Bombesin, lysophosphatidic acid, and epidermal growth factor rapidly stimulate focal adhesion kinase phosphorylation at Ser-910: requirement for ERK activation.

A rapid increase in the tyrosine phosphorylation of focal adhesion kinase (FAK) has been extensively documented in cells stimulated by multiple signaling molecules, but virtually nothing is known about the regulation of FAK phosphorylation at serine residues. Stimulation of Swiss 3T3 cells with bombesin promoted a striking increase ( approximately 13-fold) in the phosphorylation of FAK at Ser-910, as revealed by site-specific antibodies that recognized the phosphorylated state of this residue. Lysophosphatidic acid and epidermal growth factor (EGF) also stimulated FAK phosphorylation at Ser-910. Direct activation of protein kinase C isoforms with phorbol-12,13-dibutyrate (PDB) also promoted striking phosphorylation of FAK at Ser-910. Treatment with the protein kinase C inhibitor GF I or Ro 31-8220 or chronic exposure to PDB prevented the increase in FAK phosphorylation at Ser-910 induced by bombesin or PDB but not by EGF. Treatment with the ERK inhibitors U0126 and PD98059 prevented FAK phosphorylation at Ser-910 in response to all of the stimuli tested. Furthermore, incubation of activated ERK2 with FAK immunocomplexes leads to FAK phosphorylation at Ser-910 in vitro. Our results demonstrate, for the first time, that stimulation with bombesin, lysophosphatidic acid, PDB, or EGF induces phosphorylation of endogenous FAK at Ser-910 via an ERK-dependent pathway in Swiss 3T3 cells.

FAK promotes the transmission of downstream signaling by binding and recruiting signaling and adaptor proteins (3). Autophosphorylation of FAK at Tyr-397, located N-terminal to the catalytic domain, creates a binding site for the tyrosine kinase Src (35) and other downstream signaling effectors. Subsequent Src-mediated phosphorylation of FAK at Tyr-576 and Tyr-577 is important for the maximal activation of FAK and downstream signaling events (14,27). Phosphorylation at Tyr-925 within the focal adhesion targeting (FAT) domain creates a binding site for the Src homology 2 domain of the adapter protein Grb2-SOS (Ras exchange factor) complex and provides a possible mechanism of activation of the Ras/Raf/MEK/ERK pathway (36). The FAT domain binds paxillin and talin, which are responsible for targeting FAK to the focal adhesions and for promoting downstream signaling (36 -38). The importance of FAK-mediated signal transduction is underscored by experiments showing that this tyrosine kinase is implicated in embryonic development (39) and in the control of cell migration (27, 40 -42), proliferation (40,43), and apoptosis (44,45). It is increasingly recognized that overexpression of FAK is linked to the invasive properties of cancer cells (46,47).
In addition to being phosphorylated at multiple tyrosines in response to external stimuli, FAK is also phosphorylated at serine residues (48,49). Recently, Ser-722 and Ser-910, in the COOH-terminal, noncatalytic region of FAK (termed FAK-related nonkinase (FRNK)), have been identified as prominent phosphorylation sites (49). Despite the importance of FAK in signal transduction, virtually nothing is known about the regulation of these phosphorylation events. In particular, none of the previous studies demonstrated that the phosphorylation of any of these serine residues of FAK can be regulated in re-sponse to cell stimulation by either GPCR agonists or ligands of tyrosine kinase receptors.
In the present study, we demonstrate that stimulation of Swiss 3T3 cells with bombesin, LPA, PDB, or EGF induces a rapid and dramatic increase in Ser-910 phosphorylation of endogenous FAK. Treatment with the PKC inhibitor GF I or Ro 31-8220 or chronic exposure to PDB to down-regulate PKCs prevented the increase in FAK phosphorylation at Ser-910 induced by bombesin or PDB. Since bombesin or PDB induces ERK activation through a PKC-dependent pathway in Swiss 3T3 cells and because the amino acids surrounding Ser-910 conform to a consensus site for mitogen-activated protein kinases, we also examined whether the ERKs mediate FAK phosphorylation at Ser-910 in response to extracellular stimuli. Our results show that ERK inhibitors prevented FAK phosphorylation at Ser-910 in response to either bombesin or PDB. Furthermore, LPA, another GPCR agonist that induces ERK via pertussis toxin (PTx)-sensitive G i rather than G q and ligands of tyrosine kinase receptors that induce potent ERK pathway activation through PKC-independent pathways, including EGF, also stimulated a dramatic increase in FAK phosphorylation at Ser-910. In addition, activated ERK2 can directly phosphorylate FAK (or recombinant FRNK) at Ser-910 in vitro. Thus, our results show that stimulation with bombesin, LPA, PDB, or EGF stimulates phosphorylation of endogenous FAK at Ser-910 via an ERK-dependent pathway in Swiss 3T3 cells.

EXPERIMENTAL PROCEDURES
Cell Culture-Stock cultures of Swiss 3T3 cells were maintained in DMEM, supplemented with 10% fetal bovine serum in a humidified atmosphere containing 10% CO 2 and 90% air at 37°C. For experimental purposes, Swiss 3T3 cells were plated in 100-mm dishes at 6 ϫ 10 5 cells/dish in DMEM containing 10% fetal bovine serum and used after 6 -8 days when the cells were confluent and quiescent. For assays using suspended cells, confluent and quiescent Swiss 3T3 cells in 100-mm dishes were washed three times with DMEM and subsequently suspended in 10 ml of DMEM by gentle scraping. The cell suspension was then incubated for 30 min at 37°C before stimulation with the respective agonists.
Cell Stimulation with Bombesin and Other Agonists-Confluent and quiescent Swiss 3T3 cells were washed twice with DMEM, equilibrated in the same medium at 37°C for at least 30 min, and then treated with bombesin or other factors for the times indicated. We used 2 ϫ 10 6 cells grown in 100-mm dishes containing 10 ml of DMEM for each experimental condition. The stimulation was terminated by aspirating the medium and solubilizing the cells in 1 ml of ice-cold RIPA buffer containing 50 mM HEPES, pH 7.4, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 10% glycerol, 1.5 mM MgCl 2 , 1 mM EGTA, 1 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 100 mM NaF, and 1 mM phenylmethylsulfonyl fluoride.
Immunoprecipitation-Lysates were clarified by centrifugation at 15,000 rpm for 10 min. Supernatants were transferred to fresh tubes, and proteins were immunoprecipitated at 4°C for 4 h with protein A-agarose linked to polyclonal anti-FAK (C-20) antibody, as previously described. Immunoprecipitates were washed three times with RIPA buffer and extracted in 2ϫ SDS-PAGE sample buffer (200 mM Tris-HCl, pH 6.8, 1 mM EDTA, 6% SDS, 4% 2-mercaptoethanol, 10% glycerol) by boiling for 10 min and resolved by SDS-PAGE.
SDS-Polyacrylamide Gel Electrophoresis-Gel electrophoresis was performed with 8% acrylamide in the separating gel, 4% in the staking gel, and 0.1% SDS. Immunoprecipitates with paxillin were resolved with 8 and 12.5% acrylamide in the separating gel.
Generation and Overexpression of Mutant FAK-pFLAG-CMV TM -2 expression vector (Sigma) containing FAK wild type was used to perform the mutation of Ser-910 to Ala with the QuikChange site-directed mutagenesis kit (Stratagene). COS-7 cells were cultured to 40 -60% confluence on 100-mm plates, and then 6 g of DNA were transfected with Lipofectin according to the instructions from Invitrogen. Cells were then incubated for 18 h at 37°C with 5% CO 2 . Transfection medium was aspirated and replaced with 10 ml of culture medium containing 10% fetal bovine serum prior to further treatment.
In Vitro Kinase Assay-Confluent Swiss 3T3 100 mm plates were lysed in RIPA buffer, and unstimulated FAK was immunoprecipitated. The immunocomplexes were washed twice with RIPA buffer and three times with kinase buffer (30 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol). The immunocomplexes in 70 l of kinase buffer or GST-FRNK or GST-FRNKS910A in 100 l of kinase buffer were incubated with activated ERK2 (Calbiochem) and 10 M cold ATP or [␥-32 P]ATP (specific activity 490 cpm/pmol) at 30°C. The reaction was terminated by the addition of 2ϫ sample buffer and analyzed by SDS-PAGE.
GST-FRNK Fusion Protein-pFLAG-CMV TM -2 expression vector containing FAK wild type and FAK-S910A were used to create FRNK (noncatalytic domain of FAK (amino acids 693-1053)) by PCR amplification using specific oligonucleotide primers (forward primer, 5Ј-CGG-GATCCATGGAATCCAGGCGACAAG-3Ј; reverse primer, 5Ј-AGTTAG-CGGCCGCTTAGTGGGGCCTGGACTG-3Ј) containing restriction sites for BamHI and NotI, respectively (underlined). The resulting PCR products were subcloned as a BamHI-NotI fragment into the vector pGEX-4T3 (Amersham Biosciences) to generate the bacterial expression constructs pGEX-GST-FRNK and pGEX-GST-FRNKS910A. The 57-kDa GST-FRNK and GST-FRNKS910A were expressed in Escherichia coli (BL21) for 2 h, induced by isopropyl-1-thio-␤-D-galactopyranoside. The GST alone and the two fusion proteins were purified with the B-PER GST fusion protein purification kit from Pierce, according to the manufacturer's instructions. The buffer exchange was performed with an Ultrafree-0.5 centrifugal filter device (Millipore Corp.) and used immediately in the in vitro kinase assay.
Materials-Bombesin, endothelin, and LPA were obtained from Sigma. Horseradish peroxidase-conjugated donkey antibodies to rabbit (NA 934V) or mouse (NA931V), and ECL reagents were from Amersham Biosciences. FAK polyclonal Ab C-20 was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The phosphospecific polyclonal Abs to Ser-722 and Ser-910 of FAK were obtained from BioSource International (Camarillo, CA). Anti-Tyr(P) monoclonal antibody 4G10 and anti-FAK 2A7 were from Upstate Biotechnology, Inc. (Lake Placid, NY). SYPRO RED protein stain was purchased from Molecular Probes, Inc. (Eugene, OR). All other reagents used were of the purest grade available.

Bombesin Induces FAK Phosphorylation at Ser-910 -Phos-
phorylation state-specific antibodies provide a powerful strategy to analyze site-specific phosphorylations of a variety of proteins. To examine whether the phosphorylation of endogenous FAK at Ser residues is regulated by GPCR agonists, quiescent cultures of Swiss 3T3 cells were incubated with 10 nM bombesin for various times, and lysates of these agonisttreated cells were immunoprecipitated with anti-FAK antibody. The resulting immunoprecipitates were analyzed by Western blotting using a site-specific antibody that recognizes the phosphorylated state of FAK at Ser-910.
The Western blot analysis illustrated in Fig. 1 (upper panel) shows that prior to stimulation, pS910 immunoreactivity was virtually absent, indicating that Ser-910 of FAK was not phosphorylated in quiescent cultures of Swiss 3T3 cells. Upon bombesin stimulation, pS910 immunoreactivity of a single protein band in SDS-PAGE migrating at the expected apparent molecular mass for FAK (120 kDa) increased dramatically in a time-dependent manner. An increase in FAK phosphorylation, as revealed by immunoblotting with the pS910 antibody, was detected 3 min after the addition of the agonist, reached a maximum within 5-10 min, and remained relatively constant for up to 60 min. The maximal increase of FAK phosphorylation at Ser-910 induced by bombesin was 13.4 Ϯ 1.2-fold, as compared with the unstimulated level.
Bombesin stimulation of Swiss 3T3 cells promoted FAK phosphorylation at Ser-910 in a concentration-dependent manner ( Fig. 1, middle panel). Consistent with the results shown in Fig. 1 (upper panel), the pS910 antibody only very weakly recognized FAK isolated from unstimulated cells. The results presented in Fig. 1 show that bombesin stimulation of Swiss 3T3 cells induces a striking increase in FAK phosphorylation at Ser-910.
We also examined whether bombesin regulates FAK phosphorylation at Ser-722 by Western blotting using a site-specific antibody that detects the phosphorylated state of this residue in FAK. As shown in Fig. 1 (lower panel), the pS722 antibody recognized FAK isolated from unstimulated cells, indicating that Ser-722 of FAK, in contrast to Ser-910, was phosphorylated in the basal state of Swiss 3T3 cells. Furthermore, the addition of bombesin to these cells caused only small changes in the phosphorylation of Ser-722 as compared with that of Ser-910 ( Fig. 1, lower panel). These results indicate that bombesin induces a selective increase in the phosphorylation of endogenous FAK at Ser-910 in Swiss 3T3 cells.
Cytochalasin D Dissociates Bombesin-induced FAK Tyrosine Phosphorylation from FAK Phosphorylation at Ser-910 -Previously, we demonstrated that tyrosine phosphorylation of FAK and complex formation between FAK and Src in response to GPCR agonists requires an intact actin cytoskeleton (7, 16 -18, 29, 30, 50). Specifically, treatment of the cells with cytochalasin D, which caps the barbed end of actin filaments and promotes their depolymerization, inhibits the increase in FAK tyrosine phosphorylation in response to bombesin and other GPCR agonists. Here, we examined whether cytochalasin D-mediated disruption of the actin cytoskeleton interferes with the increase in the phosphorylation of FAK at Ser-910 induced by bombesin.
Quiescent Swiss 3T3 cells were exposed for 2 h to increasing concentrations of cytochalasin D and then stimulated with bombesin for another 30 min. Cell lysates were then immunoprecipitated with an anti-FAK antibody, and the immunoprecipitates were analyzed by Western blotting with either a specific anti-Tyr(P) antibody or with the antibody that recognizes phosphorylated FAK at Ser-910. As shown in Fig. 2A, treatment with cytochalasin D completely blocked the tyrosine phosphorylation of FAK induced by bombesin in a concentrationdependent manner. Maximal inhibitory effect was achieved at 2.4 M, a concentration that completely disrupts the actin cytoskeleton and the assembly of focal adhesions (7). In striking contrast, a similar treatment with cytochalasin D did not induce any significant effect on FAK phosphorylation at Ser-910 in response to bombesin (Fig. 2B). Dissolution of actin filaments can also be induced by preventing integrin-mediated organization of the actin cytoskeleton (i.e. by suspending cells in serum-free medium). To substantiate the results obtained with cytochalasin D, we also examined FAK phosphorylation at Ser-910 in suspended cells. As seen in Fig. 2B

FIG. 1. Time-course and dose response of bombesin (Bom) stimulation of FAK phosphorylation at Ser-910 and time course of FAK phosphorylation at Ser-722 in Swiss 3T3 cells.
Upper panel, confluent and quiescent Swiss 3T3 cells were treated at 37°C with 10 nM bombesin for various times as indicated and subsequently lysed in RIPA buffer. FAK phosphorylation at Ser-910 was analyzed by immunoprecipitation using anti-FAK antibody C-20 followed by Western blotting with pS910 Ab. Middle panel, confluent and quiescent cells were treated at 37°C for 30 min in the absence or presence of various concentrations of bombesin as indicated and subsequently lysed in RIPA buffer. FAK phosphorylation at Ser-910 was analyzed by immunoprecipitation using anti-FAK antibody C-20 followed by Western blotting with pS910 Ab. In all cases, the membranes were analyzed further by Western blotting using anti-FAK Ab. The position of FAKpS 910 and FAK are indicated by arrows. The autoradiograms shown are representative of at least three independent experiments. Quantification of FAK phosphorylation at Ser-910 was determined by densitometric scanning of the bands. Values shown are the mean Ϯ S.E. of at least three independent experiments and are expressed as the percentage of the maximal increase in FAK serine phosphorylation value above control (unstimulated) values. Lower panel, confluent and quiescent Swiss 3T3 cells were treated at 37°C with 10 nM bombesin for various times as indicated and subsequently lysed in RIPA buffer. FAK phosphorylation at Ser-722 was analyzed by immunoprecipitation using anti-FAK antibody C-20 followed by Western blotting with pS722 Ab. In all cases, the membranes were analyzed further by Western The findings presented in Fig. 2 indicate that, in striking contrast to the increase in tyrosine phosphorylation, bombesin induces a dramatic increase in the phosphorylation of FAK at Ser-910 through a pathway that does not require an intact actin organization and the localization of FAK to focal adhesions.
Bombesin Induces FAK Phosphorylation at Ser-910 through a PKC-dependent Pathway-Agonist stimulation of the bombesin GPCR activates the ␣-subunit of G q , which stimulates the ␤ isoforms of phospholipase C, which catalyze the production of inositol 1,4,5-trisphosphate that triggers the release of Ca 2ϩ from internal stores, and diacylglycerol that activates the classical and novel isoforms of PKC (reviewed in Ref. 51). Our previous studies demonstrated that treatment with cytochalasin D, at concentrations that disrupt the organization of the actin cytoskeleton and completely prevent the increase in the tyrosine phosphorylation of FAK, does not interfere with agonist-induced phosphoinositide hydrolysis, mobilization of Ca 2ϩ from internal stores, and activation of PKC. These findings prompted us to determine whether the increase in FAK Ser-910 phosphorylation in response to bombesin is mediated through a PKC-dependent pathway.
To determine whether direct activation of PKC increases FAK phosphorylation at Ser-910, quiescent Swiss 3T3 cells were treated with 100 nM PDB for various times and lysed. The lysates were immunoprecipitated with anti-FAK antibody, and the resulting immunoprecipitates were analyzed by Western blotting using the site-specific antibody that recognizes the phosphorylated state of FAK at Ser-910. As shown in Fig. 3, treatment with PDB caused a dramatic, time-dependent increase in FAK phosphorylation at Ser-910. PDB stimulated FAK Ser-910 phosphorylation in a concentration-dependent manner; maximal effect was achieved at 100 nM. Western blotting with anti-FAK antibody showed that the recovery of FAK from cell lysates was not altered by treatment with PDB. In addition, exposure to the membrane-permeant diacylglycerol 1-oleoyl-2-acetyl-sn-glycerol, an analog of the endogenous activator of PKC, also increased FAK phosphorylation at Ser-910 (Fig. 3B, inset). Collectively, the results presented in Fig. 3 suggested that PKC activation provides a potential mechanism

FIG. 2. Role of cytoskeleton in FAK phosphorylation at Ser-910 induced by bombesin in Swiss 3T3 cells. A, confluent and quiescent
Swiss 3T3 cells were treated at 37°C for 2 h in the absence (0) or in the presence of increasing concentrations of cytochalasin D (Cyt D), as indicated, and then stimulated without or with 10 nM bombesin for a further 30 min and subsequently lysed in RIPA buffer. Tyrosine phosphorylation of FAK was analyzed by immunoprecipitation using anti-FAK antibody C-20 followed by Western blotting with anti-Tyr(P) Ab 4G10. The membrane was analyzed further by Western blotting using anti-FAK Ab. B, confluent and quiescent cells were treated at 37°C for 2 h in the absence or in the presence of increasing concentrations of cytochalasin D, as indicated, and then stimulated without or with 10 nM of bombesin for 30 min and subsequently lysed in RIPA buffer. FAK phosphorylation at Ser-910 was analyzed by immunoprecipitation using anti-FAK antibody C-20 followed by Western blotting with pS910 Ab, and the membrane was analyzed further by Western blotting using anti-FAK Ab. An aliquot (10%) of the lysates was analyzed by Western blotting with phospho-p44/42 MAPK Ab (pERK). These membranes were further analyzed by Western blotting using anti-ERK 2 Ab. Inset, Swiss 3T3 cells held in suspension were treated at 37°C in the absence or presence of 10 nM bombesin for 30 min (or 5 min) and subsequently lysed in RIPA buffer. FAK phosphorylation at Ser-910 was analyzed by immunoprecipitation using anti-FAK antibody C-20 followed by Western blotting with pS910 Ab. An aliquot (10%) of the lysates where the suspended cells were incubated with bombesin for 5 min was analyzed by Western blotting with phospho-p44/42 MAPK Ab (pERK), and the membrane was further analyzed by Western blotting using anti-ERK 2 Ab. The positions of FAKpTyr, FAKpS 910 , FAK, pERK, and ERK2 are indicated by arrows. The autoradiograms shown are representative of at least three independent experiments. Quantification of FAK phosphorylation at Tyr and Ser-910 were determined by densitometric scanning of the bands. Values shown are the mean Ϯ S.E. of at least three independent experiments and are expressed as the percentage of the maximal increase in FAK serine or tyrosine phosphorylation value above control (unstimulated) values. leading to FAK phosphorylation at Ser-910. Consequently, the following experiments were designed to examine the role of PKC in bombesin-induced FAK Ser-910 phosphorylation.
Quiescent cultures of Swiss 3T3 cells were treated with GF I (also known as GF 109203X or bisindolylmaleimide I) or Ro 31-8220, potent inhibitors of phorbol ester-sensitive isoforms of PKC (7,(52)(53)(54), before bombesin or PDB stimulation. As shown in Fig. 4, treatment of the cells with GF I or Ro 31-8220 potently blocked FAK Ser-910 phosphorylation induced by the subsequent addition of either bombesin or PDB. In order to substantiate the results obtained with GF I and Ro 31-8220, we examined whether chronic exposure to PDB to down-regulate the classic and novel isoforms of PKC also prevents FAK phosphorylation at Ser-910 in response to bombesin and PDB. As illustrated in Fig. 4D, treatment of intact 3T3 cells with PDB for 40 h profoundly inhibited FAK Ser-910 phosphorylation induced by the subsequent addition of bombesin or PDB. Thus, the results shown in Fig. 4 indicate that FAK phosphorylation at Ser-910 in response to bombesin or PDB is mediated through a PKC-dependent pathway.
The residues surrounding FAK Ser-910 (IESPPP) do not conform to a consensus PKC phosphorylation site but rather conform to the minimal MAPK consensus site: a serine directly followed by a proline. This prompted us to consider the possibility that the ERKs could be implicated in the phosphorylation of FAK at Ser-910 in bombesin-stimulated cells. Bombesin is known to induce a striking activation of p42 MAPK (ERK-2) and p44 MAPK (ERK-1), a response mediated through PKC in Swiss 3T3 cells (43,55,56). We verified that under our experimental conditions, treatment of the cells with either GF I or Ro 31-8220 or chronic exposure to PDB not only blocked FAK Ser-910 phosphorylation but also prevented ERK activation (Fig. 4). Furthermore, in agreement with our previous results (43), we also demonstrated that bombesin-induced ERK activation, like FAK phosphorylation at Ser-910, is not prevented by treatment with cytochalasin D at concentrations that potently inhibited tyrosine phosphorylation of FAK (Fig. 2B). In addition, bombesin induced both FAK phosphorylation at Ser-910 and ERK activation in cells held in suspension (Fig. 2B, right panel). Collectively, these results prompted us to test whether bombesin induces FAK Ser-910 phosphorylation via an ERK-dependent pathway.
Bombesin Induces FAK Phosphorylation at Ser-910 through ERK-In order to determine whether FAK Ser-910 phosphorylation is mediated through an ERK-dependent pathway, we used the selective MEK inhibitor U0126. This compound has been shown to inhibit MEK1, MEK2 (57,58), and the closely related MEK5 (59, 60) but does not inhibit MEK3, MEK4, MEK6, MEK7, PKC, protein kinase A, PDK1, Raf-1, or other tested serine/threonine kinases. We have previously found that ERK1/2 activation by bombesin is abrogated by prior exposure to U0126 in a concentration-dependent manner. A detectable inhibition was seen at a concentration as low as 0.5 M, and complete blockade was obtained at 5 M U0126 (61).
In order to determine whether ERK activation is required for FAK phosphorylation at Ser-910, quiescent cultures of Swiss 3T3 cells were incubated for 1 h in the absence or presence of increasing concentrations of U0126 (0.5-10 M) and subsequently stimulated with either 200 nM PDB or 10 nM bombesin. As shown in Fig. 5, treatment with U0126 dramatically inhib- ited FAK phosphorylation at Ser-910 in a concentration-dependent fashion. A marked inhibition (ϳ60%) was observed at a concentration as low as 0.5 M, and maximal inhibition (ϳ90 -95%) was achieved at 10 M.
To substantiate the results obtained with U0126, we examined whether phosphorylation of FAK at Ser-910 induced by PDB or bombesin in Swiss 3T3 cells is also prevented by treatment with the MEK inhibitor PD98059 (62). As illustrated by the insets of Fig. 5, exposure to PD98059 (5-50 M) markedly attenuated the increase in FAK phosphorylation at Ser-910. In contrast, inhibition of a variety of other pathways including phosphatidylinositol 3-kinase with wortmannin, Rho-associated kinase with HA 1077, p70S6K with rapamycin, EGF receptor with PD168393, and disruption of the cytoskeleton with cytochalasin D did not interfere with FAK Ser-910 phosphorylation. We verified, in parallel cultures, that treatment with either GF1 or U0126 strikingly prevented the phosphorylation of FAK at Ser-910 (Fig. 6).
In order to substantiate further the specificity of the inhibitory effects of U0126, we also examined the effect of the structurally related compounds U0125 and U0124. U0125 is 10-fold less potent than U0126 in inhibiting MEK, whereas U0124 is not active. As shown in Fig. 6, treatment with U0126 at either 1 or 10 M completely prevented both FAK phosphorylation at Ser-910 and ERK activation induced by bombesin, whereas U0124 at identical concentrations did not produce any inhibitory effect. A detectable inhibition was obtained by treatment with U0125 at 10 M but not at 1 M. These results indicate that bombesin and PDB induce FAK phosphorylation at Ser-910 through a MEK-dependent pathway.
LPA Induces Phosphorylation of FAK at Ser-910 -The striking increase in FAK phosphorylation at Ser-910 in response to bombesin prompted us to examine whether this phosphorylation is also elicited by other GPCR agonists. LPA, a major bioactive lipid of serum, elicits a broad spectrum of biological responses including cell proliferation and differentiation (63). LPA binds to a seven-transmembrane domain receptor(s) and activates several heterotrimeric G proteins that are responsible for transducing LPA signals into multiple biological responses (64,65). LPA induces Ras activation, leading to stimulation of the ERKs via a PTx-sensitive pathway that involves the ␤␥ subunits of G i (66 -70).
In order to determine the effect of LPA on FAK phosphorylation at Ser-910, quiescent Swiss 3T3 cells were stimulated with 2 M LPA for various times and lysed. Cell extracts were immunoprecipitated with anti-FAK antibody, and the phosphorylation of FAK at Ser-910 was determined in the resulting immune complexes by Western blot analysis. As shown in Fig.  7A, LPA induced a striking increase in the phosphorylation of FAK at Ser-910 in a time-dependent fashion, reaching a maximum within 5 min of exposure to this agonist.
In order to determine whether the activation of the ERKs is required for FAK phosphorylation at Ser-910, quiescent cultures of Swiss 3T3 cells were preincubated for 1 h in the absence or presence of U0126, U0125, and U0124 (at 10 M) and subsequently stimulated with LPA for 5 min. As shown in Fig. 7B, treatment with U0126 completely prevented both FAK phosphorylation at Ser-910 and ERK activation induced by LPA, whereas U0124 at an identical concentration did not produce any inhibitory effect. The addition of U0125 attenuated both FAK phosphorylation at Ser-910 and ERK activation.
As indicated previously, LPA induces ERK activation through a PTx-sensitive pathway. In contrast, previous studies demonstrated that LPA-induced tyrosine phosphorylation of FAK is not prevented by treatment with PTx, indicating that it is mediated by other G protein pathways, most likely of the G 12 family. Thus, we hypothesized that treatment with PTx should produce a differential inhibition of tyrosine and serine phosphorylation of FAK. Swiss 3T3 cells were treated with PTx (1-100 ng/ml) for 4 h and then challenged with 5 M LPA. As shown in Fig. 7C, treatment with PTx dramatically inhibited FAK phosphorylation at Ser-910 and ERK activation in a dose dependent manner, but it did not interfere with FAK phosphorylation at Tyr-397, the major site of FAK autophosphorylation.
The results presented in Fig. 7 indicate that stimulation with the G i -coupled agonist LPA also strikingly increases the phosphorylation of endogenous FAK at Ser-910 via ERK in Swiss 3T3 cells.

EGF Stimulates FAK Phosphorylation at Ser-910 via ERK-
The stimulation of the EGF receptor is known to induce ERK activation via SOS-Grb2-mediated accumulation of Ras-GTP, which then recruits Raf to the plasma membrane and activates a kinase cascade comprising Raf, MEK1/2, and ERK1/2 (71,72). Interestingly, EGF receptor can form a molecular complex with FAK (73), and FAK expression is necessary for EGFstimulated cell migration (74). EGF has been shown to induce tyrosine phosphorylation of FAK in Swiss 3T3 cells (19), but the effect of this growth factor on FAK serine phosphorylation had not been investigated.
In order to examine whether EGF promotes FAK Ser-910 phosphorylation, cultures of Swiss 3T3 cells were treated with EGF at 10 ng/ml for various times (1-60 min), and then FAK phosphorylation at Ser-910 was determined by Western blotting. As shown in Fig. 8, EGF induced a striking increase in FAK Ser-910 phosphorylation, which was detectable 5 min after the addition of the factor and was nearly maximal after 10 min.
EGF induces activation of ERK pathways via Ras-dependent but PKC-independent pathways. Accordingly, treatment with the PKC inhibitor GF1 did not prevent EGF-induced FAK phosphorylation at Ser-910 (Fig. 8B). In contrast, an identical treatment with GF1 blocked bombesin-induced FAK Ser-910 phosphorylation, in agreement with the results presented in Fig. 4. The results illustrated by Fig. 8B also demonstrate that exposure to U0126 completely abrogated FAK Ser-910 phosphorylation induced by either EGF or bombesin.
To substantiate the results shown in Fig. 8B, quiescent cultures of Swiss 3T3 cells were preincubated for 1 h in the absence or presence of U0126, U0125, and U0124 (10 M) and subsequently stimulated with EGF for 5 min. As shown in Fig.  8C, treatment with U0126 completely prevented both FAK phosphorylation at Ser-910 and ERK activation induced by The membranes were further analyzed by Western blotting using anti-FAK Ab. A portion (10%) of the lysates was analyzed by Western blotting using the antibody that detects the dually phosphorylated form of ERK1/2. These membranes were further analyzed by Western blotting using ERK2 Ab. C, quiescent Swiss 3T3 cells were treated at 37°C with increasing concentrations of PTx (1-100 ng/ml) for 4 h. The cells were then stimulated with 5 M LPA for 5 min. The cells were lysed, and the extracts were analyzed by immunoprecipitation with anti-FAK (C-20) Ab followed by Western blotting with pS910 Ab or pY397 Ab. The membranes were further analyzed by Western blotting using anti-FAK Ab. A portion (10%) of the lysates were analyzed by Western blotting using the antibody that detects the dually phosphorylated form of ERK1/2. These membranes were further analyzed by Western blotting using ERK2 Ab. The positions of FAKpS 910, FAKpY 397 , FAK, pERK, and ERK2 are indicated by the arrows. EGF, whereas U0124 at an identical concentration did not produce any inhibitory effect. The addition of U0125 attenuated both FAK phosphorylation at Ser-910 and ERK activation in response to EGF. Thus, EGF induces FAK Ser-910 phosphorylation through a PKC-independent but ERK-dependent pathway.
ERK2 Catalyzes FAK Phosphorylation at Ser-910 in Vitro-The preceding results with multiple agonists prompted us to determine whether Ser-910 of FAK can serve as a direct substrate for activated ERK. In order to test this possibility, FAK was immunoprecipitated from lysates of nonstimulated Swiss 3T3 cells, and the resulting immunocomplexes were incubated in the absence or presence of purified active ERK2 and with or without ATP. After 30 min, the reactants were analyzed by Western blotting using the antibody that detects the phosphorylated state of FAK at Ser-910. As shown in Fig. 9A, Ser-910 of FAK was readily phosphorylated by ERK2 in vitro when FAK immunocomplexes were incubated with both ERK2 and ATP. FAK phosphorylated in vitro by ERK2 migrated in SDS-PAGE with the same apparent molecular mass of FAK phosphorylated within Swiss 3T3 cells treated with PDB (Fig. 9A). In contrast, no immunoreactivity of FAK Ser-910 was detected when either ERK2 or ATP were omitted from the reaction mixture. ERK2 phosphorylated FAK at Ser-910 in vitro in a concentration-and time-dependent fashion (Fig. 9B).
Although the results shown in Fig. 9 (A and B) suggested that Ser-910 of FAK can be phosphorylated by ERK, it was still possible that FAK immunoprecipitated from cells could be in a complex with an additional protein kinase that could phosphorylate Ser-910 of FAK in an ERK-dependent manner. In order to eliminate this possibility, we performed this experiment using the C-terminal, noncatalytic domain of FAK, termed FRNK, as a fusion protein (GST-FRNK), which should contain a phosphorylation site equivalent to Ser-910 of FAK. We also generated a mutant recombinant FRNK in which the equivalent of Ser-910 was converted to Ala. GST-FRNK and GST-FRNKS910A migrated at the expected apparent molecular mass of 57 kDa (Fig. 9C). As shown in Fig. 9C, recombinant activated ERK directly phosphorylated recombinant wild-type FRNK, as shown by Western blotting using the antibody that detects the phosphorylated state of FAK at Ser-910. Mutation of the putative phosphorylatable site to Ala completely blocked the phosphorylation by ERK, demonstrating that ERK phosphorylates FRNK at the expected site. This result provides a convenient additional check of the specificity of the phosphoantibody used in our study.
In order to determine the stoichiometry of phosphorylation of FRNK by ERK, GST-FRNK was incubated with activated ERK in the presence of [␥-32 P]ATP, and then the reaction mixture was analyzed by SDS-PAGE and autoradiography to determine the level of recombinant FRNK phosphorylation. As shown in Fig. 9D, ERK catalyzed a striking incorporation of [␥-32 P]P i into GST-FRNK. However, it was possible that ERK phosphorylates GST-FRNK at several sites.
In order to verify that ERK phosphorylated GST-FRNK at a residue equivalent to Ser-910 of FAK, we also performed an identical reaction using GST-FRNKS910A instead of GST-FRNK as a substrate. As shown in Fig. 9D, mutation of the putative phosphorylatable site to Ala markedly reduced the incorporation of [␥-32 P]P i into recombinant FRNK catalyzed by ERK, indicating that ERK phosphorylates FRNK at the expected site. To assess the stoichiometry of ERK-catalyzed FRNK phosphorylation at Ser-910, we subtracted the P i incorporated into GST-FRNKS910A from that incorporated into GST-FRNK and calculated that 0.8 Ϯ 0.07 mol (mean Ϯ S.E.; n ϭ 5) of phosphate were incorporated/mol of GST-FRNK.
To assess the approximate stoichiometry of FAK phosphorylation at Ser-910 in response to bombesin stimulation in intact cells, lysates prepared from control cells and from bombesintreated cells were immunoprecipitated with the antibody that recognizes FAK phosphorylated at Ser-910, and the supernatants were then immunoprecipitated with an anti-FAK antibody and analyzed by immunoblotting with anti-FAK antibody (Fig. 10A). We found that only ϳ2% of the total FAK was phosphorylated at Ser-910 in the control, unstimulated cells. In contrast, ϳ35% of the total FAK was phosphorylated at Ser-910 after 30 min of bombesin stimulation (Fig. 10A). Given that the phosphorylation of FAK Ser-910 in intact cells reflects an equilibrium between phosphorylation-dephosphorylation reactions, we did not expect the high stoichiometry achieved in vitro (ϳ0.8; see above).
FAK Phosphorylation at Ser-910 Regulates Complex Formation with Paxillin-The COOH-terminal 150 residues of FAK functions as the FAT sequence and is responsible for localizing FAK to the focal adhesions in the cell (75,76). FAT contains binding sites for the focal adhesion proteins paxillin and talin, which appear to cooperate in mediating the localization of FAK to focal adhesions. Recently, paxillin has been shown to play a critical role in development and in signaling cell motility (77). Ser-910 is located close to the paxillin-binding region of FAK, which corresponds to residues 919 -1042 of this enzyme, as determined by deletion mutagenesis. Consequently, we hypothesized that one of the functions of ERK-mediated FAK phosphorylation at Ser-910 could be the regulation of paxillin binding to FAK. To examine this hypothesis, Ser-910 of FAK was mutated to the nonphosphorylatable residue alanine (FAK-S910A), and wild type and mutant FAK were transiently expressed in COS-7 cells. Lysates of the transfected cells were subjected to immunoprecipitation with FAK antibody, and the resulting immunocomplexes were immunoblotted with antibodies that detect FAK phosphorylated at Ser-910, FAK or paxillin. As shown in Fig. 10B, FAK wild type and mutant were expressed to the same level in this system. The wild type protein was phosphorylated at Ser-910, whereas no signal was detected in the immunoprecipitates of the mutant FAK-S910A.
The salient feature of the results shown in Fig. 10B is that the association of paxillin with FAK was markedly increased in the mutant as compared with wild type FAK, implying that phosphorylated state of Ser-910 of FAK inhibits the binding of paxillin to FAK. The increase in paxillin binding to FAK-S910A as compared with wild type FAK was ϳ3.5-fold (Fig. 10). These co-immunoprecipitation results are consistent with the hypothesis that one of the functions of ERK-mediated FAK phosphorylation at Ser-910 is to modulate the association of the FAT region to its signaling partners. properties of cancer cells (47). A large body of evidence generated during the last decade has documented that a rapid increase in the tyrosine phosphorylation of the nonreceptor tyrosine kinase FAK is a prominent early event in cells stimulated by multiple signaling molecules that regulate cell proliferation, migration, and apoptosis (3)(4)(5). In striking contrast, virtually nothing is known about the regulation of FAK phosphorylation at serine residues by GPCR agonists, tumor promoters, or growth factors.
Recently, Ser-910, in the COOH-terminal region of FAK, has been identified as a prominent phosphorylation site (49). In the present study, we have used antibodies that detect the phosphorylated state of Ser-910 of FAK to elucidate whether the phosphorylation of this residue is modulated by a variety of external stimuli. Our results demonstrate that stimulation of Swiss 3T3 cells with the GPCR agonist bombesin or the potent tumor-promoting phorbol ester PDB induced a rapid and dramatic increase in FAK phosphorylation at Ser-910. Our results also indicate that the inducible phosphorylation at Ser-910 can be separated from the constitutive phosphorylation of FAK at Ser-722, implying the existence of different pathways leading to these novel phosphorylation events.
Agonist-mediated increase in FAK tyrosine phosphorylation is accompanied by Rho signaling, leading to profound alterations in the organization of the actin cytoskeleton and in the assembly of focal adhesions, the distinct areas of the plasma membrane where FAK is localized. Treatment of the cells with cytochalasin D, which disrupts actin filaments and focal adhesion assembly, prevents the increase in FAK tyrosine phosphorylation in response to multiple agents, suggesting a mechanism involving the actin cytoskeleton and the focal adhesion plaques. In contrast, the results presented here show that bombesin induces FAK phosphorylation at Ser-910 in cytochalasin D-treated cells, indicating that agonist-induced FAK phosphorylation at tyrosine residues and Ser-910 is mediated by different pathways. In agreement with this conclusion, bombesin also induces FAK Ser-910 phosphorylation in Swiss 3T3 cells placed in suspension, a condition that also impairs the increase in FAK tyrosine phosphorylation in response to bombesin.
Our previous studies demonstrated that treatment with cytochalasin D does not inhibit inositol phosphate production, Ca 2ϩ mobilization, stimulation of PKC, and ERK activation in response to bombesin or other agonists (7, 16 -18, 29, 30, 50). These findings prompted us to examine the possibility that, in contrast to tyrosine phosphorylation, FAK Ser-910 phosphorylation is mediated through a PKC-dependent pathway. We found that treatment of Swiss 3T3 cells with either the potent tumor-promoting phorbol ester PDB or the diacylglycerol analog 1-oleoyl-2-acetyl-sn-glycerol induced a striking increase in FAK Ser-910 phosphorylation. These results imply that PKC is a potential pathway leading to FAK Ser-910 phosphorylation in bombesin-stimulated cells. In agreement with this hypothesis, we demonstrate here that either treatment with PKC inhibitors or down-regulation of phorbol ester-sensitive PKC isoforms prevented FAK phosphorylation at Ser-910 in response to bombesin or PDB. These findings raise the attractive possibility that FAK integrates GPCR signals delivered through G␣ 12 and G␣ 13 , leading to Rho-mediated signaling, cytoskeletal reorganization, and phosphorylation at tyrosine residues with signals transmitted via G q , leading to PKC-dependent phosphorylation at Ser-910. These considerations provide a novel angle to the hypothesis that FAK functions as a point of convergence and integration in the action of multiple signals.
The amino acid sequences surrounding FAK Ser-910 (or equivalent) in the human, mouse, rat, chicken, or Xenopus enzymes (IESPPP) do not conform to a consensus PKC phosphorylation site but rather conform to the minimal MAPK consensus site: a serine directly followed by a proline. Consequently, we considered the possibility that the ERKs mediate the phosphorylation of FAK at Ser-910 in agonist-stimulated cells. In line with this notion, bombesin and PDB induce robust ERK activation, like FAK Ser-910 phosphorylation, through a PKC-dependent pathway in Swiss 3T3 cells. Furthermore, treatment of cells with U0126, a potent inhibitor of MEKmediated ERK activation, prevented FAK phosphorylation at Ser-910 induced by bombesin or PDB, whereas the structurally related but biologically inactive analog U0124 did not exert any inhibitory effect.
In order to substantiate the hypothesis that ERK activation mediates FAK phosphorylation at Ser-910, we examined whether this residue of FAK becomes phosphorylated in response to other stimuli that induce ERK activation through different signal transduction pathways. LPA binds to ubiquitous seven-transmembrane domain receptor(s) of the Edg/LP subfamily of GPCRs (82,83) and activates multiple heterotrimeric G proteins that are responsible for transducing LPA signals into a broad spectrum of biological responses (64,65). Specifically, LPA stimulates Ras activation, leading to stimu- Swiss 3T3 cells were then incubated for 30 min with or without 10 nM bombesin (Bom). The cells were lysed, and the extracts were immunoprecipitated (I.P.) with either the antibody that detects the phosphorylated state of Ser-910 of FAK or with an antibody that recognizes PKC⑀, used as a control. The supernatants obtained after these immunoprecipitations were then immunoprecipitated with anti-FAK (C-20 Ab). The immunoprecipitates were then analyzed by Western blotting with anti-FAK Ab (indicated by . The percentage of total FAK phosphorylated on Ser-910 was determined by densitometric scanning of the bands. Values shown are the mean Ϯ S.E. of five independent experiments. B, phosphorylation of FAK Ser-910 decreases complex formation between FAK and paxillin. FLAG-CMV-FAKwt and the mutant construct FLAG-CMV-FAK-S910A were transiently expressed in COS-7 cells. After 2 days, the cells were lysed in RIPA buffer. FAK was immunoprecipitated using the anti-FAK antibody 2A7 and washed three times. Phosphorylation was detected by Western blotting with pS910. The membranes were further analyzed by Western blotting using anti-FAK Ab and paxillin Ab. The positions of FAKpS 910 , FAK, and paxillin are indicated by the arrows. Quantification of FAK-paxillin association was determined by densitometric scanning of the bands. Values shown are the mean Ϯ S.E. of at least six independent experiments and are expressed as the -fold increase in FAK-paxillin complex formation of the mutant above the control (wild type (wt) FAK) value. lation of Raf, MEK, and the ERKs via a PTx-sensitive pathway that involves the ␤␥ subunits of G i (66 -70). In contrast, LPA induces PTx-insensitive stress fiber formation, assembly of focal adhesions, and tyrosine phosphorylation of FAK (16) via activation of G␣ 13 and Rho (84,85). The results presented here show that LPA stimulates FAK Ser-910 phosphorylation through a PTx-sensitive pathway. Similarly, EGF, which induces Raf, MEK1/2, and ERK1/2 activation via SOS-Grb2mediated accumulation of Ras-GTP (71,72), also stimulates FAK phosphorylation at Ser-910. The phosphorylation of FAK Ser-910 induced by either LPA or EGF was abrogated by treatment with U0126, whereas the inactive analog U0124 did not exert any inhibitory effect.
It is noteworthy that cell treatment with U0126 drastically attenuated FAK phosphorylation at Ser-910 at concentrations as low as 0.5-1 M, suggesting that ERK1/2 rather than ERK 5, which is inhibited at higher concentrations of U0126 (86), mediate phosphorylation of FAK at Ser-910. Furthermore, we demonstrated that incubation of activated ERK2 either with FAK immunoprecipitated from cell lysates or with recombinant FRNK promotes phosphorylation at Ser-910 in vitro. These results indicate that activated ERK can directly phosphorylate FAK at Ser-910. Collectively, our results indicate that FAK is a substrate of the ERKs and thus establish a novel link between these important signal transduction pathways.
The biological activity of FAK depends on its localization to focal adhesions, intracellular complexes that play a critical role in signal transduction. The COOH-terminal noncatalytic domain, specifically the COOH-terminal 150 residues, functions as the FAT sequence of FAK and is responsible for localizing FAK to the focal adhesions in the cell (75,76). In addition, the COOH-terminal domain serves as a docking site for a number of cytoskeletal and signaling molecules. These FAK-binding partners include talin, paxillin, GRAF (GTPase regulator associated with FAK), and p130 cas (35,38,(87)(88)(89)(90)(91)93). In addition, phosphorylation of FAK at Tyr-925 creates binding sites for Src homology 2 domain-containing signaling molecules like Grb2 (25,94,95). Based on these observations, the C-terminal domain of FAK functions in promoting subcellular localization and in the transmission of downstream signaling by binding and recruiting signaling and adaptor proteins. In this context, we hypothesize that the dramatic increase in ERK-mediated phosphorylation of FAK at Ser-910 could function in the regulation of the assembly of signaling complexes. In support of this hypothesis, we found that mutation of Ser-910 of FAK to alanine markedly increased complex formation between FAK and paxillin. An attractive possibility is that ERK-mediated FAK phosphorylation at Ser-910 induces a conformational change in the four-helix bundle structure of FAT (92,93), thereby inhibiting the ability of FAK to associate with paxillin.
In conclusion, our results identify a previously unrecognized regulation of serine FAK phosphorylation by GPCR agonists, tyrosine kinase receptors, and phorbol esters and provide a previously unrecognized mechanism by which ERK-dependent inputs could modulate FAK signaling.