3-Phosphoinositide-dependent PDK1 Negatively Regulates Transforming Growth Factor-β-induced Signaling in a Kinase-dependent Manner through Physical Interaction with Smad Proteins*

We have reported previously that PDK1 physically interacts with STRAP, a transforming growth factor-β (TGF-β) receptor-interacting protein, and enhances STRAP-induced inhibition of TGF-β signaling. In this study we show that PDK1 coimmunoprecipitates with Smad proteins, including Smad2, Smad3, Smad4, and Smad7, and that this association is mediated by the pleckstrin homology domain of PDK1. The association between PDK1 and Smad proteins is increased by insulin treatment but decreased by TGF-β treatment. Analysis of the interacting proteins shows that Smad proteins enhance PDK1 kinase activity by removing 14-3-3, a negative regulator of PDK1, from the PDK1-14-3-3 complex. Knockdown of endogenous Smad proteins, including Smad3 and Smad7, by transfection with small interfering RNA produced the opposite trend and decreased PDK1 activity, protein kinase B/Akt phosphorylation, and Bad phosphorylation. Moreover, coexpression of Smad proteins and wild-type PDK1 inhibits TGF-β-induced transcription, as well as TGF-β-mediated biological functions, such as apoptosis and cell growth arrest. Inhibition was dose-dependent on PDK1, but no inhibition was observed in the presence of an inactive kinase-dead PDK1 mutant. In addition, confocal microscopy showed that wild-type PDK1 prevents translocation of Smad3 and Smad4 from the cytoplasm to the nucleus, as well as the redistribution of Smad7 from the nucleus to the cytoplasm in response to TGF-β. Taken together, our results suggest that PDK1 negatively regulates TGF-β-mediated signaling in a PDK1 kinase-dependent manner via a direct physical interaction with Smad proteins and that Smad proteins can act as potential positive regulators of PDK1.

Transforming growth factor-␤ (TGF-␤) 2 plays a critical role in the modulation of a wide variety of biological and developmental processes (1,2). The diverse cellular responses elicited by TGF-␤ are triggered by activation of TGF-␤ receptors (type I and II), which are serine/threonine kinases. TGF-␤ receptors subsequently propagate signals through phosphorylation of intracellular signaling mediators referred to as Smads (3)(4)(5). There are three functional classes of Smad proteins (6), the receptor-regulated Smads (R-Smads), the common Smads (Co-Smads), and the inhibitory Smads (I-Smads). The R-Smads are directly phosphorylated and activated by the type I TGF-␤ receptor and undergo homotrimerization and heterodimerization with a Co-Smad (Smad 4). The activated heteromeric Smad complexes are translocated into the nucleus and cooperate with other nuclear cofactors to regulate the transcription of target genes (7). Smad-mediated signaling may be simple but it is under the control of a number of Smad-interacting proteins. Several lines of evidence have demonstrated the existence of cellular Smad regulators that interact with Smads to control the subcellular localization and the rate of R-Smad association with the TGF-␤ receptor and subsequent phosphorylation at the plasma membrane or in the cytoplasm or nucleus. Several of these Smad regulators have been identified, including the FYVE domain protein SARA (8), microtubules (9), Daxx (10), the truncated receptor-like molecule BAMBI (11), the ubiquitin ligase Smurf1 (12), the integral inner nuclear membrane protein MAN1 (13), and I-Smads (14 -16), Smad6 and Smad7. Thus, identification and characterization of additional Smad-interacting molecules should provide greater insight into the regulation of Smad-mediated signaling. In addition, growth factor-and insulin-mediated signaling pathways modulate TGF-␤ signaling through a physical interaction between PKB/Akt and Smad3 (17), suggesting a possible cross-talk between TGF-␤and PI3K/PDK1-mediated signaling pathways.
The 3-phosphoinositide-dependent protein kinase-1 (PDK1) is a member of the protein kinase A, G, and C subfamily of protein kinases with a PH domain that binds phosphoinositides such as PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 for its activity and phosphorylates Thr-308 of PKB/Akt. Phosphorylation on both Thr-308 and Ser-473 is required for maximal activation of PKB/ Akt (18 -20). Furthermore, these residues are independently phosphorylated by PDK1 (for Thr-308) and PDK2 (for Ser-473). Emerging evidence indicates that PDK1 kinase activity is controlled by several cellular proteins that interact with PDK1, including Hsp90 (21), 14-3-3 (22), protein kinase C-related kinase 2 (23), and STRAP (24). These observations strongly suggest that the PDK1-interacting proteins can regulate PDK1 activity. In this study, we show that there are direct physical and functional interactions between PDK1 and Smad proteins (Smad2, -3, -4, and -7) and that these interactions may play an important role in the regulation of both the PDK1 and Smad activities involved in PI3K/PDK1-and TGF-␤-mediated signaling.
Transient Transfection, in Vivo Interaction Assay, and Western Blot Analysis-Cells were transfected with appropriate plasmids using WelFect-Ex TM Plus (WelGENE, Daegu, Korea), according to the manufacturer's instructions. After culturing overnight, the transfected cells were incubated in the presence or absence of TGF-␤1 (100 pM) for 20 h. Cells were then washed and solubilized with lysis buffer containing 0.1% Nonidet P-40 as described (27). Detergent-insoluble materials were removed by centrifugation, and the cleared lysates were incubated with glutathione-Sepharose beads (Amersham Biosciences) and then washed three times with the lysis buffer. For Western blotting, coprecipitates or whole cell extracts were resolved by SDS-PAGE. For immunoprecipitations, cell lysates were incubated with protein A-Sepharose that had been conjugated to the appropriate antibodies (anti-Myc, anti-PDK1, anti-Smad3, anti-Smad4, anti-Akt, and anti-Bad). The immunoprecipitated proteins were electrophoresed and blotted onto polyvinylidene difluoride membranes. The membranes were immunoblotted with the indicated antibodies and then developed using an ECL detection system according to the manufacturer's instructions (Amersham Biosciences).
PDK1 Kinase Assay-To estimate PDK1-dependent serum/ glucocorticoid-regulated kinase (SGK) phosphorylation in vitro, 293T cells transiently transfected with the appropriate plasmids were washed three times with ice-cold phosphatebuffered saline (PBS) and solubilized with 100 l of lysis buffer (20 mM Hepes, pH 7.9, 10 mM EDTA, 0.1 M KCl, and 0.3 M NaCl). The cleared lysates were mixed with glutathione-Sepharose beads and rotated for 2 h at 4°C. After washing the precipitate three times with lysis buffer, and then twice with kinase buffer (50 mM Hepes, pH 7.4, 1 mM dithiothreitol, and 10 mM MgCl 2 ), the precipitate was incubated with 5 Ci of [␥-32 P]ATP at 37°C for 15 min in the presence of kinase buffer containing 500 ng of recombinant SGK (Upstate). The reactions were separated by electrophoresis and visualized by autoradiography.
Luciferase Reporter Assay-HepG2 cells were transfected using WelFect-Ex TM Plus with the p3TP-Lux or p21-Luc reporter plasmids, along with each expression vector as indicated. The cells were harvested 48 h post-transfection, and luciferase activity was measured using the Promega dual luciferase assay kit according to the manufacturer's instructions. Light emission was determined with a VICTOR TM luminometer (1420 luminescence counter, PerkinElmer Life Sciences). The total DNA concentration was kept constant by supplementing with empty vector DNA. The data were normalized to the expression levels of a cotransfected ␤-galactosidase reporter control, and experiments were repeated at least four times.
Cell Death Assay-The number of HeLa or HaCaT cells undergoing apoptosis after treatment with TGF-␤1 (HeLa, 10 ng/ml for 20 h; HaCaT, 2 ng/ml for 20 h) was quantified using the GFP system, as described previously (24). Cells grown on sterile coverslips were transfected with pEGFP, an expression vector encoding GFP, together with the indicated expression vectors. The cells were treated with TGF-␤1, at 24 h post-transfection. The cells were fixed with ice-cold 100% methanol, washed three times with PBS, and then stained with a bisbenzimide (Hoechst 33258). The coverslips were washed with PBS, then mounted on glass slides using Gelvatol, and visualized using a fluorescence microscope (Leica DM IRB, Germany). The percentage of apoptotic cells was calculated as the number of GFP-positive cells with apoptotic nuclei divided by the total number of GFP-positive cells.
Preparation of Recombinant Proteins-Recombinant glutathione S-transferase (GST) fusion vectors containing Smad3 and Smad4 were constructed by subcloning the cDNA fragments of Smad3 and Smad4 into pGEX4T-1 (Amersham Biosciences) and purified by affinity chromatography on glutathione-Sepharose 4B columns (Amersham Biosciences) as described previously (25).
FACS Analysis-HaCaT cells (2 ϫ 10 5 /60-mm dish) transfected with the indicated combinations of plasmid vectors (empty vector, PDK1, Smad3, and Smad7) and siRNA duplexes (Smad3, Smad7, PDK1, and control siRNAs) were washed with ice-cold PBS and then synchronized in G 0 /G 1 by treating with hydroxyurea (2 mM) for 20 h. The fraction of cells in each stage of the cell cycle was analyzed after 10% serum treatment for 24 h in the presence or absence of TGF-␤1 (2 ng/ml). Trypsinized cells were washed twice with ice-cold PBS and incubated at 37°C for 30 min with a solution (1 mM Tris-HCl, pH 7.5) containing 50 g/ml propidium iodide and 1 mg/ml RNase A. The cells in each phase of the cell cycle were identified using the Mod-FitLT version 3.0 (PMac) program. Flow cytometry analysis was performed using FACSCalibur-S system (BD Biosciences).
Indirect Immunofluorescence-Hep3B cells were plated and transfected with FLAG-Smads (Smad3, Smad4, and Smad7) and/or Myc-tagged wild-type and kinase-dead PDK1 constructs on sterile coverslips, placed on ice, and washed three times with ice-cold PBS prior to fixation with 4% paraformaldehyde for 10 min at room temperature. Cells were then washed with PBS, treated with 0.2% Triton X-100, and rewashed with PBS. The cells were incubated with mouse anti-FLAG (M2), diluted 1:1000 in PBS, or rabbit anti-Myc, diluted 1:200 in PBS, for 2 h at 37°C. The cells were then washed three times with PBS and incubated with Alexa Fluor-594 anti-mouse or Alexa Fluor-488 anti-rabbit secondary antibodies, diluted 1:1000 in PBS, at 37°C for 1 h. The coverslips were washed three times with PBS and then mounted on glass slides using Gelvatol. Proteins were visualized using a Leica Dmire2 confocal microscopy (Germany).

RESULTS
Identification of PDK1 as a Smad-interacting Protein-We have found previously that STRAP, a TGF-␤ receptor interacting protein, physically interacts with PDK1 in mammalian cells (24). In addition, STRAP inhibits TGF-␤ signaling by stabilizing the TGF-␤ receptor-Smad7 complex, and STRAP itself binds to Smad proteins such as Smad2, Smad3, and Smad7 (29). Based on these data, we reasoned that PDK1 might interact with Smad proteins, as well as with STRAP, in intact cells. To examine whether PDK1 directly binds to Smad proteins, we performed in vivo binding assays and coimmunoprecipitation experiments using overexpressed or endogenous proteins in 293T cells. The interaction of FLAG-tagged Smad proteins with a Myc-PDK1 fusion protein was analyzed by immunoprecipitation with an anti-Myc antibody, followed by immunoblotting with an anti-FLAG antibody. Smad2, -3, -4, and -7 were detected in the immunoprecipitate when coexpressed with Myc-PDK1 (Fig.  1A), indicating that PDK1 physically interacts with Smad proteins in cells. To confirm the interaction of PDK1 with Smad proteins in vivo, we next performed coimmunoprecipitation experiments with endogenous PDK1 and exogenous FLAGtagged Smad proteins (Fig. 1B). Endogenous PDK1 was immunoprecipitated with an anti-PDK1 antibody from cell lysates, and the binding of Smad proteins was subsequently analyzed by immunoblotting with an anti-FLAG antibody. Smad proteins were present in the PDK1 immunoprecipitate (upper panel), but not in immunoprecipitates from control lysates of cells transfected with empty vector alone (CMV). Moreover, to examine the interaction between the two endogenous proteins, immunoprecipitation of endogenous PDK1 using an anti-PDK1 antibody was performed, and the binding of the endogenous Smad proteins (Smad2, -3, and -7) was subsequently analyzed by immunoblotting with the indicated anti-Smad antibodies. As shown in Fig. 1C, endogenous PDK1 physically interacted with the endogenous Smad proteins used in 293T cells. We have further analyzed this association using other cell lines, including Hep3B cells and SK-N-BE2C cells (27), a human neuroblastoma line, and we confirmed that this association could occur in vivo (data not shown). To determine whether the Smad proteins can be substrates for PDK1 in vitro, recombinant Smad proteins (re.Smad3 and -4) were expressed in Escherichia coli and purified and then used as substrates in a PDK1 kinase assay. Extracts from 293T cells expressing GST-PDK1 and FLAG-STRAP were purified with glutathione-Sepharose beads, and incubated with [␥-32 P]ATP to allow phosphorylation of the recombinant Smad proteins. As shown in Fig. 1D, the Smad proteins were phosphorylated by PDK1 when the in vitro kinase assays were performed using re.Smad3 and -4 as substrates. However, phosphorylation of the recombinant Smad proteins was not detected in the absence of PDK1 (data not shown). In addition, we observed that the coexpres-sion of STRAP, a potential positive regulator of PDK1 (24), significantly increased the phosphorylation of Smad proteins by PDK1 (Fig. 1D, 2nd versus 3rd lane and 4th versus 5th lane). Similar results showing that the Smad proteins can be phosphorylated by PDK1 were also observed in vivo using cells coexpressing PDK1 and Smad2, -3, -4, or -7 instead of the recombinant Smad proteins (data not shown). To further confirm that the phosphorylation of Smad proteins by PDK1 occurs in vivo, FIGURE 1. Interaction between PDK1 and Smad proteins in vivo. A, vector alone (cytomegalovirus (CMV)) or FLAG-tagged Smads (Smad2, -3, -4, and -7) were cotransfected with Myc-tagged PDK1 (Myc-PDK1) into 293T cells. Myc fusion proteins were immunoprecipitated with an anti-Myc antibody (IP:␣-Myc), and complex formation between PDK1 and Smad proteins (top panel) and the amount of immunoprecipitated PDK1 (middle panel) were determined by immunoblot analyses using anti-FLAG and anti-Myc antibodies, respectively. The expression levels of FLAG-Smads in total cell lysates were analyzed by Western blot (WB) analysis using anti-FLAG antibody (bottom panel, Lysate). B, 293T cells were transiently transfected with a vector alone (cytomegalovirus (CMV)), as a negative control, or FLAG-tagged Smads (Smad2, -3, -4, and -7), lysed, and immunoprecipitated with an anti-PDK1 antibody (IP:␣-PDK1). The PDK1 immunoprecipitate was analyzed for the presence of Smad proteins by Western blot using an anti-FLAG antibody (upper panel). The expression levels of FLAG-tagged Smads in total cell lysates were analyzed by Western blot using the anti-FLAG antibody (lower panel, Lysate). C, cell lysates from parental 293T cells were immunoprecipitated with rabbit preimmune serum (preimm.) or rabbit anti-PDK1 antibody (␣-PDK1) and blotted as indicated. D, for an in vitro kinase assay, ϳ3-4 g of recombinant Smad proteins (re.Smad3 and 4) was mixed with 10 M ATP, 5 Ci of [␥-32 P]ATP, and 10 mM MgCl 2 in 20 l of kinase buffer and incubated with the coprecipitated PDK1 for 15 min at 37°C with frequent gentle mixing. E, 293T cells were transfected with either Myc-tagged WT PDK1, Myc-tagged KD PDK1, or a PDK1-specific siRNA. Lysates were immunoprecipitated with the indicated anti-Smad antibodies and probed with an anti-phospho-Ser/Thr antibody. Equal levels of the Smad3/4 proteins were precipitated (2nd and 4th panels). The down-regulation of endogenous PDK1 by PDK1-specific siRNA was confirmed by anti-PDK1 immunoblotting (right, bottom panel). Ϫ indicates transfection with the nonspecific control siRNA. The circled p-Smad3 and circled p-Smad4 indicate the position of phosphorylated Smad3 and Smad4, respectively. These experiments were independently performed at least four times with similar results.
we compared PDK1-mediated phosphorylation of Smad3/4 in cells expressing wild-type (WT) PDK1 or a kinase-dead (KD) PDK1 mutant or in the presence of a PDK1 siRNA. Expression of wild-type PDK1 produced a higher level of Smad3/4 phosphorylation, compared with cells expressing either kinase-dead PDK1 or a PDK1-specific siRNA (Fig. 1E). Taken together, our results indicate that PDK1 directly interacts with Smad proteins in vivo and that Smad proteins can be substrates for PDK1.
Mapping of the PDK1 and Smad Protein Domains Involved in the PDK1-Smad Complex Formation-To establish which regions of PDK1 are necessary for association with the Smad proteins, we generated two PDK1 deletion constructs FLAG-PDK1(PH), comprising the carboxyl-terminal pleckstrin homology (PH) domain (amino acids 411-556), and FLAG-PDK1(CA), harboring the catalytic domain (amino acids 67-359), as described previously (24), and we examined whether these constructs were able to interact with Smad proteins. Wild-type FLAG-PDK1 and FLAG-PDK1(PH), which lacks the catalytic domain of PDK1, interacted with Smad2, -3, -4, and -7 when the proteins were coexpressed in 293T cells ( Fig. 2A). However, FLAG-PDK1(CA), which contains only the catalytic domain, was unable to do so ( Fig. 2A, top panel), indicating that the interaction with Smad proteins is mediated via the carboxyl-terminal PH domain of PDK1. Next, to examine which region of Smad3 was required for binding of PDK1 in vivo, we generated four FLAG-tagged Smad3 deletion constructs (Fig. 2B, upper panel). The FLAG-MH1 (amino acids 1-136), FLAG-MH1(L) (amino acids 1-231), FLAG-MH2 (amino acids 231-425), and FLAG-MH2(L) (amino acids 136 -425) constructs were expressed in 293T cells and used for in vivo binding assays with GST-PDK1. The binding of PDK1 to the Smad3 deletion constructs (MH1(L), MH2, and MH2(L)) was readily detectable (Fig. 2B, lower top panel). However, PDK1 binding to the MH1 construct was not detected (Fig. 2B). These results suggest that the PDK1 PH domain binds to the Smad3 Linker/MH2 region.
PDK1-Smads Complex Formation Is Regulated by Insulin and TGF-␤-We assessed whether insulin and TGF-␤ can influence PDK1-Smads complex formation in cells following insulin and TGF-␤ treatment because PDK1 and Smad proteins are important intracellular mediators of insulin-and TGF-␤signaling pathways. Upon TGF-␤ treatment, the association between PDK1 and Smad2, -3, -4, and -7 was considerably decreased compared with control 293T cells not treated with TGF-␤1 (Fig. 3A, top panel, 5th, 7th, 9th, and 11th lanes versus  6th, 8th, 10th, and 12th lanes). However, exposure of 293T cells to insulin resulted in an increase in the PDK1-Smads complex formation, but this effect was inhibited by wortmannin, a PI3K inhibitor (Fig. 3B, top panel, 3rd, 6th, 9th, and 12th lanes versus 4th, 7th, 10th, and 13th lanes). These data demonstrate that the interaction between PDK1 and Smad proteins appears to be dependent on stimulation by insulin or TGF-␤, similar to our previous observations that the physical association between PDK1 and STRAP, a TGF-␤ receptor interacting protein, could be modulated by insulin or TGF-␤ stimulation (24). Given the effect of insulin on the interaction between PDK1 and Smad proteins, we conducted further assays for its biological importance of insulin in PDK1-mediated phosphorylation of Smad3/4, using a PDK1-specific siRNA. Smad3/4 phosphorylation was increased by insulin treatment (Fig. 3C, top panel, 1st versus 2nd lane), but knockdown of PDK1 decreased Smad3/4 phosphorylation in the presence of insulin (Fig. 3C, top panel, 2nd versus 3rd lane). Taken together, these data suggest that cross-talk between the PI3K/PDK1 and TGF-␤ signaling pathways occurs through Smad proteins in mammalian cells.
Smad Proteins Positively Regulate PDK1 Activity-We next examined whether the interaction between Smad proteins and . At 48 h post-transfection, the cells were incubated for 30 min with or without 100 nM wortmannin and then treated with 100 nM insulin for 20 min. The cell lysates were subjected to precipitation with glutathione-Sepharose beads (GST purification). The resulting precipitates were examined by immunoblot analysis with an anti-FLAG antibody to identify complex formation between PDK1 and Smad proteins (top panel). Equivalent amounts of GST-PDK1 were precipitated, as assessed by immunoblot analysis with an anti-GST antibody (middle panel). Expression levels of FLAG-tagged Smads were confirmed by Western blot analysis of total cell extracts using the anti-FLAG antibody (bottom panel). C, decrease in Smad3/4 phosphorylation in response to a PDK1-specific siRNA. 293T cells were transfected with the indicated siRNA duplexes (PDK1 siRNA(a) and control siRNA). At 48 h post-transfection, the cells were incubated for 20 min with 100 nM insulin. Cell lysates were immunoprecipitated (IP) with the indicated Smad antibodies (␣-Smad3 and ␣-Smad4) and blotted as indicated. Expression levels of endogenous PDK1 and ␤-actin proteins in the total cell lysates were analyzed by Western blot analysis using anti-PDK1 and anti-␤-actin antibodies (3rd and bottom panels, Lysate). These experiments were independently performed at least four times with similar results. Con, control.  (Smad2, -3, and -7) or with the positive control STRAP (24). Cell lysates were subjected to precipitation with glutathione-Sepharose beads and then the precipitates were analyzed for PDK1 kinase activity using an in vitro kinase assay with SGK as a substrate (top panel). The same blot was stripped and re-probed with the indicated antibodies to determine the expression level of precipitated PDK1 (2nd panel) and to show that equivalent amounts of substrate (His-SGK) were used in the kinase assays (3rd panel). The presence of Smad proteins and STRAP in total cell lysates was analyzed by Western blot (WB) analysis using the anti-FLAG antibody (bottom panel, Lysate). The circled p-SGK indicates the position of the phosphorylated SGK. B and C, PDK1-mediated phosphorylation of PKB/Akt and Bad. GST-Akt or GST-Bad was transiently cotransfected with Myc-PDK1 in the presence or absence of FLAG-tagged Smads (Smad2, -3, and -7). As a negative control, 293T cells were transfected with GST-Akt or GST-Bad alone. Transfected cells were precipitated with glutathione-Sepharose beads (GST purification), and the level of GST-Akt or GST-Bad phosphorylation was measured by immunoblot analysis using an anti-phospho-Thr 308-specific Akt antibody or an anti-phospho-Ser-136-specific Bad antibody (top panels). The anti-GST immunoblot for Akt or Bad (2nd panels) was prepared from the same blot. The expression levels of PDK1 and Smads in total cell lysates were analyzed by Western blot analysis using anti-Myc and anti-FLAG antibodies, respectively (3rd and bottom panels). D, effect of Smad siRNA duplexes on PDK1 kinase activity, PKB/Akt phosphorylation, and Bad phosphorylation. 293T cells were transfected with Smad siRNA duplexes (Smad3 siRNAs (a and b) and Smad7 siRNAs (a and b)). Total cell lysates were immunoprecipitated (IP) with anti-PDK1 antibody. The PDK1 immunoprecipitate was analyzed for PDK1 activity using an in vitro kinase assay with SGK as a substrate (top panel). The amounts of immunoprecipitated PDK1 and SGK in the assay were analyzed with anti-PDK1 and anti-His antibodies, respectively (2nd and 3rd panels). Total cell lysates transfected with Smad siRNA duplexes were immunoprecipitated with the indicated antibodies (IP:␣-Akt, IP:␣-Bad). The PKB/Akt and Bad immunoprecipitates were analyzed for Akt and Bad phosphorylation by immunoblot analysis using anti-phospho-Thr-308-specific Akt and anti-phospho-Ser-136-specific Bad antibodies, respectively (4th and 6th panels). The amounts of immunoprecipitated Akt and Bad were analyzed with anti-Akt and anti-Bad antibodies using the same blot (5th and 7th panels). As a control, the expression levels of endogenous Smad3 and Smad7 in the total cell lysates were analyzed by Western blot analyses using the indicated antibodies (8th panel). The expression level of endogenous ␤-actin was determined by anti-␤-actin immunoblotting (bottom panel). The circled p-SGK, circled p-Akt, and circled p-Bad indicate the position of the phosphorylated SGK, Akt, and Bad. These experiments were independently performed at least four times with similar results. PDK1 contributes to the ability of Smad proteins to modulate PDK1 kinase activity. First, PDK1 was precipitated from transfected 293T cells using FLAG-tagged Smad proteins (Smad2, -3, and -7) and STRAP (24) as a positive control, and PDK1 kinase activity was monitored by an in vitro kinase assay using SGK as a substrate, as described previously (30). Coexpression of Smad proteins with PDK1 resulted in a significant increase in PDK1 kinase activity (Fig. 4A, top panel, 1st lane versus 2nd to  5th lanes). As a control, the expression level of the precipitated PDK1 was analyzed in GST pulldown precipitates, and the amount of PDK1 in all lanes was similar (Fig. 4A, 2nd panel), indicating that the observed differences in phosphorylated SGK were not because of differences in the PDK1 expression levels in the cells. PKB/Akt, a downstream target of PDK1, has been implicated in contributing to the sequestration of Bad from the pro-apoptotic signaling pathway by Bad phosphorylation (31). To examine whether downstream targets of PDK1, such as PKB/Akt and Bad, are also affected by coexpression of Smad2, -3, and -7, we monitored PKB/Akt phosphorylation and Bad phosphorylation in cells expressing PDK1 and Smad proteins, or PDK1 alone, using immunoblotting. Coexpression of Smad proteins with PDK1 significantly induced PKB/Akt phosphorylation compared with PDK1 expression alone (Fig. 4B, top  panel, 2nd lane versus 3rd to 5th lanes). In addition, the PKB/ Akt activation induced by Smad proteins also increased Bad phosphorylation (Fig. 4C, top panel, 2nd lane versus 3rd to 5th  lanes). To confirm the physiological role of Smad proteins in regulation of the PI3K/PDK1 signaling pathway, PDK1 kinase activity, PKB/Akt phosphorylation, and Bad phosphorylation were determined in 293T cells transfected with Smad-specific siRNAs (Smad3 and Smad7 siRNAs) using SGK as a substrate or anti-phospho-antibodies, as indicated. Reducing the amount of endogenous Smad proteins in cells with sequence-specific siRNAs resulted in a dose-dependent decrease in PDK1 kinase activity (Fig. 4D, top panel), PKB/Akt phosphorylation (Fig. 4D, 4th panel), and Bad phosphorylation (Fig. 4D, 6th panel). As a control, the down-regulation of endogenous Smad proteins (Smad3 and -7) and ␤-actin was monitored by immunoblotting (Fig. 4D, 8th and bottom panels). Taken together, these findings suggest that Smad proteins positively regulate the PI3K/PDK1 signaling pathway through direct interaction with PDK1.

JOURNAL OF BIOLOGICAL CHEMISTRY 12279
presence or absence of Smad2, -3, -4, and -7. Cell lysates were precipitated with glutathione-Sepharose beads, and the binding of 14-3-3 to PDK1 was monitored by immunoblot analysis using the anti-FLAG antibody. Coexpression of Smad proteins significantly increased the dissociation of 14-3-3 from the PDK1-14-3-3 complex, compared with control cells that were not transfected with Smad proteins (Fig. 5A, top  panel, 2nd lane versus 3rd to 6th  lanes). As a control, the expression levels of PDK1 and 14-3-3 were determined, and the amount of these proteins in all lanes was similar (Fig. 5A, 2nd and bottom panels), indicating that the change in binding between PDK1 and 14-3-3 was not because of differences in PDK1 and 14-3-3 expression. We then performed RNA interference to determine the physiological role of Smad proteins in the regulation of PDK1-14-3-3 complex formation. 293T cells were transfected with control or Smad siRNAs, and the immunoprecipitation of endogenous PDK1 using an anti-PDK1 antibody was performed, and the binding of endogenous 14-3-3 to PDK1 was analyzed by immunoblotting with the anti-14-3-3 antibody. The interaction between endogenous PDK1 and 14-3-3 was significantly enhanced by Smad siRNAs compared with the control siRNA (Fig. 5B, top panel, 1st lane versus 2nd to 5th lanes). As a control, the down-regulation of endogenous Smad2, -3, -4, and -7 was monitored by immunoblotting (Fig.  5B, bottom panel). These results suggest that Smad proteins enhance PDK1 kinase activity by stimulating the dissociation of 14-3-3, a known negative regulator of PDK1, from the PDK1-14-3-3 complex.

Smad Proteins Enhance PDK1-mediated Stimulation of Cell Growth-
We have shown that the coexpression of Smad2, -3, and -7 stimulates aspects of PI3K/PDK1 signaling, including PDK1 kinase activation, PKB/Akt phosphorylation, and Bad phosphorylation (see Fig. 4). We therefore extended our analysis to investigate whether Smad proteins can stimulate serum-induced cell growth, which is one of the most important biological functions of PI3K/PDK1 signaling. We used flow cytometry analysis using HaCaT cells to monitor the percentage of cells in S phase of the cell cycle (33). As shown in Fig. 6A, HaCaT cells coexpressing PDK1 and Smad3 (or Smad7) significantly increased the percentage of cells in S phase compared with control cells expressing PDK1 alone (70 versus 54%). However, Smad proteins alone did not change the accumulation of S phase cells (ϳ42%, Smad3 and Smad7) compared with controls (ϳ45%, parental HaCaT cells (Ϫ) and vector), indicating that the increase in S phase cells when PDK1 and Smad proteins were coexpressed (ϩ PDK1, Smad3 and Smad7) is because of PDK1 activation by Smad proteins, probably through a direct interaction. To further confirm the involvement of Smad proteins in the stimulation of PDK1-mediated cell growth, flow cytometry analysis of HaCaT cells transfected with Smad3-or Smad7-specific siRNAs was performed. Reducing the amount of endogenous Smad3 or Smad7 by sequencespecific siRNAs in the presence of PDK1 resulted in a considerable decrease in the percentage of S phase cells compared with control cells expressing PDK1 alone (Fig. 6B, 49 -52% versus 42%). However, we did not observe a reduction in the percentage of S phase cells by the knockdown of endogenous Smad3 or Smad7 when endogenous PDK1 was also knocked down by siRNA treatment (Fig. 6C). Together, these results suggest that Smad proteins, such as Smad3 and Smad7, that directly interact with PDK1 play an important role in the modulation of PDK1-mediated cell growth.
PDK1 Inhibits TGF-␤-induced Transcription-Because PKB/ Akt was shown to form a complex with Smad3 and inhibit TGF-␤-induced transcription (17,34), we next tested whether PDK1, an upstream target of PKB/Akt, can also regulate TGF-␤-induced transcription. To examine the effect of increasing amounts of PDK1 on TGF-␤-induced transcription, we cotransfected HepG2 cells with PDK1 and Smad3 (or Smad7), together with the p3TP-Lux reporter plasmid containing elements from the PAI-I promoter (35) or with the p21-Luc reporter plasmid, in the presence or absence of TGF-␤1. The addition of PDK1 negatively regulated the Smad3-or Smad7induced transcription in a dose-dependent manner, suggesting that PDK1, like PKB/Akt, inhibits the TGF-␤-induced transcriptional activation (Fig. 7, A and B). We also tested whether the activity of PDK1 affects Smad3-or Smad7-induced transcription. HepG2 cells were cotransfected with wild-type PDK1 or the catalytically inactive kinase-dead PDK1 mutant, together with the p3TP-Lux reporter (Fig. 7, A and B, left panels) or the p21-Luc reporter (right panels), and luciferase assays were carried out as indicated. The wild-type PDK1 negatively regulated Smad3-and Smad7-induced transcription, in a dose-dependent manner, whereas coexpression of catalytically inactive kinase-dead PDK1 had little effect on Smad3-or Smad7-induced transcription (Fig. 7, A and B, black bars versus white  bars). These data indicate that the negative regulation of Smad3-or Smad7-induced transcription by PDK1 is dependent on its kinase activity. To examine the possibility that the effect of the inactive kinase-dead PDK1 was because of the lack of a direct physical interaction between the kinase-dead PDK1 and the Smad proteins, we carried out cotransfection experiments using GST-Smad proteins (Smad2, -3, -4, and -7) and Myc-PDK1 constructs (WT and KD). Both wild-type and kinasedead PDK1 proteins associate with Smad proteins to a similar extent (Fig. 7C), suggesting that, in addition to a direct physical interaction with Smad proteins, the kinase activity of PDK1 is necessary for the negative regulation of Smad-induced transcription. To further analyze the negative role of PDK1 in TGF-␤ signaling, we examined the effect of PDK1 siRNAs on TGF-␤-mediated gene responses in HeLa and HaCaT cells. The transfection of PDK1 siRNAs (a and b) resulted in up-regulation of TGF-␤ targets, including plasminogen activator inhibitor-1 (PAI-1), a cyclin-dependent kinase inhibitor p21 Cip1 , and Smad7, as well as down-regulation of CDK4 and cyclin D1, which are involved in TGF-␤-induced G 1 arrest (Fig. 7D). Taken together, these findings clearly suggest that PDK1 physically associates with Smad proteins and negatively regulates Smad-induced transcription.

PDK1 Modulates the Association between the Type I TGF-␤ Receptor and Smad
Proteins-To explore how PDK1 cooperates with Smad proteins in the negative regulation of TGF-␤induced transcription, we examined the effect of PDK1 on the association between T␤R1(TD), an activated type 1 TGF-␤ receptor, and Smad3 and -7, because we reasoned that PDK1 could modulate TGF-␤-induced transcription by altering the association between T␤R1(TD) and the Smad proteins. FLAG-Smad proteins (Smad3 and -7) were cotransfected with GST-T␤R1(TD) into 293T cells in the presence or absence of wildtype and kinase-dead PDK1 constructs. Compared with the control cells expressing GST-T␤R1(TD) and FLAG-Smads (Fig. 8A, top left panel, 4th and 9th lanes), the coexpression of wild-type PDK1 significantly decreased the association between T␤R1(TD) and Smad3 (ϳ57% decrease; Fig. 8A, top left panel, 4th versus 5th lanes) or increased the association between T␤R1(TD) and Smad7 (ϳ37% increase; Fig. 8B, top left  panel 4th versus 5th lane), whereas the KD PDK1 had no effect on association of the proteins (Fig. 8, A and B, top left panels, 9th  versus 10th lane). To further confirm whether the association between T␤R1(TD) and the Smad proteins is dependent on PDK1, T␤R1(TD)-Smad protein complex formation was determined in 293T cells using a PDK1-specific siRNA to knockdown PDK1 expression. Knockdown of endogenous PDK1 had an opposite effect on the association (Fig. 8, A and B, top right panels, 3rd versus 4th lanes). These data provide evidence that PDK1 negatively regulates TGF-␤ signaling through modulation of the direct interaction between the TGF-␤ receptor and Smad3 and -7.
PDK1 Modulates the Subcellular Localization of Smad Proteins-Because coexpression of wild-type PDK1 negatively regulated Smad3and Smad7-dependent transcription in a dose-dependent manner (Fig. 7), we hypothesized that PDK1 modifies the intracellular localization of Smad3, -4, and -7, which are essential to TGF-␤ signaling. Therefore, we performed an immunofluorescence microscopy analysis using Hep3B cells transfected with Smad proteins alone or together with either wild-type PDK1 or the kinase-dead mutant in the presence or absence of TGF-␤. In the absence of TGF-␤, Smad3 and Smad4 predominantly exhibited a cytoplasmic distribution, whereas Smad7 was mainly detected in the nucleus. However, TGF-␤ treatment significantly increased the nuclear localization of Smad3 and Smad4 (Fig. 9, A and B, upper panels, 1st versus 2nd lane). In contrast, the translocation of Smad7 from the nucleus to the cytoplasm was stimulated by TGF-␤ treatment (Fig. 9C, upper panel,  1st versus 2nd lane). Coexpression of wild-type PDK1 inhibited the nuclear translocation of Smad3 and Smad4 (Fig. 9, A and B, upper panels, 2nd versus 4th lane), as well as the translocation of Smad7 from the nucleus into the cytoplasm (Fig. 9C,  upper panel, 2nd versus 4th lane) in response to TGF-␤. However, the coexpression of kinase-dead PDK1 had no effect on the intracellular localization of Smad3, -4, or -7 (Fig.  9, A-C, upper panels, 2nd versus 6th  lane), consistent with the reporter  (Smad3 and -7) in the total cell lysates were determined by immunoblot analysis with anti-Myc and anti-FLAG antibodies, respectively (A and B, left, 3rd and 4th panels, Lysate). Expression level of GST-T␤R1(TD) in GST precipitates was determined by anti-GST antibody immunoblot (A and B, left, GST purification, 2nd panels). Quantification of the blots was done by band densitometry. The relative level of the complex formation was quantified by densitometric analysis, and fold increase relative to control samples cotransfected with GST-T␤R1(TD) and Smad3 (or Smad7) was calculated (A and B, left, bottom panels). 293T cells were transfected with the indicated siRNA duplexes (PDK1 siRNA and control siRNA), together with plasmid vectors expressing T␤R1(TD), FLAG-Smad3, or FLAG-Smad7, and complex formation between T␤R1(TD) and Smad3 (A, right, top panel) or T␤R1(TD) and Smad7 (B, right, top panel) was determined by anti-FLAG antibody immunoblot. The expression level of endogenous PDK1 was determined by anti-PDK1 immunoblotting (A and B, right, bottom panels). These experiments were performed in duplicate at least three times with similar results. WB, Western blot. assay data (Fig. 7). To provide further evidence that wild-type PDK1 is physiologically responsible for modulation of the intracellular localization of Smad proteins, we performed siRNA experiments using a PDK1-specific siRNA. Reducing the amount of endogenous PDK1 in cells showed a stronger effect on the translocation of Smad proteins compared with positive control cells treated with TGF-␤ in the absence of PDK1 (Fig. 9, A-C, upper panels, 2nd versus 7th lane). To verify whether the knockdown of endogenous PDK1 could alter the subcellular localization of Smad proteins, Hep3B cells transfected with wildtype or kinase-dead PDK1, together with a nonspecific control siRNA or a PDK1-specific siRNA, were treated with TGF-␤ and separated into cytoplasmic and nuclear fractions. Each fraction was analyzed by Western blot analysis. The accumulation of Smad3 and Smad4 in the nuclear fraction was significantly increased in PDK1-knockdown Hep3B cells compared with the control cells expressing a nonspecific control siRNA (Fig. 9, A and B, lower left panels, 3rd versus 4th lanes), whereas the cytoplasmic accumulation of Smad3 and Smad4 was markedly decreased (Fig. 9, A and B, lower right panels, 3rd versus 4th lanes). In contrast, a decrease in nuclear Smad3 and Smad4 was observed in cells expressing wildtype PDK1 (Fig. 9, A and B, lower left panels, 1st versus 2nd lanes), consistent with the confocal microscopy data (Fig. 9, A and B, upper panels). In addition, the opposite trend was observed for translocation of Smad7 in response to TGF-␤ under the same conditions (Fig. 9C). As a control, expression levels of exogenous and endogenous PDK1 were determined by Western blot analysis. Hep3B cells displayed a significant decrease in the amount of endogenous PDK1 after transfection of PDK1-specific siRNA (Fig.  9D, right panel). These results indicate that wild-type PDK1 prevents the normal translocation of Smad proteins in response to TGF-␤.

PDK1 Negatively Modulates TGF-␤-induced Apoptosis and Growth
Arrest-To explore the functional significance of the PDK1-Smad protein associations, based on the results described above, we first tested the effect of wild-type and kinase-dead PDK1 on TGF-␤-induced apoptosis using a GFP assay system (24). Apoptotic cells were scored by changes in nuclear morphology among GFP-positive cells after inducing apoptosis by TGF-␤ treatment, as described under "Materials and Methods." HeLa cells were transfected with an expression plasmid encoding GFP, together with wild-type or kinase-dead PDK1, and incubated in the presence or absence of TGF-␤. Approximately 48% of the HeLa cells were apoptotic following TGF-␤ treatment (Fig. 10A). Cells transfected with wild-type PDK1 showed higher apoptotic suppression (about 33% inhibition) than cells treated with TGF-␤ alone, and the effect was dose-dependent (Fig. 10A, left panel, 2nd versus 3rd and 4th  lanes). However, the inhibitory effect on TGF-␤-induced apoptosis was not observed in cells transfected with kinase-dead PDK1 (Fig. 10A, left panel, lane 2 versus lanes 5 and 6). These data are consistent with the results obtained from luciferase assays (Fig. 7) and indirect immunofluorescence studies (Fig. 9).
Furthermore, the knockdown of PDK1 with a PDK1-specific siRNA resulted in a significant and dose-dependent increase in TGF-␤-induced apoptosis, whereas a control siRNA had no effect (Fig. 10A, right panel). We extended our analysis to examine whether TGF-␤-induced apoptosis is also affected by modulating the amount of endogenous PDK1 in other TGF-␤-responsive cells, such as HaCaT and Hep3B cells. Under the conditions described above, a similar PDK1-mediated, dosedependent inhibition of TGF-␤-induced apoptosis was observed in both HaCaT and Hep3B cells (Fig. 10B and data not  shown). To examine whether TGF-␤-induced cell cycle arrest was affected by knockdown of endogenous PDK1, which might act as a negative regulator of TGF-␤ signaling, we performed a flow cytometry analysis using HaCaT cells expressing a vector control, control siRNA (Fig. 10, Con.(si)), PDK1 siRNA (PDK1(si)), Smad3, Smad3/control siRNA, and Smad3/PDK1 siRNA. As shown in Fig. 10C, ϳ30% of cells expressing Smad3 . Transfected cells were treated with TGF-␤1 (HeLa cells, 10 ng/ml; HaCaT cells, 2 ng/ml) for 20 h to induce apoptosis. Apoptotic cell death was determined using the GFP expression system, as described previously (24). GFP-positive cells were examined for the presence of apoptotic nuclei with a fluorescence microscope. The lower panels show the increasing amounts of PDK1 (WT and KD) and the knockdown of endogenous PDK1 in the cell extracts used for these assays, respectively. The data shown are the mean Ϯ S.D. of duplicate assays and are representative of at least four independent experiments. C, effect of PDK1 siRNA on TGF-␤-induced cell cycle arrest. HaCaT cells (2 ϫ 10 5 /dish) transfected with the indicated siRNA duplexes (PDK1 siRNA (a) and control siRNA) in the presence or absence of Smad3 were synchronized in G 0 /G 1 by hydroxyurea (2 mM) treatment for 20 h. Cells were collected before (0 h starvation) or after 10% serum treatment for 24 h in the absence (24 h serum stimulation) or presence (24 h serum stimulation ϩ TGF-␤1) of 2 ng/ml TGF-␤1, and the percentage of cells in the G 1 , S, or G 2 /M phases was analyzed by flow cytometry. Each experiment was repeated at least four times with similar results.
or Smad3/control siRNA accumulated in S phase after 24 h of serum stimulation in the presence of TGF-␤, whereas ϳ39% of cells accumulated in S phase after 24 h of serum stimulation in the absence of TGF-␤. However, a lower number of cells (ϳ18%) were found to be in S phase in the presence of PDK1 siRNA, compared with ϳ30% in the absence of PDK1 siRNA. In addition, this inhibitory effect was not because of the presence of the PDK1 siRNA itself, because in the absence of Smad3, the PDK1 siRNA could not effect a significant change in the percentage of cells present in S phase, compared with the control cells expressing vector or control siRNA (ϳ38 versus ϳ34%). These results indicate that PDK1 inhibits TGF-␤-induced growth arrest and suggest that TGF-␤-mediated biological functions, such as apoptosis and cell cycle arrest, are negatively regulated by PDK1.

DISCUSSION
We have recently shown that PDK1 interacts with STRAP, a TGF-␤ receptor-interacting protein, and that this interaction is involved in the activation of PDK1 activity (24). Furthermore, we have found that PDK1 could enhance the STRAP-dependent TGF-␤ transcriptional inhibition through a direct interaction, suggesting a possible cross-talk between the PI3K/PDK1 and TGF-␤ signaling pathways. Recently, cross-talk between PI3K/PDK1 and TGF-␤ signalings has been suggested in several systems (17, 34, 36 -39). TGF-␤ participated in PI3K/PDK1 activation and Akt phosphorylation in Swiss 3T3 cells and human mesangial cells (40), and LY294002, a PI3K inhibitor, blocked the Smad2 phosphorylation induced by TGF-␤ (41). Moreover, recent reports have shown that Akt, a downstream target of PDK1, physically interacts with Smad3 and suppresses TGF-␤ signaling (17,34). Based on these observations, we hypothesized that the direct interaction between PDK1 and Smads occurs in vivo.
In this study, we have shown that PDK1 physically interacts with Smad proteins (Smad2, -3, -4, and -7) and functionally suppresses TGF-␤-induced transcription. These results are analogous to our previous observations that a PDK1-STRAP association enhanced STRAP-induced inhibition of TGF-␤ signaling (24). In this context, these results suggest that PDK1 functions as a negative regulator of the TGF-␤ signaling pathway. Moreover, our data indicate that the kinase activity of PDK1 is necessary for its ability to suppress TGF-␤-induced transcription (Fig. 7) and TGF-␤-mediated biological functions such as apoptosis (Fig. 10), similar to a recent study showing that Akt, a downstream target of PDK1, associates with Smad2, -3, -4, and -7 and inhibits TGF-␤ signaling by an Akt-kinasedependent mechanism (42). As shown in this study, wild-type PDK1 modulates Smad3-induced stimulation of TGF-␤ signaling, as well as Smad7-induced inhibition of TGF-␤ signaling, but the kinase-dead form of PDK1 did not (Fig. 7). These data indicate that the suppression of TGF-␤ signaling by PDK1 occurs via a PDK1 kinase-dependent pathway. In addition, wild-type PDK1 modified the normal movement of the Smad proteins, but the kinase-dead PDK1 did not (Fig. 9). These data again support the hypothesis that PDK1 is a negative regulator of TGF-␤ signaling, and this modulation is dependent on the kinase activity of PDK1. Our present results do not support the possibility that suppression of TGF-␤ signaling by PDK1 is due only to the direct physical interaction between PDK1 and Smad proteins, because the kinase-dead PDK1 was shown to associate with Smad proteins at levels similar to wild-type PDK1, although it did not contribute to the negative regulation of TGF-␤ signaling (Fig. 7C).
In our previous report (24), we showed that PDK1 enhances the STRAP-induced inhibition of TGF-␤ signaling by stabilizing the association between TGF-␤ receptor and Smad7. Thus, it seems likely that the mechanism by which PDK1 suppresses the TGF-␤-induced transcriptional activation is by modulating the association between the TGF-␤ receptor and Smad proteins. To test this hypothesis, we examined the effects of wild-type PDK1 and kinase-dead PDK1 on Smad3 and Smad7 binding to the TGF-␤ receptor (Fig. 8). Our results show that the ability of wild-type PDK1 to suppress the TGF-␤-induced transcriptional activation correlates with PDK1-induced modulation of Smad3 and Smad7 binding to the TGF-␤ receptor. Wild-type PDK1 decreased the association between the TGF-␤ receptor and Smad3 and increased the association between the TGF-␤ receptor and Smad7, leading to the inhibition of the TGF-␤induced transcriptional activation. In contrast, no difference was found in the presence of the kinase-dead PDK1. These data indicate that the kinase activity of PDK1 is also important for modulation of complex formation between the TGF-␤ receptor and Smad proteins.
Smad3 plays a key role in the TGF-␤ signaling pathway and, upon TGF-␤ treatment, is phosphorylated by the TGF-␤ type I receptor at the SSXS motif in its carboxyl terminus and forms a complex with Smad4, a Co-Smad. The heterodimer accumulates in the nucleus to regulate transcription of target genes such as PAI-1, p21 Cip1 , Smad7, CDK4, CDK2, and cyclin D1. Recent studies have shown that cyclin-dependent protein kinases (CDK4 and CDK2) phosphorylate Smad3 and inhibit its activity (43,44). The phosphorylation sites of CDK4 and CDK2 were mapped to Thr-8, Thr-178, and Ser-212 in the Smad3 linker region. They suggest that CDK-induced Smad3 phosphorylation stimulates the G 1 to S transition. However, the underlying mechanism by which Smad3 phosphorylation inhibits the transcriptional activity of the protein remains to be determined. In addition to CDK4 and CDK2, the Smad3 linker region possesses phosphorylation sites for other protein kinases, including extracellular signal-regulated kinase (ERK), c-Jun amino-terminal kinase, p38, and calmodulin-dependent kinase II (32,(45)(46)(47)(48). The phosphorylation of Smad3 and Smad4 by PDK1, illustrated in Fig. 1, D and E, reveals one aspect of the negative regulation of TGF-␤-induced transcription, similar to the previous results (43,44) showing that Smad3 phosphorylation is important for inhibition of TGF-␤-induced transcriptional activation. However, further studies, including the identification of the PDK1 phosphorylation sites in Smad3 (or Smad4) and elucidation of the detailed mechanism by which Smad3 (or Smad4) phosphorylation inhibits TGF-␤ signaling, are required to completely elucidate the mechanism of inhibition.
Overall, our present study provides evidence that PDK1, like PKB/Akt (17,34,42), is correlated, either directly or indirectly, with the negative regulation of TGF-␤ signaling through direct interactions with Smad proteins and that Smad proteins, similar to STRAP (24), may function as adaptors coupling TGF-␤ signaling to PI3K/PDK1 signaling.