Human GLUT4/Muscle-Fat Glucose-Transporter Gene Characterization and Genetic Variation JOHN B. BUSE, KAZUKI YASUDA, TRACY P. LAY, TRACY S. SEO, ANN LOUISE OLSON, JEFFREY E. PESSIN, JOHN H. KARAM, SUSUMU SEINO, AND GRAEME I. BELL
Four overlapping DNA fragments spanning 32 kb containing the human GLUT4 facilitative glucose-transporter gene were isolated and characterized. The sequence of the GLUT4 gene (-6.3 kb) and 2.0 kb of the promoter region was determined. The sequence of the promoter revealed potential binding sites for transcription factors known to regulate gene expression in muscle cells and adipocytes. However, transfection of constructs including 2 kb of the GLUT4 promoter fused to the bacterial CAT gene into 3T3-L1 adipocytes displayed only weak promoter activity. Because insulin resistance plays a prominent role in the development of NIDDM, genetic variation in the sequence of GLUT4 also was evaluated. Oligonucleotide primer pairs were selected that allowed the protein-coding region of the human GLUT4 gene to be amplified by PCR. The sequence of the protein-coding region of the GLUT4 gene and all intron-exon junctions was determined for a single diabetic Pima Indian and was identical to that of the cloned gene and cDNA. SSCP analysis was used to screen patients with diabetes mellitus and normal, healthy nondiabetic individuals for mutations at the GLUT4 locus. In addition to the silent substitution in the codon for Asn 130 (AAC or AAT) and a Val 383 (GTCHIIe(ATC) replacement described previously, two new variants were identified. One was
From the Departments of Medicine and Biochemistry & Molecular Biology and the Howard Hughes Medical Institute, University of Chicago, Chicago, Illinois; the Department of Physiology and Biophysics, University of Iowa, Iowa City, Iowa; and the Metabolic Research Unit, University of California, San Francisco, California. Address correspondence and reprint requests to John B. Buse, MD, PhD, Department of Medicine, The University of Chicago, 5841 South Maryland Avenue, MC1027, Chicago, IL 60637. Received for publication 4 March 1992 and accepted 7 May 1992. NIDDM, non-insulin-dependent diabetes mellitus; PCR, polymerase chain reaction; RFLP, restriction fragment-length polymorphism; SSCP, singlestranded conformational polymorphism; NIDDK, National Institute of Diabetes and Digestive and Kidney Diseases; IDDM, insulin-dependent diabetes mellitus; UTP, uridine triphosphate; cpm, counts/min; PIPES, piperazine-A/,A/'bis[2-ethanesulfonic acid]; CAT, chloramphenicol acetyl transferase; DMEM, Dulbecco's modified Eagle's medium; FCS, fetal calf serum; HBS, HEPESbuffered saline; PBS, phosphate-buffered saline.
1436
a T->A substitution in intron 1 that was found in 1 of 36 NIDDM patients who were typed for this variant. The second was a Ne38S(ATT)->Thr(A£T) replacement that occurred in 1 normal individual and was not found in any of 676 other normal and diabetic subjects. A large and racially diverse group of normal and diabetic individuals also was screened for the He383 polymorphism. It occurred in both diabetic and nondiabetic subjects. There is no indication from our data that these polymorphisms are associated with NIDDM. Diabetes 41:1436-45, 1992
G
lucose transport in a|l mammalian cells is mediated by a family of facilitative glucose transporters that catalyze the diffusion of glucose across the plasma membrane (reviewed in 1). One member of this family, designated as GLUT4 or the muscle-fat glucose transporter, is expressed at high levels in brown and white adipose tissue and cardiac and skeletal muscle and is primarily responsible for mediating insulin-stimulated glucose uptake by these tissues (2-8). The precise pathophysiological mechanisms responsible for the expression of NIDDM are poorly understood. However, almost all patients with NIDDM and many of their nondiabetic family members demonstrate insulin resistance in peripheral tissues such as muscle and fat, and in liver (9-10). Mutation of GLUT4 or its altered regulation represents a potential cause of the reduced insulin responsiveness in muscle arid adipose tissue. In fact, studies in both animal models and humans suggest that decreased expression of GLUT4 mRNA and protein may be the cause of the insulin resistance in adipose tissue (11-13), and that it may contribute to the maintenance of insulin resistance seen in muscle (14-16). We and our collaborators previously reported three polymorphisms in the human GLUT4 gene. They include a Kpn\ RFLP (17), a silent polymorphism in exon 4a in the codon for Asn130 that can be detected by SSCP analysis
DIABETES, VOL. 41, NOVEMBER 1992
J.B. BUSE AND ASSOCIATES
(18-20), and-a nucleotide substitution in exon 9 that results in the replacement of Val383 with lie (19-20). The substitution in codon 383 has been reported only in patients with NIDDM, suggesting that it may contribute to the development of NIDDM. In this paper, we report the isolation and characterization of the human GLLJT4 gene and its promoter. In addition, we extended the characterization of the polymorphism in codon 383 and report two other sequence variants in the GLUT4 gene. These results will facilitate further studies of the role of GLUT4 in the pathophysiology of diabetes mellitus and of the regulation of its expression in normal and altered metabolic states.
at 30°C for 1 h. Twenty microliters of 10% sodium dodecyl sulfate and 10 jxl of proteinase K (10 mg/ml [Boehringer Mannheim, Indianapolis, IN]) were added, and the mixture was incubated for 30 min at 37°C. The mixture was phenohchloroform extracted, ethanol precipitated, rinsed, and dried. The pellets were resuspended in Sequenase stop buffer and electrophoresed on a 6% acrylamide, 7.5 M urea sequencing gel. Similar results were found with an antisense RNA probe derived from a 482-bp Alu I fragment extending from nucleotide 1728 to 2209. Primer extension was performed as described (25), using a variety of antisense primers, including the following:
RESEARCH DESIGN AND METHODS DNA from an adult diabetic Pima Indian (4011) was isolated from peripheral blood lymphocytes provided by Dr. C. Bogardus (Clinical Diabetes and Nutrition Section, NIDDK, Phoenix, AZ). Patients with NIDDM and IDDM were classified by using standard clinical criteria. The control subjects had normal fasting blood glucose values and no first-degree relative with diabetes mellitus (21, 22). Standard methods were conducted as described in Sambrook et al. (23) and as described previously (6). DNA sequencing was done by the dideoxynucleotide chain-termination procedure after subcloning appropriate DNA fragments into M13mp18 or M13mp19. Several segments were sequenced by using Sequenase (United States Biochemical, Cleveland, OH) with specific oligonucleotide primers and alkali-denatured doublestranded plasmjd templates. The sequence was confirmed on both strands and across all junctions. Isolation of the human GLUT4 gene. Human GLUT4 gene clones were isolated from the partial Hae \W-Alu I fetal human liver library of Lawn et al. (24) by highstringency hybridization with the human GLUT4 probes, phJHT-3 and phAMT-6 (6), as described previously (25). The positions of exon-intron junctions were determined by comparison of the genomic and cDNA (6) sequences, using the GT-AG rule (26). Localizing the transcription-initiation site. A genomic fragment of the human GLUT4 gene extending from bp 1709 to 2033 was cloned as a Pst \-Pvu II fragment into Sma \-Pst l-digested pBluescript (SK-); Stratagene, La Jolla, CA). To generate an antisense RNA probe, the plasmid was linearized with EcoRI and transcription driven from the T3 promoter in the presence of a-[32P]UTP. Ten micrograms of RNA isolated from cadaveric human heart or liver (National Disease Research Interchange, Philadelphia, PA) was coprecipitated with 2 x 106 cpm of probe, rinsed with 70% ethanol, and dried. The pellet was resuspended in 30 |xl of hybridization buffer (80% formamide, 0.4 M NaCI, 40 mM PIPES [pH 6.4], 1 mM EDTA) and then heated at 80°C for 10 min before it was hybridized overnight at 57°C. The reaction mixture was diluted in 270 (xl of RNase digestion buffer (300 mM NaCI, 10 mM Tris-HCI [pH 7.4], 5 mM EDTA, 2 ng/ml RNase T1 [GIBCO-BRL, Gaithersburg, MD], and 40 |xg/ml RNase A [Sigma, St. Louis, MO]) and incubated
2073AAGCCAAGGATCCGGAGAGCAGTGG2048, 2O9oATGGGACCCACAGCCACAAGCCAAGGATCCGGAGAGC2O54,
DIABETES, VOL. 41, NOVEMBER 1992
2118CCCGGAGTGACGTGCGAGGGCGGCCCCGATGGGACCC2082, 2171CTGGAGTCTCTGACTCCGGCCTGGA2147, 2226TCTCGTCTTAGAAGAGCTGGA2206.
Expression of human GLUT4 promoter-CAT constructs in 3T3-L1 adipocytes. A 2.4-kb Apa I fragment extending from —300 bp upstream of the sequence presented in this report to nucleotide 2095 was isolated, polished with T4 DNA polymerase, and subcloned into the Bgl II site of vector pCAT3M (27) to generate pApalCAT. Likewise, a 342-bp BamHI fragment (nucleotides 1722-2064) was ligated into the Bgl II site of pCAT3M to generate pBamHI-CAT. Plasmids containing these promoter constructs were isolated in both the sense (+) and antisense (-) orientation. The plasmid pSV2CAT (28), which contains the SV40 early promoter-enhancer region upstream of the CAT gene, was used as a positive control. 3T3-L1 preadipocytes (ATTC, Rockville, MD) were maintained in DMEM plus 10% FCS (GIBCO-BRL). Two days after reaching confluence, cells were treated with 0.5 mM isobutylmethylxanthine, 1 |xM dexamethasone, and 1 n-g/ml porcine insulin (all from Sigma) in DMEM plus 10% FCS (29). After 48 h, the media was changed to DMEM plus 10% FCS and 1 fjug/ml insulin. Within 24 h, >90% of the cells had developed visible lipid dropjets. At this stage, cells were transfected with CAT constructs by using a modification of a method provided by Dr. Maria Alexander-Bridges. Briefly, cells from one T75 flask were removed by gentle trypsinization and washed with 1 x HBS (25 mM HEPES [pH 7.1], 140 mM NaCI, and 0.74 mM Na2HPO4) and aliquoted into four 15-ml conical centrifuge tubes. The following procedure was used to transfect the cells in each of these tubes. Twenty micrograms of supercoiled plasmid DNA was ethanol precipitated, air dried, and dissolved in 0.5 ml of 1.0 M CaCI2. This solution was added by drop to 0.5 ml of freshly prepared 2x HBS. The DNA-calcium precipitant was incubated at room temperature for 15 min and then used to resuspend the cells. After 15 min, 11 ml of DMEM was added, and the cells were cultured in a 100-mm2 petri dish at 37°C for 6 h. The media was removed, and the cells were shocked with a solution of 30% dimethylsulfoxide in PBS for 2 min. Then, they were rinsed with PBS and cultured in DMEM plus 10% FCS
1437
HUMAN GLUT4 GENE
TABLE 1 Sequence of PCR primer pairs used to amplify exons of human GLUT4 gene Exon 1 2 3 4a 4b 5 6 6-5' 6-3' 7 8 9 9-5' 9-3' 10 10-5' 10-3'
Sense upstream primer
Antisense downstream primer
2159TCAGAGACTCCAGGATCGGT2178
2430GGATTGGA
3488ATCATGGTTCCATGTGACAT3507
3746GACAGATTTCCTGCAGACCG3727
ATCTCACAGTCA241,
3726ACGGTCTGCAGGAAATCTGT3745
3985AGCAGTGGCTCCCTTCCTGT3966
3966ACAGGAAGGGAGCCACTGCT3985
4215TGAAAGCGCAGGCATGGTGA4196
4196TCACCATGCCTGGGCTTTCA4215
ACCACTG441, AGAT 4716 TGCTCCCCACTCtCGTCTA s o g s s o o y 4946AGCTGCTGGCTCAGCTGCAG4927 5025TGCTCCCCACTCTCGTCTA5007 5308GGCAACCAAGAGCAGGACCT5289 5543CAGAGACTCACAGGACCATT5524 6264GGTGGGCTAGCTGTGTGATG6245 6115GTCCCTGGCTGAAGAGCTCG6096 6264GGTGGGCTAGCTGTGTGATG6245 7002AGAGGGTTAAAGTGCTGCAG6983 6840TGGAAGGCAGCTGAGATCTG6821 7002AGAGGGTTAAAGTGCTGCAG6983
4461CAGAAGGGCTGAGTGACCTG4480 4717TCTTGCTAGCACCTGGCTTC4736 4717TCTTGCTAGCACCTGGCTTC4736 4816CTGGCTGAGCTGAAGGATGA4835 5064TGTGGCTGGAGTAGAGGAAG5083 5350TGGGGTCAAGCTCCGACTCT5369 5949 AGAAAGGCCTCAACTGGATT 5968 5949 AGAAAGGCCTCAACTGGATT 5968 6096CGAGCTCTTCAGCCAGGGAC6115
6640TAGCTCCTGGTTGCCTGA6657 6640TAGCTCCTGGTTGCCTGA6657 6821CAGATCTCAGCTGCCTTCCA6840
4430GAGCCCCACTCTA
4735AAGCCAGGTGCTAQCA
Fragment size 271 258 259 249 234 274 308 229 209 244 193 315 166 168 362 200 181
All primer sequences are reported in the 5'-to-3' orientation. Numbers flanking the oligonucleotide sequence denote the nucleotide position at which the sequence begins and ends. Because exons 6, 9, and 10 were >300 bp in length, sequencing across the entire span was difficult. Therefore, internal primers wjthin the exon were designed both for sequencing and to improve the sensitivity of SSCP. For detecting polymorphism at codon 383 by SSCP, the primer pair 9 - 5 ' was used, though polymorphism could be detected by amplifying entire exon as well.
and 1 |xg/ml insulin for 48 h. To assay CAT activity, cells were rinsed with PBS, harvested with a rubber policeman, pelleted in a microfuge tube, and then resuspended in 150 |xl of 0.25 M Tris-HCI (pH 7.36). Cell suspensions were subjected to 3 rounds of freezethawing (dry ice and 37°C). The cellular debris was removed by centrifugation in a microfuge at 14,000 rpm for 10 min at room temperature. The CAT activity present in 10 |xl was determined as described by Gorman (30). The acetylated derivatives of [14C]chloramphenicol (Du Pont-NEN, Boston, MA) were separated by thin-layer chromatography. Amplification and sequencing of the protein-coding regions and exon-intron junctions of the GLUT4 gene. Amplification of genomic DNA and direct sequencing of the PQR products was conducted essentially as described by Kadowaki et al. (31). Briefly, PCR was performed in a volume of 100 IJLI containing 10 nnM Tris-HCI (pH 8.3); 50 mM KCI; 1.5 mM MgCI2; 0.01% gelatin; 200 IJLM each of dATP, dGTP, dCTP, and dTTP; 1.0 U of AmpliTaq DNA polymerase (Perkin-Elmer Cetus, Norwalk, CT); 0.5 |xg of genomic DNA; and 100 pmol of each oligonucleotide primer (Table 1). For amplification of exon 1, the cycle conditions were denaturation at 94°C for 1 min; annealing at 55°C for 1 min; and extension at 72°C for 2 min. For the other exons, a rapid two cycle amplification procedure was used: denaturation at 94°C for 1 min and extension at 65°C for 2 min. In both cases, 35 cycles of amplification were used. To prepare singlestranded template for DNA sequencing, a second PCR reaction was performed (35 cycles; denaturation at 94°C for 1 min; annealing at 55°C for 1 min; and extension at 72°C for 2 min) using the solution described above, except 0.5-2.0 |il of the initial PCR reaction product was used as template, and 50 pmol of only one of the two primers was included. The single-stranded products of
1438
the second PCR reaction were purified by Centricon 100 (Amicon, Danvers, MA) exclusion chromatography and then sequenced with Sequenase and a 32P-labeled primer. Because of the lengths of exons 6, 9, and 10, additional primer pairs listed in Table 1 also were used for sequencing. Variants identified by direct sequencing of PCR products were confirmed by cloning the PCR products from the first PCR reaction into M13mp18 and performing standard dideoxynucleotide sequencing on multiple templates in each orientation. All sequences reported here have been confirmed in both the sense and antisense orientations. SSCP analysis. SSCP was performed by using the method of Orita et al. (32,33). Briefly, 50 ng of genomic DNA was subjected to amplification by PCR as described above, except that the reaction volume was only 10 JJLI,' and 0.2 |xCi [a-32P]dCTP (3000 Ci/mmol [Amersham, Arlington Heights, IL]) was included in the reaction mixture. The primer pairs listed in Table 1 were used for SSCP. However, because of the large sizes of exons 6, 9, and 10, SSCP analysis also was performed with additional primer pairs as above. The 10-|xl PCR reaction was mixed with 40 |xl of 10 mM Tris-HCI (pH 8.0) and 1 mM EDTA, and then extracted with CHCI3 to remove the mineral oil. Five microliters of the diluted reaction mixture then was combined with 50 jil of a solution of 0.1% sodium dodecyl sulfate and 1 mM EDTA (pH 8.0). Two microliters of this solution was mixed with 2 |xl of a solution of 96% formamide, 20 mM EDTA (pH 8.0), 0.06% bromophenol blue, and 0.06% xylene cyanol, then was heated at 94°C for 3 min, and immediately loaded on a 3 8 x 3 1 x 0.04 cm 5% polyacrylamide (49:1, acrylamide:A/,A/'-methylene-bis-acrylamide) gel in 90 mM Tris-borate (pH 8.3), 2 mM EDTA, with or without 5% glycerol. Glycerol-containing gels were run at 20 W constant power at room temperature for 4 - 6 h, and gels
DIABETES, VOL. 41, NOVEMBER 1992
J.B. BUSE AND ASSOCIATES
Exon
H
10.4 0.8 ,
1
[>
3 4b 6 8 2 4a 5 7 9
HMHKMltHHZZh
2.4
6.2
2.4
7.4
0.8,
10
XhGLUT4-6 i
7.4
4.5
4.5 3.5
4.5
XhGLUT4-3 3.8
MGLUT4-1 6.5
,0.7
XhGLUT4-2
FIG. 1. Map of the human GLUT4-muscle-fat faciiitative glucose-transporter gene. Positions of exons and the structures of the four overlapping segments of the gene are shown. Natural EcoRI restriction sites are noted by vertical lines. Sizes (in kb) of EcoRI fragments in each cloned segment are Indicated. Please note that the 10.4-kb EcoRI fragment includes several internal EcoRI sites that have not been ordered completely.
without glycerol were run at 2 W constant power at room temperature for 9-12 h. After electrophoresis, gels were exposed to X-ray film with an intensifying screen for 12-18 hat -80°C. RESULTS Isolation and characterization of the human GLUT4 gene. The human GLUT4 gene and flanking regions were isolated as a series of overlapping DNA segments that span a region of 32 kb and include the gene, 15 kb of the 5'-flanking, and 11 kb of 3'-flanking DNA, respectively (Fig. 1). The sequence of a 8.4-kb region was determined (Fig. 2) and indicated that the human GLUT4 gene comprises 11 exons. The exon-intron organization is identical to that previously reported for the corresponding mouse gene (34). Moreover, it is similar to that of the human GLUT1 gene (25), except that exon 4 of GLUT1 corresponds to two exons in GLUT4 that we call exons 4a and 4b. A comparison of the positions of the exon-intron junctions between the GLUT1 and GLUT4 genes indicates that, except for intron 1, they are in identical locations. The start of transcription of the human GLUT4 gene was determined by RNase-protection and primer-extension strategies. The RNase-protection study (Fig. 3) localized the 5' end of the mRNA to two closely spaced regions separated by about 25 nucleotides. Primerextension studies suggest multiple transcription initiation sites (data not shown) centered about nucleotide 2033 (Fig. 2), which corresponds to the major transcriptional start site of the mouse GLUT4 gene (34). The sequence of the promoter indicates that the human GLUT4 gene lacks a TATA motif, but does contain binding sites for several other transcription factors (Fig. 2) including AP2 (35-36), SP1 (37), ETF (38), and the CCAAT binding factor, NF-1 (39-40). Also, note that consensus-binding sites for proteins activate gene expression in myocytes (MEF-2 [41]) and adipocytes (FSE2 [42]). These cisacting elements are all conserved between the human and mouse GLUT4 promoter sequences and are shown in Fig. 2. By contrast, the C/EBP binding site localized in the mouse GLUT4 promoter by DNA footprinting (34) is not present in the promoter region of the human gene. The location of the 3' end of the human GLUT4 gene is
DIABETES, VOL. 41, NOVEMBER 1992
unknown because none of the cDNA clones that were isolated contained a polyA tract. A short region of homology is found between the sequences preceding the polyadenylation sites of the mouse and rat GLUT4 transcripts (3-5) and the region around nucleotide 8076 in the human gene. However, RNA-blotting studies indicate that GLUT4 mRNA is 3.5 kb, which, if the polyA represents 200 nucleotides of this, suggests that the 3' end of the gene is located in a region about 100 bp downstream of the 3' end of the sequence shown in Fig. 2. There are two members of the Alu family (43) of repetitive sequences in the region of the human GLUT4 gene. One is located in the 5'-flanking region, about 1.8 kb upstream of the putative transcriptional initiation site, and the other is located 0.5 kb downstream of the translation termination codon. This latter Alu sequence is predicted to be in the 3'-untranslated region of the mRNA. Regarding microsatellite DNA polymorphisms, we tested the genomic clones for the presence of CA/GT repeats (44,45), and none were noted. Functional characterization of the promoter region of the GLUT4 gene. As described in METHODS, constructs were prepared in which either 2.4 (pApal-CAT) or 342 bp (pBamHI-CAT) of 5'-flanking-promoter region of the human GLUT4 gene were fused to a CAT reporter gene in the sense (+) and antisense (-) orientations. Plasmid DNA from these constructs, the CAT vector (pCAT3M), and a positive control vector containing the SV40 early promoter-enhancer fused to the CAT gene (pSV2CAT) were transfected into differentiated 3T3-L1 adipocytes. As shown in Fig. 4, the 2.4-kb and 342-bp fragments in the sense orientation had similar levels of promoter activity, which were higher than the levels produced by the same fragments in the antisense orientation. However, promoter activity of the human GLUT4 gene constructs was very low, amounting to
Four overlapping DNA fragments spanning 32 kb containing the human GLUT4 facilitative glucose-transporter gene were isolated and characterized. The sequence of the GLUT4 gene (-6.3 kb) and 2.0 kb of the promoter region was determined. The sequence of the promoter revealed potential binding sites for transcription factors known to regulate gene expression in muscle cells and adipocytes. However, transfection of constructs including 2 kb of the GLUT4 promoter fused to the bacterial CAT gene into 3T3-L1 adipocytes displayed only weak promoter activity. Because insulin resistance plays a prominent role in the development of NIDDM, genetic variation in the sequence of GLUT4 also was evaluated. Oligonucleotide primer pairs were selected that allowed the protein-coding region of the human GLUT4 gene to be amplified by PCR. The sequence of the protein-coding region of the GLUT4 gene and all intron-exon junctions was determined for a single diabetic Pima Indian and was identical to that of the cloned gene and cDNA. SSCP analysis was used to screen patients with diabetes mellitus and normal, healthy nondiabetic individuals for mutations at the GLUT4 locus. In addition to the silent substitution in the codon for Asn 130 (AAC or AAT) and a Val 383 (GTCHIIe(ATC) replacement described previously, two new variants were identified. One was
From the Departments of Medicine and Biochemistry & Molecular Biology and the Howard Hughes Medical Institute, University of Chicago, Chicago, Illinois; the Department of Physiology and Biophysics, University of Iowa, Iowa City, Iowa; and the Metabolic Research Unit, University of California, San Francisco, California. Address correspondence and reprint requests to John B. Buse, MD, PhD, Department of Medicine, The University of Chicago, 5841 South Maryland Avenue, MC1027, Chicago, IL 60637. Received for publication 4 March 1992 and accepted 7 May 1992. NIDDM, non-insulin-dependent diabetes mellitus; PCR, polymerase chain reaction; RFLP, restriction fragment-length polymorphism; SSCP, singlestranded conformational polymorphism; NIDDK, National Institute of Diabetes and Digestive and Kidney Diseases; IDDM, insulin-dependent diabetes mellitus; UTP, uridine triphosphate; cpm, counts/min; PIPES, piperazine-A/,A/'bis[2-ethanesulfonic acid]; CAT, chloramphenicol acetyl transferase; DMEM, Dulbecco's modified Eagle's medium; FCS, fetal calf serum; HBS, HEPESbuffered saline; PBS, phosphate-buffered saline.
1436
a T->A substitution in intron 1 that was found in 1 of 36 NIDDM patients who were typed for this variant. The second was a Ne38S(ATT)->Thr(A£T) replacement that occurred in 1 normal individual and was not found in any of 676 other normal and diabetic subjects. A large and racially diverse group of normal and diabetic individuals also was screened for the He383 polymorphism. It occurred in both diabetic and nondiabetic subjects. There is no indication from our data that these polymorphisms are associated with NIDDM. Diabetes 41:1436-45, 1992
G
lucose transport in a|l mammalian cells is mediated by a family of facilitative glucose transporters that catalyze the diffusion of glucose across the plasma membrane (reviewed in 1). One member of this family, designated as GLUT4 or the muscle-fat glucose transporter, is expressed at high levels in brown and white adipose tissue and cardiac and skeletal muscle and is primarily responsible for mediating insulin-stimulated glucose uptake by these tissues (2-8). The precise pathophysiological mechanisms responsible for the expression of NIDDM are poorly understood. However, almost all patients with NIDDM and many of their nondiabetic family members demonstrate insulin resistance in peripheral tissues such as muscle and fat, and in liver (9-10). Mutation of GLUT4 or its altered regulation represents a potential cause of the reduced insulin responsiveness in muscle arid adipose tissue. In fact, studies in both animal models and humans suggest that decreased expression of GLUT4 mRNA and protein may be the cause of the insulin resistance in adipose tissue (11-13), and that it may contribute to the maintenance of insulin resistance seen in muscle (14-16). We and our collaborators previously reported three polymorphisms in the human GLUT4 gene. They include a Kpn\ RFLP (17), a silent polymorphism in exon 4a in the codon for Asn130 that can be detected by SSCP analysis
DIABETES, VOL. 41, NOVEMBER 1992
J.B. BUSE AND ASSOCIATES
(18-20), and-a nucleotide substitution in exon 9 that results in the replacement of Val383 with lie (19-20). The substitution in codon 383 has been reported only in patients with NIDDM, suggesting that it may contribute to the development of NIDDM. In this paper, we report the isolation and characterization of the human GLLJT4 gene and its promoter. In addition, we extended the characterization of the polymorphism in codon 383 and report two other sequence variants in the GLUT4 gene. These results will facilitate further studies of the role of GLUT4 in the pathophysiology of diabetes mellitus and of the regulation of its expression in normal and altered metabolic states.
at 30°C for 1 h. Twenty microliters of 10% sodium dodecyl sulfate and 10 jxl of proteinase K (10 mg/ml [Boehringer Mannheim, Indianapolis, IN]) were added, and the mixture was incubated for 30 min at 37°C. The mixture was phenohchloroform extracted, ethanol precipitated, rinsed, and dried. The pellets were resuspended in Sequenase stop buffer and electrophoresed on a 6% acrylamide, 7.5 M urea sequencing gel. Similar results were found with an antisense RNA probe derived from a 482-bp Alu I fragment extending from nucleotide 1728 to 2209. Primer extension was performed as described (25), using a variety of antisense primers, including the following:
RESEARCH DESIGN AND METHODS DNA from an adult diabetic Pima Indian (4011) was isolated from peripheral blood lymphocytes provided by Dr. C. Bogardus (Clinical Diabetes and Nutrition Section, NIDDK, Phoenix, AZ). Patients with NIDDM and IDDM were classified by using standard clinical criteria. The control subjects had normal fasting blood glucose values and no first-degree relative with diabetes mellitus (21, 22). Standard methods were conducted as described in Sambrook et al. (23) and as described previously (6). DNA sequencing was done by the dideoxynucleotide chain-termination procedure after subcloning appropriate DNA fragments into M13mp18 or M13mp19. Several segments were sequenced by using Sequenase (United States Biochemical, Cleveland, OH) with specific oligonucleotide primers and alkali-denatured doublestranded plasmjd templates. The sequence was confirmed on both strands and across all junctions. Isolation of the human GLUT4 gene. Human GLUT4 gene clones were isolated from the partial Hae \W-Alu I fetal human liver library of Lawn et al. (24) by highstringency hybridization with the human GLUT4 probes, phJHT-3 and phAMT-6 (6), as described previously (25). The positions of exon-intron junctions were determined by comparison of the genomic and cDNA (6) sequences, using the GT-AG rule (26). Localizing the transcription-initiation site. A genomic fragment of the human GLUT4 gene extending from bp 1709 to 2033 was cloned as a Pst \-Pvu II fragment into Sma \-Pst l-digested pBluescript (SK-); Stratagene, La Jolla, CA). To generate an antisense RNA probe, the plasmid was linearized with EcoRI and transcription driven from the T3 promoter in the presence of a-[32P]UTP. Ten micrograms of RNA isolated from cadaveric human heart or liver (National Disease Research Interchange, Philadelphia, PA) was coprecipitated with 2 x 106 cpm of probe, rinsed with 70% ethanol, and dried. The pellet was resuspended in 30 |xl of hybridization buffer (80% formamide, 0.4 M NaCI, 40 mM PIPES [pH 6.4], 1 mM EDTA) and then heated at 80°C for 10 min before it was hybridized overnight at 57°C. The reaction mixture was diluted in 270 (xl of RNase digestion buffer (300 mM NaCI, 10 mM Tris-HCI [pH 7.4], 5 mM EDTA, 2 ng/ml RNase T1 [GIBCO-BRL, Gaithersburg, MD], and 40 |xg/ml RNase A [Sigma, St. Louis, MO]) and incubated
2073AAGCCAAGGATCCGGAGAGCAGTGG2048, 2O9oATGGGACCCACAGCCACAAGCCAAGGATCCGGAGAGC2O54,
DIABETES, VOL. 41, NOVEMBER 1992
2118CCCGGAGTGACGTGCGAGGGCGGCCCCGATGGGACCC2082, 2171CTGGAGTCTCTGACTCCGGCCTGGA2147, 2226TCTCGTCTTAGAAGAGCTGGA2206.
Expression of human GLUT4 promoter-CAT constructs in 3T3-L1 adipocytes. A 2.4-kb Apa I fragment extending from —300 bp upstream of the sequence presented in this report to nucleotide 2095 was isolated, polished with T4 DNA polymerase, and subcloned into the Bgl II site of vector pCAT3M (27) to generate pApalCAT. Likewise, a 342-bp BamHI fragment (nucleotides 1722-2064) was ligated into the Bgl II site of pCAT3M to generate pBamHI-CAT. Plasmids containing these promoter constructs were isolated in both the sense (+) and antisense (-) orientation. The plasmid pSV2CAT (28), which contains the SV40 early promoter-enhancer region upstream of the CAT gene, was used as a positive control. 3T3-L1 preadipocytes (ATTC, Rockville, MD) were maintained in DMEM plus 10% FCS (GIBCO-BRL). Two days after reaching confluence, cells were treated with 0.5 mM isobutylmethylxanthine, 1 |xM dexamethasone, and 1 n-g/ml porcine insulin (all from Sigma) in DMEM plus 10% FCS (29). After 48 h, the media was changed to DMEM plus 10% FCS and 1 fjug/ml insulin. Within 24 h, >90% of the cells had developed visible lipid dropjets. At this stage, cells were transfected with CAT constructs by using a modification of a method provided by Dr. Maria Alexander-Bridges. Briefly, cells from one T75 flask were removed by gentle trypsinization and washed with 1 x HBS (25 mM HEPES [pH 7.1], 140 mM NaCI, and 0.74 mM Na2HPO4) and aliquoted into four 15-ml conical centrifuge tubes. The following procedure was used to transfect the cells in each of these tubes. Twenty micrograms of supercoiled plasmid DNA was ethanol precipitated, air dried, and dissolved in 0.5 ml of 1.0 M CaCI2. This solution was added by drop to 0.5 ml of freshly prepared 2x HBS. The DNA-calcium precipitant was incubated at room temperature for 15 min and then used to resuspend the cells. After 15 min, 11 ml of DMEM was added, and the cells were cultured in a 100-mm2 petri dish at 37°C for 6 h. The media was removed, and the cells were shocked with a solution of 30% dimethylsulfoxide in PBS for 2 min. Then, they were rinsed with PBS and cultured in DMEM plus 10% FCS
1437
HUMAN GLUT4 GENE
TABLE 1 Sequence of PCR primer pairs used to amplify exons of human GLUT4 gene Exon 1 2 3 4a 4b 5 6 6-5' 6-3' 7 8 9 9-5' 9-3' 10 10-5' 10-3'
Sense upstream primer
Antisense downstream primer
2159TCAGAGACTCCAGGATCGGT2178
2430GGATTGGA
3488ATCATGGTTCCATGTGACAT3507
3746GACAGATTTCCTGCAGACCG3727
ATCTCACAGTCA241,
3726ACGGTCTGCAGGAAATCTGT3745
3985AGCAGTGGCTCCCTTCCTGT3966
3966ACAGGAAGGGAGCCACTGCT3985
4215TGAAAGCGCAGGCATGGTGA4196
4196TCACCATGCCTGGGCTTTCA4215
ACCACTG441, AGAT 4716 TGCTCCCCACTCtCGTCTA s o g s s o o y 4946AGCTGCTGGCTCAGCTGCAG4927 5025TGCTCCCCACTCTCGTCTA5007 5308GGCAACCAAGAGCAGGACCT5289 5543CAGAGACTCACAGGACCATT5524 6264GGTGGGCTAGCTGTGTGATG6245 6115GTCCCTGGCTGAAGAGCTCG6096 6264GGTGGGCTAGCTGTGTGATG6245 7002AGAGGGTTAAAGTGCTGCAG6983 6840TGGAAGGCAGCTGAGATCTG6821 7002AGAGGGTTAAAGTGCTGCAG6983
4461CAGAAGGGCTGAGTGACCTG4480 4717TCTTGCTAGCACCTGGCTTC4736 4717TCTTGCTAGCACCTGGCTTC4736 4816CTGGCTGAGCTGAAGGATGA4835 5064TGTGGCTGGAGTAGAGGAAG5083 5350TGGGGTCAAGCTCCGACTCT5369 5949 AGAAAGGCCTCAACTGGATT 5968 5949 AGAAAGGCCTCAACTGGATT 5968 6096CGAGCTCTTCAGCCAGGGAC6115
6640TAGCTCCTGGTTGCCTGA6657 6640TAGCTCCTGGTTGCCTGA6657 6821CAGATCTCAGCTGCCTTCCA6840
4430GAGCCCCACTCTA
4735AAGCCAGGTGCTAQCA
Fragment size 271 258 259 249 234 274 308 229 209 244 193 315 166 168 362 200 181
All primer sequences are reported in the 5'-to-3' orientation. Numbers flanking the oligonucleotide sequence denote the nucleotide position at which the sequence begins and ends. Because exons 6, 9, and 10 were >300 bp in length, sequencing across the entire span was difficult. Therefore, internal primers wjthin the exon were designed both for sequencing and to improve the sensitivity of SSCP. For detecting polymorphism at codon 383 by SSCP, the primer pair 9 - 5 ' was used, though polymorphism could be detected by amplifying entire exon as well.
and 1 |xg/ml insulin for 48 h. To assay CAT activity, cells were rinsed with PBS, harvested with a rubber policeman, pelleted in a microfuge tube, and then resuspended in 150 |xl of 0.25 M Tris-HCI (pH 7.36). Cell suspensions were subjected to 3 rounds of freezethawing (dry ice and 37°C). The cellular debris was removed by centrifugation in a microfuge at 14,000 rpm for 10 min at room temperature. The CAT activity present in 10 |xl was determined as described by Gorman (30). The acetylated derivatives of [14C]chloramphenicol (Du Pont-NEN, Boston, MA) were separated by thin-layer chromatography. Amplification and sequencing of the protein-coding regions and exon-intron junctions of the GLUT4 gene. Amplification of genomic DNA and direct sequencing of the PQR products was conducted essentially as described by Kadowaki et al. (31). Briefly, PCR was performed in a volume of 100 IJLI containing 10 nnM Tris-HCI (pH 8.3); 50 mM KCI; 1.5 mM MgCI2; 0.01% gelatin; 200 IJLM each of dATP, dGTP, dCTP, and dTTP; 1.0 U of AmpliTaq DNA polymerase (Perkin-Elmer Cetus, Norwalk, CT); 0.5 |xg of genomic DNA; and 100 pmol of each oligonucleotide primer (Table 1). For amplification of exon 1, the cycle conditions were denaturation at 94°C for 1 min; annealing at 55°C for 1 min; and extension at 72°C for 2 min. For the other exons, a rapid two cycle amplification procedure was used: denaturation at 94°C for 1 min and extension at 65°C for 2 min. In both cases, 35 cycles of amplification were used. To prepare singlestranded template for DNA sequencing, a second PCR reaction was performed (35 cycles; denaturation at 94°C for 1 min; annealing at 55°C for 1 min; and extension at 72°C for 2 min) using the solution described above, except 0.5-2.0 |il of the initial PCR reaction product was used as template, and 50 pmol of only one of the two primers was included. The single-stranded products of
1438
the second PCR reaction were purified by Centricon 100 (Amicon, Danvers, MA) exclusion chromatography and then sequenced with Sequenase and a 32P-labeled primer. Because of the lengths of exons 6, 9, and 10, additional primer pairs listed in Table 1 also were used for sequencing. Variants identified by direct sequencing of PCR products were confirmed by cloning the PCR products from the first PCR reaction into M13mp18 and performing standard dideoxynucleotide sequencing on multiple templates in each orientation. All sequences reported here have been confirmed in both the sense and antisense orientations. SSCP analysis. SSCP was performed by using the method of Orita et al. (32,33). Briefly, 50 ng of genomic DNA was subjected to amplification by PCR as described above, except that the reaction volume was only 10 JJLI,' and 0.2 |xCi [a-32P]dCTP (3000 Ci/mmol [Amersham, Arlington Heights, IL]) was included in the reaction mixture. The primer pairs listed in Table 1 were used for SSCP. However, because of the large sizes of exons 6, 9, and 10, SSCP analysis also was performed with additional primer pairs as above. The 10-|xl PCR reaction was mixed with 40 |xl of 10 mM Tris-HCI (pH 8.0) and 1 mM EDTA, and then extracted with CHCI3 to remove the mineral oil. Five microliters of the diluted reaction mixture then was combined with 50 jil of a solution of 0.1% sodium dodecyl sulfate and 1 mM EDTA (pH 8.0). Two microliters of this solution was mixed with 2 |xl of a solution of 96% formamide, 20 mM EDTA (pH 8.0), 0.06% bromophenol blue, and 0.06% xylene cyanol, then was heated at 94°C for 3 min, and immediately loaded on a 3 8 x 3 1 x 0.04 cm 5% polyacrylamide (49:1, acrylamide:A/,A/'-methylene-bis-acrylamide) gel in 90 mM Tris-borate (pH 8.3), 2 mM EDTA, with or without 5% glycerol. Glycerol-containing gels were run at 20 W constant power at room temperature for 4 - 6 h, and gels
DIABETES, VOL. 41, NOVEMBER 1992
J.B. BUSE AND ASSOCIATES
Exon
H
10.4 0.8 ,
1
[>
3 4b 6 8 2 4a 5 7 9
HMHKMltHHZZh
2.4
6.2
2.4
7.4
0.8,
10
XhGLUT4-6 i
7.4
4.5
4.5 3.5
4.5
XhGLUT4-3 3.8
MGLUT4-1 6.5
,0.7
XhGLUT4-2
FIG. 1. Map of the human GLUT4-muscle-fat faciiitative glucose-transporter gene. Positions of exons and the structures of the four overlapping segments of the gene are shown. Natural EcoRI restriction sites are noted by vertical lines. Sizes (in kb) of EcoRI fragments in each cloned segment are Indicated. Please note that the 10.4-kb EcoRI fragment includes several internal EcoRI sites that have not been ordered completely.
without glycerol were run at 2 W constant power at room temperature for 9-12 h. After electrophoresis, gels were exposed to X-ray film with an intensifying screen for 12-18 hat -80°C. RESULTS Isolation and characterization of the human GLUT4 gene. The human GLUT4 gene and flanking regions were isolated as a series of overlapping DNA segments that span a region of 32 kb and include the gene, 15 kb of the 5'-flanking, and 11 kb of 3'-flanking DNA, respectively (Fig. 1). The sequence of a 8.4-kb region was determined (Fig. 2) and indicated that the human GLUT4 gene comprises 11 exons. The exon-intron organization is identical to that previously reported for the corresponding mouse gene (34). Moreover, it is similar to that of the human GLUT1 gene (25), except that exon 4 of GLUT1 corresponds to two exons in GLUT4 that we call exons 4a and 4b. A comparison of the positions of the exon-intron junctions between the GLUT1 and GLUT4 genes indicates that, except for intron 1, they are in identical locations. The start of transcription of the human GLUT4 gene was determined by RNase-protection and primer-extension strategies. The RNase-protection study (Fig. 3) localized the 5' end of the mRNA to two closely spaced regions separated by about 25 nucleotides. Primerextension studies suggest multiple transcription initiation sites (data not shown) centered about nucleotide 2033 (Fig. 2), which corresponds to the major transcriptional start site of the mouse GLUT4 gene (34). The sequence of the promoter indicates that the human GLUT4 gene lacks a TATA motif, but does contain binding sites for several other transcription factors (Fig. 2) including AP2 (35-36), SP1 (37), ETF (38), and the CCAAT binding factor, NF-1 (39-40). Also, note that consensus-binding sites for proteins activate gene expression in myocytes (MEF-2 [41]) and adipocytes (FSE2 [42]). These cisacting elements are all conserved between the human and mouse GLUT4 promoter sequences and are shown in Fig. 2. By contrast, the C/EBP binding site localized in the mouse GLUT4 promoter by DNA footprinting (34) is not present in the promoter region of the human gene. The location of the 3' end of the human GLUT4 gene is
DIABETES, VOL. 41, NOVEMBER 1992
unknown because none of the cDNA clones that were isolated contained a polyA tract. A short region of homology is found between the sequences preceding the polyadenylation sites of the mouse and rat GLUT4 transcripts (3-5) and the region around nucleotide 8076 in the human gene. However, RNA-blotting studies indicate that GLUT4 mRNA is 3.5 kb, which, if the polyA represents 200 nucleotides of this, suggests that the 3' end of the gene is located in a region about 100 bp downstream of the 3' end of the sequence shown in Fig. 2. There are two members of the Alu family (43) of repetitive sequences in the region of the human GLUT4 gene. One is located in the 5'-flanking region, about 1.8 kb upstream of the putative transcriptional initiation site, and the other is located 0.5 kb downstream of the translation termination codon. This latter Alu sequence is predicted to be in the 3'-untranslated region of the mRNA. Regarding microsatellite DNA polymorphisms, we tested the genomic clones for the presence of CA/GT repeats (44,45), and none were noted. Functional characterization of the promoter region of the GLUT4 gene. As described in METHODS, constructs were prepared in which either 2.4 (pApal-CAT) or 342 bp (pBamHI-CAT) of 5'-flanking-promoter region of the human GLUT4 gene were fused to a CAT reporter gene in the sense (+) and antisense (-) orientations. Plasmid DNA from these constructs, the CAT vector (pCAT3M), and a positive control vector containing the SV40 early promoter-enhancer fused to the CAT gene (pSV2CAT) were transfected into differentiated 3T3-L1 adipocytes. As shown in Fig. 4, the 2.4-kb and 342-bp fragments in the sense orientation had similar levels of promoter activity, which were higher than the levels produced by the same fragments in the antisense orientation. However, promoter activity of the human GLUT4 gene constructs was very low, amounting to
