Visualizing Synaptic Specificity with GRASP

A maxim of modern neuroscience holds that the structures of neural circuits dictate their function. Circuit assembly requires each neuron to exercise remarkable precision as it selects a unique ensemble of synaptic partners from a vast array of potential targets. Visualization of individual synapses in the central nervous system is difficult in some contexts and impossible in others, and the ability to rapidly monitor connectivity in living animals would greatly facilitate deeper understanding of synaptic specificity. In this thesis, I describe the development of a method, GFP reconstitution across synaptic partners (GRASP), that allows visualization of defined synapses in vivo, and apply this method at one set of synapses in the brain of the nematode C. elegans. The principle of GRASP relies on bimolecular assembly of two GFP fragments expressed on two cells at a synapse. To this end, I appended fragments of green fluorescent protein (GFP) to the extracellular portions of transmembrane carrier proteins in apposing cells. When complementary CD4-tethered GFP fragments were brought into proximity at sites of cell contact, GFP fluorescence was observed both in vitro and in vivo. Split GFP fragments fused to the presynaptic phosphatase PTP-3A labeled synapses when expressed in connected neurons. This method detected known mutations that alter synaptic connectivity, such as syg-1 and syg-2, which affect development of synapses between HSN neurons and their postsynaptic nerve and muscle partners. These observations suggest that GRASP could aid efforts to trace behavioral circuits and investigation of the mechanisms of synaptic specificity. Additional tools based on Cre recombinase were developed to confine labeling to single cells and synapses of interest. The ability of GRASP to detect known specificity mutants prompted an investigation of synapse formation in the central nervous system of C. elegans. I generated a transgenic strain, kyIs501, in which GRASP labels synapses formed by the ASH sensory neuron onto the AVA interneuron. A genetic screen in kyIs501 identified one promising mutation, ky957, that causes loss of GRASP labeling. However, subsequent analyses revealed that ky957 is not a bona fide specificity mutant, and appears instead to be associated with alterations in the integrated kyIs501 transgene. Potential solutions to the problems raised by transgene-based approaches as well as further refinements of GRASP are discussed.