WAVE2 belongs to a family of proteins that mediates actin reorganization by relaying signals from Rac to the Arp2/3 complex, resulting in lamellipodia protrusion. WAVE2 displays Arp2/3-dependent actin nucleation activity in vitro, and does not bind directly to Rac. Instead, it forms macromolecular complexes that have been reported to exert both positive and negative modes of regulation. How these complexes are assembled, localized and activated in vivo remains to be established. Here we use tandem mass spectrometry to identify an Abi1-based complex containing WAVE2, Nap1 (Nck-associated protein) and PIR121. Abi1 interacts directly with the WHD domain of WAVE2, increases WAVE2 actin polymerization activity and mediates the assembly of a WAVE2-Abi1-Nap1-PIR121 complex. The WAVE2-Abi1-Nap1-PIR121 complex is as active as the WAVE2-Abi1 sub-complex in stimulating Arp2/3, and after Rac activation it is re-localized to the leading edge of ruffles in vivo. Consistently, inhibition of Abi1 by RNA interference (RNAi) abrogates Rac-dependent lamellipodia protrusion. Thus, Abi1 orchestrates the proper assembly of the WAVE2 complex and mediates its activation at the leading edge in vivo.

The Rho-GTPase Rac1 stimulates actin remodelling at the cell periphery by relaying signals to Scar/WAVE proteins leading to activation of Arp2/3-mediated actin polymerization. Scar/WAVE proteins do not interact with Rac1 directly, but instead assemble into multiprotein complexes, which was shown to regulate their activity in vitro. However, little information is available on how these complexes function in vivo. Here we show that the specifically Rac1-associated protein-1 (Sra-1) and Nckassociated protein 1 (Nap1) interact with WAVE2 and Abi-1 (e3B1) in resting cells or upon Rac activation. Consistently, Sra-1, Nap1, WAVE2 and Abi-1 translocated to the tips of membrane protrusions after microinjection of constitutively active Rac. Moreover, removal of Sra-1 or Nap1 by RNA interference abrogated the formation of Rac-dependent lamellipodia induced by growth factor stimulation or aluminium fluoride treatment. Finally, microinjection of an activated Rac failed to restore lamellipodia protrusion in cells lacking either protein. Thus, Sra-1 and Nap1 are constitutive and essential components of a WAVE2-and Abi-1-containing complex linking Rac to site-directed actin assembly.

Actin-crosslinking proteins organize actin into highly dynamic and architecturally diverse subcellular scaffolds that orchestrate a variety of mechanical processes, including lamellipodial and filopodial protrusions in motile cells. How signalling pathways control and coordinate the activity of these crosslinkers is poorly defined. IRSp53, a multi-domain protein that can associate with the Rho-GTPases Rac and Cdc42, participates in these processes mainly through its amino-terminal IMD (IRSp53 and MIM domain). The isolated IMD has actin-bundling activity in vitro and is sufficient to induce filopodia in vivo. However, the manner of regulation of this activity in the full-length protein remains largely unknown. Eps8 is involved in actin dynamics through its actin barbed-ends capping activity and its ability to modulate Rac activity. Moreover, Eps8 binds to IRSp53. Here, we describe a novel actin crosslinking activity of Eps8. Additionally, Eps8 activates and synergizes with IRSp53 in mediating actin bundling in vitro, enhancing IRSp53-dependent membrane extensions in vivo. Cdc42 binds to and controls the cellular distribution of the IRSp53-Eps8 complex, supporting the existence of a Cdc42-IRSp53-Eps8 signalling pathway. Consistently, Cdc42-induced filopodia are inhibited following individual removal of either IRSp53 or Eps8. Collectively, these results support a model whereby the synergic bundling activity of the IRSp53-Eps8 complex, regulated by Cdc42, contributes to the generation of actin bundles, thus promoting filopodial protrusions.

Cell migration requires the generation of branched actin networks that power the protrusion of the plasma membrane in lamellipodia. The actin-related proteins 2 and 3 (Arp2/3) complex is the molecular machine that nucleates these branched actin networks. This machine is activated at the leading edge of migrating cells by Wiskott-Aldrich syndrome protein (WASP)-family verprolin-homologous protein (WAVE, also known as SCAR). The WAVE complex is itself directly activated by the small GTPase Rac, which induces lamellipodia. However, how cells regulate the directionality of migration is poorly understood. Here we identify a new protein, Arpin, that inhibits the Arp2/3 complex in vitro, and show that Rac signalling recruits and activates Arpin at the lamellipodial tip, like WAVE. Consistently, after depletion of the inhibitory Arpin, lamellipodia protrude faster and cells migrate faster. A major role of this inhibitory circuit, however, is to control directional persistence of migration. Indeed, Arpin depletion in both mammalian cells and Dictyostelium discoideum amoeba resulted in straighter trajectories, whereas Arpin microinjection in fish keratocytes, one of the most persistent systems of cell migration, induced these cells to turn. The coexistence of the Rac-Arpin-Arp2/3 inhibitory circuit with the Rac-WAVE-Arp2/3 activatory circuit can account for this conserved role of Arpin in steering cell migration.

Proteolytic degradation of the extracellular matrix (ECM) is one intrinsic property of metastatic tumor cells to breach tissue barriers and to disseminate into different tissues. This process is initiated by the formation of invadopodia, which are actin-driven, finger-like membrane protrusions. Yet, little is known on how invadopodia are endowed with the functional machinery of proteolytic enzymes [1, 2]. The key protease MT1-MMP (membrane type 1-matrix metalloproteinase) confers proteolytic activity to invadopodia and thus invasion capacity of cancer cells [3-6]. Here, we report that MT1-MMP-dependent matrix degradation at invadopodia is regulated by the v-SNARE TI-VAMP/VAMP7, hence providing the molecular inventory mediating focal degradative activity of cancer cells. As observed by TIRF microscopy, MT1-MMP-mCherry and GFP-VAMP7 were simultaneously detected at proteolytic sites. Functional ablation of VAMP7 decreased the ability of breast cancer cells to degrade and invade in a MT1-MMP-dependent fashion. Moreover, the number of invadopodia was dramatically decreased in VAMP7- and MT1-MMP-depleted cells, indicative of a positive-feedback loop in which the protease as a cargo of VAMP7-targeted transport vesicles regulates maturation of invadopodia. Collectively, these data point to a specific role of VAMP7 in delivering MT1-MMP to sites of degradation, maintaining the functional machinery required for invasion.