I remember for a couple years, it was "lipid rafts this" and "lipid rafts that." The idea of the lipid rafts -- for the uninitiated -- was that there were microdomains in the plasma membranes of cells defined by their more hydrophobic composition. You can definitely separate these fractions from the fractions of plasma membrane (I know because I have done it), and under some circumstances proteins migrate from one compartment to another.

A paper in the latest Cell by Douglass and Vale contests the notion that the hydrophobicity of the membrane is responsible for recruiting proteins into the lipid rafts. They show through GFP labeling and observation of single molecules that protein-protein interactions are necessary for this recruitment.

The Scientist lists responses to this work:

Critics counter that Vale came to the party too late: Looking ten minutes after T-cell stimulation (as Vale did) may miss lipid-protein interactions. Barbara Baird of Cornell University says lipids play a more prominent role early on, initiating "structural rearrangements that enable the proteins to connect." Once the complex builds up, "then I think it does become protein dominated."

Protein-interaction dynamics, in turn, are influenced by their lipid environment, says Thomas Harder of Oxford University. In rafts, lipids are packed more tightly and may promote longer protein-protein interactions. Equally important is the exclusion of certain proteins from rafts. For example, in T cells the ordered lipids surrounding the TCR seem to favor kinases over phosphatases, Harder says, tipping the balance between opposing pathways.

Vale allows that if lipid rafts are as small and transient as some researchers suppose, his methods may not yet detect them. Moreover, interplay between both forces isn't out of the question. Still, he says, lipids have been overemphasized since rafts were first proposed. "What we found is a very pronounced and dominant organization of membrane domains that look to be primarily based on protein-protein interactions."

I don't know who is right on this, but I remain skeptical of lipid rafts primacy as a signaling mechanism. Actually my argument is evolutionary, as I am not certain how something like that -- which is not genetic-encoded -- might evolve.