Inactivation of phosphorylase is a major component of the mechanism by which insulin stimulates hepatic glycogen synthesis Susan Aiston 1 , Matthew P. Coghlan 2 * and Loranne Agius 1 1 School of Clinical Medical Sciences, University of Newcastle upon Tyne, The Medical School, Newcastle upon Tyne, UK; 2 Department of Vascular Biology, GlaxoSmithKline, Harlow, Essex, UK Multiple signalling pathways are involved in the mechanism by which insulin stimulates hepatic glycogen synthesis. In this study we used selective inhibitors of glycogen synthase kinase-3 (GSK-3) and an allosteric inhibitor of phosphory- lase (CP-91149) that causes dephosphorylation of phos- phorylase a, to determine the relative contributions of inactivation of GSK-3 and dephosphorylation of phos- phorylase a as alternative pathways in the stimulation of glycogen synthesis by insulin in hepatocytes. GSK-3 inhibitors (SB-216763 and Li + ) caused a greater activation of glycogen synthase than insulin (90% vs. 40%) but a smaller stimulation of glycogen synthesis (30% vs. 150%). The contribution of GSK-3 inactivation to insulin stimulation of glycogen synthesis was estimated to be less than 20%. Dephosphorylation of phosphorylase a with CP- 91149 caused activation of glycogen synthase and translo- cation of the protein from a soluble to a particulate fraction and mimicked the stimulation of glycogen synthesis by insulin. The stimulation of glycogen synthesis by phos- phorylase inactivation cannot be explained by either inhi- bition of glycogen degradation or activation of glycogen synthase alone and suggests an additional role for translo- cation of synthase. Titrations with the phosphorylase inac- tivator showed that stimulation of glycogen synthesis by insulin can be largely accounted for by inactivation of phosphorylase over a wide range of activities of phos- phorylase a. We conclude that a signalling pathway invol- ving dephosphorylation of phosphorylase a leading to both activation and translocation of glycogen synthase is a critical component of the mechanism by which insulin stimulates hepatic glycogen synthesis. Selective inactivation of phos- phorylase can mimic insulin stimulation of hepatic glycogen synthesis. Keywords: glycogen synthase kinase-3; glycogen synthesis; insulin; liver; phosphorylase. Insulin, glucose and amino acids are the major physio- logical regulators of hepatic glycogen synthesis [1–3]. The signalling pathways activated by insulin in hepatocytes [3–7] bear similarities to the mechanisms identified in other cell types [8,9]. Binding of insulin to the receptor causes phosphorylation of insulin receptor substrates 1 and 2, recruitment and activation of phosphatidylinositol 3-kinase, resulting in formation of phosphatidylinositol 3,4,5-trisphosphate, which causes recruitment of phospha- tidylinositol-dependent kinase-1 to the plasma membrane [8,9]. The latter enzyme phosphorylates and activates protein kinase B, which in turn phosphorylates and inactivates glycogen synthase kinase-3 (GSK-3). As GSK- 3 causes inactivation of glycogen synthase by phosphory- lation at three sites, inactivation of GSK-3 allows glycogen synthase to become activated by dephosphorylation. Stimulation of glycogen synthesis by insulin in hepatocytes is counteracted by inhibitors of phosphatidylinositol 3-kinase [4–6] and is associated with activation of protein kinase B and inactivation of GSK-3 [5–7]. However, the contribution of this signalling pathway to the stimulation of glycogen synthesis by insulin in hepatocytes has not been determined. An alternative mechanism for regulation of glycogen synthase is through changes in the concentration of phosphorylase a, which inhibits glycogen synthase phos- phatase activity [1] by binding to the C-terminus of the glycogen targeting protein [10]. This protein was thought to be present only in liver, but expression in human skeletal muscle has also been reported [11]. Phosphorylase kinase catalyses the conversion of inactive phosphorylase b into active phosphorylase a by phosphorylation of a serine residue at the N-terminus [1]. Metabolic conditions that decrease the concentration of phosphorylase a through either inhibition of phosphorylase kinase or activation of phosphorylase phosphatase are expected to reverse the inhibition of glycogen synthase phosphatase by phosphory- lase a (Fig. 1). Dephosphorylation of phosphorylase a is a component of the mechanism by which high glucose concentration causes activation of glycogen synthase [1]. Binding of glucose to phosphorylase a makes the enzyme a better substrate for phosphorylase phosphatase, and the decrease in phosphorylase a reverses the inhibition of glycogen synthase phosphatase [1]. Correspondence to L. Agius, School of Clinical Medical Sciences, The Medical School, University of Newcastle upon Tyne, Newcastle upon Tyne, NE2 4HH, UK. Fax: + 44 191 2220723, Tel.: + 44 191 2227033, E-mail: Loranne.Agius@ncl.ac.uk Abbreviations: GSK-3, glycogen synthase kinase-3; MEM, minimum essential medium; PTG, protein targeting to glycogen. *Present address: Cardiovascular & Gastrointestinal Department, AstraZeneca, Macclesfield, Cheshire SK10 4TG, UK. (Received 21 March 2003, revised 30 April 2003, accepted 1 May 2003) Eur. J. Biochem. 270, 2773–2781 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03648.x There is a long-standing debate as to whether inactivation of phosphorylase is a component of the mechanism by which insulin activates glycogen synthase in hepatocytes, because inactivation of phosphorylase by insulin has been observed in some but not other studies [3,6]. In freshly isolated hepatocyte suspensions, sustained inactivation of phosphorylase by insulin has generally been observed only in the presence of glucagon or other counter-regulatory hormones. However, this experimental model shows small stimulatory effects of insulin on glycogen synthesis because of the catabolic state of glycogen turnover [3]. Short-term preculture of hepatocytes with dexamethasone allows recovery from the catabolic state and restores a large stimulatory effect of insulin on glycogen synthesis similar to the stimulation that occurs in vivo [3]. The contribution of phosphorylase inactivation to the stimulation of glycogen synthesis by insulin in this experimental model has not been tested. Recently, several potent allosteric inhibitors of phos- phorylase have been identified by high-throughput screens [12–14]. Some of these compounds inhibit glycogenolysis in hepatocytes by both allosteric inhibition of phosphorylase and inactivation (conversion of phosphorylase a to b) similar to glucose [14], whereas others inhibit glycogenolysis exclusively by allosteric inhibition [15,16] or by inactivation [16]. The latter compounds are very powerful experimental tools to investigate the role of the phosphorylation state of phosphorylase in metabolic control [12,17]. We demonstrate in this study, using selective inhibitors of GSK-3 [18–20] and a selective inhibitor of phosphorylase [12] that causes dephosphorylation of phosphorylase a [16], that inactiva- tion of GSK-3 in the absence of phosphorylase inactivation is a small component of the mechanism by which insulin stimulates hepatocyte glycogen synthesis. In contrast, dephosphorylation of phosphorylase can mimic insulin action and is a major component of the mechanism by which insulin stimulates glycogen synthesis. Materials and Methods Materials CP-91149 [12] and SB-216763 [18] were gifts from Pfizer Global Research & Development, Groton Laboratories, CT, USA, and SmithKline Beecham Pharmaceuticals, Harlow, Essex, UK, respectively. The adenovirus vectors for expression of wild-type GSK-3 and S9A-GSK-3 [21] were kindly provided by M. Birnbaum, Howard Hughes Medical Institute, University of Pennsylvania, Philadelphia, PA, USA. Hepatocyte isolation and cell culture Hepatocytes were isolated by collagenase perfusion of the liver of male Wistar rats (body weight 180–280 g) obtained from B & K, Hull, UK [22]. They were suspended in minimum essential medium (MEM) supplemented with 7% (v/v) newborn calf serum, and plated in multiwell plates. After cell attachment ( 4 h), the medium was replaced with serum-free MEM containing 5 m M glucose and 10 n M dexamethasone, and the hepatocytes were cultured for 16–18 h [17]. Treatment of hepatocytes with recombinant adenoviruses At 2 h after plating, the hepatocytes were incubated for 2 h in serum-free MEM without or with various titres of either Fig. 1. Model showing two alternative mechanisms for the activation of hepatic glycogen synthase by insulin involving either inactivation of GSK-3 or inactivation of phosphorylase. Insulin phosphorylates and inactivates (–) GSK-3 by activation of protein kinase B (PKB). GSK-3 (+, dephosphorylated form) phosphorylates and inactivates glycogen synthase (GS). Dephosphorylation (activation) of glycogen synthase (GS-b) by synthase phosphatase (SP) is inhibited by phosphorylase a (Phos-a). Conversion of phosphorylase a into phosphorylase b (Phos-b) by phosphorylase phosphatase (PP) is stimulated by glucose and by CP-91149. Insulin may convert phosphorylase a into phosphorylase b by either inhibition of phosphorylase kinase (PK) or activation of phosphorylase phosphatase. This reverses the inhibition of synthase phosphatase by phosphorylase a. 2774 S. Aiston et al.(Eur. J. Biochem. 270) Ó FEBS 2003 AdCMV-GSK-3 or AdCMV-S9A-GSK-3 for expression of wild-type or mutant GSK-3, respectively [21]. The medium was then replaced with serum-free MEM containing dexamethasone, and the cells were cultured for 16–18 h as above. Overexpression of GSK-3 was confirmed by immunoblotting after 18 h of culture. Incubations with insulin and inhibitors After preculture for 16–18 h in MEM with dexametha- sone, the medium was replaced with fresh MEM without dexamethasone and with the substrates and inhibitors indicated. Parallel incubations were performed for deter- mination of glycogen synthesis, glycogen synthase and phosphorylase a activity. For determination of glycogen synthesis, hepatocyte monolayers were incubated for 3 h in MEM containing [U- 14 C]glucose (2 lCiÆmL )1 )andthe glucose concentrations indicated. Inhibitors were dissolved in dimethylsulfoxide, and control incubations contained an equivalent volume (0.1%, v/v) of dimethylsulfoxide. Radiolabelled glycogen was determined as described previously [22]. Glycogen synthesis is expressed as nmol of glucose incorporated into glycogenÆ3h )1 Æ(mg cell pro- tein) )1 . Enzyme assays and immunoblotting At the end of the incubation, the plates were snap- frozen in liquid nitrogen and stored at )80 °C until assay. For determination of glycogen synthase, cells were extracted as previously described [23], and assays were performed on the whole homogenate (unless indicated otherwise) or in the 13 000 g supernatant and pellet fractions from the incorporation of UDP[6- 3 H]glucose into glycogen in the absence or presence of 6.7 m M Glc6P, representing active and total glycogen synthase, respectively [24]. Active glycogen synthase (– Glc6P) is expressed either as mUÆ(mg protein) )1 (nmolÆmin )1 Æmg )1 ) or as the activity ratio (– Glc6P/+ Glc6P). For deter- mination of phosphorylase a, the hepatocytes were extracted as described previously [17]. Phosphorylase a was determined in the supernatant spectrometrically by coupling to phosphoglucomutase and glucose-6- phosphate dehydrogenase [23]. Phosphorylase a activity is expressed as mUÆ(mg cell protein) )1 (nmolÆ min )1 Æmg )1 ). GSK-3 was assayed as in [5]. Activity is expressed as pmol 32 P incorporatedÆmin )1 Æ(mg protein) )1 . Immunoblotting for glycogen synthase and GSK-3 was performed after fractionation of the extracts by SDS/ PAGE. After transfer of the proteins to nitrocellulose, the membrane was probed with a rabbit antibody to rat liver glycogen synthase raised against residues IP KGKKKLHGEYKN(690–703) [25] or a goat antibody to GSK-3b (Santa Cruz, Santa Cruz, CA, USA) followed by incubation with the appropriate peroxi- dase-conjugated secondary antibody (Jackson Immuno- research, West Grove, PA, USA). Immunoreactive bands were visualized using an ECL kit (Amersham Biotech). Results are expressed as means ± SEM for the number of hepatocyte preparations indicated. Statistical analysis was by Student’s paired t test. Results Insulin causes rapid and sustained inactivation of phosphorylase When hepatocytes were precultured as described in Mate- rials and Methods and then incubated in fresh MEM without dexamethasone, insulin caused a rapid and sus- tained decrease in the activity of phosphorylase a at both 5m M glucose (40% decrease) and 25 m M glucose (60% decrease). The inactivation by insulin was observed within 15 min and had a similar time course at low and high glucose (Fig. 2A). This contrasts with the activation of Fig. 2. Time course of inactivation of phosphorylase a (A) and activa- tion of glycogen synthase (B) by 10 n M insulin. Hepatocytes were incubated with the glucose concentrations indicated for 4 h and with 10 n M insulin for the time intervals indicated. Values are means ± SEM, n ¼ 4–6. *P < 0.05; **P < 0.005 inactivation by insulin relative to control. Ó FEBS 2003 Role of phosphorylase in insulin signalling (Eur. J. Biochem. 270) 2775 glycogen synthase by insulin, which was more rapid at 25 m M than at 5 m M glucose (Fig. 2B). Inhibition of GSK-3 is less effective than insulin at stimulating glycogen synthesis Inactivation of GSK-3 by insulin in hepatocytes has been demonstrated previously [5,6]. In this study we confirmed that insulin inactivates GSK-3 (control, 0.21 ± 0.04; 10 n M insulin for 10 min, 0.14 ± 0.04, n ¼ 10, pmolÆ min )1 Æmg )1 , P < 0.05). The inactivation by insulin (33%) was less than the inactivation caused by calyculin A (50 n M ), a protein phosphatase inhibitor (0.025 ± 0.01, 88%, P < 0.05), indicating that insulin causes only partial inactivation of GSK-3. To determine the role of GSK-3 inactivation in insulin action, we used the selective GSK-3 inhibitor SB-216763, an arylindolemaleimide [18]. SB- 216763 caused a concentration-dependent increase in the activity ratio of glycogen synthase (Fig. 3A), in agreement with previous findings on other cell types [18], and it had no significant effect on the activity of phosphorylase a (control, 7.0 ± 2.7; 10 l M SB-216763, 6.0 ± 2.1; 25 l M SB-216763, 5.8 ± 1.9, n ¼ 4, mUÆmg )1 ). When compared with insulin, SB-216763 caused a larger activation of glycogen synthase (93% vs. 40%, Fig. 3B) but a smaller stimulation of glycogen synthesis (28% vs. 156%, Fig. 3C). Similarly 10 m M Li + , a potent inhibitor of GSK-3 [19,20], also caused a larger activation of glycogen synthase (73% vs. 40%) but a smaller stimulation of glycogen synthesis (28% vs. 156%) than insulin. In the combined presence of insulin and either SB-216763 or Li + , the activation of glycogen synthase was greater (P < 0.05) than with GSK-3 inhibitors alone, indicating that insulin activates glycogen synthase by mechanisms additional to inactivation of GSK-3 (Fig. 3B), and it was significantly (P < 0.05) greater than with insulin alone, consistent with the partial inactivation of GSK-3 by insulin. The rates of glycogen synthesis in the combined presence of GSK-3 inhibitors and insulin were the same as with insulin alone, despite the further activation of glycogen synthase (Fig. 3C). Inactivation of phosphorylase mimics insulin stimulation of glycogen synthesis We used CP-91149, a potent selective inhibitor of phos- phorylase [12] that causes conversion of phosphorylase a to b in hepatocytes [16,17], to determine the role of dephos- phorylation of phosphorylase in the regulation of glycogen synthesis. CP-91149 (2.5 l M ) caused a similar stimulation of glycogen synthesis and inactivation of phosphorylase to that of insulin (Fig. 4). It is noteworthy that 1,4-dideoxy-1,4- imino- D -arabinitol, an allosteric inhibitor of phosphorylase [26] that does not cause dephosphorylation of phosphory- lase a [16], does not stimulate glycogen synthesis [15]. This was confirmed in parallel incubations in the present study (results not shown), indicating that stimulation of glycogen synthesis by CP-91149 is due to dephosphorylation of phosphorylase a rather than inhibition of glycogen degra- dation or cycling. Fig. 3. Effects of insulin and GSK-3 inhibitors on glycogen synthase and glycogen synthesis. Hepatocytes were incubated for 3 h in MEM containing 10 m M glucose and the additions indicated. (A) Activation of glycogen synthase by various concentrations of SB-216763. (B) and (C) Effects of 25 l M SB-216763 or 10 m M LiCl in the absence (open bars) or presence (closed bars) of 10 n M insulin on the activity of glycogen synthase and rates of glycogen synthesis. Incubation mixtures for determination of glycogen synthesis contained [U- 14 C]glucose as described in Materials and methods. The total activity of glycogen synthase assayed in the presence of Glc6P was 1.05 ± 0.16 (A) and 0.94 ± 0.14 (B) mUÆmg )1 and was not affected by the incubation conditions tested. Values are means ± SEM, n ¼ 4. *P < 0.05 relative to no insulin; #P < 0.05 relative to no inhibitor. 2776 S. Aiston et al.(Eur. J. Biochem. 270) Ó FEBS 2003 Stimulation of glycogen synthesis but not activation of glycogen synthase by insulin can be explained by inactivation of phosphorylase To determine the role of inactivation of phosphorylase in the stimulation of glycogen synthesis by insulin, we determined the combined effects of insulin and various concentrations of CP-91149. The effects of insulin on phosphorylase inactivation, glycogen synthase activation (Fig. 5A), and stimulation of glycogen synthesis (Fig. 5B) were significant at all concentrations of inhibitor tested (P < 0.05). When the activity of glycogen synthase was plotted against the respective activity of phosphorylase a, the curves for untreated (control) and insulin-treated incubations were not superimposed (Fig. 5C), indicating that activation of glycogen synthase by insulin cannot be fully explained by inactivation of phosphorylase. However, the corresponding curves for glycogen synthesis against phosphorylase a (Fig. 5D) in the absence and presence of insulin were superimposed over a wide range of activities of phosphorylase down to 40% of basal activity. This indicates that insulin stimulates glycogen synthesis predominantly by inactivation of phosphorylase over the range 3–7 mUÆmg )1 but also by additional mechanisms at activities below 3mUÆmg )1 . The stimulation of glycogen synthesis by insulin and phosphorylase inactivation are greater than can be explained by activation of glycogen synthase In view of the above findings that inactivation of phosphorylase mimics insulin stimulation of glycogen synthesis (Figs 4 and 5), whereas inactivation of GSK-3 has only a small stimulatory effect despite the large activation of glycogen synthase (Fig. 3), we tested the relative roles of GSK-3 and phosphorylase a inthesameexperimentsby modulating GSK-3 activity by adenovirus-mediated expres- sion of wild-type GSK-3 or a constitutively active mutant (S9A-GSK-3) [21,27] or by inhibiting GSK-3 activity with SB-216763 (Fig. 6). Incubations for determination of glyco- gen synthesis were performed in the absence or presence of insulin or CP-91149, and rates of glycogen synthesis were expressed relative to the corresponding activity of glycogen synthase (assayed in the absence of Glc6P). In the absence of insulin or CP-91149, the relation between the rate of glycogen synthesis and the activity of glycogen synthase was sigmoidal (Fig. 6, solid line). In cells expressing endogenous GSK-3, the rate of glycogen synthesis was at or near the plateau of the sigmoidal curve. The GSK-3 inhibitor caused a twofold increase in the activity of glycogen synthase, but had little effect on flux. Conversely, GSK-3 overexpression caused a marked decrease in both the activity of glycogen synthase and the rate of glycogen synthesis. Both insulin (dashed line) and CP-91149 (dotted line) caused an upward shift in the glycogen synthesis versus glycogen synthase curve, indica- ting that stimulation of glycogen synthesis cannot be explained by activation of glycogen synthase alone. Inactivation of phosphorylase causes translocation of glycogen synthase To test for other mechanisms that might explain the stimulation of glycogen synthesis by phosphorylase inacti- vation, we determined the effects of CP-91149 on the subcellular distribution of glycogen synthase. Previous studies showed that high glucose concentration causes translocation of glycogen synthase from a soluble to a particulate fraction by a Glc6P-dependent mechanism [28]. We therefore determined the distribution of glycogen synthase total activity (assayed in the presence of Glc6P) or protein (by immunoblotting), between the supernatant and particulate fractions in cells treated with CP-91149 or various concentrations of glucose and insulin. CP-91149 caused translocation of glycogen synthase from the super- natant to the particulate fraction, as shown by immuno- blotting (Fig. 7A) or the radiochemical assay (Fig. 7B), similar to the combined effect of 25 m M glucose and insulin (Fig. 7B). As phosphorylase a andglycogensynthasebind to the glycogen targeting protein (PTG) by a mutually exclusive mechanism [29,30], we tested for the presence of PTG in the particulate fraction. Immunoreactivity to PTG Fig. 4. Effects of insulin and CP-91149 on phosphorylase a (A) activity and glycogen synthesis (B) at various glucose concentrations. Hepato- cytes were incubated for 3 h with the glucose concentrations indicated in the absence (s) or presence of 10 n M insulin (j)or2.5l M CP-91149 (m). Values are means ± SEM, n ¼ 4. Effects of insulin and CP-91149 were significant (P < 0.05) at all concentrations of glucose. Ó FEBS 2003 Role of phosphorylase in insulin signalling (Eur. J. Biochem. 270) 2777 (at 36 kDa) was detected in both supernatant and the particulate fractions (results not shown). Discussion Inactivation of GSK-3 is considered to be a key component of the mechanism by which insulin stimulates glycogen synthesis [8,9]. However, the quantitative contribution of this mechanism compared with other signalling pathways to the stimulation of glycogen synthesis by insulin has not been evaluated. In this study we used selective inhibitors of GSK-3 [18–20] and a selective inhibitor of phosphorylase that causes dephosphorylation of the enzyme [12,16] to determine the contributions of inactivation of GSK-3 and dephosphorylation of phosphorylase to the mechanism by which insulin stimulates glycogen synthesis in liver cells. Three key conclusions can be drawn: (a) that inactivation of phosphorylase is an essential component of the mechanism by which insulin stimulates glycogen synthesis and it can account for the stimulation of glycogen synthesis by insulin over a wide range of activities of phosphorylase a;(b)that suppression of GSK-3 activity, in the absence of inactiva- tion of phosphorylase, causes a large activation of glycogen synthase but a small stimulation of glycogen synthesis; (c) that the stimulation of glycogen synthesis by inactivation of phosphorylase is associated with both activation and translocation of glycogen synthase, and that the former mechanism alone cannot explain the stimulation of glyco- gen synthesis. This suggests that translocation of glycogen synthase may be an essential component of the mechanism by which dephosphorylation of phosphorylase leads to stimulation of glycogen synthesis. GSK-3 inactivation has been implicated in the mechan- ism by which insulin stimulates glycogen synthesis on the basis of three pieces of evidence. First, inactivation of GSK-3 by insulin occurs in several cell types [8,9]. Second, GSK-3 causes phosphorylation and inactivation of glycogen syn- thase whereas inhibition of GSK-3 in intact cells causes activation of glycogen synthase [21]. Third, overexpression of a constitutively active GSK-3 mutant overrides the activation of glycogen synthase [27] and stimulation of glycogen synthesis caused by insulin in adipocytes [21]. In the present study a clear role for GSK-3 in the regulation of glycogen synthase has been demonstrated by overexpression Fig. 5. Stimulation of glycogen synthesis but not activation of glycogen synthase by insulin can be largely explained by inactivation of phosphorylase. Hepatocytes were incubated for 3 h with 15 m M glucose and the concentrations of CP-91149 indicated in either the absence (open symbols) or presence (closed symbols) of 10 n M insulin. (A) Phosphorylase a and active glycogen synthase (– Glc6P) expressed as mUÆ(mg protein) )1 .(B) Glycogen synthesis. (C) Active glycogen synthase vs. phosphorylase a. (D) Glycogen synthesis vs. phosphorylase a. Values are means ± SEM, n ¼ 4. 2778 S. Aiston et al.(Eur. J. Biochem. 270) Ó FEBS 2003 of wild-type or S9A-GSK-3, which caused marked inacti- vation of glycogen synthase, and by the GSK-3 inhibitors, which caused twofold activation of glycogen synthase. The greater activation of glycogen synthase by GSK3 inhibitors (90%) compared with insulin (40%) can be explained by the small fractional inactivation of GSK-3 caused by insulin (33%). Whereas inactivation of glycogen synthase by GSK-3 overexpression resulted in inhibition of glycogen synthesis, activation of glycogen synthase by GSK-3 inhibition had little effect on glycogen synthesis. This is explained by the sigmoidal relation between glycogen synthesis and glycogen synthase activity at endogenous activities of phosphorylase a. Thus GSK-3 inactivation in the absence of inactivation of phosphorylase had little impact on flux through glycogen synthesis in hepatocytes, despite activa- tion of glycogen synthase. Accordingly the inactivation of GSK-3 caused by insulin accounted for less than 20% of the stimulation of glycogen synthesis. The additive activation of glycogen synthase by insulin and GSK-3 inhibitors indicates that insulin activates glycogen synthase by mechanisms additional to inactivation of GSK-3. This can be explained, at least in part, by the inactivation of phosphorylase by insulin. Two pieces of evidence from the studies with the phosphorylase inhibitor (CP-91149) support a role for dephosphorylation of phosphorylase a as a critical compo- nent of the mechanism by which insulin stimulates glycogen synthesis. In the absence of insulin, the phosphorylase inhibitor CP-91149 caused a dose-dependent stimulation of glycogen synthesis that exceeded the stimulation caused by insulin, at concentrations of inhibitor that inactivated phosphorylase a by 85%. This stimulation of glycogen synthesis cannot be explained by inhibition of glycogen degradation or by a decrease in catalytic activity of phosphorylase, because it is not mimicked by another selective inhibitor of phosphorylase that inhibits glucagon- stimulated glycogenolysis [15,16] but does not cause dephosphorylation of phosphorylase a [16]. This lack of stimulation of glycogen synthesis by an allosteric inhibitor of phosphorylase [15,31] is further evidence against cycling between glycogen synthesis and degradation as shown also in other in vitro models [32]. The stimulation of glycogen synthesis by CP-91149 also cannot be explained by nonspecific effects on glucose metabolism because the Fig. 6. Relation between glycogen synthesis and active glycogen syn- thasedeterminedbymodulationofGSK-3activitybyGSK-3overex- pression or inhibition. (A) GSK-3b was determined by immunoblotting of hepatocytes that were either untreated (End) or treated with wild- type AdCMV-GSK-3 (w) or mutant AdCMV-S9A-GSK-3 (m) and cultured for 18 h. (B) Hepatocytes were either untreated (open sym- bols) or treated (closed symbols) with AdCMV-GSK-3 (w) or AdC- MV-S9A-GSK-3 (m). After 18 h of culture, they were incubated for 3 h in medium containing 10 m M glucose without (s,d)orwith10 n M insulin (h,j)or2.5l M CP-91149 (n,m) for determination of glyco- gen synthesis and active glycogen synthase (– Glc6P). Where indicated (Inh), 25 l M SB-216763 was added during the 3 h incubation to inhibit GSK-3 activity. Glycogen synthesis is plotted against the respective glycogen synthase activity. Values are means ± SEM, n ¼ 4. Fig. 7. Inactivation of phosphorylase causes translocation of glycogen synthase. Hepatocytes were incubated for 60 min with 5 m M glucose (A) or the glucose concentrations indicated (B) without (Con) or with 10 l M CP-91149 (CP) or 10 n M insulin (Ins). The cell homogenates were centrifuged at 13 000 g, and total glycogen synthase (GS) was determined in the supernatant (SN) and pellet (P) fractions by immunoblotting (A) or radiochemically (B). (A) Immunoblot and corresponding densitometry. (B) Total glycogen synthase activity (assayed in the presence of Glc6P) in the pellet as a percentage of supernatant plus pellet activity. Values are means ± SEM, n ¼ 3. Ó FEBS 2003 Role of phosphorylase in insulin signalling (Eur. J. Biochem. 270) 2779 compound had no effect on glucose phosphorylation, glycolysis, or the Glc6P content of hepatocytes [17]. The stimulation of glycogen synthesis by CP-91149 therefore shows that inactivation (dephosphorylation) of phosphory- lase can mimic the stimulatory effect of insulin. In the experiments in which the effects of insulin on glycogen synthase and synthesis were tested in the presence of various concentrations of the phosphorylase inhibitor and expressed relative to the respective activity of phos- phorylase a, the stimulation of glycogen synthesis but not the activation of glycogen synthase could be largely accounted for by inactivation of phosphorylase over a wide range of activities of phosphorylase a (> 3 mUÆmg )1 ). These findings are in agreement with the high flux control coefficient of phosphorylase on glycogen synthesis [17] and with the GSK-3 inhibitor studies that show a sigmoidal relation between glycogen synthesis and glycogen synthase activity, with a basal rate of flux at or near the plateau. The phosphorylase inactivator, CP-91149, caused both activation of glycogen synthase and translocation of the enzyme to the pellet. Previous studies have shown that high glucose concentration causes translocation of glycogen synthase in hepatocytes to the cell periphery [25] and from the supernatant to the particulate fraction [28]. This effect of glucose is abolished by inhibition of glucose phosphoryla- tion and correlates with the accumulation of Glc6P [28], suggesting a role for Glc6P in translocation of glycogen synthase. However, in some metabolic conditions, translo- cation of glycogen synthase does not correlate with Glc6P [33]. Two explanations are possible. Either there is subcel- lular compartmentation of Glc6P [34] or additional mech- anisms may be involved in mediating translocation of glycogen synthase. CP-91149 does not affect the Glc6P content of hepatocytes [17]. Although we cannot rule out subcellular changes in the concentration of Glc6P in the presence of the phosphorylase inactivator, we suggest that additional factors may be involved in translocation of glycogen synthase and that phosphorylase a itself may be an important determinant of the subcellular compartmentation of glycogen synthase through competitive binding to a common targeting protein. PTG [29] has a binding site at the C-terminus, which binds phosphorylase a and glycogen synthase by a mutually exclusive mechanism [30]. If the binding affinity of PTG is greater for phosphorylase a than for phosphorylase b, then dephosphorylation of phosphory- lase a may favour binding of glycogen synthase. Overex- pression of PTG in hepatocytes activates glycogen synthase and stimulates glycogen synthesis [35,36]. As the twofold activation of glycogen synthase by inhibition of GSK-3 has a negligible effect on glycogen synthesis (this study), the stimulation of glycogen synthesis by PTG overexpression could be explained by either combined activation of glycogen synthase and inactivation of phosphorylase [35] or combined activation of glycogen synthase and translocation or binding to PTG. The hypothesis that inactivation of phosphorylase by CP-91149 or insulin may cause binding of glycogen synthase to glycogen targeting proteins remains to be tested. Until now, studies on insulin signalling in relation to glycogen synthesis have largely focused on mechanisms leading to activation of protein kinase B and inactivation of GSK-3. We demonstrate in this study that inactivation of phosphorylase but not inhibition of GSK-3 mimics insulin stimulation of glycogen synthesis in hepatocytes and that insulin action on glycogen synthesis can be largely accoun- ted for by phosphorylase inactivation. Accordingly, studies on insulin signalling should address the mechanism that leads to dephosphorylation of phosphorylase. In hepatocyte suspensions incubated in the absence of amino acids, insulin activates protein kinase B but does not activate glycogen synthase, suggesting that activation of protein kinase B alone is not sufficient to elicit the anabolic effects of insulin [37]. On the basis of the present findings that inactivation of phosphorylase is essential for stimulation of glycogen synthesis by insulin and it is also a contributing factor to the activation of glycogen synthase, the question could be raised whether medium amino acids are either essential for, or have a permissive role in mediating, the inactivation of phosphorylase by insulin? Acknowledgements This work was supported by Diabetes UK and by the Medical Research Council. We thank Dr Morris Birnbaum for the gift of GSK-3 adenoviruses, Drs Joan Guinovart and Rez Halse for the antibodies to glycogen synthase a and PTG, respectively, and Dr Judith Treadway for CP-91149 and helpful advice. References 1. Bollen, M., Keppens, S. & Stalmans, W. (1998) Specific features of glycogen metabolism in liver. Biochem. J. 336, 19–31. 2. Meijer, A.J., Baquet, A., Gustafson, L., van Woerkom, G. M. & Hue, L. (1992) Mechanism of activation of liver glycogen synthase by swelling. J. Biol. Chem. 267, 5823–5828. 3. Agius, L., Alam, N. & Aiston, S. (2000) Short-term regulation by insulin of glucose metabolism in isolated and cultured hepatocytes. In The Hepatocyte Review (Berry, M.N. & Edwards A.M., eds), pp. 317–341. Kluwer Academic Publishers, Dordrecht, the Netherlands. 4. Carlsen, J., Christiansen, K. & Vinten, J. (1997) Insulin stimulated glycogen synthesis in isolated rat hepatocytes: effect of protein kinase inhibitors. Cell Signal. 9, 447–450. 5. Peak, M., Rochford, J.J., Borthwick, A.C., Yeaman, S.J. & Agius, L. (1998) Signalling pathways involved in the stimulation of glycogen synthesis by insulin in rat hepatocytes. Diabetologia 41, 16–25. 6. Lavoie, L., Band, C.J., Kong, M., Bergeron, J.M. & Posner, B.I. (1999) Regulation of glycogen synthase in rat hepatocytes. Evidence for multiple signaling pathways. J. Biol. Chem. 274, 28279–28285. 7. Walker, K.S., Deak, M., Paterson, A., Hudson, K., Cohen, P. & Alessi, D.R. (1998) Activation of protein kinase B beta and gamma isoforms by insulin in vivo and by 3-phosphoinositide- dependent protein kinase-1 in vitro: comparison with protein kinase B alpha. Biochem. J. 331, 299–308. 8. Vanhaesebroeck, B. & Alessi, D.R. (2000) The PI3K-PDK1 connection: more than just a road to PKB. Biochem. J. 346, 561–576. 9. Saltiel, A.R. & Kahn, C.R. (2001) Insulin signaling and regulation of glucose and lipid metabolism. Nature (London) 414, 799–806. 10. Armstrong, C.G., Doherty, M.J. & Cohen, P.T. (1999) Identifi- cation of the separate domains in the hepatic glycogen-targeting subunit of protein phosphatase-1 that interact with phosphorylase a, glycogen and protein phosphatase 1. Biochem. J. 336, 699–704. 11. Munro, S., Cuthbertson, D.J.R., Cunningham, J., Sales, M. & Cohen, P.T.W. (2002) Human skeletal muscle expresses a 2780 S. Aiston et al.(Eur. J. Biochem. 270) Ó FEBS 2003 glycogen targeting subunit of PP1 that is identical to the insulin-sensitive glycogen targeting subunit of liver. Diabetes 51, 591–598. 12. Martin, W.H., Hoover, D.J., Armento, S.J., Stock, I.A., McPherson,R.R.,Danley,D.E.,Stevenson,R.W.,Barrett,E.J.& Treadway, J.L. (1998) Discovery of a human glycogen phos- phorylase inhibitor that lowers blood glucose in vivo. Proc. Natl Acad. Sci. 95, 1776–1781. 13. Treadway, J.L., Mendys, P. & Hoover, D.L. (2001) Glycogen phosphoryase inhibitors for treatment of type 2 diabetes mellitus. Exp. Opin. Invest. Drugs 10, 439–454. 14. Bergans, N., Stalmans, W., Goldman, S. & Vanstapel, F. (2000) Molecular mode of inhibition of glycogenolysis in rat liver by the dihydropyridine derivative BAY R3401. Inhibition and inactivation of glycogen phosphorylase by an activated metabolite. Diabetes 49, 1419–1426. 15. Andersen, B., Rassov, A., Westergaard, N. & Lundgren, K. (1999) Lundgren K: inhibition of glycogenolysis in primary rat hepato- cytes by 1,4-dideoxy-1,4-imino- D -arabinitol. Biochem. J. 342, 545–550. 16. Latsis, T., Andersen, B. & Agius, L. (2002) Diverse effects of two allosteric inhibitors on the phosphorylation state of phosphorylase in hepatocytes. Biochem. J. 368, 309–316. 17. Aiston, S., Hampson, L., Gomez-Foix, A.M., Guinovart, J.J. & Agius, L. (2001) Hepatic glycogen synthesis is highly sensitive to phosphorylase activity: evidence from metabolic control analysis. J. Biol. Chem. 276, 23856–23866. 18. Coghlan, M.P., Culbert, A.A., Cross, D.A., Corcoran, S.L., Yates,J.W.,Pearce,N.J.,Rausch,O.L.,Murphy,G.J.,Carter, P.S., Cox, L., Mills, D., Brown, M.J., Haigh, D., Ward, R.W., Smith,D.G.,Murray,K.J.,Reith,A.D.&Holder,J.C.(2000) Selective small molecule inhibitors of glycogen synthase kinase-3 modulate glycogen metabolism. Chem. Biol. 7, 793–803. 19. Klein, P.S. & Melton, D.A. (1996) A molecular mechanism for the effect of lithium on development. Proc.NatlAcad.Sci.USA93, 8455–8459. 20. Stambolic, V., Ruel, L. & Woodgett, J.R. (1996) Lithium inhibits glycogen synthase kinase-3 activity and mimics wingless signalling in intact cells. Curr. Biol. 6, 1664–1668. 21. Summers, S.A., Kao, A.W., Kohn, A.D., Backus, G.S., Roth, R.A., Pessin, J.E. & Birnbaum, M.J. (1999) The role of glycogen synthase kinase-3 beta in insulin-stimulated glucose metabolism. Biol. Chem. 274, 17934–17940. 22. Agius, L., Peak, M. & Alberti, K.G.M.M. (1990) Regulation of glycogen synthesis from glucose and gluconeogenic precursors in periportal and perivenous rat hepatocytes. Biochem. J. 266, 91–102. 23. Aiston, S. & Agius, L. (1999) Leptin enhances glycogen storage in hepatocytes by inhibition of phosphorylase and exerts an additive effect with insulin. Diabetes 48, 15–20. 24. Thomas, J.A., Schlender, K.K. & Larner, J. (1968) A rapid filter paper assay for UDP-glucose-glycogen glucosyltransferase, including an improved biosynthesis of UDP 14 C-glucose. Anal. Biochem. 25, 486–499. 25. Garcia-Rocha, M., Roca, A., De La Iglesia, N., Baba, O., Fernandez-Novell, J.M., Ferrer, J.C. & Guinovart, J.J. (2001) Intracellular distribution of glycogen synthase and glycogen in primary cultured rat hepatocytes. Biochem. J. 357, 17–24. 26. Fosgerau, K., Westergaard, N., Quistorff, B., Grunnet, N., Kristiansen, M. & Lundgren, K. (2000) Kinetic and functional characterization of 1,4-dideoxy-1,4-imino- D -arabinitol: a potent inhibitor of glycogen phosphorylase with anti-hyperglycaemic effect in ob/ob mice. Arch. Biochem. Biophys. 380, 274–284. 27. Eldar-Finkelman, H., Argast, G.M., Foord, O., Fischer, E.H. & Krebs, E.G. (1996) Expression and characterization of glycogen synthase kinase-3 mutants and their effect on glycogen synthase activity in intact cells. Proc. Natl Acad. Sci. USA 93, 10228–10223. 28. Fernandez-Novell, J.M., Arino, J., Vilaro, S., Bellido, D. & Guinovart, J.J. (1992) Role of glucose 6-phosphate in the trans- location of glycogen synthase in rat hepatocytes. Biochem. J. 288, 497–501. 29. Printen, J.A., Brady, M.J. & Saltiel, A.R. (1997) PTG a protein phosphatase-1 binding protein with a role in glycogen metabolism. Science 275, 1475–1478. 30. Fong, N.M., Jensen, T.C., Shah, A.S., Parekh, N.N., Saltiel, A.R. & Brady, M.J. (2000) Identification of binding sites on protein targeting to glycogen for enzymes of glycogen metabolism. J. Biol. Chem. 275, 35034–35039. 31. Fosgerau, K., Breinholt, J., McCormack, J.G. & Westergaard, N. (2002) Evidence against glycogen cycling of gluconeogenic substrates in various liver preparations. J. Biol. Chem. 277, 28648–28655. 32. Grunnet, N., Jensen, S. & Dich, J. (1993) Absence of glycogen cycling in cultured rat hepatocytes. Arch. Biochem. Biophys. 309, 18–23. 33. Seoane, J., Gomez-Foix, A.M., O’Doherty, R.M., Gomez-Ara, C., Newgard, C.B. & Guinovart, J.J. (1996) Glucose 6-phosphate produced by glucokinase, but not hexokinase I, promotes the activation of hepatic glycogen synthase. J. Biol. Chem. 271, 23756–23760. 34. Gomis, R.R., Favre, C., Garcia-Rocha, M., Fernandez-Novell, J.M., Ferrer, J.C. & Guinovart, J.J. (2003) Glucose 6-phosphate produced by gluconeogenesis and by glucokinase is equally effective in activating hepatic glycogen synthase. J. Biol. Chem. 278, 9740–9646. 35. Berman, H.K., O’Doherty, R.M., Anderson, P. & Newgard, C.B. (1998) Overexpression of protein targeting to glycogen (PTG) in rat hepatocytes causes profound activation of glycogen synthesis independent of normal hormone- and substrate-mediated reg- ulatory mechanisms. J. Biol. Chem. 273, 26421–26425. 36. Gasa, R., Jensen, P.B., Berman, H.K., Brady, M.J., DePaoli- Roach, A.A. & Newgard, C.B. (2000) Distinctive regulatory and metabolic properties of glycogen-targeting subunits of protein phosphatase-1 (PTG, GL, GM/RGl) expressed in hepatocytes. J. Biol. Chem. 275, 26396–26403. 37. Krause,U.,Bertrand,L.,Maisin,L.,Rosa,M.&Hue,L.(2002) Signalling pathways and combinatory effects of insulin and amino acids in isolated rat hepatocytes. Eur. J. Biochem. 269, 3742–3750. Ó FEBS 2003 Role of phosphorylase in insulin signalling (Eur. J. Biochem. 270) 2781 . Inactivation of phosphorylase is a major component of the mechanism by which insulin stimulates hepatic glycogen synthesis Susan Aiston 1 , Matthew P. Coghlan 2 * and Loranne Agius 1 1 School. both activation and translocation of glycogen synthase is a critical component of the mechanism by which insulin stimulates hepatic glycogen synthesis. Selective inactivation of phos- phorylase can. doi:10.1046/j.1432-1033.2003.03648.x There is a long-standing debate as to whether inactivation of phosphorylase is a component of the mechanism by which insulin activates glycogen synthase in hepatocytes, because inactivation of