In vivo association of v-Abl with Shc mediated by a non-phosphotyrosine-dependent SH2 interaction.

A necessary downstream element of Abelson murine leukemia virus (Ab-MLV)-mediated transformation is Ras, which can be activated by the phosphotyrosine-dependent association of Shc with the Grb2-mSos complex. Here we show that Shc is tyrosine-phosphorylated and associates with Grb2 in v-Abl-transformed cells, whereas Shc in NIH3T3 cells is phosphorylated solely on serine and is not Grb2-associated. In addition, Shc coprecipitates with P120 v-Abl and P70 v-Abl, which lacks the carboxyl terminus. Surprisingly, a kinase-defective mutant of P120 also binds Shc, demonstrating that Shc/v-Abl association is a phosphotyrosine-independent interaction. Glutathione S-transferase fusion proteins were used to map the interacting domains and showed that Shc from both NIH3T3 and v-Abl-transformed cells binds to the Abl SH2 domain and that P120 v-Abl binds to a region in the amino terminus of Shc. Consistent with these data, a v-Abl mutant encoding only the Gag and SH2 regions was able to bind Shc in vivo. The unique non-phosphotyrosine-mediated binding of Shc may allow direct tyrosine phosphorylation of Shc by v-Abl and subsequent activation of the Ras pathway through assembly of a signaling complex with Grb2-mSos.

elucidated. The Bcr-Abl protein may signal Ras via direct interaction with the Grb2 protein (7), a molecule that binds the guanine nucleotide exchange factor, mSos (8,9). Localization of the Grb2-mSos complex to the plasma membrane activates Ras (9,10). Bcr-Abl-Grb2 interaction involves sequences in the Bcr domain (7,11,12). This pathway is probably not used by v-Abl because the v-Abl protein does not contain a motif similar to that involved in Bcr-Abl-Grb2 interaction.
Interactions between Shc and receptor tyrosine kinases occur when the receptors are activated by interaction with ligand; the autophosphorylation sites on the receptors provide binding sites for the Shc SH2 or amino-terminal domains (16, 24 -26). Subsequent tyrosine phosphorylation of the YVNV motif in Shc by the receptors or their associated kinases creates a binding site for the Grb2 SH2 domain. Shc then associates with Grb2 which brings mSos into the complex (27)(28)(29). Such an interaction may be involved in factor independent growth of v-Ablexpressing mast cells (30). However, association with Grb2 and evidence of stimulation of signals downstream of Ras have not been reported in that system.
Because v-Abl lacks the Grb-2 binding site found in Bcr-Abl, we investigated the possibility that Shc protein associates with v-Abl and Grb-2 in v-Abl-transformed cells. These experiments demonstrate that Shc is tyrosine-phosphorylated in Ab-MLVtransformed fibroblasts and binds to Grb2. This interaction involves an unusual non-phosphotyrosine-dependent interaction that occurs between the v-Abl SH2 domain and the Shc amino terminus. The ability of Shc to bind Grb2 and the v-Abl SH2 simultaneously suggests a model for Shc-mediated Ras activation in Ab-MLV transformation similar to that of receptor tyrosine kinases.

EXPERIMENTAL PROCEDURES
Cells and Viruses-NIH3T3 fibroblasts and v-Abl-transformed NIH3T3 cell lines were grown in DMEM (Life Technologies, Inc.) supplemented to contain 10% Cosmic Calf serum (HyClone), 2 mM Lglutamine, and 50 g/ml gentamycin (Life Technologies, Inc.). The NIH3T3 cell lines ANN-1, transformed with Ab-MLV-P120 strain (31), and 70wt, transformed with Ab-MLV-P70 (32), were described previously. Virus was obtained either by transfection of NIH3T3 cells using the calcium phosphate technique (33) followed by superinfection with Moloney murine leukemia virus (Mo-MLV) (34) or by transfection of 293T cells as described previously (35). In some cases, NIH3T3 cells were infected and grown in 0.8 mg/ml G418 (Life Technologies, Inc.) to select for cells expressing the neomycin resistance marker within the virus.
v-Abl Mutants-The P120 v-Abl kinase domain mutant (P120k Ϫ ) was created by site-directed mutagenesis in M13 phage using the modified method of Zoller and Smith described in Ref. 33. A mismatched primer replacing a G with an A at position 2070 within the kinase domain sequence of the Ab-MLV-P120 genome (36) resulted in a substitution of Asn for Asp at residue 484 of the P120 protein. The presence of the mutation was verified by dideoxy method sequencing with the Sequenase kit (Promega), and sequence containing the mutation was cloned into pANL2, a vector based on pABLSVneo (37). This vector is composed of the Ab-MLV P120 genome with a neomycin resistance gene controlled by an SV40 promoter cloned into the BamHI site at position 4355. These sequences are flanked by two complete long terminal repeats. The GagSH2 v-abl mutant was constructed by amplifying bases 621-1685 of the Ab-MLV P120 genome by PCR from pUC120 (32). The 5Ј primer included an EcoRI site and the 5Ј end of the Gag-coding region. The 3Ј primer included the 3Ј end of the SH2 coding sequence, a termination codon, and an EcoRI site. The PCR product was cloned into the pSR␣ vector (35) at the EcoRI site, and the amplified sequence was verified by dideoxy method sequencing.
Antibodies-The monoclonal antibodies, H548, directed against p12 Gag determinants present in v-Abl ((38); a gift of B. Chesebro, Rocky Mountain Laboratories, Hamilton, MT), 24 -21, directed against an epitope in the v-Abl carboxyl terminus (39), and 19 -84, directed against a v-Abl SH1 domain epitope (39), were used to detect v-Abl proteins. Anti-phosphotyrosine monoclonal antibodies were purchased from Upstate Biochemical Institute. Anti-Shc monoclonal and polyclonal antibodies as well as anti-Grb2 monoclonal antibodies were purchased from Transduction Laboratories. Anti-GST monoclonal antibodies were obtained from Santa Cruz Biotechnology. Isotype-matched monoclonal control antibody, UPC10, and RGG were generous gifts of H. Wortis (Tufts University, Boston, MA). Anti-mouse and anti-rabbit IgG alkaline phosphatase-conjugated antibodies were purchased from Promega. Anti-mouse horseradish peroxidase-conjugated antibody was obtained from Amersham.
GST Fusion Proteins-The GST fusion proteins were all constructed in the pGEX-3X vector (Pharmacia Biotech Inc.) and PCR primers for the inserts contained either a BamHI or an EcoRI linker. For the GST-ablSH2SH1 construct, PCR was performed to amplify bases 1329 -2441 of v-abl P120 using pUC120. To create the GST-Abl SH1 and GST-Abl SH2 constructs, the v-abl sequences, bases 1686 -2441 and bases 1329 -1685, respectively, were amplified from pUC120. The GST-Src SH2 was constructed by amplifying bases coding for residues 146 -251 of v-Src from pGX-4 (40). The GST-Arg SH2 construct was a gift of G. Kruh (Fox Chase Cancer Center, Philadelphia, PA), and the GST-Shc construct was a kind gift of B. Schaffhausen (Tufts University, Boston, MA). The Shc SH2 region (bases 1214 -1504 of the shc cDNA (16)), the collagen homology (CH) region of Shc (bases 777-1213), and the region coding for the amino terminus (NT) of Shc (bases 82-776) were amplified from the GST-Shc construct by PCR. The NT1 and NT2 coding regions (bases 82-340 and bases 82-586, respectively) were also amplified by PCR using the GST-Shc construct. Each PCR product was digested with BamHI and/or EcoRI and ligated into the appropriate site of pGEX-3X. The sequence of all PCR products was confirmed by dideoxy method sequencing.
To prepare GST fusion proteins, log phase Escherichia coli JM109 cells containing the pGEX-3X constructs were grown for 3 h in the presence of 50 g/ml ampicillin and 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside. The cells were pelleted and lysed in ice cold RIPA buffer (10 mM sodium phosphate, pH 7.0, 150 mM sodium chloride, 0.1% SDS, 1% Nonidet-40, 0.5% sodium deoxycholate, 2 mM EDTA, 50 mM sodium fluoride, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and 10 g/ml leupeptin (Boehringer Mannheim)) with 10 g/ml lysozyme. After 25 min on ice, the samples were sonicated (10 ϫ 1 s) and centrifuged at 14,000 rpm for 15 min. Glutathione-Sepharose beads (Pharmacia) were added to the bacterial lysates, and the samples were incubated for 1.5 h on a rotating wheel at 4°C. The beads were centrifuged and washed with RIPA buffer. To obtain purified protein cleaved from the GST moiety, the fusion proteins bound to GST beads were washed and resuspended in 50 mM Tris, pH 8.0, 150 mM NaCl, and 1 mM CaCl 2 followed by incubation with 1 unit of Factor Xa (U. S. Biochemical Corp.).
Immunoprecipitation and Western Blotting-Cells were harvested and washed once with phosphate-buffered saline. The cells were then lysed in RIPA buffer and left on ice for 25 min. The lysates were centrifuged at 14,000 rpm for 15 min, and protein concentration was determined using the BCA method kit (Pierce). Immunoprecipitations were done by incubating equivalent amounts of total cell protein with 1 g of purified antibody or serum on ice for 1 h. Protein A-or protein G-Sepharose beads (Pharmacia) were added, and the lysates were placed on a rotating wheel at 4°C for 1.5 h. The beads were washed with IP wash buffer (10 mM sodium phosphate, 150 mM sodium chloride, 1% Nonidet P-40, 2 mM EDTA, 1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride). For precipitation with GST fusion proteins, GST fusion protein/glutathione beads as described above were added to equivalent amounts of total cell protein and incubated at 4°C on a rotating wheel for 1.5 h. The beads were washed twice with IP wash buffer and once with RIPA buffer. Bound proteins were eluted by heating the beads at 95°C in SDS sample buffer for 5 min. The samples were then fractionated through a 10% SDS-polyacrylamide gel, and the proteins were transferred to a polyvinylidene difluoride membrane (U. S. Biochemical Corp.) with a Bio-Rad Transblot apparatus (Bio-Rad) according to the supplied protocol. The membrane was then blocked with phosphate-buffered saline containing 4% bovine serum albumin and 0.1% Tween 20 for at least 1 h. Western blotting was either done according to the Western-Light kit protocol (Tropix), utilizing alkaline phosphatase-conjugated secondary antibodies with a CSPD substrate (Tropix) or with the ECL detection system using horseradish peroxidase-conjugated secondary antibodies. Blots were then exposed to Kodak XAR 5 film. Blots were stripped by incubating in a solution containing 0.2 M glycine and 1% Tween 20 (pH 2.2) for 2.5 h at 80°C. After stripping, blots were washed with phosphate-buffered saline, 0.1% Tween 20, and reblocked prior to probing.
Phosphoamino Acid Analysis-Cells were phosphate-starved in phosphate-free DMEM (Flow Laboratories) supplemented with L-glutamine and 5% dialyzed fetal calf serum (HyClone) for 1 h. The cells were then incubated overnight in phosphate-free DMEM with L-glutamine, 5% dialyzed fetal calf serum, and 0.5 mCi/ml 32 P-labeled carrier-free orthophosphate (DuPont NEN). The cells were harvested, washed, and lysed in 1 ϫ PLB buffer (100 mM sodium phosphate, pH 7.5, 100 mM sodium chloride, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM sodium orthovanadate, and 1 mM phenylmethylsulfonyl fluoride). The lysates were centrifuged for 15 min at 14,000 rpm and transferred to fresh tubes. Antibody was added to the lysates, which were incubated on ice for 1 h. Protein A-Sepharose beads were added and the samples were placed on a rotating wheel at 4°C for 1.5 h. The samples were washed four times in 1 ϫ PLB, resuspended in SDS sample buffer, and eluted from the beads by heating at 95°C for 5 min. The samples were then subjected to SDS-PAGE and transferred to polyvinylidene difluoride membrane as described above. Autoradiography was used to locate the Shc bands on the membrane which were subsequently excised and used for one-dimensional phosphoamino acid analysis as described previously (41). The 32 P-labeled amino acids were visualized on a Molecular Dynamics PhosphorImager.

Shc Is Tyrosine-phosphorylated and Associates with Grb2 in
Ab-MLV-transformed Cells-To determine whether Shc is tyrosine-phosphorylated in Ab-MLV-transformed cells, lysates from ANN-1 cells, a P120 v-Abl-transformed NIH3T3 cell line, and uninfected NIH3T3 cells were precipitated with anti-Shc antibody and analyzed on a Western blot with anti-phosphotyrosine antibody (Fig. 1B). Shc protein recovered from ANN-1 cell lysates was highly phosphorylated while that recovered from NIH3T3 cells was not tyrosine-phosphorylated. The electrophoretic mobility of Shc differs between NIH3T3 and ANN-1 cell lysates (Fig. 1A). Other higher molecular weight phosphoproteins were also observed coprecipitating with Shc from ANN-1 cells (Fig. 1B). Shc from NIH3T3 cells migrates as three distinct bands, but Shc from ANN-1 cells migrates as several subspecies. A similar broadening of Shc bands has been observed after the addition of PKC activators (42). One-dimensional phosphoamino acid analysis of immunoprecipitated Shc from [ 32 P]orthophosphate-labeled cells (Fig. 2, A and B) showed the presence of phosphotyrosine and phosphoserine in ANN-1derived Shc and only phosphoserine in Shc from NIH3T3 cells. In addition, an increase in the total amount of phosphoserine was also observed in Shc from ANN-1 cells.
To determine if the tyrosine-phosphorylated Shc was able to bind Grb2, the lower portion of the blot shown in Fig. 1A was probed with anti-Grb2 antibody (Fig. 1C). Shc from ANN-1 cells binds Grb2, in contrast to Shc precipitated from NIH3T3 cells which does not. Therefore, the presence of v-Abl causes the association of Shc and Grb2, most likely through the phosphorylation of Tyr-317 of Shc. This pattern is consistent with Shc-mediated Ras activation as seen with other tyrosine kinases (16,18,20,29).
Shc Associates with v-Abl-In view of the modified state of Shc in P120 v-Abl-transformed cells, we examined whether v-Abl binds Shc in vivo. Western blots of immunoprecipitated P120 v-Abl probed with anti-Shc antibody showed that Shc binds to P120 (Fig. 3B). Conversely, when immunoprecipitations of Shc from ANN-1 cells were probed on a Western blot with anti-Abl antibody, P120 was recovered from the immune complexes (Fig. 3D). Immunodepletion experiments revealed that approximately 5% of the total Shc or v-Abl in the cell are complexed together (data not shown). This is consistent with the amount of Shc associated with a membrane-bound receptor complex during physiologic activation in at least one system (43).
To determine if the carboxyl terminus of v-Abl is required for association with Shc, lysate from cells transformed with the Ab-MLV P70, which encodes a v-Abl protein containing the Gag, SH2, and SH1 domains (32), was immunoprecipitated with anti-Abl antibodies and probed with anti-Shc antibody (Fig. 3B). P70, like P120, was found to bind Shc. Immunoprecipitation of Shc also coprecipitated P70 (Fig. 3D). Shc in P70 v-Abl-transformed cells is also tyrosine-phosphorylated (data not shown) and coprecipitates with Grb2 (Fig. 3E). Therefore, the presence of the v-Abl carboxyl terminus is not essential for interaction with Shc or Shc phosphorylation.
Most interactions of Shc with tyrosine kinases have shown that binding of Shc requires a tyrosine-phosphorylated motif (14,17,24,25,28). To assess the contribution of v-Abl kinase activity to v-Abl/Shc association, NIH3T3 cells expressing the mutant Ab-MLV-P120k Ϫ , which encodes a kinase-defective P120 v-Abl, were utilized. P120k Ϫ contains an Asp to Asn amino acid change in the kinase domain rendering it kinasedefective. Cells expressing this protein do not contain increased levels of phosphotyrosine (Fig. 4, A and C) and immunoprecipitated P120k Ϫ has no detectable kinase activity in vitro (data not shown). P120k Ϫ was immunoprecipitated from cells ex- P]orthophosphate in phosphate-free DMEM and 5% dialyzed fetal calf serum. Cells were lysed, immunoprecipitated with ␣-Shc antibody, and the precipitates were used for SDS-PAGE. A, the gel was transferred to a polyvinylidene difluoride membrane, and autoradiography was performed. B, the autoradiogram and membrane were aligned, and the portion containing both the p46 Shc and p52 Shc bands was excised for phosphoamino acid analysis as described under "Experimental Procedures." A PhosphorImager was used to examine the thin layer plate. The migration points of the cold phosphoamino acid standards and inorganic phosphate (P i ) are indicated.

FIG. 3. In vivo association of v-Abl and Grb2 with Shc. 70wt
(P70) and ANN-1 (P120) cells were lysed and immunoprecipitated for Western analysis as described under "Experimental Procedures." A, v-Abl was precipitated with ␣-Abl antibody from 70wt and ANN-1 cells, and the Western blot was probed with ␣-Abl antibody. P70 appears as a doublet due to a glycosylated isoform (32). B, the blot from A was stripped and reprobed with ␣-Shc antibody. C, Shc was immunoprecipitated with ␣-Shc antibody from 70wt and ANN-1 cells. The Western blot of the precipitates was probed with ␣-Shc antibody. D, the blot in C was stripped and probed with ␣-Abl. E, the lower portion of the C blot was probed with ␣-Grb2 antibody. ␣Abl, H548 antibody; control (A and B), UPC10 antibody, C-E, RGG. Shc, v-Abl, and Grb2 bands are indicated by arrows. * indicates Ig heavy chain. pressing the mutant virus and analyzed on Western blots for Shc coprecipitation (Fig. 4, B and D). Immunoprecipitation of P120k Ϫ showed that Shc still bound to v-Abl. Therefore, Shc binding is independent of v-Abl kinase activity.
v-Abl-Shc Interaction Involves the v-Abl SH2 Domain-To identify the domains by which v-Abl associates with Shc, GST fusions of the Abl SH1 and SH2 domains were made and bound to glutathione-Sepharose beads to precipitate cell lysates for Western analysis (Fig. 5A). Shc was precipitated from both NIH3T3 and ANN-1 cell lysates by GST-Abl SH2 and by a GST fusion consisting of both the Abl SH2 and SH1 domains. Shc did not precipitate with the GST-Abl SH1 fusion protein. Therefore, the Abl SH2 domain alone is sufficient for Shc binding. Furthermore, the ability of the Abl SH2 domain to bind Shc from NIH3T3 cell lysates shows that neither v-Abl kinase activity nor Shc tyrosine phosphorylation is required for this interaction. The amounts of Shc recovered with the GST-Abl SH2SH1 fusion are somewhat lower than might be expected based on the results obtained with the GST-Abl SH2. Because the GST-Abl SH2SH1 protein is not kinase-active in vitro (data not shown), this difference is not mediated by phosphorylation of Shc by the fusion protein. However, the presence of SH1 sequences may alter the interaction between Abl SH2 and Shc in some fashion. Although the amount of Shc binding to the GST-Abl SH2 differs between NIH3T3 and ANN-1 cell lysates (Fig. 5A), this difference is less prominent in additional experiments (data not shown). GST fusion precipitation experiments also showed that a GST-Arg (Abl-related gene) SH2 domain binds Shc as well (Fig. 5E). The Arg SH2 domain differs from Abl by only 10 residues. However, consistent with previously published results (44), a GST-Src SH2 fusion did not bind Shc under our conditions (Fig. 5E). This suggests that the in vitro binding of Shc in our system is not a general property of SH2 domains.
A mutant Ab-MLV was made to test whether the Abl SH2 domain is sufficient for interaction in vivo. The GagSH2 virus, which consists of the Ab-MLV Gag-coding region and the Abl SH2-coding region followed by a stop codon, was expressed in NIH3T3 cells. Immunoprecipitates of the GagSH2 protein were analyzed by Western blotting with anti-Shc antibodies, revealing that Shc coprecipitated with the GagSH2 protein (Fig. 5, C  and D). The Abl SH2 domain is therefore sufficient for binding of Shc both in vivo and in vitro. To examine whether the binding is direct or requires an intermediary for complex formation, the ability of the Abl SH2 to bind purified Shc was tested. Affinity-purified GST-Shc was treated with Factor Xa to release a full-length Shc protein. The GST-Abl SH2 was able to precipitate the Shc protein as observed by Western analysis with anti-Shc antibodies (Fig. 5F) despite the bacterial origin of both proteins.
Western blot analysis of GST-Abl SH2 precipitates from NIH3T3 and ANN-1 cells with anti-Grb2 antibody showed that Grb2 was precipitated with the GST-Abl SH2 in ANN-1 lysates but not in NIH3T3 lysates (Fig. 5B). This result presumably reflects the association of Grb2 with tyrosine-phosphorylated Shc in ANN-1 cells. In NIH3T3 cells, Shc and Grb2 are not associated, and, therefore, Grb2 would not be precipitated by the GST-Abl SH2 protein. Accordingly, these data suggest that binding of Shc to the Abl SH2 domain does not interfere with binding of Grb2 to Shc at tyrosine 317 and that Shc possesses independent binding sites for Grb2 and v-Abl. The potential for a v-Abl-Shc-Grb2 complex to form in Ab-MLV-transformed cells is suggestive of the Shc-mediated Ras activation described for receptor protein tyrosine kinases.
v-Abl Binds to the Amino Terminus of Shc-GST fusion proteins of the three major domains of the Shc protein, the amino terminus (NT; residues 1-232 of p52 Shc]), the collagenhomology domain (CH; residues 233-377), and the Shc SH2 FIG. 4. Coimmunoprecipitation of 120k ؊ and Shc. NIH3T3 cells expressing P120k Ϫ (P120k Ϫ ) and cells expressing wild type P120 (ANN-1) were lysed, and immunoprecipitates were analyzed by Western blotting. A, equivalent amounts of total cell lysate from P120k Ϫ and ANN-1 cells were analyzed by Western blotting with ␣-phosphotyrosine antibody. B, P120k Ϫ and wt P120 were precipitated from P120k Ϫ and ANN-1 cells, respectively, with ␣-Abl antibody. Western analysis was performed using ␣-Abl antibody. C, the blot in B was stripped and reprobed with ␣-phosphotyrosine antibody. D, the blot was stripped again and probed with ␣-Shc antibody. ␣-Abl, H548 antibody; control, UPC10.
FIG. 5. Mapping the v-Abl binding site for Shc. A, GST fusion proteins of v-Abl domains were used to precipitate lysates from NIH3T3 and ANN-1 cells as described under "Experimental Procedures." Precipitates were analyzed by Western blot with ␣-Shc antibody. Equivalent amounts of total cell lysates are included on the far left. Stripping and reprobing of the blot with ␣-GST antibody revealed equivalent amounts of GST-SH2SH1 and GST-SH1, but twice as much GST-SH2 (data not shown). B, the lower portion of A was probed with ␣-Grb2 antibody. C, GagSH2 was immunoprecipitated from NIH3T3 cells with ␣-Abl antibody and analyzed on a Western blot probed with ␣-Abl antibody. Pr65 is the pr65 Gag precursor protein encoded by Mo-MLV. ␣Abl, H548 antibody. D, the blot from C was stripped and reprobed with ␣-Shc antibody. Upon longer exposure, p46 Shc and p66 Shc are also visible (not shown). E, ANN-1 cell lysates were precipitated with equivalent amounts of the indicated GST fusion proteins and analyzed by Western blotting with ␣-Shc antibody. F, full-length Shc (0.5 g) cleaved from bacterial GST fusion protein was precipitated with either GST or GST-Abl SH2 fusion proteins and analyzed on a Western blot with ␣-Shc antibody. domain (SH2; residues 378 -473), were used to identify the region which binds v-Abl. The GST fusion proteins were bound to glutathione-Sepharose beads and used to precipitate ANN-1 cell lysates (Fig. 6A). The precipitates were analyzed on Western blots with anti-Abl antibodies to evaluate v-Abl association. Binding was observed only between v-Abl and the GST-Shc NT protein, demonstrating that the amino terminus of Shc is sufficient for interaction with v-Abl. A reciprocal direct binding experiment similar to Fig. 5F was performed with the GST-Shc NT protein and purified Abl SH2SH1 domain and showed a direct interaction as well (Fig. 6C). Further mapping of the region of the Shc amino terminus essential for v-Abl binding was accomplished by using GST-Shc NT1 and NT2, which contain the Shc residues 1-85 and 1-168, respectively. Both of these proteins precipitated v-Abl from ANN-1 cell lysates (Fig.  6B). The amount of v-Abl precipitated was proportional to the amount of GST fusion protein on the beads (data not shown). These binding results establish that Shc and v-Abl interact via the Abl SH2 and the first 85 residues of Shc. DISCUSSION Our results show that Shc is tyrosine-phosphorylated and associated with Grb2 in v-Abl-transformed cells. This association could lead to activation of Ras, an event required for v-Abl-mediated transformation. In receptor activation models, association of Shc with an activated receptor enables the receptor or a related kinase to phosphorylate Tyr-317, the Grb2 binding site on Shc (28). This modification allows the Shc-Grb2-mSos complex to assemble at the membrane in proximity to Ras (9,10,29). The binding we have documented may allow v-Abl or an associated kinase to directly phosphorylate Shc, stimulating the formation of a Shc-Grb2-Sos complex. The Gag domain of v-Abl may also provide a plasma membrane-localized platform on which the Shc-Grb2-Sos complex can assemble. Localization of this complex is a critical event for mSos function because addition of a membrane localization signal to Sos is in itself sufficient to activate Ras (10).
Some tyrosine kinases can bypass Shc and activate Ras by binding the Grb2-mSos complex directly through the Grb2 SH2 domain (8,9,27). Bcr-Abl accomplishes this by binding Grb2 to a phosphotyrosine-containing motif within the Bcr portion of the protein (7,11,12). However, v-Abl lacks a similar Grb2 binding site. Another possible site for direct Grb2 interaction with v-Abl is located within the carboxyl terminus (45). The Grb2 SH3 domains bind a proline-rich motif in the Abl carboxyl terminus in vitro. The carboxyl terminus of v-Abl also contains proline-rich motifs that bind Crk and Nck (45,46), proteins that can associate with the Ras activators, C3G and mSos (47)(48)(49). However, the efficient transformation of NIH3T3 cells by P70 (32), which lacks these proline-rich motifs, precludes any essential role for these interactions. Tyrosine phosphorylation of Shc and its association with Grb2 could allow Ras signaling in the absence of the Nck, Crk, and Grb2 binding sites.
The v-Abl/Shc interaction we have documented occurs in a phosphotyrosine independent manner and contrasts to the conventional view of SH2-mediated interactions which involve recognition of a phosphotyrosine residue in the context of three residues carboxyl-terminal to the tyrosine (50). Tyrosine phosphorylation of v-Abl is not required because both the P120k Ϫ and the GagSH2 proteins bind Shc. In addition, the GST-Abl SH2 fusion binds Shc and the GST-Shc NT binds Abl SH2SH1 in vitro; these proteins are produced in E. coli and do not contain phosphotyrosine (51,52). The possibility that tyrosine phosphorylations on Shc in vivo mediate the interaction is excluded because Shc from NIH3T3 cells binds the GST-Abl SH2 fusion and phosphoamino acid analysis of Shc recovered from those cells shows no phosphotyrosine. Shc also binds to P120k Ϫ and GagSH2 despite the absence of tyrosine phosphorylation observed on Western analysis (data not shown).
Recently, Owen-Lynch et al. (30) described the interaction of Shc with a v-Abl protein encoded by a temperature-sensitive (ts) strain of Ab-MLV at both the permissive and nonpermissive temperature. Previous descriptions of the ts strain used note the persistence of phosphotyrosine on v-Abl protein at the nonpermissive temperature (53). Our own experiments with another ts Ab-MLV yielded similar data (not shown). The persistence of phosphotyrosine at the nonpermissive temperature suggests that these ts systems may not be practical for assessing the phosphotyrosine independence of protein-protein interactions.
Although many SH2-mediated interactions involve phosphotyrosine, the number observed that occur independent of this modification is growing. Phosphotyrosine-independent association between the Abl SH2 and Bcr has been observed (54,55). Two serine-rich regions of Bcr are important for this binding, but the specific motifs involved have not been identified. SH2 domains from other proteins including phospholipase C␥, Src, and GTPase activating protein also bind to Bcr in this manner (54). The SH2 domain of the protein-tyrosine phosphatase SH-PTP2 binds its own catalytic domain in the absence of phosphotyrosine (56). Another instance of phosphotyrosine-independent binding involves interactions of the Src and Fyn SH2 FIG. 6. Mapping the Shc binding site for v-Abl. A and B, GST fusions of different regions of Shc were used to precipitate ANN-1 cell lysates as described above. Western analysis of the precipitates was done using ␣-Abl antibody. Approximately equal amounts of GST fusion proteins were used in each precipitation with the exception of GST-NT1 and NT2 which had 30 -40% less fusion protein than the GST-NT precipitation (data not shown). ␣-Abl, 24 -21 antibody. C, GST and GST-Shc NT were used to precipitate 0.5 g of Abl SH2SH1 domain cleaved from the bacterially generated fusion as described under "Experimental Procedures." The precipitates were analyzed by Western blotting. ␣Abl, 19 -84. domains with Raf, a protein which is phosphorylated on serine residues (57). It is possible that the Abl/Shc interaction is mediated by phosphoserine; however, it has not been determined whether the GST fusion proteins are phosphorylated correctly. Serine and threonine phosphorylation has been observed on bacterially produced GST fusion proteins (51).
The interaction of v-Abl with the GST-Shc NT1 protein indicates residues 1-85 of p52 Shc are sufficient for binding. p46 Shc, which also binds v-Abl, possesses only residues 46 -85 of that region. Therefore, the essential binding site must lie within the shared sequence of both p52 and p46. The region is also of interest because it contains the recently described phosphotyrosine binding domain (58). This domain binds to NPXpY motifs, including those found on polyoma middle T, EGFR, and Trk (23,25). However, this property of the Shc amino terminus is not involved in v-Abl binding because the interaction is phosphotyrosine independent. Additionally, the GST-Shc NT1 and NT2 fusion proteins precipitate v-Abl, whereas similar GST fusions cannot bind NPXpY targets without the minimal 46 -238 amino acid region (59).
Another interaction observed in the Shc amino terminus is the protein kinase C-dependent binding of the PEST-phosphatase to p52 Shc (42). In addition to inducing Shc/Grb2 association, the v-Abl interaction could also regulate the Shc signaling complex. The PEST-phosphatase binds to Shc but does not alter tyrosine phosphorylation of Shc or its association with Grb2 (42). Presumably, the PEST-phosphatase dephosphorylates other proteins in the Shc signaling complex. By binding to Shc, v-Abl could phosphorylate additional proteins in the complex. Furthermore, because the v-Abl and c-Abl SH2 domains are identical, Shc may be involved in cytoplasmic interactions that are important for normal c-Abl function.