The Urokinase-type Plasminogen Activator Receptor Mediates Tyrosine Phosphorylation of Focal Adhesion Proteins and Activation of Mitogen-activated Protein Kinase in Cultured Endothelial Cells*

Urokinase-type plasminogen activator (uPA) binds to cells via a specific glycosylphosphatidylinositol-anchored receptor. Although occupancy of the uPA receptor (uPAR) has been shown to alter cellular function and to induce gene expression, the signaling mechanism has not been characterized. Urokinase induced an increase in the tyrosine phosphorylation of multiple proteins in bovine aortic endothelial cells. In contrast, low molecular weight uPA did not induce this response. Analysis by immunoblotting demonstrated tyrosine phosphorylation of focal adhesion kinase (FAK), the focal adhesion-associated proteins paxillin and p130 cas ,and mitogen-activated protein kinase (MAPK) following the occupancy of the uPAR by uPA. Treatment of cells with phosphatidylinositol-specific phospholipase C, which cleaves glycosylphosphatidylinositol-linked proteins from the cell surface, blocked the uPA-induced tyrosine phosphorylation of FAK, indicating the requirement of an intact uPAR on the cell surface. The uPA-induced activation of MAPK was completely inhibited by genistein, but not by 4-amino-5-(4-methylphenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine, a specific inhibitor of Src family kinases. Thus, this study demonstrates a novel role for the uPAR in endothelial cell signal transduction that involves the activation of FAK and MAPK, which are mediated by the receptor-binding domain of uPA. This may have important implications for the mechanism through which uPA influences cell migration and differentiation.

Urokinase-type plasminogen activator (uPA) binds to cells via a specific glycosylphosphatidylinositol-anchored receptor. Although occupancy of the uPA receptor (uPAR) has been shown to alter cellular function and to induce gene expression, the signaling mechanism has not been characterized. Urokinase induced an increase in the tyrosine phosphorylation of multiple proteins in bovine aortic endothelial cells. In contrast, low molecular weight uPA did not induce this response. Analysis by immunoblotting demonstrated tyrosine phosphorylation of focal adhesion kinase (FAK), the focal adhesion-associated proteins paxillin and p130 cas , and mitogen-activated protein kinase (MAPK) following the occupancy of the uPAR by uPA. Treatment of cells with phosphatidylinositol-specific phospholipase C, which cleaves glycosylphosphatidylinositol-linked proteins from the cell surface, blocked the uPA-induced tyrosine phosphorylation of FAK, indicating the requirement of an intact uPAR on the cell surface. The uPA-induced activation of MAPK was completely inhibited by genistein, but not by 4-amino-5-(4-methylphenyl)-7-(t-butyl)pyrazolo[3,4d]pyrimidine, a specific inhibitor of Src family kinases. Thus, this study demonstrates a novel role for the uPAR in endothelial cell signal transduction that involves the activation of FAK and MAPK, which are mediated by the receptor-binding domain of uPA. This may have important implications for the mechanism through which uPA influences cell migration and differentiation.
Urokinase-type plasminogen activator (uPA) 1 (1) is a well characterized, highly restricted serine protease that converts the zymogen plasminogen to active plasmin. Urokinase binds with high affinity to a specific cell-surface receptor (uPAR) that has been identified in multiple cells, including bovine aortic endothelial (BAE) cells (1,2). The uPAR is a highly glycosylated 55-kDa protein linked to the plasma membrane by a glycosylphosphatidylinositol (GPI) anchor. The receptor-bind-ing domain of uPA is located in the amino-terminal fragment (ATF; residues 1-135) of the uPA molecule and does not involve the protease domain (for a review, see Ref. 3). Independently of its proteolytic activity, uPA has been shown to induce the adhesion and chemotactic movement of myeloid cells (4,5), to induce cell migration in human epithelial cells (6) and bovine endothelial cells (7), and to promote cell growth (8). These latter functions are related to the occupancy of the uPAR, which is mediated by the receptor-binding domain of uPA (3). The stimulation of basic fibroblast growth factor-induced endothelial cell migration by uPA is independent of its proteolytic activity (7).
Although devoid of membrane-spanning and intracellular domains characteristic of the G protein-coupled and tyrosine kinase receptor families (9), the uPAR has been recently shown to mediate proteolysis-independent signaling events, including protein tyrosine phosphorylation (10), diacylglycerol formation (11), and the activation of a serine kinase (6). In the human kidney epithelial tumor cell line TCL-598, the uPAR was associated with activation of the JAK1/STAT1 signaling pathway (12). A similar involvement of the JAK/STAT pathway was demonstrated for vascular smooth muscle cells (13).
It has also been reported that members of the Src kinase family and ␤ 2 -integrins are components of the uPAR complex (14). Further confirmation of the ability of the uPAR to form a complex with members of the integrin family was provided by the observation that the leukocyte ␤ 2 -integrin Mac-1 (CR3) forms a unit with the uPAR on monocytes (15). The assembly of these molecules in one receptor (uPAR) complex suggests the possibility of functional cooperation. Such a degree of functional cooperation was recently demonstrated for the uPAR and vitronectin (16); formation of a uPAR-integrin-caveolin complex resulted in inhibition of ␤ 1 -integrin-dependent adhesion to fibronectin. However, the intracellular signaling events triggered by occupancy of the uPAR that lead to effects on cellular migration and adhesion are not well characterized.
Recent studies have identified a novel intracellular proteintyrosine kinase, focal adhesion kinase (FAK), that is involved in cell adhesion and/or migration (17). FAK is localized in focal adhesion contacts and becomes rapidly phosphorylated and activated upon coupling of integrins to the extracellular matrix or integrin clustering (18). Additionally, FAK acts in concert with paxillin to localize several SH2 domain-containing pro-teins, including Src, Fyn, phosphatidylinositol 3-kinase, and the carboxyl-terminal Src kinase in focal adhesion contacts (18). Moreover, an association between FAK and Grb2 has been reported, suggesting a role for FAK in connecting integrin ligand binding to downstream signaling pathways, such as activation of the mitogen-activated protein kinase (MAPK) (19). The focal adhesion-related protein p130 cas has recently been identified as a novel SH3 domain-containing signaling molecule with a cluster of multiple putative SH2 domain-binding motifs (20). The unique structure of p130 cas indicates the possible role of p130 cas in assembling signals from multiple SH2 domain-containing molecules, including Src and Crk. Additionally, it has been shown that paxillin is a substrate for FAK kinase activity in vitro and in vivo (21,22).
The similarities between the functional effects of uPAR occupancy and FAK activation prompted us to examine the effects of uPA on endothelial tyrosine phosphorylation and specifically to characterize the effects of uPA on FAK and MAPK activation in cultured endothelial cells. Our findings demonstrate a novel role for the uPAR in endothelial cell signal transduction that involves the activation of FAK and MAPK, which may explain how uPA influences cells migration and differentiation.

EXPERIMENTAL PROCEDURES
Materials-All reagents were obtained from Sigma, unless otherwise specified. Monoclonal anti-phosphotyrosine (4G10) and polyclonal and monoclonal anti-FAK (2A7) antibodies were obtained from Upstate Biotechnology, Inc. The monoclonal anti-paxillin and anti-p130 cas antibodies were purchased from Transduction Laboratories (Lexington, KY). Genistein and PP 1 were obtained from Calbiochem. Monoclonal antibody against the uPAR (mAb 3936), recombinant human uPA, recombinant human ATF, and low molecular mass uPA were obtained from American Diagnostica Inc. (Greenwich, CT). Protein A-Sepharose was obtained from Amersham Pharmacia Biotech. Chemiluminescent detection of Western blots was performed using kits obtained from Schleicher & Schuell. The PhosphoPlus TM MAPK antibody kit was purchased from New England Biolabs Inc. (Beverly, MA).
Cell Culture-BAE cells were obtained from fresh bovine aortas and harvested using 0.1% collagenase as described previously (23). Cells of passage 1 were exclusively used in these experiments. BAE cells were grown to confluence in 100-mm tissue culture dishes, washed twice with PBS, and then incubated for 2 days in serum-free medium. For measuring MAPK, cells were incubated in serum-free medium for 2 h prior to the experiment. In a preliminary set of experiments, we found that Ͼ4 h of serum depletion resulted in a decrease in uPA-induced activation of MAPK without any apparent change in the phosphorylation of focal adhesion proteins (data not shown).
Immunoprecipitation and Western Blotting-The cells were stimulated with uPA or ATF at the indicated time periods or concentrations in serum-free medium containing 0.1% bovine serum albumin. The incubation was terminated by rapid aspiration of the medium, followed by washing twice with ice-cold phosphate-buffered saline containing 1 mM sodium orthovanadate, and then lysed on ice in 0.5 ml of Nonidet P-40 lysis buffer (50 mM HEPES, pH 7.5, 1% Nonidet P-40, 50 mM NaCl, 50 mM NaF, 10 mM sodium pyrophosphate, 5 mM EDTA, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and 10 g/ml aprotinin and leupeptin). Lysates were clarified by centrifugation at 15,000 ϫ g for 10 min, and equal amounts (1 mg) of lysate proteins were immunoprecipitated with the indicated antibody overnight, followed by the addition of protein A-Sepharose for 2 h at 4°C. Immunoprecipitates were washed five times with 1 ml of Nonidet P-40 lysis buffer, solubilized in SDS-polyacrylamide gel electrophoresis sample buffer, resolved by SDS-polyacrylamide gel electrophoresis, and transferred to membranes. The blocked membranes were then incubated with the indicated antibody, and the immunoreactive bands were visualized using chemiluminescent reagents as recommended by the manufacturer. Due to the high affinity of the monoclonal anti-paxillin antibody, a dilution of 1:10,000 was utilized.
Determination of Phosphorylated (Activated) MAPK-Activation of MAPK occurs through phosphorylation of threonine and tyrosine at the sequence T*EY* by a single upstream MAPK kinase (24). BAE cells were cultured in 6-well plates and used at subconfluence. Equal amounts of cell lysates (50 g) were directly electrophoresed on a 10% SDS-polyacrylamide gel and blotted onto a polyvinylidene difluoride membrane. The blocked membranes were then incubated with a phospho-specific MAPK antibody that detects p42 MAPK and p44 MAPK only when catalytically activated by phosphorylation at Tyr-204. The phospho-specific MAPK antibody reacts equally well with Thr/Tyrdoubly phosphorylated MAPK. The membrane was stripped in ␤-mercaptoethanol stripping buffer at 60°C for 30 min (25) and reprobed with p44/p42 MAPK antibody, which detects total MAPK protein, as a control for sample variations. Immune complexes were detected by the chemiluminescent Western detection system as recommended by the manufacturer.
Visualization of Actin Filaments-Rhodamine-phalloidin (R-415) was obtained from Molecular Probes, Inc. (Eugene, OR) and stained according to the manufacturer's protocol. In brief, BAE cells were grown to 70% confluence on coverslips. The cells were washed with PBS, fixed in 3.7% formaldehyde solution in PBS, and again washed with PBS. The coverslips were placed in a glass Petri dish and extracted with acetone at -20°C, washed with PBS, and stained with rhodamine-phalloidin for 20 min (26).

RESULTS
To investigate tyrosine phosphorylation of proteins in response to uPA occupation of the uPAR, BAE cells were incubated with 50 nM uPA, ATF, or a low molecular weight uPA for the indicated time periods and lysed. Cell lysates were immunoprecipitated with anti-Tyr(P) mAb, and the immunoprecipitates were analyzed by Western blotting with the same anti-Tyr(P) mAb. As shown in Fig. 1, uPA induced an increase in the tyrosine phosphorylation of multiple proteins including bands of 120 -130 and 70 -80 kDa. The stimulatory effect was rapid and reached its maximum at 1 min and then declined after 5 min of treatment with uPA, but remained higher than the basal level. Incubation with ATF (50 nM), an amino-terminal fragment of uPA, mimicked the effects of uPA. In contrast, low molecular weight uPA, which contains the protease domain but lacks the receptor-binding domain of uPA, had no effect on protein tyrosine phosphorylation (Fig. 1A). The phosphorylation of bands in the 70 -80-and 120 -130-kDa range did not represent a nonspecific response as these bands were not observed following immunoprecipitation with a nonspecific mouse IgG (Fig. 1B). These results suggest that the uPAinduced tyrosine phosphorylation is mediated by the receptorbinding domain of uPA.
To determine whether FAK is activated following occupancy of the uPAR, samples from uPA-treated BAE cells were immunoprecipitated with polyclonal anti-FAK antibody and analyzed by immunoblotting with anti-Tyr(P) mAb (Fig. 2). Urokinase induced a dose-and time-dependent tyrosine phosphorylation of FAK. Tyrosine phosphorylation was maximal at a uPA concentration of 50 nM in this system ( Fig. 2A). Pretreatment of the cells with phosphatidylinositol-specific phospholipase C, which cleaves GPI-linked proteins from the cell surface, blocked the effect of uPA ( Fig. 2A). Treatment of cells with cytochalasin D, which disrupts the network of actin microfilaments, prevented the tyrosine phosphorylation of FAK (Fig. 2B).
The apparent specific phosphorylation of FAK in response to uPA binding prompted further experiments to determine whether other focal adhesion-related proteins were involved in this process. As shown in Fig. 3, uPA induced time-dependent tyrosine phosphorylation of paxillin and p130 cas . The maximum effect of uPA was achieved at 5 min with 50 nM uPA. The tyrosine phosphorylation of paxillin and p130 cas coincides with tyrosine phosphorylation of FAK.
Additional experiments were performed using ATF to confirm that this uPAR-binding fragment of uPA is fully capable of inducing FAK activation. As shown in Fig. 4, ATF (50 nM) induced tyrosine phosphorylation of FAK, paxillin, and p130 cas in a pattern similar to that observed with uPA. These results strongly suggest that the action of uPA on the signal transduction to focal adhesion is mediated by the receptor-binding do-main of uPA, but independent of its catalytic domain.
We next examined the effects of uPA on the activation of MAPK in cultured endothelial cells. Cell lysates of BAE cells were analyzed by immunoblotting with phospho-specific MAPK antibody. As shown in Fig. 5A, both uPA and ATF (50 nM) induced tyrosine phosphorylation of p42 MAPK and p44 MAPK, with maximal induction present at 10 min. The phosphorylation of p42 MAPK and p44 MAPK by uPA and ATF was not due to the alteration in the amount of MAPK protein (Fig.  5A). Diisopropyl fluorophosphate-treated uPA, with no detectable amidolytic activity, also induced tyrosine phosphorylation of p42 MAPK and p44 MAPK (data not shown). The uPAinduced activation of MAPK was completely inhibited by the treatment of BAE cells with the tyrosine kinase inhibitor genistein (Fig. 5B), suggesting a tyrosine kinase-dependent pathway for the activation of MAPK by the uPAR. Furthermore, cytochalasin D, which abolished the activation of FAK in this study, also resulted in a decreased activation of MAPK (data not shown), indicating a role for cytoskeletal architecture in this response. To address the role played by members of the Src kinase family of tyrosine kinases in the activation of MAPK, the effects of the specific inhibitor PP 1 (27) were examined. PP 1 resulted in a decrease in MAPK activity at the base line (Fig.  6); however, despite the presence of PP 1 , uPA stimulation continued to result in activation of MAPK. Moreover, another Src family kinase inhibitor, herbimycin (28), did not affect the activation of MAPK by uPA treatment (data not shown). The results of these experiments indicate that uPA results in the activation of MAPK by a pathway that may involve the cytoskeleton, but not Src kinase activation.
The effects of uPA exposure on cytoskeletal arrangement were determined by staining with rhodamine-phalloidin. The cells were exposed to uPA (50 nM) for varying time intervals. Exposure to uPA resulted in a progressive alteration in the  distribution of actin (Fig. 7). Control cells typically contained a fine complex of microfilaments with the presence of a more dense peripheral band at the margins of the cells. Following exposure to uPA, peripheral bands became thinner, whereas the number of cytoplasmic stress fibers was increased (Fig. 7). DISCUSSION The uPAR has been clearly identified and characterized in human (29) and bovine (30) vascular endothelia. Urokinase has recently been shown to induce cell chemotaxis, cell adhesion/ migration, and cell growth. These effects appear to be mediated by the receptor-binding domain of uPA and can occur in the absence of catalytic activity (4 -8). The expression of uPA by migrating cells at the edge of wounded monolayers and at focal adhesion plaques has been described in keratinocytes (31), fibroblasts (32), and endothelial cells (33). In fibroblasts, uPAR occupancy by uPA restricts receptor distribution to focal contacts, whereas unoccupied uPARs remain mobile in the plane of the plasma membrane (34). In adherent monocytes exposed to a chemotactic gradient, the uPAR and uPA became highly concentrated at the leading edge of migration (5). In endothelial cells, uPA-induced cell migration is time-and concentrationdependent and does not involve extracellular proteolysis (6). Taken together, these findings suggest that uPA binding to the uPAR induces a signaling event that leads to an alteration in cell function.
Recent studies have demonstrated that a novel non-receptor tyrosine kinase, focal adhesion kinase, is involved in cell adhesion and migration (17). The mobility of cells from FAK-deficient mice has been reported to be impaired, whereas the number of focal adhesion contacts is increased, indicating that FAK may contribute to the turnover of focal adhesion contacts during cell migration (17). Using BAE cells, we have demonstrated that either uPA or ATF results in the activation of FAK, paxillin, and p130 cas . This activation pattern suggests that occupation of the uPAR by uPA facilitates adhesion signaling events and may promote cell migration by increasing the turnover of focal contacts. These data are consistent with previous observations that uPA-induced monocyte adhesion is dependent upon gene expression and protein synthesis (35).
Since the uPAR belongs to the family of GPI-anchored membrane proteins that lack transmembrane and cytoplasmic domains, it is unlikely that it can interact directly with FAK, paxillin, or p130 cas , which are cytosolic proteins localized in focal adhesions (18). Indeed, we did not observe the presence of FAK or p130 cas by immunoblotting uPAR immunoprecipitates from unstimulated or uPA-stimulated BAE cells (data not shown). Furthermore, several actin-binding proteins have been identified that co-localize with integrins in focal adhesions, including ␣-actinin, vinculin, and tensin. Talin and ␣-actinin bind to integrin cytoplasmic domains in vitro. Lewis and Schwartz (36) have recently described co-localization of talin, FAK, and actin in vivo using microbeads coated with antiintegrin antibodies. Chen et al. (37) reported that FAK associated with talin through a 48-amino acid sequence in its carboxyl-terminal domain. The uPAR is known to co-localize with the cytoskeletal components vinculin, ␣-actinin, and actin at sites of contact between cells and extracellular matrix (38). In this study, treatment of BAE cells with an inhibitor of cytoskeleton assembly, cytochalasin D, abolished the uPA-induced tyrosine phosphorylation of FAK (Fig. 2B). These results suggest that a functional linkage between the cytoskeleton and the uPAR is required for uPA-induced FAK activation. Since the uPAR has been localized to caveolae (39), which play an important role in the incorporation of proteins by the cell membrane and their intracellular uptake (40), and a functional unit of the uPAR, integrins, and caveolin regulates integrin function (16), this linkage is not surprising.
We speculate that the uPAR and cytoskeleton proteins, possibly vinculin or actinin, are bridged by at least one partner protein that contains a transmembrane segment, probably an integrin. Indeed, it has recently been shown that the uPAR is associated with the integrin Mac-1 within one receptor complex in human monocytes (15). In addition, uPA and the uPAR are co-localized with the vitronectin receptor (␣ v ␤ 3 -integrin) (3). A cooperative interaction between the uPAR and the vitronectin receptor has been reported, and transient expression of the uPAR in the human 293 cell, an embryonic kidney cell line, resulted in specific binding of vitronectin (41). Furthermore, this binding was enhanced in the presence of uPA (41), but was blocked by a chimeric protein composed of ␤ 1 -integrin transmembrane and cytoplasmic domains fused with the extracellular domain of CD4 (16). Thus, we hypothesize that the binding of uPA to the uPAR induces clustering of the uPAR to focal adhesions, where it transduces its signal by activating FAK, paxillin, and p130 cas through the cooperation of integrins and, in turn, promotes cell migration by increasing the turnover of focal adhesion contacts. There are other precedents for such a mechanism for signal transduction involving GPI-linked proteins, such as Ly-6, CD59, and CD55 (42,43). More recently, Laudanna et al. (44) reported that G protein-linked receptors of the chemoattractant subfamily can trigger adhesion through leukocyte integrins.
The pathway through which uPAR occupancy stimulates gene expression is not well understood. FAK has been reported to associate with the Grb2 adapter protein (19). Since Grb2 is linked to Ras via the Sos protein, FAK activation is linked to a well established mitogenesis pathway. Cytochalasin D treatment has been shown to prevent MAPK phosphorylation (45) and integrin-mediated activation of FAK (46,47). In this study, we showed that occupancy of the uPAR caused both activation of MAPK and tyrosine phosphorylation of FAK. The phosphorylation of FAK was apparent at an earlier time point than that of MAPK. The uPA-induced activation of MAPK was inhibited by genistein, but not by PP 1 or herbimycin, suggesting the involvement of a non-Src tyrosine kinase-dependent pathway. Furthermore, our preliminary data (not shown) indicated that cytochalasin D can decrease uPA-induced activation of MAPK in BAE cells. Although our current data do not directly address the potential relationship between FAK and MAPK in the uPAR system, these results suggest that they may be part of the same signaling pathway. MAPK is thought to play a key role in conveying signals from the cytoplasm to the nucleus, triggering gene expression. Thus, this study suggests that uPAinduced activation of MAPK may trigger the expression of cytoskeletal or cell-surface components that mediate the well established effects of uPA on cell migration.
In summary, these experiments demonstrate a role for the endothelial uPAR in the transduction of signals resulting from occupancy by the cognate ligand. They show that the catalytic domain of uPA is not involved in these responses. The uPAR appears to play a novel role in endothelial cell signal transduction through a mechanism that involves the activation of FAK and MAPK. These findings may have significant implications for the mechanism through which uPA influences cell migration and differentiation.