Activation of glycogen synthase in rat adipocytes by insulin and glucose involves increased glucose transport and phosphorylation.

Incubation of fat cells with insulin, concanavalin A, or H,Oz stimulated the incorporation of o-[U-14Clglucose into glycogen. Incubation with either concanavalin A or insulin alone activated glycogen synthase. H,O, alone did not activate glycogen synthase, but in the presence of glucose it increased the percentage of glycogen synthase Z activity. The activation of glycogen synthase by H,02 and glucose was abolished by phloridzin or cytochalasin B. These inhibitors of glucose transport also abolished the action of glucose to potentiate the effect of concanavalin A on glycogen synthase. Glucose, mannose, and Z-deoxyglucose increased glycogen synthase Z activity and potentiated the effect of insulin on glycogen synthase. Incubation of cells with 3-O-methylglucose, galactose, or fructose was without effect on glycogen synthase Z activity and did rot result in a potentiation of the effect of insulin. Compared to other hexoses, 2deoxyglucose was much more effective in activating glycogen synthase. Increases in 2-deoxyglucose-6-P produced during 5-min incubations with varying concentrations of 2deoxyglucose paralleled the increases in glycogen synthase Z activity, although phosphorylase a activity and concentrations of ATP and CAMP were unchanged. Glucose-6-P and 2-deoxyglucose-6-P (but not glucose, mannose, or 2-deoxyglucose) increased the activity of glycogen synthase phosphatase which was monitored by following the increase in endogenous glycogen synthase Z activity. These results suggest that glucose-6-phosphate formation is required for the activation of glycogen synthase by glucose. Stimulation of glucose transport that results in increased cellular concentrations of glucose-6-P is a mechanism by which insulin and other agents increase glycogen synthase I. Activation of glycogen synthase by this glucose transport-coupled and phosphorylation-dependent mechanism may involve an increase in glycogen synthase phosphatase activity. A model is presented relating this mechanism


Incubation
of fat cells with insulin, concanavalin A, or H,Oz stimulated the incorporation of o-[U-14Clglucose into glycogen. Incubation with either concanavalin A or insulin alone activated glycogen synthase. H,O, alone did not activate glycogen synthase, but in the presence of glucose it increased the percentage of glycogen synthase Z activity. The activation of glycogen synthase by H,02 and glucose was abolished by phloridzin or cytochalasin B. These inhibitors of glucose transport also abolished the action of glucose to potentiate the effect of concanavalin A on glycogen synthase.
Glucose, mannose, and Z-deoxyglucose increased glycogen synthase Z activity and potentiated the effect of insulin on glycogen synthase. Incubation of cells with 3-O-methylglucose, galactose, or fructose was without effect on glycogen synthase Z activity and did rot result in a potentiation of the effect of insulin.
Compared to other hexoses, 2deoxyglucose was much more effective in activating glycogen synthase. Increases in 2-deoxyglucose-6-P produced during 5-min incubations with varying concentrations of 2deoxyglucose paralleled the increases in glycogen synthase Z activity, although phosphorylase a activity and concentrations of ATP and CAMP were unchanged.
Glucose-6-P and 2-deoxyglucose-6-P (but not glucose, mannose, or 2-deoxyglucose) increased the activity of glycogen synthase phosphatase which was monitored by following the increase in endogenous glycogen synthase Z activity.
These results suggest that glucose-6-phosphate formation is required for the activation of glycogen synthase by glucose. Stimulation of glucose transport that results in increased cellular concentrations of glucose-6-P is a mechanism by which insulin and other agents increase glycogen synthase I. Activation of glycogen synthase by this glucose transport-coupled and phosphorylation-dependent mechanism may involve an increase in glycogen synthase phosphatase activity. A model is presented relating this mechanism and 0.5-l x 10" cells/ml were used. Labeled CO, production from D-[U-14Clglucose was measured as described previously (1). Labeled glucose incorporation into glycogen was measured after the reaction was stopped by adding 0.2 ml of 10 N KOH to 0.21 ml of cells containing 1 mM D-[U-'Qglucose (approximately lo6 cpm) and additions as indicated in Fig. 1. The polypro-10 g of adipose tissue were homogenized at 0" with a glass homogenizer (10 strokes with a Teflon pestle driven at 1000 rpml in 5 to 10 ml of buffer (10 mM EDTA and 50 mM Tris, pH 7.1). After centrifugation of the homogenate at 6000 x g for 20 min, the supernatant was collected and placed on ice.  Thomas et al. (91. The papers were dried and after radioassay, the amount of glucose incorporated into glycogen was calculated for the total cell samples after subtracting the radioactivity obtained from tubes to which KOH was added before the cells. This method provided a high degree of reproducibility and produced results equivalent to those obtained by the method we have previously used (1). Incubations of cells for enzyme assays were performed at 37" in 5 ml of medium as previously described (1). Glycogen synthase activity was assayed essentially as described by Thomas et al. (9). Glycogen synthase Z activity is expressed as a percentage of the total synthase activity (assayed in the presence of 7.2 mM glucose-6-P). Phosphorylase activity was assayed essentially by the method of Gilboe et al. (101. Phosphorylase a activity is expressed as a percentage of the total phosphorylase activity (measured in the presence of 2.0 mM AMP). The experimental conditions for assay of glycogen synthase and phosphorylase were as previously described (1).

Lectins
(2, 3) and H,O, (4, 5, 15) have been shown to share with insulin the ability to stimulate glucose oxidation and glucose transport system activity in adipocytes. We therefore investigated the effects of HzOz and concanavalin A on glycogen synthesis in fat cells to determine whether these agents also mimic the effect of insulin in stimulating the incorporation of glucose into glycogen. As shown in Fig. 1 For measurements of ATP, CAMP, and glucose-6-P, cells were incubated in 1 ml of medium plus additions at 37" for the appropriate times. For measurement of 2-deoxyglucose-6-P, cells were incubated with the appropriate concentration of 2-deoxy-n-[l-*4Clglucose (lo6 cpm/tube).
The incubations were terminated by adding 1 ml of 100 mg/ml of trichloroacetic acid and homogenizing at 0" for 15 s with a Polytron homogenizer. After centrifugation at 10,000 x g for 10 min, the supematants were removed and extracted four times with 6 ml of water-saturated ether. Glucose-6-P and ATP were assayed in the samples by the method of Williamson and Corkey (11). The concentration of CAMP in the samples was determined by radioimmunoassay using the procedures described by Harper and Brooker (12). To measure 2-deoxyglucose-6-P, samples (600 ~1) were spotted on Whatman No. 3MM paper and the amount of 2-deoxy-n-[1-14Clglucose-6-P was determined after separation from 2-deoxy-n-[l-'4C1glucose and 6-phospho-2-deoxy-n-[l-*4C]gluconate by ascending chromatography as described by Benner et al. (13). In the presence of glucose, incubation of cells at 37" for 10 min with H202 produced an increase in glycogen synthase Z activity (Fig. 2). Activation of glycogen synthase by 0.4 mM H202 was observed when the medium glucose concentration was above 1 mM. Incubation of cells with HzOz alone did not increase glycogen synthase Z activity. We have tested concentrations of H,Op between 0.05 and 8 mM for periods of incubation up to 30 min without observing activation of glycogen synthase.l Concentrations of H,O, above 4 mM actually inactivated glycogen synthase. For example, incubation of cells with 4 mM H,O, for 10 min at 37" decreased the percentage of glycogen synthasez activity from 14.8 f 1.0 to 8.7 f 1.6 (mean values -+ S.E. from seven experiments).
Unlike Hz02, concanavalin A produced an increase in glycogen synthase Z activity in the absence of glucose. The time course of glycogen synthase activation by 40 pg/ml of concanavalin A in both the presence and absence of 5 mM glucose is presented in Fig. 3. The effects of concanavalin A were maximal after 10 min of incubation and represented an increase in the percentage of glycogen synthase Z activity of For assay of glycogen synthase phosphatase activity using endogenous glycogen synthase as substrate, adipocytes prepared from 5 to ' J. C. Lawrence, and J. Larner, unpublished observations. about 1.5 times the control. The activation of glycogen synthase by concanavalin A was abolished by 40 mM a-methylmannoside.' The effect of concanavalin A was greater in the presence of 5 mM glucose at all the incubation periods tested. As previously found with insulin plus glucose (l), the effect of concanavalin A plus glucose on glycogen synthase was maximal after 10 min, but was less after longer incubation periods. Potentiation of the effect of concanavalin A by glucose was dependent upon the concentration of the hexose. As shown in Fig. 4, at concentrations of glucose above 1 mM, the effect of 40 pg/ml of concanavalin A plus glucose on glycogen synthase was greater than the additive effects of the lectin and hexose alone.
The maximal effect of concanavalin A alone was observed at a concentration of 40 yglml, and was less than the increase in glycogen synthase Z activity produced by 2.5 milliunits/ml of insulin (Fig. 5). In the absence of glucose, concanavalin A opposed the effect of insulin on glycogen synthase. In the presence of 5 mM glucose, no effect of concanavalin A over the increase in glycogen synthase Z activity produced by insulin was observed. This finding indicates that activation of glycogen synthase by concanavalin A and insulin in the presence of glucose may occur through a common pathway.
It was previously proposed that activation of glycogen synthase by insulin in the presence of glucose occurred in part as a result of stimulation of the glucose transport system and increased glucose entry into the cell (1). Because both concanavalin A and H,O, stimulate glucose transport in adipocytes, this glucose transport-coupled mechanism should also be operative with these agents in the presence of glucose. If  glucose-6-P present in the extracts was at sufficient concentrations to produce allosteric activation of glycogen synthase D that could account for this large increase in activity.' Similar controls were previously performed showing that concentrations of glucose-6-P in homogenates following incubation with glucose or insulin plus glucose are too low to produce allosteric activation of glycogen synthase D (1).
The results presented in Fig. 8 represent the percentages of glycogen synthase I activity found in fat cells following two IO-min incubation periods. In the first period, adipocytes were incubated without (A and B) or with (C and D) 0.4 mM 2deoxyglucose. This incubation was terminated by washing the cells two times and suspending them in medium not containing 2-deoxyglucose. In the second incubation period, 0.4 mM 2deoxyglucose was added to cells not previously exposed to the hexose (B) and to cells that were incubated with 2-deoxyglucase in the first period (D). Insulin (2.5 milliunits/ml) was added to cells at the start of the second incubation period. In the absence of insulin, the same percentage of glycogen synthase I activity was observed in cells that were incubated with 2-deoxyglucose only in the second incubation (B) as in cells that were incubated with 2-deoxyglucose and washed free of the hexose (C). The absolute change in glycogen synthase I activity due to insulin in cells washed free of 2-deoxyglucose (C) was equal to the increase produced by the hormone in cells not incubated with 2-deoxyglucose (A). By comparing B and D to C, it can be seen that only when 2-deoxyglucose was present in the medium with insulin were the effects of insulin and 2-deoxyglucose greater than additive. This requirement of extracellular 2-deoxyglucose is consistent with the hypothesis that the potentiation of the effect of insulin by the hexose results from a stimulation by 2-deoxyglucose uptake.
The major proportion of 2-deoxyglucose that enters adipocytes is found as 2-deoxyglucose-6-P (16). This compound accumulates as a comparatively stable metabolite in cells incubated with 2-deoxyglucose, because further metabolism occurs only to a limited extent and the phosphate group renders it relatively impermeable to the plasma membrane. These properties could explain why the effects of 2-deoxyglucase on glycogen synthase persist after removal of extracellular 2-deoxyglucose ( Fig. 8), while the effects of glucose or insulin plus glucose are rapidly reversed by washing.' The remaining experiments in this report were designed to investigate further the effects of 2-deoxyglucose and the hypothesis that phosphorylation of glucose to glucose-6-P is the link to glycogen synthase activation in the glucose transport-coupled pathway of synthase activation. by inhibitors of glucose transport. As shown in Fig. 6, incubation of cells with 10 pg/ml of cytochalasin B or 4 mM phloridzin abolished the effects of 5 mM glucose in both the absence and presence of 0.4 mM H,O, or 40 pg/ml of concanavalin A.2 Therefore, it seems likely that agents such as H202, concanavalin A, and insulin can activate glycogen synthase in the presence of glucose by stimulating glucose entry into the cell. To determine whether other hexoses could substitute for glucose, cells were incubated at 37" for 10 min with different hexoses at a concentration of 5 mM in the presence and absence of 100 microunits/ml of insulin before glycogen synthase Z activity was assayed (Fig. 7). As previously reported (l), incubation of cells with insulin alone increased the percentage of glycogen synthase I activity and glucose potentiated the effect of insulin. Mannose alone increased glycogen synthase I activity to the same extent as incubation with glucose alone. In contrast to glucose and mannose, the hexoses fructose, galactose, and 3-O-methylglucose did not increase glycogen synthase Z activity or potentiate the effect of insulin. In fact, incubation of cells with galactose and fructose partially blocked the effect of insulin on glycogen synthase.
Incubation of adipocytes for 10 min at 3'7" with 0.5 mM or 5 mM 2-deoxyglucose alone did not decrease the concentration of CAMP under conditions in which the percentage of glycogen synthase I activity was increased (Table I). These concentrations of 2-deoxyglucose were also without effect on the concentration of cyclic nucleotide observed in the presence of 100 microunitslml of insulin. As also shown in Fig. 7, incubation of cells with 2-deoxyglucase resulted in a marked increase (about five times control) in glycogen synthase Z activity. Experiments were performed (involving lo-fold dilutions of homogenates and Sephadex G-50 chromatography) that negated the possibility that 2-deoxy-2 A preparation of collagenase, different from that used in other experiments presented in this report, was used to isolate the cells in these experiments.
We feel this is the reason for the failure of concanavalin A alone to activate glycogen synthase in the experiments represented by Fig. 6.
In the experiments presented in Fig. 9, cells were incubated with 1 mM 2-deoxyglucose for periods ranging from 0 to 60 min, before the percentages of glycogen synthase Z and phosphorylase a activities and the concentrations of 2-deoxyglucase-6-P, CAMP,' and ATP were determined. It was of interest to monitor phosphorylase activity under these conditions, because it has been suggested that decreases in phosphorylase a in liver due to glucose can lead to glycogen synthase activation (1'7, 18). With 1 mM 2-deoxyglucose, the percentage of glycogen synthase Z activity was increased from about 10% The Percentage of phosphorylase a activity was slightly decreased over the 60-min incubation period. However, it should be noted that no detectable decreases in phosphorylase a activity were observed until after 12 min, at which time the percentage of glycogen synthase Z was increased about 3-fold. Thus phosphorylase inactivation did not precede synthase activation. During the time course of incubation with 1 mM 2-deoxyglucose, the concentration of 2deoxyglucose-6-P rose progressively, while the concentration of ATP remained essentially unchanged. The concentration of CAMP was also unchanged. ' The increase in glycogen synthase Z activity was linear   (19), in which large decreases (50%) deoxyglucose on phosphorylase a activity and on concentra-in the adenine nucleotide were seen with 2-deoxyglucose.
A tions of ATP and 2deoxyglucose-6-P were investigated after 5 60-min incubation period with higher concentrations of 2min, so that these parameters could be assayed under condi-deoxyglucose (10 to 20 mM) was used by these investigators tions in which initial rates of glycogen synthase activation when ATP measurements were made. Differences in experiwere observed (Fig. 10). Glycogen synthase activation was mental conditions may explain the apparently discrepant half-maximal at 0.8 mu 2-deoxyglucose and essentially maxi-results. For example, we found that when cells were incubated mal at 4 mM 2-deoxyglucose. The accumulation of 2-deoxyglu-at 37" for 10 min with 5 mM 2-deoxyglucose, the concentration case-6-P closely followed the activation of glycogen synthase.
of ATP was decreased from a control value of 15.8 + 0.6 to 12.8 The concentration of 2-deoxyglucose-6-P found at half-maxi-f 0.7 nmol/lo6 cells (mean values + S.E. from three experimal activation of glycogen synthase was about 8.5 nmol/lo6 ments).
Under these conditions, incubation with 100 microcells. In these experiments, the percentages of phosphorylase units/ml of insulin plus 5 mM 2deoxyglucose decreased cellua activity and concentrations of ATP remained essentially lar ATP from 15.7 f 0.7 (obtained with insulin alone) to 7.0 + unchanged. The failure of 2-deoxyglucose to decrease ATP 0.4 nmol/lob cells. Fat cells were incubated at 37" for 10 min with or without 5 mM glucose and 100 microunits/ml of insulin. The incubations were terminated and glycogen synthase 2 activity and the concentrations of ATP and glucose-6-P were assayed as described under "Materials and Methods."  11. Effect of glucose-6-P on the activity of glycogen synthase phosphatase using endogenous adipocyte glycogen synthase as substrate. Extracts from isolated fat cells were incubated at 30" with the indicated concentrations of glucose-6-P for 5 min. Details of this incubation are described under "Materials and Methods." Incubation of cells with 5 mM glucose or 100 microunitslml of insulin plus glucose did not alter the concentration of ATP (Table II). In the presence of glucose, insulin more than doubled the concentration of glucose-6-P.

Additions
Evidence from in vitro studies has accumulated indicating that glucose-6-P can increase the activity of glycogen synthase phosphatase from various tissue sources (20-23). However, such an effect of glucose-6-P has, to our knowledge, never been demonstrated using enzymes from fat cells. Because of problems associated with purification of glycogen synthase and phosphatase from rat adipose tissue, we investigated the effects of glucose-6-P on the activity of glycogen synthase phosphatase by monitoring the increase in endogenous glycogen synthase Z activity in extracts of isolated fat cells. Use of endogenous glycogen synthase could circumvent problems that arise from using purified glycogen synthase D from other tissues. The results of experiments in which extracts were incubated for 5 min at 30" with increasing concentrations of glucose-6-P (0.33 to 10 mM) are presented in Fig. 11. The incubations were terminated by adding 100 mM KF before a two-step procedure (described in detail under "Materials and Methods") involving ammonium sulfate precipitation and liltration through Sephadex G-25 was used to remove glucose-g P. Without glucose-6-P, incubation at 30" resulted in an increase in the percentage of glycogen synthase Z activity from about 10 to 38%. In the presence of glucose-6-P the percentage of glycogen synthase Z activity was increased (from 38 to 65% with 10 mM glucose-6-P). Addition of 100 mM KF before the incubation with glucose-6-P greatly decreased phosphatase activity. As a control, glucose-6-P was added after 5 min of incubation at 30". Because no further increase in glycogen synthase Z activity was observed, it is unlikely that glucose-g-P was carried through the procedures designed for its removal. As shown in Fig. 12, 2-deoxyglucose-6-P increased glycogen synthase phosphatase activity, although to a lesser degree than glucose-6-P (Fig. 11). We have also found that mannose-6-P increased glycogen synthase phosphatase activity, although this compound was also less effective than glucose-& P.' Because glucose, mannose, and 2-deoxyglucose are thought to enter adipocytes via a facilitated diffusion system, their intracellular concentrations should never be higher than their extracellular concentrations. Therefore, in the experiments presented in Fig. 7, 5 mM would represent an upper limit for the concentration of the free sugars within the cell. However, 5 mM glucose, 5 mM mannose, and 5 mM 2-deoxyglucose did not increase glycogen synthase phosphatase activity.'

DISCUSSION
The ability of H,O, to activate glycogen synthase only in the presence of glucose is in contrast to that of insulin, since the hormone produces an increase in glycogen synthase Z activity in the absence of medium glucose. Czech et al. (5,24) proposed that the stimulation of glucose transport by insulin and H,O, might involve a common mechanism, namely, the oxidation of key cellular sulfhydryl groups to the disulfide forms. However, the failure of H,O, alone to increase glycogen synthase Z activity supports the hypothesis that the activation of glycogen synthase by insulin in the absence of glucose does not involve oxidative events that may occur during glucose transport system activation.
Unlike H202, concanavalin A increased the percentage of adipocyte glycogen synthase in the absence of glucose (Figs. 3 to 5). One interpretation of these results is that the site of concanavalin A action is closer to the site of action of insulin than that of H,O,. Cuatrecasas' studies suggested that concanavalin A bound to a region of the insulin receptor thus preventing the binding of insulin (25). Perhaps our finding that concanavalin A reduced the activation of glycogen synthase observed with insulin relates to an effect of the lectin at the level of the insulin receptor. Binding of concanavalin A to a region of the receptor near the insulin binding site might be sufficient to promote a partial activation of glycogen synthase while preventing the "insulin-insulin receptor" interaction responsible for mediating the effects of the hormone on glycogen synthase. Evidence that concanavalin A interferes with normal insulin binding was also obtained by DeMeyts et al. who reported that concanavalin A abolished negative cooperativity between insulin receptors (26).
Both concanavalin A (Figs. 3 to 6) and H,O, (Figs. 2 and 6) increased the percentage of glycogen synthase Z activity in the presence of glucose. Because both agents are known to activate glucose transport in isolated fat cells (2%5), these findings are consistent with the hypothesis that stimulation of glucose entry into the adipocyte is a mechanism for glycogen synthase activation (1). Potentiation of the effects of these agents by glucose was blocked by inhibiting glucose transport with phloridzin or cytochalasin B (Fig. 6), which adds further evidence favoring this hypothesis.
The results presented in this report provide evidence that phosphorylation of glucose to glucose-6-P is involved in this hexose transport-coupled pathway for the activation of glycogen synthase. For example, incubation of cells with 30 methylglucose and galactose did not result in activation of glycogen synthase. The uptake of 34-methylglucose occurs in fat cells via the n-glucose transport system and its transport is stimulated in adipocytes by insulin, although the hexose is not phosphorylated (5,(27)(28)(29). Galactose transport is not well characterized in fat cells; however, in red blood cells gala&se has a high affinity for the n-glucose transport system (30). Galactose uptake is stimulated by insulin in skeletal muscle where the hexose is metabolized very slowly (31). Gala&se is also slowly metabolized in fat cells (32). The results in the present report obtained with 34-methylglucose and galactose are consistent with the idea that a certain level of cellular metabolism is necessary for glycogen synthase activation by hexoses. Mannose, a hexose that is metabolized in adipose tissue much like glucose (331, produced essentially the same effects as glucose. No activation was observed with fructose. The reason glucose and mannose, but not fructose, activate glycogen synthase and potentiate the effect of insulin on increasing glycogen synthase Z activity may relate to differences in the transport and metabolism of fructose (34). More investigation is needed to clarify this point. The strongest indication that hexose phosphorylation leads to glycogen synthase activation came from results with 2deoxyglucose. Transport of X-deoxyglucose occurs by the Dglucose transport system (35). The hexose is phosphorylated and 2-deoxyglucose-6-P accumulates in the cell because further metabolism is very limited. Czech (16) has shown that insulin can triple the concentrations of 2-deoxyglucose-6-P.
It is possible that these properties explain why 2-deoxyglucose was so effective in activating glycogen synthase and potentiating the effect of insulin to increase glycogen synthase Z activity (Figs. 7 to 10). Gilboe and Nuttal (36) reported that physiological concentrations of ATP inhibited rat liver glycogen synthase phosphatase. ATP is also required for the conversion of glycogen synthase Z to D in the protein kinase-catalyzed reaction (6). Therefore, a decrease in cellular ATP could result in glycogen synthase activation. This was considered as a possible mechanism for the increase in glycogen synthase Z activity resulting from incubation of cells with 2-deoxyglucose, since, under certain conditions, the concentrations of ATP were decreased by incubating cells with this hexose. However, this mechanism seems unlikely because glycogen synthase Z activity was increased over four times by 2-deoxyglucose under conditions in which the concentration of ATP was not detectably changed (Figs. 9 and 10). For the same reason, it is unlikely that a decrease in ATP mediates the effect of glucose to activate glycogen synthase or to potentiate the effect of insulin to increase glycogen synthase Z activity (Table II).
A decrease in CAMP could also bring about an increase in the percentage of glycogen synthase Z activity by lowering the activity of CAMP-dependent protein kinase. However, it seems unlikely that this mechanism is involved in the activation of glycogen synthase by 2-deoxyglucose, since CAMP concentrations were not detectably decreased by the hexose under conditions in which 2-deoxyglucose increased glycogen synthase Z activity (Table I). Stalmans et al. proposed that inactivation of phosphorylase is involved in the activation of hepatic glycogen synthase by glucose (17,18). They further suggested that only when levels of phosphorylase a are decreased to a certain threshold does glycogen synthase activation occur. By their model, phosphorylase inactivation should always precede glycogen synthase activation. There is no evidence that such a scheme fits the activation of glycogen synthase by 2-deoxyglucose in adipocytes, because phosphorylase inactivation clearly did not precede glycogen synthase activation (Fig. 9). Hizukuri and Takeda (20) first demonstrated an increase in glycogen synthase phosphatase activity with glucose-6-P using a partially purified enzyme preparation from bovine spleen. Kato and Bishop (21) later demonstrated a similar effect using purified enzymes from rabbit skeletal muscle. Increases in glycogen synthase phosphatase activity in the presence of glucose-6-P have also been demonstrated by Nakai and Thomas (22) using a purified heart enzyme and by Killilea et al. (23) using a highly purified rabbit liver phosphatase. These studies have not completely eliminated the possibility that glucose-6-P directly stimulates glycogen synthase phosphatase. However, as Kato and Bishop (21) suggested, the stimulation may arise from an effect of glucose-6-P to alter the conformation of glycogen synthase D, rendering it a better substrate for the phosphatase.
Regardless of the mechanism by which glucose-6-P increases glycogen synthase phosphatase activity, it seems possible that such a mechanism might be operative in adipocytes, since glucose-6-P increased the activity of fat cell glycogen synthase phosphatase (Fig. 11). Calculations suggest that intracellular concentrations of glucose-6-P following incubation of cells with glucose are in the range of concentrations that increase phosphatase activity. If the intracellular water space is assumed to be 1.5 pi/cell (3'7), then the intracellular coilcentration of glucose-6-P in cells incubated with 5 mM glucose for 10 min can be estimated at about 0.7 mM and 1.6 mM in cells incubated with insulin plus glucose (calculated from data in Table II). If the concentrations of intracellular 2-deoxyglucose-6-P resulting from incubating cells with 2-deoxyglucose are estimated using the same assumptions (from Figs. 9 and lo), these also are in the range of 2-deoxyglucose-6-P concentrations that increased phosphatase activity (Fig. 12). Estimates of the intracellular concentrations of metabolites are only approximations since the metabolites probably do not randomly distribute within the cytosol and may not be free in solution, but bound to some extent to cellular constituents . It should also be stressed that conditions of the phosphatase assay are vastly different from those present in the cellular milieu. However, glucose, mannose, and 2-deoxyglucose did not increase phosphatase activity.' These results are consistent with the proposal that hexose phosphorylation is involved in mediating the effects of glucose to activate glycogen synthase.
In summary, the results presented in this report indicate that activation of glycogen synthase by glucose occurs subsequent to the transport and phosphorylation of the hexose. Fig.  13 depicts a model for glycogen synthase activation by insulin and glucose. The interaction of insulin with its receptor triggers cellular events leading to activation of glycogen synthase by a direct, hexose transport-independent mechanism. Activation by this pathway is observed in the absence of glucose, or with glucose under conditions in which glucose transport is inhibited, as in the presence of phloridzin or cytochalasin B. This direct pathway is also activated by concanavalin A in the absence of glucose (Fig. 3). Evidence obtained from experiments with skeletal muscle (38)(39)(40)(41), liver (42), and adipose tissue (43,44) indicates that the mechanism of this direct pathway of glycogen synthase activation by insulin may involve a decrease in CAMP-dependent protein kinase activity. The insulin-insulin receptor interaction also results in the activation of the hexose transport system. In the presence of glucose, mannose, or 2-deoxyglucose, this leads to increased sugar entry and phosphorylation that results in activation of glycogen synthase by a hexose transport-coupled and phosphorylation-dependent pathway. In the presence of glucose, activation by this pathway can be effected by concanavalin A and H,02, by virtue of their effects to stimulate glucose transport. This mechanism may involve accumulation of hexose phosphate and an increase in glycogen synthase phosphatase activity.