Gene, 103 (1991) 269-274 t0 1991 Elsevier Science Publishers

GENE

B.V. All rights reserved.

269

0378-I 119191/$03.50

05003

Cloning

and characterization

(Alcohol

dehydrogenase;

of the human ADH4

gene family;

intronjexon

gene

arrangement;

promoter

region;

regulatory

elements;

transcription

start

point)

Hedvig von Bahr-LindstrGm *, Hans JGrnvall and Jan-Olov Hiiiig Department of Chemistry I, Karolinska Institutet. S-104 01 Stockholm (Sweden) Received by J.K.C. Knowles: Revised: I1 February 1991 Accepted: 12 March 1991

5 November

1990

SUMMARY

Human alcohol dehydrogenase (ADH) constitutes a set of isozymes and enzymes with different tissue and substrate specificities. The subunits are coded for by at least five gene loci, ADHl-ADH5. We now report the cloning and analysis of the human ADH4 gene coding for the class-11 ADH with rr-subunits. The gene spans a region of 21 kb and is divided into nine exons and eight introns. The arrangement is the same as for all analyzed mammalian class-1 genes, but the region covered is 50”,, larger than that in the human class-I genes. The nucleotide (nt) sequences of the exons, exon/intron boundaries and 5’- and 3’-untranslated regions were determined. The transcription start point (tsp) of the ADH4 gene was defined by primer extension and localized to a position 61 nt upstream from the ATG start codon. A TATA box and a CAAT element were identified by homology to consensus sequences for tsp. No DNA structures homologous to the glucocorticoidresponsive elements (GRE) present in the ADH2 gene were found in the upstream region ofthe ADH4 gene, but two structures with a 70”, identity to the GRE consensus sequence were found at nonhomologous locations. The difference and the overall low degree ofidentity, 41’:,, ofthe upstream regions suggest different regulatory mechanisms for the class-l and class-11 genes.

INTRODUCTION

Human ADH exists as a large number of dimeric isozymes and enzymes in various tissues. Three classes of the enzyme can be distinguished by substrate specificities and

Correspondence Karolinska

10: Dr.

Institutet,

Tel. (46-8)7287740; * Present

address:

Stockholm

(Sweden)

J.-O.

H(iiig,

Department

S-104 01 Stockholm

of Chemistry

I,

(Sweden)

Fax (46-8)338453. Kabi

Pharmacia,

Peptide

Hormones,

S-112

87

Tel. (46.8)6958119. Abbreviations:

aa, amino acid(s); ADH,

alcohol dehydrogenase;

ADH2, ADH3, ADH4 and ADHS, human genes encoding

ADHI,

r-, p-, y-, K- and

X-subunits of ADH, respectively; AMV, avian myeloblastosis virus; bp, base pair(s); GRE, glucocorticoid-responsive element(s); kb, kilobase or 1000 bp; nt, nucleotide(s); scription

start point(s).

oligo, oligodeoxyribonucleotide:

rsp, tran-

other properties (Vallee and Bazzonc. 19X3; Jiirnvall et al., 1989). In class-I, the classical liver enzyme in ethanol metabolism, the dimers consist of three different subunit types a(, p and ;I. These subunits are encoded by separate genes. ADHl-3 (Smith et al., 1973). The class-11 ADH with rr-subunits is encoded by ADH4. and the class-111 ADH with X-subunits, recently found to bc identical to glutathione-dependent formaldehyde dehydrogenase (Koivusalo et al., 1989) is encoded byADH5 (Smith, 1986). ADH2 and ADH3 are polymorphic, giving rise to subunits /?J, , 8, &, and yl, y2, respectively. The aa and cDNA sequences have been determined for all human ADH subunits, but genomic structures are only known for the different class-1 ADHs (Duester et al., 1986a; Matsuo and Yokoyama et al., 1989). For the class-III ADH, a pseudogene has recently been identified which lacks a promoter region and all introns (Matsuo and Yokoyama, 1990). The only other mammalian ADH genomic structures known arc those of

class-1 from mouse and rat (Zhang et al.. 1987: Crabb et al.,

RESULTS

AND

DISCUSSION

1989). In all mammalian ADH genomic structures analysed, the coding parts arc interrupted by eight introns and span a region of approx. lo-14 kb, suggesting that this overall arrangement is typical of the class-1 genes at large, independent of species. However, the relationships for the class-11 and class-111 genes representing different enzymes with other specificities and propertics (Vallec and Bazzone, 1983; Jiirnvall et al., 1989; Kaiser et al., 1989). are unknown. As with other proteins, some correlations between cuon/intron borders and domain borders have been detected in ADH (Duester ct al.. 1986b). The human ADHl-3 and ADHS genes have all been localized to chromosomc 4. region q21-25 (Smith, 1986). Recently, the class-11 gene, ADH4, was mapped to 4q22 (McPhearson et al., 1989) which places all human ADH genes in a cluster on chromosome 4.

t\vo (8.6 and 0.9 kb) within the coding region, one (5 kb) covering the promoter region, and one (2 kb) upstream from the gene. The insert of AADH4 : 2 (Fig. 1) is comprised of three fragments, 8.6. 3.6 and 3.5 kb. covering the coding and 3’-untranslated regions. The sum of the fragments identified from each clone is within the limits of the insert size used in the construction of the library and no clone covers the entire ADH4 gene. All EcoRI fragments from the two clones were subcloned into pEMBL8 vectors for furthcr characterization.

Structurally, the class-11 ADH constitutes a separate enzyme (Jiirnvall et al., 19X7), with kinetic properties and a substrate pocket that differs largely from those of the other classes (Eklund et al.. 1990). It is also the least extensively studied of the mammalian ADHs. The aim of the present study was to isolate and to determine the intronjexon arrangement. the tsp. and the upstream sequence of the ADH4 gene coding for the class-11 ADH. The results will be compared to the class-1 ADH genomic structures. introniexon arrangement, and regulatory elements.

I8

XADH~

(b) Introns

and exons

The intron:‘exon structure of the ADH4 gent wxs detcrmined by restriction mapping and sequence analysis of the

I '3

55

2i

20

09

of the human ADH4 gene Using three probes from a previously isolated class-11 cDNA clone (Haiig et al.. 1987) a human genomic library from fetal liver (Lawn et al., 1978) was screened. Two independent clones wcrc isolated: /IADH4 : 1 covering the S’untranslated region and part of the coding region. and iADH4 : 2 covering the 3’ end of the coding region and the 3’-untranslated region. Clone /ZADH4 : 2 was isolated in duplicate from two separate screenings with two different probes covering the 3’ part. Together, the isolated clones covered the entire ADH4 gene (Fig. 1). The insert of AADH4 : 1 (Fig. 1) is comprised of four EcoRI fragments, (a) Cloning

05

Lb

kb

t XADti4-2

Fig. I. Structure

and restriction

map of the human ADH4 gcnc. Exonsxc indicated

by blackened

boxes. The sizes of the mtrons

rclatiw positions of LADH4 : I and iADH4 : 2 arc sho\~n I.ADH4 : 1 stretches about 6 kb upstream is prrscntcd. Restriction sites used for scqucnce analysis and size determination of introns are indicated. iCharon4.4

(Lawn et al., 1978) was screened

using three restriction

arc given in kb. The

from what is shown, while the entire iADH4 : 2 A human fetal genomic library in bacteriophage

enzyme fragments from a class-11 cDNA clone (Hiiiig et al., 1987). .4 51 I-bp

RwI

fragment containing part of the 5’.untranslated region and the 5’ end of the coding region, a 398.bp Rwl fragment from the middle part of the coding region. and a 281&p A4hoII fragment from the 3’.untranslated region were purified from polyacrylamide gels by diffusion elution followed by repented ethanol

precipitation.

The purified

fragments

were lahelcd

with

[ r-“P]dATP

(Amersham)

by nick-translation

(Maniatis

et al., 19X2). The library nas

screcncd, and phage DN.4 from purified plaques was isolated as described (Maniatis et al., 1982). Inserts were liberated byEcoR1 digestion and suhcloncd into pEMBLX vectors. Ohgos unique to each cxon. \+wc synthesircd nith an Applied Biosystems 38 1A instrument and used as probes in blot hybridization (Southern.

1975) of EcoRI digests of the recombinant

phagc i DNA. 5’-end

lahelling of oligos and hybridizations

were performed

et al.. 19X7). All subclones wcrc digested with restrictIon cnqmes (Int. Riotechnologies Inc.) and analyzed by Southern-blot cxon-specific ohgos after separation of fragments by I I’,, agarohc gel electrophoresis. to estimate the intron sizes.

as described

hybridizations

(Hi@

using the

-147

ATMAAGAAA

74

GAGTTTGAAG CTTTCTTAAC TCAGAAAGAA ACTTCCAACA CAGTTTCCCA AAGAAAAATG GGCACCAAGG Met GlyThrLysG

acttttgtgttccatcacag_GCAGCCCTGGACTGTAC~CCGCAGGCTGGGGATCATGTACTTTCATT 276 LysAlaAlaI,euAspCysThrThrAlaGlyTrpGlySerCysThrPheIle

TGTGGATTTTGCCCTTGACTGTGCAGGTGGATCTGAAACCATGgtatgtatattttgtt~ttg.......3.5.kb yVa1AspPheAlaLeuAspCysAlaGlyGlySerGluThrMet 275

GGAGCCACTGACTGCCTCATCCTAGAGACTTACAT~CCGATCCAGGAAGTTATCATTG~TTGACCMGGGAGGG~~GG GlyAlaThrAspCysLeuAsnProArgAspLeuHisLysProIleGlnGluValIleIleGluLeuThrLysGlyGl

ttcctgcttgcagGTCACCCCTGGTTCGACTTGTGTTGG 189 ValThrProGlySerThrCysAlaValPheGlyLeuGlyGlyValGlyLeuSerAlaValMetG

GGCTATGGGGCTGCAATCAACAATGCCAAGgtaaatggttaaacaccaat.......5.5.kb.......actaatt GlyTyrGlyAlaAlaIleAsnAsnAlaLys 188

CAGATATCMTCTTGCCATAGATGATGATGCAAATTTTAGAGAGAGTTTGTCTGCTTGGATGTGGGTTTTCMCT frAspIleAsnLeuAlaLysIleAspAspAspAlaAsnLeuGluArgvalCysLeuLeuGlyCysGlyPheSerThr

CAGCAGGTTTACCTGCAAAGGAAAACCAGTTTACCATTCCCAGTACATTCTCTCAGTACACTGTGGTGT rSerArgPheThrCysLysGlyLysProValTyrHiSPhePheGlyThrSerThrPheSerGlnTyrThrValValS

115 rAsnLeuLysSerProAlaSerAspGlnGlnLeuMetGluspLysTh

. 2.7.kb.......tacattttctgagTMTCTCAAAAGTCCTGCTAGTGATCAACAACTRATGC

AG~TGCMGTTTTGTCTGAGTCCAAATTTTCAGgtaagCaCtCtaCaCtgttt.... ArgLysCysLysPheCySLeuSerProLeuThrAsnLeuCysGlyLysIleSe 115

cgaa......2.6.kb......catatggcctttctttctagGTGACAAAGTAATTCCACTTCCACTTTATGCACCTCTATGT 86 lyAspLysValIleProLeuTyrAlaProLeuCys

GGCCATGAGGCTGCAGGTATTGTGG4AAGTATTGGGCCAGGAGTGACC~CGTC-CCAG~attttattttattC C,lyHisG1IL4laAlaGlyIleValGluSesIleGlyProGlyValThrAsnValLysProG 86

TTGCTACCTCCCTGTGCCATACTGATGCCAGTGTTATCGATTCT~TTTGAGGGCCTAGCTTTCCCAGTGATCGTT leAlaThrSerLeuCySHisThrAspAlaSerValIleAspSerLysPheGluGlyL~~laPheProValIleVal

AGTTCGCATTCAGgtaagtggagactacccctt.......0.9.W,.......aaggatatgattgcctgtagATCA WalArgIleGln 39 40 1le1

GCAAAgtaagcaagtaagCtgtatc.......1.8.kb.......cttttctttctctctttcagGTTATTAAArGC lyLys 5 6 ValIleLysCys AAAGCAGCCATCGCCTGGGMGCAGGCRRGCCCCCTTTGCATTG~GAGGTTGMGTAGCTCCCCCC~GGCTCAT~ LysAlaAlaIleAlaTrpGluAlaGlyLysProLeuCysIleGluGluValGluValAlaPloProProLysAlaHisG1

4

AGGCCGGCA-r GGCTGTGAAT TACAGCAACA AP.GGAGX%M GGMGTGATT

GGAGAATTAA GCAACATGAA TGGTATTATT CAAAGACAGC TCATTATAGG ACACGGAACT CCCTGGCTAG

AGTACAAATG ATGTGGTMG -77

-217

AAAGAATTTA AAAAATCTTG GAGCTCACTG GGAGCAATGG GGTTGCAGCT GAAGT-

-287

GRTTATCAGG GTTGTGAAGG AGAAGAACAG GTAAGTTAAA TGGGCATTCT GAGGAGTAGA AATTTCCTTT

-357

ACTTCATAAT TATTTGTTAA TTCATCTTTA TATGTTTAAT GGGCTTTTCT CTATTATTTT ATATTTTTCA

A .RUGCTTGCT AGATTAACTA TTGATACACA AGCTTAAATA GGTAGTAAAC CACTAAATAT GCAAGGAAGT B

AGATGCCAG

ATAACTAGTG TTTATG~TA

GMTAAGACC

TGGTATTTGA TAGCACrAACA GGGAGACTAT AGTCAACAGC AATTTAA'W';

sequence

indicates

Nuclcotide

determination.

(Amersham)

of introns,

Sequence

5’.

under accession

Nos

will appear

regions. in the EMBL, GenBank

part,

the [ZSS]dATP and DDBJ

the sequence

of the coding

X5641 l-X56419.

were also on single-stranded

fragments

were used throughout

3’-untranslated

for both strands and (Pharmacia) data reported Databases

The sequence

the

nt sequence and and T7 DNA polymerase

end

This gave the complete

determination

Some restriction

and

sequences. templates.

in overlapping

beginning

resulting

into M 13mp18/19 vectors for dideoxy sequence

subcloned

1985) and exon-

method

of sequences by the dideoxy (Chen and Seeburg,

were determined plasmids

to GRE, see Fig. 4. Sequences

TATA box G. An arrow (nt 1920)

The putative

The The sizes

are shown.

(see Fig. I), the

point of the cDNA. For positions

as an outlined

from the cDNA sequence. The tsp is marked

et al., 1977) using alkali-denatured specific primers

(Sangcr

regions

gene. All nine exons

to the coding region is shown below the nt sequence.

the position of the polyadenylation with homology

ADH4

and the 5’- and 3’.untranslated

of the human

are noted, as are numbers

corresponding

of the introns,

and CAAT element are underlined.

of the introns

aa sequence

parts

Fig. 2. Nucleotide proximal

CATCTCATGT ACCCCATCCC ATATAAATGT TTTAAATAAA AATAAAGAAA AGATTTCCAC

AA'TAAAAAGA TTTTTACTGG AAAAAATCAC ATTATTT

TATATACACC TAATACCCAC AAAACTTMA

ACCCCATATT ACATGATGTG ATTATTACAC ATTGCATGCC TATATCAMA

+ TATATTTTAA TATGACTAAA AGAGTATAAT GGATTGTTTG TAACACAAAT AAATGCTTGA GGAGATGGAG

T