Structural Studies of Complexin/SNARE Interactions
Vesicular neurotransmitter release is mediated by exocytosis of synaptic vesicles at the presynaptic active zone of nerve terminals. The Ca2+-triggered release process is extremely fast, lasts less than half a millisecond, and is tightly regulated by Ca2+. Action potentials cause Ca2+ influx through voltage-gated Ca2+ channels, which in turn triggers synaptic vesicle fusion. The typical sub-millisecond latency between an action potential and neurotransmitter release and the precise timing of the release process are crucial for information processing in the nervous system. To achieve this exquisite regulation, many proteins are involved. The goal of our investigations was to delineate interactions between complexin and SNARE components that lead to the formation of a primed minimal fusion machinery. We have generated and used new constructs of short forms of synaptobrevin and complexin, as well as constructs of SNAP-25 and syntaxin that have previously been shown to be part of the minimal fusion machinery. With NMR spectroscopy, the use of the short synaptobrevin constructs has led to experimental results suggesting at least two key intermediates during the docking/priming process that are independent of complexin. We found evidence for a modular assembly of the full SNARE complex. In the absence of Syb2-CT, the N-terminal half of SNARE complex forms a four-helix bundle, while the C-terminal half, starting just after the polar layer, is disordered. In the presence of the Syb2-CT, however, both halves of the SNARE complex form a four-helix bundle. It is interesting to note that Syb2 residues 29-84 are sufficient for formation of the fully assembled SNARE complex. This evidence strongly suggests the existence of at least two intermediates during the docking priming reaction. Furthermore, with NMR spectroscopy, we have found new evidence that complexin can bind to the t-SNARE complex, in contrast to earlier evidence suggesting that complexin regulates the fully assembled SNARE complex. We demonstrated that Cpx-FL binds the t-SNARE complex SN1/SN3/Syx1a(188-259) in solution, as wa suggested for membrane-bound t-SNAREs. Note, however, that the t-SNARE complex does not contain the large complexin-binding interface provided by Syb2. Furthermore, we found that Cpx-FL also binds t-SNARE sub-complexes formed by SN1/SN3, and SN1/Syx1a(188-259), while very little binding was observed between Cpx-FL and Syx1a(188-259) alone. This finding is particularly interesting, because the cryst structure of the fully assembled SNARE complex does not suggest any binding between Cpx26-83 and either SN1 or SN3, whereas the only common component in all of the above experiments was SN1 domain.