GRD-1/PTR-11, the C. elegans hedgehog/patched-like morphogen-receptor pair, modulates developmental rate

Hedgehog (Hh) morphogens are highly conserved positive regulators of development found throughout vertebrates and invertebrates (Aspöck et al., 1999; Echelard et al., 1993; Nüsslein-Volhard and Wieschaus, 1980). Extensive study in organisms such as Drosophila melanogaster has revealed a canonical pathway wherein Hh proteins bind to Patched (Ptc) allowing shifts in gene expression via the Gli transcription factors (Ingham, 2022). However, these proteins remain relatively unexplored in the model organism Caenorhabditis elegans. In C. elegans, divergence from the canonical signaling pathway resulted in the loss of some Hh proteins, while expanding the Hh-related (Hh-r) and Patched-related (Ptr) protein families from only five members to over 80 members (Aspöck et al., 1999). Of these, only 14 have been mechanistically characterized and only two of these, wrt-10 and grl-21, are functionally associated with patched/patched-related receptors (Lin and Wang, 2017; Templeman et al., 2020). From molting to reproductive aging, all of the Hh-r proteins that have been studied serve important, non-redundant roles in C. elegans, e.g. Hh-r proteins such as qua-1 are indispensable for larval transition whereas others, such as wrt-10, govern aspects of healthspan downstream of conserved regulators such as CREB (Hao et al., 2006a,b; Templeman et al., 2020).

Activity of highly conserved, nutrient-sensing signaling pathways is integral to ensuring normal growth rate in C. elegans. For example, reduced target of rapamycin complex 2 (TORC2) signaling has pronounced effects on developmental rate (Jones et al., 2009; Soukas et al., 2009). Loss of function of the gene encoding the essential TORC2 subunit rictor extends time to adulthood at 20°C from 48 h to 72 h, reduces body size, lowers brood, increases fat and shortens lifespan (Jones et al., 2009; Soukas et al., 2009). This suggests that the kinase complex integrates nutritional signals to determine how the animal proceeds through development, via as yet uncharacterized pathways. Our previous work has shown that depletion of the essential dosage compensation complex (DCC) member dpy-21 downstream of the TORC2 effector kinase serum- and glucocorticoid-induced kinase 1 (SGK-1) was sufficient to rescue the delayed development, reduced brood and increased fat of rict-1 mutants – but not their reduced body size and shortened lifespan – via action of the histone methyltransferases SET-1 and SET-4 (Webster et al., 2013). However, the spectrum of effectors of TORC2 governance over developmental rate and healthy growth remains incompletely characterized.

In this study, we identify grd-1 and ptr-11 (hereafter, grd-1/ptr-11) as key components of a Hh-r to Ptr signaling relay that negatively regulates development downstream of TORC2. We find that, like many other Hh-r proteins, grd-1 expression is animated during molting transitions, and, surprisingly, grd-1 knockdown by RNAi leads to developmental acceleration in wild-type animals. Furthermore, grd-1 knockdown significantly rescues the slowed development of TORC2 mutants. We also identify ptr-11 as a likely recipient of grd-1 signaling. ptr-11 knockdown phenocopies the ability of grd-1 knockdown to accelerate growth in wild-type animals and to restore normal growth rate to TORC2 mutants. Importantly, augmented grd-1 expression is sufficient to delay development, and ptr-11 knockdown significantly rescues the delayed development of grd-1 overexpressor animals. Finally, we identify pqm-1 as a potential effector of the grd-1/ptr-11 signaling cascade. pqm-1 knockdown partially phenocopies grd-1/ptr-11 knockdown and pqm-1 target genes are upregulated in TORC2 loss-of-function mutants in a manner dependent upon both grd-1 and ptr-11. In aggregate, we propose that in the setting of a deficiency in TORC2 signaling, grd-1/ptr-11 may act downstream to upregulate pqm-1 activity and cause developmental delay in C. elegans.

Knockdown of Hh-r proteins in C. elegans has been reported to prompt developmental delay, arrest or defects (Cohen et al., 2021; Hao et al., 2006a,b; Zugasti et al., 2005). Aside from high-level examination of conserved domains and transcriptional analyses, the Groundhog (GRD) family is absent from previous investigations into Hh-r proteins (Aspöck et al., 1999; Bürglin, 1996; Hao et al., 2006a,b), therefore we sought to determine whether knockdown of these proteins affects worm development (Aspöck et al., 1999). We performed a small-scale screen of grd family RNAi on larval development to adulthood, as determined by the appearance of adult vulval morphology, finding that, although grd-7 knockdown does slow larval development, grd-1 knockdown accelerates development (Fig. 1A,B; Table S1 for tabular results and biological replicates). We confirmed that the Ahringer RNAi clone reduces grd-1 transcript levels by ∼90% and that an additional, non-overlapping, grd-1 RNAi clone we synthesized also accelerates development in wild-type animals (Fig. S1A,B).

Previous transcriptomic profiling of C. elegans development indicates cyclical expression of grd-1 mRNA (Hendriks et al., 2014). We confirm that grd-1 mRNA levels peak at molting, which is when peaks in mRNA encoding the cyclically expressed molting gene mlt-10 can be seen via qRT-PCR (Fig. 1C). A grd-1p::GFP promoter reporter confirms a significant increase in grd-1 expression during the L3/L4 molt and the L4/young adult (YA) molt that specifically occurs in the intestine (Fig. 1D). We also replicated previous reports of grd-1 expression in rectal epithelial cells (Aspöck et al., 1999) and further observe expression in other tissues, including head neurons, hypodermis, intestine and vulva (Fig. S1C). Only intestinal expression changes in intensity across developmental stages.

Having determined that grd-1 is necessary for normal developmental rate in wild-type animals and that it is cyclically expressed with molting, we performed an RNA-sequencing analysis of grd-1 RNAi versus vector at the L3 developmental stage in order to ascertain whether molting factors are relevant to the developmental acceleration caused by grd-1 knockdown. Indeed, the top 100 differentially expressed genes resulting from grd-1 knockdown are enriched for GO terms matching molting cycle and cuticle-related genes (Fig. 1E; Table S2).

Larval progression is gated by nutritional rheostats and heterochronic genes that tightly determine if and when the animal transitions from one larval stage to the next (Mata-Cabana et al., 2021; Moss, 2007). Based upon its developmental pattern of expression, we hypothesized that grd-1 may act downstream of major controllers of larval growth. Therefore, we assessed the effect of grd-1 RNAi on the development of several mutants in growth factor and/or nutrient sensing pathways, i.e. daf-2(e1370) (insulin-like receptor hypomorph), eat-2(ad465) (defective pharyngeal pumping which leads to caloric restriction), raga-1(ok386) (hypomorphic defects in mTOR complex 1 signaling), rsks-1(ok1255) (S6 Kinase) and rict-1(mg451) (mTOR complex 2 loss of function) (Fig. 2A-E; Table S1). As in wild type, grd-1 RNAi accelerates the L1 to YA development time of all the mutants by ∼2 h, except for rict-1 mutants, which matured ∼7 h faster (Fig. 2E). Moreover, the proportional acceleration in median time to adulthood is significantly greater in rict-1 mutants subjected to grd-1 RNAi relative to acceleration seen in wild type and other mutants assessed (Fig. 2F).

Previous work has identified the serine-threonine kinase SGK-1 as a major downstream effector of TORC2 in C. elegans and Saccharomyces cerevisiae (Aronova et al., 2008; Jones et al., 2009; Webster et al., 2013; Zhou et al., 2019). sgk-1 loss-of-function mutants phenocopy the slowed development, reduced brood, small body size and shortened lifespan of rict-1 mutants (Soukas et al., 2009; Webster et al., 2013; Zhou et al., 2019). We therefore tested whether grd-1 knockdown functions to accelerate rict-1 null development by increasing SGK-1 activity or by acting downstream of sgk-1. We find that sgk-1 null mutant development on grd-1 RNAi is accelerated by ∼7 h, similar to rict-1 mutants (Fig. 3A; Table S1). Further substantiating that grd-1 acts downstream of both rict-1 and sgk-1, grd-1 knockdown on a rict-1;sgk-1 double null mutant speeds up development by ∼7 h (Fig. 3B; Table S1). Slowed development caused by lowered TORC2 signaling in loss-of-function mutations in the essential TORC2 subunit sinh-1 (homolog of mTORC2 subunit mSin1) is similarly accelerated by grd-1 knockdown (Fig. 3C; Table S1). Finally, broader relevance of grd-1 to TORC2 biology is indicated by the fact that grd-1 RNAi also partially rescues the shortened lifespan, increased fat mass, small body size and reduced brood size of rict-1 mutants (Fig. S2A-D; Table S3) (Soukas et al., 2009).

If, as we hypothesize, grd-1 slows growth in response to inadequate TORC2/SGK signaling, increased grd-1 activity might be sufficient to slow development. In order to test this possibility, we generated transgenic C. elegans overexpressing grd-1 under the control of its native promoter. Indeed, grd-1 overexpression was sufficient to prompt slowing of developmental rate of wild-type animals to a degree similar to rict-1 mutants (Fig. 3D). We validated this result using CRISPR activation (CRISPRa) as an orthogonal method of increasing grd-1 expression; even a modest ∼2.5-fold increase in grd-1 transcript levels caused significant delay in wild-type C. elegans development (Fig. S3A,B). Although attempts were made at measuring GRD-1 protein levels via the overexpressor GFP fusion, we were unable to generate viable rict-1(mg451);grd-1::GFP and sgk-1(mg455);grd-1::GFP doubles, possibly owing to excessive GRD-1 activity from the combination of TORC2 loss of function in a GRD-1 overexpression background.

We have previously identified non-canonical activity of the dosage compensation complex (DCC) as being partially necessary for developmental delay downstream of TORC2 (Webster et al., 2013). This raised the possibility that grd-1 may be acting downstream of or in concert with the DCC. However, grd-1 transcript levels are unperturbed in both rict-1 mutants and DCC subunit dpy-21 RNAi-treated animals, suggesting that grd-1 is not transcriptionally regulated by either TORC2 or the DCC during larval development (Fig. S3C-E).

In order to probe for potential receptors acting downstream of grd-1, we performed a screen on the patched and patched-related receptor (ptc and ptr, respectively) gene families to identify phenocopiers of grd-1. Of the ptc and ptr family members tested, only ptr-11 RNAi accelerates development (Fig. 4A). Indeed, treatment of rict-1 mutants, rict-1;sgk-1 double mutants and sinh-1 mutants with ptr-11 RNAi produces quantitatively similar acceleration of growth rate as grd-1 RNAi (Fig. 4B,C, Fig. S4A; Table S1), suggesting that grd-1 and ptr-11 may function in a common genetic pathway downstream of TORC2/SGK-1 signaling. Consistent with this possibility, in slow-growing grd-1 overexpressors, ptr-11 RNAi accelerates development by ∼4 h compared with ∼1 h in wild-type animals, substantiating the notion that ptr-11 functions downstream of grd-1 (Fig. 4D; Table S1). Furthermore, similar to grd-1 overexpression, mild ∼1.7-fold induction of ptr-11 by CRISPRa causes developmental delay in wild-type worms (Fig. 4E,F).

Using an endogenously tagged ptr-11::GFP generated by CRISPR/Cas9 genome engineering, we observe ptr-11 expression in the hypodermis, head and tail neurons, dorsal and ventral nerve cords, and seam cells at both the L3 and L4 stages (Fig. S4B). As is the case for grd-1, previous transcriptomics analyses have shown that ptr-11 oscillates in expression throughout development (Hendriks et al., 2014), which we confirmed by qRT-PCR (Fig. S4C).

In order to identify potential downstream effectors of grd-1/ptr-11 signaling, we returned to our RNA-sequencing results. Analysis of these data indicates that target genes of transcription factors PQM-1 and BLMP-1 are enriched after grd-1 knockdown (Fig. 5A; Table S2). Assessment of developmental rate of both wild-type and rict-1 animals with pqm-1 and blmp-1 RNAi indicates that only pqm-1 knockdown accelerates development (Fig. 5B; Fig. S5). From our RNA-sequencing data, we identified the six pqm-1 target genes that are most decreased by grd-1 knockdown and found that two of these, C49G7.7 and ugt-43, are upregulated in rict-1 loss-of-function mutants and confirmed to be downregulated by grd-1 RNAi by qRT-PCR (Fig. 5C). This suggests that PQM-1 activity is increased by TORC2 loss of function in a grd-1-dependent manner. In support of the conclusion that grd-1 drives PQM-1 activity, pqm-1 knockdown decreases ugt-43 expression in rict-1 mutants (Fig. 5D). These results indicate that PQM-1 targets are regulated in a TORC2-dependent manner during development.

To further elucidate a connection between rict-1 and pqm-1, we examined endogenously tagged pqm-1::GFP nuclear localization at the L3 stage in wild type, rict-1 loss of function mutants and grd-1 overexpressors (grd-1oe). PQM-1::GFP shows strong nuclear localization in both rict-1 mutants and grd-1oe animals compared with wild type (Fig. 5E). In support of a TORC2/GRD-1/PTR-11/PQM-1 signaling axis, increased PQM-1 activity read out as ugt-43 expression in rict-1 mutants and grd-1 overexpression transgenics is dependent upon ptr-11, as knockdown abrogates increased ugt-43 mRNA levels in both backgrounds (Fig. 5F).

Finally, we attempted to phenocopy grd-1, ptr-11 and pqm-1 RNAi with null alleles in these genes but were unable to recapitulate developmental rescue in rict-1(mg451) and grd-1oe animals (Fig. S6A-D). Furthermore, the pqm-1(ok485) allele does not phenocopy pqm-1 RNAi in decreasing high ugt-43 mRNA levels in rict-1(mg451) worms (Fig. S6E). However, we show that the non-overlapping grd-1 RNAi clone phenocopies the Ahringer grd-1 RNAi clone in rict-1(mg451) animals (Fig. S6F), and results with CRISPRa of grd-1 and ptr-11 are internally consistent with these genes that provoke developmental slowing (Fig. S3A,B, Fig. 4E,F).

Correctly integrating nutritional signals from the environment is paramount to ensuring optimal growth rate and development. Previous work by us and others indicates that TORC2 is crucial to the governance of growth, reproduction and lifespan, regulating these processes in a diet-dependent manner (Jones et al., 2009; Soukas et al., 2009). Defects in TORC2/SGK-1 signaling lead to slow growth, substantiating the important role of this complex in ensuring environmentally appropriate organismal growth. In this article, we propose that the grd-1/ptr-11 hedgehog signaling axis may act downstream or converge on downstream effectors of TORC2 signaling (Fig. 6). Based upon data presented here, we suggest that increased grd-1 activity signals through ptr-11 to slow growth by activating pqm-1 transcriptional responses. This work not only identifies a new cognate ligand-receptor pair of Hh-Ptr in grd-1/ptr-11 but is also the first indication that TORC2 may regulate Hh signaling to negatively affect whole-organism development, growth and metabolism. Although it remains a possibility that GRD-1 is regulated by a parallel mechanism and converges on downstream effectors of TORC2 signaling, our work suggests that GRD-1/PTR-11/PQM-1 signaling executes developmental slowing under reduced TORC2 signaling conditions (Fig. 6).

The study of developmental timing in C. elegans has most extensively explored heterochronic genes required for correct developmental event sequencing, such as lin-4, lin-14 and let-7 (Ambros and Horvitz, 1984; Ambros and Moss, 1994; Reinhart et al., 2000). By comparison, there have been fewer dissections of the specific mechanisms controlling developmental rate in C. elegans. Many such investigations have focused on pathways controlling developmental progression in the absence of food or certain nutrients, revealing crucial roles for IIS signaling, TORC1, lipid metabolism, the one-carbon cycle and micronutrients in dauer formation and larval arrest (Baugh and Sternberg, 2006; Galles et al., 2018; Long et al., 2002; Watson et al., 2014; Watts et al., 2018). However, our understanding of the role of TORC2 signaling in development remains largely incomplete, limited to the DCC, to a gut-neuronal axis linking TORC2 to TGF-β signaling, and to CDC-42 induced neuronal protrusions (Alan et al., 2013; O'Donnell et al., 2018; Webster et al., 2013). Here, our results indicate a mechanism by which the Hh-r morphogens grd-1 and ptr-11 may act downstream of TORC2 to control whole-organism development and growth. Although study of grd-1 and ptr-11 activity per se, is challenging, and given that neither grd-1 nor ptr-11 is obviously regulated at the RNA or protein level downstream of TORC2, we are still able to connect TORC2 inactivation to increased GRD-1 signaling by invoking the downstream activity of PQM-1. However, it remains an unanswered question whether TORC2 is specifically modifying GRD-1 activity or whether TORC2 downregulation produces a metabolic environment in which GRD-1 activity is increased. Disentangling these possibilities will require a robust GRD-1 activity assay. Both loss of TORC2 and grd-1 overexpression lead to increased PQM-1 activity, as measured both by increased nuclear localization of the transcription factor and increased PQM-1 target gene expression. These results suggest that, under conditions of lowered TORC2 signaling, activity of grd-1/ptr-11 is induced to slow development. Although we find that the effect of grd-1/ptr-11 knockdown acts downstream of the major TORC2 effector kinase SGK-1, it remains unclear how grd-1 activity is directly regulated. The lack of grd-1 transcriptional changes in rict-1 loss-of-function mutants suggests a post-transcriptional regulatory mechanism. Attempts to make an endogenously tagged mature GRD-1 protein were not successful and have proved challenging according previous reports (Aspöck et al., 1999), so additional work is needed to define the precise mechanisms of GRD-1 activation. Notably, our data do not show a transcriptional interaction between the DCC and grd-1, indicating that TORC2 may control development through these two arms in parallel or that grd-1/ptr-11 converge downstream of TORC2.

C. elegans has a much-expanded family of Hedgehog morphogens that are thought to share a common ancestor with other phyla (Aspöck et al., 1999). Furthermore, C. elegans possesses an expanded family of patched-related receptors but lacks several canonical hedgehog signaling components, such as smoothened and a truly orthologous Gli transcription factor (Zugasti et al., 2005). Hh-r morphogens and Ptr receptors have shared biological functions in controlling developmental progression, and have been shown to interact cell non-autonomously with one another as in the case of wrt-10/ptc-1 and grl-21/ptr-24 (Lin and Wang, 2017; Templeman et al., 2020).

Here, we provide only the third example of a Hh-r-Ptr pair that functions in tandem to regulate metabolism in C. elegans. Our findings complement previous reports indicating that Hh-r proteins control important parts of organismal homeostasis such as reproductive health downstream of CREB via wrt-10/ptr-2 signaling (Templeman et al., 2020). Based upon patterns of grd-1 promoter activity, we suggest that grd-1 is produced in the intestine in a cyclical fashion during molting and that, under conditions of lowered TORC2 signaling, grd-1 function is increased to put the brakes on development. When grd-1 levels are artificially increased by overexpressing the protein under its native promoter, worm development is slowed to near rict-1 null mutant levels. Given that the intestine is both the primary tissue where grd-1 is expressed and the principal tissue of action for TORC2 in regulation of growth and metabolism (Soukas et al., 2009), it seems plausible that TORC2 controls development by modulating grd-1 activity cell-autonomously in the intestine. Proof for this hypothesis will require a better understanding of precise events occurring downstream of GRD-1 activity. As it stands, it remains a possibility that a TORC2-deficient worm may have increased GRD-1 activity independently of direct TORC2 interaction. Like other Hh-r proteins, GRD-1 is predicted to be an extracellular protein, suggesting a cell non-autonomous role for the morphogen. We identify ptr-11 as the likely receptor for grd-1, given that its knockdown significantly restores developmental rate in TORC2 mutants and grd-1 overexpression transgenics. Our endogenous ptr-11::GFP fusion indicates cell surface expression in several tissues, including the hypodermis and seam cells, but not the intestine. Definitive determination of the respective sites of action of grd-1 and ptr-11 will require additional investigation, potentially through colocalization assays such as FRET or a yeast two hybrid system. In aggregate, these results suggest that TORC2 takes stock of growth environment status and, if conditions are favorable, grd-1 activity is tamped down. However, in response to unfavorable conditions, reduced TORC2/SGK-1 signaling directly or indirectly increases grd-1 activity, likely in the intestine, and grd-1 executes a program delaying organismal growth by relaying the signal of unfavorable conditions through ptr-11.

Moreover, we identify pqm-1 as a potential downstream effector of the grd-1/ptr-11 signaling relay. Globally, PQM-1 acts antagonistically to DAF-16 in IIS signaling mutants to regulate lifespan and development (Tepper et al., 2013). Additionally, PQM-1 has been previously shown to work downstream of TORC2 to mobilize fat from the intestine to the germline at the onset of adulthood (Dowen et al., 2016). Although pqm-1 loss-of-function mutations cause slight developmental delay (Tepper et al., 2013), larval knockdown by RNAi causes developmental acceleration that partially recapitulates grd-1/ptr-11 knockdown in both wild-type animals and TORC2 loss of function. Using both increased expression of the pqm-1 target gene ugt-43 and nuclear PQM-1::GFP localization as a barometer for increased pqm-1 activity in rict-1 and grd-1oe animals, we suggest that grd-1 is activated in rict-1 mutants, which in turn activates pqm-1 to effectuate growth delay in a ptr-11-dependent manner. That said, without a direct GRD-1 activity assay, it is possible that grd-1/ptr-11 are influencing pqm-1 in parallel with TORC2. Furthermore, we fail to recapitulate the RNAi data with putative loss-of-function alleles in grd-1, ptr-11 and pqm-1. It is possible, given reports of lethality from grd-1 null mutations, that pleiotropic roles of these genes prevent mutations from phenocopying post-embryonic inactivation by RNAi. Alternatively, compensation for chronic loss of these crucial factors mutes or negates changes seen with short-term inactivation, a possibility that we favor. In support of this theory, a non-Ahringer grd-1 RNAi clone phenocopies the Ahringer grd-1 RNAi clone in rict-1(mg451) worms. Taken with RNAi data from other members of the proposed pathway, as well as findings with CRISPRa, this implies that the phenotypes observed are unlikely to be RNAi artifacts. Conclusive proof will require an orthogonal conditional loss-of-function strategy that does not rely on protein degradation, given the extracellular and membrane-bound nature of GRD-1 and PTR-11, respectively. Overall, the data support the conclusion that a grd-1/ptr-11/pqm-1 relay governs growth rate, potentially downstream of TORC2. Defining whether TORC2 directly modulates grd-1 activity, definitive proof that grd-1/ptr-11 represent a ligand-receptor pair and how ptr-11 signaling increases pqm-1 activity will require further investigation.

In summary, we define a heretofore unappreciated coordination of the TORC2 and hedgehog pathways in a signaling axis that functions to put the brakes on development when reduced TORC2 signaling indicates unfavorable environmental conditions. Future work will focus on which aspects of the nutrient milieu are sensed by TORC2, how changes that prompt developmental slowing are communicated through the GRD-1/PTR-11/PQM-1 signaling relay, and whether those changes are communicated directly from TORC2 to GRD-1 or through an independent mechanism. Better understanding of this biology will provide information about how organisms govern growth rate through ancient and complex communication between diverse signaling networks.

C. elegans animals were grown and maintained at 20°C on Nematode Growth Media (NGM) seeded with Escherichia coli OP50-1 as previously described (Soukas et al., 2009). The following strains were used: wild type N2 (Bristol), MGH266 rict-1(mg451), MGH300 sgk-1(mg455), CB1370 daf-2(e1370), DA465 eat-2(ad465), VC222 raga-1(ok386), MGH9 rsks-1(ok1255), MGH27 rict-1(mg451);sgk-1(mg455), MGH35 sinh-1(mg452), MGH629 alx88[ptr-11p::ptr-11::GFP::AID::3xFLAG], MGH531 alxIs41[grd-1p::grd-1::GFP::grd-1::grd-1 3'UTR], MIR249 risIs33[K03A1.5p::3xFLAG::SV40-NLS::dCas9::SV40-NLS::VP64::HA+unc-119(+)], MGH563 alxEx127[grd-1p::GFP], SYS573 dev60[pqm-1p::GFP::pqm-1], MGH618 rict-1(mg451);dev60[pqm-1p::GFP::pqm-1], MGH634 alxIs41[grd-1p::grd-1::GFP::grd-1::grd-1 3'UTR];dev60[pqm-1p::GFP::pqm-1], MGH696 rict-1(mg451) II;grd-1(gk428716) X, MGH694 rict-1(mg451) II; ptr-11(gk489940) I, MGH627 rict-1(mg451) II;pqm-1(ok485) II and MGH695 alxIs41[grd-1p::grd-1::GFP::grd-1::grd-1 3'UTR myo-3p::mCherry];ptr-11(gk489940) I.

Animals were synchronized by hypochlorite bleach treatment of a population of gravid adults. Animals were collected in M9 medium, centrifuged at 3300 g for 1 min and resuspended in 6 ml 1.3% bleach, 250 mM NaOH for 1 min of vigorous shaking. Bleach solution exposure and the shaking step were repeated after one wash in M9. After the second bleach step, the pellet was resuspended in minimal M9 by gentle pipetting and washed in M9 four times. Eggs were resuspended in 12 ml M9 and the solution was left to rotate overnight for 23 h at 20°C.

RNA interference (RNAi) plates were prepared with standard NGM media mixed with 5 mM isopropyl-B-D-thiogalactopyranoside and 200 µg/ml carbenicillin. All RNAi clones (except for confirmatory clones, which were constructed to reduce concerns over off-target effects; cloning primers are shown below) were isolated from the genome wide E. coli HT115 Ahringer library (Horizon Discovery) and sequence verified before use. RNAi clones were obtained by seeding onto ampicillin/tetracycline treated Luria Broth (LB) plates. RNAi clones were grown overnight at 37°C with shaking for 18 h in LB with 200 µg/ml carbenicillin. Cultures were spun down at 3300 g for 15 min, pellets resuspended in one-tenth starting LB volume and dispensed onto RNAi plates no more than 48 h before adding worms. All assays not carried out on RNAi were performed on vector HT115 seeded plates. Cloning primers used to generate a unique grd-1 RNAi clone were: unique grd-1 RNAi clone F, GAAATAGATGCTGTGATGGAAG; unique grd-1 RNAi clone R, CGGTGTTCTTCCTATCTGTC.

Synchronized L1 animals prepared as described above were dropped onto RNAi plates with vector RNAi control or RNAi directed against the gene of interest at 20°C. Time to adulthood was measuring by evaluating the appearance of the vulvar slit in hermaphrodite animals. Two or three biological replicates were performed for all conditions and data are available in Table S1.

Developmental screening was performed by evaluating the relative proportion of wild-type adults on treatment RNAi compared with vector RNAi at 49 h. Three biological replicates were pooled for analysis.

Synchronized L1 animals were dropped onto RNAi plates with the appropriate RNAi clone. After having reached young adult (YA) stage, two hermaphrodites were transferred onto five plates with the corresponding RNAi clone for a total of ten adults per condition. Adults were transferred every day for 5 days. Brood was measured as the total amount of progeny on the plates after 3 days at 20°C.

At day one of adulthood, synchronous animals were collected in M9, washed once, centrifuged at 375 g for 1 min, and resuspended in 40% isopropanol for 3 min. Fixed worms were then stained with 3 µg/ml Nile Red in 40% isopropanol for 2 h. Animals were collected in M9 supplemented with 0.01% Triton-X100, mounted on glass slides and imaged in the GFP channel for 10 ms at 5× magnification (Pino et al., 2013). Body fat mass was measured as the fluorescence intensity relative to body area in pixels using ImageJ.

Adult day one animals were mounted on 2% agarose pads in 2.5 mM levamisole and imaged by bright-field microscopy at 2.5×magnification. Body area in pixels was determined using ImageJ.

Synchronous L1 animals were dropped onto the appropriate RNAi clones and allowed to grow to L4/YA stage. 40-60 worms were then transferred to four plates with the corresponding RNAi clone supplemented with 10-50 µM 5-fluorodeoxyuridine (FUDR) for a total of ∼200 adults per condition. Animals were scored as dead or alive by movement or response to gentle prodding every other day. Data were analyzed using OASIS2 software package (https://sbi.postech.ac.kr/oasis2/).

Two grd-1 and ptr-11 guides were cloned into the L4440_BioBrick-sgRNA vector as previously described (Fischer et al., 2022). MIR249 [K03A1.5p::3xFLAG::SV40-NLS::dCas9::SV40-NLS::VP64::HA+unc-119(+)] worms were synchronized and assayed at 51 h for the ratio of L4s to YAs. Quantitative RT-PCR was performed to confirm an increase in grd-1 and ptr-11 transcript levels. Primers used to clone sgRNAs include: grd-1 guide 1 F, GTTCGGATTGAGTAGTGCAA; grd-1 guide 1 R, TTGCACTACTCAATCCGAAC; grd-1 guide 2 F, ACCGATTAGATCCAGGAAAG; grd-1 guide 2 R, CTTTCCTGGATCTAATCGGT; ptr-11 guide 1 F, TGTTTGGAAAAGGATTGAAG; ptr-11 guide 1 R, CTTCAATCCTTTTCCAAACA; ptr-11 guide 2 F, TGTATTGGGTAGGAGAGGGA; and ptr-11 guide 2 R, TCCCTCTCCTACCCAATACA.

Worms were collected in TRIzol Reagent (Invitrogen), flash frozen in liquid nitrogen and stored at −80°C before RNA isolation. Samples were lysed using metal beads and the Tissuelyser I system (Qiagen). RNA was isolated from lysate using Direct-zol RNA Miniprep kit (Zymo). cDNA samples were synthesized with the QuantiTect reverse transcription kit (Qiagen). Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) was performed with the QuantiTect SYBR Green RT-PCR kit (Qiagen) in a Bio-Rad CFX96 RT-PCR thermocycler. Fold changes were determined via the 2^−ΔΔCt method, as previously described (Livak and Schmittgen, 2001). All Ct counts were normalized to act-1 Ct counts. Primers used for qRT-PCR include: act-1 F, TGCTGATCGTATGCAGAAGG; act-1 R, TAGATCCTCCGATCCAGACG; mlt-10 F, GGCCTTGGCAGCCGTAAC; mlt-10 R, TAAGCTCCACGGATGAGGTC; grd-1 F, CCGCTCTGCTGATATAAACCACG; grd-1 R, TCATCGCAACATTCTACCGT; ptr-11 F, AGCCGCCTATCCGGTTTATT; ptr-11 R, GACACGGGTTTCATATCCAGC; C49G7.7 F, CGGAGATCGGGAAACCCTTT; C49G7.7 R, GGGCGGCAAGGAAAGTTAAA; F18E3.12 F, CAATTTCCACCACACCAGCC; F18E3.12 R, TGAGCGCAGTTGAAATCGTTG; ugt-43 F, AGTTACCGGTCATTCTCATTTAAAGTT; ugt-43 R, TGAGTGGTAAGAGAAGAGTCACA; C15B12.8 F, TTGTGACGATCCCAGGAAGC; C15B12.8 R, GTCCGGGGAGTTGGTGTATT; F33H12.7 F, TGGATTTTTGGAACACAAACGA; F33H12.7 R, CGCACCGGAAAGGTCTACTT; str-7 F, ACGCGTTTTTCGGTTTTATCCT; str-7 R, GAGGTGGAGGAACGTGTGAA.

All imaging, unless otherwise specified, was performed by mounting worms on 2% agarose pads in 2.5 mM levamisole using the Leica THUNDER Imager system. Imaging was performed within 5 min of sample preparation. Binning measurements were carried out with a minimum population of 10 worms per replicate. grd-1p::GFP activation was defined as intestinal expression.

All statistical analyses and representations were performed in either Prism 9 (Graphpad) or Bioconductor (R). LogEC₅₀ for developmental curves was determined by fitting a variable slope sigmoidal dose-response curve to each curve. The difference between the vector and grd-1 RNAi curves was calculated using the percentage difference between the logEC₅₀ of each curve.

L3 animals treated with empty vector (EV) control or grd-1 RNAi (grd-1) were collected in paired, biologically independent, triplicate experiments. RNA was collected and purified in the same way as described above for qRT-PCR. Total purified RNA samples were sent to Azenta (Genewiz) for quality control, library preparation and mRNA sequencing. Samples were first verified for RNA integrity scores greater than eight using an Agilent Tapestation 4200. Illumina library preparation was performed using polyA selection for mRNA species. Approximately 20 million paired-end, 150 bp reads were obtained per sample.

Read filtering and quasi-alignment were performed using custom UNIX/bash shell scripts on the Mass General Brigham ERISOne Scientific Computing Linux Cluster. Reads were analyzed for quality control using FastQC (www.bioinformatics.babraham.ac.uk/projects/fastqc/) and MultiQC (Ewels et al., 2016), and filtered for adapter contamination, truncated short reads or low-quality bases using BBDuk (Grigoriev et al., 2012). Trimmed cleaned reads were then quantified against the C. elegans reference transcriptome annotation (WBcel235, Ensembl Release 105) using Salmon, correcting for sequencing and GC content bias using the command parameters ‘--seqBias’ and ‘—gcBias’, respectively (Cunningham et al., 2022; Patro et al., 2017).

All statistical analysis and visualizations were performed using the R (www.r-project.org) Bioconductor (http://www.nature.com/nmeth/journal/v12/n2/abs/nmeth.3252.html) environment. Quasi-aligned transcript quantification files for each sample were collapsed into gene-level count matrices using R package tximport (Soneson et al., 2015), and paired differential expression was calculated using R package DESeq2 (Love et al., 2014) with a design formula of ‘∼Replicate+Treatment’, where ‘Replicate’ accounts for inter-replicate batch effect variation in the paired experimental samples. Genes were considered differentially expressed with a Benjamini-Hochberg False Discovery Rate (FDR) corrected P<0.05 and an absolute log₂ transformed fold change of 1.5 (Benjamini and Hochberg, 1995).

The top 100 genes contributing to variation in principal component 1 (PC1 – explanatory for the divergence seen between EV and grd-1 RNAi) were extracted and assessed for Gene Ontology (GO) term pathway over-representation using R package clusterProfiler (Wu et al., 2021). To perform transcription factor target enrichment analyses, promoter sequences 1500 bases upstream and 500 bases downstream of the transcription start site of significantly upregulated and downregulated genes were extracted using R package TxDb.Celegans.UCSC.ens11.ensGene (https://bioconductor.org/packages/release/data/annotation/html/TxDb.Celegans.UCSC.ce11.ensGene.html). Transcription factor binding sites were retrieved from the modENCODE modMine v33 database (Contrino et al., 2012) and filtered for presence in the promoter sequences of the differentially expressed gene sets. Total binding-site ratios in both upregulated and downregulated genes were tallied (defined as the number of binding sites identified for a given transcription factor divided by all transcription factor binding sites identified) and visualized using Prism 9 (Graphpad). The significance of the enrichment in observed binding-site ratio versus expected ratio was calculated using a two-tailed exact binomial test with a 95% confidence level using R function binom_test() and FDR adjusted P values were obtained using R function p.adjust().

All figures were assembled using Adobe Illustrator, Prism 9 (Graphpad) and BioRender.

We thank Dr Christopher M. Webster, Dr Yuyao Zhang, Dr Sainan Li, Adebanjo Adedoja, Luke Murphy, Ashley Duke and Samuel Doernberg for their creative input. Core services were provided by NIH/NIDDK P30 DK135043 and NIH/NIDDK P30 DK040561. Some strains were provided by the CGC, which is funded by the NIH Office of Research Infrastructure Programs (P40OD010440) and by the National BioResource Project of Japan.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201974.reviewer-comments.pdf