Stretching cells hold tight by allowing integrins to sidestep the Golgi, suggest Hans Schotman, Leena Karhinen, and Catherine Rabouille (University Medical Centre, Utrecht, Netherlands). The group uncovers a specialized section of Golgi-like plasma membrane that draws in integrins directly.

The hybrid membrane appeared in remodeling epithelial cells that cover the developing fly oocyte. The cells begin as columns that attach to each other at the sides and to the matrix on their basal side. But as the oocyte grows, the cells are stretched and flattened. The group found that these forces pulled the cells apart slightly, exposing the matrix to membrane that was previously attached to other cells.

This newly matrix-adjacent membrane soon harbored GRASP, which is normally a bona fide Golgi marker. As the epithelial cells were pulled apart, grasp RNA was translated near the exposed membrane region. Two other Golgi proteins, GM130 and Gos28, were also found in this membrane region. “Our idea is that the plasma membrane is disguising itself as the Golgi,” says Rabouille, “so that carriers from the ER fuse there instead of with the Golgi.”

These lured carriers, the group imagines, harbor integrins that attach the membrane to the matrix. Insertion of one integrin subunit into these membrane sections was indeed insensitive to inhibitors of Golgi transport. And in the absence of GRASP, the integrin was instead retained within the cell. The resulting lack of adhesion caused epithelial disorganization.

It is not clear why the cells do not use the standard Golgi trafficking pathway. Posttranslational modifications that occur at the Golgi can make integrins less sticky. Bypassing this organelle might thus create more adhesion for these highly stressed cells.

The Dictyostelium version of GRASP has been shown to drive another unusual secretion pathway, which sends proteins directly from the cytosol to the extracellular space. Rabouille's group now wants to determine how the fly bypass is activated; since integrins are thought to be mechanosensors, they might trigger their own retargeting in response to stretching.

Reference:

Schotman, H., et al.
2008
.
Dev. Cell.
14
:
171
–182.