used the differentiated insulinoma cell line INS-1 which expresses high levels of GLUT2 and show that the ex- pression of GLUT2 is regulated by glucose at the ...
THE JOURNALOF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 269,No.43, Issue of October 28. p p . 26912-26919,1994 Printed in U.S.A.
Characterization of the Murine High K , Glucose Transporter GLUT2 Gene and ItsTranscriptional Regulation by Glucose in a Differentiated Insulin-secreting Cell Line* (Received for publication, April 21, 1994, and in revised form, July 26, 1994)
GBrard WaeberS,Nancy Thompson, Jacques-Antoine Haefliger8, and PascalNicod From the Department of Internal Medicine, Internal Medicine B, Centre Hospitalier Universitaire Vaudois, 1011 Lausanne, Switzerland
In pancreatic /3-cells,the high K,,, glucose transporter GLUT2 catalyzes the first step in glucose-induced insulin secretion by glucose uptake. Expression of the transporter has been reported to be modulated by glucose either at the protein or mRNA levels. In this study we used the differentiated insulinoma cellline INS-1 which expresses high levels of GLUT2 and show that the expression of GLUT2 is regulated by glucose at the transcriptional level. By run-on transcription assays we showed that glucose induced GLUT2 gene transcription "fold in INS-1 cells which wasparalleled by a 1.7-2.3fold increase in cytoplasmic GLUT2 mRNAlevels. To determine whether glucose regulatory sequences were present in thepromoter region of GLUT2, wecloned and characterized a 1.4-kilobase region of mouse genomic DNA located 5' of the translation initiation site. By RNase protection assays and primer extension, we determined that multiple transcription initiation sites were present at positions -55, -64, and -115 from the first coding ATG and which were identified in liver, intestine, kidney, and p-cells mRNAs. Plasmids were constructed with the mouse promoter region linked to the reporter gene chloramphenicolacetyltransferase (CAT), and transiently and stably transfected in theINS-1 cells. Glucose induced a concentration-dependent increase in CAT activity which reached a maximum of 3.6-fold at 20 m~ glucose. Similar CAT constructs made of the human GLUT2 promoter region and the CAT gene displayedthe same glucose-dependentincrease in transcriptional activity when transfected into INS-1 cells. Comparison of the mouse and human promoter regions revealed sequence identity restricted to a few stretches of sequences which suggests that the glucose responsive element(s)may be conservedin these common sequences.
The GLUT2 glucose transporter isoform plays a major rolein glucose-induced insulin secretion in the pancreatic p-cell by catalyzing the uptakeof glucose into thecell (1).The regulated expression of this transporter has been studied i n v i t r oby cell culture and ina number of animal models which have an imbalanced glucose homeostasis (2-12). In diabetes, the expres-
* The costsof publication of this article were defrayed in part by the be hereby marked payment of page charges. This article must therefore "aduertisement" inaccordancewith 18 U.S.C.Section1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper hasbeen submitted to the GenBankm/EMBL DataBank with accession number(s1 X78722 $ Supported by Grants 32-31915.91 and 32-29317.91 fromthe Swiss National Science Foundation. To whom correspondence should be addressed: Dept. of Internal Medicine, Internal Medicine B, 19-533, 1011
Lausanne Switzerland. "el.:21-314-47-00; Fax: 21-314-53-90. 8 Recipient of a grant from the Max Cloetta Foundation.
sion ofthis transporteris regulated in a tissue-specific manner. as In particular, in every animal model of diabetes studied such the diabetic Zucker faKa rat, the neonatal low-dose streptozocininduced diabetic rat, the GK rat, the db/db mouse or at the onset of diabetes in the BB/W rat, there is a strong reduction in GLUT2 gene expression which is restricted to the pancreatic p-cells while its expression in liver, intestine, or kidney is unaltered or slightly increased (8-12). From these dataone could conclude that hyperglycemia of diabetic animalsis responsible for the observed decrease in expression of GLUT2 in p-cells. However, exposure of pancreatic islets to high glucose in vitro leads to an increase in GLUT2 mRNA and protein levels. In not clear what causes GLUT2 down-regudiabetes it is therefore lation. Experiments in which development of hyperglycemia was prevented by ascarbose treatment inZucker diabetic rats showed that even in the presence of normoglycemia, GLUT2 expression steadily decreased overtime (8).Taken together,the above observations suggest that in vitro glucose may directly stimulate GLUT2 gene expression whilein the diabetic state the hyperglycemia causes a decreased GLUT2 expression. To better understand the glucose-regulatedexpression of GLUT2, we studied its controlled expression by glucose in a highly differentiatedinsulinoma cell lineandinitiatedthe characterization of the promoterregion of the transportergene. We cloned and characterized the upstream regulatory region of the mouse GLUT2 gene and defined common initiation start sites in murine tissues which express GLUTS, i.e. pancreatic p-cells, liver, intestine, and kidney. The human and mouse promoters were functionally tested using a chloramphenicol acetyltransferase (CAT)' reporter system transiently or stably transfected into a highly differentiated p insulinoma cell line, INS-1 (13). By increasing the concentration of glucose in the culture medium,a concentration-dependent transcriptionalactivation of the human andmouse GLUT2 gene was shown. Similarly, the endogenous GLUT2 gene expression was asblot analsessed in INS-1 cells by nuclear run-on and Northern ysis. As observed with CAT reporter constructs driven by the murine and human promoter, GLUT2 expression is transcriptionally regulated by glucose. Therefore, the 5"regulatory regions of the human andmouse GLUT2 genes containcarbohydrate-responsive element&) responsible for the glucoseinduced gene transcriptionof GLUT2 observed in INS-1 cells. MATERIALS AND METHODS Cell Culture-The transplantable x-ray-induced rat insulinoma INS-1 cell line was kindly provided by Asfari et al. (13). INS-1 cells were grown in RPMI 1640, 10% fetal calf serum, 10 m~ HEPES, 2 mM Lglutamine, 1 mM sodium pyruvate, 100 unitdm1 penicillin, 100 pglml
The abbreviations used are: CAT, chloramphenicol acetyltransferase; MOPS, 4-morpholinepropanesulfonicacid; bp, base paids); kb, kilobase paids).
26912
26913
GLUT2 Gene Expression streptomycin, and 50 pv 2-mercaptoethanol from Life Technologies, Inc. (Life Technologies, Basel, Switzerland). The RPMISelectAmine kit (LifeTechnologies) was used for the basal RPMI media in which glucose was added to the various desired glucose concentrations. Passages 70-90 were used for the present study. When cells reached 60-70% confluency, the RPMI1640 media containing 11 mM glucosewas changed to a 2 mM glucoseRPMI1640 media for 20 h. At time 0, additional glucose was added (10 and 20 mM final concentration)for the transcription run-on and RNA expression studies. INS-1 cells for the CAT assay experiments were incubated overnight in RPMI containing no glucose then, attime 0, glucose was added to the final concentrations of 5, 10, and 20 m. Isolation of Nuclei and Cytoplasmic RNA-INS-1 cellsfromone 15-cm dish were used for each nuclear isolation. Nuclei wereisolated as described by Marie et al. (14) with the exception that 5% (w/w) Nonidet P-40 was used. Before recovering the nuclear pellet, 3 ml of the cytoplasmic fraction were collectedand RNA was isolated according to AUSubel et al. (15). Each nuclear pellet was resuspended in 150 plof 50 mM Tris, pH 8.3,40% glycerol, 5 mM MgCl,, 0.1 mM EDTA, flash frozen and stored at -70 "C. 'Ilanscription Run-on Assay--Transcription reactions were performed with 2 x lo7nuclei a t 30 "C for 30min in thepresence of 100 pCi of [IY-~~PIUTP (800 Ci/mmol, Amersham) in a final volume of 300 pl containing 10 mM Tris, pH 8.0, 5 n m MgCl,, 0.3 M KC1, 1mM ATP, 1mM CTP, 1mM GTP (Pharmacia), 0.04 pg/pl bovine serum albumin, and 0.8 unitdpl RNasin (Promega, Madison, WI). The radiolabeled RNA was isolated according to the method of Vannice et al. (16). In brief, the nuclei were diluted with 5 volumes of water, 100 pg of tRNA, and made 0.4% in SDS. An equal volume of 100 mM NaOAc, pH 5.0,20 mM EDTA was added after which the RNA was extracted with an equal volume of phenol/H,O saturated and precipitated with a final concentration of 2 M NH40Ac and 2.5 volumes of ethanol. Pellets were resuspended in water, re-precipitated, and washed. Counts were normalized for the transcripts and subsequently hybridized to filter-bound GLUT2, p-actin, and pGEM cDNA plasmids for 3 days at 42 "C. Zeta-probe membranes were used as filters (Bio-Rad)and thehydridization solution was 5 x SSPE (1 x SSPE, 0.15 M NaC1,10 m NaPO,, 1 mM EDTA), 50% formamide, 0.1 M NaPO,, pH 6.5,0.5% SDS, 1mM EDTA, 2 x Denhardt's (5 x for prehybridization), and 200pg/ml yeast tRNA. Filters were washed in 2 x SSC, 0.1% SDS at room temperature followed by two washes for 20 min in 0.2 x SSC, 0.1% SDSat 55 "C and thenexposed to Hyperfilm-MP (Amersham) at -70 "C for 7-10 days. Films were quantified using a Molecular Dynamics scanner (Sunnyvale, CAI. Northern Blot Analysis-Cytoplasmic RNAs isolated from the cells for the nuclear run-on experiments were quantified and equal amounts (10 pg) were size-fractionated on 1x MOPS, 1.2%agarose gels containing formaldehyde. Gels were transferred overnight by diffusion (10 x SSC) to GeneScreen membrane (Du Pont). Membranes were W crosslinked and baked 2 h at 80 "C. After prehybridization, the blots were hybridized with random primed (Boehringer,Mannheim, Germany) rat 5 x GLUT2 and actin probes in 5 x SSC,100 mM NaPO,,pH6.5, Denhardt's solution, 50% formamide, 10 m EDTA, 1%SDS, and 100 pg/ml yeast tRNA overnight at 42 "C.The blots werewashed in 2x SSC followed by 0.2 x SSC, 0.1% SDS at 60 "C. Directly after washing, the blots were quantified by electronic autoradiography with an Instant Imager 2024 (Packard Instrument Co.).Blotswere then exposed to Hyperfilm-MP (Amersham). Isolation and Structural Analysis of the Mouse GLUT2 PromoterThe screening of a mouse genomic library with a mouse GLUT2 cDNA probe (17) identified 4 positive bacteriophage cloneswhich together contained most of the GLUT2 gene. The 18-kb insert of clone 4 which encodes the promoter and the first three exons of GLUT2was subcloned S ' (Stratagene, La Jolla, CA). Both strands of the into a pBluescript K promoter region, exons and exon-intron boundaries were sequenced using the Sequenase sequencing kit (U. S. Biochemical Carp.). RNase Protectionand Primer Extension-Mouse tissues were homogenized in 9 ml of 4 M guanidine isothiocynate buffer with a Kinametic polytron blender (Kriens, Switzerland) and layered onto a 4-ml 5.7 M CsCl cushion. RNAs were pelleted at 33,000 rpm for 17 h in a 50 Ti rotor. Mouseislet RNA was obtained from islets isolated by the method of Gotohet al. (18). Antisense RNA probes were synthesized using subclones of the mouse promoter region (-1311 to +70 and -335 to +70 bp) as templates and T7 RNA polymerase (Promega). The RNase protection assay was carried out according to the protocol of the RPA I1 kit (Ambion,Austin, TX). Primer extension products were synthesized using an antisense primer 5'-CTTGTCTTCTGACA'I"l'GTGTG-3'which was end-labeled with [y-32PlATPand extended by reverse transcriptase (19). All probes and oligonucleotides werelabeled with radionucleotides
purchased from Amersham. The products of the RNase protection or primer extension assays were separated on a 6% polyacrylamiddurea sequencing gel. A sequencing reaction primed with the same oligonucleotide as used for the primer extension was run as a sizing marker. lkanscriptional Reporter Constructs and TransfectionStudiesPolymerase chain reaction-generated regions of the mouse promoter from -1311 to +49 or the human promoter from -1296 to +312 bp were subcloned into the SalYXbaI sites of the promoterless pCAT-basicvector (Promega). The transcriptional reporter constructs were sequenced and then transiently transfected into INS-1 cells by liposome-mediated DNA transfection (DOTAP, Boehringer). Ten pg of either construct were usually co-transfected into 1-2 x lo6 cells, with a pSV, galactosidase reporter gene (pSV,GAL) as internal control. Twenty fours hours after transfection, glucose concentrations of the medium were changed (0,5, 10, and 20 mM) and the cells incubated for an additional 24 h. The cells were harvested and then ruptured by three freeze-thaw cycles, as described by Pothier et al. (20). After removal of cellular debris, the extracts were heated for 10 min at 65 "C to destroy endogenous deacetylating activity. Protein concentrations were determined using the BCA protein assay (Pierce).CAT assays were carried out using 100-150 pg of cell extracts and the acetylated chloramphenicol was separated on a thin layer chromatography plate. The results were normalized by the value of galactosidase activity measured from the co-transfection of pSV,GAL and/or adjusted to the protein concentration of the cell extracts. The CAT-enzyme-linked immunosorbent assay system (Boehringer) was also used in some studies to quantitate the CAT protein of the transfection studies. The correct initiation start site used in the fusion mouse promoter CAT constructs was determined by primer extension of transfected INS-1 cell RNA using a complementary oligonucleotide to the CAT gene (5'-TTACGATGCCATTGGG-3'). To generate stably transfected INS-1 cells, the murine and human GLUT2 promoter CAT constructs were co-transfectedwith a pSV, neomycine plasmid (10 pg ofCAT constructs and 2 pg of the neomycine plasmid) and the neomycine-resistant clones were then studied for integration of the CAT transgenes. Statistics-All RNAor transfection studies were carried out in four or eight separate experiments, respectively. Data are expressed as mean +. S.E. and compared by the (nonparametric) Friedman test. RESULTS
Identification of the Mouse Upstream Regulatory Region of the GLUT2 Gene-By screening a mouse genomic library with a GLUT2 cDNA probe, we isolated 4 overlapping bacteriophage clones which together contained most of the GLUT2 gene. The overall gene structure and exon-intron boundaries are similar to the recently published human GLUT2 gene (21). The 18-kb insert of clone 4 contains the first threeexons of the gene and approximately 1.4 kb of the putativepromoter region of GLUT2 (Fig. LA). The nucleic acid sequence of the mouse upstream regulatory region is shown in Fig. 1B. By computer analysis, several potentialconsensus sequencesfor various transcription factors (22-28) are located within the 1.3-kb promoter region. The cis elements include a cyclicAMP responsive element (CRE = ACGTCA 6/61, an AP-1, AP-4,and a CCAAT-box-bindingtranscriptionalfactor(CTF-NF1 10/14 TGatgGTAAtCCAA). The role of these sequences in thecontrol of GLUT2 gene expression needs to be established. A potential TATAA-like motif and CAAT box are located at -34 and -99 bp, respectively, of the major initiation start site andmay be involved in mediating the basal transcriptional activity of the GLUT2 gene. Nucleic acid comparison of the human andmouse 5"regulatory region of GLUT2 gene shows approximately 50% sequence identity between the human (-398 to + 222) and the murine (-579 to -25) promoters which suggests that these conserved regions are of functional importance (Fig. 1C). Danscriptional Start Sites Are Common in GLUT2 Expressing Tissues-The initiation start sites were localized by RNase protection and primerextension assays. For the RNaseprotection assay, an antisense RNA probe corresponding to nucleotides -329 t o +70 of the mouse promoter sequence was hybridized in solution with RNAs extracted from different tissues and
26914
GLUT2 Gene Expression
A
93 bp
MSEDK
260 bp
LJ H = Hindlll B = BamHl
Sp = S p h l
FIG.1. Nucleotide sequence of the mouse GLUT2 promoter. A, diagram and partial mapping of the 18-kbinsert of bacteriophage 4 obtained by screening a mousegenomic library with a GLUT2 cDNA probe. The 18-kb insert contains the first three exons of the gene and approximately 1.4kb of sequence upstream the first coding ATG. B , sequence of the mouse promoter of GLUTS. A major initiation start site was localized at -55 bp from the first coding ATG which determined the +1 bpcoding system (arrow and T i n bold character). Two minor start sites were also localized at -64 and -115 from the ATG encoding the first methionine. Potential TATAA and CAATboxes are present at -34 and -99bp, respectively, of the major transcriptional start. Several potential consensus sites for various transacting factors have been determined by computer analysis and shown with arrows (see results for description of the putative responsive elements). C, nucleic acid comparison of the human and mouse 5'-regulatory region of GLUT2 gene, Approximately 50% sequence identity was found between the human (-398 to +222) and the mouse (-579 to -25) promoters which maysuggest that these conserved regions are of functional importance (- indicates identical sequences and A indicates gap inthe sequence comparison).
6
MOUSE GLUT2 PROMOER SEOUENCE
-1311 CGACTCaATCCCirrOGCCTGTCAGCTTGCCCACACGTATG~~CAGAGTCTCTGAGGTCCTGTC~AAATAGAGACA~
-
-1229 ATMTAGMjTAATAGAGGMGACTCTCAATACTGA~CCTGGCCTTCACACAGAGGAGTACACATATMCACATATACAG CRE -1149AGAGACAGACAOAGAOAOAOGAGAGACAGA~T~TGMT~ATGMTGMTGMTGA~GAGATGG~CG
-
& " 4 -1071 T C A G A C T M T T A A A C C C A G m C C 7 7 C C G A C C G A C C T
-887 ACTOGCTCCATCCAGATOGACTClTOTOGATGTGTGGATG~AGTGACCACAGCTGGCAGTCCCAGG~GCTACCCAAAC~~TATC
N W
-907 T C C C C C A G T A m r r G T A T C A G A T G C C A C A G G G G A T T C C T A G G M ~ C M ~ T C M G A C A C G T M ~ A ~ ~ C A C T G
-827 OAAAOCCMGGCTATAGCMGCACACTACG~TCCCMCCA~CTCTCTATGTGTAGCCMGGCTATCCTACMCTCTCAT AP-I "t -746 GTOATCCAGTGAGTCTCCMTGCACAGCCTCCCCAGTTAGTTAGAGCA~ATAGATGT~GCCACCGCATCTGGCTCCGCACTCT
-665 C A T C T T G A G A T A C A C T G A m T M T M C A G T A G ~ G C A C A ~ G C C A C T A T C T A C T G A G M C ~ C T G T G C C ~ C A -584 m A T C A C G T T A m C C C T T M T C C C C T C M C M M C A A A A T G A -502 ACAGTAGAAACCTMGACACAGAAAAGTCACAGGGmGACTCGC~MG~CATCCTCACT~TGCTGTGA~CCMCC -420 CCAGTGCAGATCCCTCCACACCTCACACAGATAACCGATGCTGCCACACTCTGTGGCCACAGAGCCCACA~CTM~
-339 CTGCAGMCAAACTGCTCTGTGCCCTCCCMCTTTGTC~CCCGACACCACTGCAT~CTCGGC~CCACAAAAGACCCCAC -257 AGGCTGCCCCTCTCAGCTGTCCCTGTCCCATTTCTGCCACAC~ATCAGGCTGAAAATGGGTCTGTCTCTGG~GTMCTT CTF-NFl -175 ATACTTATOAGACCTGCTACTGTGCTCAAGCCAAGCCACMGTCATT~GGTAAAGGGTGTA~GA~~A~ACCA~CTC -94 AGC~CTGTTAAAAAGGTCAOAACTACCTACCTCTC~~TCCTCCTCCTCCTACMTG~CCAGGTAGAGTGA~ACTCTG
- 12 G C T C ~ C A G C T ~ ~ C A ~ ~ ~ C A G T A C A G G A C C T ~ G A ~ ~ ~ ~ ~ G ~ C M S E D K +70 g t a c a g c O a c a t g e g g t c c t t t ~ u t t g g ~ g a g g ~ g t t ~ g a a t c t a t l l ~ ~ ~ ~ g a ~ a t ~ t g e c a c t ~ a l g c ~ t t ~ ~ ~ ! ~ g c t a ~ t c t t t g c n g
C mouse(-S79) human(-398) muse human
mouse human mouse
human mouse human mouse human
mouse human mouse humw mouse hunrm mouse
humon
a positive control was made from a sense RNA synthesized in vitro.Fig. 2A shows a typical RNase protection assay, where the 399-base probe protects 3 major transcripts afterRNase A and T1 digestion, labeled1,2, and 3 in thefigure. Thesetranscripts
define start sites localized at -115, -64, and -55 bases from the first coding ATG. The protected transcripts are present inthe mouse islet, liver, intestine, andkidney RNAs but not inmouse heart, whole pancreas (the endocrine pancreas representing a
GLUT2 Gene Expression
26915 pCATbasic mA1311CAT
"
GLUCOSE
0
20
0
20
mM
FIG.3. Glucose responsivenessof the mouse GLUT2 promoter. CAT activity was measured in the transiently transfected p insulinoma cell line (INS-1) exposed for 24 h to a medium without or with 20 mM glucose. The promoterless reporter CAT vector (pCAT basic) shows no glucose responsiveness, whereas 1311 bp of the mouse promoter CAT fusion construct showsa 2-3-fold increase inCAT activity with 20 mM.
1400-base antisense RNA probe (data not shown). Additional lower frequency start sites were also identified (transcripts 1 and 2 in Fig. 4A)in all tissues expressing GLUT2 which sug-2 gests thepresence of multiple initiation start sites for the RNA polymerase 11. The intensityof the protected transcripts of Fig. 2A corresponds to the abundanceof GLUT2 mRNA expressed in these tissues. The highest level being found in pancreatic islets and liver, whereas, in kidney and intestine, it is lower. -3 By primer extension (using a complementary oligonucleotide to sequence +49 to +70 of the mouse promoter sequence) one major start site was identified (data not shown), which corresponds to the-55-bp transcript (transcript number 3, Fig. 2A), and wassubsequently used to define the start site (bp +1 of the sequence shown in Fig. 1B). Mouse and Human Promoters Are Functionally Active and Glucose-responsive When Dansiently or Stably Dansfected into a Differentiated p Insulinoma Cell Line-Regions of the mouse (-1311 to +49) and human (-1296 to +312) promoters were cloned into the eukaryotic expression vector pCAT Basic (Promega) and transientlytransfected into a differentiated p insulinoma cell line, INS-1, which expresses a high level of GLUT2. Cells were then exposed for 24 h to a medium containing fetal calf serum and 0 or 20 mM glucose. As shown in Fig. 3, basal transcriptional activity, as measured by CAT assay, is higher FIG.2. Determination of the transcriptional initiation start for the mouse promoter in comparison to thepromoterless vecsites of the mouse GLUT2 gene. A, ribonuclease protection assayof mouse tissues RNAs hybridized in solution with a 399-base antisense tor pCAT Basic. Furthermore, themouse promoter CAT fusion RNAprobe that contained the -329 to +70 bp region of the promoter. No construct shows a glucose inducibility of 2-3-fold with 20 mM protected transcripts were found in thenegative controls which include glucose in comparison to the basal transcriptional rate seen tRNA, heart, and whole pancreas RNAs. A positive control was made during incubation with 0 mM glucose. The pCAT basic vector from a sense RNA synthesizied in vitro. Inalltissuesexpressing GLUT2, the probe protected 3 major transcripts defined as 1,2,and 3. did not show any change in activity when incubated with 0 or Number 3 is the strongest signal and corresponds to the start site 20 mM glucose. Since all transfections studies were done in 5% defined by primer extension, and is considered as the major initiation fetal calf serum, the0 mM glucose concentration corresponds to site for RNA polymerase 11. The intensity of the protected transcript approximately 1-1.5 mM glucose as measured in the culture corresponds to the expected amount of GLUT2 mRNA expression in these tissues: higher intensity is seen in islet and RNAs liver and lower medium at the end of each experiment. Similar transfection experiments were undertaken to deterintensity is found in the kidney and the intestineRNAs. B, the same RNA used to determine the start sites were hybridized with an actin mine if the human and themouse regulatory regions have an antisense probe to show equivalent RNA loading in the various columns identical glucose inducibility, and if so, whether it is concentrain the ribonuclease protection assay. tion-dependent. A typical CAT assay isshown in Fig. 4A where CAT activities for both human and mouse promoter constructs small percentage of the organ), or tRNA which were used as were found to increase in a concentration-dependent manner negative controls. An actin antisense RNA probe was hybrid- during incubation with glucose concentrations of 0, 5, 10, and ized to the sameRNAs to quantitate the amountof RNA (Fig. 20 mM. The transfections were repeated 6 times and CAT ac2B ). The protected transcript mapped at -55 bases (number3, tivity was normalized by protein content of the transfected cell Fig. 2 A ) is the major start site asconfirmed by primer exten- extracts andor by the activity of P-galactosidase expressed sion and several otherRNase protection assays using a larger from a co-transfected pSV, galactosidase reporter gene. Quan-
GLUT2 Gene Expression
26916 hAl296CAT pCAT-basic
A
mAl311CAT
I
20" 0
mAl311CAT
r)
hA1296-T
?
5
10
20'
I
0 rnY glucose
I
20mM glucose
I CAT I
CAT
1 10
20 Relative CAT activity
30
FIG.4. Dose-dependent transcriptional activation of the mouse and human promoters exposedto increasing concentrationsof glucose.A, a typical CAT assay of transiently transfected INS-1cells with the human(hAl296CAT) and the mouse(mAl311CAT) promoter CAT fusion constructs compared to the promoterless pCAT-basic vector (pCAT-basic).The human and mouse promoters wereglucose-responsive in a dose-dependent manner from 0 to 20 mM of glucose for a n incubation time of 24 h. The arrows designated the mono-, bi-, and triacetylated chloramphenicol forms were separated by thin layer chromatography.B, quantitative assessmentof transcriptional glucose responsiveness of the human andmouse GLUT2 promoters. Six different transfection experiments were carried out and CAT activity or protein measured by counting "C-acetylated chloramphenicol or by the CAT-enzyme-linked immunosorbent assay system. The CAT activity was normalizedby protein content of cell extracts andorby the co-transfection of a pSV, galactosidase reportergene. Glucose induced a dose-dependent transactivationof the mouse GLUT2 promoter: there was a n increase in CAT activity of 1.4-, 2.5-, and 5.1-fold a t glucose concentrations of 5, 10, and 20 mM, respectively, compared to baseline(glucose = 0 mM). Similarly, glucose induced a 1.3-, 1.8-, and 2.8-fold transactivation of the humanGLUT2 promoter when exposed to 5, 10, and 20 mM (* = p < 0.05,** = p < 0.01,and *** = p < 0.001). Data are shown as the mean 2 S.E.
titative analyses areshown in Fig. 4B. Glucose concentrations sponsive in a concentration-dependent manner (Fig. 5,A-C). GlucoseEffect on GLUT2 Gene Dunscription Rate and of 5, 10, and 20 mM induced a significant 1.3-, 1.8-,and 2.8-fold, respectively, increase in CAT activity for the humanpromoter mRNA Accumulation-GLUT2 gene expression was studied in 20 mM glucose over a 4-, 8-, and and a 1.4-, 2.5-, and 5.1-fold increase in CAT activity for the INS-1 cells exposed to 2,10, and mouse promoter when compared to the baseline (defined as 0 24-h period. Glucose-induced GLUT2 gene transcription as asmM glucose). sessed by nuclear run-on analysis (Fig. 6, A and B).A maximal The correct initiation start siteused by the mouse promoter 4- and 3.4-fold increase of GLUT2 was observed with 20 mM and CAT construct wasanalyzed by primer extension of RNA from 10 mM glucose, respectively, as compared to the2 mM condition. Cytoplasmic RNAs of the nuclear run-on experiments were the INS-1 cell transfected with mAl311CAT. The major transcriptional start sitecorresponds to the +1 bp defined in Fig. 1B collected and studied by Northern blot analysis. As expected from the dependence of GLUT2 transcription rate on the glu(data not shown). To avoid variation of transfection efficiency in between ex- cose concentration, GLUT2 mRNA accumulation increasedin a periments, we generated stably transfected INS-1 cells that concentration-dependent manner. A concentration of 20 mM integrated the murinepromoter or human promoter CAT con- glucose led to a 2.3-fold higher level of mRNA accumulation as structs together with a neomycine-resistant plasmid. Isolated compared to 2 mM glucose (Fig. 7, A and B). The experiments clones were similarly studied for CAT activity when exposed to were repeated 3 times with similar results and all data were 0,5,10, and 20 mM glucose for 24 and 48h. The glucose induc- normalized to actin. ibility of the CAT transgenes were quite similar to the transient DISCUSSION transfection observed after 24 h. After 24 or 48 h of treatment, the maximum glucose inducibility was 3.6-fold higher than the In thisstudy, we have structurallycharacterized 1.4 kb of the 0 mM conditions. Therefore, these promoters were glucose-re- mouse GLUT2 promoter and shown that this sequence shares
j:i p
s dl
GLUT2 Gene Expression
n
26917
A Glucose
0-actin -
1.9 x 1.7.T
GLUT2 pGEM
B 0
10 20 mMglucose
5
s!
10 mM
2mM
2mM
Time (hrs) o
'4
24' ' 4
8 i i
--
.'
8
20 mM 24' ' 4
8
24'
,,
1 ""
5-
E
24 hours
B
L
..e. 0
5
glucose 10
20 mMglucose
u 48 hours C
" U
10mM 2mM
20mM
FIG.6. Effect of glucose on GLUT2 gene transcription ratein INS-1 cells. A, INS-1 cells were incubated in2, 10, and 20 mM glucose for 4, 8, and 24 h. Nuclei were isolated and nascent transcripts were labeled and hybridized to the rat GLUT2 cDNA, pGEM, and p-actin cDNA. Glucose induced GLUT2 gene transcription in a concentrationdependent manner. B, quantitative assessment by laser densitometric scanning of GLUTWactin transcription measured by nuclear run-on assay. A maximal 3.4- and 4.0-fold increase in GLUT2 gene transcription was observed with 10 and 20 mM glucose concentrations, respectively. Data were normalized to the p-actin densitometricvalue. regions of identity with the previously described human promoter (21). The presence of stretches of complete identity between the species interspersed with more divergent sequences suggests that suchregions may beof functional importance in the control of gene expression. By RNase protection assay and primer extension we have defined common start sites in all mouse tissues expressing GLUTS. This finding is inconsistent with the use of alternative promoters or initiation start sitesfor cell-specific expression ofGLUTB, as has been described for another gene involved in glucose-sensing, the glucokinase gene, where two different cell-specific promoters are found in the liver and p-cells (29). In experimental diabetes, GLUT2 mRNA expression increases inliver, whereas it decreases in P-pancreatic cells (1,8-12) and it therefore seems unlikelythat the use of alternative promoters or different start sites is responsible for this difference in cell-specific regulation of GLUT2. The structurally characterized mouse GLUT2 promoter region and the human GLUT2 B'-upstream region were then functionally tested usinga CAT reporter system by transient or stable transfections into INS-1 cells. CAT activity was measured in these transfected cells after exposure to different glucose concentrations for 24 and 48h. Interestingly, a concentration-dependent induction by glucose of GLUT2 gene
I 2.5 x
T
'1
L u
I
I
u 48 hours
24 hours FIG.5. Glucose effect on stably transfected INS-1cells with human and mouse promoter CAT constructs. A, quantitative assessment of the mouse promoter CAT construct stably transfected into INS-1 cells. CAT activities were normalized to protein content.maxiA
mal 1.9-fold increase inCAT activity was observed with 20 mM glucose concentration for 24 h.B, a maximal 3.6-fold increase in CAT activity was measured with20 mM glucose concentration for 48 h using INS-1 cells stably transfected with the murine promotor CAT construct. C, quantitative assessment of the human promoterCAT construct stably transfected into INS-1cells and exposed to 0,5,10, and 20 mM glucose for 24 or 48 h.
26918
A Glucose
-
GLUT2 Gene Expression
analogy to the distal L4 element of the pyruvate kinasegene (GTGCCC). Several AT-rich stretches are present in the murine Time (hrs) o 4 a 24 '4 a 24'14 a 24' and humanpromoters but these structurally defined sequences need to be functionally tested to see if they are able toconfer glucose responsiveness on heterologous promoters. We havedemonstrated by nuclear run-on analysisthat INS-1 cells exposed to medium with 10 and20 mM glucose have a maximum 3.4- and 4-fold increase, respectively, in endoge4houn nous GLUT2 genetranscription. Furthermore, cytoplasmic Bhoun RNA of the run-on analysis haveshown a parallel induction of 24houn GLUT2 gene expressionto a maximum 1.7- and 2.3-fold for the 10 and 20 mM glucose concentrations, respectively. Therefore the endogenous GLUT2 gene is transcriptionallyregulated by glucose as is the case for the murine and humanGLUT2 promoters. The modest differences between transcription rate and mRNA expression could possibly be explainedby either a delay in the transfer of the transcripts from the nuclei to the cytoplasm andor analteration inGLUT2 mRNA stability induced by glucose. We have measuredGLUT2 mRNA half-life in INS-1 cells in 11 mM glucose and determined that it was approximately 8 h. Further work will be necessary to demonstrate a possible role of glucose in altering GLUT2 mRNA stability. Several reports have previously shown that GLUT2 gene expression is modulated by glucose in vivo and in vitro. In primary culturesof rat hepatocytes, Asano et al. (2) described a 3.2-fold increase inGLUT2 mRNA expression whenthese cells were incubated in 27.8 mM glucose for 20 h. Similar observations were recently reportedby Postic et al. (3)who have studFIG.7. Glucose effect on cytoplasmic GLUT2 mRNA expres- ied GLUT2 mRNA expression in liver in vitro and in vivo: a sion. A, cytoplasmic RNAs from the nuclear run-on were collected and maximal 4-fold increasein GLUT2 mRNA expressionwas analyzed by Northern blot. 10 pg of total RNA were hybridized with found when hepatocytes were incubated in 20 mM glucose for 24 GLUT2 and actincDNA probes. Increasing amountsof glucose induced h. In the hamster 0-cell line HIT, Inagaki et al. (4) found a 40% GLUT2 mRNAexpression in a concentrationand time-dependent manner. €?, quantitativeassessment by electronicautoradiography of increase in GLUT2 mRNA expression when these cells were GLUT2Iactin cytoplasmic RNAs. Maximum 1.7- and 2.3-fold increases incubated in 22.2 versus 11.1 mM glucose for 24 h. A 3-fold in GLUT2 mRNA expression were observed with 10 and 20 mM glucose induction in GLUT2 expression was observed in primary rat concentrations, respectively. C, Northern blot analysis of INS-1 cells exposed for 24 or 48 h to 2,10, 20 and mM glucose. The maximalglucose islets in 11.1uersus 5.5 mM glucose concentration by Yasuda et al. (5). Chen et al. (6) described a 46% increase in GLUT2 inducibility is reached at 24h and no time-dependentincreasein GLUT2 mRNA expression observed. mRNA expression in p-cells of rats maintained hyperglycemic for 5 days. More recently, a maximal 10-fold increase inGLUT2 transcription wasobserved. The murine GLUTBCAT construct expression was described by Ferrer et al. (7) when rat islets shows a maximum 2.5-fold increase and 5.1-fold increase for were incubated in 11mM glucose in comparison to 2 mM. In this the 10 and 20 mM glucose concentration, respectively, when study, a time course experiment of the effect of glucose on compared to the0 mM glucose condition. The human promoter GLUT2 has shown a 2.5-fold induction of GLUT2 mRNA only CAT transgene was similarly induced in a concentration-de- after 8 h of culture in 16.7 mM glucose. Taken together, these pendent manner.Although modest, the glucose inducibility was observations clearly demonstrate that high glucose concentrareproducible in transiently or stably transfected cells with both tions positively modulate GLUT2 gene expression in p-cells transgenes. These data suggest that the isolated murine and and hepatocytes. In diabetic animal models, a drastic P-cell-specific reduction human GLUT2 promoters contain unidentified carbohydrateresponsive elements. Several genes, such as insulin, S14, and of GLUT2 expression has been reportedfor the NIDDM Zucker pyruvate kinase, have been shown to be transcriptionally regu- fdfa,theneonatal low-dose STZ-induced diabetic rat,the lated by glucose (14, 30-35). In the rat 1 and human insulin Wistar Kyoto, GK rats, db mice, and the autoimmunediabetic genes, a n AT-rich sequence has been proposed as the specific BB rats (8-12). By cross-transplanting islets from db/db mice glucose-responsive element (34). However, in S14 and thepyru- into control mice or vice versa, the decreased GLUT2 expresvate kinase genes, E boxes have been clearly shown to be re- sion was shown to be reversible and induced by the diabetic sponsible for the carbohydrate responsiveness with a common environment of the animals(36). The pathogenic significance of core element: CACGTG. The consensus is a putative binding the decrease in p-cell GLUT2 is controversial: it has been site for MLTFNSF or other basidhelix-loop-helideucinezip- claimed that thisloss may be the primarycause of the specific per proteins. Diaz Guerra et al. (35)have recently shown that glucose-induced secretory abnormality encountered in thediathe L4 elements of the pyruvate kinasegene bind various nu- betic state (1, 8, 11)although others consider this possibility clear proteins thatinclude the ubiquitous MLTFNSF protein. unlikely (37). Whatever thepathogenic significance of the loss In the case of the murine 5"regulatory region of GLUTS, we of GLUT2 expression, it is one of the earliestbiochemical markfound no perfect consensus for the core CACGTG defined as E ers of the diabetic state. Since glucose regulates positively boxes. At -879 to -873 or -472 to -466 of the murine GLUT2 GLUT2 gene transcription in the normal state and since the sequence two CACaGGG sequences share partial analogy to carbohydrate-responsiveness of GLUT2 is selectively lost in the the L4 proximal element defined in the pyruvate kinasegene p-cell of diabetic rodents, further work to identify the cis elefactors involved in the control of (CACGGG).At -592 to -587, a GTGCCA sequence has a weak mentsandtrans-acting
0
2mM
2 mM
1
I
10 mM
20 mM
GLUT2 Gene Expression GLUT2 gene transcription is needed and this may lead to a better understanding of the pathogenic events involved in the early diabetic state. Acknowledgments-We thank G. I. Bell for sending the human GLUT2 clone and sequenceprior to its publication, andA. Asfari andC. Wollheim for the gift of INS-1 cells. We are grateful to B. Thorens for critical comments of the manuscript and M. A. Blanc for typing the manuscript. REFERENCES 1. Unger, R. H. (1991) Science 251, 1200-1205 2. Asano, T., Katagiri, H., Tsukuda, K , Lin, J. L., Ishihara, H., Yazaki, Y., and Oka, Y. (1992) Diabetes 41, 22-25 3. Postic, C., Burcelin, R., Rencurel, F., Pegorier, J.-P., Loizeau, M., Girard, J., and Leturque, A. (1993) Biochem. J . 293, 119-124 4. Inagaki, N., Yasuda, K , Inoue, G., Okamoto, Y., Yano, H., Someya, Y., Ohmoto, Y., Deguchi, K , Imagawa, K. I., Imura, H., and Seino, Y. (1992) Diabetes 41, 592-597 5. Yasuda, K., Yamada, Y., Inagaki, N., Yano, H., Okamoto, Y., Tsuji, K , Fukumoto, H., Imura, H., Seino, S., and Seino, Y.(1992) Diabetes 41, 76-81 6. Chen, L., Alam, T., Johnson, J. H., Hughes, S., Newgard, C . B., and Unger, R. H. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 4088-4092
7. Ferrer, J.,Gomis,R., Fernandez-Alvarez, J., Casamitjana,R., and Vilardell, E. (1993) Diabetes 42, 1273-1280 8. Orci, L., Ravazzola, M., Baetens, D., Inman, L., Amherdt, M., Peterson, R. G., Newgard, C. B., Johnson, J. H., and Unper. R.H. (1990) Proc. Natl. Acad. Sci. V . S. A. 87,9953-9957 9. Ohneda, M., Johnson, J. H., Inman, L. R., Chen, L., Suzuki, K I., Goto, Y., Alam, T., Ravazzola, M., Orci, L., and Unger, R. H. (1993) Diabetes 42, 1065-1072 10. Thorens, B., Weir, G. C., Leahy,J. L., Lodish, H. F.,and Bonner-Weir, S . (1990) Proc. Natl. Acad. Sei. U. S. A 87, 6492-6496 11. Johnson, J. H., Ogawa, A,, Chen, L., Orci, L., Newgard, C. B., Alam, T., and Unger, R. H. (1990) Science 250, 546-548 12. Orci, L., Unger, R. H., Ravazzola, M., Ogawa, A., Komiyia, I., Baetens, D., Lodish, H. F., and Thorens, B. (1990) J. Clin. Invest. 86, 1615-1622 13. Asfari, M., Janjic, D., Meda, P., Li, G.,Halban, P. A., and Wollheim, C. B.(1992) Endocrinology 130, 167-178 14. Marie, S., Diaz-Guerra, M. J., Miquerol, L., Kahn, A,, and Iynedjian, P. B.
26919
(1993) J . Biol. Chem. 268,23881-23890 15. Ausubel, F. M., Kingston, R. E., Moore, D. D.,Seidman, J. G . , Smith, J. A., and Struhl, K. (eds) (1987) Current Protocols in Molecular Biology, pp. 4.10.14.10.4, John Wiley and Sons, New York 16. Vannice, J. L., Taylor, J. M., and Ringold, G. M. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,4241-4245 17. Suzue, K , Lodish, H. F., and Thorens, B. (1989) Nucleic Acids Res. 17,10099 18. Gotoh, M., Maki, T., Satomi, S., Porter, J., Bonner-Weir, S., O'Hara, C. J., and Monaco, A.P. (1987) lFansplantation (Baltimore) 43,725-730 19. McKnight, S . L., and Kingsbury, R. (1982) Science 217,316324 20. Pothier, F., Quellet, M., Julien, J. P., and Gukrin, S. L. (1992) DNA Cell Bid. 11,83-90 21. Takeda, J., Kayano, T., Fukomoto, H., and Bell, G. I. (1993) Diabetes 42, 773-777 22. Locker, J., and Buzard, G. (1990) DNA Sequence-J. DNA Sequencing and Mapping 1, 3-11 23. Chiu, R., Imagawa, M., Imbra, R. J., Bockoven, J. R., and Karin, M. (1987) Nature 329,648-651 24. Bodner, M.,Castrillo, J. L., Theill, L. E., Deerinck, T., Ellisman, M., and Karin, M. (1988) Cell 56, 505-518 25. Imafawa, M., Chiu, R., and Karin, M. (1987) Cell 51,251-260 26. Lefevre, C., Imagawa, M., Dana, S., Grindlay, J., Bodner, M., and Karin, M. (1987) EMBO J 6,971-981 27. Morgan, W. D., Williams, G . T., Morimoto,R. I., Greene, J., Kingston, R. E., and Tjian, R. (1987) Mol. Cell. Bid. 7, 1129-1138 28. Habener, J. F. (1990) Mol. Endocrinol. 4, 1087-1094 29. Magnuson, M. A., and Shelton, K. D. (1989) J . Bid. Chem. 264, 15936-15942 30. Bergot, M. O., Diaz-Guerra, M. J. M., Puzenat, N., Raymondjean, M., and Kahn, A. (1992) Nucleic Acids Res. 20, 1871-1878 31. Jacoby, D. B., Zilz, N. D., and lbwle, H. C. (1989) J . Biol. Chem. 264,1762317626 32. Thompson, K. S., and lbwle, H. C. (1991) J. B i d . Chem. 266, 867M682 33. German, M. S., Moss, L. G . , and Rutter, W. J. (1990) J. Biol. Chem. 265, 22063-22066 34. Melloul,D., Ben-Neriash, Y., and Cerasi, E. (1993) Proc. Natl. Acad. Sci. U. S. A. 38653869 35. Diaz Guerra, M.-J. M., Bergot, M.-O., Martinez, A,, Cuif, M.-H., Kahn,A., and Raymondjean, M. (1993) Mol. Cell. Biol. 13, 7725-7733 36. Thorens, B., Wu, Y.-J., Leahy, J. L., and Weir, G . C. (1992) J. Clin. Invest. 90, 77-85 37. Heimberg, H., De VOS, A,,Vandercammen, A., Van Schaftingen, E., Pileleers, D., and Schuit, F. (1993) EMBO J . 12,2873-2879 ~
~~
90.
