Identification of a Skeletal Muscle-specific Regulatory Domain in the Rat GLUT4/Muscle-Fat Gene*

To identify sequences responsible for the muscle- specific expression of the rat GLUT4/muscle-fat gene, we examined the transcriptional regulation of this gene in the differentiating murine C2C 12 skeletal muscle cell line. Differentiated myofibers displayed a 4-6- fold increase in GLUT4 mRNA compared with undifferentiated myoblasts which paralleled the conversion from non-muscle &actin mRNA to muscle-specific a-actin mRNA expression. Transient transfection of pro- gressive 6’ and 3’ deletions of the GLUT4 6”flanking DNA identified a 281-base pair region located between -617 and -237 relative to the transcription start site which conferred myotube-specific expression. This re- gion increased reporter activity in the context of the GLUT4 minimal promoter in an orientation-independ- ent manner and, in addition, onto the heterologous thymidine kinase promoter. Myotube-specific expres- sion of both GLUT4 reporter constructs and the endogenous mouse GLUT4 mRNA was also observed to be thyroid hormone-dependent. Further, cotransfection of reporter constructs containing the 281-base pair GLUT4 differentiation-specific enhancer with the thyroid hormone receptor specifically increased luciferase activity in myotubes approximately 12-fold. Thus, these data demonstrate the presence of a proximal skeletal muscle-specific activation domain that is necessary for both myotube-specific GLUT4 expression and thyroid hormone responsiveness.


Recipient of Research and Career Development Award DK01822
from the National Institutes of Health. To whom correspondence should be addressed Dept. of Physiology and Biophysics, The University of Iowa, Bowen Science Bldg., Iowa City, Iowa 52242. account for some forms of peripheral insulin resistance which are associated with diabetes.
In addition to hormonal/metabolic control, the expression of GLUT4 in cardiac and skeletal muscle is developmentally regulated (13). The predominant glucose transporter in fetal muscle is the GLUTl isoform. However, during late neonatal development and within the first weeks of postnatal life there is a marked decrease in GLUTl levels with a concomitant increase in GLUT4 expression. This pattern of developmental regulation is observed with many muscle-specific genes including those encoding enzymes involved in energy metabolism as well as contractile proteins (14). Although the mechanism responsible for this isoform exchange is unknown, a number of myogenic transcription factors have been described which appear to play a major functional role in the expression of at least some muscle-specific genes. The helix-loop-helix myogenic factors, MyoD (15), myogenin (16, 17), MRF4 (18), and myf- 5 (19), have been shown to bind to specific regulatory elements in several muscle-specific genes including muscle creatine kinase (20, 21), myosin light chain (22), and the acetylcholine receptor a and 6 subunits (23,24). Although the GLUT4 gene contains several E box motifs with the potential to interact with these factors, cotransfection studies have not demonstrated any specific tram-activation by these DNA binding proteins. ' To begin to understand how tissue-specific transcription of the GLUT4 gene is regulated during skeletal muscle development, we have cloned portions of the 5"flanking sequence of the rat GLUT4 gene upstream of the luciferase reporter gene. Since the murine skeletal muscle differentiating cell line C2C12 has been used extensively to analyze musclespecific expression of a number of genes, we utilized this tissue culture model system for transient transfection assays of both 5' and 3' GLUT4 deletion reporter constructs. In this manner we have identified a muscle differentiation-specific activating domain located from -517 to -237 relative to the transcription start site of the rat GLUT4 gene. This region, which contains a thyroid hormone response element (TRE)' homology, displayed myotube-specific activation by the thyroid hormone receptor a1 (TRal) which is consistent with the observed TB induction of endogenous GLUT4 mRNA.

EXPERIMENTAL PROCEDURES
Cell Culture-The C2C12 mouse muscle cell line (American Type Culture Collection) was maintained at subconfluent densities in growth media consisting of Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 25 mM glucose and supplemented with 20% fetal calf serum. The cells were induced to differentiate by exchanging the growth medium for differentiation medium consisting J. M. Richardson and J. E. Pessin, unpublished observations. 'The abbreviations used are: TRE, thyroid response element; TRotl, thyroid hormone receptor a l ; T1, triiodothyronine; kb, kilobase(s); bp, base pair(s). of Dulbecco's modified Eagle's medium with 5.5 mM glucose supplemented with 2% horse serum following cell confluence. HeLa cells were maintained in Dulbecco's modified Eagle's medium containing 25 mM glucose plus 10% fetal calf serum. Where indicated thyroid hormone was removed from serum by treatment with AG 1-X8 ion exchange resin (Bio-Rad) as described previously (25).
Transient Transfection Studies-Proliferating myoblasts were plated 18 h prior to transfection at a density of approximately 7.5 X lo6 or 1.5 X lo6 cells/60-mm plate depending whether the cells were to be harvested as myoblasts or myotubes, respectively. The cells were transiently transfected by the calcium phosphate method (26) with a total of 12 pg of DNA consisting of equimolar amounts of the various GLUT4-luciferase (Luc) reporters brought to 10 pg with pBluescript DNA and 2 pg of the reference plasmid RSV-LacZ. Cell cultures were incubated with the precipitate for 3-4 h and then subjected to a 2-min 20% glycerol shock after which growth medium was reapplied. Twenty-four hours later differentiation medium was applied to cultures to be studied as myotubes. Cells were harvested at 24 h (myoblasts) and 72-96 h (>50% fusion into multinucleated myotubes) after transfection. In some experiments, fully differentiated fused myotube cultures were directly transfected with GLUT4 reporter constructs with similar results. Luciferase and 8-galactosidase activities were determined as described (27,28). HeLa adenocarcinoma cells were transfected in an identical manner and harvested at 24 h. All transfections were performed at least two times, in triplicate, using at least two different preparations of plasmid DNA. To correct for differences in transfection efficiencies between plates within an experiment, the luciferase activity in an extract was normalized to the 8-galactosidase activity. To allow comparison of patterns of expression among myoblasts, myotubes, and HeLa cells, the data are expressed as relative changes in luciferase activity.
In some experiments GLUT4-luciferase reporters (5 pg) were cotransfected with 5 pg of a mammalian expression plasmid containing the rat thyroid hormone receptor a1 (29), a kind gift of Dr. Howard Towle (University of Minnesota). Cotransfection with T R a l was also conducted with TREpal(5 pg), in which a palindromic variant of the rat growth hormone TRE drives transcription from the thymidine kinase promoter (30). Cells were maintained in T3-depleted medium for at least 3 days, and thyroid hormone (50 nM) was added for 12-24 h, as indicated, prior to harvest. RNA Isolation-Total RNA was isolated from cultured cells or mouse skeletal muscle tissue by the method of Chirgwin et al. (31). Poly(A+) RNA was isolated from cultured cells by the method of Badley et al. (32). Briefly, approximately 1 X 10' cells were homogenized in a buffer containing 0.2 M NaC1, 0.2 M Tris, pH 7.5, 1.5 mM MgC12,2% SDS, and 0.2 mg/ml proteinase K followed by incubation at 42 "C for 1 h. The NaCl concentration was increased to 0.5 M, and oligo(dT)-cellulose (Collaborative Research) was added to the lysate which was mixed end-over-end at 25 'C for 1 h. The oligo(dT)cellulose was then washed in buffer containing 0.5 M NaCI, 10 mM Tris, pH 7.5, followed by elution of poly(A+) RNA with water.
Northern Blot Analysis-Five pg of poly(A+) RNA or 5 pg of total RNA was fractionated on a 1% agarose-formaldehyde gel and transferred to nitrocellulose membrane (Millipore). Identical blots were hybridized with either a GLUT4 cRNA probe or an a-actin cDNA probe which recognizes both the a and 8 species. The GLUT4 probe was an antisense RNA derived from the rat GLUT4 cDNA (pSM1-1) which was linearized with EcoRV and transcribed with T7 RNA polymerase (2). Hybridization was carried out at 55 "C for 16 h in 50% deionized formamide, 5 X Denhardt's reagent, 5 X SSPE, 1% SDS and 100 pg/ml denatured salmon sperm DNA. Filters were washed three times in 1 X SSPE, 0.5% SDS followed by one wash in 0.1 X SSPE, 0.5% SDS and then treated with 10 pg/ml RNase A in 2 X SSPE at 37 "C for 15 min. The actin probe was a human muscle a-actin cDNA which was labeled with [(U-~'P]~CTP using a nick translation kit (Amersham Corp.). Hybridization was carried out at 42 "C in 50% formamide, 5 X SSC, 1 X Denhardt's solution, 100 pg/ ml salmon sperm DNA, and 0.02 M Hepes, pH 7.5. Filters were washed three times in 2 X SSC, 0.5% SDS followed by three washes in 0.2 X SSC, 0.5% SDS. The filters were then exposed to Kodak XAR film at -70 "C for 3 days (GLUT4) or 1 h (actin).
Nucleotide numbering indicates the GLUT4 sequence that is included in the GLUT4 luciferase construct. A construct containing 4.3 kb of 5"flanking DNA was generated by isolating the 3.5-kb HindIII fragment representing -813 to -4300 and inserting it into -2212Luc. A series of 3' deletion constructs was generated from -2212Luc by digestion with PuuII (-149) and either BstXI (-237), SmaI (-5011, or HindIII (-813) followed by religation. A similar 3' deletion set was generated from the -813Luc construct. To test for enhancer activity, the GLUT4 promoter EcoNI-BstXI sequence (-517 to -237) was inserted into the unique BarnHI site of TKlOBLuc, derived from TK105 (34), in which transcription of the luciferase gene is driven by the herpes virus thymidine kinase minimal promoter. In addition, the insert was ligated in either orientation into a modified Sac1 site upstream of the GLUT4 sequence in the -149Luc construct ((-5171 -237)149Luc).
Primer Extension Analysis-A synthetic oligonucleotide primer (5'-AGGATAGAATGGCGCCGGGCCTTTCTTTAT-3'), complementary to nucleotides -8 to -37 relative to the translation start site of the luciferase gene coding strand, was end labeled with [y-32P]ATP to a specific activity of 6 Ci/mmol using T4 polynucleotide kinase. The oligonucleotide was annealed as described by Ntambi et al. (35) to 10 pg of poly(A+) RNA from either untransfected C2C12 cells or C2C12 cells transfected with -2212Luc. The primer was extended with avian myloblastosis virus reverse transcriptase, and extended products were analyzed on a 5% acrylamide, 7.5 M urea sequencing gel. A sequencing ladder was generated from the -2212Luc plasmid using the same oligonucleotide as in the primer extension reaction and was employed as a size standard.

GLUT4 Gene
Expression in C2C12 Cells-The GLUT4 glucose transporter is expressed in adult skeletal muscle, heart, and adipose tissue (4). However, in fetal rat skeletal muscle the GLUT4 protein is only detectable following 21 days of gestation at which time levels account for approximately 6% of adult values (13). Subsequently, an increase in GLUT4 expression occurs with a concomitant decrease in GLUT1 levels as development proceeds into the adult state. To determine whether the differentiating murine C2C12 skeletal muscle cell line is an appropriate model for the developmental regulation of GLUT4 expression, steady-state levels of GLUT4 mRNA were determined by Northern blot analysis adult mouse hindquarter skeletal muscle, and the HeLa cell line were analyzed for the presence of GLUT4 (A) and actin ( B ) mRNA before and after treatment with differentiation medium. Poly(A+) RNA was isolated from C2C12 myoblasts ( l a n e I ) , differentiated C2C12 myotubes ( l a n e 2), HeLa cells ( l a n e 4 ) , and HeLa cells treated with differentiation medium ( l a n e 5). As a positive control, total RNA was isolated from mouse hindlimb skeletal muscle ( l a n e 3). Five pg of poly(A+) RNA (cultured cells) or 5 pg of total RNA (mouse skeletal muscle) was fractionated on a 1% agarose formaldehyde gel, transferred to nitrocellulose membrane, and hybridized with either a 32Plabeled antisense GLUT4 RNA probe (A) or a "P-labeled a-actin cDNA probe ( B ) that recognizes both the a-actin (lower band) and @-actin (upper band) mRNA isoforms as described under "Experimental Procedures." This is a representative blot that was performed independently three times. lA, lane 1 ) was increased approximately 4-fold in differentiated myotubes (Fig. IA, lane 2). The amount of GLUT4 mRNA in the myofibers was estimated to be approximately 2-5% of that present in adult mouse skeletal muscle based upon the signal intensity obtained with 5 pg of total mouse hindquarter skeletal muscle RNA (Fig. lA, lane 3). HeLa cells were observed to express substantially lower levels of GLUT4 mRNA (Fig. IA, lane 4 ) which were unchanged by treatment with the same differentiation medium that induced C2C12 myotube formation (Fig. IA, lane 5).
As an independent measure of muscle cell differentiation, an identical blot was hybridized with a probe for actin mRNA (Fig. 1B). This probe recognizes both the muscle-specific aactin (1.5 kb) as well as the non-muscle P-actin species (2.1 kb). Under these conditions, C2C12 myoblasts predominantly expressed &actin mRNA (Fig. lB, lane 1) and following myotube differentiation primarily expressed a-actin mRNA (Fig. lB, lane 2). The switch from &actin to cr-actin expression documents the phenotypic conversion of the C2C12 cells from a non-muscle (myoblast) to muscle (myofiber) state as has been described in developing muscle tissue (36,37) and in myogenic cells in culture (38,39). As expected, only the aactin mRNA species was detected in adult skeletal muscle The 5"Flanking Region of the GLUT4 Gene Confers Myofiber-specific Expression-To determine whether myotube-specific gene expression is mediated by distinct cis-acting elements within the 5'-regulatory region of the rat GLUT4 gene, three luciferase reporter constructs were prepared containing 68, 149, or 2,212 bp of upstream GLUT4 sequences relative to the transcription start site (Fig. 2A)  -2212Luc ( l a n e 1 ) or untransfected C2C12 cells (lane2) andextended with reverse transcriptase as described under "Experimental Procedures.'' The extended products were size fractionated on a 5% acrylamide, 7.5 M urea sequencing gel and exposed to x-ray film. A sequencing ladder (GATC) was generated from -2212Luc using the same primer and was employed as a size standard.
into HeLa cells and C2C12 cells that were maintained in either the myoblast state or differentiated into multinucleated myotubes. In myotubes the -2212Luc reporter displayed a 3.6-fold increase of luciferase reporter activity compared with the -149Luc construct. In contrast, expression of the -149Luc reporter was essentially identical to the expression of -2212Luc in HeLa cells and myoblasts. In all three cell types examined, deletion from -149 to -68 resulted in an additional 95% decrease in transcription activity. These data suggest that a positive regulatory region is present between -2212 and -149 which confers a differentiation-specific increase in GLUT4 transcription in C2C12 cells. An additional activating domain is present between -149 and -68 which is essential for minimal GLUT4 expression in all three cell types. Thus, the minimal GLUT4 promoter can be functionally defined by the -149 5' boundary.
T o determine whether the correct site of transcription initiation was utilized in the reporter constructs, primer extension analysis of the -2212Luc mRNA was performed (Fig.  3). RNA isolated from C2C12 myotubes transfected with -2212Luc yielded several extension products (Fig. 3, lane 1 ) that are not present in RNA from untransfected cells (Fig. 3,  lane 2). The largest and most abundant band (arrow) corresponds to an A/T base pair which represents the transcription initiation site of the mouse (40) and rat3 GLUT4 genes. Identification of a Myofiber-specific GLUT4 Activating Region-To define further the putative muscle-specific activating region within the GLUT4 upstream sequence, a series of 5' deletion constructs were prepared using convenient restriction sites (Fig. 2B), including a construct containing an additional 2100 bp of 5' sequence which extends -4300 bp upstream from the transcription start site. Transient transfection of C2C12 cells with GLUT4 promoter constructs extending from -149 bp to -4300 bp resulted in similar luciferase activities in the undifferentiated myoblasts (Fig. 4). However, myotubes displayed a 3-4-fold enhancement of luciferase activity with the -4300Luc) -2212Luc) -813Luc, and -517Luc reporter constructs compared with the -5OOLuc and -149Luc reporter constructs. These results demonstrate that the pattern of expression in C2C12 myoblasts, which express very low levels of GLUT4 mRNA, was similar for constructs spanning -149 to -4300 bp of GLUT4 upstream sequence. However, a significant increase in transcription was observed in C2C12 myotubes transfected with constructs containing a minimum of 517 bp of flanking sequence, suggesting that an activating domain resides in a region that has -517 as its 5' boundary.
To define the 3' boundary of the myotube-specific activating region, a 3' deletion set was generated which had a shared 5' end at -2212 and included -149 bp of minimal GLUT4 promoter sequence (Fig. 5A). In myotubes, internal deletion of the sequence between -149 and -237 had no significant effect on myotube-specific expression, which was similar to the intact -2212Luc reporter. However, internal deletion of -149 to -501 or -149 to -813 resulted in luciferase activities that were indistinguishable from the -149Luc reporter. Similarly, a 3' deletion set using -812 as the 5' boundary also demonstrated a loss of myotube-specific expression in the -149 to -501 internal deletion (Fig. 5B). In myoblasts, there were no significant differences in the expression of any of the 3' deletion constructs. These data demonstrate the presence of a myotube activating region that has -237 as its 3' boundary. Taken together, the results of the 5' and 3' deletion sets define the activating region as a 281-bp EcoNI-BstXI fragment (-517 to -237).
An enhancer is characterized by its ability to increase transcriptional activity of heterologous promoters in an ori- Relative Luciferase Activity/Pgal Activity FIG. 4. Expression of GLUT4 5' deletion constructs in C2C12 myoblasts and myotubes. A series of 5' deletions of the upstream region of the GLUT4 gene from -4300 to -149 bp relative to the transcription start site was generated as indicated in Fig. 1 and is numbered according to the nucleotides included in the construct. C2C12 cells were transiently transfected with the various constructs, and luciferase activities were determined as described under "Experimental Procedures." The data were expressed as luciferase activity/ &galactosidase activity f S.E. with the -149Luc construct being set to a value of 1. These results were obtained from three to nine transfection experiments, performed in triplicate, using at least two preparations of each DNA construct. entation-and distance-independent manner. To determine if the 281-bp fragment has the characteristics of an enhancer, the EcoNI-BstXI fragment was cloned into a vector containing the thymidine kinase minimal promoter fused to luciferase (TKlO5Luc). In addition, this fragment was also cloned in both orientations in front of its own promoter in the -149Luc vector (Fig. 6). The TK105Luc construct containing the EcoNI-BstXI fragment exhibited 4-fold higher luciferase activity in myofibers compared with the TK105Luc vector alone. This activation was specific for the C2C12 myotubes and did not occur in myoblasts. Similarly (-517/-237)149Luc, the 281-bp region cloned upstream of -149Luc, resulted in a 3-4-fold higher luciferase activity in both the positive and negative orientations compared with the parental -149Luc vector alone. As expected, transcriptional activation by this region only occurred in C2C12 myotubes and was not observed in myoblasts.
GLUT4 Reporter Gene Expression Is Thyroid Hormoneresponsive-Inspection of the proximal GLUT4 promoter revealed a sequence motif having strong identity with TREs described previously (41, 42). Further, digestion with SmaI (-500) bisected this putative element and resulted in reporter constructs that did not display rnyotube-specific expression ( Figs. 4 and 5). We therefore next examined the T3 dependence of GLUT4 reporter expression in the myotube cultures depleted of T3 (Fig. 7). Under these conditions, the C2C12 were morphologically observed to fuse into myotubes as well as to display the appropriate conversion of 8to a-actin (see Fig. 9B). Treatment of these myotube cultures for 24 h with 50 nM T3 resulted in a 4-and 5-fold greater expression of the -517Luc and (-517/-237)149Luc reporter genes, respectively, compared with the -5OOLuc and -149Luc constructs (Fig. 7). This pattern of expression was similar to that observed for these reporter constructs in C2C12 myotubes under standard differentiation conditions (Fig. 6). In contrast, in the absence of T3, the expression of the -517Luc and (-5171 -237)149Luc reporter genes was decreased approximately 2.5fold with no effect on the -5OOLuc and -149Luc reporters.
Since a putative TRE was contained within the myotubespecific 281-bp enhancer domain, C2C12 cultures were cotransfected with an expression vector encoding the rat TRal cDNA and the (-517/-237)149Luc, -517Luc, -5OOLuc, or -149Luc reporter constructs (Fig. 8). In addition, a reporter containing the previously characterized palindromic TREpal, which drives transcription from the TK promoter, was also included as a positive control (30). Expression of T R a l in C2C12 myotubes, either in the presence or absence of T3, had no significant effect on the reporter activity of the -149Luc or TKLuc constructs. In contrast, T R a l in the presence of T3 specifically increased the expression of the (-5171 -237)149Luc and -517Luc reporters approximately 14-and 17-fold, respectively, compared with the -149Luc reporter. T3 treatment of cells increased expression of the TREpal reporter 9-fold compared with the TKLuc reporter.
Regulation of Endogenous GLUT4 mRNA by Thyroid Hormone-To determine whether the TRal regulation of the GLUT4 reporter constructs was physiologically relevant, we next examined the expression of endogenous C2C12 GLUT4 blot relative to an identical blot that was hybridized with a cDNA probe for glyceraldehyde phosphate dehydrogenase. Five pg of poly(A+) RNA was fractionated on a 1% agarose gel, transferred to nitrocellulose membrane, and hybridized with the indicated probes as described under "Experimental Procedures." This is a representative blot that was independently performed three times.
mRNA by T3 (Fig. 9). As observed in Fig. 1, standard differentiation conditions resulted in an approximate 5-fold increase in GLUT4 mRNA in myotubes compared with myotubes (Fig. 9A, lanes 1 and 2). In contrast, differentiation of the myotubes in the absence of T3 (Fig. 9A, lane 3) resulted in a reduced expression of GLUT4 mRNA to levels similar to those observed in the myoblast cultures (Fig. 9A, lane I ) .
However, T3 treatment for 12 or 24 h resulted in a 5-fold induction of GLUT4 mRNA levels (Fig. 9A, lanes 4 and 5 ) , which was equivalent to that observed in the standard myotube cultures (Fig. 9A, lane 2). The increase in myofiber GLUT4 mRNA levels was a direct effect of T3 treatment as the Ts-depleted cultures both morphologically fused into myofibers and demonstrated the appropriate conversion of @-to a-actin (Fig. 9B). Because of the conversion from @-to aactin mRNA, densitometric quantitation of GLUT4 expression was normalized to the invariant gene glyceraldehyde-3phosphate dehydrogenase as indicated in Fig. 9C.

DISCUSSION
To define &-DNA regions that regulate tissue-specific expression of the GLUT4 gene it is necessary to identify a suitable host cell system. The murine 3T3-Ll cell line has been well documented as expressing high levels of GLUT4 mRNA and protein when differentiated into the adipocyte phenotype (6,43). To obtain a GLUT4-expressing muscle cell line, we have screened various cells including BC3H1, L6, L8, and 23A2. Although studies have observed GLUT4 mRNA and protein in the rat L6 and L8 muscle cell lines (44,45), we were unsuccessful in detecting sufficient levels of GLUT4 mRNA for transcriptional analysis in these cells. In contrast, we have observed that multinucleated C2C12 myotubes were found to express GLUT4 mRNA levels approximately 4-5fold greater than in proliferating C2C12 myoblasts. It should also be noted, however, that even following the myotubespecific induction of GLUT4 mRNA, the levels of this transcript were still substantially lower than that observed for in vivo skeletal muscle tissue. Nevertheless, since GLUT4 mRNA levels were induced in a myofiber-specific manner, we reasoned that this cell line would provide a useful system in which to study the muscle differentiation-induced expression of this gene.
Based upon the availability of a suitable model system, we undertook a systematic study to identify sequence regions that control muscle-specific GLUT4 expression. Luciferase reporter constructs containing progressive amounts of the rat GLUT4 5"flanking region were transfected into into C2C12 cells and HeLa cells. Proliferating C2C12 myoblasts and HeLa cells, which do not express appreciable amounts of GLUT4 mRNA, displayed essentially identical luciferase activity whether the reporter constructs contained -149 or -2212 bp of GLUT4 upstream sequence. However, deletion to within 68 bp of the transcription start site resulted in a 95% decrease in luciferase activity in all cell types studied, suggesting that this proximal region from -149 to -68 provides minimal promoter function. In this regard, there are several consensus sequences for general control factors within this region which include a CCAAT homology element at -94 and two GC-rich regions located at -129 and -65. 3 In contrast, a 4-5-fold myotube-specific increase in luciferase activity was observed for the -43OOLuc, -2212Luc, -813Luc, and -517Luc reporter constructs compared with the -5OOLuc and -149Luc constructs. This increase in reporter activity directly correlated with the myotube induction of endogenous GLUT4 mRNA. Importantly, primer extension analysis demonstrated that transcription from the transfected GLUT4-luciferase construct was correctly initiated and therefore represented appropriate transcriptional activation. These data establish that the 5' boundary of the myotube-specific activation region lies within -517 bp of transcription start. Further, 3' deletion analysis has defined the 3' boundary of this myotube activating region to within -237 bp of transcription start. The functional importance of this region was confirmed by its ability to activate myotube-specific expression in an orientation-and promoter-independent manner, and therefore the EcoNI-BstXI promoter region can be classified as an enhancer domain. Thyroid hormone has been well established as modulating metabolic rate and whole body glucose utilization (46, 47) as well as increasing markedly glucose uptake in primary isolated rat skeletal muscle cell cultures (48). In addition, thyroid hormone has been shown to have a positive regulatory effect on GLUT4 mRNA and protein content in the skeletal muscle of hypothyroid rats (49). Consistent with these data, we have observed T3-dependent regulation of both GLUT4 mRNA levels as well as the (-517/-237)149Luc and -517Luc reporters in C2C12 myotubes. It is of interest to note that the effects of T3 on the endogenous myofiber GLUT4 mRNA levels ( Fig. 9) were greater than those observed for the transfected GLUT4-Luc reporter genes (Fig. 7). Although we can not provide a mechanism accounting for this observation at the present time, this phenomenon has been characteristically reported for several other thyroid hormone-responsive genes (50, 51).
Within the rat GLUT4 281-bp myotube-specific enhancer region is the sequence GGGACAggctGGGACAcccGGGCCA which is highly conserved among the rat; mouse (40), and human GLUT4 genes (52). These three hexamer repeat sequences in the GLUT4 gene have close identity with vitamin D, retinoic acid, and thyroid hormone response core element consensus sequences. The thyroid hormone response element consensus sequence proposed by Leidig (G/AGGTCANs-NsG/AGGA/CCN) is composed of two direct hexamer repeats separated by any three to five nucleotides (41). Additional evidence has suggested that the thyroid hormone receptor shows a selective preference for a separation between the AGGTCA-related sequences of four nucleotides (42). The three hexamer repeat motif in the GLUT4 promoter appears to be a reasonable candidate for a functional TRE. In support of this conclusion, both the (-517/-237)149Luc and -517Luc reporters, but not the -5OOLuc and -149Luc constructs, displayed Ts sensitivity. In addition, cotransfection of these reporters with TRal demonstrated specific trans-activation in C2C12 myofibers.
In summary, we have identified a 281-bp region within the GLUT4 promoter which plays a role both in the myotubespecific activation and the thyroid hormone responsiveness of GLUT4 transcription in C2C12 cells. The specific transcription factor interactions that mediate this activation remain to be characterized and will contribute toward an understanding of the mechanisms that control GLUT4 gene expression during skeletal muscle development.