How are the many stimuli that tell cells to move translated into changes in actin polymerization? Evidence points to members of the Wiskott–Aldrich syndrome protein ( WASP) family as the interpreters but, for scientists, the language of cell movement has proved difficult to learn. Two papers in The Journal of Cell Biology provide some clues as to how a lipid and a protein collaborate to activate two WASP-family members. The details seem protein specific but the general message is the same — activation of WASPs involves stopping them from biting their own tails.

How Cdc42 and acidic phospholipids might activate WASP. Inset shows an actin-filament halo (red) surrounding a vesicle (green). Photo courtesy of Henry Higgs and Tom Pollard, Salk Institute, La Jolla, California, USA.

Actin nucleation is stimulated by the Arp2/3 complex, which is activated by WASPs. A small GTPase, Cdc42, and a phospholipid, phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P2), activate WASPs, but how they do so is controversial: recombinant WASPs often have some constitutive activity, and inactive, GDP-bound Cdc42 sometimes seems capable of activating them.

Henry Higgs and Tom Pollard sought to end this controversy — at least for the haematopoietic-cell-specific member of the family, WASP — by purifying native WASP from bovine thymus. They find that purified WASP alone has no effect on actin polymerization rates, but that micelles containing PtdIns(4,5)P2 activate polymerization through WASP. In the presence of Cdc42, PtdIns(4,5)P2 micelles, or vesicles containing either PtdIns(4,5)P2 or another acidic phospholipid, phosphatidylserine, produce halos of polymerized actin surrounding the phospholipid (see picture). This effect requires PtdIns(4,5)P2 or phosphatidyl-serine, and Cdc42 must be both GTP-bound and prenylated, indicating that it needs to be membrane associated to do its job.

Rohatgi and colleagues, working with a recombinant form of the widely expressed N-WASP, have a different story: they find that Cdc42, but not PtdIns(4,5)P2, can partly activate N-WASP, and that the two molecules synergize to activate N-WASP fully. They identify a basic region, close to the Cdc42-binding domain, that seems to bind PtdIns(4,5)P2. In actin nucleation assays, a mutant N-WASP lacking this domain remains sensitive to Cdc42 but is insensitive to the additive effects of Cdc42 and PtdIns(4,5)P2. These researchers previously showed that PtdIns(4,5)P2 stimulates actin polymerization in Xenopus egg extracts. They now show that this effect depends on N-WASP but, curiously, the deletion mutant can also translate a PtdIns(4,5)P2 signal into limited actin polymerization, albeit more slowly.

WASP's carboxyl terminus is constitutively active, suggesting an autoinhibitory mechanism. Both groups show that a separate Cdc42-binding domain can curb the activity , including the C motif and an acidic region (C and A in the figure), of the carboxyl terminus in trans. Full-length N-WASP, however, can't inhibit N-WASP's carboxyl terminus, presumably because the full-length protein is folded into its autoinhibited conformation. Higgs and Pollard find that inhibition of WASP's carboxyl terminus is relieved by GTP–Cdc42 but not by PtdIns(4,5)P2, whereas Rohatgi and co-workers find that their intermolecular inhibitory complex is regulated in exactly the same way as wild-type N-WASP.

We're still straining to understand what WASPs, Cdc42 and PtdIns(4,5)P2 are saying to each other. Perhaps WASP and N-WASP respond to their two activators slightly differently, or maybe the discrepancies are due to variations between purified and recombinant proteins. Is WASP's PtdIns(4,5)P2-binding domain equivalent to N-WASP's? And can N-WASP be activated by phosphatidylserine? Further studies should clarify how Cdc42 and acidic phospholipids unleash WASP's sting.