A diffusion barrier at the axonal initial segment (IS; the axonal area nearest the cell body) has been proposed before. But there have always been caveats: the introduced dye might have had too far to travel, or the latex bead used during testing might have cross-linked and thus immobilized its target.
The Japanese team used three single-molecule techniques to demonstrate that diffusion of a phospholipid in the IS area decreased more than 800-fold between 6 and 10 d after plating of a neuron. The decrease in other regions of the neuron was only one- to threefold.
The diffusion barrier arose coincident with the concentration of actin, ankyrin, and various transmembrane proteins in the IS area, and partial disruption of actin made the lipid mobile once again. This reminded Kusumi of his earlier results in nonneuronal cells, in which lipids appeared to “hop” between actin- dependent plasma membrane compartments.
The team thus adopted a version of this model in which actin cables provide the scaffolding, and “transmembrane proteins act like pickets for the fence,” says Kusumi. But could such a scheme explain not just slowed diffusion but a definitive barrier? The team modeled the behavior of membrane proteins and lipids and found that, sure enough, the diffusion rate decreased by a factor of 500 when the boundary coverage by the transmembrane “pickets” increased from 5% to just 25%. Thus, says Kusumi, “you don't have to close off the boundary completely. Lipids next to immobile proteins become harder to move … so the barrier function propagates.” ▪