Over 30 years ago, we observed what we interpreted to be vesicular transport in crude extracts of tissue culture cells. In subsequent years we found that we had reconstituted vesicle trafficking in the Golgi, including the process of membrane fusion. Using this assay as a guide, we purified as a required factor the NEM-Sensitive Fusion protein (NSF). This led to the purification of the Soluble NSF Attachment Factor (SNAP), which bound NSF to Golgi membranes, and then with these tools discovered that the receptors for SNAP in membranes were actually complexes of proteins (which we called SNAREs) which we envisioned could potentially partner as a bridge between membranes to contribute to the process of membrane fusion and provide specificity to it (as captured in the ‘SNARE hypothesis’ proposed at the time).
We now know that organisms have a large family of SNARE proteins that indeed form cognate partnerships in just this way, and that NSF is an ATPase that (using SNAP as an adaptor protein) disrupts the SNARE complex after fusion is complete so its subunits can be recycled for repeated use. Recombinant cognate SNAREs introduced into artificial bilayers or expressed ectopically on the outside of cells ( “flipped SNAREs”) spontaneously and efficiently result in membrane (or cell) fusion, demonstrating that the SNARE complex is not only necessary but is sufficient for fusion. There are many proteins known and rapidly being discovered which closely regulate this vital process, but the muscle – if not always the brains – is in the SNAREs. Compartmental specificity is encoded to a remarkable degree in the functional partnering of SNARE proteins, a fact which is in no way inconsistent with the emerging contribution of upstream regulatory components (like rabGTPases and tethering complexes) to domain/compartment specificity.