v-Src Induces Tyrosine Phosphorylation of Focal Adhesion Kinase Independently of Tyrosine 397 and Formation of a Complex with Src*

The non-receptor tyrosine kinase FAK plays a key role at sites of cellular adhesion. It is subject to regulatory tyrosine phosphorylation in response to a variety of stimuli, including integrin engagement after attachment to extracellular matrix, oncogene activation, and growth factor stimulation. Here we use an antibody that specifically recognizes the phosphorylated form of the putative FAK autophosphorylation site, Tyr397. We demonstrate that FAK phosphorylation induced by integrins during focal adhesion assembly differs from that induced by activation of a temperature-sensitive v-Src, which is associated with focal adhesion turnover and transformation. Specifically, although v-Src induces tyrosine phosphorylation of FAK, there is no detectable phosphorylation of Tyr397. Moreover, activation of v-Src results in a net decrease in fibronectin-stimulated phosphorylation of Tyr397, suggesting possible antagonism between v-Src and integrin-induced phosphorylation. Our mutational analysis further indicates that the binding of v-Src to Tyr397 of FAK in its phosphorylated form, which is normally mediated, at least in part, by the SH2 domain of Src, is not essential for v-Src-induced cell transformation. We conclude that different stimuli can induce phosphorylation of FAK on distinct tyrosine residues, linking specific phosphorylation events to ensuing biological responses.

Focal adhesion kinase (FAK) 1 is a non-receptor protein-tyrosine kinase that has been implicated in a variety of signaling pathways and cellular processes. It is widely expressed in most tissues, and at particularly high levels in neuronal tissue (1,2). Phosphorylation of FAK occurs in response to a wide variety of stimuli, including integrin engagement, oncogenic transformation, and growth factors or mitogenic neuropeptides (3)(4)(5). FAK was cloned simultaneously from mouse and chicken (6,7), and is distinguished from other non-receptor protein-tyrosine kinases by being devoid of the protein interaction Src homology (SH) domains, SH2 or SH3 (6,7). It consists of a central catalytic domain, which is flanked by large amino-and carboxyl-terminal regions. Furthermore, an amino-terminal sequence that can bind the ␤ 1 -integrin intracellular domain in vitro has been identified (8), and an association between ␤ 1 -integrin and FAK has been demonstrated in human keratinocytes in vivo (9). The carboxyl terminus contains two proline-rich sequences that can bind to SH3 domain-containing proteins, such as p130 cas or Graf (10,11), as well as the sequences responsible for focal adhesion targeting of FAK (FAT domain; residues 919 -1042) (12). In addition to these functional regulatory sequences, alternative splicing provides another mode of FAK regulation, with autonomous expression of the carboxyl-terminal portion of FAK (13), termed FRNK. However, while enforced overexpression of FRNK inhibits FAK-mediated cell spreading (14), its role as a physiological regulator of FAK function in vivo remains to be established.
Despite the model proposed above, the exact nature of the phosphoacceptor sites targeted by individual stimuli, the temporal sequence of events and how these are linked to cellular responses, is still unclear. To address this with respect to Tyr 397 of FAK, we generated an antibody that specifically recognized the phosphorylated form of FAK-Tyr 397 . This was used to detect phosphorylation of Tyr 397 after plating cells on fibronectin. In contrast, the Tyr 397 -phospho-specific antibody failed to react with FAK after activation of a ts v-Src mutant, although v-Src did induce phosphorylation of FAK on other tyrosine residues. In addition, a Tyr 397 to Phe 397 mutant of FAK (397F) was still tyrosine phosphorylated after v-Src activation, although direct association of the two proteins was ablated. Furthermore, vast overexpression of FAK-Y397F, which also inhibited the association of v-Src with endogenous FAK, did not affect the targeting of either v-Src or FAK to focal adhesions, v-Src-induced tyrosine phosphorylation of FAK on sites other than Tyr 397 , or cell transformation.

EXPERIMENTAL PROCEDURES
Cells and Viruses-Primary chicken embryo fibroblasts (CEF) cultures were grown in Dulbecco's modified Eagle's medium supplemented with 5% newborn calf serum, 1% chick serum, and 10% tryptose phosphate broth, and were buffered with 5% CO 2 . For FN plating experiments, dishes were coated by preincubating overnight at 4°C in a solution of 10 g/ml FN (Stratatec) in PBS before being used for cell plating. The neomycin-selectable avian retrovirus, SFCV, constructs encoding temperature sensitive (ts) LA-29 v-Src (26), or a kinase-defective variant of ts LA-29 that contains a Lys-Arg mutation at position 295, have been described previously (27). Cells expressing ts v-Src mutants were grown either at restrictive temperature (41°C), or were shifted to permissive temperature (35°C) for the times indicated. Myctagged FAK was generated by ligating a double stranded oligonucleotide (sense strand, 5Ј-gatcagcgaacaaaaactcatctcagaagaggatctgaataa-3Ј; antisense strand, 5Ј-gatcttattcagatcctcttctgagatgagtttttgttcgct-3Ј) into the BclI site at position 3186. Y397F-FAK was produced by conventional polymerase chain reaction and cloning methods, and verified by sequencing. Both Myc-tagged variants were expressed in the replication competent retrovirus RCAS.
Generation of a Phosphospecific Antibody Against Tyr 397 of FAK-For the generation of antibodies specific for the putative FAK autophosphorylation site, Tyr 397 , the following 17-amino acid peptide was synthesized SVSETDDpYAEIIDEED(C). This represents amino acids 390 -405 of avian and human FAK (1, 6) with tyrosine 397 being chemically phosphorylated. The carboxyl-terminal cysteine residue was added to facilitate coupling to a carrier protein, keyhole limpet hemocyanin. Peptides were estimated to be greater than 90% pure by high performance liquid chromatography, and composition was confirmed by amino acid analysis (synthesis and analysis of peptides was carried out by Affiniti Research Products Ltd., Exeter, United Kingdom). 200 g of conjugated peptide (molar ratio (peptide, 40:1, keyhole limpet hemocyanin) was emulsified in Freund's adjuvant (Sigma), either complete for the primary immunization, or incomplete for subsequent immunizations, as described previously (28). Serum was checked for positive reactivity and IgG was purified by caprylic acid and ammonium sulfate precipitation followed by affinity absorption, first to a Sepharose column containing the immunizing peptide in its unphosphorylated form (to remove non-phospho-specific antibodies). The column eluate was then absorbed to a Sepharose column containing the phosphopeptide, to affinity absorb phospho-specific IgG. Bound IgG was eluted under low pH conditions and subjected to dialysis against phosphate-buffered saline (PBS). (The purification procedures described were carried out according to instructions provided to us by Affiniti Research Products Ltd.) Purified IgG, termed anti-397-P, was aliquoted and stored at Ϫ70°C.
Immunoprecipitation and Western Blotting-For protein analysis, dishes of cells were washed with ice-cold PBS and either used immediately or quick frozen at Ϫ70°C. For immunoprecipitations cell monolayers were lysed in RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 5 mM EGTA, 0.1% SDS, 1% Nonidet P-40, and 1% sodium deoxycholate) with the inclusion of inhibitors (500 M sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 100 M sodium vanadate, 10 g/ml leupeptin, 10 g/ml benzamidine, and 10 g/ml aprotinin (Sigma)). Samples were sonicated and clarified by centrifugation at 21,000 ϫ g at 4°C. 800 -1000 g of protein lysates (measured by Micro BCA protein assay kit, Pierce) were immunoprecipitated with 10 l of anti-FAK rabbit polyclonal antiserum (29), 2 g of anti-Src mAb EC10 (UBI), or 2 g of anti-Myc 9E10 mAb (Sigma) overnight at 4°C. Complexes were collected by addition of a 50% (v/v) preparation of Sepharose beads coupled to Protein-A (Sigma) in RIPA buffer for a further 60 min at 4°C, followed by 5 ϫ washes with RIPA buffer and subsequent elution in sample buffer (50 mM Tris⅐HCl, pH 6.7, 2% SDS, 700 mM ␤-mercaptoethanol, 10% glycerol, and 0.1% bromphenol blue) at 100°C for 2 min. The Sepharose beads were pelleted and supernatants analyzed by SDSpolyacrylamide gel electrophoresis (30). For Western blotting, proteins were transferred to nitrocellulose using a semi-dry blotting apparatus. Unoccupied binding sites were blocked with 5% dried milk powder in PBS ϩ 0.1% Tween 20 (PBST) followed by probing with either anti-Tyr 397 (anti-397-P) purified IgG (2 g/ml) in 3% bovine serum albumin in PBST or general anti-phosphotyrosine (PY20; Transduction Laboratories) at 0.25 g/ml or FAK monoclonal antibody at 0.5 g/ml (clone 77, Transduction Laboratories). Detection was by incubation with either anti-rabbit or anti-mouse horseradish peroxidase-conjugated secondary antibody or anti-Myc conjugated to horseradish peroxidase (Invitrogen) and visualization was by enhanced chemiluminescence (ECL, Amersham Pharmacia Biotech) according to the manufacturer's guidelines.
Immunofluorescence-Cells were grown on glass coverslips, fixed at 4°C for 15 min with 3.7% formaldehyde, permeabilized with 0.5% Triton X-100, and incubated with 1:500 anti-Src EC10 (UBI), 2 g/ml anti-Myc mAb (Jackson) or 2.5 g/ml anti-paxillin mAb (Transduction Laboratories). Antibody detection was via fluorescein isothiocyanateconjugated goat anti-mouse IgG (Sigma) and/or Texas Red-conjugated goat anti-rabbit IgG (Jackson), for 45 min at room temperature. Actin stress fibers were visualized by staining with 10 g/ml phalloidin-TRITC (Sigma) for 40 min at room temperature. Fluorescence was visualized using a Bio-Rad MRC 600 confocal microscope and images were printed on a dye sublimation printer (Kodak).

Generation of a Phospho-specific Antibody Recognizing FAK-
Tyrosine 397-In order to study stimulus-induced phosphorylation of the putative autophosphorylation site in FAK, i.e. Tyr 397 , we generated an antiserum which specifically recognized Tyr 397 in its phosphorylated state. A 17-amino acid peptide encompassing residues 390 -405 of FAK, with a phosphorylated tyrosine residue at position 397 ( Fig. 1A), was coupled to keyhole limpet hemocyanin via a carboxyl-terminal cysteine residue and used to immunize rabbits (as described under "Experimental Procedures"). Antibodies present in the crude sera that recognized nonphosphorylated epitopes were removed by affinity absorption. Purified phospho-specific IgG was designated anti-397-P and tested for its ability to recognize FAK in lysates of CEF that had been plated onto FN for 60 min, conditions that promote autophosphorylation of FAK (21). Anti-397-P reacted strongly with FAK at 125 kDa (Fig. 1B, lane 2), as judged by co-migration with FAK immunoblotted with a commercial FAK-specific antibody (Fig. 1B, lanes 5 and 6). FAK specificity was supported by the lack of reactivity of anti-397-P with FAK extracted from CEF that had been in suspension for 60 min after trypsinization (Fig. 1B, lane 1), conditions that A, a schematic representation of the gene encoding FAK, indicating the peptide representing amino acids 390 -405 of the avian and human FAK sequence that was used to immunize rabbits. The tyrosine residue at position 397 was chemically phosphorylated. Also shown are the position of the focal adhesion targeting sequence (FAT) and the COOHterminal proline-rich regions (P) that mediate binding to SH3 domain containing proteins. The putative Src-specific phosphorylation sites (Tyr 407 , Tyr 576/7 , Tyr 861 , and Tyr 925 ) are indicated. B, FAK was detected by immunoblotting lysates of CEF that had been either retained in suspension (lanes 1, 3, and 5) or plated onto FN-coated dishes (lanes 2, 4, and 6) for 60 min. Membranes were probed with phospho-Tyr 397specific antibody (anti-397-P), either preincubated with excess immunizing peptide in the unphosphorylated form (lanes 1 and 2) or in the phosphorylated form (lanes 3 and 4), or probed with a FAK specific mAb to detect total cellular FAK (lanes 5 and 6).
induce rapid de-phosphorylation of FAK (15). Confirmation of the sequence and phosphorylation-specific reactivity of this antibody was obtained by blotting cell lysates with anti-397-P IgG which had been preincubated with excess immunizing peptide, either in the unphosphorylated (Fig. 1B, lanes 1 and 2) or phosphorylated form (Fig. 1B, lanes 3 and 4). Preincubation with excess Tyr 397 -containing phosphopeptide inhibited FAK reactivity after plating cells on FN (Fig. 1B, lane 4), while excess unphosphorylated peptide had no effect, indicating that purified anti-397-P IgG reacted specifically with the Tyr 397phosphorylated form of FAK.
Phosphorylation of FAK Induced by FN Differs from That Induced by the v-Src Oncoprotein-Although both FN attachment and v-Src activity stimulate tyrosine phosphorylation of FAK, the biological outcome of these two stimuli is quite different. Specifically, integrin engagement mediated by FN attachment induces focal adhesion formation and cell spreading, while v-Src induces focal adhesion turnover and cell transformation. We therefore compared FAK-Tyr 397 phosphorylation induced by either plating cells on FN or by activating a ts v-Src mutant (ts-LA29 v-Src) in the absence of FN, which induces phosphorylation, and subsequent proteolysis, of FAK, events that are visibly linked to focal adhesion disruption during transformation (31). As expected, plating of cells on FN triggered a substantial increase in FAK tyrosine phosphorylation (Fig. 2, middle panel, lanes 2-4), which coincided with an increase in the phosphorylation of Tyr 397 (Fig. 2, upper panel,  lanes 2-4). However, although FAK was also tyrosine phosphorylated by activation of ts v-Src after shift to the permissive temperature (35 o C), as detected by a general phosphotyrosine antibody (Fig. 2, middle panel, lanes 5-7), FAK-Tyr 397 failed to react with anti-397-P (Fig. 2, upper panel, lanes 5-7). This indicates that phosphorylation of FAK induced by v-Src can occur at phosphoacceptor sites that are distinct from Tyr 397 , although the latter is clearly phosphorylated after plating on FN.
Activation of ts v-Src Results in a Net Decrease in FAK-Tyr 397 Phosphorylation Stimulated by FN-Since our results indicated that different phosphorylation sites in FAK might act as recipients for signals from FN or v-Src, we next examined phosphorylation after attachment to FN when v-Src was active. CEF expressing ts v-Src, or its kinase-defective variant (ts LA29-KD), were either retained in suspension after trypsinization (S), or plated onto FN for 60 min, either at the restrictive (41°C) or permissive (35°C) temperatures. After plating on FN, specific phosphorylation of Tyr 397 occurred both at the permissive and the non-permissive temperatures and in cells expressing both kinase-active and kinase-defective v-Src (Fig. 3A, upper panel). However, we noted that ts v-Src activity at the permissive temperature was consistently associated with reduced phosphorylation of FAK-Tyr 397 , when compared with cells plated on FN at restrictive temperature (Fig. 3A,  compare lanes 4 and 5). To investigate this further, adhesion to FN and activation of v-Src were stimulated sequentially, and phosphorylation of FAK-Tyr 397 examined. Following adhesion to FN for 45 min, cells expressing ts v-Src were either switched to 35°C for various times (Fig. 3B, lanes 4 -6), or were maintained at 41°C (Fig. 3B, lanes 7-9). As expected, plating cells on FN following suspension caused a robust increase in phosphotyrosine content, which correlated with specific phosphorylation of Tyr 397 (Fig. 3B, upper and middle panels). However, following switch to the permissive temperature, there was a time-dependent decrease in the specific phosphorylation of  ). B, FAK was immunoprecipitated from CEF expressing ts LA29 v-Src. Cells were either retained in suspension for 60 min (S) or plated onto FN-coated dishes and maintained at 41°C for 30 and 45 min. Cells were then either switched to 35°C (lanes 4 -6) or maintained at 41°C (lanes 7-9) for the indicated times. Following immunoprecipitation, proteins were blotted onto nitrocellulose and probed as described above for A. FAK-Tyr 397 (Fig. 3B, upper panel, lanes 4 -6). Furthermore, although there was also some decrease in phosphorylation of FAK-Tyr 397 after 120 min at 41°C (Fig. 3B, lanes 9), indicating that phosphorylation of FAK-Tyr 397 after plating on FN was transient, the net decrease was exacerbated and was evident by 60 and 90 min when v-Src was active at 35°C (Fig. 3B, upper  panel, lanes 4 -6). Thus, v-Src-induced FAK phosphorylation is associated with reduced specific FAK-Tyr 397 phosphorylation that is normally stimulated by attachment to FN, raising the possibility that v-Src-induced tyrosine phosphorylation of FAK may antagonize integrin-induced phosphorylation.
Phosphorylation of FAK by Src Does Not Require Binding of Src to Tyr 397 -Extracellular matrix (ECM)-induced autophosphorylation of FAK on Tyr 397 creates a high affinity binding site for the SH2 domain of c-Src, and mutation (Tyr to Phe) of this residue inhibits association (21,32). These findings have led to the proposal of a model in which ECM-stimulated integrin clustering leads to accumulation of FAK at adhesion sites and autophosphorylation on Tyr 397 , facilitating the binding of c-Src, or other Src family members, via the Src SH2 domain. This binding would result in displacement of the autoregulatory, intramolecular Src-Tyr 527 -SH2 domain interaction, consequently activating Src kinase activity. Further phosphorylation of FAK on additional tyrosine phosphoacceptor sites by associated Src kinase is believed to create further potential binding sites for SH2-domain containing proteins, thus linking Src-induced FAK phosphorylation to downstream signaling pathways (33).
Our findings that FAK-Tyr 397 is not detectably phosphorylated after v-Src activation led us to investigate whether binding of v-Src to FAK, mediated by the Src SH2 domain and FAK-Tyr 397 in its phosphorylated form, was essential for v-Src to induce tyrosine phosphorylation of FAK on other sites. To address this, we used CEF expressing ts v-Src which also overexpressed Myc-tagged FAK proteins, either wt FAK or FAK in which the Tyr at position 397 had been changed by mutagenesis to Phe (Y397F-FAK). We tested whether the Y397F mutation inhibited the association of v-Src with FAK, by carrying out anti-Src immunoprecipitations from wt-and Y397F-FAK-expressing cells and probing immunoblots with anti-Myc. As expected, although expressed wt FAK co-precipitated with v-Src (Fig. 4A, top panel, lane 1), an interaction that persisted after activation of v-Src by shift to the permissive temperature (Fig. 4A, top panel, lanes 2 and 3), the Y397F-FAK mutant failed to co-precipitate with v-Src (Fig. 4A, top panel,  lanes 4 -6). Thus, consistent with the findings of others on c-Src (21,32), v-Src also requires FAK-Tyr 397 for binding. In addition, re-probing of the immunoblots with an anti-FAK mAb failed to detect any substantial binding of endogenously expressed FAK to v-Src (Fig. 4A, bottom panel, lanes 4 -6).
To determine whether v-Src-induced phosphorylation of FAK occurred in the absence of Tyr 397 and formation of the v-Src⅐FAK complex, cells expressing the Myc-tagged FAK proteins were either maintained at 41°C or shifted to 35°C for 2 h. We detected an increase in tyrosine phosphorylation of wt-FAK and to a lesser extent Y397F-FAK after activation of v-Src (Fig. 4B, top panel, lanes 2 and 4, respectively), indicating that binding of v-Src to FAK-Tyr 397 was not essential for phosphorylation of other tyrosine residues within FAK. Additional confirmation of the specificity of the anti-397-P IgG was obtained by its failure to recognize Myc-tagged Y397F-FAK immunoprecipitated from these cells with anti-Myc (Fig. 4B, middle panel,  lanes 1 and 2).
To test whether integrin-dependent tyrosine phosphorylation of FAK on sites other than Tyr 397 was dependent on prior phosphorylation of Tyr 397 , or the binding of FAK to Src family kinases, CEF overexpressing Myc-tagged wt-or Y397F-FAK were plated on FN, and tyrosine phosphorylation examined. We found that Y397F-FAK was tyrosine phosphorylated after plating of cells on FN (Fig. 4C, upper panel, lane 2), although phosphorylation was considerably delayed when compared with wt-FAK (Fig. 4C, upper panel, lanes 5 and 6). Furthermore, this delayed integrin-induced tyrosine phosphorylation of Y397F-FAK was likely due to Src family kinase activity, since phosphorylation was impaired in the presence of a Src inhibitor, PD162531 (Fig. 4C, upper panel, lanes 3 and 4), that we have previously shown to inhibit the activity of the Src family kinases in vitro and in vivo (34). Thus, as is the case  1-3) or Myc-tagged Y397F-FAK (lanes 4 -6), that had been either retained at 41°C (0 h) or shifted to 35°C (1 h, 2 h). Samples were transferred to nitrocellulose and probed with an anti-Myc mAb to detect co-precipitated FAK (upper panel), an anti-Src mAb EC10 to confirm immunoprecipitation of v-Src (middle panel), or a FAK mAb to detect any endogenously expressed FAK binding (lower panel). B, FAK was immunoprecipitated with an anti-Myc mAb from CEF-expressing ts LA29 v-Src and either Myc-tagged Y397F-FAK (lanes 1 and 2) or Myc-tagged wt-FAK (lanes 3 and 4), that had been either retained at 41°C or shifted to 35°C for 2 h. Proteins were blotted and probed with PY20 antiphosphotyrosine (upper panel), anti-397-P (middle panel), or anti-FAK (bottom panel). C, FAK was immunoprecipitated with an anti-Myc mAb from CEF expressing either Myc-tagged Y397F-FAK (lanes 1-4) or Myc-tagged wt-FAK (lanes 5 and 6). Cells were plated onto FN-coated dishes for the indicated times, either in the absence (lanes 1-2 and 5-6) or presence of 2 M PD162531, a selective Src family kinase inhibitor (lanes 3 and 4). Proteins were blotted and probed with either PY20 anti-phosphotyrosine (upper panel) or anti-FAK (bottom panel).
with v-Src, formation of a complex between c-Src SH2 and FAK, or prior phosphorylation of Tyr 397 , is not absolutely required for Src-dependent phosphorylation of FAK on sites other than Tyr 397 , but may be necessary for optimal, rapid phosphorylation after integrin engagement.
The Association of Src with FAK, and Phosphorylation of Tyr 397 , Is Not Required for Cell Transformation-As the proposed model for FAK regulation suggested a critical role for phosphorylation of FAK-Tyr 397 , and the consequent association with Src, during ECM-induced intracellular signaling, we examined whether vast overexpression of Y397F-FAK that blocks association with v-Src (Fig. 4A), and also endogenous c-Src (not shown), influenced v-Src's biological activity. Specifically, v-Src activation in CEF is accompanied, at early times, by the assembly of new focal adhesions containing the oncoprotein (27,35), and subsequently, by tyrosine kinase-induced focal adhesion disruption as cells round up during transformation (27). Thus, we tested whether mutation of Tyr 397 of FAK to Phe 397 influenced the intracellular targeting of ts v-Src to focal adhesions after switch to the permissive temperature, or subsequent focal adhesion disruption and transformation. The extent of overexpression of Myc-tagged wt-and Y397F-FAK, compared with endogenous FAK in vector-transfected controls, is shown (Fig. 5A, lanes 5-8). Double immunofluorescence with anti-Src and anti-Myc demonstrated that both exogenous FAK proteins were localized in focal adhesion structures, irrespective of v-Src activity (Fig. 5B). In addition, v-Src was translocated from the perinuclear region of the cell to focal adhesions at the periphery after shift to the permissive temperature, even in the presence of overexpressed Y397F-FAK that prevented formation of a Src⅐FAK complex (Fig. 5B, upper panels). These data indicate that both FAK and v-Src are targeted to focal adhesions under conditions where they are not associated. Moreover, overexpression of Y397F-FAK had no obvious consequences for focal adhesions formed at early times after activation of v-Src, as judged by visualization of focal adhesions containing the oncoprotein (Fig. 5B), and cells with focal adhesions containing Y397F-FAK remained spread. In addition, v-Src-induced morphological transformation that is associated with focal adhesion disruption (31) was not impaired by vast overexpression of Y397F-FAK (Fig. 5C). Staining for actin stress fibers and the focal adhesion component paxillin, demonstrated that both cells expressing wt FAK and Y397F-FAK exhibited characteristic signs of the transformed phenotype when switched to 35°C, with loss of actin stress fibers and disruption of paxillin containing focal adhesions associated with cell rounding (Fig. 5C).  2) or CEF-expressing ts LA29 v-Src alone (lanes 3 and 4), or with either Myc-tagged Y397F-FAK (lanes 5 and 6) or Myc-tagged wt-FAK (lanes 7 and 8). CEF were either maintained at 41°C or shifted to 35°C for 2 h as indicated. B, CEF cells expressing ts LA29 v-Src and either Y397F-FAK or wt-FAK were fixed and stained with rabbit anti-Src serum and mouse anti-Myc mAb to visualize exogenous Myc-tagged FAK. Subsequent detection was via appropriate secondary antibodies coupled to either fluorescein isothiocyanate (Src) or Texas Red (Myc). C, CEF expressing ts LA29 v-Src and either Y397F-FAK (lower panels) or wt-FAK (upper panels) were maintained at 41°C or were shifted to 35°C for 24 h. Altered cell morphology was visualized by phase-contrast microscopy. Cells were also fixed and stained with mouse anti-paxillin mAb or phalloidin-TRITC for visualization of actin stress fibers (small double panels). Size bars indicate 25 m. teins between 115 and 130 kDa (20,36,37). Much of this can be accounted for by FAK (6,7), which is subject to tyrosine phosphorylation following both adhesion to FN, or activation of an oncogenic variant of Src (19). Furthermore, early experiments indicated that v-Src-induced phosphorylation of FAK was specifically associated with reduced gel migration, suggesting that Src-and integrin-dependent phosphorylation of FAK were either quantitatively or qualitatively different (19). This was later supported by phospho-peptide mapping, which demonstrated that FAK-Tyr 397 is phosphorylated following integrin engagement, with further phosphorylation of FAK on Tyr 407 , Tyr 576 , Tyr 577 , Tyr 861 , and Tyr 925 likely to be mediated by c-Src, or other members of the Src family (15,16,21). Further experimentation led to a more refined model, in which cell spreading on ECM results in integrin-induced autophosphorylation of FAK-Tyr 397 , creating a binding site for the Src SH2 domain, leading to Src association and Src-mediated phosphorylation of the carboxyl-terminal Tyr residues mentioned above (33). However, this model is not necessarily relevant to activation of v-Src in adherent cells and subsequent transformation. In particular, the specific phosphorylation of FAK-Tyr 397 in vivo after v-Src activation and its role in v-Src's biological activity remained unclear.
By using an antibody that specifically recognized FAK-Tyr 397 in its phosphorylated form (characterized in Fig. 1), we confirmed that attachment to FN induces phosphorylation of FAK-Tyr 397 in vivo, but found that activation of a ts v-Src oncoprotein does not induce detectable phosphorylation of FAK-Tyr 397 , although other sites in FAK were phosphorylated (Fig. 2). This indicates that different phosphorylation states of FAK are induced by integrin engagement and activated Src, and these correlate with the distinct biological outcomes of integrin-dependent adhesion assembly during cell spreading, and v-Src-induced adhesion disruption during cell transformation. Furthermore, v-Src activity leads to a net reduction in FAK-Tyr 397 phosphorylation after attachment to FN (Fig. 3), suggesting that v-Src-and integrin-induced phosphorylation events are not only distinct, but may also be antagonistic. These findings are consistent with the idea that specific phosphorylation of FAK might act as a determinant of focal adhesion assembly or disassembly, mediated by downstream effector pathways that are engaged after stimulus-induced phosphorylation of particular tyrosine residues. If true, this implicates FAK as a pivotal switch in regulation of the cell-ECM adhesion assembly/disassembly cycle, with integrin-induced phosphorylation of Tyr 397 stimulating assembly and Srcinduced phosphorylation of one or more of the carboxylterminal tyrosine residues stimulating disassembly. In support of this mode of regulation, unregulated v-Src alters the balance of cellular FAK phosphorylation in favor of the carboxyl-terminal tyrosine residues at the expense of Tyr 397 (Fig. 3), and the consequent biological outcome is uncontrolled focal adhesion disruption and transformation (31).
In keeping with the proposed role for specificity of FAK phosphorylation as a critical determinant of focal adhesion assembly and disassembly, there is abundant genetic evidence implicating FAK in both processes. Specifically, reduced FAK phosphorylation as a result of overexpression of FRNK, impairs the ability of cells to spread on FN (14), while fibroblasts derived from FAK Ϫ/Ϫ mice retain the ability to spread on FN, but are relatively non-motile as a result of impaired focal adhesions turnover (38). Taken together, these findings imply that in cells that express FAK, it normally has a role in regulating both the assembly and disassembly of cell-ECM adhesions. Our work has indicated that v-Src-induced tyrosine phosphorylation of FAK is linked to focal adhesion disassembly during transformation and cell motility (27,31), and that v-Src's biological activity is independent of phosphorylation of FAK-Tyr 397 (Fig. 5).
Recent work has shown that association of FAK with c-Src is associated with its redistribution from a diffuse cellular location to focal adhesions (39). However, our data presented here demonstrate that at least in the case of v-Src, FAK-Tyr 397 -dependent Src association is not necessary for focal adhesion localization. We find that vast overexpression of a Y397F-FAK mutant, which inhibits Src/FAK association, had no discernible effect on the assembly of focal adhesions containing the oncoprotein. However, the inside/out signaling that induces focal adhesion assembly early after v-Src activation may be mechanistically different to the outside/in signaling that leads to adhesion assembly induced after integrin engagement and during ECM-dependent cell motility. Furthermore, although the putative signaling pathway engaged as a consequence of FAK-Tyr 397 phosphorylation, and responsible for integrin-induced focal adhesion assembly, is unclear, it is noteworthy that it is not only the Src family kinases that binds to phospho-Tyr 397 . Also binding via this site are the p85 regulatory subunit of phosphoinositol 3-kinase (40), which is required for FAK-dependent cell motility (41), and the tumor suppresser protein phosphatase, PTEN (42), that is able to de-phosphorylate FAK in response to placing cells in suspension (43)(44)(45). In addition, the adapter protein Grb7 also binds to FAK via Tyr 397 (46). The increasing number of signaling proteins now shown to bind FAK in a Tyr 397 -dependent manner indicates that there is unlikely to be a single effector pathway activated as a result of integrin-induced phosphorylation at this site.
Our results demonstrate that neither the FAK-Tyr 397 /Src-SH2 domain interaction, nor prior phosphorylation of FAK-Tyr 397 , is required for v-Src to induce tyrosine phosphorylation of FAK (Fig. 4), or focal adhesion turnover that accompanies transformation (Fig. 5). However, although there is no absolute requirement, the kinetics of Src family kinase-dependent phosphorylation of FAK that occurs at later times after integrin engagement may be influenced by the state of FAK-Tyr 397 phosphorylation (Fig. 4).
In conclusion, although the downstream consequences of individual FAK phosphorylation events are not fully understood, we have shown that integrin-and activated Src-induced phosphorylation of FAK are at least partly distinct, with only integrin engagement stimulating Tyr 397 phosphorylation. The specific phosphorylation of FAK induced by these stimuli, together with their biological correlates, suggests that FAK phosphorylation may determine the balance of focal adhesion assembly and disassembly, and consequently the rate of integrin-dependent cell motility.