Requirement for Akt (Protein Kinase B) in Insulin-induced Activation of Glycogen Synthase and Phosphorylation of 4E-BP1 (PHAS-1)*

The roles of Akt (protein kinase B) and the atypical λ isoform of protein kinase C (PKCλ), both of which act downstream of phosphoinositide 3-kinase, in the activation of glycogen synthase and phosphorylation of 4E-BP1 (PHAS-1) in response to insulin were investigated. A mutant Akt (Akt-AA) in which the phosphorylation sites targeted by growth factors are replaced by alanine was shown to inhibit insulin-induced activation of both Akt and glycogen synthase in L6 myotubes. Expression of a mutant Akt in which Lys179 in the kinase domain was replaced by aspartate also inhibited insulin-induced activation of glycogen synthase but had no effect on insulin activation of endogenous Akt. A kinase-defective mutant of PKCλ (λΔNKD), which prevents insulin-induced activation of PKCλ, did not affect the activation of glycogen synthase by insulin. Insulin-induced phosphorylation of 4E-BP1 was inhibited by Akt-AA in Chinese hamster ovary cells. However, λΔNKD had no effect on 4E-BP1 phosphorylation induced by insulin. These data suggest that Akt, but not PKCλ, is required for insulin activation of glycogen synthase and for insulin-induced phosphorylation of 4E-BP1.

Insulin exerts a variety of effects on carbohydrate, lipid, and protein metabolism (1). Although the intracellular signaling cascades that mediate these divergent actions remain to be fully characterized, phosphoinositide (PI) 3-kinase, composed of a 110-kDa catalytic subunit and a Src homology 2 domaincontaining regulatory subunit, is thought to play a central role in several metabolic effects of insulin (1)(2)(3). Inhibition of PI 3-kinase with the use of pharmacological or molecular biological strategies has revealed a variety of molecules that act downstream of this enzyme (3). One such molecule that appears to function immediately downstream of PI 3-kinase is the serine-threonine kinase Akt (also known as protein kinase B). Akt interacts through its pleckstrin homology domain with the phospholipids produced by PI 3-kinase and is thereby activated in vitro (4,5). Phosphorylation of Thr 308 and Ser 473 of Akt is also important for its activation (6,7), and PI 3-kinase appears to contribute to the regulation of a kinase, termed PDK1, that catalyzes the phosphorylation of Thr 308 of Akt (8). PDK1 is also thought to regulate members of the protein kinase C (PKC) family of enzymes (9,10), several of which, including PKC, PKC, and PKC⑀, also act downstream of PI 3-kinase (11)(12)(13)(14). In addition to mediating the effects of insulin on cytoskeletal organization (15), Rac, a member of the Rho family of small GTPases, as well as one of its direct targets, p21-activated kinase, also may play a role in insulin action downstream of PI 3-kinase (16). Furthermore, the activities of protein kinases belonging to the mitogen-activated protein (MAP) kinase superfamily, including those of extracellular signal-regulated kinase (17)(18)(19) and c-Jun NH 2 -terminal kinase (or stress-activated protein kinase) (20) isozymes, are sensitive to blockers of PI 3-kinase in intact cells, suggesting that these enzymes participate in signaling downstream of PI 3-kinase. However, despite the identification of these various molecules that act downstream of PI 3-kinase, relatively little is known of how signals diverge at this key step of insulin signaling and of which molecules are responsible for each of the specific biological activities of this hormone.
Synthesis and breakdown of glycogen are important aspects of carbohydrate homeostasis. Glycogen synthase, a key enzyme that catalyzes glycogen synthesis, is activated by insulin in a PI 3-kinase-dependent manner (21)(22)(23). Moreover, a constitutively active mutant of Akt stimulates glycogen synthase in quiescent cells (24), suggesting that the PI 3-kinase-Akt pathway mediates activation of this enzyme in response to insulin. However, demonstration of such a role for Akt would require investigation of the effects of specific inhibition of Akt on insulin-induced activation of glycogen synthase. Furthermore, it is important to study the regulation of glycogen synthase in physiologically relevant cells, such as muscle cells or hepatocytes, because such regulation may differ among cell or tissue types (24 -26).
Insulin rapidly promotes general protein synthesis in a variety of cells and tissues. Although insulin affects multiple steps of protein synthesis, initiation of mRNA translation is thought to be one of its most important targets (27). The small, heat-and acid-stable protein 4E-BP1 (also known as PHAS-1) is rapidly phosphorylated in response to insulin (28,29). Insulin-induced phosphorylation of 4E-BP1 results in a decrease in its binding affinity for eIF-4E, an essential translation initiation factor for certain mRNAs that possess an m 7 (5Ј)Gppp(5Ј)N (where N is any nucleotide) cap structure (27,29). Subsequent dissociation of eIF-4E from 4E-BP1 promotes cap structure-dependent translation initiation (29). A constitutively active mutant of Akt was shown to stimulate phosphorylation of 4E-BP1 (24,30,31). In addition, expression of a kinase-defective Akt, in which Lys 179 in the kinase domain was mutated, inhibited insulin-induced phosphorylation of 4E-BP1 (30). The relevance of this latter observation is unclear, however, because this and similar kinase-defective mutants of Akt do not exert a dominant inhibitory effect on insulin-induced activation of Akt (7,32).
We have recently shown that a mutant Akt in which the sites of ligand-induced phosphorylation are replaced with alanine (Akt-AA) inhibits the growth factor-induced activation of endogenous Akt when overexpressed in CHO cells or 3T3-L1 adipocytes (7,33). With the use of this mutant, we have also shown that Akt is essential for stimulation of bulk protein synthesis by insulin (7). In addition, we have shown that PKC is activated by insulin in a PI 3-kinase-dependent manner, and that a PKC mutant that prevents insulin-induced activation of endogenous PKC markedly, but not completely, inhibits insulin stimulation of glucose transport (14). These observations indicate that PKC acts downstream of PI 3-kinase to mediate certain biological effects of insulin. In the present study, we have investigated the roles of Akt and PKC in the insulin-induced activation of glycogen synthase and phosphorylation of 4E-BP1 in L6 myotubes and CHO cells, respectively. In particular, the requirement for these two protein kinases in the effects of insulin have been evaluated by specifically inhibiting their activities with the use of dominant negative mutants.

EXPERIMENTAL PROCEDURES
Cells and Antibodies-L6 myoblasts were maintained and induced to differentiate into myotubes as described previously (34). Polyclonal antibodies to Akt (␣panAkt) were generated against a glutathione Stransferase fusion protein containing amino acids 428 -480 of rat Akt1, as described (7). Polyclonal antibodies to PKC were generated in response to a peptide corresponding to amino acids 197-213 of mouse PKC (1). Polyclonal antibodies to MAP kinase (␣C92) (35) or to 4E-BP1 (28,29) were as described. Polyclonal antibodies specific to Akt2 (PKB␤) were obtained from Upstate Biotechnology, and those specific to Akt3 (PKB␥) (36) were kindly provided by P. Cohen (University of Dundee, Dundee, Scotland, United Kingdom). Monoclonal antibodies to the hemagglutinin (HA) epitope tag (12CA5) or to the FLAG epitope tag were obtained from Roche Molecular Biochemicals and Kodak Scientific Imaging Systems, respectively.
Construction of and Infection with Adenovirus Vectors-Adenovirus vectors encoding HA-tagged wild-type Akt (AxCAAkt-WT), HA-tagged mutant Akt in which Lys 179 in the kinase domain was replaced by aspartate (AxCAAkt-K179D), HA-tagged mutant Akt in which Thr 308 and Ser 473 were replaced by alanine (AxCAAkt-AA), a constitutively active mutant of PKC that lacks the pseudosubstrate domain (AxCA⌬PD), or a dominant negative mutant of PKC that lacks the NH 2 -terminal region (amino acids 1-235) of the wild-type protein and in which Lys 273 in the kinase domain was replaced by glutamate (AxCA⌬NKD) were as described previously (7,14). An adenovirus vector encoding FLAG epitope-tagged Akt-AA (AxCAAkt-AAFL) was constructed with the use of an adenovirus expression kit (Takara, Tokyo, Japan) and a cDNA (produced by the polymerase chain reaction) encoding Akt-AA (7) fused at its NH 2 terminus with the FLAG epitope (DYKDDDDK).
CHO cells or fully differentiated L6 myotubes were infected with adenovirus vectors at the indicated multiplicity of infection (m.o.i.) as described previously (7,35). The cells were subjected to experiments 24 -48 h after infection.
Biological Assays-L6 myotubes were deprived of serum for 16 -20 h, incubated in the absence or presence of 100 nM insulin for 10 min, and then immediately frozen with liquid nitrogen. For assay of Akt activity, the frozen cells were lysed and subjected to immunoprecipitation with antibodies to either Akt or HA, as described previously (7). The kinase activity in the resulting immunoprecipitates was then assayed as described (7) with histone 2B as substrate. For assay of PKC activity, the frozen cells were lysed and subjected to immunoprecipitation with antibodies to PKC. The kinase activity in the resulting immunoprecipitates was then assayed with myelin basic protein as substrate, as described (14). MAP kinase activity and PI 3-kinase activity were assayed in immunoprecipitates prepared with antibodies to MAP kinase or to phosphotyrosine, respectively, as described previously (35).
For analysis of 4E-BP1 phosphorylation, serum-deprived CHO cells were incubated in the absence or presence of 100 nM insulin for 10 min, scraped into a solution containing 50 mM Tris-HCl (pH 7.5), 100 mM potassium fluoride, 10 mM EDTA, 2 mM EGTA, leupeptin (10 g/ml), aprotinin (10 g/ml), and 0.1 mM phenylmethylsulfonyl fluoride (PMSF), and lysed by three cycles of freezing and thawing. After removal of debris by centrifugation in a microcentrifuge, the resulting supernatant was heated at 100°C for 7 min. The resulting precipitated material was removed by centrifugation, and the new supernatant was subjected to immunoblot analysis with antibodies to 4E-BP1. After subsequent incubation in the absence or presence of 100 nM insulin for 10 min, the cells were washed twice with KRH buffer containing 1% BSA and lysed in a solution containing 10% glycerol, 50 mM Tris-HCl (pH 7.5), 60 mM KCl, 2 mM EDTA, 1% Triton X-100, 2 mM dithiothreitol, 20 mM ␤-glycerophosphate, 10 mM sodium pyrophosphate, 50 mM NaF, 1 mM PMSF, 0.1 mM sodium orthovanadate, 50 mM p-nitrophenyl phosphate. The lysate was centrifuged (15,000 ϫ g for 20 min), and the resulting supernatant was subjected to immunoprecipitation with polyclonal antibodies to 4E-BP1. The immunoprecipitates were washed three times with a lysis buffer, then two times with a solution containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.1% sodium deoxycholate, 5 mM EDTA, 10 mM NaF, 2 mM sodium orthovanadate, 1 mM PMSF, 10 g/ml leupeptin, 10 g/ml aprotinin, 1% Nonidet P-40, then washed two times with a solution containing 200 mM LiCl, 50 mM Tris-HCl (pH 7.5), 1 mM dithiothreitol. The precipitates were then boiled in SDS sample buffer, and subjected to SDS-PAGE on a 15% gel; the incorporation of radioactivity into 4E-BP1 was visualized or quantitated with a Fuji BAS2000 image analyzer.
Glycogen synthase activity was assayed essentially as described (21). In brief, serum-deprived differentiated L6 myotubes in 60-mm culture dishes were incubated in the absence or presence of 100 nM insulin for 30 min at 37°C, and then lysed in 120 l of a solution containing 50 mM Tris-HCl (pH 7.6), 100 mM potassium fluoride, 10 mM PMSF and 30% (v/v) glycerol. After removal of insoluble material by centrifugation, 30 l of the supernatant were mixed with 60 l of a solution containing 50 mM Tris-HCl (pH 7.6), 20 mM EDTA, 25 mM potassium fluoride, glycogen (10 mg/ml), and 6.7 mM uridine diphospho-D-[U-14 C]glucose (0.05 Ci), in the absence or presence of 10 mM glucose 6-phosphate. After incubation for 30 min at 30°C, the reaction was terminated by spotting the reaction mixture onto filter paper, and glycogen was precipitated on the paper by soaking in ice-cold 70% (v/v) ethanol. Each filter was washed four times with ice-cold 70% ethanol, and the radioactivity remaining on the paper was then measured with a liquid scintillation counter. The activity ratio (the ratio of the activity in the absence of glucose 6-phosphate to that in its presence) was calculated.

Effect of Akt-AA on Insulin-induced Activation of Akt in L6
Myotubes-L6 myotubes were incubated for 10 min in the absence or presence of 100 nM insulin, lysed, and subjected to immunoprecipitation with polyclonal antibodies that recognize all three known isoforms of Akt (␣panAkt) (7). The resulting immunoprecipitates were then assayed for kinase activity with histone 2B as substrate. The amount of Akt activity in the precipitate from insulin-treated cells was about 5 times that in the precipitate from untreated cells (Fig. 1A). We have previously shown that a mutant Akt in which the sites of ligandinduced phosphorylation are replaced by alanine (Akt-AA) inhibits the activation of endogenous or transfected Akt by insulin, platelet-derived growth factor, or heat treatment in 3T3-L1 adipocytes or CHO cells (7,33). We thus examined whether Akt-AA exerts similar effects in L6 myotubes. Myotubes were infected with an adenovirus vector encoding HAtagged wild-type Akt (AxCAAkt-WT) at an m.o.i. of 0.5 plaqueforming units (pfu)/cell and, after 12 h, infected again with an adenovirus encoding FLAG-tagged Akt-AA (AxCAAkt-AAFL) at various MOIs. After an additional 20 h, the myotubes were incubated in the absence or presence of 100 nM insulin for 10 min, lysed, and subjected to immunoprecipitation with antibodies to HA. Assay of the resulting immunoprecipitates for Akt activity revealed that insulin increased the activity of HA-Akt to an extent similar to that observed for endogenous Akt (Fig.  1C). Infection of the cells with AxCAAkt-AAFL inhibited insulin-induced activation of HA-Akt in an m.o.i.-dependent manner (Fig. 1C), without affecting the amount of HA-Akt protein present in the immunoprecipitates (Fig. 1B). At an AxCAAkt-AAFL m.o.i. of 10 pfu/cell, insulin-induced activation of HA-Akt was almost completely abolished. Infection with AxCAAkt-AAFL did not affect insulin activation of MAP kinase (Fig. 1D) or of PI 3-kinase (as assayed in precipitates prepared with antibodies to phosphotyrosine) (data not shown), suggesting that the inhibitory effect of Akt-AA on Akt activation is specific. Infection of the myotubes with a control virus encoding ␤-galactosidase (AxCALacZ) at an m.o.i. of 10 pfu/cell did not affect insulin activation of Akt (data not shown). These results suggest that Akt-AA exerts a dominant negative effect on insulininduced Akt activation.
We next investigated the effect of Akt-AA on endogenous Akt activity in L6 myotubes by precipitating the endogenous protein with ␣panAkt. Because these antibodies recognize both endogenous and recombinant Akt proteins, Akt-AA was immunodepleted from cell lysates with antibodies to FLAG before endogenous Akt was immunoprecipitated with the antibodies to Akt and assayed for kinase activity. Infection of L6 myotubes with AxCAAkt-AAFL resulted in a dose-dependent increase in total Akt protein ( Fig. 2A); at an m.o.i. of 10 pfu/cell, the abundance of Akt-AA was ϳ20 times that of endogenous Akt. After three sequential rounds of immunoprecipitation of cell lysates with antibodies to FLAG, the amounts of Akt protein remaining in the supernatant were similar for infected and noninfected cells ( Fig. 2A), indicating that Akt-AA was completely removed by this procedure. The insulin-induced activation of endogenous Akt was inhibited by AxCAAkt-AAFL in an m.o.i.-dependent manner, and at an m.o.i. of 20 pfu/cell, insulin-induced activation of endogenous Akt was almost completely abolished (Fig. 2B).
We have previously shown that Akt-K179D, a mutant Akt in which Lys 179 in the kinase domain is replaced by aspartate, did not inhibit insulin-induced activation of Akt when overexpressed in CHO cells (7). We therefore examined the effect of Akt-K179D on endogenous Akt activity in L6 myotubes. Myotubes were infected with a virus encoding HA-tagged Akt-K179D (Ax-CAAkt-K179D), incubated in the absence or presence of insulin, and lysed. Akt-K179D was then immunodepleted from the cell lysates with antibodies to HA, after which endogenous Akt in the remaining supernatant was immunoprecipitated with ␣panAkt and assayed for kinase activity. Although the amount of Akt protein in the cells infected with AxCAAkt-K179D at an m.o.i. of 10 pfu/cell was similar to that in cells infected with AxCAAkt-AAFL at the same m.o.i. (Fig. 2C), activation of endogenous Akt by insulin was not affected by Akt-K179D (Fig. 2D).
Although we have previously shown that ␣panAkt recognize all three known isoforms of rat Akt transiently expressed in COS cells (7), a recent report has suggested that the previously cloned rat Akt3 (PKB␥) may be a minor splice variant of this isoform (37). We therefore examined whether ␣panAkt recognizes endogenous Akt3 in L6 myotubes. CHO cells, L6 myotubes, or 293 cells were lysed and immunoprecipitated with antibodies specific to Akt3 or with control serum, and then the precipitates were subjected to immunoblot analysis with ␣pan-Akt. Single bands, with molecular masses of ϳ60 kDa, were detected in the precipitates prepared with antibodies to Akt3, but not with control serum (Fig. 3A), indicating that ␣panAkt recognizes endogenous Akt3 expressed in these cells.
The dominant negative effect of Akt-AA was further confirmed by assay of endogenous Akt2 and Akt3 activity precipitated with antibodies specific to each isoform. Because Akt-AA was constructed from Akt1, the polyclonal antibodies to Akt2 or to Akt3 did not recognize the mutant protein (data not shown). Insulin-induced increase in endogenous Akt2 and Akt3 activity was markedly inhibited by Akt-AA (Fig. 3, B and C).
Effect of Akt-AA on Insulin-induced Activation of Glycogen Synthase-We next examined the effect of Akt-AA on the insulin-induced activation of glycogen synthase in order to determine whether Akt activity is required for this action of insulin. Incubation of L6 myotubes with 100 nM insulin resulted in a 2-3-fold increase in glycogen synthase activity within 30 min (Fig. 4A). Treatment of the cells with 100 nM wortmannin inhibited insulin-induced activation of glycogen synthase almost completely. Infection of myotubes with AxCAAkt-AA inhibited insulin stimulation of glycogen synthase activity in an m.o.i.-dependent manner (Fig. 4A). AxCAAkt-AAFL showed an essentially identical inhibition of insulin-induced activation of glycogen synthase (data not shown). Moreover, expression of Akt-K179D also inhibited activation of glycogen synthase by insulin to a marked extent (Fig. 4B). In contrast, expression of Akt-WT, the abundance of which was similar to those of Akt-AA or Akt-K179D as assessed by immunoblot analysis with antibodies to HA (Fig. 4C), had no effect on insulin activation of glycogen synthase (Fig. 4B).
Effect of Inhibition of PKC on Insulin Activation of Glycogen Synthase-PKC, one of the atypical isoforms of PKC, acts as a downstream effector of PI 3-kinase (11,14). To investigate whether PKC contributes to the activation of glycogen synthase by insulin, we examined the effect of an NH 2 -terminally truncated, kinase-defective mutant of PKC (⌬NKD) that has been shown to prevent insulin-induced activation of PKC in to immunoblot analysis with ␣panAkt (A and C, upper panels) or to three sequential rounds of immunoprecipitation with antibodies to FLAG (for AxCAAkt-AAFL) or to HA (for AxCAAkt-K179D) in order to deplete the recombinant proteins. After immunodepletion, the remaining supernatant was subjected either to immunoblot analysis with polyclonal antibodies to Akt (A and C, lower panels) or to immunoprecipitation with polyclonal antibodies to Akt; the resulting precipitates were assayed for Akt activity (B and D). Quantitative data are means Ϯ S.E. from three experiments; immunoblots are representative of three independent experiments.  (14). Because this mutant does not possess the region of PKC corresponding to the peptide that we used to generate polyclonal antibodies to PKC, it is not recognized by these antibodies (14). We were thus able to use the antibodies to assay endogenous PKC activity in L6 myotubes expressing ⌬NKD. The kinase activity in PKC immunoprecipitates from insulin-treated myotubes was about 1.5 times that precipitated from control myotubes (Fig. 5A). This activation of PKC by insulin was inhibited in an m.o.i.dependent manner by infection of the cells with a virus that encodes ⌬NKD (AxCA⌬NKD), consistent with our previous observations in 3T3-L1 adipocytes and CHO cells (14). However, infection of the myotubes with AxCA⌬NKD did not inhibit insulin-induced activation of glycogen synthase even at an m.o.i. of 10 pfu/cell (Fig. 5B), a virus dose sufficient to inhibit almost completely insulin activation of PKC (Fig. 5A). Although glycogen synthase activity in L6 myotubes infected with AxCA⌬NKD at an m.o.i. of 1 or of 3 pfu/cell was slightly higher than that in non-infected cells, glycogen synthase activity in these cells in the absence of insulin was also slightly elevated (data not shown). Thus, inhibition of PKC activity did not prevent activation of glycogen synthase by insulin. to a relatively highly phosphorylated form of 4E-BP1, appeared and, concomitantly, the band with the highest mobility (␣ band), which corresponds to a lower phosphorylation state, decreased in intensity (Fig. 6A). Pretreatment of cells with 100 nM wortmannin abolished this effect of insulin on the phosphorylation of 4E-BP1.

Effects of Various Mutant Proteins on Insulin-induced Phos
We next tested the effects of various mutant proteins on insulin-induced phosphorylation of 4E-BP1 with the use of adenovirus-mediated gene transfer. Infection of cells with wildtype adenovirus induces phosphorylation of 4E-BP1 (38,39), whereas infection with a mutant adenovirus that lacks the E1A gene does not (38). Because recombinant adenoviruses used in the present study lacked the E1A, E1B, and E3 genes and contained instead various inserted cDNAs (40), it was likely that infection with these viruses alone would not result in phosphorylation of 4E-BP1. Indeed, infection of CHO cells with AxCALacZ affected neither the basal state of 4E-BP1 phosphorylation nor the insulin-induced increase in this parameter (data not shown). Infection with AxCAAkt-AA was associated with a slight decrease in the intensity of the ␥ band and an increase in that of the ␣ band in insulin-treated cells (Fig. 6B); however, the extent of the inhibition achieved by Akt-AA was smaller than that by wortmannin (Fig. 6, compare A and B). The effect of Akt-AA on the inhibition of 4E-BP1 phosphorylation was further confirmed by in vivo labeling assays. We labeled CHO cells with [ 32 P]orthophosphate, incubated them in the absence or presence of insulin, and subjected cell lysates to immunoprecipitation with antibodies to 4E-BP1. The immunoprecipitates were then subjected to SDS-PAGE and autoradiography. Insulin induced ϳ4-fold increase in 32 P incorporation into 4E-BP1 protein and treatment of the cells with wortmannin completely abolished the effect of insulin (Fig. 7A). Infection of the cells with AxCAAkt-AA inhibited insulin-induced increase in 32 P incorporation into 4E-BP1 by ϳ50%. (Fig. 7, B  and C). These results suggest that Akt is required, at least partly, for insulin-induced phosphorylation of 4E-BP1.
We finally examined whether PKC is required for 4E-BP1 phosphorylation in response to insulin. We have recently shown that the kinase activity of a mutant PKC that lacks the pseudosubstrate domain (⌬PD) is higher than that of the wild-type enzyme (14). Infection of CHO cells with AxCA⌬PD, an adenovirus encoding ⌬PD, resulted in an increase in the phosphorylation of 4E-BP1 (Fig. 6C), suggesting that signals downstream of PKC lead to phosphorylation of this protein. However, infection of cells with AxCA⌬NKD did not prevent the insulin-induced shift in the mobility of 4E-BP1 even at an m.o.i. of 30 pfu/cell (Fig. 6D), a dose sufficient to abolish insu- lin-induced activation of PKC in CHO cells (7). Wortmannin completely inhibited the effect of insulin on 4E-BP1 phosphorylation in cells infected with AxCA⌬NKD (Fig. 6D). These results suggest that the activity of PKC is not essential for the insulin-induced increase in the phosphorylation of 4E-BP1. DISCUSSION Intracellular signaling is mediated by complex networks composed of multiple molecules, many of which exist in various isoforms and act in a redundant manner. Small, cell-permeable inhibitors of signaling components, such as wortmannin or LY294002 for PI 3-kinase and its related enzymes, as well as a variety of protein kinase inhibitors, are potent tools for dissecting signaling networks. However, highly specific inhibitors are available for only a limited number of signaling molecules. A second approach to exploring signaling networks is to examine the effects of various mutants of signaling components. Mutants that possess a catalytic activity higher than that of the corresponding wild-type enzyme, or those that are constitutively targeted to a specific intracellular compartment, sometimes mimic the signals transmitted by the wild-type proteins and can be informative as to whether such signals are sufficient to induce a particular biological action. However, overexpression of hyperactive enzymes may result in signal "overflow" and consequent induction of biological effects that are not mediated physiologically by the endogenous enzymes. In contrast, mutants that prevent the activation of an endogenous enzyme are often informative as to the requirement for the enzyme in a particular signaling cascade. However, overexpression of such a mutant protein may also inhibit other molecules by competition if they share a common upstream regulator. Thus, demonstration of the importance of a specific molecule in a specific signaling cascade requires that the effects of both types of mutants, as well as those of pharmacological inhibitors, if available, be examined.
We have now shown that activation of glycogen synthase by insulin was markedly inhibited in L6 myotubes in which insulin activation of endogenous Akt was prevented by overexpression of Akt-AA. Although insulin-induced activation of MAP kinase and PI 3-kinase in these cells appeared normal, suggesting that the effect of Akt-AA was specific, we cannot completely exclude the possibility that expression of Akt-AA also affects signaling molecules that do not normally participate in the Akt pathway. However, the present data, together with the observation that a constitutively active mutant of Akt mimics the effect of insulin on glycogen synthase activity (24), strongly suggest that Akt contributes to the physiological regulation of glycogen synthase by insulin, at least in L6 myotubes. The mechanism by which Akt transmits signals to glycogen synthase is not known. Cross et al. (41) have shown that glycogen synthase kinase (GSK) 3␤ is phosphorylated and inactivated by Akt in vitro. Moreover, van Weeren et al. (32) have reported that a mutant Akt fused with a membrane-targeted signal sequence derived from the carboxyl terminus of Ha-RAS inhibited the activity of cotransfected Akt and prevented the inactivation of GSK3␤ induced by insulin in A14 cells. Because inactivation of GSK3␤ is thought to play a key role in the dephosphorylation and activation of glycogen synthase (42), the signal from Akt to glycogen synthase may be transmitted by GSK3␤.
An Akt mutant in which Lys 179 in the kinase domain was replaced by aspartate (Akt-K179D) did not inhibit insulininduced activation of endogenous Akt in L6 myotubes, consistent with previous observations by us and other investigators (7,32). However, overexpression of this mutant markedly inhibited insulin activation of glycogen synthase in L6 myotubes. Because Akt associates with GSK3␤ in intact cells (32), Akt-K179D, when overexpressed, may compete with endogenous Akt for binding to GSK3␤ (or to other physiological substrates of Akt) and thereby prevent activation of glycogen synthase. Thus, it is possible that Akt-K179D blocks signaling downstream of Akt, and that Akt-AA blocks signaling both upstream and downstream of Akt. In this regard, growth factor-induced phosphorylation of Bcl-X L /Bcl-2-associated death factor, another Akt substrate that is important in cell survival, is blocked by expression of a mutant Akt containing a substitution at Lys 179 , and both wild-type and the mutant Akt associate with BAD in intact cells with apparently similar affinities (43).
Expression of ⌬NKD inhibited insulin-induced activation of PKC in L6 myotubes. It has been demonstrated that atypical PKC participates in signal leading to activation of MAP kinase cascade (44,45). We have recently found that expression of ⌬NKD attenuated insulin-stimulated MAP kinase activity in various cells including L6 myotubes, 2 indicating that ⌬NKD prevents a certain biological effect of insulin in these cells. However, ⌬NKD did not affect insulin activation of glycogen synthase. Furthermore, expression of ⌬PD, a constitutively active mutant of PKC (14), did not stimulate glycogen synthase activity (data not shown). These results suggest that PKC is not involved in insulin-induced activation of glycogen synthase.
Expression of constitutively active mutants of Akt has been shown to induce a shift in the electrophoretic mobility of 4E-BP1 (24,30,31). Moreover, Scott et al. (46) have shown that mTOR, also known as FRAP or RAFT, possesses protein kinase activity and directly phosphorylates 4E-BP1 in vitro (46). These investigators also showed that the kinase activity of mTOR is stimulated in response to phosphorylation of the protein by a constitutively active Akt, suggesting that the Akt-mTOR pathway transmits a signal that results in the phosphorylation of 4E-BP1. We have now shown that insulininduced phosphorylation of 4E-BP1 was inhibited by expression of Akt-AA in CHO cells, indicating that Akt indeed plays a role in this action of insulin. We have previously shown that insulin-stimulated general protein synthesis was almost completely prevented by expression of Akt-AA in CHO cells and 3T3-L1 adipocytes (7). Because insulin-induced phosphorylation of 4E-BP1 is thought to mediate the effect of insulin on protein synthesis (27,29), the inhibitory effect of Akt-AA on insulin-induced protein synthesis may, at least partly, be due to the inhibition of phosphorylation of 4E-BP1. However, it is possible that Akt regulates not only the phosphorylation of 4E-BP1 but also another pathway that leads to protein synthesis. The effects of Akt-AA on other steps of protein synthesis and turnover, including activation of initiation factors other than eIF-4E or of elongation factors as well as stabilization of mRNA or protein, warrant further investigation.
4E-BP1 is phosphorylated in vitro by various protein kinases, including MAP kinase, p38, casein kinase II, and PKC isozymes (28,47,48). We have now shown that expression of ⌬PD, a mutant PKC with a kinase activity greater than that of the wild-type enzyme (14), resulted in a shift in the electrophoretic mobility of 4E-BP1 in CHO cells. However, a dominant negative mutant of PKC (⌬NKD) had virtually no effect on phosphorylation of 4E-BP1 in response to insulin, whereas ⌬NKD inhibited insulin-stimulated PKC activity (14) as well as MAP kinase activity. 2 The simplest explanation for the latter observation is that PKC is not required for the effect of insulin on 4E-BP1 phosphorylation. Overexpression of ⌬PD might result in overflow of the activity of PKC from its physiological cellular compartment and consequent phosphorylation of nonphysiological substrates.