Genotyping of human alcohol dehydrogenases at the ADH2 and ADH3 loci following DNA sequence amplification

GENOMICS

2,209-214

(1988)

Genotyping of Human Alcohol Dehydrogenases at the ADH2 and ADH3 Loci following DNA Sequence Amplification YILING Xu, LUCINDA G. CARR, WILLIAM F. BOSRON,TING-KAI LI, AND HOWARD J. EDENBERG Departments of Biochemistry, Medicine, and Medical Genetics, Indiana University Schoolof Medicine, Indianapolis,Indiana46223 Received January 19, 1988

subunits they encode form homo- and heterodimers (Smith, 1986; Bosron and Li, 1986). Polymorphism has been observed at ADH2, producing three /3 subunits (/3i, &, and /lass) and at ADH3, producing y1 and yz subunits (Smith, 1986, Bosron and Li, 1986). The & subunit differs from both fll and fis in having a His4, instead of an Arg (Jiirnvall et al., 1984). The ,!& subunit differs from both & and /32in having a Cysss, instead of Arg (Burnell et al., 1987). These single amino acid differences have dramatic effects on the kinetic properties of the isoenxymes. The pH optimum for ethanol oxidation is 7.0 for &&, ,8.5 for &&, and 10.5 for fll&. The K, for NAD+ varies over 70fold and the V,, for ethanol oxidation varies over 40-fold among the three isozymes (Bosron and Li, 1986). We have shown that, in vitro, under near-physiological conditions of pH, temperature, and substrate and product concentrations, the ethanol oxidation rate of &,& is 4- to 7-fold faster than that of && (Burnell et al., 1987), suggesting that the kinetic differences are likely to be reflected in vim The y2 subunit differs from rl by two amino acids, Glr~~,~for Arg and ValMQfor Ile (Hiiiig et aI., 1986); there is a 2-fold difference in the V,, for ethanol oxidation (Bosron and Li, 1986). It has been suggested that the genetically determined differences in ethanol metabolic rate and/or in the susceptibility to alcoholic liver disease may be determined in part by differences in the isoforms of these alcohol-metabolizing enzymes (Li, 1983; Bosron and Li, 1986). It has, however, been difficult to study the relationship between isozyme type and alcohol elimination rate, since unambiguous determination of the phenotype requires liver biopsy. Therefore, we have been interested in developing methods to determine the genotype at these loci. Here we report the first direct determination of all known polymorphisms at both the ADH2 and the ADH3 loci, using white cell or liver DNA.

Humans are polymorphic at two of the alcohol dehydrogenase (ADD) loci important in ethanol metaboliim, ADZ?2 and ADEE% Although the coding regions of these genes are 94% identical, they produce subunits that differ greatly in kinetic properties in vitro. These differences are likely to be reflected in the pharmacokinetice of alcohol metabolism, but studies have been hampered by the need to use liver biopsy specimens to determine the ADH phenotype. This problem has now been overcome by determining the genotype at these loci using DNA that has been amplitied in uitro by the polymerase chain reaction. We report here the identiflaation of all three of the AD112 alleles and both of the ADH2 alleles. Any pair of ADH2 or ADIIS alleles can be distinguished using allelsspeciilc oligonucleotide probes directed at their single base pair difference. In addition, AD= can be distinguished from ADH2’ and ADHF by detecting a new Moe111 site created in the third exon by the single base pair alteration in ADHP. 8 19SS Acad.amic Pmn, Inc. INTRODUCTION

The first step in alcohol metabolism, the oxidation of ethanol to acetaldehyde, appears to be rate limiting (Li, 1983). This step is catalyzed by alcohol dehydrogenase (ADH; alcohol:NAD+ oxidoreductase, EC 1.1.1.1). Humans have five loci encoding ADHs (Smith, 1986; Bosron and Li, 1986); their kinetic properties in vitro suggest that at pharmacologically relevant levels of ethanol, ADHs composed of a, ~3,y, and a subunits (encoded at ADHI, ADH2, ADH3, and ADH4, respectively) are responsible for ethanol oxidation (Bosron and Li, 1986). ADHI, ADH2, and ADH3 are extremely closely related: their cDNAs are 94% identical in sequence (Ikuta et aZ.,1986; HaSg et aZ., 1986; H&den et al., 1986; von Bahr-Lindstriim et al., 1986), the genes are located in the same region of chromosome 4 (Smith et al, 1986), and the a, 8, and y

209

osss-7543/ss $3.00 Copyright Q 1988 by Academic Press, Inc. All rights of reproduction in any form reserved.

210

XU ET AL.

In order to detect the polymorphic ADH genes, the differences among their DNA sequences were identified. We recently cloned and sequenced a portion of the ADH23 gene (encoding &.) and determined that the Cys for Arg substitution at amino acid 369 is due to a single base pair change from C to T in the ninth exon (Burnell et al., 1987). Using allele-specific oligonucleotide probes (Conner et al., 1983) on enzymatitally amplified DNA from the ninth exon (Saiki et al., 1985,1986; Chehab et al., 1987), we have identified the ADH23 gene. The only 1-bp change in the ,& cDNA sequence (Ikuta et al., 1985; H&den et al., 1986) that could result in the substitution of HisdT in /3* for the Arg in & creates a new Mae111 site in ADH2’. After amplifying the region around exon 3 that contains this difference, we have identified the ADH22 allele by detection of this new restriction site. We have also determined the genotype at this location by use of allele-specific probes. The two ADH3 alleles differ by two amino acid substitutions and at least two other silent changes in the coding region (HiiGg et al., 1986). We have amplified the putative eighth exon of the ADH3 gene and detected the polymorphism by allelespecific oligonucleotide probes directed toward the single base pair substitution that results in a Val in y2 at position 349 instead of the Ile in y1 (HGiig et al., 1986). MATERIALS

AND METHODS

Liver Phenotyping Human liver samples were obtained at autopsy and stored at -70°C. ADH2 (p) andADH3 (y) phenotypes were determined by starch gel electrophoresis and agarose gel isoelectric focusing, followed by staining for ethanol-oxidizing activity (Bosron et al., 1983; Yin et al., 1984). Identification of the ADH3 phenotype in livers containing p2 subunits is difficult: isoenzymes with y2 subunits cannot be readily identified because the activity of y2y2 is much lower than that of &32, and the B2y2 isoenzyme is not resolved by isoelectric focusing. DNA Extraction

and Enzymatic

Amplification

DNA was extracted from frozen portions of livers of known phenotype after the samples were powdered in liquid nitrogen (Maniatis et al., 1982). DNA was extracted from white blood cell pellets obtained after lysis of red cells with ammonium bicarbonate (Kan and Dozy, 1978). Figure 1 shows the gene structure around exons 3 and 9 of the ADH2 gene and the putative exon 8 of the ADH3 gene, and the location and sequence of primers and probes used. The HE45 and HE39 primers at the

5’ side of ADH2 were based on sequences in the second and eighth introns of the ADH23 gene, respectively (unpublished data). Primers HE40 and HE46, at the 3’ side of ADH2, were based on sequences of ADH23 (unpublished) and ADH2’ (Duester et al., 1986). Primers HE41 and HE42, for the ADH3 gene, were based on cDNA sequences (HGbg et al., 1986). The locations of the seventh and eighth introns in ADH3 were inferred from their positions in the closely related ADH2’ and ADH23 genes (Duester et al., 1986; and unpublished data) and the related mouse Adh-1 gene (Zhang et al., 1987). HE42, the primer at the 3’ side of ADH3, includes the nearly universal GT of the putative intron. Where possible, amplification primers were chosen from regions in which ADHl, ADH2, and ADH3 differed, to minimize amplification of the other ADH genes. The sequences of the cDNAs for (Y (Ikuta et al., 1986; von BahrLindstrijm et al., 1986) and y1 and y2 (Ikuta et al., 1986; HGijg et al., 1986) have been reported, although intron sequences have not been. Thirty cycles of enzymatic amplification by the polymerase chain reaction (Saiki et al., 1985, 1986; Chehab et al., 1987) were carried out in 60-~1 reaction mixtures, using the thermostable DNA polymerase from Thermus aquaticus (New England Biolabs) under the manufacturer’s conditions. The temperature was cycled from 93 to 50 to 63°C for 1 min each. Reactions contained 2 pg of genomic DNA from either liver or white blood cells, and 60 pmol each of the two primers; 1 unit of enzyme was added at the beginning, and a second unit was added after cycle 15. Allele Detection by Oligonucleotide

Probing

After the amplification, a fraction of the reaction mixture was electrophoresed through 2% agarose (Fig. 2A). In some experiments to detect ADH23, an aliquot of the mixture was digested with BglII before electrophoresis. After denaturation, the amplified DNA in the agarose gel was transferred (Southern, 1975) to a nylon membrane for probing. Oligonucleotides were labeled with 32P using [T-~~P]ATP and polynucleotide kinase, and free nucleotides were removed by spinning through a Sephadex G-25 column (Maniatis et al., 1982). Prehybridization was performed for 4 h in 5X SSC/BO mM sodium phosphate (pH 7.0)/10X Denhardt’s solution/l% SDS/100 pg ml-l denatured salmon DNA. Overnight hybridization in the same solution, supplemented with Dextran sulfate to 10% and 32P probe at 2 X lo6 cpm ml-‘, was carried out at 55°C (HE31, HE33, HE34), 53°C (HE35), 37°C (HE43), or 40°C (HE44). The membranes were washed for 15 min at room temperature in 6X SSC/O.B% SDS, three times for 5 min in

DETERMINATION

OF HUMAN

6X SSC, and finally once for 5 min in 6X SSC at the discrimination temperature (see figure legends). The membranes were exposed to X-ray film for 1.5 to 4.5 h. Allele Detection by Restriction Digestion: ADHP

Identification

AND

A

Enzyme

For identification of ADHP, the reaction mixtures were ethanol precipitated, and samples were digested overnight with MC&II. Aliquots were electrophoresed on 8% polyacrylamide gels, stained with ethidium bromide, and photographed on Polaroid Type 57 film. RESULTS

211

ADH GENOTYPES

B

DISCUSSION

3’ non-mrlated

of ADH2’

Figure 1B shows the relevant portion of the ADH2 gene. The sequence of ADH2’ starting 15 nt upstream of exon 9 (Duester et aZ., 1986; further sequences were not reported) is identical to that of ADH23 except for the single substitution that changes the Arg,,, in ADH2’ to Cys in ADH2’ (Burnell et al., 1987). The (Y gene, ADHl, differs in 2 and 3 nt of the 20-nt probes from the ADH2 alleles, and so will not hybridize to either ADH2 probe. The sequence of ADH3, however, is identical to that of ADH2’ in this region, as demonstrated by probing unamplified genomic DNA: the & probe HE31, but not the & probe HE33, hybridized to ADH3 as well as to ADH2l (data not shown). Therefore if the ADH3 gene were also amplified, it would hybridize to the ADH2l probe and interfere with the genotyping. To avoid amplification of ADH3, we chose an amplification primer at the 3’ end that differs from the ADH3 sequence by 3 nt. After 30 cycles of amplification of either liver or blood cell DNA, a 202-bp fragment could be clearly visualized by ethidium bromide staining (Fig. 2A). Samples from livers containing only & or & subunits give strong signals with the exon 9 & (and &) probe HE31 but not with the & probe HE33 (Fig. 2, lanes 6 and 9). Conversely, the livers containing only & give strong signals with the p3 probe but not with the & probe (lanes 2 and 8). Heterozygous livers, containing both & and &, hybridized to both probes, each with about half the signal strength of the homozygotes (lanes 1 and 7). Identification of genotypes containing ADH23 alleles was thus unambiguous. Two of the blood samples (766 and 770, lanes 3 and 5, respectively) did not contain ADH23 alleles and one (768, lane 4) was heterozygous for ADH23. There was no significant hybridization when the homozygous p3 samples were probed with the ADH2’-specific oligonucleotide HE31 (Figs. 2B and D, lanes 2 and 8), demonstrating that the ADH3 gene

FIG. 1. Structure of the ADH2 and ADH3 genes in the regions amplified. (A) Exon 3 of ADHP. HE45 = AATCTTTTCTGAATCTGAACAG.HE46 = GAAGGGGGGTCACCAGG’M’GC. HE34 = GTCATCTGTGCGACAGATTC. HE35 = GAATCTGTCACACAGATGAC. (B) Exon 9 of ADH2. HE39 = TGGACTCTCACAACAAGCATGT.HE40 = TTGATAACATCTCTGAAGAGCTGA. HE31 = TGCAGTATCCGTACCGTCCT. HE33 = AGGACGGTACAGATACTGCA. The B&II site shown is not present in either ADHl or ADH3. (Cl Putative exon 8 of ADH3. HE 41 = CTTTAAGAGTAAAGAATCTGTCC. HE42 = ACCTCTTTCCAGAGCGAAGCAG. HE43 = ATAACAAATATMTACCTT. HE44 = AAGGTAAAACATTTGTI’AT. Stippled boxes represent exons, and thick lines represent intron. Vertical bars within the exons mark the site of the alteration. Vertical arrows denote restriction sites, and thick horizontal arrows denote amplification primers or allele-specific oligonucleotide probes.

was not amplified to an extent that would interfere with the determination of the ADH2 genotype. To further confirm this, we digested the amplified DNA with BgZII before electrophoresis; only the ADH2 alleles have a BglII site (Fig. 1B). The 145-bp band resulting from BgZII digestion thus contains only ADH2 alleles and gives the same pattern of hybridization as that seen for the uncut DNA (compare Fig. 2D with B and E with C). Identifiation

of ADH22

Amplification of the region surrounding the codon for amino acid 47 produced a 108-bp segment. Amplified DNA was electrophoresed and probed with the

212

XU ET AL.

had both fragments in approximately equal amounts. There was no interference by either ADHl or ADH3 alleles. Among the white cell samples, we detected one ADH2’ homozygote (Fig. 3C, lane 9) and one ADHP + ADH2l heterozygote (lane 10); three samples (lanes 6,7, and 8) contained no ADHP alleles. The combination of this test and the test for ADH23 unambiguously identifies the ADH2 alleles of all samples (Table 1). of ADH3’

Identification

and ADH32

The intronlexon boundaries of the ADH3 genes have not been reported. Homology to the ADH2 genes and the conservation of intron positions between ADH2 genes and the more distantly related mouse Adh-I gene (Duester et aZ., 1986; Zhang et al, 1987) suggested that the introns of ADH3 should coincide in position with those of related ADH genes. Using this as a guide, we amplified the putative eighth exon of the ADH3 gene, producing a 140-bp segment. The y1 probe differs from both ADHl and ADH2 sequences by 3 nt and the y2 probe differs from both by 2 nt; ,

2

3

4

5

6

7

8

9

10

FIG. 2. Identification of ADH23 in exon 9. (A) Photograph of ethidium bromide-stained gel. (B) Autoradiogram of the same gel, probed with the exon 9 @I (and &&specific oligonucleotide HE31; 6nal washing was at 68’C. (C) Autoradiogram of the same gel, probed with the Be-specific oligonucleotide HE33; final washing was at 65°C. (D) Autoradiogram of a parallel gel containing samples digested with B&II (digestion was incomplete), probed with the &-specific oligonucleotide HE31 as above. (E) Autoradiogram of the same gel as in D, probed with the &specific oligonucleotide HE33 as above. Heavy arrows, 202 bp; open arrows, 145 bp. Lanes 1 and 7, fil + 68 heterozygotes 618 and 555; lanes 2 and 8, & homozygotes 619 and 620; lane 3, white cell DNA 766; lane 4, white cell DNA 768, lane 6, white cell DNA 770; lane 6, & homozygote 485; lane 9, Is, homozygote 483.

exon 3 @I (and &J-specific probe HE34 and the &-apeciflc probe HE35 (Figs. 3A, and B). The discrimination between livers containing & and livers containing either /?i or & was excellent. Since the alteration that substitutes His4, in @z for the Arg in & and /$ also creates a new Moe111 site (Fig. lA), we tried a nonradioactive method of detecting /3zalleles. The amplified DNA was digested with M&II, electrophoresed on 8% acrylamide, and stained with ethidium bromide. As shown in Fig. 3C, lanes 1 and 5, DNA from livers containing only & or & subunits had a 95bp fragment, whereas DNA from livers containing only & subunits (lane 4) lacked this 95-bp fragment and had instead a 60-bp fragment. The DNA from heterozygous livers containing both & and & (lanes 2 and 3)

FIG. 3. Identification of ADHP in exon 3. (A) Autoradiogram of agarose gel probed with exon 3 p1 (and ~&specific oligonucleotide HE34; final washing was at 65°C. (B) Autoradiogram of the same gel probed with exon 3 &specific oligonucleotide HEl35; final washing was at 63°C. Heavy arrows in A and B denote 108 bp. (A and B) Lanes 1 and 8, pi + /3z heterozygotes 530 and 554; lanes 2 and 7, & homozygotes 529 and 483; lane 3, white cell DNA 766; lane 4, white cell DNA 768, lane 5, white cell DNA 770; lane 6, white cell DNA 32; lane 9, @Ihomozygote 485, lane 10, & homozygote 620. (C) Photograph of ethidium bromide-stained 8% polyacrylamide gels. Lane 1, & homozygote 485; lanes 2 and 3, & + j3z heterozygotee 530 and 554; lane 4, & homozygote 483; lane 5, fis homozygote 619; lanes 6-10, white cell DNAs 766,766,770,29, and 32. Light arrows in C show 95- and 60-bp bands.

DETERMINATION

OF HUMAN

therefore neither should hybridize under the conditions used. As shown in Fig. 4, there is a clear discrimination between livers containing only y1 (lanes 3 and lo), livers containing only y2 (lanes 1 and 4), and heterozygous livers containing both y1 and y2 (lane 2). Among the five samples of white cell DNA shown, four (768, 770, 29, and 32; lanes 6, 7, 8, and 9) are homozygous for ADH3l and one (766, lane 5) is heterozygous (Table 1). It is noteworthy that although the y phenotype of livers containing b2 subunits is difficult to determine, the ADH3 genotype is readily identified (Table 1).

213

ADH GENOTYPES 1

2

3

4

5

6

7

8

9

10

A

B

” ‘“”

i

:

Conclusions We can now distinguish all three known alleles at the ADH2 locus and both alleles at the ADH3 locus with a technique that requires only a small sample of DNA extracted from white blood cells. Both homoand heterozygotes can be unambiguously identified by this simple and rapid method of genotyping, eliminating the need for liver biopsies for determining ADH phenotypes. These methods of determining ADH TABLE Phenotypes

DNA

B

genotype now allow studies addressing the roles of the human ADH isozymes in the interindividual differences in pharmacological and pathological effects of alcohol.

1

and Genotypes

Phenotype

Genotype Y

ACKNOWLEDGMENTS

ADHZ

ADH3

22.

191

Liver DNAs 483 485

2,2” 191

529

292 12

530 553 554 555 557 618 619 620 781

l,l

12 193 191 193 3,3 393 191

ndb

22

191

nd nd

292 L2

22

Ll

nd 151

2,2 132

12

l,l 191

292 1,2

193

191

u

22 192

191 191

133 393 3,3

192

l,l

192

292 192

191 191

768

191 193 191

192

770

Chinese’ Chinese Amer. white’ Amer. black’ Chinese

We thank Terry E. Dailey and Colleen Harden for excellent technical assistance. This research was supported by U.S. Public Health Service Grants ROl AA06460, ROl AA02342, and P50 AA07611. L.G.C. was supported by T32 AAO7462.

2,2

191 191

Blood cell DNAs 29 32 766

FIG. 4. Identification of ADH3l and ADH32 in exon 8. (A) Autoradiogram of agarose gel, probed with the exon 8 y,-specific oligonucleotide HE43; final washing was at 45°C. (B) Autoradiogram of the same gel, probed with the yx-specific oligonucleotide HE44; final washing was at 48°C. Heavy arrows denote 140 bp. Lanes 1 and 4, yz homoxygotes 557 and 553; lane 2, yz + y1 heterozygote 618; lanes 3 and 10, y1 homoxygotes 620 and 555, lanes 5-9, white cell DNAs 766,768, 770,29, and 32.

191 191

“2,2 = homoxygous for & or yz; 1,2 = heterozygous, etc. All samples tested for both (& or /3e) vs j3sand (fil or &) vs Be. b nd = not determined, due to presence of &. ’ Gene frequencies are 65% ADHP and 95% ADH3’ for Japanese, >95% ADH2l and 50% ADH3’ for white Americans, and 85% ADH2’ + 15% ADH2’9 and 65% ADH3’ for black Americans (Ref. (l), with frequency of ADHP in Japanese corrected).

REFERENCES 1. BOSRON, W. F., AND LI, T.-K. (1986). Genetic polymorphism of human liver alcohol and aldehyde dehydrogenases, and their relationship to alcohol metabolism and alcoholism. Hepatology 8: 502-510. 2. BOSRON, W. F., MAGNES, L. J., AND LI, T.-K. (1963). Human liver alcohol dehydrogenase: ADH Indianapolis results from genetic polymorphism at the ADH2 gene locus. Biochem. Genet. 21: 735-744. 3. BURNELL, J. C., CARR, L. G., DWULET, F. E., EDENBERG, H. J., LI, T.-K., AND BOSRON, W. F. (1987). The human f13 alcohol dehydrogenase subunit differs from & by a Cys for Arg-369 substitution which decreases NAD(H) binding. Biothem. Biophys. Res. Commun. 146: 1227-1233. 4. CHEHAB, F. F., DOHERTY, M., CAI, S., KAN, Y. W., COOPER, S., AND RUBIN, E. M. (1987). Detection of sickle cell enaemia and thalassaemias. Nature (London) 329: 293-294. 5. CONNER, B. J., REyEs, A. A., MORIN, C., ITAKURA, K., TEPLITZ, R. L., AND WALLACE, R. B. (1963). Detection of sickle cell @-globin allele by hybridization with synthetic oligonucleotides. Proc. Natl. Ad. Sci. USA 80: 216-262. 6. DIJESTER, G., SMITH, M., BILANCHONE, V., AND HATFIELD, G. W. (1986). Molecular analysis of the human class I alcohol

214

I.

8.

9.

10.

11.

12.

13.

14.

XU ET AL. dehydrogenase gene family and nucleotide sequence of the gene encoding the @subunit. J. Biol. Chem. 261: 2027-2033. HBDEN, L.-O., HO&, J.-O., LA-SON, K., LAKE, M., LAGERHOLM, E., HOLMGREN, A., VALLEE, B. L., JORNVALL, H., AND VON BAHR-LINDSTROM, H. (1986). cDNA clones coding for the b-subunit of human liver alcohol dehydrogenase have differently sized 3’-non-coding regions. FEBS Lett. 194: 327-332. Ho&, J.-O., HBDEN, L.-O., LARSSON, K., JORNVALL, H., AND VON BAHR-LINDSTR~M, H. (1988). The y1 and ys subunits of human liver alcohol dehydrogenase: cDNA structures, two amino acid replacements, and compatibility with changes in the enzymatic properties. Eur. J. Biochem. 159: 215-218. IKUTA, T., FUJIYOSHI, T., KURACHI, K., AND YOSHIDA, A. (1985). Molecular cloning of a full-length cDNA for human alcohol dehydrogenase. Proc. Natl. Acad. Sci. USA 82: 2703-2707. IKUTA, T., SZETO, S., AND YOSHIDA, A. (1988). Three human alcohol dehydrogenase subuniti cDNA structure and molecular and evolutionary divergence. Proc. Natl. Acad. Sci. USA 83: 634-638. J~RNVALL, H., HEMPEL, J., VALLEE, B. L., BOSRON, W. F., AND LI, T.-K. (1984). Human liver alcohol dehydrogenase: Amino acid substitution in the &9r Oriental isoxyme explains functional properties, establishes an active site structure, and parallels mutational exchanges in the yeast enzyme. Proc. Natl. Acad. Sci. USA 81: 3024-3028. KAN, Y. W., AND DOZY, A. M. (1978). Polymorphism of DNA sequence adjacent to human 8-globin structural gene: Relationship to sickle mutation. Proc. Natl. Acad. Sci. USA 75: 5631-5635. LI, T.-K. (1983). The absorption, distribution and metabolism of ethanol and its effects on nutrition and hepatic function. In “Medical and Social Aspects of Alcohol Abuse” (B. Tabakoff, P. B. Sutker, and C. L. Randall, Eds.), pp. 47-77, Plenum, New York. MANIATIS, T., FRITSCH, E. F., AND SAMBROOK, J. (1982).

15.

16.

17. 18.

19.

20.

21.

22.

“Molecular Cloning: A Laboratory Manual,” Cold Spring Harbor Press, Cold Spring Harbor, NY. SAIKI, R. K., BUGAWAN, T. L., HORN, G. T., MULLIS, K. B., AND ERLICH, H. A. (1986). Analysis of enxymatically amplified fl-globin and HLA-DQa DNA with allele-specific oligonucleotide probes. Nature (London) 324: 163-166. SAIKI, R. K., SCHARF, S., FALOONA, F., MULLIS, K. B., HORN, G. T., ERLICH, H. A., AND ARNHEIM, N. (1985). Enzymatic amplification of fl-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 230: 1350-1354. SMITH, M. (1986). Genetics of human alcohol and aldehyde dehydrogenases. Adu. Human Gene& 15: 249-290. SMITH, M., DUESTER, G., CARLOCK, L., AND WASMUTH, J. (1985). Assignment of alpha, beta and gamma ADH encoding genes to human chromosome 4q21-4q25 through the use of DNA probes. Cytogenet. Cell. Genet. 40: 743. SOUTHERN, E. (1975). Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. MoL Biol. 98: 503-517. VON BAHR-LINDSTR~M, H., HOL)G, J.-O., HBDEN, L.-O., KAISER, R., FLEETWOOD, L., LA-SON, K., LAKE, M., HOLMQUIST, B., HOLMGREN, A., HEMPEL, J., VALLEE, B. L., AND J~RNVALL, H. (1986). cDNA and protein structure for the 01 subunit of human liver alcohol dehydrogenase. Biochemistry 25: 2465-2470. YIN, S. J., BOSRON, W. F., LI, T. K., OHNISHI, K., OKUDA, K., ISHII, H., AND TSUCHIYA, M. (1984). Polymorphism of human liver alcohol dehydrogenase: Identification of ADH2 2-l and ADH2 2-2 phenotypes in the Japanese by isoelectric focusing. Biochem. Genet. 22: 169-180. ZHANG, K., BOSRON, W. F., AND EDENBERG, H. J. (1987). Structure of the Adh-1 gene and identification of a deletion in a long alternating purine-pyrimidine sequence in the first intron of strains expressing low alcohol dehydrogenase activity. Gene 57: 27-36.