Focal Adhesion Kinase Facilitates Platelet-derived Growth Factor-BB-stimulated ERK2 Activation Required for Chemotaxis Migration of Vascular Smooth Muscle Cells*

The focal adhesion (FAK) non-receptor protein-tyrosine kinase (PTK) links both extracellular matrix/integrin and growth factor stimulation to intracellular signals promoting cell migration. Here we show that both transient and stable overexpression of the FAK C-terminal domain termed FRNK (FAK-relatednon-kinase) inhibits serum and platelet-derived growth factor (PDGF)-BB-induced vascular smooth muscle cell (SMC) migration in wound healing and in vitro Boyden Chamber chemotaxis assays, respectively. Expression of FRNK, but not a point mutant of FRNK (FRNK L1034S), disrupted the formation of a complex containing both FAK and the activated PDGF-β receptor and resulted in reduced tyrosine phosphorylation of endogenous FAK at the Tyr-397 binding site for Src family PTKs. As demonstrated using FAK-deficient and FAK-reconstituted fibroblasts, FAK positively contributed to PDGF-BB-stimulated ERK2/MAP kinase activity, and in SMCs, ERK2/MAP kinase activity was required for PDGF-BB-stimulated chemotaxis. Stable expression of FRNK but not FRNK L1034S expression in SMCs lowered the extent and duration of stimulated ERK2/MAP kinase activation at low but not at high PDGF-BB concentrations. Importantly, stable expression of FRNK in SMCs did not affect SMC morphology or proliferation in culture. Because the increased migration of vascular SMCs in response to extracellular matrix proteins and growth factors contributes to neointima formation, our results show that FAK inhibition by FRNK expression may provide a novel approach to regulate abnormal vascular SMC migration in vivo.

Several vascular diseases result from neointima formation, a process that is characterized by the accumulation of vascular smooth muscle cells (SMCs) 1 and extracellular matrix (ECM) proteins in the intima of blood vessels (1). Neointima formation is triggered upon damage to the endothelial lining by the local release of chemotactic cytokines or growth factors (2) and by increased production of ECM proteins (3). Because combinations of these factors can stimulate cell division, it was originally hypothesized that neointimal hyperplasia resulted from enhanced SMC proliferation (4). However, further studies have shown that SMC migration also plays a significant role in neointima formation (5). Growth factors acting as chemotactic agents (6) and ECM molecules acting as haptotactic factors (7) can independently stimulate SMC migration. Therefore, a common signaling component that coordinates both chemotactic and haptotactic cell migration events would be a promising target for intervention strategies.
Investigations of the molecular regulation of cell migration have demonstrated an important role for the focal adhesion kinase (FAK), a protein-tyrosine kinase (PTK) that localizes to cell-substratum contact sites also known as focal adhesions (for a review see Ref. 8). Genetic support for FAK in promoting cell motility comes from studies using FAK-null fibroblasts that exhibit refractory responses to normal motility-promoting stimuli (9). Importantly, these defects are rescued by stable re-expression of FAK in FAK-null cells (10,11). Work from a number of laboratories using a variety of cell types has provided evidence that FAK promotes cell migration potentially through the activation of multiple downstream targets such as Src family PTKs (10,12) and phosphatidylinositol 3-kinase (13,14), by the increased phosphorylation of p130 Cas (15,16) or paxillin adaptor proteins (17), or through the association with other signaling proteins such as Grb7 (18) and SHP-2 (19 -21). Although no clear consensus has emerged on the molecular mechanism(s) of how FAK promotes cell migration, studies with FAK-null cells have shown that FAK is required for both integrin-and growth factor-stimulated cell motility (10,12). FAK activity is regulated by protein-tyrosine phosphatase action (16) and inhibited by the overexpression of the FAK C-terminal domain termed FRNK (FAK-related non-kinase) (22)(23)(24). Importantly, the regulated and autonomous expression of FRNK during embryonic development is under the control of alternative intronic promoter region and therefore may function as an endogenous inhibitor of FAK in vivo (22,25) In this study, we employed the transient and stable overexpression of hemagglutinin (HA)-tagged FRNK to investigate the role of endogenous FAK in promoting platelet-derived growth factor (PDGF)-BB-stimulated vascular SMC migration. Although FRNK expression did not affect PDGF receptor-␤ (PDGFr) activation or serum-stimulated cell proliferation, FRNK interfered with the PDGF-BB-stimulated recruitment of endogenous FAK to an activated PDGFr complex, decreased FAK phosphorylation at the Tyr-397 site, and reduced the extent and duration of PDGF-BB-stimulated ERK2/MAP kinase activation. Because pharmacological inhibition of ERK activity prevented efficient PDGF-BB-stimulated SMC cell migration, our results point to an important role for FAK in relaying cell motility promoting signals from the PDGFr to MAP kinases.
Cells and Transfection-Rat aortic SMCs (ATCC, CRL-2018) were grown in DMEM, 10% calf serum (CS) containing sodium pyruvate, penicillin, and streptomycin (SMC medium). Primary mouse fibroblasts derived from FAK-null mice (FAK Ϫ/Ϫ cells) (9) and FAK-reconstituted FAK Ϫ/Ϫ cells (clone DA2) were cultured as described previously (10). SMC in 10-cm dishes (migration assay) or on collagen-coated glass coverslips in 24-well plates (wound healing assay) were transfected with the indicated amounts of DNA using Effectene (Qiagen, Valencia, CA) 36 -48 h prior to experiments. Cell transfection efficiency (ϳ15%) was determined by FACS analysis on green fluorescence protein (GFP)transfected cells. For selection of stable cell lines, hygromycin-resistant populations of cells were single-cell-sorted by FACS and expanded, and clones were analyzed by blotting for expression of HA-tagged FRNK. Control cells were transfected with the pCDNA3.1 vector and selected for growth in hygromycin (150 g/ml). PDGF-BB stimulation at the indicated concentrations was performed with serum-starved (0.5% CS for 18 h) cells, and lysates were made at the indicated time after PDGF-BB addition.
Wound Healing Assay-SMCs plated onto collagen-coated (10 g/ml) glass coverslips were transfected with GFP or GFP-FRNK (0.5 g), serum-starved (18 h in 0.5% CS), wounded with a 200-l pipette tip, washed with PBS, and incubated in SMC medium. Cells were fixed 18 or 36 h after wounding with 3.7% paraformaldehyde for 15 min, washed with PBS, and permeabilized for 10 min with 0.2% saponin in PBS containing 5% normal goat serum (blocking solution). Samples were incubated for 45 min in blocking solution containing TRITC-phalloidin (Molecular Probes), washed with PBS, embedded in Vectashield (Vector), viewed with an Olympus BX60 epifluorescence microscope, and documented using Ektachrome 400 film. Pictures were taken separately for GFP and TRITC fluorescence, and the images were merged using the program Adobe Photoshop. For assays with stably-transfected cells, migration of wounded cells was evaluated after 0 and 20 h with an inverted Nikon phase-contrast microscope and photographed with Kodak TMAX-400 film.
Chemotaxis Migration Assay-MilliCell modified Boyden chamber (Millipore, Bedford, MA) migration assays were performed as described previously (12). Briefly, serum-starved cells were detached, resuspended in migration medium (DMEM containing 0.5% BSA), and counted. Both sides of MilliCell chambers were precoated with rat tail collagen (5 g/ml in PBS) overnight at 4°C, washed with PBS, and air-dried. Chambers containing serum-starved cells (1 ϫ 10 5 cells/0.3 ml) were placed in 24-well dishes containing DMEM with 0.5% BSA with or without PDGF-BB at the indicated concentrations. Transiently transfected migratory cells on the membrane underside were identified by GFP fluorescence, and the migration of stable cell lines was visualized by Crystal Violet staining (0.1% Crystal Violet, 0.1 M borate, pH 9.0, 2% EtOH) and cell counting (cells/field using a 40ϫ objective). Background cell migration in the absence of PDGF-BB (0.5% BSA only) was less than 5% of stimulated values.
Cell Growth Assay and Immunofluorescence Staining-1 ϫ 10 3 cells/ well were seeded in 96-well plates and incubated at 37°C in complete SMC medium. After 24, 48, or 72 h, cells were fixed, stained with Crystal Violet, and air-dried. Cell-associated dye was eluted with 10% acetic acid, and absorbance values were determined at 600 nm. Immunofluorescence staining of cells attached to fibronectin-coated (10 g/ ml) glass coverslips for 2 h was performed as described previously (10).
Cell Lysis, Immunoprecipitation, and Immunoblotting-Lysates were made in modified lysis buffer containing 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS as described (28). Antibodies were incubated for 4 h at 4°C and collected on protein A (Repligen, Cambridge, MA) or protein G-plus (Calbiochem) agarose beads. Precipitated protein complexes were washed twice in Triton-only lysis buffer followed by washing in HNTG buffer (50 mM Hepes, pH 7.4, 150 mM NaCl, 0.1% Triton X-100, 10% glycerol). For immunoblotting, proteins were transferred to polyvinylidene fluoride membranes (Millipore). Blots were incubated with either 1 g/ml monoclonal or a 1:1000 dilution of polyclonal antibodies for 2 h at room temperature. Bound primary antibody was visualized by enhanced chemiluminescent detection, and subsequent reprobing of membranes was performed as described previously (28).
ERK2 in Vitro Kinase Reactions-ERK-2 was immunoprecipitated with polyclonal anti-ERK2 antibodies (Santa Cruz Biotechnologies), the IPs were washed in Triton-only lysis buffer, followed by HNTG buffer, and kinase buffer (20 mM Hepes, pH 7.4, 10% glycerol, 10 mM MgCl 2 , 10 mM MnCl 2 , 150 mM NaCl) before they were incubated for 15 min at 32°C in 30 l of kinase buffer containing 2.5 g of myelin basic protein (MBP), 20 M ATP, and 10 Ci/nmol [␥-32 P]ATP (3000 Ci/mmol). Reactions were stopped with 2ϫ SDS-polyacrylamide gel electrophoresis sample buffer, resolved by SDS-polyacrylamide gel electrophoresis, stained with Coomassie Blue, visualized by autoradiography, and the radioactivity incorporated into MBP was determined by Cerenkov counting of the excised MBP bands.
Statistical Analyses-Ordinary one-way analysis of variance was used to determine the overall significance within data groups. If a significant result was obtained by analysis of variance, the Tukey-Kramer multiple comparisons t test was used to determine significance between individual groups.

Transient FRNK Expression Inhibits Wound Healing Response and PDGF-BB-stimulated Migration of Rat Aortic
SMC-To determine whether FAK promotes vascular SMC cell migration, rat aortic SMCs were transfected with an expression vector encoding green fluorescence protein (GFP) or a GFP fusion protein of the FAK-specific inhibitor, FRNK (GFP-FRNK) and analyzed in an in vitro wound healing assay ( Fig.  1). When confluent monolayers of transfected cells were analyzed for their ability to repopulate the wounded area, nontransfected and GFP-expressing cells migrated into the open space within the first 18 h (Fig. 1A) and closed the wound within 36 h (Fig. 1C). In contrast, GFP-FRNK-expressing cells remained at the initial wound margin and did not migrate into the open space after 18 h (Fig. 1B). After 36 h, GFP-FRNKexpressing cells were not found in the previously wounded area, whereas the non-transfected cells had migrated into and closed this space (Fig. 1D).
Because matrix proteins remaining in the wound area and growth factors present in the serum-containing medium produce combined motility-promoting signals in wound healing assays, the GFP-or GFP-FRNK-transfected cells were also analyzed for migratory responses in modified Boyden chamber assays where cells migrate through membrane pores in response to defined chemotactic stimuli. PDGF is a potent chemoattractant for vascular SMCs in culture and is believed to play a major role in neointima formation in atherosclerosis and in restenosis in vivo (2). GFP-transfected SMCs migrated to-ward a PDGF-BB gradient with a maximum response at 5 ng/ml, whereas GFP-FRNK-expressing cells did not readily migrate in response to the PDGF-BB stimulation ( Fig. 2A). This inhibitory effect of FRNK was dose-dependent, as increasing amounts of transfected HA-tagged FRNK led to a strong inhibition of PDGF-BB-stimulated (5 ng/ml) SMC cell motility (Fig. 2B). These results show that transient FRNK expression in SMCs disrupts serum and PDGF-BB-stimulated motilitypromoting events.
Stable Expression of FRNK Does Not Affect SMC Viability or Morphology-Because FAK has been shown to transduce extracellular matrix-stimulated survival signals (for a review see Ref. 8), FRNK-mediated interference with endogenous FAK function could potentially result in decreased SMC cell viability and thereby affect cell motility responses. To determine the effects of FRNK overexpression on SMCs, stable cell lines were generated expressing either HA-FRNK or HA-FRNK L1034S, a point mutant of FRNK that does not bind paxillin (29). As opposed to FRNK, FRNK L1034S expression in fibroblasts does not localize to focal contacts (10, 29) and does not interfere with FAK-mediated cell motility (10,12).
Several SMC clones that expressed either FRNK (FRNK-1 and -2) or FRNK L1034S (FRNK L1034S-1 and -2) were isolated and expanded (Fig. 3A). As determined by immunoblotting whole cell lysates with antibodies to the HA-tag, FRNK and FRNK L1034S were equivalently expressed in the FRNK-1, FRNK-2, and FRNK L10234S-1 clones (Fig. 3A). Reprobing these whole cell lysates with antibodies to FAK showed that neither FRNK nor FRNK L1034S expression affected endogenous FAK expression. However, sequential reprobing this blot with antibodies to the PDGFr-␤ showed that its expression was slightly decreased in both FRNK-and FRNK L1034Sexpressing cells compared with the empty vector-transfected control (pCDNA) SMCs (Fig. 3A). Additional comparisons showed that stable FRNK expression was less than the endogenous level of FAK found in the SMCs (data not shown). Contrary to previous reports of FRNK expression-promoting cell apoptosis (30,31), stable expression of FRNK in SMCs did not detectably affect cell viability. Because FRNK expression has also been shown to inhibit cell cycle progression (32,33), serum-stimulated cell proliferation comparisons were made between FRNK, FRNK L1034S, and pCDNA control SMCs (Fig.  3B). After 48 h, no significant differences in cell growth were observed between the different cell clones. After 72 h, a significant difference in the total pCDNA control SMC cell number was observed compared with both FRNK and FRNK L1034S clones (Fig. 3B). No significant differences in serum-stimulated cell proliferation were observed between FRNK-and FRNK L1034S-expressing SMCs after 72 h.
In fibroblasts, FRNK strongly localizes to focal contact sites, whereas FRNK L1034S exhibits a perinuclear distribution  (10). Indirect immunofluorescence detection of FRNK in the SMC clone FRNK-1 showed that HA-tag staining was concentrated at the tips of cell protrusions, which correspond also to the vinculin-positive sites of focal contact formation (Fig. 3C). Importantly, no specific HA-tag staining was detected in the pCDNA control SMCs that overlapped with vinculin staining of focal contact sites (Fig. 3C). Similar to results with fibroblasts, FRNK L1034S expression in SMC clone L1034S-1 was present in a perinuclear distribution and did not significantly co-localize with focal contact sites (Fig. 3C). Importantly, no significant differences were observed in either cell morphology, vinculin staining at focal contact sites, or in the general actin cytoskeletal organization in FRNK-expressing SMC cells compared with the pCDNA and FRNK L1034S-expressing SMCs (Fig.  3C). These results show that stable expression of FRNK in SMCs was associated with focal contact sites but did not negatively affect cell viability or morphology.
FRNK Expression Inhibits Specific PDGF-BB-stimulated Tyrosine Phosphorylation Events-To determine whether the stable expression of FRNK in SMCs was able to disrupt motilitypromoting signals as did transient FRNK overexpression, anti-P.Tyr blotting analyses were performed on either serumstarved or PDGF-BB-stimulated (10 ng/ml, 5 min) pCDNA control, FRNK-, or FRNK L1034S-expressing SMC clones (Fig.  4A). Compared with normal non-transfected SMCs (data not shown), the pCDNA control SMCs exhibited an identical pattern of tyrosine phosphorylation events under starved and PDGF-BB-stimulated conditions (Fig. 4A, lanes 1 and 2). Interestingly, in both the FRNK and FRNK L1034S clones, the tyrosine phosphorylation of unknown 116-to 130-kDa protein(s) were increased under both serum-starved and PDGF-BB-stimulated conditions compared with pCDNA SMCs (Fig.  4A, lanes 3-10). Although no significant difference in PDGFr-␤ tyrosine phosphorylation upon PDGF-BB addition to the FRNK or FRNK L1034S clones was observed, the stimulated tyrosine phosphorylation of ϳ150and ϳ42-kDa proteins was noticeably reduced in the FRNK-expressing compared with the pCDNA and FRNK L1034S-expressing SMCs (Fig. 4A, lanes 4  and 6). These results show that stable FRNK but not FRNK L1034S expression in SMCs selectively disrupts the cascade of PDGF-BB-stimulated tyrosine phosphorylation to distinct downstream targets.
FAK Positively Contributes to PDGF-BB-stimulated ERK Activation at Low but Not at High Growth Factor Concentrations-Previous studies have demonstrated that FAK relays integrin-initiated signals to the activation of the ERK/MAP kinase cascade (34). To determine whether the ϳ42-kDa protein showing reduced phosphorylation after PDGF-BB stimulation in the FRNK-expressing SMC clones represents ERK2/ MAP kinase, the blot shown in Fig. 4A was reprobed with antibodies recognizing only the active and dually phosphorylated form of ERK kinases (P-ERK). As shown in Fig. 4B, activation of ERK2 in response to PDGF-BB stimulation was markedly reduced in both FRNK-expressing clones when compared with pCDNA control cells or FRNK L1034S SMC clones. Importantly, equal amounts of ERK2 were present in all samples (Fig. 4B, lower panel).
Because previous studies have shown that FAK tyrosine phosphorylation is increased after low but not high concentrations of PDGF-BB addition to fibroblasts (35), the pCDNA, FRNK-1, and FRNK L1034S-1 SMC clones were stimulated with increasing concentrations of PDGF-BB for 10 min and whole cell lysates were analyzed for ERK activity by phospho-ERK blotting (Fig. 5A). Compared with the pCDNA and FRNK L1034S-1 clones, the maximal inhibitory effect of FRNK expression was observed at low 10 ng/ml motility-promoting con- centrations of PDGF-BB, whereas at higher mitogenic 50 ng/ml concentrations of PDGF-BB, similar levels of ERK activation were observed in lysates from the FRNK-1 SMC clone compared with the pCDNA control cells (Fig. 5A). Unexpectedly, the level of PDGF-BB-stimulated ERK activity in the FRNK L1034S-1 clone at 10 ng/ml was higher than the stimulated ERK level in pCDNA control SMCs (Fig. 5A) even though PDGFr-␤ expression levels are lower in the FRNK L1034S-1 clone than in the pCDNA control SMCs (Fig. 3A). Although the mechanism(s) for increased PDGF-BB stimulated ERK activity in the FRNK L1034S-1 clones is(are) not known, the observed increased tyrosine phosphorylation of ϳ116to 130-kDa proteins in these cells (Fig. 4A) may be indicative of cellular compensatory events.
Because ERK activity in the FRNK-1 SMC clone was maximally inhibited at 10 ng/ml PDGF-BB (Fig. 5A), experiments were also performed to evaluate whether FRNK expression affected the duration of PDGF-BB-stimulated ERK activation (Fig. 5B). Serum-starved pCDNA, FRNK-1, and FRNK L1034S-1 SMC clones were stimulated with 10 ng/ml PDGF-BB, and at time points between 5 and 120 min, whole cell lysates were analyzed for ERK activity by phospho-ERK blotting (Fig. 5B). Compared with both the pCDNA and FRNK L1034S-1 SMC clones, where ERK activity persisted up to 120 min, PDGF-BB-stimulated ERK activity was decreased in the early 5-to 10-min time points and not detectably activated after 60 min in the FRNK-1 SMC clone (Fig. 5B). These combined results show that stable FRNK expression in SMCs re-  ). B, lysates from SMC stimulated for the indicated times with PDGF-BB (10 ng/ml) were analyzed as in A. C, ERK2 activation is impaired in cells that lack FAK. Whole cell lysates from serum-starved or PDGF-BB-stimulated (10 ng/ml, 5 min) FAK Ϫ/Ϫ and FAK re-expressing (DA2) fibroblasts were analyzed by ERK2 blotting. The slower migrating ERK2 band represents phosphorylated and activated ERK2. ERK2 IPs from starved or PDGF-stimulated FAK Ϫ/Ϫ , and DA2 cells were assayed for in vitro kinase activity using myelin basic protein (MBP) as a substrate. Radioactivity incorporated into MBP was determined by Cerenkov counting, and values represent the means Ϯ S.D. of three determinations. Asterisks indicate ERK2 activity significantly higher in starved FAK Ϫ/Ϫ than in DA2 cells (p Ͻ 0.05). Double asterisks indicate ERK2 activity significantly lower in PDGF-BB-stimulated FAK Ϫ/Ϫ than in DA2 cells (p Ͻ 0.01). duces the extent and duration of PDGF-BB-stimulated ERK activity.
To address the issue as to whether FRNK-mediated inhibition of PDGF-BB-stimulated ERK2 activity represents an inhibition of FAK function, analyses were performed in FAK-null fibroblasts (FAK Ϫ/Ϫ ) and FAK-reconstituted fibroblasts (clone DA2) to determine whether the presence or absence of FAK affected PDGF-BB stimulated ERK2 activation (Fig. 5C). The DA2 cells were generated from an FAK Ϫ/Ϫ cell clone, and previous studies have shown that the PDGFr is expressed equivalently in both cell lines (12). Compared with DA2 cells, FAK Ϫ/Ϫ cells displayed a higher basal level of ERK2 activity under serum-starved conditions (Fig. 5C, lanes 1 and 3) and a lower level of induced ERK2 activity after PDGF-BB stimulation (Fig. 5C, lanes 2 and 4). Although the FAK-related PTK Pyk2 is highly expressed in the FAK Ϫ/Ϫ cells (27) and may contribute to the high basal ERK activity in the absence of FAK, the DA2 cell results show that FAK positively contributes to PDGF-BB-stimulated ERK2 activation.
PDGF-BB-stimulated ERK Activation Is Required for SMC Motility-To establish the connections between FRNK expression, inhibition of PDGF-BB-stimulated ERK2 activity, and cell migration, SMCs were pretreated for 45 min with increasing concentrations of the MEK/ERK kinase inhibitor PD98059 and subsequently stimulated with PDGF-BB (Fig. 6A). Although 10 -25 M of PD98059 did not interfere with PDGF-BB-stimulated tyrosine phosphorylation of the PDGFr or FAK, PD98059 inhibited PDGF-BB-stimulated ERK activation in a concentration-dependent manner (Fig. 6A). Additionally, SMCs were either pretreated with PD98059 or a specific inhibitor for p38 kinase (SB203580) and employed in modified Boyden chamber motility assays (Fig. 6B). Although treatment of SMCs with the p38 inhibitor had no effect on PDGF-BB-stimulated cell migration (data not shown), PD98059 reduced PDGF-BB-stimulated SMC chemotaxis in a similar dose-dependent range as observed for ERK2 inhibition (Fig. 6B). Taken together, these findings support the conclusion that FRNK may disrupt stimulated SMC migration in part by interfering with the FAK-dependent aspect of PDGF-BB-mediated ERK2 activation.
Stable FRNK Expression Inhibits Wound Healing Response and PDGF-stimulated Motility of Rat Aortic SMCs-To investigate the effect of stable FRNK overexpression on SMC migration, in vitro wound healing assays were performed (Fig. 7). Control pCDNA, FRNK-1, or the FRNK L1034S-1 SMC clones were grown to the same density, and then the cell monolayer was wounded with a pipette tip (Fig. 7, A, C, and E). After 20 h in the presence of serum, pCDNA SMCs (Fig. 7, B) and the FRNK L1034S-1 SMCs (Fig. 7F) exhibited cell reorientation responses along the wounded edge margin and had migrated into the wounded area. However, after 20 h, the FRNK-1 SMCs (Fig. 7D) exhibited only limited cell reorientation responses along the wounded edge margin and did not efficiently repopulate the open space. Similar results were obtained with wound healing assays using the FRNK-2 and FRNK L1034S-2 SMC clones (data not shown).
Because stable FRNK expression in the SMCs inhibited PDGF-BB-stimulated ERK activity (Figs. 4 and 5) and pharmacological inhibition of the MEK/ERK kinase cascade disrupted PDGF-BB-stimulated SMC motility (Fig. 6), the pCDNA, FRNK-, and FRNK L1034S-expressing SMCs were analyzed for PDGF-BB-stimulated (5 ng/ml) motility responses in Boyden chamber assays (Fig. 8A). Similar to the results obtained in the transient transfection experiments (Fig. 2), stable FRNK expression led to a strong reduction (8-to 10-fold) in PDGF-BB-stimulated motility compared with FRNK L1034S and pCDNA SMCs (Fig. 8A). These results show that inhibition of FAK function by FRNK but not FRNK L1034S expression results in disrupting both PDGF-BB-stimulated ERK activity and cell motility.

FRNK Blocks the Formation of a Complex between FAK and PDGFr and Promotes FAK Dephosphorylation at Tyr-397-
Although the exact molecular connections of FAK to PDGF-BBstimulated ERK activation and cell motility remain to be defined, FAK has been shown to co-immunoprecipitate with an activated complex containing PDGFr-␤ after PDGF-BB stimulation of fibroblasts (12). To determine whether a FAK⅐PDGFr-␤ complex is formed in SMCs, serum-starved pCDNA SMCs were stimulated with PDGF-BB, FAK was iso-lated by immunoprecipitation, and FAK-associated proteins were visualized by anti-P.Tyr blotting (Fig. 8B). Compared with serum-starved cells, PDGF-BB stimulation led to the association of a ϳ190-kDa P.Tyr-containing protein with FAK in SMCs (Fig. 8B). To confirm that the ϳ190-kDa tyrosine-phosphorylated protein associated with FAK upon PDGF-BB stimulation was PDGFr-␤, reciprocal co-immunoprecipitation experiments with anti-PDGFr antibodies were performed on starved or PDGF-BB-stimulated SMCs (Fig. 8C). Under serumstarved conditions, neither FAK nor other tyrosine-phosphorylated proteins were detected in association with the PDGFr-␤. However, after PDGF-BB stimulation of SMCs, tyrosine-phosphorylated FAK was found to co-immunoprecipitate with antibodies to PDGFr-␤ but not with preimmune serum from PDGF-BB-stimulated SMC lysates (Fig. 8C). These results show that an activated complex containing tyrosine-phosphorylated FAK and PDGFr-␤ is formed after PDGF-BB stimulation of SMCs.
To test whether stable FRNK expression in SMCs may interfere with the stimulated formation of a FAK-and PDGFr-␤-containing complex, antibodies specific to the FAK N-terminal domain were used to isolate endogenous FAK from either the pCDNA control, FRNK-, or FRNK L1034S-expressing SMCs after PDGF-BB stimulation (Fig. 8D). As detected by anti-P.Tyr blotting, PDGFr-␤ co-immunoprecipitated with FAK from the pCDNA and FRNK L1034S SMCs, but not detectably with FAK from FRNK-expressing SMCs (Fig. 8D). Equal levels of FAK were isolated from the pCDNA, FRNK-, and FRNK L1034S-expressing SMCs. Significantly, FAK isolated from the PDGF-BB-stimulated FRNK-1 SMCs exhibited reduced tyrosine phosphorylation levels compared with FAK isolated from pCDNA and FRNK L1034S-1 SMCs (Fig. 8D). Previous studies in fibroblasts have shown that motility-promoting concentrations of PDGF-BB promote FAK phosphorylation at Tyr-397 and the stimulated SH2-mediated binding of Src family PTKs to this site on FAK (12,36). Using a phosphospecific antibody directed to the FAK Tyr-397 site, stable FRNK expression but not FRNK L1034S resulted in the reduced phosphorylation of FAK Tyr-397 after PDGF-BB stimulation of SMCs (Fig. 8D). Taken together, our results support the conclusion that FRNK expression in SMCs acts in a receptor-proximal fashion to uncouple the stimulated linkage between FAK and PDGFr-␤, thereby resulting in the inhibition of FAK Tyr-397 phosphorylation and the reduced propagation of downstream signaling events. DISCUSSION Accumulating evidence supports a critical role for FAK in promoting cell migration stimulated by different types of cell surface ligand receptors. In this report, we investigated the involvement of FAK in PDGF-BB-stimulated chemotaxis of vascular SMCs through the transient and stable expression of a FAK-specific inhibitor comprising the FAK C-terminal domain termed FRNK (37). FRNK encompasses the focal adhesion targeting sequences (38) and proline-rich motifs that interact with SH3-containing proteins (39,40). FRNK does not possess kinase activity, nor does it contain the FAK autophosphorylation/SH2 binding site at Tyr-397. FRNK is expressed as an independent transcript in a tissue-and developmental-specific manner from a cryptic promoter located within an intron of the FAK gene (25). Because transient FRNK expression in fibroblasts promotes the displacement of FAK from focal contact sites (22), dephosphorylation of FAK at Tyr-397 (10), and inhibition of FAK-mediated cell motility (10,17,23), FRNK is believed to function as a specific inhibitor of FAK function.
We found that both transient as well as stable FRNK expression in SMCs potently inhibited PDGF-BB-induced cell migration. Importantly, although FRNK strongly localized to focal FIG. 8. Stable FRNK expression reduces PDGF-stimulated chemotaxis and FAK⅐PDGFr associations and promotes FAK dephosphorylation at Tyr-397. A, control pCDNA, FRNK-, or FRNK L1034S-expressing SMC clones were analyzed in Boyden chamber chemotaxis assays with 5 ng/ml PDGF-BB as stimulus. After 5 h, migratory cells on the underside of the membrane were fixed, stained, and counted. Bars represent means Ϯ S.D. from two independent experiments. B, lysates were made from serum-starved SMCs that were replated onto collagen-coated (10 g/ml) dishes for 2 h at 37°C. Replated cells were then stimulated with PDGF-BB (20 ng/ml, 10 min). FAK-associated proteins were visualized by anti-P.Tyr (upper panel), and equal amounts of immunoprecipitated FAK was confirmed by FAK blotting (lower panel). C, lysates were made from either serum-starved or PDGF-BB-stimulated SMCs as described in B, and IPs were made with either preimmune serum (PI) or with anti-PDGFr-␤ serum. Sequential immunoblotting with antibodies to the PDGFr (upper panel), phosphotyrosine (middle panel), or to FAK (lower panel) were used to visualize the formation of a FAK⅐PDGFr complex after PDGF-BB stimulation of SMCs. D, N-terminal domain-specific antibodies to FAK were used in IPs with lysates from collagen-replated and PDGF-BB stimulated SMCs as in B. FAK-associated proteins were visualized by anti-P.Tyr (upper panel) and sequentially followed by anti-FAK blotting (middle panel). Whole cell lysates (50 g of protein) from collagenreplated and PDGF-BB stimulated SMCs were analyzed with phosphospecific antibodies to the FAK Tyr-397 phosphorylation site (pY397; lower panel).
contact sites in SMCs, stable expression of FRNK L1034S, which contains a point mutation disrupting the paxillin binding site, did not localize to focal contact sites and did not block FAK function in PDGF-BB-stimulated SMC migration. Importantly, although FRNK overexpression has been reported to promote cell apoptosis (31) and inhibit growth factor-stimulated cell cycle progression (32,33), stable expression of FRNK within SMCs did not affect cell viability, cell proliferation rates, or alter SMC morphology. Because stable FRNK expression in the SMCs is less than the level of endogenous FAK, it is possible that the relative levels of FRNK and FAK expression within cells may differentially influence cell survival-and cell motility-promoting signaling pathways.
Interestingly, as first observed upon PDGF-BB stimulation of fibroblasts (12), FAK associated with an activated PDGFr-␤ complex upon PDGF-BB stimulation of SMCs. FRNK, but not FRNK L1034S expression, inhibited the stimulated association of FAK with a PDGFr-␤-containing complex. Although the molecular connections linking FAK to the PDGFr-␤ remain to be determined, our results showing a strong correlation between PDGFr-␤/FAK association and cell motility in SMCs support previous studies in fibroblasts whereby a stimulated complex between FAK and the epidermal growth factor (EGF) receptor was required for efficient EGF-stimulated cell migration (12). Together, these results suggest that FAK serves as a receptorproximal coordinator of cell migration connecting both growth factor receptors and integrins with motility-promoting downstream signaling events.
In SMCs, we found that FRNK expression inhibited the magnitude and duration of PDGF-BB-stimulated ERK activity at low but not at high growth factor concentrations. We also found that FAK reconstitution of FAK Ϫ/Ϫ cells enhanced PDGF-BB-stimulated ERK activity. Although previous studies with FAK signaling have focused on its role in promoting either integrin-stimulated signaling to the ERK/MAP (8,32) or JNK/ SAP kinase (26,33) cascades, our results for the first time demonstrate the importance of FAK in the transduction of signals from a growth factor receptor to ERK2/MAP kinases.
Because the molecular mechanism(s) through which FAK facilitates PDGF-BB-stimulated ERK activation are not known, ongoing investigations are focused on the potential role of FAK in promoting the preferential localization of a motilitypromoting signaling complex or the role of FAK in facilitating efficient activation of ERK at synergistic inputs below the level of Ras activation (41). Nevertheless, studies in fibroblasts have shown that the integrity of the SH2 binding site at FAK Tyr-397 and the recruitment of Src family PTKs into a signaling complex is the first of several important signaling events necessary for PDGF-stimulated cell motility (12). Stable FRNK expression in SMCs disrupted the formation of an activated complex containing FAK and the PDGFr-␤, promoted FAK dephosphorylation at Tyr-397, and resulted in reduced PDGF-BB-stimulated ERK activity. Pharmacological inhibition of ERK activation after PDGF stimulation prevented efficient SMC motility, which is consistent with previous findings (42,43). Because growth factor-stimulated ERK2/MAP kinase activation and subsequent ERK2 connections to increased myosin light chain kinase activity can facilitate cell contractility and cell migration in a variety of cell types (44), decreased PDGF-BB-stimulated ERK activity, due to FAK inhibition, could explain the reduced cell migration observed upon transient and stable FRNK expression in SMCs.
In conclusion, our studies suggest that FAK is a promising target for future therapeutic intervention strategies. We propose that techniques to deliver FRNK to SMCs in vivo or the activation of signaling pathways regulating endogenous FRNK expression might represent novel approaches to prevent excessive vascular remodeling following endovascular manipulations.