-49 mutants also show constitutive egg-laying in liquid medium and alter the temporal pattern of egg-laying in similar ways. Together, these results provide in vivo evidence that NLP-49 peptides act through SEB-3 to modulate behaviour, and highlight the importance of neuropeptide signalling in the control of behavioural states.
This article is part of a discussion meeting issue 'Connectome to behaviour: modelling C. elegans at cellular resolution'.

Background
Neuropeptides regulate a myriad of behaviours in organisms ranging from invertebrates to humans [1][2][3][4][5]. Understanding how neuropeptide signalling between neurons drives different motor outputs not only provides important information on how behavioural states come about, but also provides insight into how neuromodulators interact with the physical (wired) connectome to drive such changes. There is increasing evidence that neuropeptides act on neuronal circuits to promote the functional importance of particular connections over others [6,7]. Caenorhabditis elegans provides an excellent system to study the role of neuropeptides in regulating behaviour owing to its ease of genetic manipulation and availability of a complete wiring diagram for its compact & 2018 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.
nervous system [8], providing a unique opportunity to investigate neuropeptide signalling in the context of the wired connectome in a whole organism.
The C. elegans genome encodes 154 predicted neuropeptide precursor genes, which can be processed to over 300 different peptides [9,10], of which at least 200 have been confirmed by mass spectrometry [11]. Neuropeptides are important signalling molecules formed from larger, inactive precursors that are processed enzymatically to produce bioactive peptides [1,12,13]. They have been implicated in a wide variety of biological processes including locomotion, sensory responsiveness, egg-laying and quiescence (reviewed in [1,14]). In C. elegans and other organisms, neuropeptides primarily exert their effects via G protein-coupled receptors (GPCRs) [5]. Based on phylogenetic analysis, GPCRs can be classified into Glutamate, Rhodopsin, Adhesion, Frizzled/ Taste2 and Secretin families [15]. The Secretin family (or family B) is one of the main receptor families for neuropeptides and neurohormones, several of which have been implicated in important biological functions [16]; for example, corticotrophin-releasing factor (CRF), which modulates anxiety and stress in humans [17].
Here, we investigated how behavioural states are affected by signalling through the C. elegans neuropeptide receptor SEB-3, a member of an orphan secretin receptor family that had already evolved in a common ancestor of bilaterian animals [5]. Using reverse pharmacology, we identified NLP-49 as a neuropeptide ligand of this ancient GPCR family. We found that NLP-49 peptides released from only a few head neurons can affect multiple motor outputs through a receptor that is broadly expressed throughout the nervous system. Specifically, our data represent the first description, to our knowledge, of a neuropeptide ligand for the SEB-3 family of secretin receptors in bilaterian animals and the biological functions of this newly identified ligand -receptor system, demonstrating a role for NLP-49 peptides and SEB-3 in arousal and egg-laying behaviours.

Experimental procedures (a) Strain and transgene information
Strains were maintained on nematode growth medium (NGM) plates seeded with E. coli strain OP50 according to standard experimental procedures [18]. Day  (b) Behavioural tests (i) Egg-laying assays Video recordings of animals freely moving on food was performed using WormTracker 2.0 software [19,20] and analysed using a custom MATLAB (MathWorks) script as described in [21,22] . At least 54 h of video obtained for multiple days  was analysed per genotype (table 1), where an individual worm was recorded for 6 h. Animals were tracked on multiple days. Egg-laying behaviour off-food was performed oby moving well-fed day 1 adult worms onto low-peptone NGM plates seeded (overnight) with 20 ml OP50 or unseeded, and eggs laid counted after 60 min. For this experiment, 12 animals were tested per condition (on food or off food) for each genotype, in at least three replicates. Egg-laying in liquid was carried out in CTX buffer (5 mM KH 2 PO 4 / K 2 HPO 4 (pH 6), 1 mM CaCl 2 and 1 mM MgSO 4 ) with six technical replicates per genotype (per biological replicate) and 10 animals per technical replicate, with eggs counted after 60 min. At least three biological replicates were conducted. Animals were transferred to 1 ml CTX buffer in each well of a 24-well plate using a mounted hair pick. Egg retention experiments were performed on animals aged 16 h post larval stage four (L4) and eggs counted using bright-field microscopy. At least three biological replicates were conducted with 15-20 animals tested per replicate. Egg retention experiments were conducted blind to the genotype of the strains tested.

(ii) Tap arousal experiments
Sixteen hours before the experiment, late L4 animals were picked onto plates seeded with 90 ml of OP50 overnight culture and left to dry at room temperature for 14-16 h. Three or four plates of 50-100 animals per plate were tested per trial, with each genotype tested in at least three trials. Multi-worm tracking was performed using a Dinolite camera (AM4113ZT) positioned above the plates, which recorded videos at 5 frames per second. An additional white LED backlight was used to improve contrast. Video recording was started 2 min after the plate was placed on the stage. Animals were recorded for 20 s before taps were applied and then for a further 5 min. Taps were applied manually to the bottom of the assay plate. Tracks were analysed using MATLAB (MathWorks) with the Multi-Worm Tracker [23]. Area under the curve (AUC) measurements were performed using GraphPad Prism, with baseline values used for each genotype being the speed measurements recorded before the stimulus was applied. AUC was determined for 10 -24 s post-stimulus to exclude the speed of the initial reversal [24].

(iii) Unsupervised behavioural tracking experiments
Recordings and analysis were performed as described in [19,20] using WormTracker 2.0 software. Briefly, tracking was carried out on 3 cm low peptone plates (standard NGM, with modifications: agar increased to 2%, peptone decreased to 0.013%). Low peptone plates were dried for 24 h at room temperature before use. On the day of tracking, plates were seeded with 20 ml of OP50 overnight culture and allowed to dry. Worms were moved to their tracking plate using a mounted eyelash hair and allowed to acclimatize for 30 min. Mutant/transgenic animals were tracked together with N2 controls. At least 17 animals were analysed per rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 373: 20170368 genotype. Recording for each strain was randomized across multiple trackers and across multiple days. Clustering analysis was performed by first analysing mutant videos with Tierpsy Tracker [25] to extract behavioural features for each worm which are averaged to create a single feature vector for each strain. Average feature vectors were then z-normalized across all the strains shown in figure 2a,c. Heat maps were produced by hierarchical clustering using complete linkage and Euclidean distance. To compare the phenotypic difference between seb-3(tm1848) and nlp-49(gk546875) with the difference observed in a large set of mutants, we used the same features as in figure 2a,c, but the z-normalization was performed across all of the strains in [19] and the two new mutants before calculating pairwise Euclidean distances. To identify features that showed significant differences between N2 and seb-3(tm1848) or nlp-49(gk546875) mutants, we performed multiple t-tests, controlling for false discovery rate (5%) using the Benjamini-Krieger -Yekutieli method. To compare the same feature between all genotypes tested, we performed a one-way ANOVA with Holm -Sidak's post-test for multiple comparisons.
Statistical analysis for all experiments was performed using GraphPad Prism v. 6.0. Exact statistical tests used are provided in the figure legends.

(c) Receptor activation assays
Assays were performed as described [26]. Briefly, seb-3/ C18B12.2 cDNA was cloned into the pcDNA3.1(þ) TOPO expression vector (Thermo Fisher Scientific). A CHO-K1 cell line (PerkinElmer, ES-000-A24) stably expressing apo-aequorin (mtAEQ) targeted to the mitochondria as well as the human Ga16 was transiently transfected with the seb-3 cDNA construct or the empty pcDNA3.1(þ) vector (not shown) using Lipofectamine (Thermo Fisher Scientific). Cells were shifted to 288C 1 day later, and collected 2 days post-transfection in BSA medium (DMEM/Ham's F12 with 15 mM HEPES, without phenol red, 0.1% BSA) loaded with 5 mM coelenterazine h (Thermo Fisher Scientific) for 4 h. A total of 25 000 cells per well were exposed to synthetic peptides in BSA medium, and aequorin bioluminescence was recorded on a MicroBeta LumiJet luminometer (PerkinElmer, Waltham, MA) for 30 s in quadruplicate. After 30 s of ligand-stimulated calcium measurements, Triton X-100 (0.1%) was added to obtain a measure of the maximum cell Ca 2þ response. BSA medium without peptides was used as a negative control and 1 mM ATP was used to check the functional response of the cells. Half-maximal effective concentration (EC 50 ) values were calculated from dose-response curves, constructed using a nonlinear regression analysis, with the sigmoidal dose-response equation (Prism v. 6.0).

Results
The C. elegans GPCR SEB-3 has been implicated in the regulation of different behavioural states, such as roaming behaviour, the control of sensory responsiveness during lethargus, and male mating behaviour [27,28]. However, the identity of ligands for this receptor remains to be discovered. To further characterize the neuromodulatory functions of SEB-3, we set out to identify its ligand(s) using a reverse pharmacology approach. We expressed SEB-3 in Chinese hamster ovary (CHO) cells together with the calcium indicator aequorin and the promiscuous Ga 16 protein. We then challenged these cells with a synthetic library of over 300 isolated and predicted C. elegans peptides and monitored receptor activation via the calcium-based aequorin read-out. A single peptide encoded by the NLP-49 precursor (NLP-49-2, SPSMGLSLAEYMASPQGGDNFHFMPSamide) activated SEB-3 in this screen (figure 1a). We next tested a concentration series to determine the ligand's potency. NLP-49-2 dose-dependently activated SEB-3 with a halfmaximal effective concentration (EC 50 ) of 62 nM (figure 1b), within the expected nanomolar range of neuropeptidereceptor interactions, but did not evoke a calcium response in cells transfected with an empty vector (not shown). The precursor gene nlp-49 (H05L03.3), previously shown to be involved in regulating fat storage [29], encodes two peptides, NLP-49-1 and NLP-49-2 (figure 1a), both of which have been identified by mass spectrometry in C. elegans [11]. To date, no receptor is known to be activated by NLP-49-1. The putative receptor for NLP-49-2, SEB-3, is a member of a largely uncharacterized family of secretin receptors of bilaterian origin, which is evolutionarily conserved from protostomes to cephalochordates, but appears to have been lost in vertebrates [5,30]. So far, SEB-3 is the only member of this family for which a ligand has been identified. Previous reports suggested that SEB-3 is similar to mammalian CRF receptors [16,27,28]. Recent phylogenetic evidence indicates that SEB-3 is not a direct orthologue of these receptors, but is most closely related to pigment-dispersing factor (PDF) receptors [5,30].
To further characterize the role of NLP-49 in vivo, we first determined the expression pattern of this neuropeptide gene using a reporter line expressing the full-length genomic DNA of nlp-49 and a 2 kb promoter region (Pnlp-49::nlp-49 gDNA::SL2-mKate2). Expression of this reporter was restricted to a few head neurons, which we identified as the SMB and AVK neurons. We also observed fainter expression in neurons that are anatomically consistent with the SAA neurons (figure 1c). In addition, consistent with a previous report [28], we found that a transgenic reporter construct for seb-3, using a 2 kb upstream region together with full intronic sequences, shows broad expression in the nervous system (figure 1d; see electronic supplementary material, table S2 for a full list of seb-3-expressing cells identified in this study).
An nlp-49 allele obtained from the Million Mutation Project [31] results in an early stop codon at position 4, leading to the loss of both NLP-49 peptides (figure 1a). To investigate whether NLP-49 peptides interact with SEB-3 in vivo, we first investigated the locomotion phenotypes of mutants lacking SEB-3 or NLP-49 peptides. We used an automated tracking system to record videos of freely moving wild-type, mutant and transgenic rescue animals and measure a large number of features related to body shape and movement dynamics [19,20]. A large-scale survey of uncoordinated mutants using the WormTracker previously indicated that these measurements provide a useful behavioural fingerprint for classifying phenotypic similarity and revealing potential functional rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 373: 20170368 relationships; for example, loss-of-function alleles of the same gene (e.g. unc-2, unc-4), or mutants affecting components of known complexes (e.g. unc-7/unc-9, unc-38/unc-63), generally showed very similar phenotypic profiles [19].
We first tracked outcrossed strains carrying loss-of-function mutations in seb-3 and nlp-49. We observed that both seb-3(tm1848) and nlp-49(gk546875) mutants showed remarkably similar locomotion phenotypes. When compared with wild-type, the seb-3 and nlp-49 strains showed similar changes for most locomotion features; by contrast, statistical analysis did not reveal any significant differences between seb-3(tm1848) and nlp-49(gk546875) mutants (multiple t-tests, Benjamini-Krieger-Yekutieli method, false discovery rate (Q) ¼ 5%) (see electronic supplementary material, table S3 for raw values). Figure 2a shows a heat map of normalized differences calculated between wild-type and each mutant. To assess whether these phenotypes were due to seb-3 and nlp-49, we also tracked seb-3(tm1848) and nlp-49(gk546875) animals rescued by transgenically re-expressing each gene under its own promoter, and compared these strains' behaviour with the wild-type. As transgenic lines were generated by microinjection, they are maintained as extrachromosomal arrays and are high-copy [32]. Therefore, we expected these transgenes to cause overexpression of SEB-3/NLP-49. Overexpression of neuropeptide precursor genes or neuropeptide receptors has previously been shown to cause gain-of-function phenotypes opposite to those seen in the respective null mutants (see, for example, [33][34][35][36]). Our data indicated that both neuropeptide and receptor rescue strains showed similar locomotion changes compared with wild-type, and that these changes were generally in the opposite direction to those seen for nlp-49 and seb-3 mutants (figure 2a). We then tested, using hierarchical clustering analysis, if locomotion profiles of nlp-49 and seb-3 mutants would cluster with each other more closely than with other strains. Indeed, we found that when computing pairwise distances between nlp-49(gk546875) and seb-3(tm1848) and all mutants in our database, nlp-49 and seb-3 mutants clustered more closely together than with previously tracked mutants [19] (figure 2b). These findings are consistent with NLP-49 and SEB-3 acting in the same molecular pathway to modulate locomotion. To identify features that are significantly different between wild-type and seb-3 or nlp-49 mutants, we performed multiple t-tests for all features between wild-type and each mutant (controlling for Q ¼ 5%). Significant effects were observed in features measuring body curvature, i.e. mid-body bend, hips bend, eigen projection 2 and eigen projection 3 (figure 2d; electronic supplementary material, figure S1B).      Figure 2d(i,ii) shows examples of specific features comparing differences in body curvature among the strains tested. Both seb-3(tm1848) and nlp-49(gk546875) show significantly smaller body bend angles compared with wildtype (see electronic supplementary material, figure S1A for representative traces; also electronic supplementary material, figure S1B), and transgenic rescue for the expression of these genes leads to an increased bend angle compared with the respective mutant strains. In addition, nlp-49(o/e) animals show generally increased bend angles compared with wild-type, and loss of SEB-3 reduces these effects. We next tested if SEB-3 and NLP-49 regulate another aspect of locomotion: touch-evoked arousal. When animals are exposed to a non-localized mechanosensory stimulus, they undergo a transient increase in locomotor activity that persists for 1-2 min [37]. To determine whether this behaviour is affected by nlp-49 and seb-3, five taps were rapidly delivered to the Petri dish housing the animals and the speed of these animals was recorded using automated tracking software [23]. To quantify this locomotor hyperactivity (arousal) following mechanosensory stimulation, we used the average area underneath the curve (AUC) for each trace for 10-24 s post-stimulus. We did not include the first 10 s post-tap as the acute effects of tap (less than 10 s) are largely due to the initial escape response comprising backward locomotion [24]. Deletion mutants of seb-3 and nlp-49, as well as seb-3;nlp-49 double mutants, showed essentially wild-type behaviour ( figure 3a,b); however, overexpression of nlp-49 led to a robust increase in arousal following the tap stimulus compared with wild-type (figure 3c,d). Even taking into account the basal increase in locomotion speed observed for nlp-49(o/e) animals, we observed a significantly greater increase in motor activity when we compared the speed observed after tap with the baseline speed recorded before tap. The increased locomotor hyperactivity seen in nlp-49(o/e) therefore seems to be due to the increase in speed post-stimulus being longer lasting than in wild-type controls.
We then tested if expression of nlp-49 from particular head neurons was specifically required for this hyper-aroused phenotype. Surprisingly, expression of nlp-49 from the AVK interneurons, but not from the SMB motor neurons, led to increased locomotor arousal (figure 3c,d), suggesting that specific inputs to AVK are required for arousal. We next asked if loss of SEB-3 would affect the hyper-arousal resulting from elevated NLP-49 peptides. Indeed, we found that loss of SEB-3 strongly attenuated the arousal effect of nlp-49 overexpression (figure 3e,f). Interestingly, although loss of SEB-3 resulted in reduced locomotor arousal after tap, it did not affect the higher baseline speed of nlp-49(o/e) transgenic animals (figure 3e). The basal increase in locomotion speed observed in nlp-49 overexpressors could therefore result from SEB-3-independent effects or increased levels of NLP-49-1 peptides. We also found that the effects of expressing nlp-49 from AVK were dependent on SEB-3, as AVK::nlp-49 transgenic animals containing a seb-3(tm1848) mutation showed significantly reduced locomotor arousal compared with AVK::nlp-49 animals containing intact SEB-3 ( figure 3g,h). These data indicate that NLP-49 and SEB-3 act together to affect locomotion in response to aversive stimuli in addition to the changes observed in unstimulated conditions. Changes in locomotion and egg-laying events appear to be coordinated [40,41] and some neuropeptides that regulate locomotion also affect egg-laying [36,42,43]. Detailed study revealed that egg-laying events tend to be clustered, with events within clusters (active phases) usually separated by approximately 20 s (intra-cluster interval). These clusters take place between long inactive periods of 20 -30 min (inter-cluster interval) ( figure 4a(i,ii)). The time constants for these intervals, and the probability of egg-laying events occurring during each cluster ( probability, p), can be inferred by obtaining maximum-likelihood estimates of the inter-and intra-cluster time constants from egg-laying interval data [21,44,45]. We found that nlp-49 and seb-3 deletion mutants show egg-laying constitutive behaviours, as indicated by a significantly higher probability of egg-laying during the active phase and longer active phase duration (figure 4a(iii,iv); table 1). Inactive state durations are not significantly affected ( figure 4a and table 1). In addition, nlp-49(gk546875) animals showed on average a shorter duration between egg-laying events during the active phase (figure 4a(iii) and table 1). Representative raster plots for these mutants are shown in electronic supplementary material, figure S2A. Consistent with a higher rate of egg-laying during the Table 1. Egg-laying parameters. Probability ( p) refers to the probability of the animal laying an egg during the active state. The intra-cluster time (s) refers to the time between egg-laying events during the active state. The inter-cluster time (min) refers to the time between active states, i.e. duration of the inactive state. The duration of the active phase (s) is inferred from p and the intra-cluster time constant. Values are provided as mean + s.e.m. Video (h) indicates the number of hours of video analysed to extract parameters for analysis. Statistical differences were determined using an unpaired t-test. n.s., not significant, ** , 0.01, *** , 0.001, **** , 0.0001. reproductive phase, single deletion mutants for both genes, as well as the double mutant, also showed significantly fewer eggs retained in the uterus as day 1 adults compared with wild-type (figure 4b). Double nlp-49;seb-3 mutants did not show an additive effect compared with single mutants, consistent with these genes acting in the same genetic pathway to modulate egg-laying (figure 4b). Conversely, overexpression of nlp-49 showed more eggs retained in the uterus (electronic supplementary material, figure S2B). Thus, nlp-49 and seb-3 loss-of-function mutants both showed detectable egg-laying constitutive phenotypes. We next investigated how NLP-49 and SEB-3 affect the modulation of egg-laying by external stimuli, such as in the presence or absence of food, or culture on liquid or solid medium. We found that neither nlp-49(gk546875) nor seb-3(tm1848) animals laid eggs in the absence of food, unlike egg-laying hyperactive mutants such as goa-1(n1134) (figure 4c). However, mutants of both genes showed increased egg-laying in liquid medium compared with wild-type controls ( figure 4d,e). Additionally, nlp-49;seb-3 double mutants did not show significantly more egg-laying than either single mutant (electronic supplementary material, figure S2C). We could rescue the mutant phenotype by transgenic re-expression of nlp-49 (figure 4d) and seb-3 (figure 4e). However, the case for seb-3 is complicated by the fact that a control line expressing mKate2 using the seb-3 promoter in a seb-3(tm1848) background, expected to behave similarly to animals carrying seb-3(tm1848) alone, showed no statistically significant difference in behaviour compared with the seb-3(tm1848);Pseb-3::seb-3 line ( p-value 0.09; figure 4e). This suggests that transgene expression using the seb-3 promoter affects egg-laying, making it difficult to conclude for certain if the rescuing effect of Pseb-3::seb-3 is due to re-expression of seb-3 or due to promoter/transgene effects. We next tested if an interaction could be observed between NLP-49 and SEB-3 by examining egg-laying behaviour in an nlp-49 over-expressing line. Although there was a trend towards a lower egg-laying rate in nlp-49(o/e) animals (Pnlp-49::nlp-49 in a wild-type background) compared with a control line expressing mKate2 using the nlp-49 promoter, this difference did not reach statistical significance. However, we also found that seb-3(tm1848);Pnlp-49::nlp-49 animals showed a similar phenotype to animals carrying the seb-3(tm1848) allele alone, which is consistent with NLP-49 and SEB-3 acting in the same pathway (figure 4f ). As FLP-1 peptides released from AVK also modulate egg-laying behaviour [22], we next tested if AVK expression of nlp-49 (using the flp-1 promoter) could rescue the mutant phenotype. Indeed, we found that transgenic expression of nlp-49 from the AVK neurons in an nlp-49(2) background showed significant rescue compared with controls ( figure 4g). Lastly, we tested if SEB-3 acts directly in egg-laying motor neurons to modulate egg-laying behaviour. SEB-3 is expressed in the HSN and VC4/5 motor neurons, consistent with this possibility (figure 4j). However, when we expressed SEB-3 in HSN using an egl-6 promoter [46], or using a cat-1 promoter, which drives expression in VC4/5 and HSN in addition to other dopaminergic and serotonergic neurons [47], in a seb-3(tm1848) mutant background, we did not observe rescue of the seb-3 mutant phenotype  (figure 4h,i), suggesting that SEB-3 expression in the egglaying circuit is not sufficient for its effect on egg-laying. Thus, SEB-3 may function in neurons apart from VC4/5 or the HSNs, or these neurons in combination with others, to modulate egg-laying behaviour.

Discussion
Here, we showed for the first time, to our knowledge, a biochemical interaction between the SEB-3 receptor and NLP-49 peptides. We also demonstrated that NLP-49/SEB-3 signalling regulates multiple behaviours in C. elegans. Unlike seb-3, which is broadly expressed, nlp-49 appears to be expressed only in a small number of head neurons: AVK, SMB and putatively the SAA neurons, which interestingly are all connected to one another via chemical synapses or gap junctions [8]. It is therefore possible that activation of one of these neurons can synchronize peptide release from other nlp-49-expressing cells. Our findings suggest that neuromodulator release from only a few neurons can affect diverse behaviours by potentially acting through a receptor that is widely present in the nervous system. This form of signalling provides another layer of  communication in addition to existing chemical synapse/gap junction connections, allowing signals to be transmitted over greater anatomical distances and longer timescales [48]. Multiple neuromodulator signals can also be coordinated to consolidate important behaviours or to provide a means by which the animal can fine-tune its responses to sensory cues [6,7]. In the case of arousal (and egg-laying), expression of nlp-49 in AVK appears to be the most important to drive this behaviour. AVK has not previously been implicated in the control of locomotion; indeed a different neuropeptide that controls arousal-like behaviour, PDF-1, appears to act primarily in a different set of neurons, AIY and ASK [49,50]. Interestingly, we have found that the receptor (FRPR-3) for another neuropeptide (FLP-20) required for arousal in response to mechanosensory signals is also expressed in AVK [37], suggesting that multiple neuropeptide systems may function together to promote arousal. We found that NLP-49/SEB-3 signalling also modulates egg-laying behaviour. Egg-laying is regulated by external conditions, such that stressful environments including thermal stress, hypertonic salt solutions and starvation tend to suppress this behaviour [22,[51][52][53][54]. nlp-49 and seb-3 deletion mutants both show an increased rate of egg-laying not only under permissive conditions, but also in a stressful environment (in liquid). Surprisingly, although SEB-3 is expressed in egg-laying motor neurons, receptor expression in these neurons was not sufficient to rescue the egg-laying phenotype, suggesting SEB-3 may function outside the egg-laying circuit to control egg-laying behaviour. Additionally, overexpression of nlp-49 promotes locomotor hyperactivity via SEB-3 in response to an aversive tap stimulus. Together, our data suggest that NLP-49/SEB-3 signalling may be stressresponsive, such that decreased signalling is indicative of a low-stress state, and increased signalling correlates with a high-stress state. The CRF peptide system in mammals is thought to mediate stress and anxiety responses (see, for example, [55,56]). Although the CRF system appears to have been lost in nematodes [5,30,57], SEB-3 belongs to the same secretin superfamily of receptors as mammalian CRF receptors, as well as invertebrate PDF receptors. PDF peptides were previously shown to also regulate arousal responses in C. elegans [50,58]. NLP-49/SEB-3 may therefore be another member of the secretin family of GPCRs, which modulates behaviour in stress conditions.
Overall, the behavioural phenotypes of nlp-49 and seb-3 loss-of-function mutants are remarkably similar. In particular, nlp-49 and seb-3 mutants differ from wild-type in many of the same locomotor features, and the nlp-49 and seb-3 rescue lines likewise show similar behavioural fingerprints with respect to locomotion. Indeed, the locomotor phenotypes of nlp-49 and seb-3 mutants cluster closer than expected relative to a large dataset of strains. Thus, our behavioural analysis supports the in vitro data indicating that NLP-49-2 peptides are ligands for the SEB-3 receptor, and demonstrates the utility of high-content behavioural phenotyping to validate ligand -receptor pairs identified through de-orphanization studies.
The C. elegans genome encodes a diverse array of neuropeptides [11]. Expression of FLP, NLP and INS peptides is widespread in the nervous system [1,9,39,59], and these peptides have been implicated in a wide variety of behaviours [1,60]. Many C. elegans neuropeptides and their receptors also have orthologues in other animals, including humans [30,57]. SEB-3 is, we believe, the first identified de-orphaned receptor of a bilaterian family of secretin PDF-like neuropeptide receptors that evolved before the split of proto-and deuterostomian animals [5]. Members of this ancient neuropeptide receptor family are evolutionarily conserved in molluscs, annelids and deuterostomian invertebrates [5]. It will be interesting to see if neuropeptides similar to C. elegans NLP-49 activate SEB-3 orthologues in these species. Furthermore, preliminary investigations on the network properties of neuropeptide signalling in the (relatively few) examples where peptidereceptor pairs have been identified in C. elegans suggest that this form of communication takes place largely beyond physical connections, i.e. in a more 'wireless' manner [48]. Large-scale de-orphanization screens (for example [61]), combined with in vivo phenotyping of receptor and peptide mutants, provide a potential avenue to build a more complete connectome of neuropeptide signalling. This multilayer connectome will prove useful as a model to understand larger nervous systems, such as the human brain. Figure 4. (Overleaf.) nlp-49 and seb-3 regulate egg-laying behaviour. seb-3(tm1848) is referred to as seb-3(2) and nlp-49(gk546875) as nlp-49 (2). (a)(i) Representative temporal pattern for egg-laying showing active phases during which eggs are laid, interspersed by longer inactive phases. The time between egg-laying events during the active phase is the inter-cluster time, and the time between active phases is the intra-cluster time; (ii) histogram of interval times between egg-laying events for wild-type animals, with the interval duration in log scale. The location of the two peaks indicated by black dashed lines corresponds to the intra-cluster and inter-cluster time constants, estimated using maximum-likelihood analysis [44]; (iii) interval times between egg-laying events for nlp-49(2) animals; (iv) interval times between egg-laying events for seb-3(2) animals. Intra-and inter-cluster time constants for mutants are indicated in green (nlp-49(2)) or blue (seb-3(2)), with wild-type shown in black for comparison. See also table 1. (b) Eggs retained in the uterus for wild-type, nlp-49(2), seb-3(2) and seb-3(2);nlp-49(2) animals, at 16 h post-L4. 3 replicates, n . 15 per replicate. The box-and-whisker plot shows the box extending from the 25th to the 75th percentile, the line in the middle indicating the median. Whiskers extend from minimum to maximum. (c) Eggs laid on agar plates in the presence or absence of food (indicated by þ or 2 beneath the graph) for wild-type, nlp-49(2) and seb-3(2) mutant animals, with goa-1 mutants as a positive control. n ¼ 3 replicates, n ¼ 12 per condition per replicate. Eggs laid in liquid (average for 10 worms) for (d) wild-type, nlp-49 (2), and nlp-49 mutants expressing a rescue transgene for nlp-49 expression under its own promoter or a control mKate2 transgene; (e) wild-type, seb-3(2), and seb-3 mutants expressing a rescue transgene for seb-3 expression under its own promoter or a control mKate2 transgene; ( f ) wild-type, seb-3 mutants, animals overexpressing nlp-49 under its own promoter or a control mKate2 transgene, and animals overexpressing nlp-49 in a seb-3(2) background; (g) wild-type, nlp-49 (2), and nlp-49 mutants containing a transgene for nlp-49 expression in AVK (Pflp-1) or a control mKate2 transgene; (h) wild-type, seb-3(2), and seb-3 mutants expressing seb-3 in HSN (Pegl-6) or a control mKate2 transgene; (i) wild-type, seb-3(2), and seb-3 mutants expressing seb-3 using a cat-1 promoter or a control gfp transgene.