3-Phosphoinositide-dependent Protein Kinase 1, an Akt1 Kinase, Is Involved in Dephosphorylation of Thr-308 of Akt1 in Chinese Hamster Ovary Cells* 210

To investigate the role of 3-phosphoinositide-dependent protein kinase 1 (PDK1) in the Akt1 phosphorylation state, wild-type (wt) PDK1 and its kinase dead (kd) mutant were expressed using an adenovirus gene transduction system in Chinese hamster ovary cells stably expressing insulin receptor. Immunoblotting using anti-phosphorylated Akt1 antibody revealed Thr-308 already to be maximally phosphorylated at 1 min but completely dephosphorylated at 5 min, with insulin stimulation, whereas insulin-induced Akt1 activation was maintained even after dephosphorylation of Thr-308. Overexpression of wt-PDK1 further increased insulin-stimulated phosphorylation of Thr-308, also followed by rapid dephosphorylation. The insulin-stimulated Akt1 activity was also enhanced by wt-PDK1 expression but was maintained even at 15 min. Thus, phosphorylation of Thr-308 is not essential for maintaining the Akt1 activity once it has been achieved. Interestingly, the insulin-stimulated phosphorylation state of Thr-308 was maintained even at 15 min in cells expressing kd-PDK1, suggesting that kd-PDK1 has a dominant negative effect on dephosphorylation of Thr-308 of Akt1. Calyculin A, an inhibitor of PP1 and PP2A, also prolonged the insulin-stimulated phosphorylation state of Thr-308. In addition,in vitro experiments revealed PP2A, but not PP1, to dephosphorylate completely Thr-308 of Akt1. These findings suggest that a novel pathway involving dephosphorylation of Akt1 at Thr-308 by a phosphatase, possibly PP2A, originally, identified as is regulated downstream from PDK1, an Akt1 kinase.

Akt, also known as protein kinase B or Rac-PK, is a Ser/Thr kinase that was originally identified as a transforming oncogene in a retrovirus from a spontaneous thymoma in an AKR mouse (1). Activation of Akt1 by growth factors, such as insulin and IGF-1, 1 has been shown to be necessary for various cellular processes including cell growth, differentiation, metabolism, and apoptosis (2)(3)(4)(5)(6). To date, two potential in vivo substrates of Akt1 have been identified, namely glycogen synthase kinase-3 (GSK3) and BAD (4). GSK3 is inhibited by Akt1, which is thought to contribute to the insulin-induced dephosphorylation (activation) of glycogen synthase and protein synthesis initiation factor eIF2B and thereby to the stimulation of glycogen synthesis and protein synthesis (4). In addition, one of the cellular targets that Akt1 may phosphorylate to protect cells from apoptosis is BAD (7,8). This protein, in its dephosphorylated form, interacts with the Bcl family member BclXL and induces apoptosis of some cells; however, BAD, which is phosphorylated by Akt1, dissociates from BclXL, instead of interacting with 14-3-3, to prevent apoptosis (9). Furthermore, in transfection-based experiments, constitutively active Akt1 also mimics other actions of insulin, such as the enhancement of glucose uptake in 3T3-L1 adipocytes (10) and L6 myotubes (11) that results in the translocation of GLUT4 from an intracellular compartment to the plasma membrane.
Recently, a Ser/Thr protein kinase (3-phosphoinositide-dependent protein kinase-1; PDK1) was identified as an enzyme that phosphorylates Akt1 at Thr-308. PDK1, which is expressed ubiquitously in human tissue, is composed of an Nterminal catalytic domain and a noncatalytic C-terminal tail containing a PH domain (15,16). Recombinant PDK1 phosphorylates Thr-308 of Akt1 directly, in a reaction that is almost completely dependent on PtdIns(3,4,5)P 3 (15,17). In previous studies, effects of PDK1 on Akt1 kinase activity were well documented (14,16,18,19), but effects of PDK1 on phosphorylation of Akt1 at Thr-308 and Ser-473 in vivo have not been fully examined.
To investigate the role of PDK1 in the insulin-stimulated phosphorylation of Akt1 at Thr-308 and Ser-473 in vivo, wildtype PDK1 (wt-PDK1) or its kinase dead mutant (kd-PDK1) was expressed using an adenovirus gene transduction system in Chinese hamster ovary cells stably expressing insulin receptor (CHO-IR). We report herein a novel mechanism of Akt1 dephosphorylation in which PDK1, an Akt1 kinase, is involved.

EXPERIMENTAL PROCEDURES
Antibodies-The rabbit polyclonal anti-phospho-Akt1 (Thr-308) and anti-phospho-Akt1 (Ser-473) antibodies were purchased from New England Biolabs, and sheep polyclonal anti-PDK1 antibody was from Upstate Biotechnology, Inc. The rabbit polyclonal anti-HA antibody and the rabbit polyclonal anti-Myc antibody were purchased from Santa Cruz Biotechnology.
Cell Culture-Chinese hamster ovary cells in which the insulin receptor is stably overexpressed (CHO-IR cells) were maintained in Ham's F-12 medium (Life Technologies, Inc.) containing 10% fetal calf serum in an atmosphere of 5% CO 2 at 37°C.
Cloning and Construct-Polymerase chain reaction (PCR) was performed to amplify cDNA of PDK1 using the human liver cDNA as a substrate and oligonucleotides based on its reported sequence (15) as primers, yielding cDNA of PDK1 encompassing the entire coding region. cDNA corresponding to HA epitope (YPYDVPDYASL) was ligated to cDNA of PDK1 to tag it with the HA epitope at its N terminus. Mutant (K111A) of PDK1 was generated by standard PCR-based strategies. By using, as a probe, amplified cDNA fragments that had been generated by PCR based on the reported sequence of mouse Akt1 (20), full-length mouse Akt1 was isolated by screening a cDNA library from MIN6 cells (21). cDNA corresponding to the Myc epitope (MEQKLI-SEEDLEF) was ligated to cDNA of Akt1 to tag it with the Myc epitope at the C termini.
Gene Transduction-The epitope-tagged PDK1, its mutant, and the epitope-tagged Akt1 were constructed by homologous recombination between the expression cosmid cassette and the parental virus genome as described previously (22,23). CHO-IR cells were incubated in Ham's F-12 medium containing the adenoviruses at a multiplicity of infection (m.o.i.) of 10 -30 plaque-forming units/cell for 1 h at 37°C, and the growth medium was then added. Experiments were performed 48 -72 h after infection. Exogenous protein expression was observed in more than 90% of CHO-IR cells.
Immunoblotting-Cells, which had been serum-starved and then incubated with or without 1 M insulin, were lysed and boiled in Laemmli buffer containing 10 mM dithiothreitol, subjected to SDSpolyacrylamide gel electrophoresis, and transferred onto nitrocellulose filters. In some experiments, cells were pretreated with 25 nM calyculin A as indicated in the text. The filters were incubated with the indicated antibodies and then with anti-rabbit or anti-sheep immunoglobulin G coupled to horseradish peroxidase. The immunoblots were visualized with an enhanced chemiluminescence detection kit (Amersham Pharmacia Biotech). The intensities of bands were quantified by NIH Image 1.62 program. Values presented are the means Ϯ S.D. of three or four separate experiments.
PDK1 Activity Assay-Cells were solubilized in ice-cold lysis buffer A containing 50 mM Tris, pH 7.4, 100 mM NaCl, 10% glycerol, 1% Nonidet P-40, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 40 mM ␤-glycerophosphate, and 50 mM NaF. Lysates were extracted by centrifugation at 15,000 ϫ g for 10 min followed by immunoprecipitation with anti-PDK1 antibody or anti-HA antibody. The Ser/Thr kinase activities of the wild-type PDK1 and its mutant were assayed in the immunoprecipitates as previously reported (24) using a peptide sequence derived from the activation loop of Akt1 (KTFCGTPEYLAPE-VRR) as a substrate (24).
Akt1 Activity Assay-Cells were serum-starved, incubated with or without 1 M insulin for indicated periods, and solubilized in ice-cold lysis buffer B containing 50 mM Tris, pH 7.4, 100 mM NaCl, 10 mM EDTA, 10% glycerol, 1% Nonidet P-40, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 40 mM ␤-glycerophosphate, and 50 mM NaF. Lysates were extracted by centrifugation at 15,000 ϫ g for 10 min followed by immunoprecipitation with anti-Myc antibody. The Ser/ Thr kinase activity of the wild-type Akt1 was assayed in the immuno-precipitates as previously reported (25) using a peptide sequence derived from GSK3 (GRPRTSSFAEG) as a substrate (25).
PP1 and PP2A Activity Assays-Cells were serum-starved, incubated with or without 1 M insulin for 5 min, and resuspended in ice-cold phosphatase extraction buffer containing 20 mM imidazole HCl at pH 7.2, 2 mM EDTA, 0.2% ␤-mercaptoethanol, 2 mg/ml glycogen, 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml each of aprotinin, leupeptin, antipain, soybean trypsin inhibitor, and pepstatin A. The cells were homogenized with by 8 passages through a 29-gauge needle and centrifuged at 1,000 ϫ g for 5 min. Supernatants were centrifuged at 100,000 ϫ g for 30 min at 4°C to separate the particulate fractions from cytosolic fractions (26). PP1 and PP2A activities were assayed in the particulate fractions and cytosolic fractions using a protein serine/threonine phosphatase assay system (New England Biolabs). The activities of PP1 and PP2A were distinguished by their differential sensitivities to okadaic acid (OA). PP1 activity was defined as the activity that was insensitive to 5 nM OA. PP2A and PP2A-like activities were defined as the activity that was inhibited by 5 nM OA. As a control, protein phosphatase activity was also measured in the presence of 1 M OA, which completely inhibits both PP1 and PP2A (27).

RESULTS
To investigate the effects of PDK1 activity on Akt1 phosphorylation at Thr-308 and Ser-473, the wild-type (wt) PDK1 or the kinase-dead (kd) mutant of PDK1 was expressed into CHO-IR cells using an adenovirus gene transduction system, followed by immunoblotting using Thr-308 and Ser-473 phospho-specific antibodies. In parental CHO-IR cells, however, it was difficult to detect constantly the phosphorylation of endogenous Akt1 at Thr-308 with immunoblotting using anti-phospho-Akt1 (Thr-308) antibody (data not shown). Therefore, the experiments were performed in CHO-IR cells coexpressing the wild-type Akt1 tagged with Myc epitope using an adenovirus gene transduction system. First, in CHO-IR cells expressing wt-Akt1 alone (A cells), those expressing wt-PDK1 and wt-Akt1 (wtPA cells), and those expressing kd-PDK1 and wt-Akt1 (kdPA cells), the Ser/Thr kinase activity of PDK1 was measured in the immunoprecipitates with anti-tag antibody ( Fig.  1A) or anti-PDK1 antibody (Fig. 1B). Exogenous PDK1 proteins were also expressed at similar levels in wtPA cells and kdPA cells (Fig. 1C). Insulin did not stimulate either endogenous or exogenous PDK1 (Fig. 1, A and B). Overexpression of wt-PDK1 markedly increased kinase activity to 6-fold (Fig. 1B), which was consistent with the marked increase (approximately 8-fold) in the expression level of wt-PDK1 as demonstrated by immunoblotting with anti-PDK1 antibody (Fig. 1D). In contrast, kd-PDK1 exhibited no significant kinase activity (Fig.  1A). In addition, expression of kd-PDK1 did not affect endogenous PDK1 activity (Fig. 1B), demonstrating that the mutant does not have a dominant inhibitory effect on the Ser/Thr kinase activity of endogenous PDK1.
Next, we examined the effects of expression of wt-PDK1 or kd-PDK1 on the phosphorylation states of Akt1 at Thr-308 and Ser-473 after 15 min of insulin stimulation. By adjusting the transfection conditions including m.o.i., wt-Akt was expressed at a level similar to that in A cells, wtPA cells, and kdPA cells, as demonstrated by immunoblotting with anti-tag antibody (Fig. 2C). Insulin-induced phosphorylation of Akt1 at Thr-308 was not observed in A cells ( Fig. 2A). Overexpression of wt-PDK1 slightly enhanced insulin-stimulated Thr-308 phosphorylation of Akt1 ( Fig. 2A). To our surprise, a marked increase (an approximately 4-fold increase compared with the value in wtPA cells) was observed in the insulin-induced phosphorylation at Thr-308 in kdPA cells 15 min after insulin addition ( Fig.  2A), despite the absence of activity of kd-PDK1 (Fig. 1A). In contrast to the change in Thr-308 phosphorylation state, insulin-stimulated Ser-473 phosphorylation of Akt1 was clearly observed in A cells, and overexpression of wt-PDK1 or kd-PDK1 had virtually no effect on its change in phosphorylation (Fig. 2B). To confirm the involvement of kd-PDK1 in the enhancement of Thr-308 phosphorylation, we examined Thr-308 phosphorylation in CHO-IR cells with different kd-PDK1 expression levels after 15 min of insulin stimulation. Different levels of kd-PDK1 expression were achieved by infection of CHO-IR cells with kd-PDK1 adenovirus at different multiplicities (Fig. 3B), whereas wt-Akt1 was expressed at similar levels in these cells in response to infection with Akt1 adenovirus at similar multiplicities (Fig. 3C). The levels of Thr-308 phosphorylation were dependent on those of kd-PDK1 expression (Fig.  3, A and B). These findings indicate that kd-PDK1 expression enhances Thr-308 phosphorylation after 15 min of insulin stimulation. An approximately 3-fold increase in the level of Thr-308 phosphorylation was observed at 30 m.o.i. as compared with that at 20 m.o.i., whereas kd-PDK1 expression was increased less than 1.5-fold. This may be attributable to the mechanism whereby kd-PDK1 functions as a dominant inhibitory mutant for dephosphorylation of Akt1 at Thr-308.
Contrary to our expectation, expression of kd-PDK1, which exhibited no kinase activity (Fig. 1A), resulted in further enhancement of Thr-308 phosphorylation compared with expression of wt-PDK1 at 15 min of insulin addition ( Fig. 2A). To clarify the mechanism underlying this unexpected observation, we examined the phosphorylation states of Thr-308 and Ser-473 at earlier time points, 1, 3, 5, 10, and 15 min after insulin addition. Immunoblotting study revealed that Thr-308 was already maximally phosphorylated at 1 min after insulin addition but almost completely dephosphorylated by 5 min in A cells (Fig. 4A). In wtPA cells, phosphorylation of Thr-308 was markedly increased (ϳ1.8-fold) at 1 min after insulin addition, compared with A cells, but Thr-308 phosphorylation was also rapidly reversed (Fig. 4B). Akt1 phosphorylation at Thr-308 occurred in kdPA cells and was similar to that observed in A cells, at 1 min after insulin addition. However, the phosphorylation was maintained at 15 min (Fig. 4C) with even higher Thr-308 phosphorylation in kdPA cells (ϳ3.8-fold) than in wtPA cells. These results suggest that overexpression of wt-PDK1 enhances Thr-308 phosphorylation of Akt1 at earlier time points, such as 1 min, and that expression of kd-PDK1 inhibits dephosphorylation but does not affect phosphorylation.
On the other hand, phosphorylation of Ser-473, which was observed at 1 min after insulin addition, was maintained even at 15 min of insulin stimulation in A cells (Fig. 4D). Overexpression of wt-PDK1 or kd-PDK1 had no significant effects on insulin-induced Ser-473 phosphorylation throughout 15 min of insulin addition (Fig. 4, E and F). Thus, PDK1 is clearly involved in Thr-308 phosphorylation, whereas Ser-473 phosphorylation appears to be accomplished through a different mechanism in vivo.
Recent findings have shown that phosphorylation of both Thr-308 and Ser-473 is necessary for full activation of Akt1 (13). To investigate the relationship between the phosphorylation states of these residues and the Ser/Thr kinase activity of Akt1, we examined the time course of Akt1 activity in these cells. In A cells, activation of Akt1 was observed at 1 min after insulin addition and was maintained even at 15 min (Fig. 5,  dotted line). Overexpression of wt-PDK1 enhanced insulin-induced activation of Akt1, and this enhanced activity was maintained for 15 min (Fig. 5, dashed line), whereas overexpression of kd-PDK1 did not appear to affect insulin-stimulated Akt1 activity (Fig. 5, solid line); Akt1 activity was increased to similar extents in A cells and kdPA cells. Comparison of the time course of Akt1 activation (Fig. 5) with that of Thr-308 phosphorylation (Fig. 4, A-C) shows clearly that phosphorylation of Thr-308 is not essential for maintaining Akt1 activity once it has been initiated.
To unravel the mechanism whereby Thr-308 is dephosphorylated, we examined the effects of a Ser/Thr protein phosphatase inhibitor, calyculin A, on the Thr-308 phosphorylation of Akt. Calyculin A is reportedly cell-permeable and equally inhibits protein phosphatase (PP) 1 and PP2A in vivo (29). When A cells were pretreated with 25 nM calyculin A for 15 min, Thr-308 phosphorylation of Akt1 was prolonged even at 15 min after insulin addition (Fig. 6A), which resembled the observation in kdPA cells (Fig. 4C). These findings suggest that expression of kd-PDK1 inhibits protein phosphatase activity, resulting in the maintenance of Thr-308 phosphorylation.
Inhibition of Thr-308 dephosphorylation by calyculin A indicates that PP1 and/or PP2A is involved in Thr-308 dephosphorylation. We examined whether PP1 or PP2A activity is affected by insulin treatment and expression of wt-PDK1 or kd-PDK1. PP1 or PP2A activity was estimated in particulate and cytosolic fractions from CHO-IR cells, those expressing wt-PDK1, and those expressing kd-PDK1, with or without insulin using appropriate concentrations of okadaic acids as described under "Experimental Procedures." No significant differences in PP1 or PP2A activity were observed in each fraction either among these cells or between the basal state and insulinstimulated states (data not shown).
Then, we further examined which phosphatase was capable of dephosphorylating Akt1 at Thr-308 in vitro. The phosphorylated Akt1, stimulated by insulin, was incubated with purified PP1 or PP2A in vitro, followed by immunoblotting with anti-phospho-Akt1 (Thr-308) antibody. Thr-308 phosphorylation completely abolished by treatment with PP2A, whereas it persisted, although at a decreased level, with PP1 treatment (Fig. 6B). Thus, phosphorylated Thr-308 of Akt1 is a more preferable substrate for PP2A than for PP1. Taken together with the findings obtained using in vivo calyculin A treatment and in vitro dephosphorylation experiments, insulin-stimulated phosphorylation of Akt1 at Thr-308 might be reversed by PP2A, although we cannot rule out the possibility that another unknown phosphatase, which is inhibited by calyculin A, is involved.

DISCUSSION
In response to growth factor stimulation, Akt1 becomes phosphorylated at two major sites, Thr-308 in the kinase domain and Ser-473 in the C-terminal tail (13). PDK1 is a recently identified protein kinase that phosphorylates Akt1 at Thr-308 in lipid vesicles containing PtdIns(3,4,5)P 3 or PtdIns(3,4)P 2 (18). Overexpression of PDK1 in 293 cells reportedly potenti- ated the IGF-1-induced phosphorylation of Akt1 at Thr-308 in vivo (15). In the present study, we showed overexpression of wt-PDK1 to potentiate the insulin-stimulated phosphorylation of Akt1 at Thr-308 at 1 min, but this was rapidly followed by dephosphorylation. In contrast, wt-PDK1 overexpression did not affect Ser-473 phosphorylation in the basal or insulinstimulated condition. Furthermore, kd-PDK1 expression did not apparently affect Ser-473 phosphorylation. Thus, PDK1 is clearly involved in Thr-308 phosphorylation, whereas Ser-473 phosphorylation appears to occur through a different mechanism in vivo. Contrary to our expectation, expression of the kinase dead mutant of PDK1 did not inhibit but rather maintained the insulin-stimulated phosphorylation of Thr-308. In addition, the levels of Thr-308 phosphorylation paralleled those of kd-PDK1 expression. Furthermore, calyculin A, an inhibitor of both PP1 and PP2A, prolonged the phosphorylation state of Thr-308. Thus, kd-PDK1 expression has a dominant inhibitory effect on Ser/Thr protein phosphatase for Thr-308 of Akt1, suggesting that PDK1, an Akt kinase, also regulates dephosphorylation of Akt1 at Thr-308, resulting in a transient Thr-308 phosphorylation in CHO-IR cells. It was reported that phosphorylation of Thr-308 of Akt1 was observed at 10 min after insulin (13) and IGF-1 (13,15) stimulation in 293 cells. We also observed the persistent phosphorylation of Thr-308 in 3T3-L1 adipocytes at 15 min after insulin stimulation (data not shown). However, in CHO-IR cells, at a more modest insulin concentration (10 nM), phosphorylation of Thr-308 was rapidly reversed, and kd-PDK1 expression prolonged Akt1 phosphoryl-ation at Thr-308 (see Supplemental Material) in a fashion similar to that observed at 1 M insulin (Fig. 4, A and C). Thus, it is unlikely that the markedly transient phosphorylation of Akt1 at Thr-308 involves inappropriate coupling of insulin receptors to intracellular targets and/or effects of insulin binding nonspecifically to related receptors. The discrepancy among cell types might be due to differences in the levels of expression and/or subcellular distribution of putative signaling molecules that transduce signals from PDK1 to the downstream phosphatase.
It appears that overexpression of PDK1 is less effective in the dephosphorylation of Thr-308 than in the phosphorylation of this residue, since the levels of Akt1 phosphorylation at Thr-308 were higher, at all times after insulin stimulation, in wtPA cells (Fig. 4B) than in A cells (Fig. 4A). Whereas PDK1 directly phosphorylates Thr-308 of Akt1, dephosphorylation of Thr-308 perhaps occurs via endogenous phosphatase activation. Therefore, this discrepancy between the effects of PDK1 on phosphorylation and dephosphorylation of Thr-308 might be attributable to endogenous phosphatase activity possibly being rate-limiting for dephosphorylation of Akt1 at Thr-308, under conditions of wt-PDK1 overexpression.
Based on biochemical characteristics, Ser/Thr protein phosphatases were initially divided into two classes; type 1 phosphatases (PP1) are inhibited by two heat-stable proteins, termed inhibitor-1 and inhibitor-2, and preferentially dephosphorylate the ␤-subunit of phosphorylase kinase, whereas type-2 phosphatases are insensitive to heat-stable inhibitors and preferentially dephosphorylate the ␣-subunit of phosphorylase kinase (30,31). Type 2 phosphatases can be further subdivided into spontaneously active (PP2A), Ca 2ϩ -dependent (PP2B), and Mg 2ϩ -dependent (PP2C) classes (32). More than 30 protein kinase activities are known to be modulated by PP2A in vitro (33), suggesting that PP2A plays a major role in kinase regulation. In fact, extensive biochemical, pharmacological, and genetic evidence suggests that PP2A controls the activities of several major protein kinase families, particularly those that belong to the AGC subgroup, the calmodulin-dependent kinases (34 -36), extracellular signal-regulated kinase/mitogen-activated protein kinases (37), cyclin-dependent kinases (38,39), and IB kinase (40). It is noteworthy that Akt1, protein kinase C, and p70 S6 kinase, which belong to AGC subgroup, are all reportedly substrates of PDK1 (18,41,42). Although previous studies have shown treatment of Akt1 with PP2A in vitro to inactivate kinase (28), its exact phosphorylation states of Thr-308 and Ser-473 in vivo remain unclear. In the present study, we showed calyculin A to inhibit dephosphorylation of Akt1 at Thr-308, which occurred in a fashion similar to that of kd-PDK1 expression. In addition, under conditions of identical enzyme activity, PP2A completely dephosphorylated Akt1 at Thr-308 in vitro, whereas Thr-308 of Akt1 was not completely dephosphorylated by PP1 despite the loose substrate specificity of the catalytic subunit of Ser/Thr protein phosphatases (32). These results suggest that phosphorylated Akt1 at Thr-308 is a more preferable substrate for PP2A than for PP1. These findings are in agreement with the recent report that PP2A is the major protein phosphatase responsible for hyperosmotic stress-induced dephosphorylation of Akt1 at both Thr-308 and Ser-473 (43). We measured the PP2A activity, but we detected no insulin stimulation of PP2A activity in CHO-IR cells. In addition, overexpression of the wild-type or kinasedead PDK1 also had no effect on PP2A activity (data not shown). However, Begum and co-workers (44,45) have recently shown that addition of insulin to rat skeletal muscle cells decreases PP2A activity by 40 -80% (44), whereas in fetal chick neurons, insulin increases PP2A activity (45). Although differences in these results may depend on cell types, these reports suggest that PP2A activity can be modulated by extracellular signals such as insulin. PP2A is a multimeric protein, and the catalytic subunit of PP2A forms a complex with an array of regulatory subunits that modulate its activity, substrate specificity, and subcellular localization (33). The phosphatase activity assay in extracts from whole cells does not necessarily reflect the PP2A activity in the specific subcellular loci. Thus, the lack of a significant change in the PP2A activity of whole cytosolic and particulate fractions does not exclude the possibility that, upon insulin stimulation, PP2A in specific subcellular loci is activated downstream from PDK1 in response to insulin in vivo. The proportion of the regulatable PP2A might be different in other cell types. Alternatively, it is also possible that, without altering phosphatase activity, a portion of PP2A changes subcellular localization upon insulin stimulation, thereby gaining access to the phosphorylated Akt1 at Thr-308, and this cellular process is inhibited by kd-PDK1.
It was reported that mutation of either Thr-308 or Ser-473 to Ala greatly decreased the activation of transfected Akt1 by insulin. In addition, Akt1 became partially active in vitro when either Thr-308 or Ser-473 was changed to Asp, and far more active when Thr-308 and Ser-473 were both mutated to Asp, suggesting a critical role of phosphorylation of both the Thr-308 and the Ser-473 residues in full activation of Akt1 (13). However, in the present study, Akt1 activation was prolonged even after dephosphorylation of Thr-308, although basal and insulin-stimulated Akt1 activities were enhanced by overexpression of wt-PDK1. Thus, a discrepancy was observed between the Akt1 phosphorylation at Thr-308 and Akt1 activity. In contrast, phosphorylation of Ser-473 was prolonged in a fashion similar to that of Akt1 activity. These findings suggest that initial activation of Akt1 depends on its phosphorylation states, including that of Thr-308, whereas the maintenance of once activated Akt1 activity is via a mechanism different from that of initial activation. Thr-308 phosphorylation is not essential for maintenance of the Akt1 activity, in which Ser-473 phosphorylation might be involved. PDK1 reportedly phosphorylates the activation loop of PKC␣ and PKC␤II, which is homologous to the sequence around Thr-308 of Akt1, and this phosphorylation is a step required for the maturation of PKC␣ and PKC␤II. This phosphorylation serves as a trigger to induce the two C-terminal autophosphorylations that are required to lock PKC in a catalytically competent conformation, resulting in full activation. Once this mature conformation is achieved, the phosphate group on the activation loop does not regulate the activity of the enzyme (46). Thus, phosphorylation of Thr-308 of Akt1, like phosphorylation of the activation loop of PKC␣ and PKC␤II by PDK1, may serve as a trigger for maturation of Akt1, although phosphorylation of Thr-308 of Akt1 does not lead to autophosphorylation of its Ser-473 (13).
It was reported that kd-PDK1 does not interfere with PKC activator (phosphatidylserine, diacylglycerol, and Ca 2ϩ )induced PKC␣ and PKC␤II activation (46). On the other hand, kd-PDK1 has a dominant negative effect on PKC activator (phosphatidylserine, phosphatidylcholine, and PtdIns(3,4,5)P 3 )-induced PKC (41) and insulin-induced p70 S6-kinase activation (42). In the present study, kd-PDK1 was shown to have a dominant inhibitory effect on dephosphorylation of Akt1 but not on phosphorylation of the same molecule or on endogenous PDK1 activity. These observations suggest that mechanisms whereby PDK1 transduces signals to downstream targets depend on the target molecule.
In summary, we have reported herein that, upon insulin stimulation, phosphorylated Thr-308 of Akt1 is rapidly dephosphorylated downstream from PDK1, an Akt kinase, and that phosphorylation of Thr-308 is not essential for maintaining the Akt1 activity once it has been achieved. Lysates were immunoblotted (Blot) with anti-phospho-Akt1 (Thr-308) antibody, and densitometry was performed on the original blots as described under "Experimental Procedures." The graph shows the intensities of bands, and data are expressed as percentages of the intensity at 15 min after insulin stimulation. Values presented are the means Ϯ S.D. of three separate experiments. The representative immunoblotting is also presented (upper panels). B, effects of PP1 or PP2A on phosphorylated Thr-308 of Akt1 in vitro. A cells were treated with 1 M insulin for 1 min. Whole cell lysates were immunoprecipitated with anti-Myc antibody. Immunoprecipitates were then incubated with or without 1.7 units/ml PP1 or PP2A for 1 h, followed by electrophoresis and immunoblotting with anti-phospho-Akt1 (Thr-308) antibody, and densitometry was performed on the original blots as described under "Experimental Procedures." The graph shows the intensities of bands, and data are expressed as percentages of the intensity after incubation without protein phosphatases. Values presented are the means Ϯ S.D. of three separate experiments. The representative immunoblotting is also presented (upper panels).