Tyrosine Phosphorylation of Missing in Metastasis Protein Is Implicated in Platelet-derived Growth Factor-mediated Cell Shape Changes
2007; Elsevier BV; Volume: 282; Issue: 10 Linguagem: Inglês
10.1074/jbc.m608448200
ISSN1083-351X
AutoresYing Wang, Kang Zhou, Xian‐Chun Zeng, Jinxiu Lin, Xi Zhan,
Tópico(s)TGF-β signaling in diseases
ResumoMissing in metastasis gene, or MTSS1, encodes an intracellular protein that is implicated in actin cytoskeleton reorganization and often down-regulated in certain types of tumor cells. In response to platelet-derived growth factor (PDGF), green fluorescent protein (GFP)-tagged murine Mtss1 (Mtss1-GFP) underwent redistribution from the cytoplasm to dorsal membrane ruffles along with phosphorylation at tyrosine residues in a time-dependent manner. Tyrosine phosphorylation of Mtss1-GFP was also elevated in cells where an oncogenic Src was activated but severely impaired in Src knock-out cells or cells treated with Src kinase inhibitor PP2. Mutagenesis analysis has revealed that phosphorylation occurs at multiple sites, including tyrosine residues Tyr-397 and Tyr-398. Mutation at both Tyr-397 and Tyr-398 abolished the PDGF-mediated tyrosine phosphorylation. Furthermore, recombinant Mtss1 protein was phosphorylated by recombinant Src in a manner dependent on Tyr-397 and Tyr-398. Efficient tyrosine phosphorylation of Mtss1 in response to PDGF also involves a coiled-coil domain, which is essential for a proper distribution to the cell leading edge and dorsal ruffles. Interestingly, overexpression of wild type Mtss1-GFP promoted the PDGF-induced dorsal ruffling, whereas overexpression of a mutant deficient in phosphorylation at Tyr-397 and Tyr-398 or a mutant with deletion of the coiled-coil domain impaired the formation of dorsal ruffles. These data indicate that Mtss1 represents a novel signaling pathway from PDGF receptor to the actin cytoskeleton via Src-related kinases. Missing in metastasis gene, or MTSS1, encodes an intracellular protein that is implicated in actin cytoskeleton reorganization and often down-regulated in certain types of tumor cells. In response to platelet-derived growth factor (PDGF), green fluorescent protein (GFP)-tagged murine Mtss1 (Mtss1-GFP) underwent redistribution from the cytoplasm to dorsal membrane ruffles along with phosphorylation at tyrosine residues in a time-dependent manner. Tyrosine phosphorylation of Mtss1-GFP was also elevated in cells where an oncogenic Src was activated but severely impaired in Src knock-out cells or cells treated with Src kinase inhibitor PP2. Mutagenesis analysis has revealed that phosphorylation occurs at multiple sites, including tyrosine residues Tyr-397 and Tyr-398. Mutation at both Tyr-397 and Tyr-398 abolished the PDGF-mediated tyrosine phosphorylation. Furthermore, recombinant Mtss1 protein was phosphorylated by recombinant Src in a manner dependent on Tyr-397 and Tyr-398. Efficient tyrosine phosphorylation of Mtss1 in response to PDGF also involves a coiled-coil domain, which is essential for a proper distribution to the cell leading edge and dorsal ruffles. Interestingly, overexpression of wild type Mtss1-GFP promoted the PDGF-induced dorsal ruffling, whereas overexpression of a mutant deficient in phosphorylation at Tyr-397 and Tyr-398 or a mutant with deletion of the coiled-coil domain impaired the formation of dorsal ruffles. These data indicate that Mtss1 represents a novel signaling pathway from PDGF receptor to the actin cytoskeleton via Src-related kinases. Growth factor-mediated cytoskeletal reorganization plays a pivotal role in cell shape changes, locomotion, endocytosis, and homeostasis. The signal transduction elicited by ligand-occupied receptors initially involves activation of several small guanosine triphosphatase proteins, including Rac, Cdc42, and Rho (1Nobes C.D. Hall A. J. Cell Biol. 1999; 144: 1235-1244Crossref PubMed Scopus (1213) Google Scholar). Activated Cdc42 or Rac binds either directly or indirectly to Wiskott-Aldrich Syndrome protein (WASP) 2The abbreviations used are: WASP, Wiskott-Aldrich Syndrome protein; CCD, coiled-coil domain; GFP, green fluorescence protein; IMD, IRSp53 and Mtss1 homology domain; IRSp53, insulin receptor tyrosine kinase substrate p53; MIM/Mtss1, missing in metastasis protein; PDGF, platelet-derived growth factor; WAVE, WASP family verprolin homologous protein; WH2, WASP homology 2; PBS, phosphate-buffered saline; MEF, mouse embryonic fibroblast; GST, glutathione S-transferase; KD, kinase-dead. or WASP family verprolin homologous protein (WAVE/Scar), resulting in recruitment of these proteins to the plasma membrane wherein they subsequently activate Arp2/3 complex, an actin nucleation factor for the branched actin filaments that constitute lamellipodia and membrane ruffles (2Stradal T.E. Scita G. Curr. Opin. Cell Biol. 2006; 18: 4-10Crossref PubMed Scopus (207) Google Scholar). On the other hand, activation of Rho promotes the assembly of non-branched actin filaments through the function of the Formin family proteins, which may be responsible for the formation of stress fibers or actin cables (3Dong Y. Pruyne D. Bretscher A. J. Cell Biol. 2003; 161: 1081-1092Crossref PubMed Scopus (123) Google Scholar). In addition to these relatively well established signaling factors, there are a few actin-associated proteins the role of which in the cytoskeletal reorganization remains to be defined. One of these proteins is MIM, which stands for missing in metastasis, or MTSS1 as recently recommended by Human Genome Organization Gene Nomenclature Committee. Previous studies have reported the existence of multiple MIM splicing transcripts, including MIM-A, the prototype of MIM that encodes only 356 amino acids; MIM-B, which encodes a protein product of 759 amino acids; and MIM-C, which contains an alternative exon and predicts a protein of 734 amino acids (4Wang Y. Liu J. Smith E. Zhou K. Liao J. Yu G.Y. Zhan X. Cancer Investig. 2007; (in press)Google Scholar). However, analysis of a variety of human cells revealed only a dominant MIM-related immuno-reactivity running at a position close to recombinant MIM-B (4Wang Y. Liu J. Smith E. Zhou K. Liao J. Yu G.Y. Zhan X. Cancer Investig. 2007; (in press)Google Scholar), indicating that MIM-B likely represents the primary protein product of MIM. We hereby designate MTSS1 as the human gene that encodes a protein product with the same amino acid sequence as MIM-B, which bears a carboxyl-terminal WASP homology 2 (WH2) domain that is known to bind to monomeric actin (5Mattila P.K. Salminen M. Yamashiro T. Lappalainen P. J. Biol. Chem. 2003; 278: 8452-8459Abstract Full Text Full Text PDF PubMed Scopus (132) Google Scholar, 6Lee Y.G. Macoska J.A. Korenchuk S. Pienta K.J. Neoplasia. 2002; 4: 291-294Crossref PubMed Scopus (144) Google Scholar), a serine-rich domain (SRD), and a proline-rich domain (PRD) (Fig. 1A). The protein product of MTSS1 is structurally related to insulin receptor substrate protein 53 (IRSp53), a multidomain scaffolding protein that has been implicated in the formation of filopodia and lamellipodia (7Yamagishi A. Masuda M. Ohki T. Onishi H. Mochizuki N. J. Biol. Chem. 2004; 279: 14929-14936Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar, 8Miki H. Yamaguchi H. Suetsugu S. Takenawa T. Nature. 2000; 408: 732-735Crossref PubMed Scopus (461) Google Scholar, 9Krugmann S. Jordens I. Gevaert K. Driessens M. Vandekerckhove J. Hall A. Curr. Biol. 2001; 11: 1645-1655Abstract Full Text Full Text PDF PubMed Scopus (311) Google Scholar). The sequence similarity between MTSS1 and IRSp53 centers on the amino-terminal region of 250 amino acids, which is also called IRSp53 and MTSS1 homology domain (IMD) (Fig. 1A). Structural and biochemical analyses have demonstrated that IMD of IRSp53 is composed of four α helices that integrate each other and form an apparent anti-parallel homodimer (10Millard T.H. Bompard G. Heung M.Y. Dafforn T.R. Scott D.J. Machesky L.M. Futterer K. EMBO J. 2005; 24: 240-250Crossref PubMed Scopus (193) Google Scholar). Expression of MTSS1 has been found to be down-regulated in a subset of tumor cells derived from advanced prostate, breast, bladder, and gastric cancers (6Lee Y.G. Macoska J.A. Korenchuk S. Pienta K.J. Neoplasia. 2002; 4: 291-294Crossref PubMed Scopus (144) Google Scholar, 11Nixdorf S. Grimm M.O. Loberg R. Marreiros A. Russell P.J. Pienta K.J. Jackson P. Cancer Lett. 2004; 215: 209-220Crossref PubMed Scopus (53) Google Scholar, 12Loberg R.D. Neeley C.K. Adam-Day L.L. Ridman Y. St. John L.N. Nixdorf S. Jackson P. Kalikin L.M. Pienta K.J. Int. J. Oncol. 2005; 26: 1699-1705PubMed Google Scholar, 13Yamashita S. Tsujino Y. Moriguchi K. Tatematsu M. Ushijima T. Cancer Sci. 2006; 97: 64-71Crossref PubMed Scopus (220) Google Scholar). The mechanism for the down-regulation of MTSS1 may involve DNA methylation (13Yamashita S. Tsujino Y. Moriguchi K. Tatematsu M. Ushijima T. Cancer Sci. 2006; 97: 64-71Crossref PubMed Scopus (220) Google Scholar, 14Utikal J. Gratchev A. Muller-Molinet I. Oerther S. Kzhyshkowska J. Arens N. Grobholz R. Kannookadan S. Goerdt S. Int. J. Cancer. 2006; 119: 2287-2293Crossref PubMed Scopus (47) Google Scholar). Although the relevance of such down-regulation to tumor progression, in particular, metastasis, has not yet been confirmed, several lines of evidence have indicated a role of MTSS1 in cell morphogenesis. Overexpression of MTSS1 triggers distinct cell shape changes such as increase in the formation of lamellipodia, membrane ruffles, and filopodia-like structures (5Mattila P.K. Salminen M. Yamashiro T. Lappalainen P. J. Biol. Chem. 2003; 278: 8452-8459Abstract Full Text Full Text PDF PubMed Scopus (132) Google Scholar, 12Loberg R.D. Neeley C.K. Adam-Day L.L. Ridman Y. St. John L.N. Nixdorf S. Jackson P. Kalikin L.M. Pienta K.J. Int. J. Oncol. 2005; 26: 1699-1705PubMed Google Scholar, 15Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (95) Google Scholar, 16Bompard G. Sharp S.J. Freiss G. Machesky L.M. J. Cell Sci. 2005; 118: 5393-5403Crossref PubMed Scopus (98) Google Scholar, 17Gonzalez-Quevedo R. Shoffer M. Horng L. Oro A.E. J. Cell Biol. 2005; 168: 453-463Crossref PubMed Scopus (54) Google Scholar). More interestingly, MTSS1 has been found to be a target of Sonic hedgehog, a well characterized morphogen that plays a critical role in the patterning of embryonic and adult tissues (18Callahan C.A. Ofstad T. Horng L. Wang J.K. Zhen H.H. Coulombe P.A. Oro A.E. Genes Dev. 2004; 18: 2724-2729Crossref PubMed Scopus (140) Google Scholar). However, how MTSS1 induces these changes is not understood. MTSS1 has been proposed to be a scaffolding protein by linking F-actin to other intracellular proteins, including Rac1 (16Bompard G. Sharp S.J. Freiss G. Machesky L.M. J. Cell Sci. 2005; 118: 5393-5403Crossref PubMed Scopus (98) Google Scholar) and δ type receptor protein tyrosine phosphatases (15Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (95) Google Scholar, 17Gonzalez-Quevedo R. Shoffer M. Horng L. Oro A.E. J. Cell Biol. 2005; 168: 453-463Crossref PubMed Scopus (54) Google Scholar). On the other hand, the potential of MTSS1 to act as a signaling protein has also been evidenced. In skin epithelial cells, MTSS1 was found to associate with Gli, a major signaling target of the Sonic hedgehog pathway, and activate the transcriptional activity of Gli within cells (18Callahan C.A. Ofstad T. Horng L. Wang J.K. Zhen H.H. Coulombe P.A. Oro A.E. Genes Dev. 2004; 18: 2724-2729Crossref PubMed Scopus (140) Google Scholar). We have recently observed that Mtss1, a murine protein sharing 96.3% identity with MTSS1, binds to cortactin through the proline-rich domain and promotes the cortactin-mediated actin assembly (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). On the other hand, Mtss1 inhibits the actin polymerization mediated by a spontaneous process or by neuronal (N)-WASP-activated Arp2/3 complex (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). Furthermore, overexpression of Mtss1 caused a reduction in the motility of NIH3T3 cells in response to platelet-derived growth factor (PDGF) and of human endothelial cells in response to sphingosine 1 phosphate (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). Although these results have indicated that Mtss1 may modulate cell morphogenesis through regulation of the dynamics of the actin cytoskeleton in response to extracellular stimuli, the link between Mtss1 and growth factor receptor signaling pathways is unknown. In the present study we attempted to explore the role of Mtss1 in the PDGF signaling pathway and found that PDGF rapidly induces tyrosine phosphorylation of ectopically expressed Mtss1 and redistribution of Mtss1 to dorsal membrane ruffles and the cell leading edge. We also identified two major phosphotyrosine sites targeted by protein tyrosine kinase Src. Overexpression of a mutant at these two tyrosine residues apparently reduced dorsal membrane ruffling mediated by PDGF. Thus, our data suggest that Mtss1 is implicated in a growth factor receptor signaling pathway that regulates cell shape changes via protein tyrosine kinases. Chemicals and Reagents—All reagents unless otherwise indicated were purchased from Sigma. G418, GFP antibody, Pro-longed Antifade kit, and Pacific Blue-coupled anti-mouse IgG antibody were from Invitrogen. Protease inhibitor cocktails were from Roche Applied Science. Recombinant Src, cortactin monoclonal antibody (4F11), Myc monoclonal antibody (9E10), phospho-tyrosine monoclonal antibody (4G10), GST polyclonal antibody, and mouse IgG were obtained from Upstate Biotechnology (Lake Placid, NY). Mtss1 polyclonal antibody was prepared as described previously (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). Rhodamine or fluorescein isothiocyanate-labeled rabbit IgG and mouse IgG were from Pierce. Fetal bovine serum and bovine calf serum were from Hyclone (Logan, UT). Dulbecco's modified Eagle's medium was from Cambrex (Walkersville, MD). Cell Culture—NIH3T3 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum. Transfection was performed using SuperFect transfection reagent (Qiagen, Valencia, CA). Mouse mammary tumor cell line 4T07, kindly provided by Dr. Gary Sahagian (Tufts University School of Medicine), was cultured in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum. Src knock-out fibroblasts (Src-/-, clone 1) and normal mouse embryonic fibroblasts (MEF, clone 8) were a gift from Dr. Phillippe Soriano (Fred Hutchinson Cancer Research Center). Embryonic fibroblasts from Src, Yes, and Fyn triple knock-out mice were obtained from American Type Culture Collection (Manassas, VA), and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Cells expressing temperature-sensitive v-Src were established by infecting 4T07 cells with ts-v-Src retrovirus followed by selection in a medium containing 2 μg/μl G418. Plasmid ts-v-Src in fpGV29 vector was a gift from Dr. Margaret Frame (Beatson Institute for Cancer Research). The procedures for virus preparation and infection have been described previously (20Li Y. Liu J. Zhan X. J. Biol. Chem. 2000; 275: 37187-37193Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar). The established cells were maintained at 39.5 °C. Plasmid, Virus Preparation, and Recombinant Protein Preparation—Kinase-dead (KD) and wild type Fgr constructs were a kind gift from Dr. Giorgio Berton (University of Verona, Italy). Mutants with deletion of CCD and Y397F/Y398F were prepared with a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) using plasmid pEGFP-C1-MIM as the template. The primers used for construction of these mutants are the following. For Y397F/Y398F, 5′-CCTTCCTGACTACGCTCATTTTTTCACCATTGGGCCCGGC-3′ and 5′-GCCGGGCCCAATGGTGAAAAAATGAGCGTAGTCAGGAAGG-3′, and for CCD deletion mutant, 5′-GATTGCCTGATAAACCCAGTGGACGCTCAGGGGAG-3′ and 5′-CTCCCCTGAGCGTCCACTGGGTTTATCAGGCAATC-3′. Mutants with either single or multiple mutations at other tyrosine residues were prepared in a similar manner. All the retroviral plasmids were prepared by insertion of the corresponding DNA fragments into the SacII and NotI sites of retroviral vector MGIN, and the mutations of the inserts were confirmed by DNA sequencing. To construct GST-tagged Mtss1 mutants containing amino acids 382-420, pEGFP-C1-MIM was amplified by PCR using the following primers, 5′-GCGGATCCCTGCTCCCTCGGGTCAC-3′ and 5′-GCGAATTCTCATCCTGGCTTTGCCCAGTC-3′. The amplified DNA fragment was inserted into BamHI and EcoRI sites of pGEX4T-2. Immunoprecipitation and Immunoblotting—Cells were extracted in lysis buffer containing 50 mm Tris-HCl, pH 7.4, 150 mm NaCl, 1 mm EDTA, 0.2 mm Na3VO4, 1% Triton X-100, and Roche protease inhibitor cocktails. The extracts were centrifuged at 14,000× rpm for 10 min at 4 °C. The clarified supernatants were incubated with appropriate rabbit polyclonal antibodies for 2 h in the presence of protein A-Sepharose. The precipitates were separated by SDS-PAGE, and the proteins on the gel were transferred to a nitrocellulose membrane, which was then blotted with appropriate antibodies. Src in Vitro Kinase Assay—Purified recombinant Mtss1 or its derivatives (3 μg) were incubated at 30 °C for 30 min in 20 μl of kinase reaction buffer (100 mm Tris-HCl, pH 7.4, 125 mm MgCl2, 25 mm MnCl2, 2 mm EGTA, 0.2 mm Na3VO4, 0.25 mm ATP, and 2 mm dithiothreitol) in the presence of 10 units of recombinant Src. The kinase reaction was terminated by adding equal volume of 2× SDS sample buffer. The sample was further analyzed by phosphotyrosine immunoblot using 4G10 antibody. Immunofluorescence—Cells were seeded onto coverslips precoated with fibronectin in a 6-well plate. The cells on coverslips were washed twice with cold PBS and fixed with 3% paraformaldehyde in PBS for 30 min, permeabilized in PBS containing 0.5% Triton X-100 for 5 min, and incubated in blocking buffer (PBS plus 5% goat serum) for 1 h, followed by incubation in blocking buffer containing primary antibody for 1 h. The cell samples were washed three times with PBS and incubated for 1 h in blocking buffer containing fluorescein isothiocyanate-coupled anti-rabbit antibody and Pacific Blue-coupled anti-mouse antibody. After three times of wash, the samples were mounted onto glass slides using Prolong Anti-fade kit and subsequently inspected under a fluorescent microscope (Nikon TE2000U) using a ×60 oil lens. Cell images were captured by digital camera Nikon 1200F and further processed by Adobe Photoshop. Protein Profile Scan—Murine Mtss1 amino acid sequence was examined for the presence of structural motifs using the Expert Protein Analysis System (ExPASy) of the Swiss Institute of Bioinformatics. PDGF Induces Redistribution and Tyrosine Phosphorylation of Mtss1-GFP—To study a potential role of Mtss1 in growth factor signal transductions, we analyzed NIH3T3 cells expressing murine Mtss1 that had been fused with green fluorescent protein (GFP) at its C terminus (Mtss1-GFP). Expression of Mtss1-GFP was driven by retroviral vector MGIN (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). Immunofluorescent analysis revealed a diffused staining of Mtss1-GFP in the cytoplasm when cells were grown in a low percent serum-containing medium (Fig. 1B, a). Upon stimulation with PDGF, Mtss1-GFP was translocated within 5 min into a membrane wave-like structure wherein cortactin (blue) and F-actin (red) were also enriched (Fig. 1B, d-f). A similar cortactin-enriched wave structure has also been previously described in PDGF-treated NIH3T3 cells (21Krueger E.W. Orth J.D. Cao H. McNiven M.A. Mol. Biol. Cell. 2003; 14: 1085-1096Crossref PubMed Scopus (181) Google Scholar). Time-lapse image analysis showed that these waves were highly dynamic (data not shown), indicating that they likely represent a type of membrane dynamics that has been previously described as dorsal ruffles where actin polymerization is intimately implicated (22Mellstrom K. Heldin C.H. Westermark B. Exp. Cell Res. 1988; 177: 347-359Crossref PubMed Scopus (132) Google Scholar). In addition to these dorsal ruffles, Mtss1-GFP was also found in the cell leading edge that outlines the newly formed lamellipodia (Fig. 1B, h). This finding is consistent with the role of Mtss1 in actin polymerization as previously reported (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). However, either in quiescent or PDGF-treated cells Mtss1-GFP did not apparently associate with stress fibers or filopodia, indicating that Mtss1 preferably binds to the cortactin-associated actin network. To study in detail about the role of Mtss1 in the PDGF signaling pathway, we examined the possibility that Mtss1 may be subjected to tyrosine phosphorylation in response to PDGF, as many phosphotyrosyl proteins are known to engage in the signal transduction of PDGF receptors (23Claesson-Welsh L. J. Biol. Chem. 1994; 269: 32023-32026Abstract Full Text PDF PubMed Google Scholar). Thus, phosphotyrosine immunoblot was carried out to analyze Mtss1-GFP in NIH3T3 cells before and after PDGF treatment. As shown in Fig. 1C, Mtss1-GFP underwent a rapid phosphorylation upon PDGF treatment, which reached to a peak within 1 min but started to decline at 10 min and retreated to the basal level after 1 h. These data indicate that tyrosine phosphorylation of Mtss1 represents an early event in the PDGF signaling cascade. Determination of Phosphotyrosine Residues of Mtss1—The predicted full-length Mtss1 protein contains a total of 15 tyrosine residues. To define specific residues responsible for the PDGF-induced phosphorylation, we prepared a series of GFP-tagged Mtss1 mutants in which individual tyrosine residues were substituted with phenylalanine and transfected each mutant into NIH3T3 cells. Transient transfectants were subjected to tyrosine phosphorylation analysis by immunoblotting. As shown in Fig. 2A, mutations at individual residues of Tyr-177, Tyr-256, Tyr-260, Tyr-394, Tyr-422, Tyr-509, Tyr-532, and Tyr-562 caused only little or minor effect on the PDGF-mediated phosphorylation. However, mutants Y397F and Y398F showed a marked decrease in the phosphorylation response to PDGF by more than 60% as compared with the wild type Mtss1-GFP. Furthermore, double mutations at Tyr-397 and Tyr-398 nearly abolished the phosphorylation (Fig. 2B), indicating that these two residues play a critical role in the PDGF-mediated phosphorylation. In contrast, mutants carrying either double mutations of Y256F/Y260F or sextuple mutations of Y177F/Y256F/Y260F/Y394F/Y422F/Y532F (Mtss1-YF6) were still effectively phosphorylated at a level similar to that of the wild type (Fig. 2B). To analyze the role of the remaining tyrosine residues, we introduced an additional mutation at each of these residues as well as Tyr-397 and Tyr-398 to the mutant Mtss1-YF6. As shown in Fig. 2C, addition of Y27F, Y130F, Y303F, and Y305F to Mtss1-YF6, respectively, resulted in little change in the PDGF-induced phosphorylation. In contrast, adding Y397F or Y398F or Y530F impaired significantly the phosphorylation of Mtss1-YF6. However, PDGF was able to induce phosphorylation of a mutant carrying Tyr-530 alone (data not shown), indicating that Tyr-530, unlike Tyr-397 and Tyr-98, may only play a role in the phosphorylation in certain contexts such as mutant YF6. Tyrosine Phosphorylation of Mtss1 Is Mainly Mediated by Src—Protein tyrosine kinase Src is known to play an important role in PDGF signal transduction as well as cytoskeletal reorganization (24Twamley-Stain G.M. Pepperkok R. Ansorge W. Courtneidge S.A. Proc. Natl. Acad. Sci. USA. 1993; 90: 7696-7700Crossref PubMed Scopus (296) Google Scholar, 25Klinghoffer R.A. Sachsenmaier C. Cooper J.A. Soriano P. EMBO J. 1999; 18: 2459-2471Crossref PubMed Scopus (647) Google Scholar, 26Wang D. Tan Y.C. Kreitzer G.E. Nakai Y. Shan D. Zheng Y. Huang X.Y. J. Biol. Chem. 2006; 281: 32660-32667Abstract Full Text Full Text PDF PubMed Scopus (22) Google Scholar). To evaluate the role of Src in Mtss1 phosphorylation, NIH3T3 cells expressing Mtss1-GFP were exposed to a selective Src inhibitor, PP2, prior to PDGF treatment. As shown in Fig. 3A, PP2 treatment completely abolished Mtss1 phosphorylation. In contrast, when cells were treated with emodin, a less selective tyrosine kinase inhibitor, only partial inhibition was observed (Fig. 3A). Tyrosine phosphorylation of Mtss1-GFP was further investigated in embryonic fibroblasts derived from Src knock-out mice (Src-/-) and those derived from normal MEFs. As shown in Fig. 3B, the level of tyrosine phosphorylation of Mtss1-GFP in Src-/- cells was significantly less than that of MEF. The remaining phosphorylation detected in Src-/- could be caused by the activity of other Src family members, which include Yes, Fyn, Lyn, Fgr, Hck, Blk and Lck. Thus, we also analyzed embryonic fibroblasts derived from triple knock-out mice at Src, Yes, and Fyn (SYF) and NIH3T3 cells expressing a KD Fgr mutant. Whereas a trace amount of phosphorylated Mtss1-GFP was still observed in PDGF-treated SYF cells (Fig. 3C), markedly reduced phosphorylation was evident in Fgr-KD cells (Fig. 3D). Furthermore, the level of Mtss1-GFP phosphorylation was significantly elevated in cells overexpressing a wild type Fgr (Fgr-WT) even in the absence of PDGF (Fig. 3D), indicating that the activity of Fgr or other closely related kinases contributes to Mtss1-GFP phosphorylation as well. We also examined the role of Syk, a non-Src-related intracellular tyrosine kinase. However, expression of either wild type Syk or a Syk-KD mutant failed to show any significant effect on the phosphorylation of Mtss1-GFP (data not shown), indicating that Src family kinases play a primary role in the phosphorylation of Mtss1 mediated by PDGF. To determine whether Mtss1 acts as a direct substrate of Src, we prepared recombinant Mtss1 proteins tagged by six histidine residues (His6) at the carboxyl terminus (Mtss1-His). Purified Mtss1-His, which was detected as multiple protein species on SDS-PAGE as a result of partial degradation (Fig. 4A), was incubated with recombinant Src kinase at 30 °C in the presence of ATP. To ensure an optimal condition used for Src, recombinant cortactin, a well characterized prominent substrate of Src (27Wu H. Parsons J.T. J. Cell Biol. 1993; 120: 1417-1426Crossref PubMed Scopus (452) Google Scholar), was also analyzed in parallel. Under this condition, both cortactin and Mtss1-His were readily phosphorylated as determined by phosphotyrosine immunoblot. Under the same condition, only a very faint band corresponding to GST was noticed in the phosphotyrosine immunoblot. Because the amount of GST used in this reaction was significantly higher than the rest of the samples (Fig. 4A), such a faint band likely represented a nonspecific reactivity of the phosphotyrosine antibody. To determine whether Src targets at Tyr-397 and Tyr-398, we also prepared a GST-tagged Mtss1 fragment corresponding to amino acids 382-420 (GST-Mtss1382-420) and its derivatives, GST-Mtss1382-420Y394F, GST-Mtss1382-420Y397F, and GST-Mtss1382-420Y398F. Although GST-Mtss1382-420 was readily phosphorylated by Src (Fig. 4B), significantly lower levels of phosphorylation were observed with mutants GST-Mtss1382-420Y397F and GST-Mtss1382-420Y398F, confirming that Tyr-397 and Tyr-398 are the direct targets of Src. A slightly reduced level of phosphorylation was also found with mutant GST-Mtss1382-420Y394F, suggesting a minor role of Tyr-394 in the phosphorylation of Mtss1 in vitro. To examine whether Src could trigger phosphorylation of endogenous Mtss1, we used murine mammary tumor 4T07 cells, in which Mtss1 is highly expressed (4Wang Y. Liu J. Smith E. Zhou K. Liao J. Yu G.Y. Zhan X. Cancer Investig. 2007; (in press)Google Scholar). 4T07 cells were infected with a retrovirus encoding ts-v-Src, a temperature-sensitive form of Src oncogene (28Catling A.D. Fincham V.J. Frame M.C. Haefner B. Wyke J.A. J. Virol. 1994; 68: 4392-4399Crossref PubMed Google Scholar). Tyrosine phosphorylation of endogenous Mtss1 was analyzed by immunoprecipitation with Mtss1 followed by phosphotyrosine immunoblot. Consistent with our previous report and others (16Bompard G. Sharp S.J. Freiss G. Machesky L.M. J. Cell Sci. 2005; 118: 5393-5403Crossref PubMed Scopus (98) Google Scholar, 17Gonzalez-Quevedo R. Shoffer M. Horng L. Oro A.E. J. Cell Biol. 2005; 168: 453-463Crossref PubMed Scopus (54) Google Scholar, 19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar), endogenous Mtss1 is resolved as two bands, with the major one at a position near 115 kDa (Fig. 4C, right panel). At the permissive temperature 35 °C, phosphorylation of Mtss1 corresponding to 115 kDa was readily detected. On the other hand, no phosphorylation was observed at the restrictive temperature 39.5 °C (left panel), a condition that had no apparent effect on the expression of Mtss1 (right panel). Thus, these data demonstrated that Mtss1, either as an exogenous or endogenous form, functions as an appropriate substrate of Src. The Coiled-coil Domain Is Required for PDGF-mediated Phosphorylation—Computer analysis has revealed that Mtss1 contains a coiled-coil domain (CCD) from amino acids 108 to 153 (Fig. 1A), which corresponds to the second α-helix of IRSp53's IMD within the interface of homodimer as indicated by a recent crystallography study (10Millard T.H. Bompard G. Heung M.Y. Dafforn T.R. Scott D.J. Machesky L.M. Futterer K. EMBO J. 2005; 24: 240-250Crossref PubMed Scopus (193) Google Scholar). To analyze whether CCD plays any functional role, we prepared Mtss1ΔCCD-GFP and analyzed the distribution of the mutant in NIH3T3 cells upon PDGF treatment. Unlike wild type Mtss1-GFP, the mutant was diffused within the cytoplasm without evident translocation into membrane ruffles or the cell leading edge after PDGF treatment (Fig. 5A). Interestingly, the mutant was also poorly phosphorylated in response to PDGF (Fig. 5B), indicating that CCD is requisite for both tyrosine phosphorylation and distribution of Mtss1 to certain cellular areas where actin polymerization actively takes place. Mtss1 Promotes PDGF-induced Dorsal Membrane Ruffling in a Tyrosine Phosphorylation-dependent Manner—To explore the physiological significance of Mtss1 phosphorylation, we examined NIH3T3 cells infected with retrovirus encoding Mtss1Y397F/Y398F-GFP. The mutant-expressing cells did not show a significant difference in morphology as compared with the wild type. Furthermore, the mutant retained the ability to translocate into the punctate structures in the cytoplasm and the cell leading edge upon PDGF stimulation (Fig. 5A). However, these cells did not establish dorsal membrane ruffles in response to PDGF as efficiently as the cells expressing Mtss1-GFP (Fig. 6A, e and f). To quantify dorsal ruffling, we measured the ruffling area of each cell at 5 min after stimulation, the time at which extensive dorsal ruffling activity was observed. This analysis indicated that dorsal ruffles on the control cells expressing GFP only had an average ruffling area of ∼24 μm2, whereas the cells expressing Mtss1-GFP had an average ruffle area of ∼50 μm2 (Fig. 6B). Further exposure to PDGF for 15 min did not significantly increase ruffle formation (Fig. 6B). In contrast, the average ruffling area on the cells expressing Mtss1Y397F/Y398F-GFP was ∼15 μm2 at 5 min, which is statistically lower than that of the cells expressing Mtss1-GFP or the control cells expressing GFP only (Fig. 6B). At 15 min, however, the average ruffling area reached to 22 μm2, which was similar to that of control cells but still significantly lower than that of the cells expressing wild type Mtss1. A similar conclusion was obtained when the dorsal ruffling was measured based on the percentage of cells that had developed significant dorsal ruffles (data not shown). We also examined dorsal ruffle formation in cells expressing Mtss1ΔCCD-GFP. These cells poorly responded to PDGF for morphological changes and developed dorsal ruffles with an average area of only 17 μm2 at 5 min and 20 μm2 at 15 min (Fig. 6A, g and h, and Fig. 6B). Therefore, Mtss1-mediated dorsal ruffling requires both tyrosine phosphorylation and the function of the CCD motif. Human MTSS1 gene has recently drawn much attention as it encodes an actin-associated protein that is frequently down-regulated in a variety of cancers. However, it is unclear whether MTSS1 acts as only a scaffolding protein or a signaling molecule for the control of cell growth and morphogenesis. In the present study, we have examined the role of murine Mtss1 in PDGF signal transduction and found that this growth factor stimulated Mtss1-GFP tyrosine phosphorylation. Several lines of evidence indicate that Mtss1 is a direct substrate of Src, a kinase that is known to be activated upon PDGF stimulation (29Kypta R.M. Goldberg Y. Ulug E.T. Courtneidge S.A. Cell. 1990; 62: 481-492Abstract Full Text PDF PubMed Scopus (480) Google Scholar). These include the inhibition of Mtss1 phosphorylation by Src inhibitor PP2 (Fig. 3A), the ability of recombinant Src to phosphorylate recombinant Mtss1 directly in vitro (Fig. 4A), and poor phosphorylation in Src knock-out cells (Fig. 3B). In addition, Mtss1-GFP phosphorylation was elevated in NIH3T3 cells overexpressing either oncogenic Src (Fig. 4C) or Fgr (Fig. 3D), a Src-related kinase. However, tyrosine phosphorylation of Mtss1-GFP was not abolished in either Src knock-out cells or cells with depletion of Src, Yes, and Fyn (Fig. 4, B and C), which are the most abundant Src family members in fibroblasts (25Klinghoffer R.A. Sachsenmaier C. Cooper J.A. Soriano P. EMBO J. 1999; 18: 2459-2471Crossref PubMed Scopus (647) Google Scholar). This suggests that other tyrosine kinases may also be involved in Mtss1 phosphorylation in cells. Consistent with the notion, we have also observed that phosphorylation of Mtss1 was significantly elevated in cells co-expressing Fyn and Syk (data not shown), suggesting that Syk or its related kinases could synergize Src kinases in phosphorylation of Mtss1. It is also possible that phosphorylation of Mtss1 is regulated by other mechanisms. It has been reported that Mtss1 interacts with protein tyrosine phosphatase receptor δ (15Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (95) Google Scholar). Interestingly, the binding motif for receptor protein tyrosine phosphatase δ is located within the central region downstream of IMD and upstream of WH2 domain (15Woodings J.A. Sharp S.J. Machesky L.M. Biochem. J. 2003; 371: 463-471Crossref PubMed Scopus (95) Google Scholar), the area that contains Tyr-397 and Tyr-398, two major phosphotyrosine residues. Therefore, it is possible that receptor protein tyrosine phosphatase δ could play a negative role in the regulation of Mtss1 during serum starvation and such inhibitory activity might be relieved upon PDGF stimulation. Characterization of the effect of PDGF on the interaction between Mtss1 and receptor protein tyrosine phosphatase δ would ultimately specify the position of receptor protein tyrosine phosphatase δ in the Mtss1 signaling pathway. The physiological significance of tyrosine phosphorylation of Mtss1 has been indicated by the observation that a mutant deficient in tyrosine phosphorylation was unable to promote efficiently the PDGF-mediated dorsal ruffling. However, the mechanism by which tyrosine phosphorylation of Mtss1 regulates dorsal ruffling is presently unclear. We have previously reported that Mtss1 binds to cortactin, a major component of the cortical actin network that is intimately involved in Arp2/3 complex-mediated actin polymerization (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). Association with cortactin becomes more evident in the dorsal membrane ruffles of NIH3T3 cells induced by PDGF (Fig. 1B). At this time we are unable to determine whether ruffle-associated Mtss1 represents exclusively the phosphorylated form as there are no available antibodies that specifically detect phosphorylated Mtss1. Also, phosphorylation-deficient mutant appears to bind to cortactin similarly as the wild type Mtss1 (data not shown), suggesting that tyrosine phosphorylation is not indispensable for cortactin binding. Nevertheless, tyrosine phosphorylation could regulate other functions of Mtss1. For example, our previous study has shown that Mtss1 inhibits actin polymerization mediated by neuronal-WASP-activated Arp2/3 complex (19Lin J. Liu J. Wang Y. Zhu J. Zhou K. Smith N. Zhan X. Oncogene. 2005; 24: 2059-2066Crossref PubMed Scopus (80) Google Scholar). Such inhibitory activity may be circumvented upon phosphorylation of Mtss1 during PDGF stimulation, resulting in increase in actin polymerization and membrane rufflings. It is interesting to note that a mutant with deletion of the CCD motif failed to be phosphorylated upon PDGF treatment (Fig. 5B). The CCD motif contains a putative F-actin binding domain and presumably locates in the interface of the homodimer, based on a recently established structural model for IRSp53 (17Gonzalez-Quevedo R. Shoffer M. Horng L. Oro A.E. J. Cell Biol. 2005; 168: 453-463Crossref PubMed Scopus (54) Google Scholar), an IMD-containing protein. A previous study with an amino-terminal-truncated mutant suggested that CCD may be involved in homodimerization as well as F-actin binding (17Gonzalez-Quevedo R. Shoffer M. Horng L. Oro A.E. J. Cell Biol. 2005; 168: 453-463Crossref PubMed Scopus (54) Google Scholar). However, we have not been able to observe any significant change of the CCD deletion mutant in the ability to form a dimer with either wild type or the mutant Mtss1 (data not shown), indicating that the CCD motif may have a function other than dimerization. Indeed, we have found that the mutant was unable to translocate into the cell leading edge in response to PDGF (Fig. 5A). Because the formation of the leading edge is dependent upon actin polymerization, this result suggests that tyrosine phosphorylation of Mtss1 may require translocation to a proper cellular compartment wherein actin polymerization actively takes place. Interestingly, overexpressing either the CCD deletion mutant or tyrosine phosphorylation-deficient mutant has shown a certain degree of inhibition of dorsal rufflings (Fig. 6B), suggesting these mutants may have a dominant negative activity. However, expression of Mtss1 in NIH3T3 cells is extremely low (4Wang Y. Liu J. Smith E. Zhou K. Liao J. Yu G.Y. Zhan X. Cancer Investig. 2007; (in press)Google Scholar), raising a possibility that these mutants may interfere with the function of other cellular proteins. Because both mutants maintain the ability to form a dimer, these mutants could bind to other IMD-containing proteins through a similar mechanism for dimerization. We thank Drs. Jeff Winkle and Dan Yu for critical review of the manuscript.
Referência(s)