Specific association of tyrosine-phosphorylated c-Cbl with Fyn tyrosine kinase in T cells.

Fyn is a Src family protein-tyrosine kinase functionally associated with the T-cell antigen receptor (TcR)/CD3 receptor complex. We have demonstrated earlier that the TcR/CD3-induced activation of Fyn results in tyrosine phosphorylation of several Fyn-associated proteins, including a protein of 116 kDa. In this report, we identify the Fyn-associated 116-kDa phosphoprotein (p116) as c-Cbl. The identity of p116 has been demonstrated by its specific reactivity with anti-Cbl and similarity of phosphopeptides generated by V8 proteolysis of phospho-Cbl and p116. We demonstrate here that the association of Fyn and c-Cbl is direct and does not require the presence of other proteins. We also demonstrate that Fyn is the Src family kinase that preferentially interacts with c-Cbl in T cells. The fraction of c-Cbl capable of coprecipitating with Fyn is increased by TcR/CD3 ligation. This increase is likely due to the involvement of Fyn SH2 in the interactions between Fyn and tyrosine-phosphorylated c-Cbl.

Fyn is a Src family protein-tyrosine kinase expressed in T lymphocytes and functionally associated with the TcR 1 /CD3 receptor complex (reviewed in Refs. [1][2][3]. Fyn appears to play a role in the activation of T cells through the TcR/CD3, based on its physical association with the TcR/CD3 (4 -6) and its enzymatic activation and tyrosine phosphorylation following ligation of the TcR/CD3 (7,8). Physiological responses of T cells to TcR/CD3 stimulation are dramatically affected by targeted disruption of the fyn gene (9,10) or overexpression of its wild-type and dominant-negative forms (11), confirming that Fyn is critically important in TcR-mediated T-cell activation. Lck is another Src-related kinase involved in TcR signaling. Although Lck appears not to be directly associated with the TcR/CD3, the CD4-Lck complex (12) seems to interact with the TcR/CD3, making Lck a crucial element of the T-cell activation pathway (13,14). Fyn and Lck appear to trigger the activation of Zap, a Syk family kinase capable of binding to the cytoplasmic sequences of the TcR/CD3 (15)(16)(17)(18)(19).
Activation of Fyn following TcR/CD3 cross-linking results in tyrosine phosphorylation of several proteins associated with Fyn (7,8). The major Fyn-associated phosphoprotein of 116 kDa (p116) becomes phosphorylated on tyrosine in vivo, as well as in vitro, following TcR/CD3 ligation (20,21). The interactions of Fyn with p116 and the other major Fyn-associated protein, p82, appear to be specific for this Src-related kinase (21,22) and mediated primarily by the SH2 domain of Fyn, although p82 can also bind to Fyn SH3 (21). Another important difference between the two Fyn-associated proteins is that the tyrosine phosphorylation of p116 is entirely TcR/CD3 ligationdependent, whereas the tyrosine phosphorylation of p82 is little affected by TcR/CD3 ligation (21). These results suggest that p116 may represent a specific physiological substrate of Fyn.
In this report, we identify the Fyn-associated p116 as c-Cbl. c-Cbl is a product of the corresponding proto-oncogene exhibiting several distinct features of a transcription factor (23)(24)(25) but whose physiological functions remain to be determined. c-Cbl protein becomes phosphorylated on tyrosine following TcR/CD3 ligation (26) and is capable of binding to Fyn, Grb2, and the p85 subunit of phosphatidylinositol 3Ј-kinase (26 -30). It has not been shown yet whether the 116-kDa protein phosphorylated in vitro in Fyn immune complexes is related to c-Cbl. The specificity and the molecular basis of c-Cbl/Fyn interactions also remain to be elucidated. Here we report that the Fyn-associated p116 protein is recognized by anti-Cbl antibodies and that this recognition is completely and specifically blocked by the c-Cbl antigenic peptide. We also report that the association between Fyn and c-Cbl does not require participation of other proteins and, therefore, is direct. Furthermore, we show in this study that Fyn SH2 is involved in the interactions between Fyn and c-Cbl following the TcR/CD3-induced tyrosine phosphorylation of c-Cbl. Finally, we demonstrate that in T cells, c-Cbl preferentially interacts with Fyn, whereas association between Lck and c-Cbl appears to be insignificant, regardless of whether c-Cbl is phosphorylated.

EXPERIMENTAL PROCEDURES
Antibodies and Peptides-Antisera to glutathione S-transferase (GST) and Src family kinases were described earlier (21). The affinitypurified polyclonal antibody against c-Cbl (C-15), the corresponding antigenic peptide, and PY20 anti-Tyr(P) monoclonal antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). OKT3 anti-CD3⑀ monoclonal antibody has been purified from ascites fluid using protein A-Sepharose. Anti-Cas antisera was kindly provided by Dr. H. Hirai, University of Tokyo. Myeloma proteins used as isotypematched antibody controls were purchased from Sigma. Rabbit IgG against mouse IgG and goat F(abЈ) 2 fragments against mouse IgG were purchased from Cappel (Durham, NC). Irrelevant Sc.A7 peptide (31) was a kind gift of Dr. D. Mosser, Temple University, Philadelphia.
Fusion Proteins-GST fusion proteins containing full-length Src family kinases and their fragments were described earlier (21). The Grb2 open reading frame was cloned from a mouse A20/2J B-cell cDNA library (Stratagene, La Jolla, CA) into pBluescript SKIIϩ vector (Stratagene) using a polymerase chain reaction-based strategy with published Grb2 sequences (32). The full-length Grb2 (amino acids 1-217), N-terminal SH3 (amino acids 5-54), SH2 (amino acids 59 -158), and * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. C-terminal SH3 (amino acids 163-208) domains, as well as the fragments containing SH2 with either N-terminal or C-terminal SH3, were amplified from pBluescript/Grb2 plasmid by DNA amplification reaction and cloned into pGEX-2T vector (Pharmacia Biotech Inc.). The GST-Grb2 fusion proteins were expressed in Escherichia coli and purified using glutathione-Sepharose columns (Pharmacia).
Cells-CEM.3-71 is a CD3 ϩ clone of the human T-cell line CEM. These cells and the way they were cultured and activated were described earlier (21).
In vitro transcription/translation was carried out using the Promega TNT-coupled reticulocyte lysate system and Tran 35 S-label reagent. Human c-Cbl cDNA in pGEM4Z vector was kindly provided by Dr. W. Langdon, The University of Western Australia, Nedlands.
Immunoprecipitation, Immune Complex Kinase Assay, and Immunoblotting-Cells were lysed, proteins were immunoprecipitated with appropriate antibodies, and immune complex kinase assay and immunoblotting were performed as described earlier (12,21,33). Where indicated, immune complexes were dissociated with SDS, and proteins were reprecipitated with the corresponding antibodies or GST-fusion proteins, followed by anti-GST (21). GST-fusion proteins were added to dissociated immune complexes in the molar amount equivalent to 1 g of GST/1 mg of total protein initially used for immunoprecipitation.
Coprecipitation with GST-Fusion Proteins-GST or GST-fusion proteins were added at 1 g of GST or equimolar amounts of GST-fusion proteins/1 mg of total T-cell lysate protein or 3 l of reticulocyte transcription/translation reaction mixture containing 60 ng of c-Cbl pGEM4Z vector. GST-containing proteins were precipitated using either glutathione-Sepharose, according to the manufacturer's recom-mendation, or anti-GST antiserum (5 l/1 mg of GST) as described above.
Phosphopeptide Mapping-Phosphoprotein bands were excised from a gel and cleaved with V8 protease from Staphylococcus aureus (Pierce) as described earlier (34).

Identification of the Fyn-associated 116-kDa Protein as c-
Cbl-We sought to determine the identity of the p116 protein, which we have previously established as associating with and undergoing phosphorylation by Fyn. Among various candidate proteins, our attention has been drawn to c-Cbl, one of the newly identified substrates of tyrosine kinases in T cells (26).
To determine whether p116 is c-Cbl, we immunoprecipitated Fyn from T-cell lysates, performed in vitro kinase assays with these immunoprecipitates, and attempted to re-precipitate p116 using several antibodies including an affinity-purified antibody against c-Cbl. The p116 protein was detected in immunoprecipitates with anti-Cbl, whereas no immunoprecipitation of p116 was observed with either nonimmune control serum or antiserum against Cas, the 130-kDa substrate of Src (35,36) (Fig. 1A, left panel, and B). To confirm the specificity of immunoprecipitation of p116 with anti-Cbl, we used the Cterminal c-Cbl peptide C-15, to which the anti-Cbl antibody was raised, to block this immunoprecipitation. The reactivity of p116 with anti-Cbl was completely blocked by low concentrations of C-15, whereas an irrelevant peptide had no effect, even at a much higher concentration (Fig. 2).
Immunoprecipitation of c-Cbl, followed by an immune complex kinase assay, demonstrated that c-Cbl and several other proteins, including a 60-kDa protein, become phosphorylated on tyrosine in c-Cbl immunoprecipitates (Fig. 1A, right panel). This indicates that a protein-tyrosine kinase is present in c-Cbl immune complexes. Considerable fractions of the 116-kDa and the p60-kDa phosphoproteins initially immunoprecipitated with anti-Cbl were re-immunoprecipitated with anti-Fyn but not with control immunoglobulins (Fig. 1A, right panel) or anti-Lck (data not shown). This result argues that Fyn is present in c-Cbl immune complexes and is capable, to a certain extent, of reassociating with c-Cbl following dissociation of their complex.
We also compared phosphopeptides generated by V8 proteolysis of the 116-kDa proteins phosphorylated in Fyn and c-Cbl immune complexes and found out that these peptides were essentially identical (Fig. 3). This result further indicates that the Fyn-associated p116 is c-Cbl. Furthermore, the pattern of phosphopeptides generated by V8 cleavage of the c-Cbl-associated 60-kDa phosphoprotein is essentially identical to the pattern of phosphopeptides generated by V8 cleavage of Fyn (data not shown), arguing that these two proteins are identical.
To provide further evidence of association between Fyn and c-Cbl, we labeled CEM.3-71 cells with 35 S-labeled amino acids and compared the pattern of anti-Cbl-immunoprecipitated 35 Slabeled proteins to the proteins phosphorylated in vitro in Fyn immune complexes from the same cells. Two major 35 S-labeled proteins specifically immunoprecipitated with anti-Cbl appeared to comigrate with phosphoproteins of Fyn immune complexes corresponding to Fyn and the Fyn-associated p116 (Fig.  4). No band comigrating with p56 lck was found specifically associated with c-Cbl in these experiments. Furthermore, we metabolically labeled CEM.3-71 cells with [ 32 P]orthophosphate, activated them by CD3 cross-linking or left unstimulated, and immunoprecipitated c-Cbl and Fyn from these cells. The first immunoprecipitation was followed by re-immunoprecipitation with antibodies to c-Cbl, Fyn, or Tyr(P). These experiments have demonstrated that Fyn is associated with a 116-kDa phosphoprotein comigrating with a 116-kDa protein double-immunoprecipitated with anti-Cbl (Fig. 5). Both of these 116-kDa protein bands are phosphorylated on tyrosine, as evidenced by their reactivity with anti-Tyr(P) antibody ( capable of binding to Grb2, the adaptor protein linking tyrosine protein kinases to the system of Ras activity regulation (26, 28 -30, 37). We determined whether the Fyn-associated p116 protein displays this property as well. This was performed by adding GST-Grb2 to the lysates of TcR/CD3-stimulated and unstimulated CEM.3-71 cells and examining the Grb2-bound proteins. A 116-kDa Grb2-associated protein, which exhibited TcR/CD3 stimulation-dependent tyrosine phosphorylation, comigrated with p116 (data not shown). This Grb2-associated 116-kDa tyrosine-phosphorylated protein has been identified earlier as c-Cbl (29,30). To further characterize the p116-Grb2 interaction, we subjected Fyn immune complexes from activated CEM.3-71 cells to kinase assays in vitro and assessed the association of 32 P-labeled phosphoproteins with GST fusions of Lck, Fyn, Grb2, and Grb2 fragments (Fig. 6). The Fyn-associated p116 was re-precipitated with full-length Fyn and Grb2. However, of the Grb2 fragments tested, only SH2SH3C exhibited the ability to bind to p116. It thus appears that a substantial part of the Grb2 protein is required for this binding. Furthermore, the interaction of p116 with full-length Lck was negligible.
Interactions of c-Cbl with Isolated SH3 Domains-Taken together, our results indicate that the p116 protein, coimmunoprecipitated with Fyn from T-cell lysates and phosphorylated in vitro in Fyn immune complexes, is c-Cbl. However, recent evidence indicates that c-Cbl can interact with a wide range of SH3 domains, including those of Grb2 and various Src family tyrosine kinases (26, 29, 30, 38 -40). In contrast, the Fyn-associated p116 appears not to bind to isolated SH3 domains of either Src family kinases (20,21) or Grb2 (Fig. 6). To determine whether c-Cbl is capable of interacting with isolated SH3 domains in our experimental system, we immunoprecipitated c-Cbl from CEM.3-71 cells, dissociated c-Cbl immune complexes with SDS, diluted the obtained supernatants with Tris/Triton X-100 buffer, and re-precipitated c-Cbl with GSTfusion proteins containing full-length and partial sequences of Fyn, Lck, and Grb2. The results of these experiments demonstrate that c-Cbl does not appreciably interact with GST-fusion proteins containing the SH3 domains of either Fyn, Lck, or Grb2, while interacting with full-length Fyn and Grb2 (Fig. 7  and data not shown). To rule out the trivial possibility that these results are due to the effect of SDS upon the structure of c-Cbl, we expressed c-Cbl cDNA in reticulocyte lysate and tested the obtained preparation of c-Cbl for its ability to bind to recombinant Fyn, Lck, and Grb2. It appears that in vitro translated c-Cbl, which has not been treated with SDS, is incapable of binding to any of the isolated SH3 domains tested in our system, while displaying binding to full-length Fyn and Grb2 and, to a very low extent, to full-length Lck (Fig. 8 and data not  shown). Therefore, c-Cbl does not interact with isolated SH3 domains in our experimental system, regardless of whether it underwent a denaturing procedure. The observed lack of c-Cbl binding to SH3 is apparently due to lower concentrations of c-Cbl and/or GST-fusion proteins in our binding assays, resulting in higher stringency of our experimental conditions toward low-affinity interactions as compared to the experiments where these interactions were detected.
The Effect of Tyrosine Phosphorylation of c-Cbl on Its Interactions with Src Family Kinases and Grb2-To determine whether the tyrosine phosphorylation of c-Cbl affects its binding to Src family kinases and Grb2, we obtained c-Cbl from unstimulated and CD3-stimulated CEM.3-71 cells by immunoprecipitation with subsequent SDS-elution/renaturation and demonstrated that c-Cbl binds to Fyn and Grb2 but not to Lck (Fig. 7). Furthermore, it appears that c-Cbl from stimulated T cells binds to GST-Fyn better that c-Cbl from unstimulated T cells, whereas the binding of c-Cbl to Grb2 does not change following CD3 stimulation (Fig. 7). The observed increase in association between c-Cbl and Fyn is likely to be caused by the interaction between the Tyr(P) residue(s) of c-Cbl and Fyn SH2, because unphosphorylated c-Cbl does not bind to Fyn SH2, whereas the TcR/CD3-induced tyrosine phosphorylation of c-Cbl renders it capable of binding to this SH2 domain (Figs. 7 and 8). It is important to note that c-Cbl is only partially tyrosine phosphorylated following CD3 cross-linking, as judged by its immunoprecipitation with anti-Cbl and anti-Tyr(P) (Fig.  5 and data not shown). Therefore, the difference between binding characteristics of unphosphorylated c-Cbl and its fully tyrosine phosphorylated form is likely to be more pronounced than the observed difference between the preparations of c-Cbl from unstimulated and CD3-stimulated T cells. DISCUSSION The results reported in this study indicate that the previously described Fyn-associated 116-kDa protein (p116) is a tyrosine-phosphorylated form of c-Cbl. We initially demon- strated association of Fyn and p116 cbl using the immune complex kinase assay of Fyn immunoprecipitates (7, 21) (Fig. 1). In addition, we showed coimmunoprecipitation of Fyn and c-Cbl from the lysates of T cells labeled with [ 35 S]methionine/cysteine or [ 32 P]orthophosphate in vivo (Figs. 4 and 5). We also demonstrated that recombinant full-length Fyn binds to c-Cbl (Figs. 6 -8). The interaction between Fyn and c-Cbl appears to be direct, based on their coprecipitation in the absence of additional proteins (Figs. 7 and 8).
Taken together with the data published earlier, our results suggest that several domains of Fyn are involved in the interaction with c-Cbl. It has been demonstrated previously that c-Cbl is capable of interacting with SH3 domains of Fyn and other Src-related kinases (26, 29, 39 -41). Although we observed no binding of c-Cbl to isolated SH3 domains, apparently due to the high stringency of our experimental system (see Figs. 6 -8), the binding of unphosphorylated c-Cbl to Fyn (Figs. 7 and 8) is consistent with the involvement of Fyn SH3 domains. It appears that the c-Cbl/Fyn complex exists in unstimulated T cells, where it is likely stabilized by SH3-mediated interactions (Fig. 4). However, the tyrosine phosphoryla-tion of c-Cbl increases its ability to bind to Fyn (Fig. 7). This increase is apparently caused by the interaction between Tyr(P) residue(s) of c-Cbl and Fyn SH2 domain and is observed once c-Cbl becomes tyrosine phosphorylated following TcR/CD3 ligation (Figs. 7 and 8). Hence, the TcR/CD3-induced tyrosine phosphorylation of c-Cbl is likely to promote formation of the complex between Fyn and c-Cbl in T cells. Although the TcR/ CD3-dependent tyrosine phosphorylation of c-Cbl does not necessarily increase the total amount of Fyn/c-Cbl complex, it may effectively change the type of association between c-Cbl and Fyn from primarily SH3-dependent in unstimulated T cells to primarily SH2-dependent in stimulated T cells. This alteration may render the Fyn SH3 domain and c-Cbl proline-rich regions available for interactions with other proteins in activated T cells.
Interestingly, the fraction of c-Cbl found in association with Fyn in CEM.3-71 cells is relatively minor (Fig. 5 and data not shown). This finding likely reflects the fact that Fyn expression in T cells is substantially lower than that of c-Cbl (7, 28) (data not shown). Furthermore, the affinity of c-Cbl/Fyn interactions may not be sufficiently high to cause quantitative binding, because the fraction of Fyn-associated c-Cbl does not exceed 20%, even in the presence of an excess of recombinant Fyn ( Fig.  7 and data not shown). In addition, we should note that coimmunoprecipitation of c-Cbl with Fyn from T-cell lysates may not exactly reflect the extent of their association in vivo, where intracellular conditions and/or compartmentalization may favor this association.
The involvement of a SH2 domain in the interaction between Fyn and c-Cbl raises the question of specificity of their association, because numerous proteins contain SH2 domains. However, we have demonstrated previously that in spite of the ability of several Src family SH2 domains to bind to p116, tyrosine-phosphorylated p116 is found only in Fyn, but not in Lck or Yes, immune complexes from T-cell lysates (21). The present study clearly demonstrates that both unphosphorylated and tyrosine-phosphorylated forms of c-Cbl exhibit preferential binding to Fyn, as compared to Lck (Figs. 1, 4, and 6 -8 and data not shown), indicating that Fyn is a specific Src family kinase that interacts with c-Cbl in T cells. This specificity, taken together with the ability of Fyn to phosphorylate c-Cbl in immune complexes in vitro, strongly argues that Fyn is capable of phosphorylating c-Cbl in T cells. However, the possibility that other protein kinases are involved in phosphorylation of c-Cbl in vivo cannot be ruled out at the moment.
It is evident that the adaptor protein Grb2 binds to c-Cbl in the cells analyzed (Figs. 6 -8). It is important that Grb2 and Fyn differ with regard to the mechanism of their association with c-Cbl. Although the Fyn SH2 domain appears to be critical for c-Cbl/Fyn interactions in activated CEM.3-71 cells, binding of Grb2 SH2 to c-Cbl has not been observed, regardless of whether c-Cbl is phosphorylated (Figs. 6 -8). Furthermore, Tyr(P) does not affect binding of Grb2 to c-Cbl (data not shown), consistent with the idea that this binding is not mediated by SH2-Tyr(P) interactions. These results are in agreement with the earlier reports indicating that c-Cbl primarily interacts with Grb2 SH3 domains (26, 28 -30, 37). In accordance with this mechanism, the ability of c-Cbl to form a complex with Grb2 is not affected by TcR/CD3 ligation (Fig. 7).
The physiological role of c-Cbl/Fyn interactions is still unclear and remains to be elucidated. The oncogenic potential of certain truncations and deletion mutants of c-Cbl (23)(24)(25) indicates that this protein may play a major role in cell activation. Tyrosine phosphorylation of c-Cbl following ligation of the TcR/CD3 (26), the B-cell receptor (38), and the Fc␥ receptor (39,41), taken together with the fact that certain oncogenic deletions of c-Cbl are highly phosphorylated on tyrosine (25), lends support to the idea that the receptor-induced tyrosine phosphorylation of c-Cbl is an important event in cell activation. Fyn is thought to play a critical role in the TcR/CD3-induced stimulation (7)(8)(9)(10)(11)19). Therefore, it is tempting to speculate that c-Cbl is a Fyn-specific adaptor protein linking this tyrosine kinase to Grb2 and thus to the system of Ras activation. However, the involvement of the SH3 domains of Grb2 in its interactions with both c-Cbl and the Ras GTP/GDP-exchange factor, mSOS (28 -30), raises doubts regarding the feasibility of a physical link between Fyn and Grb2/mSOS via c-Cbl. Failure to detect a ternary complex of c-Cbl, Grb2, and mSOS (28 -30) also argues against this possibility. However, the interactions between Fyn, c-Cbl, and Grb2 may have functions different from the regulation of Ras. Based on the finding that Grb2 SH2 is not likely to be involved in binding to Fyn or c-Cbl (Figs. 6 and 7 and data not shown; see also Refs. 26,[28][29][30]and 37), this SH2 domain may link c-Cbl-associated Grb2 to another tyrosine-phosphorylated protein. It is also possible that Fynassociated c-Cbl binds to an adaptor protein other than Grb2, such as c-Crk or CrkL (42)(43)(44). Furthermore, it has been re-ported recently that c-Cbl is associated with Zap tyrosine kinase in TcR/CD3-stimulated T cells (45). It is interesting that, unlike Src family kinases, Zap appears to be incapable of tyrosine phosphorylating c-Cbl, at least in COS cells. These results suggest that c-Cbl may function as a docking platform linking Zap to Ras activation and/or as an adaptor enhancing the ability of Fyn to phosphorylate Zap. However, further analysis is required to characterize the mechanism and the physiological functions of c-Cbl/Fyn interactions, as well as the involvement of other proteins in these interactions.