Insulin receptor substrate-2 phosphorylation is necessary for protein kinase C zeta activation by insulin in L6hIR cells.

We have investigated glycogen synthase (GS) activation in L6hIR cells expressing a peptide corresponding to the kinase regulatory loop binding domain of insulin receptor substrate-2 (IRS-2) (KRLB). In several clones of these cells (B2, F4), insulin-dependent binding of the KRLB to insulin receptors was accompanied by a block of IRS-2, but not IRS-1, phosphorylation, and insulin receptor binding. GS activation by insulin was also inhibited by >70% in these cells (p < 0.001). The impairment of GS activation was paralleled by a similarly sized inhibition of glycogen synthase kinase 3 alpha (GSK3 alpha) and GSK3 beta inactivation by insulin with no change in protein phosphatase 1 activity. PDK1 (a phosphatidylinositol trisphosphate-dependent kinase) and Akt/protein kinase B (PKB) activation by insulin showed no difference in B2, F4, and in control L6hIR cells. At variance, insulin did not activate PKC zeta in B2 and F4 cells. In L6hIR, inhibition of PKC zeta activity by either a PKC zeta antisense or a dominant negative mutant also reduced by 75% insulin inactivation of GSK3 alpha and -beta (p < 0.001) and insulin stimulation of GS (p < 0.002), similar to Akt/PKB inhibition. In L6hIR, insulin induced protein kinase C zeta (PKC zeta) co-precipitation with GSK3 alpha and beta. PKC zeta also phosphorylated GSK3 alpha and -beta. Alone, these events did not significantly affect GSK3 alpha and -beta activities. Inhibition of PKC zeta activity, however, reduced Akt/PKB phosphorylation of the key serine sites on GSK3 alpha and -beta by >80% (p < 0.001) and prevented full GSK3 inactivation by insulin. Thus, IRS-2, not IRS-1, signals insulin activation of GS in the L6hIR skeletal muscle cells. In these cells, insulin inhibition of GSK3 alpha and -beta requires dual phosphorylation by both Akt/PKB and PKC zeta.

We have investigated glycogen synthase (GS) activation in L6hIR cells expressing a peptide corresponding to the kinase regulatory loop binding domain of insulin receptor substrate-2 (IRS-2) (KRLB). In several clones of these cells (B2, F4), insulin-dependent binding of the KRLB to insulin receptors was accompanied by a block of IRS-2, but not IRS-1, phosphorylation, and insulin receptor binding. GS activation by insulin was also inhibited by >70% in these cells (p < 0.001). The impairment of GS activation was paralleled by a similarly sized inhibition of glycogen synthase kinase 3␣ (GSK3␣) and GSK3␤ inactivation by insulin with no change in protein phosphatase 1 activity. PDK1 (a phosphatidylinositol trisphosphate-dependent kinase) and Akt/protein kinase B (PKB) activation by insulin showed no difference in B2, F4, and in control L6hIR cells. At variance, insulin did not activate PKC in B2 and F4 cells. In L6hIR, inhibition of PKC activity by either a PKC antisense or a dominant negative mutant also reduced by 75% insulin inactivation of GSK3␣ and -␤ (p < 0.001) and insulin stimulation of GS (p < 0.002), similar to Akt/PKB inhibition. In L6hIR, insulin induced protein kinase C (PKC) co-precipitation with GSK3␣ and ␤. PKC also phosphorylated GSK3␣ and -␤. Alone, these events did not significantly affect GSK3␣ and -␤ activities. Inhibition of PKC activity, however, reduced Akt/PKB phosphorylation of the key serine sites on GSK3␣ and -␤ by >80% (p < 0.001) and prevented full GSK3 inactivation by insulin.

Thus, IRS-2, not IRS-1, signals insulin activation of GS in the L6hIR skeletal muscle cells. In these cells, insulin inhibition of GSK3␣ and -␤ requires dual phosphorylation by both Akt/PKB and PKC.
Most insulin effects involve tyrosine phosphorylation of insulin receptor substrates (IRSs) 1 by the receptor (1). IRSs include IRS-1 and IRS-2. These proteins feature a COOH terminus containing multiple tyrosine phosphorylation sites in various amino acid sequence motifs that bind to the Src homology 2 domain in enzymes and adapter molecules, conveying the insulin signal further downstream (2,3). In addition to the phosphorylation sites, IRS proteins contain other domains to engage activated membrane receptors. At the NH 2 terminus, the IRS proteins contain a pleckstrin homology (PH) domain (IH1 PH ). The IH1 PH is essential for the physiological interaction of IRS-1 and IRS-2 with the insulin receptor (4). In addition to the PH domain, IRS-1 and IRS-2 contain a phosphotyrosine binding (PTB) domain (IH2 PTB ), which binds the phosphorylated NPEY motif in the cytoplasmic region of the receptors for insulin, insulin-like growth factor-1, and interleukin (3,5,6). A third region encompassing residues 591-786 in IRS-2 engages the phosphorylated regulatory loop of the insulin receptor ␤-subunit (7)(8)(9). This region has been therefore termed kinase regulatory loop binding domain (KRLB) (7)(8)(9). Since IRS-1 does not contain a functional KRLB domain (9), we have previously proposed that the KRLB domain might contribute to a unique signaling potential of IRS-2 (7)(8)(9).
The signaling events by which insulin activates glycogen synthesis have become much clearer in recent years. Insulin promotes the dephosphorylation of glycogen synthase (GS) and consequent stimulation of glycogen synthesis (10 -12). Although glycogen synthase is a substrate for a large number of protein kinases, the 3a-3d cluster of phosphorylation sites are crucial to the activity of GS, and these sites are phosphorylated by glycogen synthase kinase 3 (GSK3) (13,14). Insulin inactivates GSK3 by phosphorylation of Ser-21 (GSK3␣) and/or Ser-9 (GSK3␤) (13,14) and also induces phosphorylation of the Gsubunit of the glycogen-bound form of protein phosphatase 1 (PP1) (15). These two events cooperate to activate GS in the cells (16,17), although their relative roles may vary during differentiation state of adipocytes (17). Insulin-dependent inactivation of GSK3 has been known to be dependent on Akt/ PKB (also known as related to the A and C kinase kinase) (18,19). Akt/PKB in turn was shown to be phosphorylated in response to insulin at Thr-308 and Ser-473 (20 -22), and these phosphorylation events can be blocked by inhibitors of PI 3-ki-nase (10,13,23,24). PDK-1, a phosphatidylinositol 3,4,5trisphosphate-dependent Akt/PKB kinase, was proved to phosphorylate Akt/PKB at Thr-308, which leads to a substantial but incomplete activation of Akt (21,25). The identity of the Ser-473 kinase is still unknown, but it is referred to as PDK-2, as it is expected that this kinase is also dependent upon phosphatidylinositol 3,4,5-trisphosphate. The central importance of Akt/PKB induction for insulin inactivation of GSK3 is well established (19 -21). However, recent evidence indicates that Akt/PKB activity is not sufficient for stimulation of GS in cells (26). In addition, whether induction of Akt/PKB is sufficient for insulin to inactivate GSK3 is unknown.
In the present study, we sought to identify the molecular components of the signal transduction pathway involved in insulin regulation of the glycogen synthetic machinery in skeletal muscle cells, a major target of insulin action. We demonstrate that IRS-2, not IRS-1, signals insulin activation of glycogen synthase in the L6hIR skeletal muscle cells. In these cells we show that insulin inhibition of GSK3 requires dual phosphorylation by both Akt/PKB and the atypical protein kinase C, PKC.
Cell Culture and Transfection and Cloning of the KRLB Peptide-The L6 cell clone expressing wild-type human insulin receptors (L6hIR) have been previously characterized and described with the term WT1 (30). The cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 10,000 units/ml penicillin, 10,000 g/ml streptomycin, and 2% L-glutamine in a humidified CO 2 incubator as described by Caruso et al. (30). Transient transfection of the phosphorothioate oligonucleotides and of the Lys 281 3 Trp dominant negative PKC mutant (31) was performed by the LipofectAMINE method according to the manufacturer's instruction. By using pCAGGS-␤-galactosidase as a reporter, transfection efficiency was consistently between 65 and 75%, as determined by staining with the chromogenic substrate 5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside.
IRS-2-(591-786) (IRS-2-KRLB) cDNA (8,9) was subcloned in the EcoRI and BamHI sites of the pcDNA3-Myc expression vector containing the amp r selectable marker. The construct was stably transfected in the L6hIR skeletal muscle cells using the LipofectAMINE method as in Formisano et al. (32). Individual G418-resistant clones were selected by the limiting dilution technique (G418 effective dose, 0.8 mg/ml). The expression of the KRLB peptide by the individual clones was quantitated by Western blotting.
PKC, MAPK, and PI 3-Kinase Activities-PKC activity was assayed as reported in Formisano et al. (32). Briefly, for these assays the cells were solubilized in 20 mM Tris, pH 7.5, 0.5 mM EDTA, 0.5 mM EGTA, 25 g/ml aprotinin, and 25 g/ml leupeptin (extraction buffer). Supernatants were further centrifuged at 60,000 ϫ g for 2 h, and pellets were solubilized with extraction buffer supplemented with 0.5% Triton X-100. Soluble pellets were immunoprecipitated for 18 h with isoformspecific PKC antibodies and incubated with protein A-Sepharose for 2 h. The immobilized PKC was supplemented with the lipid activators (0.32 mg/ml phosphatidylserine and 0.032 mg/ml diacylglycerol, final concentrations), and the phosphorylation reaction was initiated by adding the substrate solution (100 mM biotinylated neurogranin peptide, 0.5 mM ATP, 0.25 mM EGTA, 0.4 mM CaCl 2 , 0.1 mg/ml bovine serum albumin, 20 mM Tris, pH 7.5, 10 mM MgCl 2 , and 10 Ci/ml (3000 Ci/mmol) [␥-32 P]ATP, final concentrations). The reaction mixture was further incubated for 30 min at room temperature, and the phosphorylation reaction was terminated by adding 7.5 M guanidine hydrochloride and spotting on phosphocellulose discs. Disc-bound radioactivity was quantitated by liquid scintillation counting. Determinations of PKC␦ and PKC activities using either the acetyl-myelin basic protein (4 -14) peptide as substrate or the H-Arg-Phe-Ala-Val-Arg-Asp-Met-Arg-Gln-Thr-Val-Ala-Val-Gly-Val-Ile-Lys-Ala-Val-Asp-Lys-Lys-OH peptide (for PKC␦) or the PKC⑀ pseudosubstrate region (for PKC) provided consistent results.
For determination of PP1 activity, the cells were scraped in PP1 homogenization buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 10 mM EDTA, 10 mM Na 4 P 2 O 7 , 2 mM Na 3 VO 4 , 10% glycerol, 0.5% Triton X-100) supplemented with 0.2 mM phenylmethylsulfonyl fluoride and 25 g/ml aprotinin. The lysate was sonicated on ice and precipitated with PP1 antibodies. Immunocomplexes were immobilized on protein A-Sepharose, washed twice with 50 mM HEPES, pH 7.4, and diluted in 25 M HEPES, pH 7.4, 50 mM NaCl, 5 mM EDTA, and 4.5 nM okadaic acid. Samples were then incubated at 37°C for 15 min, and the reaction was initiated by the addition of 15 g of 32 P-labeled phosphorylase ␣ in the presence of 3 nM okadaic acid and 5 mM caffeine. Phosphate release was determined as reported by Lazar et al. (33).
Thymidine Incorporation and Glycogen Synthase Assays-The thymidine incorporation assay was accomplished as previously reported (34). Briefly, L6hIR myoblasts were seeded in six-well plates and, after 18 h, fed with Dulbecco's modified Eagle's medium supplemented with 0.25% bovine serum albumin. The cells were further incubated for 16 h in the absence or presence of 100 nM insulin followed by the addition of 500 nCi/ml [ 3 H]thymidine. 4 h later the cells were washed with ice-cold 0.9% NaCl and then with ice-cold 20% trichloroacetic acid followed by solubilization with 1 N NaOH. Radioactivity was quantitated by liquid scintillation counting. Glycogen synthase activity was assayed as described in Formisano et al. (34).

IRS-2 Block by Expression of the KRLB Peptide in L6hIR
Cells-The cDNA fragment corresponding to amino acids 591-786 of IRS-2 was transfected in L6hIR skeletal muscle cells. This fragment encodes the complete KRLB domain of IRS-2 protein (7-9). Several clones of L6hIR cells stably expressing the KRLB peptide were selected and characterized, and two of these clones expressing lower (B2) and higher levels of the peptide (F4) were studied in detail (Fig. 1A). In extracts from insulin-exposed L6hIR cells, the expressed KRLB peptide coprecipitated with insulin receptors (Fig. 1B). The co-precipitation was dependent on the concentration of insulin to which the cells were exposed (insulin EC max 100 nM, EC 50 5 nM) as well as on the KRLB peptide expression levels in the cells. In extracts from B2 and F4 cells, the KRLB peptide-insulin receptor interaction was accompanied by 75 and 95% decreased IRS-2 coprecipitation with the insulin receptor, respectively, as compared with those occurring in cells transfected with the plasmid alone (Fig. 1C). Expression of the KRLB peptide at increasing levels also caused a progressive decrease in insulindependent phosphorylation of IRS-2, with no change in the IRS-2 content of the cells (Fig. 2A). Decreased IRS-2 phosphorylation by the KRLB peptide corresponded to parallel increases in receptor co-precipitation of IRS-1 (Fig. 1C) as well as in insulin-dependent tyrosine phosphorylation of IRS-1 (Fig.  2B). As was the case for IRS-2, IRS-1 levels were unchanged in the KRLB expressors as compared with L6hIR cells transfected with the empty vector.
Insulin Action in Cells Expressing the KRLB Peptide-To address the functional consequences of blocking IRS-2-mediated insulin signaling in L6hIR cells, we have compared proliferative and glycogen synthetic responses in KRLB expressors and untransfected cells. As shown in Fig. 3 (top panel), basal (non insulin-dependent) incorporation of thymidine was increased by 25 and 40% in B2 and F4 cells, respectively, com-pared with control cells transfected with the vector alone or the untransfected cells (p Ͻ 0.001). Maximal insulin-stimulated thymidine incorporation was also increased by 18 and 40% in the B2 and F4 cell clones, respectively (p Ͻ 0.001). Increased thymidine incorporation into DNA in these cells was paralleled by 18 and 30% increased basal (p Ͻ 0.05) and a similarly sized (p Ͻ 0.05) increase in insulin-stimulated MAPK activity compared with the control cells (Fig. 3, bottom panel; p Ͻ 0.001). At variance with proliferative responses, basal glycogen synthase activity was unchanged in B2 and F4 cells (Fig. 4A). In the L6hIR control cells, insulin increased glycogen synthase activity in a dose and time-dependent fashion. Half-maximal and maximal insulin effects were achieved at 5 and 100 nM, respectively (Fig. 4A). Also, maximal stimulation was achieved within 30 min after insulin addition (Fig. 4B). Insulin-stimulated glycogen synthase activity was inhibited by 70 and 85% in B2 and F4 cells, respectively. Immunoprecipitated GSK3␣ and GSK3␤ activities toward the specific substrate phospho-glycogen synthase peptide 2 were also not different in basal F4 and in control cells (Fig. 5A). In the F4 cells, insulin, either at 1 nM or at 100 M, did not induce any significant inhibition of either GSK3␣ or GSK3␤ (Fig. 5A). At these same concentrations, insulin elicited 20 and 40% inhibition of GSK3␣ activity and 25 and 60% inhibition of GSK3␤ in the control cells (transfected with the empty vector). Phosphorylation of GSK3␣ and -␤ also exhibited no change after insulin stimulation of the F4 cells, whereas insulin stimulation increased in a dose-dependent manner in the L6hIR cells (Fig. 5A, inset). Similar results were obtained with the B2 cell clone (data not shown). Phosphatase activity in PP1 immunoprecipitates from L6hIR cells was well detectable but increased by only 25% (p Ͻ 0.05) upon insulin exposure of the cells (Fig. 5B). No significant difference in PP1 protein expression and activity were observed in the F4 compared with the control cells whether in the absence or the presence of insulin.
PKC Activation in Cells Expressing the KRLB Peptide-To elucidate the insulin-signaling events responsible for the block of glycogen synthetic responses in the KRLB peptide-expressing cells, we first analyzed IRS-1 and IRS-2-associated PI 3-kinase activity. Basal IRS-1-coprecipitated PI 3-kinase in B2 and F4 cells featured 20 and 40% increased levels compared with that in control cells (Fig. 6, top panel; p Ͻ 0.05). Maximal insulin-stimulated PI 3-kinase activity associated to IRS-1 also showed 30 and 45% increase in cells expressing the KRLB peptide (p Ͻ 0.05), paralleling IRS-1 tyrosine phosphorylation. The PI 3-kinase activity measured in IRS-2 immunoprecipitates from basal B2 and F4 cells showed 40 and 60% lower levels as compared with control cells (Fig. 6, middle panel; p Ͻ 0.001). Insulin-stimulated activity of PI 3-kinase associated to IRS-2 was inhibited by 80 and 95% in B2 and F4 cells, respec-tively. Total PI 3-kinase activity associated with tyrosine-phosphorylated proteins featured no significant differences in the B2, the F4, and the control cells, however (Fig. 6, bottom panel). It appeared, therefore, that IRS-1 and IRS-2 are redundant in transducing insulin activation of PI 3-kinase in L6hIR as in other models (32,35).
Known proteins mediating insulin signals downstream PI 3-kinase include PDK1, Akt/PKB, and PKC (13,14,36). Insulindependent activation of PDK1 and Akt/PKB was unchanged in the F4 compared with control cells (Fig. 7A). Insulin-dependent phosphorylation of key phosphorylation sites on Akt/PKB (Ser 473 and Thr 308 ) were also unchanged in the F4 compared with control cells (Fig. 7B). Similarly, protein levels of both PDK1 and Akt/PKB were unaffected by expression of the KRLB peptide. In L6hIR control cells, insulin induced PKC␣, -␤, -␦, and -in a dose-dependent fashion. In parallel with IRS-2 phosphorylation, however, maximal insulin-stimulated PKC activation (2-fold versus basal) was inhibited by 70 and Ͼ95% in the B2 and the F4 cells, respectively (Fig. 8C, bottom panel). PKC inhibition occurred with no detectable changes in the expression of PKC protein (Fig. 8A) or mRNA (data not shown) in the cells. The expression of PKC␦ was also unchanged in the KRLB peptide-expressing cells, although maximal insulin-dependent activation was reduced by 20% (p Ͻ 0.05; Fig. 8C, top panel). At variance, PKC␣ and PKC␤ expression and insulindependent activities were not significantly different in the KRLB and the control cells (Fig. 8, A and B).
The specific role of IRS-2 in insulin activation of PKC and glycogen synthetic responses was further addressed by transient transfection of IRS-2 cDNA in L6hIR cells. As shown in Fig. 9 (left panels), overexpression of IRS-2 (10-fold versus control cells) was accompanied by a constitutive increase in PKC activity levels. Insulin further increased PKC activity by 40% in cells overexpressing IRS-2. The changes in PKC activity levels caused by IRS-2 overexpression were closely paralleled by very similar changes in glycogen synthase activity and GSK3␣ and -␤ phosphorylation (Fig. 9, right panels). Thus, the block of insulin activation of PKC and of glycogen synthetic responses in cells expressing the KRLB peptide seemed to be caused by a specific block of IRS-2 signaling.

Effects of PKC Block in L6hIR
Muscle Cells-We tested the hypothesis that PKC is also involved in insulin activation of glycogen synthase in muscle cells. To address this issue, we transfected a PKC antisense oligonucleotide (PKC-AS) in wild-type L6hIR skeletal muscle cells. This antisense oligonucleotide reduced PKC protein levels by Ͼ70% in the cells without affecting those of PKC␦ (Fig. 10A). Activation of PKC by insulin was also inhibited by about 70% by the PKC antisense (Fig. 10C). In parallel, PKC-AS transfection blocked insulin inhibition of GSK3␣ and -␤ by Ͼ75% (p Ͻ 0.001; Fig.  10B). Control oligonucleotides (PKC-S. Fig. 10B, S-) neither decreased PKC (or PKC␦) levels nor GSK inhibition by insulin as compared with cells not treated with the antisense. In addition, we transfected the dominant negative Lys 281 3 Trp

FIG. 4. Glycogen synthase activity in cells expressing the KRLB peptide.
A, L6hIR, B2, and F4 cells were exposed to the indicated concentrations of insulin for 30 min. Glycogen synthase activity was then assayed as reported under "Experimental Procedures." Each data point is the mean Ϯ S.D. of duplicate determinations in four independent experiments. B, the cells were exposed to 100 nM insulin for the indicated times, and glycogen synthase activity was assayed as above. Each data point is the mean Ϯ S.D. of duplicate determinations in three independent experiments.

FIG. 5. GSK3␣ and -␤ and PP1 activities in cells expressing the KRLB peptide.
A, L6hIR and F4 cells were exposed to the indicated concentrations of insulin, lysed, and precipitated with GSK3␣ or -␤ antibodies. Kinase activity toward the phospho-glycogen synthase peptide-2 was then assayed in the immunoprecipitates, as described under "Experimental Procedures." For control, aliquots of the lysates were also blotted with Ser 21 -phospho-GSK3␣ or Ser 9 -phospho-GSK3␤ antibodies (insets). B, PP1 activity was also assayed in immunoprecipitates from lysed cells, as described under "Experimental Procedures." For control, aliquots of the lysates were blotted with PP1 antibodies. The  Similar to GSK3, insulin activation of glycogen synthase was largely prevented by PKC-AS (although not PKC-S) treatment and by transfection of the PKC-DN (Fig. 11). Different from PKC, however, inhibition of PKC␦ activity with rottlerin (3 M) had no effect on either glycogen synthase activation by insulin or GSK3␣ and -␤ inhibition (data not shown). This indicated that insulin effects on GSK3 and glycogen synthase in L6hIR skeletal muscle cells requires PKC, not PKC␦.
PKC Association and Phosphorylation of GSK3␣ and GSK3␤-To further investigate the role of PKC in activating the glycogen synthetic machinery in L6hIR cells, we sought to identify potential PKC-GSK3 interactions. Insulin pretreatment of intact L6hIR cells was found to induce co-precipitation of PKC with GSK3␣ and GSK3␤ in the cell lysates (Fig. 12A). PKC co-precipitation with GSK3␣ and -␤ was detectable by either blotting PKC precipitates with GSK3 antibodies or vice versa (data not shown). The dominant negative PKC-DN mutant did not co-precipitate with either GSK3␣ or GSK3␤ either in the absence or the presence of insulin, indicating PKC activation is necessary for its association with GSK3. In vitro, activated recombinant PKC phosphorylated purified GSK3␣ and GSK3␤ (Fig. 12, B and C). Phosphorylation also occurred using GSK3␣ and GSK3␤ preparations from L6hIR cells (data not shown). PKC phosphorylation of GSK3␣ and GSK3␤, however, did not reduce GSK3 activities toward the phospho-glycogen synthase peptide 2 substrate (Fig. 13A). This suggested that although necessary, PKC activation may not be sufficient to inhibit GSK3␣ or -␤ upon exposure of intact cells to insulin.
PKC Inhibition of GSK3␣ and GSK3␤-In most cell types, phosphorylation by Akt/PKB is believed to be necessary for insulin to restrain GSK3␣ and GSK3␤ activities (18). Consistent with those previous findings, treatment of L6hIR and F4 cells with the Akt/PKB inhibitor ML-9 reduced the insulin effect on GSK3␣ and -␤ as well as on glycogen synthase activity by Ͼ80% (p Ͻ 0.001; Fig. 14).Thus, we tested the hypothesis that phosphorylation of GSK3 by both PKC and Akt/PKB is required for fully inhibiting GSK3 activity in L6hIR cells. To address this possibility, we first phosphorylated in vitro purified GSK3␣ and -␤ with PKC. We then incubated PKC-phosphorylated GSK3 preparations with immobilized Akt/PKB from either basal or insulin-treated cells or with an active recombinant Akt/PKB. As shown in Fig. 13A, incubation of PKC-phosphorylated GSK3␣ with the activated Akt/PKB preparations reduced GSK3␣ activity by Ͼ70%. At variance, Akt/PKB phosphorylation alone elicited only a 35-40% inhibition of GSK3␣ in the absence of PKC pretreatment. No significant inhibition of GSK3␣ activity occurred upon incubation with inactive Akt/PKB both in the absence or in the presence of previous PKC phosphorylation. The same results were obtained with GSK3␤ (data not shown). At variance with Akt/ PKB, PKC did not phosphorylate Ser 21 on GSK3␣ and Ser 9 on GSK3␤ in vitro (Fig. 13B). However, a 2.5-fold increase in Akt/PKB-induced phosphorylation of Ser 21 and Ser 9 was detectable when GSK3␣ and -␤ were incubated with both PKC and Akt/PKB.
To address the relevance of PKC phosphorylation to GSK3 control by Akt/PKB in vivo, we analyzed the key Akt/PKB phosphorylation sites on GSK3␣ and -␤ in cells transfected with either PKC antisense or the dominant negative PKC-DN mutant. In both PKC antisense-transfected cells and in cells overexpressing the PKC dominant negative mutant, Ser 21 phospho-GSK3␣ and Ser 9 phospho-GSK3␤ were almost absent compared with the untransfected cells either in the absence or in the presence of insulin (Fig. 13C). This paralleled the lack of insulin effect on GSK3 activity occurring in cells when PKC expression or function is blocked. Thus, PKC phosphorylation appears to be necessary and permissive for further phosphorylation of GSK3␣ and -␤ by Akt/PKB as well as for insulin constrain of GSK3␣ and -␤ activities. DISCUSSION In the present report, we describe a novel approach to investigating IRS-2-mediated events in insulin activation of glycogen synthase in cultured cells. We have expressed a peptide corresponding to the KRLB domain of IRS-2 (amino acids 591- . The cell were then stimulated with 100 nM insulin as indicated, solubilized, and immunoprecipitated with GSK3␣ or GSK3␤ antibodies. GSK3 activity was then assayed in the immunoprecipitates, as described under "Experimental Procedures." A, for the control, aliquots of the cell lysates were blotted with PKC or PKC␦ antibodies. Filters were revealed by ECL and autoradiographed. C, alternatively, aliquots of the cell lysates were precipitated with PKCz antibodies, and precipitates were assayed for PKC activity, as outlined in the legend to 786 (7)(8)(9)) in L6hIR skeletal muscle cells. Consistent with our previous in vitro data (7)(8)(9), expression of this peptide specifically blocked IRS-2 association and phosphorylation by the active insulin receptor kinase. It appears, therefore, that the KRLB domain is necessary for enabling IRS-2 to interact with the insulin receptor in intact cells as well as in vitro. Block of IRS-2 binding to the receptor by the KRLB peptide was accompanied by increased receptor binding and phosphorylation of IRS-1, with no change in IRS-1 protein levels. The KRLB domain is unique to IRS-2. In addition to the KRLB, however, IRS-2 possesses PH and phosphotyrosine binding domains homologous to those enabling IRS-1 to interact with the insulin receptor (3). Prevention of IRS-2 binding to the receptor through the expression of the KRLB peptide may remove IRS-2 competition for PH and phosphotyrosine binding sites on the receptor, fostering binding of IRS-1. Consistent with this possibility, muscles from IRS-1 knockout mice and L6hIR cells transfected with IRS-1 ribozyme have also been reported to feature increased phosphorylation of IRS-2 (32,35).
Expression of the KRLB peptide was accompanied by increased insulin mitogenic activity through the MAPK system but block of insulin activation of the glycogen synthetic apparatus. In addition, IRS-2 overexpression constitutively activated GSK3 phosphorylation and glycogen synthase activity in L6hIR cells. Thus, IRS-1 seems to transduce insulin mitogenic effects, whereas IRS-2 is the main molecule involved in glycogen synthetic responses in L6hIR muscle cells. In these same cells, ribozyme block of IRS-1 expression prevents insulin mitogenic but not glycogen synthetic responses (32). Also, the expression of an insulin receptor mutant (IR1152) featuring constitutively increased phosphorylation of IRS-2 induces constitutive activation of the glycogen synthetic apparatus in the L6 cells (7). This effect can be prevented by expressing the KRLB peptide in the cells (data not shown). In addition to the muscle cells, IRS-2 signaling appears to be prominent in transducing insulin activation of the glycogen synthetic apparatus in isolated liver cells (37). Thus, the prominent role of IRS-2 in insulin activation of glycogen synthesis may be common to the two major tissues accomplishing this function in the organism.
Akt/PKB activation is known to represent a major mechanism leading insulin to stimulate glycogen synthase activity in cells (13,21). Accordingly, treatment of L6hIR cells with the Akt/PKB inhibitor ML-9 (38) prevents insulin stimulation of glycogen synthesis. In the present paper, however, we report that Akt/PKB phosphorylation and activation occurs at normal levels in cells expressing the KRLB peptide despite the block of insulin-stimulated glycogen synthase activity. This indicates that Akt/PKB activation is necessary but not sufficient for enabling insulin control of glycogen synthase in L6hIR cells. PDK1 and PDK2 phosphorylation of Akt/PKB in response to insulin also occurred at the same levels in cells expressing the KRLB peptide and in untransfected cells. Thus, in cells expressing the KRLB peptide, increased IRS-1 phosphorylation is paralleled by a similar increase in MAPK activity and insulin mitogenic signaling but unchanged PDK1 and Akt/PKB activities. This may occur because of the following possibilities. (i) Insulin activation of PDK1 and Akt/PKB (at variance with that of MAPK) is already saturated by IRS-1 signaling in untransfected cells; (ii) IRS-1 and IRS-2 may be redundant in signaling activation of PDK1 and Akt/PKB but not of MAPK activation, so that increased IRS-1 phosphorylation in KRLB-expressing cells exactly compensates for the lack of IRS-2 phosphorylation in activating PDK1 and Akt/PKB but not MAPK; (iii) insulin activation of PDK1 and Akt/PKB, different from that of MAPK, may not require IRS-1-or IRS-2-associated PI 3-kinase, as recently proposed by Whitehead et al. (39). These possibilities are currently under investigation in our laboratory.
At variance with PDK1 and Akt/PKB, insulin induction of PKC activity was blocked in cells expressing the KRLB peptide, indicating a major role of IRS-2 in signaling PKC activation. It is possible that expression of the KRLB peptide reduces the amount of phosphatidylinositol 3-phosphate produced after insulin stimulation of the cells and that PDK1, Akt/PKB, and PKC feature a differential need for phosphatidylinositol 3-phosphate for activation. This hypothesis is unlikely to explain the differential activation of PDK1, Akt/PKB, and PKC in KRLB and control cells, since the total amount of PI 3-kinase activity associated with tyrosine phosphoproteins was not different in the two cell types. Alternatively, activation of PKC by IRS-2-associated PI 3-kinase may occur in an intracellular compartment different from that where PDK1 and Akt/PKB are activated by IRS-1-associated PI-3 kinase. Consistent with this possibility, tyrosine-phosphorylated IRS-1 and IRS-2 are differentially located inside the cell (40,41). Other PI 3-kinase docking substrates such as IRS-4 and Grb2-associated binder (GAB) may also be differentially phosphorylated by insulin in cells expressing the KRLB peptide and contribute to the different activation of PDK1, Akt/PKB, and PKC. This latter possibility is presently under investigation in our laboratory.
We have also shown that antisense inhibition of PKC expression or block of PKC activity with a dominant negative PKC mutant prevents an insulin effect on GSK3 (both ␣ and ␤) and on glycogen synthase, as in the case of Akt/PKB block. It appears therefore that both PKC and Akt/PKB are necessary for insulin control of GSK3 and glycogen synthase in intact L6hIR cells. In these same cells, antisense block of PKC expression or inhibition of PKC activity with a dominant negative PKC mutant abolished insulin phosphorylation of the key FIG. 14. Effect of ML-9 on insulin activation of GSK3 and glycogen synthase. A, L6hIR cells were incubated with 100 M ML-9 for 10 min. During the last 5 min of incubation insulin was added at 100 nM final concentration. The cells were solubilized and blotted with phospho-Ser 21 GSK3␣ (pSer 21 -GSK3␣) or pSer 9 GSK3␤ (p-Ser 9 -GSK3␤) antibodies, as indicated. Filters were revealed by ECL and autoradiographed. The autoradiographs shown are representative of three independent experiments. B, alternatively, the cell lysates were assayed for glycogen synthase activity as described under "Experimental Procedures." The bars represent the mean Ϯ S.D. of duplicate determinations in four independent experiments. Akt/PKB phosphorylation sites in GSK3␣ and -␤ (Ser 21 and Ser 9 , respectively). These sites might undergo promiscuous phosphorylation by PKC in the L6hIR cells. This is an unlikely possibility, however, since, in vitro PKC does not phosphorylate either GSK3␣ on Ser 21 or GSK3␤ on Ser 9 . Alternatively, PKC phosphorylation of GSK3␣ and -␤ may be permissive for phosphorylation and inactivation by Akt/PKB. In vitro recombinant PKC phosphorylates GSK3␣ and -␤ but is unable to inhibit its activity. Recombinant Akt/PKB alone exhibits only modest effects on GSK3 activity. We have found, however, that PKC phosphorylation of GSK3␣ or -␤ enables Akt/PKB to further inhibit the GSK3s in vitro. This sequential phosphorylation of GSK3 by PKC and Akt/PKB was accompanied by almost complete block of GSK3 activity. Thus, in vitro full inhibition of GSK3␣ or -␤ activities requires phosphorylation by both PKC and Akt/PKB. We propose, therefore, that dual phosphorylation of GSK3␣ and -␤ by PKC and Akt/PKB may also be necessary for full inactivation of GSK3 by insulin, at least in intact L6hIR muscle cells. It appears that PKC is involved in transducing insulin action to GSK3 and glycogen synthase in addition to regulating insulin-mediated glucose uptake and general protein synthesis (42)(43)(44).
Phosphorylation of GSK3 by Akt/PKB is believed to represent a major mechanism responsible for insulin control of GSK3 and glycogen synthase activities in cells. This is also the case for the L6hIR myotubes, since insulin elicited only a slight effect on PP1 in these cells. To our knowledge, however, the present report provides the first evidence that PKC phosphorylation of GSK3 is permissive for insulin-dependent Akt/PKB regulation of GSK3. Mapping the relevant PKC phosphorylation sites on GSK3 is presently in progress in our laboratory.
A recent report by Tsujio et al. (45) shows that activation of PKC␦ rather than -is involved in insulin signaling to GSK3 in neuroblastoma cells, suggesting that different PKC isoforms may accomplish this function in different cell types. In addition, in liver PKC is not involved in insulin signaling to glycogen synthase (10). Also, the expression of the KRLB peptide in mouse liver cells inhibits insulin activation of Akt/PKB despite the unchanged induction occurring in the L6hIR muscle cells. 2 Thus, the molecular mechanisms responsible for insulin inactivation of GSK3 may feature tissue specificity.