They're either on or off. Rho GTPases regulate various cellular functions, and their ability to bind GTP or GDP, and hence their activity, is controlled by guanine nucleotide-exchange factors (GEFs) and GTPase-activating proteins (GAPs), respectively. But even the most powerful GEF can't activate a Rho GTPase that is bound to a GDP-dissociation inhibitor (GDI). Gary Bokoch's group, though, have ascertained that the serine/threonine kinase p21-activated kinase-1 (PAK1), which is regulated by the Rho GTPases Rac and Cdc42, can. And they explain how.

After noticing that RhoGDI could be phosphorylated in vitro, the authors identified the kinase in question as PAK1, and found that PAK1 and RhoGDI strongly co-immunoprecipitated — but only when PAK1 was active. The carboxyl terminus (which contains the kinase domain), but not the amino terminus (which binds to Rac and Cdc42), of PAK1 mediated the interaction, so any indirect interaction through Rac or Cdc42 seems unlikely. The RhoGDI sequence contains two consensus PAK phosphorylation sites, and in vitro labelling experiments also identified two potential phosphorylation sites — at Ser101 and Ser174 — which turned out to be phosphorylated by PAK1 in vivo.

The burning issue is what phosphorylation by PAK1 does to RhoGDI that is complexed with Rho GTPases. Bokoch's group measured 35S-labelled GTPγS incorporation by the GTPase–RhoGDI complexes with or without PAK1. Without PAK1, RhoGDI inhibits nucleotide exchange, so no [35S]GTPγS was bound. However, adding PAK1 caused phosphorylation of the complex and increased [35S]GTPγS incorporation by Rac1; this didn't occur using a non-phosphorylatable form of RhoGDI. In cells, expressing the catalytically active carboxyl terminus of PAK1 caused RhoGDI to dissociate from Rac1.

Bokoch's group thought that Cdc42 might activate PAK1, which would thereby induce Rac activation by RhoGDI dissociation. This proved correct — cells that expressed constitutively active Cdc42 and a particular non-inhibitory form of PAK1 had high levels of active Rac1; those with an inhibitory PAK1 mutant showed no increase. And the authors found that the activation of Rac1 that is induced by growth factors such as platelet-derived growth factor (PDGF) and epidermal growth factor (EGF) required PAK-induced phosphorylation of Rac1 to dissociate Rac1 from RhoGDI. So they propose a model in which this PAK-induced dissociation of RhoGDI either initiates Rac signalling (probably through Cdc42-induced PAK activation) or feeds forward to enhance Rac activation.