Mitogen-activated Protein Kinase Activation Is Not Sufficient for Stimulation of Glucose Transport or Glycogen Synthase in 3T3-Ll Adipocytes”

The role of mitogen-activated protein (MAP) kinase in the regulation of glucose metabolism has been inves-tigated by comparing the effects of insulin and epider- mal growth factor (EGF) on MAP kinase activation, glucose transport, and glycogen synthase in 3T3-Ll adipocytes. Insulin or EGF treatment for 5 min increased p42mapk and p44mapk activity to the same extent as determined by myelin basic protein kinase activity measurements and phosphotyrosine immuno- blotting. The profiles of myelin basic protein kinase activity following MonoQ chromatography of extracts obtained from cells incubated with insulin or EGF were almost identical. Insulin increased glucose transport and GLUT4 translocation to the cell surface by 15-and y-fold, respectively. EGF had no significant effect on these processes. Insulin increased the glycogen syn- thase ratio (-Glc-6-P/+Glc-6-P) by 7.5- and 3.5-fold in the presence and absence of glucose, respectively. EGF increased the ratios by only 2- and 1.3-fold, respectively. EGF did not appear to inhibit downstream of MAP kinase, because when adipocytes were incubated with insulin plus EGF, the stimulation of glucose transport and glycogen synthase was similar to that observed with insulin alone. These findings indicate that activation of the MAP kinase

The major physiological role of insulin is to regulate fuel metabolism in muscle, liver, and adipose tissue. Insulin stimulates a number of processes in these tissues that are central to the deposition of glucose into either glycogen or lipid. These processes include glucose transport, glycogen synthase activity, and acetyl-coA carboxylase activity. One of the paradoxes surrounding the regulation of metabolism by insulin is that whereas the tyrosine kinase activity of the insulin receptor is activated by insulin-dependent phosphorylation, stimulation of glycogen synthase and acetyl-coA carboxylase is mediated by insulin-dependent dephosphorylation (for review, see Ref. 1). A potential resolution to this paradox was provided by the demonstration that protein phosphatase 1 * This work was supported by National Institutes of Health Grants DK42503 and DK28312 and the Washington University Diabetes Research and Training Center. The costa 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.
II To whom correspondence should be addressed Centre for Molecular Biology and Biotechnology, University of Queensland, Brisbane 4072,  (PPl),' the major phosphatase that dephosphorylates glycogen synthase in muscle cells, is activated by phosphorylation (2). An insulin-stimulated S6 kinase I1 homolog, referred to as insulin-stimulated protein kinase, which phosphorylates the regulatory subunit of PP1 at a unique serine, has been purified from skeletal muscle (2,3). PP1 phosphorylation at this site is increased following stimulation with insulin in vivo (2). A potential link between the activation of PP1 and the tyrosine kinase signaling pathway has been provided by the observation that mitogen-activated protein (MAP) kinase activates insulin-stimulated protein kinase in vitro (3). Hence, these data imply that activation of MAP kinase leads to the activation of glycogen synthase.
Insulin as well as other polypeptide growth factors such as epidermal growth factor (EGF), platelet-derived growth factor, nerve growth factor, and fibroblast growth factor stimulate MAP kinase activity (for review, see Refs. 4

and 5).
Therefore, by examining the effects of different growth factors on MAP kinase activity and glucose metabolism in an insulinsensitive cell line, it should be possible to test the hypothesis that MAP kinase is central to the regulation of glucose metabolism. 3T3-Ll adipocytes express high levels of insulin and EGF receptors (6) and are highly responsive to insulin in terms of glucose transport, glycogen synthesis, and lipid synthesis. In the present studies we have uncoupled the activation of MAP kinase from the stimulation of glucose transport and glycogen synthase activity in these cells.

EXPERIMENTAL PROCEDURES
MateriaZ~-[y-~*P]ATP and the enhanced chemiluminescence antibody detection system were obtained from Amersham Corp. EGF was kindly provided by Dr. Linda Pike (Department of Biochemistry, Washington University). Fluorescein isothiocyanate-conjugated goat anti-rabbit was obtained from Cappel Laboratories (West Chester, PA). 3T3-Ll fibroblasts were obtained from the American Type Culture Collection (Rockville, MD). All other chemicals were obtained from Sigma.
Antibodies-The rabbit anti-glucose transporter antibodies (R820 and R493) that were used for these studies have been described previously (7). The mouse anti-phosphotyrosine antibody (4G10) was kindly provided by Dr. Thomas Roberts (Dana Farber Institute, Boston, MA). The chicken anti-MAP kinase antibody was raised against the C-terminal 14 amino acids of p44mapk. The rabbit anti-MAP kinase antibody (TR2) was raised against p42mapk and was kindly provided by Dr. Michael Weber (University of Virginia, Charlottesville, VA).
Cell Culture and Incubations-Murine-derived 3T3-Ll fibroblasts were cultured in Dulbecco's modified Eagle's medium on either tissue The abbreviations used are: PP1, protein phosphatase 1; MBP, myelin basic protein; MAP, mitogen-activated protein; EGF, epidermal growth factor; PAGE, polyacrylamide gel electrophoresis; Glc-6-P, glucose 6-phosphate; KRP, Krebs-Ringer phosphate buffer containing 2% bovine serum albumin and 2.5 mM glucose. culture dishes (10 cm or 35 mm) or ethanol-washed glass coverslips (no. 1) and differentiated as described previously (7). All experiments were performed on adipocytes 8-12 days after withdrawal from differentiation media. Adipocytes were rinsed in Krebs-Ringer phosphate buffer containing 2% bovine serum albumin and 2.5 mM glucose (KRP) and incubated in this buffer for 2 h at 37 "C, as described previously (8). In some experiments, where 2-deo~y-[~H]glucose uptake and glycogen synthase activity were measured, glucose was omitted from the KRP. Cells were then incubated for the indicated times at 37 "C in KRP containing either no additions (control), insulin M), or EGF M). Phosphotyrosine Immunoblotting-3T3-Ll adipocytes grown on 35-mm dishes were washed 2 times with 2 ml of ice-cold buffer A (66 mM Tris, pH 7.4, 1 mM sodium orthovanadate) and scraped into 0.5 ml of lysis buffer (buffer A containing 2% SDS, prewarmed to 88 OC). Lysates were boiled for 10 min, homogenized by 10 passes through a 25 G needle, and centrifuged at 10,OOO X g for 15 min. Supernatants (150 pg of protein) were subjected to SDS-PAGE and immunoblotted with the mouse monoclonal anti-phosphotyrosine antibody, 4G10. The antibody was detected using alkaline phosphatase-conjugated anti-mouse I&, followed by colorimetric detection.
Myelin Basic Protein (MBP) Kinase Assays-Cell extracts were prepared as described previously (9). Briefly, cells grown on 35-mm dishes were incubated under the appropriate conditions, washed with 3 ml of 0.15 M NaCl, and scraped from the dish into homogenization buffer (20 mM HEPES pH 7, 20 mM EGTA, 15 mM MgOAc, 1 mM dithiothreitol, 40 mM p-nitrophenyl phosphate, 0.1 mM phenylmethylsulfonyl fluoride) using a rubber policeman. Cells were homogenized at 4 "C in a 2-ml glass homogenizer with 25 strokes of a rotating pestle, followed by centrifugation at 30,000 X g for 5 min. All of the above procedures were performed at 4 "C. MBP kinase activity was measured as described previously (10). Supernatants (1 pl, approximately 5 pg of protein) were combined with 19 pl of buffer B (50 mM @-glycerophosphate, pH 7.3,7 mM NaF, 0.3 mM EDTA, 15 mM MgClz, 2 mM dithiothreitol, and 4.4 mM protein kinase inhibitor peptide (11)) containing 0.33 mg/ml MBP. The kinase reaction was initiated by the addition of 5 pl of 1 mM [y3'P]ATP (700 mCi/mmol). Following a 10-min incubation at 30 "C, reactions were terminated by spotting 10 pl of the reaction mixtures onto p81 phosphocellulose papers, which were immediately immersed in 0.85% phosphoric acid. The papers were washed 5 times for 15 min in 400 ml of phosphoric acid with stirring and once for 5 min in 95% ethanol. Papers were dried and "P was quantitated by scintillation counting.
MonoQ Chromatography-For each condition, 3 X 10-cm dishes of 3T3-Ll adipocytes were washed in 10 ml of 0.15 M NaCl and scraped into MonoQ buffer (50 mM @-glycerophosphate, pH 7.3,l mM EDTA, 1 mM dithiothreitol, 0.15 mM sodium orthovanadate) using a rubber policeman. Cells were homogenized and centrifuged at 200,000 X g for 1 h. All procedures were performed at 4 "C. Supernatants were filtered through a 0.2-pm syringe filter and loaded onto a MonoQ column (Pharmacia LKB Biotechnology Inc.), which was equilibrated in MonoQ buffer. During chromatography, the flow rate was maintained at 1 ml/min, and fractions were collected each min. By varying the concentrations of MonoQ buffer and MonoQ buffer plus 800 mM NaCl, the pumps were programed to deliver the following concentrations of NaCI: 0 mM for 10 min, 0-24 mM in 1 min, 24-200 mM in 34 min, 200-800 mM in 5 min, and 800 mM for 10 min. Fractions (10 pl) were assayed for MBP kinase activity as described above. In addition, fractions (250 pl) were precipitated with trichloroacetic acid, subjected to SDS-PAGE, and transferred to nitrocellulose. Nitrocellulose blots were incubated with a chicken antibody to MAP kinase, followed by horseradish peroxidase-conjugated anti-chicken IgG.
p42mapk Immunoprecipitation and MBP Immunoassay-Dishes (10 cm) of 3T3-Ll adipocytes were washed 2 times with buffer C (135 mM NaCl, 5.4 mM KCl, and 10 mM sodium phosphate, pH 7.4) and lysed in 0.8 ml of lysis buffer (50 mM HEPES pH 7.5, 100 mM NaCl, 2 mM EDTA, 1% Nonidet P-40, 0.1 mM sodium orthovanadate, 40 mM p-nitrophenyl phosphate, 0.1 mM phenylmethylsulfonyl fluoride). Lysates were cleared by centrifugation at 10,000 X g for 10 min. Protein A conjugated to agarose beads, which had been precoupled with 10 pl of anti-p42mapk antiserum: was added to 250 pl of cleared lysate and incubated for 2 h at 4 "C. The beads were collected by centrifugation in a microcentrifuge for 20 s, the supernatant was removed, and the beads were washed 2 times with lysis buffer and 2 ' The anti-p42mapk polyclonal antibody immunoprecipitates only the p42mapk MAP kinase isoform (M. Weber, personal communication). times with 10 mM HEPES pH 7.5, 10 mM MgOAc. The beads were then resuspended in 20 pl of substrate (MBP, 2 mg/ml), and the kinase assay was initiated by the addition of 20 pl of 200 p~ [-y-"P] ATP (25 mCi/mmol), 40 mM MgOAc, 40 mM HEPES pH 7.5. The reaction was incubated for 30 min at 30 "C and was stopped by the addition of 80 pl of 2 X Laemmli sample buffer (12). Samples were subjected to SDS-PAGE using a 15% resolving gel. The gel was dried and subjected to autoradiography, and the "P-labeled MBP was quantified.
Glycogen Synthose Assays-Dishes (35 mm) of 3T3-Ll adipocytes were washed once with buffer C (3 ml) at room temperature, scraped with a rubber policeman into ice-cold 100 mM KF, 10 mM EDTA, pH 7 (250 pl), and homogenized by 6 passes through a 22 G needle on ice. Homogenates were centrifuged at 10,000 x g for 20 min, and supernatants were assayed for glycogen synthase activity as described by Thomas et al. (13).
Other Assays-2-Deoxy-[3H]glucose uptake was measured as describedpreviously (14). Briefly, dishes (35 mm) of 3T3-Ll adipocytes were incubated under the appropriate conditions in 950 pl of KRP buffer at 37 "C. The assay was initiated by the addition of 50 p1 of 1 mM 2-deo~y-[~H]glucose (20 pCi/mmol). After 3 min, the assay was terminated by washing the dishes 3 times in ice-cold buffer C. The cells were solubilized in 1% Triton X-100, and the 3H was quantitated by scintillation counting. Levels of GLUT4 and GLUT1 at the plasma membrane were determined by the plasma membrane lawn assay as previously described (8). Fluid phase endocytosis was measured using horseradish peroxidase as a marker, as described previously (15). Protein assays were performed using the bicinchoninic acid reagent (Sigma).

Effects of Insulin and EGF on MAP Kinase Actiuatwn-
Insulin and EGF stimulated MBP kinase activity in 3T3-Ll adipocytes by 4.0-and 3.7-fold, respectively, following a 5min incubation with each ligand (Fig. 1). As shown previously

Involvement of MAP Kinases in Metabolism
slightly more rapidly than insulin-stimulated activity (Fig. 1). The effects of EGF and insulin on MBP kinase activity were not additive when the two agonists were incubated together for 5 or 15 min. In fact, the MBP kinase activity observed in the presence of insulin plus EGF was not significantly different from that observed with either agonist alone (Fig. 1). Two MAP kinase isoforms have been shown to be activated by insulin, p42mapk and p44mapk (4,5). The activation of these kinases involves phosphorylation on both threonine and tyrosine residues (4,5). To determine whether both p42mapk and p44mapk were activated in 3T3-Ll adipocytes, the phosphotyrosine content of these proteins was analyzed before and after treating cells with insulin or EGF for 5 min. The 5min time point was chosen for this experiment, as well as the experiments described below, because both insulin and EGF activated MBP kinase activity to the same extent at this time point. As shown in Fig. 2, both insulin and EGF increased the phosphotyrosyl content of two proteins of approximate M , = 44,000 and 42,000 to the same extent. Previous experiments have shown that the major M, = 42,000 protein, which is tyrosine-phosphorylated in 3T3-Ll adipocytes, is p42mapk (17). In addition, the 44-and 42-kDa tyrosine phosphorylated proteins comigrated with bands that were recognized by an anti-MAP kinase antibody (Fig. 2). These results are consistent with the assignment of these proteins as p44mapk and p42mapk, respectively. A major phosphotyrosine protein of approximate M, = 45 is phosphorylated in control cells as well as in cells treated with insulin and EGF, but this protein does not represent a MAP kinase isoform. p44mapk runs as a tight band of slightly lower apparent M, than this constitutively phosphorylated protein (Fig. 2).
To verify that the enzymatic activities of p42mapk and p44mapk were increased under these conditions, the effects of insulin and EGF on MBP phosphorylation were investigated after resolution of the two isoforms by ion-exchange chromatography. As shown in Fig. 3, the MBP kinase profiles of the columns from insulin and EGF-treated cells were almost identical. Two major peaks of activity were detected with both agonists. To determine the elution profiles of p42mapk and p44mapk, column fractions were immunoblotted with the MAP kinase antibody described in Fig. 2. Both p42mapk and p44mapk were detected in fractions 24-26, whereas only p44mapk was detected in fractions 26-29 and fractions 40-42 (data not shown). Because the peak of activity corresponding to fractions 25-28 contained both MAP kinase isoforms, it was unclear whether the activity detected in this fraction was due to activation of p42mapk, p44mapk, or both  isoforms. Therefore, to determine whether p42mapk was similarly activated by insulin and EGF, fraction 26 was immunoprecipitated with an antibody that only immunoprecipitates this isoform.' MBP kinase activity was then measured in the immunoprecipitate. Insulin and EGF activated p42mapk in this fraction by 8.2-and 6.9-fold, respectively. In contrast no significant MBP kinase activity was detected in anti-p42mapk immunoprecipitates from fractions 40-42 (data not shown).

Effects of Insulin and EGF on Glucose Transport-2-Deoxy-
[3H]glucose uptake was increased by 10.3-and 14.8-fold in adipocytes treated with insulin for 5 or 30 min, respectively (Fig. 4). In contrast, when cells were treated with EGF, there was no significant increase in 2-deo~y- [~H]glucose uptake at any of the time points analyzed, when compared with 2-deoxy-[3H]glucose measurements from untreated cells (Fig. 4). In-cubation of adipocytes with insulin and EGF together for 15 min resulted in an increase in 2-deo~y-[~H]glucose uptake that was not significantly different from that observed with insulin alone for 15 min. Furthermore, when EGF was added 15 min before a subsequent 15-min incubation in the presence of insulin and EGF, 2-deo~y-[~H]glucose uptake was not significantly different from that observed when insulin was added alone for 15 min (data not shown).
Insulin stimulates glucose transport in 3T3-Ll adipocytes by causing the translocation of two separate glucose transporter isoforms, GLUT4 and GLUT1, to the plasma membrane (7,18). Therefore, cell surface levels of both isoforms were measured after treating cells with insulin or EGF. Insulin stimulated the translocation of GLUT4 to the cell surface by 6.8-fold at 20 min (Fig. 5 ) . In contrast, EGF did not significantly affect GLUT4 translocation at any of the times indicated (Fig. 5). When insulin and EGF were added together, the levels of cell surface GLUT4 were not significantly different from the levels when insulin was added alone (Fig. 5 ) . Consistent with previous studies (7, 18), insulin stimulated GLUTl translocation by 2.5 & 0.2-fold after 20 min. EGF also significantly stimulated GLUTl translocation but by only 1.4 f 0.2-fold. Therefore, the inability of EGF to stimulate glucose uptake correlates well with the lack of an effect of EGF on GLUT4 translocation as well as the relatively small increase in GLUTl translocation observed with this growth factor.

Effects of Insulin and EGF on Glycogen Synthuse Activity-
Glucose has previously been shown to potentiate insulinstimulated glycogen synthase activity in adipocytes, due to the conversion of glucose to glucose 6-phosphate, which causes dephosphorylation and activation of the enzyme (19). Therefore, incubations were performed in either the presence or absence of glucose (Fig. 6). In the presence of glucose, insulin stimulated glycogen synthase activity by 3.3-fold after 5 min and 7.8-fold after 20 min (Fig. 6A). In contrast, EGF stimulated glycogen synthase activity by only 2-fold after 5 or 20 min of treatment (Fig. 6A). In the absence of glucose, insulin  Glycogen synthase activity was measured in the presence and absence of glucose 6-phosphate (G6P), as described under "Experimental Procedures." Total activities (+G6P) or activity ratios (-G6P/+G6P) are presented in the upper or lower panels, respectively. The results represent the mean f S.E. of three experiments. B, glycogen synthase activity from glucose-starved cells. Adipocytes were incubated in the absence of glucose for 2 h as described under "Experimental Procedures." Cells were then stimulated with insulin, EGF, or EGF and insulin together as described in A but in the continued absence of glucose. The results represent the mean f S.E. from three experiments. stimulated glycogen synthase activity by 2-fold after 5 min and 3.5-fold after 20 min, whereas EGF stimulated by only 1.3-fold at either 5 or 20 min (Fig. 6B). EGF did not significantly affect insulin-stimulated glycogen synthase activity in either the presence or absence of glucose (Fig. 6, A and B). Furthermore, when EGF was added 20 min before a subse-

Involvement of MAP Kinases in Metabolism
quent 20-min incubation in the presence of insulin and EGF, the glycogen synthase activity measured was not significantly different from that observed when insulin was added alone for 20 min (data not shown).
Effects of Insulin and EGF on Fluid Phase Endocytosis-In contrast to the metabolic effects described above, fluid phase endocytosis is a more universal effect of growth factors that is observed in many quiescent cells (20). As shown previously (21), insulin stimulates fluid phase endocytosis in 3T3-Ll adipocytes by about 2-fold. To determine whether EGF could also stimulate this process, horseradish peroxidase uptake was measured in cells treated with EGF for 20 min. As shown in Fig. 7, EGF caused a 1.4-fold increase ( p < 0.05) in fluid phase endocytosis under conditions where insulin stimulated this process by 1.8-fold.

DISCUSSION
In this study, evidence is presented that questions the role of MAP kinases in the insulin regulation of glucose metabolism in 3T3-Ll adipocytes. We have shown that a growth factor, which stimulates MAP kinase activation to the same extent as insulin, does not mimic the actions of insulin on glucose transport and glycogen synthase activity. These data indicate that activation of MAP kinase per se is not sufficient for the stimulation of these metabolic processes in adipocytes.
The rationale for the present studies was that if MAP kinase activation mediates insulin-stimulated glucose metabolism in adipocytes, then any growth factor that stimulates MAP kinase activity in these cells should have insulinomimetic effects. A significant stimulation of MAP kinase activity was observed with either insulin or EGF. There were slight differences in the time course of MAP kinase activation, but the magnitude of stimulation was virtually identical at 5 min after addition of either ligand to the cells (Fig. 1). Furthermore, the effects of insulin and EGF on the 2 major MAP kinase isoforms, p42mapk and p44mapk, were almost indistinguishable (Fig. 3). The divergence between the stimulation of MAP kinase activity and the regulation of glucose metabolism was most striking at the 5-min time point. Whereas insulin stimulated glucose transport and glycogen synthase activity by approximately 10and 2-4-fold, respectively, EGF did not significantly stimulate glucose transport and caused only a marginal stimulation of glycogen synthase (1-%fold) at this time point. We have considered the possibility that EGF could be inhibiting downstream of MAP kinases in the transduction pathways leading to activation of glucose transport and glycogen synthase. This seems unlikely because EGF did not significantly inhibit insulin-stimulated glucose transport or glycogen synthase activity.
We also considered the possibility that fibroblasts that contaminate 3T3-Ll adipocyte cultures could contribute to the EGF-stimulated MAP kinase activation, whereas the adipocytes may be responsible for the insulin-stimulated MAP kinase activation. This seems unlikely because we routinely observe greater than 90% well differentiated 3T3-Ll adipocytes in our cultures. Furthermore, we have studied the effects of EGF in primary cultures of rat adipocytes. Collagenase digestion of rat epididymal fat pads results in a uniform population of adipocytes that are not contaminated with other cell types (22). In agreement with our studies in 3T3-Ll cells, insulin and EGF activated p42mapk and p44mapk to the same extent after a 5-min incubation with each ligand, as measured by both an increase in the phosphotyrosine content of these proteins and MBP kinase assays from extracts immunoprecipitated with an antibody to p42mapk (data not shown). These data clearly indicate that EGF regulates MAP kinase activity in adipocytes.
The observation that EGF does not stimulate glucose transport in 3T3-Ll adipocytes is not surprising in view of recent studies indicating that EGF has little effect on glucose transport in rat adipocytes (23,24). However, the relative inability of EGF to stimulate glycogen synthase activity was unexpected based on two previous reports (2,25). Cohen and coworkers (2) have suggested that a pathway exists in skeletal muscle linking the MAP kinases to glycogen synthase activation. Although our data are inconsistent with this model, we cannot exclude the possibility that glycogen synthase is differentially regulated in muscle and fat. In addition, Chan et al. (25) showed that EGF stimulates glycogen synthase activity to the same extent as insulin in Swiss 3T3 cells. The most likely explanation to account for the difference in glycogen synthase regulation between the present studies and these previous reports is that there is differential regulation in different cell types. For example, in Swiss 3T3 cells, growth factor-mediated signaling may occur by a pathway that is common to all growth factors, including insulin and EGF, whereas in adipocytes, these pathways may be different.
There are two general models that could explain our observations. The first model is that MAP kinase activation may be necessary but not sufficient for the activation of glucose transport and glycogen synthase. In other words, additional signaling molecules that are not recruited by EGF may be required in concert with MAP kinase. The cellular location of effector molecules may determine the biological specificity of growth factors. For example, MAP kinase has been shown to translocate to the nucleus under certain conditions (26). Perhaps MAP kinases can also translocate to other specialized intracellular locations. In this regard, when comparing insulin signaling with that of other growth factors such as EGF, the most striking difference is the insulin-stimulated recruitment of the insulin receptor substrate-1 (27, 28). Thus, it is conceivable that insulin receptor substrate-l can confer specificity in the targeting of signaling molecules within the cell.
The second model is that MAP kinase may neither be necessary nor sufficient for glucose transport and glycogen synthase activation. If this is the case, a signaling pathway

Involvement of MAP Kinases in Metabolism 26427
must exist for insulin-stimulated glucose transport and glycogen synthase activation that is distinct from the MAP kinase pathway. In addition to MAP kinase, insulin has also been shown to regulate the activity of a number of other signaling molecules. These include the serine/threonine kinase, Raf, the small M, GTP-binding protein, Ras, the "low K," cyclic nucleotide phosphodiesterase (29), and phospha-tidylinositol3'-kinase. The correlative approach that we have adopted may prove to be a useful way of identifying signaling molecules that are specifically involved in the regulation of glucose metabolism in insulin-sensitive cell types.