GABAA receptors are the main inhibitory neurotransmitter receptors in vertebrate and invertebrate nervous systems. The major GABAA receptor in C. elegans is encoded by the
unc-49 gene. We are interested in the question of how GABAA receptors are trafficked to and maintained at synapses. We have determined that proper contact between pre- and postsynaptic cells is necessary for normal localization of GABAA receptors to postsynaptic membranes. In wild-type animals, GFP-tagged UNC-49 receptors are tightly localized to synaptic regions. In mutants with defective neuronal differentiation and axon pathfinding (
unc-3 and
unc-5), UNC-49-GFP fluorescence appears in large intracellular vesicles. Mosaic analysis revealed that UNC-5 acts in the presynaptic cell to control GABAA receptor localization in the postsynaptic cell: restoration of proper axon pathfinding in GABAergic motoneurons (by expressing UNC-5 under the control of the
unc-47 promoter) ameliorates the abnormal receptor trafficking phenotype. Thus, GABAA receptor distribution is affected by a presynaptic signal. The nature of this signal is not yet clear, however it is probably not neurotransmission per se. GABAA receptor trafficking appears normal in
unc-25 mutants, which cannot synthesize GABA, and in
unc-13 mutants, which have severe defects in all neurotransmission. The large UNC-49-GFP-containing vesicles are likely to be autophagosomes: First, they contain multiple internal membrane-bound compartments, typical of autophagosomes. Second, they are very large, typically 3-4 microns in diameter. Similar large autophagic vesicles were prominent in electron micrographs of
unc-5 mutants, and only rarely seen in the wild type. Postsynaptic material may enter the autophagy pathway following endocytosis. The putative autophagosomes are immunoreactive for RME-8, a protein found on endosomes, and required for receptor-mediated endocytosis. Moreover, electron micrographs of
unc-5 mutants show increased numbers of large endosomes compared to wild type. Autophagosomes are organelles in which excess cellular components (i.e. cytoplasm, plasma membrane and organelles) are degraded. Therefore, we propose that maintenance of postsynaptic components at the membrane may require a signal from the presynaptic cell. In the absence of this signal, postsynaptic components are removed from the plasma membrane by endocytosis, and are degraded in autophagosomes.