Our studies of synaptic signaling mutants have suggested a model, the Synaptic Signaling Network, in which the integrated activities of three G signaling pathways determine which synapses are active and which are not (Reynolds et al., 2005; Schade et al., 2005). These studies show that the Gq pathway is the core, obligatory pathway for synaptic vesicle priming, but our genetic analysis suggests that the Gs pathway converges on the Gq pathway downstream of DAG production. Interestingly, the Gs pathway is not required for synaptic vesicle priming, and animals that lack it exhibit normal levels of neurotransmitter release; however, such animals are strongly paralyzed for functional (non-developmental) reasons. Our attempts to explain these findings have led us to hypothesize that the neuronal Gs pathway transduces positional information onto the downstream part of the Gq pathway to activate optimal synapses for the locomotion behavior. This hypothesis predicts that knocking out all of the positional signals should give a phenotype similar to mutants that lack a neuronal Gs pathway. We searched for mutants with this phenotype and identified an
unc-31 null mutant which, we show, lacks the dense core vesicle priming protein CAPS. We further show that
unc-31 and Gs pathway null mutants share strikingly similar phenotypes, including strong paralysis despite exhibiting near-normal levels of neurotransmitter release. Remarkably, mutations that activate the Gs pathway bypass the requirement for UNC-31 and confer similar levels of strongly hyperactive, highly coordinated locomotion in both
unc-31 null and (+) backgrounds. The neuronal, but not the muscle, Gs pathway is sufficient to rescue the paralysis of the
unc-31 null. Further genetic analysis showed that the function of UNC-31 at synapses completely overlaps with the function of the neuronal Gs pathway, and that UNC-31 and the Gs pathway are functionally distinct from the Gq pathway on which they converge. Using cell-specific promoters, we show that UNC-31 and the Gs pathway can function together in the same neurons to regulate locomotion. Using immunostaining and confocal microscopy, we show that UNC-31 is concentrated presynaptically at cholinergic synapses in the nerve cords. Remarkably, we also found that UNC-31 is often partially or completely excluded from synaptic vesicle clusters, and it is often concentrated in subsynaptic regions containing the periactive zone. We propose that UNC-31 activity, possibly mediated by dense core vesicle release, positionally activates the presynaptic Gs pathway.