Most membrane proteins form small, organized multicopy domains. Scientists have hoped that lipid rafts might explain this compartmentalization. But Lang isn't convinced. “There are too many types of domains to explain by lipids alone,” he says. “Lipid–lipid interactions would not mediate enough specificity.”
A newer model suggests that cytoskeletal proteins form fences and pickets that corral proteins. “But then how do you get diversity?” asks Lang. “Proteins are just sorted by size. It can only explain a limited number of areas.”
Neither model explains why two very similar membrane fusion proteins, syntaxins 1 and 4, segregate into different clusters. These proteins have cytoplasmic SNARE motifs, which like to self-associate. Lang and coworkers now show that this propensity is enough to explain their specific clustering.
By characterizing syntaxin 1 clusters in detail, the authors found that most copies of syntaxin 1 were stuck in immobile clusters of ∼75 molecules, whereas a few—presumably upon release from the periphery—were freely diffusing and could exchange between clusters. The removal of cytosolic factors, including actin, did not increase their motility. But deletion of their SNARE domains did.
If SNARE association brings clusters together, a repulsive force must also limit cluster size, as revealed by computer simulations. The group imagined that repulsion might be provided by syntaxin 1's large N terminus. An in silico model of the clusters suggested that this bulky domain caused molecules near the cluster rim to bend over more, which should prevent more copies from joining. Charges caught within the oligomerized domains might also create repulsion.
“The whole thing can be explained purely by these two simple counteracting forces,” says Lang. “We predict that other proteins self-aggregate equally well.” Ion transporters and membrane-bound receptors, for example, also have domains that form oligomers.