Abstract
Neurons depend on interconnected membrane trafficking pathways that facilitate the transport, sorting and secretion of a variety of cargos, including neurotransmitters and neuropeptides. Central to many of these processes is the large GTPase dynamin, which polymerizes around membrane tubules to control membrane scission during endocytosis and endosomal sorting, and to regulate fusion pore dynamics during dense core vesicle exocytosis. Dynamin also directly binds and bundles actin filaments to regulate cytoskeletal organization. Mammalian dynamin exists as twenty-five alternative splice variants whose specific functions in these pathways remain poorly understood. By contrast, Drosophila dynamin/Shibire is encoded by a single gene that is spliced into only two major classes of isoforms, Shibire-long (Shi-L) and Shibire-short (Shi-S). These isoforms differ in the length of their C-terminal proline-rich domain (PRD), which is intrinsically disordered and binds to a variety of SH3 domain proteins and to F-actin. By endogenously-tagging these isoforms, I found Shi-S is three times more abundant than Shi-L in the nervous system. Though the two isoforms can associate with each other in fly head extracts and have similar binding partners, they exhibit both overlapping and distinct distributions relative to each other and to active zones. Moreover, by depleting Shi-L and Shi-S I found evidence for shared, distinct, and opposite functions of the isoforms at the neuromuscular junction (NMJ). First, I found that despite their unique actin-binding PRDs, Shi-L and Shi-S both bundle F-actin in vitro and regulate the dynamics of endocytic F-actin patches at the NMJ. Interestingly, this in vitro bundling activity could be differentially disrupted by SH3 domain proteins. Second, I found Shi-S promotes bulk endocytic uptake, Synaptotagmin 1 recycling, and reformation of synaptic vesicles during high intensity stimulation; however, Shi-L could compensate when overexpressed. This suggests that the abundance of Shi-S, as opposed to its unique PRD, contributes to these functions in neurons. Third, I found that Shi-L and Shi-S differentially regulate dense core vesicle cargoes depending on neuron type: Loss of Shi-L increased one neuropeptide (Dilp2) in glutamatergic Type I neurons but decreased a peptide hormone (bursicon) in peptidergic Type III neurons, whereas loss of Shi-S decreased Dilp2 in Type I neurons. Finally, I found Shi-S regulates extracellular vesicle cargo trafficking, but neither isoform is necessary for growth factor receptor trafficking. By characterizing the biochemical activities and neuronal functions of Shi isoforms, this study provides new insight into the mechanistic roles of dynamin PRDs and how two PRDs in Drosophila can accomplish the work of the multitude of dynamins in mammals.