Mechanism of Phosphorylation of Protein Kinase B/Akt by a Constitutively Active 3-Phosphoinositide-dependent Protein Kinase-1*

Phosphorylation of Thr 308 in the activation loop and Ser 473 at the carboxyl terminus is essential for protein kinase B (PKB/Akt) activation. However, the biochemical mechanism of the phosphorylation remains to be characterized. Here we show that expression of a constitutively active mutant of mouse 3-phosphoinositide-dependent protein kinase-1 (PDK1 A280V ) in Chinese hamster ovary cells overexpressing the insulin receptor was sufficient to induce PKB phosphorylation at Thr 308 to approximately the same extent as insulin stimulation. Phosphorylation of PKB by PDK1 A280V was not affected by treatment of cells with inhibitors of phosphatidylinositol 3-kinase or by deletion of the pleckstrin homol-ogy (PH) domain of PKB. C 2 -ceramide, a cell-permeable, indirect inhibitor of PKB phosphorylation, did not inhibit PDK1 A280V -catalyzed PKB phosphorylation in cells and had no effect on PDK1 activity in vitro . On the other hand, co-expression of full-length protein kinase C-re-lated kinase-1 (PRK1/PKN) or 2 (PRK2) inhibited PDK1 A280V -mediated PKB phosphorylation. Replacing alanine at position 280 with valine or deletion of the PH

that has been shown to play an important role in the regulation of a wide spectrum of cellular signaling events, including insulin-stimulated glucose transporter GLUT4 membrane translocation, anti-apoptosis, protein synthesis, and glycogen metabolism (1).
Phosphorylation of Thr 308 in the activation loop and Ser 473 at the carboxyl terminus is essential for PKB activation (2,3). Although it has been shown that phosphorylation of PKB at Thr 308 is catalyzed by 3-phosphoinositide-dependent protein kinase-1 (PDK1) (4,5), the effector of PKB Ser 473 phosphorylation remains controversial. Because PDK1 does not phosphorylate PKB at Ser 473 , it has been suggested that phosphorylation at this site is catalyzed by a distinct but yet to be characterized enzyme termed PDK2 (6). On the other hand, it has been shown that PDK1, through an interaction with a PDK1-interacting fragment (PIF) derived from the carboxyl terminus of PRK2, can switch its substrate specificity to phosphorylate PKB at Ser 473 in vitro (7). Alternatively, it has recently been reported that phosphorylation of the wild-type but not the kinase-inactive PKB at Thr 308 triggers phosphorylation at Ser 473 , suggesting that phosphorylation of PKB at Ser 473 occurs through an autophosphorylation mechanism (8).
The current model of PKB activation suggests that the enzyme undergoes mitogen-stimulated translocation to the plasma membrane, which is mediated by the binding of phosphatidylinositol 3-kinase (PI3K) products, namely PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 , to the PH domain of PKB. This binding also leads to a conformational change that exposes Thr 308 and Ser 473 to the membrane-associated and constitutively activated upstream kinases PDK1 and PDK2, respectively. This model is supported by the finding that myristoylation/palmitoylation of PKB is sufficient to induce phosphorylation of the enzyme at Thr 308 and Ser 473 (9). In addition, PDK1 purified from cells is constitutively active, and treatment of cells with mitogen does not stimulate PDK1 activity in vitro (4). However, there are some data suggesting that the PH domain of PKB is not necessary for PKB phosphorylation and activation. First, it has been shown that PKB constructs lacking the PH domain can still be activated by insulin and insulin-like growth factor-1 (10,11). In addition, the activity of a membrane-targeted PKB by myristoylation/ palmitoylation is still sensitive to the PI3K inhibitor LY294002 (9). These data suggest that activation of a PKB upstream kinase plays an important role in the phosphorylation and activation of PKB.
In the current study, we examined the phosphorylation mechanism of PKB using a constitutively active mutant of mouse PDK1, PDK1 A280V , in which the alanine at position 280 was replaced with valine. We have found that ectopic expression of PDK1 A280V resulted in a marked increase in PKB phosphorylation at Thr 308 to an extent similar to that induced by insulin stimulation. The PDK1 A280V -mediated phosphorylation of PKB was not affected by treatment of cells with the PI3K inhibitors wortmannin and LY294002, nor by ceramide, a compound known to indirectly inhibit PKB phosphorylation (12). Deletion of the PH domain of PKB had little effect on PDK1 A280V -mediated phosphorylation of PKB. On the other hand, deletion of the PH domain of PDK1 A280V inhibited the phosphorylation of PKB in cells. Taken together, our data suggest that wild-type PDK1 may not be constitutively active in cells. In addition, the PH domain of PDK1 may play both a positive and a negative role in regulating the function of the enzyme in cells. Finally, we have found that overexpression of the full-length PRK1/PKN or PRK2 inhibited PDK1 A280V -mediated PKB phosphorylation, suggesting that these enzymes play negative roles in PKB-mediated downstream signaling.

EXPERIMENTAL PROCEDURES
Cell Lines, cDNAs, and Antibodies-A Chinese hamster ovary (CHO) cell line overexpressing the human insulin receptor (CHO/IR) and the cDNA encoding human PKB␣ were gifts from Dr. Richard A. Roth. cDNA encoding mouse PDK1 and a polyclonal antibody against the protein were described previously (13). The cDNA encoding the FLAGtagged PRK1/PKN, PRK2, and the carboxyl-terminal fragment of PRK1/PKN (AF3-PKN) were gifts of Drs. Hideyuki Mukai and Yoshitaka Ono and have been described previously (14). Anti-PKB and antiphospho-PKB antibodies (Thr 308 and Ser 473 ) were obtained from New England BioLabs. Monoclonal antibody to the FLAG-tag was purchased from IBI-Kodak. Polyclonal and monoclonal anti-Myc antibodies were from BABCO and Santa Cruz Biotechnologies, Inc., respectively. The anti-GST-polyclonal antibody was generated by immunizing a rabbit with GST and was affinity-purified.
Site-directed Mutagenesis-Myc-tagged PDK1 A280V , PKB T308A , and PKB K179A were generated by single-stranded site-directed mutagenesis according to the protocol as described by Kunkel (15) using customized primers. All site-directed mutagenesis products were confirmed by restriction mapping and DNA sequencing. The kinase inactive mutant PDK1 K114G was described previously (13).
Generation of Myc-tagged and PH Domain-deleted Mutants of PDK1 and PDK1 A280V -cDNAs encoding mouse PDK1 or PDK1 A280V with a truncation in the PH domain (PDK1(⌬PH) and PDK1 A280V (⌬PH); residues 1-459) were generated by PCR, using either PDK1 or PDK1 A280V cDNA as a template. The PCR primers used were: 5Ј-GCGAATTCGA-CTTGGGGCTCATGGCC-3Ј and 5Ј-GCTCTAGACTGGTGCCAAGGGT-TTCC-3Ј, with the added restriction sites underlined. After restriction digestion with EcoRI and XbaI, the cDNA fragments were subcloned into the mammalian expression vector pcDNA3.1-Myc-His-A (Invitrogen, San Diego, CA), in-frame with a sequence encoding a Myc-tag at the 3Ј-end.
Generation of FLAG-tagged PH Domain-truncated Mutant of PKB-cDNAs encoding the PH domain-truncated mutant of PKB were amplified by PCR using human PKB␣ cDNA as a template. The forward PCR primers for full-length and the PH domain-deleted PKB were: 5Ј-CCGGAATTCTATTGTGAAGGAGGGTTGG-3Ј and 5Ј-CCGGAAT-TCCATCCAGACTGTGGCTGAC-3Ј, respectively. The reverse primer for both constructs was 5Ј-CTAGTCTAGAGTAGGAGAACTGGGG-GAAG-3Ј, with the added restriction sites underlined. The cDNA fragments were subcloned into the EcoRI/XbaI sites of pFLAG-CMV2 vector (Sigma), in-frame with the sequence encoding a FLAG-tag at the 5Ј-end.
Cell Culture, Immunoprecipitation, and Western Blot-CHO/IR cells were grown in Ham's F-12 medium (Life Technologies) supplemented with 10% newborn calf serum and 1% penicillin/streptomycin. Co-transfections of CHO/IR cells were normally performed in 60-mm plates with 2 g of each recombinant plasmid, using LipofectAMINE reagent according to the manufacturer's protocol (Life Technologies). Twenty-four hours after transfection, cells were serum-starved for 2-4 h, treated with or without 10 nM insulin, and lysed in 120 -400 l of Buffer A (50 mM Hepes, pH 7.6, 150 mM NaCl, 1% Triton X-100, 10 mM NaF, 20 mM sodium pyrophosphate, 20 mM ␤-glycerol phosphate, 1 mM sodium orthovanadate, 10 g/ml leupeptin, 10 g/ml aprotinin, 1 M microcystin-LR, and 1 mM phenylmethylsulfonyl fluoride). The homogenate was centrifuged (10,000 ϫ g, 4°C, 10 min), and the supernatants were used for immunoprecipitation or Western blot experiments. For immunoprecipitation studies, cell lysates were incubated with specific antibodies bound to protein G beads (Amersham Pharmacia Biotech) for 4 -6 h at 4°C with gentle rotation. After incubation, immunoprecipitates were washed extensively with ice-cold Buffer B (50 mM Hepes, pH 7.6, 150 mM NaCl, and 0.1% Triton X-100). Proteins bound to the beads were eluted by heating at 95°C for 4 min in SDS-PAGE sample loading buffer. The eluted proteins were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and detected with specific antibodies. The expression of various proteins was detected by Western blot of cell lysates with antibodies specific to the Myc or FLAG-tag or to the protein. Quantitation of the relative increase in PKB phosphorylation (expressed as percentage of maximum phosphorylation) was performed by analyzing Western blots using the National Institutes of Health IMAGE 1.58 program and was normalized for the amount of PKB expression in each experiment. All experiments were repeated independently at least three times to ensure the consistency of the results.
PDK1 Autophosphorylation-CHO/IR cells overexpressing Myctagged wild-type PDK1, PDK1 A280V , or PH domain-deleted mutants of PDK1 or PDK1 A280V were serum-starved for 4 h, treated with or without 10 nM insulin for 5 min, and lysed in Buffer A. Cell lysates were centrifuged at 10,000 ϫ g for 10 min at 4°C, and the proteins in the supernatants were immunoprecipitated with antibody to the Myc-tag. After washing the immunocomplexes twice with Buffer B and once with Buffer C (50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 1 mM Na 3 VO 4 , 1 mM sodium pyrophosphate, 10 mM NaF, and 1 mM phenylmethylsulfonyl fluoride), autophosphorylation of PDK1 was initiated with the addition of 30 l of Buffer C plus 2 Ci of [␥-32 P]ATP and incubated for 10 min at 30°C. Bound proteins were eluted by heating at 95°C for 4 min in SDS-PAGE sample buffer. Proteins were separated by SDS-PAGE and blotted to a nitrocellulose membrane. Phosphorylation of PDK1 was visualized by autoradiography and quantitated using the NIH IMAGE 1.58 program.
Confocal Immunofluorescence Microscopy-CHO/IR cells were seeded in 24-well plates containing coverslips and transfected with plasmids encoding Myc-tagged PDK1, PDK1 A280V , or the PH domaindeleted mutant forms of PDK1 or PDK1 A280V using LipofectAMINE (Life Technologies). 24 h after transfection, cells were serum-starved for 1 h, treated with or without 10 nM insulin for 10 min, and then fixed in phosphate-buffered saline containing 4% paraformaldehyde for 20 min at room temperature. Fixed cells were permeabilized and blocked with phosphate-buffered saline containing 0.05% Triton X-100 and 10% normal goat serum. The expression of Myc-tagged full-length or the PH domain-deleted mutants of PDK1 or PDK1 A280V was detected by incubation with the anti-Myc-antibody, followed by a rhodamine-conjugated secondary antibody (KPL). Images were captured by an Olympus Fluoview confocal fluorescence microscope.

RESULTS
Overexpression of PDK1 A280V Is Sufficient to Phosphorylate PKB at Thr 308 in Vivo-To characterize the biochemical mechanism by which PKB is phosphorylated and activated by PDK1, we generated a mutant mouse PDK1 in which the alanine at position 280 was replaced with valine (PDK1 A280V ). Mutation at the equivalent position in the Caenorhabditis elegans homologue of mammalian PDK1 has been shown to increase PDK1 kinase activity and bypass the requirement for AGE1/PI3K in transducing upstream signals to PKB (16). To test whether PDK1 A280V constitutively phosphorylated PKB in cells, we cotransfected CHO/IR cells with PKB and either wild-type, A280V, or the kinase-defective mutant (PDK1 K114G (13)) of PDK1. Western blot analysis using phospho-specific antibodies directed against either Thr 308 or Ser 473 of PKB showed little PKB phosphorylation under basal conditions or in cells coexpressing wild-type PDK1 or the kinase-defective mutant PDK1 K114G (Fig. 1A, top and second panels, lanes 1, 2, and 4). On the other hand, co-expression of PDK1 A280V with PKB resulted in a marked increase in PKB phosphorylation at Thr 308 to a level approximately equal to that attained by insulin treatment (Fig. 1A, top panel, lane 3 versus lane 5). Concomi-tantly, a modest increase of phosphorylation at Ser 473 was also observed (Fig. 1A, second panel, lane 3 versus 1). Overexpression of PDK1 A280V did not stimulate the phosphorylation of MAP kinase (data not shown).
Overexpression of Full-length PRK1/PKN or PRK2 Inhibits PDK1 A280V -mediated PKB Phosphorylation-A recent study has shown that PDK1 can phosphorylate Ser 473 of PKB upon binding the PDK1-interactive fragment (PIF) derived from PRK2 (7). To test whether PDK1 A280V was able to phosphorylate PKB at Ser 473 , we co-expressed Myc-tagged PKB and Myctagged PDK1 A280V in CHO/IR cells together with either FLAGtagged PRK2 or PRK1 (also named PKN), a PRK2-related protein that has recently been found to interact with PDK1 in intact cells (17). Our results showed that overexpression of PRK1/PKN significantly inhibited PDK1 A280V -stimulated PKB phosphorylation at Thr 308 (Fig. 1B,  It is possible that the interaction of full-length PRK1/PKN or PRK2 with PDK1 induced a conformational change that was different from that induced by the PIF fragment. To test this possibility, we co-expressed Myc-tagged PKB in CHO/IR cells together with either a control plasmid (Fig. 1C, lanes 1 and 2), Myc-tagged PDK1 A280V (Fig. 1C, lane 3), Myc-tagged PDK1 A280V plus the FLAG-tagged carboxyl terminus of PRK1/ PKN (AF3-PKN) (14) (Fig. 1C, lane 4), or Myc-tagged PDK1 A280V plus the GST-PIF fusion protein (Fig. 1C, lane 5).
Overexpression of either AF3-PKN or GST-PIF had no significant effect on the expression levels of PKB (Fig. 1C, third panel) or PDK1 (Fig. 1C, fourth panel). However, overexpression of these proteins significantly inhibited PDK1 A280V -induced PKB phosphorylation at Thr 308 (Fig. 1C,  These results are in agreement with the recent finding that binding of PIF fragment enables PDK1 to phosphorylate PKB at Ser 473 (7).
Phosphorylation of PKB at Thr 308 by PDK1 A280V Is Independent of PI3K Activation-To test whether PI3K is required for PDK1 A280V -mediated phosphorylation of PKB, we co-transfected PKB into CHO/IR cells together with either a pcDNA control plasmid or a plasmid encoding PDK1 A280V and examined PKB phosphorylation in the presence or absence of PI3K inhibitors. In the absence of PDK1 A280V , insulin treatment led to a significant increase in PKB phosphorylation at Thr 308 ( Fig.  2A, top panel, lane 1 versus lane 2). This insulin-stimulated phosphorylation of PKB at Thr 308 was blocked by either wortmannin ( Fig. 2A, top panel, lane 3) or LY294002 ( Fig. 2A, top  panel, lane 4). In agreement with our earlier findings (Fig. 1A, top panel), co-expression of PDK1 A280V with PKB resulted in a marked increase in PKB phosphorylation at Thr 308 ( Fig. 2A,  top panel, lane 5), which was only slightly enhanced after insulin treatment ( Fig. 2A, top panel, lane 6). However, neither wortmannin nor LY294002 inhibited the PDK1 A280V -stimulated PKB phosphorylation at Thr 308 ( Fig. 2A, top panel, lane 5  versus lanes 7 or 8).
Overexpression of PDK1 A280V Prevents Ceramide-induced Inhibition of PKB Phosphorylation-To further understand the mechanism of PKB phosphorylation, we investigated the phosphorylation of PKB by PDK1 in the presence or absence of C 2 -ceramide, a cell-permeable reagent that indirectly inhibits PKB phosphorylation and activity (12). Consistent with the results of others (18,19), we found that treatment of CHO/IR cells with ceramide inhibited PKB phosphorylation at both Thr 308 (Fig. 2B, top panel, lane 1 versus lanes 2 and 3) and Ser 473 (Fig. 2B, second panel, lane 1 versus lanes 2 and 3) but had no significant effect on the autophosphorylation of the IR and the tyrosine phosphorylation of IR substrate-1 (IRS-1) (data not shown). The inhibition of PKB phosphorylation by ceramide was partially prevented by overexpression of wildtype PDK1 and almost completely blocked by overexpression of PDK1 A280V (Fig. 2B, top panel, lanes 4 -9). In addition, we have found that ceramide did not inhibit PDK1 activity in vitro (data not shown).
Phosphorylation at Thr 308 and the Kinase Activity of PKB Are Important for PKB Phosphorylation at Ser 473 -It has recently been shown that phosphorylation by PDK1 in vitro triggers the phosphorylation at Ser 473 in kinase-active, but not thermally inactivated PKB (8). To test whether phosphorylation of PKB at Thr 308 is sufficient to initiate the phosphoryla-tion of PKB at Ser 473 in intact cells, we co-expressed Myctagged wild-type PKB, PKB T308A , or a kinase-inactive PKB K179A in CHO/IR cells together with either a control plasmid or Myc-tagged PDK1 A280V . Cells were treated with or without insulin, and the phosphorylation of PKB at Thr 308 and Ser 473 was examined. Stimulation of cells with insulin (Fig. 3,  top panel, lane 1 versus lane 2) or overexpression of PDK1 A280V (Fig. 3, top panel, lane 1 versus lane 3). led to a marked increase in the phosphorylation at Thr 308 of wild-type PKB. As expected, the PKB T308A mutant was not phosphorylated in the activation loop under these conditions (Fig. 3, top panel, lanes 4 -6). Inactivation of the kinase by replacing the ATP-binding site lysine residue (Lys 179 ) with alanine resulted in a 70% decrease in the insulin-or PDK1 A280V -stimulated phosphorylation at Thr 308 (Fig. 3, top panel, lanes 2 and 3 versus 8 and 9). However, replacing Ser 473 with alanine had essentially no effect on PKB phosphorylation at Thr 308 (data not shown). Consistent with the findings of others (8), a mutation at Thr 308 inhibited the insulin-stimulated phosphorylation at Ser 473 (Fig. 3, second panel, lane 5). Inactivation of PKB also resulted in a marked decrease in the insulin-stimulated PKB phosphorylation at Ser 473 (Fig. 3, second panel, lane 2 versus lane 8). However, quantitation analysis revealed that approximately 46% of Ser 473 phosphorylation remained for this kinase-inactive mutant. Together, these finding suggest that phosphorylation at Thr 308 and the kinase activity of PKB play an important role in the phosphorylation of the enzyme at Ser 473 . However, because PKB K179A was still able to undergo insulin-stimulated phosphorylation at Ser 473 in cells, a mechanism other than or in addition to autophosphorylation may play a role in the insulinstimulated phosphorylation of PKB at this site.
The PH Domain of PKB Is Not Essential for PDK1 A280Vcatalyzed Phosphorylation of PKB at Thr 308 -To test whether the PH domain of PKB is required for PDK1 A280V -catalyzed phosphorylation of PKB, we co-transfected CHO/IR cells with PDK1 A280V and either the wild-type or a PKB mutant in which the PH domain (residues 1-102) was deleted. Western blot studies showed that insulin treatment led to an increase in PKB phosphorylation at Thr 308 (Fig. 4A, top panel, lane 1  versus lane 2). This insulin-stimulated phosphorylation was abolished for the PH domain-deleted mutant PKB(⌬PH) (Fig.  4A, top panel, lane 5). On the other hand, only a partial inhibition of the PDK1 A280V -mediated phosphorylation was observed for PKB(⌬PH) compared with that of the full-length enzyme (Fig. 4A, top panel, lane 3 versus lane 6). Because the

FIG. 2. Phosphorylation of PKB by PDK1 A280V is insensitive to PI3K inhibitors and ceramide.
A, overexpression of PDK1 A280V prevents wortmannin-or LY294002-induced inhibition of PKB phosphorylation at Thr 308 . Quiescent CHO/IR cells co-expressing Myc-tagged PKB and either PDK1 or PDK1 A280V were left untreated (Ϫ) or treated (ϩ) with 50 M LY294002 or 50 nM wortmannin for 1 h, followed by treatment with 10 nM insulin for 8 min. Cells were lysed in 120 l of lysis buffer, and cell lysates were analyzed by Western blot using antibodies to phospho-Thr 308 of PKB (top panel), PKB (middle panel), or PDK1 (bottom panel). B, overexpression of PDK1 A280V prevents ceramide-induced inhibition of PKB phosphorylation. PKB was co-expressed in CHO/IR cells with a control plasmid (pcDNA) (lanes 1-3), wild-type PDK1 (lanes 4 -6), or PDK1 A280V (lanes 7-9). Quiescent cells were treated with increasing concentrations of ceramide as indicated, and then with (ϩ) or without (Ϫ) 10 nM insulin for 8 min. PH domain of PKB has been shown to be necessary for the translocation of the enzyme to the plasma membrane (1), our findings suggest that PDK1 A280V is able to phosphorylate PKB(⌬PH) in the cytosol. This is also consistent with our earlier finding that treatment of cells with PI3K inhibitors did not block PDK1 A280V -mediated phosphorylation of PKB ( Fig. 2A).
Deletion of the PH Domain of PDK1 A280V Inhibits PDK1 A280V -catalyzed Phosphorylation of PKB at Thr 308 -The PH domain of PDK1 has been shown to bind phosphoinositides and be necessary for PDK1-mediated phosphorylation and activation of PKB (20,21). To test whether the PH domain is required for PDK1 A280V -mediated phosphorylation of PKB at Thr 308 , we generated mutants of PDK1 and PDK1 A280V in which the PH domains were deleted. Co-expression of PKB with wild-type PDK1 or PDK1(⌬PH) had no significant effect on basal PKB phosphorylation at Thr 308 (Fig. 4B, top panel,  lanes 1 and 2). Conversely, co-expression of PDK1 A280V with PKB markedly increased the phosphorylation of PKB at this site (Fig. 4B, top panel, lane 3). The PDK1 A280V -stimulated phosphorylation of PKB at Thr 308 was markedly decreased by deletion of the PH domain of PDK1 A280V (Fig. 4B, top panel,  lane 3 versus lane 4). In addition, overexpression of the PDK1(⌬PH) or PDK1 A280V (⌬PH) inhibited insulin-stimulated phosphorylation of PKB at Thr 308 (Fig. 4B, top panel, lane 5  versus 6 and lane 7 versus 8). These findings indicate that the PH domain of PDK1 is important for the enzyme to interact with and phosphorylate PKB in cells.
One explanation for the inhibition of PKB phosphorylation by PDK1(⌬PH) or PDK1 A280V (⌬PH) might be that removal of the PH domain rendered the enzyme inactive. To test this possibility, we examined the autophosphorylation of wild-type and PH domain-deleted mutants of PDK1 and PDK1 A280V . Wild-type PDK1 and PDK1 A280V isolated from both serumstarved and insulin-treated cells were autophosphorylated in vitro (Fig. 4C). Deletion of the PH domain or replacing alanine at 280 with valine led to approximately a 2-and 3-fold increase in the autophosphorylation of PDK1, respectively. However, deletion of the PH domain did not further stimulate the autokinase activity of PDK1 A280V (Fig. 4C). Therefore, deletion of the PH domain increased, rather than inhibited, PDK1 catalytic activity per se.
Deletion of the PH Domain of PDK1 A280V Prevents Its Plasma Membrane Translocation-The finding that deletion of the PH domain did not affect PDK1 A280V autophosphorylation in vitro but prevented the enzyme from phosphorylating PKB in cells suggested that the PH domain plays a role in the interaction of PDK1 A280V with its substrates. To test this idea, we examined the cellular localization of PDK1, PDK1 A280V , and their PHdomain-deleted mutants by confocal immunofluorescence studies. In unstimulated CHO/IR cells, PDK1 A280V localized both at the plasma membrane and in the cytosol in a punctate manner (Fig. 5, a and b). Plasma membrane association of PDK1 A280V was also observed in cells treated with insulin (Fig. 5, c and d).
On the other hand, no significant plasma membrane immunofluorescence staining was observed for PDK1 A280V (⌬PH) before or after insulin treatment (Fig. 5, e-h). Under serum-starved conditions, wild-type PDK1 (Fig. 5, i and j) and PDK1(⌬PH) (Fig. 5, m and n) resided predominantly in the perinuclear region. Although insulin treatment led to a decrease in the perinuclear staining of the wild-type PDK1 (Fig. 5, k and l), it had no significant effect on the intracellular translocation of PDK1(⌬PH) (Fig. 5, o and p). These findings are consistent with an earlier report that the PH domain is important for PDK1 translocation to the plasma membrane (20). DISCUSSION Phosphorylation of PKB at Thr 308 in the activation loop by PDK1 is considered essential for PKB activation. However, the biochemical and cellular mechanisms by which PKB is phosphorylated by upstream kinases remain unclear.
In the present studies, we have examined the biochemical mechanism of PKB phosphorylation by using a constitutively active mutant of PDK1. Expression of the gain-of-function mu-  1-3) or ⌬PH (lanes 4 -6) PKB was co-transfected into CHO/IR cells with (ϩ) or without (Ϫ) PDK1 A280V . Quiescent cells were left untreated (Ϫ) or treated (ϩ) with 10 nM insulin for 5 min. Cell lysates were analyzed by Western blot using antibodies specific to phospho-PKB-Thr 308 (top panel), PKB (middle panel), and PDK1 (bottom panel). B, deletion of the PH domain of PDK1 or PDK1 A280V inhibits PKB phosphorylation at Thr 308 . CHO/IR cells were co-transfected with wild-type PKB and either PDK1, PDK1(⌬PH), PDK1 A280V , or PDK1 A280V (⌬PH). Cells were treated with or without insulin as described in A. Cell lysates were analyzed by Western blot with antibodies to phospho-PKB-Thr 308 (top panel), PKB (middle panel), and PDK1 (bottom panel). Results shown in A and B are representative of at least three independent experiments with similar results. C, autophosphorylation of wild-type and PH-domain-deleted mutants of PDK1 or PDK1 A280V in vitro. Myc-tagged wild-type and PH domain-deleted mutants of PDK1 and PDK1 A280V were immunoprecipitated from CHO/IR cells overexpressing these proteins using an antibody against the tag. Autophosphorylation of PDK1 was carried out as described under "Experimental Procedures." The phosphorylation of PDK1 was visualized by autoradiography and quantified by the NIH IMAGE program. Equal expression of PKB in each sample was determined by Western blot analysis of cell lysates with antibody to the Myc-tag. The mean percentage of maximal phosphorylation Ϯ S.D. from two independent experiments is shown. tant PDK1 homologue has recently been shown to bypass the requirement for AGE-1/PI3K-mediated PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 production to induce dauer arrest in C. elegans (16). In agreement with this finding, we have found that expression of PDK1 A280V was sufficient to induce significant phosphorylation of PKB at Thr 308 in intact cells (Fig. 1A). Pretreatment of cells with either LY294002 or wortmannin, or deletion of the PH domain of PKB, had essentially no effect on PDK1 A280V -mediated phosphorylation of PKB at Thr 308 (Figs. 2A and 4A). These findings suggest that activation of PDK1 is sufficient to phosphorylate PKB at Thr 308 in the cytosol. This is consistent with an earlier finding that a PKB mutant lacking its PH domain could be activated by either insulin or PDGF in several cell types (11). More recently, it has been found that insulin stimulates an increase of PKB␤ phosphorylation in the cytosol (22). Consistently, PDK1 has been found both at the plasma membrane and in the cytosol (20,21).
While this manuscript was under revision, Filippa et al. (23) reported that overexpression of ⌬PH-PDK1 activated ⌬PH-PKB in 293 cells. Based on this finding, they suggest that the PH domain of PDK1 acts as a negative regulator of its enzyme activity. In agreement with their results, we have found that deletion of the PH domain stimulated PDK1 autokinase activity in vitro (Fig. 4C). However, our results also show that the PH domain of PDK1 is necessary for PDK1-mediated PKB phosphorylation in cells, because overexpression of the PHdomain-deleted mutants of PDK1 or PDK1 A280V inhibited phosphorylation of full-length PKB (Fig. 4B). These findings suggest the PH domain plays a dual role in regulating PDK1 activity in cells. On one hand, the PH domain may be necessary for targeting the enzyme to a specific cellular location to interact with and phosphorylate its substrates. On the other hand, this domain may function as an autoinhibitory region to block the catalytic activity of PDK1 under basal conditions. This inhibitory mechanism may be important for preventing nonspecific phosphorylation and activation of PDK1 downstream substrates under non-stimulatory conditions. After growth factor stimulation, the binding of phospholipids to the PH domain may induce a conformational change that relieves the inhibitory blockage so that PDK1 becomes fully active. In addition to activating PDK1, the binding of phospholipids may also promote the plasma membrane translocation of the enzyme. Thus, the PH domain may function as a switch to regulate PDK1mediated biological events in cells.
Until recently, it has been suggested that PDK1 is constitutively active in cells and that phosphorylation and activation of its downstream substrates is mediated mainly by translocation and conformational changes of its substrates (24,25). However, our data suggest that PDK1 can be activated above its basal level both in vitro and in cells. Consistent with this, a mutation at Ala 280 or deletion of the PH domain significantly increased PDK1 autophosphorylation in vitro (Fig. 4C). In addition, overexpression of PDK1 A280V , but not wild-type PDK1, was able to phosphorylate PKB at Thr 308 to the same extent as that stimulated by insulin (Fig. 1A). Thus, although PDK1 is constitutively active in vitro, its activation may be tightly regulated in cells. This idea is consistent with a recent finding that PDK1 is activated by sphingosine both in vitro and in vivo (26). Moreover, it has also recently been shown that treatment of cells with the phosphatase inhibitors okadaic acid and calyculin A inhibited PKB phosphorylation at Thr 308 (27), suggesting that phosphorylation plays a role in regulating PDK1 activity in cells. Consistent with the findings of others (28), we have found that phosphorylation of Ser 244 in the activation loop of mouse PDK1 is essential for the autophosphorylation and kinase activity of the enzyme. 2 Ceramide has been shown to be a potent inhibitor for PKB phosphorylation in cells (18,19). Because CHO/IR cells expressing Myc-tagged fulllength PDK1 or PDK1 A280V or the PH domain-deleted mutants of these enzymes were serum-starved for 1 h and treated with or without 10 nM insulin for 10 min. Cells were then fixed and labeled with the anti-Myc antibody, followed by rhodamine-labeled second antibody. The confocal images presented here represent projections of Z-series scans. Representative fields are shown of immunofluorescence (anti-Myc) and the corresponding differential interference contrast (DIC) images. Data are representative of three independent experiments with similar results. effect on PKB catalytic activity in vitro, it has been hypothesized that its action may be due to the inhibition of an immediate upstream kinase for PKB (12). However, our data showed that ceramide-induced inhibition could be blocked by PDK1 A280V (Fig. 2B). In addition, ceramide did not inhibit PDK1 activity in vitro (data not shown). These findings suggest that the inhibition of PKB by ceramide is not mediated by direct inhibition of PDK1. These results are consistent with the findings that the inhibition of PKB phosphorylation by ceramide in cells is due to the activation of a protein phosphatase (27,29) rather than by the inhibition of the PKB upstream kinase PDK1.
It has recently been suggested that phosphorylation of PKB at the PDK2 site is mediated by autophosphorylation (8). This conclusion is supported by the findings that incubation of purified PKB with PDK1 led to phosphorylation at both Thr 308 and Ser 473 , and phosphorylation at Ser 473 requires prior phosphorylation at Thr 308 and the intrinsic catalytic activity of the enzyme (8). However, whether phosphorylation of PKB at Thr 308 is sufficient to trigger autophosphorylation at Ser 473 in intact cells is unknown. In our studies, we found that expression of a constitutively active PDK1 induced marked phosphorylation of PKB at Thr 308 and to a lesser extent Ser 473 (Fig. 1). Because PDK1 does not directly phosphorylate PKB at Ser 473 , these findings suggest that phosphorylation of Thr 308 by PDK1 is able to induce PKB autophosphorylation at Ser 473 . This idea is consistent with a recent finding that phosphorylation of Ser 473 may occur through an autophosphorylation mechanism (8). However, it should be pointed out that, in our studies, the PDK1 A280V -induced PKB phosphorylation at Ser 473 is normally less than 20% of that stimulated by insulin. In addition, it has previously been shown that a kinase-inactive, membrane-targeted mutant PKB can be phosphorylated at both Thr 308 and Ser 473 in unstimulated cells (9). Consistent with these results, we have found that a disruption of the kinase activity of the enzyme inhibited PKB phosphorylation at both Thr 308 and Ser 473 (Fig. 3). In addition, inactivation of PKB did not completely inhibit insulin-stimulated PKB phosphorylation at Ser 473 (Fig. 3). Therefore, it remains to be established whether the growth factor-stimulated stoichiometric phosphorylation of PKB at this site is mediated by autophosphorylation, by a distinct kinase, or by both.
It has recently been shown that, upon binding to the PIF fragment derived from PRK2, PDK1 is capable of phosphorylating PKB at Ser 473 (7). In agreement with this finding, we found that co-expression of PDK1 A280V with the PIF fragment led to an increase of PKB phosphorylation at Ser 473 (Fig. 1C,  second panel, lane 3 versus 5). However, we also found that co-expression of full-length PRK2, its isoform PRK1/PKN, or the carboxyl terminus of PRK1/PKN (PKN-AF3) inhibited PDK1 A280V -mediated PKB phosphorylation at both Thr 308 and Ser 473 (Fig. 1, B and C). These findings suggest the full-length PRK1/PKN and PRK2 play negative roles in PDK1-mediated phosphorylation of PKB in cells.