Cell adhesion regulates Ser/Thr phosphorylation and proteasomal degradation of HEF1

Human enhancer of filamentation 1 (HEF1), a multifunctional docking protein of the Cas family, participates in integrin and growth factor signaling pathways that regulate global cellular processes including growth, motility and apoptosis. HEF1 consists of two isoforms, p105 and p115, the larger molecular weight form resulting from Ser/Thr phosphorylation of p105HEF1. The molecular mechanisms that regulate the interconversion of the two HEF1 species as well as the function of HEF1 Ser/Thr phosphorylation are unknown. Our study reveals that cell adhesion and detachment regulate the interconversion of the two HEF1 isoforms. Experiments using various inhibitors of cytoskeletal organization indicated that disruption of actin microfilaments but not intermediate filaments or microtubules resulted in a complete conversion of p115HEF1 to p105HEF1. The conversion of p115HEF1 to p105HEF1 was prevented by inhibition of protein phosphatase 2A (PP2A), suggesting that cytoskeletal regulation of PP2A activity controlled the dephosphorylation of p115HEF1. Degradation of endogenous HEF1 was dependent on proteasomes with the p115 species of HEF1 being preferentially targeted for turnover. Dephosphorylation of HEF1 by suspending cells or disrupting actin filaments protected HEF1 from degradation. These results suggest that the adhesion-dependent actin organization regulates proteasomal turnover of HEF1 through the activity of PP2A.


Introduction
Cells adhere to matrix proteins through transmembrane receptors of the integrin family, which serve as anchors for both the extracellular matrix and the intracellular actin cytoskeleton to form a structural framework that defines cell morphology and function. Upon cell adhesion and integrin clustering, the cytoplasmic tail of the integrin receptor interacts with signaling molecules and docking proteins to form the focal adhesions that relay 'outside in' and 'inside out' signaling through the integrin receptors (Liddington and Ginsberg, 2002;Schwartz and Ginsberg, 2002). Multi-functional-domain docking proteins of the Cas family play a crucial role in adhesion signaling. Upon integrin ligation, Cas family proteins localize to focal adhesions and become tyrosine phosphorylated, resulting in the regulated recruitment of SH2-containing signaling proteins to focal adhesions (Cary and Guan, 1999;Nakamoto et al., 1997;O'Neill et al., 2000;Sakai et al., 1994). The members of the Cas family include p130Cas, Efs/Sin and human enhancer of filamentation 1 (HEF1, also known as Cas-L). HEF1 was initially identified as a human protein that promotes changes in actin organization of mammalian cells (Fashena et al., 2002;Law et al., 1996;O'Neill et al., 2000). The amino acid sequence of HEF1 is 64% similar to that of p130Cas. Both contain a number of protein-protein-interaction domains including an N-terminal SH3 domain, a substrate domain, a serine-rich domain and a helix-loop-helix motif in the C-terminus . In epithelial cells, HEF1 is post-translationally processed to p115, p105, p65, p55 and p28 protein forms. The p115 and p105 forms are generated through differential Ser/Thr phosphorylation whereas the last three low-molecular-weight forms are generated by a caspasedependent cleavage (Law et al., 1998;Law et al., 2000). The p55 species of HEF1 is produced specifically at mitosis, and associates with the mitotic spindle where it interacts with Dim1p, a G2-M spindle-regulatory protein (Law et al., 1998). The p28 form is generated from the C-terminal of HEF1 during mitosis and regulates apoptosis by mediating the disassembly of the focal adhesion . Cleavage of HEF1 to lower molecular-weight forms can be prevented by integrin clustering and focal-adhesion formation . These findings suggest that, the generation of various forms of HEF1 is carefully regulated to allow for their participation in distinct cellular processes and cell adhesion is involved in this posttranslational modification process.
HEF1 is a substrate for FAK and Src family tyrosine kinases and is thought to participate in signaling pathways regulated by integrin receptors (Kanda et al., 1999;Sattler et al., 1997;Tachibana et al., 1997). Phosphorylated HEF1 may function as a downstream effector of FAK to promote integrin-dependent cell motility van Seventer et al., 2001). Molecular genetic experiments have indicated that overexpression of HEF1 induces changes in cell morphology and motility (Fashena et al., 2002), suggesting a role for HEF1 in regulating cytoskeletal organization and integrin function. Earlier studies have proposed that HEF1 is expressed primarily in cells of lymphoid and epithelial origin (Law et al., 1996).
Human enhancer of filamentation 1 (HEF1), a multifunctional docking protein of the Cas family, participates in integrin and growth factor signaling pathways that regulate global cellular processes including growth, motility and apoptosis. HEF1 consists of two isoforms, p105 and p115, the larger molecular weight form resulting from Ser/Thr phosphorylation of p105HEF1. The molecular mechanisms that regulate the interconversion of the two HEF1 species as well as the function of HEF1 Ser/Thr phosphorylation are unknown. Our study reveals that cell adhesion and detachment regulate the interconversion of the two HEF1 isoforms. Experiments using various inhibitors of cytoskeletal organization indicated that disruption of actin microfilaments but not intermediate filaments or microtubules resulted in a complete conversion of p115HEF1 to p105HEF1. The conversion of p115HEF1 to p105HEF1 was prevented by inhibition of protein phosphatase 2A (PP2A), suggesting that cytoskeletal regulation of PP2A activity controlled the dephosphorylation of p115HEF1. Degradation of endogenous HEF1 was dependent on proteasomes with the p115 species of HEF1 being preferentially targeted for turnover. Dephosphorylation of HEF1 by suspending cells or disrupting actin filaments protected HEF1 from degradation. These results suggest that the adhesiondependent actin organization regulates proteasomal turnover of HEF1 through the activity of PP2A.
Recently, we have shown that both p105HEF1 and p115HEF1 are expressed at low levels in dermal fibroblasts, and that their expression in these cells is greatly upregulated by TGF-␤1 (Zheng and McKeown-Longo, 2002). Functions of the two HEF1 isoforms are not well understood. Recent studies using overexpression systems have defined a possible role for p115HEF1 in modulating TGF␤1 signaling. In these studies, p115HEF1 was shown to undergo ubiquitination and proteasomal degradation following complex-formation with Smad3 and the ubiquitin ligase AIP4 (Feng et al., 2004;Liu et al., 2000). The molecular pathway that governs the interconversion between p105 and p115HEF1, and the role of Ser/Thr phosphorylation in HEF1 function are unknown. In this study, we show that the interconversion between p105 and p115 HEF1 is determined by actin organization and protein phosphatase 2A (PP2A), and that the proteasomal degradation of HEF1 is regulated by HEF1 Ser/Thr phosphorylation.

Proteasome inhibitors inhibit endogenous HEF1 degradation and result in p115HEF1 accumulation
We investigated the role of the proteasome on the degradation of endogenous p115 and p105HEF1 in two cell lines, human foreskin fibroblasts (A1-F) and human osteosarcoma (MG-63). As shown in Fig. 1A, treatment of both A1-F (left panel) and MG-63 cells (right panel) with either of two proteasome-specific inhibitors, lactacystine (LAC) or N-acetyl-leu-leunorleucinal (LLnL), resulted in a large increase in the level of p115HEF1 protein. A control, N-acetyl-leu-leu-methional (ALLM) had no effect on HEF1 levels. The half-life of HEF1 protein was evaluated using cycloheximide-treated MG-63 cells (Fig. 1B). In the presence of cycloheximide, levels of both HEF1 isoforms rapidly decreased. By 8 hours after the addition of cycloheximide, p105 and p115HEF1 levels were reduced more than 90%. In the presence of the proteasome inhibitor LAC, p115HEF1 protein levels remained relatively stable, increasing slightly over the 8-hour period. By contrast, p105HEF1 protein levels decreased. The decrease in p105HEF1 was due to the conversion of p105HEF1 to p115HEF1 because total HEF1 protein was not significantly changed in the presence of proteasome inhibitor (data not shown). These data indicate that inhibition of proteasomal activity blocks HEF1 degradation, resulting in the accumulation of the p115 isoform and suggesting that Ser/Thr -coated dishes were serum-starved for 6 hours and then treated with TGF-␤1 (2.5 ng/ml) for 12 hours to induce HEF1 expression (0 hours). TGF-␤1-containing medium was removed and cell layers were washed twice with serum-free medium. Cells were treated with 20 g/ml cycloheximide (CHX) in the presence or absence of 10 M lactacystin for indicated times. Cell lysates were prepared and processed for immunoblotting with anti-HEF1 pAb. Nitrocellulose membranes were stripped and reprobed with anti-p130Cas mAb as a loading control. (2.5 ng/ml) for 4 hours to induce HEF1 expression (0 hours). Cells were then detached and suspended in serum-free medium for indicated times. (B) After suspension for 2 hours (indicated as 0 hours of adhesion), A1-F cells were plated onto FN-coated dishes for indicated times. (C) After having been treated with TGF-␤1 (2.5 ng/ml) for 4 hours, A1-F monolayers were detached and suspended in serum-free medium for 1 hour. Suspended cells were then plated onto tissue culture dishes that had been coated with polylysine (PL), fibronectin (FN), collagen (Coll), laminin (LM) or vitronectin (VN) and incubated for 2 hours. Cell lysates were prepared and processed for immunoblotting with anti-HEF1 pAb. Nitrocellulose membranes were stripped and reprobed with anti-p130Cas mAb as loading control.
phosphorylation of HEF1 might regulate its turnover within the cell.

Actin organization regulates the conversion of p115HEF1 into p105HEF1
Experiments were designed to address the role of cytoskeletal organization on HEF1 Ser/Thr phosphorylation. As shown in Fig. 2, when human dermal fibroblasts were detached and maintained in suspension, the Ser/Thr-phosphorylated isoform p115HEF1 disappeared within 2 hours. The loss of p115HEF1 was accompanied by a corresponding increase in p105HEF1 ( Fig. 2A). By contrast, p130Cas levels were unchanged by cell suspension and served as a loading control. Replating of suspended cells onto fibronectin-coated dishes resulted in a complete reappearance of p115HEF1 within 2 hours, accompanied by a concomitant decrease in p105HEF1 (Fig. 2B). These results suggest that suspension and readhesion regulate the interconversion of p105 and p115HEF1. To investigate whether the adhesion-dependent appearance of p115HEF1 correlated with ligation of a specific integrin receptor, cells were plated onto dishes that had been coated with fibronectin, collagen, laminin or vitronectin and incubated for 2 hours. As seen in Fig. 2C, all of these substrates supported the restoration of p115HEF1 protein levels -albeit not to exactly the same levels, suggesting some influence of extracellular matrix on the levels of each isoform. Adhesion of cells to polylysine was unable to support cell-spreading and did not restore p115HEF1 levels. These data suggest that the presence of p115HEF1 depends primarily on an appropriate cell morphology rather than ligation of a particular integrin receptor. To elucidate which cytoskeletal structures might be important for the maintenance of p115HEF levels, cells were incubated with reagents that selectively disrupt different cytoskeletal filaments. As shown in Fig. 3A, increasing doses of either nocodazole or colchicine, which disrupt microtubules at low concentration and microtubules as well as intermediate filaments at high concentration, did not change relative levels of p105 and p115HEF1; microtubules and intermediate filaments, however, were disrupted (Fig. 3B). Treatment with the actin-filament-disrupting drugs cytochalasin D or latrunculin A resulted in loss of p115HEF1 within two hours (Fig. 3A) without altering either microtubules or intermediate filaments (Fig. 3B). Similar to results obtained in cell suspension and replating experiments (Fig. 2), cytochalasin D or latrunculin A induced a loss of p115HEF1 that was accompanied by a corresponding increase in p105HEF1. These results suggest that the disruption of actin-filament organization activates a Ser/Thr phosphatase responsible for dephosphorylation of p115HEF1, thereby converting p115HEF1 into p105HEF1. To evaluate whether proteasomal degradation contributes to the cytochalasin-D-induced disappearance of p115HEF1, proteasome inhibitors were Journal of Cell Science 119 (1) evaluated for their effect on the cytochalasin-D-induced decrease in p115HEF1. Cells were pre-treated with the proteosomal inhibitor LAC before being treated with cytochalasin D. Consistent with results shown in Fig. 1, LAC treatment resulted in an increase in p115HEF1 (Fig. 4, lane 2).  3. The effects of cytoskeleton-disruptive agents on the conversion of p115HEF1 to p105HEF1. A1F cells were serum-starved overnight, detached and resuspended in DMEM containing 0.1% heat-inactivated BSA and 2.5 ng/ml TGF-␤1. (A) Suspended cells were then plated onto FN-coated dishes and incubated for 4 hours. They were then treated for 2 hours with the indicated concentrations of nocodazole, colchicine, cytochalasin D (Cyto D) or latrunculin A. Cells treated with vehicle only were set as control (0). Cell lysates were generated and processed for immunoblotting with anti-HEF1 pAb. Nitrocellulose membranes were stripped and reprobed with anti-p130Cas mAb as loading control. (B) Suspended cells were seeded onto FN-coated coverslips for 4 hours and treated for 2 hours with the indicated concentration of nocodazole, cytochalasin D or the same volume of vehicle as control. Cells were then triple-stained for F-actin, microtubules and vimentin. Actin was visualized using Texas-Red-conjugated phalloidin (red); microtubules were visualized by staining the cell layers with anti-␤-tubulin mAb followed by the Alexa Fluor-350-conjugated secondary antibody (blue); and intermediate filaments were stained with anti-vimentin pAb followed by Alexa Fluor-488conjugated secondary antibody (green). Adhesion regulates HEF1 degradation However, LAC had no effect on the cytochalasin-D-induced loss of p115HEF1 (Fig. 4, lane 4), suggesting that the cytochalasin-D-induced disappearance of p115 HEF1 resulted from the dephosphorylation of p115 HEF1 rather than the selective proteasomal degradation of p115 HEF1. These data indicate that the level of p115HEF1 within the cell might be regulated by an actin-cytoskeleton-dependent phosphatase.
HEF1 turnover under adhesion, suspension or cytochalasin-D-treatment conditions was evaluated to determine whether the conversion of p115HEF1 to p105HEF1 prolonged HEF1 halflife. Adherent cells in the presence or absence of cytochalasin D and suspended cells were incubated with cycloheximide to inhibit protein synthesis. HEF1 levels were detected by western blot and analyzed with densitometry (Fig. 5). After 8 hours of cycloheximide treatment, there was little detectable HEF1 in the control culture of adherent cells. By contrast, there was only a slight change in the levels of p130Cas protein, which served as a loading reference. Cells placed in suspension or treated with cytochalasin D showed a complete loss of p115HEF1 within 2 hours. Loss of p115HEF1 was associated with an increase in p105HEF1 levels, which, however, remained relatively constant during 8 hours of cycloheximide treatment (Fig. 5). In control experiments, these treatments did not affect degradation of Smad3, which has been shown to be degraded via the proteasomal pathway (Fukuchi et al., 2001), suggesting that suspension and cytochalasin D had no inhibitory effect on proteasome activity. Taken together, these data indicate that the conversion of p115 HEF1 to p105HEF1 prolongs the half life of HEF1 by removing HEF1 from proteosomal degradation pathways.
Protein PP2A-dependent dephosphorylation of p115HEF1 Preliminary experiments using 32 P-labeled myelin basic protein (MyBP) as a substrate indicated that PP1 and PP2A accounted for 95% of the total Ser/Thr-protein-phosphatase activity in human dermal fibroblasts (data not shown). Okadaic acid, calyculin A and tautomycin are potent inhibitors of PP1 and PP2A (Favre et al., 1997). To determine whether PP1 and/or PP2A are involved in the conversion of p115HEF1 to p105HEF1, selective inhibitors of each of these phosphatases were compared for their effects on the cytochalasin-D-induced loss of p115HEF1. As shown in Fig. 6, pretreatment of cells with either 10 nM calyculin A (panel A) or 200 nM okadaic acid (panel B) -both selective inhibitors of PP2A (Cohen, 1991) -was able to block the cytochalasin-D-induced dephosphorylation of p115HEF1. The dephosphorylation of p115HEF1 induced by cytochalasin D was unaffected by pretreatment of cells with 7.5 M tautomycin (Fig. 6C) -a selective inhibitor of PP1 (MacKintosh and Klumpp, 1990). These data suggest that PP2A regulates the cytochalasin-Dinduced dephosphorylation of p115HEF1. To confirm the inhibitory effects of the reagents on intracellular PP2A and PP1, the activities of PP2A and PP1 were measured in the extracts of cells that had been previously treated with various Fig. 4. Cytochalasin-D-induced conversion of p115HEF1 to p105HEF1 is not due to selective degradation of p115 HEF1. A1-F monolayers on FN-coated dishes were serum-starved overnight and pre-treated with TGF-␤1 (2.5 ng/ml) for 4 hours in the presence of 10 M lactacystin (LAC) or 50 M ALLM as indicated. 2 M cytochalasin D (CD) or equal volumes of carrier solvent DMSO (control) were then directly added into pre-treatment medium, and cells were further incubated for 6 hours. Cell lysates were generated and processed for immunoblotting with anti-HEF1 pAb. Nitrocellulose membranes were stripped and reprobed with anti-p130Cas mAb as loading control.

Fig. 5.
Cell suspension or cytochalasin D treatment inhibit HEF1 degradation. MG-63 cells on FN-coated dishes were serum-starved for 6 hours and then treated with TGF-␤1 (2.5 ng/ml) for 12 hours to induce HEF1 expression. TGF-␤1-containing medium was removed and cell layers were washed twice with serum-free medium. (A) Cell layers were then treated with 20 g/ml cycloheximide in the absence (adhesion) or presence (adhesion + CD) of 2 M cytochalasin D for indicated times, or detached and plated onto agar-coated dishes in the presence 20 g/ml cycloheximide (suspension). Times indicate hours after addition of cytochalasin D or placement in suspension. Aliquots of cell lysates containing equal amounts of protein were processed for immunoblotting with anti-HEF1 antibody. Membranes were stripped and reprobed for p130 Cas, which served as a loading control. (B) The levels of total HEF1 as shown in A were analyzed by densitometry and the time at 0 hours was set as 100%.
concentrations of inhibitor. As shown in Fig. 7A, the concentration of okadaic acid required to inhibit PP2A activity, exactly correlated with the concentration required to prevent the cytochalasin-D-induced p115HEF1 dephosphorylation. Treatment of A1-F cells with 200 nM okadaic acid, which abolished the effect of cytochalasin D on p115HEF1 dephosphorylation, resulted in almost complete inhibition of PP2A activity with little effect on PP1. The protein levels of PP1 and PP2A did not change under the cell-treatment condition (Fig. 7B). Treatment of A1-F cells with the PP1 inhibitor tautomycin induced a dose-dependent inhibition of PP1 activity but had less effect on PP2A (Fig. 7C). Although treatment of A1-F cells with 3 M tautomycin completely inactivated PP1, the dephosphorylation of p115HEF1 induced by cytochalasin D was not affected, even at 7.5 M tautomycin (Fig. 6C). The protein levels of PP1 and PP2A did not change when cells were treated with various concentrations Journal of Cell Science 119 (1) of tautomycin (Fig. 7D). These results indicate that PP2A but not PP1 is involved in the cytochalasin-D-induced dephosphorylation of p115HEF1. Taken together, these data indicate that p115HEF1 levels are regulated through a celladhesion-dependent actin-cytoskeleton organization, which regulates the activity of PP2A and that the PP2A-mediated conversion of p115HEF1 to p105HEF1 regulates the half-life of HEF1 in the cell.

Discussion
HEF1 is a Cas-family-member docking protein, which serves as a convergence point for signaling pathways that originate from integrins, growth factors and cytokines Zhang et al., 2000;Zhang et al., 1999;Zheng and McKeown-Longo, 2002). HEF1 undergoes complex posttranslational modifications, which direct its subcellular localization and promote its participation in diverse biological activities including apoptosis and motility (Fashena et al., 2002;Law et al., 1998;Law et al., 2000). HEF1 localizes to focal adhesions and has been identified as a downstream component of integrin signaling. Integrin ligation results in the tyrosine phosphorylation of HEF1, which occurs through the activity of integrin-dependent tyrosine kinases such as FAK, Src or PTK2B, also known as RAFTK Manie et al., 1997;Tachibana et al., 1997;Zheng and McKeown-Longo, 2002). These phosphorylation events are important for the docking function of HEF-1 because the phosphorylated form of HEF1 provides binding sites for the SH-2 domains of crk and zyxin family members Yi et al., 2002), which probably contribute to HEF1's role in lymphocyte migration (Ohashi et al., 1999). Ser/Thr phosphorylation of p105HEF1 results in the p115 isoform of HEF1 (Law et al., 1998;Zheng and McKeown-Longo, 2002). A distinct function for this isoform of HEF1 has been suggested based on studies, which show that p115HEF1 but not p105HEF1 forms complexes with Smad3. These complexes are cleared from the cell by proteosomal degradation, implicating HEF1 in modulating TGF-␤ signaling pathways (Liu et al., 2001). Half-life experiments indicate that there is a rapid turnover of a few hours of both the p105 and p115 species of endogenous HEF1. Treatment of either osteosarcoma or fibroblast cells with proteosome inhibitors results in a several-fold increase in the level of p115HEF1, consistent with this species of HEF1 being preferentially degraded by proteosomes (Liu et al., 2001). Our previous studies indicate that cell adhesion and cytoskeletal organization regulate the interconversion of p105 and p115 HEF1 (Zheng and McKeown-Longo, 2002). Reversible phosphorylation of proteins is an important mechanism for the regulation of cellular signal transduction (Hunter, 1995). Although the role of protein kinases in the regulation of adhesion-related signaling has been extensively studied (Howe et al., 1998;Schwartz and Assoian, 2001), the role of protein phosphatases, especially Ser/Thr phosphatases, in the transduction of cell-adhesion dependent signals is not well understood. We show here that the conversion of p115HEF to p105HEF is regulated by actin-microfilament organization because agents that disrupt intermediate filaments or microtubles did not affect the levels of either HEF isoform. The loss of p115HEF1 in response to cytochalasin D is not due to preferential degradation of p115HEF1 because proteosomal inhibitors had no effect on the conversion of p115 to p105HEF1. Treatment of cells with inhibitors of the Ser/Thr protein phosphatase PP2A, completely inhibited the conversion of p115 to p105HEF1 in response to either cell-detachment or treatment with cytochalasin D. PP2A is an abundant Ser/Thr protein phosphatase, which participates in pathways regulating cellcycle progression and apoptosis (Millward et al., 1999;Schonthal, 1998;Sontag, 2001). In this study, we demonstrate that PP2A-dependent dephosphorylation of p115HEF1 is induced by cell detachment or actin-disrupting drugs. Because p115HEF1 forms a complex with Smad3 to dampen TGF-␤ signal transduction, PP2A might be a positive regulator of TGF-␤ signaling pathways. A similar pathway has been described in HEK299 cells where cell detachment upregulates PP2A activity and attenuated IFN-␥ signaling (Ivaska et al., 2003). PP2A activity has been linked to specific integrin signaling pathways in T lymphocytes, where ␣2␤1-dependent activation of PP2A inhibits Fas-induced apoptosis (Gendron et al., 2003). The mechanism by which PP2A triggers dephosphorylation of p115HEF1 is not known. Placing cells in suspension did not result in an increase in the activity of PP2A (M.Z. and P.J.M.-L., unpublished observations). This suggests that PP2A exerts its effect on p115HEF1 following a change in the subcellular localization of either PP2A or p115HEF1. Correlations between the subcellular distribution of PP2A and changes in actin organization have been recently reported (Nakajima et al., 2004). The association of PP2A with ␤1 integrins is modulated during the differentiation of myoblasts to myotubes (Kim et al., 2004) and during cell mitosis. It is possible that the changes in actin organization that occur during mitosis and during muscle differentiation contribute to the disruption of ␤1/PP2A complexes. Adhesion of either fibroblasts or osteosarcoma cells to a range of extracellular matrices results in the conversion of p105 to p115 HEF. These results suggest that HEF1 is a downstream target of integrin dependent Ser/Thr kinases. Integrins are known to regulate the activity of several Ser/Thr kinases including MAP kinases, protein kinases B and C and integrinrelated kinases (Aplin et al., 2001;Miranti et al., 1999;Tian et al., 2002;Wu, 1999). Conversion of p105HEF1 to p115HEF1 following cell adhesion to various matrix molecules occurred over a period of 2 hours, consistent with the Ser/Thr phosphorylation of HEF1 being linked to the process of cell spreading. This therefore, suggests that the Ser/Thr kinase lies downstream of actin organization. Agents that disrupt microtubules and intermediate filaments did not affect the levels of either HEF1 isoform, indicating that actin organization is the primary effector of p105HEF1 Ser/Thr phosphorylation. The kinase responsible for the conversion of p105HEF1 to p115HEF1 has not yet been identified but kinases, such as AKT/PKB, that have been shown to be regulated by actin organization are possible candidates (Enomoto et al., 2005). Further studies are needed to elucidate the role of actin-dependent-kinases and -phosphatases in the regulation of HEF1 protein function.

Materials and Methods
Reagents TGF-␤1 was from R&D Systems. Cytochalasin D and phosphorylase b were from Sigma (St Louis, MO). Latrunculin A, nocodazole, colchicine, okadaic acid, calyculin A, tautomycin and proteasome inhibitors were from Biomol (Plymouth Meeting, PA). Anti-smad3 polyclonal antibody (pAb) was from Zymed Laboratories (South San Francisco, CA). Anti-protein phosphatase 2A monoclonal antibody (mAb) (Clone 1D6) was from Upstate Biotechnology (Lake Placid, NY). Anti-p130Cas mAb and anti-protein phosphatase 1 mAb (Clone 24) were from BD Transduction Laboratories (Lexington, KY). Anti-vimentin pAb and anti-␤-tubulin Fig. 7. PP2A, rather than PP1, is responsible for the dephosphorylation of p115HEF1 in human dermal fibroblasts. (A) The activities of PP1 and PP2A in the cell extracts obtained as described in Fig. 6 were determined by measuring the generation of free 32 P i from 32 P-phosphorylase-a and differentiated by their sensitivities to inhibitor-2 as described in Materials and Methods. The activity of control was set as 100%. The data represent the mean ± s.d. of duplicate assays from three separate experiments. (B) To ensure equal amounts of protein phosphatases in the sample shown in A, cell extracts containing the same amount of protein were processed for western blotting with anti-PP2A monoclonal antibody. Nitrocellulose membrane was stripped and reprobed with anti-PP1 monoclonal antibody. (C) The cell extracts were obtained as described in Fig. 6. The PP1 and PP2A activities in cell extracts were determined as described above, and expressed as percentage of control. The data represent the mean ± s.d. of triplicate assays from two separate experiments. (D) Cell extracts containing the same amount of protein were analyzed by western blotting with anti-PP1 monoclonal antibody. Nitrocellulose membrane was stripped and reprobed with anti-PP2A monoclonal antibody. mAb (Clone D10) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Protein phosphatase inhibitor-2 was from New England BioLabs (Beverly, MA). Dye-conjugated secondary antibodies were from Molecular Probes. Phosphorylase kinase was from ICN (Aurora, OH). Laminin was from Chemicon International, Inc (Temecula, CA). Type I collagen was from Cohesion Technologies (Palo Alto, CA). Human plasma fibronectin (FN) was purified from a fibronectin-and fibrinogenrich byproduct of factor VIII production by ion-exchange chromatography as previously described (McKeown-Longo and Etzler, 1987) and further purified by affinity chromatography with gelatin-agarose and subsequently with heparinagarose (Engvall and Ruoslahti, 1977). Anti-HEF1 antiserum was generated in rabbits with peptide [C]-EPPAETSEKANPQER linked to KLH as antigen (Zheng and McKeown-Longo, 2002). The specific IgG fraction was purified from antiserum by affinity chromatography with peptide-coupled thiopropyl-Sepharose 6B. Vitronectin was purified from human serum by heparin-Sepharose (Amersham Biosciences) affinity chromatography (Yatohgo et al., 1988).

Cell culture
Human foreskin fibroblasts (A1-F) were obtained from Lynn Allen-Hoffman (University of Wisconsin, Madison, WI) and used between passage 6 to 12. Human osteosarcoma, MG-63 cells were from ATCC. Both cell lines were maintained in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT) at 37°C in an atmosphere of 7.5% CO 2 . DMEM containing 0.1% heat-inactivated BSA was used as serum starvation medium.

Cell lysis and immunoblot analysis
Cell layers were washed with cold washing buffer (PBS containing 1 mM PMSF, 1 mM Na 3 VO 4 and 10 mM NaF) before solubilization in cell-lysis buffer (20 mM Tris-HCl, pH 7.4, 1% Triton X-100, 0.5% NP-40, 0.1 M NaCl, 40 mM NaF, 30 mM Na 4 P 2 O 7 , 2 mM EGTA, 1 mM Na 3 VO 4 , 0.5 mM PMSF and one tablet of completeprotease-inhibitor per 10 ml). After incubation on ice for 30 minutes, cell lysates were centrifuged at 20,800 g for 15 minutes at 4°C and the insoluble pellets were discarded. The protein concentration of the lysate was determined using a BCA protein assay reagent (Pierce, Rockford, IL). Aliquots of cell lysates containing equal amount of protein were subjected to SDS-PAGE on a 6.5-8%-gradient polyacrylamide gel under reducing conditions, followed by transfer to nitrocellulose membrane. Proteins were detected by western blot with a chemiluminescence reagent (Amersham Biosciences, Piscataway, NJ). Densitometric analysis was performed by using the Personal Densitometer SI and ImageQuant program (Molecular Dynamics, Sunnyvale, CA). The density-volume was converted to the amount of protein by comparison with a standard protein-density curve established by immunostaining of rabbit IgG standards.

Preparation of 32 P-labeled substrates
Myelin basic protein (MyBP) was phosphorylated with cAMP-dependent protein kinase and [␥-32 P]ATP according to the manufacturer's instructions (New England BioLabs, Beverly, MA). The conversion of phosphorylase-b to phosphorylase-a was accomplished by phosphorylase kinase and [␥ -32 P]ATP, according to the procedure described by McNall et al. (McNall et al., 1988).

Protein phosphatase assay
Cell extracts were prepared as previously described (Yan and Mumby, 1999). The activities of protein phosphatases in cell extracts were determined by measuring the release of 32 P i from 32 P-labeled substrates. Preliminary experiments indicated that the linear range of total phosphatase activity was between 0.01 g and 5 g of control-cell-extracts, with less than 20% substrate dephosphorylation. Thus, 2 g cell extract was used for phosphatase activity assays. The reaction mixture contained 50 mM Tris-HCl pH 7.0, 0.1 mM EDTA, 5 mM DTT, 0.01% Brij 35®, 2 g cell extract and 4 M 32 P-MyBP (a common substrate for Ser/Thr protein phosphatases) or 7.5 M 32 P-phosphorylase-a (a specific substrate for PP1 and PP2A) in a final volume of 50 l. Reactions were carried out in a 30°C water bath for 15 minutes and processed as previously described (McNall et al., 1988). Since 1 M okadaic acid inhibits 99% of PP1 and 100% of PP2A, the combined total activitiy of PP1 and PP2A was defined as the activity sensitive to 1 M okadaic acid (Cohen, 1991). PP1 activity was defined as the activity sensitive to 0.4 M inhibitor-2, an inhibitor specific to PP1 (Cohen, 1991). PP2A activity was obtained by subtracting PP1 activity from the total activities of both PP1 and PP2A.

Fluorescence microscopy
Cell layers on glass coverslips were fixed at 37°C for 15 minutes in 3.7% paraformaldehyde, permeabilized with 0.2% Triton X-100 for 5 minutes and blocked with 3% BSA at 4°C overnight. The cell layers were incubated at room temperature for 3 hours with primary antibody diluted in blocking buffer, followed by 1 hour with fluorescence-conjugated secondary antibody. After staining, the coverslips were mounted with Prolong Antifade according to manufacturer's instructions (Molecular Probes, Eugene, OR). The cell layers were examined using an Olympus BMX-60 microscope equipped with a cooled CCD sensi-camera (Cooke, Auburn Hills, MI). Images were acquired using Slidebook software (Intelligent Imaging Innovation, Denver, CO) and processed with Photoshop-7 program (Adobe). This work was supported by grant CA-69612 from the National Institutes of Health.