Transcriptional Regulation of the Gene for Glucose Transporter GLUT4 in Skeletal Muscle

GLUT4 glucose transporter protein and mRNA lev- els in rat skeletal muscle are decreased with streptozotocin (STZ)-induced diabetes and increased by fasting, indicating that GLUT4 expression may be regu- lated at the pretranslational level. The purpose of the present study was to determine whether GLUT4 is subject to transcriptional regulation in skeletal muscle under the altered metabolic conditions of diabetes and fasting. Nuclei were isolated from red and white por- tions of the quadriceps and gastrocnemius/plantaris muscles of control, 7-day STZ-diabetic, and 3-day fasted rats. STZ-induced diabetes resulted in a 35% reduction in GLUT4 transcription in red skeletal muscle and thus accounted for a major portion of the cor- responding 50% reduction in GLUT4 mRNA observed in red skeletal muscle. STZ-induced diabetes had no significant effect on GLUT4 transcription or mRNA in white skeletal muscle. Fasting, however, significantly increased both GLUT4 transcription (2.2-fold) and mRNA (2.9-fold) in white skeletal muscle with no change detected for either parameter in red skeletal muscle. The nearly 2-fold higher steady-state GLUT4 mRNA in red versus white skeletal muscle of control rats was not associated with any difference in basal transcription. These findings demonstrate that expres- sion of the GLUT4 glucose transporter protein in skeletal muscle is subject to regulation in vivo at the level of transcription of the GLUT4 gene. In addition,

Transcriptional Regulation of the Gene for Glucose Transporter GLUT4 in Skeletal Muscle EFFECTS OF DIABETES AND FASTING* (Received for publication, January 15, 1993, and in revised form, March 3, 1993) P. Darrell  GLUT4 glucose transporter protein and mRNA levels in rat skeletal muscle are decreased with streptozotocin (STZ)-induced diabetes and increased by fasting, indicating that GLUT4 expression may be regulated at the pretranslational level. The purpose of the present study was to determine whether GLUT4 is subject to transcriptional regulation in skeletal muscle under the altered metabolic conditions of diabetes and fasting. Nuclei were isolated from red and white portions of the quadriceps and gastrocnemius/plantaris muscles of control, 7-day STZ-diabetic, and 3-day fasted rats. STZ-induced diabetes resulted in a 35% reduction in GLUT4 transcription in red skeletal muscle and thus accounted for a major portion of the corresponding 50% reduction in GLUT4 mRNA observed in red skeletal muscle. STZ-induced diabetes had no significant effect on GLUT4 transcription or mRNA in white skeletal muscle. Fasting, however, significantly increased both GLUT4 transcription (2.2-fold) and mRNA (2.9-fold) in white skeletal muscle with no change detected for either parameter in red skeletal muscle. The nearly 2-fold higher steady-state GLUT4 mRNA in red versus white skeletal muscle of control rats was not associated with any difference in basal transcription. These findings demonstrate that expression of the GLUT4 glucose transporter protein in skeletal muscle is subject to regulation in vivo at the level of transcription of the GLUT4 gene. In addition, GLUT4 transcription is regulated in a fiber type-specific manner in response to the metabolic challenges elicited by STZ-induced diabetes and fasting.
The GLUT4 protein is expressed exclusively within skeletal muscle and adipose tissue and is a member of a multigene family which encodes for the mammalian facilitative glucose transporters (1)(2)(3). Stimulation of glucose transport in skeletal muscle by insulin and/or contraction results in the translocation and activation of GLUT4 (4,5). Since skeletal muscle is responsible for approximately 80% of postprandial glucose disposal (6, 7), regulation of GLUT4 protein expression in muscle under different metabolic conditions such as diabetes and fasting has been an active area of investigation (8,9). Rats made diabetic by streptozotocin (STZ)' treatment, in * This work was supported by National Institutes of Health Grant DK38416. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. addition to a severe reduction in glucose transport capacity (10-12), also exhibit a reduction in skeletal muscle . In contrast, the increase in muscle GLUT4 protein expression (18) found in fasted rats coincides with an enhanced insulin sensitivity seen in vitro (19).
The steady-state level of a given protein is determined by various control mechanisms operating simultaneously within the cell. Regulation occurs at multiple sites including transcription, nascent RNA processing, mRNA stability, translation, and protein processing, transport and degradation. The possibility that muscle GLUT4 protein may be subject to regulation at the pretranslational level was first suggested by Garvey et al. (13) who reported a significant decrease in GLUT4 mRNA in the quadriceps muscles of rats with STZinduced diabetes. Insulin injection in STZ-diabetic animals, however, restored both GLUT4 mRNA and protein to control levels. In contrast to diabetes, 2-3 days of fasting induces an increase in muscle GLUT4 mRNA content which returns to control levels upon refeeding (18). Skeletal muscle, however, is a heterogeneous tissue, composed of both red oxidative fibers (types I and IIA) which are high in GLUT4 protein content, and white glycolytic fibers (type IIb) which are low in GLUT4 protein content (20,21). In view of the functional and metabolic diversity of skeletal muscle, two more recent reports reexamined the effects of diabetes and fasting in both red and white skeletal muscle and found that the STZ-induced reduction in GLUT4 mRNA occurs only in red skeletal muscle while the increase in GLUT4 mRNA elicited by fasting occurs only in white skeletal muscle (12,17).
The intent of the present study was to determine whether these metabolically induced alterations in GLUT4 mRNA in skeletal muscle may be regulated at least in part at the level of transcription of the GLUT4 gene. Transcriptional analysis of various genes has been performed in vitro with nuclear extracts prepared from a wide variety of animal tissue and tissue culture cells (22)(23)(24). In skeletal muscle, however, the ability to assess in uiuo transcriptional regulation has been limited by the inability to isolate transcriptionally active nuclei under conditions typically used for cultured cells or soft tissues due to the high content of contractile protein in skeletal muscle. To circumvent this problem, Zahradka et al. (25) have recently described a technique in which the muscle is initially homogenized in a large volume of buffer which minimizes disruption and permits the isolation and purification of intact, transcriptionally active nuclei from rat skeletal muscle. We have applied slight modifications of this procedure to isolate nuclei from skeletal muscle of rats either made diabetic by STZ treatment or fasted for 3 days. Using a nuclear RNA transcript elongation (nuclear run-on) assay, we found that, in conjunction with a 50% reduction in GLUT4 mRNA, transcription originating from the GLUT4 gene was decreased 35% in red but unaltered in white skeletal muscle of STZinduced diabetic rats. In contrast, fasting elicited a 2.2-fold increase in GLUT4 transcription and 2.9-fold increase in GLUT4 mRNA content in white skeletal muscle with no change evident in red skeletal muscle. Thus these data provide evidence that GLUT4 expression in skeletal muscle is subject to regulation in vivo at the level of transcription of the GLUT4 gene in a fiber type-specific manner.

EXPERIMENTAL PROCEDURES
Materials-All radiolabeled compounds were purchased from Du Pont-New England Nuclear. Hybond-N was obtained from Amersham Corp. RQ1 DNase, RNasin, and all restriction enzymes were purchased from Promega (Madison, WI). All other chemicals unless otherwise specified were of molecular biology grade and purchased from Sigma or Fisher.
Animals and Experimental Design-Male Sprague-Dawley rats (300-350 g; CD strain, Charles River Laboratories, Wilmington, MA) were either made diabetic by a single intraperitoneal injection of 75 mg/kg STZ, fasted for 3 days, or maintained as controls. The STZtreated animals were sacrificed 7 days after injection. Blood glucose at the time of sacrifice averaged 543 k 13, 158 f 3, and 119 * 3 mg/ dl for STZ, fasted, and control animals, respectively. All animals were stunned and killed by decapitation.
RNA Isolation and Northern Blot Analysis-Upon sacrificing the rats, red and white portions of the quadriceps muscles from one hindlimb were rapidly excised, frozen with tongs cooled in liquid nitrogen, and stored at -70 "C until analysis. After powdering the muscle using a cold steel mortar and pestle, total RNA was isolated from -200 mg by the guanidinium thiocyanate-phenol-chloroform extraction method (26) as modified by Puissant and Houdebine (27). The RNA (15 pg) was denatured and size fractionated on a 1.25% agarose, 2.0 M formaldehyde gel. After staining (0.5 pg/ml ethidium bromide), the 28 S and 18 S ribosomal bands were visualized and photographed by ultraviolet transillumination to ensure that the RNA was intact and evenly loaded. The RNA was then transferred to Hybond-N, UV cross-linked, and prehybridized at 47 "C for 4 h in a solution of 50% deionized formamide, 4 X SSC (1 X SSC = 150 mM sodium chloride, 15 mM sodium citrate), 5 X Denhardt's solution (50 X Denhardt's solution = 0.1% each of bovine serum albumin, polyvinylpyrrolidone, and Ficoll), 0.1 mg/ml yeast tRNA, 50 mM sodium phosphate (pH 7.0), 0.5 mg/ml sodium pyrophosphate, and 1% SDS. Hybridizations were carried out overnight in hybridization solution (prehybridization solution with 1 X Denhardt's solution) at 47 "C using the appropriate cDNA probe at 1 X lo7 cpm/ml. All cDNA probes were labeled with [w3'P]dATP (3000 Ci/mmol) by random priming as previously described (28).' The membranes were washed for 30 min in 0.1 X SSC and 0.1% SDS at room temperature followed by 30 min at 55 "C and then subjected to autoradiography using Kodak XAR-5 film with intensifying screens. The resulting autoradiograms were quantitated by laser densitometry.
Isolation of Nuclei-Nuclei were isolated from rat skeletal muscle by a procedure described by Zahradka et al. (25) with certain modifications. Quadriceps and gastrocnemius/plantaris muscles were dissected, pooled, minced, and weighed (8-9 g). The tissue was then immediately homogenized with a Polytron homogenizer in 31 volumes (w/v) of cold (4 "c) lysis buffer (10 mM HEPES, pH 7.5, 5 mM KC1, 10 mM MgC12, 5 mM 6-mercaptoethanol) containing 0.32 M sucrose. The homogenate was passed through four layers of cheese cloth and filtered through a 100 mesh stainless-steel screen (Fisher). The nuclei were collected by centrifugation at 1000 X g for 10 min in a Soma1 GSA rotor. The nuclear pellet was resuspended by gently triturating with 35 ml of cold (4 "C) lysis buffer containing 2.2 M sucrose and centrifuged for 90 min at 80,000 X g (27,000 rpm), 4 "C in a Beckman SW28 rotor. The resulting nuclear pellet was gently rinsed with cold ~ The cDNA probes used in the present study were as follows: GLUT1, a 2.7-kb EcoRI fragment encoding the murine 3T3-Ll homolog of the HepGL/brain glucose transporter protein (57); GLUT4, a 2.8-kb EcoRI fragment encoding the 3T3-Ll homolog of the adipose/ muscle (insulin-responsive) glucose transporter protein (57); c-jos, a 1.0-kb PstI fragment of p-jos-1, ATCC No. 41040 (52); c-jun, a 1.8kb EcoRIIHindIII fragment obtained from Dr. W. W. Lamph, the Salk Institute, San Diego (58), and &actin, a 1.9-kb HindIII fragment obtained from Dr. D. W. Cleveland, The Johns Hopkins University, School of Medicine, Baltimore (59).
Run-on Transcription Analysis-The procedure used for nuclear run-on analysis is based on techniques previously described by Cornelius et al. (24). After thawing on ice, the concentration of DNA in each preparation was determined by lysing a 10-pl aliquot in 990 p1 of 0.1% SDS solution and measuring the absorbance at 260 and 230 nm (29). Equal numbers of nuclei from control, STZ, and fasted preparations were allowed to complete the synthesis of nascent RNA transcripts in a reaction mixture containing 58.7 mM HEPES, pH 7.5, 80 mM KC1, 11.7 mM NaC1, 6.5 mM dithiothreitol, 5 rnM MgC12, CTP, and 0.4 p~ [cx-~'P]UTP (250 pCi/reaction) with 40 units of RNasin (RNase inhibitor) in a total volume of 230 p1 at 2EC. After 30 min, the samples were subjected to DNase digestion for 5 min at 25 "C by the addition of 25 units of RNase-free DNase followed by the addition of 3 ml of 4 M guanidinium thiocyanate, 20 mM sodium acetate, pH 5.2,0.5% N-lauryl sarcosine, and 5% P-mercaptoethanol. The 32P-labeled RNA was isolated by centrifugation (150,000 X g) through 5.7 M CsCl (1.5 ml) in a Beckman SW55 rotor (35,000 rpm, 16 h, 18 "C).
The 32P-labeled RNA was recovered by resuspending the pellet initially in 100 pl of 10 mM Tris-HC1, pH 7.4, 5.0 mM EDTA, 1% SDS at 65 "C followed by 3 X 100 pl washes with diethylpyrocarbonate-treated water. The RNA was recovered by ethanol precipitation and subjected to DNase digestion for 30 min at 37 "C with 10 units of RNase-free DNase in 40 mM Tris-HC1, pH 7.6, 6.0 mM MgCl,, 10 mM NaCl, 1 mM dithiothreitol, and 40 units of RNasin. The samples were extracted once with chloroform/butanol (4:l). To ensure complete denaturation of the transcripts, the 32P-labeled RNA was partially hydrolyzed in 0.2 M NaOH (15 min on ice) followed by neutralization with 1 M HEPES (pH 5.2) (30). The 32P-labeled RNA was recovered by ethanol precipitation and resuspended in 1 ml of hybridization solution (see above) by heating at 65 "C for 10 min. The concentration of 32P-labeled RNA was determined by liquid scintillation spectrometry, and all samples being compared were adjusted to equal concentrations (counts/min/ml) and volume (1 ml) with hybridization solution. Hybridization was carried out using Hybond N to which 2 pg of several individual cDNAs of interest were fixed by UV cross-linking. Each cDNA was gel-purified from its plasmid vector after overnight digestion with the appropriate restriction enzyme and then denatured in 0.1 M NaOH for 30 min at 37 "C, neutralized in the presence of 10 X SSPE (1 X SSPE = 0.15 M NaCl, 0.01 M NaHP04, pH 7.4, 1.0 mM EDTA), and applied to Hybond-N using a slot-blot apparatus (Mini-fold 11, Schleicher and Schuell). All filters were prehybridized for 4 h at 47 "C in prehybridization solution (see above). Each filter for a given RNA preparation was trimmed, cut in two, and placed back to back in a Seal-a-Meal bag (Dazey Corp.) to which the 1 ml of sample hybridization solution was added. This technique permitted the detection of as many as 14 different transcripts simultaneously. Hybridization was carried out for 2.5 days at 47 "C after which the filters were washed for 30 min at 50 "C in 2 X SSC, 30 min at 37 "C in 2 X SSC containing 10 pg/ml of RNase A, and for 30 min at 55 "C in 0.1 X SSC, 0.1% SDS. After drying, the membranes were subjected to autoradiography with intensifying screens for 7-14 days. All bands were quantitated by laser densitometry.

RESULTS
It is well established that GLUT4 mRNA in skeletal muscle is reduced with STZ-induced diabetes and increased with fasting (13-18), suggesting that GLUT4 may be regulated at the pretranslational level. The primary intent of the present study was to determine whether expression of GLUT4 in skeletal muscle may at least in part be regulated in uiuo at the level of transcription of the GLUT4 gene. To this end, nuclei were isolated from whole quadriceps and gastrocnemius/plantaris muscles of control, STZ-diabetic, and fasted rats and subjected to nuclear run-on analysis. Fig. 1  typical autoradiogram in which nascent transcripts corresponding to GLUT1, GLUT4, c-{os, c-jun, and /3-actin were detected by the appropriate cDNA probes. Hybridization of the transcripts was specific as no activity was detected using nonspecific plasmid DNA ( p G E M ) . All transcripts were normalized to genomic DNA (G-DNA) to account for slight variations in RNA transcript concentration in the hybridization solution. The lower portion of Fig. 1 provides a summary of the GLUT4 transcription data from seven observations. No significant change in transcription originating from the GLUT4 gene was detected with STZ treatment. Fasting however elicited a significant 1.7-fold increase in GLUT4 transcription relative to controls.
Two more recent studies have demonstrated that the STZinduced reduction in GLUT4 mRNA occurs specifically in red skeletal muscle while the increase in GLUT4 mRNA elicited by fasting occurs only in white muscle (12, 17), suggesting that GLUT4 expression may be subject to regulation in a fiber type-specific manner. Thus, in the present study, if the effects of STZ-induced diabetes on GLUT4 transcription were specific for red muscle, nuclear run-on analysis of whole mixed muscle would limit the ability to detect any effect due to STZ treatment. To address this possibility, we reexamined the effects of STZ-induced diabetes and fasting on GLUT4 mRNA expression and transcription of the GLUT4 gene using nuclei isolated from both red and white portions of the quadriceps and gastrocnemius/plantaris muscles from a second group of rats completing the same protocol. In agreement with two previous reports (12, 17), STZ-induced diabetes resulted in a -50% decrease in steady-state GLUT4 mRNA in red skeletal muscle while no significant change was found in white skeletal muscle (Fig. 2). Nuclear run-on analysis revealed a similar pattern with STZ-induced diabetes resulting in a 35% reduction in GLUT4 transcription in red skeletal muscle with no change evident in white skeletal muscle (Fig. 3). These findings support the hypothesis that the metabolic effects of STZ treatment exert control on GLUT4 expression in red skeletal muscle primarily at the level of transcription of the GLUT4 gene.
In contrast to diabetes, fasting induced a 2.9-fold increase RED WHITE

FIG. 2. Northern blot analysis and quantification of the relative levels of GLUT4 mRNA in red and white skeletal muscle from control (C), STZ-diabetic ( D ) , and fasted (3') rats. Total
RNA was isolated from portions of red and white quadriceps muscles from control ( C ) , 7-day STZ-diabetic (D), and 3-day fasted (F) rats.
RNA (15 pg) was separated, transferred to nitrocellulose, and hybridized with a GLUT4 cDNA probe as described under "Experimental Procedures." Shown are portions of a representative autoradiogram with each lane representing a single muscle preparation from an individual animal within an experimental group. A summary of the GLUT4 mRNA densitometric analysis ( n = lO/group) is presented as mean k S.E. in the lower portion of the figure. * Significantly ( p < 0.05) different from control. . 3. Quantification of the relative levels of GLUT4 transcription in nuclei isolated from red and white skeletal muscle of control, STZ-diabetic, and fasted rats. Nuclear run-on analysis and quantification of GLUT4 transcripts was performed as described in Fig. 1 and "Experimental Procedures" and presented as mean & S.E. The data are derived from an n = 8 for the red control and red STZ-diabetic groups and an n = 3 for the other four groups.

STZ-Diabetic Fasted
* Significantly ( p < 0.05) different from control. in GLUT4 mRNA in white skeletal muscle with no change evident in red skeletal muscle (Fig. 2). These findings agree well with a recent report by Camps et al. (17) in which GLUT4 mRNA was reported to be increased by over %fold exclusively in white skeletal muscle in 48-h fasted rats. As in the STZdiabetic rats, control of GLUT4 expression was apparently exerted at the level of the gene, as transcription of GLUT4 was increased 2.2-fold in white muscle in response to fasting (Fig. 3). Fasting had no significant effect on GLUT4 transcription in red skeletal muscle, again indicating the presence of fiber type-specific regulation of GLUT4 transcription.
Although fasting appeared to up-regulate transcription of the GLUT1 gene in some experiments (Fig. I), transcription in most cases was too low to accurately quantitate. Given the importance of c-fos and c-jun in the cellular regulation of gene transcription, we examined the steady-state expression of both of these genes in red and white skeletal muscle as well as expression of the transcript encoding for the structural protein @-actin. No consistent change in either c-fos, c-jun, or 8-actin expression was observed in either STZ-diabetic or fasted animals.
Although the data in Fig. 3 is expressed relative to controls with the GLUT4 transcription rate for controls being set to 1, we noted that the absolute level of GLUT4 transcription did not appear to be different between red and white skeletal muscle from control animals. Such a finding would be in direct contrast with previous estimates of -2-5-fold higher GLUT4 protein (17, 20, 21) and -2.0-fold higher GLUT4 mRNA (17,20) in red uersus white skeletal muscle. We therefore compared GLUT4 transcription directly between nuclei isolated from red and white skeletal muscle of an additional set of animals. Despite a 2.2-fold higher steadystate GLUT4 mRNA level in red muscle, transcription of the GLUT4 gene was not different between fiber types (Fig. 4).

DISCUSSION
STZ-induced diabetes markedly decreases both maximal insulin and contraction stimulated glucose transport in skeletal muscle (10). This STZ-induced impairment in glucose transport coincides with a decrease in skeletal muscle GLUT4 protein and mRNA content (13-16), both of which are reversible with insulin treatment (13, 15). We tested the hypothesis that this -50% reduction in GLUT4 mRNA in skeletal muscle is due to an STZ-induced decrease in transcription of the GLUT4 gene. As shown in Fig. 1, our initial experiments using nuclei isolated from whole mixed muscle suggested that STZ treatment had no significant effect on transcription originating from the GLUT4 gene. Skeletal muscle, however, is a heterogeneous tissue composed of red type 1 (slow twitch, oxidative), red type IIa (fast twitch, oxidative/glycolytic), and white type IIb (fast twitch, glycolytic) muscle fibers (31). The diversity between muscle fiber types is further exemplified by the fact that insulin/contraction stimulated glucose transport capacity as well as steadystate GLUT4 protein and mRNA content are inherently higher in red relative to white skeletal muscle (17,20,21). Two recent studies have addressed the possibility that the STZ-induced effects in skeletal muscle may be fiber typespecific (12, 17). By using a protocol similar to the present study, GLUT4 protein content was found to be decreased in both red and, to a smaller extent, white skeletal muscle of STZ-diabetic animals. However, as determined by Northern analysis, a significant reduction in GLUT4 mRNA is evident only in red skeletal muscle of STZ-diabetic rats (Fig. 2) (12,  17), suggesting the presence of fiber type-specific regulatory mechanisms.
To address the possibility that the effects of STZ-induced diabetes on gene transcription may be exerted in a fiber typespecific manner, nuclei were isolated from both red and white skeletal muscle of STZ-treated animals. Transcriptional runon analysis revealed an -35% decrease in GLUT4 transcription specifically in red skeletal muscle in response to STZ treatment (Fig. 3) and could therefore account for the majority of the STZ-induced reduction in GLUT4 mRNA (Fig. 2). This effect appears to be fiber type-specific, as no changes in GLUT4 transcription or mRNA level were detected in white skeletal muscle in response to STZ treatment (Figs. 2 and 3).
Although GLUT4 transcription was reduced with STZinduced diabetes, the magnitude of change could not entirely account for the 50% reduction in GLUT4 mRNA. Likewise, the rate of GLUT4 transcription in the basal state was not higher in red skeletal muscle (Fig. 4) despite a -2-fold difference in GLUT4 mRNA. Thus, the present findings also provide indirect evidence for a post-transcription level of regulation of muscle GLUT4 expression. Other possible sites of control may include post-transcriptional processing of nascent RNA transcripts (polyadenylation, cytoplasmic transport) and/or alterations in GLUT4 mRNA stability. With respect to the former, Kent et al. (32) have recently demonstrated that the fraction of RNA in skeletal muscle present as polyadenylated RNA relative to total RNA is actually greater in skeletal muscle of diabetic rats, presumably due to a disproportionate loss in ribosomal RNA. This observation led to the conclusion that the reduction in ribosomal RNA rather than the availability of mRNA likely represents the major factor limiting overall protein synthesis in the diabetic state (32). Regarding GLUT4 mRNA stability, an alteration in GLUT4 mRNA half-life has been implicated in both the insulin and tumor necrosis factor induced decreases in GLUT4 expression in 3T3-Ll adipocytes (33, 34). Unfortunately, the determination of GLUT4 mRNA half-life in skeletal muscle either in vivo or in vitro under basal or altered metabolic states has not been possible due to the toxicity of transcriptional inhibitors in animals (35,36) and the relatively low level of GLUT4 expression in cultured myocytes (9).
It has been suggested that hyperglycemia plays a major role in the down-regulation of GLUT4 expression in STZ-induced diabetic rats (9,14,15). Down-regulation of glucose transport capacity and glucose transporter expression has been reported in L6 muscle cells in response to increasing media glucose concentration (37), providing evidence that glucotoxicity occurs in muscle cells in culture. GLUT4 expression is also reduced in specifically red skeletal muscle of the male obese Zucker diabetic fatty rat, a model of noninsulin-dependent diabetes mellitus characterized by severe hyperglycemia with relative hypoinsulinemia (38). It should also be noted that both cAMP and insulin have been shown to repress expression of GLUT4 in 40). However, the effects of insulin or counterregulatory hormones on GLUT4 expression may be quite different between adipose and skeletal muscle tissue, particularly in view of the divergent effects fasting has on GLUT4 expression between the two tissues (18).
Previous studies have established that GLUT4 protein expression is elevated 2-&fold in skeletal muscle of fasted rats (15,18). Conflicting results however have been reported regarding the effects of fasting on GLUT4 mRNA levels in skeletal muscle. In a preparation containing both soleus and gastrocnemius muscle, Charron and Kahn (18) clearly demonstrated that GLUT4 mRNA is increased 2-3-fold with fasting. However, in analyzing the effects of fasting exclusively within the soleus muscle, Bourey et al. (15) were unable to detect any change in GLUT4 mRNA despite a >2.5-fold increase in GLUT4 protein, suggesting that the effects of fasting may be differentially regulated between fiber types. As shown in Fig. 2, our findings lend support to this hypothesis in that we found a fasting induced increase in GLUT4 mRNA exclusively within white skeletal muscle, possibly indicating pretranslational regulation. Further, data derived from the nuclear run-on analysis demonstrated that initiation of transcription of the GLUT4 gene was increased 2.2-fold only in white skeletal muscle in response to fasting (Fig. 3). Thus, the induction of GLUT4 transcription appears to account for, in large part, the corresponding 2.9-fold increase in GLUT4 mRNA. These findings are consistent with the basic premise that various metabolic challenges may elicit fundamentally different adaptations in skeletal muscle, depending on the fiber-type distribution, thus implying that distinct cellular mechanisms exist between red and white skeletal muscle which mediate fiber type-specific gene expression.
The cellular events mediating the induction of GLUT4 transcription are unknown. Certainly an attractive hypothesis may be that a fasting induced increase in cellular cAMP triggers the phosphorylation and activation of a DNA binding protein specific for the promotor region of the GLUT4 gene. Such a CAMP-mediated cascade has been described in regard to the transcriptional regulation of the gene encoding the gluconeogenic enzyme phosphoenoylpyruvate carboxykinase (39). Furthermore, increased cellular cAMP has been reported to decrease GLUT4 expression in 3T3-Ll adipocytes (40), a response which mimics the effect of fasting in adipose tissue (18). A DNA binding protein capable of trans-activating the mouse GLUT4 promotor in fully differentiated 3T3-Ll adipocytes has been identified as the CCAAT/enhancer-binding protein (C/EBP) (41). The transcription factor C/EBP (C/ EBPa) is one of at least four family members (CIEBPP, C/ EBP-y, and C/EBPG) which are capable of forming both homoand heterodimers and have been implicated in regulating the expression of a variety of genes involved in cell differentiation and energy metabolism (42,43). Recent evidence has indicated that the activity of a number of C/EBP isoforms is regulated via cAMP and/or calcium-regulated phosphorylation (44)(45)(46), and that at least one family member (C/EBPG) is capable of activating transcription through the cis-acting cAMP response element of certain genes (44). However, whether the regulation of C/EBP phosphorylation and/or activity plays a significant role in regulating gene expression in muscle, either in cell culture or whole tissue, has yet to be determined. We found the transcription rate of C/EBPa and CIEBPP to be both very low and unresponsive to diabetes and fasting (data not shown), suggesting that if GLUT4 transcription is regulated by C/EBP in skeletal muscle, it is occurring through a post-transcriptional mechanism (e.g., phosphorylation).
While it is apparent that altered GLUT4 protein expression in human muscle is not responsible for the insulin resistance associated with NIDDM (47-49), certain interventions such as exercise training (50-54) and muscle creatine depletion (55), in which GLUT4 expression is increased, are also associated with corresponding increases in glucose transport capacity. Indeed, overexpression of GLUT4 in transgenic mice has recently been shown to significantly decrease serum glucose and insulin in both the fed and fasted state (56). Thus it is possible that deciphering the mechanisms regulating GLUT4 expression, particularly in skeletal muscle, may result in the development of strategies to elicit overexpression of the endogenous gene.
In summary, the results of the present study demonstrate that expression of the GLUT4 protein in skeletal muscle is subject to regulation in vivo at the level of transcription of the GLUT4 gene. Transcriptional run-on analysis of nuclei isolated from red and white portions of hindlimb skeletal muscles of STZ-induced diabetic and 3-day fasted rats revealed the presence of fiber-type specific transcriptional regulation of the GLUT4 gene. STZ-induced diabetes resulted in a 35% reduction in GLUT4 transcription and a 50% decrease in mRNA content specifically in red skeletal muscle. In con-trast, fasting elicited a 2.2-fold induction in GLUT^ transcrip-24. Cornelius, P., Marlowe, M., Lee, M. D., and Pekala, P. H. (1990) J. Biol. tion specifically in white skeletal muscle which coincided with 25. Zahradka, P., Larson, D. E., and Sells, B. H. (1989) Exp. Cell Res. 1 8 5

, a
Chem. 265,20506-20516 a 2.9:fold increase white muscle GLUT4 mRNA content. Finally, higher steady-state GLUT4 mRNA levels between red and white skeletal muscle were not associated with fiber type-specific differences in transcription of the GLUT4 gene. These data provide direct evidence that GLUT4 expression in skeletal muscle is subject to regulation in vivo at the level of transcription in a fiber type-specific manner.