Maturation of the Regulation of GLUT4 Activity by p38 MAPK during L6 Cell Myogenesis*

Insulin stimulates glucose uptake in skeletal muscle cells and fat cells by promoting the rapid translocation of GLUT4 glucose transporters to the plasma membrane. Recent work from our laboratory supports the concept that insulin also stimulates the intrinsic activity of GLUT4 through a signaling pathway that includes p38 MAPK. Here we show that regulation of GLUT4 activity by insulin develops during maturation of skeletal muscle cells into myotubes in concert with the ability of insulin to stimulate p38 MAPK. In L6 myotubes expressing GLUT4 that carries an exofacial myc-epitope (L6-GLUT4myc), insulin-stimulated GLUT4myc translocation equals in magnitude the glucose uptake response. Inhibition of p38 MAPK with SB203580 reduces insulin-stimulated glucose uptake without affecting GLUT4myc translocation. In contrast, in myoblasts, the magnitude of insulin-stimulated glucose uptake is significantly lower than that of GLUT4myc translocation and is insensitive to SB203580. Activation of p38 MAPK by insulin is considerably higher in myotubes than in myoblasts, as is the activation of upstream kinases MKK3/MKK6. In contrast, the activation of all three Akt isoforms and GLUT4 translocation are similar in myoblasts and myotubes. Furthermore, GLUT4myc translocation and phosphorylation of regulatory sites on Akt in L6-GLUT4myc myotubes are equally sensitive to insulin, whereas glucose uptake and phosphorylation of regulatory sites on p38 MAPK show lower sensitivity to the hormone. These observations draw additional parallels between Akt and GLUT4 translocation and between p38 MAPK and GLUT4 activation. Regulation of GLUT4 activity by insulin develops upon muscle cell differentiation and correlates with p38 MAPK activation by insulin.


From the Programme in Cell Biology, The Hospital for Sick Children, Toronto, Ontario M5G 1X8, Canada
Insulin stimulates glucose uptake in skeletal muscle cells and fat cells by promoting the rapid translocation of GLUT4 glucose transporters to the plasma membrane. Recent work from our laboratory supports the concept that insulin also stimulates the intrinsic activity of GLUT4 through a signaling pathway that includes p38 MAPK. Here we show that regulation of GLUT4 activity by insulin develops during maturation of skeletal muscle cells into myotubes in concert with the ability of insulin to stimulate p38 MAPK. In L6 myotubes expressing GLUT4 that carries an exofacial myc-epitope (L6-GLUT4myc), insulin-stimulated GLUT4myc translocation equals in magnitude the glucose uptake response. Inhibition of p38 MAPK with SB203580 reduces insulinstimulated glucose uptake without affecting GLUT4myc translocation. In contrast, in myoblasts, the magnitude of insulin-stimulated glucose uptake is significantly lower than that of GLUT4myc translocation and is insensitive to SB203580. Activation of p38 MAPK by insulin is considerably higher in myotubes than in myoblasts, as is the activation of upstream kinases MKK3/ MKK6. In contrast, the activation of all three Akt isoforms and GLUT4 translocation are similar in myoblasts and myotubes. Furthermore, GLUT4myc translocation and phosphorylation of regulatory sites on Akt in L6-GLUT4myc myotubes are equally sensitive to insulin, whereas glucose uptake and phosphorylation of regulatory sites on p38 MAPK show lower sensitivity to the hormone. These observations draw additional parallels between Akt and GLUT4 translocation and between p38 MAPK and GLUT4 activation. Regulation of GLUT4 activity by insulin develops upon muscle cell differentiation and correlates with p38 MAPK activation by insulin.
Insulin is the major regulator of blood glucose levels in the fed state when skeletal muscle becomes the primary consumer of glucose (1). The rate-limiting determinant of glucose utilization by muscle is its uptake mediated by glucose transporters (2). GLUT4 1 is the most abundant glucose transporter isoform in skeletal muscle and adipose tissue (3)(4)(5)(6) and is responsible for the majority of insulin-dependent glucose uptake in these tissues (7,8). It has long been recognized that insulin promotes the rapid translocation of GLUT4 glucose transporters from intracellular membrane compartments to the plasma membrane (3-5, 9, 10). Current views hold that GLUT4 translocation is entirely responsible for the increase in glucose uptake in response to insulin. However, the two widely used methods to measure GLUT4 translocation, subcellular fractionation and photoaffinity labeling of surface GLUT4, have drawbacks that compromise their ability to measure GLUT4 translocation accurately (see "Discussion"). These methods have arrived at variable conclusions regarding how closely GLUT4 translocation matches the stimulation of glucose uptake by insulin (5,6,(11)(12)(13)(14)(15)(16)(17)(18)(19). Thus, the possibility exists that the intrinsic activity of glucose transporters could be regulated in response to the hormone, and new approaches to measure GLUT4 translocation in intact cells are required (without cellular homogenization or protein immunoprecipitation).
To this end, we have used L6 muscle cells that stably overexpress GLUT4 encoding a myc epitope in its large exofacial loop (L6-GLUT4myc cells (20)). GLUT4myc can be readily detected at the cell surface of intact cells by an enzyme-linked immunosorbent-like assay (21). In these cells, GLUT4myc shows insulin-regulated behavior consistent with that of GLUT4 in 3T3-L1 adipocytes (22,23). Thus, 90% of GLUT4myc is sequestered intracellularly in the basal state, and a significant portion translocates to the cell surface in response to insulin (24). All GLUT4myc molecules are available for recycling to the cell surface (25). The exocytic and endocytic rates of GLUT4myc (24) mimic those reported for GLUT4 (26,27). The K m of glucose uptake is similar in L6GLUT4myc to that in the parental L6 cells (28,29). Most significantly, the 10% of total GLUT4myc present at the cell surface is still higher than the endogenous levels of GLUT1 or GLUT3 (by almost 100-fold (30)), and is responsible for both basal and insulin-stimulated glucose uptake. This functional preponderance was established by the nearly complete inhibition of both basal and insulinstimulated glucose uptake rates by the drug indinavir (30,31), a rather selective inhibitor of glucose influx through GLUT4 but not GLUTs 1, 3, and 8 (32,33). Hence, L6-GLUT4myc cells are uniquely suitable to make direct comparisons between glucose uptake through GLUT4 and GLUT4 translocation. The molar expression of GLUT4myc in L6-GLUT4myc cells is 5-10 fold higher than that of endogenous GLUT4 in skeletal muscle (30). Using these cells we have observed that insulin-dependent GLUT4 translocation and stimulation of glucose uptake can be segregated in time (28), by their temperature sensitivity (28), their susceptibility to inhibition by wortmannin (34), and most strikingly, by specific inhibitors of p38 MAPK. The latter (pyridinylimidazoles SB203580 and SB202190 and chemically distinct aza-azulenes A291077 and A304000) reduced the insulin response of glucose uptake without interfering with GLUT4 translocation in L6-GLUT4myc myotubes and 3T3-L1 adipocytes without directly inhibiting glucose transporters (28,35,36).
L6-GLUT4myc muscle cells undergo differentiation from myoblasts to multinucleated myotubes through multiple cell fusions (37,38). In search for information on the mechanisms responsible for the segregation of GLUT4 translocation and glucose uptake in myotubes, we examined the maturation of the insulin response of GLUT4 translocation and glucose uptake during myogenesis. We report that, in the myoblast stage, insulin-stimulated glucose uptake is lower than in myotubes and is not sensitive to inhibition of p38 MAPK. Yet, translocation of GLUT4myc is similar in both stages of cellular differentiation. Moreover, insulin stimulates p38 MAPK and its upstream activators MKK3/6 in myotubes but slightly if at all in myoblasts. These results further correlate the p38 MAPK pathway to GLUT4 activation, because both events mature during L6 cell differentiation from myoblasts into myotubes.
L6-GLUT4myc Cell Line and Cell Culture-For selective experiments, wild-type L6 myoblasts grown and differentiated into myotubes as previously reported (39) were used where indicated. Otherwise, L6 myoblasts stably expressing GLUT4myc, created (20) and characterized (40,41) as described previously, were used throughout the study. L6-GLUT4myc cells were maintained in minimal essential medium-␣ supplemented with 10% FBS in a humidified atmosphere of air and 5% CO 2 at 37°C. For experiments with myoblasts only, L6 cells were seeded in medium containing 10% FBS and used at confluence, 2 days after seeding. L6 cells were differentiated in medium supplemented with 2% FBS into myotubes within 7 days after seeding. Cells were serumdepleted for 3-4.5 h prior to all experimental manipulations. Inhibitors were administered in Me 2 SO, and the maximum concentration of the vehicle did not exceed 0.05% (v/v). This concentration of vehicle was without effect on any of the parameters measured.
2-Deoxyglucose Uptake-Following serum depletion, cells were treated with 10 M SB203580 for 20 min before the addition of insulin at the indicated concentrations and times as described in the figure legends. Following treatments, cells were rinsed and immediately used for measurement of 2-deoxyglucose uptake in the absence of inhibitors as described previously (28). For the insulin-stimulated time-course analysis, cells were grown in six-well plates and treated as indicated, and then 2-deoxyglucose uptake was measured for 30 s. Nonspecific uptake was determined in the presence of 10 M cytochalasin B, and this value was subtracted from all other values. Cell-associated radioactivity was determined by lysing the cells with 0.05 N NaOH, followed by liquid scintillation counting. Total cellular protein was determined by the Bradford method.
GLUT4myc Translocation-GLUT4myc levels at the cell surface of intact myoblasts or myotubes were measured by an antibody-coupled colorimetric assay as described (24) using the anti-myc monoclonal 9E10 as the primary antibody and a donkey anti-mouse IgG conjugated to horseradish peroxidase as the secondary antibody.
Immunoblotting and Phosphorylation of p38 MAPK, MKK3/6, and Akt-Briefly, cells in six-well plates were incubated as indicated, lysed on ice with 300 l of 2ϫ Laemmli sample buffer per well supplemented with 7.5% ␤-mercaptoethanol (v/v), protease, and phosphatase inhibitors (28). Lysates were passed 5 times through a 27-gauge syringe and heated for 15 min at 65°C. 50-g aliquots of total protein were resolved by 10% SDS-PAGE to detect phosphorylation of p38 MAPK, MKK3/6, or Akt by immunoblotting using the corresponding phospho-specific antibodies at 1:500 dilutions. Anti-p38 ␣ and anti-p38 ␤ MAPK antibodies were used at 1:1000 and 1:200 dilutions, respectively. Goat anti-rabbit IgG conjugated to horseradish peroxidase was used as secondary antibody at a 1:7000 dilution. Proteins were detected by the Enhanced Chemiluminescence method according to the manufacturer's instructions (PerkinElmer Life Sciences, Boston, MA). Immunoblots were exposed to x-ray film to produce bands within the linear range, then quantified using National Institutes of Health (NIH) Image software.
Immunoprecipitation and Assay of p38 MAPK and Akt Activities-Immunoprecipitation of p38 MAPK␣ or ␤ from 500 g (total protein) of TX-100 detergent cell lysates containing phosphatase and protease inhibitors was performed overnight at 4°C was performed as described (28). Protein concentration of the lysates was determined by the bicinchoninic acid method. p38 MAPK immunocomplexes were incubated for 30 min at 30°C in 50 l of kinase buffer supplemented with 2 g of ATF2 recombinant protein and 200 M ATP per condition. Reactions were continuously mixed on a platform shaker and were stopped by the addition of 25 l of 2ϫ Laemmli sample buffer and heating for 30 min at 65°C. Samples were sedimented (12,000ϫ g), and then 50 l of the supernatant was resolved by 10% SDS-PAGE and electrotransferred onto polyvinylidene difluoride membranes to detect phosphorylation of ATF2, using phospho-specific ATF-2 antibody.
Immunoprecipitation of Akt1, Akt2, or Akt3 from 200 g (total protein) of TX-100 detergent cell lysates containing phosphatase and protease inhibitors overnight at 4°C was performed as described (40). Akt immunocomplexes were incubated in 30 l of kinase buffer supplemented with Crosstide (150 g/condition), 5 M ATP, and 2 Ci of [␥-32 P]ATP per condition. Reactions were sedimented by centrifugation for 30 s. 25 l of the supernatants from each reaction were transferred to a Whatman P81 phosphocellulose filters and washed four times for 5 min each with 1% phosphoric acid (v/v) and once with double-distilled water. Filters were allowed to air dry before liquid scintillation counting (40).
Statistical Analysis-Statistical analysis was performed using either unpaired Student's t test or analysis of variance test (Fischer, multiple comparisons) as indicated in the figure legends.

RESULTS
Differential Gains in Insulin-stimulated Glucose Uptake and GLUT4 Translocation during Myogenesis-L6 myoblasts and myotubes were treated in parallel with 100 nM insulin for 2-15 min at 37°C, followed by determination of 2-deoxyglucose uptake or GLUT4myc translocation. In these experiments, insulin continued to be present during the 30-s uptake of [ 3 H]2deoxyglucose (see "Experimental Procedures"). In this way, the time courses of insulin-stimulated GLUT4myc translocation and glucose uptake could be effectively compared at each time of insulin incubation. In myoblasts, the maximal -fold stimulation of glucose uptake by insulin (15-min time point) was lower than the -fold stimulation in GLUT4myc translocation (1.6 Ϯ 0.1-fold compared with 2.4 Ϯ 0.1, respectively, Fig. 1A). On the other hand, in myotubes insulin-stimulated both responses more than 2-fold by 15 min (2.2 Ϯ 0.1-fold for glucose uptake compared with 2.4 Ϯ 0.1, for translocation, Fig. 1B). Interestingly, the magnitude of the GLUT4myc translocation response was comparable in myoblasts and myotubes, but the full response of insulin-stimulated glucose uptake was not realized in the myoblasts.
In myoblasts, there was no difference in the rate of stimulation of glucose uptake by insulin and GLUT4myc translocation at the early times of insulin addition. The estimated t1 ⁄2 of glucose uptake was 3.3 min and that of the arrival of GLUT4myc at the plasma membrane also was 3.3 min. In myotubes, GLUT4myc rapidly translocated to the plasma membrane with a t1 ⁄2 of 2.0 min. In contrast, the stimulation of glucose uptake in L6 myotubes was delayed by a lag of about 2 min before a significant response could be measured. Thus, maximal stimulation of glucose uptake is reached only between 10 and 15 min in the myotubes and is significantly delayed (t1 ⁄2 of 5 min) with respect to the response in myoblasts.
SB203580 Inhibits 2-Deoxyglucose Uptake in Myotubes but Not in Myoblasts-L6-GLUT4myc myoblasts or myotubes were preincubated with the pyridinylimidazole SB203580 (10 M) or Me 2 SO vehicle only for 20 min. This was followed by treatment with insulin (100 nM) for an additional 20 min in the presence of SB203580. Uptake of 2-deoxyglucose was then measured in the absence of the inhibitor. Insulin caused a significant in-crease in glucose uptake in myoblasts (1.7 Ϯ 0.1-fold above basal, p Ͻ 0.001; Fig. 2A) and myotubes (2.2 Ϯ 0.2-fold above basal, p Ͻ 0.001, Fig. 2B). Preincubation with SB203580 did not have a significant effect on either basal or insulin-stimulated glucose uptake in myoblasts ( Fig. 2A). Similarly, preincubation with SB203580 did not affect the basal rate of glucose uptake in myotubes. However, it reduced the stimulation of glucose uptake by 65% (insulin: 2.2 Ϯ 0.2-fold, insulin plus SB203580: 1.4 Ϯ 0.1-fold, p Ͻ 0.001, Fig. 2B). It is unlikely that pyridinylimidazoles act by directly binding and inhibiting GLUT4myc, because both basal and insulin-stimulated rates of glucose uptake are mediated by GLUT4myc in these cells (30,31), but only the stimulated uptake was reduced by the drug. Moreover, SB203580 was not present during the transport assay. Furthermore, SB203580 did not inhibit glucose uptake when added only to the glucose transport solution in glucose uptake assays lasting up to 30 min (28,35), confirming that the inhibition is due to an event different from direct inhibition of GLUT4. Nonetheless, to ascertain that the myc epitope is not related to the sensitivity to SB203580, the selective inhibition of insulin-stimulated glucose uptake by SB203580 in myotubes vis à vis was confirmed in wild-type L6 muscle cells (Table I). This result conforms to our previous observation that SB203580 reduces the insulin-dependent portion of glucose uptake in wild-type L6 myotubes (35) and further establishes that maturation into myotubes is required for this susceptibility to SB203580 to manifest. Hence, glucose uptake in L6-GLUT4myc cells obeys similar regulation as in parental, untransfected cells.
Insulin increased GLUT4myc at the cell surface by 2.3 Ϯ 0.2-fold in myoblasts and 2.5 Ϯ 0.2-fold in myotubes (Fig. 2, C and D). In contrast to the stimulation of glucose uptake, SB203580 did not affect GLUT4myc density at the cell surface of basal or insulin-stimulated myotubes (Fig. 2D). This suggests that pretreatment with SB203580 reduces the glucose transport activity of GLUT4myc molecules that had been fully inserted into the plasma membrane, consistent with previous observations (28,35).
As shown in Fig. 1 above, the time course and maximal response to insulin of glucose uptake in myotubes differs from that in myoblasts. The myotubes responded with an initial delay but then achieved a higher stimulation of glucose uptake than the myoblasts. Given that SB203580 lowered the maximal glucose uptake response in myotubes (see Fig. 2), we examined how inhibition of p38 MAPK may affect the time course of insulin-stimulated glucose uptake in myotubes. Preincubation of L6-GLUT4myc myotubes with SB203580 prior to determination of the insulin-stimulated time course of glucose uptake reduced the maximal stimulation of glucose uptake to the same levels as in insulin-stimulated myoblasts (Fig. 3). This observation is in keeping with the lack of inhibition of insulin action by SB203580 in myoblasts. However, the lag in stimulation was not prevented (Fig. 3). These tantalizing results raise the possibility that an event downstream of p38 MAPK is responsible for the maximal stimulation of glucose uptake and that other factors upstream or parallel to p38 MAPK contribute to the delay in stimulation of glucose uptake observed in myotubes relative to myoblasts (see "Discussion").
Insulin-induced p38 MAPK Phosphorylation Increases during Myogenesis-Activation of p38 MAPK by diverse stimuli leads to its phosphorylation by upstream kinases on tyrosine and threonine residues in the TGY motif of its regulatory domain (42). The phosphorylation of these two sites is widely used as an indication of heightened p38 MAPK activity and can be detected with anti-phospho p38 MAPK antibody that recognizes the dual phosphorylated form of p38 MAPK isoforms ␣ To avoid skewing the glucose uptake curve to the right, the time of insulin treatment shown does not include the 30 s required for the uptake assay. Data points are the mean Ϯ S.E. of five to eight experiments performed in triplicate. Insulin-stimulated glucose uptake and GLUT4myc translocation are expressed relative to the respective basal values to allow for a clearer comparison between the assays (because GLUT4myc translocation is measured in optical density units that are normalized to the control untreated value within each experiment). The basal rate of 2-deoxyglucose uptake in these experiments was 20.2 Ϯ 4.1 pmol/ min/mg of protein in myoblasts and 7.9 Ϯ 0.2 pmol/min/mg of protein in myotubes. and ␤ (42). We monitored the initial time course of p38 phosphorylation upon insulin treatment (100 nM) of myoblasts and myotubes, by immunoblotting cell lysates with anti-phospho-p38 MAPK antibody. Representative immunoblots are shown in Fig. 4A for myoblasts and Fig. 4B for myotubes. The p38 MAPK protein levels in each sample (shown below) were detected upon subsequent immunoblotting with an anti-pan p38 MAPK antibody to ensure equal sample loading (Fig. 4, A and  B). Quantification of five similar experiments was performed to determine a phospho-p38:p38 protein ratio. The results indicated that p38 MAPK phosphorylation was more responsive to insulin in myotubes (phospho-p38 MAPK was increased 2.0 Ϯ 0.1-fold at 5 min and 2.5 Ϯ 0.3-fold at 10 min relative to the myotube basal levels, Fig. 4B) than myoblasts (phospho-p38 MAPK increased to a maximum of 1.4 Ϯ 0.1-fold at 5 min, p Ͻ 0.01, Fig. 4A). These observations are consistent with the ineffectiveness of SB203580 on insulin-stimulated glucose uptake in L6 myoblasts.
Insulin-stimulated p38 MAPK ␣ and ␤ Activity Is Regulated during Myogenesis-To further examine the effect of insulin on p38 MAPK activity during myogenesis, we complemented the above results of p38 MAPK phosphorylation with measurements of the kinase activity of the two isoforms of the enzyme that are susceptible to inhibition by SB203580. p38 MAPK ␣ and ␤ were each immunoprecipitated from basal and insulintreated myoblasts and myotubes, and their activity was measured in vitro toward recombinant ATF2. Immunoprecipitates were subsequently processed by SDS-PAGE and immunoblotted for phospho-ATF2. Several experiments were quantified by densitometric scanning of immunoblots like that shown in Fig. 4 (C and D). In myoblasts, the insulin-stimulated p38 MAPK ␣ and ␤ activities were moderately elevated 3.3 Ϯ 0.6fold and 3.3 Ϯ 0.7-fold above basal, respectively (Fig. 4C). In myotubes, insulin increased p38 MAPK ␣ and ␤ activities markedly by 11.1 Ϯ 0.5-fold and 7.8 Ϯ 2.2-fold above basal, respectively (Fig. 4D). The activities measured in vitro show higher changes than those observed with the anti-phospho-p38 antibody, possibly because the anti-phospho p38 MAPK antibody also detects other isoforms of the enzyme. Although insulin was effective in stimulating p38 MAPK activity in both

TABLE I Differential effect of SB203580 on insulin-stimulated glucose uptake in wild-type L6 myoblasts and myotubes
Wild-type L6 muscle cells at the stage of myoblasts or myotubes were pretreated without or with 10 M SB203580 for 20 min, followed by treatment without or with 100 nM insulin for 20 min in the corresponding absence or presence of SB203580. Cultures were rinsed three times with HEPES-buffered saline solution without additions and uptake of 2-deoxyglucose was then determined in the absence of SB203580 or insulin. Results are in pmol/min/mg of protein. myoblasts and myotubes, p38 MAPK responded to insulin more robustly in myotubes. Notably, the stimulation of p38 MAPK ␣ and ␤ activities by hyperosmolar stress was not different in myoblasts from that in myotubes (results not shown). Basal p38 MAPK Activity Decreases during Myogenesis-Interestingly, the basal p38 MAPK phosphorylation status decreased with differentiation of L6 cells. Phospho-p38 MAPK levels in myotubes were only 45% of that seen in myoblasts (Table II). The basal activity for the ␣ and ␤ isoforms of p38 MAPK measured by in vitro kinase assays was also reduced during differentiation of L6 cells into myotubes to 43 and 58% of the activities seen in myoblasts, respectively. However, these changes in activity were not due to a change in the expression level of p38 MAPK (Table II).
Insulin-induced MKK3/6 Phosphorylation Is Regulated during Myogenesis-The dual-specificity tyrosine and serine protein kinases MKK3 and MKK6 are upstream kinases of p38 MAPK family members activated by extracellular stresses and cytokines (43). MKK3 activates the p38 MAPK ␣, ␥, and ␦ isoforms, whereas MKK6 activates the p38 MAPK ␣, ␤, ␥, and ␦ isoforms (43). MKK3 and MKK6 are themselves phosphorylated by upstream kinases, and this event can be detected by immunoblotting cell lysates with an anti-phospho-MKK3/6 antibody (this antibody does not distinguish between MKK3 and MKK6, therefore the results are presented herein as phosphorylation of MKK3/6). We analyzed the effect of insulin on MKK3/6 phosphorylation to explore if this pathway operates in response to the hormone and if it shows preference in myotubes. Insulin was unable to stimulate MKK3/6 in myoblasts within the first 10 min after its addition (Fig. 5A). In contrast, treatment of myoblasts with hyperosmolar sodium chloride increased MKK3/6 phosphorylation by 5-to 10-fold (result not shown). In these same experiments, insulin caused Akt phosphorylation in myoblasts demonstrating the effectiveness of the hormone to activate its receptor (result not shown). In contrast to the results obtained with myoblasts, insulin rapidly induced phosphorylation of MKK3/6 in myotubes within 1 min (1.8 Ϯ 0.2-fold, p Ͻ 0.05), which declined back to basal level by 10 min (Fig. 5B). These results are consistent with the development of a p38 MAPK-activating pathway during the differentiation of L6 muscle cells (Fig. 4, B and D) and suggest the possibility that insulin may activate p38 MAPK via MKK3/6. Whether there is a causal relationship between MKK3/6 and specific p38 MAPK isoforms in response to insulin is the subject of separate examination. (40, 44 -47). In L6 myoblasts, overexpression of a dominant-negative mutant of Akt1 blocked insulin-stimulated GLUT4 translocation (40). This mutant also prevented the insulin-induced acceleration of the transit of GLUT4 through the recycling endosome (25). As shown in Fig.  1, the rates by which insulin increased GLUT4 translocation in myoblasts and myotubes were strikingly similar. This observation suggests that the translocation response of the GLUT4 system is fully developed at both stages of the myogenic process. To document this possibility we measured the time course of activation by insulin for Akt isoforms 1, 2, and 3 (protein kinase B ␣, ␤, and ␥). The magnitude and activation rate of each independent isoform was remarkably similar for myoblasts and myotubes (Fig. 6). By 5 min the activities of all three isoforms of Akt activity were stimulated by at least 70% of their maximum in either myoblasts or myotubes (Fig. 6). These observations are consistent with our earlier study where IRS1associated phosphatidylinositol 3-kinase activity was equally stimulated by insulin in myoblasts and myotubes (7.8-fold in myoblasts compared with 6.4-fold in myotubes (41)). Interestingly, Akt 1 and 3 were activated ϳ7-fold by insulin, whereas Akt2 was activated only about 2-fold above basal levels. The robust response to insulin of the phosphatidylinositol 3-kinase 3 Akt axis in myoblasts and myotubes, contrasts with the developmental increase in p38 MAPK activation during myogenesis.

Activation of Each Akt Isoform by Insulin Is Comparable in Myoblasts and Myotubes-A number of studies support a role for Akt in GLUT4 translocation
GLUT4 Translocation Has Greater Insulin Sensitivity as Compared with Glucose Uptake-The time delay observed in the glucose uptake response of the myotubes in the first 2 min after the addition of insulin (Fig. 1) raised the possibility that glucose uptake may be less sensitive to the hormone compared with GLUT4 translocation. Indeed, focusing on the range of concentrations of 10 nM and below (where insulin action is deemed to occur primarily through insulin receptors (48)) we found that glucose uptake was markedly less sensitive to insulin compared with GLUT4myc translocation (Fig. 7A). At higher insulin concentrations, both curves followed relatively similar behavior (Fig. 7A, inset).
The striking difference in insulin sensitivity of GLUT4 translocation and glucose uptake prompted an examination of the insulin sensitivities of enzyme activities thought to lead to each of these phenomena, respectively, Akt and p38 MAPK. The results in Fig. 7B illustrate that Akt phosphorylation is more sensitive to insulin compared with p38 MAPK phosphorylation. These results qualitatively link Akt with GLUT4 translocation and p38 MAPK with glucose uptake at low insulin doses. DISCUSSION The development of the L6 cell line that ectopically expresses high levels of GLUT4myc has allowed us to differentiate GLUT4 translocation from GLUT4 transport activity under diverse conditions. An important feature of this cell line is that glucose uptake in the basal and insulin-stimulated states in both myoblasts and myotubes is mediated by GLUT4myc (30,31). We found that insulin-stimulated GLUT4myc translocation precedes the stimulation of glucose uptake in time, suggesting that GLUT4myc may undergo an activation step following its insertion into the plasma membrane (28). Furthermore, we identified p38 MAPK as a signal that may be involved in this activation of GLUT4, because inhibitors of p38 MAPK reduce the insulin response of glucose uptake without affecting GLUT4 translocation (28,35). This link between p38 MAPK and glucose uptake has been extended by the inhibition of insulin-stimulated glucose uptake by two unrelated families of p38 MAPK inhibitors and to a dominant-negative p38 MAPK mutant expressed in 3T3-L1 adipocytes (36).
Here we further investigate the phenomenon of GLUT4 ac-tivation and find that insulin-stimulated GLUT4 activation and p38 MAPK stimulation develop upon differentiation of L6 muscle cells from myoblasts into myotubes. In myoblasts, insulin causes a minor stimulation of the kinase and produces a submaximal stimulation of glucose uptake (1.6-fold) despite full translocation of GLUT4 to the plasma membrane (2.4-fold).
FIG. 4. Insulin-stimulated p38 MAPK phosphorylation and activity are more robust in myotubes than in myoblasts. L6-GLUT4myc myoblasts (A) and myotubes (B) were treated with or without 100 nM insulin for the indicated time periods at 37°C. Cells were lysed in Laemmli sample buffer in preparation for SDS-PAGE. 50-g aliquots of total protein from each sample were immunoblotted with either a phospho-specific p38 MAPK antibody or an antibody that recognizes all four isoforms of p38 MAPK as described under "Experimental Procedures." A representative immunoblot is shown for either myoblasts or myotubes. The amount of phosphorylated p38 MAPK or p38 MAPK protein was quantified by densitometric scanning and analysis with NIH Image software. All values are expressed as the ratio of phospho-p38 to p38 protein with a value of 1 ascribed to unstimulated cells. Results are the mean Ϯ S.E. of four to five experiments. The asterisk and double asterisk in parts A and B represent a significant difference from the measurement at time zero, p Ͻ 0.01 and p Ͻ 0.001, respectively. L6-GLUT4myc myoblasts (C) and myotubes (D) were treated with or without 100 nM insulin for the indicated time periods at 37°C. Cells were lysed with 1% TX-100 in buffer containing phosphatase and protease inhibitors in preparation for immunoprecipitation of p38␣ or p38␤ MAPK from 500 g of total protein. Enzyme activities were determined by an in vitro kinase assay using GST-ATF-2 as substrate as described under "Experimental Procedures." Kinase reactions were prepared for SDS-PAGE and immunoblotted with anti-phospho-ATF-2 antibody. Shown are representative immunoblots from the in vitro kinase reactions. All immunoblots were quantified using densitometric scanning and NIH Image. Results represent the mean Ϯ S.E. of three to four experiments performed in duplicate.
In differentiated myotubes, insulin stimulates glucose uptake maximally and stimulates p38 MAPK robustly. Similarly, the putative upstream activators of p38 MAPK, MKK3/6, are phosphorylated in response to insulin in myotubes but not in myoblasts. In contrast to the maturation of these responses, GLUT4 translocation and stimulation of the phosphatidylinositol 3-kinase-Akt signaling axis, thought to be involved in signaling GLUT4 translocation by insulin, are similar in myoblasts and myotubes. In addition, the stimulation of GLUT4 translocation and Akt activation are more sensitive to insulin than the stimulation of glucose uptake and p38 MAPK. These results segregate GLUT4 translocation with stimulation of Akt on the one hand and GLUT4 activation and stimulation of p38 MAPK on the other.
Methodological Considerations in the Analysis of GLUT4 Activity-Insulin stimulates rapidly the V max of glucose uptake into muscle and fat cells without altering the K m of the transporter proteins for glucose (49,50). An accepted explanation of this finding is a rise in the number of glucose transporters at the cell surface (GLUT4 translocation). However, the magnitude of GLUT4 translocation frequently reported in the literature does not account for the greater magnitude in insulinstimulated glucose uptake, when GLUT4 levels are assessed in isolated membranes from muscle and adipose tissue (6,(13)(14)(15)(16)(17). The discrepancy remains when glucose uptake and GLUT4 content are measured in the same vesicular membrane preparation (12,18). These results suggest that activation of GLUT4 may occur in response to insulin.
On the other hand, studies using the ATB-BMPA glucose transporter photolabel have supported the theory that GLUT4 translocation is the only mechanism by which insulin increases glucose uptake, because in many cases the magnitude of insulin-stimulated translocation approximates the increase in glucose uptake (51)(52)(53)(54). However, early studies proposed that the ATB-BMPA interaction with glucose transporters reflects not only their availability at the cell surface but also their state of activity, because the label binds to the glucose-binding site on the transporters (55,56). A similar argument has been made in cells where ATB-BMPA binding to GLUT4 or GLUT1 increases without net gain in surface transporters assessed by an alternate method of biotinylation of the transporters at the cell surface (56 -58). Moreover, changes in the K i of inhibition of glucose uptake by ATB-BMPA were noted upon treatment of brown adipocytes with norepinephrine (16). Lastly, early on during the characterization of the photolabel it was reported that cytochalasin B, a potent non-competitive inhibitor of glucose uptake that binds to the inward-facing glucose-binding site of glucose transporters, precludes exofacial binding of ATB-BMPA to the glucose transporter (55). The ensuing explanation was that the outward-facing and inward-facing glucosebinding sites are mutually exclusive. Potentially, a cytosolic protein or a post-translational modification on the glucose transporter could lock the glucose-binding site of the glucose transporter in an inward-facing position, preventing it from bringing glucose into the cell. We surmise that an insulin signal could conceivably relieve this constraint. Based on the time course of insulin-dependent GLUT4 translocation and L6-GLUT4myc differentiation Phosphorylated p38 MAPK levels were determined in cell lysates prepared from untreated myoblasts or myotubes by immunoblotting for phospho-p38 MAPK as described under "Experimental Procedures." p38 MAPK ␣ or ␤ were also immunoprecipitated from 500 g of myoblast or myotube cell lysates, and basal p38 MAPK activities were determined by an in vitro kinase assay using GST-ATF-2 as substrate as described under "Experimental Procedures." To assess changes in p38 MAPK expression, equal amounts of myoblast or myotube cell lysates were also immunoblotted with an antibody that recognizes all four isoforms of p38 MAPK, an antibody specific to p38 MAPK␣, and an antibody specific to p38 MAPK␤. All immunoblots were quantified using densitometric scanning and NIH Image software. The results were calculated as the ratio of the level seen in myotubes versus myoblasts, expressed as percent. Results represent the mean Ϯ SE of three to four experiments performed in duplicate. FIG. 5. MKK3/6 is phosphorylated in response to insulin only in myotubes. L6-GLUT4myc myoblasts (A) and myotubes (B) were treated with or without insulin (100 nM) for the indicated times, and cells were lysed using Laemmli sample buffer in preparation for SDS-PAGE. Fifty micrograms of total protein from each sample was immunoblotted with an antibody that recognizes equally the phosphorylated forms of MKK3 and MKK6 as described under "Experimental Procedures." The amounts of phosphorylated MKK3/6 were quantified using densitometric scanning and analysis with NIH Image software. Results are the mean Ϯ S.E. of four to six experiments for each myoblasts and myotubes and are expressed relative to the value in the corresponding unstimulated cells. The asterisks in B represent a significant difference from the measurement at time zero, p Ͻ 0.05. stimulation of glucose uptake, this would occur subsequent to GLUT4 insertion into the plasma membrane and would require input from insulin-stimulated p38 MAPK activity.
Correlation between p38 MAPK and Glucose Uptake-The hypothesis arises that the intrinsic activity of glucose transporters can be regulated by the cell in response to insulin, and documenting it requires methodological approaches that can compare glucose uptake and gain in surface GLUT4 in equivalent cellular preparations. In the present study, we find that inhibition of p38 MAPK with SB203580 lowers the insulin response of glucose uptake in myotubes to match the rate observed in myoblasts. This is consistent with our previous demonstration that SB203580 reduces the V max of insulinstimulated glucose uptake but not the K m in L6 myotubes (28). However, the time lag between GLUT4 translocation and glucose uptake in particular to myotubes is unaffected by inhibitors of p38 MAPK (Fig. 3), suggesting that additional mechanisms define the response to the hormone in mature cells. It is possible that this lag results from a delay in either p38 MAPK activation and/or in the removal of an inhibitor or the recruitment of an activating protein shortly after GLUT4 is inserted into the plasma membrane.
Supporting the concept that activation of GLUT4 occurs in addition to its translocation in response to insulin, these two processes also differ in their sensitivity to insulin. GLUT4 translocation occurs at slightly lower concentrations of insulin than are required to stimulate glucose uptake (Fig. 7). Both responses occur within the range of signaling through the insulin receptor. Similarly, activation of Akt/protein kinase B is more insulin-sensitive compared with p38 MAPK. We hypothesize that at some point along the insulin signaling pathway, the signals leading to the activation of Akt and GLUT4 translocation segregate from those leading to p38 MAPK activation and stimulation of glucose uptake. Indeed, activation of p38 MAPK by insulin is not down-stream of phosphatidylinosi-FIG. 6. Insulin stimulates the activities of Akt1, Akt2, and Akt3 in both myoblasts and myotubes. L6-GLUT4myc myoblasts (A) and myotubes (B) were treated with or without insulin (100 nM) for the indicated times prior to preparation of cell lysates with 1% TX-100 in buffer containing phosphatase and protease inhibitors in preparation for immunoprecipitation. Each Akt isoform was immunoprecipitated from 500 g of myoblast or myotube cell lysates using isoform-specific antibodies. Akt activities were determined in vitro by phosphorylation of Crosstide peptide substrate as described under "Experimental Procedures." Results of Akt activities are the mean Ϯ S.E. of four to six experiments in each of myoblasts or myotubes. The immunoblots were quantified using densitometric scanning and analysis with NIH Image software. The results in the main panels are expressed as the percentage of the maximal response to facilitate comparisons among the different parameters measured (the fold change in Akt phosphorylation is several times higher than the fold change in p38 MAPK phosphorylation: maximal stimulations of 11.0 Ϯ 1.8-fold and 3.3 Ϯ 0.5-fold, respectively). For clarity, only the responses to 10 nM insulin and below were illustrated. The inset in A depicts the -fold increase in insulin-mediated GLUT4myc translocation (closed squares) and 2-deoxyglucose uptake (open circles) after 20 min of exposure to a full range of insulin concentrations (up to 100 nM), expressed as -fold stimulation above basal. Results are the mean Ϯ S.E. of three to four experiments. tol 3-kinase (34). Conceivably, steps downstream of Akt and p38 MAPK or the GLUT4 molecules itself could be the site of integration of these separate signals, resulting in insulin-stimulated glucose uptake. The points of divergence and convergence are the subject of separate investigation.
Maturation of the p38 MAPK and Glucose Uptake Responses during Myogenesis-The lower stimulation of glucose uptake by insulin despite full insulin-stimulated GLUT4 translocation in myoblasts suggests that GLUT4 activation does not occur at this stage of myogenesis. Moreover, as stated above, SB203580 treatment of L6 myotubes brings down the glucose uptake response of L6 myotubes to the level of the response in myoblasts. This difference between myoblasts and myotubes correlates with the ability of insulin to phosphorylate MKK3/6 and activate p38 MAPK and with the acquisition of susceptibility to inhibitors of p38 MAPK given prior to insulin action. Similarly, p38 MAPK was activated in 3T3-L1 adipocytes (36) but not in undifferentiated 3T3-L1 fibroblasts. 2 One possible reason for the increased regulation of p38 MAPK by insulin could be the increase in insulin receptor density as L6 and 3T3-L1 cells differentiate (59,60). However, this is unlikely because activation of Akt isoforms in myoblasts is comparable to that in myotubes. Alternatively, a differentiation-dependent change in the expression of a putative p38 MAPK-regulatory protein could be the explanation. Finally, the proximity of p38 MAPK activators and p38 MAPK may be favored in myotubes. Future studies will be required to examine this possibility.
Inhibitors of p38 MAPK (the pyridinylimidazoles SB203580 and SB202190 and the aza-azulenes A291077 and A304000) reduce insulin-stimulated glucose uptake without affecting GLUT4 translocation in L6 myotubes (28,36). It is important to reiterate that the p38 MAPK inhibitors cannot directly inhibit GLUT4 glucose transporters, because they have no effect on glucose uptake when added only to the transport assay (28,35,36). In addition, basal glucose uptake, which is mediated entirely by GLUT4 in these cells (30,31), is not inhibited by preincubation with the p38 MAPK inhibitors (28,35,36). The connection between p38 MAPK and GLUT4 activation is strengthened by the similar IC 50 value of pyridinylimidazoles to reduce insulin-stimulated glucose uptake and inhibit p38 MAPK activity (28). Moreover, inducible expression of a dominant-negative mutant of p38 MAPK resulted in lower insulinstimulated glucose uptake in 3T3-L1 adipocytes (36). Thus, p38 MAPK appears to contribute to insulin-stimulated glucose uptake not only in myotubes but also in 3T3-L1 adipocytes and in isolated white and brown adipocytes and skeletal muscles (28,35,61,62). It is plausible that a lower GLUT4 activation contributes to the insulin resistance of glucose uptake in peripheral tissues that precedes or accompanies type 2 diabetes. Hence, a full understanding of the mechanism by which insulin regulates GLUT4 activation could give rise to the possibility of new therapeutic interventions to improve glucose utilization.
In conclusion, the regulation of the activity of glucose uptake by p38 MAPK in L6-GLUT4myc cells develops with differentiation of the cell from myoblasts into myotubes. Inhibiting p38 MAPK reduces the stimulation of glucose uptake in myotubes to the level observed in myoblasts. In contrast, GLUT4 translocation is equally developed in myoblasts as in myotubes. In myotubes, GLUT4 translocation and Akt stimulation had higher insulin sensitivities than stimulation of glucose uptake and p38 MAPK phosphorylation. These observations draw important parallels between Akt and GLUT4 translocation and between p38 MAPK and GLUT4 activation.