Contraction Regulation of Akt in Rat Skeletal Muscle*

The protein serine/threonine kinase Akt/protein kinase B has been recognized as a critical signaling mediator for multiple cell systems. The function of Akt in skeletal muscle is not well understood, and whether contractile activity stimulates Akt activity has been controversial. In the current study, contraction in situ, induced via sciatic nerve stimulation, significantly increased Akt Ser473 phosphorylation in multiple muscle types including the extensor digitorum longus (13-fold over basal), plantaris (5.8-fold), red gastrocnemius (4.7-fold), white gastrocnemius (3.3-fold), and soleus (1.6-fold). In addition to increasing phosphorylation, contraction in situsignificantly increased the activity of all three Akt isoforms (Akt1 > Akt2 > Akt3) with maximal activation occurring at 2.5 min and returning to base line with 15 min of contraction. Akt phosphorylation and activity were also increased when isolated muscles were contracted in vitro in the absence of systemic factors, although to a much lesser extent. The phosphatidylinositol 3-kinase inhibitors wortmannin and LY294002 fully inhibited contraction-stimulated Akt phosphorylation and activity but did not diminish contraction-stimulated glycogen synthase kinase-3 phosphorylation and glycogen synthase activity. These results demonstrate that contraction increases Akt phosphorylation and activity in skeletal muscle and that this stimulation is rapid, transient, muscle fiber type-specific, and wortmannin- and LY294002-inhibitable. Akt signaling is not necessary for the regulation of glycogen synthase activity in contracting skeletal muscle.

Insulin and contractile activity are the major regulators of glucose uptake, glycogen synthesis, and protein synthesis in adult skeletal muscle. The signal transduction mechanism by which insulin induces these metabolic responses has been the focus of intense research, and it is now well established that the activation of phosphatidylinositol 3-kinase (PI3K) 1 is critical for insulin actions, as nearly all physiological responses of mammalian cells to insulin are prevented by pharmacologic inhibition or by overexpression of dominant negative mutants of PI3K (1). In contrast, the molecular signaling mechanisms by which contractile activity leads to changes in muscle metabolic processes remain largely undefined. What is known is that insulin and contraction utilize different signaling pathways leading to glucose uptake and glycogen synthesis in skeletal muscle because contraction-stimulated glucose uptake occurs through a PI3K-independent mechanism (2)(3)(4), and contraction-stimulated glycogen synthase activation can occur in the absence of PI3K activation (5). Downstream mediators beyond PI3K in insulin signaling have been characterized (6,7); however, whether insulin and contraction signaling converge at a point that is distal to PI3K is unknown.
Akt is a serine/threonine kinase whose catalytic domain is closely related to those of protein kinase A and protein kinase C (thus also termed protein kinase B). Three isoforms of Akt (Akt1/PKB␣, Akt2/PKB␤, and Akt3/PKB␥) have been identified, all of which are expressed in skeletal muscle (8). Akt is activated by a wide variety of growth factors in a PI3K-dependent manner through translocation to the plasma membrane and phosphorylation on two regulatory sites, Thr 308 in the activation loop and Ser 473 in the hydrophobic C-terminal regulatory domain (6,9). Thr 308 is phosphorylated by 3Ј-phosphoinositide-dependent protein kinase 1 (10), whereas the mechanism of phosphorylation of the Ser 473 site remains controversial (11)(12)(13)(14). Several studies have reported that Akt can be activated in some cell systems by a mechanism independent of PI3K activation, for example in response to growth hormone treatment (15), and increases in intracellular calcium or cAMP (16,17). The mechanism for this PI3K-independent Akt activation is poorly understood.
Overexpression of constitutively active forms of Akt in insulin-responsive cells mimics the actions of insulin (18 -20), although it is not clear whether Akt is necessary for all of the effects of insulin (21). Upon stimulation by agonists, Akt is dissociated from the membrane and phosphorylates downstream targets including glycogen synthase kinase-3 (GSK-3), Bad, forkhead transcription factors, 6-phosphofructo-2-kinase, endothelial nitric-oxide synthase, mammalian target of rapamycin, and insulin receptor substrate-1 (IRS-1) (6,22). Although involvement of Akt in cellular growth and survival is established, the specific substrates for the various stimuli in vivo are unknown.
Whether muscle contraction activates Akt and utilizes this molecule as a signaling mediator is controversial. Early studies reported no Akt activation or phosphorylation with contraction (23)(24)(25) or physical exercise (5,26,27). In contrast, other studies have shown significant activation or phosphorylation of Akt in intact skeletal muscles in response to contraction (28,29). These discrepant findings suggest that contraction may regulate Akt in an intensity-and time-dependent manner, or perhaps Akt stimulation in muscle may be fiber type-specific. If contraction does increase Akt activity, the mechanism leading to activation and the biological consequences of stimulating this kinase in contracting muscle are unknown. To establish whether contraction regulates Akt activity in skeletal muscle, in situ and in vitro contractions with detailed time course experiments in multiple muscle types were performed. We report that contraction clearly stimulates Akt activity and phosphorylation in skeletal muscle and that this occurs in a timeand fiber type-specific manner. Surprisingly, contraction-induced Akt activation occurs through a wortmannin-sensitive mechanism, and furthermore, Akt does not appear to be necessary for the ability of muscle contraction to regulate glycogen synthase activity.

EXPERIMENTAL PROCEDURES
Animals-Protocols for animal use were reviewed and approved by the Institutional Animal Care and Use Committee of the Joslin Diabetes Center and were in accordance with National Institutes of Health guidelines. Male Sprague-Dawley rats obtained from Taconic (Germantown, NY) were fed standard laboratory chow and water ad libitum. Rats were fasted overnight (10:00 p.m. to 9:00 a.m.) prior to the experiment.
In Vitro Muscle Incubation-Rats weighing 60 -80 g were killed by cervical dislocation, and EDL muscles were rapidly dissected. Tendons from both ends of the muscle were tied with suture (silk 4-0) and mounted on an incubation apparatus to maintain resting length. Isolated muscles were preincubated for 30 min in Krebs-Ringer bicarbonate buffer (117 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , and 24.6 mM NaHCO 3 , pH 7.5) containing 2 mM pyruvate. Pyruvate (2 mM) was always present in the Krebs-Ringer bicarbonate buffer for the entire incubation period. After preincubation muscles were incubated for an additional 10-min period for basal, insulin, or contraction treatment. For basal and insulin treatments, muscles were incubated in Krebs-Ringer bicarbonate buffer in the presence or absence (basal) of 50 milliunits/ml insulin. For isolated muscle contraction, muscles were transferred to a tissue support with stimulating electrodes (Harvard Apparatus, Holliston, MA) and were stimulated using the following protocol: train rate, 2/min; train duration, 10 s; pulse rate, 100 Hz; pulse duration, 0.1 ms at 100 V for 1, 3, 5, or 10 min (4). Force production during the contraction was monitored with an isometric force transducer (Kent Scientific, Litchfield, CT) and recorded with a chart recorder (Kipp & Zonen, Delft, Holland). For inhibitor experiments, wortmannin (100 or 500 nM) or LY294002 (20 or 100 M), or vehicle (0.1-0.2% dimethyl sulfoxide) was present throughout the entire incubation period. The buffers were kept at 37°C throughout the experiment and were gassed continuously with 95% O 2 and 5% CO 2 . Immediately after the treatments, muscles were quickly frozen in liquid nitrogen.
Immunoblotting-Equal amounts of muscle proteins (60 -80 g, buffer A) were resolved by SDS-PAGE (8 -10% polyacrylamide), transferred to nitrocellulose membranes, and blocked in Tris-buffered saline (10 mM Tris, 150 mM NaCl) plus 0.05% Tween 20, pH 7.8 (TBST) containing either 5% non-fat dry milk or 5% bovine serum albumin for 1 h at room temperature. The membranes were incubated in TBST containing either 3% non-fat dry milk or 5% bovine serum albumin with the indicated antibodies overnight at 4°C. Membranes were probed with horseradish peroxidase-conjugated secondary antibodies (1:2,000) in TBST containing 5% non-fat dry milk for 1 h at room temperature, and antibody-bound proteins were visualized using an enhanced chemiluminescence system (PerKinElmer Life Sciences). Bands were scanned and quantitated by densitometry (Molecular Dynamics, Sunnyvale, CA).
For tyrosine phosphorylation of the insulin receptor and IRS-1, muscle lysates (buffer B) were subjected to immunoprecipitation overnight at 4°C with 5 l of polyclonal anti-insulin receptor antibody, and 8 l of polyclonal anti-IRS-1 antibody coupled to protein A-agarose (Pierce Chemical Co.). The immunoprecipitates were washed and separated by 6% SDS-PAGE, and tyrosine-phosphorylated proteins were visualized as described previously (25) using pY99 (1:1,000 dilution).
Activity Assays-For the Akt activity assay, muscle lysates (300 g, buffer A) were subjected to immunoprecipitation for 4 h at 4°C with 3 g of Akt1/2 or 3 g of Akt1, Akt2, and Akt3 isoform-specific antibody coupled to protein G-Sepharose (Amersham Biosciences, Inc.). The immune pellets were washed extensively, and in vitro kinase assays were performed as described previously (25) using Akt/SGK peptide (RPRAATF, obtained from Upstate Biotechnology) as substrate.
For the PI3K activity assay, muscle lysates (500 g, buffer A) were subjected to immunoprecipitation overnight at 4°C with anti-phosphotyrosine antibody coupled to protein G-Sepharose. The immune complexes were washed, and PI3K activity was determined as described (31).
For glycogen synthase activity, pulverized muscle (10 -20 mg) was homogenized with a Polytron in ice-cold buffer (1:20) containing 50 mM Tris-HCl, 5 mM EDTA, 100 mM NaF, pH 7.8. Glycogen synthase activity in the presence and absence of 6.7 mM glucose 6-phosphate was determined as described previously (32) and is expressed as the ratio of glycogen synthase activity in the absence of glucose 6-phosphate to that in the presence of glucose 6-phosphate.
Statistics-Data are expressed as means Ϯ S.E. Statistical analysis was undertaken using a paired Student's t test and one-way analysis of variance. When analysis of variance revealed significant differences, further analysis was performed using Tukey's post hoc test for multiple comparisons. Differences between groups were considered statistically significant when p Ͻ 0.05.

In Situ Contraction Stimulates Akt Ser 473 Phosphorylation in Rat Hind
Limb Skeletal Muscles-We initially assessed the effects of contraction on Akt Ser 473 phosphorylation in rat hind limb skeletal muscles using multiple muscle types. Fig. 1A demonstrates immunodetection of Akt phosphorylated on Ser 473 with 10 min of contraction by sciatic nerve stimulation or sham operation. Contraction significantly increased Akt Ser 473 phosphorylation in all muscle types (Fig. 1B). The largest increase in Akt phosphorylation was in the predominantly fast twitch EDL muscle (ϳ13-fold versus sham), and an intermediate elevation in phosphorylation was seen in other muscles. Soleus (predominantly slow twitch fibers) showed high basal Ser 473 phosphorylation (5.5-fold greater than EDL), whereas the fold change in Ser 473 phosphorylation with contraction was minimal (1.6-fold over sham). This high basal phosphorylation in soleus may be caused by a greater abundance of Akt1/2 protein (Fig. 1A, lower box). Contraction did not alter Akt1/2 protein expression in any muscle group (Fig. 1A). These data clearly demonstrate that sciatic nerve stimulation to produce muscle contractions causes Akt Ser 473 phosphorylation in multiple intact rat skeletal muscles.
In Situ Contraction Stimulates Phosphorylation of Akt on Ser 473 and Thr 308 with a Differential Time Course-We next determined the temporal regulation of Akt phosphorylation at Thr 308 and Ser 473 with contraction over a time range of 1-30 min. Peroneal nerve stimulation was used to produce contractions exclusively in the dorsiflexor muscle group that includes the EDL and tibialis anterior muscles. Fig. 2A depicts immunodetection of Akt phosphorylated on Ser 473 and Thr 308 using two different antibodies. Because of sequence conservation surrounding these two phosphorylation sites, antibodies specific for phospho-Thr 308 or phospho-Ser 473 appear to recognize all three Akt isoforms (33). Interestingly, although the relative changes in maximal phosphorylation level were comparable between the two residues, the contraction time required to achieve peak Akt phosphorylation was more rapid for Thr 308 than Ser 473 (1 min versus 5 min, Fig. 2B). Phosphorylation on both sites decreased to basal levels (sham) with longer time periods of contraction (15 and 30 min). Contraction did not alter Akt1/2 protein expression at any time point (data not shown). Injection of a maximal dose of insulin (10 min) caused robust increases in Ser 473 and Thr 308 compared with shamoperated muscles (Fig. 2A). These results illustrate that contraction in intact muscle causes a rapid and transient increase in Akt phosphorylation and that the increase and decrease in Thr 308 phosphorylation precede that of Ser 473 phosphorylation.
In Situ Contraction Activates Three Isoforms of Akt and Stimulates Serine Phosphorylation of GSK-3-We next determined whether contraction regulates Akt activity in an isoform-specific manner. As shown in Fig. 3A, contraction significantly increased the activity of all three isoforms, with maximal activation of 3.9-fold (Akt1), 2.5-fold (Akt2), and 1.9fold (Akt3) compared with sham-operated muscles. Temporal changes in Akt activity were similar among isoforms, and maximal activation was achieved with 2.5 min of contraction. In contrast to the relatively modest increases in Akt activation with contraction, a maximal dose of insulin (10 min) caused robust increases of ϳ20-fold (Akt1) and ϳ10-fold (Akt2), whereas similar to contraction, Akt3 was stimulated by ϳ1.5fold (data not shown).
To examine whether contraction-stimulated Akt activation results in enhanced phosphorylation of a putative downstream target in vivo, we assessed GSK-3 serine phosphorylation by immunoblotting (Fig. 3B). GSK-3 is probably the best characterized substrate of Akt and has been proposed to relay the insulin signal leading to glycogen synthesis via PI3K (34). GSK-3␣ Ser 21 phosphorylation rapidly increased with 1 min of contraction (ϳ3-fold versus basal) and tended to decrease with time (Fig. 3C). GSK-3␤ Ser 9 phosphorylation peaked at 5 min (ϳ2-fold versus basal), and this increase was sustained until 15 min. Insulin increased phosphorylation of GSK-3 (both ␣ Ser 21 and ␤ Ser 9 ) by ϳ5-fold compared with basal (Fig. 3B).
In Situ Contraction in Intact Muscle Does Not Increase Insulin Receptor and IRS-1 Tyrosine Phosphorylation-The mechanism by which Akt is activated in response to contraction is unknown. Because increased blood flow in response to contraction may facilitate delivery or binding of endogenous insulin to its receptor leading to activation of Akt, we determined whether insulin receptor and IRS-1 tyrosine phosphorylation are increased in response to contraction compared with shamoperated muscles. Muscle lysates from sham and contracted EDL-tibialis anterior complexes were immunoprecipitated with insulin receptor or IRS-1 antibody and immunoblotted with phosphotyrosine antibody. As shown in Fig. 4, A and B, there was no increase in insulin receptor or IRS-1 tyrosine phosphorylation in response to contraction at any time point measured. In fact, tyrosine phosphorylation tended to decrease with contraction. Our previous work has shown that IRS-2, which is another IRS expressed in muscle, is also not activated by contraction (25). These results demonstrate that Akt activation in response to contraction in intact muscle is not the result of enhanced insulin receptor or IRS tyrosine phosphorylation.
Muscle Contraction in Vitro Stimulates Akt Phosphorylation in the Absence of Systemic Factors-To determine whether contraction can stimulate Akt activity in the complete absence of systemic factors, we utilized an isolated muscle incubation system to produce muscle contractions in vitro. EDL muscles were preincubated in Krebs-Ringer bicarbonate buffer with 2 mM pyruvate for 30 min as a washout to recover muscles from any potential stress that may have activated Akt (i.e. stretch, hypoxia). Muscles were transferred to a new tube and subjected to 10 min of contractions produced by electrical stimulation. As shown in Fig. 5, muscle contractions increased Akt Ser 473 phosphorylation with maximal phosphorylation of ϳ3-fold above basal at 5 min. Although the time course of changes in Akt Ser 473 phosphorylation was similar to the in situ nerve stimulation experiment (Fig. 2), the magnitude of increase in Ser 473 phosphorylation with in vitro contractions was only 23% of that measured in response to in situ contractions (Fig. 2). The effects of in vitro contraction (3, 10, and 15 min, n ϭ 3/group) in soleus muscles on Akt Ser 473 phosphorylation was examined, and no significant increase was observed with contractions (data not shown).
PI3K Inhibitors Abolish Contraction-stimulated Akt Activity and Ser 473 Phosphorylation in Vitro-The mechanism through which contraction regulates Akt is not known; however, a firm link between PI3K and Akt signaling has been established in multiple cell types (6). There are also several conditions reported in which Akt is activated independently of PI3K signaling (15)(16)(17). We determined whether contraction-stimulated Akt activation is dependent on PI3K signaling using two structurally distinct inhibitors, wortmannin and LY294002. Isolated EDL muscles were contracted or insulin treated in the absence or presence of wortmannin (100 nM, 500 nM) or LY294002 (20 M, 100 M). In the presence of wortmannin, contraction-stimulated Akt Ser 473 phosphorylation and Akt1/2 activity were completely inhibited at both 100 nM and 500 nM (Fig. 6, A and  B). Insulin-stimulated Akt Ser 473 phosphorylation and activity were partially inhibited by 100 nM wortmannin and were fully inhibited at 500 nM (Fig. 6, A and C). Akt1/2 protein expression did not change with contraction, insulin, or wortmannin treatment (Fig. 6A). LY294002 at both concentrations fully blocked both contraction-and insulin-stimulated Akt Ser 473 phosphorylation and Akt1/2 activity (data not shown). Wortmannin did not affect tension production in the contracting muscles at either dose, whereas the higher dose of LY294002 (100 M) reduced the force production of the muscle (data not shown). For this reason, wortmannin was chosen for further experiments. These results demonstrate that Akt activation by both in vitro contraction and insulin treatment are inhibitable in the presence of wortmannin and LY294002, suggesting that PI3K signaling and/or wortmannin-sensitive molecules are involved in the regulation of contraction-stimulated Akt activation.
Because contraction-stimulated Akt activation is wortmannin-and LY294002-sensitive, we next tested whether PI3K activity is increased in response to contraction. We measured phosphotyrosine-associated PI3K activity with in situ contraction (EDL-tibialis anterior muscle complex) at multiple time points (1, 2.5, 5, 15, or 30 min, n ϭ 3-5) or in response to maximal insulin for 10 min. Insulin increased PI3K activity by 4-fold, whereas contraction did not increase PI3K activity at any time point measured (data not shown), even though we observed significant Akt phosphorylation and activation in the same muscle samples (Figs. 2 and 3A). In vitro contraction (10 min) also did not change phosphotyrosineassociated PI3K activity compared with basal in EDL muscles (data not shown).

Effects of Wortmannin on Contraction-and Insulin-stimulated GSK-3 Phosphorylation and Glycogen Synthase Activity-
Glycogen synthase activity has been proposed to be regulated by multiple mechanisms (35), one of which is signaling through PI3K, Akt, and GSK-3. We examined whether inhibition of contraction-and insulin-stimulated Akt activity in the presence of wortmannin affects serine phosphorylation of GSK-3 (Fig. 7). In the basal state, wortmannin decreased phosphorylation of GSK-3␣ Ser 21 and GSK-3␤ Ser 9 by 60 -70%. In vitro contraction for 10 min increased GSK-3␣ Ser 21 (8-fold above basal) and GSK-3␤ Ser 9 (3-fold) phosphorylation. In the presence of wortmannin, contraction-stimulated GSK-3 phosphorylation was partially blunted. This degree of inhibition may be because of the effects of wortmannin to inhibit basal levels of GSK-3 phosphorylation. Insulin stimulation for 10 min signif-icantly increased GSK-3␣ Ser 21 and GSK-3␤ Ser 9 phosphorylation (13-fold and 4-fold, respectively). Wortmannin partially inhibited phosphorylation at 100 nM and completely inhibited at 500 nM. Additional experiments showed that 30 min of insulin stimulation increased GSK-3␣ Ser 21 and GSK-3␤ Ser 9 phosphorylation, and this increase was also fully inhibited in the presence of wortmannin (data not shown). GSK-3␣/␤ protein expression did not change with contraction, insulin, or wortmannin treatment (Fig. 7A).
We next examined whether inhibition of Akt activity with wortmannin attenuates contraction-stimulated glycogen synthase activity in skeletal muscle (Fig. 8). In vitro contraction for 10 min increased glycogen synthase activity by ϳ3-fold, but this increase was not inhibited in the presence of wortmannin (100 or 500 nM). Insulin treatment for 30 min resulted in a modest stimulation of glycogen synthase activity (ϳ1.6-fold). Wortmannin completely inhibited this increase at a concentration of 500 nM. These data demonstrate that contraction-stimulated serine phosphorylation of GSK-3 and glycogen synthase activation are regulated largely by an Akt-independent pathway, whereas insulin-stimulated GSK-3 phosphorylation and glycogen synthase activity are PI3K/Akt-dependent. DISCUSSION The level of muscle contractile activity is an important regulator of substrate metabolism and transcriptional events in skeletal muscle, yet the molecular signaling mechanisms that regulate these responses are poorly understood. Akt has emerged as a central signaling molecule involved in multiple cell systems playing major roles in cellular growth, survival, and metabolism (6,9), and therefore, it is a logical candidate to be a critical signaling protein in contracting skeletal muscle. However, previous studies have failed to come to a consensus as to whether muscle contraction and/or physical exercise regulate the activity of this Ser/Thr kinase. The current study, which was designed to investigate the effects of contraction on Akt using multiple muscle types, detailed time courses, and multiple isoform-specific antibodies, clearly demonstrates that contraction does activate Akt in skeletal muscle. The increase in Akt phosphorylation and activity occurred in muscle contracted in situ and in isolated muscles contracted in vitro in the absence of systemic factors.
Previous studies, including our own work, failed to report an increase in Akt activity with muscle contraction or exercise (5,(23)(24)(25)(26)(27). These discrepancies may be caused by differences in the muscle studied, the time of experimentation, and/or contraction models employed. For example, Lund et al. (24) and Brozinick and Birnbaum (23) observed no effect of contraction on Akt activity using isolated soleus muscles. As shown in this and other studies, the predominantly slow twitch soleus muscle is resistant to contraction-stimulated Akt activation (29), despite abundant expression of Akt (28,36) and a robust activation with insulin stimulation (24,36). Brozinick and Birnbaum (23) also saw no effect of contraction on the predominantly fast twitch epitrochlearis muscles, but these experiments were done following 20 min of contraction, a time point where we observe that Akt phosphorylation and activity have returned to baseline. The rapid time course of Akt activation with contraction may also explain why it has been reported that 30 min of in vivo exercise does not increase Akt activity (26).
Akt can be activated by a wide variety of growth factors (6,9), and contraction causes an enhanced blood flow to contracting muscles, raising the possibility that Akt activation with contraction may be caused by an increased delivery of systemic factors to the muscle. Our results demonstrate that increased Akt activation in response to contraction can occur in the absence of endogenous insulin or other systemic factors (e.g. growth factors, cytokines). However, the magnitude of the in-crease in Akt activity and phosphorylation in isolated muscles in vitro was only about 25% of that of in situ contraction. Although it is difficult to rule out differences in contraction intensity between in situ and in vitro contractions, these data raise the possibility that systemic factors may be playing a role in enhancing Akt activity in vivo with contraction.
Despite a high degree of structural identity and a similar activation process (6,9), there appear to be distinct physiological roles for the three Akt isoforms (37)(38)(39). Recent work has demonstrated that Akt2-deficient mice have a mild impairment in insulin-stimulated glucose uptake in EDL muscles (39), whereas Akt1 knockout mice have global retardation of growth with normal whole body glucose homeostasis (37,38). Here we show that contraction increases the activity of all three isoforms of Akt and that although the time course of activation is similar, the magnitude of activation is greatest for Akt1, followed by Akt2 and then Akt3. Insulin stimulation showed a pattern of activation similar to that of Akt1 Ͼ Akt2 Ͼ Akt3, although the absolute increases in activity were much higher for Akt1 and Akt2 in response to the supramaximal dose of insulin. Akt1 is the isoform activated most robustly by insulin and contraction, two potent stimulators of glucose uptake in muscle, but Akt1 knockout mice have normal whole body glucose homeostasis. This indicates that this isoform may be not a key regulator of glucose uptake but instead may play other roles in skeletal muscle.
Our finding that contraction-stimulated Akt activation was blocked in the presence of PI3K inhibitors was unexpected because multiple studies have shown that muscle contraction (31, 40 and current study) and an acute bout of exercise (27,41) do not activate class I A PI3K associated with tyrosine-phosphorylated proteins. This raises the question of how contraction stimulates Akt activity in a wortmannin-sensitive manner when there is no detectable activation of PI3K. There are several possibilities. First, many studies have reported the dissociation of PI3K and Akt activity and that Akt activation does not necessarily correlate with the amount of PI3K recruited to tyrosine-phosphorylated proteins (27,(42)(43)(44). One example in skeletal muscle is from a study of muscle-specific insulin receptor-deficient mice, where insulin treatment increased Akt activity with no detectable activation of PI3K activity associated with tyrosine-phosphorylated proteins (27). Another example of this dissociation comes from a study in human subjects with type 2 diabetes where there was reduced insulin-stimulated PI3K activity but normal activation of Akt (42). A second possibility is that wortmannin is acting through nonspecific inhibition of some undefined upstream signaling molecules of Akt, other than PI3K. For example, wortmannin has been shown to inhibit phospholipase A 2 (45), and recent evidence demonstrated that angiotensin II stimulation of Akt activation occurs through phospholipase A 2 and independently of PI3K (46). A third possibility to explain our finding of wortmannin-inhibitable Akt activation is that class I B PI3K, which is also sensitive to wortmannin, may be involved in the regulation of contraction-stimulated Akt activation. The class I B PI3K is a heterodimer composed of a p110␥ catalytic subunit and a p101 adaptor subunit, having a distinct activation mechanism compared with class I A (1). Class I A PI3Ks include the p110␣, ␤, and ␦ catalytic subunits that are translocated via the two SH2 domains of their p85 adaptor protein to phosphorylated Tyr-Xaa-Xaa-Met motifs (1,47). In contrast, the catalytic subunit of class I B PI3K (p110␥) does not bind to p85 adaptors and is activated by the G␤␥ subunits of heterotrimeric G proteins in vitro (48,49). Northern blot analysis has shown that p110␥ mRNA is abundant in skeletal muscle (48). Thus, muscle FIG. 8. Effects of wortmannin on contraction-and insulinstimulated glycogen synthase activity in skeletal muscle in vitro. Isolated EDL muscles were incubated in the absence (Ϫ) or presence of the indicated doses of wortmannin for 30 min, and thereafter muscles were rested (Basal) for 10 or 30 min, contracted for 10 min (Contraction) (panel A), or insulin treated (Insulin) for 10 or 30 min (panel B). Wortmannin was present throughout the entire incubation period. Muscles were quickly frozen in liquid nitrogen, processed, and glycogen synthase activity was assay as described under "Experimental Procedures." Data are the means Ϯ S.E., n ϭ 4 or 5/group. # , p Ͻ 0.05 (versus basal). contraction may activate class 1 B PI3K leading to Akt activation, and this will be an important area for future investigation.
GSK-3, an established substrate for Akt, has been implicated in the control of many cellular processes including glycogen synthesis, protein synthesis, and the modulation of transcription factor activity (50). One proposed function of GSK-3 in skeletal muscle is to stimulate glycogen synthesis by promoting the dephosphorylation of glycogen synthase (19,34,51). Insulin has been shown to deactivate GSK-3 through a mechanism that involves activation of PI3K and Akt-mediated serine phosphorylation on Ser 21 of GSK-3␣ and Ser 9 of GSK-3␤ (34,52). In the current study both contraction and insulin increased serine phosphorylation of the GSK-3 isoforms, but only the insulin effect was fully inhibited in the presence of wortmannin. The minimal inhibition of contraction-stimulated GSK-3 phosphorylation with wortmannin, coincident with full inhibition of Akt, provides additional support to our previous hypothesis (5) that there are Akt-independent mechanisms for the regulation of GSK-3 in contracting skeletal muscle. In fact, GSK-3 can be phosphorylated by other kinases such as p70 S6 kinase (53), mitogen-activated protein kinase kinase-activated kinase-1 (also known as p90 rsk ) (54), and protein kinase A (55), and these proteins may play a role in GSK-3 phosphorylation with contraction.
The current results also demonstrate that Akt is not playing a role in contraction-stimulated glycogen synthase activity. Our previous work has indicated that exercise-stimulated glycogen synthase activation may be mediated through GSK-3 inactivation independently of serine phosphorylation (5), through the glycogen/sarcoplasmic reticulum-associated type I protein phosphatase (56), or most likely, through a combination of these two molecules. In addition to our novel findings of wortmannin-insensitive regulation of contraction-stimulated glycogen synthase activity, this work is the first demonstration that insulin-stimulated GSK-3 phosphorylation and glycogen synthase activity are abolished with PI3K inhibitors in skeletal muscle tissue. Thus, the PI3K/Akt/GSK-3 signaling pathway appears to play an important role in the regulation of insulinstimulated glycogen synthesis in skeletal muscle. However, it should be noted that GSK-3 alone is not sufficient to account for glycogen synthase dephosphorylation and activation by insulin (35,57), and it is possible that glycogen synthase activity was inhibited by upstream molecules other than GSK-3. In fact, wortmannin inhibits type I protein phosphatase, which is known to regulate glycogen synthase by insulin (58).
To identify the biological targets of Akt in skeletal muscle, the use of pharmacological reagents, genetic, and/or biochemical approaches will be important. Recently, it has been reported that ML-9, which was originally developed as a myosin light chain kinase inhibitor, abolishes both recombinant and endogenous Akt activity in various cell systems without attenuating insulin stimulation of insulin receptor and IRS-1 tyrosine phosphorylation (59) and PI3K activity (59,60). We tested this compound using the isolated muscle incubation system and found that ML-9 significantly inhibited insulin-stimulated tyrosine phosphorylation of the insulin receptor and severely attenuated contractile force generation in the muscle. 2 These results indicate that ML-9 cannot be used as a selective Akt inhibitor in skeletal muscle tissue. The development of specific cell-permeable inhibitors and the use of Akt knockout mice will be critical steps for further understanding the functions of Akt in contracting skeletal muscle.