A brain is composed of cells connected to one another at synapses. The output of a particular cell is determined by its synaptic inputs. Our goal is to build a mechanistic model for the synapse—to understand how the proteins assemble to perform all the functions of this nanomachine and how the output of the synapse can be modulated. To understand how the synapse functions, we are dissecting the molecular basis of synaptic function in the simplest organism with a well-defined nervous system—the nematode Caenorhabditis elegans. Models for the functions of proteins at the synapse often arise from biochemical studies characterizing the binding interactions of particular proteins. We can test the validity of these models by comparing predictions in mutants that lack a particular protein. We analyze the defects in these mutants by using electrophysiological and ultrastructural methods to determine synaptic activity and morphology. These results can then be used as a springboard for studies in the mouse, linking biochemical models of protein function and studies of the vertebrate brain.
Our first goal was to identify genes required specifically for GABA function. GABA is the primary inhibitory neurotransmitter in vertebrate and invertebrate nervous systems. We identified the genes required uniquely for GABA transmission in the worm and then sought their homologs in vertebrate genomes. Pioneering work by Steve McIntire in the laboratory of H. Robert Horvitz demonstrated that 26 of the worm's 302 neurons synthesize GABA (Figure 1). We then used a laser to kill the GABA neurons in wild-type animals and identified four behaviors mediated by GABA neurons.
By screening for mutants that had similar defects, we identified five genes required specifically for GABA neurotransmission: unc-25, unc-30, unc-46, unc-47, and unc-49 (Figure 2). In collaboration with Yishi Jin (HHMI, University of California, Santa Cruz) and Bob Horvitz, we characterized the unc-25 locus and found that it encodes the biosynthetic enzyme for GABA. Furthermore, we demonstrated that GABA transmission is not required for formation of presynaptic elements of the neuromuscular junction. My laboratory found that the unc-47 locus encodes the vesicular GABA transporter. In collaboration with Robert Edwards (University of California, San Francisco), we used the sequence of the worm protein to identify the mammalian vesicular GABA transporter. These experiments illustrate how genetic studies in the nematode can be used for gene discovery in vertebrates. We have also determined that unc-46 encodes an ancillary subunit for the transporter. We demonstrated that the unc-49 locus encodes the GABAA receptor. Finally, we cloned the exp-1 gene and found that it encodes a cationic GABA receptor—a completely novel function for the neurotransmitter GABA. In contrast to UNC-47, there is no homolog of EXP-1 in vertebrates, thus revealing that some simple nervous systems are more complex at a molecular level than the physically larger vertebrate brain. (This work was supported by a grant from the National Institutes of Mental Health.)
Our second goal is to identify genes required for the functions of all synapses. Mutants defective in synaptic transmission are easy to isolate because they are resistant to inhibitors of acetylcholinesterase. Such screens have led to the identification of a large number of synaptic proteins. We are now studying the ultrastructural and physiological phenotypes of these mutants to determine at which step these proteins function in neurotransmission.
Endocytosis. Our ultrastructural analysis has determined that several of these proteins function during endocytosis of synaptic vesicles. In collaboration with Michael Nonet and Aixa Alfonso (Washington University), we demonstrated that the clathrin-adaptor protein AP180 is required to determine the diameter of synaptic vesicles. The phosphatidylinositol phosphatase synaptojanin is required for multiple steps of endocytosis and presynaptic architecture. Moreover, we found that endophilin is likely to act as a docking protein for synaptojanin, and its proposed role in lipid metabolism may be less important. The biggest surprise came from our analysis of synaptotagmin, the calcium sensor for exocytosis; we concluded that this protein is also required for endocytosis of synaptic vesicles. Such a role had been predicted based upon its biochemical interactions with the clathrin-adaptor complex AP2.
Exocytosis. The release of neurotransmitter at synapses requires the fusion of synaptic vesicles with the plasma membrane. Synaptic vesicle exocytosis, like the fusion of membranes at all stages of vesicular transport, is mediated by SNARE proteins (Figure 3). When reconstituted into lipid bilayers, these SNARE proteins are sufficient for membrane fusion in vitro. By contrast, genetic experiments indicate that other components are also essential for fusion in vivo. For example, mutations in unc-13 in the nematode C. elegans lead to paralyzed animals. Determining the role of a protein in exocytosis requires whole-cell patch-clamp techniques. Using these techniques, we discovered that the UNC-13 protein is required for the priming of synaptic vesicles for fusion. Biochemical data suggested that the priming of vesicles involves the "opening" of an autoinhibitory domain of syntaxin, a protein required for synaptic vesicle fusion. Our studies determined that an "open" form of syntaxin could bypass the requirement for UNC-13 in exocytosis, thus demonstrating that the role of UNC-13 is to open or to stabilize the open state of syntaxin. This type of study represents one of the most useful applications of the nematode for molecular biology: the expression of specific structural variants of proteins inthe worm can test biochemical models for protein function (Figure 4). (This work was supported by a grant from the National Institute of Neurological Disorders and Stroke.)
Our third goal is to identify proteins required for neuroendocrine modulation of behavior. The behavior that we are studying is the defecation cycle in C. elegans, which consists of a series of stereotypical muscle contractions that are repeated every 50 seconds. We identified a gene that can be mutated to accelerate, retard, or eliminate the defecation cycle. Thus, this gene satisfies the criteria of the timekeeper of the biological clock that controls the defecation cycle. We cloned this gene and discovered that it encodes the IP3 (inositol 1,4,5-trisphosphate) receptor. This receptor generates calcium spikes in the intestine at 50-second intervals, and these spikes trigger the muscle contractions that initiate each defecation cycle. The intestine is thus behaving as a central clock that coordinates multiple tissues in the worm in a neuroendocrine fashion.
Recently, we have been studying the mechanism by which the intestine communicates to neurons and muscles to execute the motor program of the defecation cycle. These studies have led to the bizarre conclusion that protons can act as a neurotransmitter. First, the propagation of a calcium wave passing through gap junctions, encoded by the inx-16 gene, activates proteins that mediate the posterior body contraction. The calcium wave stimulates a calcium-activated proton transporter encoded by the pbo-4 gene. Protons then diffuse to the muscles and activate a proton-gated ion channel encoded by the pbo-5 gene. These are some of the first data suggesting that protons act as a transmitter rather than as simply an environmental signal. (This work was supported by a grant from the National Institute of Mental Health.)