Role of Tyrosine Kinase Csk in G Protein-coupled Receptor- and Receptor Tyrosine Kinase-induced Fibroblast Cell Migration*

Tyrosine kinase Csk is essential for mouse embryonic development. Csk knock-out mice died at early stages of embryogenesis (around embryonic day 10). The molecular mechanism for this defect is not completely understood. Here we report that Csk deficiency in mouse embryonic fibroblast cells blocked cell migration induced by lysophosphatidic acid through G protein-coupled receptors, by platelet-derived growth factor and epidermal growth factor through receptor tyrosine kinases, and by serum. Re-expression of Csk in these Csk-deficient cells rescued the migratory phenotype. Furthermore, deletion of Csk did not interfere with Rac activation and lamellipodia formation, but impaired the focal adhesions. Our data demonstrate a critical role for Csk in cell migration.

tributes to pathologies such as atherosclerosis, tumor invasion, and metastasis. Cell migration is a dynamic, cyclical process in which a cell extends a protrusion at its front, which in turn attaches to the substratum on which the cell is migrating (9). This is followed by a contraction that moves the cell body forward toward the protrusion, and finally the attachments at the cell rear release as the cell continues to move forward. The cycle is initiated by external signals, which are sensed and communicated to the interior of the cell by specialized receptor proteins in the cell membrane. Focal adhesion turnover (the continuous formation and disassembly of adhesions) is critical for cell migration (10). Adhesion formation takes place at the leading edge of protrusions, whereas disassembly occurs both at the cell rear and at the base of protrusions. Despite the importance of these processes in cell migration, the mechanisms that regulate adhesion formation and disassembly remain largely unknown.
Previously, we and others have shown a critical role of Csk in regulating actin cytoskeletal reorganization. However, a potential role of Csk in controlling cell migration has not been investigated. Here we have shown that deficiency of Csk blocked the MEF cell migration induced by various stimuli including lysophosphatidic acid (LPA) through its G protein-coupled receptors, by platelet-derived growth factor (PDGF) and epidermal growth factor (EGF) through their receptor tyrosine kinases, and by serum. Hence, Csk plays a general regulatory role in cell migration. Although Csk is generally considered to play a negative role in cellular signaling, our data demonstrate a positive role for Csk in cell migration. Furthermore, we have shown that Csk deficiency led to impaired focal adhesions, thus providing a possible mechanism by which Csk controls cell migration.
Fluorescence Microscopy-Preparation of samples for fluorescence microscopy was preformed as described previously (11). Cells were plated onto gelatin-coated glass coverslips. Cells were then fixed with 3.7% formaldehyde, washed three times with PBS, permeabilized with 0.1% Triton X-100 for 5 min, and then washed with PBS three times. To block nonspecific binding, the cells were incubated with a solution of PBS containing 1% bovine serum albumin for 30 min and then incubated with primary antibody diluted in 1% bovine serum albumin in PBS for 1 h. Anti-vinculin antibody (Sigma) was used at a 1:1000 dilution. Alexa Fluor 488-conjugated phalloidin (Molecular Probes) was used to visualize actin. After incubation with primary antibody, cells were washed three times with PBS and incubated with rhodamine-conjugated anti-mouse antibody (Molecu-lar Probes). The coverslips were then fixed onto slides and imaged using a Zeiss fluorescence microscope.
In Vitro Wound-healing Cell Migration Assay-Cell migration assays were performed as described previously (12,13). Cells were allowed to form a confluent monolayer in a 24-well plate coated with gelatin before wounding. The wound was made by scraping a conventional pipette tip across the monolayer. The migration was induced by adding medium supplemented with or without 10% FBS. For MEF cells, it typically took 8 -10 h for the wound to close. When the wound for the positive control closed, cells were fixed with 3.7% formaldehyde and stained with crystal violet staining solution.
Boyden Chamber Cell Migration Assay-MEF cells (5 ϫ 10 4 ) suspended in DMEM (or in DMEM ϩ 0.5% FBS when PDGF, EGF, or LPA were present in the lower chamber) were added to the upper chamber of an insert (6.5 mm diameter, 8-m pore size; Becton Dickinson), and the insert was placed in a 24-well dish containing DMEM with or without 10% FBS, 20 ng/ml PDGF, 25 ng/ml EGF, or 10 M LPA. In the case of C3 toxin treatment, 0.5 M was added to both the upper and lower chamber. Migration assays were carried out for 4 h, and cells were fixed with 3.7% formaldehyde. Cells were stained with crystal violet staining solution, and cells on the upper side of the insert were removed with a cotton swab. Three randomly selected fields (10ϫ objective) were photographed, and the cells that had migrated were counted. The migration was expressed as a percentage of migrated cells in positive control MEF cells. Percentage was calculated with the following formula: Western Blots-Western blots were performed as previously described (14). Whole cell extracts were prepared as follows. Confluent cells were harvested from 10-cm plates, washed twice with cold phosphate-buffered saline, and pellets were resuspended in 0.8 ml of lysis buffer (150 mM NaCl, 20 mM Tris, pH 7.4, l mM EDTA, l mM EGTA, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 0.2 mM sodium orthovanadate, 0.03 mg/ml leupeptin). Resuspended pellets were sonicated, centrifuged at 5000 rpm for 5 min at 4°C to remove insoluble material, and the supernatant was saved as the whole cell extract. Protein concentrations were measured by Bradford assay, and equal amounts of protein were loaded onto a gel. After SDS-PAGE, protein samples were transferred to nitrocellulose filters. Membrane filters were incubated in 1ϫ Tris-buffered saline, 5% milk for 1 h, and then incubated in primary antibody for 2 h at room temperature. Blots were washed three times with Tris-buffered saline/Tween 20 and one time with Tris-buffered saline and then incubated with secondary antibody for 2 h at room temperature. Blots were washed again, and the signal was detected with ECL (Amersham Biosciences).
Cell Proliferation Assay-Csk Ϫ/Ϫ , Csk Ϫ/Ϫ /Csk, and MEF cells were trypsinized, resuspended in DMEM with 10% FBS, and counted. 5 ϫ 10 4 cells were plated per well in a 6-well dish and grown in DMEM with 10% FBS. Each day of the assay, one well of each type of cell was trypsinized, and the cells were counted.
Protein Purification-The recombinant fusion protein TAT-C3 was purified as previously described (11). The C3 exoenzyme was introduced into cells by making a fusion protein of the active domain of C3 and TAT protein of HIV-1 (15). The pGEX-KG TAT-C3 clone was a generous gift of Dr. Michael Olson (Institute of Cancer Research, London). The recombinant fusion protein was produced in BL21 Escherichia coli. A 1-liter culture was grown to an A 600 reading of 0.6 and then induced by addition of 0.3 mM isopropyl 1-thio-␤-D-galactopyranoside for 3 h. Cells were flash frozen and then lysed by sonication in Tris-buffered saline with 5 mM MgCl 2 , 1 mM dithiothreitol. The supernatant was then incubated with glutathione-Sepharose for 2 h at 4°C. After extensive washing, the beads were cleaved with thrombin overnight at 4°C. The supernatant was collected and treated with 1 mM phenylmethylsulfonyl fluoride. The purified protein was applied to cells at a final concentration of 0.5 M.
Rho and Rac Activity Assays-Cells were treated with 20 M LPA (for Rho assay) or 20 ng/ml PDGF (for Rac assay) for 10 min and then washed with PBS. Cells were then lysed in 20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). Thirty g of GST-PBD (for Rac assay) or GST-RBD (for Rho assay) attached to beads were added to the cell lysates. Samples were incubated at 4°C for 60 min, and then the beads were washed three times with lysis buffer. SDS sample buffer was added to the beads, and the samples were boiled at 90°C for 10 min and run on 12% SDS-polyacrylamide gels. Rac and Rho were visualized by Western blot using anti-Rac antibody (clone 23A8, Upstate) and anti-Rho antibody (clone 26C4, Santa Cruz).
Statistical Analysis-Data are expressed as mean Ϯ S.D. from three experiments and analyzed by one-way analysis of variance followed by Dunnett's multiple comparison test with significance defined as p Ͻ 0.05.

Deficiency of Csk Blocks MEF Cell
Migration-To investigate a possible role of Csk in cell migration, we have used two approaches to compare the migration of Csk-deficient cells and wild-type cells. One approach is the qualitative in vitro wound-healing assay, the other the quantitative Boyden chamber assay (12,13). For the wound-healing assay, wild-type and Csk Ϫ/Ϫ MEF cells were grown to confluence. A wound (small scratch) was made in the middle of the tissue culture plate with a pipette tip. After ϳ12 h in the presence of serum, wild-type MEF cells migrated and covered the wound, yet Csk Ϫ/Ϫ cells did not migrate (Fig. 1). Therefore, serum-induced migration of Csk Ϫ/Ϫ cells was markedly reduced compared with the migration of wild-type MEF cells. These results were confirmed with Boyden chamber assays ( Fig. 2A). To determine whether this failure of serum-induced migration of Csk Ϫ/Ϫ cells was the result of an absence of Csk, we re-expressed wild-type Csk in Csk Ϫ/Ϫ cells. As shown in Figs. 1 and 2A, Csk expression restored the migration of Csk Ϫ/Ϫ cells in response to serum. Furthermore, the Csk Ϫ/Ϫ /Csk rescue cell line expressed Csk at a lower level than the wild-type MEF cells as revealed by Western blot (Fig. 2B), indicating that the Csk protein level was not unphysiologically high in Csk Ϫ/Ϫ /Csk cells.
Different cell proliferation rates of wild-type MEF cells and Csk Ϫ/Ϫ cells might affect their migratory rates in wound-healing assays because these assays were performed over ϳ12 h (the chamber assay was performed over ϳ4 h). Thus, we examined the cell proliferation of wildtype MEF cells, Csk Ϫ/Ϫ cells, and Csk Ϫ/Ϫ /Csk rescue cells. As shown in Fig. 2C, proliferation of Csk Ϫ/Ϫ /Csk cells was slower than that of MEF or Csk Ϫ/Ϫ cells, thus, the faster rate of migration exhibited by Csk Ϫ/Ϫ / Csk cells is not because of a faster proliferation rate. Also, Csk Ϫ/Ϫ cells proliferate at a faster rate than MEF or Csk Ϫ/Ϫ /Csk cells, indicating that the lack of migration exhibited by these cells cannot be attributed to a slower rate of proliferation (Fig. 2C).
Because we used serum here as an inducer and serum contains various factors, it is likely that Csk plays a general role in controlling cell migration. Indeed, LPA, PDGF, and EGF all failed to induce Csk Ϫ/Ϫ cells to migrate, whereas they induced Csk Ϫ/Ϫ /Csk cell migration (Fig. 2D). LPA acts through G protein-coupled receptors, whereas PDGF and EGF work on receptor tyrosine kinases. These results demonstrate that a deficiency of Csk blocks cell migration induced by various factors and that Csk plays a general role in controlling cell migration.
The Kinase Activity of Csk Is Required for Cell Migration-Tyrosine kinases have other structural domains in addition to their catalytic kinase domains (16). Some of their physiological functions are mediated by other structural domains, independent of their kinase activity. To investigate whether the kinase activity of Csk is required for cell migration, we made use of two kinase-dead mutations of Csk (CskR318A and CskD314N).  These mutations at the catalytic loop of Csk reduce the catalytic activity of the protein (11,17,18). As assayed with both wound-healing and chamber assays, stable lines of Csk Ϫ/Ϫ cells expressing CskR318A or CskD314N displayed impaired migration compared with Csk Ϫ/Ϫ /Csk and MEF cells (Fig. 3, A and B). The level of Csk mutant proteins in each of these cell lines is shown in Fig. 2B. Hence, these data demonstrate that the kinase activity of Csk is required for its role in cell migration.
The best studied physiological substrates for Csk are Src family tyrosine kinases. Regulation of actin cytoskeletal organization by Csk has been genetically demonstrated through Src family kinase-dependent as well as -independent pathways (6,11). Therefore, we examined whether Src family tyrosine kinases are involved in Csk-regulated cell migration. In Csk Ϫ/Ϫ cells, the specific activity of Src is high. To study whether this higher Src activity contributed to the cell migration defect in Csk Ϫ/Ϫ cells, we treated Csk Ϫ/Ϫ cells with Src family tyrosine kinase inhibitor PP2 (Fig. 3C). The underlying assumption is that this higher Src activity might be detrimental to cell migration even though normal levels of Src activity or regulation might be required for cell migration. As shown in Fig. 3C, low concentrations of PP2 increased Csk Ϫ/Ϫ cell migration, suggesting that higher activity of Src family kinases is at least partly responsible for the Csk Ϫ/Ϫ phenotype of defective cell migration. On the other hand, higher concentrations of PP2 inhibited cell migration. Hence, these data demonstrate that Src family tyrosine kinases are involved in Csk regulation of fibroblast cell migration.
Deficiency of Csk Does Not Affect Rac Activation and Lamellipodia Formation-We have previously shown that Csk acts upstream of Rho in actin stress fiber formation induced by serum, LPA, and various G proteins (11). Here we have confirmed that Csk is indeed required for LPA-induced activation of Rho (Fig. 4A). To measure Rho activity in Csk Ϫ/Ϫ cells and Csk Ϫ/Ϫ /Csk cells, a Rho activation assay (GST-RBD pull-down) was performed. Rho activity was increased in Csk Ϫ/Ϫ /Csk cells in response to LPA. However, in Csk Ϫ/Ϫ cells, there was no increase (Fig. 4A). This Csk signaling through Rho to control actin cytoskeletal reorganization could provide a mechanism by which Csk regulates cell migration. Therefore, we examined whether Rho is  involved in MEF cell migration. Clostridium botulinum C3 exoenzyme is a specific inhibitor of Rho that works by irreversible ADP-ribosylation of Asn-41 in its effector region (19). We previously used this C3 toxin to block serum-induced Rho activation and actin stress fiber formation in MEF cells (11). Treatment with C3 toxin had no effect on the migration of Csk Ϫ/Ϫ /Csk and MEF cells ( Fig. 2A), indicating that Rho activity is not required for fibroblast cell migration. This was consistent with our finding that inhibition of Rho had no effect on fibroblast cell migration and inhibition of Rho kinase (ROCK) actually increased the migration of fibroblast cells (13). In fast migrating cells such as macrophages and neutrophils, Rho and Rho kinase appear to be required for cell polarization and migration. In slow migrating cells such as fibroblasts, Rho kinase appears to inhibit migration (20).
Next, we investigated the role of Csk in Rac activation in fibroblast cells. An important step in cell migration is the activation of Rac and subsequent formation of lamellipodia at the leading edge of the cell (21). First, we tested whether Csk was essential for PDGF-induced Rac activation. The reason that we used PDGF here is that PDGF is a better stimulator of Rac activation than serum and LPA in in vitro assays. As shown in Fig. 4B, Rac activity was stimulated by PDGF in Csk Ϫ/Ϫ cells as well as in Csk Ϫ/Ϫ /Csk cells (Fig. 4B). Thus, in contrast to Rho activation, Csk is not required for Rac activation. Next, we studied whether Csk was involved in lamellipodia formation. Cells at the leading edge of a wound were fixed 3 h after wounding and stained with phalloidin to visualize actin polymers. As shown in Fig. 4C, Csk Ϫ/Ϫ cells displayed no defect in lamellipodia formation. Lamellipodia in these cells were similar to those in Csk Ϫ/Ϫ /Csk cells (Fig. 4C). After examining 300 -400 cells from three separate experiments, 80% Ϯ 1.5% Csk Ϫ/Ϫ cells, 89% Ϯ 1.9% Csk Ϫ/Ϫ / Csk cells, and 85% Ϯ 3% of MEF cells displayed membrane ruffles/ lamellipodia. These data demonstrate that Rac signaling is intact in Csk Ϫ/Ϫ cells and that Csk is not involved in lamellipodia formation.
Csk Is Involved in the Regulation of Focal Adhesions-To further investigate the mechanism by which Csk controls cell migration, we next turned our attention to focal adhesions. Cell migration requires turnover of cell adhesions to the substratum (9). Several proteins that are known to associate with Csk are involved in focal adhesion turnover such as FAK, Src, and paxillin (22). In HeLa cells, overexpressed Csk was found to localize to focal adhesions, colocalizing with FAK and talin (23). Thus, we compared the focal adhesions in Csk Ϫ/Ϫ cells and Csk Ϫ/Ϫ /Csk cells. Immunostaining of vinculin allowed the visualization of focal adhesions (Fig. 5). There were marked differences in focal adhesions in Csk Ϫ/Ϫ cells and Csk Ϫ/Ϫ /Csk cells. The focal adhesions of Csk Ϫ/Ϫ cells appeared abnormally big and mainly peripherally located, implying a defect in focal adhesion turnover. Csk Ϫ/Ϫ /Csk cells, however, displayed focal adhesions that were smaller and located throughout the cell at the ends of stress fibers (Fig. 5). Therefore, Csk deficiency impairs focal adhesion turnover. This provides a potential mechanism responsible for the defective migration of Csk Ϫ/Ϫ cells. Further investigation is needed to learn how Csk regulates focal adhesion turnover.

DISCUSSION
We have shown that a deficiency of Csk blocked MEF cell migration. Toward understanding the underlying mechanism, we found that Csk has no effect on PDGF-induced Rac activation and lamellipodia formation. However, Csk Ϫ/Ϫ cells displayed larger focal adhesions, consistent with impairment in focal adhesion turnover. Thus, Csk might critically control the cycling steps of cell migration, namely the detachment of adhesions and cell body contraction.
How does Csk control the repetitive turnover of focal adhesions during cell migration? We propose that Csk functions as a thermometer switch (Fig. 6). In our proposed model, extracellular signals such as integrin activate Src (probably through FAK). Src then phosphorylates p130 Cas . Phosphorylated p130 Cas forms a complex with Crk, in turn leading to p130 Cas -Crk-DOCK180 complex formation. In this complex, DOCK180 catalyzes the activation of Rac, leading to the formation of focal complexes. As the number of focal adhesions increases, more paxillin is accumulated at the focal complexes. Therefore, more Csk will be recruited to focal adhesions by paxillin. Csk will in turn inhibit Src activity and at the same time activate the tyrosine phosphatase PTP-PEST, starting the focal adhesion disassembly process. When the numbers of focal adhesions decrease, less paxillin and, hence less Csk, will be present at the focal complex. This will allow new focal complexes to form. This cycling of Csk association with and dissociation from the focal complex regulates the repetitive turnover of focal adhesions during cell migration.
In essence, this model is based on extensive published data. We have incorporated a major regulatory role for Csk in this focal adhesion turnover. A role for the Src family tyrosine kinases in focal adhesion turnover is supported by published data (24,25). In Csk Ϫ/Ϫ MEF cells, tyrosine phosphorylation of several focal adhesion proteins such as cortactin, tensin, FAK, and paxillin was increased (6,26). Although cortactin and tensin hyperphosphorylation is Src-dependent, FAK and paxillin hyperphosphorylation is dependent on both Src and Fyn. In addition, Src tyrosine kinase has been shown to phosphorylate p130 Cas to regulate p130 Cas interaction with Crk (27). Furthermore, Src Ϫ/Ϫ cells had reduced adhesion-induced tyrosine phosphorylation of p130 Cas and Csk Ϫ/Ϫ cells showed enhanced phosphorylation of p130 Cas (28). Src triggers translocation of Crk to focal adhesions via p130 Cas . In wild-type MEF cells overexpressing active SrcY529F, Crk localized to podosome focal adhesions. However, in p130 CasϪ/Ϫ cells expressing SrcY529F, the amount of Crk at focal adhesions was reduced (29). Moreover, expression of Csk targeted to focal adhesions by fusing Csk-GFP to the FAT domain (FA-Csk) (focal adhesion targeting domain of FAK) or a LIM domain (from paxillin) caused a loss of integrin-stimulated cell-matrix adhesion in fibroblasts. However, this effect was blocked by expressing activated SrcY529F in Src Ϫ/Ϫ cells expressing FA-Csk. In addition, loss of focal adhesions induced by FA-targeted Csk expression could be restored by expression of p130 Cas or Crk (30). Therefore, Src family tyrosine kinases are necessary for focal adhesion turnover.
PTP-PEST tyrosine phosphatase has also been implicated in focal adhesion turnover. PTP-PEST directly binds to the SH3 domain of Csk, although it is not clear what the effect of Csk on the activity of PTP-PEST is (7,31). Downstream of PTP-PEST, p130 Cas is a specific substrate for PTP-PEST in vitro and in intact cells (32). In PTP-PEST Ϫ/Ϫ cells, p130 Cas , paxillin, and FAK are hyperphosphorylated. Src and Crk SH2 domains showed increased affinity for p130 Cas (32,33). Furthermore, PTP-PEST Ϫ/Ϫ fibroblasts displayed larger and more numerous focal adhesions in comparison to MEF cells when plated on fibronectin. PTP-PEST Ϫ/Ϫ fibroblasts showed a migration defect (34). Rat1 fibroblast cells stably overexpressing PTP-PEST have lower levels of p130 Cas tyrosine phosphorylation but normal levels of tyrosine phosphorylation of most proteins. The rate of cell migration of these cells is significantly reduced and p130 Cas is not redistributed to the leading edge of the cell. A reduced association of p130 Cas with Crk was also found in these cells (35). Moreover, p130 Cas and Crk complex with DOCK180 to activate Rac, leading to the focal complex formation. It is not clear at this point whether another protein, ELMO, is also in this p130 Cas -Crk-DOCK180 complex (36,37). Also, paxillin is enriched in focal adhe-sions (38). Paxillin directly binds Csk and PTP-PEST (39 -41). We are performing experiments to test this model. Hopefully, our study on the mechanism by which Csk regulates cell migration will shed light on the mechanism of cell migration and cancer metastasis and provide possible targets for chemotherapeutic drugs and markers to help clinicians assess tumor aggressiveness. FIGURE 6. A model for Csk regulation of focal adhesion turnover. In this proposed model, extracellular signals such as integrin activate Src (probably through FAK). Src then phosphorylates p130 Cas . Phosphorylated p130 Cas forms a complex with Crk, in turn leading to the formation of the p130 Cas -Crk-DOCK180 complex. In this complex, DOCK180 catalyzes the activation of Rac, leading to the formation of focal complexes. As the number of focal adhesion increases, more paxillin is accumulated at the focal complexes. Therefore, more Csk will be recruited to the focal adhesion by paxillin. Csk will in turn inhibit Src activity and at the same time activate the tyrosine phosphatase PTP-PEST, starting the focal adhesion disassembly process. When the numbers of focal adhesions decrease, less paxillin and, hence less Csk, will be present at the focal complex. This will allow new focal complexes to form. This cycling of Csk association with and dissociation from the focal complex regulates the repetitive turnover of focal adhesions during cell migration.