NCAM drives synaptic remodeling in developing GABAergic neurons in C. elegans

Neural circuits are actively restructured during development as synapses are dismantled in some locations and assembled in others. Despite the importance of synaptic remodeling to circuit function, the underlying mechanisms are largely unknown. To investigate this question, we are exploiting the DD GABAergic motor neurons in C. elegans which undergo synaptic remodeling during larval development. In newly hatched larvae, DD presynaptic boutons are initially positioned on ventral muscles but are then relocated over a ~5 hr period to the dorsal DD neurite for input to dorsal muscles. DD remodeling is transcriptionally regulated by the Iroquois-type homeodomain protein, IRX-1. IRX-1 directs synaptic remodeling by upregulating UNC-8, a sodium epithelial channel (ENaC), which triggers a Ca2+-dependent mechanism of presynaptic disassembly. Additional downstream effectors are likely required, however, because UNC-8 dismantles a subset of presynaptic components (RAB3, v-SNARE, liprin-, endophilin) whereas IRX-1 also acts in a parallel pathway to remove additional presynaptic proteins (UNC13, ELKS, Clarinet). To identify additional IRX-1 targets, we used single-cell RNA-Sequencing (scRNA-Seq) to detect transcripts that are upregulated in remodeling DD neurons. An RNAi screen of this dataset detected a necessary role for the neural cell adhesion protein, NCAM-1, in DD synaptic remodeling. A genetic mutant of NCAM-1 impairs both the removal of ventral synapses as well as their reassembly in dorsal DD neurites, thus confirming that NCAM-1 normally promotes remodeling. Considering the established roles of NCAM in vertebrate neural development, my studies of the C. elegans NCAM homolog could reveal conserved mechanisms for synaptic plasticity that also operate in the brain. I hypothesize that NCAM-1 mediates presynaptic disassembly in remodeling GABAergic neurons in parallel to UNC-8. In Aim 1, I will test the role of NCAM-1 as a regulator of presynaptic remodeling using (A) smFISH to determine if ncam-1 is regulated by IRX-1, (B) live-cell imaging of GFP tagged Clarinet in ncam-1 mutants to determine if NCAM-1 functions in a parallel to UNC-8, and (C) GFP-tagged NCAM-1 to determine its location in remodeling DD neurons. In Aim 2, I will use genome-engineering to identify NCAM-1 structural domains that are required for presynaptic disassembly. Additionally, I will use cell-specific RNAi (csRNAi) to determine if ncam-1 function is required in DD neurons for synaptic disassembly. Finally, in Aim 3, I will determine if NCAM-1 is required for recycling of photoconverted (red) dendra2::RAB-3 from ventral to dorsal synapses in remodeling DD neurons. Because the vertebrate homolog of NCAM-1 mediates the removal of GABAergic inputs in the developing cortex, we suggest that the underlying mechanism may be conserved and thus can be molecularly defined by studies in C. elegans.