C. elegans males integrate food signals and biological sex to modulate state-dependent chemosensation and behavioral prioritization

SUMMARY

Dynamic integration of internal and external cues is essential for flexible, adaptive behavior. In C. elegans, biological sex and feeding state regulate expression of the food-associated chemoreceptor odr-10, contributing to plasticity in food detection and the decision between feeding and exploration. In adult hermaphrodites, odr-10 expression is high, but in well-fed adult males, odr-10 expression is low, promoting exploratory mate-searching behavior. Food-deprivation transiently activates male odr-10 expression, heightening food sensitivity and reducing food-leaving. Here, we identify a neuroendocrine feedback loop that sex-specifically regulates odr-10 in response to food deprivation. In well-fed males, insulin-like (IIS) and TGFβ signaling repress odr-10 expression. Upon food deprivation, odr-10 is directly activated by DAF-16/FoxO, the canonical C. elegans IIS effector. The TGFβ ligand DAF-7 likely acts upstream of IIS and links feeding to odr-10 only in males, due in part to the male-specific expression of daf-7 in ASJ. Surprisingly, these responses to food deprivation are not triggered by internal metabolic cues, but rather by the loss of sensory signals associated with food. When males are starved in the presence of inedible food, they become nutritionally stressed, but odr-10 expression remains low and exploratory behavior is increased compared to starved control males. Food signals are detected by a small number of sensory neurons whose activity non-autonomously regulates daf-7 expression, IIS, and odr-10. Thus, adult C. elegans males employ a neuroendocrine feedback loop that integrates food detection and genetic sex to dynamically modulate chemoreceptor expression and influence the feeding-vs.-exploration decision.

Graphical Abstract

INTRODUCTION

Innate drives—like those to feed, reproduce, and avoid predators—provide a flexible, modular framework for the control of behavior. Internal and external signals continuously interact with such drives, allowing behaviors to be activated, suppressed, or modified according to an animal’s history, physiology, and environment. In humans, disruption of these processes is associated with numerous neuropsychiatric conditions, including eating disorders, depression, and obsessive-compulsive disorder. Though these mechanisms also operate in simpler nervous systems, much remains unknown about how innate drives are regulated by internal and external conditions, how these conditions are represented internally, and how drive states influence behavioral choice.

One mechanism by which internal state influences behavior is through modulation of sensory perception. For example, nutritional state modulates sensory function throughout phylogeny, focusing a hungry animal’s attention on food-associated stimuli and attenuating the perception of cues that inhibit feeding [1, 2]. Biological sex and reproductive status can also influence sensory function: biological sex modulates sensory responses of C. elegans to pheromones and food [3], and female reproductive state influences the detection and processing of pheromonal and olfactory signals in both Drosophila and mice [4, 5]. However, how biological sex and reproductive state intersect with feeding status and other dimensions of internal state to modulate circuit function is largely unexplored.

State-dependent food-leaving behavior in C. elegans provides an ideal context for approaching these problems [6]. Typically, individual worms are efficiently retained by a food spot. However, well-fed adult males tend to leave a food source long before its depletion, abandoning it in search of a mate [7]. Multiple mechanisms promote this exploratory behavior, including input from male-specific sensory neurons that may tonically signal the absence of mates [8], and signaling through the PDFR-1 receptor [9]. Adult males also blunt their attraction to some food-related cues, promoting exploration over feeding [10]. This occurs in part through repression of odr-10, the chemoreceptor for the food-associated odorant diacetyl, in the AWA neurons [10–12]. This repression requires the male “state” of AWA as well as a temporal cue signaling the transition to adulthood, controlled by the heterochronic genes lep-2 (a Makorin) and lep-5 (a long non-coding RNA) [10, 13].

Food-leaving behavior in adult males is also subject to homeostatic regulation, allowing males to balance mating and feeding drives. In food-deprived males, food-leaving behavior is transiently suppressed, but is restored after several hours of re-feeding [7]. Similarly, food-deprivation transiently activates odr-10 expression in males and increases odr-10-dependent food attraction, indicating that sensory modulation is a component of this behavioral plasticity [10]. Here, we investigate the mechanisms that activate odr-10 expression upon food deprivation in adult C. elegans males, identifying a series of neuroendocrine signals that couple feeding to low odr-10 expression. Surprisingly, the signal that initiates this mechanism in males is not an internal metabolic cue, but rather the chemosensory detection of food itself. Together, these findings outline a multistep feedback mechanism that couples sensory perception to chemoreceptor expression and contributes to the homeostatic balance of behavioral drives.

RESULTS

Insulin signaling regulates odr-10 in response to feeding state in males

In adult C. elegans males, expression of odr-10 in the AWA neurons is activated by 12-18 h of food deprivation and returns to baseline after re-feeding [10] (Figures 1A, B). To identify mechanisms that underlie this, we examined odr-10 expression in males carrying mutations in pathways previously associated with feeding state. We first considered insulin/IGF-1 signaling (IIS), since daf-2, the receptor for the C. elegans IIS pathway, promotes food-leaving behavior in well-fed males [7]. Using the translational ODR-10::GFP fosmid reporter fsEx295 [10], we found that well-fed daf-2 mutant males exhibited high odr-10 expression, reminiscent of food-deprived males (Figure 1C). This upregulation was eliminated in daf-16; daf-2 double mutants, indicating a requirement for daf-16, the FoxO transcription factor that is a canonical effector of IIS [14] (Figure 1C). Loss of daf-2 also increased males’ attraction to the odr-10 ligand diacetyl (Figure 1D).

Figure 1. Insulin-like signaling regulates odr-10 expression males in response to feeding state.

(A) Representative images of ODR-10::GFP fluorescence in a well fed male, a food-deprived (“starved”) male, and a male re-fed after food deprivation. (B) ODR-10::GFP expression in WT fed and starved (stv) males and hermaphrodites, scored qualitatively on a four-point scale (see Methods). (C) ODR-10::GFP expression in male insulin-signaling mutants. (D) Chemotaxis to diacetyl (1:1000) in WT and daf-2 males. (E) ODR-10::GFP expression in fed and starved WT and daf-16 males. (F) Chemotaxis to diacetyl (1:1000) in WT and daf-16 males either well-fed or starved for 12 hours. (G) ODR-10::GFP expression in hermaphrodite IIS mutants. (H) ODR-10::GFP expression in fed and starved WT and daf-16 hermaphrodites. (I) Model for the regulation of odr-10 by food availability and IIS. In this and subsequent models, dotted lines indicate probable cell non-autonomous interactions. *0.01 < p < 0.05; **0.001 < p < 0.01; ***p < 0.001. Dotted gray brackets indicate p > 0.05. See also Figure S1.

To ask whether food availability represses odr-10 through IIS, we examined odr-10 expression in daf-16 males. Unlike wild-type, daf-16 males exhibited no detectable change in ODR-10::GFP expression upon food deprivation (Figure 1E). Similarly, starvation elicited a large increase in diacetyl attraction in wild-type males but had only a small effect in daf-16 mutants (Figure 1F). Thus, food deprivation upregulates odr-10 through decreased IIS and increased daf-16 activity. To ask whether daf-16 is simply a permissive activator necessary for all odr-10 expression, we examined larval males, which exhibit high, hermaphrodite-like odr-10 expression [10, 13]. We found that larval odr-10 expression was largely unaffected by daf-16 loss (Figure S1A), indicating a specific requirement for daf-16 in the activation of odr-10 by food deprivation (Figure 1I).

The response of odr-10 to food deprivation differs by sex

Our previous work indicated that the induction of odr-10 by food deprivation occurred only in males [10]. However, by qRT-PCR, we observed a significant increase in odr-10 mRNA upon food deprivation in hermaphrodites (Figure S1B). Furthermore, odr-10 mRNA levels were recently found to increase in daf-2 hermaphrodites [15]. Therefore, we considered the possibility that the ODR-10::GFP reporter used previously, kyIs53, might be prone to ceiling effects, or might lack important cis-acting regions. Indeed, when we examined hermaphrodites carrying the ODR-10::GFP fosmid reporter fsEx295, which contains additional regulatory sequence and has lower baseline expression than kyIs53, we found that food deprivation did elicit an increase in GFP intensity (Figure 1B). Thus, odr-10 expression is sensitive to food availability in both sexes. While the significance of increased odr-10 expression in hermaphrodites is unclear, it might make food detection more efficient or help override signals preventing the consumption of lower-quality food.

Despite this qualitatively similar response to food deprivation, we observed sex differences in the underlying mechanisms. Unlike males, daf-2 hermaphrodites exhibited no apparent change in ODR-10::GFP (Figure 1G). Consistent with this, daf-16 was not necessary for starvation-induced activation of odr-10 in hermaphrodites (Figure 1H). The apparent discrepancy between our findings and the qRT-PCR results of Hahm et al. [15] might indicate an mportant role for post-transcriptional regulation of odr-10 in hermaphrodites; further studies are needed to evaluate this and other possibilities. Nevertheless, these findings demonstrate that IIS is not the primary mediator of odr-10 activation by food deprivation in hermaphrodites and that the sexes differ in the mechanisms that link feeding state to odr-10.

odr-10 is likely a direct target of daf-16

To understand how IIS regulates odr-10, we asked whether daf-16 acts cell-autonomously. Consistent with this possibility, we detected DAF-16::GFP in the nuclei of AWA in daf-2 males (Figure 2A). Further, expression of daf-16f cDNA from the AWA-specific promoter Pgpa-4Δ6 [10] was sufficient to restore odr-10 expression in starved daf-16 males (Figure 2B). In contrast, intestine-specific expression of daf-16 had no effect on odr-10 (Figure S2A, B).

Figure 2. DAF-16 acts cell-autonomously, and likely directly, to regulate odr-10.

(A) DAF-16::GFP and AWA::mCherry expression. Upper panels show individual channels; lower panel shows a merged, pseudocolored image with a high-magnification inset. To bring about nuclear localization of DAF-16, this strain carried a daf-2 mutation and was grown at 25°C. (B) ODR-10::GFP expression in WT and daf-16 fed and starved (stv) males, with or without (–) AWA::daf-16f transgenes. Results are shown for two independent lines. (C) Diagram depicting the odr-10 genomic locus (upper) and odr-10 transcriptional reporters (middle and lower) with wild-type and mutant versions of the putative DAF-16 binding site. (D) Representative images of Podr-10^WT::GFP and Podr-10^ΔDAF-16::GFP in fed and starved males. Yellow ovals indicate the location of AWA. (E) Quantification of GFP fluorescence in Podr-10^WT::GFP and Podr-10^ΔDAF-16::GFP (two independent lines) in fed and starved males. *0.01 < p < 0.05; **0.001 < p < 0.01; ***p < 0.001. Dotted gray brackets indicate p > 0.05. See also Figure S2.

These results raised the possibility that odr-10 might be a direct transcriptional target of DAF-16. Supporting this, a putative DAF-16 binding element exists 809 bp upstream of the odr-10 start codon [15]. Therefore, we generated two transcriptional reporters, one containing roughly 1 Kb of the wild-type odr-10 promoter (Podr-10^WT) and one in which this element was disrupted by changing three conserved nucleotides (Podr-10^ΔDAF-16) (Figure 2C). We observed significant induction of Podr-10^WT::GFP expression upon food deprivation, but no change in Podr-10^ΔDAF-16::GFP (Figure 2D, E). Well-fed hermaphrodites continued to express the mutant reporter, demonstrating that this element is not generally required for activity (Figure S2C). We did not observe any obvious change in DAF-16::GFP localization in AWA in starved males (not shown), but the widespread expression of this fusion protein may have obscured partial nuclear translocation. Together, these results indicate that starvation-mediated odr-10 activation in males occurs at least in part at the transcriptional level and that DAF-16 likely activates odr-10 directly.

TGFβ signaling likely acts upstream of IIS to regulate odr-10 in males

We also considered a role for daf-7, a BMP/TGFβ-family ligand produced by the ASI neurons. daf-7 is involved in many food-related aspects of C. elegans development and physiology, including dauer entry, fertility, fat metabolism, satiety quiescence, immune response, and lifespan [16], and, in larval hermaphrodites, daf-7 expression in the ASI neurons decreases in response to nutritional stress [17–19]. daf-7 also regulates the expression of several chemoreceptor genes in adult hermaphrodites, though odr-10 was not among these [20]. Further, a recent study found that daf-7 is male-specifically expressed in the ASJ neurons and promotes food-leaving behavior [21].

We found that well-fed daf-7 mutant males displayed significant upregulation of ODR-10::GFP (Figure 3A). This regulation appears to be male-specific, as we detected no change in ODR-10::GFP in daf-7 mutant hermaphrodites (Figure 3A). To further explore this, we genetically masculinized the nervous system of hermaphrodites by expressing the male-specific sex determination regulator fem-3 [22, 23]. As expected [10], this was sufficient to reduce expression of ODR-10::GFP to male-typical levels (Figure 3B). Further, odr-10 expression in these hermaphrodites became sensitive to daf-7 loss (Figures 3A, B). To ask whether the male-specific expression of daf-7 in ASJ explains the male-specificity of the daf-7 phenotype, we examined odr-10 expression in males in which ASJ was genetically feminized by expression of an active form of the hermaphrodite-specific sex determination regulator tra-2 [24]. This reduces (but does not eliminate) ASJ expression of daf-7 [21]. We found that this significantly increased odr-10 expression, though not to the level of daf-7 mutants (Figure 3C). We also genetically ablated ASJ; this significantly increased odr-10 expression in males but had no effect in hermaphrodites (Figures 3D, E). Together, these results indicate that daf-7 production by ASJ partly underlies the sex-specificity of its phenotype, but that other sex differences, possibly in ASI, also have important roles. This is consistent with the work of Hilbert et al. [21], who showed that food-leaving behavior can be regulated by production of daf-7 in either ASJ or ASI. It may be the absolute amount of daf-7, rather than its specific source, that is important for regulation of odr-10.

Figure 3. odr-10 is sex-specifically regulated by daf-7 TGFβ.

(A) ODR-10::GFP expression in WT and daf-7 males and hermaphrodites. (B) ODR-10::GFP expression in WT and daf-7 hermaphrodites, with and without pan-neural masculinization by Prab-3::fem-3(+). (C) ODR-10::GFP expression in WT and ASJ-feminized (ASJ^f) males. Two independent lines are shown. (D) ODR-10::GFP expression in WT and ASJ-ablated (ASJ⊗) males. Two independent lines are shown. (E) ODR-10::GFP expression in WT and ASJ-ablated hermaphrodites. Two independent lines are shown. (F) ODR-10::GFP expression in WT, daf-7, daf-16, and daf-16; daf-7 mutant males. (G) ODR-10::GFP expression in WT and daf-8 mutants carrying the indicated daf-8 expression constructs. Two independent lines are shown for AWA::daf-8. (H) Model for the regulation of odr-10 by daf-7 and daf-2 (IIS) signaling. In this and following models, the dashed gray line indicates the possibility that daf-7 signaling could act in parallel to IIS. *0.01 < p < 0.05; **0.001 < p < 0.01; ***p < 0.001. Dotted gray brackets indicate p > 0.05. See also Figure S3.

Next, we examined the relationship between IIS and daf-7 signaling. Crosstalk between these pathways has been found previously: several insulin genes are transcriptionally regulated by daf-7 signaling and daf-7 mutants show increased nuclear localization of DAF-16 [25–28]. Consistent with this, daf-16 was required for the strong induction of odr-10 in daf-7 mutant males, such that ODR-10::GFP intensity was only slightly higher in daf-16; daf-7 double mutants than in daf-16 single mutants (Figure 3F). This suggests that daf-7 primarily regulates odr-10 expression by modulating IIS, with secondary effects through other mechanisms (Figure 3H). We attempted to examine DAF-16::GFP localization in daf-7 males, but, for unknown reasons, virtually all males of this strain irreversibly entered dauer, even when cultured at 16°C. Thus, while daf-7 appears to function primarily upstream of IIS, we cannot rule out parallel action of these pathways.

Next, we sought to determine the cellular site of TGFβ pathway action. First, we found that males carrying a mutation in daf-8, a SMAD downstream of daf-7 [16], phenocopied daf-7 mutants (Figure 3G). Further, this phenotype was rescued by expression of daf-8 cDNA under the control of its own promoter or a pan-neural promoter. However, we observed no rescue by expressing daf-8 in AWA itself or in RIM and RIC, neurons known to be modulated by daf-7 [17] (Figure 3G). Thus, daf-8 likely acts in the nervous system, but its activity in AWA, RIM, or RIC is not sufficient for the feeding-state regulation of odr-10 in males.

Food perception, rather than internal metabolic state, regulates odr-10 in males

During food deprivation, animals experience metabolic stress as well as the absence of food itself. To explore the relative roles of these, we uncoupled physiological state from the detection of food cues. First, we fed worms E. coli OP50, the standard laboratory diet, pre-treated with the cytokinesis inhibitor aztreonam. This generates long chains of bacteria that cannot be ingested by C. elegans [29]. Confirming this, we observed no accumulation of GFP inside nematodes cultured on aztreonam-treated GFP-labeled E. coli, even though GFP was clearly visible inside animals exposed to control bacteria (Figure S3). Moreover, animals cultured on this inedible food exhibited distended intestines, similar to animals deprived of food (Figure S3). Despite this metabolic stress, however, these animals displayed only a very modest increase in ODR-10::GFP (Figure 4A). We then re-exposed food-deprived animals to either control or aztreonam-treated bacteria for 24 h. In both cases, ODR-10::GFP expression returned to baseline levels, even though worms cultured on inedible bacteria presumably remained in a starved state (Figure 4B). Thus, while metabolic stress may play some role in odr-10 activation, a change in chemosensory environment appears to be its primary driver.

Figure 4. Sensory perception, not a metabolic cue, regulates odr-10 and daf-7 in males.

(A) ODR-10::GFP expression in males cultured on E. coli OP50, without food, or on aztreonam-treated OP50 for 16-18 h. (B) ODR-10::GFP expression in well-fed males, food-deprived males, and food-deprived males after 24 h recovery on control or aztreonam-treated OP50. (C) ODR-10::GFP expression in hermaphrodites cultured on OP50, without food, or on aztreonam-treated OP50 for 16-18 h. (D-E) ODR-10::GFP expression in males (D) and hermaphrodites (E) cultured on OP50, without food, or on heat-killed OP50 for 16-18 h. (F) ODR-10::GFP expression in WT and eat-2 males and hermaphrodites. (G-H) Pdaf-7::GFP fluorescence intensity in ASJ (G) and ASI (H) on males cultured on OP50, without food, or on aztreonam-treated OP50. (I) Pathway for the regulation of odr-10 by TGFβ and IIS in response to external food signals in males. *0.01 < p < 0.05; **0.001 < p < 0.01; ***p < 0.001. Dotted gray brackets indicate p > 0.05. See also Figure S4.

Unlike males, adult hermaphrodites exposed to aztreonam-treated bacteria displayed full induction of ODR-10::GFP, comparable to food-deprived hermaphrodites (Figure 4C). Thus, the role of food signals in odr-10 regulation differs by sex, with hermaphrodites relying primarily on internal signals of nutritional stress. This suggested that reducing food quality or intake might increase odr-10 expression in hermaphrodites but not in males. To test this, we grew animals on a low-quality food source, heat-killed E. coli, for 16-18 h. This did not alter male odr-10 expression, but hermaphrodite expression was significantly increased (Figures 4D, E). To reduce food intake, we used eat-2 mutants, which have slow pharyngeal pumping [30]. Again, this had no detectable effect on male odr-10 expression, but expression in hermaphrodites was significantly elevated (Figure 4F). These findings reinforce the idea that odr-10 is regulated primarily by food-derived sensory signals in males and by internal physiological signals in hermaphrodites.

tax-2/tax-4-dependent sensory signals repress odr-10 expression via daf-7 and IIS

Previous work showed that the expression of daf-7 in ASJ in adult males depends on the well-fed state, but that expression in ASI does not [21]. To ask whether food signals regulate odr-10 through daf-7 signaling, we examined Pdaf-7::GFP expression in animals cultured on aztreonam-treated E. coli. As expected, food deprivation decreased Pdaf-7::GFP in ASJ, but had little effect in ASI (Figure 4G, H). In contrast, culturing animals on inedible food did not reduce daf-7 expression in either neuron (Figure 4G); rather, for unknown reasons, ASJ daf-7 expression increased in males (Figure 4G). This suggests that, in males, production of DAF-7 by ASJ represents the detection of food-associated cues rather than the level of physiological satiety. Further, reduced daf-7 expression appears to couple decreased food perception to odr-10 activation in adult C. elegans males.

These findings imply that disruption of sensory function should alter odr-10 expression. Therefore, we examined animals carrying mutations in tax-2 and tax-4, subunits of a cyclic nucleotide-gated cation channel required for the function of a subset of C. elegans chemosensory neurons [31, 32]. Supporting this hypothesis, ODR-10::GFP was significantly upregulated in well-fed tax-2 and tax-4 mutant males (Figure 5A). We detected modestly increased odr-10 expression in tax-2 and tax-4 hermaphrodites, though this effect did not reach statistical significance for tax-4 (Figure 5B). tax-4 is not expressed in AWA in in either sex, nor is it required for AWA-mediated chemosensory behavior [31–33]; the same is true for tax-2, with the exception that, to our knowledge, its expression has not been examined in males. Regardless, this strongly suggests that sensory perception regulates odr-10 non-cell-autonomously. Further, while chemosensory signals are not the primary modulator of hermaphrodite odr-10 expression, they might have a role in its regulation.

Figure 5. Signals from TAX-2/TAX-4 neurons regulate odr-10 expression in males.

(A-B) ODR-10::GFP expression in tax-2, tax-4, daf-16, and daf-16; tax-4 males (A) and hermaphrodites (B). (C) Representative images of Pdaf-7::GFP expression in WT and tax-4 males. Yellow ovals indicate the locations of ASI and ASJ. (D) ODR-10::GFP (kyIs53) expression in WT and tax-4 males with the indicated tax-4 expression constructs. Two separate lines are shown for ASI+ASJ rescue. (E-F) Pdaf-7::GFP fluorescence in ASJ (E) and ASI (F) in tax-4 males with the indicated tax-4 expression constructs. Two separate lines are shown for ASI+ASJ rescue. (G) Food-leaving behavior of well-fed males (“OP50”), starved males (“no food”), and males starved in the presence of inedible food (“OP50+Az”). At the indicated time point, each plate was examined to determine the farthest distance between the food spot and the worm’s track, using the four categories shown in the cartoon to the right. “F” indicates the location of the food spot. (H) Pathway for the regulation of odr-10 by food signals, TGFβ signaling, and IIS in males. *0.01 < p < 0.05; **0.001 < p < 0.01; ***p < 0.001. Dotted gray brackets indicate p > 0.05. See also Figure S5.

If sensory signals from tax-4-expressing neurons repress odr-10 expression via daf-7 and IIS, the tax-4 phenotype should depend on daf-16. Indeed, we found that odr-10 expression in daf-16; tax-4 double mutants was significantly lower than in tax-4 single mutants, in both males and hermaphrodites (Figure 5A, B). Interestingly, previous work has shown that the expression of daf-7 is reduced in tax-4 mutant hermaphrodites [34], suggesting that the increase in odr-10 expression we observed in tax-4 males could result from reduction in daf-7 expression in ASJ and/or ASI. Consistent with this, we found that Pdaf-7::GFP expression in tax-4 males was strongly reduced in ASI and was virtually undetectable in ASJ (Figure 5C, E, F). Together, this suggests that removal from food decreases the activity of a subset of tax-4 neurons, leading to reduced daf-7 production by ASJ (and perhaps also ASI), reduced IIS pathway activity, and increased odr-10 expression.

The TAX-2/TAX-4 channel is required for sensory transduction in roughly 12 classes of C. elegans head sensory neurons, including ASI and ASJ [31, 32]. Although rescue of tax-4 under its own promoter reduced the elevated odr-10 expression of tax-4 mutant males, we observed no such effect when tax-4 was expressed specifically in ASJ, ASI, or both simultaneously (Figures 5D and S4A). Consistent with this, ASI/ASJ-specific expression of tax-4 did not restore Pdaf-7::GFP expression in tax-4 mutants (Figure 5E, F). Thus, sensory perception likely regulates daf-7 non-cell-autonomously, and chemosensation in ASI and ASJ alone is not sufficient to repress odr-10 in well-fed males.

We also examined tax-2(p694), an unusual allele that preserves wild-type function in AWB, AWC, ASI, ASJ, ASK, and ASG [31, 35]. Unlike tax-2 null mutants, tax-2(p694) males showed no increase in odr-10 expression (Figure 5A), indicating that one or more of this group of neurons is sufficient to transduce the sensory signals that repress odr-10. Examining these classes individually, we found that odr-10 expression was unchanged in ceh-36 mutant males (in which AWC does not properly differentiate) and in animals in which AWB or ASK was genetically ablated (Figure S5A, B, C). Furthermore, ASG-specific expression of tax-4 failed to reduce odr-10 expression in tax-4 males (Figure S5D). While we cannot rule out the possibility that tax-2 and tax-4 act in male-specific tail neurons to regulate daf-7 and odr-10, the observation that tax-2 and tax-4 mutants have similar daf-7 expression defects in both sexes strongly argues against this. Together, these results indicate that daf-7 expression reflects the integration of chemosensory signals transduced by multiple tax-2/tax-4-positive neurons (Figure 5G).

Sensory function and physiological status act together to modulate male behavioral prioritization

Finally, we wished to explore the role of this mechanism in the feeding-vs.-exploration decision. First, we measured food-leaving behavior in wild-type, tax-4, and daf-16; tax-4 males. In tax-4 mutants, low daf-7 and high odr-10 expression should inhibit food-leaving [10, 21]. However, food detection is probably also disrupted in these animals, making it difficult to predict their behavior. We found that food-leaving was modestly reduced in tax-4 males (Figure S5E), suggesting that these two effects counteract each other. However, because odr-10 activation in tax-4 mutants requires daf-16, food-leaving might decrease in tax-4; daf-16 males. Indeed, at 24 hr, these males exhibited slightly more food-leaving behavior than did tax-4 males (Figure S5E). However, because constitutive loss of tax-4 and daf-16 likely causes broad changes in worm behavior and physiology, these results should be interpreted cautiously.

We also examined food-leaving behavior in males cultured in the presence of inedible bacteria. Even though these animals are metabolically starved, our model predicts that the failure to upregulate odr-10 (and perhaps other food-detection functions) should cause these animals to exhibit some exploratory behavior. We found that food deprivation caused markedly reduced food-leaving, regardless of whether food cues were present during starvation, indicating that the starved state itself is a strong inhibitor of exploration (Figure 5G). However, males that experienced food cues during starvation exhibited significantly more exploration than did males that were isolated from food cues (Figure 5G). Thus, altered food perception has an important role in the behavioral reprioritization caused by food deprivation.

DISCUSSION

In this work, we have identified a multistep neuroendocrine mechanism that contributes to the homeostatic regulation of behavior in C. elegans males. A key target of this mechanism is the food-associated chemoreceptor odr-10 [10–12], whose expression is activated by food deprivation via DAF-7/TGFβ and IIS. Surprisingly, odr-10 activation in males is not driven by interoceptive signals of hunger or physiological stress; rather, the absence of sensory cues associated with food triggers this mechanism. While the sensory modulation of chemoreceptor expression has precedent in C. elegans [20, 36–38], our work shows that C. elegans males use this strategy to monitor food availability and set the optimal level of odr-10-dependent food attraction.

In our proposed model (Figure 6), multiple food-associated signals are detected by a small number of TAX-2/4-positive chemosensory neurons, including AWB, AWC, ASG, ASI, ASJ, and ASK. We were unable to ascribe specific roles to individual members of this group, suggesting that that adult males use multiple sensory streams to assess the presence of food. This is not surprising: bacterial food sources emit complex and variable chemical signatures [11, 39] and other cues, such as local oxygen concentration, can be indicators of food abundance e. These sensory signals then converge to regulate expression of DAF-7 in ASJ and probably also ASI. In hermaphrodites, daf-7 expression in ASI has been repeatedly shown to represent favorable environmental conditions, and its expression is known to be regulated by sensory signals [17–19]. Our results indicate that male-specific daf-7 expression in ASJ also responds to sensory information, and does so by integrating signals from AWB, AWC, ASK, and/or ASG, and possibly also ASI and ASJ itself, integrating distributed sensory information into a single environmental-state quality signal.

Figure 6. A sex-specific chemosensory feedback loop couples food detection, TGFβ signaling, and IIS to expression of the chemoreceptor odr-10 in adult C. elegans males.

The model shows the proposed mechanism by which chemosensory information about food availability is transmitted through a neuroendocrine loop to regulate odr-10 in AWA. The gray path between ASJ/ASI and AWA represents the possibility that daf-7 signaling may act independently of IIS. This would have to take place through an intermediate cell, as daf-8 does not appear to act in AWA. Pathways at the bottom depict the genetic architecture of the mechanism. For simplicity, the possibility of daf-16-independent functions of daf-7 signaling are not indicated. See Discussion for details.

While ASJ is an important source of DAF-7 with respect to odr-10 regulation, our experiments indicate that ASI-derived DAF-7 has roles in this process as well. This is consistent with the observation that DAF-7 release by either ASJ or ASI can regulate male food-leaving behavior [21]. Nevertheless, the clear male-specificity of the daf-7 mutant phenotype, the food-sensitive expression of daf-7 expression in ASJ, and the requirement for the male state of ASJ indicates that expression in this neuron plays an important role in males. Evaluating the mechanisms by which other sex differences contribute to the regulation of odr-10 in males will be an important focus of future studies.

While we believe the DAF-7 signal acts to modulate neural function to regulate odr-10, this does not appear to happen in RIC or RIM, the canonical sites of daf-7 action [16, 17], or AWA. Because DAF-7 likely acts upstream of IIS, a simple model would be that DAF-7 acts on neurons to regulate the production and/or release of one or more insulin-like peptides (ILPs). Though daf-7 and IIS pathways act largely independently in the dauer development decision [19], daf-7 can also act upstream of IIS [25–28, 40]. C. elegans has 40 ILPs [14]; further research is needed to identify the relevant ligands and the site of their production. Further, we cannot rule out the possibility that DAF-7 acts in parallel to IIS to regulate odr-10.

Finally, one or more ILPs act on AWA to regulate DAF-16, which in turn directly activates odr-10. While daf-16 is best known for its role in activating stress response genes and reprogramming metabolism, our results, together with those of others [38, 41, 42] show that IIS can also regulate chemosensory function. The extent to which other aspects of AWA physiology are influenced by IIS remains unknown, but it seems unlikely that IIS regulates only odr-10. Interestingly, AWA has recently been found to play a male-specific role in detecting volatile pheromones released by hermaphrodites [43]. One intriguing question is whether the mechanism we describe here also regulates pheromone detection by AWA, perhaps blunting it upon food deprivation.

Why might males and hermaphrodites use different mechanisms to couple feeding status to odr-10 expression? Differences in the ways the sexes invest resources may offer insight. In males, sperm production is relatively cheap, but generating higher levels of motor activity may be more demanding [24]. In hermaphrodites, oocyte production is resource-intensive, requiring continuous synthesis of many macromolecular components. Interestingly, recent work has demonstrated dramatic sex differences in the effect of nutritional stress on C. elegans: dietary restriction has essentially no effect on male lifespan, despite its well-known lifespan-extension effect in hermaphrodites [44]. Because of these differences in physiology, nutritional state may be a more critical internal variable in hermaphrodites than in males. In hermaphrodites, numerous signaling mechanisms have been implicated in the regulation of metabolism and behavior by internal physiological cues [45], providing a set of candidate pathways that could be examined for roles in regulating odr-10.

Why males use sensory rather than metabolic cues to regulate odr-10 is unclear. This would seemingly allow rapid assessment of environmental state, but behavioral plasticity in food-leaving behavior and odr-10 expression occurs over hours, not seconds or minutes. The integration of multiple chemosensory streams by daf-7 may occur over a longer time scale, providing a time-averaged signal of environmental conditions. Further, the time needed for signaling downstream of daf-7 also likely contributes to the delay between removal from food and the increase in ODR-10 abundance. Interestingly, males also use information about food presence when deciding whether to copulate with a hermaphrodite: in the absence of food, males mate poorly, possibly to avoid investing in progeny destined to be born into an unfavorable environment [46]. Sensory signals are integral to this, as inedible food provides an environment nearly as permissive for mating as does control food [47]. The relationship of these cues to those that regulate odr-10 are unknown, but they seem likely to overlap.

The findings in this work contribute to a growing appreciation of the importance of chemosensory signals in regulating behavior, metabolism, and internal state across phylogeny. In C. elegans, chemosensory cues have long been known to regulate entry into the dauer state, a stress-resistant larval stage triggered by pheromones and the absence of food; daf-7/TGFβ and IIS pathways are the primary regulators of this decision [14, 16]. In adult hermaphrodites, chemosensory cues can alter gene expression, physiology, behavioral state, and lifespan [17, 38, 48–51]. Recent work in other systems has led to similar conclusions: Drosophila lifespan is modulated by chemosensory function [52], and in mice, sensory detection of food regulates energy homeostasis and the activity of feeding circuits [53, 54]. By establishing an important role for such signals in the male-specific regulation of chemosensory function, our work shows that chemosensory signals contribute to the state-dependent balance between feeding and mating drives, and that they do so at least in part by influencing chemosensory function itself.

STAR METHODS

RESOURCE AVAILABLITY

Lead Contact

• Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Douglas Portman (douglas.portman@rochester.edu).

Materials Availability

• Plasmids and nematode strains generated in the course of this work are freely available to interested academic researchers through the Lead Contact.

Data and Code Availability

• Source data obtained in the current study have not been deposited in a public repository but are available from the Lead Contact on request.

• This study did not generate code.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

All C. elegans strains used were grown on E. coli OP50 using standard methods [55] and maintained at 20°C. Unless otherwise stated, all strains contained him-5(e1490) in order to obtain a high number of males; this is considered the wild-type for the purposes of these studies. Strains carrying mutations in daf-2, daf-7, or daf-8 were maintained at 15°C to prevent dauer entry. In these cases, all paired control strains were also maintained at 15°C. Animals were transferred to 25°C (for daf-2 strains) or 20°C (for daf-7 and daf-8 strains) as L4 larvae; scoring of one-day adults was carried out the following day. Unless otherwise stated, all animals were sex-segregated as L4 larvae and scored as one-day-old adults.

METHOD DETAILS

Starvation experiments

One-day-old adults were washed at least three times with M9 buffer and then transferred to NGM plates with either E. coli OP50 or no food. For GFP quantification experiments, animals were scored after 16-18 h. For behavioral assays, animals were tested after 12 h.

RNA isolation and Quantitative RT-PCR

To isolate RNA, one-day old adult hermaphrodites were collected and either fed or starved for 18 hours and then flash frozen. After cuticle disruption, total RNA was isolated using the RNeasy plus Mini kit (Qiagen) and samples were treated with DNase I (Invitrogen). cDNA was generated with the SuperScript III First-Strand Synthesis System (Invitrogen). Quantitative RT-PCR was performed on a Bio-Rad MyiQ2 icycler with iQ SYBR Green Supermix (BioRad) using 2 μl of cDNA with reactions in triplicate. Cycling conditions were: 1 cycle of 95°C for 5 min; 45 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 45 s, followed by melting curve analysis to verify product specificity. Standard curves were generated for each primer using three serial dilutions of cDNA to determine an efficiency estimate for each run. Threshold cycles (Ct) were determined by the Bio-Rad iQ5 optical system software, which were used with efficiency estimates to calculate relative odr-10 gene expression by the Pfaffl analysis method [56]. Two reference genes, cdc-42 and Y45F10D.4, were used for normalization [57]. odr-10 expression in starved hermaphrodites is shown normalized to levels detected in well-fed hermaphrodites.

Behavioral Assays

Diacetyl attraction assays were performed as described [10, 22, 58]. For food-leaving assays, one-day old adult males were placed on plates with either no food, OP50, or OP50 pre-treated with aztreonam, for 18 hours. Single males were then placed on a 7 μl spot of OP50 in the center of a 9-cm plate as described [7]. At 1, 2, 3, 6, and 24 hr, the tracks were examined to determine the farthest distance animals had traveled from the edge of the food patch. Based on this distance, animals were classified in one of four categories: never left food, minor excursion (<1 cm from the food), major excursion (> 1 cm but <3.5 cm), or left (>3.5 cm).

Aztreonam-treated E. coli

E. coli OP50 was treated with aztreonam as previously described [46]. The day before experiments, E. coli OP50 was grown in LB medium at 37°C with shaking to log-phase growth. Aztreonam was then added to a final concentration of 10 μg/ml and cultures were incubated for 3 additional hours without shaking. Bacteria were then seeded onto freshly made (same-day) NGM plates containing aztreonam (10 μg/ml). To avoid transferring untreated bacteria, worms were washed at least 3 times with M9 buffer before being placed on aztreonam plates. Animals were scored after 16-18 h. For paired control experiments, starved animals were plated on unseeded NGM plates containing 10ug/ml aztreonam to control for any effects of aztreonam on gene expression.

Heat killed E. coli

Liquid cultures of E. coli OP50 were incubated at 75°C for 1 h. Heat-killed bacteria were then plated and used for experiments on the following day.

Molecular biology and generation of transgenic strains

All cDNAs were amplified from total RNA extracted from him-5(e1490) cultures. The resulting cDNAs were cloned into pDONR221 and recombined via Multisite Gateway Cloning (Invitrogen). The Podr-10ΔDAF-16::GFP transgene was generated by Gibson assembly [59]. Plasmids were injected into animals at a concentration of 20-50 ng/ul of rescue construct and 100 ng/ul of Pelt-2::GFP or Pmyo-3::mCherry co-injection marker. See Key Resources Table for primers used to generate transgenes.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Bacterial and Virus Strains
E. coli OP50	CGC	OP50
E. coli OP50-GFP(pFPV25.1)	CGC	OP50-GFP
Chemicals, Peptides, and Recombinant Proteins
Aztreonam	Sigma-Aldrich	A6848
Sodium Azide	Sigma-Aldrich	S2002
Experimental Models: Organisms/Strains
him-5(e1490) V	CGC	CB4088
pha-1(e2123) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	[10]	UR773
him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X	[10]	UR460
daf-16(mgDF47) I; him-5(e1490) V	This work	UR1060
daf-2(e1370) III; him-5(e1490) V	This work	UR288
daf-2(e1370) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1203
daf-16(mgDF47) I; daf-2(e1370) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1204
daf-16(mgDF47) I; pha-1(e2123) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1116
him-5(e1490) V; ksIs2 [Pdaf-7::GFP+rol-6(su1006)]	[13]	UR1265
daf-7(e1372) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1205
daf-7(e1372) III; daf-16(mgDF47) I; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1206
him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsIs15 [Prab-3::fem-3::SL2::mCherry::unc-54 3′ UTR + Punc-122::GFP]	This work	UR1303
him-5(e1490) V; daf-7(e1372); kyIs53 [Podr-10::odr-10::GFP] X	This work	UR912
him-5(e1490) V; daf-7(e1372); kyIs53 [Podr-10::odr-10::GFP] X; fsIs15 [Prab-3::fem-3::SL2::mCherry::unc-54 3′ UTR + Punc-122::GFP]	This work	UR1018
dpy-5 (e61) daf-8 (e1393) I; him-5 (e1490) V; kyIs53 [Podr-10::odr-10::GFP] X	This work	UR915
dpy-5 (e61) daf-8 (e1393) I; him-5 (e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx572 [Pdaf-8::daf-8cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP]	This work	UR1286
dpy-5 (e61) daf-8 (e1393) I; him-5 (e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx573 [Prab-3::daf-8cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP]	This work	UR1287
dpy-5 (e61) daf-8 (e1393) I; him-5 (e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx476 [Pgpa-4Δ6::daf-8cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 1]	This work	UR910
dpy-5 (e61) daf-8 (e1393) I; him-5 (e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx574 [Pgpa-4Δ6::daf-8cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 2]	This work	UR1288
dpy-5 (e61) daf-8 (e1393) I; him-5 (e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx575 [Ptdc-1::daf-8cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP]	This work	UR1289
daf-2(e1370) III; him-5(e1490); zIs356 [Pdaf-16::daf-16::GFP + rol-6(su1006)] ; fsEx400 [Pgpa-4Δ6::odr-10g::SL2mCherry::unc-54-3′UTR+Punc-122::GFP]	This work	UR1290
daf-16(mgDF47) I; pha-1(e2123) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]; fsEx534[Pnhx-2::daf-16f cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line1]	This work	UR1138
daf-16(mgDF47) I; pha-1(e2123) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]; fsEx535[Pnhx-2::daf-16f cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line2]	This work	UR1139
daf-16(mgDF47) I; pha-1(e2123) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]; fsEx532[Pgpa4Δ6::daf-16f cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line1]	This work	UR1136
daf-16(mgDF47) I; pha-1(e2123) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]; fsEx533[Pgpa4Δ6::daf-16f cDNA::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line2]	This work	UR1137
him-5(e1490) V; fsEx513[Podr-10::GFP + Pmyo-3::mCherry line 1]	This work	UR1100
him-5(e1490); fsEx514[Podr-10ΔDAF−16::GFP + Pmyo-3::mCherry line 1]	This work	UR1101
him-5(e1490); fsEx571[Podr-10ΔDAF−16::GFP + Pmyo-3::mCherry line 2]	This work	UR1284
tax-4(ok3771) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1201
daf-16(mgDF47) I; tax-4(ok3771) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1202
tax-4(ok3771) III; him-5(e1490) V	This work	UR1324
daf-16(mgDF47) I; tax-4(ok3771) III; him-5(e1490) V	This work	UR1325
tax-4(ok3771) III; him-5(e1490) V; ksIs2 [Pdaf-7::GFP+rol-6(su1006)]	This work	UR1196
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X	This work	UR1194
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx550[Ptax-4::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP]	This work	UR1222
tax-4(ok3771) III; him-5(e1490) V; ksIs2 [Pdaf-7::GFP+rol-6(su1006)]; fsEx550[Ptax-4::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP]	This work	UR1291
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx553 [Pgpa-4::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 1]	This work	UR1234
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx554[Pgpa-4::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 2]	This work	UR1235
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx555 [Pssu-1::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 1]	This work	UR1236
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx556 [Pssu-1::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 2]	This work	UR1237
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx557 [Pgpa-4::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pssu-1::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt2::GFP line 1]	This work	UR1238
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx558 [Pgpa-4::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pssu-1::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt2::GFP line 2]	This work	UR1239
tax-4(ok3771) III; him-5(e1490) V; ksIs2 [Pdaf-7::GFP+rol-6(su1006)]; fsEx557 [Pgpa-4::tax-4c::SL2::mCherry + Pssu-1::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt2::GFP line 1]	This work	UR1292
tax-4(ok3771) III; him-5(e1490) V; ksIs2 [Pdaf-7::GFP+rol-6(su1006)]; fsEx558 [Pgpa-4::tax-4c::SL2::mCherry + Pssu-1::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 2]	This work	UR1293
tax-2(p691) I; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1294
tax-2(p694) I; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1295
eat-2(ad465) II; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1296
ceh-36(ky640) X; pha-1(e2123) III; him-5(e1490) V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1297
him-5(e1490) V; Is[Pstr-1::mCasp1]; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1298
him-8(e1489) IV; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1299
him-8(e1489) IV; qrIs2[Psra-9::mCasp1] V; fsEx295 [ODR-10::GFP fosmid + pBx1]	This work	UR1300
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx576 [Pops-1-::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 1]	This work	UR1301
tax-4(ok3771) III; him-5(e1490) V; kyIs53 [Podr-10::odr-10::GFP] X; fsEx577 [Pops-1-::tax-4c::SL2::mCherry::unc-54 3′ UTR + Pelt-2::GFP line 2]	This work	UR1302
him-5(e1490) V; jxEx100[ptrx-1::ICE + pofm-1::GFP line 1]; fsEx295	This work	UR1339
him-5(e1490) V; jxEx101[ptrx-1::ICE + pofm-1::GFP line 2]; fsEx295	This work	UR1340
him-5(e1490) V; fsEx482[pssu-1::tra-2(ic)::mCherry unc-54 3′UTR + pelt-2::GFP line 1]; kyIs53	This work	UR1341
him-5(e1490) V; fsEx483[pssu-1::tra-2(ic)::mCherry unc-54 3′UTR + pelt-2::GFP line 2]; kyIs53	This work	UR1342
Oligonucleotides
See Table S1 for oligonucleotide information
Recombinant DNA
Pssu-1::tax-4c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Pgpa4Δ6::daf-16f c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Pnhx-2::daf-16f c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Pdaf-8::daf-8c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Prab-3::daf-8c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Ppga-4Δ6::daf-8c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Ptdc-1::daf-8c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Pops-1::tax-4c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Podr-10::GFP	This work	N/A
Podr-10ΔDAF−16::GFP	This work	N/A
Pelt-2::GFP	Portman laboratory	N/A
Pmyo-3::mCherry	Portman laboratory	N/A
Ptax-4::tax-4c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Pgpa-4::tax-4c::SL2::mCherry::unc-54 3′ UTR	This work	N/A
Software and Algorithms
Axiovision	Carl Zeiss	N/A
FIJI	[72]	N/A
ApE, A Plasmid Editor	M Wayne Davis	N/A
Prism 8	GraphPad Software	N/A
Other
Zeiss Axioplan 2	Carl Zeiss	N/A

QUANTIFICATION AND STATISTICAL ANALYSIS

For quantification of ODR-10::GFP, Podr-10::GFP, and Pdaf-7::GFP, animals were mounted on a 4% agarose pad and immobilized with levamisole. GFP fluorescence was observed using a 63x PlanApo objective on a Zeiss Axioplan 2. Due to the complex three-dimensional structure of the AWA cilia, accurate quantitative measurements of ODR-10::GFP fluorescence intensity were not possible. ODR-10::GFP intensity was therefore scored using a scale of 0-3 (0=absent, 1=faint, 2=moderate, 3=bright) as previously described [10, 13]. When possible, the experimenter was blinded to genotype. For quantification of Podr-10::GFP and Pdaf-7::GFP expression, the integrated density of GFP was calculated using FIJI software [60]. Images were obtained under consistent imaging conditions with the cell of interest in center of the field. Background was calculated by taking the average mean fluorescence of three random background regions within the area of the animal. Total cell fluorescence was calculated as (integrated density of GFP in the region of interest) – (mean background fluorescence x area of the region of interest).

For categorical data (e.g., fluorescence intensity of ODR-10::GFP and behavior in the food leaving assay), data were analyzed using non-parametric tests. For pairwise comparisons, Mann-Whitney tests were carried out. For multiple comparisons, Kruskal-Wallis analysis with Dunn’s correction was used. For other analyses (e.g., of diacetyl chemotaxis and quantitated GFP fluorescence), data were assumed to be normally distributed. In these cases, Welch’s t tests were used for pairwise comparisons. For multiple comparisons, we used Brown-Forsyth and Welch one-way ANOVA tests with Holm-Sidak’s correction. In all experiments, only those comparisons necessary to test the hypotheses under consideration were carried out. For this reason, graphs do not show the results of all possible comparisons. Rather, all graphs show the results of all comparisons carried out. Dashed brackets indicate comparisons for which p > 0.05. Asterisks indicate p-value ranges as follows: *0.01 < p < 0.05; **0.001 < p < 0.01; *** p < 0.001.