INTRODUCTION

Akt is a serine/threonine kinase that is stimulated by several receptor tyrosine kinases, including the receptors for insulin and PDGF¹ (1, 2, 3, 4). Recent studies indicate that Akt functions downstream of phosphatidylinositol 3-kinase (PI 3-kinase), a lipid kinase that is controlled by receptor tyrosine kinases (1, 2, 3, 4). This PI 3-kinase is a heterodimer consisting of a regulatory p85 subunit and a catalytic p110 subunit (5, 6). The p85 subunit links p110 to activated tyrosine kinase receptors via its two SH2 domains. It transmits the activating signal to the p110 subunit using the inter SH2 (iSH2) segment that lies between the two SH2 domains. The p110 catalytic subunit mediates the phosphorylation of phosphoinositides (PtdIns) at the 3-position of the inositol ring to generate PtdIns3P, PtdIns(3,4)P₂, and PtdIns(3,4,5)P₃. Three different experimental approaches place Akt downstream of PI 3-kinase. First, activation of Akt kinase activity is inhibited by wortmannin, an inhibitor of PI 3-kinases (1, 2, 3). Second, PDGF receptor mutants that fail to activate PI 3-kinase also fail to activate Akt (1, 2). Finally, a dominant-negative PI 3-kinase prevents the stimulation of Akt kinase activity by PDGF (2).

Since several independent approaches indicate that Akt functions downstream of PI 3-kinase, a model has been proposed for the mechanism of growth factor stimulation of Akt kinase. First, agonist exposure leads to activation of the p110 catalytic subunit of PI 3-kinase via interaction of the p85 subunit with tyrosine-phosphorylated proteins (5, 6). Second, the activated p110 subunit phosphorylates phosphoinositides on the 3-position of the inositol ring. Subsequently, these lipid products activate Akt. The activated Akt may then regulate the p70 S6 kinase and/or the glycogen synthase kinase-3 (2, 4) .

A central question is whether PtdIns 3-phosphates directly or indirectly activate Akt kinase. Franke et al. (1) suggest that the PtdIns 3-phosphates directly activate Akt since PtdIns3P, one of the lipid products generated by PI 3-kinase, can partially activate Akt in vitro. They further propose that the pleckstrin homology (PH) domain at the amino terminus of Akt mediates this PI 3-phosphate binding since they found this domain to be essential for PDGF-directed Akt activation in NIH 3T3 cells (1). However, we found the PH domain to be unnecessary for insulin-mediated Akt activation in CHO cells expressing either endogenous levels or overexpressed levels of insulin receptor (3). We also could not detect specific binding of PtdIns 3-phosphate to Akt.² James et al. (7) have also recently reported that they could not activate Akt in vitro with PtdIns 3-phosphate. Thus, it is not clear whether PtdIns 3-phosphates directly activate Akt.

Burgering and Coffer (2) propose that Akt is upstream of p70 S6 kinase (p70^S6K) since a constitutively active Akt increased the activity of p70^S6K in the absence of growth factor treatment. The constitutively active version of Akt used in these studies has been previously identified as the transforming component of the AKT8 virus, an acute transforming retrovirus isolated from a rodent T cell lymphoma (8). In this clone (v-Akt), the Gag polypeptide is fused in-frame to the 5-untranslated portion of the Akt gene such that all of the Akt coding sequence is retained, including the Akt PH domain. The Gag polypeptide is myristoylated, and it has been shown that fusion of Gag to Akt targets the protein to the membrane (9). It is uncertain whether the constitutive activity of Akt fused to Gag is due to the membrane targeting of the kinase or a conformational change induced by Gag. If Gag can alter the conformation of Akt, it could also alter the catalytic specificity of the Akt kinase domain. Furthermore, it is unknown whether Akt requires its PH domain to interact with putative downstream target proteins.

In the present study, we show in several cell types that a variant Akt which lacks a PH domain can be activated by either PDGF or insulin. This Akt variant became constitutively active when targeted to the membrane via the addition of the 14-amino acid src myristoylation signal (10). This activation can be attributed to an increase in the level of Akt phosphorylation. This constitutively active form of Akt can stimulate p70^S6K in the absence of growth factor treatment, indicating that the PH domain is not required for downstream substrate interactions. We have also constructed another constitutively active mutant of Akt by fusing the iSH2 domain derived from the p85 PI 3-kinase regulatory subunit (11, 12) to the amino terminus of Akt. This mutant was found to be associated with a constitutively active PI 3-kinase activity. Phosphorylation of this fusion protein still critically contributes toward its increased enzymatic activity. The sum of these results indicates that phosphorylation is the primary mechanism of Akt kinase activation.

EXPERIMENTAL PROCEDURES

Restriction and modifying enzymes were purchased from New England Biolabs, Life Technologies, Inc., and Stratagene; oligonucleotides from Operon Technologies; Sequenase and DNA sequencing kit from United States Biochemical; pET 17b expression vector and anti-T7 epitope tag monoclonal antibody from Novagen; Takara DNA ligation kit from Panvera; Protein G-Sepharose, dNTPs, and DEAE-dextran from Pharmacia Biotech Inc.; chloroquine, dimethyl sulfoxide, myelin basic protein, protein kinase A inhibitor peptide, and ATP from Sigma; LipofectAMINE and Optimem I reduced serum media from Life Technologies, Inc.; [-³²P]ATP (3000Ci/mmol) from Amersham; calf intestinal alkaline phosphatase, human recombinant insulin, PDGF BB, and the 12CA5 monoclonal antibody from Boehringer Mannheim; anti-HA HA.11 polyclonal antibody from Babco; thin layer silica gel plates (0.2-mm thickness) from E. Merck; Protein A-Sepharose from Repligen; pure PI from Avanti. The Phoenix retroviral packaging cell line and the pWZLneo retroviral vector were from Dr. Garry P. Nolan, py20 was from Dr. John Glenney, the plasmid encoding p70^S6K was from Dr. Gerry Crabtree, the TRMP cells overexpressing the PDGF receptor were from Dr. Andrius Kazlauskas, the Akt R25C and Akt 11-35 constructs were from Drs. Tsichlis and Franke (1), and the p85 cDNA was a gift of Dr. Edgar Wood.

Addition of the hemagglutinin (HA) epitope tag to Akt kinase and generation of the Akt mutant bearing alterations in the pleckstrin homology domain were generated as described (1, 3). Akt kinase was epitope-tagged with the T7 peptide by digesting with NcoI to remove the HA tag and ligating the available Klenow-blunted 3 end to sequence derived from the pET 17b vector that encoded the T7 tag.

The src myristoylation signal sequence (10) was added to Akt 4-129 using the polymerase chain reaction to generate myrAkt 4-129 and A2myrAkt 4-129. The start methionine of Akt 4-129 was removed, and the new start methionine encoded by the myristoylation signal peptide was flanked by the Kozak consensus sequence. A2myrAkt 4-129 differs from myrAkt 4-129 in that A2myrAkt 4-129 encodes an alanine at amino acid position 2, instead of a glycine.

The iSH2 segment from the p85 subunit of PI 3-kinase (11, 12) was added to the amino terminus of Akt kinase by polymerase chain reaction. The iSH2 segment extends between base pairs 1438 and 1743 of human PI 3-kinase. The amino terminus was modified to encode a start methionine surrounded by a Kozak consensus sequence. The start methionine of Akt kinase was removed. Between the iSH2 segment and Akt kinase was a glycine hinge consisting of 7 glycine residues. The point mutation K179M in Akt kinase was also introduced by polymerase chain reaction. All polymerase chain reaction-derived fragments were confirmed by sequencing.

CHO.IR cells were maintained in F12 Ham's media containing 10% newborn calf serum and 1% penicillin-streptomycin in 5% CO₂. CHO.IR cells were transiently transfected essentially as described previously (3). Following transfection, the cells were incubated for 24 h before being used in subsequent experiments.

TRMP canine kidney epithelial cells overexpressing the PDGF-R were maintained in DME H21 containing 10% fetal bovine serum and 1% penicillin-streptomycin in 5% CO₂. Transient transfections were performed in 100-mm Petri dishes using DEAE-dextran as described above for CHO.IR cells, except the cells were incubated with the DEAE-dextran/chloroquine/DNA mixture in complete media for 2.5 h and the cells were not dimethyl sulfoxide-shocked.

NIH 3T3 cells stably expressing Akt constructs were generated by retroviral infection. The Phoenix packaging cell line was transfected as described previously (13), except the final concentration of chloroquine used was 50 µM, and, after the final media change, the cells were incubated at 32 °C for 48 h to generate the viral supernatant. For the infection, 7.5 × 10⁴-10⁵ NIH 3T3 cells were plated in each well of a 6-well plate, and 1 ml of viral supernatant was added to each well, in addition to 0.5 ml of complete media and 5 µg/ml polybrene. The plates were spun at 2500 rpm for 1.5 h at room temperature. They were then incubated at 32 °C for 24 h after which the medium was changed and the cells were incubated at 37 °C for another 24 h. The cells were then split into 100-mm dishes in complete media containing 0.4 mg/ml G418. The total pool of selected cells was used in all subsequent experiments.

The Akt kinase assay and immunoblotting were performed as described previously (3), except that a single plate was used for each treatment, and each plate was lysed in 400 µl of lysis buffer. Akt was treated with alkaline phosphatase also as described (3) except the immunoprecipitates were incubated in phosphatase buffer with or without 20 units of calf intestinal alkaline phosphatase for 5 min at 37 °C with shaking.

NIH 3T3 cells were transiently cotransfected using 20 µl of LipofectAMINE (Life Technologies, Inc.) and 3 µg of the HA epitope-tagged p70^S6K plasmid and either empty expression vector or the indicated T7 epitope-tagged Akt constructs. The plates were washed once with PBS and then starved for 18 h at 37 °C by adding 5 ml of serum-free DME H21 containing 20 mM Hepes, pH 7.4, and 1% penicillin-streptomycin. At the end of the incubation period, plates were incubated for 10 min at 37 °C with 10% calf serum where indicated. The plates were washed on ice once with HEPES-buffered saline (HBS pH 7.4), and the cells were lysed in 0.5 ml of lysis buffer (50 mM HEPES, pH 7.6, 1 mM EDTA, 5 mM EGTA, 10 mM MgCl₂, 50 mM -glycerophosphate, 1 mM Na₃VO₄, 10 mM NaF, 30 mM NaPP_i, 2 mM DTT, 1 mM phenylmethylsulfonyl fluoride). The cell lysates were sonicated (Heat Systems-Ultrasonics, Inc., Model W-225R) for 40 s, spun in a Microfuge for 15 min at 14,000 rpm, 4 °C, and incubated with 4 µg of either control mouse Ig or monoclonal anti-HA antibody bound to Protein A-Sepharose. After 3 h, the beads were washed three times with cold HBS pH 7.4 and 30 µl of ribosome mixture (20 mM Tris, pH 7.5, 10 mM MgCl₂, 1 mM DTT, 5 µM ATP, 0.4 µM 40 S ribosome, and 3 µCi of [-³²P]ATP) was added and the samples were incubated for 5 min at 30 °C. The samples were mixed with 3× Laemmli sample buffer, boiled for 4 min at 95 °C, and electrophoresed on 12% SDS-polyacrylamide gels. The band corresponding to the p70^S6K 40 S ribosomal substrate was cut out and counted.

Transiently transfected NIH 3T3 cells were washed once with PBS and then serum-starved for 2 h at 37 °C by adding 3 ml of serum-free DME H21 containing 20 mM Hepes, pH 7.4, and 1% penicillin-streptomycin to each plate. At the end of the incubation, 1 mg/ml bovine serum albumin was added to every plate and 50 ng/ml PDGF BB to the plates indicated in the figure legends. The plates were incubated for 10 min at 37 °C and then washed once on ice with HBS pH 7.4. The cells were lysed in 400 µl of lysis buffer (20 mM Tris pH 8.0, 137 mM NaCl, 1 mM MgCl₂, 1 mM CaCl₂, 10% glycerol, 0.1% Nonidet P-40, 1 mM DTT, 0.4 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride) and then sonicated 30 s. After 10 min at 0 °C, the lysates were microfuged for 5 min and incubated with either anti-phosphotyrosine antibody py20, normal mouse Ig, or 12CA5 anti-HA antibody preadsorbed to Protein A-Sepharose, as described above. After 3 h at 4 °C, the beads were washed twice with buffer 1 (PBS pH 7.4, 0.1% Nonidet P-40, 1 mM DTT) and twice with buffer 2 (10 mM Tris pH 7.6, 0.1 M NaCl, 1 mM DTT) and then used in a PI 3-kinase assay as described previously (14).

RESULTS AND DISCUSSION

We previously reported that the PH domain of Akt was unnecessary for the insulin-stimulated increase in Akt kinase activity in CHO cells that express only endogenous receptors or that overexpress the human insulin receptor (CHO.IR) (3). In contrast, Franke et al. (1) reported that the Akt PH domain was necessary to achieve a PDGF-mediated increase in kinase activity in NIH 3T3 cells (1). A number of possibilities could explain these opposite results. First, the PH domain mutants used in these studies were not the same (Fig. 1A). The construct tested in CHO cells lacked amino acids 4-129 (Akt 4-129) whereas the PH domain mutants tested in NIH 3T3 cells carried either a deletion of amino acids 11-35 (Akt 11-35) or an arginine to cysteine point mutation at amino acid 25 (Akt R25C), comparable to a mutation in the PH domain of Bruton's tyrosine kinase associated with the induction of X-linked immunodeficiency in CBA/N mice (15). Conformational changes caused by the introduction of these mutations might abrogate the growth factor-stimulated increase in kinase activity. A second difference in these studies was the particular growth factor used. The Akt PH domain may not be necessary to achieve an increase in kinase activity in response to insulin, but it may be essential for a PDGF-directed increase in kinase activity. Finally, the PH domain may be required only in certain cell types to achieve a growth factor-mediated increase in Akt kinase activity.

To address these questions, we compared the insulin responsiveness of the three different PH domain mutants in CHO.IR cells. Cells were transiently transfected with either wild type Akt (Akt wt), Akt 4-129, Akt R25C, or Akt 11-35, treated or not with insulin, and the expressed Akt was immunoprecipitated and assayed in vitro for kinase activity. The kinase activities of both Akt wt and Akt 4-129 were stimulated by insulin approximately 7-fold (Fig. 2A). Akt R25C also showed an insulin-induced increase in kinase activity, although the magnitude of this response was not as great as either Akt wt or Akt 4-129. Akt 11-35 completely failed to respond to insulin (Fig. 2A), even though it was expressed at levels comparable to Akt R25C (data not shown), suggesting that this deletion mutant may have an inactivating conformational change that abolishes its ability to respond to agonist stimulation.

4-129

11-35

4-129

4-129

To test the possibility that Akt requires its PH domain to respond to PDGF but not insulin, the kinase activity of Akt 4-129 and Akt wt were compared by transiently expressing these constructs in TRMP canine kidney epithelial cells that overexpress the PDGF receptor (16). After the cells were treated without or with PDGF, Akt was immunoprecipitated and assayed in vitro for kinase activity. Both Akt 4-129 and Akt wt were stimulated approximately 4-fold by PDGF (Fig. 2B). The elevated basal activity of Akt 4-129 correlates with its higher level of protein expression compared to that of Akt wt. These results indicate that Akt is capable of responding to PDGF without its PH domain.

To determine if Akt requires its PH domain to respond to growth factors only in specific cell types, we studied the ability of PDGF to stimulate the kinase activity of Akt 4-129 and Akt wt in NIH 3T3 cells, the cell type used previously to establish a role for the PH domain in the PDGF-mediated activation of Akt (1). Stable pools of cells expressing roughly equivalent levels of these two constructs were generated by drug selection after retroviral infection of NIH 3T3 cells. In these cells, PDGF activated to the same extent the Akt wt and Akt lacking its PH domain (Fig. 2C).

Targeting Akt4-129 to the Membrane Results in a Constitutively Active Kinase

Bellacosa et al. (8) proposed that v-Akt, which contains the PH domain, may be transforming because the myristoylated Gag polypeptide that is fused in-frame to the 5-untranslated region of Akt targets the kinase to the membrane (8). To determine if membrane targeting alone, without the Gag polypeptide and the PH domain, was sufficient to confer constitutive kinase activity, we fused the 14-amino acid src myristoylation signal sequence (10) to the amino terminus of Akt 4-129 (myrAkt 4-129). As a control, we also created an Akt 4-129 construct fused to an inactive src myristoylation signal (A2myrAkt 4-129) in which the myristoylated glycine at the second amino acid position was converted to alanine (Fig. 1B).

Both forms of the enzyme were transiently expressed in either CHO.IR or NIH 3T3 cells, immunoprecipitated from either untreated or growth factor-treated cells, and the precipitates were assayed in vitro for kinase activity. Akt bearing the myristoylation signal, myrAkt 4-129, had a significantly elevated activity even in the absence of any growth factor treatment in both CHO.IR and NIH 3T3 cells (Fig. 3A). In contrast, the enzyme containing the inactive myristoylation signal, A2myrAkt 4-129, had the same basal kinase activity as Akt 4-129 (Fig. 3A). Thus, targeting Akt to the membrane confers constitutive kinase activity, even in the absence of the PH domain.

Effect of membrane targeting on Akt 4-129 kinase activity and gel migration.

The immunoprecipitated proteins assayed for their kinase activities were also separated by SDS-polyacrylamide gel electrophoresis to evaluate the samples for protein expression by immunoblotting using the anti-HA HA.11 polyclonal antibody. Akt 4-129 migrates as a doublet at a molecular mass of approximately 45 kDa (Fig. 3B). A2myrAkt 4-129 also appears as a doublet but migrates slightly slower, consistent with the addition of 14 amino acids that contain the myristoylation signal sequence. In contrast, myrAkt 4-129 migrates as a single band that is aligned with the upper band of the A2myrAkt 4-129 doublet (Fig. 3B).

Treatment of myrAkt 4-129 with alkaline phosphatase causes the single upper band to migrate as a doublet like the A2myr 4-129 (Fig. 3B), demonstrating that the upper band bears additional phosphates not present on the lower band, accounting for its slower mobility. This suggested that myrAkt 4-129 may be constitutively active because it is maximally phosphorylated, even without growth factor treatment. To test this hypothesis, HA epitope-tagged myrAkt 4-129 protein transiently expressed in NIH 3T3 cells was immunoprecipitated from cell lysates, treated with either buffer alone, alkaline phosphatase alone, or phosphatase that had been premixed with phosphatase inhibitors, and then assayed for enzymatic activity. The phosphatase treatment reduced the enzymatic activity of myrAkt 4-129 by 76%, and this phosphatase effect was prevented by the presence of phosphatase inhibitors (Fig. 3C). These results argue that the constitutive activity of myrAkt 4-129 is principally due to its hyperphosphorylation.

Since Akt has been implicated in contributing to the activation of p70^S6K (2), we wanted to determine if the constitutively active construct, myrAkt 4-129, could also activate p70^S6K, even though it lacks a PH domain. A plasmid encoding HA epitope-tagged p70^S6K was transiently cotransfected in NIH 3T3 cells with either an empty expression vector or different Akt mutants bearing T7 epitope tags. The p70^S6K protein was immunoprecipitated from cell lysates using the anti-HA monoclonal antibody, and the kinase activity was measured in vitro using the 40 S ribosomal subunit as a substrate. Cotransfection of p70^S6K with an empty vector followed by treatment of cells with 10% calf serum resulted in a 2-fold increase in p70^S6K activity compared to p70^S6K activity isolated from untreated cells (Fig. 4). The p70^S6K activity isolated from NIH 3T3 cells that had not received any additional stimulation was essentially the same whether p70^S6K was coexpressed with an empty expression vector, Akt 4-129, or A2myrAkt 4-129. However, when coexpressed with the constitutively active protein myrAkt 4-129, the p70^S6K activity was elevated approximately 2-3-fold in the absence of any additional stimulation (Fig. 4). The increase in p70^S6K activity can be attributed to coexpression with myrAkt 4-129 and not to differences in p70^S6K protein expression levels, which were determined to be the same by Western blotting of immunoprecipitated p70^S6K isolated from plates transfected in parallel (data not shown). Since a constitutively active form of Akt that lacks the PH domain can still activate p70^S6K, these results demonstrate that the PH domain is not necessary for Akt to interact with and activate a downstream target.

Effect of myrAkt 4-129 on p70

4-129

A2myr 4-129

myr 4-129

To further define the role of PI 3-kinase in stimulating the enzymatic activity of Akt, we fused the iSH2 domain of PI 3-kinase to the amino terminus of Akt wt (Fig. 1C) since it has been shown that this domain alone is sufficient to cause constitutive activation of the p110 catalytic subunit (12). To compare the activity of (iSH2)Akt to Akt wt, both plasmids were transiently transfected into NIH 3T3 cells, the cells were either left untreated or treated with PDGF, lysed, and the proteins were immunoprecipitated and assayed for kinase activity. The kinase activity of (iSH2)Akt in the absence of any PDGF stimulation was comparable to the level of activity of Akt wt isolated from cells that had been treated with growth factor and PDGF did not substantially increase its activity (Fig. 5A). The protein expression level of (iSH2)Akt was comparable to that of Akt wt, as determined by immunoblotting using the HA.11 polyclonal antibody (data not shown).

myr 4-129

To ensure that the in vitro assay was measuring the kinase activity of (iSH2)Akt, rather than some associated kinase, we compared the activity of (iSH2)Akt to that of (iSH2)Akt K179M, a kinase-deficient derivative of (iSH2)Akt in which lysine 159 in the ATP-binding site has been converted to a methionine (Fig. 1C). Transiently expressed (iSH2)Akt and (iSH2)Akt K179M were immunoprecipitated from NIH 3T3 cell lysates and assayed for kinase activity in vitro. Whereas the kinase activity of (iSH2)Akt was elevated compared to Akt wt in the absence of growth factor stimulation, the activity of (iSH2)Akt K179M was comparable to the basal activity of Akt wt (data not shown).

We hypothesized that (iSH2)Akt was associating with and activating the p110 catalytic subunit of PI 3-kinase. The constitutively activated p110 subunit, in turn, would activate Akt. To verify that (iSH2)Akt associated with a constitutively active phosphatidylinositol lipid kinase, (iSH2)Akt was transiently expressed in NIH 3T3 cells, the cells were either left untreated or treated with PDGF, lysed, and the fusion protein was immunoprecipitated and assayed for PI kinase activity. For comparison, transiently expressed myrAkt 4-129 was also immunoprecipitated from PDGF-treated cells and used in the same assay. In addition, to measure the amount of PDGF-stimulated PI 3-kinase activity, the anti-phosphotyrosine antibody py20 was incubated with untreated or PDGF-treated lysates from NIH 3T3 cells that had been transiently transfected with the empty expression vector. As expected, treatment of cells with PDGF resulted in a substantial increase in the amount of anti-phosphotyrosine-precipitable PI kinase activity (Fig. 5B). The PI kinase activity associated with (iSH2)Akt was clearly elevated compared to the levels observed in the anti-phosphotyrosine precipitates from untreated cells, and, furthermore, this associated activity was approximately the same whether or not the cells were pretreated with PDGF (Fig. 5B). In contrast, there was no PI kinase activity associated with the constitutively active myrAkt 4-129 immunoprecipitated from PDGF-treated cells (Fig. 5B).

To determine if phosphorylation still played a role in mediating the activation of (iSH2)Akt, the transiently expressed protein was immunoprecipitated without any prior additional stimulation using the anti-HA antibody. For comparison, Akt wt was also immunoprecipitated from cell lysates of transiently transfected NIH 3T3 cells that were either left untreated or treated with PDGF. Before using the immunoprecipitates in a kinase assay, antibody-bound (iSH2)Akt was either dephosphorylated using alkaline phosphatase, or treated with phosphatase that had been premixed with phosphatase inhibitors. In addition, both (iSH2)Akt and Akt wt were also mock-treated with buffer alone as a control. The kinase activity of mock-treated (iSH2)Akt from unstimulated cells was comparable to that of Akt wt isolated from cells treated with growth factor, and the activity of both was elevated compared to the basal activity of Akt wt from untreated cells. Phosphatase treatment of (iSH2)Akt reduced its kinase activity by about 60%. This reduction was prevented by premixing the phosphatase with phosphatase inhibitors (Fig. 5C). Thus, the elevated kinase activity of (iSH2)Akt is still dependent on its phosphorylation, even under conditions when the protein should be in close proximity to PtdIns lipid products generated by the constitutively active p110 PI 3-kinase.

CONCLUSIONS

In the present studies, a variant of Akt lacking its PH domain was found to be activated by both insulin and PDGF in three different cell types. These results indicate that the PH domain of Akt is not required for its activation. It is possible that the PH domain of Akt, like the PH domains of other proteins (17), may under certain conditions play a role in the membrane targeting of Akt. Such a role could explain the requirement for the PH domain of Akt in its activation in some studies (1). Evidence that membrane targeting of Akt plays a role in its activation was confirmed in the present studies by the finding that the addition of the src myristoylation signal to Akt resulted in constitutive activation of the enzyme. This activation, like that observed with insulin and PDGF, was again due to an increase in the Ser/Thr phosphorylation of the enzyme and did not require the PH domain, thereby giving rise to the hypothesis that a membrane-associated Ser/Thr kinase is responsible for the activation of Akt. The finding that the constitutively active Akt could stimulate the enzymatic activity of the 70-kDa S6 kinase indicates that the PH domain of Akt is also not required for its interaction with at least one of its downstream targets.