INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Integrin-mediated cell adhesion to the extracellular matrix (ECM) is crucial for multiple biological functions, including cell growth, differentiation, survival, cytoskeleton reorganization, migration, tumor metastasis, and embryonic development (10, 23, 26, 30, 46, 59, 68, 71). Integrins are heterodimeric transmembrane receptors composed of - and -subunits whose specificity for different ECMs are determined by their combination (26). Cell attachment to the ECM induces integrin clustering and recruitment of a number of intracellular proteins, such as focal adhesion kinase (FAK), paxillin, vinculin, talin, and p130 Cas (8, 60, 65, 68) to specialized sites of the inner cytoplasmic membrane to form focal adhesion complexes. These complexes link the actin cytoskeleton and regulate intracellular signaling pathways, thereby coordinating cell attachment with cell architecture, movement, and gene expression (68). Although the molecular mechanism of integrin-mediated signal transduction has not been well defined, tyrosine phosphorylation of several cytoplasmic proteins, including FAK, paxillin, tensin, and Cas seems to be a critical biochemical aspect of this process (5, 9, 19 54, 60, 81). On integrin-mediated tyrosine phosphorylation, these proteins bind to signaling molecules such as Crk, Grb2, Nck, and Src that contain the Src homology 2 (SH2) domain (21, 68).

Paxillin is one of the focal adhesion proteins (17, 81, 82) that was originally identified as a major tyrosine phosphorylated protein in cells transformed by v-Src and v-Crk (4, 17). Paxillin associates with signaling molecules and cytoskeletal proteins, such as FAK, Pyk2/Cak/RAFTK, c-Src, PTP-PEST, talin, and vinculin, and has been suggested to be involved in the regulation of focal adhesion dynamics (6, 11, 61). In particular, the association of FAK with paxillin is essential for focal adhesion targeting of FAK (74). In addition, paxillin is tyrosine phosphorylated following integrin stimulation by FAK and/or other kinases that are associated with FAK (66, 82). This phosphorylation creates docking sites for the SH2 domain of Crk (3) and links integrin stimulation to downstream signaling pathways through guanine nucleotide exchange proteins such as C3G (77), SOS (44), and DOCK180 (22). Thus, paxillin is involved in integrin-mediated signal transduction as a scaffold of these signaling molecules. In addition, a recent study demonstrated that the tyrosine residues at positions 31 and 118 on paxillin regulate cell migration through association with Crk in NBT-II cells (53). Paxillin also binds to the recently identified paxillin-kinase linker, p95PKL, that links paxillin to p21 GTPase-activated kinase, PAK, and the guanine nucleotide exchange protein, PIX, through the LD domain that is defined as a leucine-rich motif for association with interacting proteins (84). The formation of the paxillin-p95PKL-PAK-PIX complex has been suggested to play an important role in cytoskeletal organization (84).

Hic-5, a paxillin homologue, was originally identified as a transforming growth factor 1 (TGF-1)- and hydrogen peroxide-inducible gene by differential hybridization (69). Its expression is increased during cellular senescence of normal human fibroblasts and is decreased during immortalization of mouse embryo fibroblasts (MEFs) (28, 69). The forced expression of Hic-5 in immortalized fibroblasts induced growth retardation, senescence cell-like morphology (i.e., cells became enlarged, flattened, and well spread), and the increased gene expression of p21/WAF1/Cip1/sdi1 and ECM-related proteins such as collagen, fibronectin, and collagenase (70). These observations suggested that Hic-5 is involved in the negative regulation of cell growth, including the senescence process and TGF signal transduction. However, the molecular mechanisms of the function of Hic-5 have not been clarified. Recent studies have shown that Hic-5 is a cytoskeletal protein localized to focal adhesions in fibroblasts (15, 28, 45). In addition, Hic-5 contains four LD motifs in the N-terminal half and four LIM domains in the C-terminal half (6, 70). These regions are well conserved in paxillin (6, 70). Several proteins have been identified as Hic-5 interacting factors, including FAK (6, 15), Pyk2 (45), PTP-PEST (50), vinculin (6), and p95PKL (84). These are shared with paxillin and are considered to play important roles in integrin-mediated signal transduction and remodeling of the actin cytoskeleton. Despite their similarity, Hic-5 has some features distinct from paxillin. Unlike paxillin, Hic-5 is not phosphorylated by integrin stimulation (15) and does not have a target site for the Crk SH2 domain (80). Furthermore, the Hic-5 expression level was specifically decreased during immortalization, whereas that of paxillin tended to increase (28). Taken together, we hypothesized that Hic-5 could compete for common interaction factors with paxillin and antagonize the signaling pathways that involve paxillin.

Most adherent cells respond to ECM by adhering and then spreading out to acquire a flattened morphology. This process is mediated by integrins and involves dynamic rearrangements of the actin cytoskeleton, which are regulated by intracellular signaling pathways (58, 85). The cell spreading assay provides a convenient model for examining the formation of focal adhesions and cell migration (57). In the present study, we adopted this assay system to assess the possibility that Hic-5 competed with paxillin for integrin-mediated signal transduction. The results presented below indicated that the altered balance between paxillin and Hic-5 expression affected cell spreading and that Hic-5 overexpression inhibited cell spreading.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Antibodies. Anti-mouse Hic-5 polyclonal antibody was raised against recombinant Hic-5 as described previously (28). Anti-FAK rabbit polyclonal antibody C-20 was obtained from Santa Cruz Biotechnology (Santa Cruz, Calif.). Anti-paxillin and antiphosphotyrosine (PY20) mouse monoclonal antibodies were from Transduction Laboratories (Lexington, Ky.). Anti-p130Cas polyclonal antibody was a generous gift from Tatsuya Nakamoto, University of Tokyo. Anti-hemagglutinin (HA) (12CA5) and anti-Myc (9E10) monoclonal antibodies were purchased from Boehringer Mannheim Co. (Indianapolis, Ind.) and Upstate Biotechnology (Lake Placid, N.Y.), respectively. Horseradish peroxidase (HRP)-conjugated anti-rabbit immunoglobulin G (IgG) and anti-mouse IgG antibodies were from Amersham Pharmacia Biotech (Buckinghamshire, United Kingdom). Fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG antibody was from Dako (Copenhagen, Denmark).

Cell culture and transfection. NIH 3T3 cells were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% heat-inactivated calf serum and 50 µg of kanamycin per ml at 37°C in an atmosphere of 5% CO₂ in air. Mortal and immortal MEFs were established as described previously (28) and cultured in DMEM supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 50 µg of kanamycin per ml in an atmosphere of 5% CO₂ in air. FAK-null cells derived from FAK knockout mice were a generous gift from Tadashi Yamamoto, Institute of Medical Science, University of Tokyo. Cells were transfected with plasmid DNAs using Lipofectamine Plus reagent (Life Technologies, Inc., Rockville, Md.) according to the manufacturer's protocol.

Plasmids. For construction of the HA-tagged mouse Hic-5 expression vector (pCG-mhic-5), the NspI-ApaI fragment from the CMV/S5 described elsewhere (69) was blunted and ligated with BamHI linkers for in-frame insertion into the expression vector driven by the cytomegalovirus promoter pCG-N-BL (25). After BamHI digestion, the linker-linked cDNA fragment was inserted into the BamHI site of the pCG-N-BL vector. For HA-tagged human Hic-5, an EcoRI-HindIII fragment cut out from Myc-tagged human Hic-5/pcDNA3.1A (described below) was blunted at the EcoRI site and inserted into the expression vector pCG-N-BL digested with EcoRV and HindIII.

Cell spreading. At 48 h after transfection, NIH 3T3 cells were collected by trypsinization and washing with serum-free DMEM containing 1% bovine serum albumin (BSA). Cells resuspended in serum-free DMEM containing 1% BSA were replated on coverslips coated with 5 µg of fibronectin (Upstate Biotechnology) per cm² and placed on ice for 5 min. Subsequently, the cells were allowed to spread for 30 min at 37°C. Cells attached to coverslips were fixed with 3.7% formaldehyde in phosphate-buffered saline (PBS) for 10 min and permeabilized with 0.2% Triton X-100 in PBS for 3 min. Cells were blocked with 3% BSA in PBS containing 0.1% Tween 20, treated with anti-HA antibody (12CA5) at 1:300 or with anti-Myc antibody (9E10) at 1:300 for 1 h, and then incubated with tetramethyl rhodamine isocyanate (TRITC)-conjugated phalloidin (Sigma) at 1:300 and with FITC-conjugated anti-mouse IgG (Dako) at 1:300 for 1 h. After being washed with PBS containing 0.1% Tween 20, the cells were mounted and visualized under a low-magnification fluorescence microscope. Transfected cells were counted, and the percentage of spread cells was evaluated. Nonspread cells were defined as round phase-bright cells, whereas spread cells were defined as those that lacked a rounded shape, were not phase-bright, and had extended membrane protrusions. In each experiment, more than 150 cells were counted. At least four independent experiments were performed for each combination of experiments.

Virus infection. Retroviral expression vectors were cotransfected with the packaging vector pCL-Eco (49) into 293 cells by the calcium phosphate method. At 48 h after transfection, the supernatants were collected, filtered through 0.45-µm (pore size) filters, and then used to infect MEFs in the presence of 80 µg of Polybrene per ml. The day before infection, MEFs were trypsinized and plated to reach 50 to 80% confluence on the day of transfection. Supernatants containing viruses were poured onto MEFs and, 1 day after infection, the media were replaced with fresh media.

Cell stimulation with fibronectin. Cells were starved of serum in DMEM containing 0.5% FBS and 1% BSA for 16 h and harvested with PBS containing 0.05% trypsin and 2 mM EDTA. Trypsin was inactivated by adding 0.5 mg of soybean trypsin inhibitor per ml with 1% BSA in DMEM, and cells were collected by centrifugation, resuspended in DMEM containing 10 mM HEPES (pH 7.4) and 1% BSA, and held in suspension for 90 min at 37°C. Suspended cells were distributed onto fibronectin-coated culture dishes and incubated at 37°C for 20 min. Fibronectin-coated dishes were prepared by incubation of culture dishes with 2 µg of fibronectin per ml in serum-free DMEM for 2 h at 37°C, blocking with 0.5% BSA in PBS overnight at 37°C, and washing with PBS.

Immunoprecipitation and immunoblotting. Cells were washed with PBS and lysed in lysis buffer, and insoluble material was removed by centrifugation. For evaluation of phosphotyrosine levels, cells were lysed in modified radioimmunoprecipitation assay (RIPA) buffer (50 mM HEPES, pH 7.4; 150 mM NaCl; 1.5 mM MgCl₂; 1 mM EGTA; 1% Triton X-100; 1% sodium deoxycholate; 0.1% sodium dodecyl sulfate [SDS]; 10% glycerol; 1 mM sodium orthovanadate; 10 mM sodium pyrophosphate; 1 mM NaF; protease inhibitor mixture [Wako]). For coimmunoprecipitation, cells were lysed in HTN lysis buffer (20 mM HEPES, pH 7.4; 150 mM NaCl; 1% Triton X-100; 1 mM sodium orthovanadate; 1 mM NaF; protease inhibitor mixture [Wako]). Lysates were precleaned with normal IgG (Dako, Copenhagen, Denmark) immobilized on protein G-Sepharose (Amersham Pharmacia Biotech). Precleared lysates were incubated with antibodies immobilized on protein G-Sepharose at 4°C for 60 min. The beads were then washed four times in their respective washing buffers. For evaluation of phosphotyrosine levels, 1% Triton washing buffer (50 mM HEPES, pH 7.4; 150 mM NaCl; 1.5 mM MgCl₂; 1 mM EGTA; 1% Triton X-100; 10% glycerol; 1 mM sodium orthovanadate; 10 mM sodium pyrophosphate; 1 mM NaF; protease inhibitor mixture [Wako]) was used. For coimmunoprecipitation, HTN washing buffer (20 mM HEPES, pH 7.4; 150 mM NaCl; 0.1% Triton X-100; 1 mM sodium orthovanadate; 1 mM NaF; protease inhibitor mixture [Wako]) was used. The immunecomplexes were collected by centrifugation and resuspended in SDS-polyacrylamide gel electrophoresis (PAGE) loading dye.

In vitro kinase assay. Cells were lysed in modified RIPA buffer. Immunoprecipitations by anti-c-Src antibody were carried out as described above. After the beads were washed in modified RIPA buffer, they were washed in kinase buffer (100 mM Tris-HCl, pH 7.2; 125 mM MgCl₂; 25 mM MnCl₂; 2 mM EGTA; 250 µM sodium orthovanadate; 2 mM dithiothreitol). Beads were split equally into two two tubes: one portion was analyzed by immunoblotting with anti-c-Src antibody (Upstate Biotechnology), and the other was used in the kinase assay. Acid-denatured rabbit muscle enolase was added to 30 µl of kinase reaction buffer as a substrate. To initiate kinase reactions, 10 µl of ATP mixture (20 mM morpholinepropanesulfonic acid, pH 7.2; 75 mM MgCl₂; 500 µM cold ATP; 25 mM -glycerol phosphate; 5 mM EGTA; 1 mM sodium orthovanadate; 1 mM dithiothreitol; 10 µCi of [-³²P]ATP) was added to immunoprecipitates, and reaction mixtures were incubated for 5 min at 25°C. The reactions were stopped by adding 10 µl of 5× SDS-PAGE sample buffer (313 mM Tris-HCl, pH 6.8; 50% glycerol; 10% SDS; 0.065% bromophenol blue; 7.7% dithiothreitol) and boiled for 3 min. Reaction products were analyzed by SDS-PAGE. Gels were prepared for standard autoradiography and quantitated using BAS2000 (Fuji Film). Specific kinase activity was calculated by dividing the radioactivity of the enolase band by the area of the c-Src band determined by immunoblotting.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Hic-5 Inhibits cell spreading. To investigate the effects of the counterbalance between paxillin and Hic-5 expression on integrin functions, we compared the cell spreading on fibronectin using NIH 3T3 cells, which were either nontransfected or transfected with HA-tagged Hic-5 (LD1mHic-5, described below) or paxillin expression vectors. Cells were collected by trypsinization, replated onto coverslips coated with fibronectin, and then fixed and stained with anti-HA antibody. Nontransfected NIH 3T3 cells and paxillin-overexpressing cells began to spread within 30 min (Fig. 1A and B). In contrast, Hic-5-overexpressing cells showed significantly less spreading at this time point (Fig. 1A and B).

View larger version (40K): [in this window] [in a new window]   FIG. 1.   Effects of Hic-5 overexpression on cell spreading. (A) NIH 3T3 cells transfected with pCG-pax (HA-tagged paxillin, panels 1 and 2) or pCG-LD1mhic-5 (HA-tagged LD1mHic-5, panels 3 and 4) were allowed to spread on fibronectin-coated coverslips for 30 min, fixed with 3.7% formalin, and then immunostained with anti-HA antibody (panels 1 and 3) or stained with TRITC-conjugated phalloidin (panels 2 and 4). Arrows indicate the transfected cells. (B) NIH 3T3 cells were transfected with vector (), expression vector of paxillin (), or Hic-5 (). After 48 h, cell spreading was quantified by allowing the cells to spread on fibronectin-coated coverslips for the indicated times. Cells were stained with anti-HA antibody, and the percentages of spread cells among transfected cells stained with anti-HA antibody at each time point were calculated. Values represent the means of at least three independent experiments ± the standard deviation (SD). (C) NIH 3T3 cells transfected with pEGFP-N3 (GFP), pCG-pax (paxillin), pCG-hhic-5 (hHic-5), pCG-LD1mhic-5 (LD1mHic-5), and pCG-mhic-5 (mHic-5) were allowed to spread on fibronectin-coated coverslips for 30 min. The percentages of the spread cells were determined. Each bar represents the mean of at least three independent experiments ± the SD. (D) Equal amounts of total cell lysates from cells used in the spreading assay were subjected to SDS-PAGE and immunoblotted with anti-HA antibody. Lane 1, primary MEFs; lanes 2 to 6, NIH 3T3 cells transfected with vector (lane 2) or the expression vectors of paxillin (lane 3), hHic-5 (lane 4), LD1mHic-5 (lane 5), or mHic-5 (lane 6).

Paxillin rescues Hic-5-mediated inhibition of cell spreading. To determine whether the inhibition of spreading was caused by the altered counterbalance between Hic-5 and paxillin expression, we cotransfected the paxillin and the Hic-5 expression vectors at various ratios. The expression of these proteins was determined by Western blotting using anti-HA and anti-paxillin antibodies (Fig. 2B). Hic-5-mediated inhibition of spreading was restored by cotransfection of the paxillin expression vector at a 1:1 ratio to the level of 85.2% of control cells (Fig. 2A), indicating that the counterbalance between Hic-5 and paxillin is important for the modulation of cell spreading and that the inhibition observed in this study was not caused by the toxic effects of the overexpression of exogenous proteins. In addition, the efficiency of this reversion was dose dependent. This implied that Hic-5 has a competitive effect on the action of paxillin. Indeed, Hic-5 and paxillin share common interacting factors, including FAK, Pyk2, vinculin, talin, and PTP-PEST. In this regard, FAK and PTP-PEST have been reported to be involved in cell spreading (1, 57). Therefore, these are possible molecular targets for this competition.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Paxillin rescued Hic-5-mediated inhibition of cell spreading. (A) NIH 3T3 cells were transfected with pCG-pax (HA-paxillin) and/or pCG-LD1mhic-5 (HA-LD1mHic-5) at the ratios indicated. The transfected cells were allowed to spread on fibronectin-coated coverslips for 30 min, and the numbers of spread cells were determined in each case as described above. Each bar represents the mean of at least three independent experiments ± the SD. (B) Equal amounts of total cell lysates (20 µg of cellular protein/lane) from cells transfected with an empty vector (lane 1), pCG-pax (HA-paxillin) and/or pCG-LD1mhic-5 (HA-LD1mHic-5) at the ratios 0:1 (lane 2), 0.5:1 (lane 3), 1:1 (lane 4), or 1:0 (lane 5), respectively, were subjected to immunoblotting with anti-HA antibody (upper panel) or antipaxillin antibody (lower panel). The gels were stained with Coomassie blue to confirm the same amounts of protein loading.

LD3 domain of Hic-5 is required for inhibition of spreading. A chimeric mutant containing the N-terminal region of Hic-5 and the C-terminal region of paxillin had almost the same activity as did the wild-type Hic-5 (data not shown). Therefore, we focused on the N-terminal region of Hic-5. To determine which region of Hic-5 is required for inhibition of spreading, we used various Hic-5 deletion mutants in the spreading assay (Fig. 3E). Mutant constructs used in this study (Fig. 3E) were transiently transfected, and their expression in NIH 3T3 cells was assessed (Fig. 3C and D). Each construct expressed similar amounts of the protein.

View larger version (48K): [in this window] [in a new window]   FIG. 3.   The LD3 domain of Hic-5 was required for the inhibition of spreading. (A and C) NIH 3T3 cells were transfected with pEGFP-N3 (GFP), pCG-pax (paxillin), pCG-hhic-5 (WT Hic-5), pCG-LD3-4hhic-5 (LD3-4, HA-tagged human Hic-5 with deletion of LD3-4 domains), or pCG-LD3hhic-5 (LD3, HA-tagged human Hic-5 with deletion of LD3 domain). (B and D) Cells were transfected with an empty vector or pcDNA3, 1A-hhic-5 (Myc-tagged WT hHic-5), pcDNA3.1A-LD1-2hhic-5 (LD1-2, Myc-tagged human Hic-5 with deletion of LD1-2 domains), and pcDNA3. 1A-hic/LIM (LD1-4, Myc-tagged human Hic-5 with deletion of LD1-4 domains) in lanes 1 to 4, respectively. Cells were allowed to spread on fibronectin-coated coverslips for 30 min and stained with anti-HA (A) or anti-Myc (B) antibodies, and the numbers of spread cells were determined in each case as described above. Each bar represents the mean of at least three independent experiments ± the SD. Lysates obtained from cells transfected with the different constructs were analyzed for the expression of tagged proteins with anti-HA antibody (C) or anti-Myc antibody (D). (E) Schematic representation of recombinant proteins used in this assay. PY indicates tyrosine phosphorylation sites.

FAK seems to be involved in Hic-5-mediated inhibition of spreading. Paxillin becomes phosphorylated on tyrosine residues in response to integrin-mediated cellular events that are associated with cytoskeletal remodeling (9, 83). Recent studies identified paxillin as a potential substrate for FAK and its associated kinase Src (3, 66) and showed that paxillin needed to bind to FAK for maximal phosphorylation in response to adhesion (79). Hic-5 overexpression may prevent cell spreading by sequestration of FAK, perhaps by competition between Hic-5 and paxillin. To assess this possibility, we cotransfected the LD1mHic-5 expression vector with wild-type FAK (WT FAK) or catalytically inactive FAK (kinase-defective [kd] FAK) expression vectors. Hic-5-mediated inhibition of spreading was efficiently rescued by coexpression of WT FAK (Fig. 4). Cotransfection of catalytically inactive kd FAK also restored the inhibitory effect of Hic-5 (Fig. 4). These observations indicated that FAK is a potential target for competition between Hic-5 and paxillin and that the kinase activity itself may not be necessary to rescue cell spreading. This observation is consistent with those of a previous study that showed that overexpression of catalytically inactive FAK variants restored pp41 and pp43 FRNK-mediated inhibition of cell spreading at levels comparable to those of WT FAK (57); these authors also proposed that FAK acts as an adapter molecule that recruits Src family kinase to phosphorylate paxillin and to promote cell spreading. To determine whether this model is applicable to Hic-5-mediated inhibition of cell spreading, we coexpressed Hic-5 and v-Src in NIH 3T3 cells. As shown in Fig. 4, coexpression of Hic-5 and v-Src reversed the inhibitory effects of Hic-5 on cell spreading.

View larger version (59K): [in this window] [in a new window]   FIG. 4.   FAK was a limiting factor in Hic-5-mediated inhibition of cell spreading. Upper panel. NIH 3T3 cells were transfected with pEGFP-N3 (GFP, lane 1), pCG-pax (HA-paxillin, lane 2), pCG-LD1mhic-5 (HA-LD1mHic-5, lane 3), LD1mHic-5 plus pSR-FAK (WT FAK, lane 4), LD1mHic-5 plus pSR minus FAK K454R (kinase-defective [kd] FAK, lane 5), or LD1mHic-5 plus pSRA2 (v-Src, lane 6) and allowed to spread on fibronectin-coated coverslips for 30 min, and the numbers of spread cells were determined in each case as described in the legends to Fig. 1. Lanes 7 to 9 represent cells transfected with GFP plus FAK, GFP plus kd FAK, or GFP plus v-Src, respectively. The numbers of spread cells were counted using cells labeled with anti-HA antibody or GFP. Each bar represents the mean of at least three independent experiments ± the SD. In the lower panel, cell extracts were prepared, subjected to SDS-PAGE, and immunoblotted with antiserum indicated.

View larger version (29K): [in this window] [in a new window]   FIG. 5.   Hic-5 competed with paxillin for FAK. NIH 3T3 cells were transfected with pCG-N-BL (an empty vector; lanes 3, 7, and 12), pCG-hhic-5 (HA-tagged WT Hic-5; lanes 4, 8, and 13), pCG-LD3-4hhic-5 (HA-tagged LD3-4; lanes 5, 9, and 14), or pCG-LD3hhic-5 (HA-tagged LD3; lanes 6, 10, and 15) (A) or with an empty vector (lanes 3, 7, and 12), pcDNA3. 1A-hhic-5 (Myc-tagged WT hHic-5; lanes 4, 8, and 13), pcDNA3.1A-LD1-2hhic-5 (Myc-tagged LD1-2; lanes 5, 9, and 14), or pcDNA3.1A-hic/LIM (Myc-tagged LD1-4, lanes 6, 10, and 15). Total cell lysates were prepared, immunoprecipitated with anti-HA (A) or anti-Myc (B) antibody, subjected to SDS-PAGE, and immunoblotted with anti-FAK, anti-HA, or anti-Myc antibody. Lane 1 represents whole-cell lysate, and lanes 2 and 11 indicate the results of immunoprecipitation with control IgG. Control extracts were prepared from cells transfected with tagged WT hHic-5. (C) MEFs were infected with viruses producing EGFP (lanes 1 and 3) or Myc-tagged mHic-5 (lanes 2 and 4). Total cell lysates were prepared, immunoprecipitated with anti-FAK antibody, subjected to SDS-PAGE, and immunoblotted with antipaxillin antibody and anti-FAK antibody. Lane 5 represents a result of immunoprecipitation with control IgG. Lysate lanes contained one-fifth the amount of cellular proteins used for immunoprecipitation. The amounts of protein in each lane were corrected by cell numbers.

/

/

/

/

View larger version (21K): [in this window] [in a new window]   FIG. 6.   FAK seemed necessary for Hic-5-mediated inhibition of cell spreading. (A) MEFs derived from FAK/ mice were transfected with pCG-pax (HA-paxillin) or pCG-LD1mhic-5 (HA-LD1mHic-5). Transfected cells were allowed to spread on fibronectin-coated coverslips for the indicated times and stained with anti-HA antibody, and the percentages of spread cells among those stained at each time point were calculated. Values represent the means of at least three independent experiments ± the SD. (B) FAK/ MEFs were transfected with an empty vector (lanes 1 and 4), pCG-pax (HA-paxillin, lanes 2 and 5), or pCG-LD1mhic-5 (HA-LD1mHic-5, lanes 3 and 6) in the absence (lanes 1 to 3) or presence (Myc-WT FAK, lanes 4 to 6) of pSR-FAK and allowed to spread on fibronectin-coated coverslips for 20 min. Upper panel, the numbers of spread cells were determined in each case as described above. Each bar represents the mean of at least three independent experiments ± the SD. For the lower panel, cell extracts were Western blotted with anti-FAK or antipaxillin antiserum.

Hic-5 inhibits integrin-mediated paxillin phosphorylation. Paxillin is highly phosphorylated on tyrosine residues during various cellular events associated with cell adhesion, cytoskeletal reorganization, and growth signaling (43) and is recognized as a substrate of FAK and/or of kinases activated by FAK (66). In addition, physical interaction between paxillin and FAK is required for maximal phosphorylation of paxillin in response to cell adhesion (79). Since Hic-5 inhibited cell spreading after fibronectin stimulation, we tested whether Hic-5 influenced the integrin-mediated tyrosine phosphorylation of paxillin in the presence or absence of Hic-5. Hic-5 or EGFP as a control was expressed by the retrovirus-mediated method in MEFs, and cells were replated onto fibronectin and lysed. The lysates were immunoprecipitated with corresponding antibodies and analyzed by immunoblotting with antiphosphotyrosine antibody. Paxillin and FAK were tyrosine phosphorylated in an adhesion-dependent manner in cells infected with both of these retroviruses (Fig. 7). However, the tyrosine phosphorylation level of paxillin was reduced in Hic-5-overexpressing cells relative to controls, a result consistent with the findings reported previously (15). In addition, we reproducibly observed a modest decrease in the tyrosine phosphorylation level of FAK in our system (Fig. 7, lower panels). We also examined c-Src kinase activity in cells infected with retroviral constructs. Immortalized MEFs were infected with retroviral constructs encoding EGFP or mouse Hic-5, and cells were kept in suspension or stimulated with fibronectin to be subjected to kinase reaction using enolase as a substrate. The results shown in Fig. 8 indicate that Hic-5 reduced c-Src kinase activity significantly.

View larger version (55K): [in this window] [in a new window]   FIG. 7.   Hic-5 reduced tyrosine phosphorylation of paxillin. Immortalized MEFs were infected with retroviral constructs encoding EGFP (lanes 1, 2, 6, and 7) or mouse Hic-5 (lanes 4, 8, and 9). Cells were kept in suspension (S; lanes 1, 3, 6, and 8) or were stimulated on fibronectin (FN; lanes 2, 4, 7, and 9) for 60 min at 37°C. Cell lysates containing the same amount of protein were run directly (lanes 1 to 4) or were immunoprecipitated (lanes 6 to 9) with antipaxillin (upper panels) or anti-FAK (lower panels) antibody and then immunoblotted with antiphosphotyrosine antibody or with the corresponding antibodies. Lane 5 represents immunoprecipitate with control IgG.

View larger version (13K): [in this window] [in a new window]   FIG. 8.   Hic-5 reduced c-Src kinase activity. (A) Immortalized MEFs were infected with retroviral constructs encoding EGFP (lanes 1 to 3) or mouse Hic-5 (lanes 4 to 6). Cells were kept in suspension (Sus, lanes 1 and 4) or stimulated on fibronectin (FN; lanes 2, 3, 5, and 6) for 15 min (lanes 2 and 5) or 60 min (lanes 3 and 6) at 37°C. Cell lysates containing same amount of protein were immunoprecipitated with anti-c-Src and subjected to kinase reaction using enolase as a substrate. The reaction products were analyzed by SDS-PAGE and autoradiography (upper panel). Immunoprecipitated c-Src was immunoblotted with anti-c-Src antibody (lower panel). (B) Src kinase activity was calculated by dividing the radioactivities of the enolase bands by amounts of immunoprecipitated c-Src. Each bar represents the mean of three independent experiments ± the SD (open columns, EGFP-expressing cells; closed columns, Hic-5-expressing cells).

View larger version (26K): [in this window] [in a new window]   FIG. 9.   Hic-5 inhibited paxillin-Crk complex formation. (A) Immortalized MEFs were infected with retroviral construct encoding EGFP (lanes 1 to 4) or mouse Hic-5 (lanes 5 to 8). Cells were kept in suspension (Sus; lanes 1, 2, 5, and 6) or stimulated on fibronectin (FN; lanes 3, 4, 7, and 8) for 60 min at 37°C. Cell lysates containing the same amount of protein were incubated for 1 h at 4°C with GST-Crk SH2 fusion protein (lanes 2, 4, 6, and 8) or GST alone (lanes 1, 3, 5, and 7) Immobilized on glutathione-Sepharose 4B beads. Bound proteins were separated by SDS-PAGE and immunoblotted with antipaxillin antibody. (B) Cells were infected with retroviral constructs encoding EGFP (lanes 1 and 2) or Hic-5 (lanes 3 and 4). Whole-cell lysates from MEFs kept in suspension (Sus, lanes 1 and 3) or plated on fibronectin (FN; lanes 2 and 4), and Western blotted with antipaxillin antiserum. (C) GST fusion proteins used in this experiment were visualized by Coomassie blue staining.

Rho family small GTPases are involved in Hic-5-mediated inhibition of spreading. Crk is an adapter molecule consisting almost entirely of SH2 and SH3 domains (43). Paxillin and p130Cas, both of which are components of focal adhesion complexes (33), are two major Crk SH2-binding proteins. Numerous cellular proteins have recently been identified on the basis of interaction through the SH3 domain of Crk. Among these, DOCK180 is a downstream molecule of integrin-mediated signal transduction (22, 34). DOCK180 can directly bind to and activate Rac1, a small G protein involved in the regulation of the actin cytoskeleton (35). As discussed by Petit et al. (53), paxillin-Crk complex formation could link integrin stimulation to downstream molecules, and thus this complex may promote actin reorganization, cell spreading, and migration. To further investigate the involvement of Hic-5 in downstream signaling, we examined whether constitutively active forms of Rho family GTPases, i.e., RhoA V14, Rac1 V12, and Cdc42 V12, could bypass the spreading-deficient phenotype observed in cells overexpressing Hic-5. NIH 3T3 cells were cotransfected with the LD1mHic-5 expression vector with RhoA V14, Rac1 V12, or Cdc42 V12 expression vectors and subjected to spreading assay. Coexpression of Rac1 V12 or Cdc42 V12 resulted in the restoration of cell spreading (Fig. 10). This finding was consistent with the results of a previous study in which Rac1 and Cdc42 were reported to contribute to early-phase cell spreading (55). In contrast, transfection with RhoA V14 did not rescue LD1mHic-5-mediated inhibition of cell spreading. These findings agreed with the observation that DOCK180 elevates the GTP/GDP ratio of Rac1 but not of RhoA (35). The results obtained here suggested that Hic-5 influences the signaling pathways in which Rac1 and Cdc42 are involved during the early phase of cell spreading.

View larger version (39K): [in this window] [in a new window]   FIG. 10.   Constitutively active Cdc42 and Rac-1 rescued Hic-5-mediated inhibition of spreading. NIH 3T3 cells were transfected with pGFP (lane 1), pCG-pax (HA-paxillin, lane 2), or pCG-LD1mhic-5 (HA-LD1mHic-5, lanes 3 to 6), in the presence of pEF BOS (empty vector, lanes 3), pEF BOS RhoA DA (HA-RhoA V14, lane 4), pEF BOS Rac1 DA (HA-Rac1 V12, lane 5), or pEF BOS Cdc42 DA (HA-Cdc42 V12, lane 6). In the upper panel, cells were allowed to spread on fibronectin-coated coverslips for 30 min, and the numbers of spread cells were determined in each case as described above. Each bar represents the mean of at least three independent experiments ± the SD. In the lower panel, cell extracts were Western blotted with anti-HA antibody.

Effect of antisense expression of Hic-5 on spreading. The expression level of Hic-5 is markedly decreased during immortalization of MEFs (28, 70). To investigate the participation of endogenous Hic-5 in the early phase of cell spreading, we expressed antisense Hic-5 in mortal MEFs that contained higher levels of Hic-5. As shown in Fig. 11, mortal MEFs (nine population doublings) with increased Hic-5 expression showed a delayed spreading rate relative to that of immortalized MEFs of at least within 60 min, and the expression of antisense Hic-5 in mortal MEFs restored spreading activity, at least in part (Fig. 11A). Furthermore, the forced expression of Hic-5 caused inhibition of early-phase cell spreading as described above (Fig. 6). These observations indicated that the constitutive levels of Hic-5 expression are correlated with the cell spreading rate in the early phase. These observations suggest that Hic-5 inhibits the cell spreading rate only at the early phase and that the early and late phases of cell spreading are different processes that may involve different molecules.

View larger version (14K): [in this window] [in a new window]   FIG. 11.   Effect of antisense expression of Hic-5 on spreading abilities. (A) Mortal MEFs were transfected with an empty vector () or an antisense expression vector of Hic-5 (), together with GFP. At 48 h after transfection, cells were allowed to spread on fibronectin-coated coverslips for the indicated times and stained with rhodamine-conjugated phalloidin, and the percentages of spread cells at each time point were calculated. Values represent the means of at least three independent experiments ± the SD. The spreading ability was measured as described above. The spreading activity of immortal MEFs () was also examined. (B) Cell lysates from mortal and immortal MEFs were subjected to SDS-PAGE and Western blotting using anti-paxillin antibody.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Focal adhesions are formed following stimulation of integrins and are composed of numerous scaffold, docking proteins, as well as signal transduction molecules. The structurally related molecules paxillin and Hic-5 are colocalized in focal adhesions, thereby docking several signal transducers. We previously reported the involvement of Hic-5, a paxillin homologue, in cellular senescence (70) and the specific decrease in Hic-5 expression levels during the immortalization of MEFs (28). Despite its unique features, the molecular basis of Hic-5 function has not been fully clarified. In this study, we demonstrated that the counterbalance between Hic-5 and paxillin expression is a determinant of integrin-mediated cellular events, including cell spreading, phosphorylation of focal adhesion proteins, and downstream signaling.

Competitive effects between Hic-5 and paxillin on cell spreading. Hic-5 and paxillin share a closely related structure and common interaction factors. Paxillin has phosphotransfer sites that serve as putative target sites for the SH2 domain of Crk (3, 66), implying the involvement in downstream signals such as proliferation and cytoskeletal organization. However, unlike paxillin, Hic-5 is not phosphorylated on tyrosine by integrin stimulation (15). Furthermore, the Hic-5 expression level is specifically decreased during immortalization, whereas that of paxillin tends to be increased (28). Taken together, we hypothesized that Hic-5 may not transduce signals downstream and may antagonize paxillin and that the balance between Hic-5 and paxillin expression may influence integrin-mediated signal transduction. To address these possibilities, we overexpressed exogenous Hic-5 in immortalized mouse fibroblasts in which Hic-5 was expressed at a relatively low level and assessed the effects on cell spreading. In the present study, we show that Hic-5 overexpression inhibited cell spreading on fibronectin. This inhibitory effect was rescued by coexpression of paxillin in a dose-dependent manner (Fig. 2), suggesting that Hic-5 and paxillin compete for a common interacting factor(s), the association of which with paxillin is necessary for cell spreading.

Roles of FAK in cell spreading. Richardson and Parsons reported that the C-terminal domain of FAK (pp41 and pp43 FRNK) acts as an inhibitor of FAK by delaying cell spreading on fibronectin; reducing tyrosine phosphorylation of FAK, paxillin, and tensin, and transiently blocking the formation of focal adhesions (56). These FRNK-mediated inhibitory effects were rescued by coexpression of FAK but not by coexpression of a FAK variant lacking the paxillin-binding site (cFAK) (57), suggesting that the binding of FAK to paxillin and tyrosine phosphorylation of paxillin play important roles in integrin-mediated cytoskeletal organization leading to cell spreading. Taken together, these observations indicated that Hic-5-mediated inhibition of cell spreading is probably mediated by similar mechanisms to that in the case of pp41 and pp43 FRNK.

/

/

/

/

Significance of tyrosine phosphorylation of paxillin in signals for spreading. One of the processes that follows paxillin phosphorylation is signal transduction through the docking of SH2 proteins (38). Tyrosine phosphorylation of two major sites, Y31 and Y118, on paxillin creates binding sites for the SH2 domain of the adapter protein Crk (3, 66). Although paxillin and Hic-5 show marked homology over their entire structures, Hic-5 does not have the tyrosine residue that serves as a site for Crk binding. In fact, no interaction between Hic-5 and Crk SH2 was detected even under forced phosphorylation by treatment with pervanadate (29). In the present study, we showed that formation of the paxillin-Crk complex is prevented in Hic-5-overexpressing cells (Fig. 9). This prevention probably resulted from diminished tyrosine phosphorylation of paxillin by sequestration of FAK. A recent study indicated that phosphorylation of Y31 and Y118 on paxillin plays a critical role in the collagen-induced migration of NBT-II cells. Mutations in Y31 and Y118 impaired motility, diminished tyrosine phosphorylation and prevented formation of the paxillin-Crk complex (53).