The Integrin a 7 Cytoplasmic Domain Regulates Cell Migration, Lamellipodia Formation, and p130 CAS /Crk Coupling*

The integrin a 7 b 1 is the major laminin-binding inte- grin in skeletal, heart, and smooth muscle and is a receptor for laminin-1 and -2. It mediates myoblast migration on laminin-1 and -2 and thus might be involved in muscle development and repair. Previously we have shown that a 7 B as well as the a 7 A and -C splice variants induce cell motility on laminin when transfected into nonmotile HEK293 cells. In this study we have investigated the role of the cytoplasmic domain of a 7 in the laminin-induced signal transduction of a 7 b 1 integrin regulating cell adhesion and migration. Deletion of the cytoplasmic domain did not affect assembly of the mu-tated a 7 D cyt/ b 1 heterodimer on the cell surface or adhe- sion of a 7 D cyt-transfected cells to laminin. The motility of these cells on the laminin-1/E8 fragment, however, was significantly reduced to the level of mock-trans-fected cells; lamellipodia formation and polarization of the cells were also impaired. Adhesion to the laminin-1/E8 fragment induced tyrosine phosphorylation of the focal adhesion kinase, paxillin, and p130 CAS as well as the formation of a p130 CAS -Crk complex in wild-type a 7 B-transfected cells. CAS

The integrin ␣ 7 ␤ 1 is the major laminin-binding integrin in skeletal, heart, and smooth muscle and is a receptor for laminin-1 and -2. It mediates myoblast migration on laminin-1 and -2 and thus might be involved in muscle development and repair. Previously we have shown that ␣ 7 B as well as the ␣ 7 A and -C splice variants induce cell motility on laminin when transfected into nonmotile HEK293 cells. In this study we have investigated the role of the cytoplasmic domain of ␣ 7 in the laminin-induced signal transduction of ␣ 7 ␤ 1 integrin regulating cell adhesion and migration. Deletion of the cytoplasmic domain did not affect assembly of the mutated ␣ 7 ⌬cyt/␤ 1 heterodimer on the cell surface or adhesion of ␣ 7 ⌬cyt-transfected cells to laminin. The motility of these cells on the laminin-1/E8 fragment, however, was significantly reduced to the level of mock-transfected cells; lamellipodia formation and polarization of the cells were also impaired. Adhesion to the laminin-1/E8 fragment induced tyrosine phosphorylation of the focal adhesion kinase, paxillin, and p130 CAS as well as the formation of a p130 CAS -Crk complex in wild-type ␣ 7 B-transfected cells. In ␣ 7 B⌬cyt cells, however, the extent of p130 CAS tyrosine formation was reduced and formation of the p130 CAS -Crk complex was impaired, with unaltered levels of p130 CAS and Crk protein levels. These findings indicate adhesion-dependent regulation of p130 CAS /Crk complex formation by the cytoplasmic domain of ␣ 7 B integrin after cell adhesion to laminin-1/E8 and imply ␣ 7 B-controlled lamellipodia formation and cell migration through the p130 CAS /Crk protein complex.
During muscle repair, undifferentiated muscle precursor cells, so-called satellite cells, are activated and migrate to sites of damaged muscle along the basement membranes of preexisting muscle fibers to close the wound by proliferating and fusing (1,2). In vitro, skeletal myoblasts have been shown to migrate on laminin (LN) 1 1 (3), the laminin-1-E8 fragment that is derived from laminin-1 by elastase digestion (4), and on laminin-2 (5), but not on fibronectin (4). The major component of muscle basement membranes supporting muscle cell migration is laminin-2 (6). The migration of fibroblast-like cells in culture involves polarization of cells, formation of filopodia, lamellipodia, stress fibers, and myosin-based contractility (7). Filopodia, lamellipodia and stress fiber formation are mediated by Cdc42, Rac, and Rho, respectively, which are members of the Rho family of small GTPases (8). Cdc42, Rac, and Rho can either be activated by soluble factors like growth factors, bioactive peptides, and hormones (8) or by integrins (9,10), which can transduce signals from the extracellular matrix after clustering and ligand-induced conformational changes (11) in a hierarchical fashion (12).
Several proteins become tyrosine-phosphorylated after integrin-mediated cell attachment. Those are, among others, the focal adhesion kinase (FAK) (13), the adaptor protein p130 CAS (Crk-associated Src substrate) (14,15), and paxillin (13,16). Activation of the nonreceptor FAK controls cell migration (13,(17)(18)(19). p130 CAS is an adaptor protein, which was first identified as a highly tyrosine-phosphorylated protein in v-Src-and v-Crk-transformed cells (20 -22). p130 CAS contains an N-terminal SH3 domain, a substrate domain, a proline-rich region, and several tyrosine residues near the C terminus. p130 CAS and paxillin are both Src substrates and bind to FAK with their SH3 domains (23). The adaptor protein Crk, which was first discovered as a highly tyrosine-phosphorylated protein in Rous sarcoma-transformed cells (24), forms a complex with tyrosinephosphorylated p130 CAS (25). Molecular cloning of c-Crk (26) revealed two isoforms, designated Crk I and Crk II, with molecular masses of 40 and 28 kDa, respectively. Crk II contains one N-terminal SH2 and two C-terminal SH3 domains (26). Tyrosine-phosphorylated p130 CAS can exhibit up to 15 binding sites for the SH2 domain of Crk (24), and p130 CAS /Crk binding serves as an integrin-induced switch promoting cytokine-induced migration of COS cells (27). Moreover, p130 CAS has been shown to stimulate cell migration by overexpression in Chinese hamster ovary and tumor cells (28).
Blocking integrin ␣ 7 antibodies inhibit the migration of myoblasts on laminin-1 and laminin-2, suggesting that ␣ 7 is responsible for myoblast migration on laminin (5,29). Integrin ␣ 7 is mainly expressed in skeletal, smooth, and cardiac muscle (32), but also in some glioblastoma and melanoma cells (30,31) and in nervous tissue (32,33). The extracellular and the intracellular domains of integrin ␣ 7 undergo developmentally regulated splicing (34 -36); myoblasts express the cytoplasmic splice variant B and the extracellular splice variants X1 and X2. After myotube formation, the cytoplasmic splice variants A and C and the extracellular splice variant X2 become up-regulated. The ␣ 7 chain is post-translationally cleaved into a 97-kDa fragment and a 35-kDa fragment (sizes for the B splice variant), which contains a large piece of the extracellular part (ϳ25 kDa) (11). In the mature integrin, the fragments remain disulfide-linked. Transfection of integrin ␣ 7 into ␣ 7 -deficient cells induces cell migration specifically on laminin-1 and -2 (5,37,38).
In this study we investigated the role of the cytoplasmic domain of the ␣ 7 subunit in laminin-induced signaling. We deleted the cytoplasmic domain of ␣ 7 and transfected 293 cells with a construct encoding the extracellular splice variant X2 (␣ 7 X2⌬cyt) to elucidate the role of the ␣ 7 cytoplasmic domain in terms of heterodimer formation, surface expression, integrin ␣ 7 -mediated cell attachment, migration, and p130 CAS /Crk coupling. Deletion of the ␣ 7 cytoplasmic domain did not affect receptor assembly or activity, as assessed by the ability of the mutant receptor to confer cell attachment. In contrast, cell migration, lamellipodia formation, and formation of the p130 CAS signaling complex were reduced, highlighting a role for the ␣ 7 cytoplasmic domain in signal transduction.

MATERIALS AND METHODS
Chemicals-Chemicals were from Sigma or Roth (Karlsruhe, Germany) if not stated otherwise.
Deletion of the Integrin ␣ 7 Cytoplasmic Domain-The integrin ␣ 7 X2A expression vector pCEP4␣ 7 X2A (38) was digested with NheI and Hin-dIII, which removed the cDNA segment encoding for the cytoplasmic domain except for the first two membrane-proximal amino acid residues (Lys-Leu). Ends were filled with Klenow polymerase, and the plasmid was religated, which resulted in a stop codon after the residues Lys-Leu. Plasmid DNA was purified according to the manufacturer's instructions (Qiagen, Hilden, Germany).
Cell Culture and Transfection-293HEK-EBNA cells were obtained from Invitrogen (Groningen, Netherlands) and cultured in DMEM/F-12 (Life Technologies, Inc.) containing 5% fetal calf serum (FCS; S0215-Lot 264S, Biochrom, Berlin, Germany), 50 g of streptomycin, and 50 units of Penicillin/ml (Life Technologies, Inc.), 250 g/ml G418 (Calbiochem, Bad Soden, Germany). Cells were kept in a humidified atmosphere containing 7.5% CO 2 . For certain experiments, cells were serumstarved by washing twice in serum-free medium and keeping them in serum-free medium for 20 h. The medium was replaced with serum-free medium 2 h before experiments, and for block of protein biosynthesis cycloheximide was added at a concentration of 25 M and applied for 2 h. Trypsin was stopped with 1 mg/ml soybean trypsin inhibitor (Sigma) and 1% BSA (Sigma) in DMEM/F-12 under these conditions. For transfection, 10 6 HEK293-EBNA cells were seeded on 60-mm dishes and grown for 16 h. Cells were washed twice with PBS and once with OptiMEM (Life Technologies, Inc.). Cells were incubated with 600 l of OptiMEM containing 10 g of plasmid DNA and 15 l of Lipofectin (Life Technologies, Inc.) for 6 h and then additionally 3 ml of DMEM/F-12 were added for 16 h. Medium was changed after 48 h, and cells were selected and maintained in culture medium containing 300 g/ml hygromycin B (Roche Molecular Biochemicals, Mannheim, Germany).
Cell Migration Assay-Flasks (25 cm 2 ; Falcon) were coated with PBS-diluted PLL (20 g/ml) or LN-1/E8 fragment (2 g/ml) for 1 h at 37°C with a volume of 1.5 ml/25-cm 2 flask. Flasks were washed twice with PBS and blocked with 1% heat-denatured (30 min; 80°C) BSA (Sigma; A7030) in PBS for 30 min at 37°C and again washed twice with PBS. Cells were trypsinized, washed, and plated at a density of 2000/ cm 2 in 10 mM HEPES (pH 7.4)-buffered DMEM/F-12 containing antibiotics and 5% FCS. The flasks were allowed to equilibrate in a 7.5% CO 2 -containing atmosphere at 37°C for 1 h, the lid was closed air-tight, and flasks were placed in a thermostatic chamber at 37°C under a Zeiss ICM-405 microscope (Oberkochen, Germany). Migration was monitored by time-lapse video microscopy as described previously (37). Briefly, cells were filmed under low illumination with a CCD camera (JVC) connected to a time lapse video recorder triggered by an external timer. Pictures were taken every 2 min, and cells were recorded for Ͼ12 h. For analysis of cell migration, a set of 12 pictures in 1 h steps was imported in the McDraw program and cells were tracked manually by connecting the centers of the cell bodies of individual cells. Tracks of cells were digitalized and converted to pixels, which were converted to micrometers after calibration.
Immunofluorescence Microscopy-Cells were washed quickly three times with ice-cold PBS and fixed in 3.7% p-formaldehyde in PBS for 15 min at 4°C, washed three times in PBS and permeabilized with 0.5% Triton X-100 in PBS for 30 min at room-temperature. Samples were again washed three times with PBS and blocked for 30 min at room temperature with 3% BSA in PBS. FITC-phalloidin was diluted 1:1000 in 3% BSA in PBS and applied for 1 h at room temperature. Samples were washed three times for 5 min in PBS, mounted, and examined with a Zeiss Axiophot microscope equipped with a 63ϫ oil immersion objective (numeric aperture ϭ 1.24).
Cell Lysis, Immunoprecipitation, and Western Blot Analysis-Cell lysis was performed in two ways, depending on the application. Cells were washed once with ice-cold PBS and lysed with 1 ml of buffer/10 7 cells. Condition A (LN-1/E8 chromatography) consisted of 50 mM Noctylglucopyranoside, 300 mM NaCl, 25 mM Tris/HCl, pH 7,4, 1 mM MnCl 2 , 1 mM CaCl 2 , 1 mM N-ethylmaleimide (Merck, Darmstadt, Germany), 1 mM PMSF (phenylmethylsulfonyl fluoride; Merck). Condition B (coimmunoprecipitations) was 50 mM HEPES, pH 7.5, 150 mM NaCl, 1% Triton X-100, 1 mM sodium vanadate, 50 mM sodium fluoride, 1 mM PMSF, 5 mM EDTA. Cells were scraped into ice-cold lysis buffer and allowed to lyse for 30 min on a shaking platform at 4°C, and the lysate was spun down at 10,000 ϫ g for 15 min at 4°C. The protein concentration of the supernatant was determined by Cu 2ϩ complexation (Pierce). Samples were adjusted to equal protein concentrations with lysis buffer. For immunoprecipitations samples were adjusted to a volume of 1 ml with lysis buffer and precleared with equilibrated Protein G-Sepharose (Amersham Pharmacia Biotech, Uppsala, Sweden) for 30 min at 4°C by rotation. After centrifugation appropriate antibodies were added to the supernatant, the mixture was rotated for 1 h, and 20 -50 l of equilibrated Protein-G-Sepharose were added for an additional 1 h. If not stated otherwise, antibodies were used at a concentration of 5 g/mg of protein. Coimmunoprecipitations were performed the same way, but for 2 h. The immune complexes were retained by short pulse centrifugation, washed three times in lysis buffer (for condition A: lysis buffer containing 25 mM N-octylglucopyranoside), and finally boiled in 2ϫ SDS sample buffer (41). If not otherwise indicated, SDS-PAGE was performed with 10% gels. Proteins were transferred to nitrocellulose (0.2 m, Schleicher & Schuell, Dassen, Germany) for 1.5 h at 1 mA/cm 2 by semidry blotting and a discontinuous buffer system. Blots were stained with Ponceau S and blocked with 3% BSA in TBST (0.1% Tween 20, 25 mM Tris/HCl, pH 7.4, 150 mM NaCl) for 1 h at room temperature. Antibodies were diluted in blocking solution, and incubated with the blot for 1 h at room temperature or for 16 h at 4°C. 2 H. von der Mark, manuscript in preparation.
Blots were washed four times for 10 min each with TBST, developed with peroxidase-conjugated secondary antibodies and chemiluminescence, and exposed to Kodak XAR-5 films (Eastman Co.). For stripping, blots were washed twice with distilled water, incubated two times for 10 min with 0.1 M glycine, 0.5 M NaCl, 0.1% Tween 20, 2% ␤-mercaptoethanol, pH 2.5, neutralized extensively with TBST, and blocked again.
Cell Surface Biotinylation-Cells were washed once in PBS containing 5 mM EDTA and released from plates with 5 mM EDTA in PBS. After washing three times with ice-cold PBS, pH 8.0, cells were adjusted to 25 ϫ 10 6 /ml and nonmembrane permeable NHS-LC-Sulfobiotin (Pierce) was added to a final concentration of 0.5 mg/ml. Surface biotinylation was carried out for 30 min at room temperature on a shaking platform, and the reaction was stopped by three washes with ice-cold PBS, 10 mM Tris/HCl, pH 8.0. Cells were lysed under condition A.
LN-1/E8-Sepharose Chromatography-The LN-1/E8 fragment used throughout this study was a generous gift of Dr. Rainer Deutzmann (University of Regensburg, Regensburg, Germany). It was coupled to CNBr activated Sepharose CL 4-B according to the manufacturer's instructions (Amersham Pharmacia Biotech), which resulted in ϳ1 mg of LN-1/E8 fragment/ml of Sepharose. For purification of recombinant integrin ␣ 7 ␤ 1 complexes from transfected 293 cells, LN-1/E8-Sepharose was equilibrated in lysis buffer A by three washes and 200 l of a 1:3 suspension was added to surface-biotinylated cell extracts from 10 7 cells. The mixture was rotated for 4 h at 4°C and washed three times in lysis buffer A containing 25 mM N-octylglucopyranoside by short centrifugation at 200 ϫ g and resuspension in 1 ml of wash buffer. The recombinant protein complexes were eluted in a volume of 50 l with 5 mM EDTA in 25 mM N-octylglucopyranoside, 300 mM NaCl, 25 mM Tris/HCl pH 7,4, 1 mM N-ethylmaleimide, 1 mM PMSF by vortexing and centrifugation at 200 ϫ g. Eluted proteins were subjected to Western blot analysis and probed with streptavidin-peroxidase complex (1:5000, Amersham Pharmacia Biotech)., LN-1/E8-Sepharose was recycled by repeated washes in 1 M NaCl, 5 mM EDTA, 20 mM Tris/HCl, pH 7.4 and stored at 4°C in TBS containing 0.1% sodium azide.
Cell Attachment Assay-Plates (96 wells; Nunc, Denmark) were coated with 100 l of protein solution/well for 1 h at 37°C with coating concentrations as indicated under "Results." Plates were then washed twice with PBS and blocked with 1% heat-denatured (30 min; 80°C) BSA (Sigma; 7030). Cells were trypsinized and washed either in DMEM containing 1% BSA and 1 mg/ml trypsin inhibitor under serum-free conditions or in DMEM containing 5% FCS and kept in suspension for 30 min. 5 ϫ 10 5 cells were seeded per well and allowed to attach for 1 h at 37°C. Plates were washed three times with PBS under standardized conditions with an enzyme-linked immunosorbent assay washer (M96V; Merlin, Rotterdam, Netherlands), and the amount of attached cells was determined by measurement of lysosomal hexosaminidase activity (42). Background attachment on BSA was subtracted, and the percentage of attached cells was calculated using serially diluted cells (1:3) as standard. For antibody attachment inhibition assays, cells were incubated with 10 g/ml purified antibody on ice prior to distribution in wells. The antibody concentration was kept at 10 g/ml throughout the assay. Attachment assays were performed with internal controls, e.g. nontransfected 293 cells.
Magnetic Cell Sorting-Sorting was performed over two rounds as described previously (38). ␣ 7 -transfected cells were trypsinized shortly, washed with culture medium, and adjusted to 3.3 ϫ 10 7 cells/ml in 3C12 hybridoma supernatant diluted 1:1 with PBS (corresponding to approximately 25 g/ml specific anti-␣ 7 antibody). The mixture was rocked at 4°C for 30 min, and cells were washed twice with 4 ml of ice-cold culture medium. Cells were resuspended at a concentration of 10 8 /ml in ice-cold culture medium. Sheep pan-anti-mouse coated magnetic beads (M-450; Dynal, Oslo, Norway) were washed three times in PBS containing 0.1% heat-denatured BSA, added to the cell suspension (1.3 ϫ 10 8 beads/ml), and cells were rotated with the beads for 30 min at 4°C. Cells bound to the beads were pulled out magnetically and washed three times with ice-cold medium. Cells were resuspended in warm culture medium, and beads were removed magnetically after trypsin treatment during passaging of the cells.
Analysis of Cell Spreading-5 ϫ 10 4 cells were plated on laminin-1/ E8-coated 30-mm dishes, and three pictures of each plate were taken with a Zeiss Axiovert microscope (Kodak TMax 100 film) by a second individual blinded for the experimental condition. Films were developed, and quantitation of cell spreading was performed by calculating the percentage of spread versus total cells by a blinded, second person. Data were analyzed statistically with a two-tailed Student's t test.

Heterodimerization and Surface Expression of ␣ 7 ␤ 1 Do Not
Require the Cytoplasmic Domain of Integrin ␣ 7 X2-We reported previously that integrin ␣ 7 X2B, 3 ␣ 7 X2A, and ␣ 7 X2C induce cell migration specifically on laminin-1 and its E8 fragment (LN-1/E8) when expressed ectopically in nonmotile, integrin ␣ 7 -negative 293HEK cells (38). To elucidate the role of the ␣ 7 B cytoplasmic domain, a truncation mutant lacking the cytoplasmic domain except of the two first membrane-proximal residues (Lys-Leu) 3 was constructed and transfected in HEK293-EBNA cells. Transfected 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells were magnetically sorted with anti-␣ 7 mAb 3C12 for high surface expression levels ( Fig. 1) and scanned by FACS analysis with two different anti-␣ 7 mAbs, a nonblocking (3C12) as well as a blocking antibody (6A11). The wild-type and mutant cells were rather homogenous and displayed similar surface expression levels of integrin ␣ 7 (Fig. 1A and Table I), indicating that deletion of the cytoplasmic domain of integrin ␣ 7 X2 does not interfere with cell surface presentation. 293␣ 7 X2⌬cyt cells displayed reduced surface levels of integrin ␣ 6 , thus behaving similarly to the wild-type receptor (37). Immunoblot analysis confirmed deletion of the cytoplasmic domain (Fig. 1B, lanes 3, 6, and 9) and demonstrated that the total ␣ 7 expression level was not reduced after deletion of the ␣ 7 cytoplasmic domain (compare Fig. 1B, lanes 2 and 3). These data suggest that the truncated protein does not differ from the wild-type protein with respect to stability. However, some degradation products were observed in both 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells (Fig.  1B, lanes 2, 3, and 8) and a large part of the transfected integrin was not processed (Fig. 1B, lane 8). Both findings are probably due to overexpression of ␣ 7 .
showed essentially the same behavior as the full-length receptor. Attachment of mock-transfected cells to the laminin-1 E8 fragment is mediated by integrin ␣ 6 and switches to ␣ 7 after overexpression of ␣ 7 due to down-regulation of ␣ 6 (see Fig. 1 and Ref. 38). A LN-1/E8 fragment coating concentration of 2 g/ml was chosen for the following cell migration and biochemical experiments (Figs. 3-9) because wild-type and ⌬cyt cells attached similarly at these coating concentrations.
The Integrin ␣ 7 Cytoplasmic Domain Controls Integrin ␣ 7mediated Cell Migration and Polarization-Although surface presentation of integrin ␣ 7 and cell adhesion to LN-1/E8 did occur independently of the ␣ 7 cytoplasmic domain, its deletion affected significantly cell motility on LN-1/E8, i.e. the mean migration speed (Fig. 3A). Deletion of the integrin ␣ 7 cytoplasmic domain reduced ␣ 7 -dependent cell motility significantly to about 50% of the wild-type level (p Ͻ 0.01). Integrin ␣ 7 -mediated cell migration required furthermore the presence of serum because serum-starved cells did not migrate under serum-free conditions (Fig. 3A).
The cell attachment assays presented in Fig. 2, showing equal attachment of 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells to LN-1/E8, had been carried out under serum free-conditions, in contrast to cell migration assays. It seemed possible that the different migration rates of 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells were due to different effects of serum upon their attachment cells. To rule this out, serum-starved cells were plated in the presence of serum or 1% BSA on LN-1/E8. 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells attached similarly under either condition (Fig. 3B), indicating that the reduced cell migration was not a consequence of differences in cell attachment in the presence of serum.
To test whether serum alone accounted for the observed migration, cells were plated on PLL in the presence of serum and video images were taken 1 h after plating the cells and after 12 h (Fig. 4, A and B). Cells did neither spread nor move under these conditions (see Fig. 4, arrows), thus confirming serum alone does not induce cell migration Analysis of the cell morphology of migrating cells by microscopy revealed that 293␣ 7 X2⌬cyt cells displayed a different morphology on LN-1/E8 than cells expressing the wild-type receptor (Fig. 4). To analyze this in detail, we quantitated the percentage of spread cells plated on LN-1/E8 (Fig. 4). First, only about 40% of 293␣ 7 X2⌬cyt cells spread after 30 min in contrast to about 80% of 293␣ 7 X2B cells. After 16.5 h of adhesion to LN-1/E8, 80% of 293␣ 7 X2B cells were still spread, whereas spreading 293␣ 7 X2⌬cyt cells reached only a level of 60% (Fig. 4E). Second, despite spreading as compared with cells plated on PLL, 293␣ 7 X2⌬cyt cells remained in a more or less round shape and extended filopodia, reflecting a different organization status of the actin cytoskeleton as compared with 293␣ 7 X2B cells. These elongated and polarized (see Fig. 4, C and D). To examine changes in the actin cytoskeleton cells were fixed directly after cell migration experiments (i.e. after Ͼ12 h) and stained with FITC-phalloidin (Fig. 5). 293␣ 7 X2B cells showed extended lamellipodia, ruffled their membranes, and organized stress fibers in contrast to 293␣ 7 X2⌬cyt cells, which exhibited less lamellipodia, but showed a dramatic increase in number and size of filopodia (Fig. 5). The results from this experiment implicate that the integrin ␣ 7 cytoplasmic domain controls a pathway regulating cytoskeletal architecture.
The Integrin ␣ 7 Cytoplasmic Domain Controls ␣ 7 -initiated Tyrosine Phosphorylation-Activation of nonreceptor protein tyrosine kinases like FAK and Src and subsequent proteinprotein interactions are rapid responses of cells to attachment to ECM molecules (44) and are believed to regulate cell adhesion as well as cell migration (45). We thus examined tyrosine phosphorylation events specifically induced by integrin ␣ 7 . Adhesion of 293␣ 7 X2B cells to the LN-1/E8 fragment induced tyrosine phosphorylation of 60 -80-and 120 -140-kDa proteins already after 10 min (data not shown). To identify the tyrosinephosphorylated proteins, Triton X-100 extracts from serumstarved, suspended 293␣ 7 X2B cells and from serum-starved 293␣ 7 X2B cells plated on PLL and LN-1/E8 fragment were subjected to immunoprecipitation with antibodies against tensin, p130 CAS , FAK, vinculin, paxillin, Erk2, and Shc or antiphosphotyrosine-agarose (Fig. 6). Immunoprecipitation with anti-phosphotyrosine-agarose revealed three major tyrosinephosphorylated bands of 60 -70, 120, and 130 kDa on LN-1/E8 fragment, which were not seen when cells were kept in suspension or plated on PLL. Three of these proteins were identified as p130 CAS , FAK, and paxillin (Fig. 6B, lanes 5, 6, and 8). Tensin and vinculin were not tyrosine-phosphorylated (Fig. 6B,  lanes 4 and 7), nor were Erk or Shc (data not shown).
Integrin ␣ 7 -initiated p130 CAS Tyrosine Phosphorylation and p130 CAS /Crk Coupling Are Dependent on the Cytoplasmic Domain of ␣ 7 -Klemke and co-workers (27) reported an essential role for p130 CAS in cell migration. The observation of integrin ␣ 7 -dependent p130 CAS tyrosine phosphorylation prompted us to examine the role of the integrin ␣ 7 cytoplasmic domain in this event. Starved 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells were plated on LN-1/E8 fragment in the absence or in the presence of serum for 30 min and lysed. Lysates were immunoprecipitated with anti-p130 CAS , and immunoprecipitates were subjected to anti-phosphotyrosine blotting (Fig. 7). Fig. 7 shows that p130 CAS was not phosphorylated on tyrosine residues when 293␣ 7 X2B cells were kept in suspension or plated on poly-L-lysine. Tyrosine phosphorylation of p130 CAS was, however, induced by E8 fragment via integrin ␣ 7 and strongly enhanced after addition of serum, paralleling the effect of serum on cell migration. In contrast, 293␣ 7 X2⌬cyt cells plated on LN-1/E8 in the presence of serum showed a markedly reduced p130 CAS tyrosine phosphorylation with unaltered p130 CAS protein levels. Thus, the presence of the ␣ 7 cytoplasmic domain is necessary for full p130 CAS tyrosine phosphorylation in this system. Furthermore, the cytoplasmic domain of integrin ␣ 7 was essential for coprecipitation of a tyrosine-phosphorylated 60-kDa protein with p130 CAS , which we could not identify so far. This protein coprecipitated with p130 CAS from 293␣ 7 X2B cells but not from 293␣ 7 X2⌬cyt cells. We also examined the effect of the deletion of the ␣ 7 cytoplasmic domain on FAK tyrosine phosphorylation (Fig. 8). According to the data obtained with CAS precipitates, we also found less tyrosine phosphorylation of FAK in 293␣ 7 ⌬cyt cells plated on LN-1/E8 as compared with 293␣ 7 X2B cells.
The reduced tyrosine phosphorylation of p130 CAS in the 293␣ 7 X2⌬cyt cells suggested that p130 CAS /Crk coupling could be affected due to fewer Crk SH2-binding sites offered by p130 CAS . For examination of p130 CAS /Crk complexes, 293␣ 7 X2B and 293␣ 7 X2⌬cyt were plated on LN-1/E8 fragment for 1, 2, and 4 h in the presence of serum (conditions used for cell migration) or kept in suspension for 4 h in the presence of serum. Crk was immunoprecipitated from the standardized cell lysates and the precipitates were probed with anti-phosphotyrosine, anti-p130 CAS , and anti-Crk antibodies (Fig. 9A). p130 CAS coprecipitated with Crk in 293␣ 7 X2B cells only after cell adhesion to LN-1/E8 but not when cells were kept in suspension. However, p130 CAS was absent in Crk immunoprecipitates obtained from 293␣ 7 X2⌬cyt cells although it was present in each lysate used for Crk immunoprecipitation (Fig. 9B). Crk itself was not tyrosinephosphorylated (data not shown). Thus, our results point to an essential role for the integrin ␣ 7 cytoplasmic domain in p130 CAS / Crk coupling.

DISCUSSION
In the present study we have investigated the function of the integrin ␣ 7 X2 cytoplasmic domain by comparing 293EBNA cells expressing an integrin ␣ 7 wild-type receptor or an integrin ␣ 7 lacking the cytoplasmic domain. The cells were compared in terms of (i) ␣ 7 protein expression, (ii) ␣ 7 surface presentation, (iii) cell attachment and migration conferred by ␣ 7 , and (iv) initiation of p130 CAS /Crk signaling complexes. Both cell types assembled ␣ 7 ␤ 1 heterodimers and attached equally well on LN-1/E8. We showed that the integrin ␣ 7 B cytoplasmic domain is not required for heterodimer formation with ␤ 1 and not necessarily linked to ␣ 7 protein stability, surface expression, or receptor activation but contributes to cell spreading, migration, and intracellular signaling via p130 CAS /Crk complex formation in a serum-dependent manner.
Role of the Integrin ␣ 7 X2 Cytoplasmic Domain in Receptor Assembly and Activity-The function of the ␣ cytoplasmic domains of various other ␤ 1 integrin receptors has been extensively studied by truncation and point mutation analysis. It has been reported that particularly the two phenylalanines in the conserved GFFKR motive regulate heterodimerization and surface transport, e.g. in the case of ␣ 6 ␤ 1 (46), and protein stability in the case of ␣ 6 ␤ 4 (47,48). Moreover, deletion of the cytoplasmic domains of ␣ IIb and ␣ 1 including the GFFKR caused reduced surface expression and heterodimerization but, on the other hand, resulted also in a high integrin affinity state of  *). B, the observed difference in cell migration between 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells is not due to differences in cell attachment. 293␣ 7 X2B cells (gray bars) and 293␣ 7 X2⌬cyt cells (white bars) were allowed to attach to laminin-1 LN-1/E8 fragment (2 g/ml) for 1 h either in the presence of serum or in the presence of 1% BSA. No significant difference in cell attachment was observed under either condition.
In marked contrast, we show here that deletion of the complete ␣ 7 X2 cytoplasmic domain (including the conserved GFFKR motive) did not influence ␣ 7 X2 heterodimerization with the integrin ␤ 1 chain, correct processing of ␣ 7 or cell attachment conferred by ␣ 7 ␤ 1 . This is, on the other hand, in agreement with other studies showing that the GFFKR sequence does not influence inside-out signaling or surface presentation of ␣ 1 ␤ 1 and ␣ 5 ␤ 1 integrins (49,51), as deletion of the GFFKR motive did not alter the activation state of the affected receptors. Thus, inte-grins differ in the requirement of the GFFKR motive for heterodimerization, surface transport, and inside-out activation. This may be due to different affinities of the ␣ chains for the ␤ chain as proposed previously (47).
For splice variant, required activation with the ␤ 1 activating antibody TS2/16 to bind to laminin-1 in MCF7 cells (52), although both extracellular splice variants carried the same intracellular splice variant. In contrast, ␣ 7 X2B-␤ 1 complexes are constitutively active and bind to laminin-1 and -2 when expressed in HEK293-EBNA cells and MCF-7 cells, independently of the cytoplasmic domain (Refs. 37, 38, and 52; this report). Thus, our results are in support of a key role of both intra-and extracellular domains of integrin ␣ 7 in inside-out signaling and ligand binding.
Effect of the Integrin ␣ 7 B Cytoplasmic Domain on Cell Spreading and Migration-Cell motility on the extracellular matrix is largely dependent on integrin surface expression levels, integrin activation state, and matrix concentration (53). Therefore, quantitative biochemical analysis of the truncated receptor and analysis of cell attachment was necessary to allow a direct functional description of the role of the integrin ␣ 7 cytoplasmic domain in cell migration and signaling events. The results of these studies show that (a) ␣ 7 surface and total protein levels in wild-type and mutant transfected cells were similar, (b) processing and heterodimer formation were independent of the integrin ␣ 7 cytoplasmic domain, and (c) 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells attached similarly via integrin ␣ 7 to LN-1/E8 at coating concentrations of 2 g/ml. Taking these parameters into account, we conclude that deletion of the ␣ 7 cytoplasmic domain significantly affected cell migration.
␣ 7 -mediated cell migration and cytoskeletal reorganization was specific for the LN-1/E8 fragment. 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells plated on PLL did not display marked differences and did not migrate, even in the presence of serum. Thus, integrin ␣ 7 -mediated cell migration of transfected 293 cells is not due to endogenous proteins deposited as ligands like such as fibronectin. However, serum factors are required for integrin ␣ 7 -mediated continuous cell migration. This is consistent with the notion that sustained cell migration requires the presence of growth factors (54). Thus, integrin ␣ 7 alone is not sufficient to provide enough signals for cell migration but cooperates with soluble, so far by us unspecified serum factors.
In support of the results we obtained from our cell migration experiments we observed fewer lamellipodia, cell polarization, and stress fibers in 293␣ 7 X2⌬cyt cells plated on LN-1/E8 than in 293␣ 7 X2B cells plated on LN-1/E8. In contrast, filopodial extensions were remarkably increased in 293␣ 7 X2⌬cyt cells. Deletion of the ␣ 7 cytoplasmic domain may lead to a block in transmission of signals from cdc42 to Rac, hence from Rac to Rho, and thus in accumulation of filopodia-inducing signals in 293 cells. This would explain the lack of membrane ruffling and stress fibers. Nobes and co-workers (55) reported that cdc42 is required for cell polarization by placing lamellipodia at the leading edge, but we did not observe polarization in 293␣ 7 X2⌬cyt cells on LN-1/E8 yet observing the hallmark of cdc42 activation, filopodia. Cells may not be able to polarize due to lack of lamellipodial formations, despite extending filopodia and attaching.
Influence of the Integrin ␣ 7 X2 Cytoplasmic Domain on p130 CAS /Crk Coupling-The role of the adaptor protein p130 CAS and the p130 CAS /Crk complex in ␣ v ␤ 3 and ␣ 5 ␤ 1 integrin-mediated cell migration on fibronectin has previously been demonstrated (27,28). p130 CAS becomes tyrosine-phosphorylated after cell adhesion to fibronectin, and this allows formation of a p130 CAS /Crk complex (25). Serum factors like plateletderived growth factor, lysophosphatidic acid, and bombesin induce p130 CAS tyrosine phosphorylation as well, leading to formation of a p130 CAS /Crk complex (56). We showed for the first time that not only fibronectin, but also laminin can pro-   8. The integrin ␣ 7 cytoplasmic domain is involved in FAK tyrosine phosphorylation. 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells were serum-starved for 16 h, trypsinized, and resuspended in 1% BSAcontaining medium. Cells were recovered by centrifugation and resuspended in serum-containing medium (ϩ). Cells were kept in suspension for 30 min and then either plated on LN-1/E8 (LN-1/E8) for 30 min, or kept in suspension for another 30 min (sus). Triton X-100 extracts were prepared, and 800 g of total protein were subjected to anti-p125 FAK IP. IPs were separated by 10% SDS-PAGE, transferred to nitrocellulose, and probed with anti-phosphotyrosine (PY; clone 2C8) or anti-p125 FAK (p125FAK). Molecular mass positions are shown on the left. mote p130 CAS tyrosine phosphorylation and p130 CAS /Crk coupling, which occurs through integrin ␣ 7 in our system. We observed less p130 CAS tyrosine phosphorylation and also less p125 FAK tyrosine phosphorylation in 293␣ 7 X2⌬cyt cells even in the presence of serum, and thus we conclude that the integrin ␣ 7 cytoplasmic domain is cooperatively involved in signals mediated by soluble factors. It is well known that integrins collaborate with growth factors and integrin signaling pathways converge with those of soluble factors and their receptors like epidermal growth factor (57), platelet-derived growth factor (58), fibroblast growth factor (59), transforming growth factor ␤ (60), and vascular endothelial growth factor (61), especially in the p130 CAS of cell migration (62). Our data strongly support these observations because ␣ 7 and LN-1/E8 fragment-induced tyrosine phosphorylation of p130 CAS in particular is strongly enhanced by serum. It has been reported that p130 CAS expression levels correlate directly with cell migration (28). However, there were no different p130 CAS or Crk protein levels in 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells, and thus the difference in cell migration was not due to different p130 CAS or Crk protein levels. Our results rather suggest strongly that the reduced cell migration, caused by deletion of the ␣ 7 cytoplasmic domain, is due to impaired p130 CAS /Crk coupling and impaired p130 CAS tyrosine phosphorylation.
p130 CAS can be tyrosine-phosphorylated by Src in a FAK-dependent and -independent manner after attachment of fibroblasts to fibronectin; the created phosphotyrosine residues provide binding sites for the SH2 domain of Crk (25). Lack of a p130 CAS /Crk complex in 293␣ 7 X2⌬cyt compared with wild-type cells could be due to less p130 CAS tyrosine phosphorylation, which we did observe. Since p130 CAS is a substrate for Src and Fyn (25), less p130 CAS tyrosine phosphorylation induced by deletion of the ␣ 7 cytoplasmic domain may be due to reduced Src or Fyn kinase activity induced by the mutant versus the wild-type receptor. Accordingly, deletion of the C-terminal 23 amino acids of integrin ␣ v reduced Src activation induced by osteopontin as compared with the wild-type ␣ v ␤ 3 receptor (63).
To identify the 60-kDa phosphoprotein coprecipitating with p130 CAS , we performed coimmunoprecipitation analyses of p130 CAS with Src and Lyn, both of which being expressed at same protein levels in 293␣ 7 X2B and 293␣ 7 X2⌬cyt cells (data not shown). However, we could not detect complexes of p130 CAS with Src or Lyn, nor with FAK (data not shown), suggesting that the 60-kDa protein is neither Src nor Lyn, and that CAS tyrosine phosphorylation induced by integrin ␣ 7 in 293 cells may occur in a FAK-independent manner, as is the case in p125 FAKϪ/Ϫ cells. There, p130 CAS can be phosphorylated on tyrosine by cell adhesion kinase ␤ (64).
A mechanism explaining the failure of lamellipodia formation induced by 293␣ 7 X2⌬cyt cells on LN-1/E8 may finally be through reduction of p130 CAS /Crk induced signaling to Rac; Klemke and coworkers (27) have demonstrated that p130 CAS / Crk coupling acts in a Rac-dependent manner on cell migration. It seems possible from our morphological and biochemical data that deletion of the ␣ 7 cytoplasmic domain may reduce Rac activation in comparison with the nonmutated ␣ 7 receptor as a consequence of reduced p130 CAS tyrosine phosphorylation.