Characterization of unique functionalities in c-Src domains required for osteoclast podosome belt formation

Deletion of c-Src, a ubiquitously expressed tyrosine kinase, results in osteoclast dysfunction and osteopetrosis, in which bones harden into “stone.” In contrast, deletion of the genes encoding other members of the Src family kinase (SFK) fails to produce an osteopetrotic phenotype. This suggests that c-Src performs a unique function in the osteoclast that cannot be compensated for by other SFKs. We aimed to identify the molecular basis of this unique role in osteoclasts and bone resorption. We found that c-Src, Lyn, and Fyn were the most highly expressed SFKs in WT osteoclasts, whereas Hck, Lck, Blk, and Fgr displayed low levels of expression. Formation of the podosome belt, clusters of unique actin assemblies, was disrupted in src−/− osteoclasts; introduction of constitutively activated SFKs revealed that only c-Src and Fyn could restore this process. To identify the key structural domains responsible, we constructed chimeric Src–Hck and Src–Lyn constructs in which the unique, SH3, SH2, or catalytic domains had been swapped. We found that the Src unique, SH3, and kinase domains were each crucial to establish Src functionality. The SH2 domain could however be substituted with Lyn or Hck SH2 domains. Furthermore, we demonstrate that c-Src’s functionality is, in part, derived from an SH3–proximal proline-rich domain interaction with c-Cbl, leading to phosphorylation of c-Cbl Tyr700. These data help clarify Src’s unique functionality in the organization of the cytoskeleton in osteoclasts, required for efficient bone resorption and explain why c-Src cannot be replaced, in osteoclasts, by other SFKs. Deletion of c-Src, a ubiquitously expressed tyrosine kinase, results in osteoclast dysfunction and osteopetrosis, in which bones harden into “stone.” In contrast, deletion of the genes encoding other members of the Src family kinase (SFK) fails to produce an osteopetrotic phenotype. This suggests that c-Src performs a unique function in the osteoclast that cannot be compensated for by other SFKs. We aimed to identify the molecular basis of this unique role in osteoclasts and bone resorption. We found that c-Src, Lyn, and Fyn were the most highly expressed SFKs in WT osteoclasts, whereas Hck, Lck, Blk, and Fgr displayed low levels of expression. Formation of the podosome belt, clusters of unique actin assemblies, was disrupted in src−/− osteoclasts; introduction of constitutively activated SFKs revealed that only c-Src and Fyn could restore this process. To identify the key structural domains responsible, we constructed chimeric Src–Hck and Src–Lyn constructs in which the unique, SH3, SH2, or catalytic domains had been swapped. We found that the Src unique, SH3, and kinase domains were each crucial to establish Src functionality. The SH2 domain could however be substituted with Lyn or Hck SH2 domains. Furthermore, we demonstrate that c-Src’s functionality is, in part, derived from an SH3–proximal proline-rich domain interaction with c-Cbl, leading to phosphorylation of c-Cbl Tyr700. These data help clarify Src’s unique functionality in the organization of the cytoskeleton in osteoclasts, required for efficient bone resorption and explain why c-Src cannot be replaced, in osteoclasts, by other SFKs. The ubiquitously expressed rous sarcoma oncogene tyrosine kinase (Src) is important in cell proliferation, adhesion, and migration (1Thomas S.M. Brugge J.S. Cellular functions regulated by Src family kinases.Annu. Rev. Cell Dev. Biol. 1997; 13: 513-609Crossref PubMed Scopus (2109) Google Scholar). Despite its widespread expression in multiple cell types, the major phenotype of src−/− mice is osteopetrosis caused by defective osteoclast bone resorption (2Boyce B.F. Yoneda T. Lowe C. Soriano P. Mundy G.R. Requirement of pp60c-src expression for osteoclasts to form ruffled borders and resorb bone in mice.J. Clin. Invest. 1992; 90: 1622-1627Crossref PubMed Scopus (510) Google Scholar), suggesting that Src fulfills a unique, nonredundant function in osteoclasts. Among the Src family kinases (SFKs), only the deletion of Src results in osteopetrosis (3Lowe C. Yoneda T. Boyce B.F. Chen H. Mundy G.R. Soriano P. Osteopetrosis in Src-deficient mice is due to an autonomous defect of osteoclasts.Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 4485-4489Crossref PubMed Scopus (281) Google Scholar, 4Lowell C.A. Soriano P. Varmus H.E. Functional overlap in the src gene family: Inactivation of hck and fgr impairs natural immunity.Genes Dev. 1994; 8: 387-398Crossref PubMed Scopus (191) Google Scholar, 5Stein P.L. Vogel H. Soriano P. Combined deficiencies of Src, Fyn, and Yes tyrosine kinases in mutant mice.Genes Dev. 1994; 8: 1999-2007Crossref PubMed Scopus (259) Google Scholar). It has been shown that the mechanism by which Src deletion affects bone resorption is for the most part by altering a signaling pathway that governs the formation, turnover, and organization of podosomes in a peripheral belt in active osteoclasts attached to the bone (6Miyazaki T. Tanaka S. Sanjay A. Baron R. The role of c-Src kinase in the regulation of osteoclast function.Mod. Rheumatol. 2006; 16: 68-74Crossref PubMed Scopus (90) Google Scholar). Some degree of redundancy may exist however within the SFK family of tyrosine kinases because src−/−;hck−/− double-KO mice are more osteopetrotic than Src-deficient mice (7Lowell C.A. Niwa M. Soriano P. Varmus H.E. Deficiency of the Hck and Src tyrosine kinases results in extreme levels of extramedullary hematopoiesis.Blood. 1996; 87: 1780-1792Crossref PubMed Google Scholar). In contrast, deleting Lyn increases receptor activator of nuclear factor kappa-B ligand (RANKL)-induced osteoclastogenesis, resulting in osteopenia rather than osteopetrosis, and has little effect on the activity of mature osteoclasts (8Kim H.-J. Zhang K. Zhang L. Ross F.P. Teitelbaum S.L. Faccio R. The Src family kinase, Lyn, suppresses osteoclastogenesis in vitro and in vivo.Proc. Natl. Acad. Sci. U. S. A. 2009; 106: 2325-2330Crossref PubMed Scopus (38) Google Scholar, 9Yoon S.-H. Lee Y. Kim H.-J. Lee Z.H. Hyung S.-W. Lee S.-W. Kim H.-H. Lyn inhibits osteoclast differentiation by interfering with PLCγ1-mediated Ca2+ signaling.FEBS Lett. 2009; 583: 1164-1170Crossref PubMed Scopus (23) Google Scholar). These reports suggest that while some other SFKs compensate for the absence of Src in most tissues and Hck may partially compensate for Src’s absence in osteoclasts, Src is specifically and uniquely required for the function of mature osteoclasts, more specifically the organization of their cytoskeleton. Src plays a central role in integrin-mediated cell adhesion and the formation of podosomes, the specialized adhesion structures that are required for normal osteoclastic bone resorption (10Horne W.C. Sanjay A. Bruzzaniti A. Baron R. The role(s) of Src kinase and Cbl proteins in the regulation of osteoclast differentiation and function.Immunol. Rev. 2005; 208: 106-125Crossref PubMed Scopus (126) Google Scholar). Src is activated downstream of αvβ3 integrin, which mediates the attachment of osteoclasts to bone matrix and contributes to the migration of osteoclasts over the bone surface and the formation of the sealing zone that surrounds and delineates the extracellular bone resorbing compartment (10Horne W.C. Sanjay A. Bruzzaniti A. Baron R. The role(s) of Src kinase and Cbl proteins in the regulation of osteoclast differentiation and function.Immunol. Rev. 2005; 208: 106-125Crossref PubMed Scopus (126) Google Scholar, 11Jurdic P. Saltel F. Chabadel A. Destaing O. Podosome and sealing zone: Specificity of the osteoclast model.Eur. J. Cell Biol. 2006; 85: 195-202Crossref PubMed Scopus (269) Google Scholar, 12Nakamura I. Duong L.T. Rodan S.B. Rodan G.A. Involvement of alpha(v)beta3 integrins in osteoclast function.J. Bone Miner. Metab. 2007; 25: 337-344Crossref PubMed Scopus (165) Google Scholar, 13Teitelbaum S.L. The osteoclast and its unique cytoskeleton.Ann. N. Y. Acad. Sci. 2011; 1240: 14-17Crossref PubMed Scopus (104) Google Scholar). In this area, Src regulates the formation and organization of podosomes, the dynamic punctate actin adhesion structures that are found in osteoclasts and other highly motile cells (11Jurdic P. Saltel F. Chabadel A. Destaing O. Podosome and sealing zone: Specificity of the osteoclast model.Eur. J. Cell Biol. 2006; 85: 195-202Crossref PubMed Scopus (269) Google Scholar, 14Destaing O. Sanjay A. Itzstein C. Horne W.C. Toomre D. De Camilli P. Baron R. The tyrosine kinase activity of c-Src regulates actin dynamics and organization of podosomes in osteoclasts.Mol. Biol. Cell. 2008; 19: 394-404Crossref PubMed Scopus (152) Google Scholar). Consequently, Src-deficient osteoclasts are not only unable to resorb the bone but also 50 to 60% less mobile than WT osteoclasts (15Sanjay A. Houghton A. Neff L. DiDomenico E. Bardelay C. Antoine E. Levy J. Gailit J. Bowtell D. Horne W.C. Baron R. Cbl associates with Pyk2 and Src to regulate Src kinase activity, alpha(v)beta(3) integrin-mediated signaling, cell adhesion, and osteoclast motility.J. Cell Biol. 2001; 152: 181-195Crossref PubMed Scopus (334) Google Scholar, 16Chiusaroli R. Sanjay A. Henriksen K. Engsig M.T. Horne W.C. Gu H. Baron R. Deletion of the gene encoding c-Cbl alters the ability of osteoclasts to migrate, delaying resorption and ossification of cartilage during the development of long bones.Dev. Biol. 2003; 261: 537-547Crossref PubMed Scopus (57) Google Scholar). Our studies also indicate that Src promotes the initiation of podosome formation, the rate of actin polymerization within podosomes, the disassembly of podosomes, and the organization of the peripheral podosome belt by phosphorylating some of the many proteins that are known to be Src substrate proteins and to regulate podosome function (14Destaing O. Sanjay A. Itzstein C. Horne W.C. Toomre D. De Camilli P. Baron R. The tyrosine kinase activity of c-Src regulates actin dynamics and organization of podosomes in osteoclasts.Mol. Biol. Cell. 2008; 19: 394-404Crossref PubMed Scopus (152) Google Scholar, 17Miyazaki T. Sanjay A. Neff L. Tanaka S. Horne W.C. Baron R. Src kinase activity is essential for osteoclast function.J. Biol. Chem. 2004; 279: 17660-17666Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar). All members of the SFK family have similar domain structures with four distinct domains: unique, Src homology (SH)3, SH2, and kinase (catalytic) domains (18Levinson A.D. Oppermann H. Varmus H.E. Bishop J.M. The purified product of the transforming gene of avian sarcoma virus phosphorylates tyrosine.J. Biol. Chem. 1980; 255: 11973-11980Abstract Full Text PDF PubMed Google Scholar, 19Koch C.A. Anderson D. Moran M.F. Ellis C. Pawson T. SH2 and SH3 domains: Elements that control interactions of cytoplasmic signaling proteins.Science. 1991; 252: 668-674Crossref PubMed Scopus (1424) Google Scholar, 20Pawson T. Gish G.D. SH2 and SH3 domains: From structure to function.Cell. 1992; 71: 359-362Abstract Full Text PDF PubMed Scopus (786) Google Scholar). Unique domains contain myristoylation and palmitoylation sites that mediate membrane association, but otherwise, there is little homology between any of the SFK unique domains (21Resh M.D. Myristylation and palmitylation of Src family members: The fats of the matter.Cell. 1994; 76: 411-413Abstract Full Text PDF PubMed Scopus (582) Google Scholar). In contrast, SH3 and SH2 domains contain conserved and specific amino acid sequences that bind to proximal proline-rich (PPR) motifs and sequences containing phosphotyrosines, respectively, on other signaling proteins to form signaling complexes. The SFK SH3 and SH2 domains also bind intramolecularly to downregulate Src/SFK kinase activity (20Pawson T. Gish G.D. SH2 and SH3 domains: From structure to function.Cell. 1992; 71: 359-362Abstract Full Text PDF PubMed Scopus (786) Google Scholar, 22Ren R. Mayer B.J. Cicchetti P. Baltimore D. Identification of a ten-amino acid proline-rich SH3 binding site.Science. 1993; 259: 1157-1161Crossref PubMed Scopus (1011) Google Scholar, 23Songyang Z. Shoelson S.E. Chaudhuri M. Gish G. Pawson T. Haser W.G. King F. Roberts T. Ratnofsky S. Lechleider R.J. SH2 domains recognize specific phosphopeptide sequences.Cell. 1993; 72: 767-778Abstract Full Text PDF PubMed Scopus (2358) Google Scholar, 24Yu H. Chen J.K. Feng S. Dalgarno D.C. Brauer A.W. Schreiber S.L. Structural basis for the binding of proline-rich peptides to SH3 domains.Cell. 1994; 76: 933-945Abstract Full Text PDF PubMed Scopus (863) Google Scholar). The kinase domain phosphorylates components of the signaling complexes on tyrosine residues (25Bolen J.B. Rowley R.B. Spana C. Tsygankov A.Y. The Src family of tyrosine protein kinases in hemopoietic signal transduction.FASEB J. 1992; 6: 3403-3409Crossref PubMed Scopus (193) Google Scholar). To delineate the molecular basis for the specific role of Src in osteoclasts, we first determined the relative expression level of other Src family kinases in these cells and analyzed the effects of expressing these in src−/− osteoclasts. We found that Lyn is upregulated in src−/− osteoclasts but neither compensates for Src’s absence nor represses Src’s ability to promote cell spreading and podosome belt formation. To identify the features of Src that are the basis of the kinase’s unique role in regulating osteoclast function, we expressed chimeric Hck–Src and Lyn–Src proteins in an open conformation and determined how replacing individual Src domains by the homolog domain from these other SFKs altered osteoclast spreading and podosome belt formation. The results indicate that Src’s unique, SH3, and kinase, but not the SH2, domains are required for the specific Src functions in osteoclasts. To determine whether Src’s unique functional role in osteoclasts is only due to the particularly high Src level of expression relative to other SFKs, we first measured the expression level of all SFKs in WT osteoclasts by real-time quantitative PCR (qPCR) (Fig. 1A). Src and Lyn were more highly expressed than the other SFKs, with Lyn expression about 75% of Src expression. Fyn expression was about one third of Src expression and the expression of all the other SFKs was less than 10% of the Src level. Thus, Src, Lyn, and Fyn are the major SFKs in osteoclasts. To determine whether Src deletion affects the expression level of other SFKs, possibly attempting to compensate the absence of Src, we next measured SFK expression in src−/− osteoclasts. As shown in Figure 1A, in src−/− osteoclasts, Lyn, Hck, and Fgr were significantly upregulated. Lyn expression increased to almost 4-fold higher than the expression of Src in WT osteoclasts, and the expression of Hck increased by 10-fold, reaching the level of Lyn expression in WT osteoclasts. In contrast, expression of Fyn was significantly lower in the src−/− osteoclasts and expression of the other SFKs remained very low (Fig. 1A). Corresponding with mRNA data, protein levels of Hck and Lyn in src−/− osteoclasts were about 4- to 5-fold higher than that of WT osteoclasts (Fig. 2D, Fig. S1A). Fyn in src−/− osteoclasts was about 60 to 70% of that in WT osteoclasts (Fig. 2D, Fig. S1A). It has been reported that overexpression of Lyn suppresses osteoclast differentiation (8Kim H.-J. Zhang K. Zhang L. Ross F.P. Teitelbaum S.L. Faccio R. The Src family kinase, Lyn, suppresses osteoclastogenesis in vitro and in vivo.Proc. Natl. Acad. Sci. U. S. A. 2009; 106: 2325-2330Crossref PubMed Scopus (38) Google Scholar). Thus, it is possible that high levels of Lyn in src−/− osteoclasts can suppress mature osteoclast function. To determine whether high levels of Lyn have a negative effect on podosome belt formation, we overexpressed constitutively active LynY487F in mature WT osteoclasts using an adenovirus system. Mature WT osteoclasts were large and multinuclear and formed podosome belts (Fig. 1, B–D). Infection with the CRE, SrcY527F, and/or LynY487F adenoviruses had little effect on tartrate-resistant acid phosphatase (TRAP) staining and podosome belt formation of the WT osteoclasts. These results suggest that high levels of Lyn are not likely to lead to a disruption in actin organization. To determine which SFKs can support normal actin organization in osteoclasts, we introduced constitutively active SFKs into mature src−/− osteoclasts. Mature WT osteoclasts differentiated from spleen cells with macrophage colony-stimulating factor and RANKL spread and had a round shape (Fig. 2A). F-actin in WT osteoclasts was organized in peripheral podosome belts (Fig. 2, B and C). In contrast, src−/− osteoclasts were not well spread, had irregular forms (Fig. 2A), and failed to form podosome belts (Fig. 2, B and C), as previously described (14Destaing O. Sanjay A. Itzstein C. Horne W.C. Toomre D. De Camilli P. Baron R. The tyrosine kinase activity of c-Src regulates actin dynamics and organization of podosomes in osteoclasts.Mol. Biol. Cell. 2008; 19: 394-404Crossref PubMed Scopus (152) Google Scholar). As expected, infection of src−/− osteoclasts with CRE adenovirus did not affect cell shape or actin organization. Introduction of the constitutively active SrcY527F rescued the cell spreading and podosome belt formation of src−/− osteoclasts as previously reported (14Destaing O. Sanjay A. Itzstein C. Horne W.C. Toomre D. De Camilli P. Baron R. The tyrosine kinase activity of c-Src regulates actin dynamics and organization of podosomes in osteoclasts.Mol. Biol. Cell. 2008; 19: 394-404Crossref PubMed Scopus (152) Google Scholar, 17Miyazaki T. Sanjay A. Neff L. Tanaka S. Horne W.C. Baron R. Src kinase activity is essential for osteoclast function.J. Biol. Chem. 2004; 279: 17660-17666Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar) (Fig. 2). Interestingly, FynY528F also restored effectively the cell spreading, shape, and podosome belt formation of src−/− osteoclasts, suggesting that the failure of Fyn to compensate for the absence of Src was the consequence of low expression level (Fig. 2A). On the other hand, induction of other SFKs did not rescue podosome belt formation of src−/− osteoclasts. (Fig. 2, B–D). HckY520F fails to significantly restore the spreading and podosome belt formation in src−/− osteoclasts (Fig. 2). However, the absence of both Src and Hck results in significantly more severe osteopetrosis than the absence of Src alone (7Lowell C.A. Niwa M. Soriano P. Varmus H.E. Deficiency of the Hck and Src tyrosine kinases results in extreme levels of extramedullary hematopoiesis.Blood. 1996; 87: 1780-1792Crossref PubMed Google Scholar), suggesting that some feature of Hck may be functionally redundant with certain Src domains. To explore the contributions of Src’s individual domains to its unique ability to promote podosome belt formation, we constructed chimeras of Src and Hck and expressed them in src−/− osteoclasts. We hypothesized that the specialized SFK binding domains, especially the SH2 and SH3 domains, could determine the specificity of Src. We first swapped the SH2 and SH3 binding domains of Hck and Src to generate two complementary chimeras with the binding domains of one SFK and the catalytic domain of the other (Fig. 3A). Both chimeras had nonspecific kinase activity with a short peptide substrate, although the activity of the chimera with the Hck kinase domain had only about half the activity of the Src kinase domain chimera (Fig. 3B). Similar data were obtained by Western blotting analysis (Fig. S2). Contrary to our hypothesis that the binding domains would provide Src’s functional specificity, we found that the Hck chimera with Hck SH2-SH3 and the Src kinase domain (H2SK) was comparable to Src itself in its ability to restore cell spreading, shape, and podosome belt formation in src−/− osteoclasts when expressed at the same expression level as HckY520F. In contrast, the Src chimera with Hck kinase domain and Src SH2–SH3 (S2HK) only partially restored cell spreading and had much less of an effect on the podosome belt formation than Src although the expression level of S2HK was almost same (Fig. 3, C–F, Fig. S3). These results suggest that the Src and Hck SH2 and SH3 domains have comparable specificity for the target proteins that mediate the localization of Src during the induction of spreading and podosome belt formation, but the kinase domains differ in their ability to phosphorylate the proteins that play key roles in these processes. Despite the high level of Lyn expression, src−/− osteoclasts do not spread to a round shape or form podosome belts. Moreover, overexpression of constitutively active Lyn in mature src−/− osteoclasts did not restore cell spreading or podosome belt formation (Fig. 2). These results suggest that Lyn is largely unable to replicate Src activity in osteoclasts. Thus, replacing individual Src domains with the corresponding Lyn domains should at least partially disable Src function. We therefore constructed eight Src-Lyn chimeras to analyze the functional role of each Src domain (Fig. 4A). Like the Src–Hck chimeras, the Src–Lyn chimeras retained kinase activity with both peptide and protein substrates although target proteins were different (Fig. 4B, Fig. S4). The replacement of either the Src kinase domain (S2LK) or the Src-binding domains (L2SK) with Lyn domains eliminated much of Src’s ability to restore spreading (Fig. 4C) or podosome belt formation (Fig. 4, D and E). The near-complete inability of the chimera with the Lyn SH2 and SH3 domains (L2SK) to promote spreading or podosome belt formation suggests that Src SH2 and/or SH3 domains are required for both activities of Src. On the other hand, the chimera of the Src SH2 and SH3 with Lyn kinase domain (S2LK) was about 10-fold more active in supporting belt formation than the complementary half-chimera, indicating that the Lyn kinase domain has some ability to phosphorylate Src target protein(s) that promote belt formation when properly localized by the Src-binding domains, but still at only 25 to 30% of the activity of full-length Src. We further analyzed the contributions of the three Src-binding domains (unique, SH3, and SH2) by replacing one or two of them with the corresponding Lyn domains (Fig. 4, B–F, Fig. S5). Replacing only the SH2 domain of Src with the Lyn SH2 (L2Src) had little effect on either spreading or belt formation. In contrast, replacing the SH3 domain of Src with Lyn SH3 (L3Src) significantly reduced both spreading and belt formation, suggesting that Src specificity in osteoclast cytoskeletal organization depends in large part upon the SH3 domain. The major molecular mechanism by which c-Src functions in osteoclasts is the phosphorylation of c-Cbl, a ubiquitin E3 ligase that regulates podosome belt formation (17Miyazaki T. Sanjay A. Neff L. Tanaka S. Horne W.C. Baron R. Src kinase activity is essential for osteoclast function.J. Biol. Chem. 2004; 279: 17660-17666Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar, 26Bruzzaniti A. Neff L. Sanjay A. Horne W.C. De Camilli P. Baron R. Dynamin forms a Src kinase-sensitive complex with Cbl and regulates podosomes and osteoclast activity.Mol. Biol. Cell. 2005; 16: 3301-3313Crossref PubMed Scopus (107) Google Scholar, 27Sanjay A. Miyazaki T. Itzstein C. Purev E. Horne W.C. Baron R. Identification and functional characterization of an Src homology domain 3 domain-binding site on Cbl.FEBS J. 2006; 273: 5442-5456Crossref PubMed Scopus (19) Google Scholar, 28Purev E. Neff L. Horne W.C. Baron R. c-Cbl and Cbl-b act redundantly to protect osteoclasts from apoptosis and to displace HDAC6 from beta-tubulin, stabilizing microtubules and podosomes.Mol. Biol. Cell. 2009; 20: 4021-4030Crossref PubMed Scopus (45) Google Scholar). Thus, specificity of SFK binding to c-Cbl or SFK phosphorylation of c-Cbl may explain the molecular uniqueness of c-Src. We therefore determined whether this critical interaction was based upon the specificity of Src SH3 by examining the mechanism by which c-Src interacts and phosphorylates c-Cbl. As we reported earlier (29Yokouchi M. Kondo T. Sanjay A. Houghton A. Yoshimura A. Komiya S. Zhang H. Baron R. Src-catalyzed phosphorylation of c-Cbl leads to the interdependent ubiquitination of both proteins.J. Biol. Chem. 2001; 276: 35185-35193Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar), coexpression of Src and c-Cbl induced c-Cbl and Src degradation, with c-Cbl levels reduced by more than 90% (Fig. 5A, Fig. S6). Fyn also induced degradation of c-Cbl by about 50% when Src was not co-overexpressed (Fig. 5A, Fig. S6). In contrast, other Src family kinases had little or no effect on c-Cbl level (Fig. 5A, Fig. S6). These results suggest that the Src-induced degradation of c-Cbl could contribute to the unique effect of Src on actin organization in osteoclasts. c-Cbl’s PPR domain contains multiple sites that bind to the Src SH3 domain (27Sanjay A. Miyazaki T. Itzstein C. Purev E. Horne W.C. Baron R. Identification and functional characterization of an Src homology domain 3 domain-binding site on Cbl.FEBS J. 2006; 273: 5442-5456Crossref PubMed Scopus (19) Google Scholar, 30Szymkiewicz I. Destaing O. Jurdic P. Dikic I. SH3P2 in complex with Cbl and src.FEBS Lett. 2004; 565: 33-38Crossref PubMed Scopus (22) Google Scholar). However, mutating the PPR motifs does not completely eliminate binding to Src (27Sanjay A. Miyazaki T. Itzstein C. Purev E. Horne W.C. Baron R. Identification and functional characterization of an Src homology domain 3 domain-binding site on Cbl.FEBS J. 2006; 273: 5442-5456Crossref PubMed Scopus (19) Google Scholar), and it has been reported that mutating phosphorylated c-Cbl Tyr residues reduces c-Cbl’s binding to Src and other SFKs (31Feshchenko E.A. Langdon W.Y. Tsygankov A.Y. Fyn, Yes, and Syk phosphorylation sites in c-Cbl map to the same tyrosine residues that become phosphorylated in activated T cells.J. Biol. Chem. 1998; 273: 8323-8331Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar). To better understand the roles of c-Cbl SH3-binding and SH2-binding motifs in mediating c-Cbl’s interaction with c-Src, we made truncated fragments of the C-terminal half of c-Cbl (c-Cbl-CT) that eliminated one or more specific binding motifs (Fig. 5B) and overexpressed these five mutants with or without SrcY527F in 293 cells (Fig. 5C, Fig. S7A). The binding of the fragments to Src domains was evaluated by pulldown assays using glutathione-S-transferase (GST)–Src constructs containing Src’s entire N-terminal half (unique, SH3, and SH2 domains, Fig. 5D, Fig. S7B), SH3 domain (Fig. 5E, Fig. S7C), or SH2 domain (Fig. 5F, Fig. S7, D and E). Consistent with earlier reports, c-Cbl-CT (aa 479–906), F1 (aa 479–791, lacking the UBA domain), and F2 (aa 479–636, lacking the C-terminal phosphorylated Tyr residues and UBA domain) bound to N-terminal Src and Src SH3 in the absence of coexpressed SrcY527F, whereas F3 (aa 637–906, lacking the PPR motifs) and F4 (aa 792–906, lacking both PPR domains and phosphorylated Tyr residues) did not. Coexpressing SrcY527F to catalyze phosphorylation of the c-Cbl Tyrs increased the binding of N-terminal Src to c-Cbl-CT and F1 (and to a lesser degree F2). More importantly, Src-catalyzed phosphorylation induced the binding of N-terminal Src to F3, which lacks the SH3-binding PPR motifs, indicating that phosphorylation of one or more of the c-Cbl Tyrs creates a binding site for Src. Src-catalyzed phosphorylation had little or no effect on the binding of any of the fragments to Src SH3, suggesting that Src-catalyzed phosphorylation of c-Cbl Tyrs promoted binding to the Src SH2 domain. This was confirmed using GST-c-Src SH2 in the pull-down assay (Fig. 5F, Fig. S7E). c-Cbl CT, F1, and F3 were highly phosphorylated bound to GST-c-Src SH2 when SrcY527F was coexpressed, but no binding was observed when c-SrcY527F was absent (Fig. 5F). The binding of phosphorylated F3 was relatively weak, despite its high phosphorylation (compare the binding of CT and F3), suggesting that binding of the SH2 domain to phosphorylated c-Cbl is weak relative to the SH3–proline motif interaction. F2, which lacks the phosphorylated Tyrs, did not bind to Src SH2 whether c-SrcY527F was coexpressed or not. Three phosphorylated tyrosine residues (Tyr700, Tyr731, and Tyr774) have been reported to mediate the binding of multiple proteins to c-Cbl (32Thien C.B.F. Langdon W.Y. c-Cbl and Cbl-b ubiquitin ligases: Substrate diversity and the negative regulation of signalling responses.Biochem. J. 2005; 391: 153-166Crossref PubMed Scopus (194) Google Scholar). We therefore mutated each of these residues to phenylalanine in c-Cbl F3 (Fig. 6A) and examined the effect on the binding of the Src SH2 domain. As expected, neither F3 nor the mutants bound to GST–Src [SH3/SH2] in the absence of coexpressed SrcY527F (Fig. 6B, the left side of the top panel, Fig. S8A). When SrcY527F was coexpressed with the c-Cbl F3 constructs, F3 and F3Y700F were equivalently phosphorylated and bound to GST–Src [SH3/SH2] (Fig. 6B, the right side of the top panel, Fig. S8A). In contrast, F3Y731F and F3Y774F were much significantly less phosphorylated and failed to detectably bind to GST–Src [SH3/SH2] (Fig. 6B second panel, Fig. S8B), suggesting that phosphorylation of both Tyr731 and Tyr774 can enhance c-Cbl binding to c-Src. Next, we sought to determine which domain of Src is involved in binding to phosphorylated c-Cbl F3. GST–Src SH3 domain did not bind to both nonphosphorylated and phosphorylated c-Cbl F3 (Fig. 6C, second panel). On the other hand, GST–Src SH2 bound to F3 and F3Y700F (Fig. 6C, the top panel, Fig. S8D). These results suggest that Tyr731 and Tyr774 can enhance c-Cbl binding to the c-Src SH2 domain. The Src tyrosine kinase plays an essential role in osteoclast bone resorbing activity, promoting the formation of the sealing zone and ruffled border and organization of F-actin into the peripher