skeletal muscle is the main site of postprandial glucose disposal and, thus, plays an important role in glucose homeostasis (5). Several lines of evidence indicate that the amount of GLUT-4 glucose transporter protein in skeletal muscle is an important determinant of maximal glucose transport (9, 11, 15, 16, 18, 24, 40, 42). These studies demonstrated that the level of expression of GLUT-4 in muscle is important for the regulation of glucose homeostasis and that pharmacological intervention of muscle GLUT-4 expression might be a method to treat disease states such as diabetes (52).

Physical activity has been a primary therapeutic regimen used to treat diabetes and insulin-resistance-associated diseases. Our laboratory and others have shown that exercise increases the amount of GLUT-4 mRNA and protein in skeletal muscle (4, 8, 10, 30). Designing pharmacological interventions that mimic exercise-induced GLUT-4 expression requires a fundamental knowledge of the mechanism that mediates this effect. There is considerable evidence to suggest that this mechanism involves the activation of 5′-AMP activated protein kinase (AMPK). AMPK is activated in response to exercise and electrical stimulation (17, 45, 49). Furthermore, AMPK activity has been found to be dependent on two metabolic ratios, AMP/ATP and creatine/phosphocreatine, both of which increase during exercise (50). Artificially altering either ratio in intact muscle has resulted in a predictable effect on GLUT-4 protein levels. Chronic treatment with 5′-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR), an analog of adenosine, leads to high intramuscular levels of the AMP analog 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranosyl-5′-monophosphophate (ZMP), increased AMPK activity, and increased GLUT-4 protein levels (13, 14, 27, 31). In a similar fashion, chronic treatment of rats with β-guanidinopropionic acid (an analog of creatine) leads to creatine phosphate depletion and increased GLUT-4 expression in muscle (35). How, then, does increased AMPK activity lead to increased GLUT-4 protein levels?

In addition to the well-characterized short-term regulatory role in glucose and fatty acid metabolism (12, 21, 50), there is growing evidence that AMPK is involved in regulating gene transcription. The yeast homolog to AMPK, SNF1, plays an important role in regulating glucose-inducible genes (3, 28). In mammalian models, it has been shown that the α₂ catalytic subunit of AMPK is preferentially targeted to the nucleus (38) and that AMPK inhibits the glucose-induced transcription of l-pyruvate kinase, spot 14, and fatty acid synthase genes (7, 23). We hypothesized that the activation of AMPK may lead to increased GLUT-4 protein levels by stimulating the transcription of the GLUT-4 gene.

To identify the regulatory elements in GLUT-4 proximal promoter, previous studies showed that myocyte enhancer factor-2 (MEF-2) binding site, located from −473 to −464 on human GLUT-4 proximal promoter, is necessary but not sufficient for human GLUT-4 promoter transcriptional activity (41). A second regulatory element (domain I, from −742 to −712 on human GLUT-4 proximal promoter) has been identified, which functions cooperatively with MEF-2 domain in regulating human GLUT-4 transcription (33). Human GLUT-4 promoter transgenic studies have shown that a deletion of either the 30-bp domain I or mutation of the MEF-2 binding domain, within the context of the 895-hG4-CAT parental promoter construct, completely abolishes transgenics expression (33, 41). MEF-2 DNA binding activity was substantially reduced in nuclear extracts isolated from both heart and skeletal muscle of streptozotocin (STZ)-induced diabetic mice and completely recovered after insulin treatment, which correlated with transcriptional rate change of GLUT-4 gene (41).

The purpose of this study was to investigate the role of AMPK in regulating the transcription of the GLUT-4 gene. First, we determined the effects of an injection of AICAR on GLUT-4 mRNA levels, characterizing the time course and fiber type specificity of the response. To confirm that the increase in GLUT-4 mRNA that we observed was a result of transcriptional activation, we examined the effect of an AICAR injection on the expression of a chloramphenicol acyltransferase (CAT) reporter gene driven by 1,154 bp of the human GLUT-4 promoter. After observing the induction of the reporter gene in concert with the endogenous GLUT-4 gene, we repeated the experiment in other transgenic models in which the reporter gene was driven by 895 or 730 bp of the human GLUT-4 proximal promoter (Fig.1) to identify those regions involved in AICAR-induced transcriptional activation. The shortest transgenic construct (730) contains a fully functional MEF-2 site (−473 to −464) and a truncated domain I (−742 to −712) and supports a high level of transgene expression in muscle. Finally, we examined the DNA binding of nuclear extracts from AICAR-injected rats and mice to consensus sequences of MEF-2 and domain I, two binding sites implicated in the transcriptional regulation of the GLUT-4 gene (33, 41). The results of this study show that1) increasing AMPK activity with a single injection of AICAR leads to the transcriptional activation of the GLUT-4 promoter,2) this effect is mediated by response elements that lie within 895 bp of the promoter, and 3) this effect may be cooperatively mediated by MEF-2 through increased DNA binding activity.

EXPERIMENTAL PROCEDURES

Animals/experimental protocols.

All procedures were approved by the Institutional Animal Care and Use Committees of East Carolina University and Brigham Young University. Male Sprague-Dawley rats (Harlan, Indianapolis, IN) were housed individually with room temperature and lighting controlled (20–22°C; 12:12-h light-dark cycle) and free access to food and water.

To determine the time course of the effect on GLUT-4 mRNA level in muscle from a single AICAR injection, rats were injected subcutaneously with AICAR (1.0 mg/g body wt) or saline as previously described (14). AICAR was dissolved in 50 mg/ml saline. Injection time was 2 h after the start of the dark cycle, and animals had free access to food and water. Either 7, 13, 16, or 24 h after injection, rats were anesthetized by intravenous injection of pentobarbital sodium. Our preliminary observations indicated that AICAR had no effect on GLUT-4 or mRNA 2 and 4 h after injection (data not shown). Gastrocnemius muscles were quickly removed and frozen in liquid nitrogen for later analysis.

To determine the acute effect of AICAR injection on GLUT-4 mRNA level in different muscle fiber types, rats were given a single dose of AICAR (1.0 mg/g body wt) or saline. Sixteen hours after injection, rats were anesthetized by intravenous injection of pentobarbital sodium. The superficial white and the deep red regions of the quadriceps muscles and the soleus muscles were quickly removed and frozen in liquid nitrogen for later analysis.

To determine the chronic effect of AICAR injection on GLUT-4 protein level in different fiber type muscles, rats were given AICAR or saline injections according to the following schedule: 3 days injected, 2 days not injected, 5 days injected, 2 days not injected, 3 days injected. Rats were killed 24 h after the last injection. This intermittent injection schedule was used to minimize desensitization of AMPK responses to AICAR (51). The superficial white and the deep red regions of the quadriceps muscles and the soleus muscles were quickly removed and frozen in liquid nitrogen for later analysis.

Six- to eight-week-old mice were housed with room temperature and lighting controlled (20–22°C; 12:12-h light-dark cycle) and free access to food and water. Transgenic mice were generated as previously described (25, 41), with each transgene containing a portion of the 5′ flanking region of the human GLUT-4 gene driving the expression of the bacterial CAT reporter gene. Transgenic mice contained either 1,154 bp (1154-hG4-CAT), 895 bp (895-hG4-CAT), or 730 bp (730-hG4-CAT) of the 5′ flanking region to drive the expression of transgene CAT (Fig. 1). Although single transgenic lines for each construct were used in this study, they have been characterized with a high degree of tissue specificity and appropriate hormonal/metabolic regulation (32, 33, 41). Transgenic mice were identified by slot blot analysis of isolated tail DNA, as previously described (25). To ensure that experiments were performed with the same genetic background, we chose to use experimental animals that were offspring of a cross between transgenic mice and wild-type mice (C57bl/6J). Male mice that were heterozygous for the transgenic allele were used in studies of the human GLUT-4 promoter, whereas nontransgenic males were used for control groups and studies on murine GLUT-4 expression and electrophoretic mobility shift assays. To determine the acute effect of the AICAR injection on ZMP concentration and AMPK activity, mice were injected intraperitoneally (IP) with AICAR (1.0 mg/g body wt) or saline, and the same injection protocol was used as in the rat experiments described above. One hour after the injection, gastrocnemius muscles were quickly removed and frozen in liquid nitrogen for later analysis. To investigate the effect of AICAR treatment on GLUT-4 and CAT mRNA, mice were killed 12 h after a single IP injection of AICAR (1.0 mg/g body wt). The gastrocnemius muscles were quickly removed and frozen in liquid nitrogen for later analysis. Random samples of blood were collected from mice to check the blood glucose and lactate concentration using a Yellow Springs Instruments glucose/lactate analyzer (Yellow Springs, OH). Glucose and lactate levels were within physiological range, and no difference was found between control and AICAR group 12 h after a single AICAR injection (data not shown).

RNA isolation and Northern analysis.

RNA was isolated from quick frozen muscle samples using the TRIzol (Life Technologies, Rockville, MD) reagent according to manufacturer's instructions, as our laboratory has previously described (19). For Northern analyses, 5 μg of total RNA per samples were fractionated on a 1.25% agarose-2 M formaldehyde gel and then electrotransfered to a Hybond N+ membrane (Amersham, Piscataway, NJ). The membrane blots were cross-linked under ultraviolet light for 90 s and prehybridized with ULTRAhyb buffer (Ambion, Austin, TX) for 1 h at 42°C. Blots were then hybridized overnight at 45°C with random primed [α-³²P]dATP-labeled cDNA probes for GLUT-4 and then hybridized sequentially with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and 18S ribosomal RNA after stripping and prehybridization. The random primed [α-³²P]dATP-labeled cDNA probes were synthesized with Strip-EZ DNA kit (Ambion) and purified with a G-50 Nick column (Amersham). Northern blots were visualized by PhosphorImager and quantitated with Imagequant software (Molecular Dynamics, Sunnyvale, CA).

RNase protection assays.

Radiolabeled ([α-³²P]UTP, 800 Ci/mmol) antisense RNA probes were transcribed from the pTRI-GAPDH-mouse antisense control template (Ambion) and the Bsu 361 linearized mouse GLUT-4-CAT plasmid, p469GLUT-4-CAT (32), using Ambion's T3/T7 MAXIscript in vitro transcription kit. Briefly, RNase protection assays (RPA) were performed with the streamlined procedure of Ambion's RPA II kit as previously described (19, 20). Ten micrograms total RNA per samples and ³²P-labeled antisense probes were hybridized overnight at 45°C. Nonhybridized RNA was digested for 1 h at 37°C with 1:1,000 dilution of RNase T1/A from RPA II kit. The protected RNA fragments were separated by using 6% acrylamide 7.5 M urea gel electrophoresis and were then exposed to PhosphorImager and quantitated with Imagequant software (Molecular Dynamics). The size of the protected fragments was estimated by RNA transcripts labeled with [α-³²P]UTP from Century Marker Templates (Ambion) by using T7 RNA polymerase.

Analysis of AMP kinase activity and ZMP.

Muscles from mice killed 1 h after injection of AICAR or saline were analyzed for ZMP and AMPK (27, 49). Muscles were kept under liquid nitrogen until they were analyzed. They were first ground to powder under liquid nitrogen. Perchloric acid extracts (400 mg of tissue powder added to 2.0 ml of 6% perchloric acid) were prepared, and a 1-ml aliquot of the 6% perchloric acid extract was vortexed vigorously with 1 ml of 1,1,2-trichlorotrifluoroethane:tri-n-octylamine (1:1) to remove the perchloric acid. After centrifugation, the supernatant was stored at −70°C until analysis. ZMP was determined by using a Beckman HPLC system (supported by System Gold software) by a modification of the method described by Sabina et al. (37). Briefly, we used a Hichrom P10SAX column (0.45 × 25 cm) (DyChrom, Santa Clara, CA) with a P10SAX-10C5 guard column. At a flow rate of 2 ml/min, the elution began with 100% buffer A (5 mM NH₄H₂PO₄, pH 2.8) and 0% buffer B (750 mM NH₄H₂PO₄, pH 3.9). Buffer B was increased linearly to 9.3% over a 14-min period. Then over the next 36 min buffer B was increased to 100% in a linear gradient. AMPK activity was determined by using ammonium sulfate precipitates from homogenates prepared from the powdered (under liquid nitrogen) muscles as described previously (49).

Western immunoblot.

Muscles from rats killed 22–25 h after the tenth AICAR injection were analyzed for GLUT-4 protein content. Muscle samples were ground to powder under liquid nitrogen. A homogenate (1:9 dilution) was prepared in HEPES buffer [25 mM HEPES, 1 mM EDTA, 1 mM benzamidine, 1 mM 4-(2-aminoethyl)-benzene + sulfonyl fluoride, 1 μM leupeptin, 1 μM pepstatin, 1 μM aprotinin, pH 7.5]. Proteins of these homogenates were separated by SDS-PAGE by using 10% resolving gels (Tris · HCl ready gels, Bio-Rad, Hercules, CA). Proteins were transferred from the gel to a nitrocellulose membrane at 100 V for 60 min. The membranes were blocked with 3% BSA in 139 mM NaCl, 2.7 mM KH₂PO₄, 9.9 mM Na₂HPO₄, and 0.05% Tween 20 (PBST) and 1% sodium azide. After two 5-min washes in 139 mM NaCl, 2.7 mM KH₂PO₄, and 9.9 mM Na₂HPO₄ (PBS), membranes were incubated with GLUT-4 polyclonal antibody RaIRGT (Biogenesis, Sandown, NH) for 1 h at room temperature. After two 5-min washes in PBST and two 5-min washes in PBS, membranes were exposed to horseradish peroxidase-conjugated donkey anti-rabbit IgG (Amersham) for 1 h at room temperature. After being washed twice with PBST and twice with PBS, the membranes were incubated in enhanced chemiluminescence-detection reagent and then visualized on enhanced chemiluminescence hyperfilm (Amersham). Relative amounts of GLUT-4 were then quantified by using a Hewlett-Packard ScanJet 6200C and SigmaGel software (SPSS, Chicago, IL). Total intensity of GLUT-4 spots on the developed hyperfilm was expressed as a fraction of intensity shown by a GLUT-4 standard run on the same gel. The GLUT-4 standard was a plasma membrane fraction prepared as described previously (22).

Preparation of nuclear extracts.

Twelve hours after an IP injection of AICAR (1.0 mg/g body wt) or saline, mice and rats were euthanized, and both gastrocnemius muscles were harvested and frozen in liquid nitrogen. The frozen tissues were pulverized in liquid nitrogen and homogenized in 10 volumes of homogenization buffer A (250 mM sucrose, 10 mM HEPES, pH 7.6, 25 mM KCl, 1 mM EDTA, 10% glycerol, 0.1 mM PMSF, 0.5 mM benzamidine, 2 μg/ml each aprotinin, leupeptin, and pepstatin A, 2 mM levamisole, 10 mM β-glycerophosphate, and 1 mM sodium vanadate) with 10 strokes of a Teflon pestle. The homogenate was centrifuged 10 min at 2,400 g at 4°C. The pellet was resuspended in 10 ml of homogenization buffer A and homogenized with 10 strokes of a Dounce homogenizer with a tight fitting pestle. The homogenate was layered over one-half volume ofbuffer A-1.0 (buffer A with addition of 1.0 M sucrose) and centrifuged at 2,400 g for 10 min at 4°C. The pellet was resuspended in 2 volumes of buffer A-1.0 and repelleted by microcentrifugation at 16,000 g for 20 s. The crude nuclear pellet was then resuspended in 1 volume of nuclear extraction buffer (10 mM HEPES, pH 7.6, 325 mM NaCl, 3 mM MgCl₂, 0.1 mM EDTA, 10% glycerol, 1 mM DTT, 0.1 mM PMSF, 0.5 mM benzamidine, 2 μg/ml each aprotinin, leupeptin, and pepstatin A) and incubated on ice for 20 min, and the particulate material was removed by centrifugation at 16,000 g in a microcentrifuge for 10 min at 4°C. The supernatant was dialyzed against buffer C (25 mM HEPES, pH 7.6, 100 mM KCl, 0.1 mM EDTA, 10% glycerol, 1 mM DTT, 0.1 mM PMSF, 0.5 mM benzamidine, 2 μg/ml each aprotinin, leupeptin, and pepstatin A, 2 mM levamisole, 10 mM β-glycerophosphate, and 1 mM sodium vanadate) for 4 h. The dialysate was assayed for total protein (Bradford) and stored at −70°C.

Electrophoretic mobility shift assays.

Electrophoretic mobility shift assays were performed as previously described (41). Oligonucleotides containing the human GLUT-4 MEF-2 DNA binding site (GGGAGCTAAAAATAGCAG and its complement) and human GLUT-4 domain I DNA binding site (CTTGTCCCTCGGACCGGCTCCAGGAACCAA and its complement) were custom synthesized (GIBCO BRL). Oligonucleotides containing the OCT1 DNA binding site were commercially prepared (Santa Cruz). Oct-1 oligonucleotide binding serves as a negative control for DNA binding activity to a constitutively active transcription factor that would not be expected to be changed by the experimental treatment. Oligonucleotides were end-labeled with T4 polynucleotide kinase. Labeled probes (0.5 ng) were incubated with 5 μg of total protein isolated from nuclei in a 10 μl reaction containing 2 μg poly(dI-dC), 40 mM KCl, 5 mM MgCl₂, 15 mM HEPES, pH 7.9, 1 mM EDTA, 0.5 mM DTT, and 5% glycerol for 20 min at room temperature. For competition studies, extracts were preincubated with 20-fold unlabeled oligonucleotide as indicated for 5 min before addition of the radiolabeled probe. Preincubation was carried out for 1 h on ice before addition of the radioactive probe. Similarly, electrophoretic gel supershift studies were carried out by preincubating the nuclear extracts with 2.5 μg of either the MEF-2A, MEF-2B, or MEF-2D antibodies (a gift from Dr. Ron Prywes) or preimmune IgG. Preincubation was carried out for 1 h on ice before addition of the radioactive probe. Samples were electrophoresed on a nondenaturing 6% polyacrylamide gel (29:1 acrylamide-bis acrylamide) buffered with Tris-borate/EDTA (22 mM Tris, 22 mM boric acid, and 0.5 mM EDTA) at 300 V for 25 min at room temperature. Gels were then dried and exposed to film at room temperature.

RESULTS

The effect of AICAR on GLUT-4 mRNA levels.

Previously our laboratory reported that chronic AICAR treatment in rats led to the accumulation of more GLUT-4 protein than is found in saline-injected controls (14). To investigate whether this effect on GLUT-4 protein was the consequence of a pretranslational regulation by AICAR, GLUT-4 mRNA was examined by Northern blot analysis after a single subcutaneous injection of AICAR in rats and mice. A representative Northern blot showing GLUT-4 and GAPDH mRNA 16 h after an injection of saline or AICAR in rats is shown in Fig.2A. GAPDH levels did not differ between saline and AICAR-injected animals with or without normalization to 18S ribosomal RNA. The GLUT-4/GAPDH ratio was ∼50% higher in the AICAR-injected group of rats. Of interest was the time course of AICAR-induced increase in GLUT-4 mRNA. GLUT-4 mRNA levels were elevated (∼30%) by 7 h, peaked at 13 h (∼60%), remained elevated at 16 h (∼45%), but had returned to normal by 24 h (Fig. 2B). Similar GLUT-4 mRNA time course response was confirmed in mice with smaller group sample number (2–4 mice per each time point, data not shown). This result suggested a transient increase in the rate of transcription of GLUT-4 gene in response to AICAR.

Fiber type-specific responses of GLUT-4 to AICAR.

The gastrocnemius muscle is mixed with respect to fiber type composition. Previously, our laboratory (51) reported that 4 wk of chronic AICAR treatment led to an increase in GLUT-4 protein in red and white quadriceps, but not soleus muscle. With a 2-wk regimen of chronic AICAR treatment designed to minimize the downregulation of the AICAR response, we again observed that GLUT-4 protein levels were higher with AICAR treatment in both red (∼30%) and white (∼115%) quadriceps, but not in the soleus muscle (Fig.3). We investigated whether the acute effect of AICAR on GLUT-4 mRNA levels was also fiber type specific by examining AICAR-induced GLUT-4 expression in red and white quadriceps and soleus muscles 16 h after an injection of saline or AICAR. GLUT-4 mRNA levels increased 70% in white quadriceps and 50% in red quadriceps muscles, but no difference between saline and AICAR-injected rats in GLUT-4 mRNA was observed in the soleus (Fig.4). Thus AICAR treatment leads to an acute elevation in GLUT-4 mRNA levels, which, when treated chronically, leads to the accumulation of GLUT-4 protein, and these effects appear to differ with respect to muscle fiber type composition. The most profound effect of AICAR on GLUT-4 expression was observed in white quadriceps (primarily type IIb fibers), a less dramatic effect was observed in red quadriceps (primarily type IIa fibers), and no effect was observed in the soleus (primarily type 1 fibers).

Regulation of human GLUT-4 promoter in transgenic mice by AICAR.

Previously our laboratory observed an increase in the concentration of ZMP and the activity of AMPK in rat skeletal muscle within 60 min of a single subcutaneous injection of AICAR (14). A single subcutaneous injection of AICAR in mice also leads to detectable amounts of ZMP and increased AMPK activity in the gastrocnemius muscle of mice (Table 1). As with the rats, no detectable change was observed in mice gastrocnemius ATP or ADP concentrations in response to the AICAR injection (data not shown). We also observed similar time course responses in the elevation of endogenous GLUT-4 mRNA in AICAR-injected mice as in rats, with gastrocnemius GLUT-4 mRNA peaking ∼13 h after treatment (data not shown).

How, then, does AICAR treatment lead to an elevation in GLUT-4 mRNA levels? One possibility was that activation of AMPK leads to increased transcription of the GLUT-4 gene. To investigate this possibility, we examined the AICAR-induced elevation in GLUT-4 mRNA levels in transgenic mice that harbor a promoter-reporter construct in which the CAT reporter gene is driven by 1,154 bp of the human GLUT-4 promoter (1154 hG4-CAT). Probes specific to transgene CAT mRNA, endogenous GLUT-4 mRNA, and GAPDH mRNA (a housekeeping gene) were simultaneously hybridized to isolated total RNA from excised gastrocnemius muscle 12 h after an injection of saline or AICAR. A representative RPA for saline and AICAR-injected mice is shown in Fig.5. Protected GAPDH mRNA remained unchanged in response to AICAR injection for each transgenic line and was used to normalize CAT and endogenous mRNA levels for minor RNA loading variations. In the AICAR-treated animals, CAT mRNA and GLUT-4 mRNA were higher (>200%, ∼20%, respectively) than in saline controls. This observation indicated that the effects of AICAR were likely mediated, at least in part, by the upregulation of GLUT-4 transcription. The disproportionate increase between CAT and endogenous GLUT-4 mRNA in response to AICAR could be due to a number of reasons. Although there could be inhibitory elements residing upstream of −1,154 bp of promoter, it is also possible that there are subtle species-specific differences in the GLUT-4 promoter sensitivity to transcription factors. In addition, the CAT mRNA stability could also contribute to this observation, because the construct of the CAT gene has been modified on the 3′ end to increase CAT mRNA stability.

To investigate what portion of the promoter was responsible for mediating the AICAR-induced GLUT-4 gene transcription, this experiment was repeated in transgenic lines in which the transcription of the CAT gene was driven by 895 (895-hG4-CAT) or 730 (730-hG4-CAT) bp of the human GLUT-4 promoter. The relative amounts of CAT mRNA and GLUT-4 mRNA for saline and AICAR-injected mice of each construct are summarized in Fig. 6. In 895-hG4-CAT mice, both endogenous GLUT-4 and CAT mRNA were elevated in AICAR-injected mice compared with saline-injected controls. In contrast, the same experimental protocol in 730-hG4-CAT mice indicated that endogenous GLUT-4 mRNA levels were elevated in AICAR-injected mice compared with saline-injected mice, but CAT mRNA levels did not differ between the two groups. Thus it suggests that elements required for AICAR induced GLUT-4 gene transcription are located within 895-bp human GLUT-4 promoter.

AICAR injection increases the MEF-2 binding activity.

It has been shown recently that MEF-2 and domain I binding domains on human GLUT-4 gene promoter are two regulatory domains that function cooperatively (33, 41). Nuclear extracts were prepared from gastrocnemius muscles obtained from AICAR- and saline-treated mice and rats. Electrophoretic mobility shifts of the MEF-2, domain I, and OCT-1 oligonucleotides demonstrated an increase in binding activity of MEF-2 to GLUT-4 promoter from AICAR-treated mouse muscle nuclear extracts compared with the saline control extracts (Fig.7). In contrast, binding of domain I to GLUT-4 promoter was slightly decreased in AICAR-treated nuclear extracts. As a control, binding of OCT-1 was unchanged after AICAR treatment. Similar results were observed in rats (data not shown). Thus the increased binding of MEF-2 to the GLUT-4 promoter may contribute to AICAR-induced GLUT-4 transcription.

To identify the nature of nuclear proteins that bind to GLUT-4 MEF-2 site, we preincubated the mouse muscle nuclear extracts with MEF-2A, MEF-2B, MEF-2D, and a control IgG (Fig.8). Incubation of skeletal muscle nuclear extracts with the 18-bp MEF-2 oligonucleotide resulted in the formation of a DNA-protein complex that was not present in the absence of the nuclear extract (Fig. 8, lane 1). Pretreatment of this complex with the MEF-2A and MEF-2D antibodies resulted in the supershift of the DNA-protein complex. There was no supershift of DNA-protein complex in nuclear extracts preincubated MEF-2B antibody. Thus it appears that binding of MEF-2A or/and MEF-2D transcription factors could be involved in upregulating the GLUT-4 promoter by AICAR treatment.

DISCUSSION

The present study demonstrated that AICAR, an activator of AMPK, increases transcription of the GLUT-4 gene in muscle, and chronic treatment increases the total GLUT-4 protein. The increased transcription of GLUT-4 and its protein levels in muscle after AICAR treatment appears to be fiber type specific. The red and white hindlimb muscles both responded to AICAR treatment with increased GLUT-4 transcription and protein levels. However, in the soleus neither GLUT-4 mRNA or protein level showed difference after AICAR treatment. These findings are in agreement with our laboratory's previous study with a different AICAR treatment regime that GLUT-4 protein was elevated only in red and white quadriceps, but not soleus muscle (51). In agreement with our results, a recent study also shows that 5 days chronic treatment with AICAR increased GLUT-4 protein and mRNA levels in white muscle (1). The mechanism mediating this fiber type tissue-dependent GLUT-4 gene expression is still unclear at this time. As a muscle composed of predominantly oxidative fibers, the soleus has a higher level of GLUT-4 gene expression. Therefore one possibility is that, with relatively high initial GLUT-4 level, it could be less likely to show any further increase after AICAR treatment.

AMPK is a heterotrimeric protein consisting of one catalytic subunit (α) and two noncatalytic subunits (β and γ) (12). Two isoforms of the α subunit have been identified (α1 and α2), which have broad tissue distribution, and the highest expression level of the α2 isoform is found in skeletal muscle, suggesting a physiological role for AMPK in this tissue (39, 45). Interestingly, contraction preferentially activated the α2 isoform (45). The α2 isoform has also been observed to be preferentially targeted to the nucleus (38). After entering the cell, AICAR is phosphorylated and accumulates as ZMP, thereafter mimicking the ability of AMP to activate AMPK. AMP activates AMPK through both covalent and noncovalent mechanisms (12). Increased AMP concentrations cause AMP-activated protein kinase kinase (AMPKK), an upstream kinase, to phosphorylate AMPK, thereby increasing its activity. In addition, an increased AMP concentration allosterically activates the enzyme. ZMP mimics all effects of AMP on the AMPK system, including stimulation of phosphorylation of AMPK by AMPKK. This method of activating AMPK has the advantage that it does not disturb the levels of ATP, ADP, or AMP in the muscle (27), and therefore any changes seen are not merely due to depletion of ATP. The present study did not find any detectable change in ATP and ADP levels in muscle after a single injection of AICAR.

Infusion of AICAR in vivo has previously been shown to cause hypoglycemia in mice (47). It was suggested that the hypoglycemia was caused by inhibition of gluconeogenesis, a known effect of AMPK activation (48). Recent studies have also demonstrated that AICAR-mediated activation of AMPK stimulates skeletal muscle glucose uptake through the translocation of GLUT-4 to the cell surface in muscle tissue (22, 36). We could not exclude the possibility of a direct effect of lower glucose and insulin level or any effects associated with the accompanying metabolic acidosis on the GLUT-4 gene expression regulation. However, transcriptional run-on analysis has shown that increased levels of glucose and/or insulin increase GLUT-4 gene transcription in muscle; i.e., lower glucose and insulin would be predicted to decrease transcription (19). It is also demonstrated, by analysis of the human GLUT-4 promoter in transgenic mice, that GLUT-4 expression is downregulated in insulin-deficient STZ diabetic muscle (32). In previous study, 24 h fasting and refeeding did not show an effect on CAT and endogenous GLUT-4 mRNA level in transgenic mice with 2.4-kb human GLUT-4 promoter attached to the CAT reporter gene construct (25). The recent observation that increased expression of GLUT-4 and hexokinase is observed after in vitro incubation of isolated epitrochlearis muscles with AICAR for 18 h provides additional evidence that this is an effect of AMPK activation rather than a result of humoral changes secondary to the AICAR injection (31).

There is growing evidence to suggest that AMPK is an important component of a protein kinase cascade that monitors cellular energy charge, being activated by a rise in the cellular AMP-ATP ratio. AMPK appears to act as a “metabolic master switch” (12,50), regulating metabolism both via direct phosphorylation of metabolic enzymes and via effects on gene expression. In addition to our findings, there are a number of features that provide clues for a role of AMPK in gene regulation. It was previously reported that AMPK is structurally and functionally related to the yeast SNF1 protein kinase complex (2). The activity of SNF1 is essential for the transcriptional regulation of glucose-inducible genes (3,28). This is achieved through the phosphorylation of a transcription factor (6), depending on the medium glucose concentration. Recently, a role for mammalian AMPK in gene regulation has been confirmed as AICAR-activated AMPK is reported to inhibit phospho(enol )pyruvate carboxykinase (PEPCK) gene transcription (26) and glucose-induced transcription ofl-pyruvate kinase, Spot 14, and fatty acid synthase genes (7, 23).

Numerous studies have shown that exercise training increases muscle GLUT-4 mRNA and protein in skeletal muscle (4, 10, 16,30). Nuclear run-on analysis demonstrated that GLUT-4 transcription was increased after 3 h and returned to control values after 24 h (29). A recent study on transcriptional regulation of gene expression during recovery from exercise demonstrated that exercise induces transient increases in transcription of metabolic genes in human skeletal muscle (34). Further evidence that GLUT-4 gene transcription is increased by exercise was provided by studies of transgenic mice by using 5′-deletion analysis of the mouse GLUT-4 minigene (43,44). The present study shows that GLUT-4 gene expression is transcriptionally upregulated by AICAR-activated AMPK in muscle, providing additional evidence of AMPK involvement in regulation of muscle gene expression by exercise. Investigation of the pathway by which AMPK acts may therefore give insight into mediating mechanism of muscle contraction or exercise in regulation of gene expression of GLUT-4 and other muscle genes.

Considerable work has been done in attempts to understand the function of the GLUT-4 gene promoter. Analysis of transgenic mice with sequential 5′ deletions and 3′ internal deletions of human GLUT-4 gene promoter has suggested that the first 895 base pairs upstream of the major transcription initiation site of the human GLUT-4 gene contain the major regulatory elements of the gene (32, 33, 41). The cooperative activity between two cis elements within this 895-bp proximal promoter, domain I and the MEF-2 site, are necessary for the full function of the promoter. Domain I is located between nucleotides −742 and −712, and the MEF-2 binding domain is located between nucleotides −473 and −464 (both regions are numbered relative to the major transcription start site).

In the present study, we show that transcriptional upregulation by AICAR is fully functional in 1154 (1154-hG4-CAT) and 895 (895-hG4-CAT) transgenic mice. Both of these constructs contain domain I and MEF-2 sites. However, transgene CAT mRNA level in the −730 transgenic mice (730-hG4-CAT) did not respond to AICAR injection, whereas endogenous GLUT-4 mRNA responded to AICAR treatment compared with saline injection. These data indicate that the element involved in regulation of GLUT-4 gene transcription by AICAR-activated AMPK are located within 895 bp of the human GLUT-4 promoter. The 730-hG4-CAT transgenic construct had domain I (−742 to −712) truncated but left the MEF-2 binding domain intact. A previous study has shown that the human GLUT-4 proximal promoter deleted to position −730 (730-hG4-CAT) supported a high level of transgene expression in muscle; however, the transgene expression was not regulated in STZ-induced diabetes (32). Previous studies have shown that deletion of domain I inside the −895-bp transgenic construct (895Δ742/526-hG4-CAT) prevented transgene expression, regardless of the presence of the MEF-2 domain (33). The fact that AICAR had no effect in 730-hG4-CAT construct is most likely due to the disruption of the cooperative activity between domain 1 and MEF-2. The lack of effect of AICAR-treatment on expression of this transgenic construct is similar to previous data showing that STZ-induced diabetes did not affect expression of this construct. However, it is possible that other transcription factors within 895 bp of the gene, especially between −895 and −730, may also be involved in the AICAR-induced increase in GLUT-4 transcription in addition to MEF-2 and domain I. Studies using the transgenic minigene suggested two regions of mouse GLUT-4 promoter, possibly as exercise response elements (43, 44). One region is between −551 and −442, upstream of a MEF-2 binding site (−437/−428). Another region is between −1001 and −701 in mouse GLUT-4 gene. The differences observed between our studies and these studies may be due to species-specific differences in the gene sequences.

Because either a deletion of the domain I or mutation of the MEF-2 binding domain does not support the transgene expression, we were unable to directly study these two elements for GLUT-4 transcription upregulation by using deletional analysis. To investigate whether MEF-2 site and domain I are involved in the transcriptional upregulation of GLUT-4 gene by AICAR, we studied the effect of AICAR on DNA binding activity to MEF-2 site or domain I.

To our surprise, DNA binding activity to domain I was slightly decreased in AICAR-treated animals. At this stage, we could not explain the link between decreased DNA binding at domain I and increased GLUT-4 transcription by AICAR whereas domain I is required to function together with MEF-2 to support GLUT-4 transcription. It is possible that dynamic changes in protein-DNA interactions in the 895-bp proximal promoter results in upregulation of the GLUT-4 gene through an increased MEF-2 binding and a decreased domain I interaction.

The MEF-2 DNA binding site has been demonstrated to be necessary for human GLUT-4 gene transcription. A mutation inside the MEF-2 site ablated transgene mRNA expression in the 895-bp construct human GLUT-4 transgenic mice (41). The MEF-2 binding domain functions cooperatively with domain I to support transcription of the human GLUT-4 gene in transgenic mice (33). In addition to its role in tissue-specific expression, MEF-2 binding activity may be important for metabolic regulation of the GLUT-4 gene. MEF-2 binding activity is reduced in nuclear extracts from diabetic animals and insulin treatment of these animals returns MEF-2 binding activity to normal (41). It is also shown that MEF-2A protein levels are selectively reduced in skeletal muscle of insulin-deficient diabetic rats (41). This loss of MEF-2A expression accounts for the reduction in DNA binding activity and directly correlates with the decrease in GLUT-4 gene expression, suggesting that it plays a role in regulating the transcription rate of the GLUT-4 gene. The results in this study showed that MEF-2 binding activity to the GLUT-4 promoter was increased in AICAR-treated nuclear extracts. MEF-2 binding correlates well with the increased GLUT-4 gene transcription. For this reason, we postulate that the MEF-2 binding site is necessary and may be directly involved in regulating GLUT-4 expression by AICAR-activated AMPK.

In summary, we have demonstrated that AICAR-activated AMPK increases muscle GLUT-4 transcription in a fiber type-specific manner. Thecis elements responsible for increased human GLUT-4 transcription are within the 895 base pairs upstream of the major transcriptional start site. Furthermore, the increased MEF-2 binding activity with AICAR treatment suggested that the MEF-2 domain is an important regulatory element in the human GLUT-4 gene promoter. We hypothesize that AMPK activation associated with muscle contraction and exercise not only plays important roles in stimulation of fatty acid oxidation and glucose uptake, but also appears to be important in modulating gene expression in the muscle. Identification of the specific phosphorylation targets involved in this process will greatly enhance identification of more specific mechanisms. Further studies using AMPK null transgenic models or AMPK specific inhibitors and assessment of both the GLUT-4 MEF-2 domain and domain I under the condition of muscle contraction or exercise could provide valuable insight into the regulation of muscle gene expression.

FOOTNOTES

• This work was supported by grants from the National Institutes of Health (DK-38416 to G. L. Dohm, AR-41438 to W. W. Winder, and DK-47894 to A. L. Olson) and American Diabetes Association (RA0064 to A. L. Olson and a mentor-based postdoctoral fellowship to P. S. MacLean).

• Address for reprint requests and other correspondence: G. L. Dohm, Dept. of Biochemistry, Brody School of Medicine, East Carolina Univ., Greenville, NC 27858.

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.