Morphology was not the first suspect as a trigger for uncoating Golgi transport vesicles. Lipid metabolism was a possibility, and Antonny and others experimented in vitro with the effect of different lipid mixtures on uncoating. But in vivo evidence for changes in lipid composition during vesicle budding was lacking.
The different in vitro lipid compositions did result in different uncoating dynamics, however. Antonny suspected that some of the in vitro mixtures were mimicking a distorted lipid arrangement seen during membrane curvature. Sure enough, vesicles with constant lipid composition but decreased diameter and thus increased curvature showed two changes: more active ArfGAP1, and faster dispersal of the COP1 coat. ArfGAP1 helps Arf1 to hydrolyze its bound GTP, thus releasing COP1 components that were originally recruited by Arf1-GTP.
That process of COP1 assembly begins before the budding of a vesicle. “If the membrane is flat, the coat must stay assembled because the job is not done,” says Antonny. But then COP1 polymerization helps bend the membrane, thus increasing the spacing between lipids in the outer membrane. It may be this spacing that ArfGAP1 detects.
Arf1-GDP on the positively curved membrane will rapidly disperse, but the negative curvature at the edge of a vesicle bud should ensure that a ring of Arf1-GTP will remain as a protective rim to maintain the coat until the vesicle is fully formed. This concept of Arf1GAP1 action “remains a model, but I think it can explain many things,” says Antonny. “The key point now is to find the region in ArfGAP that is sensitive to curvature.” ▪