Regulation of Glycogen Synthesis by Amino Acids in Cultured Human Muscle Cells*

Insulin and a number of metabolic factors stimulate glycogen synthesis and the enzyme glycogen synthase. Using human muscle cells we find that glycogen synthesis is stimulated by treatment of the cells with lithium ions, which inhibit glycogen synthase kinase 3. Insulin further stimulates glycogen synthesis in the presence of lithium ions, an effect abolished by wortmannin and rapamycin. We report also that amino acids stimulate glycogen synthesis and glycogen synthase, these effects also being blocked by rapamycin and wortmannin. Amino acids stimulate p70 s6k and transiently inhibit glycogen synthase kinase 3 without effects on the activity of protein kinase B or the mitogen-activated protein kinase pathway. Thus, the work reported here demonstrates that amino acid availability can regulate glycogen synthesis. Furthermore, it demonstrates that glycogen synthase kinase 3 can be inactivated within cells independent of activation of protein kinase B and p90 rsk .

Insulin and a number of metabolic factors stimulate glycogen synthesis and the enzyme glycogen synthase. Using human muscle cells we find that glycogen synthesis is stimulated by treatment of the cells with lithium ions, which inhibit glycogen synthase kinase 3. Insulin further stimulates glycogen synthesis in the presence of lithium ions, an effect abolished by wortmannin and rapamycin. We report also that amino acids stimulate glycogen synthesis and glycogen synthase, these effects also being blocked by rapamycin and wortmannin. Amino acids stimulate p70 s6k and transiently inhibit glycogen synthase kinase 3 without effects on the activity of protein kinase B or the mitogen-activated protein kinase pathway. Thus, the work reported here demonstrates that amino acid availability can regulate glycogen synthesis. Furthermore, it demonstrates that glycogen synthase kinase 3 can be inactivated within cells independent of activation of protein kinase B and p90 rsk .
One of the key actions of insulin in the lowering of blood glucose is the stimulation of glycogen synthesis in skeletal muscle, involving activation of the enzyme glycogen synthase (GS) 1 (1). Phosphorylation decreases GS activity, and the extent of this inactivation depends on the sites phosphorylated and on the concentrations of allosteric effector molecules such as glucose-6-phosphate. Activation of GS by insulin results from dephosphorylation principally at a cluster of 3 serine residues collectively termed "site 3" located at the carboxyl terminus of GS (2). Further evidence indicates that the phosphorylation state of site 2 (or 2a or both) is also decreased in response to insulin (3). Phosphorylation at site 3 is catalyzed principally by glycogen synthase kinase 3 (GSK-3), and dephosphorylation is catalyzed by the glycogen-bound form of protein phosphatase 1 (PP1; Refs. 1,4). Both PP1 and PP2A can dephosphorylate sites 2 and 3 in vitro (2), and there is evidence for a mechanism of the glycogen-bound form of PP1 activation by insulin that is mediated by p90 rsk (5,6). This provides an attractive theory for insulin exerting this stimulatory effect on GS by activating the glycogen-bound form of PP1. Recent evidence, however, has questioned this theory. Of particular significance is the finding that PD98059, which blocks activation of p90 rsk (7), has no effect on insulin stimulation of glycogen synthesis and GS in various cell types (8 -10).
Attention has since focused on the effects of insulin on GSK-3. GSK-3 is inactivated in response to insulin or growth factors by phosphorylation at a single serine residue close to its amino terminus (11,12). Three insulin-stimulated kinases, namely p70 s6k , p90 rsk , and protein kinase B (PKB), have been shown to phosphorylate and inactivate GSK-3 in vitro (12,13). However, agents that prevent the activation of both p70 s6k and p90 rsk by insulin do not block the inactivation of GSK-3, strongly suggesting that the kinase acting under these conditions is PKB (14).
Despite accumulating evidence for a key role of PKB and GSK-3 in controlling GS, there is evidence for one or more alternative pathways contributing to the effect of insulin. First, GSK-3 does not phosphorylate GS at site 2, which is also dephosphorylated in response to insulin (3). Second, use of the immunosuppressive drug rapamycin, which prevents the activation of p70 s6k (15,16), has provided evidence for contribution of a rapamycin-sensitive pathway to the activation of glycogen synthesis and GS by insulin in skeletal muscle, adipocytes, and hepatoma cells (10,(17)(18)(19). In human muscle cells in culture, the activation of glycogen synthesis and GS by both insulin and epidermal growth factor is partly blocked by preincubation of the cells with rapamycin, which, however, is without effect on the inactivation of GSK-3 by these hormones (10,20). This suggests that two pathways are contributing to the activation of glycogen synthesis in this system, the phosphoinositide-dependent kinase 1-PKB-GSK-3 pathway and a rapamycin-sensitive pathway, which can act downstream of, or independent of, GSK-3.
To investigate this further, one approach is to activate the two pathways independently. Recent work has established that activation of p70 s6k can be regulated by the availability of amino acids. Amino acid starvation causes inactivation of p70 s6k , which can be reversed by the readdition of amino acids, a process that is rapamycin sensitive and that does not involve activation of PKB or p90 rsk (21)(22)(23). Using this approach, we report here that the rate of glycogen synthesis and the activity of GS are increased in response to amino acids, indicating that the p70 s6k pathway can control these processes independent of other signaling pathways.

EXPERIMENTAL PROCEDURES
Materials-All tissue culture trays were from Costar (Cambridge, MA). Culture media, penicillin-streptomycin, and trypsin-EDTA were from Life Technologies (Paisley, UK). Chick embryo extract was obtained from ICN (Costa Mesa, CA). D-[U-14 C]Glucose (300 mCi/mmol) * This work was supported by the British Diabetic Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Holds a studentship from the Biotechnology and Biological Sciences Research Council, United Kingdom.
Cell Culture-Human myoblasts were grown from needle biopsy samples taken from the gastrocnemius muscle of healthy subjects with no family history of type 2 diabetes and with normal glucose tolerance and normal insulin sensitivity as assessed using the short insulin tolerance test. Myoblasts were maintained in growth medium consisting of Ham's F-10 nutrient mixture containing 20% fetal calf serum, 1% chick embryo extract, 100 units/ml penicillin, and 100 g/ml streptomycin. All experiments were performed using cells between the 5th and 15th passage at Ͼ90% confluence. Before hormone treatment, cells were incubated for at least 2 h in serum-free Ham's F-10 medium or serum-free minimum Eagle's medium ( Estimation of Glycogen Synthesis-Glycogen synthesis was determined as [ 14 C]glucose incorporation into glycogen over 1 h, as described previously (20). Results are expressed as picomoles of glucose incorporated into glycogen per minute per milligram of protein.
Preparation of Myoblast Extracts-After incubation with appropriate hormones, growth factors, or other agents, cells were washed three times with ice-cold phosphate-buffered saline, and excess liquid was removed. Cells were then scraped into extraction buffer (100 mM Tris-HCl, pH 7.4, 100 mM KCl, 2 mM EDTA, 25 mM KF, 0.1% (v/v) Triton X-100, 1 mM benzamidine, 0.1 mM Na 3 VO 4, 1 g/ml pepstatin, 1 g/ml antipain, and 1 g/ml leupeptin), transferred to 1.5-ml Eppendorf tubes, and immediately frozen in liquid nitrogen. Before analysis, samples were thawed and dispersed by sonication for 1 min (Sonibath; Dawe). The protein concentration was determined by a dye binding method (26).
Determination of GSK-3 and PKB Activities-GSK-3 activity was determined as described previously (27) with modifications. Samples were assayed in a final volume of 50 l containing 10 l of total cell extract (ϳ10 g of protein) and 40 l of assay mixture (5 M protein kinase A inhibitor, 0.1 mM EGTA, 0.1 mM [␥-32 P]ATP (ϳ500 cpm/pmol), 10 mM MgAc, 50 mM Tris-HCl, pH 7.5, 0.1% 2-mercaptoethanol or 0.2 mg/ml GSK phospho-eIF2B peptide, 50 mM ␤-glycerophosphate, pH 7.5, in the presence or absence of 50 mM LiCl). After a 15-min incubation at 30°C, 40 l were spotted onto Whatman P81 phosphocellulose paper squares. After washing in 175 mM phosphoric acid with four changes, the papers were dried, and phosphate incorporation was determined by liquid scintillation counting. Enzyme activity was defined as that which catalyzes the incorporation of 1 nmol of phosphate into peptide substrate in 1 min. GSK-3 activity is that in the absence of Li ϩ minus that in its presence. PKB␣ was assayed after immunoprecipitation as described previously (28).

RESULTS
In initial experiments, the effect of the inhibitors wortmannin, rapamycin, and lithium on insulin-stimulated glycogen synthesis was investigated in cultured human myoblasts (Fig.  1). Insulin stimulated glycogen synthesis from extracellular glucose by ϳ3-fold. This activation was completely blocked by wortmannin, a selective inhibitor of phosphatidylinositol 3-kinase, whereas only 40% inhibition was observed with rapamycin, which prevents the activation of p70 s6k . Rapamycin alone had little effect on basal glycogen synthesis, decreasing it by ϳ10% (data not shown). Wortmannin alone, however, induced a decrease of ϳ70%, implying that a wortmannin-sensitive mechanism is involved in basal glycogen synthesis. This is consistent with previous work from this laboratory (10,20).
Work from several laboratories has indicated that the inactivation of GSK-3 is primarily responsible for dephosphorylation and activation of GS in response to insulin (10 -12, 29, 30). Lithium is an allosteric inhibitor of GSK-3, mimicking the effect of insulin without affecting the activity of the mitogenactivated protein (MAP) kinase pathway (31). Treatment of cells with lithium chloride for 1 h caused a 3-fold increase in basal glycogen synthesis (Fig. 1), whereas no significant difference was observed in cells treated with an equal concentration of sodium chloride to correct for osmolarity changes (data not shown). This increase was not reduced significantly by wortmannin or by rapamycin (data not shown). In the presence of lithium, insulin stimulated glycogen synthesis a further 1.5fold, an effect completely blocked by rapamycin (Fig. 1) or wortmannin (data not shown). Taken together, these data provide further support for a key role of GSK-3 and the presence of a rapamycin-sensitive, insulin-stimulated pathway in the control of glycogen synthesis.
To investigate the role of this rapamycin-sensitive pathway further, amino acids were used to activate p70 s6k selectively (21-23), and the activity of GS and glycogen synthesis was measured (Fig. 2). Myoblasts starved of amino acids for 5 h were then incubated in MEM S containing 2ϫ concentrations of amino acids for various times. GS activity was measured, and the values are shown as a ratio of the GS fractional activity in the presence over the absence of amino acids in the culture medium. GS activity increased 2.5-and 2.0-fold in response to 2ϫ concentrations of amino acids for 1 and 2 h, respectively. Glycogen synthesis increased ϳ1.5-fold in response to 2ϫ concentrations of amino acids for 2 h. Both of these responses were completely inhibited by wortmannin and rapamycin. PD98059 had no significant effect on the amino acid stimulation, inhibiting basal and stimulated values to similar extents (data not shown).
To evaluate a possible role for p70 s6k in the stimulation of glycogen synthesis, the phosphorylation of this protein was measured under identical conditions. Fig. 3A shows that after 5 h of amino acid deprivation, readdition of 1ϫ or 2ϫ concentrations of amino acids resulted in the phosphorylation of p70 s6k at Thr 389 , the phosphorylation of which is most closely related to the activation of p70 s6k in vivo (32). This phosphorylation was further increased in response to insulin. This phosphorylation in response to amino acids was completely blocked by rapamycin, indicating that the amino acid activation of p70 s6k involves the mammalian target of rapamycin (mTOR). A threshold level of amino acids is also required for the insulin-stimulated phosphorylation of p70 s6k , because the increase in phosphorylation in response to insulin was reduced by ϳ72% in cells incubated in the absence of amino acids compared with cells in the presence of amino acids (Fig. 3A), consistent with earlier findings (23,33). Phosphorylation of p70 s6k in response to amino acids is time dependent, with a significant increase in phosphorylation appearing after 10 min and persisting for at least 2 h (Fig. 3B). In all experiments endogenous levels of p70 s6k protein did not change, as determined by immunoblotting (data not shown).
To investigate whether amino acid deprivation and readdition were affecting GSK-3, its activity and phosphorylation status were established under identical conditions. Fig. 4 shows that GSK-3 was inactivated by ϳ30% in response to amino acids, an effect seen within 10 min of amino acid addition and which was reversed after 30 min. Following a recent publication showing the inactivation of GSK3 in response to exercise without corresponding phosphorylation (34), the phosphorylation status of GSK-3 was also measured. Corresponding Western blots show a direct correlation of inactivation with phosphorylation (Fig. 4). During this experiment there was no significant change in GSK-3 protein expression. Although these effects are consistent with p70 s6k causing inactivation of GSK-3 and activation of GS, it is important to exclude the involvement of PKB or the MAP kinase pathway. Importantly, there was no significant change in PKB activity under matching conditions (Fig. 5), and the phosphorylation status of MAP kinase did not alter after amino acid withdrawal and readdition (data not shown). Wortmannin completely blocked the amino acid-induced inactivation and phosphorylation (Fig. 6) of GSK-3, whereas PD98059 was without effect (data not shown). Rapamycin essentially blocks the amino acid-induced phosphorylation and inactivation of GSK-3 (Fig. 6), indicating that the mTOR-p70 s6k pathway is principally responsible for the effects on GSK-3.

DISCUSSION
Insulin stimulates glycogen synthesis by both increasing glucose uptake and activating GS. Stimulation of GS by insulin involves the net dephosphorylation of specific serine residues, termed sites 2 and 3 (35). Inactivation of GSK-3 has been shown to be principally responsible for dephosphorylation of site 3, but the factor regulating site 2 dephosphorylation re-

FIG. 2. Effect of inhibitors on the stimulation of glycogen synthase and glycogen synthesis by amino acids.
Myoblasts were incubated in the MEM S for 5 h before replacement of media with MEM S (ϪAA) or MEM S plus 2ϫ AA (2ϫAA) or after preincubation with 100 nM wortmannin (2ϫAA/W), or 100 nM rapamycin (2ϫAA/R). The effects of inhibitors on GS were determined after 1 h, and those on glycogen synthesis were determined after 2 h. A, cell extracts were prepared, and GS activity was determined as incorporation of [ 14 C]glucose from UDP-[U-14 C]glucose into glycogen in the presence of high and low glucose-6phosphate to give active and total activities. Fractional activity at time zero is ϳ0.015. B, radioactive [U-14 C]glucose was incorporated for the final hour of this incubation. Extracts were prepared, and the quantity of radioactive glucose incorporated into glycogen was calculated. Results are mean Ϯ S.E., n ϭ 8, from at least three subjects. Values are expressed as -fold activation over basal (ϪAA). Basal values were ϳ400 pmol ⅐ min Ϫ1 ⅐ mg Ϫ1 . Statistical significance compared with the value in the absence of amino acids is indicated by two stars (p Ͻ 0.01) .   FIG. 3. Effects of insulin and rapamycin on p70 s6k phosphorylation in cells cultured in the absence or presence of amino acids. Myoblasts were incubated in MEM S for 5 h before replacement of media with MEM S or MEM S plus 1ϫ AA or MEM S plus 2ϫ AA for 2 h in the absence or presence of 100 nM rapamycin (R) or insulin (I; A) or MEM S plus 2ϫ AA for the times indicated (B). Total cell extracts were standardized for protein concentration and were separated in a 10% SDS-polyacrylamide gel, transferred to a polyvinylidene difluoride membrane, and blotted with anti-phospho-p70 s6k (Thr 389 ) antibody. Similar results were obtained in four independent experiments from at least three subjects. mains unidentified. Glycogen synthesis can also be stimulated in an insulin-independent manner, for example, after exercise, a process associated with glycogen depletion (36). After glycogen depletion in cultured human muscle cells, readdition of glucose leads to the stimulation of GS by an as yet unknown mechanism but one that apparently does not involve phosphorylation and inactivation of GSK-3. 2 Therefore, the role of GSK-3 in controlling glycogen synthesis was further investigated using its allosteric inhibitor lithium ions. Incubation of human myoblasts with lithium resulted in a significant increase in glycogen synthesis in these cells, which is consistent with a major role for GSK-3 in controlling glycogen synthesis. However, the addition of insulin resulted in a further, rapamycin-sensitive increase in glycogen synthesis, indicating a contribution by an alternative insulin-and rapamycin-sensitive pathway. The obvious candidate is the p70 s6k signaling pathway, which fulfills both these criteria and for which we and others have obtained evidence previously (10,(17)(18)(19). To investigate this further we have used the findings that it is possible to activate p70 s6k in a rapamycin-sensitive manner using amino acids, without the parallel activation of PKB or MAP kinase that occurs with insulin stimulation (21)(22)(23). We show here that amino acid depletion and subsequent readdition also lead to the stimulation of GS and glycogen synthesis (Fig. 2, A  and B), effects that are completely blocked by both wortmannin and rapamycin. A key question is whether this activation involves inhibition of GSK-3. As shown in Fig. 4, GSK-3 is phosphorylated and inactivated in response to amino acids.
Three kinases, namely p70 s6k , p90 rsk , and PKB, have been shown to phosphorylate and inactivate GSK-3 in vitro (12,13). Importantly, the present work and previous data (21)(22)(23) show that amino acids failed to activate PKB or the MAP kinase pathway. However, p70 s6k is phosphorylated and presumably activated by amino acids (Fig. 3). Furthermore, GSK-3 inhibition is completely blocked by both wortmannin and rapamycin, implying that the p70 s6k pathway is primarily responsible for the inactivation of GSK-3 in response to amino acids. Very recent work on amino acid transport in rat L6 myotubes has shown that leucine inactivates GSK-3 in a rapamycin-sensitive manner without activation of either PKB or the MAP kinase pathway (37). FIG. 4. Effect of amino acids on GSK-3 activity. Myoblasts were incubated in MEM S for 5 h before replacement of media with MEM S plus 2ϫ AA for the times indicated. A, total cell extracts were standardized for protein concentration and were separated in a 10% SDS-polyacrylamide gel, transferred to a polyvinylidene difluoride membrane, and blotted with anti-GSK-3␣ or anti-phospho GSK-3␣ and -␤ (Ser 21 and Ser 9 ) antibodies. Similar results were obtained in four independent experiments from at least three subjects. B, total cell extracts were assayed for GSK-3 activity. Results are mean Ϯ S.E., n ϭ 5, from at least three subjects. Values are expressed as percent basal activity (time ϭ 0). Basal values were ϳ180 pmol ⅐ min Ϫ1 ⅐ mg Ϫ1 . Statistical significance compared with the basal value is indicated by two stars (p Ͻ 0.01).
FIG. 5. Effect of amino acids on PKB activity. Myoblasts were incubated in MEM S for 5 h before replacement of media with MEM S plus 2ϫ AA for the times indicated or treated with 100 nM insulin for 10 min (I). Total cell extracts were subject to immunoprecipitation with antibodies to PKB␣ and assayed for PKB activity. Results are mean Ϯ S.E., n ϭ 3, from two subjects. Values are expressed as percent basal activity (time ϭ 0), which was ϳ1 pmol ⅐ min Ϫ1 ⅐ mg Ϫ1 .
FIG. 6. Effect of selective inhibitors on the activity and phosphorylation of GSK-3 and p70 s6k by amino acids. Myoblasts were incubated in MEM S for 5 h before further incubation for 15 min in MEM S (ϪAA) or MEM S plus 2ϫ AA (2ϫAA). Alternatively, cells were treated with 100 nM wortmannin (2ϫAA/W) or 100 nM rapamycin (2ϫAA/R) for 15 min before the addition of MEM S plus 2ϫ AA containing these inhibitors for 15 min. Total cell extracts were standardized for protein concentration and were separated in a 10% SDS-polyacrylamide gel, transferred to a polyvinylidene difluoride membrane, and blotted with anti-phospho-p70 s6k (Thr 389 ) antibodies (A) or anti-phospho GSK-3␣ and -␤ (Ser 21 and Ser 9 ) antibodies (B). Similar results were obtained in four independent experiments from at least three subjects. Extracts were assayed for GSK-3 activity (C). Results are mean Ϯ S.E., n ϭ 10, from three subjects. Values are expressed as percent basal activity (time ϭ 0). Statistical significance compared with the value in the presence of 2ϫ AA is indicated by two stars (p Ͻ 0.01).
The transient nature of the inhibition of GSK-3 (maximal after 15 min) does not coincide with the prolonged activation of GS and p70 s6k , which persists up to 2 h ( Fig. 2A). It is possible that transient inactivation of GSK-3 triggers the activation of GS, which could then be maintained by a further rapamycinsensitive pathway, perhaps acting on GS without involvement of GSK-3. This is supported by the data in Fig. 1, which indicate that when GSK-3 is inhibited by lithium ions, insulin can further stimulate glycogen synthesis in a rapamycin-sensitive manner. This would be consistent also with earlier work on insulin effects in human muscle cells (10,20) and with work (38) that showed that rapamycin has no effect on the activation of GS in skeletal muscle up to 5 min after stimulation by insulin but has a significant inhibitory effect at longer time points.
The amino acid-stimulated activation of p70 s6k observed in the human muscle cell lines is wortmannin and rapamycin sensitive. Rapamycin acts by associating with a cellular protein, FKBP12, and this complex then binds to mTOR, thus inhibiting its kinase activity. This would therefore indicate that mTOR is required for the response to amino acids, either directly or through the regulation of a downstream element, such as p70 s6k or a protein phosphatase. Although mTOR has been shown to phosphorylate p70 s6k directly in vitro (39), recent evidence points to the existence of a phosphatase, which is activated by rapamycin, via the inhibition of mTOR (23,40). Phosphatidylinositol 3-kinase is also implicated because of its wortmannin sensitivity, but it has previously been reported that wortmannin can selectively inhibit mTOR directly by binding to the catalytic domain at concentrations of 0.1-1 M (41, 42). Therefore, it is possible that phosphatidylinositol 3-kinase does not play a role in this mechanism despite the wortmannin-sensitive nature of the process. It has also recently been reported that phosphoinositide-dependent kinase 1 can phosphorylate p70 s6k in vivo (43), but there is no evidence to date to suggest that phosphinositide-dependent kinase 1 kinase activity is increased in response to amino acids, especially because PKB activity does not increase (Fig. 5).
Because of the sustained amino acid-stimulated activation of p70 s6k and the sensitivity to both wortmannin and rapamycin, this leads to the reasoning that this enzyme is part of the signal mechanism involved in the amino acid-stimulated increase in GS and glycogen synthesis activities. It is important to point out that p70 s6k itself may not play a role, and events may stem directly from mTOR. It has recently been discovered that mTOR is involved in the regulation of PP2A (40), a phosphatase known to act on GSK-3 in vitro. This could provide evidence to support a theory whereby GSK-3 is regulated by this phosphatase in response to amino acid availability. PP2A also has the ability to dephosphorylate GS at sites 2 and 3, leading to its activation; therefore, this phosphatase may play a coordinating role in the overall process.
In summary, the work reported here demonstrates that amino acid availability can regulate glycogen synthesis. Furthermore, it demonstrates that GSK-3 can be inactivated within cells independent of activation of PKB and p90 rsk .