Shc and a novel 89-kDa component couple to the Grb2-Sos complex in fibroblast growth factor-2-stimulated cells.

A major pathway for mitogenicity is gated via the small GTP-binding protein Ras. Receptor tyrosine kinases couple to Ras through the Src homology 2 (SH2) domain protein Grb2. The activated fibroblast growth factor receptor-1 (FGFR-1) expressed in L6 myoblasts did not bind Grb2 directly, but indirectly, through the small adaptor protein Shc, which was tyrosine-phosphorylated in response to fibroblast growth factor-2 (FGF-2) stimulation. A FGFR-1 mutant in which Tyr, a known autophosphorylation site, was changed to Phe, mediated less efficient tyrosine phosphorylation of Shc. FGF-2 stimulation of mutant FGFR-1-expressing cells still allowed formation of complexes containing Shc, Grb2, and the nucleotide exchange factor Sos and mediation of a mitogenic signal. Another pool of Grb2 was found in complex with a tyrosine-phosphorylated 89-kDa component after FGF-2 stimulation. Stimulation with other growth factors did not lead to tyrosine phosphorylation of p89. As shown by “far-Western” analysis, p89 bound directly to the Grb2 SH2 domain, and this interaction was inhibited by a peptide containing the Y(P)-X-N motif. Tyrosine-phosphorylated p89 was found exclusively in the membrane fraction, indicating its role in bringing Grb2, as well as Sos, to the plasma membrane. These data support the concept of growth factor-specific coupling of Grb2 to the Ras pathway.

Fibroblast growth factors (FGFs) 1 comprise a family of heparin-binding polypeptides that, to date, include nine distinct gene products, FGF-1 to FGF-9 (1-3). FGFs are mitogenic for a wide variety of cells and are also implicated in regulation of differentiation, cell motility, and transformation. In vivo studies describe a regulatory role of FGFs in physiological processes such as neovascularization, wound healing, and development (4 -6).
FGFs bind extracellularly to high affinity transmembrane tyrosine kinase receptors and low affinity heparan sulfate proteoglycan co-receptors. The high affinity FGF receptors form a family of four structurally related members, FGFR-1 to FGFR-4 (7,8). Each of these contain two or three extracellular immunoglobulin-like loops, a characteristic stretch of acidic amino acids between the first and second loop, a single transmembrane region, and an intracellular kinase domain split by a 14 -17-amino acid kinase insert. Alternative splicing generates a multitude of structural variants that differ in ectodomain regions known to be involved in ligand binding (8).
Autophosphorylated tyrosine residues and adjacent amino acids provide specific binding sites for signal transducing proteins containing Src homology 2 (SH2) domains (9). Recruitment of signal-transducing molecules to activated tyrosine kinase receptors triggers signaling pathways, which eventually result in specific cellular responses. This far, two intracellular autophosphorylation sites have been identified in the FGFR-1, at Tyr 653 and Tyr 766 , located in the kinase domain and the C-terminal tail, respectively (10). Whereas it is possible that Tyr 653 is involved in regulation of kinase activity, phosphorylated Tyr 766 has been shown to provide a binding site for phospholipase C-␥ (PLC-␥) (11). The Y766F mutant FGFR-1 has been shown to be internalized at a decreased rate in transfected rat myoblasts (12), but still mediates a mitogenic signal (13,14) and allows neurite outgrowth when expressed in PC-12 cells (15). Thus the role of PLC-␥ in FGFR-1 signaling is not well understood; however, PLC-␥-mediated activation of PKC in FGF-stimulated cells could play a negative regulatory role, i.e. in Src-dependent downstream signaling (16). A number of other signal transduction molecules have been shown not to interact with or be affected by the activated FGFR-1. These include phosphatidylinositol 3Ј-kinase (17) and RasGAP (18).
Activation of FGFR-1 leads to a potent mitogenic signal. Conversion of the Ras protein from its inactive GDP-bound state to its active GTP-bound state appears to be a key event in intracellular signaling from several activated tyrosine kinase receptors to the nucleus. The small adaptor proteins Grb2 and Shc provide a link between growth factor receptors and Ras, via the nucleotide exchange factor Sos. Grb2 has been shown to interact directly with growth factor receptors, but also indirectly, via Shc or the tyrosine phosphatase PTP1D/Syp (19). A number of other proteins are described to interact with the Grb2 SH2 domain, such as BCR-ABL (20), focal adhesion kinase (FAK) (21), and a novel 36 -38-kDa protein in T-lymphocytes (22).
In this work we have characterized coupling of FGFR-1 to Ras via Grb2 and Shc. Whereas Grb2 did not interact directly with the FGFR-1, an indirect interaction via the adaptor protein Shc was evident. In addition, a separate pool of Grb2 was in complex with a tyrosine-phosphorylated component of 89 kDa in FGF-2 stimulated cells. We propose a role for this novel component in FGFR-1-mediated Ras activation.

MATERIALS AND METHODS
Cell Culture and Transfection-Rat L6 myoblasts were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum, 100 units/ml of penicillin, and 100 g/ml of streptomycin. The wild type and mutated FGFR-1 were ex-* 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.
‡ To whom correspondence should be addressed. Tel.: 46-18-174146; Fax: 46-18-506867. 1 The abbreviations used are: FGF, fibroblast growth factor; FGFR, fibroblast growth factor receptor; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; Grb2, growth factor receptor bound protein 2; GST, glutathione S-transferase; MAPK, mitogen-activated protein kinase; MEK, MAP kinase (or ERK) kinase; PBS, phosphatebuffered saline; PDGF, platelet derived growth factor; PLC-␥, phospholipase C-␥; PAGE, polyacrylamide gel electrophoresis; SH2, Src homology 2; Tyr(P), phosphotyrosine; Shc, Src homologous and collagen; Sos, son of sevenless. pressed in L6 cells after insertion of the cDNA into the expression vector pcDNA I/neo (Invitrogen), and transfection followed by selection with geneticin G418 sulfate (Life Technologies, Inc.). For serum starvation, subconfluent cells were incubated in medium containing 1% fetal calf serum for 18 h. FGF-2 was purchased from Farmitalia Carlo Erba, Milano, Italy. Platelet-derived growth factor BB (PDGF-BB) was a kind gift from Dr. C.-H. Heldin, and epidermal growth factor (EGF) was a kind gift from Amgen Inc., Thousand Oaks, CA. FGF-2, PDGF-BB, and EGF treatment was done by incubation with 100 ng/ml growth factor for 10 min at 37°C. The LIBE cell line is a capillary endothelial cell line established from H-2K b -tsA58 mouse brain and which exhibits temperature-sensitive features. 2 The LIBE cells are highly differentiated at both permissive and nonpermissive conditions. Antisera-A synthetic peptide based on the 16 C-terminal amino acids of the human FGFR-1 sequence was used for immunization of rabbits. Rabbit antibodies against glutathione S-transferase (GST) were raised through immunization with purified GST. Polyclonal antibodies against Sos and Syp were purchased from Upstate Biotechnology, Inc., as were polyclonal and monoclonal antibodies against Grb2. Antibodies specific for phosphotyrosine (PY20) and Shc were from Transduction Laboratories, and a polyclonal antibody against Raf1 was from Santa Cruz Biotechnology, Inc. The rabbit antiserum against PLC-␥ was a kind gift from Dr G. Carpenter, Vanderbilt University, Nashville, TN.
Immunoprecipitation-Following stimulation, culture medium was removed and cells rinsed with ice-cold phosphate-buffered saline (PBS) containing 100 M Na 3 VO 4 . Nonidet P-40 lysis buffer (20 mM Hepes, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Nonidet P-40, 100 M Na 3 VO 4 , 1% aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol) was added and cells lysed for 10 min on ice before beeing scraped and transferred to Eppendorf tubes. Dithiothreitol was excluded from the lysis buffer when the experiment was done under nonreducing conditions. Cell debris was removed by centrifugation. Supernatants were transferred to new tubes and antibody was added. The samples were mixed end-over-end at 4°C for 2 h before addition of protein A-Sepharose (Pharmacia Biotech Inc.) and a final 45-min mixing. The precipitates were washed twice in Nonidet P-40 lysis buffer and two times in PBS containing 100 M Na 3 VO 4 . The samples were subjected to SDS-polyacrylamide gel electrophoresis before electroblotting to Hybond-C extra membranes (Amersham).
In Vitro Association of GST Fusion Protein-The SH2 domain of Grb2 was expressed as a part of a GST fusion protein (23) and was a kind gift of Dr. J. Schlessinger, New York University Medical Center, New York. For control experiments, GST alone was expressed from plasmid pGEX-3X. Transfected L6 cells or LIBE cells were treated with or without growth factor and then lysed in Nonidet P-40 lysis buffer. Clarified lysates were incubated with purified immobilized GST or GST-Grb2 SH2 fusion protein on glutathione Sepharose 4B (Pharmacia) (60-l bead volume with 6 g of bound fusion protein) end-over-end for 2 h at 4°C. For peptide competition experiments, peptides at a final concentration of 50 M were incubated with the lysates for 45 min before addition of fusion protein beads. Samples were washed as described above and analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting.
Receptor Peptide Affinity Chromatography-The amino acid sequences of the synthetic peptides used in this study, either nonphosphorylated or phosphorylated on tyrosine (indicated as Y(P)), are as follows: from FGFR-1, GVSEY(463)ELPEDPRWELPR; PPGLEY(583)-CY(585)NPSHNPE; LTSNQEY(766)LDLSMPLD; LDQY(776)SPSFP-DTRSS and from PDGFR-␤, LQHHSDKRRPPSAELY(716)SNALPVG. Peptides were synthesized using Fmoc (9-fluorenylmethoxycarbonyl) chemistry, followed by plasma desorption mass spectrometric analysis as described before (24). The tyrosine-phosphorylated EGFR peptide comprising Tyr 1068 (PVPEY(1068)INQSVPK) was a kind gift from Dr. J. Downward, Signal Transduction Laboratory, Imperial Cancer Research Fund, London, United Kingdom. The synthetic peptides were coupled to AH Sepharose (Pharmacia) mainly according to the manufacturer's protocol. Briefly, the powdered AH Sepharose was suspended and washed in 0.5 M NaCl and then equlibrated in H 2 O, pH 4.5. The peptides were dissolved in H 2 O, pH 4.5, and added to the beads, whereafter N-ethyl-NЈ-(3-dimethylaminopropyl)carbodiimide hydrochloride was added. During vigorous mixing for 1 h, the pH was maintained at 4.5-6.0. The coupling reaction continued overnight, end-over-end at 4°C. Excess groups were blocked by incubating the beads with 1 M acetic acid. The beads were then washed with 0.1 M Tris-HCl, pH 8.0, 0.5 M NaCl, and 0.1 M Tris-HCl, pH 4.0, 0.5 M NaCl alternatively, three times each, before beeing equilibrated in Triton X-100 buffer (1% Triton X-100, 2 mM EDTA, 1% aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 14.4 mM ␤-mercaptoethanol in PBS). 5 g of fusion protein was incubated end-over-end with 0.125 mol of receptor peptide/100-l beads for 2 h at 4°C. After washing the beads seven times in Triton X-100 buffer, the samples were subjected to SDS-polyacrylamide gel electrophoresis and subsequent electroblotting to Hybond-C extra membranes.
Immunoblot Analysis-The membranes were blocked overnight at 4°C in PBS-T (0.2% Tween 20 in PBS) containing 5% bovine serum albumin. Primary antibodies were diluted in PBS containing 0.05% Tween 20 and 3% bovine serum albumin and incubated with the membranes for 1 h at room temperature. After four 15-min washes in PBS-T, the membranes were incubated for 1 h with the appropriate secondary antibody diluted in PBS containing 0.05% Tween 20 and 3% bovine serum albumin. After several 10-min washes in PBS-T, the bands were visualized by a chemiluminescent detection system based on a protocol described by Matthews et al. (25). Before reprobing the filters, they were stripped in 62.5 mM Tris-HCl, pH 6.7, 2% SDS containing 100 mM ␤-mercaptoethanol at 50°C for 30 min.
Subcellular Fractionation-After growth factor treatment, the cells were washed two times with ice-cold PBS containing 100 M Na 3 VO 4 before beeing scraped and pelleted in PBS at 1000 rpm. The cells were resuspended in a small volume of PBS and homogenized 20 times using a Wheaton tissue grinder. The nuclear fraction was pelleted by centrifugation at 15,000 rpm for 10 s. The supernatant was centrifuged at 160,000 ϫ g in an Airfuge (Beckman) to pellet the membrane fraction. The cytosolic and membrane fractions were adjusted to the same volume with Nonidet P-40 lysis buffer before incubation with 6 g of GST-Grb2 SH2 fusion protein/60-l beads as described above.
Raf1 in Vitro Kinase Assay-Stimulated and nonstimulated cells from T-75 tissue culture flasks were lysed in Nonidet P-40 lysis buffer. Cleared lysates were incubated with 1. DNA Synthesis-Cells were inoculated into 24-well plates at a density of 1.5 ϫ 10 4 cells/cm 2 in medium containing 10% fetal calf serum. The next day, the medium was removed, and fresh medium containing 0.5% fetal calf serum was added and the culture continued for another FIG. 1. Grb2 SH2 fusion protein does not interact with intact FGFR-1. Lysates from stimulated (ϩ) or unstimulated (Ϫ) L6 cells transfected with wild type FGFR-1 were incubated with Grb2 SH2 fusion protein coupled to glutathione-Sepharose. Immunoprecipitation of the receptor was included as a control. The subsequent filter was immunoblotted with FGFR-1 antiserum. Prec., precipitation; Ib, immunoblotting.
24 h. At this point, fresh medium containing 0.5% fetal calf serum and different concentrations of growth factor (FGF-2) was added. The cells were incubated for 20 h and then 1 Ci/ml [ 125 I]iododeoxyuridine was added and incubation continued for 4 h. After aspiration of medium, cells were fixed in 10% trichloroacetic acid and lysed with 1 M NaOH. Incorporation of 125 I-labeled deoxyuridine into DNA was measured using a ␥ spectrometer.

Direct and Indirect
Interactions between Grb2 and FGFR-1-We were interested in identifying mechanisms for coupling of FGFR-1 to Grb2. We could find no indication for binding of a GST fusion protein comprising the Grb2 SH2 domain to the intact receptor (Fig. 1), in an immunoblotting analysis of FGF-2 stimulated FGFR-1 expressing L6 myoblasts. This is in agreement with the lack of the known consensus binding site for Grb2, Y(P)-X-N (26), in the FGFR-1 amino acid sequence. Since direct binding of Grb2 to the intact FGFR-1 did not occur at detectable levels, we investigated the possibilities for indirect interactions. The SH2 domain-containing tyrosine phosphatase PTP1D/Syp has been described to mediate indirect interaction between Grb2 and the PDGF ␤-receptor (27). We failed, however, to detect association in vitro and in vivo between the activated FGFR-1 and Syp or tyrosine phosphorylation of Syp in response to FGF-2 (data not shown; see Fig. 8).
In contrast, Shc, which is known to anchor Grb2 to receptor tyrosine kinases, was tyrosine-phosphorylated in response to FGF-2 treatment of FGFR-1 expressing L6 myoblasts ( Fig. 2A). In L6 cells expressing a mutant FGFR-1, Y766F, Shc was still phosphorylated, although to a lesser extent. The reduction in extent of phosphorylation of Shc was not due to impaired kinase activity of the mutant, as compared with the wild type receptor (Fig. 2B). Analysis of Shc SH2 domain fusion protein interactions in vitro with immobilized FGFR-1 synthetic peptides, phosphorylated or not, showed that the peptide containing phosphorylated Tyr 766 provided the most efficient binding site (Fig. 2C). The mitogenic signaling capacities of the wild type FGFR-1, the mutant Y766F, and in addition, the mutant Y463F, were compared. The incorporation of 125 I-labeled deoxyuridine in serum-starved, FGF-2-stimulated L6 cells expressing wild type or mutant (Y463F or Y766F) receptors showed that both mutants were less efficient in stimulating DNA synthesis (Fig. 2D). Thus, although Tyr 766 appears to be involved in coupling FGFR-1 to mitogenic signal transduction, other pathways for mitogenicity must exist, depending e.g. on Tyr 463 for their initiation. We were, furthermore, not able to demonstrate direct interaction between the receptor and Shc in vivo (Fig. 3A). Immunoprecipitation using Shc antiserum fol- FIG. 2. A, tyrosine phosphorylation of Shc after FGF-2 stimulation. L6 cells expressing wild type and mutated FGFR-1 were incubated in the presence (ϩ) or absence (Ϫ) of 100 ng/ml FGF-2 for 10 min, lysed, and immunoprecipitated with Shc antiserum. Samples were separated by SDS-PAGE and immunoblotted with Tyr(P) antibodies. As determined by densitometric scanning, tyrosine phosphorylation of Shc proteins was reduced in cells expressing the Y766F mutant FGFR-1, to approximately 30% of the level detected in cells expressing wild type and Y463F mutant FGFR-1. B, phosphorylation of wild type and mutated FGFR-1 in FGF-2-stimulated cells. Immunoprecipitation using FGFR-1 antiserum was performed on cell lysates from FGF-2 stimulated (ϩ) or unstimulated (Ϫ) L6 transfectants. The samples were separated by SDS-PAGE and immunoblotted with Tyr(P) antiserum, as indicated. C, binding of Shc SH2 domain fusion protein to FGFR-1 synthetic peptides. 5 g of GST-Shc SH2 was incubated with a series of nonphosphorylated or phosphorylated FGFR-1 synthetic peptides coupled to AH Sepharose 4B. Material bound to the beads was separated by SDS-PAGE and immunoblotted with GST antiserum. Ip, immunoprecipitation; Ib, immunoblotting. D, mitogenic capacity of wild type and mutated FGFR-1. Serum-starved L6 cells expressing wild type (triangles), mutated Y463F (circles), or Y766F (squares) were stimulated with different concentrations of FGF-2 in the presence of 125 I-labeled deoxyuridine. The extent of incorporation of radiolabeled iododeoxyuridine was assessed using a ␥-counter and plotted as fold increase as compared with unstimulated cells of each type. lowed by immunoblotting using phosphotyrosine antibody showed instead that Shc associated with a tyrosine-phosphorylated 145-kDa component after FGF-2 stimulation (Fig. 3B). This component was not recognized by antisera against the receptor, Sos, or PLC-␥ (data not shown) and may correspond to the 145-kDa component described by Kavanaugh and Williams (28).
Shc Binds Grb2 after FGF-2 Stimulation-Although we failed to identify in vivo interactions between FGFR-1 and Shc, tyrosine-phosphorylated Shc was able to bind Grb2 in vivo in FGF-2 stimulated L6 cells expressing wild type FGFR-1 or the mutants Y463F and Y766F (Fig. 4A). As expected, Grb2 was found in constitutive complex with the nucleotide exchange factor Sos (Fig. 4B) (29). We could also identify in vivo complex formation between Shc and Sos (Fig. 4C), which in contrast was dependent on FGF-2 stimulation. That FGF-2 stimulation ini- tiated signal transduction leading to activation of Ras, and further to activation of the serine/threonine kinase Raf1, was shown by in vitro phosphorylation of kinase-inactivated MEK, which was mediated to similar extents by the wild type and mutant receptors (Fig. 4D). MEK in turn activates MAPK (30). The reduced mobility of Sos after FGF-2 stimulation (Fig. 4E), could be due to MAPK-catalyzed phosphorylation of Sos on serine and threonine residues. A main role of activated MAPK appears, however, to be to further transduce the Ras activity signal to the nuclear compartment (31).
Grb2 Interacts with Components Other than Shc after FGF-2 Stimulation-To identify other possible routes for signal transduction between FGFR-1 and Ras, we analyzed components interacting with Grb2 in vivo by immunoprecipitation using Grb2 antiserum, followed by immunoblotting using phosphotyrosine antibody. As expected (see Fig. 1) we could not detect a band corresponding to FGFR-1 in this analysis. On the other hand, an 89-kDa component was found to be tyrosine-phosphorylated and in complex with Grb2 after FGF-2 stimulation of FGFR-1 expressing L6 cells (Fig. 5A). In a time course analysis, the peak of phosphorylation of the 89-kDa component occurred at 30 min of stimulation, followed by a sustained phase that lasted for several hours (Fig. 5A). At 30 min, but not at later time points, an intense band corresponding to p52 shc , was also detected. The 89-kDa component was retained after FGF-2 stimulation by the Grb2 SH2 fusion protein coupled to glutathione-Sepharose (Fig. 5B). This interaction was specific, since no binding was seen to GST alone. The 89-kDa component was visualized as a doublet in Tyr(P) immunoblots, perhaps due to heterogeneity in phosphotyrosine content. The migration rate of p89 was not dependent on whether the samples were treated with reducing agent, or not, before electrophoresis (data not shown).
Grb2 SH2 Binds Directly to Tyrosine-phosphorylated p89 -In a competition experiment, using a panel of synthetic peptides, none of the five FGFR-1 peptides competed appreciably with the 89-kDa component for binding to Grb2 SH2 fusion protein (Fig. 6A). A peptide corresponding to the known Grb2 binding site in the EGF receptor, Y(P)1068 (26), abolished binding of the 89-kDa component to the Grb2 SH2 fusion protein. Two other components binding to Grb2 SH2, which in subsequent immunoblotting were shown to correspond to p52 shc and p66 shc , were also specifically competed out by the EGFR Y(P)1068 peptide. In addition, this peptide abolished the binding of a 40-kDa component to Grb2 SH2. These data show that an 89-kDa component binds Grb2 SH2 in vivo and in vitro by exposing a Grb2 consensus binding site Y(P)-X-N, in which the tyrosine residue is phosphorylated in response to FGF-2 stimulation. In a "far-Western" analysis, the Grb2 SH2 fusion protein bound to the 89-kDa component immobilized on a nitrocellulose filter, indicative of a direct interaction between the two proteins (Fig. 6B). It is possible that there is also direct interaction between FGFR-1 and the 89-kDa component. Several tyrosine-phosphorylated components of this size are found in complex with the receptor (Fig. 6C). However, multimeric complexes containing the FGFR-1, the 89-kDa component, and Grb2 could not be demonstrated (see Fig. 5A).
Tyrosine-phosphorylated p89 Is Localized in the Membrane-To determine the subcellular localization of p89, cells were divided into membrane and cytosolic fractions, which were solubilized, reacted with Grb2 SH2 fusion protein, and samples analyzed by phosphotyrosine antibody immunoblotting. The tyrosine-phosphorylated 89-kDa component was located exclusively in the membrane fraction after FGF-2 treatment (Fig. 7A). The Shc proteins on the other hand were predominantly located in the cytoplasm of FGF-2-stimulated cells, as determined by subsequent immunoblotting using Shc antiserum. The membrane fraction contained, in addition to the 89-kDa component, the 40-kDa component, which was shown above to bind to the Grb SH2 fusion protein (compare Fig. 6A). Ras is anchored in the plasma membrane, and it has been suggested that translocation of Sos to this compartment allows a more efficient stimulation of nucleotide exchange on Ras (32). We therefore examined complex formation between p89 and Sos and found that FGF-2 stimulation allowed this interaction to occur (Fig. 7B).
The 89-kDa Component Is Tyrosine-phosphorylated Only after FGF-2 Stimulation-We compared different growth factors for their capacity to induce tyrosine phosphorylation of the 89-kDa, Grb2 SH2-binding component. For this purpose, we used murine brain endothelial cells that endogenously express receptors for EGF, PDGF-BB, and FGF-2. As seen in Fig. 8A, Grb2 SH2 bound the 89-kDa component only after FGF-2 stimulation, but not after stimulation with EGF or PDGF-BB. In the PDGF-BB-stimulated sample, a component corresponding to PTP1D/Syp was retained by the Grb2 SH2 fusion protein.
Immunoblotting with Syp antiserum failed to reveal binding of Syp to Grb2 SH2 in the FGF-2-or EGF-stimulated samples. In addition, tyrosine-phosphorylated p52 shc was bound to Grb2 SH2, irrespective of which stimulus was used and also to some extent before growth factor stimulation. These data show that tyrosine phosphorylation of the 89-kDa protein, and hence binding to Grb2 SH2, was specific for FGF-2 stimulation.

DISCUSSION
In this work, we have characterized proteins interacting with Grb2 in FGF-2-stimulated cells. Growth factor receptors are thought to couple to Ras via Grb2 in several ways; directly, by binding of the Grb2 SH2 domain to a specific motif in the receptor, or indirectly, via binding to different adaptor proteins. Thus far, two such adaptor molecules have been thoroughly characterized, Shc (33) and PTP 1D/Syp (27). We could not detect direct interaction between FGFR-1 and Grb2, which is consistent with the lack of consensus binding site for Grb2, Y(P)-X-N, in the FGFR-1 sequence (34). We further showed that PTP 1D/Syp is not tyrosine-phosporylated in FGF-2-stimulated cells. PTP 1D/Syp is in complex with and tyrosine phosphorylated by the PDGF ␤-receptor in response to PDGF-BB, but this seems to be of consequence for Ras activation only in certain cell types (35).
In agreement with previous reports, we detected tyrosinephosphorylated Shc proteins in FGF-2-stimulated FGFR-1-ex-pressing cells (15,36,37). We failed to identify in vivo complex formation between the FGFR-1 and Shc, but tyrosine phosphorylation of Shc proteins was to some extent dependent on Tyr 766 in FGFR-1. This is the only known binding site in the receptor for SH2 domain-containing proteins, and it is required for binding of phospholipase C␥ (13,14). The sequence surrounding Tyr 766 is compatible with the known consensus sequence for Shc binding (Y-hydrophobic-X-hydrophobic) in the PDGF ␤-receptor (38). A mutant FGFR-1 in which Tyr 766 was changed to Phe still mediated tyrosine phosphorylation of Shc, indicating that also other sites in the receptor are important or that association of Shc with the receptor is not required for this adaptor protein to become phosphorylated. Binding of Shc to the PDGF ␤-receptor has been shown to involve several autophosphorylated tyrosine residues, which bind Shc with low affinity (38). Moreover, tyrosine phosphorylation of Shc in response to EGF is largely independent of individual autophosphorylation sites and occurs even when several or all autophosphorylation sites have been removed from the receptor (39,40). Recent data indicate that Shc is critically involved in Ras activation (41), although a function for Shc in other pathways cannot be excluded. The Y766F mutant FGFR-1 expressed in FIG. 6. Direct interaction between Grb2 SH2 domain and p89. A, L6 cells expressing wild type FGFR-1 were incubated in the presence (ϩ) or absence (Ϫ) of 100 ng/ml FGF-2 for 10 min and lysed. The lysates were preincubated with or without synthetic FGFR-1 peptides and an EGFR peptide comprising the known Grb2 binding site, Y(P)1068, for 45 min before addition of glutathione-Sepharose bound Grb2 SH2 fusion protein and a continued 2-h incubation. After washing, samples were separated by SDS-PAGE and immunoblotted with Tyr(P) antibody to show that only the Y(P)1068 peptide was able to compete for complex formation between Grb2 and p89. B, to show direct binding of Grb2 SH2 to p89, a far-Western analysis was performed. Cells were treated as above and lysates incubated with Grb2 SH2 fusion protein bound to glutathione-Sepharose. Samples were separated by SDS-PAGE and transferred to nitrocellulose filter. The filter was incubated with 0.1 g/ml Grb2 SH2 fusion protein for 1 h before immunoblotting with GST antiserum. C, analysis of tyrosine-phosphorylated proteins associated with FGFR-1 after FGF-2 stimulation was performed by treating cells with FGF-2 as described above and immunoprecipitation with FGFR-1 antiserum. Samples were subjected to SDS-PAGE and immunoblotting with Tyr(P) antibody. Ip, immunoprecipitation; Ib, immunoblotting.
L6 myoblasts was indeed less efficient in mediating a mitogenic signal. This could be due to impaired tyrosine phosphorylation of Shc, but also to the lack of binding and activation of PLC-␥ in cells expressing the Y766F receptor. The role of PLC-␥ in mitogenic signaling is unclear, and previous reports have shown intact mitogenic capacity of the Y766F mutant receptor (13,14). That several parallel pathways for mitogenicity exists is indicated by the fact that an Y463F mutant FGFR-1, expressed in L6 cells, also was impaired in mitogenic signaling (see Fig. 2D).
In two different cell types, L6 myoblasts expressing FGFR-1 after transfection, and murine brain endothelial cells, we found in vivo complexes containing Grb2 and an 89-kDa protein. This protein was tyrosine-phosphorylated after FGF-2 stimulation, but not after EGF or PDGF-BB stimulation. Tyrosine phosphorylation apparently creates a consensus binding site for Grb2 SH2, which, as shown by far-Western analysis, allowed direct complex formation between the 89-kDa component and Grb2.
Interestingly, cell fractionation showed that the 89-kDa component was localized exclusively in the membrane fraction in FGF-2-stimulated cells, whereas Shc proteins remained in the cytoplasm. Ras activation has been shown to depend on translocation of the Grb2-Sos complex to the plasma membrane, to allow interaction with Ras (32). A role for the 89-kDa component in Ras activation was indicated by the presence of 89-kDa protein-Sos complexes in vivo in FGF-stimulated cells. This pool appeared to be separate from the Shc-Grb2-Sos complexes, since we failed to detect the 89-kDa component in Shc immunoprecipitations. Fig. 9 shows schematically the composition of FIG. 7. A, localization of Grb2 SH2-bound p89 in the membrane fraction. L6 cells expressing wild type FGFR-1 were incubated in the presence (ϩ) or absence (Ϫ) of 100 ng/ml FGF-2 for 10 min. Cells were scraped into PBS and homogenized. Nuclei were pelleted and discarded, while the supernatants were subjected to Airfuge centrifugation to pellet the particulate (P) fraction. The particulate and soluble (S) fractions were adjusted to equal volumes in Nonidet P-40 lysis buffer and incubated with Grb2 SH2 fusion protein-bound to glutathione-Sepharose. Material bound to beads was separated by SDS-PAGE and immunoblotted with Tyr(P) antibody. B, complex formation of p89 with Sos after FGF-2 stimulation. L6 cells expressing wild type FGFR-1 were incubated in the presence (ϩ) or absence (Ϫ) of 100 ng/ml FGF-2 for 10 min, lysed, and incubated with either Sos antiserum or Grb2 SH2 fusion protein. Samples were separated by SDS-PAGE and immunoblotted with Tyr(P) antibody. Prec., precipitation; Ib, immunoblotting.
FIG. 9. Schematic representation of possible Grb2 interactions after FGF-2 stimulation. Dimerized and activated FGFR-1, with FGF-2 bound to the extracellular domain, mediates tyrosine phosphorylation of Shc through several sites in the FGFR-1. Tyrosine phosphorylation of p89 is mediated directly or indirectly by FGFR-1. Tyrosinephosphorylated, membrane-bound p89 is found in complex with Grb2-Sos, and is therefore, together with Shc, implicated in Ras activation. the different Grb2-containing complexes in FGF-stimulated cells. We also detected another tyrosine-phosphorylated protein of 40 kDa interacting with Grb2 SH2 in FGF-2-stimulated cells. This far, we have only been able to visualize the 40-kDa component in vitro.
Tyrosine phosphorylation of the 89-kDa component was seen immediately in response to FGF-2 and remained in FGF-2stimulated cells for several hours. The fact that the 89-kDa molecule is tyrosine-phosphorylated rapidly after FGF-2 stimulation argues that it is phosphorylated directly by the receptor. Whether the phosphorylation involves complex formation with the receptor is presently not clear. Tyrosine phosphorylation of 80 -90-kDa proteins in FGF-2-stimulated cells has been described previously (42,43). One of these proteins has been identified as cortactin, a cytoskeleton-associated protein (44). We do not believe that the 89-kDa component described here corresponds to cortactin or to other known signal transduction molecules. Thus, cortactin lacks a potential Grb2 binding site (45). Furthermore, as shown by immunoblotting analyses, p89 does not correspond to the SH2 domain containing transcription factors denoted "signal transducer and activator of transcription" (STAT; data not shown) (46). Other possible candidate proteins, such as the regulatory subunit p85 of phosphatidylinositol 3Ј-kinase, do not fulfil the criterion of being tyrosine-phosphorylated only after FGF-2 and not after PDGF-BB or EGF stimulation (47). Yet other signal transduction molecules with an appropriate mass, such as Eps8 (48), also lack a potential Grb2 SH2 consensus binding site. In order to further characterize the 89-kDa component, we have undertaken a purification procedure. In agreement with a role for p89 in signal transduction, we have detected only very low levels of p89 binding to Grb2 SH2, in the cell types examined this far (data not shown). We aim to extensively increase the starting material, and using microsequencing techniques, we hope to deduce the structure of p89.