ABSTRACT
Long-term survival of an animal species depends on development being robust to environmental variations and climate changes. We used C. elegans to study how mechanisms that sense environmental changes trigger adaptive responses that ensure animals develop properly. In water, the nervous system induces an adaptive response that reinforces vulval development through an unknown backup signal for vulval induction. This response involves the heterotrimeric G-protein EGL-30//Gαq acting in motor neurons. It also requires body-wall muscle, which is excited by EGL-30-stimulated synaptic transmission, suggesting a behavioral function of neurons induces backup signal production from muscle. We now report that increased acetylcholine during liquid growth activates an EGL-30-Rho pathway, distinct from the synaptic transmission pathway, that increases Wnt production from motor neurons. We also provide evidence that this neuronal Wnt contributes to EGL-30-stimulated vulval development, with muscle producing a parallel developmental signal. As diverse sensory modalities stimulate motor neurons via acetylcholine, this mechanism enables broad sensory perception to enhance Wnt-dependent development. Thus, sensory perception improves animal fitness by activating distinct neuronal functions that trigger adaptive changes in both behavior and developmental processes.
INTRODUCTION
In order for animal species to thrive, development must be responsive to environmental changes. Variations in temperature, precipitation, salinity, chemicals and mechanical loads can be extreme, and can adversely affect developmental pathways. A poorly understood aspect of how development is robust to these variations relates to adaptive changes triggered by sensory input. For example, in Drosophila, although heat exacerbates protein folding damage caused by genetic variation, it induces chaperones such as HSP90 to refold these proteins and insulate developmental pathways from damage (Prodromou, 2016; Rutherford and Lindquist, 1998). However, how stress caused by other environmental variations is offset by adaptive changes is largely unknown.
C. elegans vulval development is a good model for studying this issue as the nematode has survived for 30 million years in different climates (Braendle et al., 2008; Cutter, 2008), and the vulva, which allows mating and egg laying, develops nearly perfectly under different environmental conditions (Braendle and Felix, 2008). Vulval development begins with six ‘Pn.p’ cells (P3.p-P8.p) along the anteroposterior axis becoming vulval progenitors (Fig. 1). In P3.p and P4.p, this process is mainly driven by the Wnts CWN-1 and EGL-20, and, to a lesser extent, CWN-2, which are expressed in different locations (Gleason et al., 2006; Harterink et al., 2011; Hayashi et al., 2009; Minor et al., 2013; Modzelewska et al., 2013; Myers and Greenwald, 2007; Penigault and Felix, 2011; Song et al., 2010; Whangbo and Kenyon, 1999; Yamamoto et al., 2011). P5.p-P8.p progenitor generation is regulated by CWN-1, EGL-20 and LIN-3/EGF from the mid-body anchor cell (Eisenmann et al., 1998; Hill and Sternberg, 1992; Myers and Greenwald, 2007). P4.p-P8.p always become vulval progenitors, while in 50% of animals, P3.p receives insufficient Wnt and adopts the non-progenitor 4° fate where it fuses with the hyp7 hypodermal syncytium without dividing (Eisenmann et al., 1998). Later, LIN-3 levels rise in the anchor cell and trigger vulval induction by stimulating LET-23/EGFR in P6.p, the nearest progenitor (reviewed by Shin and Reiner, 2018). P6.p adopts the 1° fate and stimulates Notch in P5.p and P7.p, which cooperates with low levels of EGFR signaling to induce 2° fates in these cells. 1°- and 2°-fated cells generate eight and seven progeny, respectively. Uninduced progenitors (P3.p, P4.p and P8.p) acquire the 3° fate, divide once and fuse with hyp7. Generation of excess uninduced progenitors is important, as some stresses shift 1° fate induction to a progenitor other than P6.p (Braendle and Felix, 2008).
C. elegans vulval induction highlighting major Wnt- and EGF/LIN-3-producing cells. During the L2 stage, mainly CWN-1 and EGL-20 promote vulval progenitor fates in P3.p-P8.p, with LIN-3 cooperation (P5.p-P8.p). In 50% of animals, P3.p does not become a progenitor and adopts the 4° fate, resulting in fusion with hyp7 without dividing. At L3 onset, LIN-3 stimulates LET-23/EGFR on P6.p to induce a 1° vulval fate, which activates LIN-12/Notch in P5.p and P7.p. Notch and low levels of EGFR signaling induce 2° vulval fates in P5.p and P7.p. Multiple Wnts, especially CWN-1, promote robustness in vulval induction. Wnts CWN-1 (from sex myoblasts), LIN-44 and MOM-2 signal through LIN-17 and LIN-18 receptors to polarize P5.p, P7.p and their innermost progeny towards the mid-body during cell division. The Wnt-responsive POPTOP mCherry reporter is mainly active in P5.p and P7.p innermost progeny, and weakly active in P6.p descendants. Uninduced progenitors (P3.p, P4.p and P8.p) acquire the 3° fate, divide once and fuse with hyp7.
C. elegans vulval induction highlighting major Wnt- and EGF/LIN-3-producing cells. During the L2 stage, mainly CWN-1 and EGL-20 promote vulval progenitor fates in P3.p-P8.p, with LIN-3 cooperation (P5.p-P8.p). In 50% of animals, P3.p does not become a progenitor and adopts the 4° fate, resulting in fusion with hyp7 without dividing. At L3 onset, LIN-3 stimulates LET-23/EGFR on P6.p to induce a 1° vulval fate, which activates LIN-12/Notch in P5.p and P7.p. Notch and low levels of EGFR signaling induce 2° vulval fates in P5.p and P7.p. Multiple Wnts, especially CWN-1, promote robustness in vulval induction. Wnts CWN-1 (from sex myoblasts), LIN-44 and MOM-2 signal through LIN-17 and LIN-18 receptors to polarize P5.p, P7.p and their innermost progeny towards the mid-body during cell division. The Wnt-responsive POPTOP mCherry reporter is mainly active in P5.p and P7.p innermost progeny, and weakly active in P6.p descendants. Uninduced progenitors (P3.p, P4.p and P8.p) acquire the 3° fate, divide once and fuse with hyp7.
As neurons are important environmental sensors, we investigated how they may make vulval development robust. Evidence suggests they may act through behavioral and non-behavioral mechanisms. For example, neurons may constitutively produce a Wnt that helps cells such as P4.p become a vulval progenitor (Myers and Greenwald, 2007). In addition, axons of the canal-associated neurons sequester extracellular EGL-20/Wnt to refine the EGL-20 gradient and modulate vulval induction and polarity (Modzelewska et al., 2013). Neither function has been linked to muscle, the target of neuronal behavioral activity, or to secretion of neurotransmitters, a sign of sensory-dependent neuronal activation.
By contrast, certain environmental conditions have been suggested to promote vulval development by changing locomotor behavior through activation of the heterotrimeric G-protein EGL-30/Gαq in neurons (Moghal et al., 2003). In that work, constitutively active EGL-30 suppressed the underinduced vulval phenotypes (incomplete vulval development) of some EGFR pathway mutants, suggesting EGL-30 stimulates a backup signal to EGF. The behavioral model stemmed from several observations. First, EGL-30 stimulates vulval induction from ventral cord motor neurons, from where it also promotes synaptic release of acetylcholine (ACh) at neuromuscular junctions to excite muscle (Koelle, 2016; Moghal et al., 2003). Second, body-wall muscle is required for EGL-30 to promote vulval induction, as well as muscle-expressed EGL-19 calcium channels, which enhance muscle excitation (Jospin et al., 2002; Moghal et al., 2003). Third, transferring the same egfr/let-23 pathway mutants from agar to water induces hyperactive swimming and suppresses the underinduced phenotypes in an EGL-30-dependent manner (Moghal et al., 2003). This behavioral change involves increased proprioceptive activation of B-type ventral cord motor neurons, which occurs through faster bending of anterior segments in the reduced viscosity of water. This entrains the rhythms of central pattern generators into a coordinated hyperactive swimming gait (Fang-Yen et al., 2010; Fouad et al., 2018; Wen et al., 2012; Xu et al., 2018). More frequent excitation of muscle during swimming was thought to increase its production of the backup signal for vulval induction. This signal may involve Wnt as mutation of the Wnt effector bar-1/β-catenin suppresses the inductive activity of EGL-30 and water (Moghal et al., 2003), and Wnt signaling overactivation rescues some egfr/let-23 pathway mutations (Gleason et al., 2002; Modzelewska et al., 2013).
How the developmental function of EGL-30 is naturally regulated by the environment and how it affects vulval development are unknown. The proprioceptive mechanoreceptors in the motor neurons have not been identified and although the behavioral activity of EGL-30 is stimulated by neurotransmitter-coupled GPCRs (Brundage et al., 1996; Chan et al., 2012; Lackner et al., 1999; Liu et al., 2007), EGL-30 is also activated independently of GPCRs (Blumer and Lanier, 2014; Papasergi et al., 2015). For example, RIC-8, a conserved EGL-30 activator, is a GPCR-independent Gαq chaperone and guanine nucleotide exchange factor, which is regulated by non-neurotransmitter signals such as growth factors (Chan et al., 2013b; Miller et al., 2000; Tall et al., 2003; Wang et al., 2011). It also is unclear whether muscle is truly downstream or parallel to neurons.
Here, we identify ACh, a neurotransmitter secreted in response to diverse sensory modalities, as a stimulator of the developmental activity of EGL-30. This activity originates within a subset of motor neurons. However, our data support ACh acting through GPCR-EGL-30-Rho-dependent stimulation of CWN-1/Wnt production from these motor neurons, rather than through the EGL-30-diacylglycerol/calcium pathway that drives synaptic transmission and muscle excitation. This mechanism differs from the behavioral model that suggested EGL-30-stimulated muscle excitation increases production of a Wnt signal from muscle (Moghal et al., 2003). We also provide evidence for a parallel function of muscle involving a non-CWN-1 signal. These findings establish a new sensory-regulated, non-behavioral role for neurons in development. Our data suggest that the nervous system increases robustness in vulval development to a broad range of stresses, and reveal that sensory perception can coordinately promote adaptive changes in developmental processes and behavior by activating divergent pathways downstream of G-proteins in neurons.
RESULTS
EGL-30 promotes Wnt signaling during vulval induction
To investigate whether EGL-30 promotes Wnt signaling during vulval development, we used the POPTOP reporter, which consists of a minimal pes-10 promoter and seven TCF-binding sites that drive mCherry expression, and responds to β-catenin and TCF-dependent Wnt signaling (Green et al., 2008; Modzelewska et al., 2013). Wnt-driven reporter activity is not detectable during vulval progenitor generation due to high basal pes-10 promoter activity at this time. However, during vulval induction, after the first and through the second divisions of P5.p-P7.p (Pn.px and Pn.pxx stages), Wnt-dependent expression is mainly detected in the innermost progeny of P5.p and P7.p, which reflects LIN-17 and LIN-18 Wnt receptor-mediated regulation of polarity in these lineages (Green et al., 2008) (Figs 1 and 2). The egl-30(tg26) gain-of-function mutation increased the frequency of reporter expression in the P6.p lineage alone (Fig. 2), suggesting EGL-30 may act on Wnt sources distinct from those affecting LIN-17 and LIN-18 signaling in the P5.p and P7.p lineages.
EGL-30 promotes Wnt-responsive POPTOP reporter activity during vulval induction. Animals harboring the POPTOP mCherry reporter were photographed in the red fluorescence channel at the Pn.pxx stage. Sample images and quantification of the percentage of animals with detectable expression in the indicated cells are shown. Scale bars: 10 µm. *P=0.01 by a two-tailed Fisher's exact test.
EGL-30 promotes Wnt-responsive POPTOP reporter activity during vulval induction. Animals harboring the POPTOP mCherry reporter were photographed in the red fluorescence channel at the Pn.pxx stage. Sample images and quantification of the percentage of animals with detectable expression in the indicated cells are shown. Scale bars: 10 µm. *P=0.01 by a two-tailed Fisher's exact test.
EGL-30-stimulated vulval induction depends on CWN-1
EGL-30 is widely expressed in the nervous system and transgenic experiments with neuronal promoters indicate EGL-30 can stimulate vulval induction from ventral cord motor neurons (Bastiani et al., 2003; Lackner et al., 1999; Moghal et al., 2003), suggesting these cells express the target Wnt. cwn-1, cwn-2 and mom-2 expression is detected in these cells (Gleason et al., 2006; Hayashi et al., 2009; Inoue et al., 2004; Modzelewska et al., 2013; Song et al., 2010). We examined the dependency of EGL-30 on these Wnts to stimulate vulval induction in let-23(sy1) mutants, which have defective induction, but not P5.p-P7.p progenitor generation (Fig. 3A). The egl-30(tg26) gain-of-function mutation increased vulval induction in all three progenitors of let-23(sy1) mutants (Fig. 3A, Table 1), indicating that EGL-30 stimulates vulval induction post-progenitor generation. Addition of the cwn-1(ok546) null mutation to the double mutant severely reduced vulval induction in P5.p-P7.p, while having little effect on their becoming progenitors (Fig. 3A, Table 1). mom-2, but not cwn-2, mutation also mildly suppressed egl-30(tg26) activity (Table 1).
EGL-30 acts through CWN-1. (A) Left panel: incidence of P5.p-P7.p becoming a vulval progenitor. Right panel: incidence of vulval induction. (B) Incidence of the 3° fate in P3.p and P4.p in egl-30 loss-of-function mutants. P-values are relative to wild type unless otherwise indicated and are calculated using two-tailed Fisher's exact tests. For single mutants: P3.p, ***P=1.5E-5 [egl-30(n686)], 2.7E-6 [egl-30(n715)], 7.3E-15 [eg-30(ad805)], 1.6E-11 [egl-20(n585)], 2.7E-18 [cwn-1(ok546)]; and for P4.p, ***P=2.7E-6 [egl-30(ad805)], 4.7E-7 [egl-20(n585)] and 5.1E-58 [cwn-1(ok546)], and *P=0.006 [egl-30(n715)]. For comparisons between egl-20(n585) and egl-30(ad805); egl-20(n585), ***P=2.6E-9 (P4.p) and *P=0.05 (P3.p). (C) Underinduced vulval phenotypes in L4-stage mutants. Scale bars: 10 µm.
EGL-30 acts through CWN-1. (A) Left panel: incidence of P5.p-P7.p becoming a vulval progenitor. Right panel: incidence of vulval induction. (B) Incidence of the 3° fate in P3.p and P4.p in egl-30 loss-of-function mutants. P-values are relative to wild type unless otherwise indicated and are calculated using two-tailed Fisher's exact tests. For single mutants: P3.p, ***P=1.5E-5 [egl-30(n686)], 2.7E-6 [egl-30(n715)], 7.3E-15 [eg-30(ad805)], 1.6E-11 [egl-20(n585)], 2.7E-18 [cwn-1(ok546)]; and for P4.p, ***P=2.7E-6 [egl-30(ad805)], 4.7E-7 [egl-20(n585)] and 5.1E-58 [cwn-1(ok546)], and *P=0.006 [egl-30(n715)]. For comparisons between egl-20(n585) and egl-30(ad805); egl-20(n585), ***P=2.6E-9 (P4.p) and *P=0.05 (P3.p). (C) Underinduced vulval phenotypes in L4-stage mutants. Scale bars: 10 µm.
egl-30 and cwn-1 loss-of-function mutations are genetically similar
As EGL-30 is most dependent on CWN-1, we further investigated whether egl-30 regulates this Wnt. cwn-1 mutation severely and moderately reduces the frequencies at which P3.p and P4.p, respectively, become a vulval progenitor and adopt the 3° fate (Fig. 3B) (Myers and Greenwald, 2007; Penigault and Felix, 2011). Three hypomorphic egl-30 mutations reduced the frequency of 3° fate in P3.p, with the two stronger mutations also affecting P4.p (Fig. 3B). egl-30 null mutants could not be examined due to early lethality (Brundage et al., 1996). cwn-1 mutants are the only single Wnt gene mutants that are sometimes underinduced (Fig. 3C, Table 1) (Gleason et al., 2006). In four out of six underinduced cwn-1 mutants, the phenotype resulted from failure of an anterior Pn.p cell to become a progenitor (Table 2). In animals where P6.p adopted the 1° fate, P5.p was affected, and where 1° induction shifted to P5.p, P4.p was affected. In the other two underinduced mutants, there were sufficient vulval progenitors (Table 2), further supporting a separate role for CWN-1 in induction. Rare underinduced phenotypes were seen with two different egl-30 loss-of-function mutations, but never in wild-type animals (Fig. 3C, Table 1). In most cases, vulval induction, but not progenitor generation, was defective (Table 2). However, in one instance, 1° fate induction shifted to P5.p, and P4.p failed to adopt the 2° fate because it did not become a progenitor (animal 36), as seen in some cwn-1 mutants.
EGL-20 also promotes the 3° fate in P3.p and P4.p (Myers and Greenwald, 2007). Addition of the egl-30(ad805) mutation to egl-20(n585) mutants further reduced the 3° fate frequencies of both P3.p and P4.p. By contrast, in cwn-1 mutants, addition of the egl-30 mutation did not affect the 3° fate frequency of P4.p, and effects on P3.p 3° fate frequency could not be assessed as this frequency was already zero. These results, along with egl-30 being expressed in cwn-1-expressing ventral cord motor neurons, but not in rectal cells where EGL-20 is made (Bastiani et al., 2003; Gleason et al., 2006; Harterink et al., 2011; Hayashi et al., 2009; Lackner et al., 1999; Modzelewska et al., 2013; Whangbo and Kenyon, 1999), support egl-30 regulating cwn-1.
EGL-30 promotes vulval development from cwn-1-expressing cells
To test whether EGL-30 acts in CWN-1-producing cells, we expressed the egl-30(tg26) gain-of-function cDNA from the cwn-1 promoter. This promoter is strongly active in posterior body-wall muscle and ventral cord motor neurons, and weakly active in anterior motor neurons (Modzelewska et al., 2013), similar to other cwn-1 promoter fragments (Gleason et al., 2006; Hayashi et al., 2009). cwn-1-expressing motor neurons are cholinergic and include both A-type and proprioceptive B-type as co-expression with A-specific [unc-4 (Lickteig et al., 2001)], B-specific [acr-5 (Haspel et al., 2010)] and pan-A/B [acr-2 (Jospin et al., 2009)] reporters was observed (Fig. 4A,B). The motor neurons included V and D subtypes, as some processes remained ventral while others extended dorsally (Fig. 4C). The Pcwn-1::egl-30(tg26) transgene partially suppressed the let-23(sy1) underinduced vulval phenotype, which, by contrast, was not suppressed when egl-30(tg26) cDNA expression was driven by the egl-20 promoter (Table 1). The Pcwn-1::egl-30(tg26) transgene also rescued the P3.p and P4.p 3° fate defects in egl-30(ad805) loss-of-function mutants (Fig. 5A). Thus, cwn-1 upregulation could explain why transgenic expression of activated EGL-30 in A-type motor neurons stimulates vulval induction (Moghal et al., 2003).
cwn-1 is expressed in A- and B-type cholinergic motor neurons. (A) L4-stage animals co-expressing Pcwn-1::DsRed2 (dyEx8) and either Punc-4::GFP (wdIs5) or Pacr-5::yellow cameleon (rtIs29) that mark A- and B-type motor neurons, respectively. (B) L2-stage animal co-expressing Pcwn-1::YFP and Pacr-2::syd-2::RFP. The SYD-2 reporter marks focal adhesions in presynaptic termini of A- and B-type cholinergic motor neurons. (C) cwn-1 is expressed in V- and D-type motor neurons that innervate ventral and dorsal muscles, respectively. Same strain and stage as in B. Arrowheads denote motor neuron cell bodies and white arrows indicate D-type motor neuron axons. M, muscle (red arrows). Scale bars: 10 µm.
cwn-1 is expressed in A- and B-type cholinergic motor neurons. (A) L4-stage animals co-expressing Pcwn-1::DsRed2 (dyEx8) and either Punc-4::GFP (wdIs5) or Pacr-5::yellow cameleon (rtIs29) that mark A- and B-type motor neurons, respectively. (B) L2-stage animal co-expressing Pcwn-1::YFP and Pacr-2::syd-2::RFP. The SYD-2 reporter marks focal adhesions in presynaptic termini of A- and B-type cholinergic motor neurons. (C) cwn-1 is expressed in V- and D-type motor neurons that innervate ventral and dorsal muscles, respectively. Same strain and stage as in B. Arrowheads denote motor neuron cell bodies and white arrows indicate D-type motor neuron axons. M, muscle (red arrows). Scale bars: 10 µm.
EGL-30 promotes vulval progenitor fates from cwn-1-expressing neurons. (A,B) Incidence of the 3° fate in P3.p and P4.p in transgenic rescue experiments. (A) ***P=6.9E-9, *P=3.1E-2. (B) For egl-19(n582) versus wild type, **P=0.001. For cwn-1(ok546) versus cwn-1(ok546); Ex[Punc-18::cwn-1], ***P=8E-10 (P3.p) and ***P=5E-35 (P4.p). For cwn-1(ok546); Ex[Punc-18::cwn-1] versus egl-30(ad805); cwn-1(ok546); Ex[Punc-18::cwn-1], ***P=5.9E-5 (P3.p) and ***P=7E-4 (P4.p). For cwn-1(ok546) versus cwn-1(ok546); Ex[Pmyo-3::cwn-1], ***P=1.2E-13 (P3.p) and ***P=2.4E-35 (P4.p). (C) EGL-19 acts parallel to CWN-1 to promote P5.p progenitor fate. *P=0.02. P-values were calculated using two-tailed Fisher's exact tests.
EGL-30 promotes vulval progenitor fates from cwn-1-expressing neurons. (A,B) Incidence of the 3° fate in P3.p and P4.p in transgenic rescue experiments. (A) ***P=6.9E-9, *P=3.1E-2. (B) For egl-19(n582) versus wild type, **P=0.001. For cwn-1(ok546) versus cwn-1(ok546); Ex[Punc-18::cwn-1], ***P=8E-10 (P3.p) and ***P=5E-35 (P4.p). For cwn-1(ok546); Ex[Punc-18::cwn-1] versus egl-30(ad805); cwn-1(ok546); Ex[Punc-18::cwn-1], ***P=5.9E-5 (P3.p) and ***P=7E-4 (P4.p). For cwn-1(ok546) versus cwn-1(ok546); Ex[Pmyo-3::cwn-1], ***P=1.2E-13 (P3.p) and ***P=2.4E-35 (P4.p). (C) EGL-19 acts parallel to CWN-1 to promote P5.p progenitor fate. *P=0.02. P-values were calculated using two-tailed Fisher's exact tests.
EGL-30 post-transcriptionally regulates neuronal CWN-1
To confirm EGL-30 regulates neuronal cwn-1, we used the unc-18 promoter to express cwn-1 only in neurons. Endogenous UNC-18 protein is expressed in all ventral cord motor neurons and a few head neurons (Gengyo-Ando et al., 1993). When cwn-1 was driven by the unc-18 promoter, which is expressed in these cells and some tail neurons (Fig. S1) (Moghal et al., 2003), it restored the 3° fate to P3.p and P4.p in cwn-1 null mutants (Fig. 5A,B). The egl-30(ad805) mutation reduced Punc-18::cwn-1 rescuing activity, but not expression of a co-injected unc-18-driven YFP reporter (Fig. 5B, Fig. S2). These results indicate EGL-30 does not regulate the unc-18 promoter and acts post-transcriptionally on CWN-1. A post-transcriptional mechanism is also supported by egl-30 mutation not affecting total cwn-1 mRNA expression (mostly derived from muscle) or cwn-1 transcriptional reporter activity in ventral cord motor neurons (Fig. S3). We then used the myo-3 promoter, which drives high expression only in body-wall muscle (Fig. S4) (Okkema et al., 1993), to express CWN-1 in cwn-1 null mutants. Two different transgenic arrays, one that rescued P3.p and P4.p progenitor fate defects to wild type, and another that caused a gain-of-function P3.p phenotype, were insensitive to the egl-30(ad805) mutation (Fig. 5C, Fig. S5). Thus, EGL-30 does not post-transcriptionally regulate body-wall muscle CWN-1, where EGL-30 expression has not been detected (Bastiani et al., 2003; Lackner et al., 1999). These results also indicate that enforced CWN-1 expression from muscle rescues the 3° fate defects caused by egl-30 mutation, which supports EGL-30 regulating vulval progenitor fates through Wnt production.
Posterior body-wall muscle and EGL-19 calcium channels within these cells are required for EGL-30 to promote vulval induction (Moghal et al., 2003). To investigate whether muscle produces a parallel signal to EGL-30-regulated CWN-1, we constructed double mutants between cwn-1(ok546) null and egl-19(n582) hypomorphic mutations. Single egl-19 mutants had only a moderate reduction in the 3° fate frequency of P3.p (Fig. 5B); however, cwn-1; egl-19 double mutants additionally showed a trend towards a greater reduction in the P4.p 3° fate frequency and increased underinduction (Fig. 5B, Table 1). The greater underinduced phenotype was due to a significant ninefold increase in P5.p not becoming a vulval progenitor relative to single cwn-1 null mutants (Fig. 5C, Table 2). Thus, EGL-19 regulates a signal other than CWN-1 that is also involved in vulval development.
Acetylcholine regulates the EGL-30-Wnt pathway
To investigate the sensory mechanism that regulates the developmental activity of EGL-30, we searched for an upstream neurotransmitter. Acetylcholine (ACh) is the most widely used neurotransmitter in C. elegans and can reach motor neurons through multiple routes (Alfonso et al., 1993; Duerr et al., 2008; Pereira et al., 2015). It is secreted by the interneurons that synapse with the motor neurons and is produced by the A- and B-type motor neurons themselves, which can respond to humoral ACh (Chan et al., 2013a). In addition, when worms are in water, hyperactivation of proprioceptive B-type motor neurons increases ACh to drive swimming (Ghosh and Emmons, 2008; Jones et al., 2011; Matthies et al., 2006; Wen et al., 2012). ACh also stimulates other EGL-30 functions in motor neurons (Lackner et al., 1999). Mutations in cha-1, the choline acetyltransferase that catalyzes ACh biosynthesis (Alfonso et al., 1994), and unc-17, the ACh transporter that loads ACh into synaptic vesicles (Alfonso et al., 1993) (Fig. 6A), caused rare underinduced phenotypes, post-progenitor generation (Fig. 6B, Tables 1, 2). They also reduced P3.p and P4.p vulval progenitor frequencies, which were rescued by overexpressing CWN-1 in muscle (Fig. 6C), suggesting ACh regulates CWN-1 levels.
Acetylcholine regulates vulval development. (A) Regulation of acetylcholine production and exocytosis in motor neurons. ACh, acetylcholine; ChAT, choline acetyltransferase; VAChT, vesicular acetylcholine transporter; nAChR, nicotinic acetylcholine receptor (29 members); mAChR, muscarinic acetylcholine receptor (three members); GEF, guanine nucleotide exchange factor; DAG, diacylglycerol; DAGK, DAG kinase; PA, phosphatidic acid. (B) Underinduced vulval phenotypes in L4-stage mutants. Scale bars: 10 µm. (C) Incidence of the 3° fate in P3.p and P4.p in ACh pathway mutants. P-values are relative to wild type unless otherwise indicated and calculated with two-tailed Fisher's exact tests. For single mutants: P3.p, ***P=1E-5 [cha-1(p1152)], 1.5E-4 [unc-17(e245)], 7.3E-6 [gar-3(vu78)] and 8.9E-5 [gar-3(gk305)]; and for P4.p, ***P=6.6E-6 [unc-17(e245)] and **P=4E10-3 [cha-1(p1152)]. For cha-1(p1152) versus cha-1(p1152); Ex[Pmyo-3::cwn1], ***P=8.4E-23 (P3.p) and *P=0.05 (P4.p). For unc-17(e245) versus unc-17(e245); Ex[Pmyo-3::cwn-1], ***P=2.5E-27 (P3.p) and **P=0.001 (P4.p). For gar-3(gk305) versus egl-30(tg26); gar-3(gk305), *P=0.008.
Acetylcholine regulates vulval development. (A) Regulation of acetylcholine production and exocytosis in motor neurons. ACh, acetylcholine; ChAT, choline acetyltransferase; VAChT, vesicular acetylcholine transporter; nAChR, nicotinic acetylcholine receptor (29 members); mAChR, muscarinic acetylcholine receptor (three members); GEF, guanine nucleotide exchange factor; DAG, diacylglycerol; DAGK, DAG kinase; PA, phosphatidic acid. (B) Underinduced vulval phenotypes in L4-stage mutants. Scale bars: 10 µm. (C) Incidence of the 3° fate in P3.p and P4.p in ACh pathway mutants. P-values are relative to wild type unless otherwise indicated and calculated with two-tailed Fisher's exact tests. For single mutants: P3.p, ***P=1E-5 [cha-1(p1152)], 1.5E-4 [unc-17(e245)], 7.3E-6 [gar-3(vu78)] and 8.9E-5 [gar-3(gk305)]; and for P4.p, ***P=6.6E-6 [unc-17(e245)] and **P=4E10-3 [cha-1(p1152)]. For cha-1(p1152) versus cha-1(p1152); Ex[Pmyo-3::cwn1], ***P=8.4E-23 (P3.p) and *P=0.05 (P4.p). For unc-17(e245) versus unc-17(e245); Ex[Pmyo-3::cwn-1], ***P=2.5E-27 (P3.p) and **P=0.001 (P4.p). For gar-3(gk305) versus egl-30(tg26); gar-3(gk305), *P=0.008.
C. elegans has many ACh receptors (Rand, 2007) and the full spectrum coupled to EGL-30 is unknown. However, the GAR-3 muscarinic ACh receptor is co-expressed with acr-2 in all cholinergic A- and B-type motor neurons (Chan et al., 2013a). It also activates other EGL-30 functions in several cell types, including ventral cord motor neurons (Chan et al., 2012; Labed et al., 2018; Liu et al., 2007). Two gar-3 mutations, a null (gk305) and a missense (vu78), reduced the 3° fate frequency in P3.p (Fig. 6C). This phenotype was rescued by the egl-30(tg26) gain-of-function mutation (Fig. 6C), placing ACh upstream of EGL-30 for vulval progenitor generation.
We next investigated whether liquid stimulation of vulval induction, which is dependent on EGL-30 and BAR-1/β-catenin (Moghal et al., 2003), requires ACh and CWN-1. We first assessed our protocol by comparing vulval induction in let-23(sy1) mutants grown under three conditions: (1) hatch eggs and grow to the L4 stage on agar plates with food; (2) hatch eggs in liquid without food, which synchronizes worms as L1 larvae, followed by plating on food; and (3) hatch eggs in liquid, with subsequent addition of food and growth to the L4 stage in liquid. Relative to growth entirely on agar plates, hatching and synchronization only to the L1 stage in liquid mildly increased vulval induction, while continuous growth in liquid to the L4 stage had a greater effect that matched genetic activation of egl-30 by the tg26 mutation (Fig. 7, Table 1).
ACh and CWN-1 are required for liquid-stimulation of vulval induction. Animals were grown as indicated with the number (n) for each experiment indicated. (A) For comparisons between (1) plate hatch/plate growth versus (2) liquid hatch/plate growth and (2) liquid hatch/plate growth versus (3) liquid hatch/liquid growth, P-values were *P=0.02 and ***P=1.1E-11 [let-23(sy1)]; *P=0.04 and 0.20 (not significant) [let-23(sy1); gar-3(gk305)]; and 0.25 (not significant) and *P=0.009 [cwn-1(ok546) let-23(sy1)]. For comparisons under liquid hatch/liquid growth conditions between let-23(sy1) versus other strains, ***P= 2E-17 [let-23(sy1); gar-3(gk305)] and ***P=2.8E-24 [cwn-1(ok546) let-23(sy1)]. (B) For comparisons within a strain between the effects of egl-30(tg26) versus liquid hatch/liquid growth, P=0.81 (not significant) [let-23(sy1)], ***P=1.3E-12 [let-23(sy1); gar-3(gk305)], P=0.36 (not significant) [cwn-1(ok546) let-23(sy1)]. For comparisons under plate hatch/plate growth conditions between egl-30(tg26); let-23(sy1) versus other strains, P=0.07 (not significant) [egl-30(tg26); let-23(sy1); gar-3(gk305)] and ***P=8.6E-19 [egl-30(tg26); cwn-1(ok546) let-23(sy1)]. P-values were calculated using two-tailed Fisher's exact tests.
ACh and CWN-1 are required for liquid-stimulation of vulval induction. Animals were grown as indicated with the number (n) for each experiment indicated. (A) For comparisons between (1) plate hatch/plate growth versus (2) liquid hatch/plate growth and (2) liquid hatch/plate growth versus (3) liquid hatch/liquid growth, P-values were *P=0.02 and ***P=1.1E-11 [let-23(sy1)]; *P=0.04 and 0.20 (not significant) [let-23(sy1); gar-3(gk305)]; and 0.25 (not significant) and *P=0.009 [cwn-1(ok546) let-23(sy1)]. For comparisons under liquid hatch/liquid growth conditions between let-23(sy1) versus other strains, ***P= 2E-17 [let-23(sy1); gar-3(gk305)] and ***P=2.8E-24 [cwn-1(ok546) let-23(sy1)]. (B) For comparisons within a strain between the effects of egl-30(tg26) versus liquid hatch/liquid growth, P=0.81 (not significant) [let-23(sy1)], ***P=1.3E-12 [let-23(sy1); gar-3(gk305)], P=0.36 (not significant) [cwn-1(ok546) let-23(sy1)]. For comparisons under plate hatch/plate growth conditions between egl-30(tg26); let-23(sy1) versus other strains, P=0.07 (not significant) [egl-30(tg26); let-23(sy1); gar-3(gk305)] and ***P=8.6E-19 [egl-30(tg26); cwn-1(ok546) let-23(sy1)]. P-values were calculated using two-tailed Fisher's exact tests.
Addition of either the gar-3(gk305) or cwn-1 null mutation to let-23(sy1) mutants strongly attenuated liquid stimulation of vulval induction (Fig. 7A, Table 1). However, these mutations behaved differently when constitutively active EGL-30 stimulated vulval induction. egl-30(tg26); let-23(sy1); gar-3(gk305) triple mutants grown on plates had much more vulval induction than let-23(sy1); gar-3(gk305) double mutants grown in liquid (Fig. 7B, Table 1). In fact, egl-30(tg26); let-23(sy1); gar-3(gk305) triple mutants had a similar frequency of animals with vulval tissue as did control egl-30(tg26); let-23(sy1) double mutants (Fig. 7B), although the gar-3 mutation did lower the average vulval induction per animal (Table 1). By contrast, cwn-1 mutation comparably reduced the vulval induction activity of both liquid and constitutively active EGL-30 (Figs 3A and 7B, Table 1). These findings support ACh being upstream of EGL-30 during liquid stimulation of vulval induction and CWN-1 being a downstream effector.
EGL-30 may act through Wntless-mediated Wnt secretion
To investigate how EGL-30 regulates Wnt levels, we considered how EGL-30 regulates neurotransmitter release from motor neurons (Fig. 6A). A major EGL-30 effector is EGL-8/PLC-β, which increases diacylglycerol (DAG) and calcium levels that stimulate UNC-13 and UNC-64/syntaxin-dependent synaptic vesicle fusion with the plasma membrane (Koelle, 2016; Michelassi et al., 2017). A strong egl-8 loss-of-function mutation did not affect vulval induction or the 3° fate frequencies of P3.p and P4.p, or reduce the vulval induction activity of egl-30(tg26) in let-23(sy1) mutants (Fig. 8A, Table 1). In addition, a null mutation in dgk-1/DAG kinase that increases DAG levels and synaptic vesicle exocytosis (Nurrish et al., 1999), did not suppress the underinduced vulval phenotype of let-23(sy1) mutants (Table 1). Non-null mutations in unc-13 and unc-64 only weakly suppressed egl-30(tg26) vulval induction activity (Moghal et al., 2003) and, on their own, did not reduce the 3° fate frequency in P3.p and P4.p, or affect vulval induction (Fig. 8A, Table 1). Furthermore, a null mutation in unc-31/Caps (calcium activated protein for secretion), an essential component of the dense core vesicle exocytic pathway (Grishanin et al., 2004; Speese et al., 2007; Zhou et al., 2007), did not affect the 3° fate in P3.p or P4.p, or vulval induction, nor did it prevent egl-30(tg26) from suppressing the let-23(sy1) underinduced vulval phenotype (Table 1, Fig. 8A). Thus, EGL-30 does not mainly affect CWN-1 levels through neurochemical secretion pathways.
EGL-30 promotes vulval progenitor fates through MIG-14/Wntless and RhoA. (A) Incidence of the 3° fate in P3.p and P4.p in exocytic pathway mutants. P-values are relative to wild type unless otherwise indicated. For P3.p, ***P=4.6E-15; for P4.p, ***P=1.7E-53. (B) UNC-73 architecture. SPEC, spectrin repeat; PH, pleckstrin homology; SH3, SRC homology 3 domain; Ig, immunoglobulin; FN, fibronectin repeat; GEF, guanine nucleotide exchange factor. Numbers indicate amino acid positions. (C) Underinduced vulval phenotype in L4-stage unc-73 mutant. Scale bar: 10 µm. (D) Incidence of the 3° fate in P3.p and P4.p in unc-73 mutants. P-values are relative to wild type. For P3.p, ***P=8.4E-6 (e936), 6.6E-6 (gm40), 1.8E-9 (ev402) and **P=0.001 (ce362); and for P4.p, ***P=9E-9 (e936) and 5.2E-18 (gm40). (E) Incidence of the 3° fate in P3.p and P4.p in egl-30 mutants rescued with activated RhoA. ***P=9.3E-6 and **P=0.004. P-values were calculated using two-tailed Fisher's exact tests.
EGL-30 promotes vulval progenitor fates through MIG-14/Wntless and RhoA. (A) Incidence of the 3° fate in P3.p and P4.p in exocytic pathway mutants. P-values are relative to wild type unless otherwise indicated. For P3.p, ***P=4.6E-15; for P4.p, ***P=1.7E-53. (B) UNC-73 architecture. SPEC, spectrin repeat; PH, pleckstrin homology; SH3, SRC homology 3 domain; Ig, immunoglobulin; FN, fibronectin repeat; GEF, guanine nucleotide exchange factor. Numbers indicate amino acid positions. (C) Underinduced vulval phenotype in L4-stage unc-73 mutant. Scale bar: 10 µm. (D) Incidence of the 3° fate in P3.p and P4.p in unc-73 mutants. P-values are relative to wild type. For P3.p, ***P=8.4E-6 (e936), 6.6E-6 (gm40), 1.8E-9 (ev402) and **P=0.001 (ce362); and for P4.p, ***P=9E-9 (e936) and 5.2E-18 (gm40). (E) Incidence of the 3° fate in P3.p and P4.p in egl-30 mutants rescued with activated RhoA. ***P=9.3E-6 and **P=0.004. P-values were calculated using two-tailed Fisher's exact tests.
To examine the dependency of EGL-30 on a general Wnt secretion mechanism, we used a mutation in mig-14/Wntless, an essential Wnt secretion gene (Banziger et al., 2006; Bartscherer et al., 2006; Goodman et al., 2006; Najdi et al., 2012), which promotes vulval progenitor generation from unidentified neurons (Myers and Greenwald, 2007). The mig-14(ga62) mutation reduced the 3° fate frequencies of P3.p and P4.p, which were not rescued with the egl-30(tg26) gain-of-function mutation (Fig. 8A). mig-14 mutants also had a moderate underinduced phenotype due to either defects in vulval progenitor generation or to induction post-progenitor generation (Table 1, Fig. S6). The egl-30(tg26) mutation did not rescue the overall or specific induction defects that occurred when sufficient vulval progenitors were present (Table 1, Fig. S6). Thus, EGL-30 cannot promote P3.p/P4.p progenitor fate or vulval induction when MIG-14 is compromised, suggesting EGL-30 acts through Wntless-dependent Wnt secretion.
EGL-30 regulates Wnt activity through UNC-73 Rho GEF and RhoA
The other major Gαq effector belongs to a family of Rho guanine nucleotide exchange factors (GEFs), of which UNC-73/TRIO is the only C. elegans member (Lutz et al., 2007; Rojas et al., 2007; Williams et al., 2007) (Fig. 8B). UNC-73 has both Rac and Rho GEF domains (Fig. 8B), and promotes ventral migration of P blast cells (Spencer et al., 2001; Steven et al., 1998). As P blast cells are Pn.p precursors, some unc-73 mutants may appear to have defects in progenitor generation and vulval induction when they really reflect failure of P cells to migrate into the ventral epidermis. Thus, we scored both P cell migration and vulval development defects in unc-73 mutants (see Materials and Methods). The unc-73(e936) splicing mutation reduces full-length UNC-73 protein levels by 90% (Steven et al., 1998). Nineteen percent of unc-73(e936) mutants had P cell migration defects and 2% displayed an underinduced phenotype (Fig. 8C, Table 2, Table S1), which was not due to P cell migration or progenitor defects as P5.p-P7.p were always visible in the ventral epidermis and they always became vulval progenitors. The unc-73(gm40) mutation introduces an early STOP before both GEF domains (Steven et al., 1998) and caused P cell migration and underinduction phenotypes in ∼70% and 6% of animals, respectively (Table 2, Table S1). In two underinduced cases, we confirmed sufficient progenitors (Table 2). The unc-73(e936) mutation also strongly suppressed egl-30(tg26) vulval induction activity in let-23(sy1) mutants (Table 1), further confirming the role of UNC-73 in vulval induction.
In unc-73(e936) and gm40 mutants with normal P cell migration, the frequencies with which P3.p and P4.p adopted the 3° fate were reduced (Fig. 8D). The unc-73(ev402) and ce362 mutations also reduced the 3° fate frequency of P3.p (Fig. 8D). These latter mutations either delete the entire Rho GEF domain (Steven et al., 1998) or encode a missense mutation in this domain that impairs EGL-30-dependent RhoA activation (Williams et al., 2007), respectively (Fig. 8B). As neither of these mutations affected P cell migration (Table S1), these data further support a role for UNC-73 in vulval progenitor generation distinct from P cell migration, and implicate Rho as being downstream of UNC-73. To test whether Rho is downstream of EGL-30 in vulval progenitor generation, we expressed constitutively active RhoA in CWN-1-producing cells using the cwn-1 promoter in egl-30(ad805) mutants. We used a Q63L mutant of human RhoA, which is 88% identical to C. elegans RhoA (RHO-1). Q63 is conserved across species and its mutation to leucine prolongs the GTP-bound activated state (Mayer et al., 1999). RhoA(Q63L) almost completely restored the normal frequency of the 3° fate to P3.p and P4.p (Fig. 8E) in egl-30 mutants, supporting RhoA as being downstream of EGL-30 and UNC-73 in the Wnt pathway.
Rho GTPases and actin polymerization promote Wnt secretion
In motor neurons, UNC-73 and RhoA have mainly been studied for promoting synaptic vesicle fusion (Fig. 6A) (Chan et al., 2012; McMullan et al., 2006), which is not the major method by which EGL-30 stimulates CWN-1 production. However, RhoA and the Rho GEF domain of the UNC-73 ortholog TRIO also have conserved roles in stimulating polymerization and contraction of actin (Katoh et al., 2011; Seipel et al., 1999; Sit and Manser, 2011), which can either negatively or positively affect secretion depending on the cargo (Porat-Shliom et al., 2013). We therefore investigated if Rho and downstream actin polymerization promote Wnt secretion. Owing to the difficulty of quantifying Wnt secretion in C. elegans, we used a mouse cell line that secretes murine WNT3A in a Wntless/MIG-14-dependent manner, and is widely used to study Wnt secretion (Beckett et al., 2013; Maruyama et al., 2013; Neumann et al., 2009; Takada et al., 2006; Verras et al., 2008; Willert et al., 2003). To perturb Rho activity, we used the C3 exoenzyme from Clostridium botulinum, which specifically ADP ribosylates and inactivates RhoA, B and C, and has no other known targets (Vogelsgesang et al., 2007). C3 was used since a compensatory increase in RhoB occurs upon RhoA knockdown, and all three Rho proteins must be inactivated to fully depolymerize actin (Ho et al., 2008). Treatment of L-Wnt3A cells with C3 reduced WNT3A in the media by ∼50%, without diminishing its intracellular levels (Fig. 9A). To disrupt actin polymerization, we treated L-Wnt3A cells with latrunculin B, which belongs to a class of marine toxins that bind actin monomers with high specificity and rapidly depolymerize actin (Morton et al., 2000; Spector et al., 1999). Latrunculin B inhibited Wnt secretion more than C3 (Fig. 9B), which correlated with greater actin depolymerization, as assessed by phalloidin staining (Fig. 9C). These findings provide a potential explanation for how EGL-30 and UNC-73 promote CWN-1 production from motor neurons.
Rho GTPases promote Wnt secretion. Murine fibroblasts expressing murine WNT3A (L-Wnt3A) were treated with either C3 transferase (RhoA/B/C inhibitor) or latrunculin B (actin depolymerizer) for 3 days. (A,B) Secreted and intracellular WNT3A were quantified in media and cell lysates by western blotting and densitometry, with intracellular WNT3A normalized to ERK1/2 expression (representative blot shown). Data are from biological replicates and include mean±s.e.m. Significance was calculated using paired two-tailed t-tests. ***P=0.0005 (C3) and **P=0.005 (latrunculin B). (C) Visualization of actin depolymerization by rhodamine phalloidin staining. Scale bars: 10 µm.
Rho GTPases promote Wnt secretion. Murine fibroblasts expressing murine WNT3A (L-Wnt3A) were treated with either C3 transferase (RhoA/B/C inhibitor) or latrunculin B (actin depolymerizer) for 3 days. (A,B) Secreted and intracellular WNT3A were quantified in media and cell lysates by western blotting and densitometry, with intracellular WNT3A normalized to ERK1/2 expression (representative blot shown). Data are from biological replicates and include mean±s.e.m. Significance was calculated using paired two-tailed t-tests. ***P=0.0005 (C3) and **P=0.005 (latrunculin B). (C) Visualization of actin depolymerization by rhodamine phalloidin staining. Scale bars: 10 µm.
DISCUSSION
Previous work discovered that neuronal activation of EGL-30 promotes vulval induction and proposed that this activity helps make vulval development tolerant of water submersion (Moghal et al., 2003). How this EGL-30 function is naturally activated was unknown, so it was unclear whether this robustness mechanism was confined to water. In light of GPCR-independent mechanisms for activating heterotrimeric G-proteins (Blumer and Lanier, 2014; Papasergi et al., 2015), water could stimulate EGL-30 independently of neurotransmitters, possibly through a change in physiology or the unknown mechanoreceptors that proprioceptively activate B-type motor neurons (Wen et al., 2012). We find that the developmental activity of EGL-30 is regulated by ACh, which is elevated during liquid growth, as it is secreted by the proprioceptive B-type motor neurons to promote swimming (Ghosh and Emmons, 2008; Jones et al., 2011; Matthies et al., 2006; Wen et al., 2012). Mutations affecting ACh signaling perturb vulval progenitor generation and induction, and gar-3 ACh receptor mutation prevents liquid from stimulating vulval induction. As gar-3 mutation does not visibly affect locomotion, GAR-3 cannot act from muscle to activate neuronal EGL-30 through a proprioception mechanism involving muscle contraction. Instead, GAR-3, which is expressed on motor neurons (Chan et al., 2013a), is likely directly coupled to EGL-30, as reported for other EGL-30 functions (Chan et al., 2012; Labed et al., 2018; Liu et al., 2007). Consistent with this role, constitutively active EGL-30 is largely epistatic to gar-3 mutation for both vulval progenitor generation and induction. However, gar-3 mutation partially reduced the vulval induction activity of activated EGL-30, suggesting ACh may also regulate vulval induction through other pathways. This finding could explain why synaptic transmission mutations in unc-13 and unc-64 weakly suppress the vulval induction activity of EGL-30; they could reduce ACh secretion and its regulation of these other pathways.
EGL-30 was suggested to promote vulval induction from ventral cord motor neurons, including the A-type (Moghal et al., 2003). We provide additional support for this possibility. Transgenic expression of gain-of-function egl-30 from the cwn-1 promoter stimulates vulval induction. Although the cwn-1 promoter drives expression in body-wall muscle and A- and B-type motor neurons, EGL-30 does not appear to stimulate vulval induction from muscle (Moghal et al., 2003). In addition, endogenous EGL-30 expression is seen in ventral cord motor neurons, but not in body-wall muscle (Bastiani et al., 2003; Lackner et al., 1999), and egl-30 loss-of-function mutation reduces the developmental activity of neuronal, but not muscle-expressed, CWN-1.
Motor neurons and ACh are ideal for communicating information about the environment to developmental pathways (Fig. 10). Motor neurons receive broad sensory input to adapt locomotion to the environment and ACh, which is made only by neurons, is the most widely used neurotransmitter (Pereira et al., 2015). Variations in temperature, O2, CO2, salt, odorants and mechanical/proprioceptive stimuli are ultimately sensed by motor neurons through ACh levels (Koelle, 2016; Metaxakis et al., 2018; Wen et al., 2012; White et al., 1986). Most sensory information is integrated by interneurons that are synapsed with the motor neurons and use ACh to stimulate them (Chalfie et al., 1985; Metaxakis et al., 2018; Pereira et al., 2015; White et al., 1986). Motor neurons and GAR-3 ACh receptors on these cells are also activated by non-synaptic ACh (Chan et al., 2013a), which could originate from cholinergic sensory neurons and the motor neurons themselves (Pereira et al., 2015). B-type motor neurons may be an important source of ACh during liquid growth as they are hyperactivated by increased proprioceptive signaling (Wen et al., 2012). This wiring could allow a broad range of stresses that trigger adaptive locomotor behavior to increase Wnt production and help ensure the vulva develops properly. Such coordination may help ensure a developing animal reaches some reproductive capacity while attempting to escape from a suboptimal environment. Mutations affecting egl-30 and ACh signaling also caused defects in vulval progenitor generation and induction under baseline conditions, suggesting this pathway also promotes robustness to stochastic variations.
Model for sensory regulation of vulval development through EGL-30. Environmental variations are sensed through ciliated amphid, labial and phasmid neurons, and proprioceptive B-type motor neurons. Cholinergic interneurons (AVA, AVB, AVD, AVE and PVC) integrate sensory information from ciliated neurons and stimulate synaptic target motor neurons with ACh. Motor neurons may also respond to non-synaptic ACh from cholinergic sensory neurons and the motor neurons themselves. In water, reduced mechanical load increases autonomic bending frequencies induced by central pattern generators (subsets of B-type motor neurons and unknown anterior neurons), which increases the rate of proprioceptive activation of B-type motor neurons. Subsets of A/B and V/D cholinergic motor neurons expressing CWN-1 respond to changes in ACh levels through the GAR-3 muscarinic receptor that is coupled to EGL-30. EGL-30 activates RhoA and PLC-β-signaling pathways that promote Wnt secretion and synaptic ACh release at neuromuscular junctions, respectively. Motor neuron CWN-1 cooperates with CWN-1 from muscle (and possibly sex myoblasts), as well as another EGL-19-regulated signal, to promote vulval progenitor fates (especially in P3.p-P5.p) and vulval induction (mainly P6.p).
Model for sensory regulation of vulval development through EGL-30. Environmental variations are sensed through ciliated amphid, labial and phasmid neurons, and proprioceptive B-type motor neurons. Cholinergic interneurons (AVA, AVB, AVD, AVE and PVC) integrate sensory information from ciliated neurons and stimulate synaptic target motor neurons with ACh. Motor neurons may also respond to non-synaptic ACh from cholinergic sensory neurons and the motor neurons themselves. In water, reduced mechanical load increases autonomic bending frequencies induced by central pattern generators (subsets of B-type motor neurons and unknown anterior neurons), which increases the rate of proprioceptive activation of B-type motor neurons. Subsets of A/B and V/D cholinergic motor neurons expressing CWN-1 respond to changes in ACh levels through the GAR-3 muscarinic receptor that is coupled to EGL-30. EGL-30 activates RhoA and PLC-β-signaling pathways that promote Wnt secretion and synaptic ACh release at neuromuscular junctions, respectively. Motor neuron CWN-1 cooperates with CWN-1 from muscle (and possibly sex myoblasts), as well as another EGL-19-regulated signal, to promote vulval progenitor fates (especially in P3.p-P5.p) and vulval induction (mainly P6.p).
It was previously thought that motor neurons stimulate vulval induction through their behavioral activity involving muscle excitation (Moghal et al., 2003). Key to this model were the ability of EGL-30 to promote synaptic release of ACh at neuromuscular junctions (Koelle, 2016) and dependencies on muscle and muscle-expressed EGL-19, which promotes muscle excitation (Jospin et al., 2002). However, mutations in the main EGL-30-DAG/calcium pathway for synaptic transmission do not affect vulval development on their own, nor do they strongly impair the vulval induction activity of EGL-30 [here and Moghal et al., (2003)]. We propose, instead, that these neurons act by releasing CWN-1 through a different ACh-EGL-30-regulated mechanism (Fig. 10). EGL-30 cannot promote vulval development if Wnt secretion is compromised, and neither EGL-30 nor liquid stimulate vulval induction when CWN-1 levels are reduced. Given that site of action experiments place EGL-30 in cwn-1-expressing motor neurons, EGL-30 may affect vulval development by stimulating CWN-1 release from these cells. This model would explain why mutations in ACh signaling, egl-30 and cwn-1 similarly affect vulval progenitor generation and induction, and why the progenitor defects in ACh and egl-30 mutants are rescued by CWN-1 overexpression from outside the nervous system. We were also able to potentially connect EGL-30 to Wnt production within neurons. Mutation of the Rho GEF unc-73, the other major EGL-30 effector in motor neurons, also causes defects in vulval progenitor generation and induction, and prevents EGL-30 from stimulating vulval induction. Furthermore, constitutively active RhoA rescues vulval progenitor defects caused by egl-30 mutation, and in Wnt-secreting mammalian cells dependent on Wntless/MIG-14, Rho and downstream actin polymerization promote Wnt secretion. EGL-30-regulated Wnt production would cooperate with other Wnt sources to promote robustness in vulval development. Most notably, other sources would include rectal cells that produce EGL-20 and posterior body-wall muscle, which we believe secretes CWN-1 independently of EGL-30 and releases another Wnt that is regulated by EGL-19. Another intriguing source of CWN-1 would be the two sex myoblasts that are near P6.p at the time of vulval induction (Minor et al., 2013). Although GAR-3 and EGL-30 expression have not been reported in these cells, they may constitutively release CWN-1 to enhance vulval induction. Target cells most sensitive to neuronal CWN-1 would include P6.p (based on Wnt reporter activity), P3.p and P4.p, with robustness in P6.p induction potentially enhancing P5.p and P7.p induction through lateral signaling.
Regulation of extracellular Wnt levels is important, as Wnts have widely conserved roles in development, and dysregulation of their signaling is implicated in many diseases (Grigoryan et al., 2008; Holstein, 2012; Mohammed et al., 2016; Oliva et al., 2018; Petersen and Reddien, 2009; Tan and Barker, 2018). Previous work only found evidence for heterotrimeric G-proteins affecting Wnt signaling within Wnt-responding cells (Katanaev et al., 2005; Liu et al., 1999; Najafi, 2009; Salmanian et al., 2010). By contrast, our work reveals that these proteins also regulate extracellular Wnt levels. Importantly, EGL-30 is part of a pathway that allows Wnt levels to be altered in response to external signals. Previous studies have shown that extracellular signals only regulate Wnt transcription (Blyszczuk et al., 2017; Ingham and Hidalgo, 1993; Labed et al., 2018; Reddy et al., 2001; Soeda et al., 2014), with secretion being under constitutive regulation (Nusse and Clevers, 2017; Yu and Virshup, 2014). EGL-30, however, regulates CWN-1 even when cwn-1 expression is driven from the heterologous unc-18 promoter, revealing that Wnt activity can be post-transcriptionally controlled by a regulated signaling pathway. Other previous work with neurotransmitters described only transcriptional regulation of Wnt genes, and only in non-neuronal cell types (Labed et al., 2018; Soeda et al., 2014). We favor a model in which EGL-30 directly affects Wnt secretion as Rho signaling and MIG-14/Wntless are required for EGL-30 activity, and Rho promotes Wnt secretion in mammalian cells that are dependent on Wntless.
A secretion mechanism would have specificity, because Rho and actin polymerization are not required for all exocytosis. For example, glucose-stimulated insulin secretion is potentiated, rather than inhibited, by C3 and latrunculin B (Hammar et al., 2009; Kong et al., 2014; Liu et al., 2014; Thurmond et al., 2003). One mechanism could relate to Wnts sometimes being found in exosomes (Beckett et al., 2013; Chen et al., 2016; Harada et al., 2017; Korkut et al., 2009), which can be detected in actin-enriched filopodia where polymerized actin promotes docking and secretion of exosomes (Hoshino et al., 2013; Korkut et al., 2009; Sinha et al., 2016). A second mechanism could involve Golgi fission, which can be facilitated by RAB6A and GPCR-Gαq-Rho-induced polymerization of actin contractile rings (Grigoriev et al., 2011; Miserey-Lenkei et al., 2010; Zilberman et al., 2011). Although Wnts have not been reported to be in RAB6A-containing vesicles, they are associated with RAB8A, which can colocalize with RAB6A (Das et al., 2015; Grigoriev et al., 2011). A third possibility could involve assembly of actomyosin contractile rings at Wnt vesicles that have fused with the plasma membrane. Wnts are lipidated and glycosylated (Takada et al., 2006; Tanaka et al., 2002; Willert et al., 2003), and ‘sticky’ cargos, including salivary gland secretory protein in Drosophila, pulmonary surfactant in type II pneumocytes and von Willebrand clotting factor in endothelial cells, require forcible expulsion (Miklavc et al., 2015, 2012; Nightingale et al., 2011; Rousso et al., 2016; Rusu et al., 2014). In these examples, Rho and actin polymerization promote exocytosis, and Gαq-coupled GPCR signaling stimulates secretion of surfactant and the clotting factor (Miklavc et al., 2012; Rusu et al., 2014). As Rho can also be activated by GPCRs coupled to other heterotrimeric G-proteins, as well as by receptor tyrosine kinases (Bhattacharya et al., 2004; Schiller, 2006), Rho could be a focal point through which many signals regulate Wnt secretion.
Finally, as neurotransmitters are effectors of all sensory modalities (e.g. light, temperature sound, smell, taste and touch), the use of neurons to induce adaptive changes may greatly expand the range of environmental variations to which development is robust. In addition, whereas the ubiquitous heat shock response broadly affects development (Rutherford and Lindquist, 1998), neuronal mechanisms such as the one we describe, acts with specificity. The ACh-EGL-30-Rho pathway targets a specific developmental pathway, Wnt, and affects the vulva, the progenitors of which respond to Wnt and develop in the vicinity of the Wnt-secreting motor neurons. It will be interesting to search for additional neuronal mechanisms and to see to what extent the EGL-30-Rho mechanism is conserved. Indeed, Wnt is an ancient growth factor (Holstein, 2012) and the signaling cascade we describe is widely conserved. Moreover, in longer-lived adult organisms, Wnts additionally have roles in tissue homeostasis (Tan and Barker, 2018), which may be influenced by neuronal activity, especially during injury.
MATERIALS AND METHODS
C. elegans strains and nematode culture
Loss-of-function alleles include LGI: egl-30(ad805) (Brundage et al., 1996), egl-30(n715) (Trent et al., 1983), egl-30(n686) (Trent et al., 1983), unc-73(e936) (McIntire et al., 1992), unc-73(ev802) (Steven et al., 2005), unc-73(ce362) (Williams et al., 2007); LGII: let-23(sy1) (Aroian and Sternberg, 1991), cwn-1(ok546); LGIII: pha-1(e2123ts) (Granato et al., 1994); LGIV: egl-20(n585) (Trent et al., 1983), cwn-2(ok895), unc-31(e169) (Brenner, 1974); cha-1(p1152) (Rand and Russell, 1984), unc-17(e245) (Alfonso et al., 1993); LGV: egl-8(n488) (Trent et al., 1983); gar-3(vu78) (Steger and Avery, 2004), gar-3(gk305), mom-2(or42) (Thorpe et al., 1997); and LGX: dgk-1(sy428) (Jose and Koelle, 2005). Gain-of-function alleles include LGI: egl-30(tg26) (Doi and Iwasaki, 2002). Integrated transgenes include the Wnt reporter syIs187 (Green et al., 2008), the A-type motor neuron reporter wdIs5[Punc-4::GFP (Lickteig et al., 2001)], and the B-type motor neuron reporter, rtIs29 [Pacr-5::yellow cameleon (Haspel et al., 2010)]. The deletion alleles cwn-1(ok546) and cwn-2(ok895) were generated by C. elegans Gene Knockout Facility at the Oklahoma Medical Research Foundation and gar-3(gk305) was provided by the C. elegans Reverse Genetics Core Facility at the University of British Columbia. Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440).
For standard growth, C. elegans were cultured at 20°C on NGM agar plates seeded with OP50 E. coli (Brenner, 1974). For liquid experiments, fertilized eggs were isolated by removing gravid worms from NGM plates in M9 buffer (Brenner, 1974) and dissolving the worms in 20% commercial bleach/0.25 N NaOH for 20 min. Eggs were washed twice in M9 buffer and allowed to hatch in 1.5 ml microfuge tubes in M9 buffer without food, so that they eventually synchronized as early L1 stage arrested larvae. After 3 days of synchronization, part of the culture was maintained in liquid with the addition of OP50 bacteria, while an aliquot of liquid-hatched worms was placed onto standard NGM plates already seeded with OP50, as reference controls. Typical liquid cultures contained ∼100 larvae in 25 µl, with additional OP50 (5 µl of a 10-fold concentrated culture) added every 1-2 days, as needed.
Cell lines, toxins and drugs
Murine L-Wnt3A cells expressing murine WNT3A (Willert et al., 2003) were obtained from the American Type Culture Collection and were grown in DMEM/10% FBS with antibiotics. C3 exoenzyme/transferase (Rho inhibitor I, #CT04) from Clostridium botulinum was obtained from Cytoskeleton and latrunculin B was purchased from Santa Cruz Biotechnology.
Vulval development and P cell migration scoring
Animals were scored during the L4 stage using DIC optics and a Zeiss Axioimager microscope. For most studies, animals were anesthetized with 10 mM sodium azide and examined on 3% Noble agar pads. For vulval induction, the number of vulval nuclei were used to extrapolate how many vulval progenitor cells (VPCs) were induced to adopt vulval fates. A VPC generating seven or eight great granddaughters (Pn.pxxx) and no hyp7 tissue was scored as 1.0 cell induction. A VPC in which one daughter (Pn.px) fused with hyp7 and the other daughter generated three or four great granddaughters was scored as 0.5 cell induction. In wild-type animals, P5.p, P6.p, and P7.p each undergo 1.0 cell induction, resulting in a total of 3.0 cell induction. Animals with less than 3.0 cell induction were classified as underinduced.
1° and 2° fate adoption were determined based on final positions of vulval progeny. Progeny descended from 1°-fated cells are completely detached from the cuticle while those of 2°-fated cells include some that remain attached at the peripheral edges of the vulva. 3° and 4° fate adoption by P3.p and P4.p were inferred by counting total hyp7 nuclei in the ventral epidermis while using the distinctive positions of the vulva and the invariantly non-dividing Pn.p cells (P1.p, P2.p, P9.p, P10.p and P11.p) that each contribute one nucleus to hyp7 as reference points. For unc-73 mutants, animals were deemed to have possible P cell migration defects if all potential fates of P1.p-P11.p could not be accounted for at the L4 stage. In animals with wild-type vulval induction (most cases), a minimum of eight hyp7 nuclei should be observed if none of non-vulva-fated Pn.p cells divide (P1.p-P4.p, P8.p-P11.p), and more hyp7 nuclei should be present in underinduced animals. Most animals ultimately classified as P cell migration defective had fewer than eight hyp7 nuclei. In cases where eight or more hyp7 nuclei were observed, mutants could be classified as migration defective due to absence of hyp7 nuclei at positions that are clearly discernible when occupied (e.g. missing single nuclei from P1.p, P2.p, P9.p, P10.p, P11.p or both nuclei from P8.p).
Fluorescence microscopy and reporter assays
For C. elegans imaging, worms were placed on 3% Noble agar pads. Colocalization images of Pcwn-1::YFP and the pan-A/B-type motor neuron reporter Pacr-2::DsRed2 were captured on a Nikon Eclipse T1 confocal microscope, with L2-stage worms anesthetized in 500 mM sodium azide. For all other fluorescence imaging, larvae were anesthetized with 5 mM levamisole and photographed using epifluorescence and DIC on an Axioimager Z1 microscope using Axiovision software (Zeiss). All animals harboring the same reporters were photographed with the same exposure time and images were adjusted to the same brightness and contrast settings for comparison. Localization of cwn-1 promoter activity to A- and B-type motor neurons was determined by crossing the dyEx8 (Pcwn-1::DsRed2) extrachromosomal array into either the A-type motor neuron reporter strain wdIs5 [Punc-4::GFP (Lickteig et al., 2001)] or the B-type motor neuron reporter strain rtIs29 [Pacr-5::yellow cameleon (Haspel et al., 2010)]. Wnt activity was assessed with the POPTOP mCherry reporter syIs187 (Green et al., 2008), cwn-1 promoter activity with the Pcwn-1::DsRed2 reporter in the dyEx8 extrachromosomal array, and unc-18 promoter activity with the Punc-18::YFP reporter in the dyEx56 extrachromosomal array. Fluorescence intensity was quantified in the images using the densitometry function in Axiovision (Zeiss).
To visualize polymerized actin in L-Wnt3A cells, cells were cultured on coverslips coated with 48 µg/ml PureCol (Advanced Biomatrix), fixed in 4% paraformaldehyde and 320 mM sucrose in cytoskeleton buffer [10 mM MES (pH 6.0), 138 mM KCl, 3 mM MgCl2, 2 mM EGTA] for 20 min at room temperature, and stained with rhodamine phalloidin (1:150, ThermoFisher) in 3% BSA and 0.1% TX-100 in PBS for 1 h at room temperature. Images were captured on a Zeiss LSM700 confocal microscope.
Plasmids and transgenics
Construction of the plasmids pcwn-1::DsRed2, pcwn-1::YFP and pacr-2::syd-2::RFP (pDM624) has been described previously (Francis et al., 2005; Modzelewska et al., 2013). The transgenic pha-1 rescuing plasmid pBX-1 has also been described (Granato et al., 1994), and the transgenic marker plasmids pmyo-2::CFP and pmyo-2::mCherry were gifts from Andy Fire (Department of Pathology, Stanford University Medical School, USA) and Villu Maricq (Department of Biology, University of Utah, USA), respectively. To construct the egl-30(tg26) cDNA expression plasmids, the wild-type egl-30 cDNA was first amplified with primers egl30-20 5′-CTAGTCTAGAAAAAATGGCCTGCTGTTTATCCGAAG-3′ and egl30-21 5′-GCTAGTCTAGATGGCCTGCTGTTTATCCGAAGAG-3′, and following XbaI digestion, cloned into pBS. The wild-type egl-30 cDNA was then excised from pBSegl-30 and cloned into punc-18::egl-30::YFP (Moghal et al., 2003) via XbaI and KpnI digestion to yield punc-18::egl-30. In parallel, the egl-30 cDNA containing the gain-of-function tg26 mutation was amplified from pTG100.1 (Moghal et al., 2003) with egl30-20 and egl30-21 primers, and blunt-end cloned into pBS, yielding pBSegl-30(tg26). pBSegl-30(tg26) was digested with BamHI and KpnI, and the fragment containing the tg26 mutation was used to replace the corresponding wild-type region in punc-18::egl-30 to yield punc-18::egl-30(tg26). The egl-30(tg26) cDNA was then excised from this vector with XbaI and KpnI digestion and cloned into XbaI/KpnI-digested pCAN::YFP (Modzelewska et al., 2013). To make pegl-20::egl-30(tg26), the egl-20 promoter was excised from pegl-20::DsRed2 (Modzelewska et al., 2013) with SphI digestion and then used to replace the SphI CAN promoter fragment in pCAN::egl-30(tg26). To make pcwn-1::egl-30(tg26), the cwn-1 promoter was removed from pcwn-1::YFP (Modzelewska et al., 2013) with XbaI/SphI digestion and used to replace the CAN promoter in pCAN::egl-30(tg26) via the same sites.
The punc-18::YFP reporter plasmid was constructed by amplifying the unc-18 promoter with primers unc18-3 5′-AGCCCAAGCTTTGAAGGACAATGAACTAGAGGGAC-3′ and unc18-5 5′-AGCCCAAGCTTCAAAAATCCTCGTCGATGCACTCAC-3′, and cloning the 2.3 kb promoter fragment into pPD95.75YFP (from Shawn Xu, Department of Molecular and Integrative Physiology, University of Michigan Medical School, USA) via HindIII digestion. The cwn-1 cDNA was amplified by PCR from N2 cDNA using primers cwn1-15 5′-CTAGTCTAGAGAAATGCTGAAATCTACACAAGTGATC-3′ and cwn1-16 5′-TCGGGGTACCTTATAAGCATAAATACTTCTCAATTCG-3′, and was cloned into pBS. The cwn-1 cDNA corresponds to the larger 1122 bp isoform K10B4.6a, annotated at Wormbase (www.wormbase.org). To drive cwn-1 expression from the unc-18 promoter, the cwn-1 cDNA was removed from pBScwn-1 with XbaI/KpnI digestion and cloned into the corresponding sites of punc-18::YFP, yielding punc-18::cwn-1. The muscle-specific cwn-1 expression construct was generated by cloning the same XbaI/KpnI cwn-1 cDNA fragment into the XbaI/KpnI sites of pmyo-3::CFP (a gift from Andy Fire).
A human RhoA cDNA encoding the constitutively active Q63l mutant was obtained from Alan Hall (Cell Biology Program, Memorial Sloan-Kettering Cancer Center, USA) and amplified by PCR with primers mrhoA-1 5′-AGCTAGTCTAGAATGGCTGCCATCAGGAAGAAACTG-3′ and mrhoA-2 5′-ATCGCGGATCCTCACAAGATGAGGCACCCAGAC-3′. The PCR product was cloned into pCR8GW/TOPO/TA and sequenced where it was discovered that the parental clone harbored two nucleotide variations that changed two amino acids relative to the reference human sequence reported at NCBI. One change, V192I, affects the second last amino acid, which is isoleucine in both the mouse and C. elegans RhoA. Thus, this alteration was not corrected. The second nucleotide alteration encodes a F25N change and has been reported as an undocumented mutation prevalent among older RhoA cDNA clones (McMullan et al., 2006). As this mutation affects effector binding (McMullan et al., 2006), it was repaired by site-directed mutagenesis using primers rhoA-8 5′-GTGGAAAGACATGCTTGCTCATAGTCTTCAGCAAGGACCAGTTCCCAGAGGTG-3′ and rhoA-9 5′-CACCTCTGGGAACTGGTCCTTGCTGAAGACTATGAGCAAGCATGTCTTTCCAC-3′ to restore F25. Once corrected, the Q63L encoding RhoA cDNA with a STOP codon was cloned into the XbaI and BamHI sites downstream of the cwn-1 promoter and upstream of YFP in pcwn-1::YFP to generate pcwn-1::RhoA(Q63L).
Transgenic extrachromosomal arrays were generated as described previously (Mello et al., 1991). Extrachromosomal arrays included dyEx8 [pBX-1(100 ng/µl); pcwn-1::DsRed2 (60 ng/µl)], akEx1115 [pcwn-1::YFP (50 ng/µl), pacr-2::syd-2::RFP (50 ng/µl), pBX-1 (100 ng/µl)], akEx1605 [pcwn-1::egl-30(tg26) (50 ng/µl), pmyo-2::CFP (10 ng/µl), pBX-1 (100 ng/µl)], akEx1608 [pcwn-1::egl-30(tg26) (50 ng/µl), pmyo-2::CFP (10 ng/µl), pBX-1 (100 ng/µl)], akEx982 [pegl-20::egl-30(tg26) (50 ng/µl), pmyo-2::CFP (10 ng/µl), pBS (120 ng/µl)], dyEx54 [pmyo-3::YFP (20 ng/µl), pmyo-3::cwn-1 (40 ng/µl), pBX-1 (100 ng/µl)], dyEx56 [punc-18::YFP (30 ng/µl), punc-18::cwn-1 (40 ng/µl), pmyo-2::mCherry (2.5 ng/µl), pBX-1 (100 ng/µl)], dyEx58 [pmyo-2::mCherry (2.5 ng/µl), pmyo-3::cwn-1 (4 ng/µl), pBX-1 (100 ng/µl)] and dyEx50 [pcwn-1::YFP (25 ng/µl), pcwn-1::RhoA(Q63 L) (25 ng/µl), pmyo-2::mCherry (2.5 ng/µl), pBX-1 (100 ng/µl)]. The dyEx8, dyEx50, dyEx54, dyEx56, dyEx58, akEx1115, akEx1605 and akEx1608 transgenic arrays were initially generated in a pha-1 mutant background.
qRT-PCR
Gravid worms were collected from 12×60 mm plates, bleached and seeded onto fresh plates to collect L1-L2 larval stage-enriched worms. RNA was isolated, reverse transcribed into cDNA and analyzed by qRT-PCR as previously described (Modzelewska et al., 2013), using Taqman probes [cwn-1, Ce02426604_g1; tbp-1 (normalizer), Ce02448934_m1].
Wnt secretion assay
L-Wnt3A cell were seeded at 300,000 cells/well into 12-well tissue culture plates. For Rho inhibitor experiments, 24 h after seeding, 10 µg of C3 transferase were added per well in 400 µl of serum-free DMEM. After 90 min, an additional 400 μl of DMEM with 10% FBS were added per well. Three days later, media were collected and cells lysed in sample buffer [1% SDS, 10% glycerol, 80 mM Tris-HCl (pH 6.8)]. For actin depolymerization experiments, 24 h after seeding, cells were re-fed with 800 µl DMEM/5% FBS+10 µM latrunculin B or DMSO control vehicle. As the latrunculin B half-life was less than 24 h (our observations; Spector et al., 1989), fresh drug was added daily at 10 µM over 3 days. After 3 days, culture media were collected and cells lysed in sample buffer. 400 µl of recovered culture media were spun through a Nanosep centrifugal column (30 kDa MWCO, PALL) to dryness, and concentrated proteins resuspended in 60 µl of sample buffer after washing with 10 mM Tris-HCl (pH 8.0). C3 transferase did not affect cell number, as assessed by total protein recovered in cell lysates. However, latrunculin B inhibited cell proliferation without being cytotoxic, such that cell number was reduced by ∼25% relative to control cultures after 3 days. Thus, during SDS-PAGE analysis of secreted Wnt, more medium was loaded from the latrunculin B-treated wells, in proportion to the reduced protein levels in cell lysates. Protein was separated on a 10% polyacrylamide gel, and WNT3A and ERK1/2 were detected by western blotting (α-WNT3A, Millipore 09-162, 1:1000; α-ERK1/2, Millipore 05-1152, 1:1000) using fluorescent secondary antibodies and the Odyssey CLx scanner (Li-Cor). WNT3A and ERK1/2 levels were quantified by densitometry using Image Studio software (Li-Cor).
Statistical analyses
Experiments were exploratory without expectation of specific effect sizes. Thus, sample sizes were generally picked based on historical exploratory work in the field. For datasets involving comparisons of mean ranks (vulval induction, unc-18 and cwn-1 reporters), distributions were determined to not be normal by the Kolmogorov–Smirnov test, and were analyzed by two-tailed non-parametric Mann–Whitney tests. Frequencies of vulval progenitor generation (P3.p, P4.p) and POPTOP reporter expression were compared using two-tailed Fisher's exact tests.
Acknowledgements
We thank Ming-Sound Tsao and Andras Villu Maricq for support, and Sebastiao N. Martins-Filho, Luis Rene Garcia, Kugeng Huo, Elisa D'Arcangelo, Jessica Weiss, Bret Pearson and Carl Virtanen for helpful discussions. Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440, https://orip.nih.gov).
Footnotes
Author contributions
Conceptualization: K.M., J.C., N.M.; Investigation: K.M., L.B., N.M.; Writing - original draft: N.M.; Writing - review & editing: K.M., J.C., N.M.; Supervision: J.C., N.M.
Funding
This work was partially supported by the Natural Sciences and Engineering Research Council of Canada (RGPIN-20160o6644 UT to J.C.), and by funds indirectly provided from the Princess Margaret Cancer Centre, The Princess Margaret Cancer Foundation, and Ontario Ministry of Health.
Peer review history
The peer review history is available online at https://dev.biologists.org/lookup/doi/10.1242/dev.186080.reviewer-comments.pdf
References
Competing interests
The authors declare no competing or financial interests.