PINCH-1 Is an Obligate Partner of Integrin-linked Kinase (ILK) Functioning in Cell Shape Modulation, Motility, and Survival*

PINCH-1 is a widely expressed focal adhesion protein that forms a ternary complex with integrin-linked kinase (ILK) and CH-ILKBP/actopaxin/α-parvin (abbreviated as α-parvin herein). We have used RNA interference, a powerful approach of reverse genetics, to investigate the functions of PINCH-1 and ILK in human cells. We report here the following. First, PINCH-1 and ILK, but not α-parvin, are essential for prompt cell spreading and motility. Second, PINCH-1 and ILK, like α-parvin, are crucial for cell survival. Third, PINCH-1 and ILK are required for optimal activating phosphorylation of PKB/Akt, an important signaling intermediate of the survival pathway. Whereas depletion of ILK reduced Ser473 phosphorylation but not Thr308 phosphorylation of PKB/Akt, depletion of PINCH-1 reduced both the Ser473 and Thr308 phosphorylation of PKB/Akt. Fourth, PINCH-1 and ILK function in the survival pathway not only upstream but also downstream (or in parallel) of protein kinase B (PKB)/Akt. Fifth, PINCH-1, ILK and to a less extent α-parvin are mutually dependent in maintenance of their protein, but not mRNA, levels. The coordinated down-regulation of PINCH-1, ILK, and α-parvin proteins is mediated at least in part by proteasomes. Finally, increased expression of PINCH-2, an ILK-binding protein that is structurally related to PINCH-1, prevented the down-regulation of ILK and α-parvin induced by the loss of PINCH-1 but failed to restore the survival signaling or cell shape modulation. These results provide new insights into the functions of PINCH proteins in regulation of ILK and α-parvin and control of cell behavior.

Cell extracellular matrix (ECM) 1 adhesion plays important roles in a number of fundamental cellular processes including cell shape modulation, motility, and survival (1)(2)(3)(4)(5). Mammalian cell ECM adhesions (e.g. focal adhesions) are complex structures consisting of at least several dozens of structurally and functionally diverse proteins (6,7). Defining the functions of the cell-ECM adhesion proteins, therefore, is essential for understanding the molecular mechanisms underlying mammalian cell shape modulation, motility, survival, and other processes involving cell-ECM adhesions.
While previous biochemical and cell biological studies in mammalian cells and genetic studies in invertebrates are largely in agreement with each other, particularly on the role of PINCH, ILK, and parvin in the coupling of the cell-ECM adhesions to the actin cytoskeleton, given that multiple PINCH and parvin proteins and potentially other proteins sharing certain similar structural motifs are present in mammalian cells, studies in mammalian cells using genetic approaches that complement the biochemical studies in mammalian cells, and the genetic studies in invertebrate systems are required for further defining the functions of each of these proteins in mammalian cells. Toward this goal, we have used RNA interference (RNAi), a powerful approach of reverse genetics (32)(33)(34), to suppress the expression of PINCH-1 and ILK in human cells and investigated the functional consequences. Our results demonstrate that PINCH-1 and ILK are indispensable for proper control of cell shape change, motility, and survival. Furthermore, they reveal a novel function of PINCH proteins in regulation of the protein levels of ILK and ␣-parvin and suggest that PINCH-1 is an obligate partner of ILK in proper control of cell behavior.
RNA Interference-The sequences of PINCH-1 and ILK siRNA were selected based on a previously described method (32). The targeted sequences (sense sequence) that we found effectively mediate the silencing of the expression of PINCH-1 and ILK are 5Ј-AAGGUGAUGUG-GUCUCUGCUC-3Ј (PINCH-1) and 5Ј-AAGGACACAUUCUGGAAGG-GG-3Ј (ILK), respectively. The sequence of siRNA that targets ␣-parvin was as previously described (35). The 21-nucleotide synthetic siRNA duplex was prepared by Dharmacon Research. Hela cells were transfected with the PINCH-1 siRNA, the ILK siRNA, the ␣-parvin siRNA, or a 21-nucleotide irrelevant RNA duplex as a control using Oligofectamine (Invitrogen). In some experiments, caspase inhibitor Z-VAD (100 M) and/or proteasome inhibitor MG132 (Calbiochem) (10 M) were/was added into the culture medium 4 -7 h after siRNA transfection. The cells were analyzed 48 h after siRNA transfection by Western blotting with rabbit polyclonal anti-PINCH antibodies, mouse antiactin, anti-ILK (clone 65.1), anti-␣-parvin (clone 3B5), anti-FLAG (M5), or anti-paxillin (clone 349) mAbs as specified in each experiment.
Cell Spreading-Cells (as specified in each experiment) were plated in Opti-MEM I serum free medium (Invitrogen) in fibronectin (10 g/ ml)-coated 96-well plates. 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 and recorded with a digital camera. Unspread cells were defined as round cells while spread cells were defined as cells with extended processes as described (20,26,36,37). The percentage of cells adopting spread morphology was quantified by analyzing at least 300 cells from four randomly selected fields (Ͼ80 cells/field).
Cell Motility-Cell motility assay was performed as previously described (38,39). Briefly, the undersurfaces of 8-mm pore diameter Transwell® motility chamber (Costar, Cambridge, MA) were coated with 10 g/ml fibronectin. Cells (2 ϫ 10 4 ) that were suspended in 0.1 ml of DMEM containing 5 mg/ml bovine serum albumin were added to the upper chamber of the Transwell R motility chambers and incubated at 37°C under a 5% CO 2 -95% air atmosphere for 18 h. Following incubation, the cells on the upper surface of the membrane were removed. The membranes were fixed with 2% formalin, and the cells on the undersurface were stained with Gill's III hematoxylin. The cells from five randomly selected microscopic fields were counted, and cell motility was expressed as the number of the cells/mm 2 of the microscopic fields ϩ S.D.
DNA Constructs and Transfection-The DNA construct (pUSEamp(ϩ)/myr-Akt1) encoding myristoylated Akt was from Upstate Biotechnology (Lake Placid, NY). The pFLAG-PINCH-2 expression vector encoding FLAG-PINCH-2 was previously described (18). Hela cells were transfected with pUSEamp(ϩ)/myr-Akt1 (and pUSEamp(ϩ) vector lacking Akt sequence as a control) or pFLAG-PINCH-2 (and pFLAG vector lacking PINCH-2 sequence as a control) using LipofectAMINE PLUS (Invitrogen). One day after DNA transfection, the cells were transfected with siRNA (as specified in each experiment) using Oligofectamine. The cells were analyzed 2 days after siRNA transfection by Western blotting with antibodies as specified in each experiment.
IGF-1 Stimulation-Cells were transfected with the PINCH siRNA, the ILK siRNA or the control RNA as described above. One day after siRNA transfection, the siRNA transfectants and control cells were starved in serum-free DMEM for 18 -24 h and then stimulated with DMEM containing 2 ng/ml IGF-1. At the end of stimulation, cells were lysed with 1% Triton X-100 in phosphate-buffered saline containing protease inhibitors and analyzed by Western blotting.
Apoptosis-Cells were cultured in DMEM supplemented with 10% fetal bovine serum. Apoptosis was analyzed 48 h after RNA transfection using terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end-labeling (TUNEL) and caspase-3 assays. In some experiments, Z-VAD (100 M) was added into the culture medium 4 h after siRNA transfections. The TUNEL assay was performed using an In Situ Cell Death Detection kit (Roche Applied Science). The caspase-3 activities were measured using a colorimetric caspase-3 assay kit from Calbiochem.

RESULTS
RNAi Depletion of PINCH-1-To evaluate the functions of PINCH-1 in human cells, we transfected Hela cells with a small interfering RNA (siRNA) that specifically targets PINCH-1, an irrelevant small RNA as a negative control, and an ␣-parvin siRNA that we have previously shown effectively suppressed the expression of ␣-parvin (35). Transfection of the cells with the PINCH-1-specific siRNA resulted in a near total loss of PINCH-1 (Fig. 1A, compare lanes 1 and 2). Depletion of PINCH-1 drastically reduced the level of ILK (Fig. 1C, compare lanes 1 and 2) but did not alter the level of paxillin (Fig. 1E, compare lanes 1 and 2) or actin (Fig. 1B, compare lanes 1 and  2). The level of ␣-parvin in the PINCH-1-deficient cells was modestly lower than that of the control cells (Fig. 1D, compare lanes 1 and 2) but is substantially higher than cells that were transfected with the ␣-parvin siRNA (Fig. 1D, compare lanes 1  and 3). Conversely, depletion of ␣-parvin modestly reduced the levels of PINCH-1 (Fig. 1A, compare lanes 2 and 3) and ILK (Fig. 1C, compare lanes 2 and 3). Taken together, these results suggest that the PINCH-1 siRNA effectively suppresses the expression of PINCH-1 without globally altering protein expression. Furthermore, they reveal that the cellular level of ILK, and to a less extend that of ␣-parvin, are regulated by PINCH-1.
Functional Consequences of PINCH-1 Depletion-We next analyzed the functional consequences of the siRNA-mediated depletion of PINCH-1 on cell behavior. Previous studies using dominant negative forms of PINCH-1, ␣-parvin, or ␤-parvin have implicated important roles of PINCH (26) and parvin (20,21,23) proteins (or other related proteins) in cell spreading. To test whether PINCH-1 or ␣-parvin is required for cell spreading, we plated the PINCH-1 siRNA transfectants, the ␣-parvin transfectants and the control transfectants on fibronectincoated surface. As expected, a majority of the control cells began to spread within 30 min of plating ( Fig. 2, C and G). By marked contrast, most of the PINCH-1-deficient cells remained round at the same time point (Fig. 2, A and G). The cellspreading defect induced by the depletion of PINCH-1 was obvious even after prolonged incubation (2 h) ( Fig. 2B) when the control cells adapted full spread morphology (Fig. 2D). Surprisingly, depletion of ␣-parvin did no inhibit cell spreading (Fig. 2, E and F). In fact, a higher percentage of the ␣-parvindeficient cells adapted spread morphology at 30 min of plating comparing to that of the control cells (Fig. 2G). Thus, PINCH-1, but not ␣-parvin, is required for prompt cell spreading.
The finding that PINCH-1 functions in cell spreading prompted us to test whether PINCH-1 is required for cell motility, a cellular process that is critically involved in normal development as well as a variety of pathological processes. To do this, we plated the PINCH-1-deficient cells and the control cells in Transwell R motility chambers in which the undersurfaces of the membranes were pre-coated with fibronectin. As expected, the control cells migrated through the pores of the membrane within 18 h of plating (Fig. 2, I and J). By contrast, the PINCH-1-deficient cells were unable to migrate through the pores of the membrane under the identical condition ( Fig. 2, H and J), suggesting that PINCH-1 is indispensable for normal cell motility.
We have recently found that ␣-parvin is crucial for protecting cells from apoptosis (35). It was not known, however, whether PINCH-1 is involved in cell survival. To test this, we transfected Hela cells with the PINCH-1 siRNA and the control RNA, respectively, and analyzed the effect on apoptosis. The results showed that depletion of PINCH-1, like that of ␣-parvin (35), resulted in a significant increase of caspase-3 activity (Fig. 3A) and TUNELpositive cells (Fig. 3, B and C), suggesting that PINCH-1 is required for protection of cells from apoptosis. Consistent with the activation of caspase-3 induced by the depletion of PINCH-1, treatment of the cells with Z-VAD, a potent caspase inhibitor, completely blocked apoptosis induced by the depletion of PINCH-1 (Fig. 3, B and C), indicating that PINCH-1 functions in apoptosis in a caspase-dependent manner.
PKB/Akt is a signaling intermediate that is crucial for transmission of extracellular survival signals to downstream effectors including caspases (for recent reviews, see Refs. 40 -42). Upon stimulation with extracellular survival factors (e.g. IGF-1), PKB/Akt is phosphorylated at Ser 473 near the C terminus and Thr 308 in the activation loop (43), which leads to full activation of PKB/Akt and phosphorylation of a number downstream apoptosis mediators, resulting in protection of cells from apoptosis. Stimulation of control cells with IGF-1, as expected, induced phosphorylation of PKB/Akt at both Ser 473 and Thr 308 (Fig. 3D, lane 4). Depletion of PINCH-1 reduced (albeit did not eliminate) the phosphorylation of PKB/Akt at Ser 473 and Thr 308 (Fig. 3D, compare lanes 2 and 4), suggesting that PINCH-1 is involved in the regulation of the activating phosphorylation of PKB/Akt.
RNAi Depletion of ILK-We next sought to suppress the expression of ILK, which we have previously shown to form a ternary complex with PINCH-1 and ␣-parvin in mammalian cells (20,25), with a siRNA that specifically targets ILK sequence. To do this, we transfected Hela cells with the ILKspecific siRNA, and the PINCH-1 siRNA and an irrelevant small RNA as controls, respectively. Western blotting analyses of the cells showed that the level of ILK in the ILK siRNA transfectants (Fig. 4A, lane 1), like that in the PINCH-1 siRNA transfectants (Fig. 4A, lane 2), was drastically reduced, confirming that the ILK siRNA effectively suppresses the expression of ILK. In control experiments, similar levels of actin were detected in all lanes (Fig. 4B, lanes 1-3), confirming equal loading of the samples. Depletion of ILK substantially reduced the level of PINCH-1 (Fig. 4C, compares lanes 1 and 3). The level of PINCH-1 protein in the ILK-deficient cells, however, was noticeably higher than that in the PINCH-1 siRNA transfectants (Fig. 4C, compares lanes 1 and 2). Depletion of ILK, like that of PINCH-1, modestly reduced the level of ␣-parvin (Fig. 4D, lanes 1-3). These results, together with those shown in Fig. 1, indicate that PINCH-1, ILK and to a less extent required for optimal activating phosphorylation of PKB/Akt. The PINCH-1 siRNA transfectants (lanes 1 and 2) and the control RNA transfectants (lanes 3 and 4) were serum-starved (lanes 1 and 3) and stimulated with 2 ng/ml IGF-1 (lanes 2 and 4) for 10 min. The cell lysates (25 g/lane) were analyzed by Western blotting with antibodies recognizing phospho-Akt(Ser 473 ), phospho-Akt(Thr 308 ), Akt, and actin, respectively. ␣-parvin, are mutually dependent in maintenance of their protein levels in cells.
Functional Consequences of ILK Depletion-We next analyzed the effects of siRNA-mediated ILK depletion on cell behavior. Depletion of ILK severely impaired cell spreading (Fig.  5, A and B) and reduced cell motility (Fig. 5, C and D). Furthermore, we detected significant increases of caspase-3 activity and positive TUNEL staining in the ILK-deficient cells (Fig.  5, E-J). Thus, ILK, like PINCH-1, is indispensable for proper cell shape change, motility and survival. To assess the effect of ILK on activating phosphorylation of PKB/Akt, we stimulated the ILK siRNA transfectants, the PINCH-1 siRNA transfectants and the control RNA transfectants with IGF-1. As expected, stimulation of the control cells with IGF-1 induced phosphorylation of PKB/Akt at both Ser 473 and Thr 308 (Fig. 6A,  lane 6). Consistent with the results shown earlier (Fig. 3D), the Ser 473 and Thr 308 phosphorylation was reduced but not eliminated in Hela cells that were transfected with the PINCH-1 siRNA (Fig. 6A, compares lanes 2 and 6). The Ser 473 phosphorylation but not Thr 308 phosphorylation of PKB/Akt was reduced in the ILK siRNA transfectants (Fig. 6A, compare lanes  4 and 6). Depletion of PINCH-1 and ILK by transfection of Hela cells with both the PINCH-1 siRNA and the ILK siRNA resulted in a reduction of phosphorylation of PKB/Akt at both Ser 473 and Thr 308 similar to that induced by the depletion of PINCH-1 (data not shown). Taken together, these results sug-gest that ILK is required for the optimal phosphorylation of PKB/Akt at Ser 473 but not that at Thr 308 , whereas PINCH-1 is required for the optimal phosphorylation of PKB/Akt at both sites.

PINCH-1 and ILK Function in Cell Survival Not Only Upstream but Also Downstream of PKB/Akt-To test whether
PKB/Akt is the only target of PINCH-1 and ILK in cell survival, we expressed a constitutively active myristoylated PKB/Akt (44) in Hela cells. Abundant Ser 473 -and Thr 308 -phosphorylated PKB/Akt protein was detected in cells that were transfected with the PINCH-1 siRNA (Fig. 6B, lane 2), the ILK siRNA (Fig.  6B, lane 3) or the control RNA (Fig. 6B, lane 1). Importantly, despite the expression of the constitutively active PKB/Akt, depletion of PINCH-1 or ILK resulted in a significant increase of apoptosis as revealed by the appearance of high percentages of TUNEL-positive cells (Fig. 6, C-O) and significant increases of the caspase-3 activity (not shown). Thus, expression of constitutively active PKB/Akt is insufficient for protection of cells from apoptosis induced by the depletion of PINCH-1 or ILK, suggesting that PINCH-1 and ILK function in the survival pathway not only upstream of PKB/Akt but also downstream (or in parallel) of PKB/Akt.

PINCH-1 Regulates the Protein Levels of ILK and ␣-Parvin-
The foregoing experiments provide important functional evidence for crucial roles of PINCH-1 and ILK in the cellular control of shape change, motility, and survival. An additional Bar, 20 m. The percentage of ILK siRNA and control RNA transfectants that adopted spread morphology 30 min after plating was quantified (B). Data represent means ϩ S.D. C and D, cell motility was analyzed using Transwell R motility chambers as described in the Experimental procedures. The ILK siRNA transfectants and the control transfectants from five randomly selected microscopic fields were counted and cell motility was expressed as the number of the cells/mm 2 1 and 2), ILK siRNA transfectants (lanes 3 and  4), and the control RNA transfectants (lanes 5 and 6) were serum-starved (lanes 1, 3, and 5) and stimulated with 2 ng/ml IGF-1 (lanes 2, 4, and 6) for 10 min. The cell lysates (25 g/lane) were analyzed by Western blotting with antibodies recognizing phospho-Akt(Ser 473 ), phospho-Akt(Thr-), Akt, and actin, respectively. B, Hela cells were transfected with the pUSEamp(ϩ)/myr-Akt1 vector encoding myristoylated PKB/Akt (lanes 1-3) or a pUSEamp(ϩ) vector lacking Akt sequence as a control (lanes 4 -6). The myristoylated PKB/Akt and the control vector transfectants were then transfected with the PINCH-1 siRNA (lanes 2 and 5), the ILK siRNA (lanes 3 and 6) and the control RNA (lanes 1 and 4), respectively. Forty-eight hours after siRNA transfection, the transfectants that were grown in DMEM containing 10% fetal bovine serum were analyzed by Western blotting with antibodies recognizing Akt ( important and somewhat surprising finding of these experiments was that siRNA-mediated silencing of PINCH-1 expression resulted in not only depletion of PINCH-1 but also a drastic reduction in ILK protein level and a modest reduction in ␣-parvin protein level. Conversely, siRNA-mediated silencing of ILK expression resulted in not only depletion of ILK but also a substantial reduction in PINCH-1 protein level and a modest reduction in ␣-parvin protein level. To determine the mechanism underlying the co-regulation of ILK, PINCH-1 and ␣-parvin, we analyzed the mRNA levels of ILK, PINCH-1, and ␣-parvin in Hela cells that were transfected with the PINCH-1 siRNA, the ILK siRNA, the ␣-parvin siRNA and the control RNA, respectively, by reverse transcription PCR. The results showed that, as expected, the level of ILK mRNA was drastically reduced in the ILK siRNA transfectants (Fig. 7A, compare  lane 5 with lanes 2 and 3). Importantly, despite the drastic reduction of the ILK protein level (Fig. 1C, lane 1), the ILK mRNA level was not affected by the depletion of PINCH-1 (Fig.  7A, compare lane 4 with lanes 2 and 3), suggesting that PINCH-1 regulates ILK at the protein level rather than the mRNA level. The PINCH-1 mRNA was drastically reduced by transfection with the PINCH-1 siRNA (Fig. 7B, compare lane 4  with lanes 2 and 3) but not by transfection with the ILK siRNA (Fig. 7B, compare lane 5 with lanes 2 and 3). Thus, the downregulation of PINCH-1 (Fig. 4C) induced by the depletion of ILK occurs at the protein level rather than at the mRNA level. Finally, transfection with ␣-parvin siRNA depleted the mRNA of ␣-parvin (Fig. 7C, lane 6) but did not reduce the mRNA levels of ILK (Fig. 7A, compare lane 6 with lanes 2 and 3) and PINCH-1 (Fig. 7B, compare lane 6 with lanes 2 and 3). Conversely, depletion of either ILK or PINCH-1, which modestly reduced the protein level of ␣-parvin (Fig. 4D), did not reduce the mRNA level of ␣-parvin (Fig. 7C, lanes 4 and 5).
Coordinate Regulation of PINCH-1, ILK, and ␣-Parvin Levels by Proteasome-mediated Protein Degradation-The finding that PINCH-1 regulates ILK and ␣-parvin at the protein level, together with the fact that PINCH-1 forms a ternary complex with ILK and ␣-parvin in cells, raised an interesting possibility that the cellular levels of PINCH-1, ILK and ␣-parvin are coordinately regulated by proteolytic degradation systems. Because loss of PINCH-1 stimulated caspase activity (Fig. 3A), which is known to be involved in degradation of a number of cellular proteins, we analyzed the effect of Z-VAD, a potent pan-caspase inhibitor, on the down-regulation of ILK and ␣-parvin induced by the loss of PINCH-1. Treatment of the PINCH-1 siRNA transfectants with Z-VAD, which completely blocked apoptosis induced by the loss of PINCH-1 (Fig. 3, B and  C), did not prevent the loss of PINCH-1 (Fig. 7D, lane 3) or the down-regulation of ILK (Fig. 7E, lane 3) or ␣-parvin (Fig. 7F,  lane 3), suggesting that the coordinated down-regulation of PINCH-1, ILK, and ␣-parvin is not mediated by caspases. To test whether proteasome is involved in this process, we treated PINCH-1 siRNA transfectants with MG132, a potent inhibitor of proteasome (45). Because treatment of the cells with MG132 alone induced apoptosis, 3 we included Z-VAD, which effectively inhibits proteasome inhibitor-induced apoptosis (46,47), in these experiments. The results showed that treatment of the PINCH-1 siRNA transfectants with the proteasome inhibitor significantly reduced the degradation of PINCH-1 (Fig. 7D,  lane 4), ILK (Fig. 7E, lane 4) and ␣-parvin (Fig. 7F, lane 4). 3 K. Chen and C. Wu, unpublished observations. Equal protein loading was confirmed by probing the samples with an anti-actin antibody (Fig. 7G). Taken together, these results suggest that the cellular levels of PINCH-1, ILK, and ␣-parvin are coordinately regulated by, at least in part, proteasomes.

Increased Expression of PINCH-2 Prevents the Down-regulation of ILK and ␣-Parvin Induced by the Loss of PINCH-1-To
further investigate the mechanism underlying the coordinated regulation of PINCH-1, ILK, and ␣-parvin, we analyze the effect of PINCH-2, which could also form a ternary complex with ILK and ␣-parvin (18), on the cellular levels of ILK and ␣-parvin. To do this, we overexpressed FLAG-PINCH-2 in Hela cells that were transfected with the PINCH-1 siRNA. The depletion of PINCH-1 in the PINCH-1 siRNA transfectants was confirmed by Western blotting with an anti-PINCH antibody (Fig. 8A, compare lanes 1 and 3 with lane 2). The expression of FLAG-PINCH-2 in Hela cells, which normally expresses an undetectable level of PINCH-2 (Fig. 8B, lanes 1 and 2), was confirmed by Western blotting with an antibody that specifically recognizes PINCH-2 (Fig. 8B, lane 3) or FLAG (Fig. 8C,  lane 3). Probing the same samples with anti-ILK and anti-␣parvin antibodies showed that overexpression of FLAG-PINCH-2 in the PINCH-1-deficient cells prevented the downregulation of ILK (Fig. 8D, lane 3) and ␣-parvin (Fig. 8E, lane  3) induced by the loss of PINCH-1. Equal protein loading was confirmed by re-probing the membrane with an anti-actin antibody (Fig. 8F).

Restoration of ILK and ␣-Parvin Expression Is Insufficient for Rescuing the Survival Signaling and Phenotypical Defects
Induced by the Loss of PINCH-1-The restoration of the ILK and ␣-parvin levels in the PINCH-1-deficient cells allowed us to test whether the defects in PKB/Akt phosphorylation and cell survival induced by the loss of PINCH-1 reflect a direct involvement of PINCH-1 in these processes or an indirect consequence resulting from the down-regulation of ILK and ␣-parvin. To do this, we expressed FLAG-PINCH-2 in PINCH-1 siRNA transfectants and stimulated the FLAG-PINCH-2 expressing cells and the control cells with IGF-1. As expected, stimulation of the PINCH-1 expressing cells with IGF resulted in phosphorylation of PKB/Akt at both Ser 473 and Thr 308 in (Fig. 9A, lanes  2 and 6). The activating phosphorylation of PKB/Akt, however, was significantly reduced in the PINCH-1 deficient cells regardless whether FLAG-PINCH-2 was expressed (Fig. 9A, lane  8) or not (Fig. 9A, lane 4). Using a similar approach, we have found that overexpression of FLAG-PINCH-2 in PINCH-1-deficient cells failed to protect cells from apoptosis ( Fig. 9, B-D) or rescue the cell-spreading defect (data not shown) induced by the loss of PINCH-1. Thus, despite restoration of the ILK and ␣-parvin levels (Fig. 8), increased expression of FLAG-PINCH-2 is insufficient for rescuing the survival signaling and phenotypical defects induced by the loss of PINCH-1.

DISCUSSION
The results presented in this paper provide important reverse genetic evidence for crucial roles of PINCH-1 and ILK in the control of mammalian cell behavior. These results demonstrate that PINCH-1 and ILK are indispensable for at least two fundamental processes in mammalian cells. First, PINCH-1 and ILK are required for proper control of mammalian cell shape change and motility. This is consistent with the findings of genetic studies in C. elegans (17,29) and Drosophila (28,30) and therefore this function appears to be evolutionally well conserved. An important, and somewhat surprising, finding of this study is that depletion of ␣-parvin did not reduce cell spreading. This differs considerably from studies in invertebrates such as C. elegans in which PAT-6/PARVIN is essential for the coupling of integrins to actin cytoskeleton (31). The finding that ␣-parvin, but not PINCH-1 or ILK, is dispensable for mammalian cell spreading strongly suggest that other related mammalian proteins (e.g. affixin/␤-parvin) could potentially fulfill such a role in mammalian cells.
The second fundamental process in which PINCH-1 and ILK are critically involved is cell survival. This is clearly demonstrated by the findings that depletion of PINCH-1 or ILK results in a dramatic increase of apoptosis. How do PINCH-1 and ILK protect cells from apoptosis? The present studies suggest that they function in the intracellular survival signaling pathway at multiple steps. One step is at the activating phosphorylation of PKB/Akt, a signaling intermediate that is of central importance in cell survival signaling (40 -42). The functions of PINCH-1 and ILK in cell survival, however, are not limited to the regulation of the activating phosphorylation of PKB/Akt. Depletion of PINCH-1-or ILK-induced apoptosis despite the presence of abundant constitutively active PKB/Akt (Fig. 6). On the other hand, inhibition of caspases completely blocked apoptosis induced by the depletion of PINCH-1 (Fig. 3) or ILK. 4 Thus, 4 K. Chen, T. Fukuda, and C. Wu, unpublished observations.  1 and 2). One day after the DNA transfection, the cells were transfected with a PINCH-1-specific siRNA (lanes 1 and 3) or a control small RNA (lane 2). The cell lysates (20 g/lane) were analyzed by Western blotting with a rabbit polyclonal anti-PINCH-1 antibody (which also recognizes PINCH-2) (A), a rabbit polyclonal antibody that is specific for PINCH-2 (18), and mAbs recognizing FLAG (C), ILK (D), and ␣-parvin (E), respectively. The membrane used in panel B was re-probed (without stripping) with an anti-actin antibody to confirm equal loading of the samples (F). PINCH-1 and ILK function in cell survival not only by regulating the activating phosphorylation of PKB/Akt but also by transmitting survival signals from (or in parallel of) PKB/Akt to other effectors that are upstream of caspases. Increase of ILK expression is intimately associated with oncogenic transformation in a large number of cell types (see, for example, Refs. 48 -56). Our finding that PINCH-1 and ILK function in cell survival not only upstream but also downstream (or in parallel) of PKB/Akt is significant, as while it has become increasingly clear that ILK is important for regulation of activating phosphorylation of PKB/ Akt in a number of cell types including human Hela cells (this work), human 293 cells, and mouse macrophages (57), ILK is dispensable for activating phosphorylation of PKB/Akt in some other cell types (58,59). Our results suggest that even in cell types in which ILK is not required for the activation of PKB/Akt, ILK, and PINCH-1 could still play an important role in cell survival signaling by regulating caspase activities in a PKB/Aktindependent manner.
Another major finding of this study is that PINCH-1, ILK, and to a less extent ␣-parvin depend on each other for maintenance of their protein levels in cells. Based on the results presented in this paper that: 1) depletion of PINCH-1 reduces the protein levels but not the mRNA levels of ILK and ␣-parvin and vice versa, 2) inhibition of proteasomes reverses the downregulation of the PINCH-1, ILK, and ␣-parvin, and 3) increased expression of PINCH-2 blocks the down-regulation of ILK and ␣-parvin induced by the loss of PINCH-1, we propose that the assembly of the ternary PINCH-ILK-parvin complexes, which occurs prior to their localization to cell-ECM adhesions (25), prevents the proteolytic degradation of PINCH, ILK and parvin proteins. The finding that depletion of PINCH-1 in Hela cells resulted in a drastic reduction of the ILK level and vice versa suggests that majorities of PINCH-1 and ILK formed an obligate complex in these cells. On the other hand, depletion of ␣-parvin modestly reduced the levels of PINCH-1 and ILK (the reciprocal is also true), suggesting that only a modest fraction of PINCH-1 and ILK formed an obligate complex with ␣-parvin. One likely possibility is that substantial amounts of PINCH-1 and ILK proteins form a ternary complex with ␤-parvin and therefore are protected from proteolytic degradation. This is highly consistent with the findings that overexpression of dominant negative mutants of ILK or parvins that inhibit the ILK-parvin interactions (20,21,23,26), but not siRNA-mediated depletion of ␣-parvin (Fig. 2), impaired cell spreading.
The presence of structurally closely related PINCH-1 and PINCH-2 proteins in mammalian cells raises an interesting question whether they are functionally redundant. The results presented in this study suggest that while they do share certain common functions (e.g. regulation of cellular levels of ILK and ␣-parvin), they are not redundant, as overexpression of PINCH-2 fails to rescue the defects in PKB/Akt phosphorylation, survival and cell shape modulation induced by the loss of PINCH-1. The presence of two structurally and functionally related but non-redundant PINCH proteins provides a versatile system by which mammalian cells can precisely control cell morphology, motility, survival, and other fundamental cellular processes that involve the PINCH-ILK-parvin complexes.