Abstract
Activated Gαq signals through Phospholipase-Cβ (PLCβ) and Trio, a Rho GTPase exchange factor (RhoGEF), but how these two effector pathways promote synaptic transmission remains poorly understood. We used the egg-laying behavior circuit of C. elegans to determine whether PLCβ and Trio mediate serotonin and Gαq signaling through independent or related biochemical pathways. Using genetic rescue experiments, we find that PLCβ functions in neurons while Trio functions in both neurons and the postsynaptic vulval muscles. While Gαq, PLCβ, and Trio RhoGEF mutants all fail to lay eggs in response to serotonin, optogenetic stimulation of the serotonin releasing HSN command neurons restores egg laying only in PLCβ mutants. Vulval muscle Ca2+ activity remained in PLCβ mutants but was eliminated in strong Gαq and Trio RhoGEF mutants. Exogenous treatment with Phorbol esters that mimic Diacylglycerol (DAG), a product of PIP2 hydrolysis, rescued egg-laying circuit activity and behavior defects of Gαq signaling mutants, suggesting both Phospholipase C and Rho signaling promote synaptic transmission and egg-laying via DAG production. DAG has been proposed to activate effectors including UNC-13, however, we find that phorbol esters, but not serotonin, stimulate egg laying in unc-13 mutants. Together, these results show that serotonin signaling through Gαq and PLCβ modulates UNC-13 activity to promote neurotransmitter release. Serotonin also signals through Gαq, Trio RhoGEF, and an unidentified PMA-responsive effector to promote postsynaptic muscle excitability. Thus, the same neuromodulator serotonin can signal in distinct cells and effector pathways to activate a motor behavior circuit.
Introduction
Neurons communicate in circuits via synaptic transmission to initiate, sustain, and terminate behaviors. During neurotransmission, both synaptic vesicles and dense-core vesicles fuse with the presynaptic membrane, releasing neurotransmitters and neuropeptides that activate postsynaptic ion channels and G protein coupled receptors (GPCRs) (Betke et al., 2012; Geppetti et al., 2015). While much has been learned about neurotransmitter signaling pathways through ionotropic receptors, the diversity of GPCRs and their signaling pathways has complicated our understanding of how their signaling exerts changes on cell excitability and behavior. The G protein, Gαq, is one of the major G proteins expressed in all excitable cells (Offermanns, 2001; Simon et al., 1991; Wilkie et al., 1992). Activated Gαq signals through PIP2- specific phospholipase-C β (PLCβ) to generate the second messengers, inositol 1,4,5 trisphosphate (IP3) and diacylglycerol (DAG). IP3 activates the IP3 receptor to release Ca2+ from intracellular stores and activate downstream kinases, lipases, and ion channels (Berridge et al., 2000; Huang, 1989; Li et al., 2014; Mujica and Gonzalez, 2011). DAG has been shown to recruit and activate numerous effector proteins including UNC-13 and Protein Kinase C (Ananthanarayanan et al., 2003; Brose and Rosenmund, 2002; Lou et al., 2008; Maruyama and Brenner, 1991; Rozengurt et al., 1997; Silinsky and Searl, 2003; Thore et al., 2005), but whether these or other identified DAG targets function to transduce all forms of Gαq signaling in vivo remains an open question.
Genetic studies in the nematode worm C. elegans show that Gαq signaling through both PLCβ and Trio, a Rho GTPase Exchange Factor (RhoGEF), promotes synaptic transmission. Whether Trio signaling through Rho similarly promotes DAG and IP3 production, or instead acts through an independent downstream pathway, is not clear. In worms, Gαq knockouts are lethal while PLCβ or Trio RhoGEF single knockouts show strong defects in neurotransmission that disrupt locomotion and egg-laying behaviors, resembling strong Gαq loss-of-function mutants (Bastiani et al., 2003; Brundage et al., 1996; Hajdu-Cronin et al., 1999; Lackner et al., 1999). Worms missing both PLCβ and Trio RhoGEF phenocopy the larval arrest phenotype of Gαq null mutants, consistent with these two effectors relaying most or all of the relevant Gαq signaling (Williams et al., 2007). Genetic and biochemical studies showed that Gαq activation of Trio and Rho is conserved (Chhatriwala et al., 2007; Rojas et al., 2007), however, it remains unclear how PLCβ and Rho signaling promotes neurotransmitter and neuropeptide release in vivo. The larval lethality of Gαq null mutants or PLCβ / Trio RhoGEF double mutants can be rescued by the DAG- mimetic phorbol ester, PMA (Reynolds et al., 2005; Steven et al., 2005; Williams et al., 2007), suggesting Gαq signaling through both PLCβ and Trio may ultimately converge to regulate DAG levels and the activation of downstream effectors. Both PLCβ and Trio RhoGEF promote acetylcholine (ACh) release from motor neurons that control locomotion (Lackner et al., 1999; Miller et al., 1999; Williams et al., 2007), although mutations in Trio RhoGEF cause behavior defects more aligned with a function in dense core vesicle release (Hu et al., 2011). PLCβ is expressed in the nervous system and intestine, while Gαq and Trio show additional postsynaptic expression in muscles (Bastiani et al., 2003; Lackner et al., 1999; Steven et al., 1998; Steven et al., 2005). While re-expression of PLCβ or Trio in neurons rescue the locomotion behavior defects of their mutants (Williams et al., 2007), whether these proteins exclusively function in neurons to regulate egg laying, is an open question.
Using genetics, optogenetics, pharmacology, and calcium imaging techniques we have investigated how Gαq signaling through its two effector pathways regulate egg-laying circuit activity and behavior. We find that Gαq signals through PLCβ in the presynaptic neurons while Gαq signals through Trio in both neurons and muscles to promote egg laying. These results clarify that despite Gαq signaling through independent PLCβ and Trio pathways, these effectors may ultimately converge to potentiate DAG levels which promote egg-laying behavior.
MATERIALS AND METHODS
Strains
C. elegans worms were maintained at 20°C on Nematode Growth Medium (NGM) agar plates with E. coli OP50 as a source of food as described previously (Brenner, 1974). All behavior assays and fluorescence imaging experiments were performed with age-matched adult hermaphrodites aged 20 – 36 h after the late L4 stage. All strains used in this study are listed in Table 1.
Molecular biology and transgenes Vulval muscle GCaMP5 strains
Vulval muscle Ca2+ activity was recorded using GCaMP5G (Akerboom et al., 2013) which was expressed along with mCherry from the unc-103e promoter (Collins and Koelle, 2013), as previously described (Collins et al., 2016; Ravi et al., 2018a). The wild-type reporter strain, LX1918 vsIs164 [unc-103e::GCaMP5::unc-54 3’UTR + unc-103e::mCherry::unc-54 3’UTR + lin- 15(+)] lite-1(ce314) lin-15(n765ts) X was used as described previously (Collins et al., 2016). LX1918 males were crossed separately into DA823 egl-30(ad805) I, MT1434 egl-30(n686) I, JT47 egl-8(sa47) V, MT1083 egl-8(n488) V, KG1278 unc-73(ce362) I, LX1226 eat-16(tm761) I, CG21 egl-30(tg26) I, him-5(e1490) V, or KP1097 dgk-1(nu62) X hermaphrodites to generate MIA140 egl-30(ad805) I; vsIs164 lite-1(ce314) lin-15(n765ts) X, MIA139 egl-30(n686) I; vsIs164 lite-1(ce314) lin-15(n765ts) X, MIA109 egl-8(sa47) V; vsIs164 lite-1(ce314) lin-15(n765ts) X, MIA288 egl-8(n488) V; vsIs164 lite-1(ce314) lin-15(n765ts) X, MIA141 unc-73(ce362) I; vsIs164 lite-1(ce314) lin-15(n765ts) X, MIA287 eat-16(tm761) I; vsIs164 lite-1(ce314) lin-15(n765ts) X, MIA286 egl-30(tg26) I; vsIs164 lite-1(ce314) lin-15(n765ts) X, and MIA296 dgk-1(nu62) vsIs164 lite-1(ce314) lin-15(n765ts) X, respectively. The corresponding gene mutation was confirmed by phenotype, genotype, or both. Presence of vsIs164 was confirmed observing the mCherry marker, and lite-1(ce314) X was confirmed with PCR genotyping. Oligo sequences used for genotyping the corresponding mutations are shown in Table 2.
Trio RhoGEF-E transgenes
Pan-neuronal expression
The rab-3 promoter was used to drive expression of GFP alone or with Trio-RhoGEF-E. Briefly, plasmids KG#68 (rab-3p::GFP; 15 ng/µl) alone or with KG#281 (rab-3p:: unc-73e; 50 ng/µL) were injected into KG1278 unc-73(ce362) I (Williams et al., 2007). For behavior experiments, five independent GFP-expressing transgenic lines were used, from which a single transgenic line from each was kept: MIA374 unc-73(ce362) I; keyEx66 (expressing GFP alone) and MIA375 unc-73(ce362) I; keyEx67 (expressing GFP and Trio RhoGEF-E). Plasmids KG#281(rab- 3p::unc-73e) and KG#68(rab-3p::GFP) were kind gifts from Dr. Kenneth Miller.
Pan-muscle expression
Plasmid pKMC33 (rgs-1p::mCherry) was digested with NheI/KpnI and ligated with similarly digested pPD96.52 (Fire lab C. elegans Vector Kit 1999; 1608: L2534, Addgene) to generate pKMC166 (myo-3p::mCherry). Plasmid KG#281 (rab-3p::unc-73e) was digested with NheI and KpnI, and the insert was ligated into similarly digested pKMC166 to generate pPD3 (myo- 3p::unc-73e). pKMC166 (15 ng/µL) alone or with pPD3 (50 ng/µL) was injected into KG1278 unc-73(ce362) I mutants. Five independent mCherry-expressing transgenic lines were used for behavior experiments from which a single transgenic line from each was kept: MIA376 unc- 73(ce362) I; keyEx68 (expressing mCherry alone) and MIA377 unc-73(ce362) I; keyEx69 (expressing mCherry + Trio RhoGEF-E).
Neuron and muscle co-expression
Plasmids KG#68 (15 ng/µl; pan-neuronal GFP) or pKMC166 (15 ng/µl; pan-muscle mCherry) alone or with KG#281 (50ng/µl; pan-neuronal unc-73e) and pPD3 (50ng/µL; pan-muscle unc- 73e) were injected into KG1278 unc-73(ce362) I, generating five independent mCherry(+), GFP(+) transgenic lines for behavior experiments from which a single transgenic line from each was kept: MIA372 unc-73(ce362); keyEx64 expressing mCherry (muscles) and GFP (neurons) only and MIA373 unc-73(ce362); keyEx65 expressing GFP (neurons), mCherry (muscles) and TrioRhoGEF-E (both neurons and muscles).
PLCβ Transgenes
To generate a control plasmid expressing GFP in neurons, GFP coding sequences were amplified from pJM60 (Moresco and Koelle, 2004) using oligonucleotides RE-GFP-FWD / -REV, digested with NheI / KpnI, and ligated into similarly digested pGP3 bearing the rgs-1 promoter (Dong et al., 2000), generating pKMC78. An egl-8 cDNA was used to generate and express a functional GFP fusion protein in neurons. Briefly, oligonucleotides egl-8-cDNA-fwd / -rev were used to amplified egl-8 coding sequences from a plasmid bearing an egl-8 cDNA provided by Dr. Kenneth Miller (pKP309). This amplicon was digested with NheI / NcoI and ligated into a similarly digested pPD49.26 plasmid, generating pKMC193. Quickchange mutagenesis with oligonucleotides egl-8-Cterm-NotI-fwd / -rev were used to insert an in-frame NotI site near the 3’ end of the egl-8 cDNA in a divergent region of the coding sequence, generating plasmid pKMC194. Coding sequences for egl-8 bearing this NotI site were then moved to pKMC78 by digestion of pKMC194 with NheI/NcoI followed by ligation into a similarly digested pKMC78, generating pKMC195. Oligonucleotides NotI-GFP-FWD / -REV were used to amplified GFP coding sequences from pKMC78, digested with NotI, and ligated into a similarly digested pKMC195, generating pKMC196. A strain bearing the egl-8(sa47) mutation was generously provided by Dr. Joshua Kaplan and backcrossed four times to N2 wild-type animals to generate LX1225 egl-8(sa47) V. MT8189 lin-15(n765ts) males were mated to LX1225 to generate LX1287 egl-8(sa47) V; lin-15(n765ts) X hermaphrodites that were kept at 15°C prior to injection. Plasmids expressing GFP alone (pKMC78; 5 ng/µl) or egl-8 CDNA fused to GFP (pKMC196; 5 ng/µl) were injected along with pL15EK (50ng/µl) into LX1287 hermaphrodites. For behavior experiments, five independent GFP-expressing lines were used from which a single transgenic line (vsEx679 [GFP] and vsEx680 [EGL-8::GFP], respectively) was kept.
Vulval muscle Channelrhodopsin-2 strains
N2 males were crossed into MIA229 keyIs48 [ceh-24p::ChR2::unc-54 3’UTR + lin-15(+)], lite- 1(ce314), lin-15(n765ts) X (Kopchock et al., 2021) to produce F1 heterozygous males, which then were crossed separately into MIA211 unc-73(ce362) I; lite-1(ce314) lin-15(n765ts) X, MIA299 egl-30(ad805) I; lite-1(ce314) lin-15(n765ts) X, MIA303 egl-8(n488) V; lite-1(ce314) lin- 15(n765ts) X, or MIA307 egl-8(sa47) V; lite-1(ce314) lin-15(n765ts) X hermaphrodites to generate vulval muscle specific Channelrhodopsin-2 (ChR2) expressing transgenic lines MIA248 unc-73(ce362) I; keyIs48; lite-1(ce314) X, MIA301 egl-30(ad805) I; keyIs48; lite- 1(ce314) X, MIA305 egl-8(n488) V; keyIs48; lite-1(ce314) X, and MIA309 egl-8(sa47) V; keyIs48; lite-1(ce314) X. The presence of lite-1(ce362) was confirmed by genotyping as above, and the presence of the ChR2 transgene was confirmed by rescue of the lin-15(n765ts) multi- vulva (Muv) phenotype.
HSN Channelrhodopsin strains
ChR2 was expressed in the HSNs from the egl-6 promoter via an integrated transgene (Emtage et al., 2012). This transgene was crossed into Gαq signaling mutants as follows. N2 males were crossed into LX1836 wzIs30 IV; lite-1(ce314) lin-15(n765ts) X to generate heterozygous F1 males, which were then crossed separately into MIA211 unc-73(ce362) I; lite-1(ce314) lin-15(n765ts) X, MIA299 egl-30(ad805) I; lite-1(ce314) lin-15(n765ts) X, MIA303 egl-8(n488) V; lite-1(ce314) lin-15(n765ts) X, or MIA307 egl-8(sa47) V; lite-1(ce314) lin-15(n765ts) X hermaphrodites to generate MIA247 unc-73(ce362) I; wzIs30 IV; lite-1(ce314) X, MIA300 egl- 30(ad805) I; wzIs30 IV; lite-1(ce314) X, MIA304 wzIs30 IV; egl-8(n488)V; lite-1(ce314) X, and MIA308 wzIs30 IV; egl-8(sa47) V; lite-1(ce314) X. The presence of lite-1(ce362) was confirmed by PCR genotyping, and the wzIs30 transgene was confirmed by rescue of the lin-15(n765ts) Muv phenotype.
Behavior assays
Quantification of egg accumulation was performed as described (Chase et al., 2004). Staged adults were obtained by picking late L4 animals and culturing them 24-30hr at 20°C. Each animal was placed in 7 µL of 20% hypochlorite (bleach) solution and eggs were counted after animals had dissolved. Numbers of eggs and any internally hatched L1 animals were combined.
Pharmacological assays
Egg laying in response to exogenous serotonin was performed as described (Banerjee et al., 2017). Individual staged adult animals were placed in 100 μl of either M9 buffer alone, or M9 containing 7.5 mg/ml serotonin (creatinine sulfate monohydrate salt, Sigma-Aldrich #H7752) or M9 containing 10 µM PMA (Phorbol-12-myristate-13-acetate, Calbiochem #524400) in a 96-well microtiter dish. After 1 hour, the number of released eggs and L1 larvae in each well were counted. Since egg-laying defective animals sometimes release one or two progeny in response to mechanical stimulation when they are first picked into the well, animals were only were recorded as responding if they laid 3 or more progeny. For calcium imaging, NGM plates containing PMA or control (equivolume ethanol solvent) were prepared as described (Reynolds et al., 2005). Age-matched adult worms (n>10) from each genotype were placed on separate PMA or control NGM plates at room temperature for 2 h, and subsequently transferred onto an agar chunk which was placed between the glass slide and a glass cover slip for Ca2+ activity recording (Ravi et al., 2018b). The unused plates were kept at 4 °C for future use.
Optogenetic assay
All-trans retinal (Sigma Aldrich, R2500) was resuspended in ethanol (100%) to make 100 mM solution and added to a warmed culture of OP50 bacteria grown in B Broth media to a final concentration of 0.4 mM. Individual NGM agar plates were seeded with 200 μl of freshly prepared +ATR food, and were grown in the dark for ∼24 hours prior to use. In all photo- stimulation experiments, control animals were grown in the absence of ATR. Animals were exposed to blue light using a Leica M165FC stereomicroscope equipped with an EL6000 metal halide light source generating 8–16 mW/cm2 (depending on the magnification used) of ∼470 ± 20 nm blue light through a GFP filter. The 30 s on/off sequence was programmed and controlled using the Optogenetics TTL Pulse Generator (OTPG-4, Version 3.3) triggering a SHB1 series shutter controller (ThorLabs; 170712-1).
Microscopy Ratiometric Ca2+ imaging
Vulval muscle Ca2+ activity was performed in freely behaving adult animals at 24-30 h past the late L4 larval stage, as described previously (Collins and Koelle 2013; Collins et al. 2016; Ravi et al. 2018a). Worms co-expressing GCaMP5G and mCherry under the unc-103e promoter transgene vsIs164 were mounted beneath the chunk of agar over the glass slide. Ca2+ activity was recorded through an 20X Apochromatic objective (0.8 NA) mounted on an inverted Zeiss Axio Observer.Z1. A Colibri.2 LED illumination system was used to excite GCaMP5 at 470 nm and mCherry at 590 nm for 10 msec every 50 msec. GFP and mCherry fluorescence emission channels were separated using a Hamamatsu W-VIEW Gemini image splitter and recorded simultaneously for 10 min with an ORCA-Flash 4.0 V2 sCMOS camera at 256 / 256-pixel resolution (4x4 binning) at 16-bit depth. The motorized stage was manually controlled using a joystick to maintain the freely behaving animal in the field of view. Image sequences were exported to Volocity software (Quorum Technologies Inc.) for segmentation and ratiometric analysis. Ca2+ transient peaks from ratio traces were detected using a custom MATLAB script, as described (Ravi et al. 2018a).
Experimental design and statistical analysis
Sample sizes for behavioral assays followed previous studies (Chase et al., 2004; Collins et al., 2016). Statistical analysis was performed using Prism 8 (GraphPad). Ca2+ transient peak amplitudes, widths, and inter-transient intervals were pooled from multiple animals (typically ≥ 10 animals per genotype). All statistical tests were corrected for multiple comparisons (Bonferroni for one-way ANOVA or Fisher’s exact; Dunn for Kruskal–Wallis). Each figure legend indicates individual p-values.
Results
Trio RhoGEF acts both in neurons and muscle to drive egg-laying behavior
Prior work has shown that Gαq signaling (Fig. 1A) through PLCβ and Trio RhoGEF promotes neurotransmitter release and locomotion (Brundage et al., 1996; Miller et al., 1999; Williams et al., 2007), however, whether Gαq and its effectors similarly regulate other C. elegans behavior circuits is less well established. To address this uncertainty, we chose to examine the neural circuit driving egg-laying behavior (Fig. 1B). Egg-laying behavior in C. elegans is regulated by small motor circuit with defined neurons and muscle connectivity (Cook et al., 2019; White et al., 1986). The HSNs are serotonergic command motor neurons (Fig. 1B) that initiate the egg-laying active state and promote the excitability of the vm1 and vm2 egg-laying vulval muscles (Waggoner et al. 1998; Emtage et al. 2012; Collins et al. 2016) while innervating ventral cord neurons release acetylcholine to regulate muscle contraction (Kim et al., 2001; Kopchock et al., 2021; Waggoner et al., 2000a). We first examined the steady-state accumulation of eggs in the uterus as a proxy for changes in egg-laying circuit activity and behavior. As previously shown, animals bearing mutations in the EAT-16 RGS protein, which inhibits Gαq signaling (Hajdu- Cronin et al., 1999), or gain-of-function mutations in Gαq itself, (Doi and Iwasaki, 2002) showed a significant increase in egg laying resulting in a significant reduction in egg accumulation compared to the ∼12 ± 3 embryos retained in wild-type animals (Fig. 1C-E; Fig. 1I). Conversely, animals bearing mutations which reduce Gαq signaling through its effectors showed the opposite phenotype. Animals bearing an early nonsense mutation predicted to be a PLCβ null mutant, [egl-8(sa47)], (Lackner et al., 1999; Williams et al., 2007) accumulated an average of 21 eggs, a significant increase (Fig. 1G). Animals bearing a missense mutation in the RhoGEF domain of Trio, unc-73(ce362) (Williams et al., 2007), showed an even stronger increase in egg accumulation, accumulating 31 eggs and closely resembling animals bearing loss-of-function mutations in Gαq itself (Fig. 1F-1I). Together, these results confirm that Gαq and its effectors PLCβ and Trio RhoGEF are required for egg-laying behavior in C. elegans and that loss of the Trio RhoGEF branch causes a stronger behavior impairment compared to loss of PLCβ.
Previous work has shown that PLCβ is expressed in neurons and in the intestine while Gαq and Trio are expressed in nearly all excitable cells including neurons and muscles (Bastiani et al., 2003; Brundage et al., 1996; Lackner et al., 1999; Miller et al., 1999; Steven et al., 1998). To understand where Gαq and its effectors function to regulate egg laying, we used cell-specific promoters to express cDNAs encoding PLCβ or Trio RhoGEF in either all neurons, in the body wall and egg-laying vulval muscles, or in both neurons and muscles. We found that transgenic expression of PLCβ from a neuron-specific promoter in PLCβ null mutants was sufficient to rescue their defects in egg laying (Fig. 1J) and acetylcholine (ACh) release as measured by restoration of sensitivity to aldicarb, a cholinesterase inhibitor (data not shown). Previous work has indicated the presence of eight transcript variants of Trio (A, B, C1, C2, D1, D2, E, and F), which are differentially expressed in C. elegans (Steven et al., 1998; Steven et al., 2005). Transgenic expression of Trio RhoGEF-E in neurons is fully sufficient to rescue the locomotion defects of Trio RhoGEF mutants (Williams et al., 2007). To explore whether Trio RhoGEF acts similarly in neurons for egg laying, we used a pan-neuronal promoter to express Trio RhoGEF- E and measured egg accumulation in these animals. We observed a partial, but significant reduction in the number of eggs retained in Trio RhoGEF mutants (∼36 eggs) compared to control Trio mutant animals (∼42 eggs; Fig. 1K). In contrast, transgenic expression of Trio RhoGEF-E in the egg-laying vulval muscles from a muscle-specific promoter showed a much greater rescue of egg accumulation (∼25 eggs), and this rescue of egg laying was improved to nearly wild-type levels when Trio RhoGEF was expressed in both neurons and muscles (∼19 eggs; Fig. 1K). Together, these results confirm that Gαq signals through both PLCβ and Trio RhoGEF function in neurons to regulate egg-laying behavior. Our results also suggest the Gαq effector Trio also functions in the postsynaptic vulval muscles for proper regulation of egg laying, a finding consistent with previously results regarding Gαq (Bastiani et al., 2003).
Serotonin signals through Gαq, Trio, and PLCβ to promote egg laying
Previous studies have shown that serotonin released from the HSNs signals through G-protein coupled serotonin receptors expressed on the vulval muscles (Bastiani et al., 2003; Dempsey et al., 2005; Fernandez et al., 2020; Tanis et al., 2008; Xiao et al., 2006). The vulval muscles are also innervated by cholinergic ventral cord motor neurons (Cook et al., 2019; White et al., 1986) whose release of ACh is regulated by serotonin and G protein signaling (Nurrish et al., 1999). To test whether serotonin promotes egg laying by activating receptors that couple to Gαq, we measured how serotonin signaling from the HSNs drives egg laying in Gαq signaling mutants. Serotonin promotes egg laying in hypertonic M9 buffer, a condition that normally inhibits egg laying in both wild-type and HSNs-deficient egl-1(dm) mutants, which developmentally lack the HSNs (Fig. 2B). Consistent with the previous results (Bastiani et al., 2003; Brundage et al., 1996; Trent et al., 1983), nearly 78% of wild-type and 70% of HSN-deficient egl-1(dm) animals laid eggs in response to serotonin compared to only 13% of mutant animals lacking Gαq (Fig. 2B). Serotonin response was similarly and significantly reduced to 23% and 3% in PLCβ and Trio RhoGEF mutant animals, respectively (Fig. 2B). The distinct responses of animals without HSNs and animals lacking Gαq signaling indicate that Gαq, PLCβ and Trio act at least in part outside of HSNs to promote egg laying in response to exogenous serotonin. To determine where Trio RhoGEF deficiency caused serotonin insensitivity, we measured egg laying in animals expressing Trio RhoGEF in either neurons, muscles, or both. Transgenic expression of Trio RhoGEF in neurons failed to rescue egg laying in response to serotonin (Fig. 2C), but expression of Trio RhoGEF in muscles, or in both neurons and muscles, restored egg laying to of animals (Fig. 2C), suggesting Trio mediates vulval muscle serotonin signaling. Together, these results indicate that Gαq, PLCβ and Trio RhoGEF function outside of the HSNs to drive egg laying in response to serotonin. The fact that PLCβ mutants are resistant to serotonin but unlike Trio mutants, show hyperactive egg-laying behavior upon pan-neuronal overexpression, suggests PLCβ acts presynaptically, either downstream or parallel to the HSNs, possibly to regulate ACh release, while Trio RhoGEF functions downstream of HSN, likely functioning postsynaptically in the egg-laying vulval muscles.
Optogenetic stimulation of the HSNs and vulval muscles suggests cellular specificity of Gαq effectors for egg-laying
Optogenetic stimulation of Channelrhodopsin-2 (ChR2) expressed in either the HSNs (Emtage et al., 2012) or vulval muscles (Kopchock et al., 2021) can drive egg laying. To test whether and how Gαq and its effectors signal downstream to mediate this response, we expressed ChR2 in HSNs in Gαq and effector mutants and measured egg laying during 30s of exposure to blue light. Blue light stimulation of the HSNs drove the release of ∼3 eggs in wild-type animals, which was reduced to essentially zero in Gαq [(egl-30(ad805)] and Trio RhoGEF [unc-73(ce362)] mutants (Fig. 3A), consistent to our previous results showing Trio RhoGEF acts downstream of the HSNs in the postsynaptic vulval muscles. In contrast, stimulation of the HSNs in PLCβ null mutants [egl-8(sa47) and egl-8(nu488)] drove egg release, releasing ∼4 and ∼6 embryos respectively in 30 s (Fig. 3A). To test whether the failure of egg laying in Gαq and Trio RhoGEF mutants was caused by developmental defects in the circuit, rather than excitability deficits, we expressed and stimulated ChR2 in the vulval muscles. Blue light exposure drove the release of ∼4 eggs in 30 s in wild-type animals. Both PLCβ [egl-8(sa47) and egl-8(nu488)] and Trio RhoGEF mutants [unc-73(ce362)] laid a similar number of eggs as wild-type control animals after blue light stimulation (Fig. 3B). The egl-30(ad805) Gαq mutant laid slightly fewer eggs (∼3) but this was not significantly different than wild type. Thus, the failure of Gαq and Trio mutants to lay eggs in response to exogenous serotonin or optogenetic stimulation of the HSNs does not arise from some intrinsic defect in vulval muscle contractility, but rather a specific deficiency in muscle electrical excitability. These results support the conclusion from our rescue experiments (Fig. 1J-1K) that Gαq signals through PLCβ and Trio RhoGEF in distinct cells and through unique mechanisms to promote egg-laying circuit activity and behavior.
Gαq and Trio are required for vulval muscle Ca2+ activity
Egg laying in C. elegans is a two-state behavior where ∼20 minute inactive states are punctuated by ∼2-minute active states with high levels of Ca2+ activity in the egg-laying circuit driving release of 3-5 eggs (Collins et al., 2016; Waggoner et al., 1998). Loss of Gαq signaling in egl- 30(n686) animals causes a significant reduction in spontaneous and serotonin-induced vulval muscle Ca2+ transients in immobilized animals (Shyn et al., 2003). We therefore tested whether these Gαq dependent Ca2+ activity defects were shared in PLCβ and Trio RhoGEF mutants. We expressed the genetically encoded Ca2+ reporter, GCaMP5, along with mCherry in the vulval muscles of animals with either increased or decreased Gαq signaling and performed ratiometric imaging in freely behaving animals. As shown in Figure 4A, the normal two-state pattern of vulval muscle Ca2+ activity is lost in animals bearing strong loss-of-function Gαq or Trio RhoGEF mutations. In both the egl-30(ad805) Gαq and unc-73(ce362) Trio mutants, we failed to observe either the strong egg-laying Ca2+ transients where both the vm1 and vm2 muscles contract or even the smaller amplitude rhythmic ‘twitch’ Ca2+ transients that are localized to the vm1 muscles (Fig. 4A), strongly reducing the observed Ca2+ transient frequency (Fig. 4B). In contrast, the weaker egl-30(n686) Gαq mutant and two different PLCβ null mutants showed grossly normal vulval muscle Ca2+ activity during both egg-laying and twitch transients (Fig. 4A-C). The elevated pattern of vulval muscle Ca2+ activity in weak egl-30(n686) Gαq and PLCβ mutants we observed resembled egl-1(dm) animals lacking the HSNs (Collins et al., 2016). These animals still enter and leave infrequent egg-laying active states, possibly driven by the accumulation of eggs in the uterus (Ravi et al., 2018a). DAG Kinase-θ (DGK-θ) is thought to reduce DAG signaling through conversion to phosphatidic acid (Fig. 1A). Loss of the DGK-θ/DGK-1 increases neurotransmitter release and egg laying, likely through elevation of DAG levels and activation of downstream effectors (Jose and Koelle, 2005; Miller et al., 1996; Nurrish et al., 1999). To test if dgk-1 mutants have increased egg-laying muscle activity, we performed Ca2+ imaging in dgk-1(nu62) mutants. Somewhat surprisingly, DAG Kinase mutant animals did not show a significant increase in vulval muscle Ca2+ activity (Fig. 4A-C). Like PLCβ, DGK-1 is expressed in neurons (Nurrish et al., 1999), suggesting alterations of IP3 and /or DAG levels in neurons may affect the frequency of egg-laying active states without altering the overall pattern or strength of vulval muscle Ca2+ activity within those active states. Conversely, animals with increased Gαq signaling either from a gain-of-function egl-30(tg26) mutations or eat-16(tm761) null mutations showed a significant increase in the amplitude of vulval muscle Ca2+ transients (Fig. 4B and 4C). There was also an increase in the frequency of vulval muscle Ca2+ transients in egl-30(tg26) mutants, but this effect did not rise to the level of significance in eat-16(tm761) mutants. Together, these results indicate Gαq signaling through Trio RhoGEF, but not PLCβ, promotes vulval muscle activity that drives twitching and egg-laying Ca2+ transients during egg-laying active states.
DAG mimetics restore muscle activity and egg laying to Gαq signaling mutants
How does serotonin signaling through Gαq promote vulval muscle activity? Previous results have shown that DAG mimetic Phorbol esters restore locomotion and animal viability to Gαq null mutants and to PLCβ, Trio RhoGEF double mutants (Williams et al., 2007). However, whether phorbol esters rescued egg-laying behavior in Gαq or effector single mutants was not reported, raising questions as to whether Gαq rescue by phorbols esters was a general feature or was restricted to specific effector signaling pathways or behaviors (Fig. 5A). We measured the egg- laying responses of wild-type animals and Gαq signaling mutants to Phorbol 12-myristate 13- acetate (PMA). As shown in Figure 5B, PMA treatment strongly stimulated egg laying in wild- type animals and all mutants with reduced Gαq signaling (≥80% animals laying eggs). Like serotonin (Fig. 2B), PMA also rescued egg laying in egl-1(dm) mutant animals lacking the HSNs (Fig. 4B), but only PMA rescued egg laying in Gαq and effector signaling mutants. These results are consistent with PMA acting downstream of both serotonin release from the HSNs and its subsequent signaling through Gαq. We next imaged how exogenous PMA affected vulval muscle Ca2+ activity. We performed 10-minute GCaMP5 recordings of wild-type or Gαq signaling mutants treated for two hours with 10 µM PMA (Movies 1-3). Quantitation of Ca2+ transients showed that exogenous PMA increased vulval muscle Ca2+ activity in wild-type animals, as shown by the more frequent rhythmic twitch Ca2+ transients (Movie 1). PMA also restored both rhythmic twitch and egg-laying Ca2+ transients to strong Gαq and Trio RhoGEF signaling mutants, as measured by both an increase in Ca2+ transient peak amplitude and frequency (Fig. 5C-5E). While PMA treatment did not restore vulval muscle Ca2+ transient frequency in Gαq and Trio mutants to wild-type levels, Ca2+ transient frequency was significantly improved in Gαq mutant animals (Fig. 5C-5D). We did record more Ca2+ transients in Trio mutants after treatment with PMA, but there was no significant difference to the vehicle control. We interpret differences observed in PMA rescue of egg laying (Fig. 5B) and vulval muscle Ca2+ (Fig. 5C-D) as being a result of the shortness of the 10-minute Ca2+ recording period that followed PMA treatment. Egg- laying active states typically occur every ∼20 minutes in wild-type animals (Waggoner et al., 1998) and every 12-15 minutes in hyperactive egg-laying behavior mutants that resemble the effects of PMA treatment (Ravi et al., 2020; Waggoner et al., 2000b). The M9 buffer assays allow egg-laying events to accumulate over one hour without having to capture specific events. While we did observe nearly all Gαq and Trio mutant animals as now being able to lay eggs, this was not always captured during the Ca2+ recording. Together, these studies show that DAG-mimetic phorbol esters restore muscle contractility defects of Gαq and Trio RhoGEF mutants, suggesting DAG production is a major and necessary consequence of Gαq signaling.
DAG promotes egg laying independent of UNC-13 and Protein Kinase C
How do Phorbol esters such as PMA rescue vulval muscle Ca2+ activity and egg laying? Previous results have shown that DAG and PMA binds to C1 domain-containing proteins such as mUNC- 13/UNC-13 and Protein Kinase C (PKC) to regulate their activity (Betz et al., 1998; Huang, 1989; Konig et al., 1985; Newton, 2001; Silinsky and Searl, 2003).To test if DAG regulates egg-laying behavior through activation of UNC-13 or PKC, we tested how PMA and serotonin affect egg laying in mutants lacking these proteins (Fig. 6). Mutants that eliminate axonal and synaptic UNC-13 show reduced egg laying, accumulating an average of 22 eggs compared to 15 seen in wild-type animals (Fig. 6A). Null unc-13(s69) and loss of function unc-13(e51) mutants have much less severe egg-laying defects compared to Gαq signaling mutants, accumulating ∼22 and ∼19 eggs (Fig. 6A) compared to >30 eggs (Fig. 1C-I). These results suggest that Gαq promotes synaptic transmission and egg-laying through an UNC-13 independent pathway. If PMA activation of vulval muscle Ca2+ and egg laying were mediated by UNC-13, then mutations that eliminate this protein should reduce that response. However, unc-13 mutants still laid eggs in response to PMA (Fig. 6B). Because DAG production may be a conserved biochemical consequence of Gαq signaling through both PLCβ and Trio RhoGEF pathways, and both effector pathways have functions in neurons, we hypothesized that serotonin might promote egg laying via modulation of UNC-13 activity. In support of this model, unc-13 mutants failed to lay eggs in response to serotonin (Fig. 6B). To determine whether Gαq promotes egg laying through Protein Kinase C, we analyzed egg accumulation in animals bearing predicted null mutations in different PKC isoforms (Edwards et al., 2012; Hyde et al., 2011; Okochi et al., 2005; Tabuse, 2002; Tabuse et al., 1989). PKC-1 has previously been shown to promote neuropeptide transmission (Sieburth et al., 2007). While neuropeptides signal to promote egg laying (Avery et al., 1993; Brewer et al., 2019; Jacob and Kaplan, 2003; Kass et al., 2001), PKC-1 (nPKC-ε) null mutants showed a grossly normal egg accumulation of 14∼17 eggs (Fig. 6C). Animals bearing predicted null mutants of novel and conventional PKCs such as nPKCδ/θ (TPA-1) and cPKCα/β (PKC-2) orthologs also show no significant differences in egg accumulation (Fig. 6C), suggesting that disruption of PKC signaling does not strongly affect egg-laying behavior. Like UNC-13, If PMA activation of vulval muscle Ca2+ and egg laying were mediated by PKCs, then mutations that eliminate PKCs, should reduce that response. However, PKC mutant animals still laid eggs in response to PMA (Fig. 6C) suggesting PKCs do not mediate the egg-laying response to PMA. Together, these results show that serotonin signals through Gαq via UNC-13 to promote neurotransmitter release, however, there are additional targets of DAG that function in parallel to UNC-13 in neurons and/or downstream of UNC-13 in the vulval muscles to promote egg laying.
Discussion
In this study we demonstrated the cellular specificity of Gαq signaling as it regulates egg-laying circuit activity and behavior using molecular genetics, optogenetics, pharmacology, and calcium imaging techniques. We found that Gαq effectors PLCβ and Trio RhoGEF differentially act in neurons and muscles to promote synaptic transmission and egg-laying behavior. PLCβ signals presynaptically in neurons while Trio signals both presynaptically in neurons and postsynaptically in muscles. Although Gαq, PLCβ, and Trio mutants all fail to lay eggs in response to serotonin, optogenetic stimulation of HSNs fully rescued egg laying in PLCβ but not Gαq or Trio RhoGEF mutants, suggesting hyperactivation of the HSNs may be able to bypass PLCβ signaling in some cases. Recent work has shown that the HSNs releases NLP-3 neuropeptides which can promote egg laying even in tph-1 mutants lacking serotonin (Brewer et al., 2019). Because pan-neuronal expression of PLCβ was able to rescue both egg-laying and cholinergic transmission defects of PLCβ null mutants, we propose that Gαq signals through PLCβ to promote acetylcholine release from the VA, VB, and VC neurons, which also innervate the vulval muscles alongside the HSNs (Cook et al., 2019; White et al., 1986). Consistent with this model, we have recently shown that VC synaptic transmission is important for egg laying in response to serotonin (Kopchock et al., 2021). Previous studies have shown that Trio acts in the presynaptic motor neurons to regulate locomotion behavior (Hu et al., 2011; Steven et al., 2005; Williams et al., 2007). Our studies show that Trio overexpression in either neurons or muscles alone is insufficient to restore wild-type level of egg-laying behavior, but expression in both is sufficient. These results further indicate that Gαq signaling through PLCβ and Trio RhoGEF fulfill distinct functions during locomotion and egg-laying behaviors.
How do Gαq, PLCβ and Trio-RhoGEF signaling promote egg-laying behavior? Earlier studies have suggested that Rho ortholog RHO-1 in C. elegans regulates synaptic activity in a mechanism that involves the G12 family protein GPA-12 (Hiley et al., 2006; Lutz et al., 2005). Activated RHO-1 also directly binds to and inhibits the DGK-1 diacylglycerol kinase expressed in neurons that signals to reduce DAG available to bind effectors (Hiley et al., 2006; McMullan et al., 2006). Our data demonstrate that Gαq signaling regulates postsynaptic vulval muscle activity mainly through Gαq-Trio pathway as Gαq, and Trio mutants show a similarly strong reduction in vulval muscle Ca2+ activity. Because muscle activity defects in Gαq and Trio mutants can be restored by the DAG-mimetic PMA, we suggest that insufficient levels of DAG are responsible for the circuit activity and behavior defects of Gαq and Trio RhoGEF mutants. In the absence of PLCβ, how would parallel Gαq signaling through Trio RhoGEF and RHO-1 generate DAG? Besides PLCβ (EGL-8), C. elegans expresses four other PLC orthologs: PLCε (PLC-1), PLC-2, PLCγ (PLC-3), and PLCδ (PLC-4) (Vázquez-Manrique et al., 2008). In vitro studies with cultured mammalian cells show that small G proteins like Rho can bind to and activate PLCε (Seifert et al., 2008; Wing et al., 2003). Genetic evidence in C. elegans suggests a model that both PLCε and PLCγ likely function downstream of Gαq to promote cell activity (Kunitomo et al., 2013; Yu et al., 2013), but whether Trio activation of Rho ultimately acts through these PLCs to produce DAG is not clear. One approach to test if these other PLCs mediate Rho signaling would be to perform genetic epistasis experiments. However, loss of Rho-1 causes lethality (Jantsch- Plunger et al., 2000; McMullan and Nurrish, 2011) and loss of PLCε and PLCγ cause sterility defects, limiting our ability to measure differences in egg laying. Imaging or biochemical experiments documenting changes in PIP2 (Stauffer et al., 1998) and/or DAG (Ohno et al., 2017; Tewson et al., 2012) could allow for directs tests of whether Gαq signaling through Trio and Rho changes PIP2 and DAG levels in vivo.
Previous studies have shown that the DAG mimetic phorbol esters promote both synaptic vesicle and dense core vesicle release from neurons and neurosecretory cells (Silinsky and Searl, 2003). DAG and phorbol esters activate many effectors including mUNC-13 in the brain and Protein Kinase C (PKC) in nearly all cells (Betz et al., 1998; Huang, 1989). In C. elegans, treatment of exogenous PMA promotes hypersensitivity to the paralytic effects of aldicarb (Sieburth et al., 2007). Double mutants of both PKC-1 null and UNC-13 (H17K) show a greater magnitude of resistivity to phorbol esters compared to either mutant alone suggesting that PMA acts in part through these effectors to regulate ACh release (Sieburth et al., 2007; Silinsky and Searl, 2003). Our data show that animals lacking either UNC-13 or single PKC isoforms still show a robust egg-laying response to PMA (Fig. 6), suggesting DAG and PMA promote muscle activity via other targets. In vitro studies show ROCK (Rho-associated coiled-coil kinase) activation in PMA-induced apoptosis and macrophage differentiation (Chang et al., 2006; Yang et al., 2017). ROCK has a predicted C1 domain that mediates protein interaction with DAG and might bind and be similarly activated by PMA (Xiao et al., 2009). In C. elegans, RHO-1 signals through LET-502/ROCK to phosphorylate non-muscle myosin light chain (Shimizu et al., 2018). Gαq also promotes neurotransmitter release via additional kinase targets including SEK-1 Mitogen-Activated Protein Kinase Kinase in the p38 MAPK pathway and KSR-1 in the ERK MAPK pathway (Coleman et al., 2018; Hoyt et al., 2017). KSR-1 is particularly interesting in that its N-terminus shares sequence similarly with C1 domains that might mediate regulation by DAG. Loss of KSR-1 and other ERK MAPK components also suppress the loopy locomotion defects caused by gain-of-function mutations in Rho-1 (Coleman et al., 2018). Taken together, our work is consistent with a model where additional DAG-sensitive effectors act downstream of Gαq to promote muscle contractility during egg laying.
Apart from direct activation of C1 domain containing effectors, emerging evidence indicates DAG can regulate neural circuit activity via endocannabinoid-mediated retrograde signaling. In pyramidal cells of the mouse hippocampus, postsynaptic calcium influx through voltage-gated calcium channels activates PLCβ1 to generate DAG (Hashimotodani et al., 2006; Hashimotodani et al., 2007; Soltesz et al., 2015). Diacylglycerol lipase-α (DAGLα) can convert DAG to Arachidonoylglycerol (2-AG), which signals from dendrites in a retrograde manner through presynaptic endocannabinoid receptors to inhibit neurotransmitter release (Hashimotodani et al., 2013; Hashimotodani et al., 2005; Soltesz et al., 2015; Tanimura et al., 2010; Wettschureck et al., 2006). In C. elegans, 2-AG activates the NPR-19 endocannabinoid receptor ortholog that couples to Gαo to modulate serotonin transmission, pharyngeal, feeding, and locomotory behaviors (Oakes et al., 2019; Oakes et al., 2017; Pastuhov et al., 2016). We have recently shown that feedback of egg accumulation alters vulval muscle Ca2+ activity, which subsequently signals to regulate bursting in the HSNs (Ravi et al., 2018b; Ravi et al., 2020), These results support a model where stretch-dependent feedback of egg accumulation stimulates postsynaptic vulval muscle Ca2+ signaling. This Ca2+ would then activate PLCs to generate DAG and 2-AG which signal to modulate HSN activity, serotonin release, and egg laying. The genetic and experimental accessibility of the C. elegans egg-laying circuit should allow us to determine if conserved G proteins like Gαq act generally to drive neural circuit activity via changes in DAG, subsequent activation of effectors, and retrograde messengers like 2-AG.
Movie 1: GCaMP5:mCherry fluorescence ratio in the vulval muscles of wild-type animals in the absence or presence of 10 µM DAG analogue, PMA. Blue indicates low Ca2+, and red indicates elevated Ca2+.
Movie 2: GCaMP5:mCherry fluorescence ratio in the vulval muscles of strong loss of function Gαq mutant, egl-30(ad805), in the absence or presence of 10 µM DAG analogue, PMA. Blue indicates low Ca2+, and red indicates elevated Ca2+.
Movie 3: GCaMP5:mCherry fluorescence ratio in the vulval muscles of strong loss of function Trio mutant, unc-73(ce362), animals in the absence or presence of PMA. Blue indicates low Ca2+, and red indicates elevated Ca2.
Acknowledgements
This work was funded by grants from the NIH (R01-NS086932) and NSF (IOS-1844657) to KMC. PD is supported by AHA predoctoral fellowship Award (20PRE35210233). We thank Dr. Kenneth Miller for sharing plasmids and strains. Some of the strains used in this study were provided by the C. elegans Genetics Center, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). We thank Drs. Julia Dallman, Laura Bianchi, and Athula Wikramanayake along with members of the Collins lab for helpful discussions and feedback on the manuscript.