Supervillin, the largest member of the villin/gelsolin/flightless family, is a peripheral membrane protein that regulates each step of cell motility, including cell spreading. Most known interactors bind within its amino (N)-terminus. We show here that the supervillin carboxy (C)-terminus can be modeled as supervillin-specific loops extending from gelsolin-like repeats plus a villin-like headpiece. We have identified 27 new candidate interactors from yeast two-hybrid screens. The interacting sequences from 12 of these proteins (BUB1, EPLIN/LIMA1, FLNA, HAX1, KIF14, KIFC3, MIF4GD/SLIP1, ODF2/Cenexin, RHAMM, STARD9/KIF16A, Tks5/SH3PXD2A, TNFAIP1) co-localize with and mis-localize EGFP-supervillin in mammalian cells, suggesting associations in vivo. Supervillin-interacting sequences within BUB1, FLNA, HAX1, and MIF4GD also mimic supervillin over-expression by inhibiting cell spreading. Most new interactors have known roles in supervillin-associated processes, e.g. cell motility, membrane trafficking, ERK signaling, and matrix invasion; three (KIF14, KIFC3, STARD9/KIF16A) have kinesin motor domains; and five (EPLIN, KIF14, BUB1, ODF2/cenexin, RHAMM) are important for cell division. GST fusions of the supervillin G2-G3 or G4-G6 repeats co-sediment KIF14 and EPLIN, respectively, consistent with a direct association. Supervillin depletion leads to increased numbers of bi- and multi-nucleated cells. Cytokinesis failure occurs predominately during early cytokinesis. Supervillin localizes with endogenous myosin II and EPLIN in the cleavage furrow, and overlaps with the oncogenic kinesin, KIF14, at the midbody. We conclude that supervillin, like its interactors, is important for efficient cytokinesis. Our results also suggest that supervillin and its interaction partners coordinate actin and microtubule motor functions throughout the cell cycle.

In migrating cells, the cytoskeleton coordinates signal transduction and redistribution of transmembrane proteins, including integrins and growth factor receptors. Supervillin is an F-actin- and myosin II-binding protein that tightly associates with signaling proteins in cholesterol-rich, 'lipid raft' membrane microdomains. We show here that supervillin also can localize with markers for early and sorting endosomes (EE/SE) and with overexpressed components of the Arf6 recycling pathway in the cell periphery. Supervillin tagged with the photoswitchable fluorescent protein, tdEos, moves both into and away from dynamic structures resembling podosomes at the basal cell surface. Rapid integrin recycling from EE/SE is inhibited in supervillin-knockdown cells, but the rates of integrin endocytosis and recycling from the perinuclear recycling center (PNRC) are unchanged. A lack of synergy between supervillin knockdown and the actin filament barbed-end inhibitor, cytochalasin D, suggests that both treatments affect actin-dependent rapid recycling. Supervillin also enhances signaling from the epidermal growth factor receptor (EGFR) to extracellular signal-regulated kinases (ERKs) 1 and 2 and increases the velocity of cell translocation. These results suggest that supervillin, F-actin and associated proteins coordinate a rapid, basolateral membrane recycling pathway that contributes to ERK signaling and actin-based cell motility.

During cell migration, myosin II modulates adhesion, cell protrusion and actin organization at the leading edge. We show that an F-actin- and membrane-associated scaffolding protein, called supervillin (SV, p205), binds directly to the subfragment 2 domains of nonmuscle myosin IIA and myosin IIB and to the N-terminus of the long form of myosin light chain kinase (L-MLCK). SV inhibits cell spreading via an MLCK- and myosin II-dependent mechanism. Overexpression of SV reduces the rate of cell spreading, and RNAi-mediated knockdown of endogenous SV increases it. Endogenous and EGFP-tagged SV colocalize with, and enhance the formation of, cortical bundles of F-actin and activated myosin II during early cell spreading. The effects of SV are reversed by inhibition of myosin heavy chain (MHC) ATPase (blebbistatin), MLCK (ML-7) or MEK (U0126), but not by inhibiting Rho-kinase with Y-27632. Flag-tagged L-MLCK co-localizes in cortical bundles with EGFP-SV, and kinase-dead L-MLCK disorganizes these bundles. The L-MLCK- and myosin-binding site in SV, SV1-171, rearranges and co-localizes with mono- and di-phosphorylated myosin light chain and with L-MLCK, but not with the short form of MLCK (S-MLCK) or with myosin phosphatase. Thus, the membrane protein SV apparently contributes to myosin II assembly during cell spreading by modulating myosin II regulation by L-MLCK.

Tumor cells use actin-rich protrusions called invadopodia to degrade extracellular matrix (ECM) and invade tissues; related structures, termed podosomes, are sites of dynamic ECM interaction. We show here that supervillin (SV), a peripheral membrane protein that binds F-actin and myosin II, reorganizes the actin cytoskeleton and potentiates invadopodial function. Overexpressed SV induces redistribution of lamellipodial cortactin and lamellipodin/RAPH1/PREL1 away from the cell periphery to internal sites and concomitantly increases the numbers of F-actin punctae. Most punctae are highly dynamic and colocalize with the podosome/invadopodial proteins, cortactin, Tks5, and cdc42. Cortactin binds SV sequences in vitro and contributes to the formation of enhanced green fluorescent protein (EGFP)-SV induced punctae. SV localizes to the cores of Src-generated podosomes in COS-7 cells and with invadopodia in MDA-MB-231 cells. EGFP-SV overexpression increases average numbers of ECM holes per cell; RNA interference-mediated knockdown of SV decreases these numbers. Although SV knockdown alone has no effect, simultaneous down-regulation of SV and the closely related protein gelsolin reduces invasion through ECM. Together, our results show that SV is a component of podosomes and invadopodia and that SV plays a role in invadopodial function, perhaps as a mediator of cortactin localization, activation state, and/or dynamics of metalloproteinases at the ventral cell surface.

The villin-type "headpiece" domain is a modular motif found at the extreme C-terminus of larger "core" domains in over 25 cytoskeletal proteins in plants and animals. Although headpiece is classified as an F-actin-binding domain, it has been suggested that some expressed fusion-proteins containing headpiece may lack F-actin-binding in vivo. To determine the intrinsic F-actin affinity of headpiece domains, we quantified the F-actin affinity of seven headpiece domains and three N-terminal truncations, under identical in vitro conditions. The constructs are folded and adopt the native headpiece structure. However, they show a wide range of affinities that can be grouped into high, low, and nonspecific-binding categories. Computer models of the structure and charged surface potential of these headpiece domains suggest features important for high F-actin affinity. We conclude that not all headpiece domains are intrinsically F-actin-binding motifs, and suggest that the surface charge distribution may be an important element for F-actin recognition.

In previous equilibrium binding studies, Dictyostelium discoideum plasma membranes have been shown to bind actin and to recruit actin into filaments at the membrane surface. However, little is known about the kinetic pathway(s) through which actin assembles at these, or other, membranes. We have used actin fluorescently labeled with N-(1-pyrenyl)iodoacetamide to examine the kinetics of actin assembly in the presence of D. discoideum plasma membranes. We find that these membranes increase the rate of actin polymerization. The rate of membrane-mediated actin polymerization is linearly dependent on membrane protein concentrations up to 20 micrograms/ml. Nucleation (the association of activated actin monomers into oligomers) appears to be the primary step of polymerization that is accelerated. A sole effect on the initial salt-induced actin conformational change (activation) is ruled out because membranes accelerate the polymerization of pre-activated actin as well as actin activated in the presence of membranes. Elongation of preexisting filaments also is not the major step of polymerization facilitated by membranes since membranes stripped of all peripheral components, including actin, increase the rate of actin assembly to about the same extent as do membranes containing small amounts of endogenous actin. Acceleration of the nucleation step by membranes also is supported by an analysis of the dependence of polymerization lag time on actin concentration. The barbed ends of membrane-induced actin nuclei are not obstructed by the membranes because the barbed end blocking agent, cytochalasin D, reduces the rate of membrane-mediated actin nucleation. Similarly, the pointed ends of the nuclei are not blocked by membranes since the depolymerization rate of gelsolin-capped actin is unchanged in the presence of membranes. These results are consistent with previous observations of lateral interactions between membranes and actin filaments. These results also are consistent with two predictions from a model based on equilibrium binding studies; i.e., that plasma membranes should nucleate actin assembly and that membrane-bound actin nuclei should have both ends free (Schwartz, M. A., and E. J. Luna. 1988. J. Cell Biol. 107:201-209). Integral membrane proteins mediate the actin nucleation activity because activity is eliminated by heat denaturation, treatment with reducing agents, or proteolysis of membranes. Activity also is abolished by solubilization with octylglucoside but is reconstituted upon removal or dilution of the detergent. Ponticulin, the major actin-binding protein in plasma membranes, appears to be necessary for nucleation activity since activity is not reconstituted from detergent extracts depleted of ponticulin.

Activation of human platelets with thrombin transiently increases phosphorylation at (558)threonine of moesin as determined with phosphorylation state-specific antibodies. This specific modification is completely inhibited by the kinase inhibitor staurosporine and maximally promoted by the phosphatase inhibitor calyculin A, making it possible to purify the two forms of moesin to homogeneity. Blot overlay assays with F-actin probes labeled with either [32P]ATP or 125I show that only phosphorylated moesin interacts with F-actin in total platelet lysates, in moesin antibody immunoprecipitates, and when purified. In the absence of detergents, both forms of the isolated protein are aggregated. Phosphorylated, purified moesin co-sediments with alpha- or beta/gamma-actin filaments in cationic, but not in anionic, nonionic, or amphoteric detergents. The interaction affinity is high (Kd, approximately 1.5 nM), and the maximal moesin:actin stoichiometry is 1:1. This interaction is also observed in platelets extracted with cationic but not with nonionic detergents. In 0.1% Triton X-100, F-actin interacts with phosphorylated moesin only in the presence of polyphosphatidylinositides. Thus, both polyphosphatidylinositides and phosphorylation can activate moesin's high-affinity F-actin binding site in vitro. Dual regulation by both mechanisms may be important for proper cellular control of moesin-mediated linkages between the actin cytoskeleton and the plasma membrane.

No abstract provided.

Ponticulin is a 17-kD glycoprotein that represents a major high affinity link between the plasma membrane and the cortical actin network of Dictyostelium. To assess the role of ponticulin in pseudopod extension and retraction, the motile behavior of two independently generated mutants lacking ponticulin was analyzed using computer-assisted two- and three-dimensional motion analysis systems. More than half of the lateral pseudopods formed off the substratum by ponticulin-minus cells slipped relative to the substratum during extension and retraction. In contrast, all pseudopods formed off the substratum by wild-type cells were positionally fixed in relation to the substratum. Ponticulin-minus cells also formed a greater proportion of both anterior and lateral pseudopods off the substratum and absorbed a greater proportion of lateral pseudopods into the uropod than wild-type cells. In a spatial gradient of cAMP, ponticulin-minus cells were less efficient in tracking the source of chemoattractant. Since ponticulin-minus cells extend and retract pseudopods with the same time course as wild-type cells, these behavioral defects in ponticulin-minus cells appear to be the consequence of pseudopod slippage. These results demonstrate that pseudopods formed off the substratum by wild-type cells are positionally fixed in relation to the substratum, that ponticulin is required for positional stabilization, and that the loss of ponticulin and the concomitant loss of positional stability of pseudopods correlate with a decrease in the efficiency of chemotaxis.

Rac1 is a small G-protein in the Ras superfamily that has been implicated in the control of cell growth, adhesion, and the actin-based cytoskeleton. To investigate the role of Rac1 during motile processes, we have established Dictyostelium cell lines that conditionally overexpress epitope-tagged Dictyostelium discoideum wild-type Rac1B (DdRac1B) or a mutant DdRac1B protein. Expression of endogenous levels of myc- or GFP-tagged wild-type DdRac1B had minimal effect on cellular morphologies and behaviors. By contrast, expression of a constitutively active mutant (G12-->V or Q61-->L) or a dominant negative mutant (T17-->N) generated amoebae with characteristic cellular defects. The morphological appearance of actin-containing structures, intracellular levels of F-actin, and cellular responses to chemoattractant closely paralleled the amount of active DdRac1B, indicating a role in upregulating actin cytoskeletal activities. Expression of any of the three mutants inhibited cell growth and cytokinesis, and delayed multicellular development, suggesting that DdRac1B plays important regulatory role(s) during these processes. No significant effects were observed on binding or internalization of latex beads in suspension or on intracellular membrane trafficking. Cells expressing DdRac1B-G12V exhibited defects in fluid-phase endocytosis and the longest developmental delays; DdRac1B-Q61L produced the strongest cytokinesis defect; and DdRac1B-T17N generated intermediate phenotypes. These conditionally expressed DdRac1B proteins should facilitate the identification and characterization of the Rac1 signaling pathway in an organism that is amenable to both biochemical and molecular genetic manipulations.

In novel, low-speed sedimentation assays, highly purified, sonicated Dictyostelium discoideum plasma membrane fragments bind to F-actin beads (fluorescein-labeled F-actin on antifluorescein IgG-Sephacryl S-1000 beads). Binding was found to be (a) specific, since beads containing bound fluorescein-labeled ovalbumin or beads without bound fluorescein-labeled protein do not bind membranes, (b) saturable at approximately 0.6 microgram of membrane protein per microgram of bead-bound F-actin, (c) rapid with a t1/2 of 4-20 min, and (d) apparently of reasonable affinity since the off rate is too slow to be measured by present techniques. Using low-speed sedimentation assays, we found that sonicated plasma membrane fragments, after extraction with chaotropes, still bind F-actin beads. Heat-denatured membranes, proteolyzed membranes, and D. discoideum lipid vesicles did not bind F-actin beads. These results indicate that integral membrane proteins are responsible for the binding between sonicated membrane fragments and F-actin on beads. This finding agrees with the previous observation that integral proteins mediate interactions between D. discoideum plasma membranes and F-actin in solution (Luna, E.J., V. M. Fowler, J. Swanson, D. Branton, and D. L. Taylor, 1981, J. Cell Biol., 88:396-409). We conclude that low-speed sedimentation assays using F-actin beads are a reliable method for monitoring the associations between F-actin and membranes. Since these assays are relatively quantitative and require only micrograms of membranes and F-actin, they are a significant improvement over other existing techniques for exploring the biochemical details of F-actin-membrane interactions. Using F-actin beads as an affinity column for actin-binding proteins, we show that at least 12 integral polypeptides in D. discoideum plasma membranes bind to F-actin directly or indirectly. At least four of these polypeptides appear to span the membrane and are thus candidates for direct transmembrane links between the cytoskeleton and the cell surface.

Regions of plasma membrane involved in Dictyostelium discoideum intercellular adhesion resist solubilization with the nonionic detergent Triton X-100. Electron microscopy shows that these regions of the plasma membrane adhere to each other, forming many bi- and multilamellar structures. NaDodSO4/polyacrylamide gels of these regions contain major polypeptides at 225 kDa (residual myosin), 105 kDa, 88 kDa, 84 kDa, 47 kDa (residual actin), and 34 kDa. These membranes contain a subset of the total plasma membrane proteins, as analyzed by labeling of electrophoretically fractionated and blotted membrane proteins with radioiodinated Con A and by electrophoresis of membrane proteins from surface-labeled cells. Antibodies specific for gp80, a glycoprotein implicated in intercellular adhesion, intensely stain the 88-kDa and 84-kDa bands. Since these membrane regions resist Triton extraction, they appear to be stabilized by protein-protein interactions. Such stabilizing interactions may involve multivalent linkages with adjacent cells, or associations with intracellular actin and myosin, or both. Since these membranes appear to represent regions of intercellular contact, we call them "contact regions."

The binding between sonicated Dictyostelium discoideum plasma membrane fragments and F-actin on Sephacryl S-1000 beads was found to be competitively inhibited by myosin subfragment-1. This inhibition is MgATP-sensitive, exhibits a Ki of approximately 5 X 10(-8) M, and is reciprocal, since membranes inhibit the binding of 125I-heavy meromyosin to F-actin on beads. These experiments demonstrate that membrane binding and S-1 binding to F-actin on beads are mutually exclusive and, therefore, that the membrane fragments bind predominantly to the sides, rather than to the ends, of the actin filaments. This conclusion is supported by electron micrographs that show many lateral associations between membrane fragments and bead-associated actin filaments. Such lateral associations could play an important role in the organization and lateral movement of membrane proteins by the cytomusculature.

We have cloned and sequenced ponticulin, a 17,000-dalton integral membrane glycoprotein that binds F-actin and nucleates actin assembly. A single copy gene encodes a developmentally regulated message that is high during growth and early development, but drops precipitously during cell streaming at approximately 8 h of development. The deduced amino acid sequence predicts a protein with a cleaved NH2-terminal signal sequence and a COOH-terminal glycosyl anchor. These predictions are supported by amino acid sequencing of mature ponticulin and metabolic labeling with glycosyl anchor components. Although no alpha-helical membrane-spanning domains are apparent, several hydrophobic and/or sided beta-strands, each long enough to traverse the membrane, are predicted. Although its location on the primary sequence is unclear, an intracellular domain is indicated by the existence of a discontinuous epitope that is accessible to antibody in plasma membranes and permeabilized cells, but not in intact cells. Such a cytoplasmically oriented domain also is required for the demonstrated role of ponticulin in binding actin to the plasma membrane in vivo and in vitro (Hitt, A. L., J. H. Hartwig, and E. J. Luna. 1994. Ponticulin is the major high affinity link between the plasma membrane and the cortical actin network in Dictyostelium. J. Cell Biol. 126:1433-1444). Thus, ponticulin apparently represents a new category of integral membrane proteins that consists of proteins with both a glycosyl anchor and membrane-spanning peptide domain(s).

The binding of native, 125I-Bolton-Hunter-labeled actin to purified Dictyostelium discoideum plasma membranes was measured using a sedimentation assay. Binding was saturable only in the presence of the actin capping protein, gelsolin. In the presence of gelsolin, the amount of actin bound at saturation to three different membrane preparations was 80, 120, and 200 micrograms/mg of membrane protein. The respective concentrations of actin at half-saturation were 8, 12, and 18 micrograms/ml. The binding curves were sigmoidal, indicating positive cooperativity at low actin concentrations. This cooperativity appeared to be due to actin-actin associations during polymerization, since phalloidin converted the curve to a hyperbolic shape. In kinetic experiments, actin added as monomers bound to membranes at a rate of 0.6 microgram ml-1 min-1, while pre-polymerized actin bound at a rate of 3.0 micrograms ml-1 min-1. Even in the absence of phalloidin, actin bound to membranes at concentrations well below the normal critical concentration. This membrane-bound actin stained with rhodamine-phalloidin and was cross-linked by m-maleimidobenzoyl succinimide ester, a bifunctional cross-linker, into multimers with the same pattern observed for cross-linked F-actin. We conclude that D. discoideum plasma membranes bind actin specifically and saturably and that these membranes organize actin into filaments below the normal critical concentration for polymerization. This interaction probably occurs between multiple binding sites on the membrane and the side of the actin filament, and may be related to the clustering of membrane proteins.

Supervillin is a 205-kDa F-actin binding protein originally isolated from bovine neutrophils. This protein is tightly associated with both actin filaments and plasma membranes, suggesting that it forms a high-affinity link between the actin cytoskeleton and the membrane. Human supervillin cDNAs cloned from normal human kidney and from the cervical carcinoma HeLa S3 predict a bipartite structure with three potential nuclear localization signals in the NH2-terminus and three potential actin-binding sequences in the COOH-terminus. In fact, throughout its length, the COOH-terminal half of supervillin is similar to segments 2-6 plus the COOH-terminal "headpiece" of villin, an actin-binding protein in intestinal microvilli. A comparison of the bovine and human sequences indicates that supervillin is highly conserved at the amino acid level, with 79.2% identity of the NH2-terminus and conservation of three of the four nuclear localization signals found in bovine supervillin. The COOH-terminus is even more conserved, with 95.1% amino acid identity overall and 100% conservation of the villin-like headpiece. Supervillin mRNAs are expressed in all human tissue tested, bu are most abundant in muscle, bone marrow, thyroid gland, and salivary gland; comparatively little message is found in brain. Human supervillin mRNA is approximately 7.5 kb; this message is especially abundant in HeLa S3 cervical carcinoma, SW480 adenocarcinoma, and A549 lung carcinoma cell lines. The human supervillin gene (SVIL) is localized to a single chromosomal locus at 10p11.2, a region that is deleted in some prostate tumors.

Dictyostelium discoideum plasma membranes isolated by each of three procedures bind F-actin. The interactions between these membranes and actin are examined by a novel application of falling ball viscometry. Treating the membranes as multivalent actin-binding particles analogous to divalent actin-gelation factors, we observe large increases in viscosity (actin cross-linking) when membranes of depleted actin and myosin are incubated with rabbit skeletal muscle F-actin. Pre-extraction of peripheral membrane proteins with chaotropes or the inclusion of Triton X-100 during the assay does not appreciably diminish this actin cross-linking activity. Lipid vesicles, heat-denatured membranes, proteolyzed membranes, or membranes containing endogenous actin show minimal actin cross-linking activity. Heat-denatured, but not proteolyzed, membranes regain activity when assayed in the presence of Triton X-100. Thus, integral membrane proteins appear to be responsible for some or all of the actin cross-linking activity of D. discoideum membranes. In the absence of MgATP, Triton X-100 extraction of isolated D. discoideum membranes results in a Triton-insoluble residue composed of actin, myosin, and associated membrane proteins. The inclusion of MgATP before and during Triton extraction greatly diminishes the amount of protein in the Triton-insoluble residue without appreciably altering its composition. Our results suggest the existence of a protein complex stabilized by actin and/or myosin (membrane cytoskeleton) associated with the D. discoideum plasma membrane.

Membrane trafficking is central to construction of the cell plate during plant cytokinesis. Consequently, a detailed understanding of the process depends on the characterization of molecules that function in the formation, transport, targeting, and fusion of membrane vesicles to the developing plate, as well as those that participate in its consolidation and maturation into a fully functional partition. Here we report the initial biochemical and functional characterization of patellin1 (PATL1), a novel cell-plate-associated protein that is related in sequence to proteins involved in membrane trafficking in other eukaryotes. Analysis of the Arabidopsis genome indicated that PATL1 is one of a small family of Arabidopsis proteins, characterized by a variable N-terminal domain followed by two domains found in other membrane-trafficking proteins (Sec14 and Golgi dynamics domains). Results from immunolocalization and biochemical fractionation studies suggested that PATL1 is recruited from the cytoplasm to the expanding and maturing cell plate. In vesicle-binding assays, PATL1 bound to specific phosphoinositides, important regulators of membrane trafficking, with a preference for phosphatidylinositol(5)P, phosphatidylinositol(4,5)P(2), and phosphatidylinositol(3)P. Taken together, these findings suggest a role for PATL1 in membrane-trafficking events associated with cell-plate expansion or maturation and point to the involvement of phosphoinositides in cell-plate biogenesis.

One-dimensional and two-dimensional peptide-mapping techniques are used to identify the protein which gives rise to the 72,000 dalton alpha-chymotryptic fragment previously shown to be the membrane attachment site for spectrin. Peptide maps of the 72,000 dalton fragment are very different from maps of Bands 1, 2, 2.9, 3, 3.1, 4.1, and 4.2 and very similar to maps of the apparently closely homologous polypeptides, Bands 2.1, 2.2, 2.3, and 2.6. Limited proteolysis of erythrocyte membranes is shown to generate Band 3', another polypeptide which has been associated with spectrin-binding activity. Peptide maps of Band 3' are very similar to maps of Band 2.1, suggesting that Band 3' is also a proteolytic fragment of Band 2.1. It is concluded that Band 2.1 and possibly some or all of the other, related polypeptides which electrophorese in the 2 region is (are) the spectrin-binding protein(s) of the human erythrocyte.

Actin-binding membrane proteins are involved in both adhesive interactions and motile processes. We report here the purification and initial characterization of p205, a 205-kD protein from bovine neutrophil plasma membranes that binds to the sides of actin filaments in blot overlays. p205 is a tightly bound peripheral membrane protein that cosediments with endogenous actin in sucrose gradients and immunoprecipitates. Amino acid sequences were obtained from SDS-PAGE-purified p205 and used to generate antipeptide antibodies, immunolocalization data, and cDNA sequence information. The intracellular localization of p205 in MDBK cells is a function of cell density and adherence state. In subconfluent cells, p205 is found in punctate spots along the plasma membrane and in the cytoplasm and nucleus; in adherent cells, p205 concentrates with E-cadherin at sites of lateral cell-cell contact. Upon EGTA-mediated cell dissociation, p205 is internalized with E-cadherin and F-actin as a component of adherens junctions "rings." At later times, p205 is observed in cytoplasmic punctae. The high abundance of p205 in neutrophils and suspension-grown HeLa cells, which lack adherens junctions, further suggests that this protein may play multiple roles during cell growth, adhesion, and motility. Molecular cloning of p205 cDNA reveals a bipartite structure. The COOH terminus exhibits a striking similarity to villin and gelsolin, particularly in regions known to bind F-actin. The NH2 terminus is novel, but contains four potential nuclear targeting signals. Because p205 is now the largest known member of the villin/gelsolin superfamily, we propose the name, "supervillin." We suggest that supervillin may be involved in actin filament assembly at adherens junctions and that it may play additional roles in other cellular compartments.