A synaptic vesicle, up close and personal.


The first molecular level model of an organelle is described by Shigeo Takamori, Reinhard Jahn (Max-Planck Institute, Göttingen, Germany), and colleagues. Proteomics, quantitative measurements, and molecular modeling give them a view of an average synaptic vesicle.

The first surprise is that the vesicle is almost smothered in proteins. The group started by identifying 410 proteins associated with purified brain synaptic vesicles, but “if you want to put 400 proteins in the model it exceeds the amount of surface area available,” says Takamori. They suspect that many of the proteins are passengers on only a few vesicles, so instead they measured the copy number of the more abundant proteins. By putting just these 27 proteins in the model, they accounted for almost 70% by mass of the vesicles' protein complement.

Even with these few proteins, “the vesicles are really protein dominated,” says Takamori. Transmembrane domains take up approximately 20% of the surface, with their attached globular domains shading far more. If each protein was surrounded by a monomeric collar of phospholipids, the majority of phospholipids would not be free. The amount of protein and lipid clustering is not yet known, however.

The model “provides a point of departure for understanding the molecular mechanism of trafficking, docking, and fusion of a trafficking organelle,” says Takamori. The synaptic vesicle and its proteins have been studied in detail, but quantitative information has been lacking. The new model suggests that quantities of certain fusion proteins are large enough not to be rate limiting. The presence of more than the expected diversity of other proteins, such as Rabs, is still to be explained as either contamination or signs of unexpected side-trips by the vesicles.


Takamori, S., et al.