Regulation of Glycogen Synthase in Rat Hepatocytes

We examined the signaling pathways regulating glycogen synthase (GS) in primary cultures of rat hepatocytes. The activation of GS by insulin and glucose was completely reversed by the phosphatidylinositol 3-kinase inhibitor wortmannin. Wortmannin also inhibited insulin-induced phosphorylation and activation of protein kinase B/Akt (PKB/Akt) as well as insulin-induced inactivation of GS kinase-3 (GSK-3), consistent with a role for the phosphatidylinositol 3-kinase/PKB-Akt/GSK-3 axis in insulin-induced GS activation. Although wortmannin completely inhibited the significantly greater level of GS activation produced by the insulin-mimetic bisperoxovanadium 1,10-phenanthroline (bpV(phen)), there was only minimal accompanying inhibition of bpV(phen)-induced phosphorylation and activation of PKB/Akt, and inactivation of GSK-3. Thus, PKB/Akt activation and GSK-3 inactivation may be necessary but are not sufficient to induce GS activation in rat hepatocytes. Rapamycin partially inhibited the GS activation induced by bpV(phen) but not that effected by insulin. Both insulin- and bpV(phen)-induced activation of the atypical protein kinase C (ζ/λ) (PKC (ζ/λ)) was reversed by wortmannin. Inhibition of PKC (ζ/λ) with a pseudosubstrate peptide had no effect on GS activation by insulin, but substantially reversed GS activation by bpV(phen). The combination of this inhibitor with rapamycin produced an additive inhibitory effect on bpV(phen)-mediated GS activation. Taken together, our results indicate that the signaling components mammalian target of rapamycin and PKC (ζ/λ) as well as other yet to be defined effector(s) contribute to the modulation of GS in rat hepatocytes.

zyme in glycogen synthesis, by insulin has received increasing attention. GS is regulated by a complex interplay of diurnal, nutritional, and hormonal factors (reviewed in Refs. 1-3) that ultimately modulate the phosphorylation state and hence the activity of the enzyme. Whereas there have been numerous studies of GS activation in skeletal muscle, the modulators of GS in liver have been less completely evaluated. Studies to date indicate that glucose and insulin are the major physiologic effectors of hepatic GS activation (1,2). It appears that glucose and/or its metabolite glucose 6-phosphate activate hepatic GS by physically associating with critical regulatory enzymes upstream of GS or with GS itself (2,4,5), although other mechanisms have been suggested (6). The mechanisms by which insulin effects activation of GS and stimulation of glycogen synthesis in liver are poorly defined, although data have been obtained indicating that modulation of PP1-G (1, 7) as well as PKB/Akt (8) and GSK-3 (9) may be involved.
The peroxovanadium compounds, insulin-mimetic agents that activate the IRK by inhibiting an IRK-associated proteintyrosine phosphatase (21,22), mimic a range of insulin actions in target tissues including liver (23)(24)(25)(26)(27) and thus provide tools for amplifying our understanding of the control mechanisms operating on biologic processes influenced by insulin. In the present study, we sought to identify, in primary rat hepatocyte cultures, the molecular components and signal transduction pathways involved in the regulation of GS by both insulin and bpV(phen). Our results demonstrate that neither PKB/Akt activation nor GSK-3 inactivation is sufficient for GS activation and hence raise a question about the role suggested for this pathway in insulin-mediated GS activation. We demonstrate the involvement of two other pathways in the modulation of GS activation, accessed by bpV(phen) but not insulin, a rapamycinsensitive component that is not p70 s6k , and wortmannin-sensitive activation of PKC (/).

EXPERIMENTAL PROCEDURES
Materials-Porcine insulin was from Lilly. Wortmannin was purchased from Sigma, rapamycin and microcystin-LR were obtained from Calbiochem, and the mitogen-activated protein kinase/extracellular signal-regulated kinase kinase inhibitor PD98059 (28) was from New England Biolabs Inc. (Beverly, MA). Peroxovanadium bpV(phen) was synthesized and purified as reported previously (21). Collagenase was from Worthington. Cell culture media, antibiotics, and protein phosphatase assay system kit were form Life Technologies, Inc. (Life Technologies, Burlington, Ontario, Canada), and Vitrogen-100 was from Collagen Corp. (Toronto, Canada). [U-14 C]UDP-glucose and [␥-32 P]ATP were purchased from NEN Life Science Products, and okadaic acid was from Moana Bioproducts (Honolulu, HI). The GSK-3␤ antibody was kindly provided by Dr. J. R. Woodgett (Ontario Cancer Institute, Toronto, Canada). Anti-PKC (/) and goat anti-PKB␣/Akt1 antibodies as well as protein G-Sepharose were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-p70 s6k antibody, S6 peptide KKRNRTLTK, phosphoglycogen synthase peptide-2, and Crosstide were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY), and myristoylated PKC (/) pseudosubstrate peptide was obtained from Quality Controlled Biochemicals (Hopkington, MA). The Phosphoplus ® PKB (Thr 308 and Ser 473 ) antibody kit, including rabbit phosphospecific and total PKB/Akt antibodies, was purchased from New England Biolabs, Inc. (Beverly, MA). Protein A-Sepharose CL-4B was from Amersham Pharmacia Biotech (Montreal, Canada). All other reagents and chemicals were obtained from Sigma or Roche Molecular Biochemicals (Laval, Quebec, Canada) and were of the highest grade available.
Cell Culture and Stimulations-Hepatocytes, isolated from 120 -140-g fed male Harlan Sprague Dawley rats (Charles River, St-Constant, Canada) by in situ liver perfusion with collagenase, were seeded onto six-well plates (Corning Costar Corp., Cambridge, MA) coated with collagen (Vitrogen-100) at a density of 5 ϫ 10 5 cells/cm 2 . Primary cultures were incubated for 16 h in Dulbecco's modified Eagle's medium/Ham's F-12 (DMEM/F-12) containing 10% fetal bovine serum, 10 mM HEPES, 20 mM NaHCO 3 , 500 IU/ml penicillin, and 500 mg/ml streptomycin. For all studies, cells were rinsed twice in serum-free medium and serum-deprived for 3 h. For studies with insulin and bpV(phen), serum-free medium consisted of DMEM/F-12, which contains 18 mM glucose. For studies with glucose, serum-free medium consisted of glucose-free DMEM. Both serum-free media were supplemented with 1.25 mg/ml Fungizone, 0.4 mM ornithine, 2.25 mg/ml L-lactic acid, 25 nM selenium, and 10 nM ethanolamine, as described previously (27). The inhibitors wortmannin (100 nM), rapamycin (200 nM), and PD98059 (30 M) or vehicle were added to the cultures 45 min prior to stimulation with insulin, bpV(phen), and glucose, at the times and doses indicated in the figure legends. The PKC (/) pseudosubstrate (80 M) was added 90 min prior to cell stimulation. Following stimulation, cells were rinsed twice with ice-cold phosphate-buffered saline, pH 7.4, and resuspended in appropriate buffers as described below.
GS Activity Assay-GS activity was assessed according to the method of Thomas et al. (29), with modifications. Hepatocytes were resuspended in 300 l of 50 mM glycylglycine, pH 7.4, 100 mM sodium fluoride, 20 mM EDTA, 0.5% glycogen, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 10 g/ml leupeptin, 10 g/ml aprotinin, 10 g/ml pepstatin A. Cell suspensions were sonicated for 10 s and centrifuged at 12,000 ϫ g for 5 min. Supernatants (20 l) were added to 100 l of assay buffer (25 mM glycylglycine, pH 7.4, 0.275 mM [ 14 C]UDPglucose (0.12 Ci/ml, 0.44 Ci/mol), 1% glycogen, 1 mM EDTA, and 10 mM sodium sulfate), the reaction mixture was incubated for 30 min at 30°C, and 90 l was spotted on 2-cm 2 ET-31 (Whatman) filters. Filters, washed in 66% (v/v) ethanol, once for 30 min at 4°C and twice for 30 min at room temperature, were immersed in acetone, dried, and counted for incorporated 14 C radioactivity. Total GS activity (a plus b) was determined as described for GS a except that the assay buffer lacked sodium sulfate and contained 5 mM glucose 6-phosphate.
Phosphorylase Activity Assay-Phosphorylase activity was determined in the same supernatants in which GS activity was assessed by quantitating inorganic phosphate release from glycogen in the presence of an excess of glucose 6-phosphate (30). Phosphorylase a activity was assayed in the presence of caffeine (30). Twenty-five microliters of supernatant and 25 l of assay buffer (100 mM glucose 1-phosphate, 2% glycogen, 300 mM sodium fluoride, 1 mM caffeine, pH 6.1) were combined and incubated for 45 min at 30°C. Inorganic phosphate levels were determined spectrophotometrically according to the method of Sapru et al. (31). Total phosphorylase activity (a plus b) was measured in the absence of caffeine but in the presence of 5Ј-AMP and ammonium sulfate according to Stalmans and Hers (32).
Phosphorylase Phosphatase Activity Assay-32 P-Labeled phosphorylase a was prepared using a kit from Life Technologies Inc. (protein phosphatase assay system), according to the manufacturer's protocol. Phosphorylase phosphatase activity was determined by in vitro dephosphorylation of 32 P-labeled phosphorylase a according to the method of Cohen et al. (33) with some modifications. Cells were resuspended in 300 l of buffer A (20 mM imidazole-HCl, pH 7.6, 0.1% ␤-mercaptoethanol, 0.1 mM EDTA, 1 mg/ml bovine serum albumin, 2 mg/ml glycogen, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 10 g/ml leupeptin, 10 g/ml aprotinin, 10 g/ml pepstatin A, and 10 g/ml soybean trypsin inhibitor), sonicated for 10 s, and centrifuged at 2000 ϫ g for 3 min. Twenty microliters of supernatants (60 g of protein) were incubated with 20 l of okadaic acid (2 nM) in buffer A for 5 min at room temperature. The phosphorylase phosphatase reaction, initiated by adding 60 g of 32 P-labeled phosphorylase a, was carried out for 10 min at 30°C and terminated by adding 200 l of 20% ice cold trichloroacetic acid. The reaction tubes were kept on ice for 10 min and centrifuged at 12,000 ϫ g for 5 min, and free 32 P in the supernatants was counted in scintillation fluid. Nonenzymatic hydrolysis of 32 P from substrate accounted for less than 4% of the total amount released in the samples.
p70 s6k Kinase Assay-Hepatocytes were serum-starved in glucosefree DMEM for 3 h and stimulated with insulin and glucose at concentrations and for times indicated in the legend to Fig. 2. p70 s6k was assayed as described previously (27).
PKB/Akt Phosphorylation Assay-Western blot analysis of Thr 308 and Ser 473 of PKB/Akt was carried out using the Phosphoplus PKB antibody kit (New England Biolabs Inc.) according to the manufacturer's protocol. Gels were reprobed with a rabbit anti-PKB/Akt antibody (provided in the kit), and data were corrected for total PKB/Akt levels after densitometric scanning with a Bio-Rad model GS-700 imaging densitometer.
PKC (/) Activity Assay-48-h serum-starved hepatocytes were lysed in 50 mM Tris-HCl, pH 7.5, 0.15 M NaCl, 1% Triton X-100, 20 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml aprotinin. Cell lysates were centrifuged at 2000 ϫ g for 30 min, and supernatants (0.5 mg of protein) were incubated with mild agitation for 2 h at 4°C with 1 g of anti-PKC (/) antibody and for an additional hour in the presence of 50 l of 50% protein A-Sepharose. Immobilized immune complexes were recovered by centrifugation at 12,000 ϫ g for 2 min, washed four times with lysis buffer containing 0.5 M NaCl, and washed two times with 35 mM Tris-HCl, pH 7.5, 10 mM magnesium chloride, 1 mM EGTA, 2 mM sodium orthovanadate (kinase buffer). PKC (/) activity was assayed in 40 l of kinase buffer containing 60 M [␥-32 P]ATP (2 Ci) and 4 g of myelin basic protein for 30 min at room temperature. Reactions were stopped by the addition of Laemmli sample buffer, and phosphorylated myelin basic protein was resolved by SDS-polyacrylamide gel electrophoresis and quantitated by autoradiography.

Effect of Insulin, bpV(phen), and Glucose on Hepatocyte GS
Activation- Fig. 1A shows that insulin at a dose of 100 nM stimulates GS activity in our primary rat hepatocyte cultures by 170% of basal level. Whereas PD98059, an inhibitor of mitogen-activated protein kinase/extracellular signal-regulated kinase kinase, the upstream activator of p44/42 MAP kinases, did not inhibit insulin's stimulatory effect, wortmannin completely inhibited GS activation induced by insulin. bpV-(phen), at a dose producing IRK activation comparable with that observed with 100 nM insulin (22), stimulated GS to a significantly greater degree (250% of basal level (p Ͻ 0.001 compared with insulin). Here too, we observed that wortmannin entirely inhibited GS activation by bpV(phen). Thus, GS activation by both insulin and bpV(phen) is PI 3-kinase-dependent. As with insulin, PD98059 did not produce inhibition of bpV(phen)-induced GS activation.
Rapamycin, an inhibitor of p70 s6k activation (35,36), has been shown to inhibit insulin-induced GS activation in some but not all cell types (37)(38)(39). We found that rapamycin had only a modest (17%), and statistically not significant, inhibitory effect on insulin-induced GS activation (Fig. 1B). In contrast, rapamycin reduced the magnitude of bpV(phen)-induced GS activation by 40%, down to the level observed with insulin alone.
Glucose is also a major physiologic regulator of hepatic GS (reviewed in Ref. 2). We found that wortmannin partially inhibited glucose-mediated GS activation (data not shown) as previously reported by others (6). Fig. 1B shows that glucoseinduced GS activation was also partially reversed by rapamycin. These data strongly suggest that in addition to allosteric regulation, glucose effects GS activation by a signal transduction-based mechanism. Glucose did not activate p70 s6k (Fig. 2), suggesting that a mTOR-dependent event(s), other than p70 s6k activation, is involved in mediating GS activation by glucose.
The Role of Glycogen Phosphorylase Inactivation and Phosphorylase Phosphatase Activation-We sought to evaluate the role of phosphorylase and its phosphatase as downstream signaling elements leading to GS activation by insulin and bpV-(phen). Active glycogen phosphorylase exerts an inhibitory effect on hepatic GS phosphatase (PP1-G) activity and hence opposes GS activation (4). Inactivation of GS phosphatase by phosphorylase appears to be an important mechanism leading to GS activation by glucose in liver (4). Inactivation of liver phosphorylase by insulin in vivo has also been demonstrated and proposed to account for the activation of GS by the hormone (40,41). As observed in Fig. 3, insulin treatment of hepatocytes did not significantly inhibit phosphorylase activity, whereas glucose was, as predicted, inhibitory. Interestingly, phosphorylase activity was significantly inhibited by bpV(phen). Whereas rapamycin substantially reversed the inhibitory effect of bpV(phen), no such reversal of the glucose inhibitory effect was observed (Fig. 3). Thus, phosphorylase inactivation can be mediated by mTOR and/or mTOR-dependent signaling event(s) and may contribute to bpV(phen)-induced GS activation. It is unlikely that p70 s6k is involved, since both insulin and bpV(phen) activate p70 s6k in a rapamycinsensitive manner (27) but only bpV(phen)-induced GS activation showed sensitivity to rapamycin. Glucose-mediated phosphorylase inactivation is not sensitive to rapamycin; hence, its activation of GS involves a mTOR-dependent effector distinct from phosphorylase.
Since dephosphorylation and inactivation of phosphorylase is effected by phosphorylase phosphatase (4), we examined the effect of insulin and bpV(phen) on the activity of this enzyme. While bpV(phen) augmented phosphorylase phosphatase activity, insulin did not (Table I). This is consistent with the effect of these agents on phosphorylase activity. Rapamycin blocked bpV(phen)-induced phosphorylase phosphatase activation (Table I). Taken together, these data indicate that 1) neither phosphorylase phosphatase nor phosphorylase are directly involved in the acute activation of GS by insulin in liver, and 2) mTOR is a critical signaling element for bpV(phen)-mediated phosphorylase phosphatase activation and phosphorylase inactivation. (3) and is considered to be a major regulator of GS in several cell types including liver cells (1). We investigated the effect of insulin and bpV(phen) on GSK-3 activity and found that both agents were comparably inhibitory (Fig. 4). Rapamycin and PD98059 were without effect on both insulin-and bpV(phen)-mediated GSK-3 inactivation (Fig. 4). Although wortmannin treatment resulted in the complete reversal of insulin-induced GSK-3 inactivation, it only partially reversed bpV(phen)-mediated GSK-3 inactivation (Fig. 4).

Insulin and bpV(phen) Modulation of PKB/Akt and GSK-3 Activities-The serine/threonine protein kinase GSK-3 inhibits GS by phosphorylation
GSK-3 is a direct substrate for PKB/Akt in vitro (42), and there is evidence that PKB/Akt phosphorylates and inactivates GSK-3 in vivo (43,44). Phosphorylation of both Thr 308 and Ser 473 of PKB/Akt are required for its full activation (45). In primary rat hepatocytes, maximal PKB/Akt activation in response to insulin occurred at 2 min (8). In contrast, bpV(phen) induces maximal IRK-dependent signaling events at 20 min (21,22). We thus assessed the phosphorylation of Thr 308 and Ser 473 of PKB/Akt by Western blotting using phosphospecific antibodies toward Thr 308 and Ser 473 as well as PKB/Akt immunoprecipitable activity at these respective times in insulinand bpV(phen)-treated hepatocytes. We observed that insulininduced phosphorylation of Thr 308 and Ser 473 (Fig. 5A) and activation (Fig. 5B) of PKB/Akt were completely abolished by wortmannin as reported in many cell types (reviewed in Ref. 46). bpV(phen)-mediated phosphorylation of Thr 308 and Ser 473 and activation of PKB/Akt were 3-4-fold greater than that observed for insulin. This ratio was maintained after correcting for total PKB/Akt levels measured by immunoblotting the gels with a PKB/Akt antibody (Fig. 5A, bottom). In contrast to insulin, wortmannin did not inhibit Thr 308 and Ser 473 phosphorylation and only minimally inhibited (30%) the activity of PKB/Akt following bpV(phen) treatment (Fig. 5, A and B). Because wortmannin has previously been shown to inhibit insulin-and bpV(phen)-activated PI 3-kinase to the same extent in primary rat hepatocytes (27), then bpV(phen)-induced PKB/Akt phosphorylation and activation must be largely independent of PI 3-kinase. (47,48) and are involved in the regulation of insulin-stimulated glucose transport (17,49,50) and protein synthesis (18). We investigated whether one such isoform, PKC (/), plays a role in signaling to GS. We measured PKC (/) activity in response to insulin and bpV(phen) in cultured rat hepatocytes. Fig. 6 shows that insulin and bpV-(phen) activated PKC (/) by 1.5-and 2-fold, respectively, and that this activity was suppressed to below basal levels by the pseudosubstrate inhibitor of atypical PKC isoforms. Wortmannin reduced insulin-and bpV(phen)-induced PKC (/) activity to basal levels, consistent with a requirement for PI 3-kinase for its activation.

Role of PKC (/) in the Regulation of GS by Insulin and bpV(phen)-Evidence suggests that atypical isoforms of PKC act downstream of PI 3-kinase
Since PKC (/) and GS activation are both wortmannin-   sensitive, we investigated whether PKC (/) plays a role in insulin-and bpV(phen)-mediated signaling to GS. Fig. 7 shows that a PKC (/) pseudosubstrate peptide did not inhibit the stimulation of GS by insulin. In contrast, a marked inhibitory effect was noted in bpV(phen)-treated cells. Furthermore, the combination of the PKC (/) pseudosubstrate and rapamycin resulted in an additive inhibitory effect on bpV(phen)-induced GS activation, thus suggesting that PKC (/) and mTOR are on distinct signaling pathways leading to GS activation by bpV(phen). DISCUSSION In the present study, we evaluated the modulation of GS in primary rat hepatocytes by glucose, insulin, and the insulinmimetic compound bpV(phen). Previous work demonstrated the allosteric actions of glucose and/or its primary metabolite glucose 6-phosphate on the key enzymes associated with GS activation or with GS itself in liver (2,4). The present study showed that a signaling-based mechanism contributes to glucose-mediated GS activation. Thus, as previously observed (6), wortmannin partially reversed glucose-induced GS activation (data not shown), implicating PI 3-kinase in glucose signaling to GS. However, glucose did not activate PI 3-kinase activity when assayed in phosphotyrosine immunoprecipitates (data not shown). Because the lipid products of PI 3-kinase are potent inhibitors of hepatic glucose-6-phosphatase (51), the inhib-itory effect of wortmannin may be explained by a reduction in the levels of such lipid products, resulting in augmented glucose-6-phosphatase activity and hence reduced amounts of glucose 6-phosphate available for GS activation. A novel finding was the demonstration that glucose-induced GS activation is sensitive to rapamycin. A role for amino acids in the control of mTOR function has recently been proposed (52). Our data suggest a role for yet another nutrient, glucose, in modulating mTOR function. The rapamycin-sensitive glucose effector(s) remains to be identified, but it does not appear to be p70 s6k (Fig. 2) or phosphorylase (Fig. 3).
We examined the roles of PI 3-kinase, MAP kinase, and p70 s6k in effecting insulin-mediated GS activation in the cultured hepatocyte system. Wortmannin blocked insulin-induced GS activation, indicating a role for PI 3-kinase, as observed previously in various cell lines (38,39,53,54), rat adipocytes (55), and rat hepatocytes (8,56). PD98059, an inhibitor of mitogen-activated protein kinase/extracellular signal-regulated kinase kinase, the upstream activator of p44/42 MAP kinases, had no effect on insulin-mediated GS activation, excluding a role for MAP kinases. This is consistent with earlier work on 3T3-L1 adipocytes, cultured muscle cells, and rat diaphragm (37,39,57,58), and by Peak et al. (8) on primary rat hepatocytes. Previous studies have shown that rapamycin, an inhibitor of p70 s6k activation (35,36), partially or completely inhibits insulin-induced GS activation in rat diaphragm muscle in vitro (37), 3T3-L1 adipocytes (38), and human myoblasts (39). However, we failed to observe an inhibitory effect of rapamycin on insulin-induced GS activation in rat hepatocytes, consistent with the lack of effect of this inhibitor on insulinmediated hepatic glycogen synthesis (8). Thus, the stimulation  top and middle, respectively). Gels were reprobed with a rabbit anti-PKB/Akt antibody for total PKB/ Akt levels (A, bottom). The phosphorylation of Thr 308 and Ser 473 , corrected for total PKB/Akt, was increased by bpV(phen) 3.0-and 3.1-fold over that with insulin. Following wortmannin, these levels were 2.9and 2.8-fold those of insulin. bpV(phen)-stimulated PKB/Akt activity, corrected for total PKB/Akt, was 4.5-and 3.2-fold that with insulin alone in the absence and presence, respectively, of wortmannin. The data are representative of experiments on three independent hepatocyte preparations. of hepatic GS requires PI 3-kinase but not MAP kinase or p70 s6k activation.
We sought to define downstream effectors of PI 3-kinase involved in GS activation. GSK-3 is a multisubstrate serine/ threonine protein kinase considered to be a major regulator of hepatic GS in vivo (1), which inhibits GS by phosphorylation (3). Hence, inactivation of GSK-3 should promote GS activation. Insulin has been shown to inactivate GSK-3 in several cell types including hepatocytes (8,9,39,42,59,60). We found that insulin induced GSK-3 inactivation and that this was completely reversed by wortmannin. Furthermore, PKB/Akt phosphorylation and activation by insulin was also completely reversed by wortmannin. These findings support the view that the PI 3-kinase/PKB-Akt/GSK-3 axis is involved in effecting insulin-mediated GS activation in hepatocytes as has been inferred in skeletal muscle and adipocytes (61).
Our studies with bpV(phen) raise reservations about the significance of the above observations for understanding the mechanism of insulin action. Furthermore, since the activation of GS by bpV(phen) was significantly greater than that effected by insulin for the same level of IRK activation (22), we anticipated seeing the involvement of processes other than those delineated above. As in the case of insulin, wortmannin entirely reversed bpV(phen)-induced GS activation. This observation effectively excluded a role for MAP kinases, since bpV-(phen) strongly activates MAP kinases in the presence of wortmannin (27). Despite the abrogation of bpV(phen)-induced GS activation by wortmannin, GSK-3 remained markedly inactivated (Fig. 4) and PKB/Akt remained powerfully activated (Fig. 5). Studies by Peak et al. demonstrated that epidermal growth factor antagonizes insulin-induced glycogen synthesis (62) despite its ability to inactivate GSK-3 and activate, albeit much less markedly, PKB/Akt in primary rat hepatocytes (8). Thus, the entrainment of the sequence involving PKB/Akt and GSK-3 may play a necessary role but is clearly insufficient for GS activation in hepatocytes. The observation that bpV(phen)induced phosphorylation and activation of PKB/Akt were only minimally affected by wortmannin raises interesting mechanistic considerations. PKB/Akt activation appears to be effected when its pleckstrin homology domain interacts with phospholipid products of PI 3-kinase to induce a conformational change in PKB/Akt, rendering Thr 308 and Ser 473 accessible for phosphorylation and full activation of the enzyme (15). Because bpV(phen) is a powerful protein-tyrosine phosphatase inhibitor, our results point to a critical role for a proteintyrosine phosphatase(s) in attenuating PKB/Akt phosphorylation and activation. In light of the recent demonstration that the protein-tyrosine phosphatase PTEN is a lipid phosphatase that dephosphorylates phosphatidylinositol 3,4,5-trisphosphate (63,64) and that PKB/Akt activity and phosphorylation are constitutively elevated in PTEN-deficient mouse embryonic fibroblasts (65), it is possible that bpV(phen) inhibits PTEN, leading to sustained phosphoinositide 3-phosphate levels and hence PKB/Akt activity. It is also possible that bpV(phen) inhibits a protein-tyrosine phosphatase(s) involved in negatively regulating the activities of 3-phosphoinositide-dependent protein kinases 1 and 2, which have been described as the enzymes responsible for phosphorylating PKB/Akt at Thr 308 and Ser 473 , respectively (15). In keeping with the current model for PKB/Akt activation (reviewed in Ref. 66), this would mean that in the presence of wortmannin, sufficient PI 3-kinase lipid products accumulate due to PTEN inhibition so as to induce the conformational change in PKB/Akt necessary for Thr 308 and Ser 473 phosphorylation.
We sought to identify the PI 3-kinase-dependent downstream protein kinase(s) that were inactivated by wortmannin and hence could be implicated in modulating GS activation by bpV(phen). PKC (/) is activated in vitro by phosphatidylinositol 3,4,5-trisphosphate (47) and has been shown to lie downstream of PI 3-kinase in rat adipocytes (17) and L6 muscle cells (49). Our observation of a reversal by wortmannin of insulinand bpV(phen)-induced PKC (/) activation in rat hepatocytes is consistent with a role for PI 3-kinase lipid products in PKC (/) activation in hepatocytes. Whereas the addition of the PKC (/) pseudosubstrate inhibitor had no effect on insulininduced GS activation, it significantly suppressed bpV(phen)induced GS activation. Thus, PKC (/) is not the downstream kinase activated by insulin. The pseudosubstrate studies implicate the involvement of PKCs in bpV(phen)-induced GS activation. We know that the pseudosubstrate inhibits both typical and atypical PKCs (see Ref. 17). However, activation of typical PKCs leads to GS inactivation (1), and their inhibition should hence augment GS activation by bpV(phen). This is clearly not the case, as seen in Fig. 7. Furthermore, since wortmannin fully inhibits bpV(phen)-induced GS activation, it must subsume that portion which is PKC-dependent and hence implicates atypical PKCs (i.e. PKC (/)) whose activation is inhibited by wortmannin.
Since the PKC (/) pseudosubstrate only effected partial inhibition (ϳ35%) of bpV(phen)-induced GS activation, we explored the role of other pathways in this process. Unlike the case with insulin, rapamycin significantly reduced the magnitude of bpV(phen)-stimulated GS activity, pointing to the involvement of mTOR-dependent event(s) in bpV(phen) action. Indeed, our data demonstrate that the activation of PKC (/) combined with a rapamycin-sensitive step constitutes most of the stimulation of GS by bpV(phen). Prior in vivo studies implicated a role for phosphorylase inactivation in insulin-mediated hepatic GS activation (40,41). However, the confounding effects of circulating glucose levels rendered interpretation difficult. In the present study, we show that, in hepatocytes incu- bated with insulin, GS was activated, while phosphorylase activity was not significantly diminished. This agrees with earlier studies performed on rat hepatocytes (67,68) and indicates that phosphorylase inactivation is unlikely to be involved in the activation of hepatic GS by insulin. In contrast to these findings, we observed that bpV(phen) significantly modulated both phosphorylase phosphatase and phosphorylase and that the changes effected by bpV(phen) were rapamycin-sensitive. Indeed, this might explain the greater ability of bpV(phen) compared with insulin to stimulate GS. How mTOR signals to these enzymes is at present unclear, but p70 s6k is not involved, since its activation by insulin and bpV(phen) is both similar in magnitude and sensitive to rapamycin (Ref. 27 and data not shown). Effectors downstream of mTOR other than p70 s6k have been described (69), and mTOR itself possesses serine kinase activity (70) such that it may directly regulate phosphorylase phosphatase and/or phosphorylase by phosphorylation.
In summary, we have presented data indicating that the activation of PKB/Akt and the inhibition of GSK-3 in a PI 3-kinase-dependent manner may be necessary for insulin-induced GS activation but cannot be sufficient. We suggest that an additional pathway needs to be activated to realize the effect of insulin. Our observations with bpV(phen) identified mTOR and PKC (/) as other PI 3-kinase-dependent signaling components that contribute to GS activation in hepatocytes. They do not appear to be the additional postulated component activated by insulin in the course of achieving GS activation. Further work is required to identify the downstream effector(s) involved.