Characterization of PINCH-2, a New Focal Adhesion Protein That Regulates the PINCH-1-ILK Interaction, Cell Spreading, and Migration*

Integrin-linked kinase (ILK) is a multidomain protein that plays important roles at cell-extracellular matrix (ECM) adhesion sites. We describe here a new LIM-do-main containing protein (termed as PINCH-2) that forms a complex with ILK. PINCH-2 is co-expressed with PINCH-1 (previously known as PINCH), another member of the PINCH protein family, in a variety of human cells. Immunofluorescent staining of cells with PINCH-2-specific antibodies show that PINCH-2 localizes to both cell-ECM contact sites and the nucleus. Deletion of the first LIM (LIM1) domain of PINCH-2 abolished the ability of PINCH-2 to form a complex with ILK. The ILK-binding defective LIM1-deletion mutant, unlike the wild type PINCH-2 or the ILK-binding competent LIM5-deletion mutant, was incapable of localizing to cell-ECM contact sites, suggesting that ILK binding is required for this process. Importantly, the PINCH-2-ILK and PINCH-1-ILK interactions are mutually exclusive. Overexpression of PINCH-2 significantly inhibited the PINCH-1-ILK interaction and reduced cell spreading and migration. These results identify a novel nuclear and focal adhesion protein that associates with ILK and reveals an important role of PINCH-2 in the regulation of the PINCH-1-ILK interaction, cell shape change, and migration. Antibodies, and Other Reagents— Human embryonal kidney 293 cells, human rhabdomyosarcoma (RD) cells, and mouse C2C12 cells

motif, and a C-terminal kinase domain that exhibits significant homology to other protein kinase catalytic domains. One of the major functions of ILK is to mediate multiple protein-protein interactions at cell-ECM adhesion sites. In previous studies, we have found that ILK binds to PINCH-1 (12), a widely expressed focal adhesion protein consisting of five LIM domains. The ILK-PINCH-1 interaction is mediated by the N-terminal ANK domain of ILK and the second zinc finger of the N-terminalmost LIM domain (LIM1) of PINCH-1 (13). In addition to interacting with PINCH-1, ILK is capable of interacting with several other focal adhesion proteins including ␤ 1 integrins (14), CH-ILKBP (15) (also known as actopaxin (16) or ␣-parvin (17)), affixin (18) (also known as ␤-parvin (17)), and paxillin (19) through its C-terminal domain. Recently, we have shown that ILK binds to both PINCH-1 and CH-ILKBP simultaneously, resulting in the formation of a PINCH-1⅐ILK⅐CH-ILKBP complex in mammalian cells (15).
Recent biochemical, cell biological, and genetic studies have provided strong evidence for an important role of the PINCH-1⅐ILK⅐CH-ILKBP complex in integrin-mediated cell-ECM interactions (9 -11, 20). For example, inhibition of the PINCH-1⅐ILK⅐CH-ILKBP complex formation in mammalian cells by overexpression of dominant negative PINCH-1 or ILK mutants significantly impaired cell shape change and migration (21). In genetic model systems such as Caenorhabditis elegans, mutations in pinch-1/unc-97 (22), ilk/pat-4 (23), or ch-ilkbp/pat-6 (46) all resulted in a Pat phenotype similar to that of ␤-integrin/pat-3 or ␣-integrin/pat-2, which is characterized by defects in dense body and M-line (similar to cell-ECM adhesion sites in mammalian cells) assembly (24,25).
While it is increasingly clear that the PINCH-1⅐ILK⅐CH-ILKBP complex plays an important role in the coupling of the ECM to the actin cytoskeleton, how cells regulate the assembly and functions of the PINCH-1⅐ILK⅐CH-ILKBP complex was not known. Search of the genomic sequence of C. elegans has revealed that there exists a gene (termed pin-2) encoding a protein that is structurally related to PINCH-1/UNC-97 (22). A partial sequence of a putative PINCH-1-related human protein (termed as PINCH-2) has been predicted based on cDNA sequence analyses of human EST clones (22). It was not known, however, whether PINCH-2 is co-expressed with PINCH-1 in human cells. Furthermore, neither the molecular activities nor the cellular functions of PINCH-2 were known. In this study, we have cloned and characterized PINCH-2. Our results show that PINCH-2 localizes to both cell-ECM adhesion sites and the nucleus and provide strong evidence for an important role of PINCH-2 in the regulation of the PINCH-1⅐ILK⅐CH-ILKBP complex formation, cell shape change, and migration.
PINCH-2 Mammalian Expression Vector Construction and DNA Transfection-DNA fragments encoding the full-length or mutant forms of PINCH-2 (as specified in each experiment) were generated by PCR and cloned into the EcoRI/SalI sites of the pFLAG-CMV-2 vector (Sigma). DNA fragments encoding the FLAG-tagged full-length or mutant forms of PINCH-2 were cloned into the pEGFP-C2 expression vector (Clontech). To express the full-length or mutant forms of PINCH-2 in mammalian cells, cells were transfected with the FLAG or GFP expression vectors encoding the full-length or mutant forms of PINCH-2 using LipofectAMINE PLUS (Invitrogen) as described (15,21). The expression of the FLAG-and/or GFP-tagged full-length or mutant forms of PINCH-2 in the transfectants was confirmed by Western blotting with anti-FLAG and anti-GFP antibodies.
Adenoviral Expression Vector Construction and Infection-The adenoviral expression vector encoding FLAG-tagged full-length human PINCH-2 was generated following a previously described protocol (21). Briefly, human PINCH-2 cDNA was cloned into the SalI/XbaI sites of the pAdTrack-CMV shuttle vector. The shuttle vector plasmid was linearized with PmeI, purified, and mixed with supercoiled pADEsay-1. The vectors were transferred into Escherichia coli BJ5183 by electroporation. The bacteria were placed in 1 ml of LB broth, Lennox (Fisher) at 37°C for 1 h. The bacteria were then inoculated onto agar containing LB broth supplemented with 50 g/ml kanamycin. After 16 -20 h growth, colonies were picked and grown in 2 ml of LB broth containing 50 g/ml kanamycin. Clones were screened by digestions with restriction endonucleases PacI and BamHI. The positive plasmids were transformed into DH10B cells by electroporation for large-scale amplification. The plasmid DNA was digested with PacI, ethanol-precipitated, and was used to transfect 293 cells with LipofectAMINE PLUS. The transfected cells were harvested 10 days after transfection. The cells were lysed by three cycles of freezing in a methanol/dry ice bath and rapid thawing at 37°C, and the lysates containing the recombinant adenovirus were collected. The control adenoviral expression vector encoding ␤-galactosidase was kindly provided by Drs. Tong-Chuan He and Bert Vogelstein (Howard Hughes Medical Institute, The Johns Hopkins Oncology Center, Baltimore, MD). The recombinant adenoviral expressing vectors were used to infect REF-52 cells. The infection efficiency was monitored by the expression of GFP encoded by the viral vectors and typically reached 80 -90% within 48 h. The expression of the FLAG-PINCH-2 in the infected cells was confirmed by Western blotting with monoclonal anti-FLAG antibody M5.
Generation and Purification of Anti-PINCH-2 Antibodies-Polyclonal anti-PINCH-2 antisera were generated by immunizing rabbits with a keyhole limpet hemocyanin-conjugated peptide containing PINCH-2 sequence AQPKATDLNSA. To purify anti-PINCH-2 antibodies, rabbit anti-PINCH-2 antisera were passed through an affinity column containing agarose beads coupled with the PINCH-2 peptide (1 mg of peptide/ml of beads). After collecting the flow-through fractions, the column was washed with phosphate-buffered saline. Anti-PINCH-2 antibodies were eluted from the affinity column with 100 mM glycine buffer (pH 2.9) and dialyzed against phosphate-buffered saline.
Immunofluorescent Staining-Immunofluorescent staining was performed as described (15,21). Briefly, cells (as specified in each experiment) were plated in complete medium on fibronectin-coated coverslips or Lab-Tek 8-chamber culture slides and incubated at 37°C under a 5% CO 2 , 95% air atmosphere for at least 6 h. The cells were fixed with 3.7% paraformaldehyde in phosphate-buffered saline, and stained with affinity purified rabbit polyclonal antibodies and/or mouse monoclonal antibodies as specified in each experiment. The primary antibodies were detected with secondary Rhodamine Red TM -conjugated anti-rabbit IgG antibodies and fluorescein isothiocyanate-conjugated anti-mouse IgG antibodies, respectively. In some experiments, cells were dually stained with affinity purified rabbit anti-PINCH-2 antibodies and tetramethylrhodamine-labeled phalloidin (to visualize the actin cytoskeleton). The rabbit anti-PINCH-2 antibodies were detected with fluorescein isothiocyanate-conjugated anti-rabbit IgG antibodies in these experiments.
Immunoprecipitation-Cells (as specified in each experiment) were cultured in complete medium in 60-or 100-mm culture plates. Cell monolayers were rinsed twice with phosphate-buffered saline and directly lysed on the plates with 1% Triton X-100 in 50 mM Tris-HCl (pH 7.4) containing 150 mM NaCl, 10 mM Na 4 P 2 O 7 , 2 mM Na 3 VO 4 , 100 mM NaF, and protease inhibitors. The protocol for immunoprecipitation with monoclonal anti-CH-ILKBP antibody or affinity purified rabbit anti-PINCH-2 antibodies was described previously (15,21). Briefly, cell lysates (500 g) were mixed with 500 l of hybridoma culture supernatant containing monoclonal anti-CH-ILKBP antibody 1D4 or 20 g of affinity purified rabbit anti-PINCH-2 antibodies. Control immunoprecipitates were prepared by substitution of the 1D4 supernatant with an equal volume of unconditioned culture supernatant or by substitution of the affinity purified rabbit anti-PINCH-2 antibodies with an equal amount of irrelevant rabbit IgG. The samples were incubated for 3 h, mixed with 40 l of UltraLink immobilized protein G (Pierce), and then incubated for an additional 1.5 h. The beads were washed four times and the proteins bound were released from the beads by boiling in SDS-PAGE sample buffer for 5 min. The samples were analyzed by Western blotting with antibodies to CH-ILKBP, ILK, PINCH-1, or PINCH-2 as specified in each experiment.
For immunoprecipitation of FLAG-tagged full-length or mutant forms of PINCH-2, cells expressing the FLAG-tagged PINCH-2 proteins and control cells lacking the FLAG-tagged PINCH-2 proteins were lysed as described above. The cell lysates (500 g) were mixed with 40 l of anti-FLAG antibody M2-conjugated agarose beads (Sigma). The immunoprecipitated FLAG-tagged PINCH-2 proteins and the associated proteins were released from the beads by boiling in 60 l of SDS-PAGE sample buffer for 5 min and analyzed by Western blotting (20 l/lane) with antibodies as specified in each experiment.
Cell Spreading-Cells (as specified in each experiment) were plated in Opti-MEM I serum-free medium (Invitrogen) in wells of 96-well plates that were coated with 0.1 g/ml fibronectin. The plates were incubated at 37°C under a 5% CO 2 , 95% air atmosphere and the cell morphology was observed under an Olympus IX70 fluorescence microscope equipped with an Hoffman Modulation Contrast system. The images were recorded with a DVC-1310C Magnafire TM digital camera (Optronics). Unspread cells were defined as round cells, whereas spread cells were defined as cells with extended processes as described (15,21,27,28). The percentage of cells adopting spread morphology was quantified by analyzing at least 300 cells from three randomly selected fields (Ͼ100 cells/field) (15,21,27,28).
Cell Migration-Cell migration was assessed by the ability of the cells to migrate into a cell-free area as previously described (21,29,30). Briefly, cells were plated in complete medium on 24-well plates and grown for 24 h to reach confluence. The monolayers were then wounded by scratching with a plastic pipette tip. After washing, the cells were incubated in complete media for the indicated times and observed under an Olympus IX70 microscope equipped with an Hoffman Modulation Contrast system. Images of three different segments of the cell-free area were recorded with a DVC-1310C Magnafire TM digital camera (Optronics), and the distances traveled by the cells at the front in three different segments of the wound were measured.  1A, lane 3) was used as a template, confirming the specificity of the reactions. The cDNA obtained from the human RD cells was sequenced. It encodes a protein encompassing the previously described partial PINCH-2 sequence (22). The full-length human PINCH-2 consists of five LIM domains, each of which contains a sequence motif CXXCX 16 -18 C/HXXC/HXX-CXXCX 16 -18 CXXC/H/D (the full-length nucleotide sequence and the deduced protein sequence of PINCH-2 derived from the human RD cells have been deposited into GenBank TM (accession number AF484961)). BLAST search of the human genome data base indicates that the gene encoding PINCH-2 is located within the human chromosome 2q14.3 region. Human PINCH-2 is 82% identical to human PINCH-1 at the amino acid sequence level. Analyses of DNA sequences suggest that they are not splicing variants but are instead encoded by two different genes. To test whether PINCH-2 and PINCH-1 are co-expressed in human cells, we carried out reverse transcription-PCR on poly(A) ϩ RNA from human RD cells using primers specific for PINCH-1. A major cDNA with the predicted PINCH-1 size was amplified (Fig. 1B, lane 2). In control experiments, no DNA was amplified with the PINCH-1 primers when the template was omitted (Fig. 1B, lane 6), or when an irrelevant cDNA (Fig. 1B, lane 3) or PINCH-2 cDNA (Fig. 1B, lane 4), which was readily amplified with the PINCH-2 primers (Fig. 1A, lane 4), was used as a template. Taken together, these results suggest that PINCH-2 and PINCH-1 transcripts are co-expressed in human RD cells.

Cloning and Expression of PINCH-2-To
To facilitate studies on PINCH-2, we generated an anti-PINCH-2 antibody by immunizing rabbits with a peptide containing a specific PINCH-2 sequence. Western blotting analyses of lysates of Chinese hamster ovary cells transfected with expression vectors encoding GFP-and FLAG-tagged human PINCH-2 or PINCH-1 showed that the rabbit anti-PINCH-2 antiserum recognized the recombinant human PINCH-2 ( Fig.  2A, lane 2) but not the PINCH-1 protein (Fig. 2A, lane 1). In control experiments, no protein from the Chinese hamster ovary transfectants was recognized by preimmune rabbit serum under the same conditions (Fig. 2B). The expression of the Western blotting with the anti-PINCH-2 antiserum. The results show that the anti-PINCH-2 antiserum recognized a protein band with an apparent molecular weight similar to the calculated molecular weight of human PINCH-2 protein (38,917.03) (Fig. 2D, lane 1). Binding of the antibody to the ϳ39-kDa endogenous protein was completely blocked by incu- bation of the antiserum with the PINCH-2 peptide (Fig. 2D, lane 2), indicating that the ϳ39-kDa band represents the endogenous human PINCH-2 protein. Several additional protein bands with higher or lower molecular weights were also detected (Fig. 2D, lane 1). However, they unlikely represent proteins specifically recognized by the anti-PINCH-2 antibody, as they were not blocked by the PINCH-2 peptide (Fig. 2D, lane 2). To confirm that the anti-PINCH-2 antiserum recognizes PINCH-2 but not PINCH-1, we reprobed the membrane that had been probed with the anti-PINCH-2 antiserum (Fig. 2D,  lane 1) with an antibody recognizing PINCH-1. The result showed that the PINCH-1 protein, which migrated faster than PINCH-2 (the calculated molecular weight of PINCH-1 ϭ 37,251.32) and was not recognized by the anti-PINCH-2 antiserum (Fig. 2D, lane 1), was readily detected with the anti-PINCH-1 antibody in the same cell lysates (Fig. 2D, lane 3). These results confirm the specificity of the anti-PINCH-2 antiserum. Furthermore, they show that PINCH-2 and PINCH-1 are co-expressed in human WI-38 cells. Western blotting analyses of other human cell types including RD cells, A431 epidermoid carcinoma cells, IMR-90 lung cells, HT-1080 fibrosarcoma cells, embryonal kidney 293 cells, and primary glomerular mesangial cells revealed that PINCH-2 is also co-expressed with PINCH-1 in these cells (data not shown). To do this, we purified the anti-PINCH-2 antibodies with an affinity column containing the PINCH-2 peptide (Fig. 3A). The affinity purified antibodies specifically recognized PINCH-2 (Fig. 3A, lane 2), whereas the flow-through fractions failed to recognize PINCH-2 (Fig. 3A, lane 3). Immunofluorescent staining of human cells with the affinity purified anti-PINCH-2 antibody showed that PINCH-2 localizes to both focal adhesions and the nucleus (Fig. 3B). No focal adhesion or nuclear staining was observed with the flow-through fraction (Fig. 3C), confirming the specificity of the PINCH-2 immunofluorescent staining. Double staining of the cells with the affinity purified rabbit anti-PINCH-2 antibodies and mouse anti-ILK antibody 65.1 showed that PINCH-2 is co-clustered with ILK in focal adhesions (Fig. 3, D and E). Interestingly, although abundant PINCH-2 was detected in the nucleus (Fig. 3D), ILK was absent in the nucleus (Fig. 3E). As expected, the clusters of PINCH-2 at the focal adhesions (Fig. 3F), which are docking sites for actin stress fibers at the cell-extracellular matrix adhesion sites, are co-aligned with actin filaments (Fig. 3G).
In control experiments, no ILK was detected in the control immunoprecipitates obtained with irrelevant rabbit IgG (Fig.  4F, lane 3). These experiments demonstrate that PINCH-2, like PINCH-1 (15), forms a complex with ILK in human cells.
The N-terminal-most LIM Domain (LIM1) of PINCH-2 Mediates the Association with ILK-We next sought to identify the domain of PINCH-2 that mediates the interaction with ILK. Because an ILK-binding site is located within the PINCH-1 LIM1 domain (12,13,31), we hypothesized that the PINCH-2 LIM1 domain, which contains a sequence almost identical to that of the PINCH-1 ILK-binding sequence, mediates the interaction with ILK. To test this experimentally, we expressed GFP-and FLAG-tagged LIM1-deletion mutants of PINCH-2, and the wild type and the LIM5-deletion mutants of PINCH-2 as controls, in mammalian cells. The expression of the GFPand FLAG-tagged wild type (Fig. 5A, lane 2), the LIM5-deletion mutant of PINCH-2 (Fig. 5A, lane 3), the LIM1-deletion mutant of PINCH-2 (Fig. 5A, lane 4), or GFP alone as a negative control (Fig. 5A, lane 1) was confirmed by Western blotting with anti-GFP antibodies. The GFP-and FLAG-tagged wild type or mutant forms of PINCH-2 (Fig. 5A, lanes 6 -8), but not GFP (Fig.  5A, lane 5), were immunoprecipitated with monoclonal anti-FLAG antibody (M2)-conjugated beads. Western blotting analyses of the immunoprecipitates showed that ILK was co-immunoprecipitated with GFP-FLAG-PINCH-2 (Fig. 5B, lane 6) or the LIM5-deletion mutant (Fig. 5B, lane 7) but not the LIM1deletion PINCH-2 mutant (Fig. 5B, lane 8), indicating that the LIM1 domain of PINCH-2 indeed mediates the interaction with ILK. Probing the immunoprecipitates with a monoclonal antibody recognizing CH-ILKBP, which is known to interact with the C-terminal domain of ILK (15), showed that CH-ILKBP was co-immunoprecipitated in the presence of ILK (Fig. 5C,  lanes 6 and 7) but not in the absence of ILK (Fig. 5C, lane 8). In control experiments, neither ILK (Fig. 5B, lane 5) nor CH-ILKBP (Fig. 5C, lane 5) was detected in immunoprecipitates from the GFP control transfectants, confirming the specificity of the co-immunoprecipitation assays. The Association with ILK Is Required for the Localization of PINCH-2 to Focal Adhesions-Because PINCH-2 associates with ILK (Figs. 4 and 5) and the two proteins are co-clustered at focal adhesions (Fig. 3), we tested whether the association with ILK is required for the localization of PINCH-2 to focal adhesions. To do this, we analyzed the subcellular localization of GFP-and FLAG-tagged wild type PINCH-2, the ILK-binding defective LIM1-deletion mutant of PINCH-2, and the ILKbinding competent LIM5-deletion mutant of PINCH-2 in mammalian cells. Consistent with the results obtained with the affinity purified anti-PINCH-2 antibodies (Fig. 3), the GFPand FLAG-tagged wild type PINCH-2 localized to focal adhesions (Fig. 6A), where ILK was clustered (Fig. 6B). The ILKbinding competent LIM5-deletion mutant, like the wild type PINCH-2, was also able to localize to focal adhesions, albeit the efficiency appeared to be lower (Fig. 6, E and F). By marked contrast, the ILK-binding defective LIM1-deletion mutant of PINCH-2 failed to localize to the ILK-rich focal adhesions (Fig.  6, C and D), indicating that localization of PINCH-2 to focal adhesions requires the association with ILK.
Overexpression of PINCH-2 Inhibits the PINCH-1-ILK Interaction-Because PINCH-2, like PINCH-1, forms a complex with ILK, we tested whether PINCH-2 plays a role in the regulation of the PINCH-1⅐ILK⅐CH-ILKBP complex formation. To do this, we transfected mammalian cells with an expression vector encoding FLAG-PINCH-2, or a FLAG vector lacking the PINCH-2 sequence as a control, and analyzed the formation of the ILK complexes by co-immunoprecipitation with a monoclonal anti-CH-ILKBP antibody. Equal amounts of ILK were co-immunoprecipitated with CH-ILKBP from both the FLAG-PINCH-2 and the vector control transfectants (Fig. 7, A and B), indicating that the ILK-CH-ILKBP interaction was not impaired by overexpression of PINCH-2. Western blotting analyses of the same immunoprecipitates with anti-PINCH-1 antibodies showed that the amount of PINCH-1 associated with ILK, however, was significantly reduced in cells overexpressing FLAG-PINCH-2 (Fig. 7C, compare lanes 1 and 2). In control experiments, FLAG-PINCH-2 was readily detected in immunoprecipitates from cells overexpressing FLAG-PINCH-2 (Fig.  7D, lane 2) but not in those from the control cells (Fig. 7D, lane  1), confirming that FLAG-PINCH-2 forms a complex with ILK and CH-ILKBP. To further analyze this, we immunoprecipitated FLAG-PINCH-2 from the FLAG-PINCH-2 expressing cells (Fig. 8A, lane 1). Western blotting analyses of the FLAG-PINCH-2 immunoprecipitates showed that ILK was co-immunoprecipitated with FLAG-PINCH-2 (Fig. 8B, lane 1). No PINCH-1, however, was co-immunoprecipitated with ILK and FLAG-PINCH-2 (Fig. 8C, lane 1), despite the presence of abundant PINCH-1 in the cell lysates (Fig. 8C, lane 3). In control experiments, no ILK (Fig. 8B, lane 2) was precipitated with the anti-FLAG antibody in the absence of FLAG-PINCH-2 (Fig. 8A,  lane 2), confirming the specificity of the co-immunoprecipitation. Taken together, these results suggest that the binding of FLAG-PINCH-2 to ILK prevents the PINCH-1 binding to ILK, resulting in a reduction of the complex formation between PINCH-1 and ILK.
Overexpression of PINCH-2 Inhibits Cell Spreading and Migration-We have previously shown that the interaction between PINCH-1 and ILK is critically involved in cell shape change and migration (21). To test whether PINCH-2 is involved in the regulation of cell shape change, we transiently transfected human embryonal kidney 293 cells, a cell line that is known for its high transfection efficiency, with an expression vector that encodes FLAG-PINCH-2 or a FLAG vector that lacks PINCH-2 sequence as a control. The expression of FLAG-PINCH-2 in the FLAG-PINCH-2 transfectants (Fig. 9A, lane  3), but not in the parental 293 (Fig. 9A, lane 1) or the control vector transfectants (Fig. 9A, lane 2) was confirmed by Western blotting with a monoclonal anti-FLAG antibody. To assess cell spreading, we plated the parental 293 cells, the FLAG-PINCH-2 transfectants, and the control transfectants on fibronectin-coated surface. Majorities of the parental 293 cells and the control transfectants exhibited spread morphology within 30 min of plating (Fig. 9, B and C). By contrast, most of the FLAG-PINCH-2 transfectants remained round (Fig. 9, B and C), suggesting that overexpression of PINCH-2 inhibits the spreading of 293 cells. Despite our efforts, we were unable to obtain stable 293 clones overexpressing FLAG-PINCH-2. To facilitate further functional studies on PINCH-2, we generated a recombinant adenovirus encoding FLAG-PINCH-2. REF-52 were infected with the FLAG-PINCH-2 adenovirus or an adenovirus encoding ␤-galactosidase as a control. Under the experimental condition used, the viral infection efficiency was ϳ80 -90%. The expression of FLAG-PINCH-2 in the FLAG-PINCH-2 adenoviral-infected REF-52 cells (Fig. 10A, lane 3) but not in the control adenoviral infected REF-52 cells (Fig.  10A, lane 2) or uninfected REF-52 cells (Fig. 10A, lane 1) was confirmed by Western blotting. Consistent with the results obtained with 293 cells, overexpression of FLAG-PINCH-2 significantly inhibited the spreading of REF-52 cells (Fig. 10B). To analyze the effect of PINCH-2 overexpression on cell motility, we wound monolayers of REF-52 cells overexpressing FLAG-PINCH-2 or those of the control cells and measured the migration of the cells into the cell-free area as previously described (21,29,30). The results showed that cells overexpressing FLAG-PINCH-2 migrated much slower than the control cells (Fig. 10, C and D). Thus, consistent with an inhibitory effect on cell spreading, overexpression of FLAG-PINCH-2 significantly reduces cell migration. DISCUSSION In this study, we have cloned and characterized PINCH-2, a new member of the PINCH family. Our results have shown that: 1) PINCH-2 is co-expressed with PINCH-1 in a variety of human cells; 2) PINCH-2 localizes to both cell-ECM contact sites and the nucleus; 3) PINCH-2 forms a complex with ILK in mammalian cells and the LIM1 domain of PINCH-2 mediates the complex formation; 4) the PINCH-2 LIM1 domain, which mediates the association with ILK, is required for the localization of PINCH-2 to cell-ECM contact sites; 5) the PINCH-2-ILK and PINCH-1-ILK interactions are mutually exclusive; and 6) overexpression of PINCH-2 inhibits the PINCH-1-ILK interaction and reduces cell spreading and migration. These results identify a new component of the cell-ECM adhesion structure and suggest an important role of PINCH-2 in the cellular control of the PINCH-1⅐ILK⅐CH-ILKBP complex formation, cell shape change, and migration.
In previous studies, we have shown that ILK, through interactions mediated by its N-and C-terminal domains, respectively, forms a ternary complex with PINCH-1 and CH-ILKBP in mammalian cells (15). The PINCH-1⅐ILK⅐CH-ILKBP complex likely represents an evolutionally conserved and functionally important complex that is involved in the coupling of ECM to the actin cytoskeleton at cell-ECM adhesion sites (11). In invertebrate organisms such as C. elegans, PINCH-1/UNC-97 (22), ILK/PAT-4 (23), and CH-ILKBP/PAT-6 (46) are co-expressed in body wall muscle cells. Null mutations or suppression of expression of any one of the three proteins (PINCH-1/ UNC-97, ILK/PAT-4, or CH-ILKBP/PAT-6) all result in a Pat phenotype similar to that of ␤-integrin/pat-3 or ␣-integrin/ pat-2, which is characterized by defects in dense body and M-line assembly (24,25). In mammalian systems, PINCH-1 (12,32), ILK (14,19,26,33,34), and CH-ILKBP (15)(16)(17) are expressed in many different types of tissues and cells. We have detected the PINCH-1⅐ILK⅐CH-ILKBP complex in all types of mammalian cells that we have analyzed, which include both primary cells such as human and rat mesangial cells and established cell lines such as mouse C2C12 myoblasts, rat embryo fibroblasts, Chinese hamster ovary cells, human 293 kidney cells, and human WI-38, and IMR-90 lung fibroblasts. 2 Consistent with the genetic studies in invertebrate systems, disruption of the PINCH-1⅐ILK⅐CH-ILKBP complex in mammalian cells with dominant negative forms of PINCH-1 or ILK impairs cell spreading and migration (21). The importance of the PINCH-1⅐ILK⅐CH-ILKBP complex in the cellular control of cell shape change and motility and the ability of the dominant negative forms of PINCH-1 or ILK to inhibit the PINCH-1⅐ILK⅐CH-ILKBP complex formation beg the question as to whether there exist cellular proteins that can compete with the binding of PINCH-1 to ILK and thereby regulate the assembly of the PINCH-1⅐ILK⅐CH-ILKBP complex. The findings described in this report suggest that PINCH-2 likely functions as a naturally occurring regulator of the PINCH-1⅐ILK⅐CH-ILKBP complex. First, PINCH-2 is co-expressed with PINCH-1 in human cells. Second, PINCH-2, like PINCH-1, can form a complex with ILK. Third, the complex formation of ILK with PINCH-2 excludes the ability of ILK to form a complex with PINCH-1. decrease of the amount of PINCH-1 associated with ILK. Thus, cells could regulate the formation of the PINCH⅐1⅐ILK-CH-ILKBP complex and consequently, the functions of the PINCH-1⅐ILK⅐CH-ILKBP complex, by controlling the cellular level of PINCH-2.
An important functional consequence of overexpression of PINCH-2 is that it significantly impairs cell spreading and migration. This is remarkably similar to that of overexpression of the ILK ANK fragment or the PINCH-1 fragment containing the LIM1 domain (21), two dominant negative inhibitors of the PINCH-1-ILK interaction, suggesting that the inhibitory effect of overexpression of PINCH-2 on cell spreading and migration is caused by, at least in part, down-regulation of the PINCH-1⅐ILK⅐CH-ILKBP complex. The results described in this paper, however, do not exclude the possibility that PINCH-2 could participate in other events that regulate cell shape change and migration. Cell shape change and migration are complex processes involving coordinated regulation of multiple proteinprotein interactions at cell-ECM adhesion sites. A common feature of many components of the cell-ECM adhesion structures is that they interact with multiple partners at the adhesion sites (35). The finding that PINCH-2 localizes to cell-ECM adhesion sites, together with the fact that PINCH-2 contains five LIM domains and only one of them (LIM1) is involved in the association with ILK, suggest that PINCH-2 could potentially interact with other components of the cell-ECM adhesion structures. Identification and characterization of other binding partners of PINCH-2 will be an important goal of future studies.
In addition to showing that PINCH-2 localizes to cell-ECM adhesion sites and regulates the formation of the PINCH-1⅐ILK⅐CH-ILKBP complex, cell spreading and migration, we have found that a substantial amount of PINCH-2 is present in the nucleus. Nix and Beckerle (36) have demonstrated that zyxin, a focal adhesion protein with three LIM domains, shuttles between the cytoplasm and the nucleus. Yang et al. (37) have shown that Hic-5, a paxillin-related focal adhesion protein with four LIM domains, also localizes to nuclei. Thus, PINCH-2 appears to join zyxin and Hic-5 as a member of the LIM protein group that localize to both focal adhesions and the nucleus. It is interesting to note, however, that at steady state, zyxin is typically undetectable in the nucleus by indirect immunofluorescent staining because of the presence of a nuclear export signal (36,38). The immunofluorescent staining of nuclear Hic-5 was also weak (37) or undetectable (39), albeit the nuclear signal was enhanced after high-salt and detergent extraction to remove cytoplasm-and focal adhesion-localized Hic-5 (37). In contrast, PINCH-2 was readily detected at steady state in the nuclei by indirect immunofluorescent staining (Fig. 3). What is the molecular basis that controls the distribution of PINCH-2 to the two different subcellular compartments? The findings that PINCH-2 is co-clustered with ILK at focal adhesions (Fig. 3) and that the mutation dissociating the PINCH-2⅐ILK complex impairs the focal adhesion localization of PINCH-2 (Fig. 6) suggest that the association with ILK is required for this process. Whereas it is clear that the LIM1 domain is involved in the focal adhesion localization of PINCH-2, the sites that are involved in the nuclear transport remain to be defined. Inspection of the PINCH-2 sequence reveals that in the C-terminal region there exists a stretch of positively charged residues and an overlapping leucine-rich sequence (LELKKRLKKLSEL) that could potentially be involved in the nuclear export. It is interesting to note in this regard that the PINCH-2 C-terminal deletion mutant (⌬LIM5), which lacks the putative leucine-rich nuclear export signal, accumulated in the nucleus (Fig. 6E), suggesting that the leucine-rich sequence could be involved in the nuclear export of PINCH-2. Additional studies, however, are needed to further define the sites that are involved in the nuclear transport of PINCH-2, as GFP-⌬LIM5 can also localize to focal adhesions (albeit with lower efficiency).
It has been well documented that cell-ECM adhesion is pivotal to the regulation of gene expression and cell cycle progression (5, 40 -42). Proteins that localizes to both the nucleus and cell-ECM contact sites likely actively participate in this process. For example, c-Abl, a nonreceptor tyrosine kinase that localizes to both the nucleus and cell-ECM contact sites (43,44), is involved in the cell cycle-dependent and DNA damageinduced gene expression (45). The finding that abundant PINCH-2 is present in the nucleus suggest that, in addition to regulating the PINCH-1-ILK interaction, cell spreading, and migration, PINCH-2 likely participates in the regulation of nuclear processes. Furthermore, the distribution of PINCH-2 in both the nuclei and focal adhesions raises an interesting possibility that PINCH-2 could be involved in linking the nuclear processes to cell shape changes. For example, it has been well described that adherent cells undergo extensive shape change during cell cycle. During M phase, mitosis, in which the nuclear envelope breaks down, is always (at least in normal cells) followed by cell rounding up and cytokinesis. Given the dual localization of PINCH-2 in both the nucleus and focal adhesions and its role in the regulation of the PINCH-1⅐ILK⅐CH-ILKBP complex and cell spreading, it will be of considerable interest to test whether PINCH-2 participates in the cellular control of cytokinesis and other processes that involve coordinated changes of cell shape.