The DRL-1 protein kinase functions in the intestine to modulate fat transport and growth

The vitellogenin genes (vit-1 through vit-6) are specifically expressed in the C. elegans intestine during the L4 larval to adult transition, coinciding with the onset of reproduction. This metabolic commitment can be precisely followed using a multi-copy Pvit-3::GFP or a single-copy Pvit-3::mCherry reporter transgene comprised of GFP or mCherry under the control of the vit-3 promoter, respectively (Figs 1A and S1A). Notably, the single-copy Pvit-3::mCherry transgene is highly sensitive and its expression closely resembles endogenous vit-3 expression, and thus, it is the primary vitellogenesis reporter used in this study.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. DRL-1 functions in the intestine to promote development, growth, and lipid reallocation.

Representative overlaid DIC and mCherry fluorescence images of day 1 adult animals expressing a single-copy Pvit-3::mCherry reporter (vit-3 promoter fused to mCherry) in (A) wild-type or drl-1 mutant animals or (B) mKate2::3xFLAG::AID::drl-1 animals following intestinal (Pges-1::TIR1) or hypodermal (Pcol-10::TIR1) degradation using 4 mM auxin (scale bars, 200 μm). (C) RT-qPCR analysis of endogenous vit gene expression in day 1 adult animals after whole-body or tissue-specific knockdown of drl-1 by RNAi. (D) Body size and (E) growth rate of wild-type, drl-1(rhd109), or the indicated drl-1(rhd109) tissue-specific rescue strains (2 independent lines each). All strains contain mgIs70 and the data are presented as (D) the median and interquartile range (*, P < 0.0001, one-way ANOVA) or (E) the mean +/− SEM of 3 independent experiments. Raw data underlying panels C, D, and E can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002320.g001

We previously performed an RNAi screen for factors that are required for proper vitellogenin expression at the onset of adulthood and identified the drl-1 gene as essential for robust Pvit-3::GFP expression [14]. The drl-1 gene encodes a protein kinase with highest similarity to the mammalian mitogen-activated protein kinase kinase kinase 3 (MAP3K3/MEKK3) protein [17]. To eliminate the possibility that drl-1 RNAi produces off-target effects that impairs vitellogenesis, we generated several drl-1 genetic mutants, which are all likely full loss-of-function alleles, using CRISPR/Cas9 gene editing and assessed Pvit-3 reporter expression. Indeed, loss of drl-1 resulted in a dramatic reduction in vitellogenin reporter expression (Figs 1A, S1A and S1B). These findings are consistent with the observation that knockdown of drl-1 reduces intestinal lipid stores [17], which could impair lipoprotein synthesis and assembly. Vitellogenin gene expression is regulated through non-cell-autonomous and cell-autonomous mechanisms via hypodermal, germline, and intestinal regulators [25], which prompted us to test where drl-1 functions in the worm to regulate vit gene expression. After introduction of an auxin-inducible degron (AID) tag into the endogenous drl-1 locus using CRISPR/Cas9 genome editing [26,27], we performed tissue-specific DRL-1 protein depletion in hypodermal or intestinal cells and assessed Pvit-3::mCherry reporter expression. DRL-1 depletion in the intestine, but not the hypodermis, markedly impaired reporter expression (Figs 1B and S2), indicating that DRL-1 functions cell-autonomously to control vitellogenesis. Moreover, knockdown of drl-1 in the intestine, but not in the hypodermis, using tissue-specific RNAi reduced the expression of the endogenous vit genes (Fig 1C). Consistently, we were able to rescue vitellogenin reporter expression in drl-1 mutant animals with a transgene that expresses drl-1 under the control of an intestinal, but not a hypodermal, promoter (S1C Fig), indicating that intestinal drl-1 expression is sufficient to restore vitellogenin expression.

Impaired lipid homeostasis can have profound effects on organismal growth and many mutants with vitellogenesis phenotypes also display gross defects in development [25]. Thus, we inspected the drl-1 mutants for growth rate and body size phenotypes, finding severe defects (S1D, S1E and S2C Figs). Consistent with our previous observations, genetic rescue of the drl-1 mutant with intestinal, but not hypodermal, drl-1 expression strongly suppressed these developmental phenotypes (Fig 1D and 1E). Together, these data demonstrate an intestinal role for DRL-1 in the regulation of lipid allocation, growth rate, and body size.

Intestinal DRL-1 and FLR-4 interact to form a presumptive protein kinase complex

The flr-4 gene encodes a MAP kinase that is required for metabolic homeostasis and proper aging, and importantly, mutants display a diet-specific life span extension that is similar to that of drl-1 mutants [17,20], suggesting that these two MAP kinases may act together to govern lipid homeostasis, growth, and development. To investigate this possibility, we first inspected the kinase domain of DRL-1 and FLR-4, as well as several other similar MAPKs, and discovered that while DRL-1 shares significant sequence similarity in the kinase domain, it lacks several of the conserved amino acids that participate in ATP binding and catalysis (Fig 2A). To address whether DRL-1 possesses kinase activity, we used CRISPR/Cas9 to generate a DRL-1 P269S mutation, which is analogous to the FLR-4 P223S hypomorphic mutation that is positioned within the activation loop of the kinase domain and yields temperature-sensitive phenotypes [19]. Similar to the flr-4(P223S) mutant, drl-1(P269S) animals display temperature-sensitive defects in vitellogenin expression (Figs 2B and S3A). Additional mutation of E253 and G254 in the kinase domain, which are required for in vitro kinase activity [17], enhanced the temperature-sensitive vitellogenesis phenotype to similar levels as the null mutants (S3A Fig). Not only is the presumptive kinase activity of DRL-1 and FLR-4 necessary for proper vit gene expression, but it is required to suppress the expression of the β-oxidation gene ech-9 and the detoxification gene ugt-18 (Fig 2B) [17], indicating that DRL-1 and FLR-4 function broadly in regulation of intestinal metabolism.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The presumptive kinase activity of DRL-1 and FLR-4 is essential for function.

(A) An amino acid alignment (MUSCLE) of the kinase domain of the indicated MAP kinases. The key residues are indicated, including those mutated in the presumptive kinase dead mutants (DRL-1 E253, G254, P269; FLR-4 P223). (B) Expression of the indicated genes as measured by RT-qPCR in day 1 adult animals grown at 20 or 25°C. (C) Body size of the indicated strains at 20 or 25°C (*, P < 0.0001, one-way ANOVA). (B, C) The drl-1(rhd109) and flr-4(ut7) alleles are presumed to be null mutations. (D, E) Longitudinal life span assays of wild-type and presumptive kinase dead mutants reared at 25°C with FUDR on E. coli (D) OP50 or (E) HT115. Log-rank test P values are reported. Raw data underlying panels B, C, D, and E can be found in S2 Data.

https://doi.org/10.1371/journal.pbio.3002320.g002

Similar to the drl-1 null mutants, flr-4 null mutants display severely reduced developmental rates and body size defects when reared on E. coli OP50, the standard laboratory diet (S3B and S3C Fig) [20,28]. Furthermore, the presumptive kinase activity of DRL-1 and FLR-4 is required to maintain proper body size (Fig 2C), suggesting that these kinases may function in a broader signaling cascade to balance metabolic needs during development and aging. Intriguingly, the FLR-4 P223S activation loop mutant exhibits a life span extension only when reared on E. coli HT115, and not OP50, suggesting that components in the diet interface with longevity pathways downstream of FLR-4 [20]. To investigate whether DRL-1 also restricts life span in a diet-dependent manner, we performed longitudinal life span assays on the flr-4(P223S) and drl-1(P269S) mutants fed different E. coli food sources. Indeed, the presumptive DRL-1 and FLR-4 kinase dead mutants are markedly long-lived on E. coli HT115 (Fig 2D and 2E). These findings argue that the kinase activity of both FLR-4 and DRL-1 are required to restrict longevity in response to diet. Consistently, overexpression of drl-1 on an extrachromosomal array is sufficient to suppress the growth rate defects induced by simultaneous loss of both drl-1 and flr-4 (S3D Fig), suggesting that DRL-1, and its potential kinase activity, can compensate for loss of flr-4. Together, these data suggest that the kinase activity of DRL-1 and FLR-4 is required to maintain overall organismal homeostasis; however, we cannot rule out the possibility one or both proteins lack kinase activity and act as a non-catalytic partner to another protein kinase.

These observations support the intriguing possibility that DRL-1 and FLR-4 act in the same tissue (i.e., the intestine), or even potentially in a protein complex, to coordinate developmental or nutritional programs. To investigate this hypothesis, we first sought to define where FLR-4 functions, focusing on the intestinal and neuronal tissues [19,20]. Using CRISPR/Cas9 to introduce an AID tag at the endogenous flr-4 locus, we performed tissue-specific depletion experiments and assessed vitellogenin expression, body size, and growth rate. Depletion of FLR-4 in the intestine, but not in neurons, abrogated Pvit-3::mCherry expression, severely reduced body size, and dramatically slowed growth rates (Figs 3A–3C and S4A). Surprisingly, post-developmental depletion of either DRL-1 or FLR-4 reduced Pvit-3::mCherry reporter expression and decreased the body size of day 2 adults (S4B–S4F Fig), indicating that these phenotypes are not solely a consequence of impaired development and are likely driven by metabolic dysfunction. Finally, intestinal depletion of either DRL-1 or FLR-4 is sufficient to confer unique responses to different E. coli food sources, including HT115-dependent effects on growth rate and life span (S5 Fig). Taken together, these data indicate that FLR-4 acts cell-autonomously, during both larval stages and adulthood, to regulate intestinal homeostasis, possibly by directly interacting with DRL-1.

To test whether DRL-1 and FLR-4 colocalize, we used CRISPR/Cas9 gene editing to generate a strain that coexpresses HA::mGreenLantern::FLR-4 and mKate2::3xFLAG::DRL-1. Although these proteins are expressed at low levels, we detected the mGL::FLR-4 and mKate2::DRL-1 proteins at the intestinal plasma membrane (Fig 3D), consistent with overexpression studies of FLR-4 and DRL-1 [29,30]. Furthermore, reciprocal co-immunoprecipitation studies demonstrated that HA::FLR-4 and FLAG::DRL-1 interact (Fig 3E), likely forming a larger protein kinase complex. These data explain why drl-1 and flr-4 mutants have similar phenotypes and provide a mechanistic basis for their role in maintaining intestinal homeostasis via MAPK signaling.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Intestinal DRL-1 and FLR-4 function in a complex to promote growth and lipoprotein production.

(A-C) Phenotypic characterization of mKate2::3xFLAG::AID::drl-1 or mNG::3xFLAG::AID::flr-4 animals carrying the Pvit-3::mCherry reporter after tissue-specific depletion with 4 mM auxin (Pges-1::TIR1, intestinal depletion; Prgef-1::TIR1, pan-neuronal depletion; Pcol-10::TIR1, hypodermal depletion). (A) Representative overlaid DIC and mCherry fluorescence images of day 1 adults with or without auxin treatment (scale bar, 200 μm), (B) body size of day 1 adults (median and interquartile range; ns, not significant, *, P = 0.003, **, P < 1 × 10⁻⁴⁵, t test), and (C) growth rate (mean +/− SEM) of animals after tissue-specific protein depletion. (D) Colocalization of mKate2::3xFLAG::DRL-1 and 3xHA::mGL::FLR-4 in intestinal cells of glo-4(ok623) animals (scale bar, 50 μm). Panels on the right show magnified images of the outlined regions and arrowheads point to areas of strong colocalization. (E) Co-immunoprecipitation of the indicated proteins after mock, anti-FLAG, or anti-HA immunoprecipitation followed by western blotting. As a positive control, immunoprecipitations were probed with the same antibodies (far right lanes). The whole cell lysate (WCL) represents 5% of the IP input. FLAG::DRL-1 was not detectable in the WCL likely due to low levels of protein expression or poor sensitivity of the anti-FLAG antibody. The co-IP/western blot experiment was performed twice with similar results. Raw data underlying panels B and C can be found in S3 Data, and raw images for panel E can be found in S1 Raw Images.

https://doi.org/10.1371/journal.pbio.3002320.g003

The FLR-2 neuropeptide hormone antagonizes DRL-1 and FLR-4 signaling

DRL-1/FLR-4 signaling promotes organismal development, growth, and aging. Yet, it is possible that pro-growth signaling through this pathway may be dynamically tuned in response to adverse environmental or nutritional conditions to temper development. It is likely that this signaling would need to be balanced by other signaling events to maintain overall homeostasis. Thus, we reasoned that in the absence of drl-1/flr-4, other pathways may actively restrict development. To identify components of this signaling axis, we performed a forward genetic screen to isolate mutations that suppress the growth and vitellogenesis defects associated with the drl-1(rhd109) mutant. We isolated over 100 mutants and initially pursued a small pilot set for further analysis. Following backcrossing and whole genome sequencing, we identified a putative null mutation in the flr-2 gene as a likely drl-1 suppressor mutation (S1 Table).

The flr-2 gene encodes a neuronally expressed secreted protein with highest similarity to the human glycoprotein hormone subunit α2 (GPA2) of thyrostimulin [31,32], as well as weaker similarity to the α subunits of other human glycoprotein hormones (i.e., FSH, LH, and TSH). The flr-2(rhd117) mutation identified in our screen, as well as a second allele flr-2(ut5), strongly suppressed the defects in vitellogenin reporter expression observed in the drl-1 and flr-4 mutants (Figs 4A and S6A). Furthermore, while loss of drl-1 impaired the accumulation of neutral lipids as revealed by reduced Oil Red O staining [17], the drl-1; flr-2 double mutants displayed wild-type fat levels (Figs 4B and S6B). Consistent with these defects in vitellogenesis and lipid homeostasis, the drl-1 mutant animals also had a dramatically smaller brood size; however, drl-1; flr-2 double mutants produced nearly as many progeny as wild-type animals (Fig 4C). Together, these data suggest that FLR-2 opposes MAPK signaling to balance intestinal resources to maintain homeostasis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Loss of flr-2 suppresses mutations in drl-1 and flr-4 when reared on specific food sources.

(A) Representative fluorescence images of Pvit-3::mCherry reporter expression (day 1 adults; scale bar, 200 μm), (B) whole animal Oil Red O staining (day 1 adults; median and interquartile range; ns, not significant, *, P < 0.0001, one-way ANOVA), (C) brood size (mean +/− SD), (D) growth rate (mean +/− SEM), and (E) body size (day 1 adults; median and interquartile range; ns, not significant, *, P < 0.0001, one-way ANOVA) for wild-type animals and the indicated mutants grown on E. coli OP50. (D, E) Strains contain the mgIs70 transgene except for the drl-1; nhr-49(ok2165) strain. (F) The flr-2(rhd117) mutation robustly suppresses drl-1 mutant growth defects when animals are reared on E. coli OP50 (left), but not E. coli HT115 (right). Longitudinal life span assays of wild-type animals and the indicated mutants reared at 25°C with FUDR on E. coli (G) OP50 or (H) HT115. Animals were L4s at day 0 and P values (log-rank test) are reported. (A-H) The drl-1(rhd109) and flr-4(ut7) alleles were used in these studies. Raw data underlying panels B, C, D, E, F, G, and H can be found in S4 Data.

https://doi.org/10.1371/journal.pbio.3002320.g004

To assess whether FLR-2 impairs growth upon loss of drl-1, we performed growth measurements in the drl-1 and drl-1; flr-2 mutants. Indeed, both flr-2 alleles suppress the slow growth rate and small body size phenotypes displayed by the drl-1 mutant (Fig 4D and 4E). Mutations in the nuclear hormone receptor nhr-49 (orthologous to human HNF4 and PPARα), which has been previously shown to suppress drl-1 RNAi phenotypes [17], failed to suppress the vitellogenesis, growth rate, and body size defects of the drl-1(rhd109) mutant (Fig 4A, 4D and 4E). It is possible that experimental differences, such as the type of food or degree of drl-1 inactivation (RNAi versus mutant), account for these discrepancies.

Expression of the FLR-2 protein is limited to a small set of neurons in the head and tail [32], suggesting that FLR-2 functions non-cell-autonomously to regulate intestinal homeostasis. Indeed, genetic rescue of flr-2 under the control of a pan-neuronal promoter partially reverses the growth and body size phenotypes of the drl-1; flr-2 double mutants (S6C and S6D Fig), suggesting that expression of flr-2 in neurons is sufficient to slow the growth of drl-1 mutant animals. To explore whether flr-2 is developmentally regulated, we inserted a HA tag into the flr-2 locus using CRISPR/Cas9 and performed a western blot analysis of HA::FLR-2 from lysates of animals at different developmental stages. While FLR-2 is expressed throughout larval development and into adulthood, expression peaks at the L3 stage and is lowest during reproduction (S7A Fig), suggesting that flr-2 expression may be repressed at adulthood to facilitate vitellogenin production. Notably, knockdown of drl-1 or flr-4 did not alter FLR-2 levels, and constitutive overexpression of flr-2 in neurons did not dramatically suppress or enhance the body size defects exhibited by the drl-1 mutant, which together, indicate that flr-2 acts in parallel, and not downstream, of drl-1 (S7B and S7C Fig). Our data demonstrate that the FLR-2 neurohormone is a potent inhibitor of growth, development, and reproduction in the absence of active DRL-1/FLR-4 signaling.

The life span and stress responses of drl-1 and flr-4 mutants are markedly different when animals are fed E. coli HT115 bacteria compared to OP50, likely due to differences in the nutritional value of the strains [33]. Consistent with these observations, genetic mutation of drl-1, as well as intestinal depletion of AID::DRL-1, more severely impairs growth of animals reared on E. coli HT115 compared to OP50 (Figs 4F and S5A). Although mutation of flr-2 suppresses the growth defects of the drl-1(rhd109) mutant on OP50, it surprisingly fails to suppress drl-1(rhd109) growth on HT115 (Fig 4F). Similarly, the flr-2 mutation fails to suppress the longevity conferred by loss of drl-1 when animals are grown on E. coli HT115 (Fig 4G and 4H). Notably, loss of flr-2, which results in short-lived animals [32], suppresses the modest life span increase displayed by drl-1 mutant animals on OP50 (Fig 4G). Our results demonstrate that while loss of flr-2 strongly suppresses the drl-1 mutant phenotypes on OP50, it fails to yield similar results on HT115, suggesting that a HT115-specific nutritional input may be acting redundantly with FLR-2 to oppose DRL-1/FLR-4 signaling.

FLR-2 acts via the G protein–coupled receptor FSHR-1 to stimulate PKA activity

The FLR-2 protein is a secreted hormone that likely acts by binding a cell surface receptor on a distal tissue, possibly the intestine, to stimulate a signaling cascade that slows animal development. We predicted that mutations in the FLR-2 receptor may also suppress the growth and vitellogenesis defects displayed by the drl-1 mutant. Indeed, sequencing of additional drl-1 suppressor mutations identified a putative null mutation in the fshr-1 gene (S2 Table), which encodes a G protein–coupled receptor with similarity to the family of glycoprotein hormone receptors that include the TSH, FSH, and LH receptors in humans [34]. The fshr-1(rhd118) mutation, as well as the well-characterized fshr-1(ok778) allele, both suppressed the vitellogenesis defects conferred by loss of either drl-1 or flr-4 (Figs 5A and S8A). Furthermore, the fshr-1 mutations suppressed the slow growth and reduced the body size of the drl-1(rhd109) mutant (Fig 5B and 5C). The partial suppression of the body size can be attributed to the fact that flr-2 and fshr-1 single mutants are smaller than wild-type animals (S6D and S8B Figs). We reasoned that if FSHR-1 is the intestinal receptor for FLR-2, then knockdown of fshr-1 specifically in the intestine would suppress the drl-1 mutation. Indeed, intestine-specific knockdown of fshr-1 suppresses the vitellogenesis defects and small body size of drl-1 mutant animals to similar levels as systemic fshr-1 knockdown (Fig 5D and 5E). Together, these data suggest that intestinal FSHR-1 mediates the effects of FLR-2 to slow developmental rate of animals when DRL-1/FLR-4 signaling is reduced.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. The intestinal GPCR FSHR-1 opposes DRL-1/FLR-4 signaling.

(A) Representative overlaid DIC and mCherry fluorescence images (day 1 adults; scale bar, 200 μm), (B) growth rate (mean +/− SEM), and (C) body size (day 1 adults; median and interquartile range; *, P < 0.0001, one-way ANOVA) for wild-type and mutant animals. (B, C) Strains contain the mgIs70 transgene. (D) Pvit-3::mCherry reporter expression (scale bar, 200 μm) and (E) body size (median and interquartile range; *, P < 1 × 10⁻¹⁹, **, P < 1 × 10⁻³⁰, t test) of day 1 adult drl-1(rhd109) animals subjected to systemic or intestine-specific RNAi. Raw data underlying panels B, C, and E can be found in S5 Data.

https://doi.org/10.1371/journal.pbio.3002320.g005

The mammalian FSH receptor, like other glycoprotein hormone receptors, signals through heterotrimeric G proteins, primarily Gα_s, to stimulate adenylate cyclase activity, cAMP production, and PKA activation [35,36]. In C. elegans, gain-of-function mutations in gsa-1 (Gα_s) or acy-1 (adenylate cyclase) suppresses the germline defects observed in the fshr-1 mutant, providing genetic evidence that this pathway functions in the worm. Thus, we predicted that loss of gsa-1, the acy-1-4 genes (adenylate cyclase), or kin-1 (PKA) would also suppress the drl-1 mutant phenotypes if FSHR-1 couples to this signaling pathway (Fig 6A). RNAi knockdown of gsa-1, acy-4, or kin-1 reactivated vitellogenin reporter expression to different degrees in the drl-1 mutant background (Fig 6B). Since FSHR-1 functions in the intestine to oppose DRL-1 signaling, we then performed intestine-specific RNAi against gsa-1, acy-4, and kin-1 in the drl-1 mutant and measured Pvit-3::mCherry expression levels and body size (Fig 6C and 6D). Indeed, inactivation of this canonical PKA activation pathway in the intestine suppressed the drl-1 mutation to levels similar to fshr-1 knockdown. Notably, knockdown of kin-1 by RNAi failed to strongly suppress the vitellogenesis and body size defects exhibited by the drl-1 mutant. To investigate the role of KIN-1 in this pathway more rigorously, we introduced AID tags into two locations within the kin-1 locus using CRISPR/Cas9 editing, resulting in AID tagging of all the kin-1 isoforms. Depletion of KIN-1 in the intestine, but not in neurons, suppressed the vitellogenesis and body size defects resulting from drl-1 knockdown (Fig 6E and 6F), which is consistent with our intestine-specific RNAi experiments. Interestingly, hypodermal depletion of KIN-1 at the L4 stage results in a highly penetrant bursting phenotype (Fig 6G), indicating that kin-1 is required for developmental programs in the hypodermis or seam cells. While the ligand for FSHR-1 has remained elusive despite its widespread roles in immunity, stress responses, and germline development [34,37,38], our results are consistent with the possibility that FLR-2 is a ligand for FSHR-1 and induces cAMP signaling and PKA activation in the intestine. Indeed, FLR-2 was recently shown to bind to the extracellular domain of FSHR-1 in response to freeze-thaw stress [39], further supporting our conclusion that FLR-2 and FSHR-1 act in the same pathway.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. FLR-2 and FSHR-1 function through intestinal Gα_s/cAMP signaling.

(A) A diagram of how FLR-2/FSHR-1 may activate PKA via Gαs/cAMP signaling. (B) Representative overlaid DIC and mCherry fluorescence images of drl-1(rhd109) day 1 adults following systemic RNAi (scale bar, 200 μm). (C) Quantification of Pvit-3::mCherry reporter expression and (D) body size of drl-1(rhd109) animals subjected to intestine-specific RNAi (day 1 adults; median and interquartile range; ns, not significant, *, P < 0.0001, one-way ANOVA). (C, D) The intestine-specific RNAi was performed for one generation apart from kin-1, which was performed for two generations. (E) Quantification of Pvit-3::mCherry reporter expression and (F) body size of kin-1::AID animals subjected to control or drl-1 RNAi in the absence or presence of 4 mM auxin (median and interquartile range; n.s., not significant, *, P < 0. 0005, **, P < 0.0001, t test). KIN-1 was either depleted in the intestine (Pges-1::TIR1) or in neurons (Prgef-1::TIR1) using a 24-hour auxin treatment that was initiated at the L4 stage. (G) Hypodermal depletion of KIN-1 at the L4 stage is lethal and animals do not reach adulthood (scale bar, 500 μm). Raw data underlying panels C, D, E, and F can be found in S6 Data.

https://doi.org/10.1371/journal.pbio.3002320.g006

FLR-2 and DRL-1/FLR-4 inversely regulate p38 signaling to tune development

It is possible that FLR-2/FSHR-1/PKA signaling functions in parallel to DRL-1/FLR-4 to differentially regulate a core developmental pathway. Intriguingly, mutations in components of the p38 MAPK pathway, including tir-1 (orthologue of the human TIR domain protein SARM1), nsy-1 (MAPKKK), sek-1 (MAPKK), or pmk-1 (MAPK), suppresses the increased life span of flr-4 and drl-1 mutants [18,20]. The p38/PMK-1 pathway functions broadly in innate immunity [21,22], response to oxidative stress [23], development [40–42], and longevity [43]. Moreover, loss of flr-4 or drl-1 results in hyperphosphorylation of PMK-1 [18,20], which may promote slower developmental rates [41]. Consistent with these previous observations, our loss-of-function mutations in drl-1 or flr-4, as well as intestinal depletion of AID::DRL-1 or AID::FLR-4 with auxin, increased the levels of active, phosphorylated PMK-1 (S9 Fig).

We hypothesized that hyperactivation of p38/PMK-1 may underlie the reduced developmental rates, small body sizes, and impaired vitellogenin production of the drl-1 and flr-4 mutants. Indeed, knockdown of the p38 pathway components tir-1, nsy-1, sek-1, or pmk-1 by RNAi restored Pvit-3:mCherry expression and increased the body size of drl-1 mutant animals to varying degrees (S10A and S10B Fig). Moreover, intestine-specific knockdown of pmk-1 partially suppressed the vitellogenesis defects, small body size, and slow growth of the drl-1 mutant (Fig 7A–7C). Intestinal knockdown of pmk-1 also partially suppressed the body size defects of animals depleted of intestinal FLR-4 (S10C Fig). Importantly, loss of flr-2 was significantly more effective in suppressing DRL-1 depletion than a null mutation in either pmk-1 or tir-1 (Fig 7D), suggesting that FLR-2/FSHR-1 may function through multiple parallel pathways to modulate growth and lipid homeostasis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. DRL-1 and FLR-2/FSHR-1 signaling balance p38/PMK-1 activity to promote growth and development.

(A) Representative overlaid DIC and mCherry fluorescence images of day 1 adults (scale bar, 200 μm), (B) body size (day 1 adults; median and interquartile range; *, P = 3 × 10⁻³⁶, t test), and (C) growth rate (mean +/− SEM) of drl-1(rhd109) animals subjected to intestine-specific RNAi. (D) Body size of AID::drl-1 day 1 adult animals following intestinal depletion of DRL-1 with 4 mM auxin (median and interquartile range; *, P < 0.0001, one-way ANOVA). (E) Body size of wild-type and the indicated mutants after control or fshr-1 RNAi (median and interquartile range; *, P < 0.01, **, P < 0.0001, t test). These mutations have been previously shown to activate p38/PMK-1 signaling. (F) Representative images of TIR-1::wrmScarlet localization in posterior intestinal cells (scale bar, 50 μm). The red channel shows the TIR-1 puncta (left panels), the green/red overlay shows yellow puncta that result from intestinal autofluorescence (right panels), and wild-type animals lacking TIR-1::wrmScarlet are included as a negative control (top panels). Magnifications of the boxed areas are displayed in the bottom images. Quantification of the (G) total number, (H) absolute size, and (I) relative size of TIR-1 puncta in day 1 adult wild-type and drl-1(rhd109) mutant animals (mean +/− SD; ns, not significant, *, P < 0.05, **, P < 0.01, ***, P < 0.0001, one-way ANOVA). Raw data underlying panels B, C, D, E, G, H, and I can be found in S7 Data.

https://doi.org/10.1371/journal.pbio.3002320.g007

Although our genetic data argue that FLR-2/FSHR-1 activates the p38 pathway, the site of this regulation is unknown. The MAP3K gain-of-function (gf) mutation, nsy-1(ums8), hyperactivates downstream p38 signaling, resulting in developmental delay and small body size [41]. Using this mutant, we tested whether fshr-1 functions genetically upstream of nsy-1. While fshr-1 RNAi strongly suppresses the drl-1(rhd109) mutation, it fails to suppress the small body size of the nsy-1 gf mutant (Fig 7E), indicating that FSHR-1 likely functions upstream of NSY-1 to regulate development. Moreover, other genetic perturbations that induce p38 hyperactivation include mutations in the neuronal developmental regulator olrn-1 or the intestinal pseudokinase nipi-3 [40,42,44]; however, knockdown of fshr-1 failed to suppress the body size defects caused by the olrn-1 or nipi-3 mutations, indicating that these pathways function independently of fshr-1 (Fig 7E).

Given that FSHR-1 functions upstream of NSY-1, we investigated whether the activity of TIR-1, the C. elegans orthologue of SARM1 (sterile alpha and TIR motif-containing 1) and upstream regulator of the p38 pathway [45,46], is modified by loss of drl-1. TIR-1 activation is triggered by a stress-induced phase transition that promotes protein oligomerization, NAD⁺ glycohydrolase activity, and stimulation of the downstream NSY-1/SEK-1/PMK-1 pathway [24,47]. Using a strain expressing TIR-1::wrmScarlet [24], we tested whether loss of drl-1 could enhance TIR-1 phase transition, which is visible as intestinal puncta that are distinct from the autofluorescent gut granules. While mutation of drl-1 stimulated a reduction in the number of TIR-1::wrmScarlet puncta, the size of the puncta was markedly increased, suggesting that loss of drl-1 likely induces TIR-1::wrmScarlet phase transition (Fig 7F–7I). This drl-1-induced oligomerization of TIR-1 was suppressed by loss of flr-2. Together, these data suggest that DRL-1/FLR-4 and FLR-2/FSHR-1 exert opposing effects on TIR-1 to govern the activity of p38 signaling and animal development.

Our results suggest that FLR-2/FSHR-1 signaling stimulates p38 activity in the absence of drl-1/flr-4; however, it remains unclear whether PKA mediates these effects downstream of FLR-2. A gain-of-function mutation in Gα_s, gsa-1(ce81), or a loss-of-function mutation in the inhibitory regulator of PKA, kin-2(ce179), hyperactivates PKA signaling [48] and reduces body size. This body size defect is partially suppressed by pmk-1 RNAi (S11A Fig), suggesting that PKA functions either upstream or in parallel to PMK-1. To directly test whether PKA activation stimulates p38, we measured TIR-1 oligomerization and PMK-1 phosphorylation in the PKA hyperactivation mutants, finding that independent activation of PKA failed to induce TIR-1 phase transition or PMK-1 phosphorylation (S11B–S11D Fig). While it is possible that PKA only acts on the p38 pathway when drl-1 or flr-4 is lost, it is more likely that FLR-2 stimulates a PKA-independent parallel pathway, possibly mediated by an alternative G protein signaling cascade, to modulate the activity of TIR-1.

While the transcription factors that act downstream of p38 signaling to regulate stress responses have been well studied in C. elegans [49], those that function in p38-regulated development are not well understood. Thus, we performed a small-scale RNAi screen to identify transcriptional regulators that function downstream of drl-1 to regulate development. Using the Pvit-3::mCherry reporter, we first assessed whether knockdown of each transcription factor could suppress the vitellogenesis defects of the drl-1 mutant. Interestingly, knockdown of pha-4, a FoxA transcription factor, partially suppressed the vitellogenesis and body size defects conferred by the drl-1 mutation (Fig 8A and 8B). PHA-4 functions broadly in C. elegans development but also has a distinct role in promoting longevity in response to dietary restriction [50]. Furthermore, pha-4 is required for the life span increase that results from loss of drl-1 or flr-4 [17,20]. Thus, we reasoned that DRL-1/FLR-4 may impact the localization of PHA-4, which we assessed using a strain that expresses an endogenously tagged PHA-4::GFP to avoid potential artifacts of PHA-4 overexpression. Indeed, depletion of DRL-1 stimulated the accumulation of PHA-4::GFP protein in the nucleus of intestinal cells (Fig 8C and 8D). Moreover, the PHA-4 nuclear accumulation was dependent on pmk-1 regardless of the food source, suggesting that DRL-1, and downstream p38/PMK-1 signaling, may regulate the nuclear translocation of PHA-4/FOXA (Figs 8C, 8D, S12A and S12B). It is possible that PMK-1 promotes the nuclear accumulation of PHA-4, where it directly represses the expression of specific developmental genes. In support of this hypothesis, PHA-4 is found at the promoters of the vit genes at low levels specifically at adulthood (S12C Fig), possibly mediated by the basal activity of PMK-1. Dynamic nuclear translocation of PHA-4 has not yet been observed in C. elegans [50,51]; however, these previous studies used PHA-4 overexpression transgenes, while our experiments employed a CRISPR/Cas9-based GFP knock-in at the pha-4 locus, which tags all isoforms. Notably, the human FOXA proteins show different abilities to dynamically shuttle between the nucleus and cytoplasm [52]. Together, our data demonstrate that DRL-1 signaling governs TIR-1 phase transition to module p38 activity and PHA-4 localization in the intestine to tune the developmental rate of C. elegans.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. PHA-4/FOXA suppresses growth upon loss of drl-1.

(A) Representative fluorescence images of Pvit-3::mCherry reporter expression (F1 animals; scale bar, 200 μm) and (B) body size (P0 animals; median and interquartile range; *, P = 7 × 10⁻¹⁷, t test) of drl-1(rhd109) day 1 adults after knockdown of pha-4 by RNAi. (B) This control RNAi data is also presented in Fig 5E. (C) GFP fluorescence images (magnifications of the outlined regions are shown below the primary image with white arrowheads indicating intestinal nuclei; scale bar, 100 μm) and (D) quantification (median and interquartile range; *, P < 0.05, **, P < 0.0001, one-way ANOVA) of PHA-4::GFP nuclear localization after intestine-specific depletion of AID::DRL-1 using 4 mM auxin in wild-type or pmk-1(km25) animals reared on E. coli OP50. (E) A model illustrating the opposing effects of DRL-1/FLR-4 and FLR-2/FSHR-1 on p38-mediated growth. Raw data underlying panels B and D can be found in S8 Data.

https://doi.org/10.1371/journal.pbio.3002320.g008

C. elegans strains

C. elegans strains were cultured on NGM media seeded with E. coli OP50 or HT115(DE3) [71]. Animals were reared at 20°C unless specified otherwise. For auxin-inducible degradation experiments, embryos were transferred to plates containing 4 mM naphthaleneacetic acid (K-NAA, PhytoTech) and grown to adulthood. All strains used in this study are listed in S1 File.

Generation and imaging of transgenic animals

Strains carrying the high-copy mgIs70[Pvit-3::GFP] transgene or the single-copy rhdSi42[Pvit-3::mCherry] transgene have been previously described [14]. The drl-1 rescue constructs were generated by fusing the col-10 promoter (chromosome V: 9,166,416–9,165,291; WS284) or the vha-6 promoter (chromosome II: 11,439,355–11,438,422; WS284) to the drl-1 cDNA (1,770 bp of coding sequence with 141 bp of 3′ UTR) via Gibson assembly to generate plasmids pRD141 and pRD142, respectively [14,72]. The resulting plasmids were microinjected into drl-1(rhd109); mgIs70[Pvit-3::GFP] animals at 20 ng/μl, along with 2.5 ng/μl pCFJ90(Pmyo-2::mCherry) and 77.5 ng/μl of 2-Log DNA ladder (New England BioLabs), to generate 2 independent strains expressing Ex[Pcol-10::mCherry::his-58::SL2::drl-1 cDNA] (DLS515, DLS516) and Ex[Pvha-6::mCherry::his-58::SL2::drl-1 cDNA] (DLS513, DLS514). For the pan-neuronal flr-2 rescue transgene, the sng-1 promoter (chromosome X: 7,325,641–7,327,607; WS284) was fused to the flr-2 gene (including the 3′ UTR) via Gibson assembly to generate the pRD157 plasmid. This plasmid, which is derived from the MosSCI-compatible pCFJ151 plasmid, was microinjected into EG6699 to generate the single-copy integrant rhdSi46[Psng-1::mCherry::his-58::SL2::flr-2] as previously described [73]. All strains carrying mgIs70 or rhdSi42 were imaged on a Nikon SMZ-18 Stereo microscope equipped with a DS-Qi2 monochrome camera.

CRISPR/Cas9 gene editing

Generation of the drl-1 deletion alleles (rhd109 and rhd110) were generated using the pha-1 co-conversion approach, as previous described [74]. All additional edits were performed by microinjection of Cas9::crRNA:tracrRNA complexes (Integrated DNA Technologies) into the germlines of C. elegans animals as previously described [75]. Large dsDNA donor molecules with approximately 40 bp homology arms on each end were prepared by PCR using Q5 DNA Polymerase (New England BioLabs) and purified using HighPrep PCR Clean-up beads (MagBio) per the manufacturers’ instructions. The DNA repair templates were melted and reannealed prior to microinjection [75]. The PCR templates used to generate the dsDNA donor molecules were pRD156, pBluescript II(linker::mKate2::TEV::linker::3xFLAG::AID); pRD160, pBluescript II(AviTag::linker::2xTEV::linker::3xHA::linker::mNeonGreen::TEV::linker::3xFLAG::AID); and pRD174, pBluescript II(AviTag::linker::2xTEV::linker::3xHA::linker::mGreenLantern::TEV::linker::3xFLAG::AID). To generate missense mutations using CRISPR/Cas9 gene editing, single-stranded oligodeoxynucleotides were used as donor molecules [75]. All CRISPR crRNA guide sequences are listed in S1 File.

RNAi experiments

E. coli HT115(DE3) strains carrying the control L4440 plasmid or individual RNAi plasmids for gene expression knockdown were grown for approximately 16 hours in Luria–Bertani medium containing ampicillin (50 μg/ml), concentrated by 20 to 30× via centrifugation, and seeded on NGM plates containing 5 mM isopropyl-β-D-thiogalactoside (IPTG) and 50 μg/ml ampicillin. Plates seeded with RNAi bacteria were maintained at room temperature overnight to induce expression of the dsRNAs. All RNAi clones were selected from the Ahringer or Ahringer Supplemental RNAi library and confirmed by Sanger sequencing, with the exception of nsy-1, which has been described previously [41], and sek-1 and pmk-1, which were generated by cloning cDNA fragments into the pL4440 vector using standard techniques. We selected 4 different kin-1 RNAi clones from the Ahringer library to initially test for their ability to suppress the drl-1 mutant phenotypes. Three of these RNAi clones caused lethality or severe developmental phenotypes, while the weakest clone permitted growth to adulthood but produced modest drl-1 suppression phenotypes, likely due to incomplete knockdown of the kin-1 transcript. Synchronized C. elegans L1 larvae were generated by bleaching gravid animals to liberate embryos followed by overnight incubation in M9 buffer. Synchronized L1s were dropped on RNAi plates, grown at 20°C until they were day 1 adults (72 to 120 hours), and processed for imaging.

Growth, life span, and brood size assays

Animals were grown on their respective E. coli food sources (OP50 or HT115) for at least 2 generations without starvation prior to being assayed growth, life span, and brood size. For developmental growth rate assays, 100 to 200 eggs were picked from plates with freshly laid embryos (16 hours or less) to new plates. Animals were scored every 24 hours for 7 days for the presence of gravid adults, which were promptly removed and recorded as having reached adulthood. For RNAi-treated animals, synchronized L1s were dropped on RNAi plates grown for up to 7 days and similarly scored. For developmental experiments employing auxin-inducible degradation of AID::DRL-1 or AID::FLR-4, freshly laid embryos were picked to plates containing 4 mM naphthaleneacetic acid and worm growth was scored as described above.

For body size measurements, L4 animals were picked to fresh plates and imaged 24 hours later on a Nikon SMZ-18 Stereo microscope equipped with a DS-Qi2 monochrome camera. Using the Fiji software [76], animals were outlined by hand and the number of pixels were measured, which was converted to a square micron value based on the known imaging settings. Body size data in mm² are presented as the median with the interquartile range. A one-way ANOVA with a Bonferroni correction for multiple testing was performed to determine whether samples were significantly different.

Life span assays were performed at 25°C in the presence of 50 μM FUDR unless otherwise noted. Briefly, well-fed L4 animals were picked to FUDR-containing plates and animals (30 animals/plate, 120 to 150 total animals) were maintained on uncontaminated plates and scored daily for survival as previously described [16]. A log-rank test was applied to determine whether survival curves were significantly different.

For brood sizes experiments, L4 animals (N2, n = 11; drl-1(rhd109), n = 23; drl-1(rhd109); flr-2(rhd117), n = 11) were picked to individual plates and transferred daily throughout the reproductive period. Progeny were counted as L3 or L4 animals and data are reported as the mean +/− the standard deviation.

Oil Red O staining

Approximately 75 L4 animals were picked to new plates, harvested 24 hours later as day 1 adults in S buffer, washed, fixed with 60% isopropanol, and stained for 7 hours with 0.3% Oil Red O as previously described [16]. Animals were mounted on 2% agarose pads and imaged with a Nikon SMZ-18 Stereo microscope equipped with a DS-Qi2 monochrome camera for intensity analyses or a Nikon Ti2 widefield microscope equipped with a DS-Fi3 for representative color images. Quantification of Oil Red O staining was performed in ImageJ by manually outlining worms and determining the mean gray value in the worm area. Each value was subtracted from 65,536, the maximum gray value for 16-bit images. The data were plotted using Prism 9 and a one-way ANOVA with a Bonferroni correction for multiple testing was performed to calculate P values.

Quantitative PCR

L1 animals were synchronized by bleaching and grown to the first day of adulthood prior to being harvested in M9 buffer, washed, and flash frozen. The total RNA was isolated using Trizol (Thermo Fisher), followed by chloroform extraction and precipitation with isopropanol. DNase treatment, followed by cDNA synthesis with oligo(dT) priming, was performed using the SuperScript IV VILO Master Mix with ezDNase Kit according to the manufacturer’s instructions (Thermo Fisher). Quantitative PCR was performed exactly as previously described [16]. All primer sequences are listed in S1 File. Expression data are presented as the mean fold change relative to a control (N2 for mutant analyses, L4440 for RNAi analyses) with the standard error of the mean (SEM) reported for 3 independent experiments.

Western blot analyses

Animals were synchronized by bleaching and the resulting L1s were grown to day 1 adults. The animals were harvested in M9 buffer, washed 3 times, and the worm pellets were snap frozen in liquid nitrogen. The pellets were resuspended in an equal volume of 2× RIPA buffer (Cell Signaling Technology) containing a 2× Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher) and homogenized for approximately 15 seconds with disposable pellet pestle (Thermo Fisher) mounted in a cordless drill (Dewalt). Samples (approximately 200 μL) were then subjected to sonication (30-second on/off cycles, 10 cycles) using a Bioruptor Pico sonication instrument (Diagenode). Whole cell lysates were cleared by centrifugation and protein concentrations were determined using the DC Protein Assay (BioRad) according to the manufacturer’s instructions. Equal amounts of protein for each sample (approximately 50 μg) were resolved by SDS-PAGE, transferred to a PVDF membrane, blocked in 5% nonfat dry milk (BioRad), and probed with either anti-phospho-p38 MAPK (Thr180/Tyr182; #9211, Cell Signaling Technology), anti-PMK-1 [77], or anti-Actin antibodies (ab3280, Abcam).

For co-immunoprecipitation (co-IP) experiments, approximately 50,000 day 1 adults were harvested, homogenized, and resuspended in lysis buffer exactly as previously described [16]. Worm lysates were sonicated with the Bioruptor Pico sonication instrument (Diagenode) and cleared by centrifugation. For co-IP of mKate2::3xFLAG::DRL-1 with 3xHA::mGL::FLR-4 from strain DLS781, the lysate was split in 3 equal parts and subjected to a mock IP, anti-HA IP (2 μg of 3F10; 11867423001, Sigma), or anti-FLAG IP (5 μg of M2; F1804, Sigma). Prior to the immunoprecipitations antibodies were bound to Protein G Dynabeads (Invitrogen) with the mock IP sample lacking antibody. Western blotting of immunoprecipitated proteins was performed as described above using anti-FLAG (F1804, Sigma) or anti-HA (11867423001, Sigma) antibodies.

EMS mutagenesis and identification of causative mutations

Mutagenesis of drl-1(rhd109); mgIs70 (DLS364) animals with ethyl methanesulfonate (EMS, Sigma-Aldrich) was performed exactly as previously described [14]. A total of 50,000 haploid genomes were screened across 6 genetically distinct pools. Suppressor mutants (approximately 130 viable strains) were selected based on increased growth rate and/or mgIs70 [Pvit-3::GFP] expression relative to the parental strain.

Suppressor strains were selected and backcrossed to the DLS364 strain 2 times. F2 recombinants from the second backcross that displayed the suppression phenotype were singled, the resulting plates were allowed to starve, the animals across all individual plates were pooled, and genomic DNA was prepared from the pool using the Qiagen Gentra Puregene Tissue Kit [78]. Whole genome sequencing libraries were prepared using the TruSeq DNA PCR-Free kit (Illumina) and sequenced on an Illumina HiSeq 4000 instrument according to the manufacturer’s instructions. Identification of candidate suppressor mutations was performed as previously described [79] using in-house scripts.

Reporter imaging and quantification

To measure Pvit-3::mCherry fluorescence, strains carrying the rhdSi42 reporter were grown to the L4 stage, picked to new plates, mounted 24 hours later as day 1 adults with 25 mM levamisole on a 2% agarose pad, and imaged with a Nikon SMZ-18 Stereo microscope equipped with a DS-Qi2 monochrome camera. Worm bodies were traced in the brightfield channel and mean intensities (gray values) were calculated in the mCherry channel using ImageJ. The data were plotted in Prism 9 and a one-way ANOVA with a Bonferroni correction was performed to calculate P values.

For TIR-1::wrmScarlet imaging, animals were reared on glo-3 RNAi for several generations using the RNAi-competent E. coli OP50(xu363) strain [80]. Knockdown of glo-3 reduces intestinal autofluorescence and facilitates TIR-1 imaging [24]. At the L4 stage, animals were picked to fresh glo-3 RNAi plates, mounted on a 2% agarose pad 24 hours later, and imaged with a CFI Apo 60X Oil TIRF objective on a Nikon Ti2 widefield microscope equipped with a Hamamatsu ORCA-Fusion BT camera. The mKate2::drl-1; mGreenLantern::flr-4; glo-4(ok623) CRISPR knock-in animals were imaged at 60×, and animals expressing the TIR1::F2A::mTagBFP2::AID::NLS transgene were imaged at 20× on the Nikon Ti2 microscope.

TIR-1 puncta were measured in the posterior intestine since these 2 to 4 epithelial cells are not obstructed by other tissues. Z-stacks were cropped to a depth of 3 to 4 μm to capture planes where the intestine can be clearly visualized. All Z-stacks were denoised, deconvoluted, and compressed into a single image using the Nikon NIS-Elements analysis software. ROIs were hand drawn around the 2 to 4 posterior-most intestinal cells and the TIR-1 puncta were quantified using the object counts feature in NIS-Elements, which separated TIR-1::wrmScarlet puncta from the autofluorescent gut granules that appear in both the FITC and mCherry channels. TIR-1 puncta counts, puncta area, and puncta area relative to the ROI area were plotted in Prism 9, and a one-way ANOVA with a Bonferroni correction was performed to calculate P values.

For PHA-4::EGFP imaging, reSi5; mKate2::3xFLAG::AID::drl-1; pha-4::EGFP animals were reared on OP50(xu363) or HT115(DE3) glo-3 RNAi bacteria in the presence of 4 mM auxin for multiple generations to reduce gut autofluorescence and deplete the DRL-1 protein. L4 animals were picked to new plates and imaged 24 hours later as day 1 adults with a 10× objective on the Nikon Ti2 microscope. All Z-stacks were denoised and deconvoluted using the Nikon NIS-Elements analysis software and a single plane was selected for analysis based upon which had the best resolved intestinal nuclei. The nuclei of the 2 visible anterior and posterior-most intestinal cells were outlined by hand in ImageJ and GFP fluorescent signal intensity measurements were obtained by recording the mean gray value of each nucleus. The mean intensity values for each anterior nucleus for the auxin-treated animals (approximately 10 individuals/sample) were divided by the mean intensity value of all the anterior nuclei for the minus auxin control animals. An identical analysis was performed for the posterior nuclei, and the data were plotted in Prism 9 and a one-way ANOVA with a Bonferroni correction was performed.