Neurotransmitter (NT) release is accomplished by a machinery that unclamps fusion in response to calcium and then fuses the synaptic vesicle and plasma membranes. These are often thought of as distinct tasks assigned to non-overlapping components. Vesicle release rates have a power law dependence on [Ca2+] with an exponent of 3-5, long taken to indicate that 3-5 Ca2+ ions bind the calcium sensor Synaptotagmin to trigger release. However, dependencies at low [Ca2+] are inconsistent with simple sequential binding to a single Ca2+ sensor followed by a final fusion step. Here we developed coarse-grained molecular dynamics simulations of the NT release machinery accounting for Synaptotagmin-mediated unclamping and SNARE-mediated fusion. Calcium-triggered unclamping and SNARE-mediated fusion emerged from simulations as contemporaneous, coupled processes. Increasing cytosolic [Ca2+], the instantaneous fusion rate increased as SNAREpins were progressively and reversibly released by dissociation of Synaptotagmin-SNAREpin complexes. Simulations reproduced the observed dependence of release rates on [Ca2+], but the power law was unrelated to the number of Ca2+ ions required. Action potential-evoked vesicle release probabilities depended on the number of transiently unclamped SNAREpins, explaining experimental dependencies of release probabilities on both unclamping and membrane-fusing machinery components. These results describe a highly cooperative NT release machinery with intrinsically inseparable unclamping and membrane-fusing functionalities.

Neurotransmitter release is accomplished by a multi-component machinery including the membrane-fusing SNARE proteins and Ca2+-sensing Synaptotagmin molecules. However, the Ca2+ sensitivity of release was found to increase or decrease with more or fewer SNARE complexes at the release site, respectively, while the cooperativity is unaffected (Acuna et al., 2014; Arancillo et al., 2013), suggesting that there is no simple division of labor between these two components. To examine the mechanisms underlying these findings, we developed molecular dynamics simulations of the neurotransmitter release machinery, with variable numbers of Synaptotagmin molecules and assembled SNARE complexes at the release site. Ca2+ uncaging simulations showed that increasing the number of SNARE complexes at fixed stoichiometric ratio of Synaptotagmin to SNAREs increased the Ca2+ sensitivity without affecting the cooperativity. The physiological cooperativity of ~4-5 was reproduced with 2-3 Synaptotagmin molecules per SNARE complex, suggesting that Synaptotagmin and SNAREs cooperate in fixed stoichiometry modules. In simulations of action potential-evoked release, increased numbers of Synaptotagmin-SNARE modules increased release probability, consistent with experiment. Our simulations suggest that the final membrane fusion step is driven by SNARE complex-mediated entropic forces, and by vesicle-tethering forces mediated by the long Synaptotagmin linker domains. In consequence, release rates are increased when more SNARE complexes and Synaptotagmin monomers are present at the fusion site.