STAT3 Deficiency in Keratinocytes Leads to Compromised Cell Migration through Hyperphosphorylation of p130 cas *

We previously reported that STAT3 plays a crucial role in transducing a signal for migration of keratinocytes (Sano, S., Itami, S., Takeda, K., Tarutani, M., Yamaguchi, Y., Miura, H., Yoshikawa, K., Akira, S., and Takeda, J. (1999) EMBO J. 18, 4657–4668). To clarify the role of STAT3 in signaling the migration, we studied the intracellular signaling pathway through an integrin receptor in STAT3-deficient keratinocytes. STAT3-deficient keratinocytes demonstrated increased adhesiveness and fast spreading on a collagen matrix. Staining with anti-phosphotyrosine antibody revealed that STAT3-deficient keratinocytes had an increased number of tyrosyl-hyperphosphorylated focal adhesions. Analyses with immunoprecipitation revealed that p130 cas was constitutively hyperphosphorylated on tyrosine residues, while other focal adhesion molecules such as focal adhesion kinase and paxillin were not. Transfection of STAT3-deficient keratinocytes with an adenoviral vector encoding the wild-type Stat3 gene reversed not only impaired migration but also the increased tyrosine phosphorylation of p130 cas . These results strongly suggest that STAT3 in keratinocytes plays a critical role in turnover of tyrosine phosphorylation of p130 cas , modulating cell adhesiveness to the substratum leading to growth factor-dependent cell migration.

We previously reported that STAT3 plays a crucial role in transducing a signal for migration of keratinocytes (Sano, S., Itami, S., Takeda, K., Tarutani, M., Yamaguchi, Y., Miura, H., Yoshikawa, K., Akira, S., and Takeda, J. (1999) EMBO J. 18, 4657-4668). To clarify the role of STAT3 in signaling the migration, we studied the intracellular signaling pathway through an integrin receptor in STAT3-deficient keratinocytes. STAT3-deficient keratinocytes demonstrated increased adhesiveness and fast spreading on a collagen matrix. Staining with anti-phosphotyrosine antibody revealed that STAT3-deficient keratinocytes had an increased number of tyrosyl-hyperphosphorylated focal adhesions. Analyses with immunoprecipitation revealed that p130 cas was constitutively hyperphosphorylated on tyrosine residues, while other focal adhesion molecules such as focal adhesion kinase and paxillin were not. Transfection of STAT3-deficient keratinocytes with an adenoviral vector encoding the wild-type Stat3 gene reversed not only impaired migration but also the increased tyrosine phosphorylation of p130 cas . These results strongly suggest that STAT3 in keratinocytes plays a critical role in turnover of tyrosine phosphorylation of p130 cas , modulating cell adhesiveness to the substratum leading to growth factor-dependent cell migration.
Signal transducers and activators of transcription (STATs) 1 are a family of latent cytoplasmic transcription factors that are activated by many cytokines and growth factors (1)(2)(3). STATs are phosphorylated on tyrosine residues by activated kinases in receptor complexes, leading to formation of homo-or heterodimers and translocation to the nucleus in which they regulate transcription. STAT3 is activated by a variety of cytokines and growth factors such as interleukin-6, epidermal growth factor (EGF), hepatocyte growth factor (HGF), plateletderived growth factor, and granulocyte colony-stimulating factor. These cytokines and growth factors regulate the biological activities of keratinocytes (4,5), suggesting that STAT3 plays a crucial role in keratinocytes. Because germ line STAT3 deletion leads to embryonic lethality (6), to elucidate the biological roles of STAT3 in the skin, we previously generated keratinocytespecific STAT3-deficient mice by conditional gene targeting using the Cre-loxP strategy (7). The Stat3 gene was disrupted under the control of a keratin 5 promoter. The mutant mice were born with no apparent abnormalities, and their epidermis and hair follicle development was normal at birth. However, wound healing was markedly retarded, and the second hair cycle was impaired in keratinocyte-specific Stat3 gene knockout mice. An in vitro study with cultured keratinocytes revealed that this phenotype was attributed to impaired migration because of the failure of STAT3 activation.
Cell migration is composed of several concerted steps (8,9). Migration is initiated with membrane protrusion (filopodia and leading edge) and adhesion to the extracellular matrix, followed by cell traction and the release of adhesions at the rear portion of the cell. These events are regulated by multiple signaling mechanisms, such as tyrosine kinase and/or phosphatase signaling (10 -14), mitogen-activated protein kinase signaling (15,16), small GTPase signaling (i.e. Rho and Rac) (17,18), and cytoskeletal reorganization (i.e. actin polymerization/ depolymerizaton, actin/myosin motor) (19,20). However, it is still undetermined whether STAT3 is involved in these signaling events.
In this report, we found that STAT3-deficient keratinocytes showed increased adhesiveness and forced spreading on a collagen matrix and that an increased number of hyperphosphorylated focal adhesions, in particular, an adaptor protein, p130 cas , was constitutively hyperphosphorylated on tyrosine residues. These results strongly suggest that intracellular signaling of STAT3 in keratinocytes modulates tyrosine phosphorylation of p130 cas and cell adhesiveness to the substratum, leading to cell migration in response to growth factors.
Cell Culture-Primary cultures of mouse keratinocytes were established from newborn to 5-day-old mice. The total skin taken from mice was treated with 250 units/ml dispase (Godo Shusei, Tokyo, Japan) overnight at 4°C, and the epidermis was peeled off from the dermis. Keratinocytes were collected upon trypsinization and washed in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. Cells were resuspended in MCDB153 medium containing 3% fetal calf serum and seeded onto type I collagen-coated dishes.
In Vitro Migration Assay-Keratinocytes were cultured in type I collagen-coated dishes until they reached confluence. After starvation for 24 h, they were treated with 10 g/ml mitomycin C for 2 h to avoid a proliferative effect on the cells. A cell-free area was introduced by scraping the monolayer with a yellow pipette tip. Cell migration to the cell-free area was evaluated in the absence or presence of 10 ng/ml EGF (Upstate Biotechnology, Inc., Lake Placid, NY) or HGF (Collaborative Biomedical Products, Bedford, MA). After 48 h, photographs were taken using a phase-contrast microscope (DIA-PHOT 300; Nikon). The number of migrating keratinocytes was counted after taking photographs of four nonoverlapping fields. Values represent the mean Ϯ S.D. of migrating cells per square millimeter beyond the front lines of the introduced wound edge. Student's t test was used for statistical analysis.
Cell Adhesion Assay-Keratinocytes collected upon trypsinization of epidermal sheets were washed in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. Cells were resuspended in serum-free MCDB153 medium at 10 6 cells/ml and added to type I collagen-coated dishes. These dishes were incubated at 37°C for the indicated times. Nonadherent cells were removed by washing with phosphate-buffered saline (PBS), and attached cells were counted. Values represent the mean Ϯ S.D. and were analyzed statistically using Student's t test.
Cell Spreading Assay-Keratinocytes were seeded onto type I collagen-coated dishes at 5 ϫ 10 5 cells/ml, and allowed to spread for the indicated times. After washing out the nonadherent cells, photographs were taken in the nonoverlapping fields under a phase-contrast microscope, and spread cells were counted. Spread cells were identified as opaque cells, exhibiting membranous protrusions. However, nonspread cells were rounded and phase-bright. Values represent the mean Ϯ S.D. and were analyzed statistically using Student's t test.
Immunofluorescence Staining-Keratinocytes (5 ϫ 10 5 cells/ml) were plated onto type I collagen-coated coverslips for 24 h in MCDB153 medium containing 3% fetal calf serum. After serum starvation for 24 h, coverslips were washed twice with PBS and fixed for 5 min with 4% paraformaldehyde containing 0.5% Triton X-100 for permeabilization, followed by fixation with 4% paraformaldehyde for 15 min. After washing with PBS, fixed cells were blocked in PBS containing 1% bovine serum albumin. Cells were stained with a 1:100 diluted anti-phosphotyrosine antibody (PY20). Cells were further incubated with a fluorescein isothiocyanate-conjugated anti-mouse secondary antibody and washed with PBS. Images were analyzed using a confocal microscope (model LSM 410; Carl Zeiss).
Immunoprecipitations and Western Blot Analysis-Keratinocytes were washed twice with PBS and lysed in a radioimmune precipitation buffer (10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.1% sodium deoxycholate, 1 mM EDTA, 1 mM sodium orthovanadate, 0.2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin). The protein concentration of the lysates was normalized before Western blotting or immunoprecipitation. For immunoprecipitation, equivalent lysates were incubated with precipitating antibodies bound to protein G-Sepharose beads (Amersham Biosciences) for 90 min at 4°C. Immunoprecipitates were washed five times with the radioimmune precipitation buffer. The immunoprecipitates or whole lysates were separated on 10% SDS-polyacrylamide gels, transferred to nitrocellulose membranes, and reacted with primary antibodies. Primary antibodies were detected using horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (ECL; Amersham Biosciences).
Construction of a Recombinant Adenoviral Vector-A recombinant adenoviral vector containing the wild-type STAT3 expression cassette (Adeno-stat3) was constructed using the circular form of the adenoviral genome cloned in a cosmid and the Cre-loxP recombination system according to Tashiro (21). The pALC cosmid and wild-type STAT3 expression cassette were ligated. The resulting pALC-STAT3 cosmid vector was purified from E. coli DH5␣ and transfected into 293 cells with the pMC1-cre plasmid. The cosmid backbone was excised by transiently expressed Cre recombinase. Recombinant adenoviruses were purified by polyethylene glycol precipitation and cesium chloride ultracentrifugation. For reintroduction of wild-type STAT3, keratinocytes were infected with Adeno-stat3 during a serum starvation period for 24 h at a multiplicity of infection of 5.

STAT3-disrupted Keratinocytes Exhibited an Increased Adhesiveness to the Collagen Matrix-
We previously demonstrated that keratinocyte-specific STAT3-disrupted mice showed retardation of skin wound healing and that migration of STAT3-deficient keratinocytes was impaired (7). Cell migration is a coordinated and complex process, including cell adhesion to the extracellular matrix and organization of the actin cytoskeleton. Therefore, we hypothesized that STAT3 is involved in a specific point in these processes. First we compared cell attachment to the extracellular matrix. Freshly isolated keratinocytes were seeded onto type I collagen-coated dishes. Then nonadherent cells were removed by washing at the indicated time points, and attached cells were counted. As shown in Fig. 1a, STAT3-disrupted keratinocytes (white bars) showed significantly increased adhesiveness to the collagen matrix as compared with control cells (black bars) at 20 min, although no difference was observed after 1 or 3 h. The adhesion of keratinocytes to type I collagen is mediated through ␣ 2 ␤ 1 and/or ␣ 3 ␤ 1 integrins, and the STAT3 signaling pathway may modulate integrin expression (22) in some cells. However, no difference in integrin expression was found between control and STAT3-disrupted keratinocytes (Ref. 7 and data not shown). Therefore, increased adhesion of STAT3-disrupted keratinocytes was not attributed to increased integrin expression.
STAT3-disrupted Keratinocytes Spread Faster on the Collagen Matrix-Next, we compared cell spreading on the substratum of STAT3(Ϫ/Ϫ) keratinocytes with controls. Twenty min after seeding, an increased number of STAT3(Ϫ/Ϫ) keratinocytes were opaque with membranous protrusions (Fig. 1b,  black arrows), which is a characteristic feature of spread cells, whereas control keratinocytes (STAT3(ϩ/Ϫ)) were mostly round and phase-bright (nonspread) (Fig. 1b, upper panel). Quantitative evaluation revealed that the number of spread cells in STAT3-disrupted keratinocytes was 5-fold that at 20 min compared with controls, although no difference in cell spreading was observed between the two kinds of keratinocytes later than 1 h (Fig. 1c). Taken collectively, these results indicate that STAT3-disrupted cells were primed for attachment and spreading on the substratum, implying that STAT3 plays a critical role in regulating subsequent cell motility through integrin/focal adhesion signaling.
An Increased Number of Tyrosine-phosphorylated Focal Adhesions in STAT3-disrupted Keratinocytes-In cultured cells, integrins and tyrosine-phosphorylated proteins are concentrated at focal adhesions, where actin cytoskeletons are connected with the extracellular matrix. Many lines of evidence have revealed that the number and/or tyrosine phosphorylation status of focal adhesions influenced adhesiveness, spreading, and cell migration (10,11,14,23). Therefore, we next examined the formation and tyrosine phosphorylation status of focal adhesions, by immunostaining with an anti-phosphotyrosine monoclonal antibody. Strikingly, individual STAT3-disrupted keratinocytes exhibited coarse and strong signals on the surface (Fig. 2, right panel, red arrowheads) compared with control cells (Fig. 2, left panel), indicating that focal adhesions of STAT3-disrupted keratinocytes were hyperphosphorylated.
p130 cas Was Constitutively Hyperphosphorylated in STAT3disrupted Keratinocytes-Previous studies showed that many tyrosine-phosphorylated proteins are associated with focal adhesions, including nonreceptor kinases, adaptor proteins, and cytoskeletal proteins (13,24,25). Among them, focal adhesion kinase (FAK; M r ϳ125,000), p130 cas (M r ϳ130,000), paxillin (M r ϳ68,000), and Src (M r ϳ60,000) were candidates for the hyperphosphorylated proteins observed in Fig. 3a. Therefore, we analyzed the tyrosine phosphorylation state of these proteins. As shown in Fig. 3b, no differences were demonstrated in protein expression or tyrosine phosphorylation levels of FAK, Src, and paxillin between control and STAT3-disrupted keratinocytes. However, p130 cas was constitutively hyperphosphorylated on tyrosine residues in STAT3(Ϫ/Ϫ) keratinocytes (Fig.  3b, black arrow). p130 cas is thought to be an adaptor protein that contains an Src homology 3 (SH3) domain, a central substrate domain, and proline-rich motifs (26). The central substrate domain has multiple phosphotyrosine motifs where Src homology 2 (SH2) domain-containing molecules recruit. The hyperphosphorylation of p130 cas in STAT3-disrupted keratinocytes may result in abnormal assembly of interacting proteins, aberrant turnover of focal adhesions, increased cell adhesiveness, and faster cell spreading, leading to impaired migration. It is possible that the hyperphosphorylation of p130 cas might be the result of down-regulation of protein-tyrosine phosphatases. Protein-tyrosine phosphatase, PTP-PEST, has been shown to interact with p130 cas , and it has been demonstrated that PTP-PEST(Ϫ/Ϫ) fibroblasts have a phenotype similar to STAT3(Ϫ/Ϫ) keratinocytes (10). Therefore, we examined expression levels of PTP-PEST. However, as shown in Fig. 3c, no difference was demonstrated in protein expression levels of PTP-PEST between control and STAT3(Ϫ/Ϫ) keratinocytes.

Rescue of the Phenotype of STAT3-disrupted Keratinocytes by Reintroduction of a Wild-type STAT3
Gene-We previously demonstrated that the migration of STAT3-disrupted keratinocytes was impaired (7). Control keratinocytes (STAT3(ϩ/Ϫ)) migrated in response to EGF and HGF (Fig. 4, a (top panels) and b (black bars)). In contrast, the migration of STAT3-disrupted keratinocytes (STAT3(Ϫ/Ϫ)) was severely compromised (Fig. 4, a (middle panels) and b (white bars)). These findings suggest that STAT3 plays a critical role in keratinocyte migration in response to growth factors that activate STAT3, including EGF and HGF. To clarify whether STAT3 deficiency was primarily responsible for the phenotype, we introduced the wild-type Stat3 gene into STAT3(Ϫ/Ϫ) keratinocytes using an FIG. 1. STAT3-disrupted keratinocytes exhibit increased adhesiveness and spreading onto the collagen matrix. a, analysis of cell attachment to the collagen matrix. Freshly isolated keratinocytes were seeded onto type I collagen-coated dishes. After incubation for 20 min, 1 h, and 3 h, nonadherent cells were removed by washing, and attached cells were counted. STAT3-disrupted keratinocytes (white bars) showed significantly stronger cell adhesiveness to the collagen matrix than control cells (black bars) at 20 min (p Ͻ 0.05, Student's t test), although no significant differences were observed at 1 and 3 h. Each bar represents the mean Ϯ S.D. of three independent experiments. N.S., not significant. b and c, analysis of cell spreading on the collagen matrix. Freshly isolated keratinocytes were seeded onto type I collagen-coated dishes, and nonadherent cells were washed out. Photographs were taken using a phase-contrast microscope after incubation for the indicated times. b, representative photographs at 20 min. Spread cells were identified as opaque cells, exhibiting membranous protrusions (black arrows). However, nonspread cells were rounded and phase-bright under a phase-contrast microscope. Bar, 20 m. c, quantitative evaluation of keratinocyte spreading on the collagen matrix. adenoviral vector. Upon reintroduction of wild-type STAT3, impaired migration of the STAT3-deficient keratinocytes was completely reversed (Fig. 4, a (bottom panels) and b (gray bars)), while an empty vector alone could not rescue the phenotype (data not shown). Furthermore, reexpression of Stat3 using an adenoviral vector normalized the tyrosine phosphorylation level in focal adhesions and p130 cas (Fig. 5, a (right  panel) and b (top panel, black arrow), respectively). These findings indicate that STAT3 abrogation primarily leads to compromised cell migration through hyperphosphorylation of p130 cas . DISCUSSION We previously reported that cell migration was severely compromised in STAT3(Ϫ/Ϫ) keratinocytes in vivo and in vitro (7). In the present study, we showed that STAT3(Ϫ/Ϫ) keratinocytes display enhanced cell adhesion and spreading. Cell migration is a coordinated, multistep processes involving 1) extension of membrane protrusions (filopodia and lamellipodia) at the leading edge, 2) formation of attachment sites to the were subjected to Western blot analysis with PY20. Approximately 60 -90-and 130-kDa proteins are hyperphosphorylated on tyrosine residues in STAT3-disrupted keratinocytes (black arrows). b, phosphorylation profiles of focal adhesion-associated proteins. Whole lysates or PY20 immunoprecipitates were subjected to Western blot analysis with anti-focal adhesion kinase (FAK), Src, paxillin, and p130 cas (Cas) antibodies. No differences were demonstrated in protein expression and tyrosine phosphorylation levels of FAK, Src, and paxillin between control and STAT3(Ϫ/Ϫ) keratinocytes. However, p130 cas was constitutively hyperphosphorylated on tyrosine residues in STAT3-disrupted keratinocytes (black arrow). c, whole lysates were subjected to Western blot analysis with an anti-PTP-PEST monoclonal antibody. No difference was demonstrated in protein expression between control and STAT3(Ϫ/Ϫ) keratinocytes.

FIG. 4.
Rescue of the impaired migration of Stat3-disrupted keratinocytes by re-introduction of wild-type STAT3. a, primary cultured keratinocytes were subjected to in vitro migration assays. Migration to the cell-free area was evaluated in the absence (nontreated; NT) or the presence of 10 ng/ml EGF or HGF. Control keratinocytes (STAT3(ϩ/Ϫ)) migrated in response to EGF or HGF (top panels). In contrast, the migration of STAT3-disrupted keratinocytes (STAT3(Ϫ/Ϫ)) was severely compromised (middle panels). Impaired migration in STAT3-deficient keratinocytes was rescued by wild type Stat3 gene transfer using an adenoviral vector (bottom panels). Bar, 200 m. b, quantitative evaluation of keratinocyte migration in response to ligands. Cell migration to the cell-free area for 48 h was assessed as described under "Experimental Procedures." Each bar represents the mean Ϯ S.D. of three independent experiments. *, p Ͻ 0.05, significantly different from the control as determined by Student's t test. extracellular matrix (focal adhesions) at the newly formed cell periphery, 3) cell traction, which might be driven by contraction of the actin cytoskeleton, and 4) the release of adhesions at the rear portion of the cell (detachment). It has been reported that cell migration depends on adhesiveness to the substratum (27). The authors have examined the relationship between maximum migration speed and cell adhesiveness and demonstrated that maximum migration occurs at an intermediate level of cell adhesiveness. The strong adhesiveness might disturb the detachment step and lead to impaired cell migration. Conversely, abnormality in the breakdown rate of focal adhesions might interfere with the detachment step and lead to an increased number of focal adhesions and strong adhesiveness to the extracellular matrix. This hypothesis is supported by the observation that FAK(Ϫ/Ϫ) fibroblasts showed an increased number of focal adhesions, strong adhesiveness, and impaired cell migration (23). In FAK(Ϫ/Ϫ) fibroblasts, no changes in the expression of ␤ 1 integrins were found, which is similar to STAT3(Ϫ/Ϫ) keratinocytes. The authors concluded that FAK might be involved in the turnover of focal adhesion during cell migration but not in the formation of focal adhesions. Considering this, it is possible that the phenotype of STAT3(Ϫ/Ϫ) keratinocytes might be the result of focal adhesion turnover failure.
The present study also demonstrated that STAT3-disrupted keratinocytes exhibited a constitutive hyperphosphorylation status in the focal adhesion molecule, particularly p130 cas . p130 cas was originally identified as a protein that was highly tyrosine-phosphorylated in v-Src-and v-Crk-transformed cells (28). The role of p130 cas signaling in cell spreading by anchoring the actin cytoskeleton to focal adhesion has been suggested by the fact that myofibrils and Z discs in cardiocytes were disorganized in p130 cas (Ϫ/Ϫ) mice and that actin stress fiber formation was severely impaired in p130 cas (Ϫ/Ϫ) primary fibroblasts (29). p130 cas has several protein interaction motifs, including an SH3 domain, a central substrate domain (tyrosine phosphorylation sites), and proline-rich motifs (30). Upon integrin stimulation by means of binding to the extracellular matrix, FAK undergoes autophosphorylation, providing a binding site for nonreceptor tyrosine kinases, such as Src (31,32). p130 cas binds to the proline-rich sequence of FAK via its SH3 domains (33) and is tyrosine-phosphorylated by FAK and/or Src (25). Tyrosine-phosphorylated p130 cas plays a role as an adaptor molecule, where SH2 domain-containing proteins such as Crk are recruited (25,34). Recent evidence indicates that this p130 cas /Crk coupling serves as a molecular switch for activating small GTPase Rac through the small GTPase-activating protein DOCK180 and leads to the rearrangement of the actin cytoskeleton (15,34). On the other hand, upon certain extracellular stimuli, protein-tyrosine phosphatases (PTP-PEST, SHP2) might also be recruited (36,37), leading to the dephosphorylation of p130 cas and other focal adhesion-associated proteins. These tyrosine phosphorylation-dephosphorylation cycles might contribute to the formation and stability of focal adhesions and reorganization of cytoskeletal proteins. In STAT3-disrupted keratinocytes, p130 cas was hyperphosphorylated on tyrosine residues (Fig. 3b), although the tyrosine phosphorylation levels of the interacting protein kinases, FAK and Src, were normal (Fig. 3b). Although the exact sites of phosphorylation in FAK and Src were not investigated, the normal phosphorylation status found in STAT3(Ϫ/Ϫ) keratinocytes suggested that the catalytic activities of these kinases were normal. It is therefore possible that dephosphorylation of p130 cas was decreased due to suppression of catalytic activities and/or failure in p130 cas targeting of protein-tyrosine phosphatases. To date, protein-tyrosine phosphatases that have been demonstrated to interact with p130 cas in vivo and/or in vitro are PTP1B, SHP2, and PTP-PEST (36 -38). Interestingly, it has been reported that PTP-PEST(Ϫ/Ϫ) fibroblasts display a phenotype similar to STAT3(Ϫ/Ϫ) keratinocytes (10). The authors demonstrated that PTP-PEST(Ϫ/Ϫ) cells had decreased motility and increased focal adhesions and that PTP-PEST(Ϫ/Ϫ) cells spread faster, while p130 cas was constitutively hyperphosphorylated.
It is still unclear why a deficiency in STAT3 leads to the hyperphosphorylation of p130 cas in keratinocytes. When STAT3(Ϫ/Ϫ) keratinocytes were reconstituted with wild-type STAT3 using an adenoviral vector, the impaired migration and the tyrosine phosphorylation level in focal adhesions and p130 cas were normalized simultaneously (Figs. 4 and 5). These findings suggest that the STAT3 defect is primarily responsible for the phenotype of STAT3(Ϫ/Ϫ) keratinocytes. STAT3 is a member of the STATs activated by a variety of cytokines and growth factors such as interleukin-6, EGF, and HGF. In response to cytokine or growth factor signals, STAT3 is phosphorylated on a tyrosine residue (Tyr-705). STAT3 has an SH2 domain immediately upstream of the phosphotyrosine residue. Since the SH2 domain can recognize and bind to the phosphotyrosine residue, STAT3 molecules form dimers via their SH2 domains and phosphotyrosines. The dimers translocate to the nucleus and bind at response elements, involving expressions of specific genes (1,39). It is therefore possible that STAT3 abrogation results in a defect in the expression of proteins involving the aberrant phosphorylation of p130 cas . However, we found no difference in the expression levels of tyrosine kinases (Fig. 3b) or a tyrosine phosphatase, PTP-PEST (Fig.  3c), between control and STAT3(Ϫ/Ϫ) keratinocytes. It is also possible that STAT3 functions as an adapter molecule, which recruits kinases and/or phosphatases to p130 cas via a tyrosine residue (Tyr-705) or an SH2 domain. Recently, several reports have suggested that STAT3 plays such roles (40, 41). Chaturvedi et al. (40) reported that STAT3 interacted with v-Src in myeloid cells. Ryu et al. (41) demonstrated that STAT3 was immunoprecipitated with FAK in neutrophils. Although we could not demonstrate the interaction between STAT3 and these kinases in keratinocytes (data not shown), it is remains to be investigated whether STAT3 interacts with other signal-transducing molecules involved in p130 cas phosphorylation status.
In summary, our results suggest that STAT3 in keratinocytes plays an important role in modulating tyrosine phosphorylation of p130 cas and cell adhesiveness to the substratum leading to cell migration.