Portal de Periódicos da CAPES

Activation of the Abl Tyrosine Kinase in Vivo by Src Homology 3 Domains from the Src Homology 2/Src Homology 3 Adaptor Nck

1999; Elsevier BV; Volume: 274; Issue: 39 Linguagem: Inglês

10.1074/jbc.274.39.27956

1083-351X

Jodi M. Smith, Steve Katz, Bruce J. Mayer,

Protein Kinase Regulation and GTPase Signaling

The nonreceptor tyrosine kinase c-Abl is tightly regulated in vivo, but the mechanisms that normally repress its activity are not well understood. We find that a construct encoding the first two Src homology 3 (SH3) domains of the Src homology 2/SH3 adaptor protein Nck can activate c-Abl in human 293T cells. A myristoylated Nck SH3 domain construct, which is expected to localize to membranes, potently activated Abl when expressed at low levels. An unmyristoylated Nck SH3 domain construct, which localizes to the cytosol and nucleus, also activated Abl but only at high levels of expression. Activation by both myristoylated and unmyristoylated Nck constructs required the C terminus of Abl; a C-terminally truncated form of Abl was not activated, although this construct could still be activated by deletion of its SH3 domain. Activation did not require the major binding sites in the Abl C terminus for Nck SH3 domains, however, suggesting that the mechanism of activation does not require direct binding to the C terminus. Activation of c-Abl by Nck SH3 domains provides a robust experimental system for analyzing the mechanisms that normally repress Abl activity and how that normal regulation can be perturbed. The nonreceptor tyrosine kinase c-Abl is tightly regulated in vivo, but the mechanisms that normally repress its activity are not well understood. We find that a construct encoding the first two Src homology 3 (SH3) domains of the Src homology 2/SH3 adaptor protein Nck can activate c-Abl in human 293T cells. A myristoylated Nck SH3 domain construct, which is expected to localize to membranes, potently activated Abl when expressed at low levels. An unmyristoylated Nck SH3 domain construct, which localizes to the cytosol and nucleus, also activated Abl but only at high levels of expression. Activation by both myristoylated and unmyristoylated Nck constructs required the C terminus of Abl; a C-terminally truncated form of Abl was not activated, although this construct could still be activated by deletion of its SH3 domain. Activation did not require the major binding sites in the Abl C terminus for Nck SH3 domains, however, suggesting that the mechanism of activation does not require direct binding to the C terminus. Activation of c-Abl by Nck SH3 domains provides a robust experimental system for analyzing the mechanisms that normally repress Abl activity and how that normal regulation can be perturbed. The nonreceptor tyrosine kinase c-Abl has been extensively studied, yet its regulation and mechanism of activation are poorly understood. Originally identified as the transforming gene of Abelson murine leukemia virus (1Goff S.P. Gilboa E. Witte O.N. Baltimore D. Cell. 1980; 22: 777-785Abstract Full Text PDF PubMed Scopus (280) Google Scholar), Abl has also been implicated in human leukemia via a specific chromosomal translocation that generates a fusion protein containing c-Abl and N-terminal sequences derived from a second locus, Bcr (2Shtivelman E. Lifshitz B. Gale R.P. Canaani E. Nature. 1985; 315: 550-554Crossref PubMed Scopus (1270) Google Scholar, 3Mes-Masson A.-M. McLaughlin J. Daley G.Q. Paskind M. Witte O.N. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 9768-9772Crossref PubMed Scopus (89) Google Scholar). Overexpressed c-Abl is catalytically inactive, nontransforming, localized predominantly in the nucleus, and lacks detectable tyrosine phosphorylation; on the contrary, various mutant forms of Abl have increased kinase activity, can transform fibroblasts and hematopoietic cells, are highly tyrosine-phosphorylated, and are localized predominantly in the cytosol and on membranes (for review see Ref. 4Raitano A.B. Whang Y.E. Sawyers C.L. Biochim. Biophys. Acta. 1997; 1333: F201-F216PubMed Google Scholar). This dramatic difference in biological activity suggests that c-Abl is tightly regulated in vivo. The nonreceptor tyrosine kinases whose regulation is best understood are the members of the Src family. Src and its relatives are regulated by a C-terminal tyrosine residue that, when phosphorylated, can bind in cis to the SH2 1The abbreviations used are:SH2Src homology domain 2SH3Src homology domain 3WTwild typeHAhemagglutinin domain of the kinase, effectively locking it in an inactive form (for review see Ref.5Cooper J.A. Howell B. Cell. 1993; 73: 1051-1054Abstract Full Text PDF PubMed Scopus (495) Google Scholar). This inactive or “closed” conformation is stabilized by interaction of the SH3 domain of Src with a linker region between its SH2 and kinase domains (6Xu W. Harrison S.C. Eck M.J. Nature. 1997; 385: 595-602Crossref PubMed Scopus (1252) Google Scholar, 7Sicheri F. Moarefi I. Kuriyan J. Nature. 1997; 385: 602-609Crossref PubMed Scopus (1047) Google Scholar). When the regulatory tyrosine is dephosphorylated, the closed conformation is unstable, the kinase domain is unleashed, and Src exhibits full kinase activity. Binding of a high-affinity ligand to the Src SH3 domain can also destabilize the closed conformation by displacing the intramolecular SH3-linker interaction (8Moarefi I. LeFevre-Bernt M. Sicheri F. Huse M. Lee C.-H. Kuriyan J. Miller W.T. Nature. 1997; 385: 650-653Crossref PubMed Scopus (541) Google Scholar). Src homology domain 2 Src homology domain 3 wild type hemagglutinin Because c-Abl closely resembles c-Src in domain structure (with the exception of the long C-terminal extension that defines the Abl family of kinases) it is reasonable to suppose that it is regulated in a similar fashion. However, several pieces of evidence suggest that this is not the case. Under most conditions c-Abl is not detectably phosphorylated on tyrosine, inconsistent with an intramolecular SH2-phosphotyrosine interaction (9Pendergast A.M. Muller A.J. Havlik M.H. Clark R. McCormick F. Witte O.N. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 5927-5931Crossref PubMed Scopus (123) Google Scholar, 10Ponticelli A.S. Whitlock C.A. Rosenberg N. Witte O.N. Cell. 1982; 29: 953-960Abstract Full Text PDF PubMed Scopus (50) Google Scholar, 11Franz W.M. Berger P. Wang J.Y.J. EMBO J. 1989; 8: 137-147Crossref PubMed Scopus (161) Google Scholar). Instead, the SH3 domain of c-Abl appears to play a more critical role in regulation, because mutations in the SH3 domain potently activate the transforming activity of c-Abl (11Franz W.M. Berger P. Wang J.Y.J. EMBO J. 1989; 8: 137-147Crossref PubMed Scopus (161) Google Scholar, 12Jackson P. Baltimore D. EMBO J. 1989; 8: 449-456Crossref PubMed Scopus (222) Google Scholar, 13Van Etten R.A. Debnath J. Zhou H. Casasnovas J.M. Oncogene. 1995; 10: 1977-1988PubMed Google Scholar). A recent study (14Barilá D. Superti-Furga G. Nat. Genet. 1998; 18: 280-282Crossref PubMed Scopus (184) Google Scholar) suggests that the negative regulation of Abl might be dominated by an intramolecular SH3-linker interaction similar to that seen in the Src structure. However, purified WT c-Abl kinase prepared from baculovirus has identical specific activity to an SH3-deleted, activated Abl mutant, inconsistent with intramolecular interactions being entirely responsible for negative regulation (15Mayer B.J. Baltimore D. Mol. Cell. Biol. 1994; 14: 2883-2894Crossref PubMed Scopus (158) Google Scholar). 2B. J. Mayer, unpublished results. It has been proposed that Abl is regulated at least in part by a titratable SH3-binding cellular inhibitor (9Pendergast A.M. Muller A.J. Havlik M.H. Clark R. McCormick F. Witte O.N. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 5927-5931Crossref PubMed Scopus (123) Google Scholar), and several candidate inhibitors have been identified (16Zhu J. Shore S.K. Mol. Cell. Biol. 1996; 16: 7054-7062Crossref PubMed Scopus (37) Google Scholar, 17Wen S.-T. Van Etten R.A. Genes Dev. 1997; 11: 2456-2467Crossref PubMed Scopus (239) Google Scholar, 18Dai Z. Pendergast A.M. Genes Dev. 1995; 9: 2569-2582Crossref PubMed Scopus (241) Google Scholar, 19Shi Y. Alin K. Goff S.P. Genes Dev. 1995; 9: 2583-2597Crossref PubMed Scopus (218) Google Scholar). The activity of endogenous c-Abl has been shown to be regulated by several different conditions and factors, most notably DNA damage (20Liu Z.-G. Baskaran R. Lea-Chou E.T. Wood L.D. Chen Y. Karin M. Wang J.Y.J. Nature. 1996; 384: 273-276Crossref PubMed Scopus (347) Google Scholar, 21Baskaran R. Wood L.D. Whitaker L.L. Canman C.E. Morgan S.E. Xu Y. Barlow C. Baltimore D. Wynshaw-Boris A. Kastan M.B. Wang J.Y.J. Nature. 1997; 287: 516-519Crossref Scopus (485) Google Scholar, 22Kharbanda S. Ren R. Pandey P. Shafman T.D. Feller S.M. Weichselbaum R.R. Kufe D.W. Nature. 1995; 376: 785-788Crossref PubMed Scopus (463) Google Scholar, 23Kharbanda S. Pandey P. Jin S. Inoue S. Bharti A. Yuan Z.M. Weichselbaum R. Weaver D. Kufe D. Nature. 1997; 386: 732-735Crossref PubMed Scopus (238) Google Scholar), integrin-mediated adhesion (24Lewis J.M. Baskaran R. Taagepera S. Schwartz M.A. Wang J.Y.J. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 15174-15179Crossref PubMed Scopus (269) Google Scholar, 25Lewis J.M. Schwartz M. J. Biol. Chem. 1998; 273: 14225-14230Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar), and the cell cycle (26Welch P.J. Wang J.Y. Cell. 1993; 75: 779-790Abstract Full Text PDF PubMed Scopus (368) Google Scholar,27Welch P.J. Wang J.Y. Mol. Cell. Biol. 1995; 15: 5542-5551Crossref PubMed Scopus (77) Google Scholar). Although these studies have shed some light on the potentialin vivo role of c-Abl, the low expression level of endogenous Abl and modest levels of activation (typically 2–4-fold) have made it difficult to study the mechanism of regulation. Toward this end, a robust system in which the activation of overexpressed c-Abl and its mutants can be experimentally manipulated would be extremely useful. Among the large number of proteins and factors that have been shown to interact with c-Abl are the SH2/SH3 adaptor proteins Grb2, Crk, and Nck, which bind via their SH3 domains to proline-rich sites in the Abl C terminus (28Feller S.M. Knudsen B. Hanafusa H. EMBO J. 1994; 13: 2341-2351Crossref PubMed Scopus (326) Google Scholar, 29Ren R. Ye Z.-S. Baltimore D. Genes Dev. 1994; 8: 783-795Crossref PubMed Scopus (291) Google Scholar). The role of these adaptors is not firmly established, but binding of the Crk adaptor has been shown to affect the processivity of Abl toward some substrates, presumably by forming a tight complex via its SH2 domain with those substrates when they are phosphorylated (30Mayer B.J. Hirai H. Sakai R. Curr. Biol. 1995; 5: 296-305Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar). The Nck adaptor consists of three SH3 domains and one SH2 domain (31Lehman J.M. Riethmuller G. Johnson J.P. Nucleic Acids Res. 1990; 18: 1048Crossref PubMed Scopus (160) Google Scholar). Its SH2 domain binds a variety of phosphorylated proteins such as receptor tyrosine kinases, IRS-1, and focal adhesion components (32Lee C.-H. Li W. Nishimura R. Zhou M. Batzer A.G. Myers J. White M.F. Schlessinger J. Skolnik E.Y. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 11713-11717Crossref PubMed Scopus (198) Google Scholar, 33Li W. Hu E. Skolnik E.Y. Ullrich A. Schlessinger J. Mol. Cell. Biol. 1992; 12: 5824-5833Crossref PubMed Scopus (155) Google Scholar, 34Park J.W. Ma M. Ruedi J.M. Smith R.M. Babior B.M. J. Biol. Chem. 1992; 267: 17327-17332Abstract Full Text PDF PubMed Google Scholar, 35Schlaepfer D.D. Broome M.A. Hunter T. Mol. Cell. Biol. 1997; 17: 1702-1713Crossref PubMed Scopus (404) Google Scholar). Nck is itself tyrosine-phosphorylated in response to many stimuli (33Li W. Hu E. Skolnik E.Y. Ullrich A. Schlessinger J. Mol. Cell. Biol. 1992; 12: 5824-5833Crossref PubMed Scopus (155) Google Scholar, 36Park D. Rhee S.G. Mol. Cell. Biol. 1992; 12: 5816-5823Crossref PubMed Scopus (98) Google Scholar, 37Meisenhelder J. Hunter T. Mol. Cell. Biol. 1992; 12: 5843-5856Crossref PubMed Google Scholar) and contains a conserved potential SH3-binding site, 3B. J.Mayer and T. Akiyama, unpublished observation. hence many possible interactions can be mediated by Nck. Although the interaction between Abl and Nck was suggested several years ago (29Ren R. Ye Z.-S. Baltimore D. Genes Dev. 1994; 8: 783-795Crossref PubMed Scopus (291) Google Scholar), recent data have led to new insights into their relationship. The Xenopus Arg protein (which is highly related to Abl) (38Kruh G.D. Perego R. Miki T. Aaronson S.A. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 5802-5806Crossref PubMed Scopus (136) Google Scholar) emerged from a screen of Xenopusproteins that bind specifically to the first two SH3 domains of Nck, and microinjection data suggest that Abl and Arg are effectors for the ability of Nck mutants to induce mesoderm patterning defects in early embryos. 4C. Adler, T. Akiyama, J. M. Smith, and B. J. Mayer, manuscript in preparation. OtherXenopus experiments demonstrated that localization of the first two SH3 domains of Nck to membranes (mimicking the relocalization of full-length Nck to sites of tyrosine phosphorylation on the membrane) induced strong mesoderm patterning effects (39Tanaka M. Lu W. Gupta R. Mayer B.J. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 4493-4498Crossref PubMed Scopus (30) Google Scholar). Here we report that expression of Nck SH3 domains can activate Abl in human 293T cells. These experiments establish a model system that will facilitate delineation of the mechanisms that underlie the regulation of Abl. All Abl genes are derived from mouse type IV c-Abl (12Jackson P. Baltimore D. EMBO J. 1989; 8: 449-456Crossref PubMed Scopus (222) Google Scholar, 40Ben-Neriah Y. Bernards A. Paskind M. Daley G.Q. Baltimore D. Cell. 1986; 44: 577-586Abstract Full Text PDF PubMed Scopus (170) Google Scholar) and were generated by polymerase chain reaction as described previously (15Mayer B.J. Baltimore D. Mol. Cell. Biol. 1994; 14: 2883-2894Crossref PubMed Scopus (158) Google Scholar). The c-Abl construct that we use here encodes authentic WT c-Abl and does not include linkers 5′ to the SH3 domain or 3′ to the SH2 domain as in earlier versions (15Mayer B.J. Baltimore D. Mol. Cell. Biol. 1994; 14: 2883-2894Crossref PubMed Scopus (158) Google Scholar). The “SF” and “SG” constructs both encode a stop codon following Arg-532 (immediately C-terminal to the catalytic domain). The AblΔPro mutant was generated by two-step polymerase chain reaction mutagenesis and contains a deletion (amino acids 536–645) encompassing the proline-rich adaptor binding region as well as the major nuclear localization signal (41Van Etten R.A. Jackson P. Baltimore D. Cell. 1989; 58: 669-678Abstract Full Text PDF PubMed Scopus (338) Google Scholar). All Abl clones are expressed via Moloney murine leukemia virus-derived retroviral vectors pGDN (15Mayer B.J. Baltimore D. Mol. Cell. Biol. 1994; 14: 2883-2894Crossref PubMed Scopus (158) Google Scholar) or pBPN (a derivative of pBABE-puro) (42Morgenstern J.P. Land H. Nucleic Acids Res. 1990; 18: 3587-3596Crossref PubMed Scopus (1903) Google Scholar). The generation of constructs expressing myristoylated and unmyristoylated Nck SH3 domains has been described (39Tanaka M. Lu W. Gupta R. Mayer B.J. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 4493-4498Crossref PubMed Scopus (30) Google Scholar, 43Lu W. Katz S. Gupta R. Mayer B.J. Curr. Biol. 1997; 7: 85-94Abstract Full Text Full Text PDF PubMed Scopus (215) Google Scholar). The Nck “PA” mutant was generated by polymerase chain reaction and contains a point mutation that changes proline 84 to alanine, eliminating the conserved potential SH3-binding site. All Nck constructs are derived from human Nck (31Lehman J.M. Riethmuller G. Johnson J.P. Nucleic Acids Res. 1990; 18: 1048Crossref PubMed Scopus (160) Google Scholar), tagged with the influenza virus HA epitope at their C termini, and expressed in the pEBB mammalian expression vector (39Tanaka M. Lu W. Gupta R. Mayer B.J. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 4493-4498Crossref PubMed Scopus (30) Google Scholar). Transfection of 293T cells was by the calcium phosphate method as described (44Tanaka M. Gupta R. Mayer B.J. Mol. Cell. Biol. 1995; 15: 6829-6837Crossref PubMed Scopus (221) Google Scholar). Cells were lysed 24–48 h post-transfection in 1 ml of kinase lysis buffer (25 mmTris, pH 7.4, 150 mm NaCl, 5 mm EDTA, 1% Triton X-100, 10 mm sodium pyrophosphate, 10 mmβ-glycerophosphate, 1 mm sodium orthovanadate, 10% glycerol, 1 mm phenylmethylsulfonyl fluoride, and 20 μl/ml aprotinin solution (Sigma)). The protein content of lysates was determined by Bradford assay (Bio-Rad). Immunoblotting of whole cell lysates was as described (15Mayer B.J. Baltimore D. Mol. Cell. Biol. 1994; 14: 2883-2894Crossref PubMed Scopus (158) Google Scholar). The following antibodies were used: for Abl, monoclonals 8E9 (PharMingen International) and 24–21 (kindly provided by A. Pendergast); monoclonal HA11 to detect Nck constructs (Berkeley Antibody Co.); monoclonal antibodies 4G10 (Upstate Biochemical, Inc.) and PY20(45) for anti-phosphotyrosine blotting. For immunoprecipitation, lysate from one-tenth of a 10-cm dish was pre-cleared with protein A-Sepharose beads (Amersham Pharmacia Biotech), then incubated on ice with 1 μg of anti-Abl antibody K12 (Santa Cruz Biotechnology, Inc.). Immune complexes were collected on protein A beads, washed three times in lysis buffer, and analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting as above. Our previous results indicated that c-Abl interacts with the first two SH3 domains of Nck, suggesting that Nck might affect Abl activity by changing its localization or binding partners. Nck has the potential to bind via its SH2 domain to tyrosine-phosphorylated proteins (for example in focal adhesions), and thereby serve to relocalize Abl in response to extracellular signals. Most tyrosine-phosphorylated proteins are associated with membranes, so we reasoned that fusion of the two N-terminal SH3 domains of Nck to a membrane targeting signal (the Src myristoylation sequence) might mimic endogenous signals and thus modulate Abl activity. We therefore tested whether a myristoylated form of the first two SH3 domains of Nck (Myr 1+2) or the unmyristoylated form (1+2) (depicted in Fig. 1) could affect the activity of overexpressed c-Abl in transiently transfected 293T cells. Preliminary experiments suggested that effects of Nck SH3 domains might be concentration-dependent, so a wide range of plasmid amounts (from 25 ng to 8 μg) was tested for each Nck construct. The in vivo activity of Abl was assayed indirectly by the presence of tyrosine-phosphorylated proteins in total cell extracts. Both myristoylated and unmyristoylated SH3–1+2 induced a dramatic increase in cellular phosphotyrosine, not only in Abl itself but also in a variety of other cellular proteins (Fig.2). Autophosphorylation of c-Abl has historically been used as a read-out for Abl activation (9Pendergast A.M. Muller A.J. Havlik M.H. Clark R. McCormick F. Witte O.N. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 5927-5931Crossref PubMed Scopus (123) Google Scholar, 11Franz W.M. Berger P. Wang J.Y.J. EMBO J. 1989; 8: 137-147Crossref PubMed Scopus (161) Google Scholar), and immunoprecipitation experiments showed that the major tyrosine-phosphorylated protein in transfected cell lysates of approximately 150-kDa molecular mass was Abl itself (data not shown). These results indicate that Nck SH3 domains can activate the normally repressed c-Abl kinase. A myristoylated SH3–1+2 construct in which the binding activity of the two SH3 domains was ablated by site-directed mutation did not activate Abl (data not shown), indicating that ligand binding is required. Although both Myr 1+2 and the unmyristoylated version activated Abl, the concentration dependence of the two was very different. In the case of Myr 1+2, maximal activation was achieved with 100–250 ng and quickly tapered off at increasing concentrations (Fig. 2). At these higher concentrations of Myr 1+2, a decrease in the total amount of Abl was usually observed, suggesting feedback regulation at the level of protein synthesis or degradation. Activation by the unmyristoylated 1+2 occurred only with very high amounts, typically requiring 10–50-fold more plasmid than the myristoylated version to activate to a comparable level (Fig. 2). The concentration-dependent activation of Abl was highly reproducible, although the amount of Nck plasmid required for maximal activation and the absolute level of Abl activation varied somewhat from experiment to experiment, most likely because of differences in transfection efficiency. The Nck 1+2 constructs were themselves tyrosine-phosphorylated, with the extent of phosphorylation correlating with Abl activity (data not shown). In similar co-transfection experiments, we also observed activation of the Abl-related kinase Arg-1B (38Kruh G.D. Perego R. Miki T. Aaronson S.A. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 5802-5806Crossref PubMed Scopus (136) Google Scholar) by both myristoylated and unmyristoylated SH3–1+2 (data not shown). As expected from the increase in in vivo Abl activity, there was a modest increase in the in vitro kinase activity of Abl immunoprecipitated from lysates of cells expressing low amounts of Myr 1+2 or high amounts of 1+2 compared with those expressing Abl alone (data not shown). When Nck SH3–1+2 constructs were transfected alone we did not observe significant increases in total cell phosphotyrosine or in the tyrosine phosphorylation orin vitro kinase activity of endogenous Abl; activation of endogenous Abl in these experiments would be difficult to detect, however, because of background from untransfected cells. To gain insight into the mechanism of activation, the domains of Abl necessary for activation by either myristoylated or unmyristoylated SH3–1+2 were investigated. The Abl constructs tested are depicted in Fig. 1. SF-Abl encodes a c-Abl protein that is truncated immediately C-terminal to the catalytic domain and is therefore lacking the known adaptor-binding sites in addition to a variety of other protein-binding sites and the nuclear localization and nuclear export signals (4Raitano A.B. Whang Y.E. Sawyers C.L. Biochim. Biophys. Acta. 1997; 1333: F201-F216PubMed Google Scholar, 28Feller S.M. Knudsen B. Hanafusa H. EMBO J. 1994; 13: 2341-2351Crossref PubMed Scopus (326) Google Scholar, 29Ren R. Ye Z.-S. Baltimore D. Genes Dev. 1994; 8: 783-795Crossref PubMed Scopus (291) Google Scholar, 46Wen S.-T. Jackson P.K. Van Etten R.A. EMBO J. 1996; 15: 1583-1595Crossref PubMed Scopus (186) Google Scholar, 47Taagepera S. McDonald D. Loeb J.E. Whittaker L.L. McElroy A.K. Wang J.Y.J. Hope T.J. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 7457-7462Crossref PubMed Scopus (264) Google Scholar). G-Abl is an SH3-deleted form of c-Abl, which has previously been shown to be constitutively active in vivo and transforming (15Mayer B.J. Baltimore D. Mol. Cell. Biol. 1994; 14: 2883-2894Crossref PubMed Scopus (158) Google Scholar), and SG-Abl is a form of G-Abl containing the identical C-terminal truncation to SF-Abl. SG-Abl is transforming in fibroblasts (though focus formation is decreased approximately 10-fold relative to full-length G-Abl), whereas SF-Abl is nontransforming, indicating that C-terminal truncation does not abrogate the SH3-mediated repression of Abl activity.2 As shown in Fig. 3, neither Myr 1+2 nor 1+2 had a significant effect on the levels of tyrosine phosphorylation induced by SF-Abl, G-Abl, or SG-Abl over a wide range of concentrations. The phosphotyrosine levels in the SF-Abl blot remain steady and equivalent to the negative control, whereas under identical conditions WT c-Abl was clearly activated (Fig. 3 A, comparelanes 3 and 7). The lack of activation of the SF mutant demonstrates that the C terminus contains sequences required for activation by Nck SH3 domains. As expected, tyrosine phosphorylation levels in cells expressing the transforming mutants G-Abl and SG-Abl were high relative to controls; they were not significantly affected, however, by the presence of Myr 1+2 or 1+2. A slight increase in phosphotyrosine in cells expressing Nck SH3 domains and G-Abl or SG-Abl was occasionally observed (Fig. 3, B and C), but this was modest and inconsistent unlike the robust activation seen in the case of WT c-Abl. The inability of Nck SH3 domains to further enhance the activity of mutagenically activated Abl suggests that its effects are not because of some general mechanism affecting phosphotyrosine metabolism, such as inhibition of phosphatase activity in the cell. It can also be seen that Nck SH3 domains elevate thein vivo activity of c-Abl to a level comparable with that of transforming mutants such as G-Abl (Fig. 3 B, comparelanes 3 and 4). The lack of activation of the C-terminally truncated SF-Abl mutant suggested that direct interaction between the Nck SH3 domains and the Abl C terminus might be involved in activation. Immediately following the catalytic region of c-Abl is a proline-rich region that has been previously shown to contain binding sites for the Crk, Grb2, and Nck SH3 domains (28Feller S.M. Knudsen B. Hanafusa H. EMBO J. 1994; 13: 2341-2351Crossref PubMed Scopus (326) Google Scholar, 29Ren R. Ye Z.-S. Baltimore D. Genes Dev. 1994; 8: 783-795Crossref PubMed Scopus (291) Google Scholar, 30Mayer B.J. Hirai H. Sakai R. Curr. Biol. 1995; 5: 296-305Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar). We have also demonstrated that this region contains the major binding sites for Nck SH3–1+2 in vitro.4 We therefore generated a construct (AblΔPro) to address whether the high affinity binding site for Nck SH3 domains is necessary to mediate Abl activation (Fig. 1). Plasmids encoding c-Abl and the AblΔPro mutant were co-transfected with increasing amounts of Myr 1+2, and activity was assayed by immunoblotting with antiphosphotyrosine antibodies as above. Surprisingly, AblΔPro was activated by Myr 1+2 to a similar extent and with similar concentration dependence as WT c-Abl (Fig.4). A more precise mutant that eliminates the high-affinity Nck-binding site but does not affect the Crk-binding site or major nuclear localization signal was also activated to a similar extent (data not shown). Although activation levels of either c-Abl or AblΔPro were similar, the association of Myr 1+2 with AblΔPro was much weaker than with WT c-Abl as detected by co-immunoprecipitation (Fig. 5).Figure 5Immunoprecipitation of Abl and Nck constructs. 4 μg of plasmid expressing WT Abl or AblΔPro mutant was transfected with 1.6 μg of plasmid expressing Myr 1+2 or the Myr 1+2 PA mutant. 100 μl of lysate was immunoprecipitated with 1 μg of polyclonal anti-Abl antibody K12. Immunoprecipitates were separated on SDS-polyacrylamide gel electrophoresis, transferred to filters, and probed with either monoclonal anti-Abl antibody 8E9 (top) or anti-HA to detect Nck (bottom).Arrowhead indicates position of Myr 1+2 and Myr 1+2 PA. Upper band in the Nck blot is nonspecific.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Because this result suggested that direct association of Nck SH3 domains with the Abl C terminus might not be required for activation, we considered other possible modes of interaction between Myr-1+2 and Abl. We noted that there is a potential type I SH3-binding site KRKPS(V/M)P located between SH3–1 and SH3–2 that is conserved in human, mouse, and Xenopus Nck.3 We reasoned that interaction between the Abl SH3 domain and this site could potentially relieve the inhibition of Abl and explain activation by the Nck constructs. To test this model we transfected 293T cells with a mutant Myr 1+2 in which one of the conserved prolines of the putative SH3-binding site was mutated to alanine (Myr 1+2 PA mutant). This mutation is expected to render this site unable to bind SH3 domains (48Mayer B.J. Eck M.J. Curr. Biol. 1995; 5: 364-367Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar). As shown in Fig. 4, the PA mutant activated both c-Abl and the AblΔPro mutant to a similar extent as WT Myr 1+2, and the PA mutation had no effect on coprecipitation of Abl and Myr-1+2 (Fig. 5). This demonstrates that the putative SH3-binding site in Nck is not required for activation. The results shown in Figs. 4 and 5 suggest that activation of c-Abl by Myr 1+2 cannot be explained simply by binding to the C terminus of c-Abl and relocalizing it to membranes. We confirmed this by constructing a myristoylated Crk SH3 domain. c-Abl binds the Crk SH3 domain with higher affinity than to the Nck SH3 domains (29Ren R. Ye Z.-S. Baltimore D. Genes Dev. 1994; 8: 783-795Crossref PubMed Scopus (291) Google Scholar)2; therefore a myristoylated Crk SH3 construct should activate c-Abl at least as well as the Nck Myr 1+2 construct if binding and relocalization were sufficient for activation. To the contrary, we found that although the Crk SH3 construct slightly activated c-Abl, the extent of activation was minimal when compared with Myr 1+2 or even to a myristoylated construct encoding the SH3–2 of Nck in the absence of SH3–1 (data not shown). Our results demonstrate that overexpression of Nck SH3 domains can potently activate the tyrosine kinase activity of the Abl kinase in transfected cells. The observation that Myr 1+2 could activate Abl was suggestive because this construct has the potential to relocalize Abl, and others have shown that a fraction of endogenous Abl translocates from the nucleus to focal adhesions upon integrin engagement, concomitant with its activation (24Lewis J.M. Baskaran R. Taagepera S. Schwartz M.A. Wang J.Y.J. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 15174-15179Crossref PubMed Scopus (269) Google Scholar, 25Lewis J.M. Schwartz M. J. Biol. Chem. 1998; 273: 14225-14230Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). To our surprise, however, we found that the major Nck SH3-binding sites in the Abl C terminus were not required for activation. Although we cannot rule out other cryptic SH3-binding sites in the C terminus that are able to partially compensate for the loss of the major Nck SH3-binding sites, the observation that the AblΔPro mutant (lacking the major SH3-binding sites) co-precipitated with Myr-1+2 much more poorly than WT Abl suggests this is not the case. We were also surprised to find that the unmyristoylated 1+2 construct (which is largely nuclear in localization 5J. M. Smith and B. J. Mayer, unpublished observation. ) could activate Abl, albeit at much higher levels of expression. This implies that increased membrane localization cannot explain activation of Abl by Nck, although membrane localization of the Nck SH3 domains clearly facilitates activation. One possibility that we considered was that the conserved potential SH3-binding site in Nck might engage the Abl SH3 domain and relieve inhibition, because activation of the Src family kinases has been observed using ligands that bind their SH3 domain (8Moarefi I. LeFevre-Bernt M. Sicheri F. Huse M. Lee C.-H. Kuriyan J. Miller W.T. Nature. 1997; 385: 650-653Crossref PubMed Scopus (541) Google Scholar,49Alexandropoulos K. Baltimore D. Genes Dev. 1996; 10: 1341-1355Crossref PubMed Scopus (221) Google Scholar). However, mutation of this putative SH3-binding site did not affect activation of Abl by Myr 1+2 (Fig. 4). Several possible mechanisms must be considered in trying to understand Nck-mediated activation of Abl. One intriguing possibility is that the action of Nck SH3 domains is indirect, the result of perturbation of signaling pathways that ultimately impinge on Abl activation. Little is known about specific stimuli that can activate c-Abl in vivo, and activation by Nck could provide important insights in this regard. Another possibility is that the Nck SH3 domains bind to and sequester an endogenous inhibitor of c-Abl that normally binds to the Abl SH3 domain (9Pendergast A.M. Muller A.J. Havlik M.H. Clark R. McCormick F. Witte O.N. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 5927-5931Crossref PubMed Scopus (123) Google Scholar). Such a mechanism would be more likely to explain the activation by large amounts of unmyristoylated 1+2; it is less likely to explain activation by Myr 1+2, given the small amounts of plasmid required for maximal activation. It is also possible that the Nck SH3 domains bind directly to the SH2-kinase domain linker of c-Abl, preventing inhibitory intramolecular interactions with the Abl SH3. If this were the case, then the much higher potency of the Myr 1+2 construct would imply that the membrane-associated pool of Abl is more susceptible to activation, or that membrane localization leads to higher local concentrations of Nck SH3 domains with respect to a pool of Abl that is also localized on the membrane. Because no single mechanism is entirely consistent with all of our experimental results, it must be kept in mind that multiple mechanisms could be involved. In this regard, our results using the C-terminally truncated Abl mutants SF and SG are quite informative (Fig. 3,A and C). The only difference between these two constructs is the presence or absence of the Abl SH3 domain, which is known to play a role in inhibiting kinase activity. The phosphotyrosine levels seen in the SG-Abl blot (Fig. 3 C) and the transforming activity of this mutant indicate that it is activated, and imply that SF-Abl could also be activated if the inhibitory effects of its intact SH3 domain were disrupted. The observation that SF-Abl is not activated by Myr 1+2 or 1+2 argues against any mechanism that only involves disruption of SH3-mediated repression. These results could be reconciled, however, if activation involves both a high local concentration of Nck SH3 domains with respect to Abl, plus a specific subcellular localization or protein interaction that is conferred by the C terminus. More specific Abl mutants will resolve the role of the C terminus in activation. Regardless of the specific mechanism(s), Nck-mediated activation provides an excellent experimental system to study the parameters governing Abl activity in vivo. It is robust and highly reproducible and allows analysis of mutants of both Abl and Nck. Further experiments will surely reveal the details of regulation and provide insights into methods for rationally modulating the activity of Abl in human disease.