In active T cells, signal transduction proteins cluster within the plasma membrane, probably to enhance signaling by concentrating the interacting proteins. Popular model mechanisms for clustering involve either lipid rafts or actin. But Douglass and Vale found that the signaling proteins themselves hold clusters together.
In their system, T cell clusters included the LAT adaptor protein, the Lck tyrosine kinase, and the CD2 costimulatory transmembrane protein. LAT and Lck are thought to be raft-localized proteins. But mutation of their raft-localizing regions did not alter LAT or Lck diffusion or clustering. By contrast, mutating LAT residues that are essential for protein–protein interactions prevented LAT clustering.
Clusters were also maintained in the absence of actin polymerization, although actin was needed for cluster formation. Douglass and Vale think that actin or actomyosin may be needed for the initial movement of proteins into clusters, but not for anchoring them together.
When tracking single molecules, the authors noticed that LAT and Lck diffused rapidly outside of clusters but became temporarily trapped, probably via protein–protein interactions, when encountering cluster sites. Nonclustering proteins were rarely trapped and were forced to navigate between clusters.
Vale notes that their findings do not rule out the existence of lipid rafts. Rather, the findings support the idea that “protein–protein interactions may be a more common mechanism for creating signaling microdomains,” he says. He hopes eventually to understand why microdomains need to be formed during T cell signaling. “Many people in the signaling field think about which molecules interact with one another,” he says, “but the issue of how the molecules are organized spatially and how this affects their function is often not addressed.”