Distinct TATA Motifs Regulate Differential Expression of Human ...

THEJOURNAL OF B[owIchL CHEMISTRY 0 1993 by T h e American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 32,Issue of November 15, pp. 24460-24466, 1993 Printed in U.S.A.

Distinct TATA Motifs Regulate Differential Expression of Human Metallothionein I Genes MT-IF and MT-IG* (Received for publication, April 22, 1993, and in revised form, July 28, 1993)

Nicholas W. Shworak$§, Tracy O’Connornll, Norman C. W. Wongn**$$,and Lashitew Gedamu$§I From the Deoartments of Woloizical Sciences. **Medicine, and Wedical Biochemistry, the University of Calgary, Calgary, Alberta T2N 1N4, Canada

In this report,we have measured the cadmium (Cd2+)induced expression of all known metallothionein I (MT-I) mRNAsin a human hepatoma cell line, Hep GZ. Among the human MT-I gene family promoters, marked sequence conservation exists; despite this, the mRNA accumulation level for each species was found to be quite unique. This differential Cd2+induction of MT-I family members provides an ideal opportunityto assess whether the characteristicresponse results from subtle isoform-specific variations in promoter structure. Accordingly, we have examined the mechanism for differential expression of two isoforms, MT-IG and MT-IF, by transient transfection into Hep G2 cells. In the presence of Cd2+,MT-IC promoter activity and endogenous mRNA level were, respectively, 4.7- and 3-fold greater than those of MT-IF. This close correlation between promoter activity and mRNA accumulation strongly suggests that differential expression occurs at the level of transcription. The difference in Cd2+-stimulatedactivity was found to be conferred by 240- and 243-base pair promoter fragments spanning nucleotides -174 to +66 and -172 to +71 of the MT-IG and MT-Jp genes, respectively. One of the most striking nonhomologies between the promoters is a single A ( T A T U ) to C (TATCAA) transversion in the TATA motifs of MT-IG and MT-IF genes, respectively. To determine whethersuch a subtle change in the TATA motif could account for themarked differences in promoter function, we constructed MT-IGTATCA and MT-IF-TAT@ promoters and measured their activities in transient transfection and cell-free transcription assays. Results of both assays showed a profound difference between the two motifs that paralleled the difference in Cd2+-stimulated MT-IG and MT-I, mRNA levels. In summary, we have shown that differential regulationof two MT-I promoters is primarily due to a single base alteration in their TATA motifs. Metallothioneins (MTs)’ are a family of phylogenetically ubiquitous proteins that have the following common features:

* This work was supported in part by an operating grant from the Medical Research Councilof Canada (to L. G.).The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely toindicate this fact. 5 Recipient of Medical Research Counciland Alberta HeritageFoundationfor Medical Researchfellowships. To whomcorrespondence should be addressed. Present address:Dept. of Biology, Bldg. E25-229, MIT, Cambridge, MA 02139. Tel.: 617-253-8803; Fax: 617-258-6553. 1) Supported by an Alberta Heritage Foundationfor Medical Research summer studentship. ff Recipient of scholarship awardsfrom the Medical Research Council and Alberta Heritage Foundation for Medical Research. $8 To whom reprints requests should be addressed:Dept. of Biological Sciences, University of Calgary, 2500 University Dr., N.W., Calgary, Alberta T2N 1N4, Canada. The abbreviations used are: MT(s),metallothionein(s);MRE(s),

low molecular weight, high cysteine content, and the ability to bind heavy metal ions(1,2). In man, theMTs are encoded by a multigene family that exist as a gene cluster on chromosome 16q22 (3, 4). The MTs are grouped into two classes, MT-I and MT-11, according to their electrophoretic mobilities. The MT-I family is encoded by five known functionalmembers: MT-IA, IB, IE, IF, and IG genes, whereas MT-I1 protein is derived from a single functional gene, MT-IIA (5-10). MT expressionprovides an attractive model for studying generegulation, because thesegenesaretranscriptionally regulated by a variety of factors including: cadmium, copper, zinc, and glucocorticoid hormone (10-15). Examination of a variety of cultured cell lines reveals that the humanMT gene isoforms are expressed in a cell type-specific fashion. Furthera differmore, within a single cell line individual genes exhibit ential response to heavy metals(8-10, 16). For example, exposure of Hep G2 cells to heavy metal ions increases relative mRNA abundance in the following order: MT-IF < MT-IG < MT-IIA (17). In contrast to the complex nature of the MT-IIA promoter (reviewed in Ref. I), the MT-I promoters are relativelysimple. They are comprised of metal response elements (MREs) that are usually 12 bp in length, a variable number of GC motifs, and a TATA element (10). Such a limited number of cis-acting elements in theMT-I promoters make them an ideal model for studying mechanismsby which metal ions induce gene activity. Although previous studies have shown that metal ion induction increases MT-I promoter activity, the absolute magnitude of response for each examined gene was different (10,17). To investigatethepotentialexplanation for variablepromoter function in response to the metal ion Cd2+,we chose to examine two of the MT-I promoters, MT-IG and MT-IF, because of the marked difference in the magnitude of their response to this ion, In an attempt to define a potential mechanism for differential regulationof MT-IGversus MT-IF, we have measured the activity of these promoters using both transient transfection and cell-free transcription assays. Our results lead us to a simple explanation that may account for the differing levels of activity between MT-IG and MT-IF promoters. EXPERIMENTAL. PROCEDURES

Cell Culture-Hep G2 cells (from the American Type Culture Collection) were maintained as anchoragedependentcultures in 75-cm2 flasks containing 25 ml of modified Eagle’s minimumessential medium (Life Technologies,Inc.) supplemented with10%fetal calf serum, 0.15% sodium bicarbonate, 1 mM glutamine, 1 mM pyruvate, 100 units/ml penicillin, and 100 &ml streptomycin media. Cd2+ inductionwas performed 7 h prior to harvesting by the addition of 2 PM CdCl,. These conditions are optimal for MT induction (18).Cultures were subcultured 1:3 every 68-72 h in order to maintain logarithmic growth and were incubated at 37 “C in the presence of 5% COz.

metal responsive elementb); bp, base pair(s); CAT, chloramphenicol acetyltransferase; NFT-11, neomycin phosphotransferase-11; Pipes, 1,4piperazinediethanesulfonic acid.

24460

MT-I Gene Vitro Danscription in

24461

TABLEI Characteristics of oligonucleotides used for primerextension analysis ~~

Oligonucleotide sequence ~~

Size

Tda

bases

(C) 50

Species detected

Extended lengthh

P-Actin MT-1, MT-1, MT-1, MT-IF MT-1, MT-IIA

91,94 73, 76 69, 72 76,80 71, 74 75 77, 80

Ref.

~

dATCATCCATGGTGAGCT dCATTTCGAGCCGCGGT dCATTGTGGAGCCAAGAC 52 dGGGTCCATTTCGAGCAA dCATTGCAAGCCGAGGA dCATTGCAACCCGAGGC dGGATCCATGGCGAGCT

17 16 17 17 16 16 16

52 52 50 52

52

bases

20 6 8 7 9 10 5 __

Td values were calculated by the Wallace rule (24). The references cited provide data that indicate the anticipated extension product sizes. promoters with TATA box mutations were created by cloning the approQuantitation of RNA by Primer Extension-Cytoplasmic RNA was isolated from Hep G2 cellsby the procedure describedby Favaloro (19). priate 5' regions back onto the deleted mutated promoters. Dansient Dansfection-Plasmid DNA was prepared by cesium chloPrimer extension analysis employed MT isoform specific oligonucleotides (Table I) that were radiolabeled at the 5' end (21) usingT, poly- ride double banding. Hep G2 cells were plated at a density of 275,000 nucleotide kinase (Pharmacia LKB Biotechnology Inc.) and HPLC-pu- cells/60-mm plate. After 60-h growth, cultures werere-fed with 2 ml of rified [ Y - ~ ~ P ] A T (4500 P Ci/mmol; ICN). To control for variationsin media. Following 1-2-h incubation, 200 p1of Ca,(PO,),-DNA precipisample manipulation all reactions contained a P-actin-specific oligo- tate was added to each plate. The DNA mixture contained 0.70 pmol of nucleotide that waslabeled with lower specific activity [y-32P1ATP(450 a pMEVlRderivativecontaining an MT-I promoter, 0.70 pmol of Ci/mmol). Primer extension analysis was performed by modifying the pMEV39R, and sufficient pMEVOR to adjust the totalDNA mass to 10 100 fmol of a singleMT procedure of Williams and Mason (22). In brief, pg. After 3-h incubation, the media was removed and cells were glycerol of isoform-specific primer, 100 fmol of P-actin-specific primer, and 30 pg shocked for 60 s with 15%glycerol in 25 mM HEPES, 1.4mM NaZHP04, RNA were mixed in 2pl of 10 mM Pipes (pH 6.4), 0.4M NaCI, and then and 140 mM NaCl (pH 7.06). Monolayers were then rinsed with Hanks' annealed a t 46 "C for 18 h under mineral oil. To this mixture, we then balanced salt solution (Life Technologies, Inc.), and fed media containadded 10 pl of 50 mM Tris-CI (pH 8.3),6 mM MgCI,, 1 mM dithiothreitol, ing 5 mM of HEPES (pH8.1).After 6 h, 2 p~ CdCI, was addedto half of 1 mM of each dNTP, with 10 units of reverse transcriptase (Molecular the samples and after an additional 7 h, cells were harvested. Genetics Resources), and the reaction was incubated for 2 h a t 46 "C. CAT and NPTIIAssays-Cell extracts were preparedas described in Reaction products were resolved on a 6% denaturing polyacrylamide de Wet et al. (28), and protein concentration was determined using gel. Gel regions containing full-length extension products were isolatedbovine serum albumin as a standard (29). CAT activity was determined andcountedin Aquasol-2 (DuPont).Relativeexpressionwasdeterby the procedure of Gorman et al. (30) and was expressed as picomoles mined by dividing the amount of radioactivity in the band correspond- of chloramphenicol acetylated permidpg of protein. NPT-I1 assays (31 ) ing to the MT isoform by that in the p-actin band. The various MT wereperformed on 20 pg samples of cell extract, and activity was isoform-specific primers were designed to be of a similar size and Td expressed as picomoles of kanamycin phosphorylated per 45mid20 pg (Table I), so that hybridization to the corresponding mRNAs would of protein. Promoter activity was expressedas units of relative activity occur with equal efficiency. Descending chromatography on DE-81 paby obtaining the CATNPT-I1 ratiofor each sample. per (Whatman) (23) was used confirm to that all primers were radiolaCell-free in Vitro Danscription-To measure activity of the MT-I probeled to comparable specific activities. moters we inserted fragments into the G-free cassette vector, p(C2AT)19 Dansfection Plasmid Templates-Maps describing the vectors used (32). This plasmid was prepared by Sal1 digestion followed by endin this report are displayed inFig. 3. Transfection studiesemployed the constructs pMEVOR, derivatives of pMEVlR which contained MT-I pro- filling, EcoRI digestion, and dephosphorylation. MT-I DNA fragments (33)with the appromoters, and pMEV39R. The detailed construction and characterization were synthesized using polymerase chain reaction as templates. TheMT-IG\179and MT-IF\177 of these vectors shall be described elsewhere., pMEVOR was used as priate pMEVlR derivatives carrier DNA in transfection experiments. pMEVlR derivatives contain-promoter fragments were amplified from the MT-IG\240 or MT-IF\243 ing MT-I promoters were also used to create single-stranded DNA tem- templates using the SP6 sequencing primer in conjunction with the primers UTAYFGTGGTGGACCCCAACA or UTATTGGAGTGplatesrequired for site-directedmutagenesis of MT-I promoterseGAGCCCAACA, respectively. The amplified fragments were digested quences. pMEV39R, which contains the MT-IG\240 promoter (Fig. 4) to driving theneo gene, served as a transfection efficiency control and was with EcoRI prior to ligation. The resulting constructs were sequenced confirm that mutations hadnot occurred during the polymerase chain co-transfected along with individual pMEVlR derivatives. MT-I promoters (Fig.4) were insertedas blunt end fragments into thereaction amplification process. Transcriptionallycompetenthepatonuclearextractwasprepared SmaI site of pMEVlR so as to drive the CAT gene. MT-I,\677 was obtained from phMT-I, (IO) as an ApaI-AuaI fragment (-611 to +66) from male Sprague-Dawley rats as described previously (34, 35). Male from CharlesRiver which was subsequently end-filled. Partial digestion of the MT-IG\677 rats,(standardpathogen-free),wereobtained Breeding Company (St. Constance, Quebec). Transcription reactions fragment with BbuI removed the upstream 437 bp to createMT-I,\240 (-174 to +66). MT-IF\489 (-418 to +71) was isolated from pMT-IF5' contained 60 pg of nuclear extract,0.1 pg of the internal control plasmid (provided by C. Sadhu; ICOS Corporation, Bothell, WA) by XbaI-SstI pAdML19O(361, and 1.5 pg of a p(CzAT)19 derivative containing a n digestion and then blunt-ended using T4 polymerase (Pharmacia). To MT-I promoter. Reactions were performed essentially as described by obtain MT-IF\243(-172 to +71), MT-IF\489wasSacI-digested and blunt- Gorski etal. (34) exceptthat KC1 and MgCI, concentrations were50 and ended by T4 polymerase. 9.5 m, respectively. Point mutations were created in MT-IF\243 and MT-I,\240 promoter fragments using the uracil strand selection method (25). Mutagenesis RESULTS reactions were performed according to the method of Zoller and Smith Characteristics of MT-I Expression in Hep G2 Cells Tkeated (26) except that reactions were performed with only the mutagenic oligonucleotide. The oligonucleotides dCCTGACTAT@AAGCAGCCG, with and without Cd2+-To determine whether quantitative differences in expression exist among members of the human dCCTGACTAT&AAGCAGCGG, and dACCCAGGACCGC=GC@E a MT-I multigene family, we measured therelative expression of GCPGACAGGC were used to create a TATCAbox in-M-1,\240, TATAA box in MT-IF\243, and a GCbox in MT-IF\243, respectively. all known functional MT genes in the human hepatoma cell Underlined characters indicate mutated positions. Mutated promoter line, Hep G2. In these experiments, cytoplasmic RNA from fragments were verified by nucleotide sequencing (27), removed from control and Cd2+-treated HepGZ cells was subjected to primer themutation vector by SstI-Sal1digestion,andreinsertedinto pMEVIR. Thislatter procedure insured that the structure of the parent extension analysis so as to quantitate MT-IA, MT-IB,MT-IE, vector was not altered by the mutagenesis reaction. Undeleted MT-I MT-IF, MT-Ic, and MT-11, mRNA accumulation relative to that N. W. Shworak and L. Gedamu, manuscript in preparation.

of p-actin (Fig. 1).Cells cultured in the absence of Cd2+exhibited weak expression of MT-I,, MT-I,. MT-I,. and MT-11, that

MT-I Gene Danscription in Vitro

24462

A 105 .

b.0

91

78

MT primer:

78 758

0:

4 4

75

4

a.

None

m

z

1

0:

E

2:

0.

Relative expression: 0.97 68 4.2 13.7 Frc. 1. Quantitationof human MT cytoplasmic mRNk Primer extension analysis was performed on cytoplasmic RNAobtained from control and Cd2+-treated HepG2 cells a s described under “ExperimentalProcedures.” Reactions contained an MT isoform-specific primer and a p-actinspecific primer labeled tolow specific activity. The small arrows indicate the expected position for the variousMT extension products, whereas the large arrow denotes the expected position of the p-actin extension product. Cd”-stimulated MT mRNA accumulation is expressed in units of relative expression which was determined by standardization to the p-actin extension product. The low background of the p-actin primer is indicated by two reactions which contained only the &actin primer labeled to high specific activity. Extension reactions are flanked by DNA size markers that were created by 5‘ end-labeling Sau3A-digested pGEM3. A is a 2-h exposure, whereas B is a 15-h exposure. -260-270-280-290 m-1: -300-310 -350 -320-330-340 A Ccgt GAg aGGGGtGGGGTGGAGGCGaccGaAAcGCCAaGgctg GGTtCCCGGAACgCgcgGGGaccaGgGTgGaAGGCAACtTCGGgGAAAC GGG E CcgcgGAgcaGGGGtGGGGTGGAGaCGcccGcgAcGCCAaGgctg GGgtCCCGGAACaCgctGGGaggaGgGTgGaAGGCAACCTCGGgGAAACTGGG F gaaggaccggGGcGgGGccgGGgGGCGgggcgAAgGCCAgGatctccaGGTaCCCGGAACcCcaaGGGgcaqGtGTaGcAGGCAAtCTtGGcGAAACTGGG G CtagaGAcctGGGcgttGGTGaAGGtGaggGgAAaGCCAgaatactggGGTaggCGG~CgCcaaGGGqqcqGtGTaGcAGGCAACCTCaG GgAgCTGGG

220

-230

-240 A

E

F G

-140

-250

-150

-160

B A

E F G

-90 B A

E

F G

, ,, -170

-180

-210 -190

-200

AAaGGCGacCgGGAcctC GGGgAcGCCcCgtACCc cGGGcGtAaAct CaCTCCCGCGttAGCGGGCgCCAAAgcggggagGGGgtGGTCCcGtGGTcc GGGaaGGTCCcGCGGTGc AAaGGCGGcCtGGActtC GGGAAcaCCGCgtACCtgcGGGGGCACAg CcCaCCCGaGcGAaCGGGCtCCAAA GGGacGGTCC GCGGTGt AAqGGCGGgCaGGAgggCaGGGAA GCCGCtcACCa GGCACAaagcgcCTCCCGCttGAGCGGaCtCCAAA GGGgccGTCCtGCGGTGt AAgGcCtGgC AggggaGaGAA GCCGCacAC GGGGGCACAg CtCTCCCtCGcGAGCCGGtgCCAAA -100 -120

-130

-80

,7?

GGGCaGTGCaG

- -[q -I;[ -6!,

x

-50

ACTCAGCa GGGGtcacTGCA AGGGCc aGGCGG GGCCTCTGCaCCCcaCCa ACTCAGC GGGGCGGGcGCA gGcGCa FGGCGG GtCCTt TGCGtCCGGCCC ACTCAGC GGGcCGGGTGCA gGcGCG GaGCtG GGCCTC TGCGCCgGGCCC ACag GC GGGGCGGGTGCA AaGaCG GGGCGG GGtCTC TGCGCCCGGCCC ACTCAGC GGGctGGGTGCA AGGGCG GGGCGG GGCgTC TGCGCCCGGCCC

-

-110

ccGCaGcGCtG GGGCgGTGCGG

-20

-10

TATAAAGgAGCaGCCaGCTcCTGGGCTCC CC TCCCCcGAC tCttTCCCCTGAC cATAAA AGCAGCcGCtGGCTGCTGGGCcCt g aCCtCcGtC TATAAA tagAGCaGCCaGtTGCaGGGCTCC TATcAA AGCAGCgGCCGGCTGtTGGGgTCC CC TCCCCTGAC TATAAA AGCAGCcGCtGGCTGtTGGGCTCC Cg TCtCCTGAC

-

Frc. 2. Comparison of 5‘-untranscribedregions from the functional human MT-I genes. Sequences of the 5’-untranscribed regions of the known functional MT-I genes were obtainedfrom the references cited in Table I and were aligned with the assistance of the program NUCALN (37). Theregion of MT-IRdistal to-128 is not displayed as it isextremely nonhomologous. Homologous nucleotides are in uppercase letters, whereas to the MT-IGsequence. Within thisregion, MT-Io and MT-IFare 69% nonhomologous nucleotides are inlowercase letters. The numbering is relative homologous. Three typesof recognizable putative regulatory elements displayhomologous localization: a TATA box (double-line box), four MREs (single-line box, arrows indicate orientation), and a GC box (thick-line box). Additional GC boxes (underlined) are located throughout thedisplayed region. The location of the reciprocal TATA box mutation a t -27 is indicated by x, whereas theMT-Io GC3 region is in boldface.

was only detected by prolonged autoradiography of primer extension products. In the presence of Cd2+,mRNA levels were readily detected for all MT-I family members except MT-IB. Although a similar fold rise in MT-IE,MT-IF, and MT-IG mRNA was observed in response to Cd”, the absolute level of mRNA accumulation for each MT isoform was different. The relative order of expression from weakest to strongest was: MT-IB < MT-IA < MT-IF < MT-IG < MT-IIA 5 MT-IE. Together these results indicate that in Hep G2 cells, all detectable membersof the MT-I gene family are induced by Cd2+, but each isoform responds with a unique level of mRNA accumulation. Differential Expression of MT-I Genes Is Due to Promoter Activity-To determine whetherdifferential regulation of MT-I

isoforms could originate from the promoters of these genes, we compared the nucleotide sequences immediately flanking the 5’ end of all MT-I genes (Fig. 2). This analysisrevealed a striking sequence conservation among all MT-I gene promoters. Findings of note included: 1) sequence homology is most pronounced among the MT-I genes with the highest expression levels (MTIE, MT-IF, and MT-I,); 2) the most invariable region occurs in the sequence-spanning nucleotides -1 to -170 of the genes; and 3) within this region the composition and distribution of regulatory elements is almost identical. The profound conservation of the MT-I regulatory array argues against the possibility that MT-I isoform differential expression results from gross alterations in regulatory element

MT-I Gene 7Fanscription Vitro in composition. This observation prompted us to consider an alternative hypothesis thatsubtle differences within specific regulatory elements may be responsible for the varying levels of MT-I expression. To test this possibility, we focused exclusively on the MT-IG and MT-IF genes, because their promoters share strikingsequence homologies (Fig. 2) butdiffer markedly in their activities. Initially, we constructed two templates, MT-1~\677 andMTIF\489, which, respectively, contain MT-IG and MT-IF promoters fused to the CAT reporter gene of the expression vector pMEVlR (Fig. 3). Fig. 4 schematically displays the promoter regions of the parental constructs (MT-1~\677 and MT-I,\489) and derivatives of these templates.To standardize for transfectional variations, MT-IG and MT-IF constructs were co-transfected with pMEV39R (Fig. 3), a plasmid that expresses the reporter enzyme NPT-11 regulated by the MT-IG\240promoter (described below). In Hep G2 cells transfected with MT-1~\677or MT-I,\489, CAT activity was Cd2+-inducible (Fig. 5 ) . In the presence of Cd2+, CAT activity in cells carrying MT-1~\677was 4.7-fold higher than those containing MT-IF\489. Similarly, with Cd2+ induction endogenous MT-IG mRNA accumulation was %fold greater than MT-IF mRNA (Fig. 1).The close correlation between Cd2+ stimulated promoteractivities and endogenous mRNA levels suggests thatdifferential expressionof MT-IGand MT-IF is predominantly attributable to differences in activity between the two promoters. Differences in Activity between the W o Promoters Is Not Due to Lengths of the Sequences Examined-In the preceding study, we analyzed MT-IG and MT-IF promoter fragments that,respectively, encompassed 611 and 418 bp of 5"flanking sequence (Fig. 4). To eliminate the possibility that this difference in length wasresponsible for the increasedactivity of MT-1~\677, we constructed two templates containing5"flanking sequences of similar length,sequence homology, and composition of regulatory elements. Thepromoter regions distal to positions -172 and -174 were deleted from the parental templates(MT-IG\677 and MT-IF\489) togeneratethetruncatedconstructs, MTI ~ \ 2 4 0and MT-1~\243, respectively (Fig. 4). Removal of the MT-IG distal region produced a slight increase (1.3-fold) in basal promoter activity, but Cd2+-induced activity was essentially unchanged (Fig. 5 , compare MT-1~\677with MT-IG\240).In contrast, deletion of the MT-IF distal region, which contains four putatively functional GC boxes (Fig. 4), decreasedbasal expression (6.5-fold) but did not significantly affect Cd2+-inducedpromoter activity (Fig. 5 , compare MT-IF\489with MT-IF\243).Although the basal activity of MT-IF\243 was extremely low, we consider it tobe detectable, since it was2-fold greater than the FIG.3. Vectors employed in cotransfectionanalysis. The three vectors used for co-transfection posses very similar structures. All plasmids contain the ampicillin resistance gene (Amp),an origin of replication (ori), and the filamentous phage fl intergenic region ( P I Zg).MT-I promoters were inserted into the SmaI site of the CAT expression vector pMEVlR. The promoter insertion site is expandedtoindicateuniquerestriction sites.The co-transfection control vector, pMEV39R, expresses the neo gene product (NPTZZ) underthecontrol ofMTIG\240 promoter.Thereportergenesin pMEVlR andpMEV39R each contain a3' SV40 polyadenylation signal (polyA). pMEVOR lacks a reporter gene and was included to maintain a constant mass of transfected DNA.

24463

CAT activity of the promoterless vector, pMEVlR (results not shown). Together, these findings indicate that Cd2+-induced activity is mediated by the proximal promoter regions of both MT-IG and MT-IF. Most importantly, this analysis of comparable fragments indicates that differential promoter activity results from sequence differences within the truncated promoter fragments. A Subtle Difference between M T I p and MTIG Promoters Is Responsible, in Part, forthe Differential Regulation of the W o Genes-Next, we attempted to further identify MT-IG promoter regions that mediate theelevated expression level of this gene. As shown in Fig. 2, there is 80% sequence homology between the proximal promoter regions of MT-IG and MT-IF. Moreover, the two promoters are composed of a similar arrayof regulatory elements that direct transcription.Despite this intense structural similarity, the MT-IGand MT-IF promoters are functionallydifferent. Specifically, in Cd2+-treated cells the relative CAT activity of MT-IGQ40 was 6-fold higher than MT-I&243 (Fig. 5 ) . These observations promptedus to speculate thatsuch marked difference in promoter activity may be due to subtle differences between the two promoters. Although nonhomologous nucleotides are scattered throughout the 5"flanking sequences, we focused on motifs of known regulatory elements and noted a single base pair difference between the TATA elements of the two genes. The MT-IG TATA motif consists of the canonical element TATAAA (38, 391, whereas the MT-IF gene has an infrequent (40)variant, TATCAA (Fig. 2). The only difference between the two motifs being the A + C transversion at the underlined position. Next, we tested the possibility that this transversion mediatesdifferential regulation by measuring the activity of two constructs, MT-IG\240-TATCA and MT-IF\243-TATAA, which contain a point mutation in the TATA box. These plasmids were generated by using site-directed mutagenesis to, respectively, replace the wild type MT-IG TATAmotif with TATCA and substitute the wild type MT-IF counterpart with TATAA. Relative to the wild type MT-IG\240 promoter, basaland Cd2+-inducedactivities ofMT-IGQ40-TATCA were bothdecreased by 5- and 4-fold, respectively (Fig. 5). In contrast, cells carrying the mutant construct MT-IF\243-TATAA, relative to those transfected with the wild type promoter MT-IF\243, had increased basal andCd2+-induced activitiesof 5.5- and 2.6-fold, respectively. In these experiments, the TATA motif imparted to the mutant promoter differential regulation which mimicked the response of the gene from which the TATA element originated. Although large changes were observed, the A + C mutation failed to completely lower MT-IG\240activity to thelevel of the wild type MT-I,\243 construct.Similarly, the CA

3319,O

Vector used as carrier DNA

Vector for determinationof experimental promoter activity

Co-transfection standardization vector

24464

MT-I Gene ?Fanscription i n Vitro Promoters for transfection. Apa I, -61 1

MT-l~\677

1

2

, ,

Bbvl A

U

Ebv I

A I?

AVBI, +66

Bbvl -174

MT-I ~ \ 2 4 0 sac I

Xba I, -418

FIG.4. Schematic representation of MT-I promoters. Thepromoters MTIG\677, MT-IG\240, MT-IF\489, and MTIF\243 and their point mutation-containing derivatives were inserted into pMEVlR,andtheresultantconstructs were usedfor transfectional analysis. The MT-IG\179and MT-IF\177 promoters and theirpointmutation-containingderivatives were inserted intop(C,AT), and the resultant templates were used for in vitro transcription analysis.

I

MT-I ~ \ 4 8 9

n n w w

A

Bal31, +71 Sst I

A

A

v

u

W

MT-I F Q43

Promoters for in vitro transcription. Bb I -174

MT-I ~ \ 1 7 9

MT-I \177

50 bp TATA box GCbox

A 0

MRE

mutation did not completely increase MT-IF\243 activity up to TATA Motif Variants Affect in Vitro Promoter Activity-The the level of MT-IG\240. These observations demonstrate that preceding studies relied exclusively on the use of cell transfecdifferential expressionof the MT-IG and MT-IF genes is in large tions to examine promoter activity. One way t o verify these part mediated by their respective TATA motifs. Thus, it ap- observations is to determine whether theTATA motif variants pears thata single nucleotidedifference between the two TATA function in a similar fashion using an entirelydifferent assay. motifs plays an important role in governing both basal and Therefore, we measured MT-IG and MT-IF promoter activity in Cd2+-inducible expression of these genes. a cell-free transcription assay. In order to use this transcription Although the TATA motif plays a major role in differential system, we constructed plasmids containing the promoters of regulation of the two genes, it is unlikely to be the sole respon- interest attached to a 377-bp reporter "G-free" cassette (32). sible factor. With this in mind, we attempted to localize addi- Thetemplates, MT-IG\179 and MT-IF\177 (Fig. 4), contain tional sequences that mightalso be involved in thisprocess. For MT-IG and MT-IF promotersequences spanning nucleotides this study we focused on the area of highest sequence diver- -174 to +5 and -172 to +5, respectively; two additional temgence between the two proximal promoters, the segment situ- plates containing reciprocal transversions in the TATA motifs ated between MREb and MREc which we designate as GC3 were also constructed. The transcriptional activity of all con(Fig. 2). Although the GC3 region of MT-IG contains a GC box, structs was assessed using rat hepatonuclear extracts. Cellthe MT-IF segment does not. Therefore,we mutated thisregion free transcription reactions contained a MT-I promoter temin the MT-IF\243 promoter so that it was identical to the se- plate, and as an internal control, the adenovirus 2 major late quence found from -104 to -93 in MT-1~\240 (Fig. 2). Transfec- promoter attached to a truncated 200-bp G-free cassette. The MT-I and tional analysis of this mutantshowed that the basal and Cd2+- differing lengths of the G-free cassette attached to the induced promoter activities were increased by factors of 2.3- adenovirus 2 major late promoters give rise toRNA products of and 1.3-fold, respectively (Fig. 5 , compare MT-1~\243with MT- distinct size that are readily resolved by denaturing gel elecIF\243-GC3). These results show that the GC3 region has a trophoresis, thus autoradiographs display two discrete bands minimal quantitativeeffect on Cd2+-modulated promoter activ- (Fig. 6). The presence of a 380-base radiolabeled transcript indicates ity, and its contribution is much less than that of the TATA that theMT-I promoters were activeand initiated transcription motif. Effect of TATA Motif Variant Is a n Intrinsic Property Znde- at the predicted site. Consistent with data from transfection pendent of Distal Sequences-If the TATA motif is a major de- studies, MT-IG\179 was more active than MT-IF\177. However, terminant of differential promoter activity, then the effect of an A + C mutation in the TATA motif of MT-1~\179reduced the variantsshould notbe altered by the inclusion of additional activity to undetectable levels. In contrast, thereciprocal muupstream sequences. Hence, we examined the effect of mutat- tation in MT-1~\177 increased activity substantially (Fig. 6). ing theTATAmotif in thefull-length constructs,MT-1~\677 and Although we have attempted toinduce MT-I promoter activity not clear at this time, MT-IF\489. The A + C change in the MT-IG\677 TATA motif with theaddition of Cd2+,the results are labile decreased basal and Cd2+-inducedactivity by a factor of 5- and possibly because the required transcription factods) are 4-fold, respectively (Fig. 5, compare MT-1~\677with MT-1~\677- under our present extraction procedure. Alternatively, the exTATCA). The reciprocal mutation inMT-IF\489 increased basal tracts may not be totally free of copper or zinc ions, which will and Cd2+-induced activity by a factor of 1.5- and 2.7-fold, re- also induce MT promoter activity. Metal induction studies will spectively (compare "l?-IF\489 with MT-I,\489-TATAA). These have to await the development of a reconstitution assay. Nevobservations confirm our speculation that theTATAmotif plays ertheless, the activity of wild type and mutated MT-IG and a dominant role in regulatingdifferential expressionof the two MT-IF promoters in the cell-free transcription system closely mimicked their respective activityin the transient transfection promoters.

MT-I Vitro Gene Dunscription in 90

90

24465

Promoter:

MT-IG\179

MT-I ~ \ 1 7 7

W Cd70

70

El Cd+

Variant:

TATAA

50 30

" ._ .* m 20 0 > c

MT

+ @

TATCA

TATCA

.

TATAA

- 4

AdPML

T

a,

> ._ 15

a, U

10

5

0

Size: Promoter: Variant:

I

MT-IF TATCA

MT-IG

I

MT-IFTA;AAMT-IG

FIG.6. In vitro transcriptional activity of MT-I promoters. In vitro transcription reactions were performed as described under "Experimental Procedures." The large triangle indicates the position of the transcript derived from the control vector pAdML19O. The small triangle denotes the position of transcripts derived from MT-I promoter containing vectors. Identical results were obtained with two different plasmidpreparationsusingthreedifferenthepatonuclearextract preparations.

TATA element plays a key role in MT-I gene expression. Furthermore, recent studies of the sheep MT-I multigene family showed that thetwo most highly expressed members both contain canonical TATA boxes (TATAAA), whereas theweak sheep FIG.5.Activity of transfected MT-I promoters. Cells were trans- MT-IB has an unusual variant (TACAAA) (41,42). Thus,TATA fected with 0.7 pmol of a MT-I promoter-bearing derivativeof pMEVlR (CAT expressionvector), 0.7 pmol of thestandardization vector box variants may also contribute to the differential regulation pMEV39R (NPT-I1 expression vector), and sufficientpMEVOR to bring of MT-I genes in other species. the totalDNA mass to10 pg. Cultures were either untreated or induced This assertion is consistentwithobservations,indicating with 2 p~ Cd2+. Promoter strength was expressed in units of relative activity by standardizing CAT levels to NPT-I1 levels as described under that metal-induced expression is mediated through the TATA "Experimental Procedures." Results are the mean of at least two box. Garrity andWold (43) have found that two classes of tranindependent transfections performed in duplicate with two different scripts are initiated from distinct locations within the mouse plasmid preparations. MT-I promoter. TATA-dependant mRNA are initiated at +1, whereas TATA-independent mRNA are initiated further upassay. This observation serves toconfirm that theTATA motifs stream. Cd2+ induction required a functional TATAbox and in therespective genes are strong determinantsof differential generated transcripts originating at +l. Furthermore, a funcpromoter activity. tional TATA-binding site is required for basal and copper induced expression of the Saccharomyces cerevisiae MT gene (44). DISCUSSION Although the above studies demonstrate thatmetal regulated In this report,we attempted to gain insight into how heavy metal ions regulate gene activityby examining, in HepG2 cells, MT expression requires a functional TATA element, our findboth the basal andCd2+-inducedexpression of all MT-I multi- ings indicate that subtleTATA variants can result in differengene family members. We used primer extension analysis to tial MT-I expression. In summary, we have measured the variation in expression show, for the first time, that mRNA levels for all known MT-I family members, MT-IA, MT-IB, MT-IE, MT-IF, and MT-IG were among all known members of the humanMT-I multigene famdifferent in thepresence or absence of Cd2+.Results revealed a ily. These variations occur despite the marked sequence homoldistinct profile of Cd2+-induced expression that followed the ogy displayed by all MT-I promoters. In the case of human order: MT-IB < MT-IA < MT-IF < MT-IG < MT-IIA 5 MT-IE. The MT-IG and MT-IF, differential expression is in largepart attribvariations in geneactivity observed for MT-I family members, uted to a single base transversion within the TATA box. These provides an ideal system for dissecting the mechanisms of dif- findings demonstrate that differential MT-I gene expression can result from differential functioning of a common set of ferential gene regulation following Cd2+stimulation. With this inmind, we focused on MT-IGand MT-IF and found regulatory components. Further analysis of this phenomenon that theTATA motif plays an important role in regulating the should reveal how MT-I transcription rate is defined by transcription factor interactions. activity differences between these genes for both basal and Cd2+-induced states. Interestingly, several other MT-I genes Acknowledgments-We thank Cecilia Po for her excellent technical possess variant TATA motifs. For example, the human MT-IA assistance, RichardPon for synthesizing oligonucleotides, and Drs. Bill promoter has the variant CATAAA (6).Since the Cd2+-stimu- Aird and Mike Simons for critically reviewing the manuscript. In parlated level of MT-IA expression is much lower than for MT-IF ticular the many thoughtful discussions with Drs. Kostas Iatrou and (Fig. l),this provides additional support for the idea that the Randy Johnston are gratefully appreciated. (r

MT-I Vitro Gene Dunscription in

24466 REFERENCES

1. Hamer, D. H. (1986) Annu. Reu. Biochern. 55, 913-951 2. Kagi, J. H. R., and Nordberg, M. (eds) (1979) Metallothionein, Birkhauser, Basel, Switzerland 3. Karin, M., Eddy, R. L., Henry, W. M., Haley, L. L., Byers, M. G., and Shows, T. B. (1984) Proc. Natl. Acad. Sci. U . S. A . 81, 54965498 4. Schmidt, C . J., Hamer, D. H., and McBride, 0.W. (1984) Science 224, 11041106 5. Karin, M., and Richards, R. I. (1982) Nature 299, 797-802 6. Richards, R. I., Heguy, A,, and Karin, M. (1984) Cell 37, 263-272 7. Schmidt, C. J., Jubier, M.F., and Hamer, D. H. (1985) J . Biol. Chern. 260, 7731-7737 8. Heguy, A,, West, A., Richards, R. I., and Karin, M. (1986) Mol. Cell. Biol. 6, 2149-2157 9. Varshney, U., .Jahroudi, N., Foster, R., and Gedamu, L. (1986) Mol. Cell. Bid. 6,2637 10. Foster, R., Jahroudi, N., Varshney, U., and Gedamu, L. (1988) J . Biol. Chem. 263, 11528-11535 11. Karin, M., Haslinger, A,, Holtgreve, H., Richards, R. I., Krauter, P., Westphal, H. M., and Beato, M. (1984)Nature 308, 513-519 12. Sadhu, C . , and Gedamu, L. (1989) Mol. Cell. Biol. 9, 5738-5741 13. Searle, P. F., Davidson, B. L., Stuart, G. W., Wilkie, T.M., Norstedt, G., and Palmiter, R. D. (1984) Mol. Cell. Biol. 4, 1221-1230 14. Hager, L. G., and Palmiter, R. D. (1981) Nature 291, 340-342 15. Karin, M., Haslinger, A,, Holtgreve, H., Cathala, G., Slater, E., and Baxter, J. (1984) Cell 36,371-379 16. Schmidt, C. J., and Hamer, D. H. (1986) Prm. Natl. Acad. Sci. U. S. A. 83, 334G3350 17. Sadhu. C., and Gedamu, L. (1988) J. Biol. Chem. 263, 2679-2684 18. Jahroudi, N. (1988)Studies on ihe Regulaison ofHuman MetallothiominGene Expression. PhD.thesis,TheUniversity of Calgary,Calgary,Alberta, Canada 19. Favaloro, J., Treisman, R., and Kamen, R. (1980)Methods Enzymol. 65, 718749 20. Ng. S.-Y., Gunning, P., Eddy, R., Ponte, P., Leavitt. J., Shows, T., and Kedes, L. (1985) Mol. Cell. Biol. 5, 2720-2732 21. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A

LaboratoryManual, pp. 122-123, Cold SpringHarhorLaboratory, Cold Spring Harbor, NY 22. Williams, J. G., and Mason, P. J. (1985) in Nucleic acidHybridization:A Practical Approach (Hames, B. D., and Higgins,S. J., eds) pp. 139-160, IRL Press, Oxford 23. Falaschi, A., Adler, J., and Khorana, H. G. (1963) J . Biol. Chern. 238, 30803085 24. Wallace, R. B., and Miyada, C. G. (1987) Methods Enzymol. 152,432-442 25. Kunkel, T.A. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 4 8 8 4 9 2 26. Zoller, M. J., and Smith, M. (1984) DNA ( N . Y ) 3, 4 7 9 4 8 8 27. Sanger, F., Nicklen, S., and Coulson,A. R. (1977)Proc. Nail. Acad. Sci.U. S. A. 74,5463-5467 28. de Wet, J. R., Wood, K. V.,De Luca, M., Helinski, D. R., and Subramani, S. (1987) Mol. Cell. Biol. 7, 725-737 29. Bradford, M.M. (19761Anal. Biochem. 72,248-254 30. Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2, 1044-1051 31. McDonnell, R. E., Clark, R. D., Smith, W. A., and Hinchee. M. A. (1987) Plant Mol. Biol. Reporter 5, 380-386 32. Sawadogo, M., and Roeder, R. G. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 43944398 33. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B., and Erlich, H. A. (1988) Science 239, 4 8 7 4 9 1 34. Gorski, K., Carneiro, M., and Schibler, U.(1986) Cell 47, 767-776 35. Lichtsteiner, S., Wuarin, J., and Schibler, U. (1987) Cell 51, 963-973 36. Lichtsteiner, S., and Schibler, U. (1989) Cell 57, 1179-1187 37. Wilbur, W. J., and Lipman, D. J. (1982) Proc. Natl. Acad. Sci. U. S. A. 80, 726730 38. Chen, W., and Struhl, K. (1988)Prm. Natl. Acad. Sci. U. S. A. 85, 2691-2695 39. Wobbe. C. R., and Struhl, K. (1990) Mol. Cell. Biol. 10, 3859-3867 40. Breathnach, R., and Chambon, P. (1981) Annu. Rev. Biochern. 50,349-383 41. Peterson, M. G., Hannan, F., and Mercer, J. F. B. (1988) Eur. J . Biochern. 174, 417-424 42. Peterson, M. G., and Mercer, J. F.B. (1988) Eur. J . Biochem. 1 7 4 , 4 2 5 4 2 9 43. Garrity, P. A,, and Wold, B. J. (1990) Mol. Cell. Biol. 10, 56465654 44. Kambadur, R., Culotta, V.,and Hamer,D. (1990) Proc. Natl. Acad. Sci.U. S. A . 87,9168-9172