A Novel Actin Bundling/Filopodium-forming Domain Conserved in Insulin Receptor Tyrosine Kinase Substrate p53 and Missing in Metastasis Protein
2004; Elsevier BV; Volume: 279; Issue: 15 Linguagem: Inglês
10.1074/jbc.m309408200
ISSN1083-351X
AutoresAkiko Yamagishi, Michitaka Masuda, Takashi Ohki, Hirofumi Onishi, Naoki Mochizuki,
Tópico(s)Microtubule and mitosis dynamics
ResumoInsulin receptor tyrosine kinase substrate p53 (IRSp53) has been identified as an SH3 domain-containing adaptor that links Rac1 with a Wiskott-Aldrich syndrome family verprolin-homologous protein 2 (WAVE2) to induce lamellipodia or Cdc42 with Mena to induce filopodia. The recruitment of these SH3-binding partners by IRSp53 is thought to be crucial for F-actin rearrangements. Here, we show that the N-terminal predicted helical stretch of 250 amino acids of IRSp53 is an evolutionarily conserved F-actin bundling domain involved in filopodium formation. Five proteins including IRSp53 and missing in metastasis (MIM) protein share this unique domain and are highly conserved in vertebrates. We named the conserved domain IRSp53/MIM homology domain (IMD). The IMD has domain relatives in invertebrates but does not show obvious homology to any known actin interacting proteins. The IMD alone, derived from either IRSp53 or MIM, induced filopodia in HeLa cells and the formation of tightly packed parallel F-actin bundles in vitro. These results suggest that IRSp53 and MIM belong to a novel actin bundling protein family. Furthermore, we found that filopodium-inducing IMD activity in the full-length IRSp53 was regulated by active Cdc42 and Rac1. The SH3 domain was not necessary for IMD-induced filopodium formation. Our results indicate that IRSp53, when activated by small GTPases, participates in F-actin reorganization not only in an SH3-dependent manner but also in a manner dependent on the activity of the IMD. Insulin receptor tyrosine kinase substrate p53 (IRSp53) has been identified as an SH3 domain-containing adaptor that links Rac1 with a Wiskott-Aldrich syndrome family verprolin-homologous protein 2 (WAVE2) to induce lamellipodia or Cdc42 with Mena to induce filopodia. The recruitment of these SH3-binding partners by IRSp53 is thought to be crucial for F-actin rearrangements. Here, we show that the N-terminal predicted helical stretch of 250 amino acids of IRSp53 is an evolutionarily conserved F-actin bundling domain involved in filopodium formation. Five proteins including IRSp53 and missing in metastasis (MIM) protein share this unique domain and are highly conserved in vertebrates. We named the conserved domain IRSp53/MIM homology domain (IMD). The IMD has domain relatives in invertebrates but does not show obvious homology to any known actin interacting proteins. The IMD alone, derived from either IRSp53 or MIM, induced filopodia in HeLa cells and the formation of tightly packed parallel F-actin bundles in vitro. These results suggest that IRSp53 and MIM belong to a novel actin bundling protein family. Furthermore, we found that filopodium-inducing IMD activity in the full-length IRSp53 was regulated by active Cdc42 and Rac1. The SH3 domain was not necessary for IMD-induced filopodium formation. Our results indicate that IRSp53, when activated by small GTPases, participates in F-actin reorganization not only in an SH3-dependent manner but also in a manner dependent on the activity of the IMD. Insulin receptor tyrosine kinase substrate p53 (IRSp53), 1The abbreviations used are: IRSp53, insulin receptor tyrosine kinase substrate p53; CRIB, Cdc42/Rac interactive binding; MIM, missing in metastasis protein; SH3, Src homology 3; WWB, WW domain-binding motif; WASP, Wiskott-Aldrich syndrome protein; WH2, WASP homology 2; IMD, IRSp53/MIM homology domain; GFP, green fluorescent protein; GST, glutathione S-transferase; IRTKS, insulin receptor tyrosine kinase substrate; aa, amino acids. 1The abbreviations used are: IRSp53, insulin receptor tyrosine kinase substrate p53; CRIB, Cdc42/Rac interactive binding; MIM, missing in metastasis protein; SH3, Src homology 3; WWB, WW domain-binding motif; WASP, Wiskott-Aldrich syndrome protein; WH2, WASP homology 2; IMD, IRSp53/MIM homology domain; GFP, green fluorescent protein; GST, glutathione S-transferase; IRTKS, insulin receptor tyrosine kinase substrate; aa, amino acids. also known as brain-specific angiogenesis inhibitor 1-associated protein 2, is a multifunctional adaptor protein enriched in the central nervous system (1Abbott M.A. Wells D.G. Fallon J.R. J. Neurosci. 1999; 19: 7300-7308Crossref PubMed Google Scholar, 2Oda K. Shiratsuchi T. Nishimori H. Inazawa J. Yoshikawa H. Taketani Y. Nakamura Y. Tokino T. Cytogenet. Cell Genet. 1999; 84: 75-82Crossref PubMed Scopus (72) Google Scholar, 3Yeh T.C. Ogawa W. Danielsen A.G. Roth R.A. J. Biol. Chem. 1996; 271: 2921-2928Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar). The protein contains a unique N-terminal 250-amino acid stretch, a half-Cdc42/Rac interactive binding (CRIB) motif, a proline-rich domain, a Src homology 3 (SH3) domain, and a WW domain-binding motif (WWB). IRSp53 is directly regulated by Rho family small GTPases Rac1 and Cdc42 and provides a molecular link between these GTPases and the actin cytoskeleton regulators Wiskott-Aldrich syndrome protein (WASP) family verprolin homologous protein 2 (WAVE2) and mammalian enabled (Mena), which are involved in the formation of lamellipodia (4Miki H. Yamaguchi H. Suetsugu S. Takenawa T. Nature. 2000; 408: 732-735Crossref PubMed Scopus (457) Google Scholar, 5Takenawa T. Miki H. J. Cell Sci. 2001; 114: 1801-1809Crossref PubMed Google Scholar) and filopodia (6Govind S. Kozma R. Monfries C. Lim L. Ahmed S. J. Cell Biol. 2001; 152: 579-594Crossref PubMed Scopus (138) Google Scholar, 7Krugmann S. Jordens I. Gevaert K. Driessens M. Vandekerckhove J. Hall A. Curr. Biol. 2001; 11: 1645-1655Abstract Full Text Full Text PDF PubMed Scopus (309) Google Scholar). Active Cdc42 binds to the half-CRIB motif (6Govind S. Kozma R. Monfries C. Lim L. Ahmed S. J. Cell Biol. 2001; 152: 579-594Crossref PubMed Scopus (138) Google Scholar, 7Krugmann S. Jordens I. Gevaert K. Driessens M. Vandekerckhove J. Hall A. Curr. Biol. 2001; 11: 1645-1655Abstract Full Text Full Text PDF PubMed Scopus (309) Google Scholar), whereas Rac1 binds to the unique N-terminal domain (8Miki H. Takenawa T. Biochem. Biophys. Res. Commun. 2002; 293: 93-99Crossref PubMed Scopus (64) Google Scholar). The association of Rac1 or Cdc42 is proposed to liberate the C-terminal SH3 domain masked intramolecularly by its N terminus, thereby allowing the SH3 domain to interact with its binding partners (4Miki H. Yamaguchi H. Suetsugu S. Takenawa T. Nature. 2000; 408: 732-735Crossref PubMed Scopus (457) Google Scholar, 7Krugmann S. Jordens I. Gevaert K. Driessens M. Vandekerckhove J. Hall A. Curr. Biol. 2001; 11: 1645-1655Abstract Full Text Full Text PDF PubMed Scopus (309) Google Scholar, 9Alvarez C.E. Sutcliffe J.G. Thomas E.A. J. Biol. Chem. 2002; 277: 24728-24734Abstract Full Text Full Text PDF PubMed Scopus (21) Google Scholar). Thus, the SH3 domain is thought to be essential for IRSp53-mediated actin reorganization. However, the N-terminal half of IRSp53 lacking the SH3 domain was reported to induce neurite outgrowth in a neuroblastoma cell line (6Govind S. Kozma R. Monfries C. Lim L. Ahmed S. J. Cell Biol. 2001; 152: 579-594Crossref PubMed Scopus (138) Google Scholar) and filopodia in B16 melanoma cells (10Nakagawa H. Miki H. Nozumi M. Takenawa T. Miyamoto S. Wehland J. Small J.V. J. Cell Sci. 2003; 116: 2577-2583Crossref PubMed Scopus (111) Google Scholar), suggesting that IRSp53 promotes actin reorganization independently of SH3 domain-mediated intermolecular interactions. Recently, a novel monomeric actin-binding protein, missing in metastasis protein (MIM), containing a WASP homology 2 (WH2) domain in the C terminus, was reported in human and mouse (11Lee Y.G. Macoska J.A. Korenchuk S. Pienta K.J. Neoplasia. 2002; 4: 291-294Crossref PubMed Scopus (141) Google Scholar, 12Mattila P.K. Salminen M. Yamashiro T. Lappalainen P. J. Biol. Chem. 2003; 278: 8452-8459Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 13Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (92) Google Scholar) and found to share the unique N-terminal domain with IRSp53 (13Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (92) Google Scholar). We found that the N-terminal domains of MIM and IRSp53 also share other characteristic features; the predicted secondary structures are almost purely helical (see Ref. 9Alvarez C.E. Sutcliffe J.G. Thomas E.A. J. Biol. Chem. 2002; 277: 24728-24734Abstract Full Text Full Text PDF PubMed Scopus (21) Google Scholar for IRSp53), and the estimated isoelectric points are around 9. MIM induces actin cytoskeleton reorganization in cultured cells. This activity is not dependent on the C-terminal half (13Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (92) Google Scholar), suggesting that the N-terminal half containing the IRSp53 homologous domain plays a key role in actin reorganization. Here we show that IRSp53 and MIM belong to an evolutionarily related protein family sharing a well conserved N-terminal helical domain (IRSp53/MIM homology domain (IMD)) as a key constituent. We investigated the role of the IMD in actin reorganization. Our results indicate that the IMDs of IRSp53 and MIM induce filopodia in cultured cells and form tightly packed F-actin bundles in vitro. The filopodium forming activity of the IMD in full-length IRSp53 is regulated by small GTPases. Thus, upon association with active Rac1 or Cdc42, IRSp53 can induce actin cytoskeleton reorganization by dual mechanisms: the SH3-mediated recruitment of F-actin regulators and the action of the novel actin bundling domain in the N terminus. Both mechanisms may work synergistically or additively in controlling cortical actin dynamics. Data Base Search—Proteins homologous to IRSp53 and MIM were identified in the GenBank™ data base using BLAST on the National Center for Biotechnology Information web site and on the GenomeNet web site. The details of sequences thus retrieved are described in the supplemental data. The Clustalw engine at the GenomeNet Web site was used to align amino acid sequences and to construct phylogenetic trees. The secondary structures of proteins were predicted by 3D-PSSM (14Kelley L.A. MacCallum R.M. Sternberg M.J. J. Mol. Biol. 2000; 299: 499-520Crossref PubMed Scopus (1120) Google Scholar). Plasmids—The IRSp53 expression vector pEF-BOS-Myc-IRSp53 (human isoform 1) was kindly donated by Dr. Miki (4Miki H. Yamaguchi H. Suetsugu S. Takenawa T. Nature. 2000; 408: 732-735Crossref PubMed Scopus (457) Google Scholar). cDNAs encoding full-length IRSp53 (amino acids (aa) 1-521), IRSp53-ΔSH3 (aa 1-364), IRSp53-IMD (aa 1-250), and IRSp53-ΔIMD (aa 251-521) (see Fig. 1B were amplified by PCR and inserted into pEGFP-C1 (Clontech), pCXN2-FLAG (15Niwa H. Yamamura K. Miyazaki J. Gene (Amst.). 1991; 108: 193-199Crossref PubMed Scopus (4593) Google Scholar), and pGEX-4T3 or 6P3 (Amersham Biosciences) vectors. The DNA fragments encoding IRSp53 where Arg was substituted for Trp413 or Ala for both Phe427 and Pro428 in the SH3 domain, hereafter referred to as IRSp53-W/R or IRSp53-FP/AA, were amplified by PCR and ligated into pEGFP-C1. cDNA of Rac1V12, Rac1N17, Cdc42V12, or Cdc42N17 was subcloned into pIRM21, an expression vector expressing FLAG-tagged protein and internal ribosomal entry site-driven dsFP593 (16Nagashima K. Endo A. Ogita H. Kawana A. Yamagishi A. Kitabatake A. Matsuda M. Mochizuki N. Mol. Biol. Cell. 2002; 13: 4231-4242Crossref PubMed Scopus (84) Google Scholar). A cDNA clone encoding the N-terminal fragment (aa 1-430) of human MIM was obtained by PCR from a human brain cDNA library (Clontech). The cDNA encoding C-terminal MIM (aa 400-755, KIAA0429) was obtained from the Kazusa DNA Research Institute. The full-length MIM cDNA was amplified through overlap PCR using these N- and C-terminal cDNAs as templates. The cDNAs encoding the full-length human MIM (aa 1-755), MIM-IMD (aa 1-250), and MIM-ΔIMD (aa 251-755) were inserted into pEGFP-C1, pCXN2-FLAG, and pGEX-4T3 or 6P3 vectors. Cells and Transfection—HeLa cells and 293T cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and2mml-glutamine. HeLa cells and 293T cells were transfected using LipofectAMINE (Invitrogen) according to the manufacturer's protocol. Antibodies and Immunofluorescence Analysis—Rhodamine-conjugated phalloidin and Alexa546-conjugated anti-mouse IgG were purchased from Molecular Probes (Eugene, OR); anti-FLAG M2 antibody was from Sigma-Aldrich. HeLa cells transfected with the plasmids indicated in the figures and cultured for 15-18 h were fixed with 2% formaldehyde in phosphate-buffered saline and permeabilized with 0.1% Triton X-100 in phosphate-buffered saline. The cells transfected with plasmids expressing GFP-tagged proteins were counterstained with rhodamine-phalloidin. The cells transfected with both GFP-tagged protein-expressing vectors and FLAG-tagged small GTPase-expressing vectors were immunostained with anti-FLAG M2 antibody followed by an Alexa546-conjugated anti-mouse IgG. Fluorescence images were obtained using a confocal microscope (BX50WI, Fluoview, Olympus, Tokyo, Japan) with a water immersion objective lens (LUMPlanFl 60×, 0.90 W). To show the entire cell morphology in detail, all of the cell images shown were extended focus images reconstructed from a series of optical sections taken at 0.2-0.3-μm intervals. F-actin Binding and Bundling Assays—F-actin was prepared from rabbit skeletal muscle as described (17Spudich J.A. Watt S. J. Biol. Chem. 1971; 246: 4866-4871Abstract Full Text PDF PubMed Google Scholar). Glutathione S-transferase (GST) fusion proteins of various fragments of IRSp53 and MIM (see Fig. 1B) were expressed in BL21-Star (DE3) cells (Invitrogen), purified using glutathione-Sepharose (Amersham Biosciences), and then buffer-exchanged into F buffer (25 mm Hepes, pH 7.5, 100 mm KCl, 0.2 mm CaCl2, 2 mm MgCl2, 2 mm EGTA, 0.2 mm ATP, 1 mm dithiothreitol) containing 0.1% C12E8 (Nikko Chemicals, Tokyo, Japan). For binding assays, purified GST-fused fragments were clarified by centrifugation at 400,000 × g for 15 min to remove any aggregates, mixed with F-actin in the F buffer, and incubated for 30 min on ice. The final concentration of the GST fusions and F-actin were 1.2 and 5 μm (as for G-actin), respectively. The mixture was then centrifuged as above, and equal aliquots of the supernatant and the pellet were analyzed by SDS-PAGE followed by Coomassie Blue staining. For quantitative analysis of F-actin binding and bundling, the IMDs were cleaved out from the GST fusions expressed by pGEX-6P3 vectors using PreScission Protease (Amersham Biosciences) and further purified by cation exchange chromatography (Resource S; Amersham Biosciences). To quantify F-actin binding, increasing amounts of F-actin were incubated with 2 μm IRSp53-IMD or MIM-IMD in the F buffer for 3 h at room temperature. The samples were then centrifuged and analyzed as above. The protein bands were quantified by densitometry (Personal Densitometer SI; Amersham Biosciences). For quantitative bundling assay, increasing amounts of the IMDs were incubated with 1 μm F-actin in the F buffer for 1 h at room temperature. The supernatant and the pellet were separated by low speed centrifugation (10,000 × g for 30 min) and analyzed as above. Observation of Actin Bundles—For fluorescence microscope observation, a fixed concentration of F-actin (final concentration, 1.2 μm) was mixed with variable concentrations of the GST-fused fragments (0.24 to 12 μm). After incubation for 30 min on ice in F buffer, F-actin was stained with rhodamine-phalloidin for 15 min on ice. The mixtures were applied to poly-l-lysine-coated glass coverslips and incubated for 20 min at room temperature. The adherent material was washed with F buffer and observed with the confocal laser scanning microscope using an oil immersion objective lens (PlanApo 60×, 1.40 oil). For negative staining of actin filaments and bundles, the rhodamine-phalloidin-stained specimens described above were diluted 10 times with F buffer, placed onto a carbon-coated mesh, and stained with 2% uranyl acetate. For observation of thin sectioned specimens, actin bundles formed after incubation for 1 h on ice were packed by centrifugation and fixed in 2.5% glutaraldehyde in 0.1 m cacodylate buffer, pH 7.4, and then sequentially incubated with 0.1% aqueous tannic acid and 0.2% uranyl acetate (18Svitkina T.M. Bulanova E.A. Chaga O.Y. Vignjevic D.M. Kojima S. Vasiliev J.M. Borisy G.G. J. Cell Biol. 2003; 160: 409-421Crossref PubMed Scopus (583) Google Scholar), postfixed in 0.5% aqueous OsO4, dehydrated, and embedded in Epon 812. Thin sections stained by lead citrate were examined with a CM 120 electron microscope (Philips Electronics, Eindhoven, The Netherlands) equipped with a multiscan cooled charge-coupled device camera (model 791; Gatan, Pleasanton, CA). Cross-linking of Proteins—One μm of purified IRSp53-IMD, MIM-IMD, and chymotrypsinogen A (Amersham Biosciences) as the control were cross-linked in 0.1 m 2-morpholinoethanesulfonic acid, pH 5.0, at room temperature. The reaction was started by the addition of 4 mm 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide and was stopped by the addition of 50 mm Tris-HCl, pH 8.0, at the time points indicated in Fig. 7A. Immunoprecipitation—293T cells were washed with phosphate-buffered saline and lysed in lysis buffer (100 mm NaCl, 25 mm Hepes, pH 7.5, 2 mm MgCl2, 2 mm EGTA, 0.5% Triton X-100, containing protease inhibitor mixture; Roche Applied Science). The lysates were precleared by centrifugation at 100,000 × g for 10 min, followed by immunoprecipitation with a rabbit anti-GFP antibody and protein A-Sepharose beads (Amersham Biosciences). The immunoprecipitates were subjected to SDS-PAGE and immunoblotting with antibodies as indicated in Fig. 7B. The proteins reacting with primary antibodies were visualized by an enhanced chemiluminescence system (Amersham Biosciences) for detecting peroxidase-conjugated secondary antibodies and analyzed with an LAS-1000 system (Fuji Film, Tokyo, Japan). The N-terminal Helical Domain Is Evolutionarily Conserved in IRSp53 Family Proteins and MIM Family Proteins—IRSp53 and MIM share the N-terminal stretch of 250 amino acids (22% identical, 18% similar), whereas the remaining parts of the molecules show only marginal similarity. To explore whether this similarity is based on real homology, we searched the GenBank™ data base for proteins having similar sequences. First we found three more genes encoding homologous N-terminal sequences in the human genome: insulin receptor tyrosine kinase substrate (IRTKS), the hypothetical gene FLJ22582, and ABBA-1 (Fig. 1A; see the supplemental table for details). IRTKS and FLJ22582 are IRSp53-related proteins containing an SH3 domain in the C-terminal half. However, both of them lack the half-CRIB motif found in IRSp53, and FLJ22582 further lacks the WWB (PPPXY) (Fig. 1B). ABBA-1 (GenBank™ accession number AB115770) is a MIM-related protein that possesses a WH2 domain in the C terminus. Further data base searches have shown that each of these five proteins has a putative ortholog in chicken and zebra fish, indicating that they are well conserved through vertebrate evolution (Supplemental Figs. 1-4). As pointed out previously (9Alvarez C.E. Sutcliffe J.G. Thomas E.A. J. Biol. Chem. 2002; 277: 24728-24734Abstract Full Text Full Text PDF PubMed Scopus (21) Google Scholar, 13Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (92) Google Scholar), related proteins are also found in invertebrates (Caenorhabditis elegans M04F3.5 protein and Drosophila melanogaster CG32082 protein). In an amino acid sequence alignment of the N-terminal region of these proteins (Fig. 1A), clusters of basic amino acids, proline, glycine, and clusters of hydrophobic amino acids are well conserved. There is a signature sequence of ALXEE(R/K)(R/G)RFCX0-1F(I/L) in the C-terminal half of the stretch. As expected from the number of basic amino acid clusters in this domain, the estimated isoelectric points are highly basic, ranging from 8.5 for human MIM to 9.2 for human FLJ22582. The identity of the N-terminal domain is further supported by predicted secondary structures. The domains are almost purely helical, 82-87% for the IRSp53-related proteins, 96-100% for MIM-related proteins, and intermediate contents of 89 and 87% for M04F3.5 and CG32082, respectively. Helix-breaking amino acid residues at the four breaking sites of the IRSp53-related proteins are also conserved in MIM/ABBA family proteins (asterisks in Fig. 1A). Thus, all of these proteins appear to have a common segmentation pattern of helices, helix I-V (Fig. 1A). Although human IRSp53 lacks helix V, it is predicted to be present in the chicken ortholog. A phylogenetic tree (Fig. 1C) based on the alignment of the IMDs shows that the vertebrate IRSp53/MIM family is divided into two major groups: the IRSp53 subfamily and the MIM/ABBA subfamily. The putative invertebrate homologs are positioned between them. The tree of the IMDs exactly reflects the hierarchy of domain composition of these proteins. The IRSp53 subfamily members contain an SH3 domain, and the MIM/ABBA subfamily proteins contain a WH2 domain. The vertebrate SH3-containing subfamily is further divided into three groups according to the presence or absence of the WWB and the half-CRIB motif. These data suggest that the IRSp53/MIM family originated from a common ancestor and diverged through evolution. This hypothesis is supported by the fact that IRTKS and FLJ22582 but not M04F3.5 or CG32082 share highly homologous C termini with the MIM/ABBA subfamily members (supplemental Fig. 6). Our analyses suggest the presence of an evolutionarily conserved IRSp53/MIM family and that the IMDs are the key components for the functional roles of proteins belonging to this family. The IMDs of IRSp53 and MIM Induce Filopodia in HeLa Cells—To explore the functional roles of the IMD, we first examined the morphological effects of ectopic expression of the IRSp53-IMD and the MIM-IMD in HeLa cells. The cells expressing the GFP-tagged IMD of IRSp53 formed numerous long filopodia that were F-actin-rich as demonstrated by rhodamine-phalloidin staining (Fig. 2, panels a, a′, b, and b′). The MIM-IMD also induced filopodia, but they were reduced in length (Fig. 2, panels c, c′, d, and d′). In addition, MIM-IMD promoted the formation of microvillus-like protrusions on the apical cell surface. IRSp53-IMD and MIM-IMD localized to and occasionally were concentrated in these protrusions (arrows in Fig. 2, panels c′ and d′). Both IMDs appeared not to be associated with stress fibers. There were no obvious signs of enhanced lamellipodial activity or disruption of stress fibers in these IMD-expressing cells. GFP used as a negative control did not induce any morphological changes (Fig. 2, panels e and f). Truncated fragments of IMD, IRSp53-N-IMD (aa 1-161), and IRSp53-C-IMD (aa 105-250) could not stimulate filopodium-formation (data not shown). These data indicate that both IMDs are capable of inducing filopodia in cells. Because IRSp53 and MIM represent the most divergent members of the vertebrate IRSp53/MIM protein family (Fig. 1C), the filopodium inducing activity of the IMD is likely to be conserved in all family members. IMD Does Not Act Upstream of Rac1 or Cdc42 for Filopodium Formation—Actin cytoskeletal reorganization is often a hallmark of Rho family GTPases. Previous reports have shown that Rac1 binds to the N terminus of IRSp53 (8Miki H. Takenawa T. Biochem. Biophys. Res. Commun. 2002; 293: 93-99Crossref PubMed Scopus (64) Google Scholar) and Cdc42 binds to the aa 202-305 fragment containing the half-CRIB motif (6Govind S. Kozma R. Monfries C. Lim L. Ahmed S. J. Cell Biol. 2001; 152: 579-594Crossref PubMed Scopus (138) Google Scholar). Therefore, we examined whether Cdc42 or Rac1 activation was involved in IMD-induced filopodium formation. The formation of numerous filopodia induced by IRSp53-IMD was not perturbed by the co-expression of dominant negative Cdc42 or Rac1 (Fig. 3). There was no quantitative difference in the ratio of filopodium forming cells among HeLa cells transfected with IRSp53-IMD alone, those transfected with IRSp53-IMD and Cdc42N17, and those transfected with IRSp53 and Rac1N17 (Fig. 3, panel d), suggesting that the IMD itself is not regulated by these small GTPases. This result also excludes the possibility that the domain functions upstream of these small GTPases. The Filopodium-inducing IMD Activity of Wild-type IRSp53 Is Regulated by Cdc42 and Rac1—The common mechanism of effecter activation by Rho family GTPases appears to be dependent on the disruption of intramolecular autoinhibitory interactions. Cdc42-induced conformational changes have also been demonstrated for the molecule containing a half or semi-CRIB motif, Par6 (19Garrard S.M. Capaldo C.T. Gao L. Rosen M.K. Macara I.G. Tomchick D.R. EMBO J. 2003; 22: 1125-1133Crossref PubMed Scopus (134) Google Scholar). First we found that GFP-tagged IRSp53-WT or GFP-tagged IRSp53-ΔSH3, when expressed in moderate levels, could not induce filopodia (Fig. 4A, panels a-c). As reported earlier (7Krugmann S. Jordens I. Gevaert K. Driessens M. Vandekerckhove J. Hall A. Curr. Biol. 2001; 11: 1645-1655Abstract Full Text Full Text PDF PubMed Scopus (309) Google Scholar, 10Nakagawa H. Miki H. Nozumi M. Takenawa T. Miyamoto S. Wehland J. Small J.V. J. Cell Sci. 2003; 116: 2577-2583Crossref PubMed Scopus (111) Google Scholar), cells expressing very high amounts of IRSp53 often formed dendritic extensions accompanied with severe retraction of the cell body. As noted in the legend to Fig. 2, these cells were omitted from our analyses. Next, we examined whether the IMD function was regulated by Cdc42 and Rac1 in IRSp53-WT and in IRSp53-ΔSH3 containing the half-CRIB motif. Co-expression of the active Cdc42 with these IRSp53 constructs led to massive formation of wavy filopodia (IRSp53+Cdc42 phenotype as shown in Fig. 4A, panels d and e) that was clearly distinguishable from straight filopodia induced in cells co-expressing GFP and active Cdc42 (Cdc42 phenotype as shown in Fig. 4A, panel f). A similar level of filopodium induction mixed with Rac1-dependent enhanced lamellipodia activity (Fig. 4A, panel i) was induced by the co-expression of active Rac1 (Fig. 4A, panels g and h). These results suggest that the SH3 domain is not necessary for IMD-dependent filopodium formation. Our results also suggest that the filopodium-inducing IMD activity in wild-type IRSp53 is regulated by Cdc42 and Rac1. The central region of IRSp53 containing the half-CRIB motif appears to be essential for this regulation, as previously suggested for the regulation of the SH3 domain (7Krugmann S. Jordens I. Gevaert K. Driessens M. Vandekerckhove J. Hall A. Curr. Biol. 2001; 11: 1645-1655Abstract Full Text Full Text PDF PubMed Scopus (309) Google Scholar, 9Alvarez C.E. Sutcliffe J.G. Thomas E.A. J. Biol. Chem. 2002; 277: 24728-24734Abstract Full Text Full Text PDF PubMed Scopus (21) Google Scholar). To further confirm that the IMD-induced filopodium formation is independent of SH3-binding molecules, we used two nonfunctional SH3 mutants, IRSp53-W/R and IRSp53-FP/AA (7Krugmann S. Jordens I. Gevaert K. Driessens M. Vandekerckhove J. Hall A. Curr. Biol. 2001; 11: 1645-1655Abstract Full Text Full Text PDF PubMed Scopus (309) Google Scholar). Both mutants could induce filopodia when expressed with active Cdc42 (Fig. 4B). Although active Cdc42 alone induced filopodium formation in the majority of cells (Fig. 4B, panel d, the blank segment in the stacked bar graph reflecting the Cdc42 phenotype), exaggerated filopodium formation, an indication of IMD activity, occurred only when Cdc42 was co-expressed with IRSp53-WT and IRSp53-SH3 mutants (Fig. 4B, panel d, the filled segment in the stacked bar graph reflecting the IRSp53+Cdc42 phenotype). Thus, IRSp53 can promote filopodium formation independently of SH3-mediated intermolecular interactions. In Vitro F-actin Bundling Activity of IMD—The filopodium promoting activity of the IMDs of IRSp53 and MIM in cultured cells led us to examine whether these IMDs have F-actin binding and bundling activity. We examined F-actin binding/bundling activity of the GST-fused IMD and other fragments and also tag-free purified IMDs in vitro. As shown in Fig. 5A, GST-fused IRSp53-IMD, IRSp53-ΔSH3, and MIM-IMD but not GST were co-sedimented with F-actin in a high speed assay (total binding). To exclude the possible contribution of GST-tag or contaminating bacterial proteins to F-actin binding and bundling, the activities of purified tag-free IMDs (Fig. 5B, left panel) were examined. In the high speed assays, the IMDs of IRSp53 and MIM bound to F-actin in a concentration-dependent and saturable manner (Fig. 5B, right panel). The apparent half-maximum concentrations of F-actin for IMD binding were almost the same (0.5 μm), irrespective of the variation between the maximum extents of these IMDs, suggesting that both IMDs have roughly the same affinity to F-actin. Low levels of the maximum extent of bound IMDs, about 30% for IRSp53-IMD and 20% for MIM-IMD, can be explained by improper protein folding of the bacterially made IMDs or their denaturation during the purification process. The GST fusions capable of F-actin binding induced thick F-actin bundles (Fig. 5C). Although GFP-tagged IRSp53-ΔSH3 required activation by Rac1 or Cdc42 for filopodium formation, GST-fused IRSp53-ΔSH3 alone could induce F-actin bundling. It is possible that the bacterially made protein may not be folded properly to form the self-inhibitory conformation. IRSp53-ΔSH3 showed stronger bundling than IRSp53-IMD or MIM-IMD; however, different levels of bundling activity among these proteins may simply reflect differences in stability of these fusio
Referência(s)