VAV-1 acts in a single interneuron to inhibit motor circuit activity in Caenorhabditis elegans

Abstract

The complex molecular and cellular mechanisms underlying neuronal control of animal movement are not well understood. Locomotion of Caenorhabditis elegans is mediated by a neuronal circuit that produces coordinated sinusoidal movement. Here we utilize this simple, yet elegant, behavior to show that VAV-1, a conserved guanine nucleotide exchange factor for Rho-family GTPases, negatively regulates motor circuit activity and the rate of locomotion. While vav-1 is expressed in a small subset of neurons, we find that VAV-1 function is required in a single interneuron ALA to regulate motor neuron circuit activity. Furthermore, we show by genetic and optogenetic manipulation of ALA that VAV-1 is required for the excitation and activation of this neuron. We find that ALA signaling inhibits command interneuron activity by abrogating excitatory signaling in the command interneurons that is responsible for promoting motor neuron circuit activity. Togther, our data describe a novel neuromodulatory role for VAV-1-dependent signaling in the regulation of motor circuit activity and locomotion.

Introduction

Vav proteins are a family of evolutionarily conserved molecules that contain a series of protein-protein interaction domains, including SH2 and SH3 domains, and a Dbl homology domain that acts as a Rho GTPase family guanine nucleotide exchange factor (GEF). The first identified Vav protein, Vav1, is highly expressed in the immune system, whereas Vav2 and Vav3 are ubiquitous and are highly expressed in the nervous system ^1-4. While the function of Vav in the immune system has been well studied, the understanding of Vav proteins in the nervous system is limited. Nevertheless, Vav proteins have been implicated in axonal growth and guidance, cerebellar development, and plasticity ^3-6. Furthermore, Vav2 and Vav3 single knockout mice exhibit evidence of sympathetic neuron hyperactivity, including elevated release of noradrenaline, adrenaline and dopamine ^7,8, suggesting Vav proteins have a neuromodulatory role in the nervous system.

To explore the molecular mechanisms underlying the regulation of behavior, we are studying locomotion in the model system Caenorhabditis elegans. C. elegans locomotion is governed by a small network of command interneurons that are postsynaptic to glutamatergic sensory neurons and presynaptic to the motor neurons that drive locomotion (Fig 1a) ^9-12. Excitatory signaling from the sensory neurons to the command interneurons promotes locomotion ^13,14. This simple locomotory control circuit and the fact that the C. elegans genome encodes only one Vav family member, vav-1 ¹⁵ make C. elegans a valuable experimental system to help dissect the role of Vav proteins in the nervous system. Previously, vav-1 was shown to regulate several rhythmic behaviors ^15-17. Here, we show that loss of vav-1 function disrupts normal motor circuit activity. While vav-1 mutants exhibit normal sinusoidal locomotion and nervous system development, they display increased locomotory velocity and elevated motor circuit activity. Moreover, by restoring VAV-1 expression in a single interneuron (ALA), the elevated motor circuit activity of vav-1 mutants can be returned to a wild type level. Lastly, we have found evidence that VAV-1-dependent signaling in ALA opposes the activity of the command interneurons that directly promote motor circuit activity and animal locomotion (Fig 1a). Together, our results suggest that VAV-1 has a crucial role in regulating a neural circuit required for controlling motor neuron activity, which controls the rhythm of sinusoidal locomotion.

Figure 1. vav-1 mutant worms are hyperactive.

(a) Schematic of the locomotory control circuit (motor circuit) utilized by C. elegans. Glutamatergic sensory neurons (SN) stimulate the activity of a set of command interneurons (CN), which includes AVA, AVD, AVE and PVC (GluR, glutamate receptor). Activation of the command interneurons promotes locomotion by direct regulation of the motor neurons (MN). VAV-1 acts in ALA to inhibit motor circuit activity by blocking command interneuron activity. (b) vav-1 mutants crawled faster than wild type (WT) animals on an agar surface seeded with bacteria, which was rescued by VAV-1 expression under control of the full vav-1 promoter. n = 5-10 animals in 7 (vav-1; vav-1 rescue), 9 (vav-1), or 11 (WT) replicate assays. Data are represented as mean +/− SEM (*** p < 0.001, * p < 0.05 using one-way ANOVA). (c) WT and vav-1 mutants exhibited characteristic coordinated forward sinusoidal locomotion. Anterior is to the right Scale bar = 100μm.

Results

vav-1 mutants have an elevated rate of locomotion

In C. elegans, loss of function mutations in vav-1 result in severely defective pharyngeal pumping, which leads to larval lethality, since animals cannot feed. Selective expression of vav-1 in pharyngeal tissue completely restores pharyngeal activity and animals develop into sexually mature adults ¹⁵. However, in rescued vav-1 mutant animals, referred to as vav-1 mutants from this point forward, defects in other tissues are observed, including abnormal ovulation, fertilization and defecation cycle timing ¹⁵. While documenting these adult phenotypes, we observed that vav-1 mutants are more active than wild type worms. To quantify the locomotory behavior of vav-1 mutants, we measured the crawling speed of animals moving on an agar surface seeded with bacteria. Consistent with our observations, vav-1 mutants show an increase in crawling velocity compared to wild type animals (Fig. 1b). To determine if this elevated locomotion was the result of disrupted VAV-1 function, we introduced a wild type vav-1 genomic construct into vav-1 mutants and found that this rescued the elevated locomotion phenotype (Fig. 1b). Importantly, although vav-1 mutants travel faster than wild type worms, they maintain a normal sinusoidal mode of locomotion and normal posture (Fig. 1c), indicating coordinated motor output.

vav-1 mutants have elevated motor circuit activity

Since Vav proteins are expressed in the nervous system, we hypothesized that VAV-1 acts in the nervous system to regulate motor activity. In order to investigate the role VAV-1 may have in the nervous system, we employed a pharmacological assay using the small molecule aldicarb. Aldicarb is an acetylcholine (ACh) esterase inhibitor, which prevents the breakdown of the excitatory neurotransmitter ACh at synapses of neuromuscular junctions (NMJs) ¹⁸. In C. elegans, exposure to aldicarb leads to progressive paralysis over time ¹⁸. Moreover, mutations that increase synaptic transmission cause hypersensitivity to aldicarb, and those that decrease synaptic transmission cause resistance to the drug ¹⁸. Strikingly, we found that vav-1 mutants are hypersensitive to aldicarb and become paralyzed more rapidly than wild type animals, indicating excessive synaptic transmission (Fig. 2a). Like the hyperactive locomotion detected in vav-1 mutants, the increased sensitivity to aldicarb can be rescued by expression of wild type vav-1 genomic fragment in vav-1 mutant animals (Fig. 2a). Importantly, aldicarb hypersensitivity is consistent with the increased rate of locomotion observed in vav-1 mutants, suggesting that VAV-1 negatively regulates motor circuit activity. Thus, to test whether vav-1 activity is required in the nervous system, we selectively expressed the vav-1 cDNA in the nervous system of vav-1 mutants using a pan-neuronal promoter from the aex-3 gene ¹⁹. Accordingly, we found that neuronal expression of vav-1 can rescue the aldicarb hypersensitivity caused by loss of vav-1 (Fig. 2b). Similarly, we found that knocking down VAV-1 expression by neuronal-specific RNA interference (RNAi) caused aldicarb hypersensitivity similar to that of vav-1 mutant worms (Fig. 2c). These data suggest that vav-1 activity is required in the nervous system to inhibit motor circuit activity.

Figure 2. vav-1 mutants exhibit elevated motor neuron excitability.

(a) and (b) Animals were paralyzed with 1mM aldicarb. (a) vav-1 mutants were hypersensitive to aldicarb, and vav-1 expression driven by the full vav-1 promoter (vav-1; vav-1 rescue) rescued this hypersensitivity. (b) Rescue of vav-1 mutant aldicarb hypersensitivity by a pan-neuronal promoter, aex-3 (vav-1; neuronal rescue) is shown. (c) Neuronal-specific RNAi of vav-1 caused hypersensitivity in a 0.5mM aldicarb paralysis assay, similar to the effect of vav-1 mutation. Control: empty vector (ev). For (a)-(c), n= 15-20 animals in ≥ 5 replicate assays. (d) and (e) Cholinergic neurons or GABAergic neurons were specifically activated using channelrhodopsin (ChR2). (d) vav-1 mutant worms contracted significantly more than WT upon cholinergic stimulation (percent change of pre-stimulus worm length). n = 48 (+ATR) or 27 (−ATR controls) animals per genotype. (e) vav-1 mutant worms relaxed significantly more than WT upon GABAergic stimulation (percent change of pre-stimulus worm length). n = 37 (+ATR) or 27 (−ATR) animals per genotype. Data are displayed as mean +/− SEM (*** p <0.001 using two-tailed T tests). (f) Animals were paralyzed with 100μM levamisole. vav-1 mutants showed a normal muscle response to levamisole. Aldicarb and levamisole paralysis curves display the mean proportions of moving or paralyzed worms, +/− SEM (statistical significance determined by two-way ANOVA and Bonferroni post-tests; see Methods for detailed analysis).

In order to further explore the neuronal function of VAV-1, we utilized transgenic animals that express channelrhodopsin-2 (ChR2) specifically in cholinergic (excitatory) and GABAergic (inhibitory) motor neurons. ChR2s are light-gated ion channels that nonspecifically allow the flow of cations into a cell, leading to electrical excitation ²⁰. Thus, activation of ChR2 in cholinergic neurons results in a visible and measurable body wall muscle contraction as quantified by the percent change in overall length of the animal (Fig. 2d)^20,21. In contrast, activation of ChR2 in the GABAergic neurons results in a visible and measurable body wall muscle relaxation (Fig. 2e)²¹. Activation of cholinergic ChR2 in the vav-1 mutant background resulted in enhanced muscle contraction compared to wild type animals (Fig. 2d). Additionally, we found that activation of GABAergic ChR2 resulted in enhanced relaxation in the vav-1 mutant background compared to the wild type background (Fig. 2e). These findings suggest that vav-1 mutants have hyperexcitable motor neurons, and further indicate that VAV-1 may play an inhibitory role in regulating motor circuit activity.

However, it is possible that the vav-1 mutant aldicarb hypersensitivity and excessive response to ChR2 activation are caused by muscle cell defects. For instance, heightened muscle cell response to neurotransmitters could explain these phenotypes. Therefore, we tested the sensitivity of vav-1 mutants to levamisole, an ACh receptor agonist. Similar to aldicarb, levamisole causes paralysis over time. Levamisole activates postsynaptic muscle cells, so paralysis occurs independently of presynaptic (neuronal) defects ¹⁸. Analysis of levamisole treated animals revealed that vav-1 mutants responded like wild type animals (Fig. 2f), indicating normal activity of muscle in vav-1 mutants. Taken together, our data indicate that VAV-1 acts in the nervous system to regulate motor circuit activity.

Since internal metabolic signals resulting from being well-fed or starved can influence the rate of locomotion of C. elegans ²², and mutations in genes affecting feeding can lead to aldicarb hypersensitivity ²³, it is possible that vav-1 mutants do not feed as well as wild type animals, resulting in enhanced locomotion and aldicarb sensitivity. However, this is unlikely because mutant vav-1 animals develop at a similar rate as wild type animals; we regularly obtain synchronized populations of adult wild type and vav-1 animals for use in assays described here. Furthermore, we investigated fat accumulation in adult vav-1 mutants by Oil-Red-O staining ²⁴ and found that vav-1 mutants were indistinguishable from wild type animals (Supplementary Fig. 1). These data indicate that vav-1 mutants develop at a similar rate and have similar metabolic activity as wild type animals.

vav-1 is expressed in different classes of neurons

To determine which neurons may require VAV-1 function for normal locomotion and motor circuit activity, we used a standard reporter gene fusion approach ²⁵ to express fluorescent proteins (GFP and mCherry) under control of the vav-1 promoter ¹⁵. For this analysis, we used the same genomic sequence for the vav-1 promoter that rescued the elevated crawling speed and aldicarb hypersensitivity of vav-1 mutants when driving expression of wild type vav-1 (Fig. 1b and 2a). The fluorescent proteins were targeted to the nucleus by a nuclear localization signal to aid in cell identification. We observed fluorescence in expected tissues, such as the pharynx and intestine ¹⁵, as well as in a small set of neurons. In five independent reporter lines generated, fluorescence was regularly observed in the ALA interneuron. To confirm the identity of this cell, we crossed a vav-1 mCherry reporter line into animals carrying an ALA neuron marker, IDA-1::GFP , and observed co-localization (Fig. 3a). Interestingly, the ALA interneuron has been implicated in regulating sleep-like behavior ²⁶. Additionally, we observed variable reporter expression in cholinergic motor neurons in the ventral nerve cord (Fig. 3b), which are required for locomotion, but did not observe expression in GABA motor neurons of the ventral nerve cord. Rather, the vav-1 reporter was expressed in two GABA motor neurons in the head, known as RME dorsal and RME ventral, which regulate head oscillations (Fig. 3c). Several cell types were found to be negative for vav-1 reporter expression. We found that our vav-1 reporters did not show co-expression with markers of dopaminergic neurons (Pdat-1::GFP) ²⁷, command interneurons of the locomotory circuit, which directly activate motor neurons (Pglr-1::RFP)²⁸, or with a marker of a small set of sensory neurons (AWC and ASE neurons, marked by Pceh-36::RFP)²⁹ (Supplementary Fig. 2). In sum, fluorescent vav-1 reporters are expressed in a small subset of neurons that comprise multiple cell types, including an interneuron, cholinergic motor neurons, and a small subset of GABAergic motor neurons.

Figure 3. VAV-1 is expressed in multiple neuronal cell types.

(a, b, c) Expression of mCherry, tagged with a 2x-nuclear localization signal, under control of the vav-1 promoter, is shown. (a) vav-1 reporter expression co-localized with a GFP marker for the ALA interneuron. Anterior is to the left. (b) vav-1 reporter expression was found in cholinergic motor neurons of the ventral nerve cord. Shown are two posterior cholinergic motor neurons (VA11 and VA12) and their colocalization with cytoplasmic cholinergic GFP reporter. Arrows mark these two cell bodies. Anterior is to the left. (c) Only two GABAergic neurons (RME dorsal and ventral, head motor neurons) expressed the vav-1 reporter. Asterisks mark vav-1 expression in the pharynx, a dashed arrow marks the ALA neuron, and solid arrows mark the 2 RME neurons that express the vav-1 reporter and co-localize with a GABA neuron specific GFP marker. Anterior is up. Scale bars = 5μm.

VAV-1 is not required for neuronal development

Previous studies have indicated that Vav proteins in Drosophila and in mice are important for neuronal development ^3-5. Specifically, these studies have shown that Vav proteins are required for axonal guidance and cell migration events in the developing nervous system. Since defects in production or release of the inhibitory neurotransmitter GABA can cause aldicarb hypersensitivity due to increased excitability at the neuromuscular junction (NMJ) ^30,31, we reasoned that improper development (e.g. differentiation, axon guidance or cell migration) of GABAergic neurons or impaired GABA release could explain the aldicarb hypersensitivity phenotype of vav-1 mutants (Fig. 2a). Though we did not observe vav-1 reporter expression in ventral nerve cord GABAergic motor neurons in adult animals, we cannot rule out embryonic expression or non-cell autonomous roles in GABAergic neuronal development. Thus, to examine GABAergic motor neuron development and synapse formation, we utilized existing transgenic strains carrying either a soluble GFP or a GFP synapse marker (GFP-tagged synaptobrevin, GFP::SNB-1), each driven by a GABAergic neuron-specific promoter ³². These transgenes were introduced into the vav-1 mutant background. Analysis of the soluble GFP marker in vav-1 mutants indicated that all 26 GABAergic neurons differentiated and were present in their proper locations, and circumferential axon projections (commissures) were present and did not display defects in connectivity (Fig. 4a and 4b). Furthermore, using the GFP::SNB-1 marker ³², we could identify and characterize presynaptic regions between motor neurons and adjacent muscle cells in the dorsal nerve cord (at the NMJ), which appear as fluorescent puncta (Fig. 4c). Comparing the distribution and fluorescence intensity of mutant and wild type synaptic puncta can reveal defects in synaptic vesicle organization and kinetics ³³. As indicators of synaptic development and function, we used four parameters to analyze the synapses of vav-1 mutants: the number of synaptic puncta per 20μm of the dorsal nerve cord (density), the area of puncta, the fluorescence intensity of puncta, and the interpunctal fluorescence intensity (intensity of the axon between synaptic puncta). From these analyses, we found vav-1 mutant animals had a normal number of GABAergic synaptic puncta per length of dorsal nerve cord (Fig. 4d, left Y axis), and the puncta were of normal size (Fig. 4d, right Y axis), which indicates normal development. Moreover, fluorescence intensity of puncta (Fig. 4e, left Y axis) and fluorescence intensity of the dorsal nerve cord between puncta (interpunctal intensity)(Fig. 4e, right Y axis) in vav-1 mutants were indistinguishable from wild type animals, suggesting normal synaptic activity.

Figure 4. Loss of VAV-1 does not lead to defects in neuronal development or synapse formation.

(a) Representative confocal micrographs of WT (top panel) and vav-1 mutant (bottom panel) animals expressing GFP in all GABAergic neurons (oxIs12 [Punc-47::GFP]). In WT (top panel), arrows and lines point to GABA neuron representative cell bodies and circumferential axons (commissures), respectively. In vav-1 mutants (bottom panel), these features are also present. Dashed arrows indicate fluorescent coelomocytes, a marker of the vav-1 mutant strain that should be ignored. Asterisk marks non-specific posterior intestinal GFP. Anterior is to the left, and scale bars = 100μm. (b) Quantification of the number of GABA neuron cell bodies and commissures indicates that vav-1 mutants showed grossly normal GABA neuron development. n= 13 (WT) and 19 (vav-1). (c) Representative images of GABAergic neuron synapses and (f) cholinergic synapses in the dorsal nerve cord of WT (left) and vav-1 mutant (right); fluorescent puncta are synaptobrevin tagged with GFP at the neuromuscular junction (NMJ). GABAergic animals are juIs1 [Punc-25::GFP::snb-1], and cholinergic are nuIs152 [Punc-129::GFP::snb-1]. (d) and (g) Quantification of density of fluorescent puncta (left Y axis) and the mean area of puncta (right Y axis). For (d), n= 10 (WT) and 11 (vav-1). For (g), n= 11 (WT) and 16 (vav-1). (e) and (h) Quantification of fluorescence intensity of puncta (left Y axis), and of nerve cord between puncta (right Y axis). For (e), n= 10 (WT) and 11 (vav-1). For (h), n= 28 (WT) and 36 (vav-1). (i) The structure of the ALA neuron in WT and vav-1 mutants is shown using a fluorescent ALA reporter analyzed by confocal microscopy (Pida-1::ida-1::GFP). In the left panels, solid arrows point to the ALA neuron cell body and a dashed arrow points to one ALA axonal projection. The right panels show the same axons projecting to the tails of these animals. Asterisks mark GFP associated with coelomocytes and rectal epithelial cells, and can be ignored. n= 3(WT) and 11 (vav-1). Quantified data are displayed as mean +/− SEM and were analyzed by two-tailed Student's T tests (no significant differences were found).

To further investigate GABA motor neuron function, we utilized a pharmacologic approach. Pentylenetetrazol (PTZ), a GABA_A receptor antagonist, induces seizure like behaviors in mutants with defects in GABA signaling ³⁴. We exposed vav-1 mutants to PTZ and found that, unlike mutants with defects in GABA production, vav-1 mutants did not display seizure-like behavior (Table 1). These data indicate that GABA motor neuron activity is not impaired in vav-1 mutants. This conclusion is further supported by the response of vav-1 mutants to levamisole. Indeed, mutations that disrupt GABAergic motor neuron function cause levamisole hypersensitivity due to faulty inhibitory neurotransmission to the body wall muscle ^30,31, whereas we observed a normal response to levamisole in vav-1 mutants (Fig. 2f). Together, these data indicate that GABAergic neurons in vav-1 mutant animals develop normally and form proper synapses.

Table 1.

Response of vav-1 mutants to 6mM PTZ

Genotype	% Animals Displaying Anterior Seizure 30 min	% Animals Displaying Anterior Seizure 60 min
WT	0	0
unc-25 (e156)/GAD*	70	87
vav-1 (ak41)	0	0

^*unc-25 encodes a glutamic acid decarboxylase

Since vav-1 reporter expression was observed in the cholinergic motor neurons, we next investigated the organization of these neurons and their synapses in vav-1 mutants. First, we analyzed the development of the cholinergic nervous system using a cholinergic-specific soluble mCherry (Punc-17::mCherry)³⁵ in the vav-1 mutant background. From this analysis, consistent with their normal sinusoidal locomotion, we found that cholinergic motor neurons of vav-1 mutants have normal cell positions, axon development and morphology, and were indistinguishable from wild type animals (Supplementary Fig. 3). To further examine cholinergic neuron development, we investigated cholinergic synapse formation using a transgenic strain that expresses the same GFP-tagged synaptobrevin as in Fig. 4c, but specifically in cholinergic motor neurons (Punc-129::GFP::SNB-1) ³³. We found that, like GABA motor neuron synapses, synapses at cholinergic NMJs are present and appear to be organized normally in vav-1 mutants (Fig. 4f). Examination of puncta number (Fig. 4g, left Y axis), the size of individual synaptic puncta (Fig. 4g right Y axis), as well as the fluorescence intensity of puncta (Fig. 4h, left Y axis) and interpunctal intensity (Fig. 4h, right Y axis) in vav-1 mutants revealed no differences from wild type animals, indicating that VAV-1 does not have a role in cholinergic motor neuron synaptic development.

As vav-1 reporters are expressed in the ALA interneuron, we also investigated the development of this neuron in vav-1 mutants. Using a fluorescent ALA reporter, IDA-1::GFP, we found that the ALA cell body is visible in the dorsal ganglion adjacent to the nerve ring of wild type and vav-1 animals (Fig. 4i, left 2 panels). Additionally, axonal projections that extend laterally on both sides of the animal to the posterior end were also present in wild type and vav-1 animals (Fig. 4i, right 2 panels). Therefore, collectively, these data indicate that VAV-1 is not required for normal neuronal development, but may instead play a role in neuronal function.

VAV-1 is required in ALA and requires GEF function

In order to determine which neuronal cells require VAV-1 for proper modulation of motor circuit activity, we expressed VAV-1 in different neuronal cell types using specific promoters. Since we observed strong expression with our vav-1 reporter in ALA and variable but consistent expression in the ventral cord cholinergic motor neurons, which directly coordinate locomotion, we used the following promoters to drive vav-1 expression in these neurons: the ver-3 promoter to drive expression in the ALA neuron ²⁶ and the unc-17 promoter to drive expression in cholinergic motor neurons ³⁶. To test for rescue, these neuronal transgenic strains were then crossed into the vav-1 mutant background. For simplicity, the two genotypes are thus referred to as vav-1; ALA rescue and vav-1; cholinergic rescue. These animals were assayed for aldicarb sensitivity in comparison to wild type and vav-1 mutants. Interestingly, expression of VAV-1 in cholinergic neurons caused heightened aldicarb sensitivity in the vav-1 mutant (Fig. 5a). Surprisingly, expression of VAV-1 in a single interneuron, ALA, returns the aldicarb sensitivity of vav-1 mutants to a wild type level (Fig. 5b). Thus, these data indicate that VAV-1 function in the ALA interneuron is sufficient for normal motor circuit activity. Since VAV-1 is a conserved GEF for the Rho GTPase family ^1,15, we examined whether VAV-1 requires an active GEF domain to regulate motor circuit activity. To accomplish this, a mutant vav-1 construct that lacks GEF function ¹⁵ was introduced into the vav-1 mutant background. We then tested the ability of GEF-dead VAV-1 to rescue the vav-1 mutant aldicarb hypersensitivity, and found that unlike the wild type construct (Fig. 2a), the mutant GEF construct could not restore a normal aldicarb response (Fig. 5c). Therefore, these data suggest that VAV-1 activates a Rho/Rac GTPase to mediate its neuromodulatory function. Furthermore, since we observed an elevated locomotion rate in vav-1 mutants, we measured the rate of locomotion of vav-1; ALA rescue animals and found that expression of VAV-1 in the ALA neuron restores normal locomotory speed to vav-1 mutants (Fig. 5d). We also confirmed that VAV-1 lacking GEF activity cannot rescue the elevated crawling speed of vav-1 mutants (Fig 5d). These data indicate that VAV-1 in a GEF-dependent manner is required in ALA for normal motor circuit activity.

Figure 5. VAV-1 modulates motor neuron activity via the ALA neuron.

Specific neuronal subsets were targets for VAV-1 expression in vav-1 mutant animals to assay for rescue of aldicarb hypersensitivity (0.5mM). (a) Expression of VAV-1 in cholinergic neurons did not rescue aldicarb hypersensitivity of vav-1 mutants. (b) Expression in the ALA neuron rescued aldicarb hypersensitivity. (c) The GEF-inactive VAV-1 (Pvav-1::vav-1 GEF-dead) was not capable of rescuing vav-1 mutant aldicarb hypersensitivity. In (a)-(c), n= 15-20 animals in ≥ 5 replicate assays. Data are represented as mean +/− SEM (statistical significance determined by two-way ANOVA and Bonferroni post-tests; more detailed information in Methods). (d) Elevated crawling speed of vav-1 mutants was rescued by VAV-1 expression in ALA, but not by expression of GEF-dead VAV-1 under its own promoter. n≥ 20 animals in 5 replicate assays. Data are represented as mean +/− SEM (*p <0.05, ** p <0.01 using one-way ANOVA and Bonferroni post-tests).

ALA inhibits motor circuit activity

While we have found that VAV-1, via its Rho/Rac GTPase exchange activity, regulates motor circuit activity from the ALA neuron, the role(s) of VAV-1 in ALA, and of ALA in regulation of motor circuit activity, are not clear. ALA has been shown to electrically couple to the interneuron RID in the nerve ring and form chemical synapses with the command interneurons AVA, AVE (in the nerve ring) and possibly PVC (in the posterior ganglion)³⁷. Thus, ALA has few synaptic partners and does not directly synapse onto motor neurons. Additionally, it is not clear what neurotransmitter(s) ALA releases. However, it has been proposed that the ALA neuron is peptidergic ³⁸; ALA expresses proteins involved in dense core vesicle-mediated signaling ³⁹. To gain insight into the function of VAV-1 in ALA, we examined the subcellular localization of a functional VAV-1::GFP expressed under its own promoter in ALA. We found that VAV-1::GFP is localized to the cell body of ALA, but not in the nucleus. Also, we did not observe VAV-1::GFP localization in the axons along the length of the animal (Fig. 6a). These data indicate that VAV-1 does not associate with synaptic or dense core vesicles or their release sites within ALA and suggest that VAV-1 may not have a direct role in regulating vesicle release from the ALA neuron. However, these observations do not address a potential distal role of VAV-1 in neuropeptide or neurotransmitter release from ALA.

Figure 6. VAV-1 is required for excitation of ALA, which inhibits motor circuit activity.

(a) Representative confocal image of functional VAV-1::GFP subcellular localization (under control of its own promoter). The left image shows VAV-1::GFP expression in the head, including the pharynx (bracket) and the ALA neuron (white dashed box). The right image shows VAV-1::GFP expression in the cell body of ALA; there was no observable fluorescence in the ALA nucleus or axons. (b) Schematic of strategy used to drive ALA-specific expression of tetanus toxin (TeTx) or ChR2*(C128S), using a promoter-intersectional, inverted Cre-Lox system. (c) Interfering with vesicle release using ALA-specific TeTx induced aldicarb hypersensitivity, like vav-1 mutation (0.5 mM aldicarb). (d) Activation of ALA using ChR2*(C128S) induced aldicarb resistance (0.5 mM aldicarb). (e) Activation of ALA using ChR2*(C128S), in the vav-1 mutant background, rescues aldicarb response to wild type sensitivity (0.5 mM aldicarb). Blue stripes indicate 1-minute pulses of low intensity blue light. n= 15-20 animals in ≥ 5 replicate assays. Data are represented as mean +/− SEM (statistical significance determined by two-way ANOVA and Bonferroni post-tests; more detailed information in Methods).

To address whether vesicle release from ALA is involved in its modulation of motor circuit activity, we drove ALA-specific expression of tetanus toxin light chain, which blocks release of vesicles containing neurotransmitters and neuropeptides by cleaving the vesicle protein synaptobrevin ^40-43. To accomplish this, we exploited a cell-specific, inverted Cre-Lox expression system ^44,45. This approach involves one promoter driving expression of a floxed inverted tetanus toxin (TeTx) and GFP construct, and a second promoter driving Cre. Thus, by promoter overlap the expression of TeTx occurs in only one cell (Fig. 6b). We confirmed by fluorescence microscopy that GFP, and hence TeTx, was expressed solely in the ALA neuron (Supplementary Fig. 4a). Utilizing this strain, we found that blocking vesicle release by TeTx in ALA induces aldicarb hypersensitivity in otherwise wild type animals (Fig. 6c) similar to vav-1 mutants. These data suggest that ALA releases vesicles, perhaps containing neuropeptides or another neurotransmitter, to inhibit motor circuit activity.

Then, we performed a complimentary experiment, where we photoactivated the ALA neuron using ChR2. We used the same promoter-intersectional, inverted Cre-Lox strategy to drive expression of a slowly inactivating variant of ChR2*(C128S) ^46,47 specifically in ALA (Fig. 6b, Supplementary Fig. 4b). Due to the length of time needed to complete an aldicarb assay, we chose this particular ChR2*(C128S) variant so that ALA activation could be prolonged beyond the duration of the blue light stimulus, allowing us to investigate the aldicarb sensitivity of ALA-activated animals over the course of hours. Animals were exposed to a 1-minute pulse of low intensity blue light before, and every 40 minutes during, the aldicarb assay. Strikingly, we found that ALA-photoactivated animals were aldicarb resistant, indicating that, consistent with the ALA::TeTx results, the ALA neuron provides an inhibitory input to the motor circuit (Fig. 6d). In order to determine if VAV-1 may play a role in regulating the activation state of ALA, we tested whether the aldicarb hypersensitivity of vav-1 mutants could be rescued by photoactivating ALA. We found that, indeed, the vav-1 mutant aldicarb response is returned to a wild type sensitivity, albeit not to the level of aldicarb resistance observed in ALA::ChR2*(C128S) animals (Fig. 6e). Nevertheless, our results suggest that the role of VAV-1 in ALA is to regulate the activity of this neuron, and the effect of vav-1 loss can be bypassed by photoactivating ALA.

VAV-1 abrogates command interneurons signaling

To further investigate the role of VAV-1-dependent signaling from ALA in mediating motor circuit activity, we investigated the potential role of the command interneurons in regulating motor neuron activity. The command interneurons AVA, AVE and PVC are postsynaptic to ALA ³⁷ and promote forward and backward locomotion by directly regulating motor neurons ⁹. AVA, AVE and PVC have been shown to express glutamate receptors glr-1 and nmr-1 ⁴⁸. Since these neurons promote locomotion, we hypothesized that VAV-1-dependent signaling in ALA inhibits the activity of the AVA, AVE and PVC command interneurons, thus inhibiting motor neuron activity and locomotion. In support of this hypothesis, activated mutations in glr-1 result in aldicarb hypersensitivity ⁴⁹. Therefore, we reasoned that if the command interneurons are a link between ALA and regulation of motor neuron activity, ablating these neurons should block the effect of vav-1 loss on aldicarb sensitivity (Fig. 7a). Similarly, if we remove or reduce glutamatergic (excitatory) signaling in the command interneurons, we should decrease motor circuit activity (Fig. 7a). Conversely, we hypothesized that if glutamate receptor signaling in the command interneurons is downstream of VAV-1 in ALA, activating glutamate receptor signaling should produce aldicarb hypersensitivity that is not synergistic with vav-1 mutant hypersensitivity (Fig. 7a).

Figure 7. VAV-1 acts in ALA to inhibit the command interneurons of the locomotory circuit to oppose motor circuit activity.

(a) Schematic depicting strategies to determine potential influence of command interneurons (CN) on activity of the motor circuit, and their role in relation to VAV-1 in ALA. We used ablation of the CN with miniSOG (top panel), reduction of glutamate signaling in the CN using glutamate receptor (GluR) mutants (middle panel), and activation of glutamate signaling in the CN using gain-of-function GluR mutants (bottom panel). (b) Ablating the CN with miniSOG does not affect aldicarb sensitivity in a wild type background, but prevents aldicarb sensitivity in the vav-1 mutant background. (c) A double glutamate receptor mutant, nmr-1; glr-1, shows wild type aldicarb response, and double nmr-1; glr-1 mutation prevents aldicarb hypersensitivity of vav-1 mutants. (d) Gain-of-function GluR mutants glr-1 (nd38) and Pnmr-1::Lurcher are aldicarb hypersensitive, like vav-1 mutants, and a double glr-1 (nd38); vav-1 mutant is not more aldicarb hypersensitive than either mutant alone. n= 15-20 animals in ≥ 5 replicate assays. Data are plotted as mean +/− SEM (statistical significance determined by two-way ANOVA and Bonferroni post-tests; more detailed information in Methods).

To test these hypotheses, we first killed the command interneurons using an existing transgenic strain carrying cell-specific miniSOG (mini singlet oxygen generator), allowing for light-inducible cell ablation ⁵⁰. We found that ablating the command interneurons (including AVA, AVD, AVE, PVC) alone did not alter aldicarb sensitivity, but returned the aldicarb response of vav-1 mutants to wild type sensitivity, indicating that these neurons are downstream of vav-1 in the ALA neuron in regulating motor circuit activity (Fig. 7b).

Then, we generated a glr-1; nmr-1 double loss of function mutant and a glr-1; nmr-1; vav-1 triple loss of function mutant and tested aldicarb sensitivity. The glr-1; nmr-1 mutant showed a normal response to aldicarb (Fig. 7c). However, we found that introduction of the glr-1 and nmr-1 mutations into the vav-1 mutant background completely restore a normal aldicarb response in vav-1 animals (Fig. 7c). These data suggest that VAV-1-dependent signaling in ALA is required to abrogate excitatory glutamatergic signaling in the command interneurons and, thus, inhibit the motor circuit. Finally, we examined the aldicarb sensitivity of an activated glr-1 mutant reported to be aldicarb hypersensitive, glr-1(nd38) ⁴⁹, as well as a second activated glr-1 mutant that is equivalent to the δ2 glutamate receptor subunit mutation found in the Lurcher mouse ⁵¹. Importantly, this second activated glutamate receptor is under control of the nmr-1 promoter, so is expressed in fewer cells than glr-1, specifically including the command interneurons and 2 others neurons (AVG and RIM). We found that, indeed, both of these activated glutamate receptor mutants are aldicarb hypersensitive to a similar extent as vav-1 mutants, and the glr-1(nd38); vav-1 double mutant is not significantly more aldicarb hypersensitive than either single mutant alone (Fig. 7d). In support of our hypotheses, these data indicate that glutamate receptor excitatory signaling in the command interneurons is likely downstream of the inhibitory effect of VAV-1 activity in ALA on the motor circuit.

Discussion

Previously, VAV-1 was shown to be a critical regulator of several rhythmic activities in C. elegans (e.g. pharyngeal pumping, ovulation and the defecation motor program) ^15-17. Here, we show that the rhythmic behavior of locomotion is also abnormal in vav-1 mutants; these animals show hyperactive locomotion due to loss of vav-1 in the nervous system. Accordingly, vav-1 mutants are hypersensitive to aldicarb, an ACh esterase inhibitor, which suggests heightened ACh release from the motor neurons. Furthermore, photoactivation of the cholinergic motor neurons or GABAergic motor neurons in vav-1 mutants results in excessive muscle contraction or excessive muscle relaxation, respectively. Intriguingly, we found that expression of vav-1 in a single interneuron ALA could rescue the locomotory and aldicarb hypersensitivity phenotypes of vav-1 mutants. These data indicate that the ALA interneuron, in a VAV-1-dependent manner, negatively regulates motor activity. Subsequently, we found evidence that ALA is an excitable neuron that when photoactivated can repress locomotory circuit activity. Additionally, photoactivation of ALA was found to bypass the loss of VAV-1 function suggesting that VAV-1 is required for ALA signaling. Moreover, we found that vesicular release from ALA is crucial for repression of motor circuit activity.

ALA is a single interneuron that is positioned in the anterior nerve ring of the head. It has two processes that extend to the posterior which run along the lateral nerve cords until they reach the tail. Additionally, ALA sends a short process into the dorsal nerve cord. Although it is believed to be peptidergic³⁸, no peptides or neurotransmitters have yet been shown to have a role in ALA. Additionally, ALA has few synaptic partners; ALA exhibits electrical coupling to the RID interneuron and chemical synapses with the command interneurons AVA, AVE and PVC ^37,52. Thus, ALA does not have a direct connection with any motor neurons. How does VAV-1 in ALA regulate motor neuron activity? Interestingly, the command interneurons are integral components of the locomotory control circuit ^9,10,37, which directly promotes the activity of the motor neurons^13,14 (Fig. 1a). The motor neurons themselves are thought to be the central pattern generator that produces rhythmic sinusoidal crawling ⁵³ and the command interneurons promote their activity ^13,14. Thus, the command interneurons provide an ideal neural circuit in which ALA could participate to negatively regulate motor neuron activity and locomotion. In support of this notion, we have found that the aldicarb hypersensitivity phenotype of vav-1 mutants can be rescued by ablation of the command interneurons or removal of excitatory glutamate receptors from the command interneurons. Additionally, constitutive activation of excitatory glutamate signaling in the command interneurons leads to aldicarb hypersensitivity similar to vav-1 mutants and introduction of the vav-1 mutation into the activated glutamate signaling background does not enhance the aldicarb hypersensitivity of the single mutants, suggesting they act in a common pathway. Together, the neural connectivity of ALA with the command interneurons and our genetic and pharmacological data support a model where ALA in a VAV-1 dependent manner acts to abrogate command interneuron function in promoting motor neuron activity. Previously, ALA has been found to have a role in behavioral quiescence (suppression of pharyngeal pumping and locomotion during lethargus) and it is required to suppress egg-laying activity after strong mechanical stimulation ^26,54. However, it is unclear how ALA suppresses these motor activities. Further studies will be required to investigate whether the motor circuit described here is involved in these other ALA mediated behaviors.

Additionally, we found that the GEF activity of VAV-1 is required for regulating motor circuit activity. Vav proteins, in addition to the Dbl homology domain, which contains the GEF activity, contain several protein-protein interaction domains, and some studies have revealed GEF independent activity for Vav proteins, suggesting Vav proteins can act as molecular adaptor proteins ^1,55-57. Nevertheless, our findings indicate that the role of VAV-1 in the regulation of motor circuit activity requires Rho/Rac GTPase function. The C. elegans Rho-family GTPases are widely expressed in most, if not all, neurons ^58,59 and VAV-1, in vitro, can promote the exchange of GDP for GTP and activate all the Rho GTPase family members in C. elegans ¹⁵. Thus, one or multiple Rho-family GTPases could mediate the effects of VAV-1 in ALA.

Mammalian Vav proteins have been reported to be highly expressed in the nervous system; however, their role in neurons is not well understood. Nevertheless, analysis of Vav2 and Vav3 knockout mice has revealed a remarkable parallel with the C. elegans vav-1 mutant (e.g. nervous system hyperactivity) suggesting a conserved mechanism. Indeed, both Vav2 and Vav3 knockout mice display sympathetic nervous system hyperactivity, which results in cardiovascular and renal dysfunction ^7,8. However, unlike C. elegans vav-1 mutants, which do not show any detectable nervous system developmental defects, Vav3 is required for proper GABAergic axon guidance events in the ventrolateral medulla, which is a brainstem region that regulates sympathetic nervous system activity ⁴. Thus, the function of Vav3 in the mammalian nervous system has diverged or has an expanded role compared to VAV-1 in C. elegans in the regulation of nervous system activity. Nevertheless, the mechanism underlying the elevation in sympathetic neuronal activity in Vav2 knockout mice is not known ⁴. Moreover, human Vav3 was recently identified as one of the top candidate genes in a Japanese genome-wide association study of schizophrenia ^60,61, indicating the importance of understanding the role of Vav proteins in the nervous system. In conclusion, our studies on VAV-1 in the C. elegans nervous system have demonstrated a novel role of VAV-1 in the regulation of a locomotory control circuit and indicate that C. elegans provides a powerful system to further explore the neuromodulatory role of Vav function in a developmentally intact nervous system.

Methods

C. elegans strains and maintenance

Animals were grown at ~20°C on nematode growth media (NGM) plates seeded with OP50 Escherichia coli. Experiments were performed at ~24°C. Strain maintenance was performed as described ⁶². N2 Bristol was the wild type strain. All analyses were carried out on young adult hermaphrodite worms. Strains were synchronized using hypochlorite treatment ⁶² and allowed to grow (~2.5 days at 20°C) to the young adult stage. The mutant and transgenic strains used in this study were: glr-1 (ky176), nmr-1 (ak4), unc-25 (e156), glr-1 (nd38), akEx52 [Pnmr-1::glr-1 (A/T], vav-1(ak41); akEx162 [Ppha-4::vav-1::GFP], vav-1(ak41) takIs5 [Ppha-4::vav-1::GFP], vav-1(ak41); akEx405 [Pvav-1::vav-1::GFP, Psur-5::GFP], vav-1(ak41); akls76 [Paex-3::vav-1::GFP], vav-1; akEx87 [Pvav-1::vav-1::GFP], vav-1 (ak41); takEx81 [Pvav-1::vav-1 L288Q::GFP, Pttx-3::GFP], takEx5 [Pvav-1::2xNLS GFP], takEx6 [Pvav-1::2xNLS mCherry], takEx65 [Punc-17::vav-1::GFP, Pttx-3::GFP], takEx67 [Pver-3::vav-1::GFP, Pttx-3::GFP], takEx136 [Prab-3::lnverse TeTx-SL2-GFP, Pver-3::Cre, Pttx-3::GFP], takEx132 [Prab-3::lnverse ChR2 (C128S)::GFP, Pver-3::Cre, Pttx-3::GFP], inIs181; inIs182 [Pida-1::ida-1::GFP], vtIs1 [Pdat-1::GFP], otIs151 [Pceh-36::RFP], odIs6 [Pglr-1::RFP], juIs1 [Punc-25::GFP::snb-1], nuIs152 [Punc-129::GFP::snb-1, Pttx-3::RFP], oxEx1088 [Punc-17::halorhodopsin::GFP], oxIs12 [Punc-47::GFP] , nuIs321 [Punc-17::RFP], sid-1(pk3321) him-5(e1490) V; lin-15B(n744) X; uIs72[Pmyo-2::mCherry, Punc-119::sid-1, Pmec-18::mec-18::GFP], zxIs3 [Punc-47::ChR2(H134R)::YFP], zxIs6 [Punc-17::ChR2(H134R)::YFP], juEx3771 [Pnmr-1-tomm-20N::miniSOG; Pnmr-1-mCherry].

Transgenic strains and constructs

We used a standard PCR fusion-based technique ²⁵ to generate fluorescent vav-1 reporters containing the full vav-1 promoter (~8.4 kb upstream sequence), amplified from pRF35 (vav-1 genomic clone)¹⁵, followed by a 2xSV40-NLS (second-generation nuclear localization signal), and either a ~1.2kb fragment encoding GFP [S65C], amplified from pPD135.83 (L4796), or a ~1kb fragment encoding mCherry, amplified from pDM1284 63, followed by the let-858 3’UTR.

All plasmids generated to drive vav-1 cDNA expression in specific neuronal subsets are derivatives of pRF84 [Ppha-4::vav-1::GFP]. The unc-17 (~3.5 kb) promoter and the ver-3 (~3 kb) promoter were amplified by PCR, using published primers ^26,36. Promoters were amplified from either genomic worm DNA or corresponding fosmids. Primers were used to flank each promoter with an SphI and an NheI restriction site, making them compatible to replace Ppha-4 in pRF84.

To generate ALA specific expression of tetanus toxin light chain (TeTx) and ChR2*(C128S), we used floxed/inverted vector strategy previously described ^44,45. Briefly, the following constructs were custom synthesized by Genscript: Pver-3::nCre, Prab-3::lox2272loxP::inverse-TeTx::SL2::GFP::inverse-lox2272loxP and Prab-3::lox2272loxP::inverse-ChR2*(C128S)::GFP::inverse-lox2272loxP (Fig 6b). The ver-3 promoter (3kb promoter ²⁶) drives expression in ALA, the pharyngeal saucer muscle cell and the anal sphincter muscle ^26,64. The rab-3 promoter (1.2kb promoter ³⁶) displays pan-neuronal expression ⁶⁵. Lox sequences ⁴⁵ and nCre sequence⁴³ were obtained from published sources. The temporal and spatial overlap of these promoter elements is only in ALA.

Transgenic strains were generated by microinjection following standard protocols ⁶⁶. The injected strain for fluorescent vav-1 reporters was lin-15(n765ts). The host strain for all other transgenes was wild type (N2). Listed are the injection concentrations of expression constructs of interest and coinjected DNAs: takEx6 [Pvav-1::2xNLS mCherry at 100 ng/μL, pJM23 at 20 ng/μL, DNA ladder at 5 ng/μL], takEx65 [Punc-17::vav-1::GFP at 5 ng/μL, Pttx-3::GFP at 85 ng/μL, DNA ladder at 10 ng/μL], takEx67 [Pver-3::vav-1::GFP at 5 ng/μL, Pttx-3::GFP at 85 ng/μL, DNA ladder at 10 ng/μL], takEx75 [Punc-25::vav-1::GFP at 5 ng/μL, Pttx-3::GFP at 85 ng/μL, DNA ladder at 10 ng/μL], takEx81 [Pvav-1::vav-1 L228Q::GFP at 5 ng/μL, Pttx-3::GFP at 80 ng/μL, DNA ladder at 15 ng/μL], takEx132 [Prab-3::inverse ChR2*C128S-SL2-GFP at 25 ng/μL, Pver-3::Cre at 25 ng/μL, Pttx-3::GFP at 50 ng/μL, DNA ladder at 25 ng/μL], takEx136 [Prab-3::inverse TeTx-SL2-GFP at 5 ng/μL, Pver-3::Cre at 25 ng/μL, Pttx-3::GFP at 62.5 ng/μL, DNA ladder at 25 ng/μL].

Briefly, integration of the Ppha-4::vav-1 extrachromosomal array was carried out by treating vav-1(ak41); akEx162 [Ppha-4::vav-1, Punc-122::GFP] ¹⁵ with 3800 Rads (gamma-irradiation). F2 animals were screened for 100% transmission of the Punc-122::GFP marker. Two integrants were isolated using this method, takIs3 and takIs5, and were subsequently outcrossed 10x using N2 wild type background.

RNA interference

RNAi by feeding was carried out essentially as previously described ⁶⁷. In brief, L4 animals that are specifically sensitive to RNAi in the nervous system (sid-1(pk3321) him-5(e1490) V; lin-15B(n744) X; uIs72) ⁶⁸ were raised on bacteria expressing double stranded RNA specific for vav-1 and empty vector. F1 adult progeny were subjected to aldicarb assays.

Locomotion Assays

Crawling assays were conducted on NGM plates seeded with OP50. Animals were analyzed as synchronized young adults on their growth plates, or young adults selected from mixed-age populations and moved to fresh seeded NGM plates. Each plate was placed on an Olympus dissecting microscope equipped with a Sony CCD camera for 2-5 minutes before starting a video recording of locomotion. For Figure 1b, ~5 animals were included in each 5-minute recording, and 7-11 replicate plates per strain were analyzed. For Figure 5d, an estimated 20-50 animals per plate were included in each 1-min recording, and 5 replicate plates per strain were analyzed. Wormlab software (MBF Bioscience) was used to calculate the average crawling speed.

Pharmacological assays

Aldicarb and levamisole assays were performed with the experimenter blind to the strains. NGM plates containing either 1mM or 0.5mM (for differentiation between more hypersensitive strains) aldicarb were used in this study. Aldicarb assays were conducted as previously described ¹⁸. Briefly, worms were gently prodded (with a platinum wire pick) three times between the vulva and tail, and if after any prod the worm responded by crawling, it was counted as not paralyzed. Worms that crawled off the plate or exhibited vulval prolapse were removed from the assay. The ratio of worms still moving to total worms per plate was plotted for every 20 or 30-minute time point. As previously observed ³¹, we noted day-to-day variation in the aldicarb assays. Nevertheless, all aldicarb assay paralysis curves are the summary of at least 5 assays, with ~15-20 worms per genotype per plate.

Levamisole plates were made with NGM containing 100μM drug. Worms were deemed paralyzed if they failed to respond by crawling when the agar plate was dropped (from a height of a few cm) while viewing on the dissecting microscope. Plates were dropped 3 times, with a short pause of ~5 seconds in between drops. The ratio of worms paralyzed to total worms per plate was plotted for every 30-minute time point. Levamisole paralysis curves are the summary of at least 4 assays, with ~15 worms per genotype per plate.

PTZ (pentylenetetrazole) convulsion assays were performed as previously described ³⁴. Worms were exposed to 6mM PTZ on NGM plates and assayed for seizure-like behavior at 30 and 60 minutes. 10-25 worms per genotype were tested, divided into 3 assays.

Imaging and Analysis

Animals were immobilized using 300μM sodium azide on 2% agarose pads. Soluble fluorophore imaging to investigate neuronal structure (GABAergic GFP, cholinergic mCherry, ALA neuron GFP, or functional VAV-1 GFP) was performed with a Zeiss LSM510 META-NLO confocal microscope. All other imaging was conducted with an inverted Zeiss AxioObserver microscope equipped with a Zeiss AxioCam MRm (for imaging GABA GFP::SNB-1) or Andor Clara CCD camera (for imaging cholinergic GFP::SNB-1). When comparing WT to vav-1 mutants, 10-25 animals were analyzed per genotype.

For gross morphological analysis of GABAergic neurons (oxIs12 [Punc-47::GFP]), cholinergic neurons (nuIs321 [Punc-17::mCherry]), and the ALA neuron (inIs181; inIs182 [Pida-1::ida-1::GFP]), Z-stacks of entire worms were collected using a 10x objective. For quantitative analysis, z-stacks of overlapping fields of view were collected, using a 40x objective and 3μm-thick optical slices. Image J software was used to compress z-stacks into maximum intensity projection images. Using these images and z-stacks, the total number of neuronal cell bodies, commissures and axonal projections were counted and analyzed. The data are plotted as the mean numbers of cell bodies or commissures +/− SEM.

For neuron identification, we investigated co-localization of known neuron reporter fluorescence with vav-1 reporter fluorescence. Adobe Photoshop was used to overlay the fluorescent images of a vav-1 reporter and specific neuron reporter co-expressed in one animal.

All fluorescent synaptic vesicle marker imaging was conducted with a 63x oil objective. Images were taken of the dorsal nerve cord (DNC) in animals with the DNC oriented toward the objective. A 20μm-long region of interest (ROI), located 20-40μm posterior from the vulva, was selected from each image. All intensity measurement values are in arbitrary fluorescence units, and are corrected for mean background fluorescence within the worm (background was measured for each image, near the DNC). Image J software was used to analyze GFP GABAergic synaptic vesicle distribution (juIs1 [Punc-25::GFP::snb-1]). Synaptic fluorescent puncta above a set intensity threshold were automatically detected using Image J, and puncta size is displayed in pixels. Metamorph software was used to analyze GFP cholinergic synaptic puncta (nuIs152 [Punc-129::GFP::snb-1]). Selection of ROIs and analysis of fluorescence was performed with experimenter blind to the strains. A Wacom Bamboo tablet and stylus were used to draw freehand selections around apparent puncta, interpunctal regions of the DNC, and around a background ROI for each image. Metamorph software was used for fluorescence intensity and area measurements.

Optogenetics

Optogenetic experiments were performed essentially as described ²¹. All-trans retinal (ATR) plates were NGM plates seeded with 100 μl of an E. coli (OP50) solution containing 0.67 mM ATR and allowed to grow overnight. Control plates were seeded with 100 μl OP50 solution lacking ATR. Worms carrying cholinergic [zxIs6 (Punc-17::ChR2)] or GABAergic channelrhodopsin [zxIs3 (Punc-25::ChR2)] were grown for 96 hours, or synchronized and grown from the L1 larval stage for 54 hours, on +/− ATR plates at 20°C, and resulting young adult worms were picked individually to fresh, seeded NGM plates for analysis. For each worm, a 10-second video was recorded using a fluorescence dissecting microscope and Sony CCD camera. The recording was started while the worm was mid-crawl, the blue light turned on within three seconds, and left on for the duration of the video (light intensity was ~80 mW cm⁻² from an X-Cite 120 excitation light source). From each video, one frame was selected before blue light exposure; post-light exposure, the most contracted (for zxIs6) or most relaxed (for zxIs3) frame was selected. Using Image J and a Wacom Bamboo tablet and stylus, a freehand line was traced down the center of each worm, from nose tip to posterior-most point of the intestine. The length of each worm before and after exposure to blue light was measured. The difference between these lengths was taken, and then divided by the starting worm length. The resulting values were either percent contraction or percent relaxation, for zxIs6 or zxIs3, respectively. Worm length measurements were performed blind except for –ATR control experiments, in which worms did not visibly contract or relax.

Optogenetic stimulation of ALA neuron for aldicarb assay

Prolonged photoactivation of the ALA neuron was achieved by a promoter-intersectional strategy (described above) used to drive ALA-specific expression of a slowly-inactivating ChR2*(C128S) ⁴⁷. Transgenic ALA::ChR2*(C128S) animals were raised from synchronized L1 larva to young adults on NGM plates seeded with OP50 bacterial solution supplemented with 0.67 mM ATR. Before beginning the aldicarb assay, each strain (including wild type, vav-1 mutant, and non-ATR-treated control animals) was exposed to blue light (at ~1 mW cm⁻²) for 60 s. Immediately, animals were transferred to aldicarb plates and the experimenter assessed paralysis every 20 min, blind to the strains. After each 40 min interval, each aldicarb plate (bearing 15-20 animals) was again exposed to blue light (~1 mW cm⁻²) for 60 s to ensure prolonged activation of the channelrhodopsin. This light intensity was chosen based on previous studies using this channelrhodopsin variant (C128S) in muscle cells or neurons ^45,47, and the timing of blue light exposures was chosen in an effort to maintain activation of the channelrhodopsin while minimizing a potential decrease in function due to repeated blue light stimuli, which has been observed in muscle ⁴⁷. Additionally, 60 s of blue light stimulus per plate was about the maximum duration possible due to the time restraints imposed by the aldicarb assay (i.e. at every 20 minute time point, ~ 10 min was required to test all strains for paralysis).

Command interneuron ablation with miniSOG

Animals expressing miniSOG (mini singlet oxygen generator) expressed in the command interneurons (juEx3771 [Pnmr-1-tomm-20N::miniSOG; Pnmr-1-mCherry]) ⁵⁰ were exposed to blue light for 30 min (~50 mW cm⁻²) as synchronized L1 larva in clear PCR tubes without caps, holding an estimated 150 animals in M9 buffer in each of 4 wells. Command interneuron-ablated worms were then plated onto NGM seeded with OP50 and grown to the young adult stage, along with synchronized wild type and vav-1 mutant animals, for subsequent aldicarb assays. Ablation was confirmed by loss of fluorescence in the command interneurons.

Oil-Red-O staining

Synchronized wild type and vav-1 mutant adult animals were stained using 60% Oil-Red-O to analyze fat content, as previously described ²⁴. To permeabilize the cuticle, worms were resuspended and gently rocked in 120 μl of PBS and 120 μl of 2x MRWB buffer containing 2% paraformaldehyde (PFA) for 1 hour at room temperature. 2x MRWB buffer: 160 mM KCl, 40 mM NaCl, 14 mM Na2EGTA, 1 mM spermidine-HCl, 0.4 mM spermine, 30 mM Na-PIPES pH 7.4, 0.2% β-mercaptoethanol. Samples were resuspended in 60% isopropanol for 15 minutes to dehydrate. Animals were then stained by rocking overnight at 20°C in 1 mL of freshly-diluted and filtered (0.2 μm filter) 60% Oil-Red-O in water. Animals were washed with 200 μL 1x PBS with 0.01% Triton X-100 to remove dye and mounted on 2% agarose pads for imaging. Whole slide imaging at 40x resolution was conducted using a Hamamatsu NanoZoomer and NDP.scan software (Version 2.3).

Statistics

Any comparison made between two groups (for example: WT versus vav-1) was performed using a two-tailed Student's t-test. For comparisons between three or more groups, we used one-way ANOVA followed by Bonferroni post-tests. For pharmacological assay statistical analysis, we performed two-way ANOVA followed by Bonferroni post-tests to determine differences between paralysis curves over time. The following relevant statistical comparisons were made:

In Fig. 1c, vav-1 is significantly different than WT at 60 (p <0.05), 90 (p<0.01), and 120 min (p<0.05). vav-1 is different than vav-1;vav-1 rescue at 90 (p<0.001), 120 (p<.01), and 150 min (p<0.05). WT is not significantly different than vav-1; vav-1 rescue. In Fig. 1d, vav-1 is different than WT at 60 (p<0.01), 90 (p< 0.001), and 120 (p<0.01) min. vav-1 is different than vav-1; neuronal rescue at 60 (p<0.01) and 90 min (p<0.05). vav-1;neuronal rescue is different than WT only at 120 min (p<0.05). In Fig. 1e, empty vector RNAi is different than gar-2 RNAi, positive hypersensitive control, at 90 (p<0.001), 120 (p<0.001), and 150 min (p<0.001). Empty vector RNAi is different than vav-1 RNAi at 90 (p<0.001), 120 (p<0.001), and 150 min (p<0.001). In Fig. 1h, WT, vav-1, and vav-1; vav-1 rescue are not significantly different from each other.

In Fig. 5a, WT and vav-1 are different at 60 (p<0.01), 90-120 (p<0.001), and 150 min (p<0.05). WT and vav-1; cholinergic rescue are different at 60-120 (p<0.001) and 150 min (p<0.01). vav-1 and vav-1; cholinergic rescue are different at 90 (p<0.01) and 120 min (p<0.05). In Fig. 4b and 4c, WT and vav-1 are different at 120 (p<0.01) and 150 min (p<0.001). In Fig. 5b, WT is not significantly different from vav-1; ALA neuron rescue. Paralysis of vav-1 versus vav-1; ALA neuron rescue is different at 150 (p<0.01) and 180 min (p<0.05). In Fig. 5c, WT and vav-1 are different at 60 (p<0.05), 90 (p<0.001), and 120 min (p<0.01). WT and vav-1; GEF-dead rescue are different at 90 (p<0.05), 120 (p<0.01), and 150 min (p<0.05). vav-1 and vav-1; GEF-dead rescue are not significantly different from each other.

In Fig. 6c, WT and vav-1 are different at 120 (p<0.01), 140 (p<0.01), 160 (p<0.05), and 180 min (p<0.05). WT and ALA::TeTx are different at 100 (p<0.05), 120 (p<0.05), 140 (p<0.001), and 160 min (p<0.05). vav-1 and ALA::TeTx are not statistically different. In Fig. 6d, WT and ALA::ChR2 +ATR are different at 100 (p<0.05), 120 (p<0.0001), 140 (p<0.001), and 160 min (p<0.05). ALA::ChR2 –ATR and ALA::ChR2 +ATR are different at 120 (p<0.001), 140 (p<0.05), and 160 min (p<0.05). WT and ALA::ChR2 –ATR are not statistically different. In Fig. 6e, WT and vav-1 are different at 80 (p<0.05), 100 (p<0.01), and 120 min (p<0.05). WT and vav-1; ALA::ChR2 are different at 80 (p<0.0001), 100 (p<0.0001), and 120 min (p<0.01). vav-1 and vav-1; ALA::ChR2 +ATR are different at 120 (p<0.01) and 140 min (p<0.05). The following paralysis curves were not statistically different: WT and vav-1; ALA::ChR2 +ATR , as well as vav-1 and vav-1; ALA::ChR2 –ATR.

In Fig. 7b, WT and vav-1 are different at 80 to 220 (p<0.001) and 240 min (p<0.05). The following paralysis curves are not different at any time point: WT and miniSOG, and WT and vav-1 ; miniSOG. In Fig. 7c, WT and vav-1 are different at 80 (p<0.05), 100-180 (p<0.001), and 200 min (p<0.05). WT and nmr-1; glr-1; vav-1 are different only at 220 min (p<0.05), and WT and nmr-1; glr-1 are not significantly different. nmr-1; glr-1 and nmr-1; glr-1; vav-1 mutants are not significantly different at any point. In Fig. 7d, WT and glr-1 (nd38) are different at 80-200 (p<0.001) and 220 min (p<0.01). WT and Pnmr-1::Lurcher are different at 80 (p<0.05), 100-200 (p<0.001), and 220 min (p<0.01). WT and vav-1 are different at 80-200 (p<0.001) and 220 min (p<0.05). WT and glr-1 (nd38); vav-1 are different at 60-200 (p<0.001) and 220 min (p<0.01). glr-1 (nd38) and Pnmr-1::Lurcher are different at 80 and 120 min (p<0.05). Pnmr-1::Lurcher and glr-1 (nd38); vav-1 are different at 60 (p<0.01), 80 (p<0.001), 100 (p<0.01), 120 (p<0.001), and 140 min (p<0.05). The following paralysis curves are not statistically different at any time point: glr-1 (nd38) and vav-1, glr-1 (nd38) and glr-1 (nd38); vav-1, vav-1 and glr-1 (nd38); vav-1, Pnmr-1::Lurcher and vav-1.