Actin takes its hat off to dynamin.

A major player in membrane traffic, the large GTPase dynamin, is known in the actin dynamics field for its indirect association with the actin cytoskeleton via actinbinding proteins (ABPs) like cortactin. In this issue of The EMBO Journal, Gu et al extend this picture by revealing a direct interaction between dynamin and F-actin. They further show that oligomerized dynamin kicks off the gelsolin cap at the barbed ends of filaments. This study not only provides an interesting link between actin remodelling and membrane dynamics via dynamin, but also sheds light on the long-standing mystery of how barbed ends are liberated from the high-affinity capping protein gelsolin. For most people, dynamin belongs to the cell membrane dynamics world, and is particularly recognized for its role in endocytosis where it combines membrane deformation and fission activities (Praefcke and McMahon, 2004). Recent work has fleshed out a molecular model for dynamin function, whereby the dynamin helix undergoes a unique mechanical sequence of constriction and torsion followed by depolymerization in order to break membranes (Lenz et al, 2009). While it is clear that dynamin is also involved in actin dynamics, only indirect interactions with filamentous actin have been reported in which the proline-rich domain (PRD) of dynamin interacts with the SH3 domains of ABPs such as cortactin (Orth and McNiven, 2003; Mooren et al, 2009). In this scenario, it is unclear how the nucleotide load of dynamin could influence actin dynamics and how actin dynamics could interfere with the GTPase activity of dynamin. Surprisingly late in the long history of dynamin–actin interaction, and thus even more remarkably, Gu et al show not only a direct interaction between these two major players of intracellular dynamics, but also that the dynamics of actin and dynamin are coupled. First, Gu et al identify an F-actin-binding site in dynamin and show that binding is increased or decreased in predictable ways by mutating specific charged residues. WT dynamin expressed in vitro or a mutant lacking its PRD domain are both active for actin filament binding, providing further evidence that this interaction is direct and not mediated by