Fig. 2. A hypothetical model for the activity cycle of cofilin in crawling tissue cells. Some motile cells, such as neutrophils, regulate cofilin by dephosphoylation of an inactive phosphorylated pool upon chemotactic stimulation (not shown). By contrast, others, such as carcinoma cells, might maintain the majority of cofilin prior to stimulation in a dephosphorylated yet inactive state (cf_i) generated through interaction of cofilin with PtdIns(4,5)P₂ and/or formation of cofilin–G-actin heterodimers. Following EGF stimulation, a PLC-dependent step releases activated cofilin (cf_a), which then associates with F-actin to promote F-actin severing. This leads both to polymerization and depolymerization, the balance being determined by the relative availability of G-actin. Cofilin is rescued from the cofilin–G-actin heterodimer by two mechanisms. (1) Phosphorylation by LIM kinase (LIMK) or TES kinase (TESK) turns off the actin-binding activity of cf, releasing G-actin and phospho-cofilin. Cofilin phosphatases such as PP1, PP2A or SSH (slingshot) can then replenish the pool of dephosphorylated cofilin. (2) CAP can bind to the cofilin–G-actin heterodimer and release free cofilin and G-actin. The freed cofilin can bind to PtdIns(4,5)P₂ to form an inhibitory complex that is released locally by EGF-stimulated receptors to begin the activity cycle again. Cofilin may bind directly to PtdIns(4,5)P₂ or through another protein (X). CAP is a candidate for X since it regulates cofilin location in vivo. Localized activation of cofilin by PtdIns(4,5)P₂ hydrolysis causes local actin polymerization and protrusion, and sets the direction of movement.